Research Collection

Doctoral Thesis

Medium-chain-length poly(R-3-hydroxyalkanoates): frombiosynthesis towards medical applications

Author(s): Furrer, Patrick Christian

Publication Date: 2008

Permanent Link: https://doi.org/10.3929/ethz-a-005705895

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.

ETH Library

Diss. ETH No. 17654

Medium-chain-length poly([R]-3-hydroxyalkanoates):

from biosynthesis towards medical applications

A dissertation submitted to

ETH ZURICH

for the degree of

Doctor of Sciences

presented by

Patrick Christian Furrer

Dipl. Chem. University of Basel

born January 19, 1976

citizen of Attinghausen, UR

accepted on the recommendation of

Prof. Dr. S. Panke, examiner

Prof. Dr. T. Egli, co-examiner

Dr. M. Zinn, co-examiner

2008

2

For my parents

3

Table of contents

Summary 6

Zusammenfassung 8

Chapter 1: Poly([R]-3-hydroxyalkanoates): biosynthesis,

recovery, and their potential for biomedical appl ications 11

1.1 Introduction 12

1.2 Biosynthesis 13

1.3 Fermentative production 18

1.4 Material properties 20

1.5 Recovery methods 22

1.5.1 Chemical digestion of non PHA-biomass 23

1.5.2 Standard method: solvent extraction and

precipitation of PHA in non-solvents 26

1.5.3 Special approaches for the recovery of PHA 28

1.6 Purification of PHA 29

1.6.1 Sources and characterization of contaminations 30

1.6.2 Methods to purify extracted PHA 32

1.7 Potential applications of PHA in medicine and pharmacy 34

1.7.1 PHA as drug carrier 34

1.7.2 PHA as scaffold material in tissue engineering 35

1.8 Conclusions and outlook 37

4

Chapter 2: Quantitative analysis of bacterial mediu m-chain-length

poly([R]-3-hydroxyalkanoates) by gas chromatograp hy 39

2.1 Introduction 40

2.2 Materials and methods 42

2.3 Results and discussion 45

2.4 Conclusions 54

2.5 Acknowledgement 55

2.6 Supplementary information 56

Chapter 3: Biosynthesis of medium-chain-length

poly([R]-3-hydroxyalkanoate): endotoxin reduction

by phosphorus limitation 61

3.1 Introduction 62

3.2 Materials and methods 64

3.3 Results and discussion 67

3.4 Conclusions 77

3.5 Acknowledgements 78

Chapter 4: Efficient recovery of low endotoxin medium-chain-le ngth

poly([R]-3-hydroxyalkanoate) from bacterial bioma ss 79

4.1 Introduction 80

4.2 Materials and methods 82

4.3 Results and discussion 86

4.4 Conclusions 94

4.5 Acknowledgement 95

5

Chapter 5: Biocompatibility of untreated and chemically modifi ed

medium-chain-length poly(3-hydroxyalkanoates) 97

5.1 Introduction 98

5.2 Materials and methods 100

5.3 Results and discussion 104

5.3.1 Physical and chemical properties of PHAs 104

5.3.2 Cytotoxicity according to the eluate test 105

5.3.3 Cell culturing on PHA coatings 107

5.4 Conclusions 112

Chapter 6: General conclusions 115

References 119

Curriculum vitae 141

6

Summary

Poly(3-hydroxyalkanoate) (PHA) is a class of biodegradable polyesters that is

accumulated by a large number of bacteria as carbon storage compound under

nutrient limited growth conditions. PHA with side-chains of medium-chain-length

(mcl-PHA) represents a promising class of polymers because of its broad spectrum

of physical properties which are determined by the monomeric composition. Mcl-PHA

has a wide variety of potential applications in medicine and pharmacy due to its

versatility and biodegradability. However, a major breakthrough of mcl-PHA in the

biomedical field has not taken place yet. An important reason for this is that

biosynthetic products must fulfil strict requirements to be approved for medical

applications. In particular biosynthesis with Gram-negative bacteria is delicate,

because endotoxins from the cell membrane can contaminate the product.

In the first part of this doctoral thesis a method for the quantitative analysis of

mcl-PHA was developed, to optimise the fermentative production and the

downstream processing (DSP) of mcl-PHA (chapter 2). Standard methods for the

quantification of short-chain-length PHA proved to be inappropriate for the analysis of

mcl-PHA. It could be shown that the transesterification catalyzed by protic acids was

not quantitative for mcl-PHA under typical conditions. This was due to slow reaction

kinetics and the formation of side-products in case of mcl-PHA with functionalised

side-chains. An improved method using boron trifluoride as catalyst was developed

to quantitatively methanolyse different mcl-PHAs (recovery > 94%) and to analyse

the resulting methyl esters by gas chromatography.

The fermentative production and the DSP of mcl-PHA were investigated in the

second part of this doctoral thesis with the objective to generate mcl-PHA of high

purity and minimal endotoxic contamination. Phosphorus limited culturing conditions

were applied to Pseudomonas putida GPo1 in the chemostat to reduce the endotoxic

activity of lipopolysaccharides which act as endotoxins (chapter 3). Indeed, the

combination of phosphorus, carbon and nitrogen limitation drastically changed the

cell composition. The content of phospholipids and the endotoxic activity of dried

biomass were decreased by up to 90% under phosphorus limitation compared to

conditions with excess phosphorus. Hence, the endotoxic contamination of mcl-PHA

caused by the fermentation process could be drastically decreased.

7

To obtain mcl-PHA of high purity, the recovery from biomass was optimised

(chapter 4). Classical non-solvent precipitation was replaced by applying

temperature-controlled extraction and precipitation. N-hexane was found to be a

suitable solvent to recover poly(3-hydroxyoctanoate) and copolymers thereof with

this technique. A purity of >97% (w/w) and endotoxicities around 15 EU g-1 PHA were

obtained. Additional re-dissolution in 2-propanol combined with temperature-

controlled precipitation resulted in an enhanced purity of close to 100% (w/w) and a

minimal endotoxicity of 2 EU g-1 PHA.

In the third part of this doctoral thesis the biocompatibility of different mcl-PHAs

was studied with fibroblasts (chapter 5). Cytotoxicity experiments with aqueous

eluates of mcl-PHA as well as cell-culturing experiments on mcl-PHA coatings

showed that these polymers caused no toxic effect, when an effective purification

procedure, such as repeated dissolution and precipitation, was applied. Cell

adhesion was in fact limited on the non-polar surfaces of mcl-PHA, but chemical

derivatisation of the polymer surface clearly improved its hydrophilicity and cell

adhesion. However, the functionality of fibroblasts cultured on mcl-PHA surfaces

seemed to be limited as a reduced collagen production was measured.

The results achieved in this work show that an optimised DSP is of immense

importance to efficiently recover mcl-PHA of high purity which can be used for

biomedical applications. Furthermore, the activity of endotoxins can significantly be

reduced during biosynthesis of mcl-PHA which facilitates the DSP considerably. For

most biomedical applications chemical modification of hydrophobic mcl-PHA is

required to assure favourable interactions with cells.

8

Zusammenfassung

Poly(3-hydroxyalkanoat) (PHA) ist eine Klasse von biologisch abbaubaren

Polyestern, die von einer Vielzahl von Bakterien unter Nährstoff-limitierten

Wachstumsbedingungen als Kohlenstoffspeicher akkumuliert wird. PHA mit

Seitenketten von mittlerer Kettenlänge (mcl-PHA) repräsentiert eine vielver-

sprechende Klasse von Polymeren wegen ihres breiten Spektrums an physikalischen

Eigenschaften, welche durch die monomerische Zusammensetzung bestimmt

werden. Mcl-PHA besitzt eine grosse Vielfalt von möglichen Anwendungen im

medizinischen und pharmazeutischen Bereich aufgrund ihrer Vielseitigkeit und

biologischen Abbaubarkeit. Allerdings fehlt ein entscheidender Durchbruch von mcl-

PHA im biomedizinischen Bereich immer noch. Ein wichtiger Grund dafür ist, dass

biosynthetische Produkte strenge Anforderungen erfüllen müssen, um für

medizinische Anwendungen zugelassen zu werden. Speziell die Biosynthese mit

Gram-negativen Bakterien ist kritisch, weil Endotoxine der Zellmembrane das

Produkt verunreinigen können.

Im ersten Teil dieser Doktorarbeit wurde eine Methode zur quantitativen Analyse

von mcl-PHA entwickelt, um die fermentative Produktion und die Aufarbeitung von

mcl-PHA zu optimieren (Kapitel 2). Standardmethoden für die Quantifizierung von

PHAs mit kurzen Seitenketten erwiesen sich als ungeeignet für die Analyse von mcl-

PHAs. Es konnte gezeigt werden, dass die durch protische Säuren katalysierte

Umesterung nicht quantitativ ablief für mcl-PHAs unter typischen Bedingungen.

Grund dafür war die langsame Reaktionskinetik und die Bildung von Nebenprodukten

im Fall von mcl-PHAs mit funktionellen Seitenketten. Eine verbesserte Methode mit

Bortrifluorid als Katalysator wurde entwickelt um verschiedene mcl-PHAs quantitativ

zu methanolysieren (Wiederfindung >94%) und um die resultierenden Methylester

gaschromatografisch zu analysieren.

Die fermentative Produktion und die Aufarbeitung von mcl-PHA wurden im

zweiten Teil dieser Doktorarbeit untersucht mit dem Ziel mcl-PHA von hoher Reinheit

und minimaler endotoxischer Kontamination herzustellen. Phosphorlimitierte

Wachstumsbedingungen wurden angewendet für Pseudomonas putida Gpo1 im

Chemostat, um die endotoxische Aktivität der Lipopolysaccharide, welche als

Endotoxine wirken, herabzusetzen (Kapitel 3). Tatsächlich wurde die

Zellzusammensetzung durch kombinierte Phosphor-, Kohlenstoff- und Stickstoff-

9

limititation drastisch verändert. Der Phospholipidgehalt und die endotoxische Aktivität

der trockenen Biomasse wurden um bis zu 90% herabgesetzt unter

Phosphorlimitation verglichen mit Bedingungen mit Phosphor im Überschuss. Somit

konnte die Kontamination von mcl-PHA mit Endotoxinen, verursacht durch den

Fermentationsprozess, drastisch gesenkt werden.

Um mcl-PHA von hoher Reinheit zu erhalten, wurde der Aufreinigungsprozess

aus der Biomasse optimiert (Kapitel 4). Die klassische Fällung des Polymers mit Hilfe

eines Nicht-Lösungsmittels wurde ersetzt durch die Anwendung der temperatur-

kontrollierten Extraktion und Fällung. N-Hexan erwies sich als geeignetes Lösungs-

mittel, um Poly(3-Hydroxyoktanoat) und dessen Copolymere mit dieser Technik

aufzureinigen. Eine Reinheit von >97% (w/w) und eine endotoxische Aktivität von

ungefähr 15 EU g-1 PHA wurden damit erreicht. Zusätzliches Lösen in 2-Propanol

kombiniert mit temperaturkontrollierter Fällung resultierten in einer erhöhten Reinheit

von nahezu 100% (w/w) und einer minimalen endotoxischen Aktivität von 2 EU g-1

PHA.

Im dritten Teil dieser Doktorarbeit wurde die Biokompatibilität von verschiedenen

mcl-PHAs anhand von Fibroblasten untersucht (Kapitel 5). Zytotoxizitätstests mit

wässrigen Eluaten der mcl-PHAs und Zellkulturexperimente auf mcl-PHA Schichten

zeigten, dass diese Polymere keinen toxischen Effekt verursachten, wenn eine

effektive Aufreinigungsprozedur wie das wiederholte Lösen und Fällen angewendet

wurden. Die Zelladhäsion auf unpolaren mcl-PHA Oberflächen war limitiert, doch

durch chemische Derivatisierung der Polymeroberfläche konnte die Hydrophilie und

die Zelladhäsion deutlich verbessert werden. Allerdings schien die Funktionalität von

Fibroblasten, welche auf mcl-PHA Oberflächen kultiviert wurden, eingeschränkt zu

sein, da eine reduzierte Produktion von Kollagen gemessen wurde.

Die Resultate dieser Arbeit zeigen, dass ein optimiertes Aufreinigungsverfahren

von immenser Wichtigkeit ist, um mcl-PHA effizient und mit hoher Reinheit zu

erhalten, welches für biomedizinische Zwecke eingesetzt werden kann. Ausserdem

kann die Aktivität der Endotoxine während der Biosynthese von mcl-PHA signifikant

reduziert werden, was das Aufreinigungsverfahren wesentlich vereinfacht. Für die

meisten biomedizinischen Anwendungen wird eine chemische Modifikation des

hydrophoben mcl-PHAs benötigt, um günstige Interaktionen mit Zellen zu

gewährleisten.

10

11

Chapter 1

Poly([R]-3-hydroxyalkanoates):

Biosynthesis, recovery, and their potential

for biomedical applications

Patrick Furrer, Sven Panke and Manfred Zinn. 2007. Accepted for

publication in the Handbook of Natural-based Polymers for Biomedical

Applications.

Introduction PHA 12

1.1 Introduction

New biomaterials of the third generation are needed for particular medical

applications such as vascular implants, heart valves and cardiovascular fabrics

(Hench and Polak 2002; Hubbell 1995; Ueda and Tabata 2003). They are designed

to stimulate specific cellular responses at the molecular level and to combine the

concepts of bioactive and resorbable materials. They are supposed to help the body

heal itself by prompting cells to repair their own tissues. Today’s polymeric and

biodegradable systems used in medicine are mainly based on poly(lactic acid) (PLA),

on poly(glycolic acid) (PGA), and on their co-polymers (Gomes and Reis 2004).

Other biodegradable polymers have been proposed, but could not enter the market

yet, due to lacking FDA approval (Gomes and Reis 2004).

One of the candidates is poly([R]-hydroxyalkanoate) (PHA), a class of

biodegradable and biocompatible polyesters with many potential applications in the

medical field, such as heart valve scaffolds (Sodian et al. 2000a; Sodian et al. 2002),

pulmonary conduits (Stock et al. 2000), sutures, screws, bone plates, repair patches,

stents, bone marrow scaffolds, and many others over the last years as recently

reviewed by Chen and Wu (Chen and Wu 2005). PHA is composed of 3, 4, or rarely

5-hydroxy fatty acid monomers, which form linear polyesters. The general structure

of poly([R]-hydroxyalkanoate) is shown in Fig. 1.1. PHAs can be separated into 3

classes according to the size of comprising monomers: short-chain-length PHAs (scl-

PHA) with monomers of 3-5 carbon atoms, medium-chain-length PHAs (mcl-PHA)

with monomers of 6-14 carbon atoms and long-chain-length PHAs (lcl-PHAs) with

monomers of more than 14 carbon atoms. PHA is produced as reserve material by

many archae and eubacteria in aerobic and anaerobic habitats. PHA accumulation

occurs when microorganisms experience a metabolic stress, such as limitation by an

essential nutrient in excess of a suitable carbon source (Lageveen 1986, Lee 1996b,

Schlegel and Gottschalk 1962, Ward et al. 1977, Zinn and Hany 2005).

O

O

n

R

m

m=1-3

Fig. 1.1: General structure of poly([R]-hydroxyalkanoate).

13 Chapter 1

The purity of PHA is a crucial factor for sophisticated applications. It is

determined by the fermentation process and the downstream processing. The

production strain and the selective recovery procedure are essential factors for

obtaining PHA of high quality. Although there has been a continuous progress in

downstream processing of PHA for bulk applications (more than 50 patents have

been filed in the past 40 years) further improvements for high-quality PHAs and in

particular for mcl-PHAs are to be expected.

The restricted availability of PHAs has been limiting research significantly. Poly(3-

hydroxybutyrate) (PHB) and poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV)

are the only PHA-polymers that are currently commercially available (e.g. from

Metabolix and Biomer). At present, polymers based on poly(4-hydroxybutyrate)

(P4HB) are developed for medical use and one of them has been approved by FDA

recently (Rizk 2005). In 2008, Telles, a joint venture between Metabolix and Archer

Daniels Midland, plans to start the production of PHA under the name MirelTM at an

annual rate of 50 kt.

PHAs offer a wide range of physical properties due to their chemical diversity and

due to chemical modifications of functional groups following biosynthesis.

1.2 Biosynthesis

PHAs are a class of polyesters produced as reserve materials by many archae and

eubacteria (Gram-negative and Gram-positive, see Table 1.1) in aerobic and

anaerobic habitats. Up to date, more than 300 microorganisms are known to

synthesize PHA (Suriyamongkol et al. 2007). PHA accumulation is generally

triggered when the microorganisms experience a metabolic stress, such as limitation

by an essential nutrient (e.g., nitrogen, phosphorus or oxygen) and a concomitant

excess of a suitable carbon source (Anderson and Dawes 1990). PHA is stored

intracellularly in form of granules as a carbon and energy storage compound. The

PHA granules are surrounded by a phospholipid monolayer and several proteins like

the PHA polymerase and depolymerase and structural proteins called phasins are

part of the granule surface (see Fig. 1.2). Upon carbon starvation or a change of the

environmental pH (Ruth et al. 2007), intracellular PHA depolymerases, which are

attached to the granule, release 3-hydroxyalkanoic acids.

Introduction PHA 14

Table 1.1: Important PHA-producing genera, substrates and resulting PHAs.

Genus Gram stain

Substrates PHA Ref.

Alcaligenes - Fatty acids PHB Doi et al. 1992, Fukui and Doi 1998

Azospirillum - Fatty acids, saccharides

PHB Itzigsohn et al. 1995

Bacillus + Fatty acids (C2-C6), saccharides, lactones

PHB, PHBV, PHBHx, P4HB

Caballero et al. 1995, Chen et al. 1991, McCool et al. 1996, Tajima et al. 2003, Valappil et al. 2007

Beijerinckia - Glucose PHB Jendrossek et al. 2007

Brevundimonas - Saccharides PHB Silva et al. 2007

Burkholderia - Saccharides PHB, PH4PE Rodrigues et al. 2000

Caulobacter - Saccharides PHB Qi and Rehm 2001

Chromobacterium - 1,3-propandiol Scl- and mcl-PHA

Kimura et al. 2002

Clostridium + Acetate PHB Emeruwa and Hawirko 1973

Corynebacterium + Saccharides PHBV Haywood et al. 1991

Cupriavidus (former Wautersia)

- Saccharides, fatty acids

PHB, PHBV, P4HB

Kimura et al. 1999, Volova et al. 2006, Yan et al. 2003

Haloferax - Saccharides PHB, PHBV Koller et al. 2007, Lillo and Rodriguezvalera 1990

Methylobacterium - Alcohols, fatty acids

PHB, PHBV Yezza et al. 2006

Nocardia + Fatty acids, saccharides

PHB, PHBV Alvarez et al. 1997

Pseudomonas - Fatty acids, saccharides

PHHp - PHDd Kessler and Palleroni 2000

Rhizobium - Saccharides PHBV Lakshman et al. 2004

Rhodobacter - Fatty acids PHB, PHBV Kranz et al. 1997

Rhodococcus + Saccharides, fatty acids

PHBV, PHPi Füchtenbusch et al. 1998, Füchtenbusch and Steinbüchel 1999

Sphingopyxis - Saccharides PHBV Godoy et al. 2003

Staphylococcus + Saccharides PHB Szewczyk and Mikucki 1989

Streptomyces + Saccharides PHB Manna et al. 1999

PHB: Poly(3-hydroxybutyrate), PHBV: Poly(3-hydroxybutyrate-co-3-hydroxyvalerate), PHPi: Poly(3-hydroxypivalate), P4HB: Poly(4-hydroxybutyrate), PHBHx: Poly(3-hydroxybutyrate-co-3-hydroxyhexanoate), P3HB-4HB: Poly(3-hydroxybutyrate-co-4-hydroxybutyrate), PHHp: Poly(3-hydroxyheptenoate), PHDd: Poly(3-hydroxydodecanoate), PH4PE: Poly(3-hydroxypent-4-enoate)

15 Chapter 1

Fig. 1.2: Structure of PHA- granules accumulated by bacteria (Zinn et al. 2001).

PHAs have the general chemical structure depicted in Fig.1.1. The high

stereoselectivity of the producing enzyme machinery guarantees complete

stereospecificity (all chiral carbon atoms in the backbone are in the R configuration),

which is essential for the biodegradability and biocompatibility of PHA (Bachmann

and Seebach 1999; Hocking et al. 1996). The type of bacterium and the growth

conditions determine the chemical composition of PHAs and the molecular weight,

which typically ranges from 2x105 to 3x106 Da (Byrom 1987).

Research has focused on the substrate specificity of the PHA polymerase (Rehm

2003). Four major classes of PHA polymerases have been proposed with respect to

their primary structures, their substrate specificities and their subunit composition

(Rehm 2007) (see Table 1.2). It was found that the substrate specificity of the PHA

polymerase and the supply of cells with a particular substrate control the monomeric

composition of the resulting PHA. In general, different pathways for the biosynthesis

of PHA are possible (see Fig. 1.3). Acetyl-CoA activated precursors of PHA are

either synthesized in the course of anabolism through de novo fatty acid synthesis or

stem in the course of catabolism from the β-oxidation of fatty acids that are supplied

to cells.

Introduction PHA 16

Fig. 1.3: Metabolic routes of PHA biosynthesis. PhaA, β-ketothiolase; PhaB, NADPH-dependent acetoacetyl-CoA reductase; PhaC, PHA synthase; PhaG, 3-hydroxyacyl-ACP-CoA transferase; PhaJ, (R)-enoyl-CoA hydratase; FabD, malonyl-CoA-ACP transacylase; FabG, 3-ketoacyl-ACP reductase.

Table 1.2: The four classes of polyester synthases (Rehm 2003).

HA-CoA: coenzyme A thioester of [R]-hydroxy fatty acids of variable length

17 Chapter 1

To date, more than 150 different hydroxyalkanoic acids are known as PHA

constituents (Rehm 2007), but only few of the corresponding PHAs have been

produced in large quantities and well characterized (Kessler and Witholt 1998;

Witholt and Kessler 1999). As a consequence, little is known about the chemical and

mechanical properties of the larger number of these polymers. To date, PHA

monomers with straight, branched, saturated and unsaturated side chains have been

identified (see Fig. 1.4) (Steinbüchel and Valentin 1995). Of special interest are

functionalized groups in the side chain that allow further chemical modification, e.g.

halogens, carboxy, hydroxy, epoxy, phenoxy, cyanophenoxy, nitrophenoxy,

thiophenoxy, and methylester groups. The size of the monomer and its functional

group considerably influence properties like the bioplastic’s melting point, the glass

transition temperature, or the crystallinity, and therefore determine its final

application.

HO OH

n

HO

R

OH

O O

HO OH

O

HO OH

O

HO

X

OH

n O

HO OH

n O

m

HO OH

O

O

n

Halogen and cyano substituentsX=F, Br, Cl, CN

Epoxy substituent

Linear and branched alkyl substituentsR=alkyl substituent (C1-C12)

Terminal double bond

Cyclohexane substituent Aromatic substituent

Substituent with internal double bonds

OH

O

HOOH

OHO

4-hydroxy fatty acid 5-hydroxy fatty acid

R

R

Fig. 1.4: Monomers found in poly([R]-3,4 and 5-hydroxyalkanoates).

Introduction PHA 18

1.3 Fermentative production

Most PHA-producing bacteria start to accumulate PHA when their cell growth is

impaired by the limitation of an essential nutrient (e.g., N, P, Mg, K, O or S) in the

presence of excess carbon source. It is therefore important to use a suitable

fermentation strategy to enhance the production of PHA. The PHA content is thereby

the most important parameter for an efficient and economic downstream processing.

For fed-batch cultures a two step process is usually employed. In the first phase

the cells are grown to a desired cell concentration and in the second phase PHA

biosynthesis is triggered by a nutrient limitation. For continuous cultures, cell growth

and PHA accumulation are controlled by the ratio of the carbon source to the limiting

nutrient as well as by the dilution rate.

PHB has been efficiently produced with Alcaligenes latus in a two-step fed-batch

fermentation. Nitrogen limitation was chosen as the best strategy as it allowed the

greatest enhancement of PHB production. A cell concentration of 111.7 g L-1 and a

PHB content of 88 w% were reached resulting in the productivity of 4.94 g PHB L-1 h-1

(Wang and Lee 1997a). With Cupriavidus necator (former Wautersia eutropha and

Ralstonia eutropha) a PHB content of 76 w% was obtained at a cell concentration of

164 g L-1 resulting in a productivity of 2.42 g L-1 h-1 under nitrogen limitation (Kim et

al. 1994b).

PHBV has been successfully produced with C. necator. The two-step fed-batch

strategy was applied using nitrogen or phosphorus limitation in the second step. The

mole fraction of HV in the co-polymer could be controlled by changing the ratio of

glucose to propionic acid in the feed. PHBV contents of up to 75 w% and cell

concentrations of up to 158 g L-1 were obtained under nitrogen limitation (Kim et al.

1994a). However, with an increasing HV mole fraction, the PHA content decreased.

P3HB4HB could be produced in fed-batch cultures under nitrogen limitation. Cell

concentration of 34-49 g L-1 and PHA contents of 39-50 w% were reached with 4HB

mole fractions of 1.6-25.2 mol% (Kim et al. 2005).

For the efficient production of scl-PHAs, recombinant E. coli and C. necator have

been intensively investigated. In contrast to natural PHA-producing bacteria, PHA

accumulation by recombinant E. coli was not triggered by nutrient limitation. In a fed-

batch fermentation PHB was produced with recombinant E. coli reaching a cell

19 Chapter 1

concentration of 204 g L-1 with a PHB content of 77 w% and a productivity of

3.2 g L-1 h-1 (Wang and Lee 1997b). A higher PHB productivity of 4.63 g L-1 h-1 was

obtained with recombinant E. coli harboring an optimally designed plasmid containing

the PHA biosynthesis genes from A. latus (Choi et al. 1998). PHBV with 11 mol% HV

could be produced with a concentration of 159 g L-1 and a productivity of

2.88 g L-1 h-1 with recombinant E. coli (Choi and Lee 1999b). Recombinant C. necator

was investigated for the production of PHB and PHBV but the final cell

concentrations and PHA contents were only slightly increased compared to the

parent strain (Lee et al. 1996; Park et al. 1995; Valentin and Steinbuchel 1995).

Although efforts have been made to use recombinant E. coli and recombinant

Pseudomonas for the production of mcl-PHA, few fermentation processes were

based on recombinant strains (Prieto et al. 1999).The production of mcl-PHA was

extensively investigated with Pseudomonas species (Diniz et al. 2004; Eggink et al.

1992; Hartmann et al. 2006; Hoffmann and Rehm 2004; Hori et al. 1994; Huijberts

and Eggink 1996; Kim et al. 2000). Various alkanes, alkanoic acids as well as

glucose, fructose and glycerol were used as substrates for the production of mcl-PHA

(Huijberts et al. 1992; Timm and Steinbüchel 1990). Especially with structurally

related carbon sources, such as alkanes and aliphatic acids efficient mcl-PHA

synthesis occurs (Lageveen et al. 1988). But these carbon sources are poorly water

miscible or/and toxic to bacteria at rather low concentration (Sun et al. 2007). Hence

the concentration of these carbon sources has to be well controlled.

For fed-batch cultures the two-step fermentation strategy has often been applied.

Using octanoic acid as substrate, PHA contents up to 75 w% at a cell concentration

of 55 g L-1 and a productivity of 0.63 g L-1 h-1 were obtained with Pseudomonas

putida GPo1 under nitrogen limitation (Kim 2002). With Pseudomonas IPT 046 cell

concentrations of up to 50 g L-1 with a PHA content of 63 w% and a productivity of

0.8 g L-1 h-1 were reached under phosphate limitation using a mixture of glucose and

fructose as carbon source (Diniz et al. 2004). When oleic acid was used as substrate

for the cultivation of Pseudomonas putida KT2442, a cell concentration of 141 g L-1

with a PHA content of 51 w% and a productivity of 1.91 g L-1 h-1 were obtained under

phosphate limitation (Lee et al. 2000b).

Continuous fermentation has been optimized for the efficient production of mcl-

PHA (Hartmann et al. 2006; Huijberts and Eggink 1996; Prieto et al. 1999). It has

Introduction PHA 20

been shown that the PHA content decreased with increased specific growth rate.

Thus, a compromise between PHA content and productivity is required in a single-

stage continuous process (Sun et al. 2007). A two-stage continuous process was

developed to overcome this limitation (Jung et al. 2001). Cell densities of 18 g L-1

with a PHA content of 63 w% and an overall volumetric productivity of 1.06 g L-1 h-1

were obtained.

1.4 Material properties

Interestingly, the material properties of PHA are similar within one class as shown in

Table 1.3. They are strongly dependent on the monomeric composition, in particular

on the monomers’ side-chain and on the distance between the ester linkages in the

backbone. In general, PHA is water insoluble, biodegradable, and biocompatible. It

can be degraded at a moderate rate (3–9 months) by many microorganisms into

carbon dioxide and water using their own secreted PHA depolymerases (Jendrossek

2001). Its primary breakdown products, 3-hydroxyacids, are naturally found in

animals and humans.

Table 1.3: Physical properties of short-chain-length (scl) and medium-chain-length (mcl) PHAs.

Scl-PHA Mcl-PHA

Crystallinity [%] 40-80 20-40

Glass transition temperature [°C] 2 -36

Melting point [°C] 80-180 30-80

Elongation to break [%] 6-10 300-450

Density [g cm-3] 1.25 1.05

Scl-PHAs are typically crystalline thermoplasts. The homopolymer PHB is a

relatively stiff and brittle bioplastic, which can be processed by melt extrusion. PHA

co-polymers composed of primarily HB with a fraction of longer chain monomers,

such as HV, HHx or HO, are more flexible and less brittle thermoplasts. Their

structure is shown in Fig. 1.5. They can be used in melt-extrusion processes to form

a wide variety of products including containers, bottles, razors, and materials for food

21 Chapter 1

packaging. PHB and PHBV was used as water-proof film on the back of diaper

sheets (Martini et al. 1989). PHBV is more flexible, more impact resistant and was

marketed under the trade name Biopol™ by ICI/Zeneca and later on by Monsanto till

1995. Co-polymers consisting of 3-hydroxybutyrate and a 3-hydroxyalkanoic acid

with at least 6 carbon atoms, for instance PHBHx was developed by Procter and

Gamble and commercialized under the name Nodax. In analogy to PHBV, PHBHx is

less crystalline than PHB and more flexible.

O

O

O

n

O

O

n

poly(3-hydroxybutyrate-co-3-hydroxyvalerate)

poly(4-hydroxybutyrate)

O

m

O

O

O

npoly(3-hydroxybutyrate-co-3-hydroxyhexanoate)

O

m

Fig. 1.5: Chemical structures of commercially important scl-PHAs: Poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV), poly(3-hydroxybutyrate-co-3-hydroxyhexanoate) (PHBHHx) and poly(4-hydroxybutyrate) (P4HB).

Common mcl-PHAs are less crystalline and more elastomeric. Their properties

largely depend on their side-chains. For instance, the glass transition temperature of

a mcl-PHA containing linear chains up to seven carbon atoms and aromatic groups

varies between -39 °C and -6 °C with an increasing content of aromatic side-chains

(Hartmann et al. 2004). The physical properties of mcl-PHAs containing functional

groups in the side-chains can be adapted through chemical modification. Thereby,

the field of possible applications is considerably extended. Through the insertion of

hydroxy groups in the side-chains of mcl-PHA, the polymer’s polarity and solubility in

polar solvents could be drastically changed (Eroglu et al. 2005; Lee et al. 2000a;

Renard et al. 2005). Epoxidation and crosslinking of mcl-PHAs with unsaturated side-

chains increased the polymer’s tensile strength and Young’s modulus by a factor of 4

and 39, respectively (Ashby et al. 2000). Polyhedral oligomeric silsesquioxane

(POSS) were linked to the side-chains of unsaturated mcl-PHAs thereby increasing

the crystallinity and rising the melting point up to 120 °C (Hany et al. 2005).

Introduction PHA 22

Very little is known about lcl-PHA. It has been shown that the ability to polymerize

long-chain-length PHAs is limited to Pseudomonas putida KTOY06 (Liu and Chen

2007) and probably this is valid for most PHA accumulating bacteria.

1.5 Recovery methods

Recovery processes can be divided into two categories: solvent extraction and

chemical digestion. In the former one PHA is extracted from biomass by using an

organic solvent like methylene chloride. In the latter case the rest biomass is

digested by applying chemicals like sodium hypochlorite or hydrolytic enzymes.

Through cross-flow filtration or centrifugation PHA is separated then from cell debris.

The main problem of chemical digestion is severe degradation of PHA, resulting in a

reduction of the molecular weight.

To certify polymers for medical applications, very demanding specifications have

to be fulfilled. Considerable amounts of impurities like proteins, surfactants and

endotoxins are not tolerated by regulatory agencies. The presence of bacterial

endotoxins in medical polymers is one of the biggest concerns of suppliers (Williams

et al. 1999). Up to now only few methods for depyrogenation of PHA were described

in literature, although this aspect is very important for medical applications. More

details are given in section 1.6.

Fig. 1.6: PHA recovery through chemical digestion (A) and solvent extraction (B).

23 Chapter 1

1.5.1 Chemical digestion of non PHA-biomass

Digestion of non-PHA biomass is performed in aqueous environment without or with

only small amounts of organic solvents (Fig. 1.6A). The suspended cells are usually

lysed to release the PHA granules, which form an aqueous suspension. For cell lysis

various techniques used in biotechnology can be applied as shown in Table 1.4. The

granule envelope protects the PHA to a certain degree from chemical degradation.

All the residual biomass constituents are rendered water soluble by chemical

modification. This can be done by using chemical agents like peroxides,

hypochlorites and ozone or by using hydrolytic enzymes. After cell disruption and

digestion of residual biomass, PHA granules can be separated from cell debris by

conventional methods like centrifugation or cross-flow filtration. While PHB granules

have a density of about 1.2 g cm-1, mcl-PHA granules possess a density close to that

of water (Preusting et al. 1993) and therefore do not settle in aqueous suspensions

(Marchessault et al. 1995). Hence, ultracentrifugation or cross-flow filtration has to be

applied (deKoning et al. 1997; Yasotha et al. 2006). Although the PHA granules are

not physically dissolved in H2O, they form a stable suspension (latex). Once the PHA

granules are isolated further purification steps usually follow. In the following we will

focus the discussion to the first step, the degradation of residual biomass, because

this is crucial for the further processing.

Introduction PHA 24

Table 1.4: Methods to lyse bacterial cells for the recovery of PHA.

Physical

methods

Ref. Chemical

methods

Ref.

Bead milling

Tamer et al. 1998a,

Tamer et al. 1998b

Reducing or oxidizing

agents

Ramsay et al. 1992,

Tamer et al 1998a

Homogenization Ling et al 1997b,

Tamer et al. 1998a

Detergents

de Koning et al. 1997

Ultrasonication

Hwang et al. 2006 Lytic enzymes Kim et al. 1996

Thermal shock

Cumming et al.

1994

Solvents (osmotic

shock)

de Koning and Witholt

1997, Escalona et al.

1996

Supercritical fluid

treatment

Castor and Hong

2003, Hejazi et al.

2003

pH-shock Tamer et al. 1998b

Cell debris digestion with hypochlorite (Berger et al. 1989; Hahn et al. 1994;

Hahn et al. 1995; Lee and Choi 1998; Ling et al. 1997b; Middelberg et al. 1995;

Tamer et al. 1998a; Taniguchi et al. 2003) was one of the earliest methods for

degrading residual biomass. Most cell constituents are oxidized by hypochlorite and

thus become water soluble. However, PHA was shown to become severely degraded

with a decrease in molecular weight Mw from e.g. 1’200 kDa to 600 kDa under

optimized conditions (Berger et al. 1989; Hahn et al. 1994; Hahn et al. 1995; Ramsay

1990). In addition, it is difficult to eliminate residues of sodium hypochlorite

completely in the subsequent steps. Hypochlorite digestion could be improved by

combining it with surfactants or organic solvents and resulted in PHA of higher purity

with less polymer degradation (Bordoloi et al. 2003; Hahn et al. 1994; Ramsay 1990).

Alternative oxidizing agents like hydrogen peroxide or ozone were investigated to

oxidize residual biomass. Peroxides were especially useful to degrade nucleic acids

and to decrease the viscosity of the lysate (Greer 1994; Greer 1997; Liddell and

Locke 1997). In combination with enzymes and surfactants, the efficiency of rest

25 Chapter 1

biomass degradation with peroxides could be increased (George and Liddell 1997).

Treatment of PHA containing biomass or partially purified PHA with ozone yielded an

enhanced level of purity in combination with other purification steps. The resulting

polymer was basically odor-free (Horowitz and Brennan 1999). Furthermore,

coloration of PHA was drastically reduced with ozone or peroxide. It has not clearly

been described in literature if PHA was degraded by hydrogen peroxide or ozone.

But a certain degradation has to be expected due to the strong reactivity of these

agents.

Simple alkaline digestion of suspended biomass provided PHB with a purity of

85 % or higher, but a minor degradation of PHA has also been observed (Choi and

Lee 1999a). Combination with chelates and surfactants improved cell disruption and

solubilisation of the rest biomass, and therefore PHA of higher purity was obtained.

The use of enzymes is an alternative to harsh chemical agents. Cell fragments

can be degraded more selectively and the degradation of PHA is negligible. Enzyme

cocktails consisting of various hydrolytic enzymes like proteinases, nucleases,

phospholipases, lysozymes and others in combination with surfactants and chelate

formers have been applied (Byrom 1987; George and Liddell 1997; Holmes and Lim

1990; Kim et al. 1996; Ramsay et al. 1992; Yamamoto et al. 1995). Combined with

heat treatment, PHA purities of up to 95% could be achieved (deKoning and Witholt

1997). Heat accelerated the degradation of residual biomass, while surfactants and

chelates improved the solubility of all components. Mcl-PHA with a purity of 93 %

was obtained by enzymatic digestion combined with ultrafiltration (Yasotha et al.

2006).

Generally the isolation of the PHA-granules is much easier with bacteria that

have a high PHA content (>60 % in dry mass). Only one or two steps are necessary

to achieve a purity of around 90 %. If the PHA content is below 60%, the separation

process is more complicated. For this case a combined method with enzymes and

reducing agents like sodium dithionite was effective (Schumann and Müller 2003).

Cell components like proteins, nucleic acids and polysaccharides were decomposed

without drastically changing PHA properties.

To avoid the problem of a dramatic increase in viscosity caused by the liberation

of DNA, a nuclease-encoding gene was integrated into the genome of PHA

Introduction PHA 26

producing bacteria (Boynton et al. 1999). Thereby the amount of enzymes or

chemicals necessary for the digestion of the rest biomass was reduced. The lysate

viscosity was significantly reduced without affecting the PHA production or the strain

stability.

Major disadvantages of the enzymatic digestion are the high costs of hydrolytic

enzymes and the additional purification steps necessary to reach a high degree of

purity. Moreover surfactants can hardly be separated from PHA.

The purification of PHA granules would be much more efficient if they were

secreted into the medium. Separation from the cells could be done by centrifugation

without digesting the residual biomass. The production of extracellular PHA has been

proposed (Sabirova et al. 2006). However, this seems to be irrational as excreted

PHA granules would hardly be accessible to the bacteria for a later use of PHA.

1.5.2 Standard method: solvent extraction and prec ipitation of PHA in non-

solvents

As discussed before, the alternative method to chemical or enzymatic digestion of

the residual biomass is the selective extraction of PHA from biomass with an organic

solvent. Most recovery procedures based on the extraction with organic solvents

follow the steps shown in Fig.1.6B. Bacterial cells are collected by centrifugation or

filtration and dried to remove water that inhibits an effective extraction. Pre-treating

dry biomass with methanol can be effective to remove some of the lipids and coloring

impurities (Gorenflo et al. 2001; Jiang et al. 2006). PHA is extracted from the dried

biomass with an organic solvent under stirring and in some cases heating. The

resulting suspension is filtered or centrifuged to remove particulate cell debris and

subsequently PHA is either precipitated with a non-solvent, or obtained by

evaporation of the solvent. In some cases the crude PHA thus obtained is washed

with a non-solvent. Repeated extraction, precipitation and washing steps result in a

higher purity.

Solvent extractions normally use large amounts of solvents, typically 5-20 times

of the dry weight of biomass and similar amounts of non-solvents for precipitation.

Solvent recycling is energy consuming, especially for solvents with high boiling

points. The high viscosity of even diluted PHA solutions (e.g. 5% w/v) impedes

27 Chapter 1

extensive solvent savings and limits economical optimization. Soxhlet extraction was

used to work with reduced volumes of solvent (Valappil et al. 2007).

The solubility of PHA is strongly dependent on the polymer composition, the

molecular weight and on the temperature and pressure (Terada and Marchessault

1999). The extraction of scl-PHA was usually performed with chlorinated solvents like

methylene chloride, chloroform, 1,2-dichloroethane, 1,1,2-trichloroethane and

1,1,2,2-tetrachloroethane (Baptist 1962b; Barham 1982; Holmes et al. 1980;

Numazawa et al. 1987; Ramsay et al. 1994; Schmidt et al. 1986; Stageman 1985;

Vanlautem and Gilain 1982; Vanlautem and Gilain 1987) because most chlorinated

solvents are capable of dissolving PHB at a relatively high concentration, but only

little of the rest biomass is dissolved as well. Preferentially the extraction was carried

out at elevated temperatures to increase the solubility and to reduce the viscosity of

the polymer solution and the extraction time. To reach high temperatures without

evaporation of the solvent, pressurization has been applied (Kurdikar et al. 1998).

PHA degradation at high extraction temperatures has been observed, especially

when water was present in solution. At temperatures above 200 °C degradation

occurs also in the absence of water by a non-radical random chain scission reaction

(cis-elimination) (Lee et al. 2001). Recently, it has been shown that even at moderate

temperatures thermal degradation occurs via E1cB mechanism if carboxylate groups

are present (Kawalec et al. 2007). To precipitate the dissolved polymer, different non-

solvents like methanol, ethanol, water, or ether have been used (Baptist 1962a;

Holmes 1980; Kessler and Witholt 1998; Numazawa et al. 1987; Stageman 1985).

Thereby, ether could only be applied for scl-PHA as it dissolves mcl-PHAs. PHB with

a purity of 99% could be obtained with precipitation in water (Hrabak 1992). Mixtures

of chlorinated solvents with a non-solvent were applied to extract PHB from biomass

at high temperatures and to precipitate the polymer by cooling to room temperature

(Schmidt et al. 1986). Azeotrope-building chlorinated solvents (e.g., 1,1,2-

trichloroethane) were used for simultaneous azeotropic distillation and PHA

extraction from an aqueous suspension of microorganisms (Vanlautem and Gilain

1987). Thereby water was removed from the cell suspension as a minimum boiling

azeotrope. The remaining chlorinated solvent was used to extract PHA. Hence a

preceding drying of the cell suspension could be avoided.

Introduction PHA 28

In an effort to prevent the disadvantages of halogenated solvents, which are

known to pose health risks and environmental problems, alternative solvents like

cyclic carbonic acid esters (Lafferty 1977), methyl lactate (Metzner et al. 1997), ethyl

lactate (Metzner et al. 1997; Sela and Metzner 1996), acetic acid (Rapthel et al.

1993a; Rapthel et al. 1993b; Runkel et al. 1993a), acetic acid anhydride (Runkel et

al. 1993b; Schmidt et al. 1985), n-methylpyrrolidone (Schumann and Mueller 2001),

tetrahydrofuran (Matsushita et al. 1995) and mixtures of non-halogenated solvents

(Noda and Schechtman 1997) were investigated for the extraction of PHB. Co-

polymers with PHB were also extracted by non-chlorinated solvents like acetone

(Gorenflo et al. 2001), ethyl acetate (Chen et al. 2001), butyl acetate (Walsem et al.

2004), methyl isobutylketone (Walsem et al. 2004) and cyclo-hexanone (Walsem et

al. 2004).

Mcl-PHA has the advantage to be soluble in a broader spectrum of solvents than

PHB. Even at room temperature typical mcl-PHAs are soluble in acetone, THF or

diethyl ether. A patent from Firmenich S.A. describes an extraction method with non-

halogenated solvents at room temperature tailored for polyhydroxyoctanoate

(Ohleyer 1993). More recently solvents like acetone were used for extracting mcl-

PHA from biomass (Jiang et al. 2006).

A combined extraction-filtration method has been described using a circular

filtration system (Horowitz 2000), where the aqueous slurry was diafiltrated and an

organic solvent like acetone was continuously added. In the beginning the bacterial

cells were rejected, but at a certain acetone to water ratio of about 9:1 (w/w) the cells

were lysed, mcl-PHA was dissolved and passed the filtration membrane.

The recycling of solvent and non-solvent mixtures is time and energy

consuming. The utilisation of non-solvents can be avoided by temperature induced

extraction and precipitation, described in chapter 4.

1.5.3 Special approaches for the recovery of PHA

Supercritical fluid extraction (SFE) with carbon dioxide has been proposed for the

recovery of PHA (Williams et al. 1999). After extraction, carbon dioxide evaporates

immediately and a drying step is not required. However, published data are

conflicting. Pure PHB is soluble up to 8.01 g L-1 in supercritical CO2 at a temperature

29 Chapter 1

of 348 K and a pressure of 355 bar (Khosravi-Darani et al. 2003). In contrast to these

results, Seidel and Hampson showed that lipids, pigments and ubiquinones were

extracted from biomass containing PHB or PHO without dissolving the polymer under

similar conditions (Hampson and Ashby 1999; Seidel et al. 1991). In these reports

PHA was extracted from the purified biomass with an organic solvent, e.g. chloroform

(Hampson and Ashby 1999; Seidel et al. 1991). Consequently, it remains uncertain

whether SFE can be efficiently used to extract PHA from biomass.

A process for recovering PHA from biological material by comminution and air

classification has been patented (Noda 1995; Noda 1998). In principle no chemicals

or solvent is needed for this process. The dried biomass was defatted and grinded so

that the diameter of most particles was below 100 µm. An air stream was used to

suspend the particles and classify them according to their weight or size. At

appropriate stream velocities the particles were separated into a coarse and a fine

fraction. The fine fraction was then subjected to further purification steps to reach a

PHA purity of 80% or higher. So far, it has not been described whether this process

is successfully applied for the recovery of PHA at a larger scale.

Furthermore dissolved-air flotation was recently used to separate mcl-PHA

granules from the fermentation broth (van Hee et al. 2006). Flotation is potentially

cheap and is common in wastewater treatment. Flotation separates particles

according to their affinity to the air/liquid interface which is generated by bubbling air

(or another gas) through a liquid phase (e.g. water). Hydrophobic particles are

transported with the air-bubbles to the surface, whereas more hydrophilic ones

remain in the aqueous phase. The cells were pre-treated with enzymes to release the

PHA granules. Selective aggregation and flotation of mcl-PHA granules could be

triggered by adjusting the pH at around 3.5. A purity of 86% was obtained after three

consecutive batch flotation steps.

1.6 Purification of PHA

For applications in the medical field, PHA of high purity is needed. In particular

biologically active contaminants like proteins and lipopolysaccharides (LPS) have to

be reduced to a very low level as they could induce immunoreactions. The U.S. Food

Introduction PHA 30

and Drug Administration (FDA) requires the endotoxin content of medical devices not

to exceed 20 U.S. Pharmacopeia (USP) endotoxin units (EU) per device, except for

those devices that are in contact with the cerebrospinal fluid. In this case the content

must not exceed 2.15 USP EU per device. LPS act as endotoxins whereby minute

quantities can have severe effects in contact with blood and trigger immunoreactions.

Particularly for PHA derived from fermentation of Gram-negative bacteria,

contamination with endotoxins is a serious problem, as LPS are part of the outer

membrane. During cell lysis and product recovery, LPS are liberated from the outer

membrane and contaminate PHA. Therefore, PHA for use in medical devices has to

be carefully purified from such endotoxins.

Standard techniques used for the purification of PHA are re-dissolution and

precipitation, washing with a non-solvent, purification by chromatography, treatments

with chemical agents and filtration. The purity of PHA can also be increased by

washing the biomass before extracting PHA or by aqueous digestion to remove

major impurities as described before.

1.6.1 Sources and characterization of contamination s

The spectrum of possible PHA-contaminants from residual biomass is broad. But

basically, the extraction solvent determines which impurities will be carried over. For

instance, proteins and DNA have been frequently detected when PHA was recovered

by aqueous chemical digestion. Solvent extraction with non-polar solvents or with

solvents of average polarity is more susceptible to co-extract lipids and coloring

substances. Notably polar organic solvents like acetone and 2-propanol co-extract

chromophores and give the polymer a yellow to brownish color (see Fig. 1.7).

Lipopolysaccharides have been detected with aqueous digestion as well as with

solvent extraction. They are soluble in water, but their amphiphilic character renders

them to some extent soluble in non-polar solvents.

Besides lipids, proteins and lipopolysaccharides also antifoam agents from

fermentations, surfactants and hydrolytic enzymes from the purification procedure

are contaminants of the polymer. The most common impurities found in PHA are

summarized in Table 1.5. In summary the nature of contaminants is determined by

the biosynthesis as well as by the downstream processing.

31 Chapter 1

Table 1.5: Common contaminants found in PHA.

Contaminant Amount in

PHA

Ref. Recovery

method

Microorganism PHA

Lipids 0.1-3 mol%a Sevastianov et al.

2003

Ethanol-KOH

wash +

chloroform-

ethanol

extraction

R. eutropha

B5786

PHB,

PHBV

Proteins 0-9 w% Ling et al. 1997a,

Ling et al. 1997b

Homogenization

+ centrifugation

E. coli

PHB

n.d Seidel et al. 1991 Extraction M. rhodasianum PHB

n.d Williams et al.

1999

Aq. digestion n.d. PHO

DNA 0.03-2.2 w%

Ling et al. 1997b Homogenization

+ centrifugation

E. coli PHB

0.6-2.2 w% Ling et al. 1997a Homogenization

+ centrifugation

E. coli PHB

LPS >120 EU g-1 Williams et al.

1999

N.d. N.d. PHB

1-104 EU g-1 Lee et al. 1999 NaOH-digestion E. coli PHB

Antifoam

agents

< 1 w% Unpubl. results Chloroform-

ethanol

extraction

P. putida GPo1 PHO

SDS N.d. de Koning and

Witholt 1997

SDS

solubilisation +

enzym.

treatment

P. putida GPo1 PHO

UV absorbers N.d. Jiang et al. 2006 Acetone

extraction

P. putida

KT2440

mcl-

PHA

N.d. not defined, a) relative to the PHA monomers

Introduction PHA 32

Fig. 1.7: Coloring impurities co-extracted from biomass. PHO extracted with A) hexane, B) methylene chloride and C) acetone. The solvent was evaporated after filtration.

1.6.2 Methods to purify extracted PHA

Repeated dissolution and precipitation was commonly applied to reach a purity of

close to 100%. The efficiency of this method is dependent on several parameters like

the concentration of the polymer solution and the temperature but usually large

amount of non-solvents were used. Jiang et al. (Jiang et al. 2006) showed that the

concentration of UV absorbing molecules is considerably decreased by dissolution of

mcl-PHA in acetone and precipitation in cold methanol as shown in Fig. 1.8. Scl-PHA

was repeatedly dissolved in chloroform and precipitated in ethanol to remove

biologically active substances. Traces of fatty acids from C6 to C18 were eliminated

and the hemocompatibility increased (Sevastianov et al. 2003). The authors propose

that these fatty acids originated from lipopolysaccharides.

Fig. 1.8: UV spectra of crude extracted and purified mcl-PHA in chloroform. 1: acetone extract precipitated in methanol, 2-4: additional cycles of dissolution and precipitation (1-3 times). The absorption bands at 275 and 241 nm were attributed to aromatic amino acids and nucleic acids (Jiang et al. 2006).

33 Chapter 1

Temperature controlled dissolution and precipitation in 2-propanol was used to

purify certain mcl-PHA like PHO. A purity of nearly 100 % and an endotoxicity of

below 10 EU g-1 polymer was obtained with this method as shown in chapter 4.

To reduce the endotoxin content further to acceptable values, e.g. < 10 EU g-1 of

PHA, a treatment with an oxidizing agent such as hydrogen peroxide or benzoyl

peroxide was used successfully (Williams et al. 2001). Destruction of LPS by

applying basic conditions was also successful (Lee et al. 1999; Sevastianov et al.

2003). The concentration of base and the treatment time (see Fig. 1.9) are crucial for

an ideal detoxification. Otherwise, the destruction of LPS is incomplete or PHA is

depolymerised and its molecular weight reduced as previously mentioned.

Fig. 1.9: Endotoxic activity of PHB recovered by 1.2N NaOH digestion at 30 °C for various durations (Lee et al. 1999).

Further methods can be used for purification such as washing of PHA with a non-

solvent, purification by chromatography, filtration and treatment with endotoxin

removing agents, however, accurate data about their efficacy are not available. For

the protein purification in aqueous solutions, e.g., cationic endotoxin removal agents

have been shown to be very successful (Zhang et al. 2005).

Introduction PHA 34

1.7 Potential applications of PHA in medicine and p harmacy

PHA has the potential to become an important compound for medical applications

(Williams et al. 1999; Zinn et al. 2001). Biocompatibility and slow biodegradability are

thereby essential properties. The changing PHA composition also allows favorable

mechanical properties as shown in Table 1.3. In vitro cell experiments and in vivo

studies have focused on PHB, PHBV, P4HB, PHBHx and PHO. An overview of the

potential applications of PHA is given in Table 1.6. In the following we discuss the

applications in the field of drug delivery and tissue engineering. The discussion is

restricted to these fields, because the most promising research was done in these

areas.

Table 1.6: Potential applications of PHA in medicine and pharmacy.

Type of application Products Type of PHA

Wound management Sutures, skin substitutes, nerve cuffs, surgical meshed, staples, swabs

Scl, mcl

Vascular system Heart valves, cardiovascular fabrics, pericardial patches, vascular grafts

Mcl

Urology Urological stents Scl, mcl

Orthopaedy Scaffolds for cartilage engineering, spinal cages, bone graft substitutes, meniscus regeneration, internal fixation devices (e.g. screws)

Scl, (mcl)

Dental Barrier material for guided tissue regeneration in periodontosis

Scl, mcl

Computer assisted tomo-graphy and ultrasound imaging

Micro- and nanospheres for anticancer therapy

Scl, mcl

Drug delivery Chemoembolizing agents, micro- and nanospheres for anticancer therapy

Scl, mcl

1.7.1 PHA as drug carrier

PHAs became candidate material as drug carriers in the early 1990s due to their

inherent biocompatibility (Pouton and Akhtar 1996). Microspheres of PHB loaded

with rifampicin were investigated for their use as chemoembolizing agent (agent for

the selective occlusion of blood vessels) (Kassab et al. 1999; Kassab et al. 1997).

35 Chapter 1

The drug release of all microspheres was very rapid with almost 90% of the drug

released within 24h. The drug release rate could be controlled by the drug loading

and the particle size. A similar behavior was described by Sendil and coworkers for

PHBV supplemented with tetracycline (Sendil et al. 1999). PHBV was further

investigated as an antibiotic-loaded carrier to treat implant-related and chronic

osteomyelitis (Yagmurlu et al. 1999). The antibiotic sulbactamcefoperazone was

integrated into PHBV rods and implanted into a rabbit tibia that was artificially

infected by S. aureus. The infection subsided after 15 days and was nearly

completely healed after 30 days.

In search of an efficient transdermal drug delivery system a PHO-based system

with a polyamidoamine dendrimer was examined. Tamsulosin was used as the

model drug. The dendrimer was found to act as weak permeability enhancer. By

adding the dendrimer, the dendrimer-containing PHA matrix achieved the clinically

required amount of tamsulosin permeating through the skin model (Wang et al.

2003).

1.7.2 PHA as scaffold material in tissue engineerin g

PHBV was chosen as a temporary substrate for growing retinal pigment epithelium

cells as an organized monolayer before their subretinal transplantation. The surface

of the PHBV film was hydrophilized by oxygen plasma treatment to increase the

attachment of D407 cells to the polymer surface. The cells grew to confluency as an

organized monolayer. Hence, PHBV films can be used as temporary substrates for

subretinal transplantation to replace diseased or damaged retinal pigment epithelium

(Tezcaner et al. 2003).

An interesting approach is the implantation of biodegradable supporting scaffolds

that are seeded with tissue-engineered cells. This approach was exemplified by

Sodian and co-workers (Sodian et al. 2000c) who used PHO and P4HB for the

fabrication of a tri-leaflet heart valve scaffold. A porous surface was achieved with the

salt leaching technique resulting in pore size between 80 and 200 µm. The scaffold

was seeded with vascular cells from ovine carotid artery and subsequently tested in a

pulsatile flow bioreactor. The cells formed a confluent layer on the leaflets.

Introduction PHA 36

In another study, the native pulmonary leaflets were resected with the use of

cardiopulmonary bypass, and segments of pulmonary artery were replaced by

autologous cell-seeded heart valve constructs. All animals survived the procedure

without receiving any anticoagulation therapy. The tissue engineered constructs were

covered with tissue and no thrombus formation was observed. It was concluded that

tissue engineered heart valve scaffolds fabricated from PHO can be used for

implantation in the pulmonary position with an appropriate function for 120 days in

lambs (Sodian et al. 2000a).

The same group demonstrated that PHO and P4HB have thermoprocessible

advantages over PGA who has better property for ovine vascular cell growth (Sodian

et al. 2000b; Sodian et al. 2000c).

Vascular smooth muscle cells and endothelial cells from ovine carotid arteries

were seeded on P4HB scaffolds to study autologous tissue engineered blood vessels

in the descending aorta of juvenile sheep. Up to 3 months after implantation, grafts

were fully patent, without any signs of dilatation, occlusion or intimal thickening. A

confluent luminal endothelial cell layer was observed. In contrast, after 6 months graft

displayed significant dilatation and partial thrombus formation, most likely caused by

an insufficient elastic fiber synthesis (Opitz et al. 2004).

PHBHx was found to be a suitable biomaterial for osteoblast attachment,

proliferation and differentiation from bone marrow cells. The cells on PHBHx

scaffolds presented typical osteoblast phenotypes: round cell shape, high alkaline

phosphotase (ALP) activity, strong calcium deposition, and fibrillar collagen

synthesis. After incubation for 10 days, cells grown on PHBHx scaffolds were

approximately 40% more than that on PHB scaffolds and 60% more than that on PLA

scaffolds. ALP activity of the cells grown on PHBHx scaffolds was up to about 65

U g-1 scaffolds, 50% higher than that of PHB and PLA, respectively (Wang et al.

2004).

Similarly, it was observed that chondrocytes isolated from rabbit articular

cartilage proliferated better on PHB scaffolds blended with PHBHx than on pure PHB

scaffolds. Chondrocytes proliferated on the PHB-PHBHx scaffold and preserved their

phenotype up to 28 days (Deng et al. 2003).

37 Chapter 1

1.8 Conclusions and Outlook

Polyhydroxyalkanoates with a wide range of physical properties are accessible

through biosynthesis in bacteria. It has been shown that they have a potential in

several medical applications. Unfortunately, inappropriate downstream processing of

the polymer resulted in contamination of PHAs by bacterial cell compounds and

therefore may have affected first studies in a negative way. Improved purification

methods have been described in the last years which were successful in reducing

pyrogenic contaminations. Recently, a type of PHA, P4HB, obtained the approval of

the U.S. Federal Drug Administration for application as a suture material. It is to be

expected that more PHAs will follow because material properties of PHAs can

already be tailored for particular applications during biosynthesis or later on by

chemical and physical modifications.

Introduction PHA 38

39

Chapter 2

Quantitative analysis of bacterial medium-

chain-length poly(R-3-hydroxyalkanoates) by

gas chromatography

Patrick Furrer, Roland Hany, Daniel Rentsch, Andreas Grubelnik, Katinka Ruth,

Sven Panke, Manfred Zinn. 2007. Journal of Chromatography A, 1143, 199-206.

Quantitative analysis of mcl-PHA 40

2.1 Introduction

Polyhydroxyalkanoates (PHAs) are a class of biodegradable and biocompatible

polyesters with many potential applications in the medical field (Chen et al. 2005). In

particular, the scientific and industrial research on elastomeric PHA with medium-

chain-length side chains (mcl-PHA, containing 3-hydroxyalkanoate monomers with

chain lengths from C6-C14) has been intensified in the last years (Chen et al. 2005;

Sodian et al. 2000; Stock 2000; Williams et al. 1999; Zinn et al. 2001).

PHAs are produced by microorganisms under unbalanced growth conditions

(Anderson and Dawes 1990). In order to control and optimize the conditions of

fermentation and down-stream processing, the quantitative analysis of PHA in

biomass and of purified polymer is of great importance. Different methods have been

applied to determine the content of PHA in biomass and to analyze the monomeric

composition of PHA-copolymers. After the discovery of poly(3-hydroxybutyrate)

(PHB), a short-chain-length PHA (scl-PHA), Lemoigne saponified the extracted PHB,

and determined the amount of PHB by gravimetry after applying a complicated

treatment (Lemoigne 1926). Slepecky and Law converted PHB with concentrated

sulphuric acid to crotonic acid and estimated the content via its UV absorption at 235

nm (Slepecky and Law 1960). Recently, Fourier transform infrared spectroscopy

(FTIR) has been applied to determine the content of PHA in cell suspensions (Jarute

et al. 2004; Randriamahefa et al. 2003). However, all these methods lack the

specificity to discriminate between different monomers and hence they can not be

used to determine the monomeric composition of PHA copolymers. 1H and 13C NMR experiments were applied to scl- and mcl-PHA without chemical

derivatisation of the polymer (Doi et al. 1995; Doi et al. 1986; Jan et al. 1996). For

quantitative measurements, NMR is limited to purified PHA. For scl-PHA, 1H NMR

measurements are sufficient to elucidate their composition, whereas for mcl-PHA

time-consuming 13C-NMR measurements are necessary. The need for faster

analytical methods for the determination of PHA in biomass directly forced the

development of chromatographic methods such as high performance liquid

chromatography (HPLC) or gas chromatography (GC). This requires a quantitative

depolymerisation of the polymer, usually combined with a derivatisation. Ion-

exchange HPLC with conductivity detection was applied for the analysis of digested

41 Chapter 2

poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV) in activated sludge

(Hesselmann et al. 1999). However, the use of GC is more common due to the

simple detection by flame ionisation (FID). The first derivatisation methods for GC

were developed for PHB (Braunegg et al. 1978; Riis and Mai 1988). Sulphuric acid in

methanol and later hydrochloric acid in n-propanol were used as reagent for the

transesterification, forming methyl and propyl esters, respectively. In contrast to the

alkaline hydrolysis (Wallen and Rohwedder 1974), which led to a mixture of 3-

hydroxyalkanoic acid methyl esters and 2-alkenoic acid methyl esters, the acidic

transesterification resulted in the ideal case in only one methyl ester per component.

In addition, it was considerably faster because extraction and transesterification

proceeded in one step. Certain modifications have been introduced for the acidic

transesterification method of PHA in biomass, like an extended reaction time for the

analysis of PHB (Huijberts et al. 1994; Jan et al. 1995), and an increased

concentration of sulphuric acid for scl-PHA (Oehmen et al. 2005) and mcl-PHA

(Lageveen et al. 1988; Lee and Choi 1995). For mcl-PHA, sulphuric acid in methanol

was used (Huijberts et al. 1994; Lageveen et al. 1988; Lee and Choi 1995). However,

sulphuric acid is of limited use as a general catalyst for transesterification reactions,

due to further decomposition of the 3-hydroxy esters (Oehmen et al. 2005; Riis and

Mai 1988), e.g. by acid catalyzed elimination.

So far, the transesterification of mcl-PHA has not been studied in detail. In this

work, we show that the methods developed for scl-PHA can not be adapted to mcl-

PHA without reservation. This is because the reaction kinetics differ significantly

between scl- and mcl-PHAs and side products occur for certain side-chain

functionalized mcl-PHAs, e.g. olefinic mcl-PHAs. Both effects lead to a significant

underestimation of the effective PHA content and to an incorrect copolymer

composition. The aim of this work was to develop a novel method for the accurate

quantification of mcl-PHA by GC for purified polymer as well as for mcl-PHA in

biomass directly. We found that the Lewis acid boron trifluoride (BF3) in methanol is

an effective transesterification reagent without significant formation of side-products.

We attribute this mainly due to the avoidance of water and protic acids. BF3 in

methanol is well known for the derivatisation of fatty acids (Lopez-Lopez et al. 2001;

Moffat et al. 1991; Rotzsche 1991; Shantha et al. 1993). The transesterification with

BF3-methanol was quantitative under optimized conditions for all mcl-PHAs analyzed

Quantitative analysis of mcl-PHA 42

in this study, and the copolymer composition could be confirmed by NMR

measurements.

2.2 Materials and methods

PHA derivatisation. The recovery and the composition of different purified PHAs

and of PHAs in biomass directly was determined by GC after their depolymerisation

and derivatisation to methyl or propyl esters. Each sample was derivatised twice for

independent analysis. The propanolysis according to Riis and Mai was applied to a

set of scl- and mcl-PHA as described by Riis and Mai (Riis and Mai 1988).

Further, this method was adapted for mcl-PHA to improve the recovery of propyl

esters. The following conditions were used: About 10 mg of polymer was transferred

into a 10 mL Pyrex tube. One mL of methylene chloride containing 10 mg mL-1 of

2-ethyl-2-hydroxybutyric acid as internal standard was added and the polymer was

let to dissolve at room temperature for one hour. One mL of a 20/80 (v/v) mixture of

HCl (37%) and n-propanol was added, the tube was tightly sealed, and the mixture

was vigorously shaken. The tube was placed in an oven at 80 °C for 16 h.

Subsequently, it was cooled to room temperature and 2 mL of demineralized water

was added. The mixture was vigorously shaken and after phase separation, the

aqueous (upper) phase was removed. The organic phase was dried over Na2SO4,

neutralized by adding Na2CO3, and filtered through a 1.0 µm nylon filter. In contrast

to the original method (Riis and Mai 1988), we chose a lower reaction temperature

but an extended reaction time. Furthermore, methylene chloride was used as solvent

instead of 1,2-dichloroethane and 2-ethyl-2-hydroxybutyric acid was used as internal

standard instead of benzoic acid. Additionally, the organic phase was dried and

neutralized to remove residual water and acid.

A similar transesterification procedure was developed using boron trifluoride

(BF3) as catalyst. In contrast to the transesterification catalyzed by hydrochloric acid,

the transesterification catalyzed by BF3 did not work with n-propanol. Therefore

methanol was used instead, forming methyl esters. About 10 mg of polymer was

transferred into a 10 mL Pyrex tube. One mL of methylene chloride containing 10

mg mL-1 of 2-ethyl-2-hydroxybutyric acid as internal standard was added. The

polymer was let to dissolve at room temperature for one hour. One mL of BF3 (1.3 M)

43 Chapter 2

in methanol was added, the tube was tightly sealed and the mixture was vigorously

shaken. The tube was placed in an oven and kept at 80 °C for 20 h. The subsequent

extraction and drying procedure was similar to the procedure described above,

except that the sample was extracted twice with 1 mL of a saturated NaCl solution for

a complete removal of the catalyst. Method optimization was carried out with 50 mg

of polymer and corresponding volumes of solvent and reagent to decrease the

measurement uncertainty. The measurement uncertainty was calculated from the

standard deviation of the transesterification (n=8) combined with uncertainties

inherent to the analytical system. Thereby, the mass of derivatised PHA was

important. The transesterification of larger masses (50 mg instead of 10 mg)

decreased the measurement uncertainty from ± 7.2% to ± 4.0%.

For the direct analysis of PHA, 50 mg biomass was derivatised in the same way

as described for 10 mg of purified PHA. For comparison, 5 g of biomass was

extracted twice with 100 mL methylene chloride for 24 h. The PHA solution was

filtrated, concentrated to 20% (v/v) and the polymer was precipitated by adding 200

mL of ethanol at 0 °C. The polymer was dried under vacuum at 40 °C for two days.

Analytical methods. For gas chromatographic analysis, one µL of the ester solution

was analyzed on a GC (GC 8575 Mega 2, Fisons Instruments, Rodano, Italy)

equipped with a Supelco SPB-35 (30 m x 0.32 mm) column with a film thickness of

0.25 µm (Supelco, Bellefonte, USA) at a split ratio of 1:10 and an initial temperature

of 80 °C. The temperature was raised with a rate of 10 °C min-1 to 240 °C. For

quantification purposes by FID, known amounts of pure 3-hydroxybutyric, -valeric, -

hexanoic, -octanoic, -nonanoic and -undecanoic acid were derivatised and measured

to calculate their response factors. 3-Hydroxyalkenoic acids were not available

commercially and their response factors were assumed to be similar to those of their

corresponding saturated alkanoic acids. The response factor of 3-hydroxy-6-

heptenoic acid was calculated by linear interpolation. 2-Ethyl-2-hydroxybutyric acid

(Sigma-Aldrich, Buchs, Switzerland) was used as internal standard because of its

chemical similarity to the 3-hydroxyalkanoic acids and because its elution time is

different to the 3-hydroxyalkanoic acids measured. Tetradecane (Sigma-Aldrich,

Buchs, Switzerland) was used as second internal standard. 3-Hydroxyalkanoic acids

were purchased from Larodan (Malmö, Sweden).

Quantitative analysis of mcl-PHA 44

NMR experiments in solution were performed on a Bruker AV-400 spectrometer

at 297 K using a 5 mm broad-band probe. For PHBV samples, typically 5 mg polymer

were dissolved in 0.7 mL CDCl3, for purified mcl-PHA samples, 30 mg per 0.7 mL

CDCl3 were used. NMR samples of whole cells were prepared by stirring 50 mg

biomass in 1 mL CDCl3 for 5 h followed by filtration. Chemical shifts are given in ppm

relative to the remaining signals of chloroform as internal reference (1H NMR: 7.26

ppm; 13C NMR: 77.0 ppm). Proton (1H) NMR spectra were recorded at 400.13 MHz

with a 9.6 µs 90° pulse length, 4460 Hz spectral width, 64 k data points, and 24

scans with a relaxation delay of 20 s were accumulated. Carbon (13C) NMR spectra

at 100.61 MHz with 1H WALTZ decoupling were recorded for the mcl-PHA samples

with the following parameters: 3.2 µs 45° pulse length, 26250 Hz spectral width, 64 k

data points, 5000 scans, relaxation delay 10 s, and decoupling field 2.5 kHz. 13C

NMR chemical shifts of PHA monomer units are compiled in the Supplementary

Information.

Propyl and methyl esters and side-products were analyzed by HPLC on a C18

Nucleosil column (2 x 250 mm, 3 µm, 100 Å, Macherey-Nagel Inc., Easton, USA).

Separation was achieved using a linear gradient from 10% to 80% acetonitrile in 20

min. 0.1% acetic acid in double distilled water was used as mobile phase. The flow

rate was 0.2 mL min-1 with injection volumes of 7.5 µL. Sample compounds were

identified by mass spectroscopy with an APCI ion source and an ion trap mass

detector (esquire HCT, Bruker Daltonics, Bremen, Germany).

To describe the monomeric composition of PHAs, the notation Cn:x was used,

where n is the number of carbon atoms and x indicates the position of the double

bond.

PHA biosynthesis. The following PHAs were produced by cultivation of Cupriavidus

necator for scl-PHA and with Pseudomonas putida GPo1 for mcl-PHA as described

elsewhere (Hartmann et al. 2006; Zinn et al. 2003) (see Table 2.1): Poly(3-

hydroxybutyrate) (PHB), from a carbon feed of 100% butyric acid; poly(3-

hydroxybutyrate-co-3-hydroxyvalerate) (PHBV), from a carbon feed of 80% (mole

mole-1) butyric acid and 20% (mole mole-1) valeric acid (PHBV-a), and 100% valeric

acid (PHBV-b); poly(3-hydroxyoctanoate-co-3-hydroxyhexanoate) (PHO), from a

carbon feed of 100% octanoic acid; poly(3-hydroxy-10-undecenoate-co-3-hydroxy-8-

45 Chapter 2

nonenoate-co-3-hydroxyoctanoate-co-3-hydroxy-6-heptenoate-co-3-hydroxy-

hexanoate) (PHOUE), from a carbon feed of 46% (mole mole-1) octanoic acid and

54% (mole mole-1) 10-undecenoic acid; poly(3-hydroxy-10-undecenoate-co-3-

hydroxy-8-nonenoate-co-3-hydroxy-6-heptenoate) (PHUE), from a carbon feed of

100% 10-undecenoic acid; and poly(3-hydroxyundecanoate-co-3-hydroxynonanoate-

co-3-hydroxyheptanoate-co-3-hydroxypentanoate) (PHUA), from a carbon feed of

100% undecanoic acid. Polymers were purified by repeated extraction and

precipitation with methylene chloride and ethanol, respectively.

2.3 Results and discussion

The propanolysis method developed by Riis and Mai (Riis and Mai 1988) was applied

to various scl- and mcl-PHA. The data obtained with GC-FID were compared with the

results from NMR as shown in Table 2.1.

The monomeric composition of PHBV copolymers was determined from

quantitative 1H NMR spectra. The signals of the methyl groups of HB units at 1.28

ppm and of HV units at 0.90 ppm are well separated, and the ratio between these

intensities was used to calculate the polymer composition. The overall intensities of

the methine HB- and HV- resonances at 5.26 and 5.16 ppm such as that of the

backbone methylene protons resonating at 2.54 ppm were used to cross-check the

individual HB and HV fractions.

The copolymer composition of mcl-PHAs was also investigated with NMR

spectroscopy. The combined fractions of (C6:0 + C8:0) units and (C7:6 + C9:8 +

C11:10) units were readily obtained from 1H NMR spectra by integration of the

overlapping methyl group signals at 0.87 ppm and the olefinic proton signals at 5.78

ppm. For standardization, the intensities of the backbone –CH protons resonating at

5.17 ppm was set to 100 (mol%). The signals of the different saturated and

unsaturated PHA units overlap in the 1H NMR spectrum, but the resonances are

separated in the corresponding 13C NMR spectrum (Supplementary Information

Table 2.3). In these spectra, the intensity of the backbone -CH carbons resonating at

69.8 – 71.3 was then set to 100 (mol%). This resulted in relative intensities of 99 -

103 for the overlapping carbonyl carbons at 168.8 – 169.8 ppm and of 96 - 101 for

the backbone methylene carbons at 39.1 ppm. From this good agreement between

Quantitative analysis of mcl-PHA 46

the intensities for backbone carbons C1, C2 and C3 we concluded that a relaxation

delay of 10 s between individual pulses was sufficient to yield quantitative 13C NMR

data, and the intensities of resolved carbon signals of different monomer units were

used to determine the individual fractions. For example for sample PHOUE,

integration of the resonances at 13.9, 22.5 and 31.5 ppm for C8:0 yielded an intensity

of (32.9 ± 1.7). Similarly, fractions of C6:0 = (7.1 ± 0.8), C7:6 = (12.7 ± 0.2), C9:8 =

(38.2 ± 0.3) and C11:10 = (9.1 ± 0.4) were determined. The sum of C6:0 and C8:0

units is 40 mol%, close to the 41.3 mol% determined from the 1H NMR spectrum.

Also for the sum of unsaturated monomer units, the agreement between the 13C

NMR (60 mol%) and 1H NMR (58.7 mol%) data is satisfactory. NMR results were

converted to % w/w for direct comparison with GC data.

The original propanolysis method was quantitative for the transesterification of

PHB and PHBV containing a low amount of valerate. Nearly 100% of the polymer

was recovered as propyl ester. However, for PHBV with a high amount of valerate

and for PHAs with longer side-chains the reaction was not quantitative and the

recovery of propyl esters fell below 50% for PHU. The reaction time was not sufficient

for a complete reaction of mcl-PHA and side-products were observed for PHA with

unsaturated side-chains.

Therefore we tried to improve the HCl-propanol method for PHO and PHUE, two

common mcl-PHAs (Bear et al. 1999; Dufresne et al. 2001). Experiments with these

two co-polymers revealed that the maximum recovery of 97% for PHO, and 55% for

PHUE was reached after 16 h of propanolysis at 80 °C. Increasing the reaction

temperature or reaction time decreased the recovery of PHUE due to the formation of

side-products. Already after 16 h side-products were observed for PHUE in the GC-

chromatogram as shown in Fig. 2.1a. In contrast to PHUE the recovery of PHO

reached a plateau after 16 h and no considerable side-products were observed. The

improved HCl-propanol transesterification method (16 h at 80 °C, [HCl] = 1.2 M) was

quantitative for scl- and mcl-PHAs with saturated side-chains. Table 2.1 shows that

the recovery had significantly decreased for PHA containing unsaturated side-chains,

due to the formation of side-products (see Fig. 2.1a). Thus HCl-propanol is an

effective reagent for the transesterification of all PHAs with saturated side-chains

under these conditions.

47 Chapter 2

Table 2.1: Monomeric composition and recovery (% w/w) of different scl- and mcl-PHA analyzed with different GC-transesterification methods and with 1H NMR for scl–PHA and 13C-NMR for mcl-PHA, respectively. For GC-measurements with 10 mg of PHA a measurement uncertainty of ± 7.2% was determined.

PHAa Analysis C4:0 C5:0 C6:0 C7:0 C8:0 C9:0 C11:0 C7:6 C9:8 C11:10 Total

NMR 100

BF3-MeOH 45 45

HCl-PrOH1 98 98

PHB

HCl-PrOH2 100 100

NMR 81 19

BF3-MeOH 34 12 46

HCl-PrOH1 79 18 97

PHBV-a

HCl-PrOH2 78 19 97

NMR 35 65

BF3-MeOH 17 48 65

HCl-PrOH1 33 65 98

PHBV-b

HCl-PrOH2 29 59 88

NMR 14 86

BF3-MeOH 11 90 101

HCl-PrOH1 12 85 97

PHO

HCl-PrOH2 9 64 73

NMR 6 32 11 40 11

BF3-MeOH 4 31 10 39 15 99

HCl-PrOH1 5 30 3 21 6 65

PHOUE

HCl-PrOH2 4 25 3 18 5 55

NMR 23 60 17

BF3-MeOH 22 59 17 98

HCl-PrOH1 6 41 8 55

PHUE

HCl-PrOH2 8 30 7 45

NMR 1 33 50 16

BF3-MeOH 2 34 49 15 100

HCl-PrOH1 2 32 51 14 99

PHUA

HCl-PrOH2 2 22 33 9 66 aFermentation conditions see materials and methods. 1Transesterification method with HCl-propanol adapted for mcl-PHA: 16h, 80 °C, [HCl]=1.2 M. 2Transesterification method with HCl-propanol developed by Riis and Mai for PHB.

Quantitative analysis of mcl-PHA 48

Fig. 2.1: GC-FID-chromatograms of PHUE transesterified with the adapted HCl-propanol method of this work (a) and with BF3-methanol (b).

We suspected the C7:6-monomer to excessively undergo side-reactions during

propanolysis and to form the side-products observed by GC. To confirm this

hypothesis and to elucidate the structure of the side-products, 3-hydroxy-6-heptenoic

acid produced by fermentation and separated by column chromatography (Ren et al.

2005) was propanolysed for 2 days with HCl-propanol at 100 °C to quantitatively

convert it into side-products. LC-MS measurements showed that these products had

identical retention times and mass spectra as the dominant side-products from PHUE

obtained by transesterification with HCl-propanol. Two products with a molecular

mass of 186 u were detected with LC-MS, both possessing the same fragmentation

pattern. NMR experiments revealed that the two products were (2-S, 5-R) and (2-R,

5-R) diastereomers of (5-methyl-tetrahydrofuran-2-yl)-acetic acid propyl ester as

shown in the supplementary information, Fig. 2.8.

These side-products were generated by the intramolecular cyclisation of propyl

3-hydroxy-6-heptenoate. We propose that a secondary cation is formed, catalysed by

protic acids and that the furane ring is built by an intramolecular nucleophilic attack,

as shown in Fig. 2.2A. The chiral centre in position 5 is in R configuration due to

biosynthesis whereas the chiral centre in position 2 is generated during cyclisation.

The ratio between the (2-S, 5-R) and (2-R, 5-R) diastereomers is 69:31 according to 1H NMR. The chirality in position 5 induces asymmetry in the new chiral centre in

position 2. The formation of pyrane analogs was not observed, since the formation of

an intermediate primary cation is less favoured. Likewise, the formation of 7- and 9-

0 5 1 0 1 5

C 1 1 :1 0

C 9 :8

C 7 :6Rel

ativ

e in

tens

ity [a

.u.]

T im e [m in ]

a

b

S id e p ro d u c tsC 7 :6

C 9 :8

C 1 1 :1 0

In te rn a l s ta n d a rd

49 Chapter 2

membered rings from propyl 3-hydroxy-8-nonenoate and propyl 3-hydroxy-10-

undecenoate was not observed. The addition of H2O to the double bond led to further

side-products. This reaction most likely proceeds again through the secondary

cation. Subsequently, H2O acts as nucleophile to generate a secondary alcohol as

shown in Fig. 2.2B. According to the NMR data for PHU, 32% and 53% of the C9:8-

and the C11:10-propyl esters were lost by this and other side-reactions, whereas

74% of the C7:6-monomer was lost mainly through cyclisation.

O

O

nHO

O

O

O

O

O

H

H+,∆T

-H+

n-propanol

A)

O

O

n

H+, ∆Tm

O

O

n

m

H

O

O

n

m

H

OH

H2O, ∆T

-H+

B)

Fig. 2.2: Side-reactions of PHUE during transesterification with HCl-propanol. A) Cyclisation of propyl 3-hydroxy-6-heptenoate. B) Water addition to unsaturated side-chains (m = 1-3).

Improved results for the transesterification of mcl-PHAs were obtained with

BF3-methanol as derivatisation reagent. PHO and PHUE were used to optimize the

conditions of this reaction. Best results were obtained for a reaction time of 20 h at 80

°C with a BF3-concentration of 0.65 M. Similar to the transesterification with HCl-

propanol, PHO was quantitatively methanolysed with BF3-methanol without forming

side-products. After 8 h at 100 °C and 20 h at 80 °C the maximal recovery was

reached as shown in Fig. 2.3. Further reaction time did not reduce the recovery

substantially. showed a similar reactivity, but the conversion to side-products was

Quantitative analysis of mcl-PHA 50

observed when the reaction time was further elongated, as shown in Fig. 2.4. At the

maximum recovery at 80 °C only negligible peaks of side-products were observed in

the chromatogram and about 98% of the polymer could be recovered as methyl

esters.

The influence of the boron trifluoride concentration in the reaction mixture was

studied for PHO and PHUE derivatised at 80 °C for 20 h. The optimum boron

trifluoride concentration was 0.65 M for both polymers. Increasing this concentration

to 0.87 M reduced the recovery of PHO from 100 to 53% (w/w) and the recovery of

PHUE from 98 to 61% (w/w) (Fig. 2.5 and Fig. 2.6). Similarly, a decrease in the

concentration of BF3 to 0.16 M reduced the recovery of PHO to 34% (w/w) and the

recovery of PHUE to 24% (w/w).

Low recoveries can also be caused by steps following the derivatisation of PHA.

Jan et al. (Jan et al. 1995) described the loss of short 3-hydroxyalkanoic acid methyl

esters during aqueous extraction, leading to incorrect results. To exclude this

phenomenon for mcl-PHA, the aqueous phases of the extraction from PHO and

PHUE were dried and analyzed, but no methyl esters could be detected. Hence,

3-hydroxyhexanoic methyl ester and the corresponding monomers with longer chains

did not migrate to the aqueous phase.

To study the possible influence of the polymer concentration on the recovery of

the BF3-transesterification, the concentration of PHO was varied between 0.1 - 21.4

mg mL-1. The recovered ratio for C8:0 to C6:0 was 8.9 (w/w) in all cases. However,

the total recovery decreased slightly for small concentrations. For concentrations

above ~3 mg mL-1, the recovery was always greater than 97% (w/w), and was

approximately (93.5 ± 1)% (w/w) in the concentration range 0.1 – 1.5 mg mL-1. This is

attributed to adsorptive effects of glass surfaces and of the GC injection system.

Similar trends were observed for the BF3-transesterification of PHU. This

concentration was varied between 0.1 – 19.7 mg mL-1. For example for C7:6, the

mean value of recovery was 20% (w/w), with maximum deviations of +3.4% (w/w) for

a concentration of 19.7 mg mL-1, and -3.5% (w/w) for 0.2 mg mL-1.

51 Chapter 2

0 25 500

50

100R

ecov

ery

[% w

/w]

Derivatisation time [h]

abcd

e

f

Fig. 2.3: BF3-methanol transesterification of PHO at 80 °C (�) and 100 °C (�), [BF3] = 0.65 M. a) sum of monomers at 80 °C, b) sum of monomers at 100 °C, c) 3-hydroxyoctanoate at 80 °C, d) 3-hydroxyoctanoate at 100 °C, e) 3-hydroxyhexanoate at 100 °C, f) 3-hydroxyhexanoate at 80 °C. A measurement uncertainty of ± 4% was determined for 50 mg of PHA.

0 25 500

50

100

c

a

b

hgfe

Rec

over

y [%

w/w

]

Derivatisation time [h]

d

Fig. 2.4: BF3-methanol transesterification of PHUE at 80 °C (�) and 100 °C (�), [BF3]=0.65 M. a) sum of monomers at 80 °C, b) sum of monomers at 100 °C, c) 3-hydroxy-8-nonenoate at 80 °C, d) 3-hydroxy-8-nonenoate at 100 °C, e) 3-hydroxy-6-heptenoate at 80 °C, f) 3-hydroxy-10-undecenoate at 80 °C, g) 3-hydroxy-10-undecenoate at 100 °C, h) 3-hydroxy-6-heptenoate at 100 °C. A measurement uncertainty of ± 4% was determined for 50 mg of PHA.

Quantitative analysis of mcl-PHA 52

0.1 0.5 0.90

50

100

ab

c

Rec

over

y [%

w/w

]

Concentration of BF3 [M]

Fig. 2.5: BF3-methanol transesterification of PHO at 80 °C for 20 h with variable concentrations of BF3. a) sum of monomers, b) 3-hydroxyoctanoate, c) 3-hydroxyhexanoate. A measurement uncertainty of ± 4% was determined for 50 mg of PHA.

0.1 0.5 0.90

50

100

Rec

over

y [%

w/w

]

Concentration of BF3 [M]

a

b

cd

Fig. 2.6: BF3-methanol transesterification of PHUE at 80 °C for 20 h with variable

concentrations of BF3. a) sum of monomers, b) 3-hydroxy-8-nonenoate, c) 3-hydroxy-10-undecenoate d) 3-hydroxy-6-heptenoate. A measurement uncertainty of ± 4% was determined for 50 mg of PHA.

53 Chapter 2

The transesterification with BF3-methanol was quantitative for all mcl-PHAs

and the data were in good agreement with the results obtained by NMR as shown in

Table 2.1. In contrast to mcl-PHAs, scl-PHAs were only recovered to 45-65% (w/w).

This loss is due to the partitioning of short-chain methyl esters between the organic

and the aqueous phase during extraction as described previously (Huijberts et al.

1994; Jan et al. 1995). In this work, NaCl-saturated water (Niskanen et al. 1978) was

used instead of pure water, but this could not inhibit an extensive loss of short-chain

methyl esters in our experiments. Hence, methanolysis should not be applied to scl-

PHA when an aqueous extraction is necessary to remove reagents. In case of PHUE

negligible amounts of side-products were observed. Compared with the HCl-

propanolysis, more than three times the amount of C7:6-monomer, and increased

amounts of C9:8 and C11:10 monomers were detected. The recovery improved from

55 to 98% (w/w).

The quantitative determination of PHA in biomass is of particular importance.

Therefore, the transesterification method with BF3-methanol was applied to biomass

with different contents of PHOUE. Equal amounts (50 mg) of these biomass samples

were methanolysed to study if other biomass components interfere with the methyl

esters from PHA. A small number of weak additional peaks was observed, but they

did not superpose the peaks of interest. The monomeric composition of PHOUE

remained nearly constant for samples 1-4 as shown in Table 2.2. For instance, the

relative amount of C8:0 in the polymer varied between 43.4% and 42.1% (w/w). Only

in sample 5, larger deviations were observed.

To determine the PHA content alternatively, large samples of biomass (5 g) were

extracted twice with methylene chloride and the polymer was precipitated with

ethanol. The mass of PHOUE was determined by gravimetry after drying under

vacuum. PHA contents between 3.4% and 16.8% (w/w) and between 2.9% and

15.9% (w/w) were determined by gravimetry and GC, as shown in Table 2.2. The

correlation with the gravimetric determination showed a relative difference between

0% and 5.5% except at the lowest PHA concentration where the difference was

14.7% (Table 2.2). The extraction recovery was higher than the one determined by

GC. This is reasonable since non-polar cell components like membrane lipids can be

co-extracted and lead to an overestimation of the PHA content determined by

gravimetry. Especially at low PHA contents this effect became marked. Further, we

Quantitative analysis of mcl-PHA 54

observed that biomass with low contents of PHA comprises higher amounts of non-

polar extractable impurities.

Table 2.2: Comparison between the quantitative analysis of mcl-PHA in biomass with GC after methanolysisa and the gravimetric determination after extractionb. Five biomass samplesc with different contents of PHOUE were analyzed to determine the relative monomeric composition (% w/w) and the PHA content (% wPHA/wbiomass).

Sample C6:0

GC

C8:0

GC

C7:6

GC

C9:8

GC

C11:10

GC

PHA

content

GC

PHA

content

gravimetry

Relative

difference

PHA

content

1 5.0 43.4 10.1 32.1 9.4 15.9 16.8 5.4

2 5.0 43.3 10.0 32.5 9.2 12.0 12.7 5.5

3 5.1 42.9 10.2 32.6 9.2 9.8 9.8 0.0

4 5.3 42.1 10.5 33.3 8.8 5.7 5.7 0.0

5 3.5 41.4 10.3 34.5 10.3 2.9 3.4 14.7

a) 50 mg of biomass was methanolysed with BF3-methanol (at 80 °C for 20 h). The measurement uncertainty is ≥ 7%. b) 5 g of biomass was extracted twice with methylene chloride, the solution was concentrated and precipitated with an excess of ethanol. The measurement uncertainty is around 5%. c) Sample 1 and 5 were produced by continuous fermentation under different conditions. Samples 2-4 were obtained by mixing sample 1 and 5 with varying ratios.

2.4 Conclusions

GC analysis is a common and efficient tool to quantify bacterial PHA. This study

however emphasised the need for novel derivatisation methods for the analysis of

mcl-PHA. A common transesterification method (Riis and Mai 1988) using HCl-

propanol was shown to be inadequate for the quantitative analysis of mcl-PHA due to

a serious underestimation of the PHA content. Although the adaptation of this

method was successful for mcl-PHA with saturated side-chains, another

derivatisation method was needed for mcl-PHA with unsaturated side-chains, to

avoid an excessive formation of side-products. The main side-products of PHUE

55 Chapter 2

derivatised with HCl-propanol were identified as the diastereomers of propyl 5-

methyltetrahydrofuran-2-acetate formed by cyclisation of propyl 3-hydroxy-6-

heptenoate, and as the products of water addition at the double bond. Other protic

catalysts like sulphuric acid will produce the same side-products and hence generate

incorrect results.

The formation of these side-products was avoided by using the Lewis-acid BF3 in

a water-free solvent. Transesterification with BF3-methanol was applied successfully

to all mcl-PHA investigated in this study. This novel method is well suited for the

detection of mcl-PHA in biomass where other analytical methods fail. In comparison

with the adapted HCl-propanol transesterification, only a few non-interfering

additional peaks from other lipids of biomass appeared in the spectrum, which

simplifies the analysis of data considerably.

NMR spectroscopy has been widely used to investigate various aspects of PHA,

including copolymer compositions, linkage sequences or metabolic pathway studies.

However, the measurement time for a single quantitative 13C spectrum was

approximately 16 hours; therefore, NMR spectroscopy is not suitable for the routine

determination of PHA copolymer composition. We also stirred whole cells in CDCl3

for 5 hours and measured NMR spectra directly after filtration. In these spectra,

several unidentified signals appeared that overlapped with PHA resonances. This

means that the dissolution of PHA from biomass is not selective and that also the

time-consuming extraction and purification steps can not be avoided for NMR

analysis of PHA. This is in contrast to the GC method with BF3-methanol presented in

this work that was used to quantify PHA in biomass directly. This method is

convenient to derivatise large amounts of samples in parallel and to measure them in

a short time, since one GC-measurement requires only about 20 minutes.

2.5 Acknowledgment

We gratefully acknowledge the financial support by the Empa R&D fund. We also

thank Thomas Ramsauer for assisting with polymer extractions, Elisabeth Michel for

supporting GC analyses, and Matthias Nagel for chemical advices.

Quantitative analysis of mcl-PHA 56

2.6 Supplementary information

Table 2.3: 13C NMR chemical shift assignments of the incorporated monomer units 3-hydroxypentanoate (C5:0), 3-hydroxyhexanoate (C6:0), 3-hydroxyoctanoate (C8:0), 3-hydroxy-6-heptenoate (C7:6), 3-hydroxy-8-nonenoate (C9:8), and 3-hydroxy-10-undecenoate (C11:10) in mcl-PHA.

δδδδ(13C) relative to δδδδ(CDCl3) = 77.0 ppm

Position C5:0 C6:0 C7:0 C8:0 C9:0 C11:0 C7:6 C9:8 C 11:10

1 168.8 – 169.8

2 38.7 39.1

3 69.8 – 71.3

4 26.7 35.9 33.5 33.6 33.8 33.8 32.9 33.8 33.8

5 9.3 18.3 27.1 24.7 25.0 25.1 29.2 24.5 25.0

6 13.8 22.4 31.5 29.0 29.5* 137.2 28.5 28.9*

7 13.9 22.5 31.7 29.4* 115.3 33.5 29.2*

8 13.9 22.5 29.2* 138.5 28.8*

9 14.0 31.8 114.6 33.7

10 22.6 139.0

11 14.1 114.2

* Positions not assigned

OO

O

8

176

2

3 4

510

9

Fig. 2.7: Numbering of carbon atoms used for the signal assignments of compound 1.

57 Chapter 2

Table 2.4: Chemical shifts of the furanoyl isomers.

Position a) δδδδ(1H) / ppm δδδδ (13C) / ppm

1a 1b 1a 1b

1 1.21 1.23 21.2 21.4

2 4.11 3.97 74.9 75.7

3 1.49/2.05 1.48/1.99 33.6 32.7

4 1.61/2.16 1.63/2.06 32.1 31.2

5 4.40 4.22 74.8 75.4

6 2.43/2.61 2.46/2.64 41.1 41.3

7 171.4 171.4

8 4.05 4.05 66.0 66.0

9 1.65 1.65 22.0 22.0

10 0.94 0.94 10.3 10.3

a) For atom numbers, see Fig. 2.7.

Quantitative analysis of mcl-PHA 58

Table 2.5: 1H, 13C-HMBC correlations used for NMR shift assignments and resolved 1H,

1H

coupling constants.

Compound HMBC correlations (w = weak) 1 J(H,H) / Hz

1a H-(1) → C-(2, 3); H-(3) → C-(1, 4); H-

(4) → C-(3, 6); H-(5) → C-(2, 3(w),

7(w)); H-(6) → C-(4, 5, 7); H-(8) → C-

(7, 9, 10); H-(9) → C-(8, 10); H-(10)

→ C-(8, 9)

J(1,2) = 6.1; J(5,6a) = 6.6; J(5,6b)

= 6.8; J(6a,6b) = 14.9; J(8,9) = 6.8;

J(9,10) = 7.5

1b H-(1) → C-(2, 3); H-(3) → C-(1,

4(w)); H-(4) → C-(3(w), 5(w)); H-(5)

→ C-(7(w)); H-(6) → C-(4, 5, 7); H-

(8) → C-(7, 9, 10); H-(9) → C-(8, 10);

H-(10) → C-(8, 9)

J(1,2) = 6.1; J(5,6a) = 6.7; J(5,6b)

= 6.8; J(6a,6b) = 15.1; J(8,9) = 6.8;

J(9,10) = 7.5

(S)

O(R) O

O

H3C

(R)

O(R) O

O

H3C

H

H

1a

1b

Fig. 2.8: Chemical structures with assignments of the absolute configuration of the mixture of isomers (1a and 1b) of (5-methyl-tetrahydro-furan-2-yl)-acetic acid propyl ester.

The signals at 4.40 and 4.22 ppm in the 1H NMR spectrum of the mixture of

isomers (1a and 1b, respectively) do not interfere with other resonances. Relative

amounts of 69% (1a) and 31% (1b) were determined from the integration of these

signals. The relative configurations were established with a series of 1D NOESY

spectra. When the pair of doublets at 1.2 ppm was selected, only H-(5) of 1a showed

the proximity to the corresponding methyl group, together with H-(2) for both

diastereomers at 4.11 and 3.97 ppm, respectively (see Fig. 2.9). Selective excitation

59 Chapter 2

at 4.40 ppm (H-(5) of 1a) showed an NOE to H-(1), whereas the analogous

experiment performed with H-(5) of 1b did not show an effect on the methyl group of

this isomer (spectra not shown).

4.14.24.34.4 ppm

a)

b)

4.14.24.34.4 ppm

a)

b)

Fig. 2.9: Sections of 1H NMR spectra of the 1a/1b mixture of diastereomers: a) 1D-NOESY experiment with selection of H-(1) at 1.2 ppm and b) normal 1H NMR spectrum.

Experimental

The 1H and 13C NMR spectra were recorded at 400.13 (100.61) MHz on a Bruker

Avance-400 NMR spectrometer. The 1D 1H and 13C NMR spectra and the 1H,1H and 1H,13C 2D correlation experiments were performed at 297 K using a 5 mm broadband

inverse probe with z-gradient (100% gradient strength of 10 G cm–1) and 90° pulse

lengths of 6.7 µs (1H) and 14.9 µs (13C). The chemical shifts were referenced

internally using the resonance signals of CDCl3 at 7.26 (1H) and at 77.0 (13C) ppm.

The gradient selected HSQC (Davis et al. 1992) (HMBC; (Bax and Marion 1988))

experiments were performed with selection of 1H,13C coupling constants of 145 (10)

Hz, gradient strengths of 80 : 20.1 : 11: –5 (50 : 30 : 40.1), applying a carbon

decoupling field of 3.1 kHz for the HSQC experiments (GARP decoupling; (Shaka et

al. 1985)). Before processing the data matrices of 1024 x 512 were zero filled to 1024

x 1024. The HSQC-TOCSY (Palmer et al. 1992) spectra were recorded with the

selection of 1J (1H,13C) = 145 Hz and a 27 µs 90° pulse length for the TOCSY transfer

with a total mixing time of 110 ms, applying the above mentioned carbon decoupling

Quantitative analysis of mcl-PHA 60

conditions, data matrices and processing conditions. The 1H,1H connectivity’s and the

relative configuration of the isomers were established by DQF-COSY and by 1D

NOESY experiments (mixing time of 1s, 50 ms selective 180° inversion pulse with

Gaussian shape) (Stott et al. 1997).

61

Chapter 3

Biosynthesis of medium-chain-length

poly(3-hydroxyalkanoate): endotoxin reduction by

phosphorus limitation

Biosynthesis of mcl-PHA under P-limitation 62

3.1 Introduction

PHAs with side chains of medium-chain-length (mcl-PHA, monomers with 6-14 car-

bon atoms) possess a wide range of physical and chemical properties due to their

manifold monomeric compositions. They are synthesized by microorganisms under

appropriate growth conditions, such as nitrogen limitation (Witholt and Kessler 1999).

The production of mcl-PHAs is well established with Gram-negative bacteria but its

contamination with lipopolysaccharide (LPS) from the outer membrane restrains the

application in the medical field (Valappil et al. 2007; Williams et al. 1999). LPS, also

referred to as endotoxin, represents one of the microbial molecular signals responsi-

ble for the activation of the innate immune system (Trent et al. 2006). It is recognised

by receptor TLR4 (toll-like receptor 4) that is present on many cell types like macro-

phages and dendritic cells. Even nanogram amounts of LPS cause fever, septic

shock and other immune responses when injected intravenously into humans.

Hence, its concentration in medical products has to be very low, e.g. the endotoxic

activity has to be smaller than 20 endotoxic units (EU) per medical device (U.S. De-

partment of health and human services 1997).

To obtain mcl-PHA of medical purity, several purification steps are necessary for

its recovery from biomass and special procedures have to be applied for endotoxin

removal (Lee et al. 1999; Williams et al. 2001). This complicates the downstream

processing and increases the production costs. Preventing the synthesis of LPS or

reducing their endotoxic activity during the production of mcl-PHA would drastically

simplify the recovery procedure. However, LPS are essential for the survival of

Gram-negative bacteria as they augment the rigidity of the cell wall and protect bac-

teria for instance from bile salts, gut enzymes and antibiotics (Holst 1995). Among

Gram-negative bacteria, only mutants of Neisseria meningitidis were viable without

LPS, presumably due to the presence of capsular polysaccharide (Bos and Tom-

massen 2005; Dixon and Darveau 2005; Raetz et al. 2007).

Indeed, the biological activity of LPS and in particular its endotoxic activity is

largely dependent on the structure of lipid A, the lipophilic part of LPS (Dixon and

Darveau 2005; Trent et al. 2006). Reducing the degree of acylation or phosphoryla-

tion drastically decreases the biological activity of LPS as shown by Rietschel et al.

(Fig. 3.1, (Rietschel et al. 1994)). Certain bacteria adapt the structure of lipid A to the

environmental growth conditions. Yersinia pestis for instance reduces the degree of

63 Chapter 3

acylation of lipid A during invasion of mammals to avoid the host’s immune defence

(Knirel et al. 2005). Furthermore, it has been observed that some pathogenic organ-

isms lack one or both phosphate groups of lipid A (Trent et al. 2006). But so far, it

has been unknown whether non-pathogenic bacteria can be stimulated by environ-

mental conditions to reduce the degree of phosphorylation of lipid A.

Fig. 3.1: Lipid A – structure activity relationships. Shown are independent chemical changes of the lipid A structure of Escherichia coli and the factor by which the modified structure is less biologically active than the reference lipid A (Rietschel et al. 1994).

In this work we investigated whether the endotoxic activity of crude PHA can be

lowered by phosphate-limited growth conditions such that the phosphorylation of lipid

A is reduced. This is of great importance because the endotoxic activity of lipid A is

decreased by a factor of 100 or more when one phosphate group is removed

(Rietschel et al. 1994). To realize such a scenario, we used carbon and nitrogen lim-

ited conditions ((C,N)-limitation) as starting point for the cultivation of Pseudomonas

putida GPo1 to induce PHA accumulation (Hartmann et al. 2006). Next, the phos-

phate concentration in the mineral medium was reduced stepwise to reach phospho-

rus limitation (P-limitation) as part of a triple (C,N,P)-limitation. Drastic changes in the

cell composition were observed with P-limitation.

Modification Reduction factor

1a 102

1b 102

1a + 1b Non toxic

2 107

3a 102

3a + 3b 107

4 102

5 101

Biosynthesis of mcl-PHA under P-limitation 64

3.2 Materials and methods

Biosynthesis of mcl-PHA

Polyhydroxyoctanoate (PHO) consisting of 87% (w/w) 3-hydroxyoctanoate and 13%

(w/w) 3-hydroxyhexanoate was produced by balance-controlled continuous cultiva-

tion of P. putida GPo1 (ATCC 29347) in a 3 L bioreactor (KLF 2000, Bioengineering,

Wald, Switzerland) with a working volume of 2.6 L and a dilution rate (D) of 0.125 h-1.

The preculture was prepared as described elsewhere (Durner et al. 2000; Hartmann

et al. 2006). An aliquot of 200 mL of a shake-flask culture was used to inoculate the

bioreactor containing 2.4 L of fermentation medium.

For continuous cultivation, the following medium was used (per liter): 0.80 g of

EDTA-Na2·2H2O, 3.78 g of (NH4)2SO4, 1.00 g of MgSO4·7H2O, 0.28 g of

FeSO4·7H2O, 0.03 g of CaCl2·2H2O and 1.05 g of MOPS was added as pH buffer.

The carbon to phosphorus ratio (C/P) in the mineral medium was adjusted by the ad-

dition of KH2PO4 to the medium (e.g. 1.17 g of KH2PO4 for C/P = 30 g g-1), whereas

the carbon source octanoic acid was pumped separately to the culture broth (result-

ing in a carbon feed concentration of 8.0 g C L-1). As a consequence the carbon to

nitrogen ratio (C/N) in the medium feed was fixed to 10 g g-1. For fermentations with

nitrogen excess the concentration of (NH4)2SO4 was increased to 11.34 g L-1 result-

ing in a C/N of 3.3 g g-1. Furthermore, 1 L of the medium was supplemented with 2

mL of trace element stock solution that contained per liter: 12.22 g of MnCl2·4H2O,

1.27 g of CoCl2·6H2O, 2.00 g of CuCl2·2H2O, 7.50 g of ZnSO4·7H2O, and 0.5 g of

Na2MoO4·2H2O in 1M HCl. The medium was sterilized by filtration. Octanoic acid

(95 % w/w) was pumped directly into the culture vessel by using a dosimat (Metrohm,

Herisau, Switzerland). The dissolved oxygen concentration was maintained between

30 and 50% relative to the maximum oxygen concentration in the aerated culture

medium before inoculation. Poly(propylene glycol) (Mw = 4000) was added to oc-

tanoic acid to 4% (v/v) to avoid excessive foam formation. For each phosphate con-

centration, at least five volume exchanges were allowed to reach steady state. Two

measurements were made within 24 h.

The fermentation broth was concentrated by centrifugation at 4’500 x g for 20

min. The paste was lyophilized for 48 h, ground, and the dry biomass was stored in a

closed container at -25 °C. The supernatant was filtrated (PES filter, 0.22 µm) and

stored at -25 °C.

65 Chapter 3

Analytical methods

PHA determination. The PHA content in biomass was determined in duplicate by

GC after propanolysis of biomass as described in Chapter 2.

Endotoxin detection. The endotoxic activity in aqueous solutions was determined

with the PyroGene recombinant factor C assay (Lonza Bioscience, Visp, Switzer-

land). For the endotoxic activity of biomass, 1 g of dried biomass was washed twice

with 50 mL of 0.9% NaCl by vortexing and centrifugation to remove free LPS. The

pellet was freeze dried. Twenty mg washed biomass was suspended in 10 mL buffer

(2.3 g SDS, 11.5 mL glycerol, 12.4 mL 0.5 M Tris-HCl and 76.1 mL H2O adjusted to

pH = 6.8 with HCl), heated up to 95 °C for 10 min to lyse the cells and then shaken at

150 rpm at room for 24 h. The suspension was filtrated (PES filter, 0.22 µm) and di-

luted in three steps (1:100, 1:100 and 1:500) to a total dilution of 1:5x106. One hun-

dred µl of this solution were used to measure the endotoxic activity with the recombi-

nant factor C assay. For the endotoxic activity of PHA, 200 mg of washed biomass

was extracted with 5 mL of methylene chloride under stirring at room temperature for

20 h. The suspension was centrifuged, filtrated (Nylon filter, 0.45 µm) and 3.5 mL of

the resulting polymer solution was poured into a small Petri dish (d=4 cm) that had

been depyrogenized previously at 200 °C for 6 h. The solvent was slowly evaporated

(3 days) to form PHA films with a homogenous surface. Endotoxins were extracted

from the surface with 3 mL of endotoxin-free water at room temperature for 24 h. An

aliquot of 100 µl of this solution was then used to measure the endotoxic activity.

The recombinant factor C of the Pyrogene assay is activated by endotoxin-

binding, and the active moiety created then acts to cleave a synthetic substrate,

which results in the generation of a fluorogenic compound. After one hour of incuba-

tion, the fluorescence at 440 nm can be measured by exciting at 380 nm and the en-

dotoxic activity is determined relative to an endotoxin standard. One endotoxic unit

(EU) corresponds to about 0.1 ng of standard endotoxin.

Biomass concentration. Cells were collected on pre-weighed polycarbonate filters

(0.2 µm, Nucleopore, Sterico AG, Dietikon, Switzerland). The filters were first washed

with 10 mM MgCl2, dried overnight at 105 °C and weighed. An aliquot of 3 mL of cell

suspension was filtered through the filter and the filter was dried again at 105 °C

overnight. The mass difference was used to calculate the concentration of dry bio-

mass in the culture.

Biosynthesis of mcl-PHA under P-limitation 66

Optical density. The cell suspension was diluted 1:100 with deionised water and the

optical density was determined at 450 nm.

Analyses of substrate concentrations in the culture supernatant. Ammonium

was measured by using a photometric ammonium test (Spectroquant, Merck Darm-

stadt, Germany). The measurement range was linear between 0.01 and 3 mg N L-1

(present as ammonium). If necessary, the samples were diluted with deionised water.

Octanoic acid was measured by gas chromatography after derivatisation to propyl

esters. An aliquot of 5 mL of methylene chloride containing 1 mg mL-1 of 2-ethyl-2-

hydroxybutyric acid as internal standard was added to 2 mL of supernatant in a 20

mL Pyrex tube. Five mL of a 20/80 (v/v) mixture of HCl (37%) and n-propanol was

added, the tube was tightly sealed, and the mixture was vigorously shaken. The tube

was placed in an oven at 80 °C for 16 h. The subsequent steps were similar to the

derivatisation of PHA as described above.

Phosphate was determined according to the ascorbic acid molybdane blue

method (Chen et al. 1956). An aliquot of 10 mL of supernatant was mixed with 5 mL

of freshly prepared reagent (containing sulphuric acid, ammonium molybdate and

ascorbic acid) and supplemented with deionised water to 50 mL. The mixture was

shaken and the absorbance at 800 nm was measured after 15 minutes of reaction at

room temperature. The measurement range was linear between 0.1 and 100 mg P

L-1 (present as phosphate), for higher concentrations dilution was necessary.

Analyses of biomass. Total phosphorus was determined as follows: About 300 mg

of dried biomass and 3 mL of HNO3 (65%) were added to a 20 mL cone. An aliquot of

2 mL of distilled water and 1 mL of H2O2 (30%) were added to a 30 mL vessel and

the cone was placed therein. The sample was digested in a microwave digester

(Milestone, MLS 1200 Mega, Sorisole, Italy) for 15 minutes with a stepwise increase

from 100 to 500 W. After cooling, the clear solution was filled up with distilled water to

a volume of 10 mL. The solution was diluted between 1:10 and 1:100 with 2% HNO3

and the emission was measured with ICP-OES (Perkin Elmer, Optima 3000,

Schwerzenbach, Switzerland) at 213.6 and 214.9 nm. Standard solutions were pre-

pared from a 1.00 mg P L-1 standard solution (Alfa Aesar, Karlsruhe, Germany) for

external calibration.

Phospholipids were extracted with chloroform-methanol according to method of

Folch (Folch et al. 1957). Quantitative analysis was done by a colorimetric method

based on the formation of a complex between phospholipids and ammonium

67 Chapter 3

ferrothiocyanate (Stewart 1980). Phospholipids extracted from 200 mg of dried bio-

mass were dissolved in 2 mL of chloroform. To an aliquot of 0.1 mL, 1.9 ml of chloro-

form and 2 mL of thiocyanate reagent was added. The mixture was vortexed for 1

min and centrifuged for 10 min at 2’000 x g. The lower layer was removed and the

light absorbance of this solution was measured at 488 nm. 1,2-Dipalmitoyl-sn-

glycero-3-phosphoethanolamine was used for calibration. The linear range was be-

tween 1.5 and 30 mg L-1.

3.3 Results and discussion

General observations

The adaptation of P. putida GPo1 to P-limitation was generally slow and a drastic de-

crease of the phosphorus concentration in the mineral medium feed (e.g. from 200 to

100 mg P L-1) was not tolerated by the cells. Excessive foam generation as well as

biofilm formation were observed that soon led to instable process conditions. Even

with small changes of the phosphorus concentration, wall growth could not be pre-

vented completely. C/P ratios higher than 90 g g-1 lead to instable cultures. One pos-

sible reason is that the specific growth rate of the cells was decreased by very low

phosphorus concentrations, so that octanoic acid could not be consumed completely,

resulting in a substrate inhibition finally leading to a wash-out of the culture. Under

phosphorus limitation, the cell shape became more rod-like and the brown colour of

the fermentation broth brightened.

Nutrient limitations

The accumulation of PHA in bacteria can be triggered by nutrient limitations such as

nitrogen limitation. Carbon limitation is commonly applied to prevent substrate inhibi-

tion. For the production of mcl-PHA with P. putida GPo1 the combined (C, N)-

limitation strategy has successfully been applied (Durner et al. 2000; Hartmann et al.

2006; Ruth et al. 2007).

Nutrient limitation has often been described using the kinetic model of Monod

(Monod 1949), although this model neglects bacterial adaptations to prolonged nutri-

ent limitation (Ferenci 1999). The concentration of the limiting nutrient (s) in the cul-

ture broth can be calculated with Monod’s equation if the specific growth rate (µ), the

maximum specific growth rate (µm) and the half saturation coefficient (Ks) are known:

Biosynthesis of mcl-PHA under P-limitation 68

[1]

For chemostat cultures at steady state the specific growth rate is equal to the di-

lution rate and consequently the concentration of the limiting nutrient is constant.

However, the Monod model is only valid for single nutrient limitation. A simple kinetic

model for dual and triple-limited fermentations does not exist. Hence a simple ap-

proximation was used to define kinetic nutrient limitation. Here limitation was defined

as the state where the residual concentration of a nutrient becomes virtually constant

at a low level. A further decrease of the nutrient concentration in the medium does

not affect the residual concentration significantly.

The measured concentration of a nutrient is in fact compromised by method of

measurement. During the time between sample drawing and centrifugation, the bac-

teria continue to consume the residual nutrients in the medium. The measured nutri-

ent concentrations are consequently lower than at the time point of sampling. Ex-

trapolating from the observed nutrient consumption rates, a decrease in phosphorus

concentration of about 0.7 mg L-1 has to be considered for a concentration of about 7

g biomass L-1 for a delay of 2 minutes between sampling and centrifugation. Simi-

larly, a reduction of 34 mg/L has to be taken into account for carbon and 3.4 mg/L for

nitrogen for a period of 2 minutes.

The limiting (measured) concentrations of nutrients obtained were 0.3 mg P L-1, 3

mg N L-1 and 10 mg C L-1. The values for phosphorus and nitrogen appeared reason-

able as similar values, 0.2 mg P/L and 2.2 mg N/L could be calculated with Ks values

from E. coli and P. putida PGA1 (17 µM for phosphate and 460 µM for ammonium,

respectively) using the Monod equation (Annuar et al. 2007; Owens and Legan

1987). According to the experiments, phosphorus limitation started at a C/P = 65

g g-1. Further, it was observed that nitrogen became non-limiting at a C/P = 75 g g-1

and octanoic acid at a C/P = 85 g g-1. According to these boundaries four regimes of

nutrient limitation could be distinguished:

- Dual (C,N)-limited regime: C/P = 30-65 g g-1 (I)

- Triple (C,N,P)-limited regime: C/P = 65-75 g g-1 (II)

- Dual (C,P)-limited regime: C/P = 75-85 g g-1 (III)

- Single P-limited regime: C/P = 85-95 g g-1 (IV)

s=KS

m−

69 Chapter 3

0.1

1

10

100

1000

1

10

100

1000

10000

20 30 40 50 60 70 80 90 1000.1

1

10

100

Con

cent

ratio

n of

nitr

ogen

[mg

L-1]

B

Con

cent

ratio

n of

car

bon

[mg

L-1]A

I II III IV

C/P [g g-1]

Con

cent

ratio

n of

phos

phor

us [m

g L-1

]

Fig. 3.2: Residual concentration of nutrients in the supernatant and their original concentration in the medium of fermentation 1. A) Free carbon from octanoic acid in medium (�), free carbon from octanoic acid in supernatant (�), free nitrogen (present as NH4

+) in medium (�), free nitrogen (present as NH4+) in supernatant

(�). B) Free phosphorus (present as phosphate) in medium (�), free phosphorus (present as phosphate) in supernatant (�), bound phosphorus in supernatant (�). Limitation regimes: I) (C,N)-limited, II) (C,N,P)-limited, III) (C,P)-limited, IV) (P)-limited.

Interestingly up to 6 mg P L-1 were found in extracellular substances under P-

limiting conditions (Fig. 3.2B). Likewise, in the N-limited regime (C/P = 30-70 g g-1) a

substantial amount of nitrogen was found as extracellular protein (140-200 mg pro-

tein L-1, corresponding to ~24-34 mg N L-1). Hence, the quantity of extracellular sub-

stances containing nitrogenous compounds was non-negligible. More detailed analy-

ses to identify the compounds were not carried out.

Biosynthesis of mcl-PHA under P-limitation 70

Biomass production and PHA accumulation

The optical density (OD) and the cell dry weight (CDW) increased towards higher C/P

and reached a maximum at C/P = 80 g g-1 (Fig. 3.3A). This increase was mainly due

to enhanced PHA accumulation as the concentration of residual biomass remained

nearly constant until C/P = 70 g g-1 and only slightly decreased towards higher C/P.

The PHA content of dried biomass increased with P-limitation (C/P > 60 g g-1) from

10 w% to 25 w% from regime I to regime III (Fig. 3.3B).

0

5

10

0

10

20

30

20 30 40 50 60 70 80 90 1000

10

20

Dry

wei

ghts

[g L

-1]

B

Opt

ical

den

sity

(45

0nm

) [-

]A

PH

A c

onte

nt [%

w/w

]

C/P [g g-1]

I II III IV

Fig. 3.3: A) Optical density (OD) measured at 450 nm (�), cell dry weight (CDW) (�), residual biomass weight (�), and PHA weight (�) and B) PHA content in dried biomass (�) of fermentation 1. Limitation regimes: I) (C,N)-limited, II) (C,N,P)-limited, III) (C,P)-limited, IV) (P)-limited.

Phosphorus and phospholipid in biomass

The phosphorus content in residual biomass and the content of phospholipids de-

creased with P-limitation (Fig. 3.4). Phosphorus in residual biomass was reduced un-

der P-limited conditions to virtually 50% of the maximum value, whereas the content

71 Chapter 3

in phospholipids was decreased to 10% of its maximum value. We therefore could

confirm that membrane phospholipids play a key role in the phosphorus pool of P.

putida GPo1, as has been observed for P. diminuta, P. aeruginosa and P. fluores-

cens (Minnikin and Abdolrah.H 1974; Minnikin et al. 1974; Yuasa 2002). For these

species, most of the membrane phospholipids were substituted under strong

P-limitation by nitrogen or sulphur containing lipids (Minnikin and Abdolrah.H 1974;

Minnikin et al. 1974; Yuasa 2002). Interestingly, under conditions of excess phos-

phate (C/P = 30 and 40 g g-1) the phospholipid content was also decreased relative

to the observed maximum at C/P of 50 g g-1. This may be due to a shift in the phos-

pholipid composition with excess phosphate eventually causing an underestimation.

20 30 40 50 60 70 80 90 1000

2

4

0

2

4

Con

tent

of p

hosp

holip

ids

in

resi

dual

bio

mas

s [%

w/w

]

C/P [g g-1]

Con

tent

of p

hosp

horu

s in

re

sidu

al b

iom

ass

[% w

/w]

I II III IV

Fig. 3.4: Content of phosphorus (�) and phospholipids (�) in residual biomass of fermentation 1. Limitation regimes: I) (C,N)-limited, II) (C,N,P)-limited, III) (C,P)-limited, IV) (P)-limited.

Biosynthesis of mcl-PHA under P-limitation 72

Endotoxins in biomass and extracted PHA

The endotoxic activity of biomass was considerably reduced with P-limitation, as

shown in Fig. 3.5A. Because the measurement of endotoxic activities suffered from a

large uncertainty, the fermentation experiment was repeated twice for independent

measurements. Large variations were observed between the three independent ex-

periments, but the same trend was clearly found for each experiment – the endotoxic

activity dropped between 12- and 20-fold. One reason for these variations might be

LPS that were released from the cell membrane, partially adhering to biomass during

centrifugation. It has been observed that special fermentation conditions trigger the

release of endotoxins (Svensson et al. 2005). Biomass was washed twice during

work-up to remove free LPS, but this procedure might not have been quantitative.

For fermentation 1 (F1) the endotoxic activity of biomass dropped from 1.46 x 106

(C/P = 30 g g-1) to 1.22 x 105 EU mg-1 biomass (C/P = 90 g g-1). Fermentation 2 (F2)

and 3 (F3) showed endotoxic activities of 2.35 x 106 and 3.55 x 106 EU mg-1 biomass

(C/P = 40 g g-1) and 1.04 x 105 and 2.62 x 105 EU mg-1 biomass (C/P = 90 g g-1), re-

spectively. An analogous trend was measured for PHA extracted from biomass al-

though here the effect was more pronounced. Methylene chloride was chosen as a

standard solvent because it is suitable for the extraction of all mcl-PHAs and is often

used for this purpose. Again the endotoxic activities of PHA from three independent

fermentation experiments varied considerably, but showed the same trend (Fig.

3.5B). In particular for the regime III and IV PHA showed a very low endotoxic activity

of 18, 66 and 28 EU g-1 PHA for F1, F2, and F3, respectively. In general, the en-

dotoxic activity of PHA was considerably reduced, e.g. by a factor of about 300 from

C/P = 30 g g-1 to CP = 90 g g-1 for F1, whereas the endotoxic activity of biomass was

reduced only by a factor of 12 for F1.

The reason for the decreased endotoxic activity of biomass and of PHA produced

under phosphorus limitation can be explained either by structural changes of lipid A,

most apparently from biphosphorylated lipid A to the mono- or non-phosphorylated

form (compare Fig. 3.1), or by a reduction of the concentration of LPS in the outer

membrane. Although the latter is rather unlikely because of the necessity for mem-

brane stabilisation and protection, we have so far been unsuccessful in confirming

the structural change of lipid A by separation of lipid A from other cell components

with thin layer chromatography and analysis by mass spectrometry. Thus, more de-

tailed investigations are required for clarification.

73 Chapter 3

100000

1000000

20 30 40 50 60 70 80 90 10010

100

1000

10000

End

otox

icity

[EU

mg-1

bio

mas

s]

End

otox

icity

[EU

g-1 P

HA

]

C/P [g g-1]

A

B

I II III IV

Fig. 3.5: A) Endotoxic activity of dried biomass B) Endotoxic activity of PHA extracted from dried biomass with methylene chloride. Results were obtained from three independent fermentation experiments. Fermentation 1 (�), fermentation 2 (�) and fermentation 3 (�). Limitation regimes: I) (C,N)-limited, II) (C,N,P)-limited, III) (C,P)-limited, IV) (P)-limited.

The consequences of P-limitation seem to be similar for phospholipids and LPS.

In contrast to phospholipids, the reduction of the endotoxic activity started already

when the concentration of free phosphorus was low (0.3-5 mg P L-1) but not limiting

and continued with P-limitation.

Yield coefficients

Carbon and nitrogen yield coefficients for residual (PHA-free) biomass were hardly

affected by P-limitation (Table 3.1). The carbon yield coefficient varied between 0.69

and 0.81 g g-1 and the nitrogen yield coefficient between 7.60 and 8.27 g g-1 for F1.

The phosphorus yield coefficient rose from 30 to 62 g g-1 for a change of C/P from 30

Biosynthesis of mcl-PHA under P-limitation 74

to 90 g g-1 for F1. F2 and F3 showed similar results. The carbon yield coefficients for

F2 and F3 varied only slightly as well as the nitrogen yield coefficients. The phospho-

rus yield coefficients increased by about 100% for a change of C/P from 40 to 90

g g-1 (F2 and F3).

By using the modified stoichiometric approach (Egli 1991), dual and triple-limited

growth regimes can be estimated. According to this approach growth limiting nutri-

ents are completely converted into biomass under nutrient limitation. The ratio of

these nutrient concentrations in the medium feed (Cf, Nf, and Pf, respectively) equals

the inverse ratio of the corresponding growth yield coefficients (YX/C, YX/N, and YX/P,

respectively). The growth yield coefficients for total biomass can be replaced in this

calculation by growth yield coefficients for residual biomass since this leads to the

same result. Under triple (C,N,P)-limitation the following equations must hold:

[2]

[3]

The yield ratios of F1 are shown in Table 3.1. According to the stoichiometric ap-

proach, triple limitation had to be found between C/P = 40 and 70 g g-1. For F2 and

F3 triple limitation was reached from C/P = 50 to 70 g g-1. This estimation was in

clear contrast to the previous determination of nutrient-limited growth regimes based

on the kinetic approach, where the triple-limited regime was determined from C/P 65

to 75 g g-1. The stoichiometric P-limitation was reflected in the decreasing phospho-

rus content in biomass (Fig. 3.4). P-limitation according to the kinetic approach went

together with a considerable increase of the PHA content in biomass. This was simi-

lar to N-limitation, where higher C/N ratios enhance PHA accumulation (Durner et al.

2000).

Hence, the stoichiometric approach indicated physiological adaptations to re-

duced concentrations of phosphorus. Effective limitation in the sense of very low re-

sidual phosphorus concentrations in the supernatant and enhanced PHA accumula-

tion was only determined with the kinetic approach.

Cf

P f

≈YX /P

C,N,P−lim.

YX /CC,N,P−lim.

Nf

Pf

≈YX /P

C,N,P−lim.

YX /NC,N,P−lim.

75 Chapter 3

Table 3.1: Yield factors for residual biomass and yield ratios for the determination of dual

and triple-limited growth regimes of fermentation 1. Dual limitation is reached

when the yield ratio equals the ratio of the corresponding nutrients in the feed

medium (C/P or N/P). For triple limitation both yield ratios have to fulfil this

condition.

Regime C/P N/P Y X/C YX/P YX/N YX/P/YX/C YX/P/YX/N

[g g-1] [g g-1] [g g-1] [g g-1] [g g-1] [g2 g-2] [g2 g-2]

I 30 3.0 0.79 30.1 7.93 38.0 3.8

I 40 4.0 0.77 31.9 7.73 41.2 4.1

I 50 5.0 0.76 39.2 7.60 51.6 5.1

I 60 6.0 0.81 48.8 8.11 60.3 6.0

II 70 7.0 0.80 57.6 7.99 72.1 7.2

III 80 8.0 0.76 62.1 8.07 81.7 7.7

IV 90 9.0 0.69 62.0 8.27 90.5 7.5

YX/Y: growth yield factor for residual biomass, YX/Y/YX/Z: yield ratio.

Cell physiological aspects with different nutrient limitations

Nitrogen limitation was applied in the previous experiments combined with C- and or

P-limitation because it was considered to be essential for PHA accumulation. To see

if the N-limitation played a crucial role for the PHA accumulation and for the en-

dotoxic activity, the concentration of nitrogen in the mineral medium was increased

threefold (resulting in a C/N = 3.3 g g-1) to ensure excess nitrogen. Two steady

states, C/P = 60 and C/P= 80 g g-1, were analysed and compared with the results ob-

tained with C/N = 10 g g-1 (see Table 3.2).

Excess nitrogen (C/N = 3.3 g g-1) induced an enhanced production of biomass of

8.3 g L-1 and an increased consumption of phosphorus causing P-limitation for C/P =

60 g g-1. PHA accumulation in dried biomass was similar to the one under N-limited

conditions. The endotoxic activity of PHA and biomass, however, was about 2.5

Biosynthesis of mcl-PHA under P-limitation 76

times higher under nitrogen in excess than under nitrogen limitation. Interestingly, the

content of phospholipids was reduced by more than 50% with excess nitrogen.

Excess carbon (30 mg L-1) and excess nitrogen (1’740 mg L-1) was measured in

the supernatant for C/P = 80 g g-1 and C/N = 3.3 g g-1 and hence this steady state

was only P-limited. The cell dry weight was reduced, the phospholipid content

strongly reduced and the PHA content was increased by about 50%. In comparison,

for a C/P = 80 g g-1 and a C/N = 3.3 g g-1, significant amounts of free nitrogen (48 mg

L-1) could be measured in the supernatant and hence this growth state was (C,P)- but

not N-limited. Concomitantly PHA accumulation was more efficient at a C/N = 10

g g-1. Significant differences were observed for the endotoxic activity of biomass and

PHA which was about 8- respectively 126-times smaller with C/N = 10 g g-1 com-

pared to the fermentation with excess nitrogen (C/N = 30 g g-1).

Concluding, increasing C/P from 60 to 80 g g-1 did raise the endotoxic activity

with C/N = 3.3 g g-1 to some extent. Thus, PHA accumulation was higher and the ac-

tivity of endotoxins was lower with a high carbon to nitrogen ratio. Hence, a low N/P

ratio (6 and 8 g g-1) was favourable over high N/P values (18 and 24 g g-1) for P-

limited conditions.

These effects can be explained by the reduced flexibility of cells with additional

nitrogen limitation, compared to a state with nitrogen in large excess, forcing the bac-

teria to efficiently use phosphorus and to accumulate mcl-PHA from carbon feed. For

instance the synthesis of proteins and membrane lipids lacking phosphorus (e.g. or-

nithine-based lipids) will be facilitated with excess nitrogen compared to N-limited

conditions. One could argue that both fermentations with C/P = 80 g g-1 were non-

nitrogen limited and should give similar results. But it seems that a large nitrogen ex-

cess has an inhibiting effect on the cell growth at this C/P ratio.

Some wall growth leading to biofilm formation was observed with P-limitation in-

dependent from N-limitation. It has been observed for P. putida in a laminar flow cell

reactor that biofilm accumulation rate was highest at C/P =100 g g-1 (Rochex and Le-

beault 2007). This confirms our observation that biofilm formation increased with a

higher C/P. Calcium crosslinking with extracellular polysaccharides is supposed to be

important for the matrix of biofilms and can be inhibited by increased precipitation of

calcium phosphate (Turakhia and Characklis 1989). But this is only possible with ex-

cess phosphate and hence biofilm formation seems to be an intrinsic problem of P-

limitation.

77 Chapter 3

Table 3.2: Comparison of different nutrient limitations with C/N = 10 and C/N = 3.3 g g-1.

Limited

nutrients

C/P C/N OD CDW CDW-

PHA

PHA

content

Phospholipid

content

Endotoxins

biomass

Endo-

toxins

PHA

[g g-1] [g g-1] [-] [g L-1] [g L-1] [w%] [w%] EU mg-1

biomass

EU g-1

PHA

C, N 60 10 25.6 7.3 6.5 10.7 4.4 4.2 x 105 828

C, P 60 3.3 26.6 8.3 7.3 11.5 2.0 10.9 x 105 2208

C, P 80 10 31.7 8.2 6.1 25.6 1.0 2.2 x 105 21

P 80 3.3 26.9 7.5 6.2 16.5 1.0 17.6 x 105 2661

OD: optical density measured at 450nm, CDW: cell dry weight, CDW-PHA: residual biomass.

3.4 Conclusions

P. putida GPo1 cultivated continuously under P-limitation combined with carbon and

nitrogen limitation drastically changed its cell composition. The amount of phospho-

lipids in dried biomass was decreased by about 90% under P-limitation compared to

a state without P-limitation. In parallel, the endotoxic activity of biomass was reduced

to 10% of its initial value. Furthermore, P-limitation combined with N-limitation con-

siderably increased the PHA accumulation. With combined (C,P)-limitation and ex-

cess nitrogen the PHA accumulation was significantly lower and the endotoxic activ-

ity was notably increased. This clearly demonstrated that combined (C,N,P)-limitation

leads to a better performance of the culture than (C,P)-limitation with nitrogen ex-

cess.

Consequently, by selecting appropriate fermentation conditions, the contamina-

tion of PHA with endotoxins can be significantly reduced. However, P-limitation can-

not be extended arbitrarily as the cell growth rate becomes restricted. Further, P-

limitation of P. putida GPo1 is prone to biofilm formation, which can limit the stability

and reproducibility of fermentation experiments and should be avoided.

Biosynthesis of mcl-PHA under P-limitation 78

Overall, cultivation under P-limitation could be a promising technique for fermen-

tations with Gram-negative bacteria to reduce the product contamination by endoto-

xins.

3.5 Acknowledgments

We gratefully acknowledge the financial support by the Empa R&D fund. We also

thank Ernst Pletscher and Dominik Noger for assisting with fermentations and for ad-

vices and Kathrin Grieder for ICP-OES measurements.

79

Chapter 4

Efficient recovery of low endotoxin medium-chain-

length poly([R]-3-hydroxyalkanoate) from bacterial

biomass

Patrick Furrer, Sven Panke, and Manfred Zinn. 2007. Journal of Microbiological

Methods 69, 206-213.

Efficient recovery of mcl-PHA 80

4.1 Introduction

New biomaterials, in particular degradable ones are needed for various medical

applications (Hubbell 1995; Ueda and Tabata 2003). Today’s biomaterials of the third

generation (Hench and Polak 2002) are designed to stimulate specific responses at

the molecular level and to combine the concepts of bioactive and resorbable

materials. Today’s polymeric and biodegradable systems used in medicine are

mainly based on poly(lactic acid) (PLA), on poly(glycolic acid) (PGA), and on their co-

polymers (Gomes and Reis 2004). Other biodegradable polymers have been

proposed, but could not enter the market yet, due to lacking FDA approval (Gomes

and Reis 2004).

One of the candidates is polyhydroxyalkanoate (PHA), a class of biodegradable

and biocompatible polyesters with many potential applications in the medical field,

such as heart valve scaffolds (Sodian et al. 2000a; Sodian et al. 2002), pulmonary

conduits (Stock et al. 2000), sutures, screws, bone plates, repair patches, stents,

bone marrow scaffolds and many others over the last years as recently reviewed by

Chen and Wu (Chen and Wu 2005). In particular the scientific and industrial research

on elastomeric PHA with medium-chain-length side-chains (mcl-PHA, containing 3-

hydroxyalkanoate monomers from C6-C14) has been intensified in the last years

(Chen et al. 2005; Sodian et al. 2002; Sodian et al. 2000b; Stock et al. 2000; Williams

et al. 1999; Zinn et al. 2001).

Mcl-PHA is synthesized by fluorescent pseudomonads belonging to the rRNA

homology group I (Huisman et al. 1989) and serves as an intracellular carbon and

energy storage compound. Its biosynthesis can be triggered by appropriate growth

conditions such as phosphorus and nitrogen limitation (Witholt and Kessler 1999).

Mcl-PHA is stored in distinct granules that are coated by amphiphilic phospholipids

and proteins (Pötter and Steinbüchel 2004).

Along with cell growth, all of the Gram-negative production strains produce

lipopolysaccharides (LPS) as an integral part of the outer membrane (Petsch and

Anspach 2000). LPS are pyrogenic and their concentration in a medical product is

strictly regulated. The US-American Food and Drug Administration has given clear

81 Chapter 4

regulations on the maximum endotoxin content of medical devices, namely 20 EU

per medical device (U.S. Department of health and human services 1997).

There are currently two well-established approaches for recovering PHA: solvent

extraction and aqueous digestion of non-PHA materials. The former one uses an

organic solvent to extract PHA from dried biomass, whereas the latter one involves

chemical agents or enzymes to lyse bacteria and thus liberating the PHA granules

forming a latex (Choi and Lee 1999; de Koning et al. 1997; de Koning and Witholt

1997; Ling et al. 1997; Taniguchi et al. 2003; Yu and Chen 2006). To purify and

depyrogenize PHA in latex form, oxidizing agents like hydrogen peroxide and sodium

hypochlorite as well as sodium hydroxide were applied (Lee et al. 1999; Sevastianov

et al. 2003; Williams et al. 1999). However, oxidizing agents may result in polymer

degradation as was reported for the treatment with sodium hypochlorite (Taniguchi et

al. 2003) and usually do require additional, expensive processes like

ultracentrifugation and microfiltration (de Koning and Witholt 1997; Marchessault et

al. 1995). In general, the PHA material obtained by this method is of low purity

(Kessler et al. 2001) and therefore not suitable for medical applications.

Various organic solvents can be used to extract mcl-PHA from dried biomass

(Chen et al. 2001; Kessler et al. 2001; Williams et al. 1999). The lyophilized biomass

is commonly extracted with a chlorinated solvent and the resulting PHA solution is

filtered (see Fig. 4.1A). In order to separate PHA from co-extracted impurities (e.g.

lipids and proteins), the polymer solution is usually mixed with a non-solvent like

methanol which induces precipitation of the polymer, leaving most contaminants in

solution. The extraction conditions and further processing steps strongly affect the

recovery and purity of the resulting PHA. Although polymer purities above 90% (w/w)

are easily obtained by a solvent extraction followed by a non-solvent precipitation,

the amount of endotoxins can still reach high levels (Sevastianov et al. 2003).

Typically values of more than 100 EU g-1 of polymer were found in a commercially

available PHA, namely poly(3-hydroxybutyrate) (Williams et al. 1999).

In this study, we investigated an alternative method for the purification of mcl-

PHA to obtain a polymer of high purity and low endotoxicity, namely by the heat

driven extraction with a solvent and the selective precipitation induced by

temperature reduction (see Fig. 4.1B). Special attention was paid to minimize the

contamination with endotoxins.

Efficient recovery of mcl-PHA 82

Fig. 4.1: Schematic of PHA recovery: A) extraction and non-solvent precipitation, B) extraction and temperature-controlled precipitation.

4.2 Materials and Methods

Biosynthesis. Polyhydroxyoctanoate (PHO) consisting of 87% (w/w)

3-hydroxyoctanoate and 13% (w/w) 3-hydroxyhexanoate was produced by

continuous cultivation of Pseudomonas putida GPo1 (Durner et al. 2000) in a 16 L

bioreactor (Bioengineering, Wald, Switzerland). The dilution rate was set to 0.1 h-1,

the ammonium content of the minimal medium was 800 mg N L-1 and the carbon feed

rate was adjusted to obtain a carbon to nitrogen ratio of 12 g g-1. The fermentation

broth was then concentrated with a continuous centrifuge (Cepa Z61H, Lahr,

Germany) at 17’500 x g and a flow rate of 50 L h-1. The paste was lyophilized for

48 h, ground, and the dry biomass was stored in a closed container at -25 °C until

extraction.

Analysis: Gel permeation chromatography (GPC). Molecular weights (number-

average (Mn) and weight-average (Mw)) were determined by gel permeation

chromatography (TDA 302, Viscotek, Waghausel-Kirrlach, Germany) equipped with a

viscometer, RI- and light-scattering detector. The system was calibrated by using 2

polystyrene standards (Polycal, Viscotek) with known molecular weights and intrinsic

viscosities (standard 1: Mw=115 kDa, Mn=112 kDa, IV=0.519, standard 2: Mw=250

kDa, Mn=100 kDa, IV=0.843). Forty mg of PHO-sample was dissolved in 10 mL of

THF over night, filtered through a 1.0 µm nylon filter, and aliquots of 100 µL were

83 Chapter 4

chromatographed with pure THF as solvent phase through three GPC columns

(Mixed bed, Viscotek) at 35 °C and a flow rate of 1 mL min-1.

Gas chromatography (GC). The purity of PHO and its composition with respect to

alkanoate monomers were determined by gas chromatography. About 50 mg of

polymer was transferred into a 20 mL Pyrex tube. Five mL of methylene chloride

containing 10.0 mg mL-1 of 2-ethyl-2-hydroxybutyric acid as internal standard and 5

mL of a 20/80 (v/v) mixture of HCl (37%) and n-propanol were added for the

propanolysis at 80 °C for 16 h. Subsequently, the sample was cooled on ice and 8

mL of demineralized water was added. The mixture was vigorously shaken, the

phases were separated, and the aqueous phase was removed. The organic phase

was dried over Na2SO4, neutralized by adding Na2CO3, and filtered through a 1.0 µm

nylon filter. One µL of this sample was injected into the GC (GC 8575 Mega 2, Fisons

Instruments, Rodano, Italy) onto a Supelcowax-10: 30 m x 0.32 mm column with a

film thickness of 0.5 µm (Supelco, Bellefonte, USA) at a split ratio 1:10 and an initial

temperature 120 °C, which was raised at 10° min-1 to 280 °C. Detection was

performed by FID detection.

Limulus amebocyte lysate (LAL)-test. The endotoxicity was measured using a

chromogenic LAL endpoint assay (QCL-1000, Cambrex, New Jersey, USA). About

300 mg of PHO were placed in a 15 mL pyrogen-free centrifuge tube (Corning, New

York, USA) and heated up to 80 °C for 24 h to obtain a smooth polymer surface.

Afterwards 6 mL of pyrogen-free water (Cambrex) was added. The tubes were

placed in a rotary shaker and incubated for 24 h at 37 °C under constant shaking at

180 rpm to extract the endotoxins from the polymer. The aqueous endotoxin-solution

was directly used for the LAL-test or diluted 1:10 or 1:100 if the concentration of

endotoxins was high. Each analysis was carried out at least twice.

Experimental setup. For the recovery of PHO, we applied the solvent extraction

method with precipitation by a non-solvent (standard extraction, Fig. 4.1A), and

temperature-controlled precipitation, which does not require a non-solvent (Fig.

4.1B).

Standard extraction. Five gram of dry biomass containing about 30% (w/w) PHO

were extracted with 75 mL of different organic solvents - methylene chloride

(CH2Cl2), ethyl acetate, acetone, n-hexane, tetrahydrofurane (THF) and 2-propanol of

purum grade or higher (Fluka, Buchs, Switzerland) - for 24 h at room temperature

Efficient recovery of mcl-PHA 84

under intense stirring. The biomass to solvent ratio of 1:15 (g mL-1) had been

identified as optimal in a preliminary test (data not shown) and was applied for all

extractions. The suspension was filtered through a regenerated cellulose membrane

filter (1.0 µm, Schleicher & Schuell Microscience GmbH, Dassel, Germany) with a

pressure filtration unit (200 mL stainless steel, Sartorius AG, Goettingen, Germany)

by applying an overpressure of typically 3 bars to obtain a clear PHO solution. The

polymer solution was either evaporated or concentrated by rotary evaporation to 1/5

of the initial volume and then precipitated at -20 °C by adding 10 times the volume of

methanol. Finally the polymer was dried under vacuum at 40 °C for 48 h. The

resulting polymers had molecular weights of about 200 kDa (Mw) and 100 kDa (Mn)

and hence a polydispersity of 2.0.

Temperature dependent solubility. About 1 g of pure PHO was dissolved in 100

mL organic solvent at 40 °C under intense stirring. The solution was cooled to 0 °C in

an ice bath under gentle stirring. After 4 hours the solution was decanted. The

precipitate was dried and the solvent was evaporated from the solution to determine

the ratio between precipitated and dissolved PHO.

Temperature-controlled extraction and precipitation . Five gram of dried biomass

was extracted with 75 mL of n-hexane at different temperatures under intense stirring

for 24 h. The suspension was filtered at the same temperature as the extraction took

place. In order to study the influence of temperature on the solubility of PHO in n-

hexane, the polymer solution was kept at a temperature between +20 °C and -10 °C

under gentle stirring for 24 h. The precipitated polymer was decanted and dried

under vacuum at 40 °C for 48 h.

Re-dissolution and precipitation. For further purification, dried PHO (extracted with

n-hexane at 50 °C and precipitated at 0 °C) was re-dissolved in 2-propanol (2%

(w/v)) at 45 °C under intense stirring for 8 h. PHO was precipitated by cooling the

solution to a temperature between 30 °C and 0 °C under gentle stirring for 24 h.

Finally the polymer was dried under vacuum at 40 °C for 48 h.

Calculation. The purity and the recoveries of PHO were calculated according to the

following equations:

85 Chapter 4

100Em

xOH-3m

EP ⋅=∑ [% w/w] [1]

100Pm

xOH-3m

PP ⋅=∑ [% w/w] [2]

100Bm

xOH-3m

PHOX ⋅=∑ [% w/w] [3]

where PE is the purity of the extract, PP the purity of the precipitate, XPHO the PHO

content in the dried biomass, mE is the mass of the extract, mP the mass of the

precipitate, mB the mass of the biomass, and m3-OH x are the masses of the individual

dehydrated 3-hydroxy acids (as present in the polymer) determined with GC as

described above.

100PHOBm

EPEm

ER ⋅⋅

⋅=

X [% w/w] [4]

100EPEmPPPm

PR ⋅⋅

⋅= [% w/w] [5]

The recovery is either calculated for the extract relative to the total mass of PHO

in the biomass (RE), or for the precipitate relative to the mass of extracted PHO (RP).

All measurements were carried out in duplicates. The bars in Fig. 4.2, Fig. 4.3

and Fig. 4.5 represent the uncertainty of measurement for the purity and the recovery

(±5% and ±6%, respectively) and the maximum standard deviation for the

endotoxicity (±10%). The uncertainty of measurement was calculated based on

additional experiments not described in this paper.

Efficient recovery of mcl-PHA 86

4.3 Results and discussion

Standard extraction and precipitation. Biomass was extracted according to the

conventional technique (Fig. 4.1A) to find solvents suitable to extract PHO selectively

with a good recovery. As shown in Table 4.1, methylene chloride was the most

efficient solvent at room temperature. 2-propanol and n-hexane were rather poor

PHO extracting solvents. However, ethyl acetate, tetrahydrofurane, and acetone had

recovery rates around 80% and provided PHO of purities > 80%. Especially ethyl

acetate and acetone extracts contained less than 10% impurities. Furthermore n-

hexane, methylene chloride, ethyl acetate, and acetone resulted in PHO of low

endotoxicity.

Table 4.1: Recovery (RE), purity (PE or PP) and endotoxicity of PHO extracted from dried biomass with different solvents.

Solvent evaporated Precipitated in MeOH

Solvent R E

[% w/w]

PE

[% w/w]

Endotoxicity

[EU g -1]

RE

[% w/w]

PP

[% w/w]

Endotoxicity

[EU g -1]

n-hexane 53 93 320 49 100 102

CH2Cl2 86 78 43 83 100 10

Ethyl

acetate

80 92 16 78 99 10

THF 80 84 589 77 100 10

Acetone 83 92 75 77 100 36

2-propanol 23 66 115 12 84 30

The subsequent precipitation of PHO with methanol drastically increased the

purity of the polymer. Purities close to 100% could be reached (Table 4.1). At the

same time, only small amounts of PHO, usually less than 5%, were lost.

87 Chapter 4

Temperature-controlled precipitation. To identify solvents capable of dissolving

PHO in a temperature dependent manner (Fig. 4.1B), PHO solutions of 1% (w/v) in

organic solvent were prepared. The concentration of 1% (w/v) was chosen because it

is typical for the concentration reached after biomass extraction with an organic

solvent at a ratio 1:15 (w/v). The polymer solutions were cooled down to 0 °C and

sol- and gel-phase were separated by decantation.

There were only two solvents with sufficiently different solubilities at 40 oC and

0 oC: 2-propanol and n-hexane (Table 4.2). Nearly all dissolved polymer could be

recovered from 2-propanol whereas in n-hexane about 26% of the polymer remained

in solution. The other solvents showed no polymer precipitation when the

temperature was adjusted to 0 °C, consequently they are not suitable for a

temperature induced precipitation of PHO. Eventually polymer precipitation would

have taken place at significantly lower temperatures but lower purities had to be

expected due to co-precipitation of impurities present in the solution.

Of the two solvents with a suitable temperature-dependent behavior, 2-propanol

is much less suitable for extracting PHO since it co-extracted a lot of cellular

compounds and thus resulted in PHA of low purity (Table 4.1). In contrast, n-hexane

selectively extracted PHO from biomass and was therefore suitable for a combined

extraction / temperature-controlled precipitation.

Table 4.2: Fractions of precipitated PHO after cooling the polymer solutions to 0 °C.

Solvent Precipitated polymer

[% w/w]

n-hexane 74

CH2Cl2 0

Ethyl acetate 0

THF 0

Acetone 0

2-propanol 97

Efficient recovery of mcl-PHA 88

Temperature optimization of PHO extraction. After having identified n-hexane as

the solvent with the most promising properties, we optimized the extraction step with

respect to temperature. PHO was extracted with n-hexane at temperatures between

20 °C and n-hexane’s boiling point at 69 °C. The recovery increased considerably

with rising extraction temperature. A plateau was reached at 50 °C (Fig. 4.20). The

purity of the resulting PHO slightly decreased with increasing temperature, but still

remained close to 90% at the boiling point. In contrast, the endotoxicity increased

dramatically with the extraction temperature from 10 to >1’000 EU g-1 polymer (Fig.

4.2). Hence the co-extraction of LPS from biomass is much more pronounced at

higher extraction temperatures. The molecular weight of PHO increased moderately

with increasing temperature, indicating that high molecular weight PHO dissolved

better above 20 °C than below (Table 4.3). Consequently, the extraction with n-

hexane was very efficient above 50 °C, but due to the increased contamination with

endotoxins, an extraction temperature below 50 °C was more reasonable.

In subsequent experiments, dry biomass was extracted with n-hexane at 50 °C

for 24 h. This temperature was chosen in order to clearly document a potential

endotoxin reduction effect of the precipitation.

0

25

50

75

100

10 20 30 40 50 60 70Extraction temperature [ °C]

Pur

ity a

nd re

cove

ry [%

w/w

]

1

10

100

1000

10000

End

otox

icity

[EU

g-1

pol

ymer

]

Recovery

PurityEndotoxicity

Fig. 4.2: Recovery, purity and endotoxicity for PHO extracted with n-hexane at different temperatures.

89 Chapter 4

Table 4.3: Weight averaged molecular weight (Mw), number averaged molecular weight (Mn) and polydispersity (D) of PHO extracted from biomass with n-hexane at different temperatures.

Extraction

temperature

[°C]

Mw

[kDa]

Mn

[kDa]

D

[-]

20 190 96 1.97

30 202 104 1.95

40 193 98 1.97

50 209 113 1.86

60 221 114 1.95

69 212 111 1.92

Temperature optimization of PHO precipitation. To investigate the optimum

precipitation temperature of PHO from n-hexane, the polymer solution was cooled

after filtration to temperatures between +20 and -10 °C. First precipitations were

observed at 20 °C. The amount of precipitated polymer increased further with

decreasing temperature and may have reached a maximum between 5 and 0 °C.

Surprisingly, precipitation became less efficient at a temperature below 0 °C as

shown in Fig. 4.3. Apparently, PHO formed a dispersion with n-hexane at

temperatures below 0 °C and precipitated very slowly. About 80% of the extracted

PHO could be precipitated at best. However, an improved recovery could be

obtained when the polymer solution was concentrated previously, e.g. by solvent

evaporation.

The purity of the resulting polymer was close to 100% for precipitation

temperatures down to 5 °C, and declined for lower temperatures (Fig. 4.3). This

indicates that non-polar contaminants co-precipitated significantly below 0 °C. The

endotoxicity increased at low temperatures (Fig. 4.3) in parallel with the observed

precipitation of other impurities.

Efficient recovery of mcl-PHA 90

0

25

50

75

100

-20 -10 0 10 20 30

Precipitation temperature [ °C]

Pur

ity a

nd re

cove

ry [%

w/w

]

1

10

100

1000

End

otox

icity

[EU

g-1

pol

ymer

]

RecoveryPurityEndotoxicity

Fig. 4.3: Recovery, purity and endotoxicity for PHO extracted with n-hexane at 50 °C, precipitated at different temperatures (20 °C: endotoxicity not determined).

Gel permeation chromatography (GPC) showed that particularly the high

molecular weight fractions of the polymer precipitated at higher temperatures and

that the low molecular weight fractions remained in solution as shown in Fig. 4.4.

This led to a decrease of the polydispersity (Mw/Mn) from 1.86 to 1.39 at 10 °C (Table

4.4). This is important since decreasing the amount of low molecular weight species

prevents a rapid and extensive release of mono- or oligomers in medical (implants)

and pharmaceutical (drug carriers) applications. For some polymers in medical

applications it has been reported that the release of low molecular weight

components led to a drastic decrease of the pH near the implant surface (Agrawal

and Athanasiou 1997) or often caused inflammations (Eckelt et al. 2004).

Since the goal was to produce low endotoxin containing PHA at a high yield, it

was concluded that the optimal precipitation temperature was 0 °C taking into

account that the polydispersity was somewhat larger (D = 1.56).

91 Chapter 4

Fig. 4.4: Normalized molecular weight distribution of PHO extracted with n-hexane at 50 °C, precipitated at different temperatures.

Table 4.4 : Weight averaged molecular weight (Mw), number averaged molecular weight (Mn) and polydispersity (D) of PHO extracted with n-hexane at 50 °C, precipitated at different temperatures (starting material: Mw=209 kDa, Mn=113 kDa, D=1.86).

Precipitation

temperature*

[°C]

Mw

[kDa]

Mn

[kDa]

D

[-]

10 363 262 1.39

5 233 157 1.48

0 228 146 1.56

-5 204 121 1.70

-10 217 125 1.73

*Molecular weight analysis could not be carried out for the sample at 20 °C because of

insufficient polymer recovery.

2-propanol re-dissolution and precipitation. In order to further improve the purity

of PHO above 97%, we investigated the re-dissolution of the purified polymer in

2-propanol followed by another temperature-controlled precipitation, because

2-propanol was very efficient in co-extracting cellular impurities (see above). Thus we

Efficient recovery of mcl-PHA 92

proposed that the sequential usage of n-hexane and 2-propanol for temperature-

controlled precipitation has a beneficial effect on the purity of PHA since impurities

may have a different affinity to these solvents. PHO obtained from n-hexane

extraction at 50 °C and precipitated at 0 °C was re-dissolved in 2-propanol at 45 °C

to a concentration of 20 g L-1. The solution was then cooled to temperatures between

30 and 0 °C. The PHO recovery increased with decreasing temperatures as shown in

Fig. 4.5. At 25 °C already 80% of the polymer had precipitated. At 0 °C nearly 100%

of the PHO was recovered. The purity was close to 100% for all temperatures. The

endotoxicity was lowered to less than 10 EU g-1 polymer at all temperatures, except

at 0 °C (Fig. 4.5). A temperature of 10 °C resulted in the lowest endotoxicity of PHO

with 2 EU g-1 polymer. GPC measurements again showed a fractionation effect

according to the molecular weight (Fig. 4.6 and Table 4.5). High molecular weight

fractions precipitated at elevated temperatures, whereby the low molecular weight

fractions remained in solution. The polydispersity was decreased from 1.50 to a

minimum of 1.26.

0

25

50

75

100

-5 0 5 10 15 20 25 30 35

Precipitation temperature [ °C]

Pur

ity a

nd re

cove

ry [%

w/w

]

1

10

100

End

otox

icity

[EU

g-1

pol

ymer

]

Recovery

Purity

Endotoxicity

Fig. 4.5: Recovery, purity and endotoxicity for PHO re-dissolved in 2-propanol at 40 °C and precipitated at different temperatures.

93 Chapter 4

Fig. 4.6: Normalized molecular weight distribution of PHO re-dissolved in 2-propanol and precipitated at different temperatures.

Table 4.5: Weight average molecular weight (Mw), number average molecular weight (Mn) and polydispersity (D) of PHO re-dissolved in 2-propanol and precipitated at different temperatures (starting material: Mw=228 kDa, Mn=146 kDa, D=1.56).

Precipitation

temperature

[°C]

Mw

[kDa]

Mn

[kDa]

D

[-]

30 321 255 1.26

25 260 199 1.31

20 244 184 1.33

15 234 159 1.47

10 231 156 1.48

5 239 159 1.50

0 227 156 1.46

Efficient recovery of mcl-PHA 94

4.4 Conclusions

Temperature-controlled extraction and precipitation can be effectively used to

increase the purity and reduce the endotoxin content of microbial PHO samples. For

this reason, it represents a valuable alternative to the classical approach where the

PHA-solution is mixed with an excess of non-solvent for precipitating mcl-PHA. Our

method yields in a polymer of high quality, is very simple and omits the

disadvantages of a non-solvent precipitation (e.g. separation and recycling of

solvents and non-solvents).

N-hexane was useful for a quantitative extraction of PHO at a temperature

around 50 °C. By decreasing the extraction temperature to 40 °C or below, the

recovery decreased moderately but the PHA contained much less endotoxins.

Precipitation induced by cooling led to PHO of high purity and low endotoxicity,

leaving most impurities in solution. However, a certain loss of polymer can not be

prevented except through a pre-concentration of the polymer solution. Especially the

low molecular weight fractions of PHO hardly precipitated.

The temperature dependent capacity to dissolve PHO made 2–propanol a

promising candidate for PHO extraction from biomass, but it dissolved other cell

components as well, which made it less suitable for the first extraction step.

However, its increased polarity relative to n-hexane led us to reason that it might be

an excellent solvent to re-dissolve PHO and selectively remove endotoxins. Indeed,

endotoxin concentrations of lower than 10 EU g-1 of polymer could be reached.

Additionally, the polymer polydispersity and the molecular weight can be controlled

by a temperature-controlled precipitation in order to achieve a more homogenous

degradation.

The solubility of various mcl-PHA in organic solvents can differ notably and

therefore the presented method might not be applicable to all mcl-PHAs without

adjustments. Recently, the method could be applied to a co-polymer produced from

octanoic and 10-undecenoic acid, where the relative composition was 50:50 w%

(Hartmann et al. 2006).

Concluding, we present a simple and very efficient method to produce PHO that

complies with the endotoxin requirements of the FDA for biomedical applications

95 Chapter 4

such as implants and drug delivery systems (U.S. Department of health and human

services 1997).

4.5 Acknowledgement

We gratefully acknowledge the financial support by Empa. We also thank Ernst

Pletscher and René Hartmann for their help with fermentations, Thomas Ramsauer

for assisting with polymer extractions and Manfred Schmid and Elisabeth Michel for

supporting GPC and GC analyses.

Efficient recovery of mcl-PHA 96

97

Chapter 5

Biocompatibility of untreated and chemically

modified medium-chain-length

poly(3-hydroxyalkanoates)

The results of this chapter were obtained in collaboration with:

Katharina Maniura and Philipp Spohn – cytotoxicity experiments

Sarah Cartmell – cell culturing on PHA coatings

Sarah Rathbone – cell culturing on PHA coatings

Patrick Furrer has carried out the following parts: PHA biosynthesis, PHA

extraction and purification, chemical derivatisation of PHA, solvent casting and

interpretation of cell culturing experiments.

Biocompatibility of mcl-PHA 98

5.1 Introduction

During the last two decades, significant advances have been made in the

development of biocompatible and biodegradable materials for biomedical

applications. The development of new biocompatible materials includes

considerations that go beyond the absence of toxic effects: the stimulation of specific

responses of tissue cells at the molecular level is the aim of so-called third

generation biomaterials (Hench and Polak 2002).

The chemical composition and surface structure of biomaterials are mainly

responsible for the interaction with biological systems. In the presence of serum, it is

thought that attachment of cells is mediated through adsorbed proteins and the

underlying surface chemistry controls protein adsorption. Through a protein layer, cell

adhesion and spreading can be prevented or promoted (Mrksich 2000). Physical and

chemical techniques can be used to modify the surface properties and hence to

control protein adsorption and cell adhesion. For instance plasma treatment (Qu et

al. 2005; Tezcaner et al. 2003), UV radiation (Shangguan et al. 2006; Viktor N.

Vasilets et al. 2004), collagen immobilisation (Tesema et al. 2004) and chemical

derivatisation with hydroxide (Rouxhet et al. 1998) have been used. Consequently,

the wettability or hydrophilicity of surfaces, as indicated by the water contact angle, is

of paramount importance for their interaction with cells. For osteoblasts, good

proliferation was obtained for fibronectin-coated surfaces with a water contact angle

of 50-90 ° (Kennedy et al. 2006).

Poly(3-hydroxyalkanoate) (PHA), a class of biodegradable polyesters, has the

potential to become an important biomaterial for medical applications (Williams et al.

1999; Zinn et al. 2001). Due to lower crystallinity and increased elasticity, mcl-PHA

offer physical properties that were quite different from scl-PHAs and give new

possibilities for medical applications such as scaffolds for tissue engineering and as

carrier for drug delivery. Functional groups in the side-chains of mcl-PHA allow to

regulate its properties. In addition, chemical modification of functionalized PHAs

enables further tailoring of the chemo-physical properties to suit particular

applications (Hazer and Steinbuchel 2007). For instance, the side-chain alkenes of

mcl-PHA were quantitatively transformed into hydroxyl groups with borane

99 Chapter 5

tetrahydrofurane (Renard et al. 2005). The hydrophilicity of the polymer increased

and the modified PHA was soluble in polar solvents like ethanol.

In vitro and in vivo studies have so far focused on short-chain-length PHA (scl-

PHA) with monomers of 3-5 carbon atoms and rather few studies were carried out

with medium-chain-length PHA (mcl-PHA) with monomers of 6-14 carbon atoms

(Chen and Wu 2005; Sodian et al. 2000a; Sodian et al. 2000b; Sodian et al. 2000c).

The combination of polyhedral oligomeric silsesquioxane (POSS) molecules and

polymers holds great promises for the future of biomedical devices, especially for

cardiovascular interfaces (Kannan et al. 2005). It has been shown that POSS

nanocomposites, unlike certain carbon nanotubes (Cui et al. 2005) are

cytocompatible and hence suitable for tissue engineering (Kannan et al. 2005;

Punshon et al. 2005). Further, It has been postulated that pendant nanocages

containing silicon atoms form foci with increased surface free energy thus promoting

the formation of an endothelial cell layer (endothelialization) and repelling coagulant

proteins (Kannan et al. 2005; Silver et al. 1999). The combination of POSS to mcl-

PHA would be of particular interest because the cell adhesion and proliferation on a

mcl-PHA matrix could be drastically enhanced.

In this work different mcl-PHAs with saturated and unsaturated side-chains were

produced and purified until the contamination with biological residues, in particular

with endotoxins was very low. The biocompatibility of these PHAs was studied.

Further, chemical modification was applied to hydrophobic mcl-PHAs with

unsaturated side-chains to increase their hydrophilicity and to improve cell adhesion

and proliferation. This was done by introducing carbonyl and carboxyl groups via

reaction with ozone or by covalently binding POSS to the unsaturated side-chains.

Biocompatibility of mcl-PHA 100

5.2 Materials and Methods

Biosynthesis. Mcl-PHAs were produced by continuous cultivation of Pseudomonas

putida GPo1 (ATCC 29347) as previously described (Ruth et al. 2007). Undecanoic,

10-undecenoic acid or a mixure of octanoic and 10-undecenoic acid were used as

carbon source for the production of poly(3-hydroxyundecanoate) (PHUA), poly(3-

hydroxy-10-undecenoate) (PHUE) and poly(3-hydroxyoctanoate-co-3-hydroxy-10-

undecenoate) (PHOUE). A 16 L laboratory bioreactor (Bioengineering, Wald,

Switzerland) with a working volume of 10 L was used. The dilution rate was set to

0.1 h-1, the ammonium content of the minimal medium was 800 mg N L-1 and the

carbon feed rate was adjusted to obtain a carbon to nitrogen ratio of 12 g g-1. The

fermentation broth was then concentrated with a continuous centrifuge (Cepa Z61H,

Lahr, Germany) at 17’500 x g and a flow rate of 50 L h-1. The paste was lyophilized

for 48 h, ground, and the dry biomass was stored in a closed container at -25 °C until

PHA extraction.

PHA extraction . PHUA, PHUE and PHOUE were extracted from dried biomass with

methylene chloride under stirring at room temperature for 24 h. After filtration, the

polymer solution was concentrated to one fifth of its original volume, cooled to 0 °C

and a tenfold excess of ethanol was added to induce precipitation. The polymer was

separated from the liquid phase by decantation and dried under vacuum. For

purification, the polymers were re-dissolved in methylene chloride, cooled to 0 °C and

a 10 fold excess of ethanol was added for precipitation. The solvent was decanted

and the PHA dried under vacuum.

Chemical derivatisation. PHOUE was linked with polyhedral oligomeric

silsesquioxane (POSS) via a free radical addition reaction as described elsewhere

(Hany et al. 2005). 1.9 g PHOUE, 14.0 g mercaptopropyl-isobutyl-POSS (POSS-SH)

and 0.27 g azobisisobutyronitrile (AIBN) were dissolved in 100 mL toluene and

heated to 75 °C for 18 h under nitrogen atmosphere. The solution was cooled to 0 °C

and 1 L of methanol was added for precipitation. To remove unreacted POSS-SH,

the polymer was dissolved in 200 mL methylene chloride and 1.5 L methanol was

added at room temperature for the precipitation of PHOUE-POSS. This procedure

was repeated four times for a complete removal of POSS-SH. The precipitate was

dried under vacuum.

101 Chapter 5

PHUE coatings in Petri dishes were treated with ozone generated with an

ozoniser (Certizon 200, Sander, Uetze-Eltze, Germany) supplied with pure oxygen.

The treatment time was set to 15 min for complete derivatisation of the double bonds

on the polymer surface based on contact angle measurements.

Solvent casting. Small (d=4 cm) and large (d=19.5 cm) glass Petri dishes were

coated with PHA for biocompatibility experiments after they had been depyrogenised

by heating to 200 °C for 6 h. To this end, the polymer was dissolved in

dichloromethane at a concentration of 2.5% w/v. The solution was filtrated (Nylon

filter, 0.45 µm) and cast into depyrogenised Petri dishes. The polymer solution was

slowly allowed to evaporate for 3 days at room temperature under an atmosphere

nearly saturated with dichloromethane to avoid the formation of bubbles. After that,

the polymer coatings were further dried under vacuum for 18 h at 40 °C. The

resulting polymer coating had a thickness of about 100 µm. Petri dishes with PHA

coatings were sterilised by x-ray previous to cell culturing.

PHA coating purification. For the cell culturing on PHA coatings, PHOUE-POSS

and PHUE treated with ozone (PHUE-O3) were purified to remove impurities of the

chemical derivatisation from the surface: PHOUE-POSS coatings were extracted with

3 mL of pure water for 3 h and with 3 mL of ethanol for 3 h and dried under vacuum.

PHUE-O3 was only dried under vacuum for 18 h to remove volatile impurities.

Endotoxin measurement. The endotoxicity was measured using a chromogenic

limulus amebocyte lysate (LAL) endpoint assay (QCL-1000, Cambrex, New Jersey,

USA). About 100 mg PHA was dissolved in 3 mL methylene chloride and solvent cast

into a depyrogenised glass Petri dish (d=4 cm). The solvent was slowly evaporated

under a nearly saturated atmosphere as described above the polymer coatings were

further dried under vacuum at 40 °C for 18 h. Three mL of pyrogen-free water

(Cambrex) was added to the Petri dish and incubated at 37 °C for 24 h to extract the

endotoxins from the polymer. The aqueous endotoxin-solution was directly used for

the LAL-test. Each analysis was carried out at least twice.

Water contact angle. The water contact angle was determined using a Krüss G10

contact angle measuring system (Hamburg, Germany) and pure water (Chromasolv,

Sigma-Aldrich). Three slides were coated for each polymer and two measurements

with droplets of 30 µL were made on each slide.

Biocompatibility of mcl-PHA 102

Gel permeation chromatography (GPC). Molecular weights (number-average Mn

and weight-average Mw) were determined by gel permeation chromatography (TDA

302, Viscotek, Waghausel-Kirrlach, Germany) equipped with a viscometer, RI- and

light-scattering detector. The system was calibrated by using 2 polystyrene standards

(Polycal, Viscotek) with known molecular weights and known intrinsic viscosities

(standard 1: Mw = 115 kDa, Mn = 112 kDa, IV=0.519, standard 2: Mw = 250 kDa, Mn =

100 kDa, IV=0.843). Then, 40 mg of PHA was dissolved in 10 mL of THF over night,

filtered through a 1 µm nylon filter, and aliquots of 100 µL were chromatographed

with pure THF as solvent phase through three GPC columns (Mixed bed, Viscotek) at

35 °C and a flow rate of 1 mL min-1.

Gas chromatography. Quantitative analysis of PHA has been performed by

methanolysis catalyzed by BF3 as described in chapter 2.

Cytotoxicity according to ISO10993-5

To test if a material releases biologically active substances to an aqueous

environment, the material was incubated for 3 days in water. Fibroblasts of type 3T3

were cultured with these extracts on 96 well plates. After 5 days the cell cultures

were analysed to determine the influence of the extract.

Extraction of water soluble PHA contaminations. The polymer coatings prepared

by solvent casting in Petri dishes (d=19.5 cm) had a thickness of about 10 µm and a

surface of 300 cm2. An aliquot of 25 mL of water containing 1% penicillin-

streptomycin-neomycin (PSN) antibiotics (Invitrogen, Basel, Switzerland) was added

to the Petri dish to extract substances from the PHA coating and incubated for 72 h

at 37 °C. The pH was initially adjusted to 7.4 with NaHCO3. The aqueous extract was

diluted with water and supplemented with 800 µL of Dulbecco’s minimal essential

medium (DMEM, Invitrogen), 300 µL NaHCO3 7.5%, 400 µL of foetal bovine serum

(FBS, Invitrogen) and 80 µL of PSN antibiotics to obtain solutions containing 0, 10,

20, 50 and 70 % of the aqueous extract.

Cell culturing. Each well was incubated with 200 µL of cell suspension containing

about 7500 3T3 cells per mL for 24 h. The medium was gently removed and 200 µL

of the diluted extract was added. After 5 days of incubation at 37 °C and a gas

atmosphere containing 5% CO2, DNA from permeabilised cells was measured with

103 Chapter 5

the Hoechst 33258 DNA stain (Invitrogen) and total protein was determined with the

bicinchonin acid protein assay kit (Pierce, Nr. 23225, Rockford Illinois) according to

the manufacturers protocols. To measure the metabolic activity, the mitochondrial

activity was determined by the MTT assay (Mosmann 1983) and the activity of

lysosymes was measured with the neutral red assay (Borenfreund et al. 1985).

Cell culturing on PHA coatings

Cell culture. Mouse connective tissue fibroblasts were bought (cell line L929 –

ECACC, 85011425, lot 04H008, Sigma, Poole, UK) and cultured in 88% DMEM high

glucose (Biosera L0106, East Sussex, UK), supplemented with 10% FCS (504002,

batch S1900-Biosera), 1% antibiotics (streptomycin and penicillin L00100, Biosera),

1% L-glutaraldehyde (G7513, Sigma-Aldrich), and incubated at 37 °C with a gas

atmosphere containing 5% CO2 in T175 culture flasks (Starstedt T175 vented red cap

83.1813.002, Leicester, UK). The cells were passaged 2 times and seeded at a

density of 100’000 cells/well into the small PHA-coated Petri dishes and into the wells

of Costar polystyrene (PS) 6-well plates (Corning, No. 3516, Massachusetts, USA)

and incubated to the required time point (7 days and 14 days). The medium was

changed every 4 days throughout the incubation period.

The cell viability was measured with the live/dead assay kit (Invitrogen,

L3224). To quantitate double-stranded DNA in solution, the PicoGreen® assay was

used (Invitrogen, P7589). Cells adhered to PHA and PS and cells in suspension were

measured separately. The manufacturers’ protocols were used for these two assays.

The hydroxyproline content was considered as an indicator for the collagen

production. It was determined as described elsewhere (Kivirikko et al. 1967).

Probability values of less than 0.05 (P<0.05) were considered to be significantly

different (2-tailed student T-test, n=5).

Biocompatibility of mcl-PHA 104

5.3 Results and discussion

5.3.1 Physical and chemical properties of PHAs

The purity of all PHAs was >95 w% according to GC measurements (Table 5.1). The

endotoxic activity of the polymer coatings was decreased below the detection limit

and was below 1 endotoxic unit per gram (EU g-1). The molecular weights were

similar for PHUA, PHUE, PHOUE and PHUE-O3 (Mw ≅ 200 kDa) and only for

PHOUE-POSS clearly increased, which was due to the linked POSS groups. The

water contact angle of untreated PHAs (PHUA, PHUE and PHOUE) was between

104° and 109°, whereas the POSS-derivatised PHOUE and the ozone-treated PHUE

showed lower values of 95° and 73° respectively. The monomeric composition of all

PHAs prior to chemical derivatisation is given in Table 5.2.

Table 5.1: Extraction, purification and derivatisation techniques applied for the production of different mcl-PHAs and their physical properties.

PHAs Extraction Purification Derivati-

sation

Contact

angle

Mw Mn Purity Endotoxicity

solvent-

non solvent

solvent-

non solvent

[°] [kDa] [kDa] [w%] [EU g-1]

PHUA CH2Cl2-EtOH CH2Cl2-EtOH 2x - 109 168 91 >95 <1

PHUE CH2Cl2-EtOH CH2Cl2-EtOH 2x - 104 231 115 >95 <1

PHOUE CH2Cl2-EtOH CH2Cl2-EtOH 2x - 108 255 151 >95 <1

PHOUE-

POSS

CH2Cl2-EtOH CH2Cl2-EtOH 2x,

CH2Cl2-EtOH2 4x

POSS-

SH

95 910 420 >95 <1

PHUE-

O3

CH2Cl2-EtOH CH2Cl2-EtOH 2x Ozone 73 231 115 >95 <1

2 Purification after chemical derivatisation

105 Chapter 5

Table 5.2 : Monomeric composition of mcl-PHAs determined by GC after transesterification.

PHAs Monomeric composition [mole%]

C5:0 a C6:0 C7:0 C8:0 C9:0 C11:0 C7:6 C9:8 C11:10

PHUA 2.8 33.8 49.0 14.4

PHUE 24.8 62.9 12.3

PHOUE 5.1 32.5 11.7 38.1 12.6

PHOUE-POSS 5.1 32.5 11.7 38.1 12.6

PHUE-O3 24.8 62.9 12.3

a) The notation Cn:x was used to denote the 3-hydroxy acids, where n is the number of

carbon atoms and x indicates the position of the double bond (0: no double bond present).

5.3.2 Cytotoxicity according to the eluate test

A critical point for the application of PHA (and other polymers) in the medical field is

the effect of water soluble substances contained in the polymer. To proof that no

toxic substances can diffuse from the polymer matrix to an aqueous environment the

cytotoxicity test according to ISO 10993-5 is commonly applied. Accordingly, PHA

coatings were extracted with water for 3 days and 3T3 fibroblasts were incubated

with diluted extracts for 5 days and finally the proliferation was determined. For pure

PHAs without chemical derivatisation a toxic effect in the aqueous eluate could

hardly be seen (Fig. 5.1A - C) in contrast to chemically derivatised PHAs where cell

proliferation was strongly reduced Fig. 5.1D - E).

Biocompatibility of mcl-PHA 106

0

50

100

0

50

100

0

50

100

0

50

100

0 20 40 60 800

50

100

Extract concentration [% v/v]

Rel

ativ

e co

ncen

trat

ion

[%]

Rel

ativ

e co

ncen

trat

ion

[%]

Rel

ativ

e co

ncen

trat

ion

[%]

Rel

ativ

e co

ncen

trat

ion

[%]

Rel

ativ

e co

ncen

trat

ion

[%]

A

B

C

D

E

Fig. 5.1: Cytotoxic activity of aqueous extracts from PHUA (A), PHUE (B), PHOUE (C), PHOUE-POSS (D) and PHUE-O3 (E). The toxic activity was determined with 3T3-fibroblasts by � DNA, � MTT, � protein and � neutral red assays.

107 Chapter 5

For PHUA, and PHUE no toxic effect could be seen with the DNA, MTT, protein

and neutral red assay (Fig. 5.1A - B). Their values varied between 90 and 115%

relative to the control, but there was no increase in the effect with increasing

concentration. Hence, these polymers appear to be free of biologically active

impurities that could be solubilised in an aqueous environment. For PHOUE a

cytotoxic effect was measured at the highest extract concentration where the values

for DNA, MTT and protein were reduced to about 80% of the control (Fig. 5.1C).

Apparently, PHOUE contained small amounts of biologically active impurities.

For PHOUE-POSS the cell proliferation and activity was decreased after 5 days

to about 60% (Fig. 5.1D). This toxic effect was attributed to side-products formed

during chemical derivatisation. Although the derivatised PHA was thoroughly purified

by dissolution-precipitation until no impurities of low molecular weight could be

detected by GPC, high toxicity was still measured.

Even more drastically the relative values of DNA, MTT, protein and neutral red

decreased for PHUE-O3. They reached around 20% at the highest extract

concentration (Fig. 5.1E). It has been reported that formaldehyde and small amounts

of formic acid are generated during oxidation of alkenes with ozone (Dubowski et al.

2004; Eliason et al. 2003). Adsorption of these products on the polymer surface, or

trapping in the polymer combined with slow diffusion can explain the toxic effect we

measured. Further purification of PHOUE-POSS and PHUE-O3 was required after

chemical derivatisation and was applied before cell culturing on PHA coatings.

5.3.3 Cell culturing on PHA coatings

Mouse connective tissue fibroblasts were seeded directly on the PHA coatings to

analyse their interaction with the polymer surface. Polystyrene (optimised for cell

attachment) was used as control material.

Unmodified PHAs: PHUA, PHUE and PHOUE

Proliferation had occurred during the 7 days according to the Picogreen assay. About

90%, 230% and 60% more cells were found on the PHUA, PHUE and PHOUE

coating, respectively in comparison to the control (Fig. 5.2). Most cells were attached

to the mcl-PHA surface and only few cells (~ 2%) were suspended in the medium.

Biocompatibility of mcl-PHA 108

Consequently, the loss of cells during medium change was negligible. This

observation was similar for all mcl-PHA coatings.

0

600000

1200000

Control PHUA PHUA coating Control PHUE PHUE coating Control PHOUE PHOUE coating

Cel

l num

ber

per

wel

l

Fig. 5.2: Cell numbers of adhered cells on PHUA, PHUE and PHOUE coatings and corresponding controls (control PHUA, control PHUE and control PHOUE) after 7 days, determined with the Picogreen assay.

The results of the live/dead assay for PHUA and PHUE showed that very few

viable cells were present on these PHA coatings in comparison to the control (Fig.

5.3a-c). This is in contrast to the measurements made with the Picogreen assay. The

cell film on the PHA coating was very confluent before the staining procedure.

However, before the live/dead staining was carried out, the cell density was not

evenly distributed on the PHUE coating but found in patches. The surface properties

of the PHA coating may have prevented the cells from adhering strongly, where they

became dislodged and washed away during washing steps of the live/dead assay.

Still, cell adhesion was strong enough to prevent a significant loss of cells during

routine medium change as shown with the Picogreen assay.

Contrarily, the live/dead assay for PHOUE showed that this coating had a high

density of viable cells, comparable to the cell density of the control (Fig. 5.3d and e).

This indicates that the surface properties were suitable for providing strong cell

attachment.

109 Chapter 5

Fig. 5.3: Fluorescence microscope images of live/dead-stained mouse connective tissue fibroblasts on PHUA after 7 days. (a) PHUA coating, (b) PS control for PHUA and PHUE, (c) PHUE coating, (d) PHOUE coating and (e) PS control for PHOUE. Viable cells were stained green.

The results of the hydroxyproline assay, which is used as an indicator for the

collagen production, showed that there was a higher concentration of hydroxyproline

in the PS control than in the coatings of PHUA, PHUE and PHOUE as exemplified for

PHUA in Fig. 5.4. The hydroxyproline concentration was reduced to 35% for PHUA

and PHUE and to 50% for PHOUE compared to the reference. However, for the

PHUE coating there was a considerable error in the hydroxyproline concentration.

This assay specifically measures the amount of hydroxyproline produced by

a b

c

d e

Biocompatibility of mcl-PHA 110

functional cells in the form of collagen. Hydroxyproline makes up for up to 13% of

collagen. However, the cells on PHA coatings seemed to be significantly less

functional than the control cells. Possibly, degradation products of PHA had effected

gene expression of certain extracellular matrix genes, such as switching off the gene

that produces hydroxyproline.

The results of these assays suggest that the PHUA, PHUE and PHOUE coatings

had no toxic effect and were supporting some cell growth with limited collagen

production. The surface properties of PHUA and PHUE may not suitable to allow

strong cell adherence. Only for PHOUE good cell attachment was observed. The

reason for improved cell adhesion was not clear. It could not be attributed to the

unsaturated side-chains because the cell adhesion was weak for PHUE, which

contained more unsaturated side-chains. Possibly the surface of PHOUE had an

increased roughness, thereby easing cell adhesion.

0

5

10

15

Control PHUA PHUA coating Control PHOUE-POSS PHOUE-POSS coating

Hyd

roxy

prol

ine

conc

entr

atio

n (

µg m

L-1

)

Fig. 5.4: Hydroxyproline assay for cells on PHUA and PHOUE-POSS coatings and corresponding PS controls (control PHUA and control PHOUE-POSS) after 14 days. P = 9.45 x10-10 and 0.029, respectively.

Chemically modified mcl-PHAs: PHOUE-POSS and PHOUE- O3

Similarly to PHOUE, a high cell number was observed with the Picogreen assay for

the PHOUE-POSS coating. In contrast to this, the cell number on the PHOUE-O3

coating was nearly identical to the one of the control and hence smaller compared to

all other PHA coatings.

111 Chapter 5

For PHOUE-POSS many viable cells were observed with the live/dead assay

(Fig. 5.5a). The density was comparable to the control. Hence, the cell adhesion was

strong, similar to that of PHOUE. In analogy, the live/dead assay showed a high cell

density for PHUE-O3 (Fig. 5.5c). Consequently, strong cell adhesion was obtained by

treating PHUE with ozone.

Fig. 5.5: Fluorescence microscope images of live/dead-stained mouse connective tissue fibroblasts on PHOUE-POSS (a), PHUE-O3 (c) and corresponding PS controls (b and d) after 7 days.

The results of the hydroxyproline assay for PHOUE-POSS showed that the

control had again a significantly higher hydroxyproline concentration than the coating

(Fig. 5.4). Hence, the functionality of fibroblasts on untreated and modified PHA

coatings seemed to be reduced. Surprisingly, the POSS derivatisation of PHOUE did

not improve cell adhesion, proliferation and collagen production considerably. For

PHUE-O3, the results of the hydroxyproline assay showed a reduced hydroxyproline

concentration, but they were not very significant because the error of the control was

big (data not shown).

a b

c d

Biocompatibility of mcl-PHA 112

The ozone derivatisation drastically decreased the water contact angle of the

surface and increased the cell adhesion compared to the untreated PHUE. In parallel

to this the cell proliferation was clearly decreased. The PHUE-O3 coating showed

similar properties to PS optimized for cell culturing, if the results from the

hydroxyproline were not considered.

According to our results, the PHAs can be divided into two groups according to

their suitability for cell attachment, proliferation and function:

PHOUE, PHOUE-POSS and PHUE-O3: These biomaterials provided good cell

attachment regarding the live/dead staining. The cell proliferation was equal or better

on these coatings than on PS. The collagen production was reduced with these

PHAs.

PHUA and PHUE: Neither coating provided very good cell attachment (live/dead

stain) and most of the cells became dislodged during the staining procedure. Both

coatings supported proliferation resulting in cell numbers higher than on the PS

control. The cells produced lower amounts of collagen on both coatings in

comparison to controls.

5.4 Conclusions

Cytotoxicity experiments as well as cell-culturing experiments on PHA coatings

showed that biosynthesised mcl-PHA with different chemical compositions are suited

for medical applications, when an effective purification procedure such as repeated

dissolution-precipitation was applied. Cell adhesion was limited on the non-polar

surfaces of mcl-PHA and hence a derivatisation of the PHA surface was required to

improve its hydrophilicity and to optimise the cell attachment. For long term

applications the type of chemical modification may be of crucial importance. Surface-

modified PHAs like PHUA-O3 could loose the polar groups during degradation and

consequently the surface would become more hydrophobic. Homogeneous PHAs on

the other hand will not change their surface polarity drastically during degradation.

Possibly, the cleavage of ester bonds during degradation will increase the surface

polarity through the formation of additional carboxyl and hydroxyl groups. Hence,

long time studies are needed to elucidate these questions. For applications where

113 Chapter 5

cell attachment has to be avoided, mcl-PHA with saturated side-groups could be well

suited.

Our measurements indicated that the collagen production of fibroblasts was

reduced on all PHA coatings, compared to the cells on PS. Hence the functionality of

fibroblasts on PHA coatings appeared to be reduced. The reason for this observation

was indistinct and further measurements with another assay are needed to confirm

this result. Furthermore the collagen type should be determined to identify eventual

changes.

Biocompatibility of mcl-PHA 114

115

Chapter 6

General conclusions

General conclusions 116

The use of polymeric materials for the administration of pharmaceutical and as bio-

medical devices has increased dramatically over the last ten years (Shikanov et al.

2005). Important biomedical applications of biodegradable polymers are in the areas

of drug delivery systems and in the form of implants and devices for fracture repairs,

surgical dressings, artificial heart valves, vascular grafts, and organ regeneration. It

has been concluded that PHA has the potential to become an important compound

for biomedical applications (Valappil et al. 2007; Williams et al. 1999; Zinn et al.

2001) and the results of this thesis further confirm this conclusion.

The aim of this work was to adapt the properties of mcl-PHA for medical applica-

tions by optimising its fermentative production and downstream processing (chapter

3 and 4). PHO was used as a model for the class of mcl-PHA. To analyze and opti-

mize the production of mcl-PHA, a new method for the quantitative analysis by gas

chromatography (GC) was developed (chapter 2). Finally, the biocompatibility of mcl-

PHA films was investigated in vitro using mouse fibroblasts and the polymers were

chemically modified to enhance their properties (chapter 5).

Quantitative analysis of PHA

Nuclear magnetic resonance spectroscopy (NMR) is a common tool to analyse and

quantify bacterial PHA. But NMR analyses are restricted to pure PHA and are very

time-consuming, whereas GC measurements are cheap and need less time. Estab-

lished methods for the transesterification of PHA in biomass with protic acids were

shown to be inadequate for the quantitative analysis of mcl-PHA due to an excessive

formation of side-products. This problem could be solved by using the Lewis-acid BF3

in water-free methanol. However, the development of a fast method that can be used

to derivatise and analyse PHA-containing biomass in a short time (e.g. 1 h) remains

a challenge. For PHB-containing biomass the derivatisation time was drastically de-

creased to about 4 minutes by using microwave heating (Betancourt et al. 2007).

Presumably, microwave heating can likewise be used for the accelerated transesteri-

fication of mcl-PHA. Alternatively, other derivatisation reactions like the reductive de-

polymerisation by LiAlH4 could be applied to mcl-PHA. Ideally, the PHA concentration

in a fermenter is measured continuously in situ, but a reliable method has not been

developed yet.

117 Chapter 6

PHA biosynthesis – possibilities to reduce the endo toxic contamination

P-limitation was successfully applied in this work to the fermentative production of

mcl-PHA to reduce the endotoxic contamination. The confirmation of the structural

change of lipid A was unsuccessful and it would be interesting to investigate it. How-

ever, P-limitation could not be extended arbitrarily in the chemostat as the cell growth

rate became restricted. Possibly, P-limitation would be more efficient in a two-stage

fed batch or two-stage continuous fermentation with P-limitation in the second stage.

In this case the endotoxic activity would only decrease if the structure of Lipid A is

changed during P-limitation. Recent experiments with Francisella novicida showed

that certain microorganisms own specific phosphatases for lipid A (Wang et al. 2004;

Wang et al. 2006). If similar phosphatases were also present in P. putida, the phos-

phorylation of lipid A could be modified during the adaptation of cells to altered cultur-

ing conditions. Otherwise, the genes coding for these phosphatases (LpxE and LpxF)

could be overexpressed in PHA producing bacteria like P. putida to detoxify the lipid

A, independently from culturing conditions.

Another alternative to avoid the contamination with LPS is the production of PHA

with Gram-positive bacteria. PHAs were produced with the genera Bacillus and

Streptomyces (Valappil et al. 2007). However, the ability to synthesize mcl-PHA with

Gram-positive bacteria appears to be rather limited (Valappil et al. 2007).

PHA recovery

PHA extraction with organic solvents and precipitation with non-solvents results in a

polymer of high quality, but this standard method is rather complicated and produces

large amounts of mixtures of solvent and non-solvent. Temperature controlled extrac-

tion and precipitation (TCP) was found to be a good alternative to the classical ap-

proach. It could be used to extract and precipitate PHO and co-polymers of PHO effi-

ciently. The recovery of PHA is basically complicated by the fact that PHA is stored

intracellularly. It would be simplified a lot if the PHA granules were secreted to the

medium. The cells could be separated from the supernatant by centrifugation and the

PHA granules would be recovered with the supernatant in high purity. However, the

transport of huge PHA granules through the membrane of intact cells appears to be

unrealistic. Recently, the production of extracellular PHA with Alcanivorax borkumen-

sis SK2 has been claimed (Sabirova et al. 2006), but it was not shown that extracel-

General conclusions 118

lular PHA did not originate from lysed cells. Hence, serious doubts remain and further

research is needed.

Towards biomedical applications

Cytotoxicity experiments with aqueous extracts from mcl-PHA as well as cell-

culturing experiments on films of mcl-PHA confirmed that these polymers were bio-

compatible when mcl-PHA of high purity was used. Cell adhesion could be improved

by chemical treatments of the non-polar PHA surface. However, the functionality of

mouse connective tissue fibroblasts was limited. The reason for this observation re-

mained obscure and further investigation is needed.

An important but restraining aspect is the processing of mcl-PHA. For most mcl-

PHA processing is difficult due to their elastomeric and sticky properties. Hence, ap-

propriate processing techniques have to be developed for a successful commercial

application. Again, chemical derivatisation may help to increase the crystallinity of

PHA, thereby facilitating its processing.

The results obtained in this doctoral thesis show, that optimized fermentation and

downstream processing generate mcl-PHA of high quality with very low endotoxic

contamination that can be used in the biomedical field. The broad range of different

mcl-PHAs combined with chemical modification allows adjusting the chemical and

physical properties to the target application. Thus, a basis for successful biomedical

applications of mcl-PHA was established with this work.

References 119

References

Agrawal, C. M., Athanasiou K. A. 1997. Technique to control pH in vicinity of biodegrading

PLA-PGA implants. J. Biomed. Mater. Res. 38: 105-114.

Alvarez, H. M., Kalscheuer R., Steinbüchel A. 1997. Accumulation of storage lipids in

species of Rhodococcus and Nocardia and effect of inhibitors and polyethylene glycol. Fett-

Lipid 99: 239-246.

Anderson, A. J., Dawes E. A. 1990. Occurence, metabolism, metabolic role, and industrial

uses of bacterial polyhydroxyalkanoates. Microbiol. Rev. 54: 450-472.

Annuar, M. S. M., Tan I. K. P., Ibrahim S., Ramacha ndran K. B. 2007. Production of

medium-chain-length poly(3-hydroxyalkanoates) from crude fatty acids mixture by

Pseudomonas putida. Food Bioprod. Process. 85: 104-119.

Ashby, R. D., Foglia T. A., Solaiman D. K. Y., Liu C. K., Nunez A., Eggink G. 2000.

Viscoelastic properties of linseed oil-based medium chain length poly(hydroxyalkanoate)

films: effects of epoxidation and curing. Int. J. Biol. Macromol. 27: 355-361.

Bachmann, B. M., Seebach D. 1999. Investigation of the enzymatic cleavage of

diastereomeric oligo(3-hydroxybutanoates) containing two to eight HB units. A model for the

stereoselectivity of PHB depolymerase from Alcaligenes faecalis T-1. Macromolecules 32:

1777-1784.

Baptist, J. N. W. R. Grace & Co., 1962a. Process for preparing PHB. US patent 3036959.

Baptist, J. N. W:R. Grace & co., 1962b. Process for preparing PHB. US patent 3044942.

Barham, P. J. Imperial Chemical Industries Ltd., UK, 1982. Extraction of poly(ß-

hydroxybutyric acid). EP patent 58480.

Bax, A., Marion D. 1988. Improved resolution and sensitivity In H-1-detected heteronuclear

multiple bond correlation spectroscopy. J. Magn. Reson. 78: 186-191.

Bear, M. M., Mallarde D., Langlois V., Randriamahef a S., Bouvet O., Guérin P. 1999.

Natural and artificial functionalized biopolyesters. II. Medium-chain length poly-

hydroxyoctanoates from Pseudomonas strains. J. Environ. Polym. Degrad. 7: 179-184.

Berger, E., Ramsay B. A., Ramsay J. A., Chaverie C. , Braunegg G. 1989. PHB recovery

by hypochlorite digestion of non-PHB biomass. Biotechnol. Tech. 3: 227-232.

References 120

Betancourt, A., Yezza A., Halasz A., Van Tra H., Ha wari J. 2007. Rapid microwave

assisted esterification method for the analysis of poly-3-hydroxybutyrate in Alcaligenes latus

by gas chromatography. J. Chromatogr. A 1154: 473-476.

Bordoloi, M., Borah B., Thakur P. S., Nigam J. N.; Council of scientific & industrial

research, India. 2003. Process for the isolation of polyhydroxybutyrate from Bacillus

mycoides RLJ B-017. US patent 2003027293.

Borenfreund, E. and Puerner, J. A. 1985. Toxicity determined in vitro by morphological

alterations and neutral red absorption. Toxicol. Lett. 24, 119-124.

Bos, M. P., Tommassen J. 2005. Viability of a capsule- and lipopolysaccharide-deficient

mutant of Neisseria meningitidis. Infect. Immun. 73: 6194-6197.

Boynton, Z. L., Koon J. J., Brennan E. M., Clouart J. D., Horowitz D. M., Gerngross T.

U., Huisman G. W. 1999. Reduction of cell lysate viscosity during processing of poly(3-

hydroxyalkanoates) by chromosomal integration of the staphylococcal nuclease gene in

Pseudomonas putida. Appl. Environ. Microb. 65: 1524-1529.

Braunegg, G., Lefebvre G., Genser K. F. 1998. Polyhydroxyalkanoates, biopolyesters from

renewable resources: Physiological and engineering aspects. J. Biotechnol. 65: 127-161.

Braunegg, G., Sonnleitner B., Lafferty R. M. 1978. A rapid gas chromatographic method

for the determination of poly-3-hydroxybutyric acid in microbial biomass. Eur. J. Appl.

Microbiol. Biotechnol. 6: 29-37.

Byrom, D. 1987. Polymer synthesis by microorganisms: technology and economies. Trends

Biotechnol. 5: 246-250.

Caballero, K. P., Karel S. F., Register R. A. 1995. Biosynthesis and characterization of

hydroxybutyrate-hydroxycaproate copolymers. Int. J. Biol. Macromol. 17: 86-92.

Castor, T. P., Hong G. T.; Aphios Corporation, USA, 2003. Method of fractionation of

biologically derived materials using critical fluids. US patent 6569640.

Chen, G. Q., Konig K. H., Lafferty R. M. 1991. Occurrence of poly-D(-)-3-

hydroxyalkanoates in the genus Bacillus. FEMS Microbiol. Lett. 84: 173-176.

Chen, G. Q., Wu Q. 2005. The application of polyhydroxyalkanoates as tissue engineering

materials. Biomaterials 26: 6565-6578.

Chen, G. Q., Wu Q., Wang Y. W., Zheng Z. 2005. Application of microbial polyesters-

polyhydroxyalkanoates as tissue engineering materials. Asbm6: Advanced Biomaterials Vi. p

437-440.

121 References

Chen, G. Q., Zhang G., Park S. J., Lee S. Y. 2001. Industrial scale production of poly(3-

hydroxybutyrate-co-3-hydroxyhexanoate). Appl. Microbiol. Biotechnol. 57: 50-55.

Chen, P. S., Toribara T. Y., Warner H. 1956. Microdetermination of phosphorus. Anal.

Chem. 28: 1756-1758.

Choi, J. I., Lee S. Y. 1999a. Efficient and economical recovery of poly(3-hydroxybutyrate)

from recombinant Escherichia coli by simple digestion with chemicals. Biotechnol. Bioeng.

62: 546-553.

Choi, J. I., Lee S. Y. 1999b. High-level production of poly(3-hydroxybutyrate-co-3-

hydroxyvalerate) by fed-batch culture of recombinant Escherichia coli. Appl. Environ. Microb.

65: 4363-4368.

Choi, J. I., Lee S. Y., Han K. 1998. Cloning of the Alcaligenes latus polyhydroxyalkanoate

biosynthesis genes and use of these genes for enhanced production of poly(3-

hydroxybutyrate) in Escherichia coli. Appl. Environ. Microb. 64: 4897-4903.

Cui, D., Tian F., Ozkan C. S., Wang M., Gao H. 2005. Effect of single wall carbon

nanotubes on human HEK293 cells. Toxicol. Lett. 155: 73-85.

Cumming, R. H., Rees P., Watson J. S.; Zeneca Ltd., UK, 1994. Process for the separation

of solid materials from microorganisms. WO patent 9411491.

Davis, A. L., Keeler J., Laue E. D., Moskau D. 1992. Experiments for recording pure-

absorption heteronuclear correlation spectra using pulsed field gradients. J. Magn. Reson.

98: 207-216.

de Koning, G. J. M., Witholt B. 1997. A process for the recovery of poly(hydroxyalkanoates)

from Pseudomonads. 1. Solubilization. Bioproc. Eng. 17: 7-13.

de Koning, G. J. M., Kellerhals M., vanMeurs C., Wi tholt B. 1997. A process for the

recovery of poly(hydroxyalkanoates) from Pseudomonads. 2. Process development and

economic evaluation. Bioproc. Eng. 17: 15-21.

Deng, Y., Lin X. S., Zheng Z., Deng J. G., Chen J. C., Ma H., Chen G. Q. 2003.

Poly(hydroxybutyrate-co-hydroxyhexanoate) promoted production of extracellular matrix of

articular cartilage chondrocytes in vitro. Biomaterials 24: 4273-4281.

Diniz, S. C., Taciro M. K., Gomez J. G. C., Pradell a J. G. D. 2004. High-cell-density

cultivation of Pseudomonas putida IPT 046 and medium-chain-length polyhydroxyalkanoate

production from sugarcane carbohydrates. Appl. Biochem. Biotechnol. 119: 51-69.

Dixon, D. R., Darveau R. P. 2005. Lipopolysaccharide heterogeneity: Innate host responses

to bacterial modification of lipid A structure. J. Dent. Res. 84: 584-595.

References 122

Doi, Y., Kawaguchi Y., Koyama N., Nakamura S., Hira mitsu M., Yoshida Y., Kimura H.

1992. Synthesis and degradation of polyhydroxyalkanoates in Alcaligenes eutrophus. FEMS

Microbiol. Rev. 103: 103-108.

Doi, Y., Kitamura S., Abe H. 1995. Microbial synthesis and characterization of poly(3-

hydroxybutyrate-co-3-hydroxyhexanoate). Macromolecules 28: 4822-4828.

Doi, Y., Kunioka M., Nakamura Y., Soga K. 1986. Nuclear magnetic resonance studies on

poly(ß-hydroxybutyrate) and a copolyester of ß-hydroxybutyrate and ß-hydroxyvalerate

isolated from Alcaligenes eutrophus H16. Macromolecules 19: 2860-2864.

Dubowski, Y., Vieceli J., Tobias D. J., Gomez A., L in A., Nizkorodov S. A., McIntire T.

M., Finlayson-Pitts B. J. 2004. Interaction of gas-phase ozone at 296 K with unsaturated

self-assembled monolayers: A new look at an old system. J. Phys. Chem. A 108: 10473-

10485.

Dufresne, A., Reche L., Marchessault R. H., Lacroix M. 2001. Gamma-ray crosslinking of

poly (3-hydroxyoctanoate-co-undecenoate). Int. J. Biol. Macromol. 29: 73-82.

Durner, R., Witholt B., Egli T. 2000. Accumulation of poly[(R)-3-hydroxyalkanoates] in

Pseudomonas oleovorans during growth with octanoate in continuous culture at different

dilution rates. Appl. Environ. Microbiol. 66: 3408-3414.

Eckelt, J., Haase T., Loske S., Wolf B. A. 2004. Large scale fractionation of macro-

molecules. Macromol. Mater. Eng. 289: 393-399.

Eggink, G., de Waard P., Huijberts G. N. M. 1992. The role of fatty acid biosynthesis and

degradation in the supply of substrates for poly(3-hydroxyalkanoate) formation in

Pseudomonas putida. FEMS Microb. Rev. 103: 159-164.

Egli, T. 1991. On multiple-nutrient-limited growth of microorganisms, with special reference

to dual limitation by carbon and nitrogen substrates. Antonie van Leeuwenhoek 60: 225-234.

Eliason, T. L., Aloisio S., Donaldson D. J., Cziczo D. J., Vaida V. 2003. Processing of

unsaturated organic acid films and aerosols by ozone. Atmos. Environ. 37: 2207-2219.

Emeruwa, A. C., Hawirko R. Z. 1973. Poly-ß-hydroxybutyrate metabolism during growth

and sporulation of Clostridium botulinum. J. Bacteriol. 116: 989-993.

Eroglu, M. S., Hazer B., Ozturk T., Caykara T. 2005. Hydroxylation of pendant vinyl groups

of poly(3-hydroxy undec-10-enoate) in high yield. J. Appl. Polym. Sci. 97: 2132-2139.

Escalona, A. M., Varela F. R., Gomis A. M.; Repsol Quimica S.A. Madrid, 1996. Procedure

for extraction of polyhydroxyalkanoates from halophilic bacteria which contain them. US

patent 5536419.

123 References

Ferenci, T. 1999. 'Growth of bacterial cultures' 50 years on: towards an uncertainty principle

instead of constants in bacterial growth kinetics. Res. Microbiol. 150: 431-438.

Folch, J., Lees M., Stanley G. H. S. 1957. A simple method for the isolation and purification

of total lipids from animal tissues. J. Biol. Chem. 226: 497-509.

Füchtenbusch, B., Fabritius D., Waltermann M., Stei nbüchel A. 1998. Biosynthesis of

novel copolyesters containing 3-hydroxypivalic acid by Rhodococcus ruber NCIMB 40126

and related bacteria. FEMS Microbiol. Lett. 159: 85-92.

Füchtenbusch, B., Steinbüchel A. 1999. Biosynthesis of polyhydroxyalkanoates from low-

rank coal liquefaction products by Pseudomonas oleovorans and Rhodococcus ruber. Appl.

Microbiol. Biotechnol. 52: 91-95.

Fukui, T., Doi Y. 1998. Efficient production of polyhydroxyalkanoates from plant oils by

Alcaligenes eutrophus and its recombinant strain. Appl. Microbiol. Biot. 49: 333-336.

George, N., Liddell J. M.; Monsanto Company, USA; George, Neil; Liddell, John

Macdonald. 1997. Separating polyester particles from fermentation broth. WO patent

9722654.

Godoy, F., Vancanneyt M., Martinez M., Steinbüchel A., Swings J., Rehm B. H. A. 2003.

Sphingopyxis chilensis sp. nov., a chlorophenol-degrading bacterium that accumulates

polyhydroxyalkanoate, and transfer of Sphingomonas alaskensis to Sphingopyxis alaskensis

comb. nov. Int. J. Sys. Evol. Micr. 53: 473-477.

Gomes, M. E., Reis R. L. 2004. Biodegradable polymers and composites in biomedical

applications: from catgut to tissue engineering - Part 1 - Available systems and their

properties. Int. Mater. Rev. 49: 261-273.

Gorenflo, V., Schmack G., Vogel R., Steinbüchel A. 2001. Development of a process for

the biotechnological large-scale production of 4-hydroxyvalerate-containing polyesters and

characterization of their physical and mechanical properties. Biomacromolecules 2: 45-57.

Greer, W.; Zeneca Ltd. 1994. Peroxide degradation of DNA for viscosity reduction. WO

patent 9410289.

Greer, W.; Zeneca Ltd. 1997. Peroxide degradation of DNA for viscosity reduction. US

patent 5627276.

Hahn, S. K., Chang Y. K., Kim B. S., Chang H. N. 1994. Optimization of microbial poly(3-

hydroxybutyrate) recovery using dispersions of sodium hypochlorite solution and chloroform.

Biotechnol. Bioeng. 44: 256-261.

References 124

Hahn, S. K., Chang Y. K., Lee S. Y. 1995. Recovery and characterization of poly(3-

hydroxybutyric acid) synthesized in Alcaligenes eutrophus and recombinant Escherichia coli.

Appl. Environ. Microbiol. 61: 34-39.

Hampson, J. W., Ashby R. D. 1999. Extraction of lipid-grown bacterial cells by supercritical

fluid and organic solvent to obtain pure medium chain-length polyhydroxyalkanoates. J. Am.

Oil Chem. Soc. 76: 1371-1374.

Hany, R., Hartmann R., Bohlen C., Brandenberger S., Kawada J., Lowe C., Zinn M.,

Witholt B., Marchessault R. H. 2005. Chemical synthesis and characterization of POSS-

functionalized poly [3-hydroxyalkanoates]. Polymer 46: 5025-5031.

Hartmann, R., Hany R., Geiger T., Egli T., Witholt B., Zinn M. 2004. Tailored biosynthesis

of olefinic medium-chain-length poly [(R)-3-hydroxyalkanoates] in Pseudomonas putida

GPo1 with improved thermal properties. Macromolecules 37: 6780-6785.

Hartmann, R., Hany R., Pletscher E., Ritter A., Wit holt B., Zinn M. 2006. Tailor-made

olefinic medium-chain-length poly[(R)-3-hydroxyalkanoates] by Pseudomonas putida GPo1:

Batch versus chemostat production. Biotechnol. Bioeng. 93: 737-746.

Haywood, G. W., Anderson A. J., Williams D. R., Daw es E. A., Ewing D. F. 1991.

Accumulation of a poly(hydroxyalkanoate) copolymer containing primarily 3-hydroxyvalerate

from simple carbohydrate substrates by Rhodococcus sp. ncimb 40126. Int. J. Biol.

Macromol. 13: 83-88.

Hazer, B., Steinbüchel A. 2007. Increased diversification of polyhydroxyalkanoates by

modification reactions for industrial and medical applications. Appl. Microbiol. Biotechnol. 74:

1-12.

Hejazi, P., Vasheghani-Farahani E., Yamini Y. 2003. Supercritical fluid disruption of

Ralstonia eutropha for poly(ß-hydroxybutyrate) recovery. Biotechnol. Progr. 19: 1519-1523.

Hench, L. L., Polak J. M. 2002. Third-generation biomedical materials. Science 295: 1014-

1018.

Hesselmann, R. P. X., Fleischmann T., Hany R., Zehn der A. J. B. 1999. Determination of

polyhydroxyalkanoates in activated sludge by ion chromatographic and enzymatic methods.

J. Microbiol. Methods 35: 111-119.

Hocking, P. J., Marchessault R. H., Timmins M. R., Lenz R. W., Fuller R. C. 1996.

Enzymatic degradation of single crystals of bacterial and synthetic poly(ß-hydroxybutyrate).

Macromolecules 29: 2472-2478.

125 References

Hoffmann, N., Rehm B. H. A. 2004. Regulation of polyhydroxyalkanoate biosynthesis in

Pseudomonas putida and Pseudomonas aeruginosa. FEMS Microbiol. Lett. 237: 1-7.

Holmes, P. A., Wright L. F., Alderson B., Senior P. J.; Imperial Chemical Industries Ltd.,

UK. 1980. A process for the extraction of poly(3-hydroxybutyric acid) from microbial cells. EP

patent 0015123.

Holmes, P. A., Lim G. B.; Imperial Chemical Industries PLC, 1990. Separation process. US

patent 4910145.

Holst, O. 1995. Endotoxin antagonists: possible candidates for the treatment of Gram-

negative sepsis. Angew. Chem. Int. Ed. Engl. 34: 2000-2002.

Hori, K., Soga K., Doi Y. 1994. Effects of culture conditions on molecular weights of poly(3-

hydroxyalkanoates) produced by Pseudomonas putida from octanoate. Biotechnol. Lett. 16:

709-714.

Horowitz, D.; Metabolix, Inc., USA, 2000. Methods for purifying polyhydroxyalkanoates. WO

patent 2000068409.

Horowitz, D. M., Brennan E. M.; Metabolix, Inc., USA, 1999. Methods for separation and

purification of polyhydroxyalkanoates from biomass using ozone treatment. WO patent

9951760.

Hrabak, O. 1992. Industrial production of Poly-ß-hydroxybutyrate. FEMS Microbiol. Rev.

103: 251-255.

Hubbell, J. A. 1995. Biomaterials in tissue engineering. Nat. Biotechnol. 13: 565-576.

Huijberts, G. N. M., Eggink G. 1996. Production of poly(3-hydroxyalkanoates) by

Pseudomonas putida KT2442 in continuous cultures. Appl. Microbiol. Biotechnol. 46: 233-

239.

Huijberts, G. N. M., Eggink G., de Waard P., Huisma n G. W., Witholt B. 1992.

Pseudomonas putida KT2442 cultivated on glucose accumulates poly(3-hydroxyalkanoates)

consisting of saturated and unsaturated monomers. Appl. Environ. Microbiol. 58: 536-544.

Huijberts, G. N. M., Vanderwal H., Wilkinson C., Eg gink G. 1994. Gas-chromatographic

analysis of poly(3-hydroxyalkanoates) in bacteria. Biotechnol. Tech. 8: 187-192.

Huisman, G. W., de Leeuw O., Eggink G., Witholt B. 1989. Synthesis of poly(3-

hydroxyalkanoates) is a common feature of fluorescent Pseudomonads. Appl. Environ.

Microbiol. 55: 1949-1954.

References 126

Hwang, K. J., You S. F., Don T. M. 2006. Disruption kinetics of bacterial cells during

purification of poly-ß-hydroxyalkanoate using ultrasonication. J. Chin. Inst. Chem. Eng 37:

209-216.

Itzigsohn, R., Yarden O., Okon Y. 1995. Polyhydroxyalkanoate analysis in Azospirillum

brasilense. Can. J. Microbiol. 41: 73-76.

Jan, S., Roblot C., Courtois J., Courtois B., Barbo tin J. N., Seguin J. P. 1996. H-1 NMR

spectroscopic determination of poly(3-hydroxybutyrate) extracted from microbial biomass.

Enzyme Microb. Technol. 18: 195-201.

Jan, S., Roblot C., Goethals G., Courtois J., Court ois B., Saucedo J. E. N., Seguin J. P.,

Barbotin J. N. 1995. Study of parameters affecting poly(3-hydroxybutyrate) quantification by

gas chromatography. Anal. Biochem. 225: 258-263.

Jarute, G., Kainz A., Schroll G., Baena J. R., Lend l B. 2004. On-line determination of the

intracellular poly(β-hydroxybutyric acid) content in transformed Escherichia coli and glucose

during PHB production using stopped-flow attenuated total reflection FT-IR Spectrometry.

Anal.Chem. 76: 6353-6358.

Jendrossek, D. 2001. Microbial degradation of polyesters. Adv. Biochem. Eng. Biotechnol.:

Springer-Verlag, Berlin Heidelberg. p 293-325.

Jendrossek, D., Selchow O., Hoppert M. 2007. Poly(3-hydroxybutyrate) granules at the

early stages of formation are localized close to the cytoplasmic membrane in Caryophanon

latum. Appl. Environ. Microbiol. 73: 586-593.

Jiang, X., Ramsay J. A., Ramsay B. A. 2006. Acetone extraction of mcl-PHA from

Pseudomonas putida KT2440. J. Microbiol. Meth. 67: 212-219.

Jung, K., Hazenberg W., Prieto M., Witholt B. 2001. Two-stage continuous process

development for the production of medium-chain-length poly(3-hydroxyalkanoates).

Biotechnol. Bioeng. 72: 19-24.

Kannan, R. Y., Salacinski H. J., Butler P. E., Seif alian A. M. 2005. Polyhedral oligomeric

silsesquioxane nanocomposites: The next generation material for biomedical applications.

Accounts Chem. Res. 38: 879-884.

Kassab, A. C., Piskin E., Bilgic S., Denkbas E. B., Xu K. 1999. Embolization with

polyhydroxybutyrate (PHB) microspheres: In-vivo studies. J. Bioact. Compat. Polym. 14: 291-

303.

127 References

Kassab, A. C., Xu K., Denkbas E. B., Dou Y., Zhao S ., Piskin E. 1997. Rifampicin carrying

polyhydroxybutyrate microspheres as a potential chemoembolization agent. J. Biomater. Sci.

Polym. Ed. 8: 947-961.

Kawalec, M., Adamus G., Kurcok P., Kowalczuk K., Fo ltran I., Focarete M. L., Scandola

M. 2007. Carboxylate induced degradation of PHB. Biomacromolecules 8: 1053-1058.

Kennedy, S. B., Washburn N. R., Simon C. G., Amis E . J. 2006. Combinatorial screen of

the effect of surface energy on fibronectin-mediated osteoblast adhesion, spreading and

proliferation. Biomaterials 27: 3817-3824.

Kessler, B., Palleroni N. J. 2000. Taxonomic implications of synthesis of poly-ß-

hydroxybutyrate and other poly-ß-hydroxyalkanoates by aerobic Pseudomonads. Int. J. Syst.

Evol. Micr. 50: 711-713.

Kessler, B., Westhuis R., Witholt B., Eggink G. 2001. Production of microbial polyesters:

fermentation and downstream processes. Babel W, Steinbüchel A, editors. Berlin: Springer.

159-182 p.

Kessler, B., Witholt B. 1998. Synthesis, recovery and possible application of medium-chain-

length polyhydroxyalkanoates: A short overview. Macromol. Symp. 130: 245-260.

Khosravi-Darani, K., Vasheghani-Farahani E., Yamini Y., Bahramifar N. 2003. Solubility

of poly(ß-hydroxybutyrate) in supercritical carbon dioxide. J. Chem. Eng. Data 48: 860-863.

Kim, B. S. 2002. Production of medium chain length polyhydroxyalkanoates by fed-batch

culture of Pseudomonas oleovorans. Biotechnol. Lett. 24: 125-130.

Kim, B. S., Lee S. C., Lee S. Y., Chang H. N., Chan g Y. K., Woo S. I. 1994a. Production of

poly(3-hydroxybutyric-co-3-hydroxyvaleric acid) by fed-batch culture of Alcaligenes

eutrophus with substrate control using online glucose analyzer. Enzyme Microb. Tech. 16:

556-561.

Kim, B. S., Lee S. C., Lee S. Y., Chang H. N., Chan g Y. K., Woo S. I. 1994b. Production of

poly(3-hydroxybutyric acid) by fed-batch culture of Alcaligenes eutrophus with glucose-

concentration control. Biotechnol. Bioeng. 43: 892-898.

Kim, D. Y., Kim Y. B., Rhee Y. H. 2000. Evaluation of various carbon substrates for the

biosynthesis of polyhydroxyalkanoates bearing functional groups by Pseudomonas putida.

Int. J. Biol. Macromol. 28: 23-29.

Kim, J. S., Lee B. H., Kim B. S. 2005. Production of poly(3-hydroxybutyrate-co-4-

hydroxybutyrate) by Ralstonia eutropha. Biochem. Eng. J. 23: 169-174.

References 128

Kim, K., Sonn Y., Lee M., Jung S.; Cheil Synthetics Inc., S. Korea. 1996. Preparation

process of poly-3-hydroxybutyrate. KR patent 9609065.

Kimura, H., Ohura T., Takeishi M., Nakamura S., Doi Y. 1999. Effective microbial

production of poly(4-hydroxybutyrate) homopolymer by Ralstonia eutropha H16. Polym. Int.

48: 1073-1079.

Kimura, H., Yamamoto T., Iwakura K. 2002. Biosynthesis of polyhydroxyalkanoates from

1,3-propanediol by Chromobacterium sp. Polym. J. 34: 659-665.

Kivirikko, K. I ., Laitinen O ., Prockop D.J . 1967. Modifications of specific assay for

hydroxyproline in urine. Anal. Biochem. 19: 249.

Knirel, Y. A., Lindner B., Vinogradov E. V., Kochar ova N. A., Senchenkova S. N.,

Shaikhutdinova R. Z., Dentovskaya S. V., Fursova N. K., Bakhteeva I. V., Titareva G. M.,

Balakhonov S. V., Holst O., Gremyakova T. A., Pier G. B., Anisimov A. P. 2005.

Temperature-dependent variations and intraspecies diversity of the structure of the

lipopolysaccharide of Yersinia pestis. Biochemistry 44: 1731-1743.

Koller, M., Hesse P., Bona R., Kutschera C., Atlic A., Braunegg G. 2007. Potential of

various archae- and eubacterial strains as industrial polyhydroxyalkanoate producers from

whey. Macromol. Biosci. 7: 218-226.

Kranz, R. G., Gabbert K. K., Locke T. A., Madigan M . T. 1997. Polyhydroxyalkanoate

production in Rhodobacter capsulatus: genes, mutants, expression, and physiology. Appl.

Environ. Microb. 63: 3003-3009.

Kurdikar, D. L., Strauser F. E., Solodar A. J., Pas ter M. D.; Monsanto Co., USA. 1998.

High temperature PHA extraction using PHA-poor solvents. WO patent 9846783.

Lafferty, R. M., Agroferm A.G., Switzerland, 1977. Use of cyclic carbonic acid esters as

solvent for poly(ß-hydroxybutyric acid). US patent 4140741.

Lageveen, R. 1986. Oxidation of aliphatic compounds by Pseudomonas oleovorans. PhD

thesis, Rijksuniversiteit Groningen.

Lageveen, R. G., Huisman G. W., Preusting H., Ketel aar P., Eggink G., Witholt B. 1988.

Formation of polyesters by Pseudomonas oleovorans: Effect of substrates on formation and

composition of poly-(R)-3-hydroxyalkanoates and poly-(R)-3-hydroxyalkenoates. Appl.

Environ. Microbiol. 54: 2924-2932.

References 129

Lakshman, K., Rastogi N. K., Shamala T. R. 2004. Simultaneous and comparative

assessment of parent and mutant strain of Rhizobium meliloti for nutrient limitation and

enhanced polyhydroxyalkanoate (PHA) production using optimization studies. Process

Biochem. 39: 1977-1983.

Lee, E. Y., Choi C. Y. 1995. Gas chromatography mass spectrometric analysis and its

application to a screening procedure for novel bacterial polyhydroxyalkanoic acids containing

long chain saturated and unsaturated monomers. J. Ferment. Bioeng. 80: 408-414.

Lee, I. Y., Kim G. J., Choi D. K., Yeon B. K., Park Y. H. 1996. Improvement of

hydroxyvalerate fraction in poly(ß-hydroxybutyrate-co-ß-hydroxyvalerate) by a mutant strain

of Alcaligenes eutrophus. J. Ferment. Bioeng. 81: 255-258.

Lee, M. Y., Lee T. S., Park W. H. 2001. Effect of side chains on the thermal degradation of

poly(3-hydroxyalkanoates). Macromol. Chem. Phys. 202: 1257-1261.

Lee, M. Y., Park W. H., Lenz R. W. 2000a. Hydrophilic bacterial polyesters modified with

pendant hydroxyl groups. Polymer 41: 1703-1709.

Lee, S. Y. 1996a. Plastic bacteria? Progress and prospects for polyhydroxyalkanoate

production in bacteria. Trends Biotechnol. 14: 431-438.

Lee, S. Y. 1996b. Bacterial polyhydroxyalkanoates. Biotechnol. Bioeng. 49: 1-14.

Lee, S. Y., Choi J. I. 1998. Effect of fermentation performance on the economics of poly(3-

hydroxybutyrate) production by Alcaligenes latus. Polym. Degrad. Stabil. 59: 387-393.

Lee, S. Y., Choi J. I., Han K., Song J. Y. 1999a. Removal of endotoxin during purification of

poly(3-hydroxybutyrate) from Gram-negative bacteria. Appl. Env. Microb. 65: 2762-2764.

Lee, S. Y., Wong H. H., Choi J. I., Lee S. H., Lee S. C., Han C. S. 2000b. Production of

medium-chain-length polyhydroxyalkanoates by high-cell-density cultivation of Pseudomonas

putida under phosphorus limitation. Biotechnol. Bioeng. 68: 466-470.

Lemoigne, M. 1926. Produits de deshydration et de polymerisation de l'acide β-oxybutyric.

Bull. Soc. Chem. Biol. 8: 770-782.

Liddell, J. M., Locke T. J.; Zeneca Limited, 1997. Production of plastics materials from

microorganisms. US patent 5691174.

Lillo, J. G., Rodriguezvalera F. 1990. Effects of culture conditions on poly(ß-hydroxybutyric

acid) production by Haloferax mediterranei. Appl. Environ. Microb. 56: 2517-2521.

Ling, Y., Williams D. R. G., Thomas C. J., Middelbe rg A. P. J. 1997a. Recovery of poly-3-

hydroxybutyrate from recombinant Escherichia coli by homogenization and centrifugation.

Biotechnol. Tech. 11: 409-412.

References 130

Ling, Y., Wong H. H., Thomas C. J., Williams D. R. G., Middelberg A. P. J. 1997b. Pilot-

scale extraction of PHB from recombinant Escherichia coli by homogenization and

centrifugation. Bioseparation 7: 9-15.

Liu, W., Chen G. Q. 2007. Production and characterization of medium-chain-length

polyhydroxyalkanoate with high 3-hydroxytetradecanoate monomer content by fadB and

fadA knockout mutant of Pseudomonas putida KT2442. Appl. Microbiol. Biot.

Lopez-Lopez, A., Castellote-Bargallo A. I., Lopez-S abater M. C. 2001. Comparison of two

direct methods for the determination of fatty acids in human milk. Chromatographia 54: 743-

747.

Manna, A., Banerjee R., Paul A. K. 1999. Accumulation of poly (3-hydroxybutyric acid) by

some soil Streptomyces. Curr. Microbiol. 39: 153-158.

Marchessault, R. H., Morin F. G., Wong S., Saracova n I. 1995. Artificial granule

suspensions of long side-chain poly(3-hydroxyalkanoate). Can. J. Microbiol. 41: 138-142.

Martini, F., Perazzo L., Vietto P.; W. R. Grace & Co.-Conn., 1989. Sheets materials of HB

polymers. US patent 4826493.

Matsushita, H., Yoshida S., Tawara T.; Mitsubishi Gas Chemical Co, Japan. 1995.

Extraction of poly(3-hydroxybutyric acid) from microorganisms. JP patent 07079788.

McCool, G. J., Fernandez T., Li N., Cannon M. C. 1996. Polyhydroxyalkanoate inclusion-

body growth and proliferation in Bacillus megaterium. FEMS Microbiol. Lett. 138: 41-48.

Metzner, K., Sela M., Schaffer J.; Buna Sow Leuna Olefinverbund GmbH, Germany, 1997.

Agents for extracting polyhydroxyalkanoic acids. WO patent 9708931.

Middelberg, A. P. J., Lee S. Y., Martin J., William s D. R. G., Chang H. N. 1995. Size

analysis of poly(3-hydroxybutyric acid) granules produced in recombinant Escherichia coli.

Biotechnol. Lett. 17: 205-210.

Minnikin, D. E., Abdolrah.H. 1974. Replacement of phosphatidylethanolamine and acidic

phospholipids by an ornithine-amide lipid and a minor phosphorus-free lipid in Pseudomonas

fluorescens Ncmb 129. FEBS Letters 43: 257-260.

Minnikin, D. E., Abdolrah.H, Baddiley J. 1974. Replacement of acidic phospholipids by

acidic glycolipids in Pseudomonas diminuta. Nature 249: 268-269.

Moffat, C. F., McGill A. S., Anderson R. S. 1991. The Production of artefacts during

preparation of fatty acid methyl esters from fish oils, food products and pathological samples.

HRC J. High Resolut. Chromatogr. 14: 322-326.

Monod, J. 1949. The growth of bacterial cultures. Annu. Rev. Microbiol. 3: 371-394.

131 References

Mosmann, T. 1983. Rapid colorimetric assay for cellular growth and survival: application to

proliferation and cytotoxicity assays. J. Immunol. Meth. 65, 55-63.

Mrksich, M. 2000. A surface chemistry approach to studying cell adhesion. Chem. Soc. Rev.

29: 267-273.

Niskanen, A., Kiutamo T., Raisanen S., Raevuori M. 1978. Determination of fatty acid

compositions of Bacillus cereus and related bacteria - rapid gas chromatographic method

using a glass capillary column. Appl. Environ. Microbiol. 35: 453-455.

Noda, I.; Procter and Gamble Co., USA, 1995. Process for recovering

polyhydroxyalkanoates using air classification. WO patent 9533064.

Noda, I.; The Procter & Gamble Company, USA, 1998. Process for recovering

polyhydroxyalkanoates using air classification. US patent 5849854.

Noda, I., Schechtman L. A.; Procter and Gamble Company, USA. 1997. Solvent extraction

of polyhydroxyalkanoates from biomass. WO patent 9707230.

Numazawa, R., Miyamori T., Sakimae A., Onishi H.; Mitsubishi Rayon Co., Ltd., Japan.

1987. Separation and purification of poly-ß-hydroxybutyric acid from cell extracts. JP patent

62205787.

Oehmen, A., Keller-Lehmann B., Zeng R. J., Yuan Z. G., Keller E. 2005. Optimisation of

poly-ß-hydroxyalkanoate analysis using gas chromatography for enhanced biological

phosphorus removal systems. J. Chromatogr. A 1070: 131-136.

Ohleyer, E.; Firmenich S. A., Switzerland, 1993. Extraction of poly-(ß)-hydroxyoctanoate

from microbial biomass. WO patent 9311656.

Opitz, F., Schenke-Layland K., Cohnert T. U., Starc her B., Halbhuber K. J., Martin D. P.,

Stock U. A. 2004. Tissue engineering of aortic tissue: dire consequence of suboptimal

elastic fiber synthesis in vivo. Cardiovasc. Res. 63: 719-730.

Owens, J. D., Legan J. D. 1987. Determination of the Monod substrate saturation constant

for microbial growth. FEMS Microbiol. Rev. 46: 419-432.

Palmer, A. G., Cavanagh J., Byrd R. A., Rance M. 1992. Sensitivity improvement in 3-

dimensional heteronuclear correlation NMR-Spectroscopy. J. Magn. Reson. 96: 416-424.

Park, J. S., Park H. C., Huh T. L., Lee Y. H. 1995. Production of poly(ß-hydroxybutyrate) by

Alcaligenes eutrophus transformants harboring cloned Phbcab genes. Biotechnol. Lett. 17:

735-740.

Petsch, D., Anspach F. B. 2000. Endotoxin removal from protein solutions. J. Biotechnol.

76: 97-119.

References 132

Pötter, M., Steinbüchel A. 2004. Poly(3-hydroxybutyrate) granule-associated proteins:

Impacts on poly(3-hydroxybutyrate) synthesis and degradation. Biomacromolecules 6: 552-

560.

Pouton, C. W., Akhtar S. 1996. Biosynthetic polyhydroxyalkanoates and their potential in

drug delivery. Adv. Drug. Deliver. Rev. 18: 133-162.

Preusting, H., Kingma J., Huisman G., Steinbüchel A ., Witholt B. 1993. Formation of

polyester blends by a recombinant strain of Pseudomonas oleovorans: different poly(3-

hydroxyalaknoates) are stored in separate granules. J. Environ. Polym. Degrad. 1: 11-21.

Prieto, M. A., Kellerhals M. B., Bozzato G. B., Rad novic D., Witholt B., Kessler B. 1999.

Engineering of stable recombinant bacteria for production of chiral medium-chain-length

poly-3-hydroxyalkanoates. Appl. Environ. Microb. 65: 3265-3271.

Punshon, G., Vara D. S., Sales K. M., Kidane A. G., Salacinski H. J., Seifalian A. M.

2005. Interactions between endothelial cells and a poly(carbonate-silsesquioxane-bridge-

urea)urethane. Biomaterials 26: 6271-6279.

Qi, Q. S., Rehm B. H. A. 2001. Polyhydroxybutyrate biosynthesis in Caulobacter crescentus:

molecular characterization of the polyhydroxybutyrate synthase. Microbiol.-SGM 147: 3353-

3358.

Qu, X. H., Wu Q., Liang J., Qu X., Wang S. G., Chen G. Q. 2005. Enhanced vascular-

related cellular affinity on surface modified copolyesters of 3-hydroxybutyrate and 3-

hydroxyhexanoate (PHBHHx). Biomaterials 26: 6991-7001.

Raetz, C. R. H., Reynolds C. M., Trent M. S., Bisho p R. E. 2007. Lipid A modification

systems in Gram-negative bacteria. Annu. Rev. Biochem. 76: 295-329.

Ramsay, B. A. 1990. Recovery of poly-3-hydroxyalkanoic acid granules by a surfactant-

hypochlorite treatment. Biotechnol. Techn. 4: 221-226.

Ramsay, B. A., Ramsay J., Berger E., Chavarie C., B raunegg G.; Ecole Polytechnique

Montreal, 1992. Separation of poly-ß-hydroxyalkanoic acid from microbial biomass. US

patent 5110980.

Ramsay, J. A., Berger E., Voyer R., Chavarie C., Ra msay B. A. 1994. Extraction of poly-3-

hydroxybutyrate using chlorinated solvents. Biotechnol. Techn. 8: 589-594.

Randriamahefa, S., Renard E., Guérin P., Langlois V . 2003. Fourier transform infrared

spectroscopy for screening and quantifying production of PHAs by Pseudomonas grown on

sodium octanoate. Biomacromolecules 4: 1092-1097.

133 References

Rapthel, I., Lehmann O., Runkel D., Mayer T., Rauch stein K. D.; Buna A.G., Germany,

1993a. Manufacture of colorless polyhydroxyalkanoates by extraction of bacterial biomass

with acetic acid containing ß-butyrolactone. DE patent 4215860.

Rapthel, I., Lehmann O., Runkel D., Mayer T., Rauch stein K. D., Schaffer J.; Buna A.G.,

Germany, 1993b. Manufacture of colorless polyhydroxyalkanoates by extraction of bacterial

biomass with acetic acid containing acetic anhydride. DE patent 4215861.

Rehm, B. H. A. 2003. Polyester synthases: natural catalysts for plastics. Biochem. J. 376:

15-33.

Rehm, B. H. A. 2007. Biogenesis of microbial polyhydroxyalkanoate granules: a platform

technology for the production of tailor-made bioparticles. Curr. Issues Mol. Biol. 9: 41-62.

Ren, Q., Grubelnik A., Hoerler M., Ruth K., Hartman n R., Felber H., Zinn M. 2005.

Bacterial poly(hydroxyalkanoates) as a source of chiral hydroxyalkanoic acids.

Biomacromolecules 6: 2290-2298.

Renard, E., Poux A., Timbart L., Langlois V., Gueri n P. 2005. Preparation of a novel

artificial bacterial polyester modified with pendant hydroxyl groups. Biomacromolecules 6:

891-896.

Rietschel, E. T., Kirikae T., Schade F. U., Mamat U ., Schmidt G., Loppnow H., Ulmer A.

J., Zahringer U., Seydel U., Dipadova F., Schreier M., Brade H. 1994. Bacterial endotoxin

- molecular relationships of structure to activity and function. FASEB J. 8: 217-225.

Riis, V., Mai W. 1988. Gas chromatographic determination of poly-ß-hydroxybutyric acid in

microbial biomass after hydrochloric acid propanolysis. J. Chromatogr. 445: 285-289.

Rizk, S.; Tepha, Inc., 2005. Non-curling polyhydroxyalkanoate sutures. WO patent

2006015276.

Rochex, A., Lebeault J. M. 2007. Effects of nutrients on biofilm formation and detachment

of a Pseudomonas putida strain isolated from a paper machine. Water Res. 41: 2885-2892.

Rodrigues, M. F. D., Valentin H. E., Berger P. A., Tran M., Asrar J., Gruys K. J.,

Steinbüchel A. 2000. Polyhydroxyalkanoate accumulation in Burkholderia sp.: a molecular

approach to elucidate the genes involved in the formation of two homopolymers consisting of

short-chain-length 3-hydroxyalkanoic acids. Appl. Microbiol. Biot. 53: 453-460.

Steinbüchel A. 2000. Polyhydroxyalkanoate accumulation in Burkholderia sp.: a molecular

approach to elucidate the genes involved in the formation of two homopolymers consisting of

short-chain-length 3-hydroxyalkanoic acids. Appl. Microbiol. Biot. 53: 453-460.

References 134

Rotzsche, H. 1991. Gas chromatographic analysis of fatty acid salts. J. Chromatogr. 552:

281-288.

Rouxhet, L., Legras R., Schneider Y. J. 1998. Interactions between a biodegradable

polymer, poly(hydroxybutyrate-hydroxyvalerate), proteins and macrophages. Macromol.

Symp. 130: 347-366.

Runkel, D., Lehmann O., Mayer T., Rauchstein K. D., Schaffer J.; Buna A.G., Germany.

1993a. Manufacture of colorless polyhydroxyalkanoates by extraction of bacterial biomass

with acetic acid. DE patent 4215862.

Runkel, D., Lehmann O., Mayer T., Rauchstein K. D., Schaffer J.; Buna A.G., Germany,

1993b. Manufacture of colorless polyhydroxyalkanoates by extraction of bacterial biomass

with acetic anhydride. DE patent 4215864.

Ruth, K., Grubelnik A., Hartmann R., Egli T., Zinn M., Ren Q. 2007. Efficient production of

(R)-3-hydroxycarboxylic acids by biotechnological conversion of polyhydroxyalkanoates and

their purification. Biomacromolecules 8: 279-286.

Sabirova, J. S., Ferrer M., Lunsdorf H., Wray V., K alscheuer R., Steinbüchel A., Timmis

K. N., Golyshin P. N. 2006. Mutation in a "tesB-like" hydroxyacyl-coenzyme A-specific

thioesterase gene causes hyperproduction of extracellular polyhydroxyalkanoates by

Alcanivorax borkumensis SK2. J. Bacteriol. 188: 8452-8459.

Schlegel, H.G. and Gottschalk, G., 1962. Poly-beta-hydroxybuttersäure, ihre Verbreitung,

Funktion und Biosynthese. Angew. Chem. 74: 342-347.

Schmidt, J., Biedermann W., Schmiechen H.; Akademie der Wissenschaften der DDR,

Ger. Dem. Rep., 1985. Extraction of poly-ß-hydroxybutyric acid from bacterial biomass. DD

patent 229428.

Schmidt, J., Schmiechen H., Rehm H., Trennert M.; Akademie der Wissenschaften der

DDR, Ger. Dem. Rep., 1986. Verfahren zur Gewinnung von PHB aus getrockneten

Biomassen. DD patent 239609.

Schumann, D., Mueller R. A.; UFZ Umweltforschungszentrum Leipzig-Halle G.m.b.H.,

Germany, 2001. Method for obtaining polyhydroxyalkanoates (PHA) or the copolymers

thereof. WO patent 2001068892.

Schumann, D., Müller R. A.; UFZ Umweltforschungszentrum Leipzig-Halle G.m.b.H.,

Germany, 2003. Method for obtaining polyhydroxyalkanoates (PHA) and the copolymers

thereof. WO patent 0168892.

135 References

Seidel, H., Voigt B., Roethe K. P., Rosahl B., Moth es S.; Akademie der Wissenschaften,

Germany, 1991. Supercritical fluid extraction in polyhydroxybutyrate and ubiquinone

extraction for bacteria. DD patent 294280.

Sela, M., Metzner K.; Buna Sow Leuna Olefinverbund Gmbh, Germany, 1996. Use of ethyl

lactate as an extractant for poly(hydroxyalkanoic acids). DE patent 19533459.

Sendil, D., Gursel I., Wise D. L., Hasirci V. 1999. Antibiotic release from biodegradable

PHBV microparticles. J. Control. Rel. 59: 207-217.

Sevastianov, V. I., Perova N. V., Shishatskaya E. I ., Kalacheva G. S., Volova T. G. 2003.

Production of purified polyhydroxyalkanoates (PHAs) for applications in contact with blood. J.

Biomater. Sci. Polym. Ed. 14: 1029-1042.

Shaka, A. J., Barker P. B., Freeman R. 1985. Computer-optimized decoupling scheme for

wideband applications and low level operation. J. Magn. Reson. 64: 547-552.

Shangguan, Y. Y., Wang Y. W., Wu Q., Chen G. Q. 2006. The mechanical properties and in

vitro biodegradation and biocompatibility of UV-treated poly(3-hydroxybutyrate-co-3-

hydroxyhexanoate). Biomaterials 27: 2349-2357.

Shantha, N. C., Decker E. A., Hennig B. 1993. Comparison of methylation methods for the

quantitation of conjugated linoleic acid isomers. J. AOAC Int. 76: 644-649.

Shikanov, A., Kumar N., Domb A. J. 2005. Biodegradable polymers: an update. Israel J.

Chem. 45: 393-399.

Silva, J. A., Tobella L. M., Becerra J., Godoy F., Martínez M. A. 2007. Biosynthesis of

poly-β-hydroxyalkanoate by Brevundimonas vesicularis LMG P-23615 and Sphingopyxis

macrogoltabida LMG 17324 using acid-hydrolyzed sawdust as carbon source. J. Biosci.

Bioeng. 103: 542-546.

Silver, J. H., Lin J.-C., Lim F., Tegoulia V. A., C haudhury M. K., Cooper S. L. 1999.

Surface properties and hemocompatibility of alkyl-siloxane monolayers supported on silicone

rubber: effect of alkyl chain length and ionic functionality. Biomaterials 20: 1533-1543.

Slepecky, R. A., Law J. H. 1960. A rapid spectrophotometric assay of alpha, ß-unsaturated

acids and ß-hydroxy acids. Anal. Chem. 32: 1697-1699.

Sodian, R., Hoerstrup S. P., Sperling J. S., Daebri tz S., Martin D. P., Moran A. M., Kim

B. S., Schoen F. J., Vacanti J. P., Mayer J. E. 2000a. Early in vivo experience with tissue-

engineered trileaflet heart valves. Circulation 102: 22-29.

References 136

Sodian, R., Hoerstrup S. P., Sperling J. S., Martin D. P., Daebritz S., Mayer J. E.,

Vacanti J. P. 2000b. Evaluation of biodegradable, three-dimensional matrices for tissue

engineering of heart valves. Asaio J. 46: 107-110.

Sodian, R., Loebe M., Hein A., Martin D. P., Hoerst rup S. P., Potapov E. V., Hausmann

H. A., Lueth T., Hetzer R. 2002. Application of stereolithography for scaffold fabrication for

tissue engineered heart valves. Asaio J. 48: 12-16.

Sodian, R., Sperling J. S., Martin D. P., Egozy A., Stock U., Mayer J. E., Vacanti J. P.

2000c. Fabrication of a trileaflet heart valve scaffold from a polyhydroxyalkanoate

biopolyester for use in tissue engineering. Tissue Eng. 6: 183-188.

Stageman, J. F.; Imperial chemical industries, 1985. Extraction process. US patent

4562245.

Steeghs, L., de Cock H., Evers E., Zomer B., Tommas sen J., van der Ley P. 2001. Outer

membrane composition of a lipopolysaccharide-deficient Neisseria meningitidis mutant.

EMBO Journal 17: 6937–6945.

Steinbüchel, A., Valentin H. E. 1995. Diversity of bacterial polyhydroxyalkanoic acids.

FEMS Microbiol. Lett. 128: 219-228.

Stewart, J. C. M. 1980. Colorimetric determination of phospholipids with ammonium

ferrothiocyanate. Anal. Biochem. 104: 10-14.

Stock, U. A., Nagashima M., Khalil P. N., Nollert G . D., Herden T., Sperling J. S., Moran

A., Lien J., Martin D. P., Schoen F. J., Vacanti J. P., Mayer J. E. 2000a. Tissue-

engineered valved conduits in the pulmonary circulation. J. Thorac. Cardiovasc. Surg. 119:

732-740.

Stott, K., Keeler J., Van Q. N., Shaka A. J. 1997. One-dimensional NOE experiments using

pulsed field gradients. J. Magn. Reson. 125: 302-324.

Sun, Z. Y., Ramsay J. A., Guay M., Ramsay B. A. 2007. Fermentation process

development for the production of medium-chain-length poly-3-hyroxyalkanoates. Appl.

Microbiol. Biot. 75: 475-485.

Suriyamongkol, P., Weselake R., Narine S., Moloney M., Shah S. 2007. Biotechnological

approaches for the production of polyhydroxyalkanoates in microorganisms and plants - A

review. Biotechnol. Adv. 25: 148-175.

Sutcliffe, I. C., Shaw N. 1991. Atypical lipoteichoic acids of Gram-positive bacteria. J.

Bacteriol. 173: 7065-7069.

137 References

Svensson, M., Han L., Silfversparre G., Haggstrom L ., Enfors S. O. 2005. Control of

endotoxin release in Escherichia coli fed-batch cultures. Bioproc. Biosyst. Eng. 27: 91-97.

Szewczyk, E., Mikucki J. 1989. Poly-ß-hydroxybutyric acid in Staphylococci. FEMS

Microbiol. Lett. 61: 279-284.

Tajima, K., Igari T., Nishimura D., Nakamura M., Sa toh Y., Munekata M. 2003. Isolation

and characterization of Bacillus sp. INT005 accumulating polyhydroxyalkanoate (PHA) from

gas field soil. J. Biosci. Bioeng. 95: 77-81.

Tamer, I. M., Moo-Young M., Chisti Y. 1998a. Disruption of Alcaligenes latus for recovery of

poly(ß-hydroxybutyric acid): comparison of high-pressure homogenization, bead milling, and

chemically induced lysis. Ind. Eng. Chem. Res. 37: 1807-1814.

Tamer, I. M., Moo-Young M., Chisti Y. 1998b. Optimization of poly(ß-hydroxybutyric acid)

recovery from Alcaligenes latus: combined mechanical and chemical treatments. Bioprocess

Eng. 19: 459-468.

Taniguchi, I., Kagotani K., Kimura Y. 2003. Microbial production of poly(hydroxyalkanoate)

from waste edible oils. Green Chem. 5: 545-548.

Terada, M., Marchessault R. H. 1999. Determination of solubility parameters for poly(3-

hydroxyalkanoates). Int. J. Biol. Macromol. 25: 207-215.

Tesema, Y., Raghavan D., Stubbs J. 2004. Bone cell viability on collagen immobilized

poly(3-hydroxybutrate-co-3-hydroxyvalerate) membrane: Effect of surface chemistry. J. Appl.

Pol. Sci. 93: 2445-2453.

Tezcaner, A., Bugra K., Hasirci V. 2003. Retinal pigment epithelium cell culture on surface

modified poly(hydroxybutyrate-co-hydroxyvalerate) thin films. Biomaterials 24: 4573-4583.

Timm, A., Steinbüchel A. 1990. Formation of polyesters consisting of medium-chain-length

3-hydroxyalkanoic acids from gluconate by Pseudomonas aeruginosa and other fluorescent

Pseudomonads. Appl. Environ. Microbiol. 56: 3360-3367.

Trent, M. S., Stead C. M., Tran A. X., Hankins J. V . 2006. Diversity of endotoxin and its

impact on pathogenesis. J. Endotoxin Res. 12: 205-223.

Turakhia, M. H., Characklis W. G. 1989. Activity of Pseudomonas aeruginosa in biofilms -

effect of calcium. Biotechnol. Bioeng. 33: 406-414.

U.S. Department of health and human services, F. D. A. 1997a. Guidance for industry.

Rockville. 54 p.

U.S. Department of health and human services, F. D. A. 1997b. guideline on validation of

the Limulus amebocyte lysate test. Rockville. 54 p.

References 138

Ueda, H., Tabata Y. 2003. Polyhydroxyalkanoate derivatives in current clinical applications

and trials. Adv. Drug Deliv. Rev. 55: 501-518.

Valappil, S. P., Boccaccini A. R., Bucke C., Roy I. 2007a. Polyhydroxyalkanoates in Gram-

positive bacteria: insights from the genera Bacillus and Streptomyces. Anton. Leeuw. Int. J.

G. 91: 1-17.

Valappil, S. P., Peiris D., Langley G. J., Hemiman J. M., Boccaccini A. R., Bucke C., Roy

I. 2007b. Polyhydroxyalkanoate (PHA) biosynthesis from structurally unrelated carbon

sources by a newly characterized Bacillus spp. J. Biotechnol. 127: 475-487.

Valentin, H. E., Steinbüchel A. 1995. Accumulation of poly(3-hydroxybutyric acid co-3-

hydroxyvaleric acid co-4-hydroxyvaleric acid) by mutants and recombinant strains of

Alcaligenes eutrophus. J. Environ. Polym. Degrad. 3: 169-175.

van Hee, P., Elumbaring A., van der Lans R., van de r Wielen L. A. M. 2006. Selective

recovery of polyhydroxyalkanoate inclusion bodies from fermentation broth by dissolved-air

flotation. J. Colloid Interf. Sci. 297: 595-606.

Vanlautem, N., Gilain J.; Solvay & Cie., 1982. Process for separating poly-ß-

hydroxybutyrates from a biomass. US patent 4310684.

Vanlautem, N., Gilain J.; Solvay & Cie., 1987. Process for extracting poly-ß-

hydroxybutyrates by means of a solvent from an aqueous suspension of microorganisms. US

patent 4705604.

Viktor N. Vasilets, Artem V. Kuznetsov, Sevastianov V. I. 2004. Vacuum ultraviolet

treatment of polyethylene to change surface properties and characteristics of protein

adsorption. J. Biomed. Mater. Res. A 69A: 428-435.

Volova, T. G., Trusova M. Y., Kalacheva G. S., Kozh evnicov I. V. 2006. Physiological-

biochemical properties and the ability to synthesize polyhydroxyalkanoates of the glucose-

utilizing strain of the hydrogen bacterium Ralstonia eutropha B8562. Appl. Microbiol. Biot. 73:

429-433.

Wallen, L. L., Rohwedder W. K. 1974. Poly-ß-hydroxyalkanoate from activated sludge.

Environ. Science. Technol. 8: 576-579.

Walsem, H., Zhong L., Shih S.; Metabolix, 2004. Polymer extraction methods. US patent

2004013204.

Wang, F. L., Lee S. Y. 1997a. Poly(3-hydroxybutyrate) production with high productivity and

high polymer content by a fed-batch culture of Alcaligenes latus under nitrogen limitation.

Appl. Environ. Microb. 63: 3703-3706.

139 References

Wang, F. L., Lee S. Y. 1997b. Production of poly(3-hydroxybutyrate) by fed-batch culture of

filamentation-suppressed recombinant Escherichia coli. Appl. Environ. Microb. 63: 4765-

4769.

Wang, X. Y., Karbarz M. J., McGrath S. C., Cotter R . J., Raetz C. R. H. 2004a. MsbA

transporter-dependent lipid A 1-dephosphorylation on the periplasmic surface of the inner

membrane - Topography of Francisella novicida LpxE expressed in Escherichia coli. J. Biol.

Chem. 279: 49470-49478.

Wang, X. Y., McGrath S. C., Cotter R. J., Raetz C. R. H. 2006. Expression cloning and

periplasmic orientation of the Francisella novicida lipid A4 '-phosphatase LpxF. J. Biol. Chem.

281: 9321-9330.

Wang, Y. W., Wu Q. O., Chen G. Q. A. 2004b. Attachment, proliferation and differentiation

of osteoblasts on random biopolyester poly(3-hydroxybutyrate-co-3-hydroxyhexanoate)

scaffolds. Biomaterials 25: 669-675.

Wang, Z. X., Itoh Y., Hosaka Y., Kobayashi I., Naka no Y., Maeda I., Umeda F.,

Yamakawa J., Kawase M., Yagi K. 2003. Novel transdermal drug delivery system with

polyhydroxyalkanoate and starburst polyamidoamine dendrimer. J. Biosci. Bioeng. 95: 541-

543.

Ward, A.C., Rowley, B.I. and Dawes, E.A. 1977. Effect of oxygen and nitrogen limitation on

poly-β-hydroxybutyrate biosynthesis in ammonium-grown Azotobacter beijerinckii. J. Gen.

Microbiol. 102: 61-68.

Williams, S. F., Martin D. P., Gerngross T., Horowi tz D. M.; Metabolix, Inc. Cambridge,

MA, 2001. Removing endotoxin with an oxidizing agent from polyhydroxyalkanoates

produced by fermentation. US patent 6245537.

Williams, S. F., Martin D. P., Horowitz D. M., Peop les O. P. 1999a. PHA applications:

Addressing the price performance issue I. Tissue engineering. Int. J. Biol. Macromol. 25:

111-121.

Witholt, B., Kessler B. 1999a. Perspectives of medium-chain-length

poly(hydroxyalkanoates), a versatile set of bacterial bioplastics. Curr. Opin. Biotechnol. 10:

279-285.

Yagmurlu, M. F., Korkusuz F., Gursel I., Korkusuz P ., Ors U., Hasirci V. 1999.

Sulbactam-cefoperazone polyhydroxybutyrate-co-hydroxyvalerate (PHBV) local antibiotic

delivery system: In vivo effectiveness and biocompatibility in the treatment of implant-related

experimental osteomyelitis. J. Biomed. Mater. Res. 46: 494-503.

References 140

Yamamoto, O., Myata Y., Yanagi S.; Denki Kagaku Kogyo Kk, Japan. 1995. Separation and

purification of poly(hydroxyalkanoates) from microorganisms using surfactants. JP patent

07079787.

Yan, Q., Du G. C., Chen J. 2003. Biosynthesis of polyhydroxyalkanoates (PHAs) with

continuous feeding of mixed organic acids as carbon sources by Ralstonia eutropha.

Process Biochem. 39: 387-391.

Yasotha, K., Aroua M. K., Ramachandran K. B., Tan I . K. P. 2006. Recovery of medium-

chain-length polyhydroxyalkanoates (PHAs) through enzymatic digestion treatments and

ultrafiltration. Biochem. Eng. J. 30: 260-268.

Yezza, A., Fournier D., Halasz A., Hawari J. 2006. Production of polyhydroxyalkanoates

from methanol by a new methylotrophic bacterium Methylobacterium sp. GW2. Appl.

Microbiol. Biot. 73: 211-218.

Yu, J., Chen L. X. L. 2006. Cost-effective recovery and purification of polyhydroxyalkanoates

by selective dissolution of cell mass. Biotechnol. Prog. 22: 547-553.

Yuasa, J. 2002. Pseudomonas aeruginosa under phosphorus-poor culture conditions.

Microb. Environ. 17: 75-81.

Zhang, J. P., Qun Wang T. R. S., William E. Hurst, and, Sulpizio T. 2005. Endotoxin

removal using a synthetic adsorbent of crystalline calcium silicate hydrate. Biotechnol. Progr.

21: 1220-1225.

Zinn, M., Hany R. 2005. Tailored material properties of polyhydroxyalkanoates through

biosynthesis and chemical modification. Adv. Eng. Mat. 7: 408-411.

Zinn, M., Weilenmann H. U., Hany R., Schmid M., Egl i T. 2003. Tailored synthesis of

poly([R]-3-hydroxybutyrate-co-3-hydroxyvalerate) (PHB/HV) in Ralstonia eutropha DSM 428.

Acta Biotechnol. 23: 309-316.

Zinn, M., Witholt B., Egli T. 2001. Occurrence, synthesis and medical application of

bacterial polyhydroxyalkanoate. Adv. Drug Del. Rev. 53: 5-21.

141

Curriculum Vitae

1976, 19. January Born in Bern, Switzerland

1991-1996 Grammar school, Bern

1996 Matura type C

1996-1999 Chemistry studies at ETH-Lausanne

1999-2002 Chemistry studies at university of Basel

2002 Diploma in chemistry, with physics as minor subject

2002-2003 Trainee in polymer analytics, Ciba specialty chemicals

Basel

2003-2007 PhD studies in bioprocess engineering at Empa St. Gallen

and ETH Zurich

2008 Project manager at Sika Technology