microbial cell factories

7/21/2019 microbial cell factories

1/36

Microbial Cell Factories

Review

Yeast cell factories for fine chemical and API productionBeate Pscheidt1 and Anton Glieder*1,2

Address: 1Research Centre Applied Biocatalysis GmbH, Petersgasse 14/3, 8010 Graz, Austria and 2Institute of Molecular Biotechnology,

Graz University of Technology, Petersgasse 14/2, 8010 Graz, Austria

E-mail: Beate Pscheidt [email protected]; Anton Glieder* [email protected]

*Corresponding author

Published: 07 August 2008 Received: 26 May 2008

Microbial Cell Factories 2008, 7 :25 doi: 10.1186/1475-2859-7-25 Accepted: 7 August 2008

This article is available from: http://www.microbialcellfactories.com/content/7/1/25

2008 Pscheidt and Glieder; licensee BioMed Central Ltd.

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0 ),

which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

This review gives an overview of different yeast strains and enzyme classes involved in yeast whole-

cell biotransformations. A focus was put on the synthesis of compounds for fine chemical and API

(= active pharmaceutical ingredient) production employing single or only few-step enzymatic

reactions. Accounting for recent success stories in metabolic engineering, the construction and use

of synthetic pathways was also highlighted. Examples from academia and industry and advances in

the field of designed yeast strain construction demonstrate the broad significance of yeast whole-cell

applications. In addition to Saccharomyces cerevisiae, alternative yeast whole-cell biocatalysts are discussed

such as Candida sp., Cryptococcus sp., Geotrichum sp., Issatchenkia sp., Kloeckera sp., Kluyveromyces sp.,

Pichiasp. (includingHansenula polymorpha = P. angusta), Rhodotorulasp.,Rhodosporidiumsp., alternative

Saccharomyces sp., Schizosaccharomyces pombe, Torulopsis sp., Trichosporon sp., Trigonopsis variabilis,

Yarrowia lipolyticaandZygosaccharomyces rouxii.

BackgroundThe advent of yeast whole-cell biocatalysis coincided withthe development of first technologies for the humansociety. Some thousand years BC, when organizedagriculture had appeared, the development of yeast-basedtechnology started. Today, artistic records (e.g. in tombs ofwealthier members of the ancient Egyptian society) andarchaeological data provide an insight into the daily life inancient Egypt [1,2] and teach us that bread and also beer

were central parts of the Egyptians diet [3]. However, nomore than approximately 200 years ago, leading scientists

recognized yeasts as the cause of fermentation and startedto examine them because of their economic importanceand one morphologic advantage compared to some othermicroorganisms, their large cells [4]. Understanding thescientific basis for alcoholic fermentation was financiallywell supported by the alcoholic fermentation industriesand governments at that time. Basically, research on yeasts

greatly contributed to the development of microbiology,biochemistry and also biocatalysis. Berthelots and EmilFischers research results on the utilization of differentsugars by yeasts were central to studies on enzymes andtheir specificity. Especially Fischers lock and key modelpublished in 1894 [5] provided the basis for subsequentconcepts. These include for example the theory about theenzyme-substrate complex developed by Henri [6] andMichaelis & Menten [7], respectively, Haldanes concept of

substrate activation representing the idea of selectivebinding energy which leads either to transition statestabilization or substrate destabilization [8], and theinduced fit theory described by Koshland [9,10]. Earlyyeast research, which was comprehensively reviewed byBarnett [4], also contributed to basic biochemical knowl-edge including the understanding of metabolic pathways,Monods concepts of enzyme induction[11] and basicfindings on cell cycles [12,13].

Page 1 of 36(page number not for citation purposes)

BioMedCentral

Open Acces
mailto:[email protected]:[email protected]://creativecommons.org/licenses/by/2.0http://www.biomedcentral.com/http://www.biomedcentral.com/http://www.biomedcentral.com/http://www.biomedcentral.com/http://www.biomedcentral.com/info/about/charter/http://www.biomedcentral.com/info/about/charter/http://www.biomedcentral.com/http://creativecommons.org/licenses/by/2.0mailto:[email protected]:[email protected]


2/36

Most of this early research was based on Saccharomycescerevisiaeand therefore, the term yeastwas and is oftentaken as a synonym for S. cerevisiae. However, yeastsbelong to a group of eukaryotic microorganisms,predominantly unicellular and phylogenetically quite

diverse [14]. They are assigned to two taxonomic classesof fungi, the ascomycetes and the basidiomycetes[14,15]. Their classification is mostly based on pheno-typic characters such as morphology, the ability to utilizevarious exogenous compounds and modes of vegetativereproduction, namely budding or fission. Some yeastsalso form sexual states which differ from those of otherfungi as they are not enclosed in fruiting bodies [15]. Atthe end of the 20th century molecular methods havebecome increasingly popular in order to estimate geneticrelation among yeasts. Recently, Hibbett et al. [16]attempted to create a consensus higher-level classifica-tion for the Fungi for general use. This broad-based

consensus classification was necessary in order toprevent confusion and loss of information as caused byrepetitive renaming of several yeast strains in the past[17]. An overview of yeast genera was provided byWalker [18] or Boekhout and Kurtzman [15,19], forexample. A comprehensive phylogenetic relationshipamong sequenced fungal genomes, including morethan 20 yeast genomes, was recently depicted byScannell et al. [20].

Although yeasts, especiallySaccharomyces cerevisiae, havebeen employed in synthetic organic chemistry since thebeginning of the 20th century [21-23], scientists devoted

to classical organic chemistry often hesitated to considerbiological systems for their synthetic problems [24]. Inthe 1970s, biocatalysis started booming and up to nowthe number of publications on biotransformations hasbeen rising exponentially. Remarkable findings led to abetter understanding of biological systems and conse-quently to their increased application for chemicalconversions, especially in the field of organic synthesis[25].

In general, two major synthetic technologies based onbiocatalytic reactions were described, namely fermenta-tion and enzymation (alternative names for enzyma-

tion:

microbial transformation

,

microbial conversion

,biotransformation , bioconversion) [24,25]. Fermenta-

tion was considered to be a biological method resultingin products which are the result of the complexmetabolism of microorganisms starting with inexpensivesimple carbon and nitrogen sources. As a consequence,living or even growing cells were a prerequisite for thistechnology and fermentation was regarded to alwaysresult in natural products [25]. On the other hand,enzymation was specified not to necessarily requireliving cells, as cells were only important for the enzyme s

production and were themselves regarded as simple bagof enzymes or catalysts [25]. Furthermore, enzymationwas described to be a one or few-step conversion of amore complex substrate into a product [24,25]. How-ever, Yamada and Shimizu [25] already stated that it was

not always possible to clearly distinguish between thetwo categories.

In recent years, the way for the generation of designedmicroorganisms was paved by an increasing number ofsequenced genes and even whole genomes, new bioin-formatic tools providing the basis for analyzing thiswealth of information, biochemically well-characterizedbiosynthetic pathways and well-established and facilegenetic engineering techniques. These approachesinclude for example, the construction of syntheticpathways for the production of structurally complex,natural products like isoprenoids or polyketides and

novel variations thereof. Furthermore, even minimumgenome factories could enter the field of biocatalysis infuture. Currently, first examples of such cell factories arecreated in which unnecessary or harmful genes aredeleted and only genes necessary for industrial produc-tion are present [26]. For this approach, up to now, threespecies were selected, namely two bacteria (Escherichiacoli [27] and Bacillus subtilis [28]), and the fission yeastSchizosaccharomyces pombe[29]. With these developmentsin mind, one has to admit that the idea of whole-cellbiocatalysts being black boxes already started to fade.

The following chapters will now give an overview of

different yeast strains and enzyme classes involved inyeast whole-cell biocatalysis. We will provide examplesfrom academia and industry and especially focus onrecent advances in the field of designed yeast strains forwhole-cell biocatalytic applications. Thereby, we willalso include synthetic pathways to structurally complexcompounds, a methodology which lies in betweenclassical fermentation and enzymation. Thus, we willinclude both, one- and multi-step enzymatic reactions ina native or engineered environment, starting with simpleor complex substrate molecules.

Review1. Chemical reactions catalyzed by wild-typeyeast whole-cell biocatalysts

Especially bakers yeast (= Saccharomyces cerevisiae) wasregarded to be ideal for chemists looking for a stereo-selective biocatalyst, which should eventually lead tochiral intermediates in the synthesis of enantiomericallypure compounds [30]. It is nonpathogenic, inexpensive,simple to grow at laboratory and large scale and the cellscan be stored indefinitely in dried form [24]. Forchemical synthesis, however, the chemical repertoire of

Microbial Cell Factories2008, 7:25 http://www.microbialcellfactories.com/content/7/1/25



3/36

yeast whole-cell biocatalysts is of major importance. Inthe following, an overview is given, outlining the mainenzymatic reactions performed by wild-type yeaststrains, and positive and negative aspects of thesewhole-cell biocatalysts are discussed.

1.1 Reduction of C=O-bonds

The asymmetric reduction of carbonyl-containing com-pounds by yeast, in particular Saccharomyces cerevisiae,depicts probably the most thoroughly investigated classof whole-cell biotransformations. One of the first reportson this topic was the reduction of furfural to furfurylalcohol by Windisch [31] and Lintner [32] at thebeginning of the 20th century. The first comprehensiveoverview of reduction reactions catalyzed by yeast waspublished in 1949 [33]. Since that time, many differentsubstrates containing carbonyl moieties were subjected

to yeast bioreduction and the most important achieve-ments were summarized in reviews and book chapters,partly focusing onSaccharomyces cerevisiae[30,34,35] butalso on biocatalysts in general, including alternativeyeasts [24,25,36-43]. The investigated substrate spectrumis huge, including a variety of functional groups assubstituents of the ketone moiety (for example hetero-cyclic-, hydroxyl-, sulfur-, cyano-, and azido-groups, ordifferent halogenides) and even derivatives such as silyl-or germyl-groups were found to be accepted [24].Generally, Saccharomyces cerevisiae reduces simple ali-phatic and aromatic ketones according to Prelogs rule[44] resulting in the corresponding (S)-alcohols [45]

(Figure 1). However, this should not always be general-ized and caution should be exercised in particular, whenPrelogs rule is applied to whole cells [34].

A major advantage of whole-cell (redox)-biocatalysts isthe availability of all the necessary cofactors andmetabolic pathways for their regeneration. Furthermore,

cheap carbon sources (e.g. glucose, saccharose) can beemployed as auxiliary substrates. Finally, the actualbiocatalyst and cofactors are well protected within theirnatural cellular environment which makes the catalyticsystem more stable [24]. However, employing wild-type

yeast strains as whole-cell biocatalysts also includesdistinct drawbacks: Most of the interesting substrates arenon-natural and toxic to living organisms. Therefore,they must be used in diluted systems at low concentra-tions (mostly 0.3% per volume) [24,46]. Only smallfractions of the auxiliary substrate are used for cofactorrecycling, the majority is metabolized. This results inlarge amounts of biomass and by-products whichimpede product recovery tremendously, especially ifthe product is not secreted to the reaction medium.Transport phenomena into and out of the cell could beencountered which may even influence specificity[24,46]. Eventually, comparable results are only possible

if exactly the same culture is used for repetitivebiotransformations since different strains of the samemicroorganism could have different specificities [24,46].Another major limitation of wild-type yeast strains is thepresence of a large number of different dehydrogenaseswith overlapping substrate specificities but oppositestereoselectivities [47]. The elucidation of the completegenome sequence of Saccharomyces cerevisiae [48,49]finally made scientists aware of the huge variety ofavailable oxidoreductases. Earlier, empirical findingshelped to develop methods to improve the selectivityof yeast whole-cell biotransformations. These techniquesincluded substrate modification [50], changes in cultiva-

tion conditions or the application of different carbonsources [51], the use of inhibitors [52] or the use of two-phase systems [53,54]. Recently, also water immiscibleionic liquids were used as biocompatible solvents foryeast whole-cell biocatalysis in order to provide asubstrate reservoir and an in situ extracting agent toincrease chemical yields [55].

Figure 1Asymmetric reduction of ketones according to Prelogs Rule[44].




4/36

If these techniques were not successful, alternativemicroorganisms had to be screened. In addition toSaccharomyces cerevisiaealso alternative yeasts were foundwhich provided valuable biocatalysts for asymmetriccarbonyl reduction. Some were efficient enough to be

employed for industrial applications by different com-panies (Table 2and Additional File2, Section 2).

Some recent advances in the biocatalytic reduction ofcarbonyl-containing compounds are summarized inTable 1(Additional file 1) and include for example thefinding of new activities forYarrowia lipolytica[56]. Lagoset al. [56] screened different yeast strains for theenantioselective production of a halohydrin precursorfor (S)-propranolol synthesis. Yarrowia lipolytica 1240(Spanish type culture collection CECT, Valencia) restingcells gave the anti-Prelog-enantiomer (S)-1-chloro-3-(1-naphthyloxy)propan-2-ol with 87% yield and 99% ee. In

addition,Pichia mexicana11015 (CECT, Valencia) restingcells were found to give 85% yield and 95% ee for the(R)-enantiomer. Dehli and Gotor [57] discovered thatresting cells of Saccharomyces montanus CBS 6772performed the bioreduction of 2-oxo cyclopentanecarbonitriles to the corresponding cis-hydroxynitriles in93% ee, and high de and chemical yields.

As cheap alternative, Saccharomyces cerevisiaeis still quiteoften used for laboratory-scale bioreductions. Enderset al. [58] described the efficient asymmetric totalsynthesis of (-)-callystatin A, a potent cytotoxic polyke-tide from a marine sponge, employing a combination of

chemical and biocatalytic methods. Therefore, tert-butyl6-chloro-3,5-dioxohexanoate was subjected to driedbakers yeast in a biphasic system (water/XAD-7 adsorberresin) and was regio- and enantioselectively reduced tothe (5R)-hydroxy keto ester in 94% enantiomeric excess[58]. Furthermore, Bertau and Burli [59] reported theapplication of Saccharomyces cerevisiae cells for thesynthesis of (2S,5S)-hexanediol. Glucose was used asauxiliary substrate for cofactor regeneration and with25 mmol of hexanedione as substrate complete conver-sion with > 99% ee, 96% de and 75% yield was achieved.Servi and coworkers [60] showed that whole cells ofGeotrichum candidumCBS 233.76 (95% yield, > 98% ee;

4 g/L, with absorbing resin Amberlite XAD-1180) andRhodotorula mucillaginosa CBS 2378 (88% yield, > 99%ee; 1 g/L) reduced 3,4-dichlorophenacyl chloride to the(S)- and (R)-alcohol, respectively.

Matsuyama et al. [61] screened several yeast strains forthe large scale production of (R)-1,3-butanediol, as theproductivity of already reported asymmetric microbialreduction processes (for example with S. cerevisiae) wasnot satisfactory. Starting with 4-hydroxy-2-butanone,Candida arborea IAM 4147 and Issatchenkia scutulata IFO

10070 showed excellent ee (99%) but low yield, whereasKluyveromyces lactis IFO 1267 gave high yield and goodenantiomeric excess (93%) for (R)-1,3-butanediol. Incontrast, Candida parapsilosis IFO 1396 performed thebest for (S)-1,3-butanediol production (98% ee, 60%

yield). In recent times, a new yeast isolate, namelyCandida tropicalis PBR-2 MTCC 5158 was reported toenantioselectively reduce acetophenone and severalsubstituted analogues to the corresponding (S)-alcoholswith an enantiomeric excess of > 97%, mostly even >99% [62]. As the natural aroma compounds 2-pheny-lethanol (2-PE) and 2-phenylethylacetate (2-PEAc) are ofhigh industrial value, Etschmann and Schrader [63]lately improved their production by employingKluyver-omyces marxianus CBS 600 for a growth-associatedproduct formation starting with L-phenylalanine. 26.5g L1 2-PE and 6.1 g L1 2-PEAc were obtained in theorganic phase which consisted of polypropylene glycol

1200, used as in situ extractant. This corresponded tospace-time yields of 0.33 (2-PE) and 0.08 g L1 h1

(2-PEAc) surpassing the results of a previously reportedS. cerevisiae process (Table 1; Additional File 1) [64].Finally, Nakamura et al. [41] reviewed the use ofGeotrichum candidum whole-cell preparations for thesynthesis of chiral secondary alcohols with high enan-tioselectivities (up to > 99% ee).

In addition to these and also future findings in the fieldof selective yeast carbonyl-bioreductions, a strong trendfor the construction of so-called designer-bugs namelywhol e-cell bi ocatalysts co-expressin g all requir ed

enzymes can be observed [65]. Most often E. coli isemployed as recombinant host [66-69] due to the wealthof oxidoreductases found in yeasts, in particular, inSaccharomyces cerevisiae. Nowadays, the increasing num-ber of annotated DNA sequences helps to identifyenzymes involved in bioreductions and provides theinformation needed for the rational design of recombi-nant whole-cell biocatalysts [47].

1.2 Reduction of C=C-bonds

In 1933, the first flavin-dependent redox enzyme wasdiscovered in Brewers bottom yeast by Warburg and

Christian [70]. Enzymes of this family are also known asold yellow enzymesbecause of their color derived fromthe flavin cofactor. Typical substrates [24,71] are alkeneswhich are activated by electron-withdrawing substitu-ents. Such substrates are reduced at the expense of NAD(P)H leading to enantiomerically pure alkanes. Therebyup to two chiral carbon centers are created (Figure 2).

In the past, most asymmetric bioreductions of activatedC=C bonds employing enoate reductases were per-formed with whole cells, since problems with external


http://-/?-http://-/?-


5/36

Table 1: Examples of wild-type yeast whole-cell biocatalysts for the reduction of C=O bonds and comparison to S. cerevisiae, if pro

Whole-Cell Biocatalyst Substrate Product Perfor

Candida parapsilosis IFO 1396CH3 OH

O

4-hydroxy-2-butanone

CH3 OH

OH

(S)-1,3-butanediol

C.p

CH3 OH

OH

(R)-1,3-butanediol

Candida arborea IAM 4147 C.aIssatchenkia scutulata IFO 10070 I.s.Kluyveromyces lactis IFO 1267 K.l

Candida tropicalis PBR-2 MTCC 5158 S

O

NCH3

CH3

N,N-dimethyl-3-keto-3-(2-thienyl)-1-propanamine

S

OH

NCH3

CH3

(S)-N,N-dimethyl-3-hydroxy-3-(2-thienyl)-1-propanamine

C.t.: > 84

Geotrichum candidum CBS 233.76 withAmberlite XAD-1180

O

Cl

Cl

Cl

3,4-dichloro--chloroacetophenone

OH

Cl

Cl

Cl

S

G.c.:

OH

Cl

Cl

Cl

R

Rhodotorula mucillaginosa CBS 2378 R.m.

Saccharomyces cerevisiaeb S.c

Saccharomyces montanus CBS 6772 O

CN

2-oxo cyclopentanecarbonitrile

OH

CN

2-hydroxy cyclo-pentane carbonitrile

S.m.: eci

Saccharomyces cerevisiae(Type II from Sigma) S.c.: eci

MicrobialCellFactories2008,

7:25

http://www.microbialcellfactories.com/content/7/1/25


6/36

Table 1: Examples of wild-type yeast whole-cell biocatalysts for the reduction of C=O bonds and comparison to S. cerevisiae, if pro

Saccharomyces cerevisiae(dried baker's yeast)CO2tBu

O

O

Cl

CO2tBu

OH

O

Cl

S

Saccharomyces cerevisiaeCH3

CH3

O

O

2,5-hexanedione

CH3

CH3

OH

OH

(2S,5S)-hexanediol

S.c.: 75

Pichia mexicana CECTc 11015

O

O Cl

1-chloro-3-(1-naphthyloxy)propan-2-one

O

OHH Cl

(R)-1-chloro-3-(1-naphthyloxy)propan-2-ol

P.m.:

Saccharomyces cerevisiae CECTc 1317 S.c

O

HOH Cl

(S)-1-chloro-3-(1-naphthyloxy)propan-2-ol

Yarrowia lipolytica CECTc 1240 Y.l.: 8

Saccharomyces cerevisiae (Type II, Sigma) S.c

Kluyveromyces marxianus CBS 600

COOH

NH2

L-phenylalanine

OH

2-phenylethanol(2-PE)

O O

CH3

2-phenylethylacetate(2-PEAc)

+

K.m.:2-PE: 2STY:

2-PEAc: 6STY

Saccharomyces cerevisiae GIV 2009

d

S.c.:2-PE: 24.0 g STY

aconv. = conversion; bbaker's yeast from Distillerie Italiane, Eridania group; cSpanish type culture collection CECT, Valencia; dThe wild type strGivaudan Ltd. (Dbendorf, Switzerland); eorganic phase = Polypropylene glycol 1200 [63]; fSTY = space-time yield; calculated from t= 0 until product concentration was reached; gorganic phase = oleic acid [64].


7:25



7/36

cofactor recycling and the enzymes sensitivity to tracesof oxygen were encountered [24,72]. With whole-cellbioreductions often excellent stereoselectivities wereachieved. However, chemoselectivity is often poor,regarding especially competitive C=C- and C=O-bondreductions. Whole-cells do not only provide enoatereductases but also alcohol dehydrogenases which bothdepend on the same nicotinamide cofactor. Uncouplingis hardly possible and the relative rates of alcoholdehydrogenases are comparable to those of enoatereductases leading to undesired by-products [71]. Never-theless, successful examples for asymmetric bioreduc-

tions ofa,b-unsaturated ketones exist, for example, withS. cerevisiae cells: Careful reaction control duringoxoisophorone bioreduction led to the main product((R)-2,2,6-trimethylcyclohexane-1,4-dione) in > 80%yie ld. The unw ant ed by- pro ducts wer e kept to aminimum. The product also known as (R)-levodione was produced on a 13 kg-scale [73]. It is industrially usedfor 3-hydroxycarotenoid production [74] (Table 2 andAdditional File2).

In general, the most prominent whole-cell biocatalystemployed was again bakers yeast (S. cerevisiae) whichwas tested with a great variety of differently substituted

alkene substrates, such as e.g. differenta,b-unsaturatednitroalkenes. It tolerated many different functionalgroups like alkyls and (substituted) aryls attached tothe nitroalkene moiety [75,76]. However, also alterna-tive yeasts were described to be active. Geotrichumcandidum[77],Rhodotorula rubra[78] andRhodosporidiumsp. [78] were found to possess enoate reductases, activefor diverse non-natural a,b-unsaturated carbonyl- andcarboxyl-compounds, respectively [71]. Candida sp.,Rhodotorula sp. and Torulopsis sp. were also shown to beactive on a,b-unsaturated nitroalkenes [79]. Currently,

the use of isolated enzymes or the construction ofdesigner bugs provides the possibility to reduce sidereactions significantly. Desired yeast enoate reductases(e.g. from Saccharomyces cerevisiae [80], Saccharomycescarlsbergensis [81] or Candida macedoniensis [82]) were

recently cloned and overexpressed in E. coli. Kataokaet al. [82] even coexpressed glucose dehydrogenase(GDH) for efficient cofactor regeneration. More detailswere outlined and reviewed by Faber [24] and Stuermeret al. [71], respectively.

1.3 Oxidation and racemization reactions

Yeast alcohol oxidases were shown to be responsible forthe oxidation of methanol and other primary alcohols[83-85]. However, they did not oxidize secondaryalcohols. In general, reports on oxidation reactionsperformed by yeasts are quite rare. Considering espe-

cially the oxidation of secondary alcohols, one has torecognize that instead of creating a chiral center, it isdestructed. The reaction is therefore regarded to be oflimited synthetic use except for the regioselectiveoxidation of polyols where chemical methods are ofteninadequate [24]. However, the recent, increasingdemand for deracemization processes, leading to asingle stereoisomer in 100% yield, evoked the need forclean racemization protocols [86]. Not to forget, theoxidation of sulfides can result in chiral sulfoxides whichhave been widely used in organic synthesis as asym-metric auxiliary groups to control the stereochemicaloutcome of the reaction at nearby centers [87,88].

In 1979, Patel et al. [89] described the application of cellsuspensions of Candida utilis ATCC 26387, Hansenulapolymorpha ATCC 26012, Pichia sp. NRRL-Y-11328,Torulopsis sp. and Kloeckera sp. for the oxidation ofsecondary alcohols. 2-Propanol, 2-butanol, 2-pentanol,and 2-hexanol were oxidized to the correspondingmethyl ketones. Furthermore, Saccharomyces cerevisiaewas reported to catalyze the selective oxidation ofsulfides to sulfoxides. Beecher et al. [90] investigatedthe oxidation ofp-tolyl sulfide to the R-sulfoxide (92%ee) which was used for the preparation of the mevinicacid-type hypocholestemic agent. Buist et al. [91]

employed bakers yeast for the enantioselective sulfox-idation of a fatty acid analogue (~70% ee). In this study,

the authors presumed that a desaturase could beresponsible for the transformation, but did not verify it.

In order to satisfy needs for sophisticated racemizationprotocols, Nestl et al. [86,92] screened various microbialcells for the biocatalytic racemization of functionalizeda-hydroxyketones. Although whole lyophilized cells ofGeotrichum candidum DSM 6401, Candida parapsilosisDSM 70125, or Kluyveromyces lactis DSM 3795 were

Figure 2Enoate reductases perform the NAD(P)H-dependent,asymmetric bioreduction of activated C=C bonds.A cofactor recycling system is required for an economicprocess. Asterisks (*) indicate chiral centers, X depicts anactivating group such as carbonyl-, carboxyl-, imide- ornitro-group [24,71].


http://-/?-http://-/?-


8/36

Table 2: Industrial biotransformations [74] employing wild type yeast whole-cell biocatalysts ordered by E.C. numbers

Yeast strain Enzyme name(E.C. No.)

Reaction Product

Zygosaccharomyces rouxii Alcohol NAD+

oxidoreductase(1.1.1.1)

O

O

CH3

O

O

O

CH3

OH

E

suspended whole cells

3,4-methylenedioxyphenylacetone (S)-(3,4-methylenedioxyphenyl)-

2-propanol

LY 500164 = Befor treating am

sc

Geotrichum candidum Dehydrogenase,NADPH-dependent(1.1.X.X)

OCl

O O

CH3

E


+ NADPH, H+

OCl

OH O

CH3

4-chloro-3-oxo-butanoic acidmethyl ester

(S)-4-chloro-3-hydroxy-butanoic acid methyl ester

chiralb-hydroxycholesterol an

HMG-Co

Candida sorbophila Dehydrogenase(1.1.X.X)

N

O

NH

ONO2

E


N

OH

NH

ONO2

2-(4-nitro-phenyl)-N-(2-oxo-2-pyridin-3-yl-ethyl)-acetamide

(R)-N-(2-hydroxy-2-pyridin-3-yl-ethyl)-2-(4-nitro-phenyl)-acetamide

(R)-amino alcohob-3-agonist

therapy, ahypertension

disease and thetype I

Pichia methanolica Reductase (1.1.X.X)OEt

O

CH3

OE

whole cells OEt

O

CH3

OH

ethyl 5-oxo-hexanoate ethyl 5-(S)-hydroxyhexanoate

CNCH3

OE

whole cellsCN

CH3

OH

5-oxohexanenitrile 5-(S)-hydroxyhexanenitrile

Ethyl-5-(S)-hydr(S)-hydroxyhe

build

Candida boidinii(E2) [andE. coliwith cloned E1]a

or Pichia pastoris (E2 +cloned E1)

b

E2: FDH (1.2.1.2)[E1: PheDH(1.4.1.20) fromThermoactinomycesintermedius]

O

OHOOC

O O

OHOOC

NH2E1

E2

NADH NAD+

CO2 HCO

2NH

4

5-(1,3-dioxolan-2-yl)-2-oxopentanoic acid

(S)-2-amino-5-(1,3-dioxolan-2-yl)-pentanoic acid

(S)-2-Amino-5-pentanoic acid/

for OmapatrilatNEP i

Trigonopsis variabilisATCC 10679

D-Amino acidoxidase (1.4.3.3)

OH

NH2

O

OHE1

O2

OH

O

O

O-

OH

NH2

O

OH

+

(rac)-6-hydroxynorleucineL-6-hydroxynorleucine

Na+

L-6-Hydroxyintermediate

vasopeptidase antihypertensiv

inh


7:25



9/36

Table 2: Industrial biotransformations [74] employing wild type yeast whole-cell biocatalysts ordered by E.C. numbers (Continued)

Baker's yeast (=Saccharomyces cerevisiae)

Reductase (1.X.X.X)

O

O

CH3

CH3CH3

O

O

CH3

CH3CH3

+ NADH, H+

- NAD+

E

oxoisophorone (R)-2,2,6-trimethylcyclo-hexane-1,4-dione


The correspointermediate

natural 3-hy(e.g. cryptoxant

other terpen

Cryptococcus laurentii(E1) [and Achromobacterobae (E2)]a

E1: Lactamase(3.5.2.11) and [E2:Racemase(5.1.1.15)]a

NH

O

NH2

2

NH

O

NH2E1

suspended whole cells +NH2

COOH

NH2D,L--amino--

caprolactamD--amino--caprolactam

L-lysine

E2


L-Lysine/nusup

Saccharomyces cerevisiae Pyruvatedecarboxylase(4.1.1.1)

HO

+OmolassesE

whole cells

OHCH3

O

yeast

acetaldehyde benzaldehyde (1R)-phenylacetylcarbinol

(PAC)

PAC pseudoephed

asthma, hay fbronchodil

deco

Candida rugosa Enoyl-CoA hydratase(4.2.1.17)

COOHE

H2O

COOH

OH

butyric acid (R)--hydroxy-n-butyric acid

whole cells

COOH

CH3E

H2O

COOH

CH3

OH

isobutyric acid (R)--hydroxy-isobutyric acid

whole cells

(R)-b-Hydroxy(R)-b-Hydroxy-

synthons fintermedia

(ACE-inhibit

Rhodotorula rubra L-Phenylalanineammonia-lyase(4.3.1.5)

COOH

+ NH3E

suspendedwhole cells

COOH

NH2

trans-cinnamic acid ammonia L-phenylalanine

L-phenylalanine/aspartame, pa

chiral building bB s

aBacterial whole-cell biocatalyst; bIn this case, also engineered P. pastoris cells were employed; cACE = angiotensin-converting enzyme; NEP = n


7:25



10/36

identified to racemize a variety of the employed targetsubstrates, the activities of yeast strains were modestcompared with bacteria and fungi. Traces of diketones asside products let them assume that dehydrogenaseenzymes were responsible for the racemization reaction

[86].

Matsuyama et al. [61] performed the oxidative kineticresolution for the large-scale production of (R)-1,3-butanediol employing Candida parapsilosis IFO 1396.Starting with 20 kg (rac)-1,3-butanediol, 258 kg cells,465 kg water and 7.5 kg calcium carbonate resulted in~3.1 kg of (R)-1,3 butanediol in high chemical (~99%)and optical purity (94% ee). A subsequent enzymepurification step elucidated (S)-1,3-butanediol dehydro-genase (CpSADH) to be responsible for producing (R)-1,3-butanediol from the racemate. CpSADH was thenalso overexpressed inE. coliyielding in increased specific

activities and a strain with good racemization propertiesfor the production of (R)-1,3-butanediol (95% yield,94% ee) without the need for cofactor regeneration.Recently, Titu and Chadha [93] described the biocataly-tic preparation of optically pure alkyl 3-(hetero-2-yl)-3-hydroxypropanoates by deracemization. Candida para-psilosis ATCC 7330 was employed and various opticallypure products resulted in high ee (8999%) and yields(5875%). These compounds represent important chiralintermediates for the synthesis of pharmaceuticals suchas duloxetine, tetrahydropyrans or heteroarylaminoalk-anols, among others.

In most cases, however, the enzyme or enzyme-setcatalyzing racemization reactions remains unknown.Their identification and employment in a non-naturalenvironment (biotransformation with solubilized orimmobilized enzyme(s) or a recombinant whole-cellbiocatalyst) could further reduce the amount of unde-sired side-reactions, improve optical purities and alsoprovide a platform for enzyme engineering approaches.Examples ofyeast designer bugswith enzymes for chiralsulfoxide production are given in Section 3.

1.4 Hydrolase reactions

Saccharomyces cerevisiae was also tested for hydrolasereactions. However, first yeast-catalyzed deacylationreactions were regarded to be undesired side-reactionswhen Mamoli et al. [94] was studying whole cells ofbakers yeast for the stereoselective reduction of estrogenester. Later, the whole variety of hydrolysis reactionsperformed by Saccharomyces cerevisiae was investigatedand comprehensive reviews on the properties of protei-nases [95], lipases and esterases [34] were given. Whatwas encountered quite fast was the complex reactioncontrol when whole microbial cells were employed [24].

To overcome the problems resulting from the metabo-lism of fermentingSaccharomyces cerevisiae, lyophilizedresting cells were employed. Thus, for example theacetate of pantolactone the chiral precursor for vitaminB5 synthesis was nicely resolved [96]. In addition,

differently substituted 1-alkyn-3-yl acetates were con-verted with high ee (9196%) to the corresponding (S)-alcohols and acceptable enantioselectivities (E ~46100)were reached [24].

Many yeast hydrolases especially from Candida sp. wereemployed in industrial biotransformations. However,they were mostly used as immobilized enzyme prepara-tions due to perturbing unspecific hydrolases availablein whole cells. Only were other fungal and somebacterial strains industrially employed as whole cellsfor hydrolase reactions [74]. Immobilized whole cells ofFusarium oxysporum were for example used for the

production of D-pantoic acid starting with (rac)-panto-lactone [97]. Suspended whole cells of Comamonasacidovorans catalyzed the conversion of (rac)-2,2-dimethylcyclopropanecarboxamide to the corresponding(R)-carboxylic acid and the (S)-carboxamide remained[98]. In addition, whole cells ofBacillus brevis [99,100],Pseudomonas sp. [74] and Arthrobacter sp. [101] wereemployed i.a. for hydantoinase catalyzed reactions.

Furthermore, Rhodotorula sp., Rhodosporidium sp. andTrichosporon sp. were found to give best enantioselec-tivities for the hydrolysis of monosubstituted oxiranes.All showed preferred enantiopreference for the (R)-form

and regioselectivity for the sterically less hindered carbonatom [24]. Recently, again a highly selective epoxidehydrolase from Rhodosporidium paludigenum CBS 6565was observed during whole-cell biotransformations,subsequently cloned and further characterized [102].

1.5 Formation of C-C-bonds

From a synthetic point of view, carbon-carbon bondforming reactions are highly interesting as new asym-metric carbon centers can be formed. However, this typeof yeast whole-cell biotransformation was and is limitedto only a few examples [30,34]. The most prominent and

industrially employed acyloin condensation performedby Saccharomyces cerevisiae yields in the formation of(1R)-phenylacetyl carbinol, a chiral synthon of D-ephedrine (Table 2 and Additional File 2). Crout et al.[103] succeeded in finding out the detailed reactionpathway, involving pyruvate decarboxylase [104]. Notonly benzaldehyde can be subjected to bakers yeastmediated acyloin condensations. Fuganti et al. [105,106]also investigated the conversion of a,b-unsaturatedaldehydes giving optically active diols and used thistype of reaction for the production of the C-14


http://-/?-http://-/?-


11/36

chromanyl moiety ofa-tocopherol (vitamin E) [107]. Aninteresting L-threonine aldolase was isolated fromCandida humicola. However, only the isolated enzymewas characterized [108].

2. Application of yeast whole-cell biocatalysis in industry

One of the first industrial processes combining micro-biological and chemical synthesis was the acyloin-typecondensation [109] of benzaldehyde resulting in (1R)-phenylacetylcarbinol which is subsequently converted to(1R, 2S)-ephedrine and (1R, 2S)-pseudoephedrine,respectively. For that purpose, Saccharomyces cerevisiaewhole cells were used in an aqueous medium [110]. Theproducts find application for the treatment of asthma,hay fever and as a bronchodilating agent and deconge-stant [74,110].

In addition to this application, two more lyase-basedbiocatalytic approaches with whole cells were employedin industry: Candida rugosa enoyl-CoA hydratase cata-lyzes for example one of three biotransformation stepsfrom butyric acid to (R)-b-hydroxy-n-butyric acid[100,111], namely the enantioselective introduction ofa hydroxy-group at theb-position of thea,b-unsaturatedacid which resulted after dehydrogenation. The chiralproduct is produced with high optical purity (> 98% ee)and a space-time yield of 510 gL1d1. It is used forthe production of a carbapenem intermediate [74]. Thesame whole-cell biocatalyst is also capable of convertingisobutyric acid to (R)-b-hydroxy-isobutyric acid, in this

case with 98% yield, 97% ee and again a space-time yieldof 510 gL1d1 [100,111,112]. (R)-b-hydroxy-iso-butyric acid is a chiral synthon for the synthesis ofcaptopril, an inhibitor of angiotensin converting enzyme[74]. Secondly, suspended whole cells of Rhodotorularubrapossess an L-phenylalanine ammonia lyase whichperforms the selective transformation of trans-cinnamicacid and ammonia to L-phenylalanine. The reaction isperformed in aqueous solution at 25C and pH 10.6 andresults in 85.7% yield. The L-phenylalanine is commonlyapplied for the production of artificial sweeteners likeaspartame and in parenteral nutrition. It can also be usedfor the synthesis of macrolide antibiotics such as

rutamycin B [74,99,100].

Most yeasts which are industrially employed for whole-cell biocatalysis however are used for asymmetricreductions. Eli Lilly for example employed whole cellsof Zygosaccharomyces rouxii for the enantioselectivereduction of 3,4-methylenedioxy-phenylacetone to thecorresponding (S)-alcohol. Since the substrate was toxicto the microorganism, it was provided in an adsorbedform on the resin XAD-7. This allowed elevated reactionc o n ce n tr a ti o n s a n d h i gh l e v el s o f v o lu m et r ic

productivity. The process showed a space-time yield of75 gL1d1 and a high optical purity (> 99.9% ee)[113-116]. The strain Geotrichum candidum SC5469 wasdescribed to be employed by Bristol-Myers Squibb forthe stereoselective reduction of 4-chloro-3-oxo-butanoic

acid methyl ester to the corresponding (S)-3-hydroxy-alcohol [74]. This conversion yielded the product in 95%yield and 99% ee and thus provided a useful chiralbuilding block, e.g. for the synthesis of a cholesterolantagonist that inhibits hydroxymethyl glutaryl CoAreductase [117,118]. Furthermore, a second Candidasp.,namely Candida sorbophila, was employed for theproduction of a chiral (R)-amino alcohol (Entry 3 inTable 2 and Additional File 2) with high optical purity(> 98% ee; 99.8% after purification) and high chemicalpurity (95%) [119]. The resulting (R)-amino alcohol isan important chiral synthon for the synthesis of a b-3-agonist which can be used for obesity therapy and the

treatment of associated type II diabetes, coronary arterydisease and hypertension [74,120].

The yeastPichia methanolicawas applied by Bristol-MyersSquibb because of its beneficial reductase for thereduction of ethyl 5-oxo-hexanoate and 5-oxo-hexaneni-trile to the corresponding (S)-alcohols. The reactionresulted in high yield (9095%) and high enantiomericexcess (> 95%) [121].

Chiral intermediates for the production of the antihy-pertensive drug, Omapatrilat, were produced by twodifferent biocatalytic approaches (Entry 5 inTable 2and

Additional File 2 ): In both cases the enzyme phenyla-lanine dehydrogenase from Thermoactinomyces interme-dius (TiPheDH) was employed for the reductiveamination step and a yeast formate dehydrogenase(FDH) was required for NADH regeneration. For thefirst, heat-dried cells of recombinantE. coli with clonedTiPheDH were used together with heat-dried Candidaboidinii cells, as a source of FDH. The conversion resultedin 96% yield and > 99% ee. Secondly, dried recombinantPichia pastoris cells containing endogenous FDH andoverexpressing TiPheDH were employed. The applica-tion of the second biocatalyst resulted in slightlyimproved yield (97.5%) and also high enantiomeric

excess (> 99%) for the production of (S)-2-amino-5-(1,3-dioxolan-2-yl)-pentanoic acid [122]. Regardingcofactor regeneration, NAD+ produced during the reac-tion was in both cases regenerated to NADH by theoxidation of the auxiliary substrate formate to CO2[123].

A resolution technique is the basis for the whole-cell biocatalytic application of Trigonopsis variabilisATCC 10679 cells. (rac)-6-Hydroxynorleucine, whichresults after hydrolysis of commercially available


http://-/?-http://-/?-http://-/?-http://-/?-


12/36

5-(4-hydroxybutyl)-hydantoin, is converted to the ketoa-cid leaving the L-enantiomer. This is then isolated by ionexchange chromatography and can be used as chiralsynthon, again, for the synthesis of Omapatrilat [43], theproduction of a vasopeptidase inhibitor, and the

synthesis of C-7 substituted azepinones which representpotential intermediates for other antihypertensivemetalloproteinase inhibitors [74]. The remaining 2-keto-6-hydroxyhexanoic acid was further converted toL-6-hydroxynorleucine by reductive amination catalyzedby beef liver glutamate dehydrogenase (8992% yield, >99% ee, 3 h) [124]. Bacillus megaterium glucose dehy-drogenase was used for NAD+ recycling.

Furthermore, Saccharomyces cerevisiae was industriallyemployed because of its ene-reductase which belongsto the old yellow enzyme family. The substrate oxoiso-phorone was subjected to fermentative reduction with

bakers yeast in an aqueous solution of saccharose.Periodically, substrate and saccharose were added in

order to avoid toxic levels of oxoisophorone. Theproduct 2,2,6-trimethylcyclohexane-1,4-dione resultedin 80% yield and < 97% ee and was used as intermediatefor the synthesis of natural 3-hydroxycarotenoids such aszeaxanthin, cryptoxanthin and other structurally relatedcompounds. The space-time yield of the process wasgiven with 2.8 gL1d1 [73,74].

Finally, a hydrolase-based reaction employing wholecells ofCryptococcus laurentii was used by the Japanesecompany Toray Industries Inc. for the synthesis of

L-lysine [99,100,125]. For that purpose actually acombination of cells from Cryptococcus laurentii withintrinsic L-lysine lactamase and the bacterium Achromo-bacter obae with a lactame racemase were used. L-Lysineresulted with > 99.5% ee and was applied as nutrient andfood supplement. Alternatively, the combination of cellsfrom the yeast Candida humicola and the bacteriumAlcaligenes faecalis were used. However, this process wastotally replaced by highly efficient fermentation methods[74]. All above described industrial applications of yeastwhole-cell biocatalysts are depicted in Table 2 andAdditional File 2.

3. Recombinant Yeasts for Whole-Cell Biocatalysis

Remarkable progress in the field of genetic engineeringmade a variety of yeast strains, in particular Sacchar-omyces cerevisiae, susceptible to modifications by recom-binant DNA technology. In the mid-1980s, a number ofexpression systems based on alternative yeasts appearedwhich eliminated some initial limitations to S. cerevisiaesuch as low yields or a high extent of undesirablehyperglycosylation. Advantages and disadvantages ofsuch nonconventional yeasts for recombinant protein

production were reviewed in detail elsewhere [126-130].Prominent established examples are Hansenula polymor-pha [131], Pichia pastoris [132-134], Kluyveromyces lactis[135], Yarrowia lipolytica [136], Candida boidinii [128],and Schizosaccharomyces pombe [137], among others.

In order to circumvent disadvantages of wild-type whole-cell biocatalysts, different groups started to constructdesigner yeasts to improve first of all single-stepbiocatalytic syntheses. Methods for gene knockout andoverexpression of heterologous enzymes, offered alter-native approaches to classical techniques aiming atimproving the whole-cell biocatalysts selectivity (seeSection 2). Catalytic activities of competing host-ownenzymes were eliminated and desired catalytic propertiescould be reinforced, respectively. First, simple singlemutants with one deleted or overexpressed gene were tho rou ghly studied but soon combina tor ial

approaches enlarged the spectrum of valuable whole-cell biocatalysts. Of course, the identification of com-plete genome sequences facilitated investigations andelucidated the wealth of the host-own enzyme spectrum.Since the publication of the complete genome sequenceof S. cerevisiae in 1996 [48,49], more than 30 yeastgenomes have been completely or partly sequenced anddeposited in public databases (e.g. NCBI: http://www.ncbi.nlm.nih.gov/sites/entrez?db=genome). This infor-mation is now available and ready to be exploited forthe construction of valuable whole-cell factories for i.a.biocatalytic applications.

At the same time, a second branch of science, namelymetabolic engineering, emerged in the field of geneticengineering and in 1991, Bailey [138] extensivelyreviewed this new field for the first time. In the followingyears, metabolic engineering has developed rapidly andconsequently resulted in various reviews [139-142] andtextbooks [143,144] explaining basic principles andmethodologies. The many existing examples of meta-bolic engineering were classified in several categorieswhich include heterologous protein production, exten-sion of substrate range, pathways leading to newproducts, pathways for the degeneration of xenobiotics,engineering of cellular physiology for process improve-

ment, elimination/reduction of by-product formationand improvement of yield or productivity [142]. Allapproaches have two important parts in common: first,the identification of the most promising target(s) forgenetic manipulation by solid strain analysis andsecond, the construction of genetically modified cells[141]. As several categories of metabolic engineering andthe construction of recombinant strains for whole-cellbiocatalysis target at improving similar characteristics ofa cell, it is not always possible to draw a clear linebetween these two disciplines. However, the variety of


http://-/?-http://www.ncbi.nlm.nih.gov/sites/entrez?db=genomehttp://www.ncbi.nlm.nih.gov/sites/entrez?db=genomehttp://www.ncbi.nlm.nih.gov/sites/entrez?db=genomehttp://www.ncbi.nlm.nih.gov/sites/entrez?db=genomehttp://-/?-


13/36

recent reports on metabolic engineering of yeasts for theefficient production of top value added chemicals (e.g.ethanol, glycerol, xylitol, succinic acid and other organicacids) [145] would go beyond the scope of this review.Therefore, we refer to existing comprehensive overviews

in this field of research [141,142,146-150].

In the following we focus on engineered yeast platformstrains for the production of mostly chiral precursors forthe pharmaceutical, food or feed industry, therebyincluding single- and multi-step biocatalytic reactions.Finally, we highlight recent pathway engineering effortsleading to structurally complex natural products, a fieldwhich indicates the possible catalytic scope of engi-neered yeasts.

3.1 Organic single or few-step transformations catalyzed by

engineered yeast cellsEmploying non-engineered yeast strains for whole-cellbiotransformations has often resulted in low chemicaland optical purities. As selective reduction and oxidationreactions are the most thoroughly investigated applica-tions of whole-cell yeast biocatalysts, the first recombi-nant yeast strains were also engineered for this class ofbiotransformations. Several groups subjected S. cerevisiaeto step-by-step gene knockout and overexpression ofvarious reductases and studied the catalytic properties ofthe resulting engineered yeast strains in detail (Table 3and Additional File3).

In 1 98 5, Shieh et a l. [151] already succeeded inimproving the optical purity for ethyl (R)-4-chloro-3-hydroxybutanoate production starting with the prochiralb-keto ester. They employed theS. cerevisiaeATCC 26403mutant lacking a competing b-keto reductase andachieved 90% ee compared to 57% ee for the wild-typestrain [151]. Fifteen years later, Rodrguez et al. [152]altered the levels of three S. cerevisiae oxidoreductases,namely fatty acid synthase (Fasp), aldo-keto reductase(Ypr1p), and a-acetoxy ketone reductase (Gre2p),respectively. They first created a set of first-generation yeast strains deleting or overexpressing only one singleenzyme at a time. Based on subsequent results, asecond-generation

set of yeast strains was constructed,combining gene deletion and overexpression. Several

b-keto esters were used as substrates and among theengineered strains not only significantly improvedstereoselectivities compared to the wild-type but alsoengineered strains with inverted selectivities were found[152] (see for example Table 3 and Additional File 3,Entry 2, Substrate d and e). However, this study alsorevealed that additional yeast proteins exist which mightparticipate in the reduction of some of the subjectedb-keto ester substrates and deteriorate especially

enantiopurities. Later, forty-nine open reading framesencoding for potential reductase activities were eluci-dated in the complete S. cerevisiae genome [68].However, up to now not all corresponding reductaseshave been investigated in detail. Some selective alter-

native S. cerevisiae reductases were identified andexa mined by Gor wa- Gr aus lund a nd cowork er s[153,154]. Overexpressing the NADPH-dependentreductases Ara1p, Ypr1p and YMR226c, respectively,led to yeast strains with increased yield ( 48%) andenantiomeric excess ( 87%) for the conversion ofdiacetyl to (S)-acetoin (wild-type: 42% yield, 80% ee)[153,154]. Furthermore, Johanson et al. [155] improvedthe asymmetric reduction of a bicyclic diketone to(1R,4S,6S)-6-hydroxy-bicyclo[2.2.2]octane-2-one whichis used as intermediate for the production of transitionmetal-based chiral chemical catalysts [156]. A combina-tion of reaction and strain engineering (overexpression

of the S. cerevisiae reductase YDR368w in S. cerevisiae

Table 3and Additional File3) led to 84% yield, > 99%ee and 97% de compared to initially described resultswith 52% yield, 95% ee, and 84% de after chromato-graphy and crystallization [157]. Recently, Kratzer et al.[158] succeeded in overcoming limitations of bakersyeast in the asymmetric reduction ofa-keto esters usingalso a combinatorial approach. They improved theenzyme by exchanging the amino acid residue trypto-phane 23 with phenylalanine (CtXR-W23F) whichresulted in an up to eightfold higher NADH-dependentactivity compared to the wild-type enzyme employing aseries of aromatica-keto esters as substrates [159]. The

strain S. cerevisiae was engineered to efficiently over-express CtXR-W23F by introducing multiple copies ofthe expression cassette (strain S.c. WF2). Finally, thereaction itself was optimized performing it underanaerobic conditions in the presence of EtOH. TherebyNADH-regeneration was preferred. This was accom-plished by the yeasts own NADH-dependent alcoholdehydrogenase (ADH) and NAD(P)H-dependent alde-hyde dehydrogenase (AlDH). Thus, the action ofendogenous NADPH-dependent yeast reductases withdiverse enantiopreferences was suppressed [15 8].a-Hydroxy esters were obtained in acceptable yieldsand high enantiopurities ( 99.4% ee, see Table 3 and

Additional File 3, Entry 6).

Exploiting the advantage ofS. cerevisiae to be GenerallyRegarded As Safe (GRAS) for human consumption,Farhi et al. [160] engineered this yeast for the improvedsynthesis of the food flavoring methyl benzoate. Thiswas achi eved by expressing the Antirrhinum majusbenzoic acid methyltransferase (BAMT) under the con-trol of the copper-inducible CUP1 promoter. A yield of~1 mg methyl benzoate per liter of culture was achievedafter 24 h, starting with benzoic acid. In addition, it was


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


14/36

Table 3: Examples for one- to two step enzymatic reactions employing engineered yeast whole-cell biocatalysts (+ overexpression

Whole-CellBiocatalyst

Engineering+ overexpression- deletion

Substrate Product

Saccharomycescerevisiae

- b-keto reductase

OC2H5

Cl

OO

OC2H5

Cl

OOH


A: +Fasp, -Gre2p, -Ypr1pIV

R1

O

OR3

O

R1

OH

OR3

O

R1

OH

OR3

O

+

(3R) (3S)B: -Fasp, +Gre2pIV

a: R1 = Me; R3 = Me b: R1 = Et; R3 = Me c: R1 = Me; R3 = Et d: R1 = Et; R3 = Et e: R1 = n-Pr; R3 = Et


A: - Gre2p, + Ypr1pIV

CH3

O

OEt

O

R2

a: R2 = Me

CH3

OH

OEt

O

R2

CH3

OH

OEt

O

R2

syn(2R, 3S) anti(2S, 3S)

B: + Gre2P, - Ypr1pIV b: R2 = Et c: R2 = allyl

d: R2 = propargyl


A: +Ymr226cIII

R1

R2

O

O

R1

R2

OH

O

(S)

B: +Ara1pIII R1 = Me; R2 = Me C: +Ypr1pIII


+ reductase gene YDR368wV

O

O

bicyclo[2.2.2]octane-2,6-dione

O

OH

(1R,4S,6S)-6-hydroxy-bicyclo[2.2.2]octane-2-ol


7:25



15/36

Table 3: Examples for one- to two step enzymatic reactions employing engineered yeast whole-cell biocatalysts (+ overexpression


+ multiple copies ofCandidatenuis xylose reductase variantW23F R O

O

O

CH3

-keto ester

a: R = H

R O

OH

O

CH3

(R)- or (S)-hydroxy ester

b: R = Ph c: R = o-Cl-Ph d: R = m-Cl-Ph e: R = p-Cl-Ph f: R = p -CN-Ph


+Antirrhinum majusbenzoic acidmethyltransferase (BAMT)

OH

O

benzoic acid

OMe

O

methyl benzoate

Pichia pastoris + Rhodotorula glutinis epoxidehydrolase

O

(rac)-styrene oxide

O

(S)-styrene oxide

Schizo-saccharomycespombe

+ human CYP11B1-V78I+ electron transfer proteinsadrenodoxin (Adx) &adrenodoxin reductase (AdR)

O

CH3 H

OOH

OH

HH

CH3

11-Deoxycortisol

O

CH3 H

OH

OOH

OH

HH

CH3

Hydrocortisone


+ human CYP2D6

CH3

O

CH3

N

4-methyl--pyrrolidinobutyrophenone(MPBP)

O

CH3

NOH

4-hydroxymethyl--pyrrolidinobutyro-phenone (HO-MPBP)


+ Trigonopsis variabilis D-AminoAcid Oxidase (multi copies)- catalase

Cephalosporin C a-ketoadipyl-7-cephalosporanicacid and glutaryl-7-amino-cephalosporanic acid

Icarbon source: glucose; IIcarbon source: galactose; IIIAra1p = S. cerevisiae reductase, NADPH-dependent; Ypr1p = S. cerevisiae reductase, NADchain dehydrogenase ORF [153]; IVFasp = S. cerevisiae fatty acid synthase; Gre2p = S. cerevisiae a-acetoxy ketone reductase; Ypr1p = S. cerevisi2-methylbutyraldehyde reductase (NCBI accession number: NP 010656); VIWhole-cell bioreductions ofa-keto esters (10 mM) under anaerobicn.d. = not determined.


7:25



16/36

demonstrated that yeast strains expressing BAMT weremore tolerant towards benzoic acid. Several groups alsofocused on engineeringS. cerevisiaein order to modulateand improve the formation of antioxidants and aromacompounds during wine fermentation. In 2000, Gonz-

lez-Candelas et al. [161] used a transgenic yeast strain forthe improved production of resveratrol, an antioxidantwhich was reported to possess cancer chemoprotectiveproperties [162 ,163 ]. They expressed the Candidamolischiana bglN gene encoding for a b-glucosidase inthe industrial wine yeast S. cerevisiae T73 (CECT1894)and yielded elevated contents of both trans- and cis-resveratrol ( 0.75 M) compared to the wild-type( 0.25 M) [161]. Further engineering concepts forimproved resveratrol production with recombinant yeaststrains are discussed in detail in chapter 3.4. Subse-quently, Genovs et al. [164] optimized the productionof recombinant Candida molischiana BGLN enzyme in

S. cerevisiae and tested the purified BGLN enzyme forvinification experiments. The efficient release of terpenesand alcohols from Muscat wine glycosides was observed.Smit et al. [165] tried to develop wine yeasts withoptimized decarboxylating activity on phenolic acids.Therefore, two different phenolic acid decarboxylases(PADC from Bacillus subtilis and p-coumaric acid dec-arboxylase (PDC) from Lactobacillus plantarum) wereexpressed in S. cerevisiae under the control of theconstitutive S. cerevisiae phosphoglyceratekinase I genepromoter and terminator (PGK1P and PGK1T). In mostof the Chardonnay and Riesling wines, the vinificationdone by the recombinant strains resulted in slightly

increased 4-ethylphenol, 4-vinylphenol and 4-vinyl-guaiacol concentrations. However, first results of indus-trial yeast fermentation analysis showed very highconcentrations of volatile phenols close to and evenhigher than the ideal concentration. Thus, furtherimprovements are necessary including a better regulationof the recombinant enzymes overproduction or the pre-evaluation of the available concentration of precursor namely phenolic acid components in the grape juice.Cordente et al. [166] engineered S. cerevisiae in order tooptimize concentrations of volatile esters which repre-sent the largest and most important group of wineflavoring components produced during fermentation.

They overexpressed the major mitochondrial and perox-isomal carnitine acetyltransferase (CAT2p) from S.cerevisiae but instead of increased levels of ester compounds reduced concentrations were observed. Theauthors hypothesized that the overexpression of Cat2pfavors the production of acetylcarnitine and CoA andtherefore limits the precursor for ester formation [166].

Finally, Stewart and coworkers tested a Saccharomycescerevisiae strain expressing the cyclohexanone monooxy-genase from the bacterium Acinetobacter sp. NCIB 9871

with a variety of substituted cycloalkanones [167-170] andseveral sulfides, dithianes and dithiolanes [88], respectively(Table 4and Additional File 4). Thereby, they combinedthe broad substrate tolerance of the enzyme with bakersyeast tolerance for relatively high concentrations of organic

compounds, providing a stable environment for theenzyme. Furthermore, no enzyme purification or addi-tional NADPH-regeneration system was necessary and theenzyme from this potentially pathogenic bacterial strain(class II) [88] became easily accessible. Enzyme inductionin the Acinetobacter strain also required the addition ofcyclohexanol which complicated product isolation [68].Table 4 and Additional File 4 depicts detailed results ofenantioselective Baeyer-Villiger oxidations with the abovedescribed designer yeast. Although high optical puritieswere achieved with the majority of investigated substrates,it is still clear that yields must be improved significantly inorder to allow the biocatalysts commercial application

[68].

In addition to S. cerevisiae strains, also alternative yeastswere genetically modified to exhibit beneficial oxidor-eductase properties (Table 3 and Additional File 3).Rhodotorula glutinis epoxide hydrolase was for exampleoverexpressed in Pichia pastoris leading to a 10-foldelevated activity toward (R)-styrene oxide conversioncompared to the R. glutinis wild type. Kinetic resolutionof racemic styrene oxide yielded in the (S)-styrene oxidewith 98% ee and 36% yield [171]. A recombinantSchizosaccharomyces pombestrain coexpressing the humancytochrome P450 CYP11B1 mutant V78I and the

electron transfer proteins adrenodoxin (Adx) and adre-nodoxin reductase (AdR) showed a 2.5-fold higher 11b-hydroxylation activity than the parental strain for theproduction of hydrocortisone [172]. Furthermore, Peterset al. [173] developed a Schizosaccharomyces pombe strainexpressing human CYP2D6 and evaluated its principlefeasibility for the synthesis of 4-hydroxymethyl-a-pyrrolidinobutyrophenone (56% yield, 98.5% purity Table 3 and Additional File 3). Other yeast whole-cellbiocatalysts which were engineered to display enzymeson their surface are described in section 3.3.

3.2 Cofactor regeneration using genetically engineered yeastsAs far as organic biotransformations are concerned, themost common application of yeast whole-cell biocata-lysts is for asymmetric reductions. For these syntheticallyuseful reactions cofactor-dependent enzymes arerequired. Considering the cost of NAD(P)+ and NAD(P)H, their stoichiometric application is not economic-ally feasible unless an efficient regeneration method isemployed which allows the use of only low amounts[174]. In general, whole-cell biocatalysts provide thecheapest cofactor regeneration system preventing the


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


17/36

Table 4: Examples for one step enzymatic reactions employing the designer yeast made by Stewart and coworkers [88, 16

overexpressing the Acinetobactersp. NCIB 9871 cyclohexanone monooxygenase. The data is chronologically ordered

Entry Substrate Product Substrate/Product Code

Performance:yield (ee)

1 O

R

O

O

R

a: R = Me a: 83% ( 98%)

b: R = Et b: 74% ( 98%)c: R = Pri c: 60% ( 98%)d: R = Pr d: 63% (92%)e: R = allyl e: 62% (95%)

1

2

O

R

O

RO

O

R

+

1 a-f 2 a-f

a: R = Me a: 50% (49%) b: R = Et b: 79% (95%) c: R = Pri c: 41% ( 98%) d: R = Pr d: 54% (97%) e: R = allyl e: 59% ( 98%) f. R = n-butyl f: 59% ( 98%)

1 2

3

O

R

O

O

R

1 a-f 3 a-f

O

O

R

2 a-f

O

O

R

a: R = Me a: 71% ( 98%) 60 % ( 98

b: R = Et b: 18% (70%) 20% (70%c: R = Pri c: N.R.II I

d: R = Pr d: 11% ( 98%) I

e: R = allyl e: 15% (97%) I

f. R = n-butyl f: 37% (56%) I

1

4

O

RO

O

R

+

O

R

1 a-d 2 a-d

a: R = n-Bu a: 18% ( 98%) b: R = n-Hex b: 32% ( 98%) c: R = n-Oct c: 25% ( 98%) d: R = n- d: 39% ( 98%) C11H23

1III 2III

5

O

R

O

R

O

O

R

O

O

R

+

2 a-b 3 a-b

+

1 a-b

a: R = n-Pr a: 27% (13%) -IV (33%

b: R = n-Hex overall yield: 44%; ratio

b: 54% (29%) (60%)overall yield: 20%; ratio

6 RS

CH3

RS

CH3

O-

H

+

a: R = Ph a: 95% (> 99%)

b: R = But b: 47% (99%)c: R = Bun c: 53% (74%)

MicrobialCellFactories2

008,

7:25



18/36

Table 4: Examples for one step enzymatic reactions employing the designer yeast made by Stewart and coworkers [88, 16overexpressing the Acinetobactersp. NCIB 9871 cyclohexanone monooxygenase. The data is chronologically ordered (Continued)

7

SS

R1 R2

SS

R1 R2

O-

:SS

R1O

O

R2

1 a-b 2 a-b

+

a: R1 = HR2 = H

1a: 18% (90%)(2a: 19% yield)

b: R1 = HR2 = Ph

1b: 30% (30%)(no sulfone 2bdetected)

8

S S

R2R1

S S

R2R1

O-

:S S

R2R1O

O

1 a-c 2 a-b

+

a: R1 = HR2 = H

1a: 20% (75%)[2a: 45% yield]

b: R1 = HR2 = CH3

1b: 16% (20%)[2b: 15% yield

76% ee]c: R1 = HR2 = Ph

1c: 74% (20%)[no sulfone 2cdetected]

9

S S

R2R1 S S

R2R1O

-

:S S

R2R1O

O

+

1 2

R1 = CH3 1: 84% (48%)

R2 = CH3 2: 10%

IThis isomer was not detected in the product mixture [168];IIno oxidation detectable [168];IIIdue to the low enantioselectivities of the reactions, no absolute configuration was assigned [ 170];

IVvalue not given [170].


7:25



19/36

laborious procedure of enzyme isolation and takingadvantage of prolonged enzyme stability and cofactorrecycling from cheap auxiliary substrates. However, thereexist several disadvantages when employing wild-typewhole-cell biocatalysts (see Section 1.1).

Recently, improvements in cofactor regeneration systemswere comprehensively reviewed focusing on pyridinenucleotide regeneration [175], cofactor regeneration withgenetically engineered bacteria [176] and new methods forthe improved regeneration of ATP/NTP, sugar nucleotidesand PAPS (3-phosphoadenosine-5-phosphosulfate)which is involved in sulfuryl transfer reactions [177]. Inmost cases however isolated enzymes and geneticallymodified bacteria were employed. Regarding again S.cerevisiae, only traditional approaches were followed,using the yeast for its supply of reducing power andencountering serious drawbacks when trying to freely

combine it with desired non-natural substrates. Stewartand coworkers [47,68,168] were among the first scientistswho constructed genetically engineeredS. cerevisiae strainsand thereby reduced undesirable dehydrogenase back-ground reactions by deleting single reductase genes. Thus,it became again more attractive to utilize the S. cerevisiae-own cofactor regeneration system for the asymmetricreduction of selected ketones. However, considering allattempts it seems that the potential of S. cerevisiae aswhole-cell redox-biocatalyst has not been fully exploitedyet. The huge number of existing S. cerevisiae genesencoding for reductases suggests a more complex strategysuch as temporarily silencing of metabolic pathways either

by reaction engineering as indicated by Kratzer et al. [158](see Chapter 3.1, Table 3 and Additional File 3) or byregulation of gene expression, a perhaps more challengingapproach.

At the same time, metabolic engineers also investigated theimpact of cofactor engineering on the redox metabolism ofS. cerevisiae. Nissen et al. [178] expressed for example thecytoplasmic transhydrogenase fromAzotobacter vinelandiiinS. cerevisiae CBS8066. Transhydrogenases, which were notfound in yeasts [179,180], catalyze the hydrogen transferbetween the two cofactor systems NADH/NAD+ andNADPH/NADP+. Nissens results indicated that when

introducing a transhydrogenase in yeast, the conversionof NADP(H) and NAD+ to NADP+ and NADH was favored[178]. This was already observed by Anderlund et al. [181]who constructed recombinant S. cerevisiae strains expres-sing a membrane-bound E. coli transhydrogenase. Incontrast, Nielsen and coworkers [182] also succeeded inconstructing genetically engineeredS. cerevisiaestrains witha modified ammonium assimilation pathway for increasedNADPH availability. Therefore, the NADPH-consumingglutamate dehydrogenase (GDH1) was deleted and analternative pathway consisting of an NADH-consuming

GDH2 or the ATP-dependent glutamine synthetase(GLN1) and the NADH-dependent glutamate synthase(GLT1) was introduced. The major redox alteration of theengineered strain was shown by a reduced flux through thepentose phosphate pathway during aerobic growth on

glucose. Finally, this switch in cofactor requirement forammonia assimilation resulted in both an increase inethanol yield (10%) and a significant reduction of theglycerol yield (~40%) [183]. Furthermore, S. cerevisiaestrains with a decreased intracellular NADH pool weregenerated by expressing theLactobacillus lactisH2O-formingNADH oxidase [184]. A major consequence of the thusprovided NAD+ excess led to a significant reduction inethanol yield (15%). This strategy could provide a route toreduce and adjust the ethanol content of fermentedbeverages like beer or wine [184]. The overexpression ofalternative oxidases could also result in valuable strains forNAD+ regeneration.

Recently, the methylotrophic yeast Pichia pastoris wasengineered for efficient NADH regeneration in ourlaboratory (F. S. Hartner, personal communication).The concept included the decrease in MeOH assimilationby deleting two dihydroxyacetone synthase genes.Thereby, the existing methanol utilization pathway generating two molecules of NADH while methanol isirreversibly oxidized to CO2 was enhanced. Thisapproach led to improved space-time yields and specificconversion rates for the conversion of acetoin to 2,3-butanediol catalyzed by the overexpressed S. cerevisiaebutanediol dehydrogenase. In this case, methanol was

used as a cosubstrate for NADH regeneration but also asan inducer for the expression of the heterologous catalystand the endogenous NADH regeneration pathway.

Furthermore, engineeredS. cerevisiaecells overexpressinghexokinase or glucokinase were reported to almostcompletely convert adenosine (130 mM) to ATP [185],providing a possible ATP regeneration system [25].

Again E. coliand the yeastS. cerevisiae are the most oftenand thoroughly studied organisms regarding attempts tooptimize cofactor recycling by genetic engineering.However, especially methylotrophic yeasts provide inter-

esting alternatives for engineering endogenous cofactorrecycling pathways.

3.3 Transport limitations and displaying enzymes

on the surface of yeast whole-cell biocatalysts

Considering whole-cell processes, it would be ideal if thestarting material was transported into the cell and ifproducts were released without any limitation. The produc-tion rate would then only be dependent on the cellsmetabolic functions including the catalysts kinetic


http://-/?-http://-/?-


20/36

limitations. In reality, bioprocesses are often drasticallylimited by barrier functions of the cell wall or membrane. Inorder to circumvent these limits, permeabilization methodsare common practice. Recently, Chen [186] reviewedstandard permeabilization protocols including efficient

methods for yeasts such as solvent [187-189], detergent[190] or alkaline [191] treatment and air-drying at 42C[192], respectively. However, classical permeabilizationmethods often show serious shortcomings causing extensivedamage to membrane systems, even cell lysis [193], whichmakes cofactor regeneration or the reuse of the catalystimpossible. Furthermore, downstream processing can becomplicated. Recently, molecular engineering approachesprovided an alternative route to solve the cell permeabilityissue in a way thatcan be better predicted [186]. Either outermembrane structures were engineered a method whichwas up to now only used for bacterial cells or enzymeswere displayed on the cell surface. In the following the later

method will be discussed in more detail.

Molecular displaying techniques allow the targeting ofheterologous proteins to the surface of yeast strains. Theyare covalently linked to the internal, skeletal layer ofb-1,3-and b-1,6 glucan complexed with chitin and protrude tothe outer layer of glycoproteins [194]. This techniqueprovides several advantages: The anchoring to the cellsurface can eliminate purification and separation processesand result in increased biocatalyst stability. Benefits ofS.cerevisiae which has GRAS status and is well suited for theexpression of proteins from eukaryotes can be combinedwith economic aspects like facilitated biocatalyst recycling

due to the self-immobilization of the recombinantenzymes, simple cultivation and fast growth to high celldensities. Furthermore, mass transport problems of sub-strate and/or product across the cell membrane can beprevented as the enzyme, necessary for catalysis, isdisplayed on the cellssurface [195,196]. This is especiallyimportant when polymeric compounds are used assubstrates. In spite of these benefits, current disadvantagesof whole-cell biocatalysts with enzymes anchored to theircell wall should not be neglected such as currentlimitations to one- or few-step reactions, the restriction tosimple biotransformations independent of costly cofactorsand the loss of the protective subcellular environment

providing optimal ionic, pH- and redox-conditions forenzymatic reactions. However, one or the other disadvan-tage could be solved by further genetic engineeringapproaches such as co-displaying of a cofactor-dependentoxidoreductase and an enzyme for cofactor recycling or theengineering of the outer membrane structure of yeast cellsto improve the conditions in the vicinity of the displayedenzyme.

In particular, Japanese scientists succeeded in developingthe first enzyme-displaying yeast cells and investigated

their application for whole-cell biocatalysis focusing onhydrolases. They studied for example the optical resolutionof (RS)-1-phenylethanol by enantioselective transesterifica-tion with vinyl acetate [197]. Therefore,S. cerevisiaeMT8-1cells were employed which displayed the pro-region of

Rhizopus oryzaelipase (ProROL) by fusing the flocculationfunctional region of the cell-wall protein Flo1p to thelipases N-terminus [198]. The conversion resulted in the(R)-1-phenylethyl acetate with high yields and enantio-meric excess (> 93%) [197]. Furthermore, Kondo andcoworkers [199] applied S. cerevisiae cells displayingRhizopus oryzae lipase (ROL) for the optical resolution of(R,S)-1-benzyloxy-3-chloro-2-propyl monosuccinate. Aftera reaction time of 16 h, the ROL-displaying yeast hadhydrolyzed the (R)-compound with an eep-value of 95.5%and a conversion of 50.2%, whereas the employment ofsoluble R. oryzae lipase (F-AP15, Amano enzyme Inc.,Japan) showed quite poor ee and conversion (58.7% eep,

35.9% conv., at 12 h) [199]. The stability of ROL-displaying yeast cells was confirmed by reusing thebiocatalyst eight times without significant activity loss.The displaying-technique was also combined with enzymeengineering: Shibamoto et al. [200] constructed a surface-displayed combinatorial library of R. oryzae lipase in S.cerevisiaewhich resulted in new biocatalysts with increasedactivity for the hydrolysis and methanolysis, respectively, ofsoybean oil. Methanolysis yields in methyl esters constitut-ing a potential biodiesel fuel. Recently, Kaya et al. [201]reported the production of isoflavone aglycones fromisoflavone glycosides employing S. cerevisiae cells. Threedifferentb-glucosidases from Aspergillus oryzaewere indivi-

dually displayed on the yeast cell surface. The engineeredyeast strain with the b-glucosidase BGL1 (gene ID:AO090003001511 from the A. oryzae RIB40 genome)exhibited the highestb-glucosidase activity and first whole-cell biotransformations were performed to produce theisoflavone aglycones daidzein, glycitein and genistein.Furthermore, Kim et al. [202] described the productionof cyclofructans from inulin. They engineered S. cerevisiaeto display cycloinulooligosaccharide fructanotransferasefrom Paenibacillus macerans on the cell surface. As majorproduct, cycloinulohexaose was detected along withcycloinuloheptaose and cycloinulooctaose as minor pro-ducts. Yeast surface display was also used by Wittrup and

coworkers for the FACS-based selection of horseradishperoxidase variants with enhanced enantioselectivitytoward L- and D-tyrosinol, respectively [203].

Another lipase was displayed by Ueda and coworkers[204]. They developed S. cerevisiae cells displaying amutated form ofCandida antarctica lipase B which wasconstructed on the basis of the primary sequences of theCALBs from C. antarctica CBS 6678 and C. antarctica LF058, respectively.a-Agglutinin was used as anchor protein.The newly generated biocatalyst displayed high thermal




21/36

stability (T1/2 (60C) = 30 min) and > 6-fold higheractivity for the hydrolysis of p-nitrophenyl butyratecompared to previously reported CALB mutants generatedby Zhang et al. (CALB-23G5: 86 nmol/min/g protein)[205]. However, despite of several advantages such as for

example straight-forward library construction and screen-ing, whole cells displaying enzymes still possess their ownpool of wild-type enzymes that might act on providedsubstrates able to diffuse through the cell membrane.Considering especially esterolytic and lipolytic activitieswhich are quite common in all organisms, this might leadto undesirable by-products and a significant reduction ofchemical and optical yield.

A great benefit of the enzyme displaying technique is theprevention of mass transport problems of the substrateacross the cell membrane. This makes polymeric com-pounds amenable as substrates for whole-cell biocatalysis.

In the following some examples are given describing thesuccessful degradation of different polymeric substrates byrecombinantS. cerevisiae cells displaying enzymes on theircell surface. More details were summarized by Ueda andTanaka [206] and Wu et al. [207], respectively. Fukudaet al. [196] produced chitooligosaccharides from a chitosanpolymer employing yeast cells withPaenibacillus fukuinensischitosanase on the yeast surface. The products are aremarkable resource for the development of e.g. functionalfood and diverse materials such as artificial skin [196].Fujita et al. [208] constructed a Saccharomyces cerevisiaewhole-cell biocatalyst displayingTrichoderma reeseixylanaseII for the degradation of xylan. Furthermore, they devel-

oped a yeast strain (S. cerevisiae) which co-displayed threetypes of cellulolytic enzymes: Trichoderma reesei endoglu-canase II, T. reesei cellobiohydrolase II and Aspergillusaculeatus b-glucosidase 1 [209]. The generated whole-cellbiocatalyst performed the saccharification and subsequentfermentation of amorphous cellulose and produced 0.45 gethanol per gram of carbohydrate consumed (88.5%yield).

Although the cell wall ofSaccharomyces cerevisiae is theone which is the best characterized among yeasts[210,211], targeting proteins is also applicable toalternative variants. Kluyveromyces lactis, for example,

was used to test this and cell-associated a-galactosidaseactivity was detected [194]. Recently, Jiang et al. [212]reported for the first time the displaying of an enzymeon Pichia pastoris KM71, namely the lipase LipB52 fromPseudomonas fluorescensB52. EngineeredP. pastorisstrainsfeatured a slightly improved thermal stability comparedto a Saccharomyces cerevisiae EBY100-pLHJ026 straindisplaying the same lipase. Furthermore, Yue et al.[213] developed a new plasmid for the display ofproteins on Yarrowia lipolytica cells providing the basisfor applications in the field of whole-cell biocatalysis.

3.4 Synthetic pathways in yeasts

The construction and employment of designed strainsfornatural product biosynthesis clearly indicate the potentialof yeast engineering. In recent years, remarkable progresspaved the way for the design of whole-cell biocatalysts

capable of producing structurally complex natural productsand novel variations thereof. An exponentially increasingnumber of sequences was elucidated. New bioinformatictools providing the basis for analyzing this load ofinformation were developed and sophisticated tools forengineering heterologous hosts were established.

Natural products like isoprenoids, flavonoids or polyketidesrepresent structurally complex compounds which are oftenroutinely made in nature but require long and elaboratesynthesis routes when classical chemical methods areemployed. Compounds of all three classes have attractedattention because of their great commercial potential. The

class of isoprenoids includes for example carotenoids whichare valuable antioxidants and food and feed additives,steroids which are widely used as drugs constituting anti-inflammatory, contraceptive and anticancer agents, orterpenoids like the well-known diterpenoid taxol with itsexcellent activity, also against a range of cancers. Flavonoidswhich derive from the phenylpropanoid pathway possessanti-allergenic, anti-inflammatory, and anti-oxidant activ-ities in humans, and polyketides are applied for thetreatment of cancer (adriamycin), cardiovascular diseases(lovastatin), immunosuppression (rapamycin, tacrolimus)or infectious diseases (erythromycin, tetracycline) [214].

As the structural complexity makes the chemical synthesis ofnatural products quite difficult, fermentation is regarded tobe an economically feasible alternative to produce pharma-ceutically useful compounds for commercial purposes.Most native organisms producing complex polyketide- orisoprenoid-derived compounds do not provide high-levelsof these substances as they tend to grow slowly. In addition,they are often not amenable to genetic manipulation. Thus,alternative heterologous systems with faster growth andestablished genetic engineering techniques seem to providean interesting alternative with respect to volumetricproductivity. The relatively homogenous processes towardsthe final products are another great benefit of natural

product pathways. Precursors and enzymes of naturalproduct synthesis pathways are closely related and a kindof modular system can be established in order to producedifferent natural but also novel compounds with therapeu-tic properties. Improvements of manufacturing processesare thus possible at a more rapid pace. Essential for theunderstanding of the synthetic pathways are not onlygenetic techniques for their manipulation but also themolecular understanding of every individual biocatalyticstep, hence the information gained by simple one-stepbiochemical reactions.




22/36

So far,Saccharomyces cerevisiaerepresented again an idealhost for synthetic pathway engineering due to its rapidgrowth, advanced fermentation techniques and wellestablished genetic tools [215]. In the following someexamples are given employing S. cerevisiae, but also

Candida utilis which was engineered for carotenoidproduction (Table 5: Additional File5 and 6: AdditionalFile6).

Carotenoids represent one of the first classes of naturalcompounds produced by recombinant yeast strains.Gene clusters responsible for the synthesis of carotenoidswere isolated from carotenogenic bacteria, including forexampleErwiniasp. andAgrobacterium aurantiacum. Theirfunction was elucidated [216,217], and essential geneswere introduced to S. cerevisiae and C. utilis, respectively[218-220]. S. cerevisiae and C. utilis are capable ofaccumulating ergosterol as their principle isoprenoid

compound. Obviously, this strongly suggested thepossibility of redirecting the pathway partly to carote-noid production. Finally, engineered C. utilis strains[218,219] surpassed recombinant S. cerevisiae strains[221-223] in the production ofb-carotene and lycopene(Table 5and Additional File5). Actually,C. utilis strainsperformed even better than engineered E. coli, especiallyfor lycopene production (C. utilis: up to 7.8 mg/g CDW[218]). EngineeredE. coli strains were found to accumu-late more than 1 mg/g CDW of lycopene,b-carotene andzeaxanthin, and more than 0.5 mg/g CDW of astax-anthin [220].

In general, the pivotal intermediate of all isoprenoidcompounds produced in a cell is farnesyl pyrophosphate[224], which also represents the branching point of thediverse isoprenoid biosynthetic pathways (Figure 3).Eukaryotic cells possess a variety of isoprenoid com-pounds, which are involved in a vast array of cellularprocesses such as post-translational modification ofproteins (prenylation) and tRNA modification. Theyrepresent compounds of the lipid bilayer, electrontransporters during respiration, cofactors involved inenzyme catalysis, and hormones for signal transduction.

In 1999, Dimster-Denk et al. [225] investigated the

regulation of the isoprenoid pathway in S. cerevisiae.They evaluated the expression of

microbial cell factories

Documents

Transcript of microbial cell factories