Research Collection

Doctoral Thesis

Structural and functional characterization of the multidrug ABCtransporter Sav1866 from Staphylococcus aureus

Author(s): Dawson, Roger John Peter

Publication Date: 2007

Permanent Link: https://doi.org/10.3929/ethz-a-005429459

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. 17117

Structural and Functional Characterization of the Multidrug ABC

Transporter Savl866 from Staphylococcus aureus

A dissertation submitted to

ETH Zurich

for the degree of

Doctor of Sciences

presented by

ROGER JOHN PETER DAWSON

Dipl.-Chem. Univ. TU München

born 09.05.1977

citizen of

Great Britain

accepted on the recommendation of

Prof. Dr. Kaspar Locher, examiner

Prof. Dr. Rudolf Glockshuber, co-examiner

2007

Acknowledgements

It is a pleasure for me to express my gratitude to the people who supported me during

my time at ETH Zurich.

I am particularly thankful to Prof. Kaspar Locher (ETH Zurich) for giving me the

opportunity to perform this exciting work in his laboratory and to be integrated in an

excellent and challenging research environment. I would like to thank him very much

for his continuous support and the numerous inspiring discussions.

I would like to thank Prof. Rudi Glockshuber (ETH Zurich) for his support as

co-examiner of this thesis and his generous support with equipment, space and

expertise.

I thank Dr. David Sargent (ETH Zurich) for his help with the derivatisation of crystals

with xenon noble gas and Dr. Clemens Schulze-Briese, Dr. Ehmke Pohl and Dr.

Takashi Tomizaki for their assistance with synchrotron data collection at the Swiss

Light Source.

1 would very much like to thank the members of the Locher Lab (Kaspar Hollenstein,

Rikki Hvorup, Sabina Gerber, Birke Goetz, Dr. Mireia Cornelias, Christian Flogaus,

Martina Niederer and Dominik Frei) for all the help, stimulating and motivating

discussions and especially for the excellent work atmosphere!

1 would like to thank many members of the institute (Institute of Molecular Biology

and Biophysics, ETH Zurich), especially Monika Marti, Gabriela Arm and Markus

Hildbrand for the support and help.

Finally, I dedicate special thanks to the people that belong to me, my friends and

family, who always encouraged and supported me throughout all those years...

Table of Contents

Acknowledgements ii

Table of contents iii

Abbreviations iv

Summary vi

Zusammenfassung vii

Chapter 1 : Introduction 1

Chapter 2: Recombinant Expression, Purification and Activity

Screening of Bacterial MDR-ABC transporters 19

Chapter 3: Expression, Purification, Crystallization and

Crystallographic Analysis of 5. aureus Savl866 31

Chapter 4: Structural Characterization of the Bacterial MultidrugABC transporter Savl 866 50

Chapter 5: Functional Reconstitution of Savl866, Basal and

Drug-stimulated ATPase Activity 71

Chapter 6: Conclusions and Outlook 82

Publications 85

Curriculum Vitae 86

m

Abbreviations

2-I-ADP 2'-iodo-adenosine diphosphate

A600 optical density at a wavelength of 600 nm

ABC ATP-binding cassette

ADP adenosine diphosphate

Amp ampicilin

AMP-PCP adenyl 5'-ß,y-methylene diphosphate

AMP-PNP adenosine 5'-ß,y-imido triphosphate

ATP adenosine triphosphate

B. subtilis Bacillus subtilis

C-terminal carboxy-terminal

C12E8 octaethylene glycol monododecyl ether

C10E5 pentaethylene glycol monodecyl ether

CMC critical micelle concentration

DDM n-dodecyl-ß-D-maltopyranoside

DM n-decyl-ß-D-maltopyranoside

E. coli Escherichia coli

ECL extracellular loop

EDTA ethylenediaminetetraacetic acid

BMP ethyl mercury phosphate

FOS12 FOS-choline 12

FOS14 FOS-choline 14

G3P glycerol-3-phosphate

H. influenzae Haemophilus influenzae

HEPES N-(2-hydroxyethyl)-piperazine-N '-2-ethanesulfonic acid

ICL intracellular loop

IPTG isopropyl-ß-D-thiogalactopyranoside

J^m Michaelis-Menton constant

LDAO n-dodecyl-N,N-dimethylamine-N-oxide

L. lactis Lactococcus lactis

MAD multiple-wavelength anomalous diffraction

M. jannaschii Methanococcusjannaschii

MDR multidrug resistance

IV

MFS major facilitator superfamily

MIRAS multiple isomorphous replacement with anomalous scattering

MPD 2-methyl-2,4-pentanediol

NBD nucleotide-binding domain

NiNTA nickel-nitrilotriacetic acid

N-terminal amino-terminal

OG octylglucoside

PCR polymerase chain reaction

PDB protein data bank

PEG polyethylene glycol

S. aureus Staphylococcus aureus

S. typhimurium Salmonella typhimurium

SDS-PAGE sodium dodecyl sulfate Polyacrylamide gelelectrophoresis

SUV small unilamelar vesicle

TM transmembrane

TMD transmembrane domain

Tris tris(hydroxymethyl)aminomethane

V. cholerae Vibrio cholerae

V

Summary

ATP binding cassette (ABC) transporters are integral membrane proteins that actively

transport chemically diverse substrates across the lipid bilayers of cellular

membranes. Clinically relevant examples like human Mdrl contribute to multidrug

resistance of cancer cells by catalyzing the extrusion of cytotoxic compounds used in

cancer therapy. Prokaryotic homologs of human multidrug ABC transporters exhibit

overlapping substrate specificity and LmrA from L. lactis has been found to

functionally substitute for human Mdrl when overexpressed in lung fibroblast cells.

The basic ABC transporter architecture comprises two transmembrane domains that

provide a translocation pathway, and two cytoplasmic, water-exposed nucleotide-

binding domains (NBDs) that hydrolyze ATP. Bacterial multidrug ABC proteins are

generally expressed as half-transporters' that contain one TMD fused to a NBD and

dimerize to form the full transporter. Functionally important residues are highly

conserved among the NBDs, suggesting that ABC transporters share a common

mechanism of coupling ATP hydrolysis to substrate transport. However, despite the

large body of genetic, biochemical, and structural evidence, a common mechanism

has not emerged.

This thesis presents the first crystal structure of a multidrug ABC transporter -

Savl866 from Staphylococcus aureus - at high resolution in a physiologically

relevant conformation. In a homology screening approach, Savl866 was selected and

purified in detergent solution. Purified Savl866 exhibited basal ATPase activity that

could be specifically inhibited by vanadate. When reconstituted in liposomes, the

ATPase activity could be stimulated by typical substrates of LmrA and Mdrl (cancer

drugs and a fluorescent dye). The structure of Savl866 defines the architecture of an

ABC exporter in the outward-facing, ATP-bound state suggesting that ABC

transporters likely follow an 'alternate access and release' mechanism. The

nucleotide-binding domains tightly sandwich nucleotide at the shared interface in the

closed ATP-bound conformation that is coupled to the outward-facing conformation

of the transmembrane domains. A central cavity, likely resembling an extrusion

pocket, is exposed to the outer leaflet of the membrane and the extracellular space but

is shielded from the cytoplasm. The transmembrane helix topology is consistent with

Mdrl as revealed by cross-linking and electron microscopic imaging.

vi

Zusammenfassung

ABC transporter durchspannen zelluläre Membranen und ermöglichen den aktiven

Transport chemisch verschiedener Substrate. Humane Vertreter der Familie (z.B.

Mdrl) sind von klinischer Bedeutung und tragen zur Resistenz von Krebszellen gegen

Chemotherapeutika bei. Prokaryotische Homologe der 'multidrug' ABC Transporter

zeigen eine überlappende Substratspezifität, und LmrA aus L, lactis substituierte

funktionell für Mdrl bei seiner Überexpression in humanen Fibroblastzellen. ABC

Transporter bestehen allgemein aus zwei Transmembran-domänen (TMDs), die einen

Transportkanal bilden, und zwei zytoplasmatischen Nukleotid-bindende Domänen

(NBDs), die ATP hydrolysieren. Prokaryotische 'Multidrug' ABC Transporter

werden generell als Halbtransporter exprimiert und bestehen aus zwei, aus TMD und

NBD zusammengesetzten, Domänen. Die funktionelle Einheit ist dabei dimer. Die

NBDs enthalten hochkonservierte, funktionswichtige Aminosäuren; ein Indiz für

einen einheitlichen Mechanismus zur Kopplung der ATP-hydrolyse mit dem

Transport von Substraten. Trotz jahrelanger Forschung auf genetischer,

biochemischer und struktureller Ebene, konnte noch kein den ABC Transportern

gemeinsamer Mechanismus beschrieben werden.

Diese Dissertation zeigt die erste hochaufgelöste Kristallstruktur eines 'Multidrug'

ABC transporter - Savl866 aus S. aureus - in einer physiologisch relevanten Konfor¬

mation. Savl866 wurde aus verschiedenen Homologen selektiert und in Detergenz-

lösung gereinigt. Das gereinigte Protein zeigte ATPase Aktivität, welche spezifisch

mit Vanadat inhibiert werden konnte. Die ATPase Aktivität von Savl866 in

Liposomen konnte mit typischen Substraten von L. Lactis LmrA und humanem Mdrl

(Chemotherapeutika und einem Farbstoff) stimuliert werden. Die Struktur von

Savl866 definiert die Architektur eines ABC Transporters im ATP-gebundenen, nach

aussen geöffneten Zustand und verdeutlicht, dass ABC Transporter nach dem

'alternate access and release' Mechanismus funktionieren. Die NBDs binden

Nukleotide an den Bindestellen in der ATP-gebundenen Konformation und koppeln

diesen Zusand mit dem nach aussen geöffneten Zustand der TMDs. Eine vom

Zytoplasma abgeschirmte Öffnung ist sichtbar, die zur äusseren Lipidschicht und zum

extrazellulären Raum hin geöffnet ist. Die Orientierung der Transmembranhelices

entspricht der von humanem Mdrl.

vn

Chapter 1: Introduction

ATP-binding Cassette Transporter Proteins

ATP binding cassette (ABC) transporters comprise one of the largest families of

proteins and are located in the lipid membranes of cells and organelles1. ABC

transporters are crucial for numerous important biological processes in all domains of

life2 and form as integral membrane proteins selective passages between separated

cell compartments. Since their discovery3' 4, increasing attention has been turned to

the field of ABC transporters because of their involvement in the uptake and efflux of

a wide variety of substrates in prokaryotes and eukaryotes. In bacteria, ABC

transporters are mostly involved in the uptake of essential nutrients5'7. Typical

substrates are ions, sugars, amino acids and vitamins. Each importer system has a

cognate substrate binding protein that specifically administers and delivers the

substrates to the membrane embedded ABC transporter. In eukaryotes, predominantly

exporters facilitate the transport of antigenic peptides from the cytosol into the

endoplasmic reticulum for antigen presentation8'9 and play important roles in human

tissue like the liver, blood-brain-barrier2'10 and intestine. Mutations in several of the

approximately 50 human ABC transporter genes lead to malfunctions of their proteins

and form the molecular basis for severe human diseases like cystic fibrosis ' '.

Several of the human ABC transporters are involved in the transport of cytotoxic

drugs and agents used in chemotherapy and are associated with multidrug resistance

of human cancer cells.

ABC Transporters and Multidrug Resistance

Reports about drug resistant cells were published early in 195014 but evidence for the

involvement of ABC transporters in drug resistance was produced much later in

197315. It is known now that acquired drug resistance is observed in bacteria and

human cancer cells and is caused by multidrug (MDR) efflux pumps that are

overexpressed in cytoplasmic membranes. Clinically relevant examples for these

primary active multidrug transport proteins (MDR-ABC transporters) are human

Mdrl and human MRP1 that are both expressed in the plasma membrane of cancer

cells. Physiologically, MDR-ABC transporters are suggested to have a pivotal role in

host detoxification16 and some of the MDR-ABC transporters have been proposed to

1

act as lipid transporters but physiological substrates have not been directly identified.

In prokaryotes, multidrug transporters often function as H+/drug exchangers but

multidrug ABC transporters contribute to drug and antifungal resistance18, 19. What

relates the bacterial with the human ABC transporter systems, is the ability to

translocate a wide variety of neutral and positively charged hydrophobic compounds20

that are functionally and structurally not related. Examples are cancer drugs like

anthracyclines, Paclitaxel and Vinca alkaloids2' or chemical compounds like

fluorescent dyes (Figure 1.1 a-e). The substrates are non-physiological substrates but

have been shown to be transported by bacterial and human ABC transporters. Similar

pharmacological characteristics have been observed for human Mdrl and LmrA from

Lactococcus lactis, for example. When expressed in human lung fibroblast cells, the

bacterial ABC transporter was targeted to the plasma membrane and was able to

functionally substitute for the human multidrug ABC transporter .

The Structure of Multidrug ABC Transporters

ABC transporters are considered primary active transporters that couple the

translocation of substrate to the conversion of adenosine triphosphate (ATP) '

.The

energy generated by binding and hydrolysis of ATP empowers the import or export of

substrates across membranes against a concentration gradient. The minimal functional

transporter unit comprises two transmembrane domains (TMDs) that provide the

translocation pathway, and two cytoplasmic, water-exposed nucleotide-binding

domains (NBDs) that bind and hydrolyze ATP1.

The diversity of the native protein assembly ranges from four individual polypeptide

chains to the fusion of all four domains into a single polypeptide chain. Additional

components are reported for some ABC transporters like regulatory domains and

ABC importers typically require periplasmic binding proteins to acquire substrate

(Figure 1.2a-d). The NBDs bind to the cytosolic surface of the TMDs and contain

highly conserved sequence motifs that resemble the most characteristic features of

this superfamily4' 7' 23. Functionally important residues are located within these

sequence motifs, suggesting that ABC transporters share a common mechanism of

coupling ATP hydrolysis to substrate transport24'25.

2

a b

H /H

H

/v

ri /

ï l-

OH H rf

n

)

c

^ i

ï iîçuk I I îïiiKpori suhsti ttcs of fht nmltidiii (MDR) ABC lîinspottus luniin

Vkii i mdl laUr- I miA FypRal subsUdtts ire shi t. tnctr drii^, a. P ic)itixt,! b,

tht v««f/ UKiloui VinbUi^fwt c, tin mthi K^dsne Ooxoiubsun md She 'luoie luîX

(lyLsd RhothmitK 12 3-ind t. Horchst Hil2

MDR ABC Iranspoitcrs sluie a sigmfiutntlv hiyjier sequence similanî) m (ht,

transmembrane domains indicating eonrnon evolutionary loots'

I he Typaal iour-

ciomain oigani/ation oi ABC iransportcrs is observed litres lypitaiiy encode lor

1

half- and lull -transporters with fused TMD-NBD domains (t <c I mrA ABtG2)or

all foi» domains fused to a >ingîc polypeptide chain (eg Vidi 1 ) While tue hau¬

ts ansportcr configuration is found in piokarvotes and eukaiyotcs in homo- and

heteiodimmc arrangements iuli-tiansportcrs aie cxclusnely ob~.cn cd ni

ciikaryotes borne MDR-ABC tianspoitcrs contain additional icgiilatory domains oi

transmembrane hdaes in the f MDs hke the Multidrug Résistante Piolein i (MRP1 )

t'if»«n 1 2 Ss.hc.mt of the ABC Irsnspoitct domain organ!/ ilioi Tht minimal

functions! unit eornpnses ioiu edit dorn uns two »ansmembranc domains

\I\lDs) and two mieleofkie bmd'ng domains (NIMX) (ht toiofed bIob> inchoate

a, hmdsnü piotoirt dependant (Blub giey) / Kill B importer BtuC D waîh four

individiidüv expressed ssibuniK <2xBluC blue 2\liUiI) or tngei b, binding

piofciti dependant (RbsB dark gte\) nbost fnipoitoi RbsAC with two idcnficai

TVIDs (Rbst blue) uid iwo fused NBDs to one polvpeptuie ( ham fRbsA ted) t,

huHfirt rrndtiUriif. ABC lull ttmspoitu \kiri (Mus ) with ill lout dorn sins kised

!o one polvpeptKL chain d. heletodiiiietie bun an antigenic peptide hsif

tunsportei Fapl !"ap2 (bkie org mu ) with two diikicnt subtimts e it h ton-astiipg

oi in N luminal Î MI) that is jused to »( itrmind NBD

1 ransmembranc Domains

I he bis» multidrug ABC. transporta «rchitectuie comprises two hydropnobic

transmembrane domains (IMDs) tha' piovide s translocation pathway for the

transport substrates I he V and C-termini of tru Î MDs are both located in the

4

cytoplasm «ntl each fMD consists s.! six ti iki^a1 ùansnicmbïaiio helices that

lias use llis membrane (I îgure î >) The individual membrane heln.es ait connected

via two iihiJLcllulai loops m the cytoplasm ànà truu exttacciiuhi loops Lhrt ait

sxposcd to the periplasm or extracellular space Soin* muHîdntg ABC îiansporttï •>

contain additional membrane helices to the eoie anani*emet]t ol twelve I V! a helices

like the Multidrug Resistance Protein 1

I - ? * \v , 1 \ *\ "r 1

I igiirt ! 1 Itpcio^v i lucid ot I rrA Schemii t itpicseni iljo« ol une suhuriit oi

thv hopiedioivr <. npilfidiu AHC I tit ti tnsposfer fron / f m I sch si burnt

toi sists <>' »i \ tei nm tl TWO that is fused to i t tuns mi NBP I Ik

Ii m JocrfUoti ptfhvviv <.f ins fh< neitihi me (gic\) Hid urn i>>Ss < i Iwt Inns. si\

fr tn-^mcmbi tue (TVf) jit.lii.es in the dimu « teheed fiom K!

I x^ept foi the vitamin B5 impoilei fiooi î ioh Btut D}and the close homolog

III 1470/1°

(I iguie- I 4a b) no experimental strut, an a! data at high nsohition is

tl\ lilabit that punodes inhumation about the topology ol the iransmcrsbiaiie helites

oi She FMDABD uiteitate m \B( tiarspoiieis in tiis ease ot BtuCt) and SI! 1470 1

a tore ol <wcntv instead of twehe transmembrme helices spans the membrane

sndiutmir the hnktiona! and sttuttu ai vanetv ot the ditferent tianspoii wstcnis

fuitheimore neatly no sequence sinnltnt) within die transmembrane, sci2'lients ot

MDR ABC bansportci-s compared to othei ABC. transporters s1-onsen ed

*>

1 sgsm î 1 Ki'iboii itp estntafion of the AB( tmnoiLi prolans â / col Ht K f) it

i1 A is.solution tnd the dost, hu ïioUu» I) // ufltiahM 1ÏI14^0 i it > -1 K

i„solution u tht micicolüc tie,, stiL, Both piotems consist «I lui» indi\idiid!

dnm »us fvo identical FMD isirt«. mannt) md two identic si \B!)^

'bn^htoMj t orwjît) Viuebi tnc bound incs He iodic jlcci Silt ti m-dot ition

[Hti v, i\ i1- formed h) ">() fxiO) Ï Vi ft h hecs tin! spin (It munbniK md

dist nttîv üiikusit l'on ABC cxi ortcis with tust, of 12 TM hJtce«

Within the MDR VBf" tiansposfc! family sequence snrulantits and identities a>e

piescnl but still lather iov\ Given the view that MDR \BC ttanspottcis NBDs and

I MIX ha\e developed horn uimmon ancestei scries the low ovctai! sequence

similarities hkelv reflect an cuduhoika"), adaptation to the bioad diversity ui the

substt ato ifowevcr it was suggested that a ectam memhiane topology tan he

established m MDR-ÀBC tumspoiîeis on the basis ol alternativ«, amino acid

combinations ma süuctuul constraints that ongniate Jrom the common evoiutionaiy

roots Independent nom sequence analysis results it could he demonstratt d that the

nnpoitant sites foi the suhstsate fuinspoit aie located within die tiansmembrane

domains of the MDR- \Bt ttanspoitcrs No wed detined snbstiale binding sites

ha\e been identified in A,B( exporters but the binding oi up io two -.ubstiates to the

human MDR- \B( uanspoUeis Mch Î and MRP? has been observed"'

In addition

6

residues from transmembrane hehces [Ml. TM4-IM6, and 1 MIO-TM 12 have been

implicated in dnig binding to human Mdri'9and residues from multiple helices ha\c

been found to contribute H> substrate binding to Fapi/T*

Although low resolution electron microscopy studies of human Mdri' *'

and the

close MDR-ABC transport« homolog CUR' levealcd the overall shape of the

protein and cross-linking experiments on Mdri identified important heh\-hehx

interactions,the exact topology of the transmembrane helices and the locations of

contact residues to the NBDs al the shared mteifaee remained elusive

!\ ticieotide-hinding Domains

She nucleotide-binding domains aie the motor domains of the ABC transporte;

machinery and provide the energy lor the transport of substrate The two hydrophihe

domains are typically located in the cytoplasm at the membrane boundary and hind to

their cognate transmembrane domains at the cytosohe lace The typical ABC cassette

subunit consists of the catalytic core ATPase domasn (ReeA-hke domain) that

contains the P-loop (Walker A) and die a-hchtal (1 »-ATPase like) subdoniam that

contains the ABC signature sequence'"' Ï he N'BDs exhibit a much high« degiee of

eonseivation than the IMDs m ABC exporters and importais I his is mostly because

o! highly conserved sequence mohls that are involved in binding and hydrolysis of

AIP aï the shared inteifaee between two NRDs* Pxamples are the mechanistically

iclevant residues like the P-loop (Walker A), ABC signatuie sequence (LSGUQ),

Q-loop and A loop In addition, residues from the f)~!oop and Q-loop are involved in

NBD-NRD domain mteiaetions'^ that stabih/e the dimer and NBD-1VID domain

interactions, respeclivelv" At the shared interlace. P-loop and \BC signature

sandwich two AIP molecules, pixtapo.se each other and additionally coordinate a

catalytic water molecule and a eo-factoi ion close to the hydroiysabie y-phosphate oi

the ATP Hie resulting nead-to-taii arrangement of the NBDs has been widely

accepted as the physiological télevant eontormatton on the basis of struclural and

biochemical'1 '

data i\ sgisrc 1 5b) I he sandwiching' of two AIP molecules

between the P-loop of one and the AB( signatuie sequence of the othei NBD cieales

a direct link between the Swo sites and provides the molecular basis of the observed

eooperafivHy in AT P binding and hydrolysis"'%

Biochemical analysis of the A FPaiu

activity of ARC transporter proteins demonstrated a dependency from the pi I, with

activity in a pi I range of about o~9s2 "A

7

s

I igun 1 ^ Ribbon iu~»rcsent<ifiop ol Sh< AB<" uissUtt dimus 4. Sop vjvw »n îi v

\HJ) oi ir>sfif s1îmc!uic of (he ni! it) ABC impeilo // uifhumcK H!!4"0 1 in

11^. nutitoini htt LOîifour ilior f ipui stau ) b, t sysUi! stiutttut ol fist \BD

dtmu oi M icmtiasthii Vlf07^6 in the ATP inchvi h dimes <,*.lo'-.cd sink) The

nutkotuk landing motifs i'ioop (V\ i!ku A biuv) smi \\M. snjvAux., stquti tt

(sup it the, shaitd mtuiut dit hsLhiiyhted ih notils au iloses in ths, pit sente

< I \ I T <b ill tiiu! stidA than m its ibstnt

In the nucicotide-itet stale the two niicieotide~bmdin|> domains separate m the

functional üanspoitu h\ seveial Angstroms7',0 '

the\ »cvei dissociate and

iheiefoie exhibit an increased solvent accessibilityv

(I igure i 5a) In the absence

oi AFP the cHiehcal subdomaui is slutted lcldtne to the A I Fast sobdomatn as

obscivcd in hieh resolution tivsta! stiuetmes if "was suggested that the relative

moscmen* ueates energy that is likely liansmiUcd to the IMDs to induce the

touiormation.ii changes needed foi the tianslocalion leactson in the ti\s<al structure

oi the impoitci Btu( I), a cleft m each of the NBIX was observed that iiaiboi> one ol

the two a-heheal sketches oi the L loop of the f MÜs I he second I loop hchx in this

motif tighih inlet acts with the NBI) luriace The Î -loop was suppested hut not

expelimontaiiv proven to lepiesent a geneial inleitace between NHPs and FVIDs ol

\B( importers and exporfers

\B( 1 ransporiet s and Mechanistic Aspects

f o understand the mechanism oi membrane piotems it is important to experimental^

investigate oioiecühtr cLtails of the iunctîon and structure and io combine the

informai«on in a mechanistic model The mechanistic scheine leniesents a simplified

N

explanation for how the 'molecular machine' works and provides the basis for future

investigations.

uytopiasn I inside'

a bLdLlOif >-> (jJF

I ,i m«" h u ir t'

Pt r pldsrïi iuuîmJH

Figure 1.6: Scheme of the 'alternating access and release mechanism' on the

examples of a, the lactose permease symportcr LacY and b, the inoiganic

phosphate (PJ /glycerol-3-phosphate antiporter GlpT (depicted from59) c, The

scheme presents the original drawing ofthe simple allostenc model ofa membrane

pump (depicted from60)

Active transport of substrates across membranes was theoretically approached in the

1960s and described as an allosteric rearrangement of two alternative conformations.

The energy required for the conversion would be derived from chemical reactions,

photochemical reactions or from the free energy stored in electrochemical gradients or

pressure. The model is known as 'alternating access and release mechanism'

(Figure 1.6c) and gives a simple explanation for unidirectional transport across

membranes. Experimental data obtained for major facilitator superfamily (MFS)

proteins like LacY and GlpT from E. coli59 appeared to be consistent with the model.

Driven by an electrochemical gradient, MFS proteins translocate substrates against a

concentration gradient. The proteins expose a substrate-binding site that is

alternatively accessible to one side of the membrane or the other. This is possible

because the MFS proteins adopt an inward- and an outward-facing conformation, both

being convertible energetically facilitated by substrate binding (Figure 1.6a,b).

ABC transporters have been studied extensively on a genetic and functional level but

limited structural information, mostly of the nucleotide-binding domains, is available.

It was generally accepted that ATP binding and hydrolysis in the nucleotide-binding

domains provided the energy for the translocation of substrates24, 2Sbut how the

energy was transmitted to the translocation pathway (TMDs) was not understood. A

9

uniform mechanism has never emerged but several mechanistic models tried to

explain the molecular function of ABC exporters. One example is the alternating two-

site (two-cylinder engine) transport model that described the oscillation of two

configurations within one ABC transporter containing a high-affinity and a low-

affinity binding site61. The model was mainly based on functional data and had the

intrinsic problem that structural data of the full transporter unit was missing.

Emerging structural data of isolated nucleotide-binding domains allowed early

speculations about the conformational changes within the NBDs that are transmitted

to the transmembrane domains.

<rt

I AIM'P ADpi AIP i

: tb*

'dp -r w

aip'adp p^aup-p i !! I

ou •• itra tn 1

Figure 1.7: Alternating two-site transport model (according to61). Rectangles

represent the TMDs; circles, squares and hexagons the NBDs in different

conformations. The ATP-bound state (circle) is associated with a high-affinity

binding site, the ADP/Pi-bound state with an occluded binding site and the ADP-

bound state with a low-affinity binding site as indicated by a black ball, grey ball

and a black ellipsoid. Arrow a: initial substrate-binding event; arrow b: second

substrate-binding event. The scheme explains the conversion of a low-affinity

binding site to a high-affinity binding site in combination with different

conformations of the nucleotide-binding domains within one transporter. Arrow 1

and 2: reversed relationship of the two domains or cylinders' in the dinier during

transport. ATP hydrolysis in one 'cylinder' is coupled to substrate efflux in the

same 'cylinder' and to ATP binding at the other 'cylinder'. A high-affinity inside-

facing substrate binding site is converted to a low-affinity binding site.

In 2001 and 2003 - before the start of the thesis - two crystal structures of full ABC

exporters, the lipid A transporter MsbA from E. coli and V. cholerae have been

10

,62, 63

published that indicated a dissociation of NBDs in a nucleotide-free state'

''

.This

'flip-flop' model described a large chamber formed by two TMDs that was open to

the cytoplasm in the absence of nucleotide. Upon binding of substrate, the binding of

nucleotide would be induced and the NBDs would align and dimerize. Further support

for their mechanistic model was provided by the same group in 2005 with a third

structure of MsbA from S. typhimurium64 but already at that time the structures have

been criticized as non-physiologic due to improperly arranged NBDs65. Experimental

proof for a complete dissociation of the NBDs of an intact ABC transporter in a native

environment has never been provided.

Figure 1.8: MsbA 'flip-flop' mechanism (according to' ). Dissociated NBDs in the

nucleotide-free conformation oscillate between conformational states with rotating

NBDs. Upon binding of substrate (1) induces binding of ATP and the

dimerization. The substrate flips spontaneously to its new location (2) and is

released to the outer leaflet of the membrane. ADP and Pi diffuse away and reset

the transporter to the dissociated state (3).

11

Objective of the thesis

The goal of the thesis was to determine the first high resolution crystal structure of a

bacterial multidrug ABC exporter in a physiological relevant state. Since much is

known about ABC transporters, this was needed for the understanding of the

mechanism of substrate transport that is coupled to the binding and hydrolysis of

ATP. At the beginning of the thesis, two crystal structures of a complete bacterial

ABC exporter - MsbA from E. coli and from V. cholerae - were available and later,

in 2005, a third structure of MsbA from S. typhimurium was published. However, the

three crystal structures caused confusion about their correctness and physiological

relevance and following the results of this thesis all three structures have been

retracted.

The thesis describes the development of a method to screen for active and pure

multidrug ABC exporter protein in detergent applying a homology screening

approach (Chapter 2). Several multidrug ABC exporter proteins have been expressed

and purified and the most promising ABC exporter - Savl866 from S. aureus -

selected. In chapter 3, the expression, purification, crystallization and structure

determination of Savl866 is demonstrated in detail and subsequently the structure of

Savl866 is presented in Chapter 4. The structural model of Savl866 is described and

discussed in a functional context. The value of the structure for the understanding of

the mechanism is highlighted by comparing the results with data from the literature

and with the structures of MsbA. In Chapters, the multidrug ABC transporter

Savl866 is characterized on a functional level. The ATPase activity of Savl866 is

investigated in the presence and absence of putative transport substrates.

References

1. Holland, I. B., Cole, S. P. C, Kuchter, K. & Higgins, C. F. ABC proteins: from

bacteria to man (Academic Press, Amsterdam, 2003).

2. Dean, M. The genetics of ATP~binding cassette transporters. Methods Enzymol

400, 409-29 (2005).

3. Ames, G. F. & Roth, J. R. Histidine and aromatic permeases of Salmonella

typhimurim. JBacteriol 96, 1742-9 (1968).

4. Higgins, C. F., Hiles, I. D., Salmond, G. P., Gill, D. R., Downie, J. A., Evans, I. J.,

Holland, I. B., Gray, L., Buckel, S. D., Bell, A. W. & et al. A family of related

ATP-binding subuuits coupled to many distinct biological processes in bacteria.

12

Nature 323, 448-50 (1986).

5. Ames, G. F., Mimura, C. S. & Shyamala, V. Bacterial periplasmic permeases

belong to a family of transport proteins operating from Escherichia coli to human:

Traffic ATPases. FEMS Microbiol Rev 6, 429-46 (1990).

6. Ames, G. F. & Lecar, Fl. ATP-dependent bacterial transporters and cystic fibrosis:

analogy between channels and transporters. FasebJb, 2660-6 (1992).

7. Higgins, C. F. ABC transporters: from microorganisms to man. Annu Rev Cell

Ä/o/8,67-113(1992).

8. Kelly, A., Powis, S. H., Kerr, L. A., Mockridge, I., Elliott, T., Bastin, J.,

Uchanska-Ziegler, B., Ziegler, A., Trowsdale, J. & Townsend, A. Assembly and

function of the two ABC transporter proteins encoded in the human major

histocompatibility complex. Nature 355, 641-4 (1992).

9. Lankat-Buttgereit, B. & Tampe, R. The transporter associated with antigen

processing: function and implications in human diseases. Physiol Rev 82, 187-204

(2002).

10. Gottesman, M. M., Fojo, T. & Bates, S. E. Multidrug resistance in cancer: role of

ATP-dependent transporters. Nat Rev Cancer 2, 48-58 (2002).

11. Sheppard, D. N. & Welsh, M. J. Structure and function of the CFTR chloride

channel. Physiol Rev 79, S23-45 (1999).

12. Gadsby, D. C, Vergani, P. & Csanady, L. The ABC protein turned chloride

channel whose failure causes cystic fibrosis. Nature 440, 477*83 (2006).

13. Gottesman, M. M. & Ambudkar, S. V. Overview: ABC transporters and human

disease. JBioenerg Biomembr 33, 453-8 (2001).

14. Burchenal, J. H., Robinson, E., Johnston, S. F. & Kushida, M. N. The induction of

resistance to 4-amino-N 10-methylpteroylglutamic acid in a strain of transmitted

mouse leukemia. Science 111, 116 (1950).

15. Dano, K. Active outward transport of daunomycin in resistant Ehrlich ascites

tumor cells. Biochim Biophys Acta 323, 466-83 (1973).

16. Szakacs, G., Paterson, J. K., Ludwig, J. A., Booth-Genthe, C. & Gottesman, M.

M. Targeting multidrug resistance in cancer. Nat Rev Drug Discov 5, 219-34

(2006).

17. Nikaido, H. Prevention of drug access to bacterial targets: permeability barriers

and active efflux. Science 264, 382-8 (1994).

18. Kontoyiannis, D. P. & Lewis, R. E. Antifungal drug resistance of pathogenic

13

fungi. Lancet 359, 1135-44 (2002).

19. Sanglard, D. & Odds, F. C. Resistance of Candida species to antifungal agents:

molecular mechanisms and clinical consequences. Lancet Infect Dis 2, 73-85

(2002).

20. Seelig, A., Blatter, X. L. & Wohnsland, F. Substrate recognition by P-

glycoprotein and the multidrug resistance-associated protein MRP1: a

comparison. IntJ Clin Pharmacol Ther 38, 111-21 (2000).

21. Allen, J. D., Brinkhuis, R. F., van Deemter, L., Wijnholds, S. & Schinkel, A. H.

Extensive contribution of the multidrug transporters P-glycoprotein and Mrp 1 to

basal drug resistance. Cancer Res 60, 5761-6 (2000).

22. van Veen, H. W., Callaghan, R., Soceneantu, L., Sardini, A., Konings, W. N. &

Higgins, C. F. A bacterial antibiotic-resistance gene that complements the human

multidrug-resistance P-glycoprotein gene. Nature 391, 291-5 (1998).

23. Hyde, S. C, Emsley, P., Hartshorn, M. J., Mimmack, M. M, Gileadi, U., Pearce,

S. R., Gallagher, M. P., Gill, D. R., Hubbard, R. E. & Higgins, C. F. Structural

model of ATP-binding proteins associated with cystic fibrosis, multidrug

resistance and bacterial transport. Nature 346, 362-5 (1990).

24. Schneider, E. & Hunke, S. ATP-binding-cassette (ABC) transport systems:

functional and structural aspects of the ATP-hydrolyzing subunits/domains. FEMS

Microbiol Rev 22, 1 -20 ( 1998).

25. Holland, I. B. & Blight, M. A. ABC-ATPases, adaptable energy generators

fuelling transmembrane movement of a variety of molecules in organisms from

bacteria to humans. JMol Biol 293, 381-99 ( 1999).

26. Dassa, E. & Bouige, P. The ABC of ABCS: a phylogenetic and functional

classification of ABC systems in living organisms. Res Microbiol 152, 211-29

(2001).

27. Biemans-Oldehinkel, E., Doeven, M K. & Poolman, B. ABC transporter

architecture and regulatory roles of accessory domains. FEBS Lett 580, 1023-35

(2006).

28. Konings, W. N. & Poelarends, G. J. Bacterial multidrug resistance mediated by a

homologue of the human multidrug transporter P-glycoprotein. IUBMB Life 53,

213-8 (2002).

29. Locher, K. P., Lee, A. T. & Rees, D. C. The E. coli BtuCD structure: a framework

for ABC transporter architecture and mechanism. Science 296, 1091 -8 (2002).

14

30. Pinkett, H. W., Lee, A. T., Lum, P., Locher, K. P. & Rees, D. C. An Inward-

Facing Conformation of a Putative Metal-Chelate-Type ABC Transporter. Science

(2006).

31. Ames, G. F.-L. & Higgins, C. F. The organization, mechanism of action, and

evolution of periplasmic transport systems. Trends in Biochemical Sciences 8, 97-

100(1983).

32. Ames, G. F. Bacterial periplasmic transport systems: structure, mechanism, and

evolution. Annu Rev Biochem 55, 397-425 (1986).

33. Dey, S., Ramachandra, M., Pastan, I., Gottesman, M. M. & Ambudkar, S. V.

Evidence for two nonidentical drug-interaction sites in the human P-glycoprotein.

Proc Natl Acad Sei USA94, 10594-9 (1997).

34. Ambudkar, S. V., Dey, S., Hrycyna, C. A., Ramachandra, M., Pastan, I. &

Gottesman, M. M Biochemical, cellular, and pharmacological aspects of the

multidrug transporter. Annu Rev Pharmacol Toxicol 39, 361-98 (1999).

35. Pleban, K., Kopp, S., Csaszar, E., Peer, M., Hrebicek, T., Rizzi, A., Ecker, G. F. &

Chiba, P. P-glycoprotein substrate binding domains are located at the

transmembrane domain/transmembrane domain interfaces: a combined

photoaffinity labeling-protein homology modeling approach. Mol Pharmacol 67,

365-74 (2005).

36. Higgins, C. F. & Linton, K. J. The ATP switch model for ABC transporters. Nat

Struct Mol Biol 11, 918-26 (2004).

37. Shapiro, A. B. & Ling, V. Positively cooperative sites for drug transport by P-

glycoprotein with distinct drug specificities. Eur JBiochem 250, 130-7 (1997).

38. Zelcer, N., Huisman, M. T., Reid, G., Wielinga, P., Breedveld, P., Kuil, A.,

Knipscheer, P., Schellens, J. H., Schinkel, A. H. & Borst, P. Evidence for two

interacting ligand binding sites in human multidrug resistance protein 2 (ATP

binding cassette C2). JBiol Chem 278, 23538-44 (2003).

39. Loo, T. W. & Clarke, D. M. Recent progress in understanding the mechanism of

P-glycoprotein-mediated drug efflux. J Memhr Biol 206, 173-85 (2005).

40. Nijenhuis, M. & Hammerling, G. J. Multiple regions of the transporter associated

with antigen processing (TAP) contribute to its peptide binding site. J Immunol

157,5467-77(1996).

41. Rosenberg, M. F., Velarde, G., Ford, R. C, Martin, C, Berridge, G., Kerr, I. D.,

Callaghan, R., Schmidlin, A., Wooding, C, Linton, K. J. & Higgins, C. F.

15

Repacking of the transmembrane domains of P-glycoprotein during the transport

ATPase cycle. Embo J20, 5615-25 (2001).

42. Lee, J. Y., Urbatsch, J. L., Senior, A. E. & Wilkens, S. Projection structure of P-

glycoprotein by electron microscopy. Evidence for a closed conformation of the

nucleotide binding domains. J Biol Chem 277,40125-31 (2002).

43. Rosenberg, M. F., Callaghan, R., Modok, S„ Higgins, C. F. & Ford, R. C. Three-

dimensional structure of P-glycoprotein: the transmembrane regions adopt an

asymmetric configuration in the nucleotide-bound state. J Biol Chem 280, 2857-

62 (2005).

44. Awayn, N. H., Rosenberg, M. F., Kamis, A. B., Aleksandrov, L. A., Riordan, J. R.

& Ford, R. C. Crystallographic and single-particle analyses of native- and

nucleotide-bound forms of the cystic fibrosis transmembrane conductance

regulator (CFTR) protein. Biochem Soc Trans 33, 996-9 (2005).

45. Stenham, D. R., Campbell, J. D., Sansom, M. S., Higgins, C. F., Kerr, I. D. &

Linton, K. J. An atomic detail model for the human ATP binding cassette

transporter P-glycoprotein derived from disulfide cross-linking and homology

modeling. Faseb J17, 2287-9 (2003).

46. Hung, L. W., Wang, I. X., Nikaido, K., Liu, P. Q., Ames, G. F. & Kim, S. H.

Crystal structure of the ATP-binding subunit of an ABC transporter. Nature 396,

703-7 (1998).

47. Hopfner, K. P., Karcher, A., Shin, D. S., Craig, L., Arthur, L. M., Carney, J. P. &

Tainer, J. A. Structural biology of Rad50 ATPase: ATP-driven conformational

control in DNA double-strand break repair and the ABC~ATPase superfamily.

Cell 101, 789-800 (2000).

48. Smith, P. C, Karpowich, N., Milien, L., Moody, J. E., Rosen, J., Thomas, P. J. &

Hunt, J. F. ATP binding to the motor domain from an ABC transporter drives

formation of a nucleotide sandwich dimer. Mol Cell 10, 139-49 (2002).

49. Hrycyna, C. A., Ramachandra, M., Ambudkar, S. V., Ko, Y. H., Pedersen, P. L.,

Pastan, T. & Gottesman, M. M. Mechanism of action of human P-glycoprotein

ATPase activity. Photochemical cleavage during a catalytic transition state using

orthovanadate reveals cross-talk between the two ATP sites. J Biol Chem 273,

16631-4(1998).

50. Chen, J., Sharma, S., Quiocho, F. A. & Davidson, A. L. Trapping the transition

state of an ATP-binding cassette transporter: evidence for a concerted mechanism

16

of maltose transport. Proc Natl Acad Sei U S A 98, 1525-30 (2001).

Sl.Vergani, P., Lockless, S. W., Nairn, A. C. & Gadsby, D. C. CFTR channel

opening by ATP-driven tight dimerization of its nucleotide-binding domains.

Nature 433, 876-80 (2005).

52. Davidson, A. L., Laghaeian, S. S. & Mannering, D. E. The maltose transport

system of Escherichia coli displays positive cooperativity in ATP hydrolysis. ./

Biol Chem 271,4858-63 (1996).

53. Senior, A. E. & Bhagat, S. P-glycoprotein shows strong catalytic cooperativity

between the two nucleotide sites. Biochemistry 37, 831-6 (1998).

54. Zaitseva, J., Jenewein, S., Jumpertz, T., Holland, I. B. & Schmitt, L. H662 is the

linchpin of ATP hydrolysis in the nucleotide-binding domain of the ABC

transporter HlyB. Embo J 24, 1901-10 (2005).

55. Chen, J., Lu, G., Lin, J., Davidson, A. L. & Quiocho, F. A. A tweezers-like

motion ofthe ATP-binding cassette dimer in an ABC transport cycle. Mol Cell 12,

651-61 (2003).

56. Mannering, D. E., Sharma, S. & Davidson, A. L. Demonstration of

conformational changes associated with activation of the maltose transport

complex. JBiol Chem 276, 12362-8 (2001).

57. Mourez, M., Hofnung, M. & Dassa, E. Subunit interactions in ABC transporters: a

conserved sequence in hydrophobic membrane proteins of periplasmic permeases

defines an important site of interaction with the ATPase subunits. Embo J 16,

3066-77(1997).

58. Cotten, J. F., Ostedgaard, L. S., Carson, M. R. & Welsh, M. J. Effect of cystic

fibrosis-associated mutations in the fourth intracellular loop of cystic fibrosis

transmembrane conductance regulator. J Biol Chem 271, 21279-84 (1996).

59. Abramson, J., Kaback, H. R. & Iwata, S. Structural comparison of lactose

permease and the glycerol-3-phosphate antiporter: members of the major

facilitator superfamily. Curr Opin Struct Biol 14, 413-9 (2004).

60. Jardetzky, O. Simple allosteric model for membrane pumps. Nature 211, 969-70

(1966).

61. van Veen, H. W., Margolles, A., Müller, M., Higgins, C. F. & Konings, W. N.

The homodimeric ATP-binding cassette transporter LmrA mediates multidrug

transport by an alternating two-site (two-cylinder engine) mechanism. Embo J 19,

2503-14(2000).

17

62. Chang, G. & Roth, C. B. Structure of MsbA from E. coli: a homolog of the

multidrug resistance ATP binding cassette (ABC) transporters. Science 293, 1793-

800 (2001).

63. Chang, G. Structure of MsbA from Vibrio cholera: a multidrug resistance ABC

transporter homolog in a closed conformation. J Mol Biol 330, 419-30 (2003).

64. Reyes, C. L. & Chang, G. Structure of the ABC transporter MsbA in complex

with ADP.vanadate and lipopolysaccharide. Science 308, 1028-31 (2005).

65. Davidson, A. L. & Chen, J. Structural biology. Flipping lipids: is the third time the

charm? Science 308, 963-5 (2005).

18

Chapter 2: Recombinant Expression, Purification and Activity

Screening of Bacterial MDR-ABC transporters

Abstract

Several bacterial multidrug ABC transporter proteins were expressed in E. coll,

solubilized, purified in detergent solution and analyzed for ATPase activity. Protocols

for the individual preparation steps had to be established to produce large amounts of

protein suitable for crystallographic experiments. To find optimal solubilization and

purification conditions various detergents were applied and the ATPase activity of the

individual MDR-ABC transporters in the different detergents was analyzed. Although

several zwitterionic detergents produced stable protein with yields of >0.6 mg

protein/g wet cell pellet for all transporters, ATPase activity assays of purified and

reconstituted protein suggested an irreversible inactivation of the transporter.

Extraction with non-ionic detergents DDM and Ci2Eg yielded -0.4 mg protein/g wet

cell pellet but demonstrated V04-inhibitable, basal ATPase activity for two MDR-

ABC transporters. Further analysis investigating the stability of the homodimeric

proteins showed that only C^Es-solubilized and purified S. aureus Savl866 remained

stable at higher concentrations proving suitable for macromolecular crystallography.

Introduction

Macromolecular crystallography of membrane proteins requires experimental efforts

because of their dynamic nature in detergent. The replacement of the native lipid

environment with detergent is very often the reason for the loss of function, fast

aggregation of the target proteins or inhibition of the formation of well-ordered three-

dimensional crystals. This is mostly influenced by the properties of the various

detergents in aqueous solution and specific interactions with the membrane protein.

Although generalizations are reported in the literature1, 2, the work with non-

characterized membrane proteins and detergents is still a challenging process. Tn

detergent solution, membrane proteins are more fragile and protocols for stabilization

and purification have to be established and optimized for each and individual

membrane protein.

The aim of this approach was the expression, solubilization and purification of stable

and active multidrug ABC transporter homologs for crystallographic purposes. Yields

19

of more than 0.1 mg protein/g wet cell pellet of pure and active membrane protein had

to be achieved and the stability in detergent solution granted at concentrations of

15 mg/ml. On the basis of standard molecular biological methods, protocols for the

expression, solubilization and purification had to be established. As an important part

for the screening process, the basal ATPase activity of the MDR-ABC transporter

homologs was analyzed. Numerous different detergents had to be applied to find the

appropriate buffer conditions at each of the purification steps. To improve the

likelihood of success3, a homology screening approach was performed4'5 and several

homodimeric multidrug ABC 'half-transporter' genes from different bacterial sources

were carefully selected and screened.

Materials and Methods

Cloning of Bacterial MDR-ABC Transporters on the example S. aureus Savl866

The gene encoding full-length Staphylococcus aureus Savl866 was amplified from

genomic DNA (ATCC no. 700699D-S) by polymerase chain reaction. The primers

SAV1866N19 (5'-CGA TCA TAT GAT TAA ACG ATA TTT GCA ATT TG-3') and

SAV1866C19 (5'-CGA TGG ATC CTT ATA AGT TTT GAA TGC TAT ATA AAT GC-3')

yielded a DNA fragment with a 5'-NdeI and a 3'-BamHI single restriction site that

was purified from a 1 % (w/v) agarose gel with the Qiaex II gel extraction kit

(Qiagen) and subsequently cloned into a pCR-Blunt 1I-TOPO vector with the Zero

Blunt TOPO PCR Cloning Kit (Invitrogen). Escherichia coli DH5a chemical ultra-

competent cells were prepared essentially as described in6 with slight modifications.

Transformed cells were cultured and the plasmid DNA isolated using the QIAprep

Spin Miniprep Kit (Qiagen). The plasmid product was fractionally digested with the

restriction enzymes Ndel and BamHI (NEB), purified and the DNA fragment was

ligated into a similarly digested, modified pet 19b expression vector (Novagen)7"9. The

resulting plasmid contained the strong E. coli T7 phage promoter controlled by the

L8-UV5 lac operator followed by full-length Savl866 with amino-terminal

decahistine affinity tag. The analogous cloning strategy was applied for all other ABC

transporter homologs tested in this study. These are LmrA from Lactococcus lactis,

YvcC, YwjA and YgaD from Bacillus subtilis and Sav0643 from S. aureus. All

plasmids were verified by DNA sequencing (Microsynth) and results were analyzed

using Chromas Lite 2.0 and Clustal W(http://www.ebi.ac.uk/clustalw). Physico-

20

chemical parameters were calculated with ProtParam (http://www.expasy.org/tools/

protparam.html) and sequences were derived from GenBank, Swiss-Prot and

TrEMBL using the ExPASy proteomics server10.

Recombinant Expression of MDR-ABC Transporters in E. colt

All MDR-ABC transporters analyzed in this study were overexpressed in E. coli

BL21-codon plus (DE3)-RIPL expression strain (Stratagene). Competent, untreated

cells were stored for a maximum of 3 months at -80 °C in a Revco chest freezer

(Thermo Scientific). Expression levels decreased dramatically in cells stored for

longer than three months when cultured in a fermentor. Expression cultures invariably

originated from fresh transformations and single colony picking.

Initial expression tests were performed for all transporters in Terrific Broth rich

medium supplemented with glycerol, glycine or glucose at 1 % (w/v) concentration as

an additional carbon source. Cells from fresh transformations (LBAmp plates) were

cultured at 37 °C and target proteins expressed by induction with 0.4 mM isopropyl-

ß-D-thiogalactopyranoside (IPTG) at temperatures of 37 °C or 20 °C. The production

of protein was followed by SDS-PAGE using Coomassie staining and western-

blotting using the antibodies anti His5 (Qiagen) and alkaline phosphatase (Rockland)

for transport protein detection.

For large scale protein production, cells from fresh transformations (LBAnipGiucosc

plates) were cultured in Terrific Broth medium supplemented with 1 % (w/v) glucose

in 500 ml volumes. At an optical density (A6oo) of ~1, cells were transferred to a 10

liter Techfors S fermentor (Infors) and cultured at 37 °C. The expression of the

Savl866 gene was induced with 0.4 mM IPTG at an optical density of 12-15 for

1.5 hours. Cells were harvested by centrifugation and stored as pellets at -80 °C.

Extraction and Purification of Detergent-solubilized MDR-ABC Transporters

The subsequent procedures were invariably performed at 4 °C unless specified

differently. Frozen cell pellets were resuspended by stirring in a buffer containing

50 mM Tris (HCl) pH 7.5 and 500 mM sodium chloride and disrupted using a

M-110L microfluidizer (Microfluidics) at 15000 psi external pressure. Subsequently,

the cell suspension was directly ultracentrifuged at 100,000xg and resulting crude cell

membranes were resuspended in a buffer containing 50 mM Tris (HCl) pH 7.5,

25 mM imidazole (HCl) pH 8.0 and 600 mM sodium chloride. Alternatively, a buffer

21

containing 100 mM sodium phosphate (NaH2P04) pH 8.0, 20 mM imidazole (HCl)

pH 8.0 and 200 mM sodium chloride was used. Crude membranes corresponding to

10 g of wet cell pellet were resuspended in a volume 20 ml and snap-frozen in liquid

nitrogen. For the extraction tests, 20 ml of membrane suspension was thawed,

detergent to a final concentration of 1 % (w/v) was added and the mix was stirred for

1-2 hours at 4 °C and room temperature, respectively. The solubilization mix was

sonicated 3 times for 1 minute with a small sonication tip and moderate sonication

strokes. Solubilized protein was separated from the cell debris by centrifugation at

40000xg and purified in a one-step procedure using 15 ml NiNTA superflow resin

(Qiagen), a XK16 chromatography setup (GE Healthcare) and standard

chromatography techniques. Pure and biochemically stable protein was transferred by

desalting with a HiPrep Desalting 26/10 column (GE Healthcare) into a buffer

containing 50 mM Hepes (NaOH) pH 8.0, 150 mM sodium chloride, 5 mM

ß-mercaptoethanol and the corresponding detergent in concentrations above the

critical micelle concentrations (CMC). The proteins were concentrated to 15 mg/ml

using an Amicon Ultra-15 concentrator unit (Millipore) with a molecular cutoff of

100 kDa. The concentration was determined by UV-absorption spectroscopy and the

individual extraction and purification steps were analyzed by SDS-PAGE, western-

blot and size exclusion chromatography using a Superdex 200 10/300 GL column (GE

Healthcare).

ATPase Activity Assay

The ATPase activity of detergent solubilized and purified MDR-ABC transporters

was measured at room temperature according to11 in reaction volumes of 350 ul.

Reactions of 1 uM protein were prepared in desalting buffer and incubated at room

temperature for 3 minutes. ATP hydrolysis was initiated adding 2 mM ATP and

10 mM MgCl2 to the mixture and stopped by SDS inhibition. Therefore 50 ul samples

were removed at various time points and added to 50 u.1 of a 12 % (w/v) SDS solution.

For ATP hydrolysis inhibition, 1 mM freshly boiled sodium ortho-vanadate was

added to the solution. Inorganic phosphate was assayed by a modified molybdate

method12. The effect of zwitterionic detergents FOS12, FOS14 and LDAO on

Savl866 and YvcC ATP hydrolysis was assayed adding various sub-CMC and CMC

concentrations of the detergents to the assay buffer solution.

22

Reconstitution of detergent-purified ABC Transporters into Proteoliposomes

Purified protein was reconstituted as described earlier" based on previously

developed protocols13. Chloroform dissolved Egg yolk L-R-phosphatidylcholine

(Avanti Polar Lipids) and E. coli polar lipid extract (Avanti Polar Lipids) were mixed

at a ratio of 1:3 and the organic solvent was evaporated at room temperature under an

argon stream. Residual solvent was removed by lyophilization and subsequently the

mixture was hydrated in a round bottom flask by water bath sonication in a buffer

containing 50 mM Tris (HCl) pH 7.5 and 150 mM NaCl. The final lipid concentration

was 20 mg/ml. The turbid suspension was extruded with a mini extruder (Avanti Polar

Lipids) with 400 nm extrusion filters, flash frozen in liquid nitrogen and stored at -

80 °C.

Lipids were thawed and pre-incubated with 0.14 % (w/v) Triton X-100 for 1 hour at

room temperature. Freshly purified protein in 50 mM Tris (HCl) pH 7.5, 150 mM

NaCl, 0.5 mM EDTA and detergent at concentrations over the CMC was

supplemented with 0.14 % (w/v) Triton X and added to yield a molar protein to lipid

ratio of about 1:50. The mixture was equilibrated at room temperature for 1 hour with

gentle agitation. The final lipid concentration of the mixture was 4 mg/ml. The

detergent was removed by adding BioBeads SM2 (BioRad) at 40 mg/ml wet weight

and the solution was agitated at room temperature for 15 minutes. Similar aliquots of

Biobeads were added four times for the incubation periods of 15 min, 30 min,

overnight and 60 min at 4 °C according to11' 14. Biobeads were removed after each

step. The turbid suspension was ccntrifuged at 184000xg for 20 min in a TLA 120.1

rotor and the proteoliposome pellet washed by gentle resuspension and re-

centrifugation. The final buffer contained 25 mM Hepes (NaOH) pH 7.5 and 150 mM

NaCl. Proteoliposomes were flash-frozen at 20 mg/ml lipids in liquid nitrogen and

stored at -80 °C.

Results

Recombinant Expression of MDR-ABC Transporters

All bacterial multidrug ABC transporters were recombinantly expressed as N-terminal

Hisio-tag fusion proteins in E. coli BL21(DE3)-R1PL strain in small scale expression

tests using Terrific Broth media containing glucose, glycine or glycerol. The

induction was started at optical density values of-1 at 37 °C and 20 °C. Expression

23

levels were generally low and protein could only be visualized by the Western-blot

method. Maximum expression was reached after 45 minutes at 37 °C and after 2.5

hours at 20 °C and cells stopped growing as indicated by constant optical density

values. No significant differences in the expression levels were visible and 37 °C was

chosen as standard expression temperature.

83kDa ->. 0' 20' 60' 120' 0' 20' 40' 60' 0' 20' 40' 60'

62 kDa.,,*^,!iî&**W

- *%4%y«»«IW**"Wft> llrt^v ff*

"

47.5 kDa

32.5 kDa^

25 kDa —,

b c d

Figure 2.1: Time course of the ABC transporter expression in E. coli BL21(DE3)-

RILP upon induction with IPTG. Harvested cells were cracked by sonication in the

presence of SDS. After centrifugation, supernatants were used and proteins

separated on a 12.5 % (w/v) SDS-PAGE gel and transferred to a nitrocellulose

membrane for Western-Blot analysis (Anti His5 antibody), a, Molecular mass

standard P77702 (NEB) b, B. subtilis YvcC (BmrA) in TBG,y,m at 20 °C and c,

S. aureus Savl866 in TB0iycCT0| at 37 °C d, B. subtilis YwjA in TBciucose at 37 °C

The Western Blot analysis revealed controlled expression in glucose rich medium

only. The /ac/-regulated genes were efficiently repressed and the expression quickly

and efficiently induced upon addition of IPTG (Figure 2.1a-d). While cell culturing in

the fermentor in TBoiydne was limited to optical densities (A6oo) of 4-5, culturing in

TBoiycerot resulted in a 2- to 3-fold decrease in the protein yield because of a leaky

expression and negative influences of expressed protein on the cell viability. Cultures

in TBgiucosc were stable, expression yields reproducibly high and up-scaling to larger

media volumes and higher optical density values was successfully performed. Cells

were cultured at 37 °C to Aöoo values of 12-15 for an induction of 1.5 hours. Although

using the identical protocols, fermentor cultures in TBgiucosc systematically provided

~70 % of the protein yields of shaking flask cultures (data not shown). Another effect

dramatically limiting the yield was observed in the fermentor using cells from -80 °C

stocks that were frequently exposed to room temperature. Fermentor cultures

24

exceeding A60o values of ~4 showed decreased expression levels of more than 30 %.

Therefore, purchased cells were only stored for a maximum of three months and

handled with care.

Extraction and Purification of Active MDR-ABC Transporters

The extraction and purification of the MDR-ABC transporter homologs was based

standard molecular biological methods to screen for the optimal type of detergent and

buffer conditions to produce stable and active protein. The developed method

included four consecutive steps: the preparation of membranes, the extraction and

purification of protein with detergent and the ATPase activity analysis of the

detergent-purified protein.

ATPase activity in vitro was reported for ABC transporters to be optimal in a pH

range of 6-915"17. Therefore, the solubilization from crude membranes and purification

of the MDR-ABC transporter homologs with various detergents was performed at pH

7.5-8.0. Other parameters like salts were only slightly varied to minimize the amount

of screening parameters (Table 2.1).

Soluble protein impurities from the crude membrane extract were removed during the

one-step NiNTA affinity chromatography purification with a washing buffer

containing 50 mM imidazole (Figure 2.2a, 2.3a). The extraction efficiency of the

selected detergents was crucial for the initial detergent screening. The solubilization

efficiency of the detergents at 1 % (w/v) concentration was judged after the one-step

purification by SDS-PAGE analysis. Preliminary extraction tests showed that

detergents with short hydrophobic carbon chain tails (n<10) only minimally extracted

the transport proteins (Table 2.1). Maximum solubilization of >0.6 mg protein/g wet

cell pellet was observed at 4 °C and 1.5 hours of stirring without sonication with

zwitterionic detergents, like FOS-choline 12 (FOS12), FOS-choline 14 (FOS14;

Figure 2.2a), n-dodecyl-N,N-dimethylarnine~N-oxide (LDAO) and Anzergent® 3-12.

25

LrnrA YwjA YvcC Savl866

L, lactis B. subtilis B, subtilis S. aureus

detergent DM, UDM, DDM, CHAPS, CHAPSO CHAPS, CHAPSO DM, UDM, DDM,

with TDM CxKj, CrE5, C«E(, QE4, CbE<;, QF.&, TDM

low CHAPS, CHAPSO Cm-sucrose CioE-i, C12EK CHAPS, CHAPSO

solubilization CgH4, CijBï, CxEfi, C-I1RSO Cm-sucrose CsE4, CgEs

efficiency C10E5, C^Es OCi, NO C8-HESO Cm-sucrose

(no further Cymal4-7 FOSiO OG, NG CrHESO

investigation) Cio-sucrosc TritonX OG, NG

CrHESO C,4-DAO FOSIO

Co-DAO, LDAO, Cymal4Cu-DAO

OÜ, NG

TritonX

detergents C,o-,C,2-DMG DM, UDM, DDM, DM, UDM, DDM, CkF-6, C^fjEf, C12EK

with Anzergent®-3-12, -3- TDM TDM Cymal4-7

high 14 C10E5, C12ES Cih» C10-DAO, LDAO,

solubilization FOSIO,-12,-14 Cymal4-7 Cymal5-7 Cu-DAO

efficiency Ci0-DAO, LDAO, Co-DAO, LDAO TritonX

C14-DAO Cm-, C|?-DMCi C,0-, C.2-DMG

TritonX Anzergent®-3-12, -3- AnzergenW<>-3-12,-3-C10-, C,2-DMO 14 14

Anzcrgcnt®-3-12,-3-14

FOSI2, -14

FOSIO,-12,-14 FOS12,-14

& DDM F0S12, FOS14 DDM DDM

biochemical FOS12,FOS14 LDAO CI2E„

stability FOS12.FOS14 FOS12,FOS14

& _ - DDM DDM

positive C,2ES

ATPase

activity

assay

& . _ _ C12E8

stability>15 mg/ml

Table 2.1: Detergents used in the homology screening approach arc listed according

to their chemical properties for the bacterial multidrug ABC transporter homologs.

Tests on YgaD from B. subtilis and Sav0643 from S, aureus using Ci2Es for the

solubilization, purification and stability tests are not listed.

Nonionic detergents like Triton® X-100, Thesit®, n-dodecyl-ß-D-maltopyranoside

(DDM), pentaethylene glycol monodecyl ether (C10E5) and octaethylene glycol

monododecyl ether (C^Eg) extracted less efficient than zwitterionic detergents (0.1-

0.4 mg protein/g wet cell pellet (Table 2.1). All detergent-solubilized and purified

proteins were analyzed by size exclusion chromatography probing the structural

integrity of the homodimeric protein. Stable and non-aggregated protein was observed

for all transporters using FOS12 and FOS14 (Figure 2.2b), for B. subtilis YvcC using

LDAO and DDM and for S. aureus Savl866 using DDM and Ci2E8 (Table 2.1).

26

175 kDa

X3kDa

62 kDa

47.5 kDa

32.5 kDa

25 kDa

16.5 kDa

a

Figure 2.2: Analysis of FOS-choline 14-solubilizcd, NiNTA purified and desalted

B subtilis YvcC a, by a 12.5% (w/v) SDS-PAGH: left lane' molecular mass

standard, right lane purified YvcC b, by analytical gelfiltration of the desalted

protein using a Superdex 200 10/300 GL and a buffer containing 50 mM Hepes

(NaOH) pH 8.0, 150 mM NaCl and 0.01 % (w/v) FOS14. The maximum of the

elution signal corresponds to an elution volume of-10 75 ml.

Sodium chloride at concentrations above 500 mM proved to stabilize the detergent-

solubilized proteins. Subsequent ATPase activity assays with detergent-purified and

reconstituted MDR-ABC transporter protein demonstrated V04-inhibatable, basal

ATPase activity for B. subtilis YvcC in DDM and S. aureus Savl866 in DDM and

CnEs. The use of zwitterionic detergents for the solubilization invariably yielded

inactive protein. Neither the detergent exchange nor the reconstitution into

proteoliposomes of FOS-choline purified YvcC resulted in ATPase activity (not

shown) suggesting an irreversible inactivation of the transporter. Further evidence

was provided by ATPase activity assays of C^Es-purified Savl866 and DDM-

purified YvcC showing a clear inactivation in the presence of sub-CMC and CMC

concentrations of FOS12, FOS14 and LDAO (not shown). Among the three positively

27

tested MDR-ABC transporters, only Savl866 in C]2E8 remained stable during the

concentration process up to concentrations exceeding 15 mg/ml proving suitable for

macromolecular crystallography with yields of 0.4 mg protein/g wet cell pellet

(Figure 2.4a,b).

175kDafif IT

83 kDa _~

62 kDa

47.5 kDa . „

32.5 kDa

25 kDa

16.5 kDa»

Figure 2.3: Analysis of C|2E8-solubilizcd, purified and concentrated S. aureus

Savl866 at 21.5 mg/ml concentration by a, 12.5 % (w/v) SDS-PAGE: molecular

mass standard and concentrated protein b, analytical size exclusion

chromatography of desalted, concentrated protein using a Superdex 200 10/300 GL

and a buffer containing 2 mM Tris (HCl) pH 8.0, 600 mM NaCl, 0.5 niM EDTA

(NaOH) pH 8.0 and 0.01 % (w/v) C,2E8. The maximum of the Savl866 dimer

clution signal corresponds to an clution volume of~11.75 ml.

Discussion

The homolog screening approach was performed to screen for a MDR-ABC

transporter proteins suitable to serve as a model system for both functional

investigation and macromolecular crystallography. With the aim to first investigate

the structure and later functionally characterize the protein, selected multidrug ABC

transporter homologs were expressed, solubilized and purified in several detergents.

i

F

L Sav1866 dimer -^î

bc

oCOCM

*j

tu

c

o

j

o I n-,^ *^

«> i -i

"

£>n » A P

<

i

; h

*V

CO

?

r j

. - ,

?Vi \injection / '

1 void /

J1—

Elution volume'

28

Protocols for the individual steps were successfully established and optimized to

produce sufficient amounts of stable and active protein. All MDR-ABC transporter

homologs could be expressed in E. coli in TBoiuco.se medium but only few detergent-

transporter combinations demonstrated to be stable in aqueous buffer and exhibited

V04-inhibitable ATPase activity. In the majority of the cases, mostly with

zwitterionic detergents, stable detergent-solubilized protein and reconstituted fractions

were not active suggesting an irreversible inactivation of the molecules. From a pool

of identically prepared and tested ABC transporter homologs, the Staphylococcus

aureus multidrug ABC transporter homolog Savl866 remained stable and active at

15 mg/ml concentration in CnE% detergent. Furthermore, the protein could be

produced in amounts sufficient for crystallographic analysis with yields exceeding

0.1 mg protein/g wet cell pellet.

References

1. Helenius, A. & Simons, K.. Solubilization of membranes by detergents. Biochim

Biophys Acta 415, 29-79 (1975).

2. Tanford, C. & Reynolds, J. A. Characterization of membrane proteins in detergent

solutions. Biochim Biophys Acta 457, 133-70 (1976).

3. Locher, K. P., Lee, A. T. & Rees, D. C. The E. coli BtuCD structure: a framework

for ABC transporter architecture and mechanism. Science 296, 1091-8 (2002).

4. Kendrew, J. C, Parrish, R. G., Marrack, J. R. & Orlans, E. S. The species

specificity of myoglobin. Nature IIA, 946-9 (1954).

5. Campbell, J. W., Duee, E., Hodgson, G., Mercer, W. D., Stammers, D. K.,

Wendell, P. L., Muirhead, H. & Watson, H. C. X-ray diffraction studies on

enzymes in the glycolytic pathway. Cold Spring Harb Symp Quant Biol 36, 165-

70(1972).

6. Inoue, H., Nojima, H. & Okayama, H. High efficiency transformation of

Escherichia coli with plasmids. Gene 96, 23-8 (1990).

7. Studier, F. W, & Moffatt, B. A. Use of bacteriophage T7 RNA polymerase to

direct selective high-level expression of cloned genes. J Mol Biol 189, 113-30

(1986).

8. Rosenberg, A. H„ Lade, B. N., Chui, D. S., Lin, S. W., Dunn, J. J. & Studier, F.

W. Vectors for selective expression of cloned DNAs by T7 RNA polymerase.

Gene 56, 125-35 (1987).

29

9. Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. W. Use of T7 RNA

polymerase to direct expression of cloned genes. Methods Enzymol 185, 60-89

(1990).

10. Gasteiger, E., Gattiker, A., Hoogland, C, Ivanyi, I., Appel, R. D. & Bairoch, A.

ExPASy: The proteomics server for in-depth protein knowledge and analysis.

Nucleic Acids Res 31, 3784-8 (2003).

11. Borths, E. L., Poolman, B., Hvorup, R. N., Locher, K. P. & Rees, D. C. In vitro

functional characterization of BtuCD-F, the Escherichia coli ABC transporter for

vitamin B12 uptake. Biochemistry 44, 16301-9 (2005).

12. Chifflet, S., Torriglia, A., Chiesa, R. & Tolosa, S. A method for the determination

of inorganic phosphate in the presence of labile organic phosphate and high

concentrations of protein: application to lens ATPases. Anal Biochem 168, 1-4

(1988).

13. Poolman, B., Doeven, M. K., Gecrtsma, E. R., Biemans-Oldehinkel, E., Konings,

W. N. & Rees, D. C. Functional analysis of detergent-solubilized and membrane-

reconstituted ATP-binding cassette transporters. Methods Enzymol 400, 429-59

(2005).

14. Patzlaff, J. S., van der Heide, T. & Poolman, B. The ATP/substrate stoichiometry

of the ATP-binding cassette (ABC) transporter OpuA. J Biol Chem 278, 29546-51

(2003).

15. Davidson, A. L., Laghaeian, S. S. & Mannering, D. E. The maltose transport

system of Escherichia coli displays positive cooperativity in ATP hydrolysis. J

Biol Chem 271,4858-63 (1996).

16. Zaitseva, J., Jenewein, S., Jumpertz, T., Holland, I. B. & Schmitt, L. H662 is the

linchpin of ATP hydrolysis in the nucleotide-binding domain of the ABC

transporter HlyB. Embo J 24, 1901-10 (2005).

17. Wang, Z., Stalcup, L. D., Harvey, B. J., Weber, J., Chloupkova, M., Dumont, M.

E., Dean, M. & Urbatsch, I. L. Purification and ATP hydrolysis of the putative

cholesterol transporters ABCG5 and ABCG8. Biochemistry 45, 9929-39 (2006).

30

Chapter 3: Expression, Purification, Crystallization and Crystal-

lographic Analysis of S. aureus Savl866

Abstract

The multidrug ABC transporter Savl866 from S. aureus was expressed as an

N-terminal Hisio-fusion protein in E. coli, purified in C^Es and crystallized in the

presence of adenosine diphosphate (ADP). Crystals were of sufficient size and quality

and diffracted synchrotron radiation to 3.0 Â resolution. Experimental phase

information was obtained to 3.3 Â resolution using five different derivative crystals

and the structure was solved by multiple isomorphous replacement with anomalous

scattering (MIRAS). A structural model of Savl866 in complex with ADP was build

and refined to 3.0 Â resolution with good refinement statistics. Another structure of

Savl866 was obtained by soaking native crystals with AMP-PNP. ADP was partially

replaced by the AMP-PNP as evidenced by clear electron density for the y-phosphate

moiety.

Introduction

High resolution crystal structures depict -- with exceptions - 'snap shot views' of

physiologically relevant conformations of soluble and membrane proteins. In the past,

structural models derived from crystals structures provided crucial information about

mechanistic aspects of the target proteins and largely contributed to the understanding

of their mode of action. The mechanism ofABC transporter proteins is not understood

due in part to the difficulties in obtaining well-diffracting three-dimensional crystals

of full ABC transporters. Factors limiting the production of such crystals are the low

expression yields and a successful extraction and purification of active protein in

detergent solution. Furthermore, the presence of the detergent complicates the

crystallization due to its unpredictable properties in aqueous solution and in complex

with the transport protein. In many cases, crystal contacts are restricted to solvent-

exposed regions and crystallization only occurs under conditions close to the cloud

point of the detergents.

Active Savl866 protein in detergent solution was subjected to crystal screening with

the aim to solve the structure of a bacterial multidrug ABC transporter at high

resolution and to find a model system, powerful enough to serve for structural and

31

functional investigations. Initial crystals obtained during the crystal screening phase

were analyzed with standard crystallographic methods and synchrotron radiation.

Based on the preliminary crystallographic data, crystal size and quality was optimized

and protocols for the treatment of the crystals developed (post-growth-treatment). The

protocols included the optimization of the crystal harvesting, stabilization and cryo-

protection. In addition, protocols for the derivatisation of native crystals with heavy

atoms had to be established. In order to solve and refine the structure of Sav 1866, a

strategy for the structure determination had to be developed using standard

crystallography programs.

Materials and Methods

Recombinant Expression of the MDR-ABC Transporter Savl866 S. aureus

Staphylococcus aureus Sav 1866 was overexpressed in E. coli BL21-codon plus

(DE3)-RJPL expression strain (Stratagene) by fermentation at 37 °C in terrific broth

medium supplemented with 1 % (w/v) glucose. Cells, stored for a maximum of 3

months at -80 °C were freshly transformed, cultured in 500 ml of medium to an

optical density (A60o) of ~1. Subsequently, cells were transferred to a 10 liter Techfors

S fermenter (Infors) and grown to optical cell densities of 12-15. The expression was

induced with 0.4 mM IPTG and cells were harvested after 1.5 hours by centrifugation

and stored at -80 °C.

Recombinant Expression of Selenomethionine-labeled S. aureus Savl866

Selenomethionine-labeled Staphylococcus aureus Savl866 was produced in shaking

flasks using E. coli BL21-codon plus (DE3)-RIPL expression strain (Stratagene).

Cells from fresh transformations were cultured at 37 °C in 50 ml of terrific broth rich

medium supplemented with 1 % (w/v) glucose to Amo of 1. 100 ml ofpre-warmed M9

medium supplemented with 0.5 % glucose and 0.5 mg/1 vitamin Bi hydrochloride was

inoculated with 1 ml of the TBAmpGiucose suspension and cultured at 37 °C to A6oo of

0.5. Subsequently, 12 times 2 1 of pre-warmed M9AmpGiucose medium in 5 1 shaking

flasks was inoculated with 12x4 ml of M9AmpGiucosc and cells were grown at 37 °C and

95 rpm. At Aeoo of 0.6, 12x40 ml of an aqueous solution of L-lysine, L-threonine, L-

leucine, L-isoleucine, L-valine, L-phenylalanine at final concentrations of 200 mg/1

and selenomethionine at a final concentration of 100 mg/1, was added and incubated

32

for 30 minutes. The expression was induced with 0.2 mM of IPTG and cells were

harvested after 1.5 hours by centrifugation and stored at -80 °C.

Extraction and Purification of Recombinants Expressed S. aureus Savl866

All subsequent procedures were performed at 4 °C unless specified differently. Frozen

cell pellets were resuspended by stirring in a buffer containing 50 mM Tris (HCl)

pH 8.2 and 500 mM sodium chloride and disrupted using a M-l 1 OL microfluidizer

(Microfluidics) at 15000 psi external pressure. Pure membranes were produced by

centrifugation at 40,000xg for 30 minutes to separate from cell debris and subsequent

ultracentrifugation at 100,000xg. The resulting membrane pellet was resuspended in a

buffer containing 100 mM sodium phosphate (NaH2P04) pH 8.0, 200 mM sodium

chloride, 15 % (w/v) glycerol, 20 mM imidazole (HCl) pH 8.0 and stored at -80 °C.

Batches of membrane suspension were thawed in a water bath at room temperature

for about 10 minutes and supplemented with 0.1% (w/v) n-dodecyl-ß-D-

maltopyranoside (DDM, Anatrace) and 1 % (w/v) polyoxyethylene-8-dodecylether

(anapoe-CnEg, Anatrace) for solubilization at 4 °C for 1.5 hours. All subsequent

buffers contained 0.01 % (w/v) C12EX anagrade as detergent. The solubilized

membrane proteins were purified in a one-step procedure, loaded onto a NiNTA

superflow affinity column (Qiagen) with a XK26 setup (GE Healthcare), washed with

50 mM imidazole, and Savl866 was eluted with 200 mM imidazole at low flow rates

of 3 ml/min. The buffer was exchanged to 10 mM Tris (HCl) pH 8.2, 100 mM NaCl

by desalting using a HiPrep Desalting 26/10 column (GE Healthcare), 1 mM of

buffered adenosine diphosphate (Tris (HCl) pH 7.5) was added and the protein

concentrated to 15mg/ml using an Amicon Ultra-15 concentrator unit (Millipore)

with a molecular cutoff of 100 kDa.

Crystallization of Savl866 in Complex with ADP and AMP-PNP

Staphylococcus aureus Savl866 was crystallized at 15 mg/ml concentration in the

presence of 1 mM ADP by vapor diffusion in sitting drops using 24-well Cryschem

plates (Hampton Research) and Crystal Clear sealing tape (Hampton Research). Initial

screening was performed at 4 °C, 10 °C and 20 °C against reservoirs essentially

combining the influences of polyethylene glycols, ion strength, pH, anions and cations

on the protein. The only commercial screen used was the PEG/Ion screen (Hampton

Research). In the final crystallization phase, Savl866 was pipetted at 4 °C and

33

crystallized at 10 °C against a reservoir containing 20 % (w/v) polyethylene glycol

6000, 50 mM Li3-citrate, 150 mM K3-citrate, 100 mM Na2HP04 and 3 raM MgCl2.

Reservoir solutions were pipetted at room temperature, cooled and stored at 4 °C

overnight.

The protein to reservoir ratio in the sitting drop was 2:1. Crystals of good quality were

harvested at 4 °C and moved with a capillary to a solution consisting of a 2.3 fold

concentrated mix of reservoir and desalting buffer with fixed values of 125 mM

sodium chloride and 17.5 % (w/v) PEG 6000. The crystals were incubated overnight

at 4 °C and subsequently cryo-protected in 15% (w/v) glycerol by a step-wise

concentration increase before flash-freezing in liquid nitrogen. Selenomethionine and

xenon derivative crystals were analogously produced either using selenomethionine-

labeled protein or incubating cryo-protected native crystals for 5 minutes at 10 bar

xenon pressure before flash-freezing in liquid nitrogen. 2'-iodo-adenosine

diphosphate^'-I-ADP), ethyl mercury phosphate (EMP) and TaeBriV derivative

crystals were produced by soaking of native crystals before cryo-protection with the

individual compounds dissolved in harvest solution. 2'-I-ADP was exchanged over

the course of 2 days at a concentration of 0.5 mM, Ta6Bri4 was soaked for 2 days in a

saturated solution and EMP was soaked at a concentration of 10 mM for 3 days.

AMP-PNP-bound Savl866 crystals were produced using ADP crystals of good

quality and size. These were incubated for 10 days in a harvesting solution containing

1.3 mM adenosine 5'-ß,y-imido triphosphate (AMP-PNP) to replace ADP. The

solution was exchanged after 5 days. AMP-PNP-bound Savl866 crystals were

prepared and cryo-protected in 15 % (w/v) glycerol before flash-freezing in liquid

nitrogen analogous to the ADP crystals.

Structure Determination of Savl866 in Complex with ÂDP

Diffraction data were exclusively collected at 100 K using synchrotron radiation at the

protein crystallography beamline X06SA PX at the Swiss Light Source (SLS) and

processed with Denzo and Scalepack .The structure was solved by multiple

isomorphous replacement with anomalous scattering using data from xenon (collected

at a wavelength of 1.54004 Â), Ta6BrI4 (collected at X= 1.25447 Â),

selenomethionine (collected at X = 0.97894 Â), 2 ~iodo-ADP (collected at X - 1.06994

Â), and ethyl mercury phosphate (EMP, collected != 1.00799 Â) derivative crystals

(Table 3.1).

34

Native Xenon Ta6BrM SelenoMet 2-I-ADP EMP

Data Collection

Space group C2 C2 C2 C2 C2 C2

Cell dimensions

a, b, c (Â) 161.279 161.391 160.895 161.399 160.621 161.510

103.955 103.839 104.163 103.886 105.045 104.074

181.013 181.948 181.937 181.644 181.245 181.546

a,ß,y O 90.000 90.000 90.000 90.000 90.000 90.000

97.987 97.471 97.849 97.501 98.021 97.672

90.000 90.000 90.000 90.000 90.000 90.000

Resolution (Â) 30-3.0 30-3.3 30-3.8 30-3.3 30-3.1 30-3.3

'HYm or AmcrKC 8.1 (56.9) 10.9 (45.7) 10.8 (27.8) 8.1 (35.1) 7.9 (45.9) 9.9(53.3)liai 20.9(2.18) 14.5(2.00) 13.3(3.96) 16.7(2.83) 19.5(1.99) 15.0(2.40)

Completeness (%) 99.1 (87.9) 98.0 (85.9) 95.2 (73.9) 97.6 (79.2) 98.0 (79.9) 100(100)

Redundancy 7.4 (5.3) 7.1(5.1) 6.6 (5.5) 6.0 (4.4) 6.2(5.1) 6.4 (5.9)

Refinement

Resolution (Ä) 20-3.0

No. reflections 54627/4176

"work/ "free 0.254/0.272

No. atoms

Protein 9170

ADP 54

Na 2

Water 16

B-factors

Protein 102.4

Ligand 82.7

Ion 110.5

Water 56.8

R.m.s deviations

Bond lengths (Â) 0.0097

Bond angles Q 1.4

"Highest resolution shell is shown in parenthesis.

Table 3.1: Data collection and refinement statistics for Sav1866 in complex with

ADP (PDB ID: 2HYD).

The individual wavelength values were determined in individual x-ray absorption

scans. Native data was collected at a wavelength of 1.07252 Â. Initial phases were

obtained using SOLVE and were used to calculate anomalous cross-Fourier maps

using programs from the CCP4 suite4. Additional heavy atom positions were refined

using SHARP5 and the FOM of centric/acentric reflections was 0.32/0.26 overall for

phasing from 30-3.0 Â resolution. Solvent flattening and non-crystallographic

averaging was performed using Solomon' and DM7. Model building was performed

using O8. Chain tracing was aided by the known positions of methionines from the

selenomethionine data, and the location of ADP was indicated by the presence of the

35

iodine signal from the 2'-iodo-ADP data (Figure 3.4b). The refinement was carried out

using CNS9 to 3.0 Â resolution and except for crystal contact regions in the

extracellular loops and a few residues with evidently different density in the two

subunits, strict two-fold non-crystallographic symmetry was imposed. Ramachandran

analysis was performed using Procheck10.

Structure Determination of Savl866 in complex with AMP-PNP

Diffraction data were collected using synchrotron radiation at beamline X06SA of the

Swiss Light Source (SLS) and processed with Denzo and Scalepack2. The structure

was solved by molecular replacement using the structural model of Savl866 [PDB

code: 2HYD] and CNS 9. Model modifications in O8 using CNS-generated omit maps

and further refinement steps in CNS resulted in the structural model of Savl866 in

complex with AMP-PNP. Strict two-fold non-crystallographic symmetry was

imposed during the refinement to 3.4 Â resolution except for crystal contact regions in

the extracellular loops and a few residues with evidently different electron density.

Ramachandran analysis was performed using Procheck10.

36

Native

Data Collection

Space group C2

Cell dimensions

a,b,c(Â) 160.961

104.447

181.391

a, p, y (°) 90.000

98.229

90.000

Resolution (A) 30-3.4

Ä„oryU* 9.5(52.9)

liai 13.7(2.77)

Completeness (%) 100 ( 100)

Redundancy 5.2 (5.3)

Refinement

Resolution (A) 20-3.4

No. reflections 54627/4176

Äwork/JRftee 0.254/0.272

No. atoms

Protein 9168

ADP 62

Na 4

Water 20

B-factors

Protein 124.6

Ligand 77.1

Ion 94.9

Water 73.2

R.m.s deviations

Bond lengths (A) 0.0106

Bond angles Ç) L5

"Highest resolution shell is shown in parenthesis.

Table 3.2: Data collection and refinement statistics for Savl866 in complex with

AMP-PNP (PDB ID: 20NJ).

Results and Discussion

Extraction and Purification of S. aureus Savl866

Full-length Staphylococcus aureus Savl866 (residues 1-578) was expressed as an

N-terminal Hisio-fusion protein and extracted with 0.1 % (w/v) DDM and 1 % (w/v)

Anapoe-CnEs from purified membranes. The detergent was added at 4 °C to a turbid

membrane suspension to prevent from aggregation observed for starting temperatures

below 0 °C in previous solubilization experiments. Optimal extraction of

homodimeric, non-aggregated protein was observed at pH 7.5-8.0 in the presence of

37

glycerol and sodium phosphate at 4 °C for 1.5 hours. Maximal sodium phosphate

concentrations were applied close to the point of precipitation and an additional

requirement for ion strength in the mix compensated with sodium chloride. The

separation of the protein solution from debris by centrifugation at 40,000xg was

incomplete but did not interfere with the purification by NiNTA affinity

chromatography. The purification was performed using Ci2Eg (analytical grade)

buffers at low flow rates to maximize the loose binding of homodimeric Savl866 and

to minimize the binding of aggregated Savl866. Low concentrations of 50 mM

imidazole competed with Savl866 on the NiNTA superflow resin but proved to be

sufficient to remove mostly soluble protein impurities. Imidazole at elution

concentrations of 200 mM harmed Savl866 with the time but could efficiently be

removed by immediate buffer exchange (desalting). Savl866 protein was

concentrated to 15 mg/ml in a desalting buffer of low ion strength. A minimum of

100 mM of sodium chloride was needed to prevent from aggregation and to guarantee

the reproducibility of the preparation. However, Savl866 was extremely sensitive to

overconcentration in C^Es and during the concentration process the protein had to be

frequently mixed. Addition of nucleotide, ADP, during the concentration stabilized

the protein. Size exclusion chromatography confirmed the homodimeric nature of the

concentrated protein and the absence of aggregates.

Crystallization of S. aureus Savl866

Crystallization screening experiments were exclusively performed by vapor diffusion

using sitting drop plates and scaling tape. The screening was performed at 4 °C, 10 °C

and 20 °C with concentrated Savl866 in C12E8 at ~15 mg/ml concentration against

reservoirs containing various salts and salt mixes at different pH values and

polyethylene glycols as exclusive précipitants". Interestingly, polyethylene glycols up

to molecular weights of 2000 did not induce crystal or precipitation formation but

tended to conserve the protein. For example, no precipitation was observed in 40 %

(w/v) PEG 400 after one month. Further crystallization experiments using PEG 4000

yielded needle-like crystals after 3 days at 20 °C. The protein to reservoir drop ratio

was 1:1 and the reservoirs contained 300 mM lithium and potassium citrate at ~pH 7

(Figure 3.1a). Larger crystals were observed by screening in a pH range of 6-9 at

10 °C using additives - salts and organic compounds, respectively, that influence

crystal growth12 - and two new crystals forms were observed. While hexagonal

38

crystals occasionally occurred at phase separation, boundaries (not shown), box-

shaped crystals were reproducibly crystallized at a protein to reservoir drop ratio of

2:1. The box-shaped crystals appeared after 5 days and matured to full size (150-

200 u.m) within 2-3 weeks in a reservoir containing 20 % (w/v) polyethylene glycol.

6000, 50 mM Li;5-eitratc, 1.50 mM K3-citratc, 100 mM Na.2HP04 and 3mM MgCI2

(Figure 3.1b). The crystals were formed in wells of 4.5 pi drop volume that showed

granulate but not heavy precipitation at day one. The crystallization of Savl866 has

most likely occurred in a metastable phase close to the point of liquid-.li.quid pliasc

separation. Reservoirs approximately resembling the final concentrations in the

crystal well showed a clear température- and polyethylene giycol-dependenee on the

citrate contend at. the concentrations used (Figure 3.2). Liquid-liquid phase separation

occurred at lower concentrations of PEG 6000 at 10 °C than at 4 °C, likely facilitating

crystallization of Sa.vl866 at 10 °C.

Figure 3.1: Staphylococcus aureus Savl866 with bound ADP a, initial crystals at

23 % PEG 6000, 300 mM Li3/K3-citrate, Ï .67 % citric acid and 3 miV! JVSgCh b,

we!!-diffracting three-dimensional crystals of Sav!866 in complex, with ADP at

20 % (w/v) poiyethylene glycol 6000, 50 mM Li3-citratc, Î 50 mM Kreitrate,

100 mM Na2HPO,i and 3 mM IVlgCb,

Positive or no effects on the crystal growth were only observed for very few additives

like lithiuin chloride, sodium chlonde and potassium, phosphate but they did not prove

to be useful in the optimization crystallization phase. Non-ionic detergents like DDM

did not influence crystal formation significantly but lowered, the overall diffraction

quality of the crystals. It has to be pointed out that crystal quality and size could be

improved during the optimization of the crystal screening in combination with an

39

optimization of the extraction, purification and concentration protocols for Sav 1866

protein.

16 17 18 19 20 21 22 2.1 24

ruiicciiti'illiiin of PF(i 6000 |"/„ (\v/v)|

Figure 3.2: PEG 6000 - citrate - phase diagram at 4 °C, 10 °C and 20 °C at 0.01 %

(w/v) Ci2E8. The individual points represent the points of liquid-liquid phase

separation of incubated reservoir. The circle indicates the reservoir concentration

used to derive box-shaped Savl866 crystals.

Crystals grown at 10 °C were incubated for about 10 minutes at 4 °C and 5-10 ul of

harvest solution were gently added to the crystal well for an overnight incubation.

While flawed crystals suffered from razor-thin cracks after the treatment, flawless

crystals were selected and transferred with a capillary to 15 ul of fresh harvest

solution. Precipitate was gently removed and crystals were quickly and reproducibly

cryo-protected. Therefore, glycerol was added by pipetting with increasing amounts

up to 15 % (v/v)13 and snap-frozen in liquid nitrogen. Different cryo-protectants were

tested in order to dehydrate the crystal and stabilize the lattice for the flash-freezing in

liquid nitrogen14' 15. Alcohols serving as cryo-protectants proved incompatible

because of the immediate precipitation of the detergent-solubilized protein. Because

of precipitation of the cryo-protectants themselves, cryo-salts were not tested. The use

of sugars like maltose, sucrose, 2-methyl-2,4-pentanediol (MPD) or triethylene glycol

to polyethylene glycol 2000 resulted in an immediate disintegration of the crystals.

Furthermore, the crystal growth was completely inhibited in the presence of any cryo-

protectant. Therefore, glycerol remained the only cryo-protectant applicable. A less

pronounced but crucial decrease in the crystal quality was observed for the treatment

40

with glycerol for more than 30 minutes and for a non-rushing crystal transfer from the

final cryo-solution to the liquid nitrogen filled Dewar vessel.

Derivatisation of native Savl866 crystals

Derivatisation experiments were performed to produce crystals containing heavy

atoms suitable for solving the phase problem in protein X-ray crystallography.

Experimentally, Savl866 native crystals containing ADP were produced with

selenomethionine-labeled protein, soaked with heavy atom compounds or subjected to

noble gas diffusion experiments. Several derivative crystals of high quality were

observed, diffraction data collected and the structure of Savl866 in complex with

ADP solved by multiple isomorphous replacement with anomalous scattering

(MIRAS).

Selenomethionine derivative crystals were produced analogous to native crystals.

Labeled Savl866 crystallized under the same reservoir conditions but proved less

soluble and crystallized at much lower protein concentrations of about 7-10 mg/ml.

The crystals proved to be more fragile than native crystals but provided a comparable

quality after cryo-protection in 15 % glycerol.

Xenon derivatisation was performed using the noble gas diffusion technique16' 17.

Xenon noble gas is water soluble and able to form reversible but specific interactions

with proteins under moderate xenon pressure. Cryo-protected native crystals of

Savl866 could be successfully treated with xenon noble gas at 10 bars pressure for 5

minutes at 4 °C. The crystals were unaffected by the procedure but only rapidly

frozen crystals (>1 minute) proved to contain enough xenon to contribute to phasing.

Furthermore, 50 % of the crystals were lost during pressure release caused by fluid

cavitation of the xenon saturated, detergent containing harvesting solution. Crystals

were analyzed within 3 days because of the known lattice damage effects caused by

the formation ofxenon clathrates18.

Crystals were soaked with heavy atom compounds before cryo-protection. Several

heavy atom compounds frequently used for derivatisation in macromolecular

crystallography were tested at various concentrations and incubation times. From the

commercially available heavy atom screens (Hampton Research) only ethyl mercury

phosphate (EMP) bound within 3 days of soaking to the Savl866 ADP crystals

without damaging the crystal lattice. The non-commercial and water-soluble

compound Ta^Brn was soaked for only 1 day as a saturated solution because of

41

apparent crystal damages already after 2 days of exposure. At lower soaking

concentrations, the green staining of the crystals with TaoBr^ was not observed.

2'-iodo-adenosine diphosphate was soaked for 2 days at 0.5 mM concentration to

replace the ADP in the crystals and to verify the presence of ADP in the crystal.

Soaking of native crystals with ATP at a concentration of 1.3 mM preserved the

crystals but did not succeed due to the hydrolysis of ATP in the crystal. Crystals

treated with ATP were analyzed but were found to contain ADP. Therefore, binding

of adenosine 5'-ß,Y-imido triphosphate (AMP-PNP), with the ß,y-imido group

preventing hydrolysis, at 1.3 mM was investigated and crystals incubated at 4 °C for

10 days. The harvest solution containing AMP-PNP was exchanged after 5 days and

crystals were finally prepared analogous to native ADP crystals.

Structure Determination of Savl866 in complex with ADP and AMP-PNP

Crystallographic data was collected at the beamline X06SA PX at the Swiss Light

Source (SLS) and processed with Denzo and Scalepack2. The use of synchrotron

radiation was crucial for the work with the ABC transporter crystals because the

intensity and the brilliance of home source x-ray beams was not sufficient. Crystals

were analyzed at cryogenic temperatures (100 K) and native datasets of crystals

containing ADP and AMP-PNP were collected at the wavelengths of 1.07252 Â and

1.00004 À (Table 3.1, 3.2, Figure 3.3). In order to collect datasets for the derivative

crystals, x-ray absorption spectra in the range of the absorption edge of the

corresponding heavy atom were recorded to determine the optimal wavelength for the

data collection. The peak wavelengths for 2'-I-ADP, Ta<,Bri4 and selenium were

experimentally determined and datasets collected at 1.06994 Â, 1.25447 Â (Ta L-III

absorption edge) and 0.97894 Â (Se K.). The large anomalous scattering of mercury at

wavelengths shorter than the L-III edge was utilized and datasets were recorded at

1.00799 Â19' 20. With the experimental setup at X06SA PX wavelengths

corresponding to the absorption edges of Xe (L-I/I1/III) could not be reached and

datasets were recorded at a wavelength of 1.54002 Â. All datasets were recorded with

oscillation steps of 0.5 ° and oscillation ranges of about 180 ° for acceptable

redundancies in the range of 5-7. The exposure times varied along with the

attenuation with corresponding filters and the observed decay in scattering efficiency

of the crystals. A single multiple-wavelength anomalous diffraction (MAD)

experiment was performed using a selenomethionine-labeled crystal. The datasets

42

collected did not provide additional information and were not compatible with the

previous experiments and were omitted in the structure determination process.

Figure 3.3: High resolution native diffraction pattern of Savl866 crystals with in

complex with ADP at 3 0 A resolution The dataset was recorded at a wavelength

of 1 07252 Â at X06SA PX bcamhne at the SLS (Vilbgen) The Savl866 crystals

belonged to the space group C2 with the unit cell parameters of a = 161 Â,

6=105Â,< = 181 Â and/?-98°

The structure of Sav1866 in complex with ADP was phased by multiple isomorphous

replacement with anomalous scattering (MIRAS) using data from xenon, Ta<,Bri4,

selenomethionine, 2'~iodo-ADP and ethyl mercury phosphate derivative crystals

(Table 3.1). Sav 1866 crystals belonged to the space group C2 with one molecule

(Sav1866)2 in the asymmetric unit. Excellent experimental density was visible for the

entire molecule (residues 1 to 578) including loops connecting transmembrane helices

(Figure 3.4a). Side chain position could be located for most of the important residues

and model building was further aided by the presence of selenomethionine sites

serving as marker positions (Figure 3.4b). Except for extracellular loops engaged in

lattice contacts and a few residues with distinct side chain conformations at the

subumt interface, no differences are visible between the two non-crystallographically

related subunits of (Sav8166)2.

43

Figure 3.4: a, Experimental electron density map ( 1 a) of the transmembrane region

of S. aureus Savl866 at 3.1 Â resolution with the Ca trace in red and b,

experimental anomalous cross-Fourier maps of Sc (yellow, 7 o), Ta6Br,4 (green,

5 er), I (purple, 8 a) and xenon (red, 15 a). Savl866 C„ trace in white. The figures

were prepared using Dino (www.dino3d.org)

Initial phases were obtained using SOLVE3 and were used to calculate anomalous

cross-Fourier maps using programs from the CCP4 suite4. Additional heavy atom

positions were refined using SHARP5 and the FOM of centric/acentric reflections was

0.32/0.26 overall for phasing from 30 to 3.0 Â resolution. Solvent flattening and non-

crystallographic averaging was performed using Solomon6 and DM7 with best results

for a crystal solvent content of 64 % as parameter for the calculation. Theoretical

calculations following the method of Matthews resulted in a value of 78.74 % solvent

content21. Model building was performed using Ox. Chain tracing was aided by the

known positions of methionines from the selenomethionine data, and the location of

ADP was indicated by the presence of the iodine signal from the 2'-iodo-ADP data

(Figure 3.4b). The refinement was carried out using CNS9 to 3.0 Â resolution with

good refinement statistics of RWOrk and Rfree values of 25.40 % and 27.23 % and an

expected average protein B-factor value of 102.4 (Table 3.1). Individual B-factors

(Debye-Waller factors), as a measure of the atomic thermal motion and disorder, were

above average in the model of Savl866 in the membrane spanning parts and below

average at mechanistically relevant sites. Locations of highest thermal motion were

44

exclusively found in the extracellular loop regions mostly at the ones not contributing

to crystal contacts. The strict two-fold non-crystallographic symmetry imposed on the

two subunits of Savl866 was released for crystal contact regions in these regions and

a few residues with evidently different density and beneficial effects for the

refinement were observed. A final Ramachandran analysis was performed using

Procheck10 and revealed 86.0% in the most favorable, 14.0% in the additional

allowed, and no residues in the generously allowed or disallowed regions after very

few manual adjustments using O (Figure 3.5). The atomic coordinates of Savl866 in

complex with ADP and the corresponding structure factors of the native high

resolution dataset were deposited at the protein databank (PDB) with the PDP

identifier 2HYD23

Plot statistics

Residues in most ftisnuri'd regions jA.lt.LjResidues in iidihlumjl jHouctl r^mms U& '-I1!KesiJtu's m ^ncuni^lj ,illirt\i,(t ifjjirtiis [ .1, 1"', ! p|Residues in disallowed regions

Miniher ol non-dyeme and non-proline residues

Number oi cnd-reuduosd'vil (jI\ it nit I'ri'l

Numhn ni jityiiiicrLsidu^CilKuvnastiMiijilLS)Numhei ot proline residues

louil number uf refill tit*.

So 0%

14 0",,

0 U<"n

Figure 3.5: Ramachandran Plot of Sav 1866 in complex with ADP (PDP ID: 2HYD).

Analysis of the crystal packing of Sav 1866 crystals showed that exclusively polar

surfaces of the MDR-ABC transporters contact one another. The crystals consisted of

two-dimensional planes of Sav1866 molecules that were interconnected by the

residues of a flexible linker region that fuses the transmembrane domain and the

nucleotide binding domain of one subunit of Sav 1866. While these contacts appeared

to be the weakest of the crystal contacts (Figure 3.6a), strong crystal contacts were

formed within the planes. The extracellular loops, ECL1 and ECL2, of one molecule

interacted with the analogous loops of another transporter and tight contacts were

observed at each of the nucleotide binding domains (Figure 3.6b).

45

Figure 3.6: Crystal packing in the box-shaped Savl 866 protein crystals (space group

C2). a, view along the planes demonstrating interplane crystal contacts of the

linker region b, crystal contacts of the transporter molecules within the planes.

The structure of Savl866 in complex with AMP-PNP was solved by molecular

replacement using the structural model of Sav 1866 in complex with ADP. Manual

placement of the structural model was the starting point and the model refined against

the native dataset of the AMP-PNP crystal. Refinement was performed using CNS,

model modifications in O using CNS-generated omit maps and further refinement

steps in CNS resulted in the structural model of Sav 1866 in complex with AMP-PNP.

Strict two-fold non-crystallographic symmetry was imposed analogously to Sav 1866

with bound ADP. The structure was refined to 3.4 Â resolution with good refinement

statistics of Rwork and Rfree values of 25.39 % and 27.76 %, respectively and a

Ramachandran analysis yielded 85.9 % in the most favorable, 15.1 % in the additional

allowed regions and no residues in the generously allowed or disallowed regions. A

B-factor analysis revealed a slightly higher average B-factor for the AMP-PNP

structure compared to the ADP structure with regions of high B-factors in the

extracellular loop regions. A structural comparison and superposition of AMP-PNP-

bound Savl 866 with the ADP-bound structure of Savl866 revealed a root mean

square deviation of 0.097 Â for 1116 residues (of 1156 visible residues), which

demonstrates the overall similarity of the two models. Minor conformational

46

differences existed at the nucleotide binding sites, most notably in the LSGGQ motifs

and the x-loops, and the Gln422 from the Q-loop. However, because of the limited

resolution, these changes may be insignificant. ADP was indeed partially replaced as

evidenced by clear electron density for the y-phosphate moiety (Figure 3.7a,b) and

extra electron density close to where magnesium or sodium ions are observed in high

resolution crystal structures of isolated nucleotide binding domains24' 2\ This suggests

that AMP-PNP had indeed replaced ADP, and that a cation, most probably sodium,

was also bound at the nucleotide binding site.

a

*>& "frr

Figure 3.7: a, Electron density contoured at 3 5 a (red) of y-phosphate of soaked

AMP-PNP replacing the ADP in Savl866 crystals a, side view and b, bottom

view of the protein. C0 traces of the two submits of Savl 866 are in yellow and

green. AMP-PNP is in ball-and-stick Pictures were produced using PyMol26

ADP stabilized Savl 866 during the concentration in detergent solution. This indicated

that the interaction of ADP with the sequence motifs at the shared interface of the

NBDs was sufficient to trap the transporter in the observed confonnation. The protein

conformation did not change upon replacement of ADP in native crystals with AMP-

PNP suggesting that the conformational equilibrium of Savl866 was shifted in

detergent solution. A comparison of the observed conformation with crystals

structures of isolated NBDs with bound ATP demonstrated that Savl 866 in complex

with AMP-PNP reflected a physiologically relevant conformation (Chapter IV).

47

ki.

à

/ s

References

1. Schneider, G. & Lindqvist, Y. Ta6Brl4 is a useful cluster compound for

isomorphous replacement in protein crystallography. Acta Crystallogr D Biol

Crystallogr 50, 186-91 (1994).

2. Otwinowski, Z. & Minor, W. Processing of X-ray diffraction data collected in

oscillation mode. (ed. Carter, C. W.) (Academic Press, New York, 1997).

3. Terwilliger, T. C. & Berendzen, J. Automated MAD and MIR structure solution.

Acta Crystallogr D Biol Crystallogr 55, 849-61 (1999).

4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol

Crystallogr 50, 760-3 (1994).

5. de la Fortelle, E, & Bricogne, G. Maximum-likelihood heavy-atom parameter

refinement for multiple isomorphous replacement and multiwavelength

anomalous diffraction methods. Methods Enzymol 276, 472-494 (1997),

6. Abrahams, J. P. & Leslie, A. G. Methods used in the structure determination of

bovine mitochondrial Fl ATPase. Acta Crystallogr D Biol Crystallogr 52, 30-42

(1996).

7. Cowtan, K. D. & Main, P. Phase combination and cross validation in iterated

density-modification calculations. Acta Crystallogr D Biol Crystallogr 52, 43-8

(1996).

8. Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M. Improved methods for

building protein models in electron density maps and the location of errors in

these models. Acta Crystallogr A 47 (Pt 2), 110-9 (1991).

9. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-

Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R.

J., Rice, L. M., Simonson, T. & Warren, G. L. Crystallography & NMR system: A

new software suite for macromolecular structure determination. Acta Crystallogr

D Biol Crystallogr 54, 905-21 (1998).

10. Laskowski, R. A., Moss, D. S. & Thornton, J. M. Main-chain bond lengths and

bond angles in protein structures. J Mol Biol 231, 1049-67 ( 1993).

11. McPherson, A., Jr. Crystallization of proteins from polyethylene glycol. J Biol

Chem 251, 6300-3 (1976).

12. Cudney, R., Patel, S., Weisgraber, K., Newhouse, Y, & McPherson, A. Screening

and optimization strategies for macromolecular crystal growth. Acta Crystallogr D

48

Biol Crystallogr 50,414-23 (1994).

13. Sousa, R. Use of glycerol, polyols and other protein structure stabilizing agents in

protein crystallization. Acta Crystallogr D Biol Crystallogr 51,271-7(1995).

14. Garman, E. F. & Schneider, T. R. Macromolecular crystallography. J Appl Cryst

30,211-237(1997).

15. Rodgers, D. W. in Methods in Enzymology 211-203 (Academic Press, London,

1997).

16. Schütz, M., Prange, T. & Fourme, R. On the preparation and x-ray data collection

of isomorphous xenon derivatives. J Appl Cryst 27, 950-960 (1994).

17. Schütz, M., Fourme, R. & Prange, T. Use of noble gases xenon and krypton as

heavy atoms in protein structure determination. Methods Enzymol 374, 83-119

(2003).

18. Sauer, O., Schmidt, A. & Kratky, C. Freeze-trapping isomorphous xenon

derivatives of protein crystals. JAppl Cryst 30, 476-486 (1997).

19. Hendrickson, W. A. Determination of macromolecular structures from anomalous

diffraction of synchrotron radiation. Science 254, 51-8 (1991).

20. Dumas, C, Duquerroy, S. & Janin, J. Phasing with mercury at 1 A wavelength.

Acta Crystallogr D Biol Crystallogr 51, 814-8 (1995).

21. Matthews, B. W. Solvent content of protein crystals. J Mol Biol 33, 491-7 (1968).

22. Kleywegt, G. J. & Jones, T. A. Homo crystallographicus—quo vadis? Structure 10,

465-72 (2002).

23. Dawson, R. J. & Locher, K. P. Structure of a bacterial multidrug ABC transporter.

Nature 443, 180-5(2006).

24. Hopfner, K. P., Karcher, A., Shin, D. S., Craig, L., Arthur, L. M., Carney, J. P. &

Tainer, J. A. Structural biology of Rad50 ATPase: ATP-driven conformational

control in DNA double-strand break repair and the ABC-ATPase superfamily.

Cell 101, 789-800 (2000).

25. Smith, P. C, Karpowich, N., Milien, L., Moody, J. E., Rosen, J., Thomas, P. J. &

Hunt, J. F. ATP binding to the motor domain from an ABC transporter drives

formation of a nucleotide sandwich dimer. Mol Cell 10, 139-49 (2002).

26. DeLano, W. L. (DeLano Scientific LLC, San Carlos, CA, USA, 2002).

49

Chapter 4: Structural Characterization of the Bacterial

Multidrug ABC Transporter Savl866

Abstract

This chapter describes the architecture and structure of Savl866 with bound

nucleotide in detail. The structure reveals twelve transmembrane helices in an

arrangement that is consistent with cross-linking studies and electron microscopic

imaging ofhuman multidrug resistance protein Mdrl, but critically different from that

reported for the bacterial lipid flippase MsbA. The observed, outward-facing

conformation reflects the ATP-bound state, with the two nucleotide-binding domains

in close contact. The two transmembrane domains form a central cavity, presumably

the drug translocation pathway that is shielded from the inner leaflet of the lipid

bilayer and from the cytoplasm, but exposed to the outer leaflet and the extracellular

space. In agreement with data from the literature, the conformation observed suggests

that ABC transporter proteins likely follow the concept of an 'alternate access and

release' mechanism under the control of ATP binding and hydrolysis.

Introduction

The crystallographic analysis of Savl866 revealed the structure of a bacterial

multidrug ABC transporter at high resolution. The observed conformation of Savl866

was analyzed in detail and compared to genetic, biochemical and structural data from

the literature. The analysis aimed at proving the physiological relevance of the

structural model, at defining the architecture of a multidrug ABC transporter and at

discovering the mechanism of ATP-driven substrate translocation by ABC

transporters.

The Architecture and Domain Structure of Savl866

The asymmetric unit of the crystal contained the functional unit of Savl 866 composed

of a dimer of two elongated subunits related by two-fold molecular and non-

crystallographic symmetry (Figure 4. la,b). Normal to the membrane, the transporter

appears elliptical with dimensions of 65 Â by 55 Â exhibiting a large cavity open to

the extracellular space. The overall shape of the molecule is consistent with that of

human Mdrl studied at low resolution by electron microscopy1' 2. In the third

50

dimension, viewed parallel to the membrane, the transporter is 120 Â long. Each

subunit consists of an N-terminal transmembrane domain (residues 1 to 320) and a C-

terminal nucleotide-binding domain (residues 337-578).

Figure 4.1: Ribbon representation of the Savl866 structure a, the subunits of the

homodimer are colored yellow and green with bound ADP m ball-and-stick The

grey box depicts the likely location of the lipid bilayer based on the

hydrophobicity of the protein surface b, 90° rotation of a around the vertical

2-fold molecular and non-crystallographic axis. The TM helices of one subunit

(green) are numbered. '1CL' denotes intracellular loops between TM helices,

'ECL' extracellular loops, 'N-ter' and 'C-ter' the N- and C-termini.

The two subunits reveal a considerable twist and embrace each other, with both the

transmembrane and the nucleotide-bmding domams tightly interacting. The

polypeptide that links the domains (residues 321 to 336) emanates from

51

transmembrane helix 6 (TM6) and wraps around the distal side of the NBDs

(Figure 4.lb), similar to the analogous linker observed in crystal structure of human

Tapl3.

The transmembrane domains span the membrane and are about 75 Â long. According

to the electrostatic potential surface, the native lipid bi layer appears to cover most of

the upper half of the TMDs. Each transmembrane domain traverses the lipid bilayer

six times to yield twelve transmembrane helices (TM) for the homodimeric

transporter, in agreement with the canonical ABC exporter topology4. The twelve TM

helices are clustered in the cytoplasmic part of the transmembrane domain and form a

tightly packed core (Figure 4.2b). Around the middle of the membrane, bundles of

TM helices diverge into two discrete 'wings' that point away from one another

towards the extracellular space (Figure 4.2a). The observed cavity likely provides an

outward-facing conformation. The individual wings' do not represent individual

transmembrane domains. Each "wing' consists of helices TM1-TM2 from one subunit

and TM3-TM6 from the other (Figure 4.2a).

Figure4.2: Cylinder representation of the TM helices of Savl866. a, Top view

reveals the 'wings' formed by the helices and b, bottom view highlight clustering

at the shared interface to the NBDs of Savl866. Yellow and green indicates the

two subunits of the homodimer and the helices arc numbered.

The transmembrane segments are connected by long intracellular and short

extracellular loops (ICLs and ECLs). The ICLs extend the helical secondary structure

beyond the lipid bilayer and protrude approximately 25 Â into the cytoplasm.

Remarkably, the helices TM1-TM3 are closely related to TM4-TM6 by approximate

52

two-fold rotation around a symmetry axis parallel to the membrane plane, with a root

mean square deviation of 5.7 Â for 146 Ca positions (residues 9-154 and 175-320)

(Figure 5.3b). Rotational symmetry of subsets of transmembrane helices may indicate

gene duplication and has been observed in the structures of aquaponns5 and major

facilitators1, but has not previously been reported for ABC exporters.

Figure 4.3: Savl866 TM helix topology a, Cylinder representation of the TMDs of

Savl866 in yellow and green, b, Schematic of the TM helix topology of one TMD

and demonstration of the approximate 2-fold rotation symmetry of the helices TM

1-3 (yellow) to TM 4-6 (orange) within one subunit of Savl866

The arrangement of the transmembrane helices observed for Savl866 is consistent

with recent cross-linking data that identified neighboring helices in human Mdrl.

With respect to the nucleotide-binding domains, the structure of Savl866 revealed

close structural agreement with that of MJ0796 and clear electron density was

evident at the P-loops and the ABC signature motifs of opposing subunits for ADP or

AMP-PNP, respectively. In the AMP-PNP structure extra electron density is visible

adjacent to the y-phosphate of the AMP-PNP close to where magnesium or sodium

ions were observed in the crystal structures of Mj'0796 or Rad50. This suggested that

AMP-PNP had indeed replaced ADP during the soaking of native crystals, and that a

cation like sodium was also bound at the nucleotide binding site. At the present

resolution, the conformations in the structural models of Savl866 are

indistinguishable from the conformations observed in the high resolution crystal

53

structures of MJ0796 or Rad509 with trapped ATP. The crystal structure demonstrates

that the NBDs of Savl 866 with bound ADP or AMP-PNP expose the conserved ATP

binding and hydrolysis motifs10 at the shared interface in a"

head-to-tail'

arrangement.

This has been widely accepted as physiologically relevant based on structural ' and

biochemical data11"13. The ATP hydrolysis sites in Savl866 are formed at the shared

interface and AMP-PNP is tightly sandwiched between the P-loop of one NBD and

the ABC signature motif of the other. A direct link between the two sites is created

that provides the molecular basis of the observed cooperativity in ATP binding and

hydrolysis14.

Figure 4.4: Ribbon representation of the NBD dimer of Savl866 in complex with

AMP-PNP. a, side view on the ABC cassette dimer and b, top view The

conserved sequence motifs P-loop (blue), ABC signature sequence (red; LSGGQ)

and the D-loop (orange) tightly bind AMP-PNP in ball-and-stick at the shared

interface in a 'head-to-tail' arrangement analogous to MJ0796

A comparison of the Savl 866 structure with bound nucleotide with the structure of

the E. coli vitamin Bi2 transporter BtuCD, crystallized in the absence of nucleotide15

revealed differences in the proximity of the sequence motifs. Whereas in Savl 866 the

ATP-bound conformation is observed, with AMP-PNP (or ADP) tightly sandwiched

between P-loop and ABC signature sequence, BtuCD revealed a substantial

separation of the analogous motifs in a nucleotide-free state (Figure 4.4). A

comparison of the NBDs of Savl 866 with the NBDs of another full ABC importer at

high resolution (2.4 Â), the BtuCD homolog H11470/1 from Haemophilus fulgidus16,

54

analogously demonstrated the differences. HI 1470/1 was crystallized in the inward-

facing, nucleotide free conformation. In contrast, the ATP-bound conformation of

Figure 4.5: Superposition of the NBDs of Savl866, MJ0796 and BtuCD. Both

P-loops of the NBD dimers were used for superposition, but for clarity only one

nucleotide binding site is shown in stereo. The C„ positions of conserved residues

GxxGxGKST (P-loop) and LSGGQ (ABC signature motifs) are depicted as

spheres in the backbone traces. Bound nucleotides arc in ball-and-stick.

Savl866 is consistent with the transmembrane domains adopting an outward-facing

conformation. The coupling of the ATP-bound state of the NBDs to the outward-

facing conformation of the TMDs is consistent with structural and biochemical data

from the literature highlighting the physiological relevance of the Savl866

conformation. The crystallization of Savl866 with ADP and its replacement with

AMP-PNP, that did not alter the conformation, indicates that the purification and

crystallization conditions, in particular the presence of C^Es, shifted the

55

conformational equilibrium of Savl866 towards the ATP-bound state. This is

consistent with the observation that addition of ADP stabilized the protein during the

concentration and that the ATPase activity of Savl866 was reduced in the presence of

DDM, C10E5 and a lipid environment (see Chapter V). Evidence from the literature

supported the observations. Experiments on ABC transporters in an altered chemical

(detergent) or lipid environment (lipid composition) demonstrated a clear influence of

the surfactants or lipids on the binding and hydrolysis of ATP (Km values and ATPase

rates) of the ABC transporters17"19.

Transmission interface

Conformational changes generated by ATP binding and hydrolysis are transmitted

from the nucleotide-binding domains to the transmembrane domains through non-

covalent interactions at the shared interface. The TMDs of Savl866 contribute to this

interface mainly through the intracellular loops ICL1 and ICL2. This is consistent

with genetic data that implicated ICL1 of human Mdrl to interact with the NBDs,

and with mutational studies and sequence comparisons that suggested ICL2 of

Tap 1/2, as well as ICL4 of CFTR, to provide similar crucial contacts'

.In Savl866,

both intracellular loops contain short helices oriented roughly parallel to the

membrane plane and providing the bulk of the contacts (Figure 4.6b). The so-called

'coupling helices', the name was first introduced in,are likely involved in the

transmission of mechanistically critical conformational changes. Conformational

changes generated in the NBDs are propagated by residues around the Q-loop to the

'coupling helices' to the transmembrane helices. As an anchor for the transporter

protein, small conformational changes at the NBDs might be amplified at the

transmission interface resulting in larger conformational changes at the extracellular

loops (e.g. to close the large cavity). Unexpectedly, the intracellular loops of Savl866

reach across and primarily contact the nucleotide-binding domain of the opposite

subunit (Figure 4.6), analogous to domain swapping observed in various enzymes .

While coupling helix 1 is in contact with the NBDs of both subunits, coupling helix 2

interacts with that of the opposite subunit exclusively. Such swapping has not been

anticipated for ABC exporters (Figure 4.8), but is not inconsistent with the available

biochemical or genetic data. By contrast, the transmembrane domains of the ABC

importer BtuCD and the close homolog H11470/1 contact only one nucleotide-binding

domain, resulting in a large gap at the center of the four protein domains'5, 16. The

56

architecture of the ICLs and coupling helices observed in Savl866 is likely conserved

among ABC exporters, as suggested by significant sequence similarity in the

corresponding regions (Figure 4.6b).

At the NBD-TMD interface similar contacts are provided by the nucleotide binding

domains in Savl866 and in the importers BtuCD and H11470/1. The NBDs exhibit a

cleft that forms the socket for ICL2 in Savl866 or for the second L loop helix in the

importers. The cleft is primarily lined with residues around the Q-loop ,as was also

observed in BtuCD and HU470/115'16. In Savl866, two prominent exceptions include

the conserved residues Tyr391 and Glu473. The latter appears particularly intriguing,

as it interacts with both intracellular loops and is part of a previously unrecognized,

short sequence motif (TBVGERG) that appears conserved in ABC export proteins

only (Figure 4.6b). The so-called 'x-loop', first introduced in22, highlights the

apparent function in cross-linking the ICLs. Because the x-loops precede the ABC

signature motifs, they likely respond to ATP binding and hydrolysis and may transmit

conformational changes to the intracellular loops by alternatingly engaging in, and

releasing, the cross-link upon ATP binding and hydrolysis. The absence of this motif

in binding protein-dependent, bacterial ABC importers suggests a distinct coupling

mechanism compared to ABC exporters.

57

a

«. f <

s. aux

s. tph

L. Jac

B. sub

H. sap

H, sap

H. sap

H. sap

TMD IGL1

1866

HbbA

LmrA

YvcC

Tapi

Tap2

Mdrl

Mdrl

TMD ICL2

S.aUL . _1S66

S,tph. _HfibAI.Jar» imrA

B. £u£>. _Y\ffC

ff.Sdp. Tflpl

H.sap. Tap2

//. sap. Mdrl

H.sap. Mdrl

NBD

s.aur._i866

s,tph. Msbd

L.iac._LmrA

B.sufc._YvcC

H,sap. Tapi

h.sap._Tap2

ff.sap._MdrlH. çap._Mdrl

f oiioltn i heii>

TM2-t h -*MWfc - -

91 KILYDIRKKLYNHLQALSPIFYÄHNQV-

95 KWMTMRRRLFGHMMGMPVAFFDKQST-

102 SVVKNLRTRVWDKMIHLPVKYFDEVKT-

"* 3

-QQVIS&VINDVEQT 131

-«JT1LSRITYQSEÇ>V 13i

-QEMSSRLANPTTQV 143

94 KIlbGLRELLWKKLIKLFVSYFDTHÄS—EETVSUVÏMBTMVV 134

255 IIVHSHLQGEVFUAVLRQErErFQQNQT~-ŒNIMS#VTEBI"S'lL 295

220 RJNLRÏREQI.F'SSLLRQI» Gl-'FQhTKI —SFLNS|M,SSjWTT,M 760

142 R01HK1RKQ1 FHA1MRQLIGWFDVHUV--QELNTftLIDOVSKl 182

783 ILTKRLRYMVFRSMLRODVSWFDDPKNTTGALTTItLANDAAQV 825

coupling he'iy t'

TM4-+ tHIMH f.flUHrV * f-TMS192 RSQALAEVpCFLHERVQciSV^SÏ'AIEDNEAKNFDKKNTHFL 234

1<56 MQHTMGQVTTSAEQMLKGHKFyLIFGGQEVETKRFDKVSNKMR 240

703 RQDSLAWFQGTASESLSETRr^KSSNAEKQASKKAEHDVNALY 245

19"> TQDETARFTGU.MQILPEIRIJÖ'KASNAEDVEYGRGKMGJSSLF 23/

356 VRESLAKSSQVAIEALSAMFT^RSFANEFGEAQKFRFKLQFTK J96

321 IüDAVARAGQWREAVGGLÖT/^RSFGAEEHEVCRYKEALEQCR 363

243 ellayakagavaeêvlaalr^iafggqkkeleryhknleeak 285

886 dkkelegsgkiatëaienfrtVvsltqeqkfehmyaqslqvpy 928

-iMIC-»»3 74 GMSGGGKSTLIH^IPRF§[.376 GHSGSGKSTIASÏ,ITRFÏ.382 GFSGGGKSTIFsijLERFY.1/4 gpsgggktîlfk&lerfIk:.b38 GPNGSGKSTVAA&LQNIjX.503 GPHGSGKSTVAA&LQHI#.

42 8 GNSGCGKSTTVQjliïMQRLÎÏ.1070 gssgcgkstwqllerfY.

\ ki >p ART '-\^ri ufp

èMl / . . GYDTEVG^RGVKLSGGQItQ 484

//..GLDTIIGENGVLLSGGöWl 4 8/

//. .QLHTEVGERGVKISGGQttij} 493

//..QFDTEVGERGIMLSGGQRÔ 485

/ / . . GYDTEVDEAGSQLSGGQ^O 64")

//..GIYTDVGEKGSQLAAGQKQ 613

//..KFDTLVGERGAQLSGGQAÛ 537

// . .KYSTOVGDKGTSLSGG0»01182

Figure 4.6: Transmission interface, a, Close-up view ofSav1866 with one subunit in

green ribbon, the other in yellow coil representation. For clarity, only the ICL

regions of the transmembrane domains are shown. Black spheres depict the first

and last C„ positions of the coupling helices'. Side chains of the conserved

Glu473 and Tyr391 are in ball-and-stick. b, Sequence alignment of Sav1866 with

the sequences of bacterial and human ABC transporters. Residues of high and

significant conservation arc highlighted in dark grey and light grey, respectively.

Relevant motifs are indicated. Residues of the ICLs interacting (4 Â cutoff) with

the NBD of the same subunit arc indicated green, residues contacting the opposite

NBD yellow. Residues of the NBD interacting with the TMD are highlighted in

red.

58

Substrate translocation pathway

A large cavity is present at the interface of the two transmembrane domains

(Figure 4.7). The cavity is shielded from the cytoplasm by a cluster of all twelve TM

helices that converge at the level of the membrane boundary and protrude 25 A into

the cytosol. Although shielded from the cytoplasm, the cavity reaches beyond the

intracellular membrane boundary. It is accessible from the outer leaflet and exposed

to the extracellular space. At the level of the inner leaflet, the cavity features a

hydrophilic surface that is primarily lined with polar and charged amino acid side

chains from helices TM2-TM5, with a slight surplus of negative charges and no

significant hydrophobic patches. At the level of the outer leaflet, the translocation

pathway is lined with residues from helices TM1, TM3, and TM6. The outward-

facing conformation and the predominance of polar and charged amino acids - a

hydrophilic character of the cavity - indicate a mode of substrate release. This also

suggests that rather than a high-affinity binding site, little or no affinity for

hydrophobic drugs is probable. This is consistent with the reported outward-facing

conformation of the ATP-bound state of human Mdrl, as revealed by electron

microscopy1' 24. In the presence of ATP, binding of substrates to human Mdrl was

indeed significantly reduced25. Similarly, LmrA from L, lactis exposed a high-affinity

binding site to the interior of the cell in the absence of nucleotide, but in the ATP-

bound state, the high affinity binding site was occluded and a low affinity binding site

was accessible from the extracellular space26.

It was suggested that substrate binding sites in ABC exporters are located in the

transmembrane domains27"29. So far, no well-defined substrate-binding sites could be

identified in ABC

59

Figure 4.7: Substrate translocation pathway, a, Backbone traces of the Sav 1866

subunits (green and yellow) in two orientations, rotated by 90 ° around a vertical

axis. The molecular surface of the central cavity is colored according to its

calculated electrostatic potential. The rest of the transporter is in light grey. Lines

indicate the approximate position of the two leaflets of the lipid bilaycr. Access

from the cavity to the outer leaflet of the lipid bilayer is visible in the right panel

from the front and back, between the wings' formed by the TMDs. b, Cavity at

the level of the inner leaflet viewed from the extracellular side. The TM helices are

numbered and the cavity is shown as a grey surface. C, Same as b but at the level

of the outer leaflet. Note that due to helix bending, different subsets of helices line

the cavity when compared to b.

exporters30 but the Sav 1866 architecture of Sav 1866 suggests that residues from all

transmembrane helices contribute to the surface of the translocation pathway. In

addition, residues from transmembrane helices TM1, TM4-TM6, and TM10-TM12

have been implicated in drug binding to human Mdrl,whereas residues from

60

multiple helices have been found to contribute to substrate binding to Tap 1/2.The

analogous residues in Savl866 indeed point towards the translocation pathway.

Transport mechanism

The simplest scheme of the transport mechanism invokes two states: An inward-

facing conformation with the substrate binding site accessible from the cell interior,

and an outward-facing conformation with an extrusion cavity exposed to the external

medium. The structure of Savl866 reveals that tight interaction of the nucleotide

binding domains in the ATP-bound state is coupled to the outward-facing

conformation of the transmembrane domains. In this conformation, bound substrates

may escape into the outer leaflet of the lipid bi layer or into the aqueous medium

surrounding the cell, depending on their hydrophobicity. Hydrolysis of ATP is

expected to return the transporter to an inward-facing conformation, again granting

access to the binding site from the cell interior. Indeed, the ABC importers HI 1470/1

from H. influencae clearly revealed an inward-facing conformation that is coupled to

a nucleotide-free state of the NBDs. Although differences exist between the TMDs of

Savl866 and H11470/1, both structures reveal striking similarities of the NBDs and

the way they are assembled to the TMDs via structurally conserved helices.

ABC transporters follow the concept of an 'alternating access and release' mechanism

that was first postulated for major facilitator transport proteins33, with the distinction

that ATP binding and hydrolysis, rather than substrate acquisition, may control the

conversion of one state into the other. Most ABC transporters bind and hydrolyze two

ATP molecules in each reaction cycle34'35. However, the number of bound substrates

can vary, and depending on the molecular mass, charge and structure, up to two

substrates have been found to enter the binding pockets of human Mdrl and MRP236'

.MDR-ABC transporters or exporters in general, might be exceptional because two

small substrates may be transported with an apparent stoichiometry of one hydrolyzed

ATP per transported substrate and single larger substrates with a stoichiometry of

two. Binding protein dependant ABC transporter systems, like the osmoregulated

ABC transporter OpuA from L. lactis34, BtuCD or HI 1470/1 have a transport

stoichiometry of two ATPs per transported substrate.

The architecture of the TMD-NBD interface of Savl866 defies the widely used ABC

exporter schematic shown in (Figure 4.8a). Rather than aligned side-by-side, the two

subunits of Savl866 are intricately associated (Figure 4.8b) and the main interface of

61

the intracellular loops is with the nucleotide-binding domain of the opposite subunit.

Furthermore, helices from both TMDs form the two 'wings' consisting of bundles of

helices embedded in the membrane (Figure4.la). Given these constraints, the two

subunits are unlikely to move independently and their maximum separation during the

reaction cycle is therefore limited. Our results challenge mechanistic models that

suggest that NBDs dimerize upon binding ATP and dissociate upon completing a

transport cycle30,38'39. Purified, isolated NBDs might tend to dissociate, but in the

context of the full assembly other observations were made. Electron microscopic

studies of full-length Mdrl and CFTR2,40 and the crystallographic analysis of BtuCD

and H11470/115'16 have demonstrated close proximity of the NBDs in the absence of

nucleotide.

Figure 4.8; ABC exporter schemes, a, earlier cartoons incorrectly depict two

compact halves (dark & light grey) arranged side-by-sidc, suggesting a

dissociation during the transport cycle, b, The schematic of Savl866 in the

observed, outward-facing, ATP-bound conformation emphasizes the domain

swapping and subunit twisting of two subunits (yellow & green). The grey box

indicates the lipid bilayer and the arrows indicate the release of bound drug into

the extracellular space.

Compared to Savl866, the ABC signature motifs of the nucleotide-free state of

BtuCD are separated by an additional ~5 A from the P-loops (Figure 4.5) - this is

similar for HI 1470/1. This is a conformational rearrangement that reduces the

interface between opposing nucleotide binding domains. Solvent accessibility of the

nucleotide-binding site in the bacterial maltose transporter is indeed increased in the

absence of ATP41 and the energetic coupling between side chains at the NBD

interface of CFTR is decreased.The 'power stroke' of ABC transporters is triggered

62

by binding of ATP, not hydrolysis, and might consist of moderate conformational

rearrangements that originate at the interface of the two nucleotide-binding domains,

propagate to the shared interface of NBDs and TMDs and amplify towards the

transmembrane domains up to the ECLs.

Physiologically and clinically relevant mutations occur in various human ABC

transporters of known substrates or proven implication in disease. The structure of

Savl866 provides useful information for the interpretation of the relevance of certain

mutations. One striking example is the deletion mutation À508 of CFTR, responsible

for 70% of cystic fibrosis cases42'43. The analogous residue in Savl866 is revealed as

Leu437 by superposition with the nucleotide-binding domain of wild type and À508

CFTR44. Leu437 is part of a cluster of aromatic and hydrophobic side chains at the

interface of ICL2 and the NBD. Given this structural context, the absence of the

phenylalanine side chain in CFTR A508, despite limited impact on the protein

backbone, likely destabilizes the analogous packing of residues at the interface of the

NBD1 and 1CL4, thus contributing to the observed misassembly of the mutant CFTR

protein.

Comparison of the Crystal Structures of Savl866 and MsbA

The crystal structure of Savl866 was compared to the three crystals structures of the

close ABC transporter homolog MsbA, the bacterial lipid A transporter. MsbA

molecules from E. coli, V. cholerae and S. typhimurium have been crystallized each in

a different conformation. When directly compared, all three crystal structures

contradicted the results obtained for Savl866 revealing crucial differences in the

transmembrane helix arrangement. The superposition of the Savl866 structure with

the structure of MsbA from S. typhimurium (2005) highlighted that none of the

transmembrane helices of Savl866 were oriented similar to the transmembrane

helices in MsbA (Figure 4.9a). The helix topology of Savl866 was consistent with

recent cross-linking data that identified neighboring helices in human Mdrl7. In

addition, the conformation of the mechanistically relevant sequence motifs in the

NBDs were, at the present resolution, indistinguishable from that observed for

MJ07968. Therefore, the mismatch in the transmembrane domains of MsbA and

Savl866 was suspected to originate from a problem during the structure

determination process of the MsbA structures.

63

A more detailed analysis of the MsbA structure revealed that, upon inversion of the

MsbA structure from 5*. typhimurium using crystallographic software, the inverted

structure could be superimposed on Savl866 with a rough match in the TMDs. Six

transmembrane helices of one subunit of MsbA, now left-handed due to the inversion,

and four TM helices from one and two TM helices from the other subunit of the

Savl866 molecule were apparently similar in their arrangement, but in fact the

connectivity and the orientations of the helices were different (Figure 4.9b). This

indicated a crystallographic mistake that might have occurred during the structure

Figure 4.9: Superposition of S aureus Savl866 and S typhimurium MsbA The

protein d trace of the TMDs of a single subunit of Msba (PDB TD' 1Z2R; purple)

and Savl866 (PDB ID: 2HYD; green) are shown, a, The two transporters were

manually superimposed resulting m no agreement in the helix arrangements b,

MsbA was inverted using the coordinate manipulation program pdbset,], and

superimposed manually on Savl866 The TMD of one subunit of inverted MsbA

is shown in purple. Subunit A of Sav1866 is shown m green and the helices TM1

and TM2 are shown in yellow. The approximate superposition of the TM helices is

evident.

64

determination process. Upon publication of the structure of Savl866 in 2006 the three

structures of MsbA45-47 were retracted and the crystallographic problem was

explained4""50.

During the structure determination, a computer program has changed the signs of the

anomalous differences in the experimentally derived anomalous data and produced

incorrect structure factors. Further used for the calculations, the structure factors

created incorrect electron density maps and, as a consequence, the structure was

incorrectly built.

In addition to the inherent inversion problem, the authors admitted in the retractions

that the wrong topology in the transmembrane domains was the result of breaks in the

electron density in the original structure of MsbA (2001) at the positions of the loop

regions. Transmembrane helices were incorrectly assigned and the problems not

recognized due to the low resolution of the electron density maps in all three

structures. However, the correctness of the three structures of MsbA has been

contested already before, the reason being an aberrant conformation of the nucleotide-

binding domains52. The functionally relevant sequence motifs in the nucleotide-

binding domains have not been found in the correct conformation.

Conclusions

The structure of Savl866 in complex with ADP or AMP-PMP defines, at high

resolution, the architecture of a bacterial MDR-ABC transporter in the outward-facing

conformation. The arrangement observed in Savl866 indicates an 'alternating access

and release' mechanism for ABC transporters. A single substrate translocation

pathway exposed to the extracellular space is evident and the arrangement of the

canonical twelve transmembrane helices forming the core of ABC exporter proteins is

visualized. The structure of Savl866 demonstrates the intricate association of two

homodimeric subunits with a domain-swapped NBD-TMD interface. The results

agree well with genetic, biochemical, and structural data of bacterial and human

homologues, and the observed conformation likely reflects a physiologically relevant

state. Furthermore, the structure serves as a model system of bacterial and human

homologues of ABC transporters and may initiate the rational design of drugs aimed

at interfering with the extrusion of agents used in chemotherapy.

65

References

1. Rosenberg, M. F., Callaghan, R., Modok, S., Higgins, C. F. & Ford, R. C. Three-

dimensional structure of P-glycoprotein: the transmembrane regions adopt an

asymmetric configuration in the nucleotide-bound state. J Biol Chem 280, 2857-

62 (2005).

2. Lee, J. Y., Urbatsch, I. L., Senior, A. E. & Wilkens, S. Projection structure of P-

glycoprotein by electron microscopy. Evidence for a closed conformation of the

nucleotide binding domains. JBiol Chem 277, 40125-31 (2002).

3. Gaudet, R. & Wiley, D. C. Structure of the ABC ATPase domain of human TAP 1,

the transporter associated with antigen processing. Embo J 20, 4964-72 (2001).

4. Holland, I. B., Cole, S. P. C, Kuchler, K. & Higgins, C. F. ABC proteins: from

bacteria to man (Academic Press, Amsterdam, 2003).

5. Stroud, R. M., Savage, D„ Miercke, L. J., Lee, J. K., Khademi, S. & Harries, W.

Selectivity and conductance among the glycerol and water conducting aquaporin

family of channels. FEBSLett 555, 79-84 (2003).

6. Abramson, J., Kaback, H. R. & Iwata, S. Structural comparison of lactose

permease and the glycerol-3-phosphate antiporter: members of the major

facilitator superfamily. Curr Opin Struct Biol 14, 413-9 (2004).

7. Stenham, D. R., Campbell, J. D., Sansom, M. S., Higgins, C. F., Kerr, I. D. &

Linton, K. J. An atomic detail model for the human ATP binding cassette

transporter P-glycoprotein derived from disulfide cross-linking and homology

modeling. FasebJ 17, 2287-9 (2003).

8. Smith, P. C, Karpowich, N., Milien, L., Moody, J. E., Rosen, J., Thomas, P. J. &

Hunt, J. F. ATP binding to the motor domain from an ABC transporter drives

formation of a nucleotide sandwich dimer. Mol Cell 10, 139-49 (2002).

9. Hopfner, K. P., Karcher, A., Shin, D. S., Craig, L., Arthur, L. M., Carney, J. P. &

Tainer, J. A. Structural biology of Rad50 ATPase: ATP-driven conformational

control in DNA double-strand break repair and the ABC-ATPase superfamily.

Cell 101, 789-800 (2000).

10. Schneider, E. & Hunke, S. ATP-binding-cassette (ABC) transport systems:

functional and structural aspects of the ATP-hydrolyzing subunits/domains. FEMS

Microbiol Rev 22, 1-20 (1998).

11. Hrycyna, C. A., Ramachandra, M., Ambudkar, S. V., Ko, Y. H., Pedersen, P. L.,

66

Pastan, I. & Gottesman, M. M. Mechanism of action of human P-glycoprotein

ATPase activity. Photochemical cleavage during a catalytic transition state using

orthovanadate reveals cross-talk between the two ATP sites. J Biol Chem 273,

16631-4(1998).

12. Chen, J., Sharma, S., Quiocho, F. A. & Davidson, A. L. Trapping the transition

state of an ATP-binding cassette transporter: evidence for a concerted mechanism

of maltose transport. Proc Natl Acad Sei USA9K, 1525-30 (2001).

13. Vergani, P., Lockless, S. W., Nairn, A. C. & Gadsby, D. C. CFTR channel

opening by ATP-driven tight dimerization of its nucleotide-binding domains.

Nature 433, 876-80 (2005).

14. Senior, A. E. & Bhagat, S. P-glycoprotein shows strong catalytic cooperativity

between the two nucleotide sites. Biochemistry 37, 831-6 (1998).

15. Locher, K. P., Lee, A. T. & Rees, D. C. The E. coli BtuCD structure: a framework

for ABC transporter architecture and mechanism. Science 296, 1091-8 (2002).

16. Pinkett, H. W., Lee, A. T., Lum, P., Locher, K. P. & Rees, D. C. An Inward-

Facing Conformation of a Putative Metal-Chelate-Type ABC Transporter. Science

(2006).

17. Borths, E. L., Poolman, B„ Hvorup, R. N., Locher, K. P. & Rees, D. C. In vitro

functional characterization of BtuCD-F, the Escherichia coli ABC transporter for

vitamin B12 uptake. Biochemistry 44, 16301-9 (2005).

18. Modok, S., Heyward, C. & Callaghan, R. P-glycoprotein retains function when

reconstituted into a sphingolipid- and cholesterol-rich environment. J Lipid Res

45,1910-8(2004).

19. Urbatsch, I. L. & Senior, A. E. Effects of lipids on ATPase activity of purified

Chinese hamster P-glycoprotein. Arch Biochem Biophys3\6, 135-40 (1995).

20. Currier, S. J., Kane, S. E., Willingham, M. C, Cardarelli, C. 0., Pastan, I. &

Gottesman, M. M. Identification of residues in the first cytoplasmic loop of P-

glycoprotein involved in the function of chimeric human MDR1-MDR2

transporters. J Biol Chem 267, 25153-9 (1992).

21. Gotten, J. F., Ostedgaard, L. S., Carson, M. R. & Welsh, M. J. Effect of cystic

fibrosis-associated mutations in the fourth intracellular loop of cystic fibrosis

transmembrane conductance regulator. J Biol Chem 271, 21279-84 (1996).

22. Dawson, R. J. & Locher, K. P. Structure of abacterial multidrug ABC transporter.

Nature 443, 180-5(2006).

67

23. Liu, Y. & Eisenberg, D. 3D domain swapping: as domains continue to swap.

Protein Sei 11, 1285-99 (2002).

24. Rosenberg, M. F., Velarde, G., Ford, R. C, Martin, C, Berridge, G., Kerr, I. D.,

Callaghan, R., Schmidlin, A., Wooding, C, Linton, K. J. & Higgins, C. F.

Repacking of the transmembrane domains of P-glycoprotein during the transport

ATPase cycle. Embo J 20, 5615-25 (2001).

25. Ramachandra, M., Ambudkar, S. V., Chen, D., Hrycyna, C. A., Dey, S.,

Gottesman, M. M. & Pastan, I. Human P-glycoprotein exhibits reduced affinity

for substrates during a catalytic transition state. Biochemistry 37, 5010-9 (1998).

26. van Veen, H. W., Margolles, A., Müller, M., Higgins, C. F. & Konings, W. N.

The homodimeric ATP-binding cassette transporter LmrA mediates multidrug

transport by an alternating two-site (two-cylinder engine) mechanism. Embo J 19,

2503-14(2000).

27. Dey, S., Ramachandra, M., Pastan, I., Gottesman, M. M. & Ambudkar, S. V.

Evidence for two nonidentical drug-interaction sites in the human P-glycoprotein.

Proc Natl Acad Sei USA 94, 10594-9 (1997).

28. Ambudkar, S. V., Dey, S., Hrycyna, C. A., Ramachandra, M., Pastan, I. &

Gottesman, M. M. Biochemical, cellular, and pharmacological aspects of the

multidrug transporter. Annu Rev Pharmacol Toxicol 39, 361-98 (1999).

29. Pleban, K., Kopp, S., Csaszar, E., Peer, M., Hrebicek, T., Rizzi, A., Ecker, G. F. &

Chiba, P. P-glycoprotein substrate binding domains are located at the

transmembrane domain/transmembrane domain interfaces: a combined

photoaffinity labeling-protein homology modeling approach. Mol Pharmacol 67,

365-74 (2005).

30. Higgins, C. F. & Linton, K. J. The ATP switch model for ABC transporters. Nat

Struct Mol Biol 11, 918-26 (2004).

31. Loo, T. W. & Clarke, D. M. Recent progress in understanding the mechanism of

P-glycoprotein-mediated drug efflux. JMembr Biol 206, 173-85 (2005).

32. Nijenhuis, M. & Hammerling, G. J. Multiple regions of the transporter associated

with antigen processing (TAP) contribute to its peptide binding site. J Immunol

157,5467-77(1996).

33. Jardetzky, O. Simple allosteric model for membrane pumps. Nature 211, 969-70

(1966).

34. Patzlaff, J. S., van der Heide, T. & Poolman, B. The ATP/substrate stoichiometry

68

of the ATP-binding cassette (ABC) transporter OpuA. J Biol Chem 278, 29546-51

(2003).

35. Sauna, Z. E. & Ambudkar, S. V. Evidence for a requirement for ATP hydrolysis at

two distinct steps during a single turnover of the catalytic cycle of human P-

glycoprotein. Proc Natl Acad Sei USA91, 2515-20 (2000).

36. Shapiro, A. B. & Ling, V. Positively cooperative sites for drug transport by P-

glycoprotein with distinct drug specificities. Eur JBiochem 250, 130-7 (1997).

37. Zelcer, N., Huisman, M. T., Reid, G., Wielinga, P., Breedveld, P., Kuil, A.,

Knipschecr, P., Schellens, J. H., Schinkel, A. H. & Borst, P. Evidence for two

interacting ligand binding sites in human multidrug resistance protein 2 (ATP

binding cassette C2). J Biol Chem 278, 23538-44 (2003).

38. Moody, J. E., Milien, L., Bums, D., Hunt, J. F. & Thomas, P. J. Cooperative,

ATP-dependent association of the nucleotide binding cassettes during the catalytic

cycle of ATP-binding cassette transporters. J Biol Chem 277, 21111-4 (2002).

39. Chen, J., Lu, G., Lin, J., Davidson, A. L. & Quiocho, F. A. A tweezers-like

motion of the ATP-binding cassette dimcr in an ABC transport cycle. Mol Cell 12,

651-61 (2003).

40. Awayn, N. H., Rosenberg, M. F., Kamis, A. B., Aleksandrov, L. A., Riordan, J. R.

& Ford, R. C. Crystallographic and single-particle analyses of native- and

nucleotide-bound forms of the cystic fibrosis transmembrane conductance

regulator (CFTR) protein. Biochem Soc Trans 33, 996-9 (2005).

41. Mannering, D. E., Sharma, S. & Davidson, A. L. Demonstration of

conformational changes associated with activation of the maltose transport

complex. J Biol Chem 276, 12362-8 (2001).

42. Kerem, B. S., Zielenski, J., Markiewicz, D., Bozon, D., Gazit, E., Yahav, J.,

Kennedy, D., Riordan, J. R., Collins, F. S., Rommens, J. M. & et al. Identification

of mutations in regions corresponding to the two putative nucleotide (ATP)-

binding folds of the cystic fibrosis gene. Proc Natl Acad Sei USA 87, 8447-51

(1990).

43. Tsui, L. C. The spectrum of cystic fibrosis mutations. Trends Genet 8, 392-8

(1992).

44. Lewis, H. A., Zhao, X., Wang, C, Sauder, J. M., Rooney, L, Noland, B. W.,

Lorimer, D., Kearins, M. C, Conners, K., Condon, B., Maloney, P. C, Guggino,

W. B., Hunt, J. F. & Emtage, S. Impact of the deltaF508 mutation in first

69

nucleotide-binding domain of human cystic fibrosis transmembrane conductance

regulator on domain folding and structure. J Biol Chem 280, 1346-53 (2005).

45. Chang, G. & Roth, C. B. Structure of MsbA from E. coli: a homolog of the

multidrug resistance ATP binding cassette (ABC) transporters. Science 293, 1793-

800(2001).

46. Chang, G. Structure of MsbA from Vibrio cholera: a multidrug resistance ABC

transporter homolog in a closed conformation. J Mol Biol 330,419-30 (2003).

47. Reyes, C. L. & Chang, G. Structure of the ABC transporter MsbA in complex

with ADP.vanadate and lipopolysaccharide. Science 308, 1028-31 (2005).

48. Chang, G., Roth, C. B., Reyes, C. L., Pornillos, 0., Chen, Y. J. & Chen, A. P.

Retraction. Science 314, 1875 (2006).

49. Miller, G. Scientific publishing. A scientist's nightmare: software problem leads to

five retractions. Science 314, 1856-7 (2006).

50. Miller, C. Pretty structures, but what about the data? Science 315, 459 (2007).

51. The CCP4 suite: programs for protein crystallography. Acta Crystallogr D Biol

Crystallogr 50, 760-3 (1994).

52. Davidson, A. L. & Chen, J. Structural biology. Flipping lipids: is the third time the

charm? Science 308, 963-5 (2005).

70

Chapters: Functional Reconstitution of Savl866, Basal and

Drug-stimulated ATPase Activity

Abstract

ABC transporters couple the translocation of substrates across cellular membranes to

the binding and hydrolysis of ATP. In human tumor cells, multidrug (MDR) ABC

transporter proteins like human Mdrl facilitate the extrusion of agents used in

chemotherapy contributing to multidrug resistance. To complement the structural

studies, the basal and drug-stimulated ATPase activity of the MDR-ABC transporter

homolog Savl866 was measured. The protein exhibited V04-inhibitable basal

ATPase activity in C^Ea and in proteoliposomes with hydrolysis rates of 341 ±6

nmol ATP/(mg protein • min) and 63 ± 5 nmol ATP/(mg protein • min), respectively.

In the presence of the cancer drugs Doxorubicin, Vinblastine, and the fluorescent dye

Hoechst 33342 a 1.3-, 2.1- and 3.6~fold increase of the ATPase activity of Savl866 in

liposomes was measured. The results indicate that Savl866 features all the

characteristics of a MDR-ABC transporter.

Introduction

Eukaryotic multidrug (MDR) efflux pumps like Mdrl or MRP2 are ABC transporters

that have pivotal role in host detoxification'. In tumor cells, they cause multidrug

resistance against agents used in chemotherapy2. Prokaryotic MDR-ABC transporters

contribute to drug and antifungal resistance3, 4but have been found to functionally

substitute for their eukaryotic family members5. MDR-ABC transporters transport a

wide variety of neutral and positively charged, hydrophobic compounds like vinca

alkaloids and anthracyclines6 and couple their translocation to the binding and

hydrolysis of ATP.

The ATPase activity of Savl866 protein was investigated to complement the

structural studies on the model system and to provide experimental proof that

Savl866 is a MDR-ABC transporter. The basal ATPase activity of Savl866 protein

was determined in detergent and in proteoliposomes. In addition, the stimulation and

inhibition of the ATPase activity of reconstituted Savl866 in proteoliposomes was

analyzed in the presence of various known substrates of the multidrug ABC

transporters LmrA fromZ,. lactis and human Mdrl.

71

Materials and Methods

Expression and Purification

S. aureus Savl866 was purified in Ci2E8 as detergent as described in Chapter II. The

protein was concentrated in 10 mM Tris pH 8.2, 100 mM NaCl to 15 mg/ml using an

Amicon Ultra-15 concentrator unit (Millipore) with a molecular cutoff of lOOkDa.

Concentrated protein was analyzed by size exclusion chromatography using a

Superdex 200 10/300 GL column (GE Healthcare).

Functional Reconstitution of Savl866 Protein into Proteoliposomes

Purified Savl866 was reconstituted based on previously developed protocols .

Chloroform dissolved Egg yolk L-R-phosphatidylcholine (Avanti Polar Lipids) and

E. coli polar lipid extract (Avanti Polar Lipids) were mixed at a ratio of 1:3 in a

round-bottomed flask and the organic solvent was evaporated at room temperature

under an argon stream. Residual solvent was removed by lyophilization and the

mixture was subsequently hydrated in a buffer containing 50 mM Tris (HCl) pH 7.5

and 150 mM NaCl by ultra sonication. The final lipid concentration was 20 mg/ml.

The turbid suspension was extruded through a 400 nm membrane using a mini

extruder (Avanti Polar Lipids), flash frozen in liquid nitrogen and stored at -80 °C.

Lipids were thawed and pre-incubated with 0.14 % (w/v) Triton X-100 for 1 hour at

room temperature. Freshly purified and concentrated Savl866 at 10 mg/ml in 50 mM

Tris (HCl) pH 7.5, 150 mM NaCl, 0.5 mM EDTA and 0.01 % (w/v) octaoxyethylene-

n-dodecylether (Ci2E8, Anatrace) was supplemented with 0.14 % (w/v) Triton X and

mixed with the lipids (4 mg/ml) to yield a molar protein to lipid ratio of-1:50. The

mixture was equilibrated at room temperature for 2 hours by gentle agitation. The

detergent was removed by adding BioBeads SM2 (BioRad) at 40 mg/ml wet weight

and the solution was agitated at room temperature for 15 minutes. Similar aliquots of

Biobeads were added four times for the incubation periods of 15 min, 30 min,

overnight and 60 min at 4 °C according to8' 9. Previously added Biobeads were

removed after each step. The turbid suspension was centrifuged at 184000xg for 20

min in a TLA 120.1 rotor and the proteohposome pellet washed by gentle

resuspension and recentrifugation. The final buffer contained 25 mM Hepes (NaOH)

pH 7.5 and 150 mM NaCl. Proteoliposomes were flash-frozen at 20 mg/ml lipids in

liquid nitrogen and stored at -80 °C.

72

Functional Assay - Drug-stimulated ATPase Activity of Savl866

The ATPase activity of detergent-purified Savl866 (C^Eg) and reconstituted Savl 866

in proteoliposomes was measured at room temperature in reaction volumes of 350 ul8.

Reactions of 1 uM protein were prepared in a buffer containing 25 mM Hepes

(NaOH) pH7.5 and 150 mM sodium chloride with or without 0.01 % (w/v) C,2EX,

respectively, and incubated at room temperature for 3 minutes. ATP hydrolysis was

initiated by addition of 2 mM ATP and 5 mM MgCb to the mixture and stopped by

SDS inhibition. Samples of 50 ul volume were removed at various time points and

added to 50 u.1 of a 12 % (w/v) SDS solution. For inhibition, 1 mM freshly boiled

(5 min at 95 °C) sodium ortho-vanadate was added to the solution. Inorganic

phosphate was assayed as described earlier10. The effect of putative substrates on the

basal ATPase activity was assayed by addition of various concentrations of drug to

the reaction mixtures.

Results and Discussion

Basal ATPase Activity ofS. aureus Savl866

All ABC transporters exhibit significant ATPase activity in the absence of the

transport substrate. So far, the only exception is the osmoregulated ABC transporter

OpuA from L. lactis that tightly couples the hydrolysis of ATP to substrate

translocation9. In order to complement the structural studies, the ATPase activity of

Savl866 protein was analyzed in detergent solution (C^E«) and after reconstitution

into proteoliposomes9 (Figure 5.1 a,b). Therefore, Savl866 protein was purified in

CnEg as described in Chapter 2 and purified protein was reconstituted into

proteoliposomes.

Reconstituted ABC transporters can be oriented in two different ways in sinle

unilamellar vesicles (SUVs): right-side-in or inside-out. Assays on the E. coli vitamin

Bi2 importer BtuCD clearly demonstrated 93 % of the reconstituted protein to be in

the inside-out orientation8. BtuCD inserts with its membrane spanning domains first

and exposes the large hydrophilic nucleotide binding domains at the vesicle surface.

Savl866 shares the overall domain organization of BtuCD and likely incorporates

inside-out to a similar extend. Only NBDs exposed at the surface of SUVs are

accessible for ATP and bind and hydrolyze ATP (Figure 5.1a).

73

So far, reconstitution protocols were reported to be optimal using Triton"5 X-100 'but

reconstitutions have also been performed with detergents like n-dodecyl-

ß-D-maltopyranoside (DDM). Savl866 was purified in C|2ES and Triton® X-100 was

used to pre-condition the lipids (1 h preincubation) and to initiate the insertion

reaction. Reconstituted protein was analyzed by SDS-PAGE and the protein

concentrations determined. A loss of about 20 % of protein was estimated and,

although Biobeads might have absorbed lipids as well, no losses of lipids were

assumed. An average molecular weight value of -920 Da for the lipid mixture was

used for the calculations.

Figure 5.1: Schematic representation of the experimental setup for the m vitro

assays investigating basal ATPase activity of Savl866. a, Detergent-purified and

reconstituted protein to proteoliposomes (SUVs) with most of the transporters in

the inside-out orientation b, C^Ea-solubilized and purified Savl 866 protein.

In a first experiment, 40 % of the Savl 866 protein at 0.5 mg/ml was reconstituted at

room temperature for 1 hour with a final molar protein to lipid ratio of >I :2000. When

concentrated to ~8 mg/ml a 10-fold increase of the reconstituted protein in the

liposomes and a molar protein to lipid ratio of ~ 1:200 was observed (Figure 5.2a).

The reconstitution rate increased 2-fold applying concentrated protein at 10 mg/ml

and a doubled incubation time of 2 hours for the insertion reaction. The reconstitution

rate increased by 2-fold to a value of 80 % (Figure 5.2b). The final protein to lipid

ratio was estimated to be ~1:100 judged by SDS-PAGE analysis.

74

a b12 3 4 5 6 7

A^"^ ^^^^ ^v |

Figure 5.2: SDS-PAGE analysis of the reconstitution of C,2Es-purificd Savl866

mediated by Triton© X-100. a, Savl866 protein at 0.5 mg/ml concentration (1)

and the reconstituted protein (2); Protein at 8 mg/ml concentration (3) and the

reconstituted protein (4). If the reconstitution efficiency was 100 %, the intensities

of the bands in (2) and (4) would match those of (1) and (3), respectively, b,

Savl866 protein at 10 mg/ml concentration (5) and reconstituted protein derived

from reconstitutions with incubation times of 2 hours (6) and 1 hour (5).

Measurements of the basal ATPase activity of Savl866 in detergent and of

reconstituted Savl866 to proteoliposomes was performed at protein concentrations of

1 |xM in the presence of 5 mM Mg2+ and 2 mM ATP. An inorganic phosphate assay

was applied that was insensitive for detergent and most of the chemicals used in

biological buffers10. The Km of Savl866 for ATP was not experimentally determined

but was assumed to be similar to the Km values of 0.4-0.8 mM for importers and

exporters and therefore significantly below the ATP concentration applied for the

assay12"15. In C,2ES, Savl866 hydrolyzed ATP with a rate of 341 ± 6 nmol ATP/(mg

protein - min), whereas in proteoliposomes, the rate decreased to 63 ± 5 nmol

ATP/(mg protein • min). The reported hydrolysis rates were within the order of

magnitude of hydrolysis rates reported for MDR-ABC transporters like LmrA16 and

Mdrl15' lv. Reconstituted LmrA to proteoliposomes hydrolyzed ATP with a rate of

200 ± 20 nmol ATP/(mg protein • min) and reconstituted Mdrl hydrolyzed ATP

depending on the lipid composition of the liposomes with a rales between

62 ± 9 nmol ATP/(mg protein - min) and 104 ± 37 nmol ATP/(mg protein • min).

In addition, the ABC transporter specific inhibition of the basal ATPase activity with

ortho-vanadate was also observed for Savl866. Surprisingly, the inhibition was

complete in detergent solution but the degree of completeness in proteoliposomes was

not evident from the measurements. Other inhibitors like AMP-PNP at 25 uM or

AMP-PCP at 300 uM concentration - non-hydrolyzable analogs of ATP - did not

show any inhibition, likely due to the lower affinity of AMP-PNP when compared

with that of ATP12' 14. An increase of the basal ATPase activity of reconstituted

75

Savl866 in proteoliposomes was observed upon addition Ci2Eg detergent to the

reaction solution.

The dynamic detergent environment of Savl866 was replaced with E. coli

Upïd/Eggyolk phophatidylcholine by reconstitution, and reproducibly, reversibly

caused a 6-fold decrease of the basal hydrolysis rate of the protein. This demonstrates

a dramatic influence on the ATPase activity caused by the altered chemical

environment. Effects of surfactants on transport proteins were reported for C^Es on

membrane embedded (Na^K^-ATPases18 and OG on hamster Mdrl in plasma

membranes19. E. coli BtuCD, purified in four different detergents, exhibited

significantly different hydrolysis rates and varying inhibition behavior in chemically

different environment of the four detergents . Analogously, in Sav 1866, the ATPase

activity of Ci2Ex-purified protein was lower in the presence of C10E5 or DDM and in

DDM-purified Savl866 (not shown). Addition of zwitterionic detergents completely

abolished the ATPase activity. Furthermore, the purification of Sav 1866 in FOS12,

FOS14 and LDAO inactivated the protein irreversible although the protein appeared

to remain biochemically stable. A reconstitution of FOS12-purified protein did not

recover the ATPase activity.

It was reported that reconstituted hamster Mdrl into proteoliposomes exhibited

different levels of basal ATPase activity in liposomes of different lipid

compositions15, 19. It was suggested that the lipid composition in the proteoliposomes

had directly affected binding and hydrolysis of ATP (Km and ATPase rate). Similar to

surfactants, specific lipids or surfactants and their mixes, have the power to shift

conformational equilibria of MDR-ABC transporters. It could be shown in these

experiments that the chemically altered environment at the transmembrane domains

directly affected the binding and hydrolysis of ATP in the nucleotide-binding

domains. Further experiments on Mdrl demonstrated an influence of sphingolipids or

cholestérols on substrate transport and drug resistance. An increase of the modulation

of the ATPase activity of reconstituted human Mdrl by substrates was observed in

sphingolipid-based liposomes15 and elevated levels of sphingolipids demonstrated in

multidrug resistant cancer cell lines20.

76

-, , , , , , , , j-> J-, .'

i i » 1 ' r'

» Ml Zll 111 411 II * II) I* 211

linn |tllill| turn {mini

Figure 5.3: Basal ATPase activity of Savl866 protein and its ABC transporter

characteristic inhibition of the ATP hydrolysis reaction with sodium ortho-

vanadatc (1 mM). The error bars indicate a 3-fold redundant measurement.

Savl866 ATPase activity was assayed with a, CnEx-purified protein and b, with

C^Eii-purified and reconstituted protein.

Drug-stimulation of the Basal ATPase Activity of Savl866 Protein

The ATPase activity of MDR-ABC transporters is sensitive to substrates. In many

cases the stimulation or inhibition of the ATPase activity was reported. Therefore,

these effects are experimentally used to analyze the characteristics of the transport

protein in cellular membranes (in vivo) and proteoliposomes (in vitro). Reconstituted

Savl866 in proteoliposomes was used and the hydrolysis rates were measured in the

presence or the absence of putative substrates. Substrates were selected from a pool of

known transport substrates of the MDR-ABC transporter homologs L. lactis LmrA

and human Mdrl '. Drugs at different concentrations were incubated with the reaction

mix for 5 minutes prior to the addition of ATP for the start of the hydrolysis reaction.

For the applied cancer drugs Doxorubicin (anthracycline) and Vinblastine

(vinca alkaloid) and for the fluorescent dye Hoechst 33342, clear stimulation patterns

were observed (Figure 5.4). With increasing amounts of substrate, an increase of the

stimulation of the ATPase activity has been observed. While for the Hoechst 33342

dye a 3.6-fold stimulation at a concentration of 100 uM was observed, Doxorubicin

and Vinblastine stimulated 1.3-fold and 2.1-fold, respectively, at the maximally

applied concentration of 1 mM (Figure 5.4).

jHypr*

——hocclKt 33342

—•— vinblastine

—— doxoiubicin

i

ï

j—! , ! 1 ! 1 ! , ( , ,_!0 20111 400 600 800 1000

L-<>ut:?iitralmil »f tiansport Mibstiale \\iM\

Figure 5.4: Drug-stimulated ATPase activity of the reconstituted Sav1866 protein

into liposomes in the presence of the fluorescent dye Hoechst 33342 and the

cancer drugs Vinblastine and Doxorubicin at various concentrations. Initial

ATPase rates were determined in triplicates by linear regression of time points

similar to the experiments shown in Figure 3.4b.

Stimulation of less than 1.3-fold was measured for the cancer drug Daunorubicin

(anthracycline) and Rhodamine 123, Rhodamine 6G, Reserpine and Actinomycin D.

A biphasic pattern was observed for Rhodamine 123, Rohodamine 6G, Reserpine and

Actinomycin D demonstrating a stimulation and inhibition of the ATPase activity at

various substrate concentrations (Figure 5.5a,b). Verapamil, a Mdrl blocking agent,

inhibited the ATPase activity at low micromolar concentrations.

The stimulation and inhibition of the ATPase activity of reconstituted Savl866 by

typical MDR-ABC transporter substrates demonstrated that Savl866 is a MDR-ABC

transporter .The molecular mechanism of stimulation and inhibition by substrates

and inhibitors is not understood but a strong dependency on the relative

concentrations of the transporter in proteoliposomes and the drugs is apparent. To

make a molecular characterization even more problematic, in some cases, known

substrates are reported not to contribute to stimulation21.

3.5-I-1 V.

JL\t

g 3.ft-

1N

41(A

Cu

H 2.5-

1

re

;| z.o-

I1-5- ^1.0- u —I

'

78

1 lit

I (I*

£ 1 (III

^ If"

I (I *>U -

t <IK*

"

tl Nil

tiili limn ol !i,IH\|iml \nl)Mi IU |mM| loiiU[Hi moil oi K<.sti|>iiK ||iM|

Figure 5.5: Stimulation and inhibition of the ATPase activity of reconstituted

Savl866 into liposomes by a, Rhodamine 123 and Rhodamine 6G and b,

Rcserpine at various concentrations. Initial ATPase rates were determined in

triplicates by linear regression of time points similar to the experiments shown in

Figure 3.4b.

Conclusion

Savl866 exhibited V04-inhibitable basal ATPase activity in C^Eg and in

proteoliposomes typical for MDR-ABC transporters. Various detergents influenced

the ATPase activity of reconstituted Savl866 in proteoliposomes suggesting that the

chemical environment of the transport protein influences directly on binding and

hydrolysis of ATP. In the absence of direct evidence for transport, assays quantifying

the effect of putative substrates on the ATPase activity of reconstituted Savl866

provided valuable information. Transport substrates of MDR-ABC transporters like

Mdrl and LmrA when used in the ATPase activity assays caused the stimulation or

inhibition of the ATPase activity. The observed data indicated a direct influence on

the Savl866 protein and the coupling of the hydrolysis reaction. Therefore,

Doxorubicin, Vinblastine and Hoechst 33342 may also belong to the group of

transport substrates of Savl866. The experimental data complements the genetic and

structural data and suggests that Savl866 is a homolog of the MDR-ABC transporter

proteins.

References

1. Szakacs, G., Paterson, J. K.., Ludwig, J. A., Booth-Genthe, C. & Gottesman, M.

79

M. Targeting multidrug resistance in cancer. Nat Rev Drug Discov 5, 219-34

(2006).

2. Gottesman, M. M. & Ambudkar, S. V. Overview: ABC transporters and human

disease. JBioenerg Biomembr 33, 453-8 (2001 ).

3. Kontoyiannis, D. P. & Lewis, R. E. Antifungal drug resistance of pathogenic

fungi. Lancet 359, 1135-44 (2002).

4. Sanglard, D. & Odds, F. C. Resistance of Candida species to antifungal agents:

molecular mechanisms and clinical consequences. Lancet Infect Dis 2, 73-85

(2002).

5. van Veen, H. W., Callaghan, R., Soceneantu, L., Sardini, A., Konings, W. N, &

Higgins, C. F. A bacterial antibiotic-resistance gene that complements the human

multidrug-resistance P-glycoprotein gene. Nature 391, 291-5 (1998).

6. Seelig, A., Blatter, X. L. & Wohnsland, F. Substrate recognition by P-

glycoprotein and the multidrug resistance-associated protein MRP1: a

comparison. Int J Clin Pharmacol Ther 38,111-21 (2000).

7. Poolman, B., Doeven, M. K., Geertsma, E. R., Biemans-Oldehinkel, E., Konings,

W. N. & Rees, D. C. Functional analysis of detergent-solubilized and membrane-

reconstituted ATP-binding cassette transporters. Methods Enzymol 400, 429-59

(2005).

8. Borths, E. L., Poolman, B., Hvorup, R. N., Locher, K. P. & Rees, D. C. In vitro

functional characterization of BtuCD-F, the Escherichia coli ABC transporter for

vitamin B12 uptake. Biochemistry 44, 16301-9 (2005).

9. Patzlaff, J. S., van der Heide, T. & Poolman, B. The ATP/substrate stoichiometry

of the ATP-binding cassette (ABC) transporter OpuA. J Biol Chem 278, 29546-51

(2003).

10. Chifflet, S., Torriglia, A., Chiesa, R. & Tolosa, S. A method for the determination

of inorganic phosphate in the presence of labile organic phosphate and high

concentrations of protein: application to lens ATPases. Anal Biochem 168, 1-4

(1988).

ll.Knol, J., Sjollema, K. & Poolman, B. Detergent-mediated reconstitution of

membrane proteins. Biochemistry 37, 16410-5 (1998).

12. Liu, C. E., Liu, P. Q. & Ames, G. F. Characterization of the adenosine

triphosphatase activity of the periplasmic histidine permease, a traffic ATPase

(ABC transporter). J Biol Chem 272, 21883-91 (1997).

80

13. Steinfels, E., Orelle, C, Fantino, J. R., Dalmas, O., Rigaud, J. L., Denizot, F., Di

Pietro, A. & Jault, J. M. Characterization of YvcC (BmrA), a multidrug ABC

transporter constitutively expressed in Bacillus subtilis. Biochemistry 43, 7491-

502 (2004).

14. Urbatsch, I. L., al-Shawi, M. K. & Senior, A. E. Characterization of the ATPase

activity of purified Chinese hamster P-glycoprotein. Biochemistry 33, 7069-76

(1994).

15. Modok, S., Heyward, C. & Callaghan, R. P-glycoprotein retains function when

reconstituted into a sphingolipid- and cholesterol-rich environment. ,/ Lipid Res

45, 1910-8(2004).

16. Vigano, C, Grimard, V., Margolles, A., Goormaghtigh, E., van Veen, H W.,

Konings, W. N. & Ruysschaert, J. M. A new experimental approach to detect

long-range conformational changes transmitted between the membrane and

cytosolic domains of LmrA, a bacterial multidrug transporter. FEBS Lett 530,

197-203 (2002).

17. Ramachandra, M, Ambudkar, S. V., Chen, D., Hrycyna, C. A., Dey, S.,

Gottesman, M. M. & Pastan, I. Human P-glycoprotein exhibits reduced affinity

for substrates during a catalytic transition state. Biochemistry 37, 5010-9 (1998).

18. Huang, W. H., Kakar, S. S. & Askari, A. Mechanisms of detergent effects on

membrane-bound (Na+ + K+)-ATPase. J Biol Chem 260, 7356-61 (1985).

19. Urbatsch, I. L. & Senior, A. E. Effects of lipids on ATPase activity of purified

Chinese hamster P-glycoprotein. Arch Biochem Biophys 316, 135-40 (1995).

20. Lavie, Y., Cao, H., Bürsten, S. L., Giuliano, A. E. & Cabot, M. C. Accumulation

of glucosylceramides in multidrug-resistant cancer cells. J Biol Chem 271, 19530-

6(1996).

21. Sharom, F. J. The P-glycoprotein efflux pump: how does it transport drugs? ./

MembrBiol 160, 161-75 (1997).

81

Chapter 6: Conclusions and Outlook

Crystal structures are important prerequisites for the elucidation of molecular

mechanisms. Even though the purification and crystallization of membrane proteins is

notoriously difficult, the resulting information is of high value. It provides a

fundamental understanding of the architecture and function of the protein. The risk to

fail finding an ABC transporter protein that is stable and active in detergent solution

was minimized by applying a homology screening approach. Although the function of

the various ABC transporter homologs was not known, the generality of the approach

permitted a rather fast identification of a crystallizable protein. Kendrew was

screening for crystals of myoglobins from different biological sources suitable for

detailed study by x-ray diffraction in 1954. The approach in this thesis was to screen

for multidrug ABC transporter proteins from different biological sources suitable for

crystallization. The prerequisite for this was to discover a state of a transport protein

in detergent solution that corresponds to its native state in lipid membranes and to

chemically conserve it for the crystallization. Over sixty years after the homology

screening approach was used for the screening of myoglobin homologs, it is still

useful and has lead to a structural model of a multidrug ABC transporter.

The structure of the previously unstudied protein Savl866 from S. aureus describes

the first ABC exporter structure after the retraction of the three ABC exporter

structures from MsbA. It defines the architecture of a bacterial ABC transporter in the

physiological relevant outward-facing, ATP-bound conformation and suggests an

'alternating access and release' mechanism under the control of binding and

hydrolysis of ATP. Direct evidence is provided for a coupling of the ATP-bound state

of the nucleotide-binding domains to the outward-facing conformation of the

transmembrane domains. In essence, the observed conformation is the result of an

induction of conformational changes in the nucleotide-binding domains upon binding

of ATP and the transmission of these to the transmembrane domains to reveal the

outward-facing extrusion cavity. This action characterizes the 'power stroke' in ABC

transporters and the Savl866 structure directly associates it to ATP-binding, not to

hydrolysis. Therefore, ATP-binding empowers the major conformational changes that

are suggested to be responsible for the translocation of substrates against a

concentration gradient. The subsequent hydrolysis of ATP and the release ofADP and

inorganic phosphate converts the ABC transporter back to an inward-facing state that

82

likely exhibits one or more high-affmity substrate binding sites. The binding of

substrate might trigger the binding of ATP molecules for the next hydrolysis and

transport cycle.

Unexpectedly, the part of the Savl866 protein that connects the NBDs and the TMDs,

the transmission interface, was observed in a domain swapped arrangement. The

'coupling helices' are crucial for the propagation of the mentioned conformational

changes during the 'power stroke' and are important for the structural integrity of

Savl866. The intricate assembly seems to allow for a separation of the nucleotide-

binding motifs in the NBDs but permits their dissociation. Minor conformational

changes in the NBDs are propagated and amplified to rearrange the transmembrane

helices that form the large cavity and the 'wings' that create the opening to the outer

leaflet of the membrane. Therefore, transported substrates may be repelled by the

hydrophilic surface of the extrusion pocket to the outer leaflet of the membrane or to

the extracellular space.

The general architecture ofSavl866 appears structurally conserved in ABC exporters

because of the core arrangement of twelve transmembrane helices and the significant

conservation of the 'coupling helices'. Given that structural constraints exist and that

a certain topology can be established on the basis of alternative amino acid

combinations, the structure of Savl866 might be useful for the interpretation of

clinically relevant mutations in human ABC transporters causing disease.

To understand the details of the mechanism of ATP-dependant transport, it is essential

to crystallize an ABC exporter in the inward-facing conformation that likely exposes

the substrate binding sites. One way to achieve this involves the screening and the

analysis of ABC exporter homologs in the absence of nucleotides. Another way

involves the engineering of Sav 1866 protein or other ABC transporter homologs. The

conformational equilibrium of Savl866 appeared to be shifted in detergent solution

and the outward-facing conformation dominated. Therefore, mutations and covalent

cross-unking at crucial positions in the protein may help to shift the equilibrium

favoring the inward-facing conformation.

In addition to the characterization of the inward-facing state by means of

crystallography, information about ABC transporters in a more native environment

has to be collected. Therefore, reconstituted wild-type and mutant Sav1866 protein in

proteoliposomes has to be produced and analyzed with different techniques. One of

these techniques uses photoreactive substrate analogs of multidrug ABC transporters

83

that bind and subsequently link covalently to residues in the substrate binding sites.

Afterwards, the residues are identified and located in the high-resolution structure of

either the structure in the outward-facing state or a future structure in the inward-

facing state. As a consequence, this might initiate the rational design of drugs aimed

at interfering with the extrusion of agents used in chemotherapy.

84

Publications

Dawson, R. J. P. & Locher, K. P.

Structure of the Multidrug ABC Transporter Savl 866 from Staphylococcus aureus in complex with

AMP-PNP.

FEBS lett. 581, 935-938 (2007)

Dawson, R. J. P. & Locher, K. P.

Structure of a Bacterial Multidrug ABC Transporter.Nature 443, 180-185(2006)

Dawson, R., Mullcr, L., Dchncr, A., Klein, C, Kessler, H., Buchner, J,

The N-terminal Domain of p53 is Natively Unfolded.

J. Mol. Biol. 332, 5, 1131-1141 (2003)

Dawson, R.

Strukturelle Charakterisierung von p53-FragmentenDiploma thesis, Department of Chemistry, Technical University of Munich (2003)

85

Curriculum vitae

Personal information

Name Dawson Roger John Peter

Address Universitätsstrasse 15, CH-8006 Zürich

Telephone +41 76 5860023

E-mail roger.dawson@moI.biol.ethz.ch

Nationality British

Date of birth 09.05.1977

Education - Studies

Oct 2003 - current PhD Student

Group of Prof. Dr. Kaspar Locher, Institute of Molecular Biologyand Biophysics, Swiss Federal Institute of Technology Zurich

(Switzerland)Structural and functional analysis of multidrug ABC transporters

Apr 2003 - Jun 2003 Research Associate

Group of Prof. Dr. Johannes Büchner, Institute for Organic

Chemistry and Biochemistry, Technical University of Munich

(Germany)

Analysis of the tumor suppressor protein p53

Sep 2002 - Mar 2003 Diploma Student

Group of Prof. Dr. Johannes Büchner, Institute for Organic

Chemistry and Biochemistry, Technical University of Munich

(Germany)Structural characterization of human p53-fragmcnts

NOV 1997 - Mar 2003 Studies

Chemistry - Technical University of Munich (Germany)

Practical Training

Oct 2001-Nov 2001 Trainee

Group of Prof. Dr. Robert Hubcr, Department for Structural

Research, Max-Planck-Institute for Biochemistry (Munich -

Germany)Structural analysis of the proteolytic system HslUV

Mar 2001 Research Trainee

Group of Prof. Dr. Johannes Büchner, Institute for OrganicChemistry and Biochemistry, Technical University of Munich

(Germany)Production and purification of the heat shock protein Hsp26

Mar 2001 Apr 2001 Trainee

Group ofProf. Dr. Horst Kessler, Institute for Organic Chemistry and

Biochemistry, Technical University ofMunich (Germany)

Structure prediction of the transmembrane protein Colicin ImmEl

86