University of Groningen

Structure and function of substrate-binding proteins of ABC-transportersBerntsson, Ronnie Per-Arne

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2010

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Berntsson, R. P-A. (2010). Structure and function of substrate-binding proteins of ABC-transporters. s.n.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 03-02-2021

Chapter 5

Selenomethionineincorporation in proteinsexpressed in Lactococcus lactis

Ronnie P-A Berntsson*, Nur Alia Oktaviani*, Fabrizia Fusetti, Andy-Mark WH Thunnissen, Bert Poolman and Dirk Jan Slotboom

(*) Shared first authorship

Published 2009 in Protein Science 18, 1121-7

5.1 Abstract

Lactococcus lactis is a promising host for (membrane) protein overproduction.Here, we describe a protocol for incorporation of selenomethionine intoproteins expressed in L. lactis. Incorporation efficiencies of selenomethion-

ine in the membrane protein complex OpuA (an ABC transporter) and the solubleprotein OppA, both from L. lactis, were monitored by mass spectrometry. Both pro-teins incorporated selenomethionine with high efficiencies (>90%), which greatlyextends the usefulness of the expression host L. lactis for X-ray crystallographypurposes. The crystal structure of ligand-free OppA was determined at 2.4 Aresolution by a semi-automatic approach using selenium SAD phasing.

86 Chapter 5

5.2 Introduction

Membrane protein structures are under-represented in the PDB, and account forless than 1% of the database entries. A major bottleneck in studying membraneproteins is the difficulty to produce the molecules in their native state in sufficientquantities. It is often worthwhile testing more than one expression host becauseproteins that express poorly in one host may be produced in large amounts in otherorganisms. Over the past decade the Gram-positive bacterium Lactococcus lactis hasbeen developed as a robust expression host for (membrane) proteins (Geertsma andPoolman, 2007; Geertsma, Groeneveld, et al., 2008; Morello et al., 2008; Kunji et al.,2003). Protocols for cloning, expression and quality control of membrane proteinsin L. lactis are available (Geertsma and Poolman, 2007; Geertsma, Groeneveld, etal., 2008). The organism has proven to be particularly suitable for overexpressionof complex, multisubunit proteins (such as ABC and TRAP transporters) andmammalian membrane proteins, often when E. coli failed to deliver (Kunji etal., 2005, 2003; Geertsma and Poolman, 2007; Quick and Javitch, 2007; Mulliganet al., 2009; Monne et al., 2005). To fully exploit the potential of the host forproduction of proteins to be used in crystallographic studies, we have developeda protocol for incorporation of selenomethionine (SeMet) into expressed proteins,which can be used for experimental phasing. Two proteins from L. lactis were testedfor SeMet incorporation: The oligopeptide-binding protein OppA without lipidanchor (termed OppA*) was produced in the cytoplasm of L. lactis (Lanfermeijeret al., 1999), and the ABC transporter for glycine betaine (OpuA), consisting of thetwo subunits OpuAA and OpuABC, was produced in the membrane (Heide andPoolman, 2000). Using mass spectrometry and X-ray crystallography, we show thatL. lactis can effectively incorporate SeMet in overexpressed proteins and produceamounts sufficient for crystallization and structure determination.

5.3 Results

5.3.1 Selenomethionine incorporation

L. lactis NZ9000 cells containing the expression plasmids were cultivated in chem-ically defined medium (CDM) to an OD600 of 1.5, after which the medium waschanged to CDM containing SeMet instead of methionine, and expression was

Selenomethionine incorporation in proteins expressed in L. lactis 87

induced. OpuA was purified as previously described (Mahmood et al., 2006) anddetails of OppA* purification are presented in the methods section. The extent ofSeMet incorporation in OppA* and OpuA was measured by mass spectrometry.Purified proteins with and without incorporated SeMet were digested with trypsinand peptides were analyzed by MALDI-TOF MS. Selenium has 6 different naturallyoccurring isotopes (74Se 76Se 77Se 78Se 80Se 82Se), 5 of which have an abundance ofmore than 5%, making Se-Met containing peptides readily distinguishable frompeptides containing methionine [32S (95%) and 34S (4.3%)] on the basis of their m/zvalue and the distribution of isotope masses (Fig. 5.1A). Peptides were matched tothe amino acid sequence of each protein, resulting in the identification of trypticpeptides covering 56%, 42% and 29% of the protein sequences of OppA*, OpuAAand OpuABC, respectively. For OppA* the identified peptides contained 9 out ofthe total 14 (seleno)methionines, and for the OpuA complex 18 out of 38. Theratio of peak areas from SeMet-containing and wild-type peptides was calculated toestimate the incorporation efficiency. OppA* and OpuAA/OpuABC had 93% and94% SeMet incorporated, respectively. MALDI mass spectrometry was also used todetermine the masses of the intact SeMet substituted and wild-type proteins. Themass of OppA* increased from 64923 Da to 65526 Da upon SeMet incorporation;that of OpuAA from 45634 Da to 46312 Da and the mass of OpuABC from 62949 Dato 63915 Da, indicating incorporation efficiencies of 92%, 96% and 89%, respectively(Fig. 5.1B-C).

5.3.2 Crystallization and Structure Determination

OppA* was stripped of any co-purified peptide ligands by guanidium chloridetreatment, which generates ligand-free OppA* (Lanfermeijer et al., 1999), and theprotein was crystallized by vapour diffusion as detailed in the method section.SeMet-substituted OppA* yielded crystals belonging to space group P21, with 1protein molecule in the asymmetric unit and a solvent content of 42%. Single-wavelength anomalous diffraction (SAD) data were collected at the K-absorptionedge of selenium on beamline BM16 at the ESRF, Grenoble. The data were ofsufficient quality to allow the structure to be built in a semi-automated mannerby SAD phasing using the Auto-Rickshaw server (Panjikar et al., 2005) (Fig. 5.2).Thirteen out of the 14 possible selenium sites were identified and used for phas-ing, confirming the high efficiency of SeMet incorporation determined by mass

88 Chapter 5

Figure 5.1. Selenomethionine incorporation in OppA* and OpuA. A: Mass spectra of the trypticpeptides ALANDPEILLMDEAFSALDPLIR and VGFVVPSYMNVNSIEDLTNQANK from OppA*containing SeMet (upper panel) or methionine (lower panel). B and C: Mass spectra of intact OppA*(panel B) and OpuAA/OpuABC (panel C). Continuous lines represent methionine-containing proteins,dotted lines represent proteins with SeMet incorporated. In panel B, the triply charged peaks of OppA*are shown. In panel C the triply charged peaks of OpuABC (m/z values of 20,984 and 21,306), and thedoubly charged peaks of OpuAA (m/z 22,818 and 23,157) are shown.

spectrometry. Relevant crystallographic statistics are given in Table 5.2. The refinedmodel contains the continuous polypeptide trace from Asn18 to Ala576. Residues1-17 and 577-590 could not be built due to weak or missing electron density. Similarto other bacterial substrate binding proteins OppA* consists of two domains withα/β folds connected by a hinge. OppA* is in an open conformation with thetwo domains separated from each other, leaving the binding site, which does notcontain ligand, accessible to the solvent (Fig. 5.3). The biological relevance of thestructure was discussed in Chapter 2 and 3.

Selenomethionine incorporation in proteins expressed in L. lactis 89

Figure 5.2. Experimental electron density and model of OppA*. A: Stereo view of the anomalousdifference Fourier map superimposed on a backbone trace of OppA*. Selenomethionine residues arehighlighted in a balls-and-sticks representation (orange). The Se anomalous difference Fourier mapwas calculated between 61 and 2.4 A and contoured at 5σ is shown in blue. 13 out of a total 14 SeMetwere visible in the map, SeMet at position 1 was in a disordered region of the protein. Peak heightsare as follows: MSE-178, 36.3σ; MSE-530, 35.5σ; MSE-62, 32.7σ; MSE-45, 32.6σ; MSE-555, 22.0σ; MSE-457, 21.3σ, MSE-562, 20.2σ; MSE-448, 19.8σ, MSE-182, 17.8σ; MSE-278, 17.5σ; MSE-305, 17.3σ; MSE-300,14.3σ; MSE-542, 11.7σ. No noise peaks were visible at a 5σ cutoff. B: Electron density (2Fo-Fc mapcontoured at 1.5σ) showing part of OppA*, visualizing the good quality of the experimentally derivedphases. The protein is shown in a stick representation.

90 Chapter 5

Figure 5.3. Ribbon representation of OppA* in the open unliganded conformation. The refined modelcontains a continuous polypeptide trace from Asn18 to Ala576. The two α/β domains are made upof residues 18-300 and 541-576 (domain I, green and blue), and residues 301-540 (domain II, orange).Domain I has two sub-domains: domain Ia (residues 18-82, 220-300 and 541-570, green), and domain Ib(residues 83-219, blue). The hinge-region corresponds to the segments that link domain Ia and domainII, i.e., the blue to orange transitions.

5.4 Discussion

Selenomethionine incorporation in proteins expressed in L. lactis expands the utilityof the organism as protein production host. The first expression host used forSeMet incorporation was Escherichia coli (Doublie, 1997), but in the last decadeSeMet has been incorporated into proteins produced in Baculovirus (W. Chen andBahl, 1991; Bellizzi et al., 1999), Saccharomyces cerevisiae (Bushnell et al., 2001), Pichiapastoris (Larsson et al., 2002; Xu et al., 2002) and mammalian cells (Wu et al., 1994;Lustbader et al., 1995). The SeMet incorporation efficiency of L. lactis is comparableto that of E. coli, and superior to that of the eukaryotic expression systems (Doublie,1997). L. lactis has proven to be a well-suited host for (eukaryotic) membraneprotein over-expression (Kunji et al., 2005, 2003; Geertsma and Poolman, 2007;Quick and Javitch, 2007; Mulligan et al., 2009; Monne et al., 2005). The possibilityto efficiently incorporate SeMet into those proteins makes the organism a completeprotein production host for crystallography purposes.

Selenomethionine incorporation in proteins expressed in L. lactis 91

5.5 Experimental procedures

5.5.1 Selenomethionine incorporation

For inducible production of OppA* in L. lactis NZ9000 (Ruyter et al., 1996), theoppA* gene (Lanfermeijer et al., 1999) was inserted into vector pNZcLIC, usingligation-independent cloning (Geertsma and Poolman, 2007). Production of theABC transporter OpuA in L. lactis NZ9000 from plasmid pNZOpuAhis was doneas previously described (Mahmood et al., 2006). A 10 mL preculture was grownovernight at 30C in M17 medium (Oxoid), supplemented with glucose (0.5%) andchloramphenicol (5 µg/ml). The preculture was used to inoculate 1 L of chemicallydefined medium (CDM) composed of 10 mL vitamin mix, 10 mL metal mix, 10mL base mix, 50 mL amino acid mix, 1 mL MnSO4 × H2O, 0.25 g cysteine-HCland 919 mL basic medium (see Table 5.1 for the exact composition of the CDM),and containing 5 µg/ml chloramphenicol and 1% glucose. The culture was grownin a 2 L bioreactor (Applikon Biotechnology) at 30C, pH 6.5, stirred at a speed of200rpm, until the OD600 reached a value of 1.5. The cells were spun down in sterilecentrifuge tubes for 10 min, 10000g, 30C, washed with 150 mL buffer (50mM KPipH 7.0, preheated to 30◦C), spun down again and subsequently resuspended tothe same OD600 in 1 L CDM, with SeMet (0.84 mM) instead of methionine, andcontaining 5 µg/ml chloramphenicol plus 1% glucose. The culture was put backin the bioreactor and incubated for 20 minutes before inducing the expression with0.05% (v/v) of culture supernatant of the nisin A producing strain NZ9700 (Ruyteret al., 1996). The cells were grown until the OD600 reached a value of 3 - 5 and thenharvested. Cells were resuspended to an OD600 of∼ 200 in 50 mM Tris-HCl pH 7.8,300 mM NaCl (buffer A), frozen in liquid nitrogen and stored at -80◦C.

5.5.2 Purification of OppA*

Frozen cells were thawed at room temperature, and diluted 2-fold with bufferA. Subsequently, 100 µg/mL deoxyribonuclease type I, 10 mM MgSO4 and 0.1mM phenylmethanesulfonyl fluoride was added. The cells were broken by twoconsecutive passes in a cell disruptor (Constant Systems Ltd) at 39,000 psi, 5◦C.Unbroken cells and cell debris were removed by ultracentrifugation, 267.000g, 4◦Cfor 75 min. The supernatant was collected (resulting in 50 mL lysate per literculture), and 0.5 mL Ni2+-sepharose resin (Amersham Biosciences) was added per

92 Chapter 5

Table 5.1. Composition of chemically defined mediuma

Amino acid mix, 20x solution (ad-justed to pH 7.0, filter sterilized,heated to 50◦C)

mg/L Vitamin mix, 100x solution (ad-justed to pH 7.0, filter sterilized)

mg/L

L-alanine 237.5 nicotinic acid 200L-glutamine 390 thiamine dichloride 100L-asparagine 350 riboflavin 100L-arginine 125 Ca-(D+)pantothenate 100L-lysine 437.5 K-p-aminobenzoate 1000L-isoleusine 212.5 D-biotin 1000L-methionine 125 folic acid 100L-phenylalanine 275 vitamin B12 100L-serine 337.5 orotic acid 500L-threonine 225 2-deoxythymidine 500L-trypthophane 50 inosine 500L-valine 325 DL-6,8 thioctic acid 250glycine 175 pyridoxamine dichloride 500L-histidine 150 pyridoxal-chloride 200L-leucine 475 Base Mix, 100x solution (in 0.2 M

NaOH)g/L

L-proline 675 adenine 1L-aspartate 0 uracil 1Metal mix, 100x solution (in H2O) g/L xanthine 1MgCl × 6H2O 20 guanine 1CaCl × 2H2O 5 Basic medium (set pH to 6.4 with

HCl, autoclaved)g/L

FeCl × 4H2O 0.5 Tyrosine (in boiling H2O) 0.29ZnSO × 7H2O 0.5 KH2PO4 2.5CoCl × 6H2O 0.3 K2HPO4 3CuSO × 5H2O 0.02 (NH4)3-citrate 0.6Separately Na-Acetate 1MnSO4 × H2O

Per liter CDMAmino acid mix 50 mLVitamin mix 10 mLMetal mix 10 mLBase mix 10 mLMnSO4 × H2O 1 mLBasic medium 919 mL

a The Vitamin mix, Amino acid mix and the Base mix can be stored at -20◦C. The Metal mix hasto be made fresh. The Basic medium should be autoclaved and stored in the dark.

50 mL lysate. The mixture was incubated for 1 h at 4◦C in buffer A supplementedwith 15 mM imidazole and 10% (v/v) glycerol. Subsequently, the resin was pouredinto a disposable column (BioRad) and washed with buffer A containing 50 mMimidazole, for 20 column volumes (CV). Bound peptides were removed fromOppA* by partially unfolding the protein while bound to the resin. The followingwash steps were performed (all in buffer A with 15 mM imidazol): 40 CV with2M Guanidine-HCl (GndHCl), 4 CV with 1.5 M GndHCl, 4 CV with 1 M GndHCl,4 CV with 0.5 M GndHCl and finally 8 CV with 0 M GndHCl. The protein was

Selenomethionine incorporation in proteins expressed in L. lactis 93

eluted with 20 mM Na-MES, pH 6.0, 300 mM NaCl, 500 mM imidazole, pH 6.0, 2CV. Purified SeMet-OppA* was concentrated to 0.5 ml in spin concentrators with a30 kDa cut-off (Vivaspin with PES membrane from Sartorius), and further purifiedon a Superdex 200 10/300 GL size exclusion column (Amersham Biosciences) in20 mM Na-MES, pH 6.0, 150 mM NaCl. Fractions containing OppA* were pooled,concentrated 10-fold, and diluted such that the final buffer composition was 10mMNa-MES, pH 6.0, 10mM NaCl, and finally concentrated again to 11 mg/mL ofprotein. The final yield was about 1 mg SeMet-OppA* per liter culture, which iscomparable to that of OppA* produced in L. lactis grown in normal CDM.

5.5.3 Purification of OpuA

OpuA was purified as described (Mahmood et al., 2006). For SeMet-OpuA, thefinal yield was 50 µg of pure protein from 10 mg total membrane protein, which issimilar to the yields of OpuA produced in L. lactis grown in normal CDM.

5.5.4 Mass spectrometry analysis of SeMet substituted proteins

10 µg of purified protein was subjected to digestion for 18 h with 125 ng porcinetrypsin (MS-grade, Promega, Leiden, The Netherlands) in 200 mM triethylammo-nium bicarbonate, pH 8.0, 10% acetonitrile. The peptide digests were mixed 1:1v/v with a solution of α-cyano-4-hydroxycinnamic acid matrix (5 mg/ml in 50%ACN and 0.1% TFA, LaserBio Labs), spotted directly onto stainless steel MALDItargets and analyzed using either a 4700 or 4800 Proteomics Analyzer MALDI-TOF/TOF mass spectrometer (Applied Biosystems). For the digested peptidesamples, the MALDI-TOF/TOF was operated in reflector positive ionization modein the m/z range 700-4000. MS peak-lists were extracted by GPS Explorer Software,version 3.5 (Applied Biosystems), applying default parameters. A list of theoreticaltryptic peptides, obtained with the program GPMAW, version 7.10sr2 (Lighthouse),allowing for one missed cleavage, was used to interpret the MS spectra, estimatingan average increase in m/z of 47 Da for each SeMet). For mass spectrometryanalysis of the intact OppA*, the protein was mixed 1:1 v/v with a solution ofα-cyano-4-hydroxycinnamic acid matrix (5 mg/ml in 50% ACN and 0.1% TFA,LaserBio Labs), spotted onto stainless steel MALDI targets and analyzed usingeither a 4700 or 4800 Proteomics Analyzer MALDI-TOF/TOF mass spectrometer(Applied Biosystems) in the linear mode. For mass spectrometry analysis of the

94 Chapter 5

intact OpuA complex the purified protein was spotted using the ultra thin layermethod to increase the signal to noise levels (Cadene and Chait, 2000).

5.5.5 Crystallization and structure determination

Crystals of SeMet-substituted OppA* were grown by vapor diffusion in hangingdrops. The drops consisted of 1 µL protein (10 mg/mL OppA*, 9 mM Na-MES,pH 6.0, 9 mM NaCl and 1 mM peptide (YGGFL)) and 1 µL reservoir solution(0.2 M NaCl, 0.1 M Na-Hepes, pH 7.0, 20% PEG 6000). Diffracting crystals wereobtained after 24 hours of incubation at room temperature. Crystals were soakedin mother liquor supplemented with 42% PEG 6000 for 30 s and then flash cooled inliquid nitrogen. Single-wavelength dispersion (SAD) data to 2.38 A resolution werecollected at the selenium K-absorption edge (wavelength: 0.979 A) on beamlineBM16 at the ESRF, Grenoble, using a single crystal (0.15 mm x 0.15 mm x 0.2 mm).Data processing and reduction were carried out using HKL2000 (Otwinowski andMinor, 1997).

Data was submitted to Auto-Rickshaw, and the structure was solved using theSAS protocol of Auto-Rickshaw: the EMBL-Hamburg automated crystal structuredetermination platform (Panjikar et al., 2005). The input diffraction data wereprepared and converted for use in Auto-Rickshaw using programs of the CCP4suite (Collaborative Computational Project Number 4, 1994). FA values werecalculated using the program SHELXC. Based on an initial analysis of the data, themaximum resolution for substructure determination and initial phase calculationwas set to 2.38 A. 13 heavy atoms out of the maximum number of 14 were foundusing the program SHELXD (Schneider and Sheldrick, 2002). The correct handfor the substructure was determined using the programs ABS (Hao, 2004) andSHELXE (Schneider and Sheldrick, 2002). The occupancy of all substructure atomswas refined and initial phases were calculated using the program MLPHARE(Collaborative Computational Project Number 4, 1994). Density modification andphase extension were performed using the program DM (Cowtan, 1994). 39.8% ofthe model was built using the program ARP/wARP (Perrakis et al., 1999). Auto-Rickshaw returned the partially built model, which was subsequently resubmittedto Auto-Rickshaw using the MRSAD protocol, combining SAD data with molecularreplacement of the partially built model. The calculated electron density wasof sufficient quality to let ARP/wARP build 89% of the residues, which Auto-Rickshaw returned. A few cycles of refinement using Refmac5 (Murshudov et al.,

Selenomethionine incorporation in proteins expressed in L. lactis 95

1997), interspersed with manual model building using Coot (Emsley and Cowtan,2004), were necessary to complete the model. Water molecules were placed auto-matically in Fo-Fc Fourier difference maps at a 3 σ-cutoff level, and validated toensure correct coordination geometries using Coot. Relevant statistics of the datacollection, phasing and model refinement are given in 5.2.

Table 5.2. Data collection and refinement statistics for OppA-SeMet

Data collectionSpace group P21Cell dimensionsxxxa, b, c (A ) 39.6, 122.3, 59.4xxxα β γ (◦) 90, 104, 90Wavelength (A ) 0.979Resolution range (A) 61-2.38Rsym 0.079 (0.178)I/σ (I) 6.1 (4.0)Completeness (%) 99.6 (100.0)Redundancy 6.0SAD phasingResolution range (A) 19.8-2.38Cullis R factor 0.67FOMacentric 0.272RefinementResolution range (A) 61-2.38Number of reflections 20965Rwork/R f ree 0.203/0.257No. atomsProtein 4355Water 290Average B-factors (A 2)Protein 30.048Water 37.7R.m.s deviationsBond lengths (A ) 0.019Bond angles (◦) 1.362

a The number in parentheses corresponds to the highest resolution shell

96 Chapter 5

Accession codes

The coordinates have been deposited in the Protein Data Bank with accession code3FTO.

Acknowledgments

The authors thank Dr. Hjalmar Permentier for help with mass spectrometryanalysis, and the ESRF for providing excellent beam line facilities. This workwas supported by grants from Marie Curie Early Stage Training (EST, to RB), TheNetherlands Organisation for Scientific research (NWO, Vidi grant to DJS), EU(EMEP & EDICT programs) and the Netherlands Proteomics Centre (NPC).