+ All Categories
Home > Documents > Structure and Function of the Oxidoreductase DsbA1 from Neisseria meningitidis

Structure and Function of the Oxidoreductase DsbA1 from Neisseria meningitidis

Date post: 11-Nov-2023
Category:
Upload: monash
View: 0 times
Download: 0 times
Share this document with a friend
10
Crystal Structure of LipL32, the Most Abundant Surface Protein of Pathogenic Leptospira spp. Julian P. Vivian 1,2 , Travis Beddoe 1,2 , Adrian D. McAlister 1 , Matthew C.J. Wilce 1 , Leyla Zaker-Tabrizi 1,2 , Sally Troy 1 , Emma Byres 1,2 , David E. Hoke 3 , Paul A. Cullen 3 , Miranda Lo 3 , Gerald L. Murray 3 , Ben Adler 2,3 and Jamie Rossjohn 1,2 1 The Protein Crystallography Unit, Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia 2 ARC Centre of Excellence in Structural and Functional Microbial Genomics, Monash University, Clayton, Victoria 3800, Australia 3 Department of Microbiology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia Received 19 December 2008; received in revised form 10 February 2009; accepted 12 February 2009 Available online 21 February 2009 Spirochetes of the genus Leptospira cause leptospirosis in humans and animals worldwide. Proteins exposed on the bacterial cell surface are implicated in the pathogenesis of leptospirosis. However, the biological role of the majority of these proteins is unknown; this is principally due to the lack of genetic systems for investigating Leptospira and the absence of any structural information on leptospiral antigens. To address this, we have determined the 2.0-Å-resolution structure of the lipoprotein LipL32, the most abundant outer-membrane and surface protein present exclusively in pathogenic Leptospira species. The extracellular domain of LipL32 revealed a compact, globular, jelly-rollfold from which projected an unusual extended β-hairpin that served as a principal mediator of the observed crystallographic dimer. Two acid-rich patches were also identified as potential binding sites for positively charged ligands, such as laminin, to which LipL32 has a propensity to bind. Although LipL32 shared no significant sequence identity to any known protein, it possessed structural homology to the adhesins that bind components of the extracellular matrix, suggesting that LipL32 functions in an analogous manner. Moreover, the structure provides a framework for understanding the immunological role of this major surface lipoprotein. © 2009 Elsevier Ltd. All rights reserved. Edited by I. Wilson Keywords: Leptospira; LipL32; jelly-roll fold; outer-membrane protein Introduction Leptospirosis is caused by infection with the spirochete bacterium, Leptospira. The organism has a broad host range and is thought to be the most widespread zoonotic agent worldwide. The main- tenance hosts involved in transmission to humans are mainly rodents, dogs, or cattle that harbor pathogenic Leptospira spp. in their proximal renal tubules, resulting in long-term urinary shedding of bacteria into the environment. Human infection occurs when leptospires enter via mucosal surfaces or broken skin following contact with infected animals or contami- nated soil or water. 1 Once in the body, leptospires are found in multiple organs and tissues. The disease ranges from mild symptoms to more serious compli- cations including jaundice, pulmonary hemorrhage, and renal and hepatic failure, which may prove fatal. 24 Pathogenic Leptospira spp., previously classi- fied into the single species Leptospira interrogans, are now differentiated into at least 12 species, with L. interrogans and Leptospira borgpetersenii being the main causes of human disease worldwide. More than half *Corresponding authors. The Protein Crystallography Unit, Department of Biochemistry and Molecular Biology, School of Biomedical Sciences, Monash University, Clayton, Victoria 3800, Australia. E-mail addresses: travis. [email protected]; jamie.rossjohn@med. monash.edu.au. J.P.V. and T.B. contributed equally to this work. Abbreviations used: ECM, extracellular matrixc; Se- Met, selenomethionine; SC, shape complementarity; PDB, Protein Data Bank. doi:10.1016/j.jmb.2009.02.038 J. Mol. Biol. (2009) 387, 12291238 Available online at www.sciencedirect.com 0022-2836/$ - see front matter © 2009 Elsevier Ltd. All rights reserved.
Transcript

doi:10.1016/j.jmb.2009.02.038 J. Mol. Biol. (2009) 387, 1229–1238

Available online at www.sciencedirect.com

Crystal Structure of LipL32, the Most Abundant SurfaceProtein of Pathogenic Leptospira spp.

Julian P. Vivian1,2†, Travis Beddoe1,2⁎†, Adrian D. McAlister1,Matthew C.J. Wilce1, Leyla Zaker-Tabrizi1,2, Sally Troy1, Emma Byres1,2,David E. Hoke3, Paul A. Cullen3, Miranda Lo3, Gerald L. Murray3,Ben Adler2,3 and Jamie Rossjohn1,2⁎

1The Protein CrystallographyUnit, Department ofBiochemistry and MolecularBiology, School of BiomedicalSciences, Monash University,Clayton, Victoria 3800,Australia2ARC Centre of Excellence inStructural and FunctionalMicrobial Genomics, MonashUniversity, Clayton, Victoria3800, Australia3Department of Microbiology,School of Biomedical Sciences,Monash University, Clayton,Victoria 3800, Australia

Received 19 December 2008;received in revised form10 February 2009;accepted 12 February 2009Available online21 February 2009

*Corresponding authors. The ProteiUnit, Department of Biochemistry anSchool of Biomedical Sciences, MonClayton, Victoria 3800, Australia. [email protected]; jamiemonash.edu.au.† J.P.V. and T.B. contributed equalAbbreviations used: ECM, extrace

Met, selenomethionine; SC, shape coProtein Data Bank.

0022-2836/$ - see front matter © 2009 E

Spirochetes of the genus Leptospira cause leptospirosis in humans andanimals worldwide. Proteins exposed on the bacterial cell surface areimplicated in the pathogenesis of leptospirosis. However, the biological roleof the majority of these proteins is unknown; this is principally due to thelack of genetic systems for investigating Leptospira and the absence of anystructural information on leptospiral antigens. To address this, we havedetermined the 2.0-Å-resolution structure of the lipoprotein LipL32, themost abundant outer-membrane and surface protein present exclusively inpathogenic Leptospira species. The extracellular domain of LipL32 revealed acompact, globular, “jelly-roll” fold from which projected an unusualextended β-hairpin that served as a principal mediator of the observedcrystallographic dimer. Two acid-rich patches were also identified aspotential binding sites for positively charged ligands, such as laminin, towhich LipL32 has a propensity to bind. Although LipL32 shared nosignificant sequence identity to any known protein, it possessed structuralhomology to the adhesins that bind components of the extracellular matrix,suggesting that LipL32 functions in an analogous manner. Moreover, thestructure provides a framework for understanding the immunological roleof this major surface lipoprotein.

© 2009 Elsevier Ltd. All rights reserved.

Edited by I. Wilson

Keywords: Leptospira; LipL32; jelly-roll fold; outer-membrane protein

Introduction

Leptospirosis is caused by infection with thespirochete bacterium, Leptospira. The organism has abroad host range and is thought to be the mostwidespread zoonotic agent worldwide. The main-

n Crystallographyd Molecular Biology,ash University,ail addresses: travis.

.rossjohn@med.

ly to this work.llular matrixc; Se-mplementarity; PDB,

lsevier Ltd. All rights reserve

tenance hosts involved in transmission to humans aremainly rodents, dogs, or cattle that harbor pathogenicLeptospira spp. in their proximal renal tubules,resulting in long-term urinary shedding of bacteriainto the environment. Human infection occurs whenleptospires enter via mucosal surfaces or broken skinfollowing contact with infected animals or contami-nated soil or water.1 Once in the body, leptospires arefound in multiple organs and tissues. The diseaseranges from mild symptoms to more serious compli-cations including jaundice, pulmonary hemorrhage,and renal and hepatic failure, which may provefatal.2–4 Pathogenic Leptospira spp., previously classi-fied into the single species Leptospira interrogans, arenow differentiated into at least 12 species, with L.interrogans and Leptospira borgpetersenii being themaincauses of human disease worldwide. More than half

d.

Table 1. Data collection and refinement statistics

Native KAuBr4

Data collectionSpace group P3221 P3221Cell dimensions (Å) a=b=125.9,

c=95.9a=b=124.7,

c=95.4Resolution (Å) 50–2.0

(2.11–2.0)50–2.8

(2.95–2.80)Total observations 1,270,906 377,219Unique observations 59,453 20,492Rsym

a 0.078 (0.724) 0.089 (0.822)Rpim

b 0.023 (0.273) 0.027 (0.311)I/σI 41.9 (2.4) 30.7 (2.1)Completeness (%) 100 (100) 95.7 (77.7)Redundancy 21.4 (20.3) 18.4 (9.9)

RefinementResolution 50–2.0Reflections in working set 54,787Rcryst

c 0.185Rfree

c 0.218Total protein atoms 3520Total waters 447Bond lengths (Å) 0.011Bond angles (°) 1.314Ramachandran analysis (%)Most favored 97.3Additionally allowed 2.7Average B-factor (Å2)Main chain 47.5Side chain 48.8Waters 53.5

Values in parentheses represent those of the highest-resolutionshell.

a Rsym=∑(∣Ii− Imean∣)/∑(Ii).b Rpim=√(1/(n−1)) ·∑(∣Ii− Imean∣)/∑(Ii).c Rcryst=∑hkl∣∣Fo∣−∣Fc∣∣/∑hkl∣Fo∣ for all data excluding the 5%

that comprised the Rfree used for cross-validation.

1230 Crystal Structure of a Leptospiral Antigen

of the 250 serovars of Leptospira are contained in thesetwo species. The global importance of this disease ishighlighted by estimates that leptospirosis causessevere complications for 1million people annually.5–7

Immunity to Leptospira in humans is mediatedprimarily by antibodies directed towards lipo-polysaccharide.8 This immunity is serovar specificand has led to a search for conserved cell surfaceproteins thatmight elicit cross-serovar immunity.Onesuch candidate is LipL32, the most abundant outer-membrane protein9 and the most abundant surface-exposed protein.10 LipL32 is a lipoprotein, with theprotein component remaining external yet anchoredto the outer membrane through fatty acids covalentlyattached to the amino-terminal cysteine residue.There are several lines of evidence that suggest thatLipL32 may be important in pathogenesis. The lipL32gene is present only in pathogenic species, where itdisplays a high degree of conservation.11 Cellularassays indicate that LipL32 induces an inflammatoryresponse in cultured renal cells12,13 and a promyelo-cytic cell line expressing Cd14. In vivo studies showthat LipL32 is a major target of the antibody responsein animals and humans and is expressed in thekidneys of infected animals.11,14,15 The function ofLipL32 may be affected by posttranslational eventssuch as proteolytic cleavage, with studies suggestingthat the C-terminus of a subset of LipL32 proteins isshed while the N-terminus remains attached to thecell surface.9 LipL32 might be involved in thepathogenesis of leptospirosis through extracellularmatrix (ECM) binding activity identified in two recentreports.15,16 Protein truncation studies were used tomap this activity,with one report determining that the72C-terminal amino acidswere sufficient for binding,while the other showed that the 85 C-terminal aminoacids were sufficient and also inhibited binding to theintact protein. A LipL32 ortholog in the ubiquitousmarine bacterium Pseudoalteromonas tunicata alsoshowed C-terminal ECM binding activity and cross-reacted with anti-LipL32 antiserum.16 These studiessuggest that C-terminal elements within orthologousLipL32 proteins function as ECM-binding domains indiverse bacteria.Lastly, LipL32 has been shown in some studies to

confer partial protection against infection in animalmodels.17–19 Interestingly, anti-LipL32 antibodiesproduced in response to naturally acquired infectiondo not appear to be protective. A recent studyfocused on LipL32 domains that are recognizedduring human infection and showed that theimmunoglobulin M response was limited to a C-terminal domain during acute and convalescentphases.15 The immunoglobulin G response changedduring infection, with most patient sera recognizinga central region of the protein during convalescenceand 5 out of 12 patients responding to the C-terminusin the acute and convalescent phases. These findingssuggest that differential immune recognition ofLipL32 domains during infection and convalescencemay be a factor in immunity to Leptospira.The lack of sequence similarity between LipL32 and

any other functionally characterized protein has

hindered progress in determining structural domainsthat relate to function. Therefore, structural determi-nation of LipL32 is important for future work inunderstanding the potential role of LipL32 in immu-nity to infection, characterization of the pro-inflam-matory response, and interaction with ECM proteins.In this article, we describe the high-resolution crystalstructure of LipL32, thereby providing fundamentalinsight into its role in infection and immunity.

Results

Expression, purification, and crystallization ofLipL32

To gain insight into the structure–function rela-tionship of leptospiral antigens, we expressed theentire mature form of LipL32 in Escherichia coli andpurified it to homogeneity (see Materials andMethods). While this domain produced extremelylarge crystals of defined morphology, they diffractedvery poorly (approximately to 4 Å resolution) andthe diffraction pattern exhibited diffuse thermalscattering, indicating potential disorder within thecrystal lattice. Accordingly, we undertook a limitedproteolysis approach in an attempt to “shave off”any flexible regions of LipL32. Of the proteases that

1231Crystal Structure of a Leptospiral Antigen

we investigated, cleavage with V8 protease yieldedoptimal results and provided a fragment that wasmore amenable for structural studies. The V8protease cleaved LipL32 within the N-terminaltobacco etch virus protease recognition site and 11residues from the C-terminus. This yielded a proteincomprising 11 vector-encoded residues at the N-terminus and residues 2–242 of the mature LipL32,as confirmed by N-terminal sequencing and massspectrometry (data not shown). The V8-cleavedLipL32 crystallized readily and diffracted to 2.0 Åresolution and belongs to space group P3221, withunit cell dimensions a=b=125.9 Å and c=95.9 Å,which is consistent with two LipL32 molecules perasymmetric unit (see Table 1). Given that sequenceanalysis suggested that there were no structuralhomologues of LipL32, we next sought to label theLipL32 with selenomethionine (Se-Met). However,the Se-Met-labeled LipL32 expressed extremelypoorly (0.3 mg/l), and the resulting crystals werevery small, thereby precluding structure determina-tion via Se-Met multiwavelength anomalous diffrac-tion phasing. Accordingly, we determined thestructure of LipL32 via the single-isomorphousreplacement method (see Materials and Methods),which resulted in a readily interpretable electrondensity map. The structure of LipL32 was subse-quently refined to 2.0 Å resolution to an R-factor andRfree of 18.5% and 21.8%, respectively. The modelcontains residues 6–242 of the mature protein, withresidues 2–5 and 140–146 being uninterpretable inthe electron density map and 447 water molecules.LipL32 crystallized as two protomers within the

Fig. 1. Crystal structure of LipL32. (a) Cartoon representatfrom the N-terminus to the C-terminus. The secondary-strucomprises a two-strand N-terminal dimerization domain, witharranged in aβ-hairpin. The core of the protein is a jelly roll comβ-6 (residues 73–79), β-8 (residues 115–120), β-9 (residues 151–β-12 (residues 205–214). LipL32 has three short parallel β-stran(residues 52–54), β-7 (residues 92–94), and β-13 (residues 225short helices, with helices α-1 (residues 96–100) and α-2 (residuand α-5 (residues 182–184). Distal to the dimerization interfacethe C-terminal helix α-6 (residues 230–241). (b) Topology diag

asymmetric unit, which were virtually identicalwith each other (rmsd, 0.47 Å over all Cα positions),and unless explicitly stated, structural analysis willbe confined to one monomer.

Overall structure of LipL32

LipL32 adopted a compact, globular fold of overalldimensions of ≈29 Å×50 Å×56 Å, is rich in β-sheet(38%), and possesses a few peripheral α-helices (17%),and the “top” and “bases”of LipL32possess a numberof meandering loops that serve to connect the majorsecondary structural elements (Fig. 1a). While theseloops are generally long, the mobility of the loops isrestricted via interloop contacts, and as such, themajority of the molecule displays limited mobility,with an average B-factor of≈48Å2. The core of LipL32comprised a mixed β-sheet that packed against anantiparallel β-sheet and adopted a “jelly-roll” fold ofeight β-strands (β-3, β-5, β-6, β-8, β-9, β-10, β-11, andβ-12) (Fig. 1b). The interior of this β-sandwichcontained a large number of hydrophobic interactions,characterized by small aliphatic residues towards thetop of the molecule and a cluster of aromatic residuesat the base of the core. Prior to the core of themoleculeis the N-terminal region (which represents the site ofattachment to the bacterial outer membrane) thatcontains a β-hairpin that runs approximately ortho-gonal to and extends approximately 35Å from the β-sheet of the jelly roll. On the other face of LipL32 aretwo C-terminal helical regions that interdigitate in anantiparallel manner. These two helices (residues 126–134 and 230–241) pack against the core β-sheet

ion of the LipL32 structure. The structure is color rampedcture elements of the protein are labeled. The structurethe β-1 (residues 11–20) and β-2 (residues 27–33) strandsposed of strandsβ-3 (residues 36–44),β-5 (residues 58–67),154), β-10 (residues 162–166), β-11 (residues 189–196), andds that form extensions to the core jelly roll including β-4–227). Flanking the poles of the molecule are two sets ofes 103–107) located opposite helices α-4 (residues 174–177)is a helical domain comprising α-3 (residues 126–134) andram of the LipL32 structure colored similarly to (a).

Fig. 2. Representation of the electrostatic potential on the surface of LipL32. Twoviews of the LipL32monomerwith thecharge distribution calculated by theAdaptive Poisson–Boltzmann Solver22 mapped onto the surface of the protein. For thepurposes of modeling the charge distribution, the aspartate-rich loop (residues 140–146, sequence: KLDDDDD) that wasnot ordered in the crystal structure was modeled. The regions of electronegative and electropositive charge are labeled.

Table 2. Interactions at the dimeric interface of LipL32

Chain A Chain B Interaction

Thr30 Thr30 H-bondThr82 Tyr159 H-bondTyr159 Thr82 H-bondAsn160 Tyr159 H-bondGly6 Tyr159 NonbondedLeu7 Tyr159 NonbondedLys11 Val28 NonbondedLeu16 Asp19, Val28 NonbondedSer17 Asp19 NonbondedAsp19 Leu16, Ser17 NonbondedIle21 Leu32 NonbondedPro22 Leu16 NonbondedVal28 Lys11, Leu16, Pro33 NonbondedThr30 Asp19, Thr30 NonbondedIle79 Tyr159 NonbondedPro81 Phe112 NonbondedThr82 Phe112, Tyr159 NonbondedGly83 Phe112 NonbondedGlu104 Pro200 NonbondedSer107 Phe112, Pro200 NonbondedMet108 Pro200 NonbondedPro109 Pro109, His110, Trp111 NonbondedHis110 Pro109 NonbondedTrp111 Pro109 NonbondedPhe112 Pro81, Thr82, Gly83, Ser107 Nonbonded

1232 Crystal Structure of a Leptospiral Antigen

(predominantly via hydrophobic interactions) that iscentered on Trp66 packing against Pro126, Ile129, andAla240. Above and below the Trp residue are polarinteractions that further serve to tether the helicalregion to the core of the molecule. Surface-exposedelectrostatic or hydrophobic patches can often beindicative of regions that mediate protein–proteinand/or protein–ligand interactions.20,21 Notably,LipL32 possessed a cluster of surface-exposed aro-matic residues towards the base of the molecule,which included Phe112, His156, Tyr159, and Tyr198.Additionally, LipL32 possessed two patches ofmarked electronegative charge, with electronegativepatch 1 dominated by residues Glu231, Glu232,Glu241, and Glu242, while patch 2 is dominated byGlu119 (Asp142, Asp143, Asp144, Asp145, andAsp146 modeled for electrostatic calculations),Asp147, and Asp148. Additionally, LipL32 possesseda region of positive charge that included residuesLys58, Lys59, Lys199, and Lys204 (Fig. 2). Accord-ingly, LipL32 possessed a compact globular fold witha number of surface-exposed features that indicated apotential capacity to bind other components orinteract with itself.

LipL32 as a dimer

Although LipL32 was purified as a monomer bygel filtration, the two monomers within the asym-metric unit were observed to form a crystallographicdimer, with the monomers related by an approx-imate 2-fold rotation. This higher-order assembly isconsistent with the propensity for LipL32 to oligo-merize on the cell surface (data not shown).Although crystallographic dimers can often be anartifact of crystallization, the observed contactswithin the crystalline lattice were suggestive of abiologically relevant LipL32 dimer, as was alsosuggested by the PISA web server‡.23 The interface

‡http://www.ebi.ac.uk/msd-srv/prot_int/pistart.html

was extensive (circa 1250Å2), exhibited high shapecomplementarity (SC) (SC index=0.66), and wascharacterized by a large number of hydrophobicinteractions and only four hydrogen bonds (Table 2).The dimer is formed such that the N-terminalregions are close to the dyad axis, while the C-terminal helices are located at the external face of thedimer. There are two main sites of interaction at thedimer interface, mediated by the unusual β-hairpinand the base of the molecule (Fig. 3). Firstly, theβ-hairpin that was observed to project from thecore of the molecule wraps around its counterpartin the other protomer. While the top of the respective

Tyr159 Gly6, Leu7, Pro8, Ile79, Thr82, Asn160 NonbondedAsn160 Tyr159, Asn160 NonbondedPro200 Glu104, Met108 Nonbonded

Fig. 3. Dimerization of LipL32. (a) and (b) are cartoon representations of the LipL32 dimer. (a) View of the LipL32dimer looking perpendicularly to the dimer interaction surface. The monomers of LipL32 are colored green and brown.(b) Another view of the LipL32 dimer looking perpendicularly to the dimer interaction surface except that it is rotated 90°about the x-axis relative to (a). (c) View of the dimerization interface of LipL32. The figure is a surface representation of theLipL32 monomer with some of the important residues involved in dimer–dimer contacts labeled. The residues formingnonbonded contacts are colored green, and the atoms involved in forming hydrogen bonds are colored blue.

1233Crystal Structure of a Leptospiral Antigen

β-hairpins are not involved in mediating dimercontacts, the residues proximal and distal to the tipare involved in a number of van der Waalsinteractions, centered on Thr27. While Thr27 inter-acted with its equivalent residue, Ile21 packedagainst Leu32; Val28 packed against Leu16 andPro33; Leu16 interacted with Asp19, the latter ofwhich also forms van der Waals interactions withSer17. Secondly, the base of the molecule, which isrich in aromatic residues, forms a cluster ofhydrophobic contacts at the interface. Tyr159wedged within a bulged region of β-strand 8,packing against Ile79, while the TyrOH H-bonds tothe main chain of Thr82. The main chain of Tyr159H-bonds to Asn160, the latter of which also formsvan der Waals interactions with its counterpart. Thehydrophobic cluster is extended with Phe112 inter-acting with Pro81, the main chain of Thr82, Gly83,and Ser107.

Taken together, the features at the dimer interface,namely, an extensive hydrophobic interface exhibit-ing high SC, suggest that the crystallographicLipL32 dimer may represent a biologically relevantLipL32 dimer.

Structural comparisons

While LipL32 possessed no significant sequencesimilarity to any protein in the Protein Data Bank(PDB), we undertook automated structural compar-isons (DALI server) against the PDB database, as,often, structural conservation is stronger thansequence conservation. The DALI-based searchrevealed that LipL32 exhibited significant structuralhomology to proteins that also possess a jelly-rollfold despite the low level of sequence identity (circa8–13%). Among the significant hits, LipL32 showedmatches to a domain from calpain (Z score, 8.3;

Fig. 4. Comparison of LipL32 with the structural homologue collagenase collagen-binding domain 3B fromClostridium histolyticum. Structural superposition of LipL32 with the collagen-binding domain 3B was performed withSSM.26 In (a) and (b), LipL32 is represented in cartoon format in green, and the collagen-binding domain 3B is representedin cartoon format in light pink. The two orthogonal views (a and b) show the structural similarity of the core β-sheets ofthe proteins. (c) Surface representation of collagen-binding domain 3B in the orientation depicted in (b). The residuescomprising the hydrophobic surface shown to be important for collagen binding are highlighted in red. (d) Surfacerepresentation of LipL32 in the orientation depicted in (b). Highlighted in orange are residues that comprise the putativecollagen binding site of LipL32. These residues form a hydrophobic surface positioned similarly to that of the structuralhomologue collagen-binding domain 3B.

1234 Crystal Structure of a Leptospiral Antigen

rmsd, 3.5 Å over 124 Cα atoms) and collagenase (Zscore, 6.0; rmsd, 2.5 Å over 87 Cα atoms). Interest-ingly, the collagenase domain binds the ECMcomponent collagen,24 and the calpain domain isthought to localize calpain to the membrane.25 TheECM binding site of collagenase is characterized byhydrophophic interactions, with mutagenesis stu-dies identifying a cluster of tyrosine residues, with aleucine, an arginine, and a threonine residuecontributing to a hydrophobic surface.24 Structuralsuperposition of LipL32 and the collagenase revealsa similarly located hydrophobic surface on LipL32,comprising residues Tyr62, Tyr151, Tyr198, Trp115,Leu53, Val54, and Arg117, as shown in Fig. 4.Accordingly, LipL32 possesses homology to struc-tures that bind to ECM components. Recently, asequence homologue of LipL32 has been reportedfrom P. tunicata (PTD2-05920).16 This homologue,which shares 44% sequence identity with LeptospiraLipL32, is from a marine bacterium that inhabits aprimitive chordate (Ciona intestinalis) that has ECMcomponents.27 The high level of sequence identityacross the entire extracellular domain is indicative ofP. tunicata PTD2-05920 possessing a similar fold, andmoreover, as the electrostatic patches, including the

string of aspartate residues between 142 and 148 andthe hydrophobic surface-exposed patches includingresidues Phe112, His156, Tyr159, and Tyr198, areconserved, it is likely to have similar bindingproperties to itself and other components. It hasrecently been shown that PTD2-05920 can bind ECMproteins.16

Discussion

The array of cell-surface adhesion moleculesfound on pathogenic species of Leptospira (includingLigA, LigB, and other lipoproteins) is of funda-mental importance to our understanding and pre-ventive treatment of leptospirosis. These moleculesare key to the invasion and adhesion of bacteria tothe host tissue, a process that is central to thepathogenesis of the organism. Furthermore, the cell-surface localization of these adhesins makes themideal targets for vaccine development. Of theseadhesins, LipL32 stands out as a prime candidate forvaccine development as it is the most abundant cell-surface protein in pathogenic Leptospira and is themajor target of the immune system during human

1235Crystal Structure of a Leptospiral Antigen

infection. Despite the important role of LipL32, ourcurrent understanding of its functional role wasunclear. To address this, we have determined thecrystal structure of LipL32 to 2 Å resolution, the firstcrystal structure of a leptospiral adhesin.It has been shown recently that LipL32 has the

capacity to interact in vitro with multiple ECMbinding partners including plasma fibronectin,collagen IV, and laminin.15,16 Fibronectin andlaminin have defined cell- or bacteria-bindingdomains where collagen provides multiple sites ofinteractions along its triple-helical structures.28

Electrostatic analysis of the LipL32 dimer hasidentified two noncontiguous electronegativepatches arranged in tandem. It is tempting tospeculate that these patches have a role in bindingECM components with sites of known complimen-tary charge, especially laminin. Similarly, a hydro-phobic surface identified by comparison with thestructural homologue collagen-binding protein 3Brepresents a putative site for collagen binding. Otherstudies have mapped ECM interaction sites to the C-terminal region of LipL32 comprising residues 166–253. Mapping this C-terminal fragment onto the

Fig. 5. Antibody and ECM-binding domains of LipL32. (a ain binding assays with ECM proteins. The LipL32 structure is swere subcloned to produce a C-terminal fragment for bindAnalysis of this C-terminal fragment, with the benefit of the cryfragment would not fold as it does in the full-length protein. This likely to occur via linear segments. (c) Epitope 1. The LipL32residues 132–158 shown in pink.29 (d) Epitope 2. The LipL32 sresidues 162–186 shown in pink.29

crystal structure of LipL32 reveals that it comprisesthe large loop between β-10 and β-11 (includinghelices 4 and 5) and the large loop region between β-strands 12 and 13. It also contains β-strands 11, 12,and 13 and helix 6 (Fig. 5a and b). As this C-terminaldomain would be expected to have limited tertiaryand secondary structure in isolation, the proposedsite of interaction would, therefore, be a linearstretch of amino acids. There are three surface-exposed regions spanning residues 168–185, 214–222, and 230–242 that could mediate the interactionwith ECM proteins. Linear peptides have previouslybeen shown to inhibit binding to laminin by theTp0751 adhesin from the spirochete Treponemapallidum, the causative agent of syphilis.30 Theobserved affinity of LipL32 binding to ECM proteinsis relatively weak with a Kd of ≈7–10 μM. Therefore,the existence of the observed dimer may be a meansof increasing the avidity for the ECM proteins,through an increase in the local concentration ofbinding sites.16

The multiple ECM binding by LipL32 has alsobeen shown in other leptospiral adhesins such asLigA, LigB, and the family of Len proteins.31,32 This

nd b) Two views of the C-terminal fragment of LipL32 usedhown in cartoon representation in green. The residues thating assays (residues 166–253)15,16 are shown in orange.stal structure of LipL32, suggests that, in all likelihood, thiserefore, binding of this C-terminal section to ECMproteinsstructure is shown in cartoon representation in green withtructure is shown in cartoon representation in green with

1236 Crystal Structure of a Leptospiral Antigen

redundancy in ECM adhesion would convey aselective advantage to Leptospira during the initialinfection stage as 95% of human patients withleptospirosis produce antibodies against LipL32.14

Many bacterial adhesins can be effective vaccinecandidates, such as Hap adhesin from Haemophilusinfluenzae and collagen adhesin from Staphylococcusaureus.33,34 In addition, peptide inhibitors of severalbacterial pathogens such as Streptococcus mutans,Porphyromonas gingivalis, and Pseudomonas aerugi-nosa have been used to prevent attachment of thesebacteria to the corresponding ligand.35–37 Variousstrategies have been used in these studies to inhibitadhesion. These include vaccination with recombi-nant protein, therapeutic administration of selectedpeptides, and passive administration of antibodiesraised against peptide sequences that mediateadhesion.38 Recently, human antibody epitopeshave been characterized for LipL32.29 They consistof residues 132–158 and 162–186 (Fig. 5c and d). Thepeptide epitope 162–186 corresponds to one of theproposed interaction loops of LipL32. In addition,this peptide also lies within the C-terminal domain,which is the most immunogenic region of LipL32.15

The crystal structure of LipL32 will facilitatefuture studies that aim to elucidate the interactionof Leptospira with host tissue components and itsrole in the pathogenesis of leptospirosis.

Materials and Methods

Protein expression and purification

DNA encoding residues 2–253 of mature LipL32 wasamplified via PCR from L. interrogans serovar Lai strainL391 genomic DNA using the primers 5′-ATAGCGGCCG-CAGGTGCTTTCGGTGGTCTG-3′ and 5′-GCCA-CCTTTCGGTACCTTTTTAACC-3′. The resultant PCRproduct was digested with NotI and KpnI restrictionenzymes and cloned into modified pQE30 expressionvector (Qiagen) containing a tobacco etch virus proteaserecognition site between the histidine tag and start of themature protein. The recombinant protein was produced inE. coli BL21 (DE3) cells grown in LB broth supplementedwith 100 μg/ml ampicillin, grown with shaking at200 rpm at 37°C to an OD600 (optical density at 600 nm)of 0.5. At this point, IPTG was added to a concentration of0.5 mM and the cells were grown for a further 4 h at 37 °Cwith shaking at 200 rpm. The cells were then harvested bycentrifugation at 6000g for 15 min and resuspended in20 mM Tris, pH 8.0, 300 mM NaCl, and 10 mM imidazole,pH 8.0, prior to lysis by French press. Cellular debris werecleared by centrifugation at 30,000g for 30 min, and theresultant supernatant was loaded onto a nickel-Sepharosecolumn pre-equilibrated in 20 mM Tris, pH 8.0, and300 mM NaCl. The column was washed with 6 columnvolumes of 20 mM Tris, pH 8.0, 300 mMNaCl, and 10 mMimidazole, pH 8.0, prior to elution of the bound proteinwith 5×1 column volumes of 20 mM Tris, pH 8.0, and400 mM imidazole, pH 8.0. The eluted fractions werepooled, and further purification proceeded by gel-filtra-tion chromatography using an S75 16/60 column equili-brated in 10 mM Tris, pH 8.0, and 150 mMNaCl. Fractionscontaining LipL32 were pooled, and the solution was

diluted 1:3 with 10 mM Tris, pH 8.0, prior to loading ontoa Mono Q 5/50 (GE Healthcare) anion-exchange column.Pure LipL32 was collected from the unbound fraction. Theprotein was then dialyzed against 10 mM Tris, pH 8.0, and150 mM NaCl prior to concentration to 10 mg/ml. Thissolution was used for crystallization trials. The proteinwas proteolyzed with V8 protease following elution fromthe size-exclusion column to improve the diffractionquality of the LipL32 crystals. V8 protease (Sigma) wasadded at a 1/20 ratio (w/w) to LipL32 for 80 min and thenreapplied to the S75 16/60 column equilibrated in 10 mMTris, pH 8.0, and 150 mM NaCl to separate V8 proteasefrom Lipl32. The purified protein was concentrated to10 mg/ml.

Crystallization and X-ray diffraction data collection

Initial crystallization trials centered on the full-lengthconstruct of LipL32, including the N-terminal His6purification tag. Clusters of very thin needle-like crystalswere produced using the hanging-drop vapor-diffusionmethod at 294 K. One μl of protein solution (10 mM Tris,pH 8.0, and 150 mM NaCl) was mixed with 1 μl ofprecipitant solution (2.0 M ammonium sulfate and 0.1 MTris, pH 7.8) and suspended above 1 ml of precipitantsolution. These crystals were improved by addition of 4%1,4-butandiol; however, diffraction never exceeded 4 Å.Subsequent crystallization trials were based on purifiedLipL32 proteolyzed with V8 protease. The hanging-dropvapor-diffusion method was used with 1 μl of proteinsolution (10 mM Tris, pH 8.0, and 150 mM NaCl) mixedwith 1 μl of precipitant solution (2.0 M sodium malonate,0.1 M sodium cacodylate, pH 6.2, and 4% γ-butyrolac-tone). The drops were suspended above 1ml of precipitantsolution and incubated at 294 K. Typically, crystalsappeared within 3 days and grew to dimensions0.4 mm×0.4 mm×0.3 mm. The crystals were mounted innylon loops and then flash-cooled in a nitrogen stream at100 K without additional cryoprotectant. Native X-raydiffraction data were collected in-house in 0.25° oscilla-tions on an R-AXIS IV++ detector at a distance of 150 mm.CuKα radiation was generated by a Rigaku RU-H3RHBrotating anode generator equipped with Osmic focusingmirrors (Auburn Hills, MI). Derivative crystals wereproduced by soaking in precipitant solution with 2 mMpotassium tetrabromoaurate (III) added for 20 min. Thecrystals were then back-soaked 3 times in precipitantsolution prior to flash-cooling. Derivative data werecollected using 0.5° oscillations at beamline 3BM1 of theAustralian Synchrotron from a single crystal mounted350 mm from a Quantum 4 CCD detector. The crystalswere of space group P3221, with unit cell dimensionsa=b=125.9 Å and c=95.9 Å. The native data diffracted tobeyond 2.0 Å, and the derivative, to 2.8 Å resolution. Thedata were processed and scaled with DENZO39 andSCALEPACK.39 The data collection statistics are summar-ized in Table 1.

Structure determination and refinement

The two gold atoms in the asymmetric unit werepositioned from the anomalous diffraction signal usingSHELXD.40 The initial heavy-atom positions were refined,and single-wavelength anomalous diffraction phases werecalculated with BP3.41 Phase improvement was thenperformed with SOLOMON41 using solvent flipping. Aninitial model was traced into the resultant density-modified map using TEXTAL.42,43 This model was

1237Crystal Structure of a Leptospiral Antigen

subsequently built and refined using the native data to2.0 Å resolution with the automated package ARP/wARP,44 which was able to successfully build ∼90% ofthe structure. The model was improved by iterativerounds of manual building in Coot45 and maximum-likelihood-based refinement in REFMAC.41,46 Strict non-crystallographic symmetry restraints were applied in theearly rounds of refinement and were subsequentlyloosened and then removed for the final round ofrefinement. The refinement statistics are summarized inTable 1.

Accession code

The atomic coordinates and observed structure factorshave been deposited in the Research Collaboratory forStructural Bioinformatics PDB under accession code 2ZZ8.

Acknowledgements

We thank the Australian Synchrotron staff for theirassistance in data collection at the AustralianSynchrotron, Australia. This work was supportedby the Australian Research Council Centre ofExcellence in Structural and Functional MicrobialGenomics and the National Health and MedicalResearch Council (NHMRC) of Australia. J.R. is anAustralian Research Council Federation Fellow, J.P.V.and G.L.M. are NHMRC Peter Doherty Fellows, T.B.is an NHMRC Career Development Award Fellow,and P.A.C. is an NHMRC C. J. Martin Fellow.

References

1. Levett, P. N. (2001). Leptospirosis. Clin. Microbiol. Rev.14, 296–326.

2. Silva, J. J., Dalston, M. O., Carvalho, J. E., Setubal, S.,Oliveira, J. M. & Pereira, M. M. (2002). Clinicopatho-logical and immunohistochemical features of thesevere pulmonary form of leptospirosis. Rev. Soc.Bras. Med. Trop. 35, 395–399.

3. Simpson, F. G., Green, K. A., Haug, G. J. & Brookes, D.L. (1998). Leptospirosis associated with severe pul-monary haemorrhage in Far North Queensland. Med.J. Aust. 169, 151–153.

4. Trevejo, R. T., Rigau-Perez, J. G., Ashford, D. A.,McClure, E. M., Jarquin-Gonzalez, C., Amador, J. J.et al. (1998). Epidemic leptospirosis associated withpulmonary hemorrhage—Nicaragua, 1995. J. Infect.Dis. 178, 1457–1463.

5. Dolhnikoff, M., Mauad, T., Bethlem, E. P. & Carvalho,C. R. (2007). Pathology and pathophysiology ofpulmonary manifestations in leptospirosis. Braz. J.Infect. Dis. 11, 142–148.

6. Dolhnikoff, M., Mauad, T., Bethlem, E. P. & Carvalho,C. R. (2007). Leptospiral pneumonias. Curr. Opin.Pulm. Med. 13, 230–235.

7. Isturiz, R. E., Torres, J. & Besso, J. (2006). Globaldistribution of infectious diseases requiring intensivecare.Crit. Care Clin. 22, 469–488; ix.

8. Midwinter, A., Faine, S. & Adler, B. (1990). Vaccinationof mice with lipopolysaccharide (LPS) and LPS-

derived immuno-conjugates from Leptospira interro-gans. J. Med. Microbiol. 33, 199–204.

9. Cullen, P. A., Cordwell, S. J., Bulach, D. M., Haake,D. A. & Adler, B. (2002). Global analysis of outermembrane proteins from Leptospira interrogans serovarLai. Infect. Immun. 70, 2311–2318.

10. Cullen, P. A., Xu, X., Matsunaga, J., Sanchez, Y., Ko,A. I., Haake, D. A. & Adler, B. (2005). Surfaceome ofLeptospira spp. Infect. Immun. 73, 4853–4863.

11. Haake, D. A., Chao, G., Zuerner, R. L., Barnett, J. K.,Barnett, D., Mazel, M. et al. (2000). The leptospiralmajor outer membrane protein LipL32 is a lipoproteinexpressed during mammalian infection. Infect. Immun.68, 2276–2285.

12. Naiman, B. M., Alt, D., Bolin, C. A., Zuerner, R. &Baldwin, C. L. (2001). Protective killed Leptospiraborgpetersenii vaccine induces potent Th1 immunitycomprising responses by CD4 and gammadelta Tlymphocytes. Infect. Immun. 69, 7550–7558.

13. Naiman, B. M., Blumerman, S., Alt, D., Bolin, C. A.,Brown, R., Zuerner, R. & Baldwin, C. L. (2002).Evaluation of type 1 immune response in naive andvaccinated animals following challenge with Leptos-pira borgpetersenii serovar Hardjo: involvement ofWC1(+) gammadelta and CD4 T cells. Infect. Immun.70, 6147–6157.

14. Flannery, B., Costa, D., Carvalho, F. P., Guerreiro,H., Matsunaga, J., Da Silva, E. D. et al. (2001).Evaluation of recombinant Leptospira antigen-basedenzyme-linked immunosorbent assays for the ser-odiagnosis of leptospirosis. J. Clin. Microbiol. 39,3303–3310.

15. Hauk, P.,Macedo, F., Romero, E. C., Vasconcellos, S. A.,de Morais, Z. M., Barbosa, A. S. & Ho, P. L. (2008).LipL32, the major leptospiral lipoprotein: the C-terminus is the primary immunogenic domain andmediates interaction with collagen IV and plasmafibronectin. Infect. Immun. 76, 2642–2650.

16. Hoke, D. E., Egan, S., Cullen, P. A. & Adler, B. (2008).LipL32 is an extracellular matrix-interacting protein ofLeptospira spp. and Pseudoalteromonas tunicata. Infect.Immun. 76, 2063–2069.

17. Branger, C., Chatrenet, B., Gauvrit, A., Aviat, F.,Aubert, A., Bach, J. M. & Andre-Fontaine, G. (2005).Protection against Leptospira interrogans sensu latochallenge by DNA immunization with the geneencoding hemolysin-associated protein 1. Infect.Immun. 73, 4062–4069.

18. Seixas, F. K., da Silva, E. F., Hartwig, D. D., Cerqueira,G. M., Amaral, M., Fagundes, M. Q. et al. (2007).Recombinant Mycobacterium bovis BCG expressing theLipL32 antigen of Leptospira interrogans protectshamsters from challenge. Vaccine, 26, 88–95.

19. Seixas, F. K., Fernandes, C. H., Hartwig, D. D.,Conceicao, F. R., Aleixo, J. A. & Dellagostin, O. A.(2007). Evaluation of different ways of presentingLipL32 to the immune system with the aim ofdeveloping a recombinant vaccine against leptospiro-sis. Can. J. Microbiol. 53, 472–479.

20. Clements, C. S., Reid, H. H., Beddoe, T., Tynan, F. E.,Perugini, M. A., Johns, T. G. et al. (2003). The crystalstructure of myelin oligodendrocyte glycoprotein, akey autoantigen in multiple sclerosis. Proc. Natl Acad.Sci. USA, 100, 11059–11064.

21. Marland, Z., Beddoe, T., Zaker-Tabrizi, L., Lucet, I. S.,Brammananth, R., Whisstock, J. C. et al. (2006).Hijacking of a substrate-binding protein scaffold foruse in mycobacterial cell wall biosynthesis. J. Mol. Biol.359, 983–997.

1238 Crystal Structure of a Leptospiral Antigen

22. Holst, M. & Saied, F. (2003). Multigrid solution of thePoisson–Boltzmann equation. J. Comput. Chem. 14,105–113.

23. Krissinel, E. & Henrick, K. (2007). Inference ofmacromolecular assemblies from crystalline state. J.Mol. Biol. 372, 774–797.

24. Wilson, J. J., Matsushita, O., Okabe, A. & Sakon, J.(2003). A bacterial collagen-binding domain withnovel calcium-binding motif controls domain orienta-tion. EMBO J. 22, 1743–1752.

25. Moldoveanu, T., Hosfield, C. M., Lim, D., Elce, J. S.,Jia, Z. & Davies, P. L. (2002). A Ca(2+) switch alignsthe active site of calpain. Cell, 108, 649–660.

26. Krissinel, E. & Henrick, K. (2004). Secondary-structurematching (SSM), a new tool for fast protein structurealignment in three dimensions. Acta Crystallogr., Sect.D: Biol. Crystallogr. 60, 2256–2268.

27. Holmstrom, C., James, S., Neilan, B. A., White, D. C. &Kjelleberg, S. (1998). Pseudoalteromonas tunicata sp.nov., a bacterium that produces antifouling agents.Int. J. Syst. Bacteriol. 48, 1205–1212.

28. Cossart, P., Boquet, P., Normark, S. & Rappuoli, R.(2005). Cellular Microbiology, 2nd edit. ASM Press,Washington, D.C.

29. Lottersberger, J., Vanasco, N. B., Viarengo, G.,Guerrero, S. A., Frank, R. & Tonarelli, G. G. (2007).Epitope mapping of Lipl32 protein of pathogenicleptospires. The 5th Meeting of the InternationalLeptospirosis Society, Quito, Ecuador.

30. Cameron, C. E., Brouwer, N. L., Tisch, L. M. &Kuroiwa, J. M. (2005). Defining the interaction of theTreponema pallidum adhesin Tp0751 with laminin.Infect. Immun. 73, 7485–7494.

31. Choy, H. A., Kelley, M. M., Chen, T. L., Møller, A. K.,Matsunaga, J. & Haake, D. A. (2007). Physiologicalosmotic induction of Leptospira interrogans adhesion:LigA and LigB bind extracellular matrix proteins andfibrinogen. Infect. Immun. 75, 2441–2450.

32. Stevenson, B., Choy, H. A., Pinne, M., Rotondi, M. L.,Miller, M. C., Demoll, E. et al. (2007). Leptospirainterrogans endostatin-like outer membrane proteinsbind host fibronectin, laminin and regulators ofcomplement. PLoS ONE, 2, e1188.

33. Liu, D. F., Mason, K. W., Mastri, M., Pazirandeh, M.,Cutter, D., Fink, D. L. et al. (2004). The C-terminalfragment of the internal 110-kilodalton passengerdomain of the Hap protein of nontypeable Haemophi-lus influenzae is a potential vaccine candidate. Infect.Immun. 72, 6961–6968.

34. Nilsson, I. M., Patti, J. M., Bremell, T., Hook, M. &Tarkowski, A. (1998). Vaccination with a recombinant

fragment of collagen adhesin provides protectionagainst Staphylococcus aureus-mediated septic death.J. Clin. Invest. 101, 2640–2649.

35. Cachia, P. J., Glasier, L. M., Hodgins, R. R., Wong,W. Y., Irvin, R. T. & Hodges, R. S. (1998). The use ofsynthetic peptides in the design of a consensussequence vaccine for Pseudomonas aeruginosa. J. Pept.Res. 52, 289–299.

36. Kelly, C. G., Younson, J. S., Hikmat, B. Y., Todryk, S.M.,Czisch, M., Haris, P. I. et al. (1999). A synthetic peptideadhesion epitope as a novel antimicrobial agent. Nat.Biotechnol. 17, 42–47.

37. Lee, K. K., Wong, W. Y., Sheth, H. B., Hodges, R. S.,Paranchych, W. & Irvin, R. T. (1995). Use of syntheticpeptides in characterization of microbial adhesins.Methods Enzymol. 253, 115–131.

38. Kelly, C. G. & Younson, J. S. (2000). Anti-adhesivestrategies in the prevention of infectious disease atmucosal surfaces. Expert Opin. Invest. Drugs, 9,1711–1721.

39. Otwinowski, Z., Borek, D., Majewski, W. & Minor, W.(2003). Multiparametric scaling of diffraction intensi-ties. Acta Crystallogr., Sect. A: Found. Crystallogr. 59,228–234.

40. Schneider, T. R. & Sheldrick, G. M. (2002). Substruc-ture solution with SHELXD. Acta Crystallogr., Sect. D:Biol. Crystallogr. 58, 1772–1779.

41. The CCP4 suite: programs for protein crystallography.Acta Crystallogr., Sect. D: Biol. Crystallogr. 50. (1994).,760–763.

42. Ioerger, T. R., Holton, T., Christopher, J. A. &Sacchettini, J. C. (1999). TEXTAL: a patternrecognition system for interpreting electron densitymaps. Proc. Int. Conf. Intell. Syst. Mol. Biol.130–137.

43. Ioerger, T. R. & Sacchettini, J. C. (2003). TEXTALsystem: artificial intelligence techniques for auto-mated protein model building. Methods Enzymol.374, 244–270.

44. Cohen, S. X., Morris, R. J., Fernandez, F. J., Ben Jelloul,M., Kakaris, M., Parthasarathy, V. et al. (2004).Towards complete validated models in the nextgeneration of ARP/wARP. Acta Crystallogr., Sect. D:Biol. Crystallogr. 60, 2222–2229.

45. Emsley, P. & Cowtan, K. (2004). Coot: model-buildingtools for molecular graphics. Acta Crystallogr., Sect. D:Biol. Crystallogr. 60, 2126–2132.

46. Murshudov, G. N., Vagin, A. A. & Dodson, E. J. (1997).Refinement of macromolecular structures by themaximum-likelihood method. Acta Crystallogr., Sect.D: Biol. Crystallogr. 53, 240–255.


Recommended