Protein sorting to the cell wall envelope of Gram-positive bacteria

(1)

Review

Protein sorting to the cell wall envelope of Gram-positive bacteria

Hung Ton-That

1

, Luciano A. Marraffini, Olaf Schneewind*

Committee on Microbiology, University of Chicago, 920 East 58th Street, Chicago, IL 60637, USA Received 24 October 2003; received in revised form 28 April 2004; accepted 28 April 2004

Available online 31 May 2004

Abstract

The covalent anchoring of surface proteins to the cell wall envelope of Gram-positive bacteria occurs by a universal mechanism requiring sortases, extracellular transpeptidases that are positioned in the plasma membrane. Surface protein precursors are first initiated into the secretory pathway of Gram-positive bacteria via N-terminal signal peptides. C-terminal sorting signals of surface proteins, bearing an LPXTG motif or other recognition sequences, provide for sortase-mediated cleavage and acyl enzyme formation, a thioester linkage between the active site cysteine residue of sortase and the C-terminal carboxyl group of cleaved surface proteins. During cell wall anchoring, sortase acyl enzymes are resolved by the nucleophilic attack of peptidoglycan substrates, resulting in amide bond formation between the C-terminal end of surface proteins and peptidoglycan cross-bridges within the bacterial cell wall envelope. The genomes of Gram-positive bacteria encode multiple sortase genes. Recent evidence suggests that sortase enzymes catalyze protein anchoring reactions of multiple different substrate classes with different sorting signal motif sequences, protein linkage to unique cell wall anchor structures as well as protein polymerization leading to the formation of pili on the surface of Gram-positive bacteria.

Keywords:Surface protein; Sortase; Transpeptidation reaction; Pilus biogenesis; Heme-iron transport; Hyphae formation

1. Introduction

Invading bacterial pathogens first attach to and then colonize host tissues in order to mount successful infections. Surface proteins and pili are instrumental for these processes in all bacterial pathogens as they provide specific receptor – ligand interactions with tissue factors that determine both bacterial host range and site of infection [1 – 3]. Some bacterial adhesins, for example surface proteins that interact with extracellular matrix molecules of their host (fibronec-tin, collagen, laminin, fibrinogen or others), have been classified as microbial surface components recognizing adhesive matrix molecules (MSCRAMMs); full reviews of MSCRAMMs have been published elsewhere [4]. The

importance of these molecules during the pathogenesis of infection has warranted a rigorous examination of their assembly within the bacterial envelope. While surface proteins of Gram-negative bacteria are assembled in the outer membrane, Gram-positive bacteria predominantly uti-lize their cell wall as an organelle for anchoring and display of adhesive molecules [5]. Six different mechanisms of protein anchoring to the Gram-positive bacterial envelope are currently known [6,7]. (i) Sortase-mediated linkage of sorting signal bearing surface proteins to the cell wall envelope [8]. (ii) Binding of surface proteins to choline containing teichoic acids (inStreptococcus pneumoniae)[9]. (iii) Binding of surface proteins to lipoteichoic acids (in

Listeria monocytogenes) [10]. (iv) Insertion of surface proteins into the plasma membrane via an alpha-helical membrane anchor structure (ActA of L. monocytogenes)

[11]. (v) Insertion of proteins into the plasma membrane via an N-terminal diacyl-glycerol modification (BlaZ of Staph-ylococcus aureus)[12,13]. (vi) Reassociation of anchorless adhesins and invasins (in Streptococcus pyogenes) [14]. This review focuses exclusively on the role of sortase

* Corresponding author. Tel.: 9060; fax: +1-773-834-8150.

E-mail address:[email protected] (O. Schneewind). 1

Present address: Department of Molecular, Microbial and Structural Biology, University of Connecticut Health Center, 263 Farmington Ave., Farmington, CT 06030, USA.

(2)

enzymes in the envelope assembly of surface proteins with an emphasis on recent observations.

2. The Gram-positive envelope as a scaffold for surface protein anchoring

Gram-positive bacteria are surrounded by a rigid cell wall, also referred to as ‘‘murein’’ or ‘‘peptidoglycan’’

[15]. The cell wall envelope functions as a physical barrier that protects bacteria from their environment and as a rigid exoskeletal element that prevents bacterial rupture in low osmolar environments such as host tissues. Sjo¨quist et al. [16] presented the first evidence that the bacterial cell wall serves as scaffold for surface protein anchoring. These authors showed that protein A

of S. aureus could be released from the bacterial surface by treatment of staphylococci with lysostaphin, a glycyl-glycine endopeptidase that cleaves the penta-glycyl cross-bridge of the cell wall [17,18]. Lysozyme, an N-acetyl-muramidase that cuts the glycan strands

[19], releases protein A molecules as a spectrum of fragments with different mass due to the presence of linked peptidoglycan fragments of different sizes

[16,20,21].

Although bacterial peptidoglycan structure varies from one species to another, several structural or functional elements are conserved(Fig. 1) [15,22]. The glycan strands of all bacterial peptidoglycan consist of repeat disaccharide units, N-acetylglucosamine – (h1 – 4)-N-acetylmuramic acid

(GlcNAc – MurNAc)[23]. Glycan chains are cross-linked by short cell wall peptides, generating a three-dimensional molecular network that maintains the integrity of the bacte-rium[24,25]. The amino group of the first amino acid in the peptide moiety is linked to the carboxyl group of the lactic acid of N-acetylmuramic acid via an amide bond (Fig. 1) [26].Table 1summarizes common features of some Gram-positive bacterial peptidoglycans.

Peptidoglycan biosynthesis occurs in the bacterial cy-toplasm, membrane and extracellular cell wall compart-ments [5,27]. After synthesis in the cytoplasm, nucleotide precursors form a phosophodiester bond with an undecap-renol carrier (lipid I) and then N-acetylglucosamine is attached to generate lipid II [27]. Finally, penicillin bind-ing proteins catalyze the polymerization of lipid II sub-units via transglycosylation and transpeptidation reactions, hence generating the cross-linked peptidoglycan that con-stitutes the main component of the bacterial cell wall [27]. Lipid II is also the substrate for sortase [28,29]. The enzyme utilizes the free amino group of the peptidoglycan cross-bridge to form an amide bond with the carboxyl group produced by the cleavage of the cell wall sorting signal of the sorted protein. In this regard, the sortase reaction could be considered a transpeptidation between lipid II and surface protein. This protein-containing lipid II is incorporated into the growing peptidoglycan by the action of the penicillin binding proteins, resulting in the concomitant incorporation of the surface protein into the cell wall (Fig. 2) [30].

Fig. 1. A diagram of Gram-positive bacterial peptidoglycan. Peptidogly-can consists of glyPeptidogly-can chains, cell wall peptides, and cross-bridges (CB). Generally, glycan chains are composed of a repeating disaccharide, GlcNAc – MurNAc. The glycan chain is linked to peptide moieties via the lactyl linkage of MurNAc to the alanine residue of the tetrapeptide, with the sequence L-Ala-iGlu-DAA-D-Ala, where DAA represents a diamino acid such as L-lysine or m-diaminopimelic acid. The peptidoglycan chains are connected by cross-bridges (CB) whose structure varies in different bacterial species, connecting the diamino acid at position three (L-Lys orm-Dpm) to theD-Ala at position four in

a neighboring subunit.

Table 1

Peptidoglycan of Gram-positive bacteria

Species Glycan strand Peptide moiety Cross-bridge (CB)n References

Corynebacterium diphtheriae GlcNAc – MurNAc L-Ala-D-iGlu-m-Dpm-D-Ala m-Dpm [15]

Bacillus anthracis GlcNAc – MurNAc L-Ala-D-iGlu-m-Dpm-D-Ala m-Dpm [15,68]

Clostridium perfringens GlcNAc – MurNAc L-Ala-D-iGlu-m-Dpm-D-Ala Gly [69]

Enterococcus faecalis GlcNAc – MurNAc L-Ala-D-iGlu-L-Lys-D-Ala (L-Ala)2 [70] Lactococcus lactis GlcNAc – MurNAc L-Ala-D-iGlu-L-Lys-D-Ala D-Asp [15,71]

Listeria monocytogenes GlcNAc – MurNAc L-Ala-D-iGlu-m-Dpm-D-Ala mDpm [15,33]

Staphylococcus aureus GlcNAc – MurNAc L-Ala-D-iGlu-L-Lys-D-Ala Gly5 [21,32] Streptococcus pyogenes GlcNAc – MurNAc L-Ala-D-iGlu-L-Lys-D-Ala (L-Ala)2 [5,72] Streptomyces coelicolor GlcNAc – MurNAc L-Ala-D-iGlu-L,L-Dpm-D-Ala Gly [15]

(3)

3. Sorting signals of Gram-positive bacterial surface proteins

The C-terminal cell wall sorting signal of staphylococcal protein A encompasses a 35-residue peptide with an LPXTG motif, followed by a hydrophobic domain and a positively charged tail [20]. In addition, surface proteins also contain an N-terminal signal peptide that promotes their translocation across the bacterial membrane [31]. It is presumed that the hydrophobic and positively charged residues of the cell wall sorting signal retain the translocated protein within the plasma membrane until the substrate is recognized by sortase[8] (Fig. 3).

Mutations that truncate the sorting signal cause the secretion of mutant protein A molecules into the extracel-lular medium [8]. In contrast, mutations that remove or substitute residues within the LPXTG motif abolish sortase-mediated cell wall linkage without secretion of mutant protein A [8]. When fused to the C-terminus of polypep-tides, the cell wall sorting signal alone is sufficient to cause cell wall anchoring of hybrid proteins that are initiated into the secretory pathway ofS. aureusvia an N-terminal signal peptide[13,20,32 – 34]. Moreover, sorting signals from one bacterial species can be functional in another microbe[20]. When they fail to do so, however, either mutations that alter the distance between the LPXTG motif and the charged tail

Fig. 2. Biosynthesis pathway of bacterial cell wall assembly. Peptidoglycan synthesis begins with the formation of UDP-MurNAc from UDP-GlcNAc and phosphoenolpyruvate. The UDP-MurNAc molecule acquiresL-Ala,D-iGlu, DAA (L-Lys orm-Dpm), andD-Ala-D-Ala by a series of enzymatic reactions to generate Park’s nucleotide. Lipid I molecule is generated from the Park’s nucleotide and undecaprenylpyrophosphate. Modification of lipid I by UDP-GlcNAc generates the lipid II molecule, which then acquires the cross-bridge (DAA) at theq-amino group of lysine orN-amino group of diaminopamelic acid. Lipid II becomes the precursor for cell wall synthesis after it has been translocated across the membrane, serving as substrate for cell wall synthesis and protein sorting process (modified after Navarre and Schneewind[5]).

(4)

or mutations that affect residues within the charged tail of the sorting signal can restore function of cell wall sorting signals in heterologous bacteria [20]. Sorting signals have been observed in several hundred predicted gene products, identified via genome sequencing of Gram-positive bacteria

[5,35 – 37]. While most of these sorting signals carry the LPXTG motif, in many other cases they harbor variations of the sequence motif.Table 2lists some of these non-LPXTG motifs in the context of the complete cell wall sorting signal. If a surface protein gene is located in the same transcrip-tional unit with a sortase gene, it is generally presumed that the two genes encode enzyme/substrate pairs [37]. This hypothesis has been corroborated for S. aureus IsdC-SrtB andL. monocytogenesSvpA-SrtB[35,38](vide infra).

4. Sortase

A genetic screen for S. aureus mutants that failed to anchor a reporter protein to the bacterial cell wall resulted in the identification of the sortase gene, named srtA (surface protein sorting A) [39]. Sortase homologs have been found in all available Gram-positive bacterial genomes and in most cases, more than one sortase gene has been identified

[36,37]. Sortase enzymes typically encompass an N-termi-nal sigN-termi-nal peptide which also serves as a membrane anchor (type II membrane protein) [40]. It is conceivable that membrane-anchored sortase may be positioned in immedi-ate vicinity of protein translocation sites as these enzyme are expected to scan polypeptide sequences for the presence of

Fig. 3. The cell wall sorting pathway in Gram-positive bacteria. Cell wall synthesis begins in the bacterial cytoplasm via the assembly of nucleotide linked wall peptides. After the precursor is transferred to undecaprenylpyrophosphate, the lipid II molecules are translocated across the cytoplasmic membrane (1). Surface proteins are synthesized as precursors in the bacterial cytoplasm bearing an N-terminal signal peptide and a C-terminal sorting signal. The sorting signal is comprised of an LPXTG sequence motif, followed by a hydrophobic domain (black box) and tail of positively charged residues (boxed +). Following cleavage of the N-terminal signal peptide, the hydrophobic domain and charged tail of the precursor molecule retain surface proteins in the secretory pathway. The precursor is substrate for cleavage by sortase, a membrane-anchored transpeptidase, thereby generating an acyl enzyme intermediate and the cleaved sorting signal (2). The acyl-enzyme intermediate, a thioester bond between the thiol of sortase and the carboxyl group of threonine at the C-terminal end of surface proteins, is resolved by the nucleophilic attack of the amino group of the cross-bridge (CB) within lipid II precursor (3). Surface proteins linked to lipid II may be incorporated into the cell wall envelope by the transglycosylation and transpeptidation reactions that generate mature cell wall (4, 5).

Table 2

Different types of sorting signals of surface proteins in Gram-positive bacteria

Organism C-terminal sorting signal Protein

A. naeslundii LPLTG ANGMLILTASGAALLMIAVGSVLVARYRERKRNRDLAA FimA

C. diphtheriae LPLTG GSGRIAITIGIVGLLVALASYVLSRRKDNR SpaA

L. monocytogenes LPTTG DSDNALYLLLGLLAVGTAMALTKKARASK InlA

L. monocytogenes LPEAG RRKAEILTLAAASLSSVAGAFISLKKRK PAB

S. aureus LPKTG LTSVDNFISTVAFATLALLGSLSLLLFKRKESK IsdA

S. pyogenes LPASG DKREASFTIVALTIIGAAGLLSKKRRDTEEN SfbII

C. diphtheriae LAFTG ASILGLLAIATISTLIGIALLRTRRAEKKG SpaB

C. diphtheriae LALTG VQIIGLVLAAVALMGAGLLMLLITKKRKQEG SpaE

C. diphtheriae LGNTG ANVLGIAALGIALAIAGFLVQRRKKNEENG SpaI

B. anthracis NPKTG DEARIGLFAALILISGVFLIRKVKLSK IsdC

(5)

recognition motifs (LPXTG or otherwise). Based on the presence or absence of topogenic sequences, sortases can be divided into two classes. The major class of sortases encompasses enzymes with an N-terminal signal peptide/ membrane anchor (class I). The second, minor class is formed by sortases with an N-terminal signal peptide and a C-terminal membrane anchor (class II)[41]. Enzymes of the second class are mainly found encoded in the genomes of corynebacteria, enterococci and streptococci. All sortase homologs contain an active site signature motif, LXTC with conserved leucine (L), threonine (T), and cysteine (C) residues [41,42]. In addition to the active site cysteine, histidine 120 of sortase A is also absolutely required for catalysis[43]. Staphylococcal sortase A, a class I enzyme, contains 206 amino acid residues [42]. NMR experiments revealed the three-dimensional structure of SrtA as an eight-strandedh-barrel structure with strandsh7 andh8 forming the floor of a hydrophobic depression that builds the active site[42]. The wall of this catalytic pocket is constructed by residues located in loops connecting strands h3 –h4, h2 –

h3, h6 –h7, and h7 –h8. A putative active site cysteine – histidine ion pair is located within this pocket, where the LPXTG motif of a surface protein is cleaved [43,44]. Recently, the X-ray structure ofS. aureussortase B revealed the presence of an arginine residue in the near vicinity of the catalytic cysteine which is highly conserved among all sortases [45]. Formation of a Cys – Arg catalytic dyad represents an alternative possibility for a reaction mecha-nism requiring enzyme nucleophilic attack at the scissile peptide bond[45].

Owing to the presence of the aforementioned signal peptide and cell wall sorting signal processing mechanisms, four distinct surface protein species can be detected during the sorting process. The P1 precursor is synthesized in the cytoplasm with an N-terminal signal peptide and C-terminal sorting signal. The P2 precursor is generated by cleavage of the N-terminal signal peptide but retains the C-terminal sorting signal. The P3 precursor, i.e. the product of the sortase reaction, represents surface protein linked to lipid II. Finally, the mature, anchored form of surface proteins is generated after incorporation of P3 precursors into the cell wall envelope via transglycosylation and transpeptidation reactions. Staphylococci lacking thesrtA gene accumulate the P2 precursor of protein A in the bacterial plasma membrane[39].S. aureusstrains contain genes for 18 – 22 different surface proteins with C-terminal LPXTG sorting signals [36,46]. Deletion of the srtA gene abolishes the function of all of these sorting signals and abrogates anchoring and display of an entire class of surface proteins

[35,39]. Sortase A mutant staphylococci also display re-duced virulence in several different animal models of disease, revealing a plethora of important functions of surface proteins during the pathogenesis of staphylococcal diseases[40,47,48]. A secondS. aureussortase gene,srtB, plays no role in the anchoring of LPXTG-type surface proteins. Sortase B cleaves an NPQTN sorting signal that

has been observed at the predicted C-terminus of IsdC, a heme binding surface protein[35]. It has been proposed that SrtB-mediated anchoring of IsdC leads to a unique cell wall anchor structure, which is distinct from the anchor structures of SrtA-catalyzed sorting reactions[49]. Current knowledge suggests that IsdC may be the only sorting substrate of sortase B inS. aureus(vide infra). InL. monocytogenes, it was recently found that sortase B catalyzes the anchoring of the SvpA surface protein [38]. SvpA encompasses an NAKTN cell wall sorting motif and its structural gene is transcribed together with thesrtBgene in the same operon. Sortase B mutant listeria do not display SvpA on the bacterial surface, a phenotype that could be complemented by the expression of plasmid-encodedsrtB[38]. The anchor structure of SvpA and the role of the NAKTN sequence in this process remain, however, to be determined.

The genome of S. pyogenes, another Gram-positive bacterial human pathogen, possesses four different sortase genes [37]. Deletions in streptococcal srtA abrogate the anchoring of LPXTG containing surface proteins such as M6 protein, protein F, C5a peptidase (ScpA), and GRAB

[50]. Surprisingly, the anchoring of another LPXTG-type surface protein, T6, is not affected by the deletion of the

srtA gene. However, inactivation of the streptococcal srtB

gene abolished the surface display and anchoring of T6 protein without affecting that of M6 protein, protein F, C5a peptidase (ScpA), or GRAB[50]. The mechanism whereby two different sortase enzymes distinguish between surface protein substrates that carry the same LPXTG motif is still unknown. One explanation for these observations is that the

tee6gene (encoding the T6 protein) may be expressed under conditions when srtAis not activated. Alternatively, strep-tococcal sortases may distinguish substrates by properties other than (LPXTG) motif sequences.

The expression of some sortase genes is coordinately regulated during infection. ThesrtBgene ofS. aureusserves again as a paradigm. Staphylococcal srtAis constitutively expressed, whereas srtB expression is only induced under iron starvation, a condition that bacteria encounter when entering their mammalian hosts[35,49]. The iron-regulated surface determinant locus, isd, encompasses two genes encoding sortase A substrate polypeptides (LPXTG motif containing surface proteins) as well as the srtB and isdC

genes. It is not immediately obvious why staphylococci require two sortase enzymes under iron starvation condi-tions as the two enzymes simultaneously anchor different polypeptides to the cell wall envelope. It is presumed that SrtB-mediated anchoring can provide for a designated anchor structure, subcellular location and/or function of IsdC in heme-iron transport.

5. Sortase catalysis

Sortase cleaves the LPXTG motif between the threonine (T) and the glycine (G) residues both in vivo and in vitro,

(6)

capturing cleaved polypeptide as a thioester-linked acyl enzyme at its active site cysteine residue[34,51,52]. Catal-ysis is completed by the nucleophilic attack of the amino group of cell wall cross-bridges within lipid II [28 – 30,32,51]. In the absence of a specific nucleophile for the sorting reaction, S. aureus sortase A catalyzes a slow hydrolysis of peptides, cleaving the peptide bond between the threonine and the glycine of the LPXTG motif. The presence of specific nucleophiles that mimic the physiolog-ical substrate of the sorting reaction (tri- and pentaglycine in the case of S. aureus) leads to an increase in the catalytic rate[52,53]. Current structural biology studies with sortase enzymes aim at defining the binding sites for polypeptide (LPXTG) and peptidoglycan (lipid II) substrates and pro-vide structural comparisons between different types of sortases with different function. Sorting reactions are inhibited by organic mercurials and methylmethane thiosul-fonates [30,51]. The latter reagent forms a disulfide bond with the sulfhydryl group of the active site cysteine, thus generating an acyl enzyme intermediate. Consequently, methylmethane thiosulfonate-inhibited sortase can be reac-tivated with the reducing agent dithiothreitol (DTT). Thio-ester-linked acylenzyme intermediates of sortase are sensitive to hydroxylaminolysis, releasing surface protein threonine hydroxamate into the extracellular medium and preventing cell wall anchoring [51]. Taken together, these observations corroborate the hypothesis that sortases func-tion as transpeptidases that catalyse the cleavage and linkage of surface protein to envelope substrates.

6. Pilus assembly in Gram-positive bacteria

Proteinaceous filaments on microbial surfaces, named pili or fimbriae, provide adhesive functions and play im-portant roles during the establishment of respiratory, urinary, periodontal and intestinal infectious diseases[54 – 57]. The structure, function, and biogenesis of pili in Gram-negative bacteria have been revealed. In contrast to Gram-negative pathogens, pili of Gram-positive bacteria have been less well studied[3].Actinomyces naeslundii, a human pathogen that can be isolated from supragingival dental plaque, forms fimbriae, short filaments on the bacterial surface that are composed of FimP (type I fimbriae) or FimA (type II fimbriae)[58]. Yeung[58]showed that thefimP and fimA

genes encode major protein subunits ofA. naeslundiitype I and type II fimbriae, respectively. Studies by Yeung and Ragsdale [59] revealed chromosomal regions that are in-volved in the biogenesis of A. naeslundii T14V type I fimbriae. An operon encompassing seven open reading frames was identified with the gene order (5Vto 3Vcoding sequence): orf3-orf2-orf1-fimP-orf4-orf5-orf6. orf1 and

fimP encode protein products whose primary sequences display striking similarity to the topogenic sequences of cell wall-anchored surface proteins ofS. aureus. The fimA

gene is located elsewhere on the actinomycetal chromosome

and theorf365gene is located immediately adjacent tofimA

[60]. An orf365 mutant strain failed to assemble type II fimbriae [60]. Instead, the mutant accumulated monomeric FimA subunits in the bacterial envelope that contained uncleaved C-terminal sorting signals. A similar result was observed whenorf4was disrupted; FimP and Orf1 subunits accumulated in the envelope, and the fimbrial structure was not assembled[59]. After the identification of staphylococ-cal srtA, orf4 and orf365 were assigned as sortase-like genes, which presumably cleave the C-terminal sorting signals of fimbrial precursors proteins in a pathway that penultimately leads to the formation of fimbriae. Orf4 appears to be responsible for the assembly of type I pili but not of type II fimbriae, whereas Orf365 fulfills the reciprocal role for type II fimbriae [3,58]. In contrast to staphylococci, where sortase anchors monomeric surface proteins to the cell wall envelope, Actinomyces sortases appear to polymerize fimbrial precursors into a supramo-lecular structure that may be composed of repeating sub-units. Such a scenario is thought to require the formation of novel sortase-mediated peptide bonds, as the cleaved sorting signal is presumably attached to another fimbrial subunit (vide infra). This hypothesis does not preclude the possibil-ity that one or more fimbrial subunits of Actinomyces are anchored to the bacterial peptidoglycan in a manner resem-bling staphylococcal sortase reactions.

Li et al. [61] examined A. viscosus ATCC19246 and identfied a fimbrial operon that encompasses orfA, fimP,

orfB, andorfC, whereorfBencodes the sortase homolog and

fimPand orfApilin subunits. These authors speculated that the prepilin peptidase-like OrfC may cleave the signal peptide of FimP and OrfA during secretion, releasing both proteins to become substrates of the sortase OrfB. Wu and Fives-Taylor [3]investigated fimbrial formation in Strepto-coccus parasanguis,Streptococcus sanguis, and Streptococ-cus salivarius.S. parasanguisassembles adhesive filaments containing Fap1, which harbors a cell wall sorting signal. A precipitation or nucleation mechanism has been proposed for the assembly of fimbriae in streptococci [3]. According to this model, the anchored form of Fap1 may be targeted to the bacterial cell wall and may then serve as a scaffold for polymerization of soluble Fap1. This polymerization may be terminated when FimA interacts with soluble Fap1. The mechanism underlying this process is still unknown.

Recently, the assembly of pili inCorynebacterium diph-theriae was examined [62]. Corynebacterial pili were first revealed in electron microscopic studies by Yanagawa and Honda[63]. Homology searches with the translated genome sequence of C. diphtheriae NCTC13129, using S. aureus

SrtA and LPXTG sorting signals as queries in BLAST searches, revealed three chromosomal gene clusters that harbored nine surface protein genes encoding N-terminal signal peptides and C-terminal sorting signals (spaA – I,spa

for sortase-mediated pilus assembly) as well as five sortase genes (srtA – E)(Fig. 4A). A sixth sortase gene (srtF) was located elsewhere on the chromosome. Using

(7)

immunogold-labeling and electron microscopy, three proteins (SpaA, SpaB and SpaC) were shown to be components of coryne-bacterial pili. SpaA, the major pilin protein, is distributed uniformly along the pilus shaft, while SpaB is observed at regular intervals and SpaC seems positioned at the pilus tip

(Fig. 4B – E). Assembled pili are released from the bacterial surface by treatment with murein hydrolase, suggesting that pilus fibers may be anchored to the cell wall envelope. All three pilin subunit proteins are synthesized as precursors carrying N-terminal signal peptides and C-terminal sorting signals. Deletion of srtA, but not of other sortases (srtB – F), abolished pilus assembly. Three arguments support the hypothesis that corynebacterial pili are composed of

cova-lently linked protein subunits and tethered to the bacterial cell wall. (i) Pili are left intact under harsh chemical conditions that would dissociate Gram-negative pili (me-chanical disruption, boiling in SDS, and treatment with 70% formic acid). (ii) Pili are released from the coryne-bacterial surface by enzymatic cleavage of the peptidogly-can. (iii) Pilus assembly is a sortase (transpeptidase)-mediated process [62].

Sequence alignment of FimA, FimP (A. naeslundii), SpaA, SpaD, and SpaH (C. diphtheriae) revealed a pilin motif sequence with a conserved lysine residue [62]. Sub-stitution of this lysine with alanine or arginine abolished pilus assembly, as did amino acid substitutions within the LPXTG motif. The pilin-specific sortase enzyme is pro-posed to cleave SpaA precursor protein between the threo-nine and the glycine of sorting signals (LPXTG motif), resulting in the formation of a thioester linkage between the carboxyl of threonine and the enzyme sulfhydryl [62]. In vivo, the acyl-enzyme intermediate may be resolved by the nucleophilic attack of the side chain amino group of lysine (K) within the pilin motif of SpaA to generate amide bonds with the carboxyl group of threonine (within the LPXTG motif). Alternatively, the nucleophilic attack of the amino group within themeso-diaminopimelic acid cross-bridge of corynebacterial lipid II would lead to cell wall anchoring of the surface proximal subunit of mature pili, thereby termi-nating the assembly process(Fig. 5).

Pilus assembly is proposed to occur by a mechanism of ordered cross-linking, whereby pilin-specific sortase enzymes cleave precursor proteins at sorting signals and involve the side chain amino groups of pilin motif sequen-ces to generate links between pilin subunits. This covalent tethering of adjacent pilin subunits appears to have evolved in many Gram-positive pathogens such as A. naeslundii,

Fig. 5. A proposed model of sortase-catalyzed pilus assembly. Sortase is proposed to catalyze the formation of linkages between SpaC and SpaA, and SpaA subunits, a mechanism requiring the nucleophilic attack of the amino group of lysine at the thioester-linked acyl-enzyme (S – CMO), generated by the cleavage of sorting signals at the LPXTG motif. SpaA pili are presumably linked to the cell wall envelope. (Modified after Ton-That and Schneewind[62]).

Fig. 4. Pilus components ofC. diphtheriae. (A) Diagram of three gene clusters found in the chromosome of C. diphtheriae NCTC13129 that encode sortase genes (srtA – E) and sortase-mediated pilus assembly genes (spaA – I). Predicted promoters as well as the direction of transcription are shown with arrows. Similarities betweenspagene products are indicated as shared patterns. (B – E). Corynebacterial pili are stained with specific antiserum (a-SpaA,a-SpaB, ora-SpaC) and IgG-conjugated gold particles, either 12 nm (B – C) or 12 nm (a-SpaB anda-SpaC, D and E) and 6 nm (a -SpaA, D and E). Samples were viewed by transmission electron microscopy. Bars indicate a distance of 0.2Am. (Modified after Ton-That and Schneewind[62]).

(8)

Clostridum perfringens,Enterococcus faecalis, Streptococ-cus agalactiae, and S. pneumoniaethat encode sortase and pilin subunit genes with sorting signals and pilin motifs

[62]. Thus, the assembly of corynebactial pili resembles the sorting mechanism of surface proteins in Gram-positive bacteria, in which sortase A is the catalyst.

7. Protein sorting and iron acquisition inS. aureus

A chromosomal locus namedisd(iron-responsive surface determinants) is involved in heme-iron transport of S. aureus [35,49]. The isd locus is comprised of three tran-scriptional units. The sortase B gene (srtB) is located in a cistron with genes encoding heme-iron transport (isdE,

isdF) and heme-iron binding proteins. One of these binding proteins, IsdC, is synthesized as a precursor with a NPQTN type C-terminal sorting signal that is cleaved and tethered to the cell wall envelope via a sortase B-dependent mecha-nism. Two other surface proteins, encoded byisdBandisdA, carry LPXTG motif sorting signals and are cell wall-anchored by sortase A. These factors first bind hemoglobin (IsdB) and presumably remove bound heme-iron (IsdA, IsdB). Heme-iron is then transported across the cell wall envelope (IsdA, IsdC) and plasma membrane (IsdD, IsdE and IsdF). A heme oxygenase, IsdG, cleaves the tetrapyrrol ring of heme in the bacterial cytoplasm and liberates iron for various biosynthetic pathways [64]. In this model, sortase SrtA and SrtB function to assemble components of the

heme-iron transport machinery to unique locations within the cell wall envelope[49]. The LPXTG anchor structures of IsdA and IsdB promote receptor function for hemoglobin and heme-iron on the staphylococcal surface, whereas the NPQTN structure may function in heme-iron passage across the bacterial cell wall (Fig. 6).

InL.monocytogenes, in spite of the genomic similarities between the isd gene cluster of listeria and staphylococci, there is, however, no evidence involving sortase B and its substrate, SvpA (homolog of IsdC), in iron uptake [38]. Moreover, using the mouse model of listerial infection, the

srtB mutant strain displayed no measurable defect in viru-lence. Further investigation will be required to address the function of listerial sortase B, as iron acquisition strategies of L. monocytogenes, which assumes a predominantly intracellular life cycle, seem to differ markedly from the mechanisms used by the extracellular pathogenS. aureus.

8. Surface protein and hyphae development in Streptomyces coelicolor

S. coelicolor, a Gram-positive soil-dwelling bacterium, is a representative for a group of filamentous bacteria which are the main source of natural antibiotics [65]. Their life cycle is quite complex, beginning with the formation of a feeding (submerged) mycelium, which produces antibiotics. Mycelium differentiates into aerial hyphae that then septate into spore chains. The hyphal surface becomes hydrophobic,

Fig. 6. A model of an iron-acquisition apparatus mediated byisdlocus. Schematic illustrates the deduced location of Isd proteins and their binding activities. Heme is removed from presumed bacterial surface receptors (IsdA and IsdB) and transferred to the cell wall receptor IsdC and the membrane translocators (IsdD/IsdE/IsdF). Once heme-iron is transported into the cytoplasm, the hemoxygenase IsdG cleaves heme and releases iron. (Courtesy of S.K. Mazmanian and E. Skaar).

(9)

a property that promotes aerial hyphae outgrowth into the air and facilitates the dispersal of spores. Recently, two groups have identified a set of six genes encoding proteins that are likely involved in aerial hyphae formation and which were named chaplins [66,67]. All chaplins are syn-thesized as precursors with an N-terminal leader sequence and a short hydrophobic domain (chaplin domain) of about 40 residues. Chaplins D – H are very short polypeptides, whereas chaplins A – C are longer and carry a C-terminal sorting signal (LAXTG motif). Deletion of six chaplin genes (chpA – H) hindered the formation of aerial hyphae; the defect could be complemented by the addition of exogenous chaplin proteins. A model for aerial hyphae assembly has been proposed, in which surface proteins ChpA – D are thought to function as a scaffold for other chaplins. As ChpE and ChpH are expressed during mycelium formation, these proteins are presumed to form a film that reduces water surface tension and allows hyphal protrusion into the air. All chaplins, which are expressed in the aerial hyphae, are presumed to diffuse to the outer surface and to assemble a layer of fibrils that confers hydrophobicity to the aerial hypha. It is suggested that chaplins A – C are anchored to the peptidoglycan via sorting signals with a LAXTG motif, allowing the smaller chaplins DH to be immobilized and nucleated on the cell surface by interaction with their chaplin domains(Fig. 7) [66].

9. Conclusions

One can view the cell wall envelope of Gram-positive bacteria as a surface organelle with anchored proteins that require unique mechanisms and enzymatic machines for localization. Sortases are membrane transpeptidases that catalyze amide bond exchange mechanisms, using sorting signals of precursor proteins and cell wall envelope compo-nents as substrates. Sortases are strategically positioned on the outside of bacterial cells and catalyze amide bond exchange reactions with newly translocated polypeptides,

reactions that are fueled by peptide bond cleavage without direct consumption of high energy phosphates. Sortases are also involved in the polymerization of protein subunits that assemble into pili. A wide variety of sortase enzymes recognize cognate sorting signals that allow Gram-positive bacteria to regulate protein composition and functional display within the cell wall envelope in response to envi-ronmental signals or specific host signals during infection. Thus, the identification of inhibitors for sortase-catalyzed surface protein anchoring may provide compounds that are useful for the therapy of Gram-positive bacterial infections.

Acknowledgements

This work was supported by grants from the United States Public Health Service and the National Institute of Allergy and Infectious Diseases, Infectious Disease Branch, awards AI38897 and AI52474, to Olaf Schneewind.

References

[1] T.J. Foster, D. McDevitt, FEMS Microbiol. Lett. 118 (1994) 199 – 205. [2] M.A. Kehoe, in: J.-M. Ghuysen, R. Hakenbeck (Eds.), Bacterial Cell

Wall, Elsevier, Amsterdam, 1994, pp. 217 – 261.

[3] H. Wu, P.M. Fives-Taylor, Crit. Rev. Oral Biol. Med. 12 (2001) 101 – 115.

[4] J.M. Patti, B.L. Allen, M.J. McGavin, M. Hook, Annu. Rev. Micro-biol. 48 (1994) 585 – 617.

[5] W.W. Navarre, O. Schneewind, Microbiol. Mol. Biol. Rev. 63 (1999) 174 – 229.

[6] P. Cossart, R. Jonquieres, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 5013 – 5015.

[7] G.S. Chhatwal, Trends Microbiol. 10 (2002) 205 – 208.

[8] O. Schneewind, P. Model, V.A. Fischetti, Cell 70 (1992) 267 – 281. [9] J.-V. Ho¨ltje, A. Tomasz, J. Biol. Chem. 250 (1975) 6072 – 6076. [10] R. Jonquieres, H. Bierne, F. Fiedler, P. Gounon, P. Cossart, Mol.

Microbiol. 34 (1999) 902 – 914.

[11] C. Kocks, E. Gouin, M. Tabouret, P. Berche, H. Ohayon, P. Cossart, Cell 68 (1992) 521 – 531.

Fig. 7. A model of aerial hyphae formation involved in chaplins and rodlins. In the submerged growth, hyphae secrete ChpE and ChpH, which form a layer at the water – air interface to lower surface tension that allows the hyphae to grow into the air. It is proposed that the surface proteins ChpA – C, which are secreted as monomers, assemble into fibrillar structures by interacting with other chaplins (D – H). The chaplin structure may allow the formation of the rodlet layer composed of RdlA and RdlB. (Modified after Claessen et al.[66].

(10)

[12] J.B.K. Nielsen, J.O. Lampen, J. Bacteriol. 152 (1982) 315 – 322. [13] W.W. Navarre, S. Daefler, O. Schneewind, J. Bacteriol. 178 (1996)

441 – 446.

[14] V. Pancholi, V.A. Fischetti, J. Exp. Med. 186 (1997) 1633 – 1643. [15] K.H. Schleifer, O. Kandler, Bacteriol. Rev. 36 (1972) 407 – 477. [16] J. Sjo¨quist, J. Movitz, I.-B. Johansson, H. Hjelm, Eur. J. Biochem. 30

(1972) 190 – 194.

[17] C.A. Schindler, V.T. Schuhardt, Proc. Natl. Acad. Sci. U. S. A. 51 (1964) 414 – 421.

[18] J. Sjo¨quist, B. Meloun, H. Hjelm, Eur. J. Biochem. 29 (1972) 572 – 578. [19] J.H. Hash, M.V. Rothlauf, J. Biol. Chem. 242 (1967) 5586 – 5590. [20] O. Schneewind, D. Mihaylova-Petkov, P. Model, EMBO 12 (1993)

4803 – 4811.

[21] W.W. Navarre, H. Ton-That, K.F. Faull, O. Schneewind, J. Biol. Chem. 273 (1998) 29135 – 29142.

[22] J.-M. Ghuysen, Bacteriol. Rev. 32 (1968) 425 – 464.

[23] J.-M. Ghuysen, J.L. Strominger, Biochemistry 2 (1963) 1119 – 1125. [24] J.-M. Ghuysen, D.J. Tipper, C.H. Birge, J.L. Strominger,

Biochemis-try 4 (1965) 2245 – 2254.

[25] D.J. Tipper, J.L. Strominger, Proc. Natl. Acad. Sci. U. S. A. 54 (1965) 1133 – 1141.

[26] D.J. Tipper, J.-M. Ghuysen, J.L. Strominger, Biochemistry 4 (1965) 468 – 473.

[27] J.L. Strominger, K. Izaki, M. Matsuhashi, D.J. Tipper, Fed. Proc. 26 (1967) 9 – 18.

[28] A. Ruzin, A. Severin, F. Ritacco, K. Tabei, G. Singh, P.A. Bradford, M.M. Siegel, S.J. Projan, D.M. Shlaes, J. Bacteriol. 184 (2002) 2141 – 2147.

[29] A.M. Perry, H. Ton-That, S.K. Mazmanian, O. Schneewind, J. Biol. Chem. 277 (2002) 16241 – 16248.

[30] H. Ton-That, O. Schneewind, J. Biol. Chem. 274 (1999) 24316 – 24320. [31] T. Bae, O. Schneewind, J. Bacteriol. 185 (2003) 2910 – 2919. [32] H. Ton-That, K.F. Faull, O. Schneewind, J. Biol. Chem. 272 (1997)

22285 – 22292.

[33] G. Dhar, K.F. Faull, O. Schneewind, Biochemistry 39 (2000) 3725 – 3733.

[34] W.W. Navarre, O. Schneewind, Mol. Microbiol. 14 (1994) 115 – 121. [35] S.K. Mazmanian, H. Ton-That, K. Su, O. Schneewind, Proc. Natl.

Acad. Sci. U. S. A. 99 (2002) 2293 – 2298.

[36] D. Comfort, R.T. Clubb, Infect. Immun. 72 (2004) 2710 – 2722. [37] M.J. Pallen, A.C. Lam, M. Antonio, K. Dunbar, Trends Microbiol. 9

(2001) 97 – 101.

[38] H. Bierne, C. Garandeau, M.G. Pucciarelli, C. Sabet, S. Newton, F. Garcia-del Portillo, P. Cossart, A. Charbit, J. Bacteriol. 186 (2004) 1972 – 1982.

[39] S.K. Mazmanian, G. Liu, H. Ton-That, O. Schneewind, Science 285 (1999) 760 – 763.

[40] S.K. Mazmanian, G. Liu, E.R. Jensen, E. Lenoy, O. Schneewind, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 5510 – 5515.

[41] S.K. Mazmanian, H. Ton-That, O. Schneewind, Mol. Microbiol. 40 (2001) 1049 – 1057.

[42] U. Ilangovan, H. Ton-That, J. Iwahara, O. Schneewind, R.T. Clubb, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 6056 – 6061.

[43] H. Ton-That, S.K. Mazmanian, L. Alksne, O. Schneewind, J. Biol. Chem. 277 (2002) 7447 – 7452.

[44] K.M. Connolly, B.T. Smith, R. Pilpa, U. Ilangovan, M.E. Jung, R.T. Clubb, J. Biol. Chem. 278 (2003) 34061 – 34065.

[45] Y. Zong, S.K. Mazmanian, O. Schneewind, S.V. Narayana, Structure 12 (2004) 105 – 112.

[46] S.K. Mazmanian, O. Schneewind, in: A. Sonenshine, R. Losick, J. Hoch (Eds.),Bacillus subtilisand other Gram-Positive Bacteria, ASM Press, New York, 2002, pp. 57 – 70.

[47] I.M. Jonsson, S.K. Mazamanian, O. Schneewind, M. Vendrengh, T. Bremell, A. Tarkowski, J. Infect. Dis. 185 (2002) 1417 – 1424. [48] W.J. Weiss, E. Lenoy, T. Murphy, L. Tardio, P. Burgio, S.J. Projan, O.

Schneewind, L. Alksne, J. Antimicrob. Chemother. 53 (2004) 480 – 486.

[49] S.K. Mazmanian, E.P. Skaar, A.H. Gasper, M. Humayun, P. Gornicki, J. Jelenska, A. Joachimiak, D.M. Missiakas, O. Schneewind, Science 299 (2003) 906 – 909.

[50] T.C. Barnett, J.R. Scott, J. Bacteriol. 184 (2002) 2181 – 2191. [51] H. Ton-That, G. Liu, S.K. Mazmanian, K.F. Faull, O. Schneewind,

Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 12424 – 12429.

[52] H. Ton-That, H. Mazmanian, K.F. Faull, O. Schneewind, J. Biol. Chem. 275 (2000) 9876 – 9881.

[53] X. Huang, A. Aulabaugh, W. Ding, B. Kapoor, L. Alksne, K. Tabei, G. Ellestad, Biochemistry 42 (2003) 11307 – 11315.

[54] M.E. Pichichero, M. Loeb, Anderson, D.H. Smith, Lancet 2 (1982) 960 – 962.

[55] S.J. Hultgren, S. Abraham, M. Caparon, P. Falk, J.W. St Geme 3rd, S. Normark, Cell 73 (1993) 887 – 901.

[56] K.H. Thelin, R.K. Taylor, Infect. Immun. 64 (1996) 2853 – 2856. [57] M.K. Yeung, in: V.A. Fischetti, R.P. Novick, J.J. Ferretti, D.A. Portnoy,

J.I. Rood (Eds.), Gram-Positive Pathogens, ASM Press, Washington, DC, 2000, pp. 583 – 593.

[58] M.K. Yeung, Crit. Rev. Oral Biol. Med. 10 (1999) 120 – 138. [59] M.K. Yeung, P.A. Ragsdale, Infect. Immun. 65 (1997) 2629 – 2639. [60] M.K. Yeung, J.A. Donkersloot, J.O. Cisar, P.A. Ragsdale, J. Bacteriol.

66 (1998) 1482 – 1491.

[61] T. Li, M.K. Khah, S. Slavnic, I. Johansson, N. Stromberg, Infect. Immun. 69 (2001) 7224 – 7233.

[62] H. Ton-That, O. Schneewind, Mol. Microbiol. 50 (2003) 1429 – 1438. [63] R. Yanagawa, E. Honda, Infect. Immun. 13 (1976) 1293 – 1295. [64] E.P. Skaar, A.H. Gasper, O. Schneewind, J. Biol. Chem. 279 (2004)

436 – 443.

[65] D.A. Hodgson, Adv. Microb. Physiol. 42 (2000) 47 – 238.

[66] D. Claessen, R. Rink, W. de Jong, J. Siebring, P. de Vreugd, F.G. Boersma, L. Dijkhuizen, H.A. Wosten, Genes Dev. 17 (2003) 1714 – 1726.

[67] M.A. Elliot, N. Karoonuthaisiri, J. Huang, M.J. Bibb, S.N. Cohen, C.M. Kao, M.J. Buttner, Genes Dev. 17 (2003) 1727 – 1740. [68] A. Fouet, S. Mesnage, Curr. Top. Microbiol. Immunol. 271 (2002)

87 – 113.

[69] M.W. Pickett, N. Weiss, D.J. Kelly, FEMS Microbiol. Lett. 122 (1994) 7 – 12.

[70] A. Bouhss, N. Josseaume, A. Severin, K. Tabei, J.E. Hugonnet, D. Shlaes, D. Mengin-Lecreulx, J. Van Heijenoort, M. Arthur, J. Biol. Chem. 277 (2002) 45935 – 45941.

[71] V.M. de Ambrosini, S. Gonzalez, G. Perdigon, A.P. de Ruiz Holgado, G. Oliver, Chem. Pharm. Bull. (Tokyo) 44 (1996) 2263 – 2267. [72] H. Labischinski, H. Maidhof, in: J.-M. Ghuysen, R. Hakenbeck