Mycoplasma hyopneumoniae Surface Proteins Mhp385 and Mhp384Bind Host Cilia and Glycosaminoglycans and Are EndoproteolyticallyProcessed by Proteases That Recognize Different Cleavage MotifsAnia T. Deutscher,†,‡ Jessica L. Tacchi,§ F. Chris Minion,∥ Matthew P. Padula,§ Ben Crossett,⊥

Daniel R. Bogema,†,‡ Cheryl Jenkins,† Tracey A. Kuit,‡ Mark J. Walker,‡,# and Steven P. Djordjevic*,†,§

†NSW Department of Primary Industries, Elizabeth Macarthur Agricultural Institute, Camden NSW 2567, Australia‡School of Biological Sciences, University of Wollongong, Wollongong NSW 2522, Australia§The ithree Institute, University of Technology Sydney, Broadway NSW 2007, Australia∥Department of Veterinary Microbiology and Preventive Medicine, Iowa State University, Ames, Iowa 50011, United States⊥School of Molecular Bioscience, University of Sydney, Sydney NSW 2006, Australia#School of Chemistry and Molecular Biosciences and the Australian Infectious Diseases Research Centre, University of Queensland,Brisbane Qld 4072, Australia

*S Supporting Information

ABSTRACT: P97 and P102 paralogues occur as endoproteo-lytic cleavage fragments on the surface of Mycoplasmahyopneumoniae that bind glycosaminoglycans, plasminogen, andfibronectin and perform essential roles in colonization of ciliatedepithelia. We show that the P102 paralogue Mhp384 is efficientlycleaved at an S/T-X-F↓X-D/E-like site, creating P60384 andP50384. The P97 paralogue Mhp385 is inefficiently cleaved, withtryptic peptides from a 115 kDa protein (P115385) and 88 kDa(P88385) and 27 kDa (P27385) cleavage fragments identified byLC−MS/MS. This is the first time a preprotein belonging to theP97 and P102 paralogue families has been identified by massspectrometry. The semitryptic peptide 752IQFELEPISLNV763

denotes the C-terminus of P88385 and defines the novel cleavagesite 761L-N-V↓A-V-S766 in Mhp385. P115385, P88385, P27385, P60384, and P50384 were shown to reside extracellularly, though it isunknown how the fragments remain attached to the cell surface. Heparin- and cilium-binding sites were identified within P60384,P50384, and P88385. No primary function was attributed to P27385; however, this molecule contains four tandem R1 repeats withsimilarity to porcine collagen type VI (α3 chain). P97 and P102 paralogue families are adhesins targeted by several proteases withdifferent cleavage efficiencies, and this process generates combinatorial complexity on the surface of M. hyopneumoniae.

KEYWORDS: Mycoplasma hyopneumoniae, paralogue, endoproteolytic cleavage, cilium adhesin, heparin-binding protein

■ INTRODUCTIONM. hyopneumoniae is a genome-reduced, geographically wide-spread pathogen that causes porcine enzootic pneumonia (PEP), achronic, nonfatal respiratory disease that inflicts significanteconomic losses on swine production worldwide. The M. hyo-pneumoniae genome is approximately 900 kb (∼700 codingsequences) and has limited coding capacity compared to othermedically significant Gram-positive pathogens to which it isphylogenetically most closely related, such as Clostridium andStreptococcus.1 Mycoplasmas are thought to have evolved by aprocess of degenerative evolution from the low guanine andcytosine (G + C) Firmicutes.2 Genes that encode enzymes for cellwall biosynthesis, the tricarboxylic acid cycle, and many anabolicbiosynthetic pathways are absent from the genome. In M. hyo-pneumoniae, glycolysis and the associated “pyruvate roundhouse”,are the principal means by which adenosine triphosphate (ATP)

is derived.3,4 Analysis of theM. hyopneumoniae strain 232 genomeshowed that 182 of 692 (∼26%) open reading frames could beassigned to a paralogous gene family.1 Prominently featuredparalogue families include ABC transporters with ATP binding sitemotifs and the P97 and P102 paralogue families.Ciliostasis, cilial shedding, and epithelial cell death are

pathological hallmarks in the development of PEP.5−8

Adherence to target host cells is a critical, early step in patho-genesis and is a complex multifactorial process involvingadhesins that directly target host cell receptors, extracellularmatrix (ECM) components, and host molecules that act as abridge between microbe and host receptors.9−12 Members ofthe P97 and P102 paralogue families in M. hyopneumoniae play

Received: November 8, 2011Published: January 9, 2012

Article

pubs.acs.org/jpr

© 2012 American Chemical Society 1924 dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−1936

critical roles in binding porcine cilia, highly sulfated glycosamino-glycans (GAGs), fibronectin, and plasminogen.11,13−21 Interactionswith heparinase sensitive GAGs displayed on the surface ofporcine epithelial cilia provide a mechanism by which M. hyo-pneumoniae overcomes the mucociliary escalator during the earlystages of colonization.16 In addition, interactions with various hostmolecules enable the pathogen to remain within the host asinfection proceeds and epithelial shedding occurs.Reverse transcriptase−polymerase chain reaction (RT-PCR)

studies performed on M. hyopneumoniae taken from bron-choalveolar lavage fluid of experimentally M. hyopneumoniae-challenged pigs confirmed that mRNA encoding almost allmembers of the P97 and P102 paralogue families are simul-taneously expressed.22 Members of the P97 and P102 paraloguefamilies are targets for post-translational endoproteolyticcleavage by unknown proteases,13−15,17−19,21,23 a feature thatdistinguishes them from all other previously described genefamilies in the Mollicutes. All P97 and P102 paralogues are theproducts of gene fusion events.1 The processing of theseproteins generates endoproteolytic cleavage fragments compris-ing novel C-terminal sequences and N-terminal sequences thatshare sequence similarity to other paralogue family members.All of these fragments are displayed on the surface ofM. hyopneumoniae cells.14,15,17−19,23 Initial studies of thearchetype cilium adhesin P97 indicate that this moleculeundergoes extensive processing,23 but further work is requiredto understand this complexity. Recently, we showed that severalmembers of the P97 and P102 paralogue families possess asequence motif S/T-X-F↓X-D/E that delineates the site ofproteolytic cleavage.15 Cleavage occurs after the phenylalanine(F) residue. A negatively charged amino acid [glutamic acid (E)or aspartic acid (D)] typically occupies the +2 position, and analcohol-containing residue (S/T) is often found in the −3position. Cleavage efficiency seems to be influenced by thecleavage motif residing within disordered regions spanning atleast 40 amino acids.15 Our prevailing hypothesis is that a singleprotease is responsible for cleavage at the S/T-X-F↓X-D/E site.However, at least one other protease may play an importantrole in processing members of the P97 and P102 paraloguefamilies. In strain J, an inefficient cleavage event, whichseparates the C-terminal 230 amino acids that comprise P28from the P97 homologue MHJ_0194, occurs at the site 862T↓NTNTSKF869.23 This sequence is structurally unrelated to theone described above; therefore, its cleavage is thought to beperformed by a different protease.Our proteome studies indicate that most representatives of

the P97 and P102 paralogue families are very highly expressedduring broth culture. Given the limited capacity of M. hyo-pneumoniae to generate ATP, it is significant that considerableenergy is expended to express, secrete, and extensively pro-teolytically process members of the P97 and P102 paraloguefamilies. To gain a better understanding of the surfacetopography of M. hyopneumoniae, it is important to preciselymap the cleavage events in the P97 and P102 paralogue familiesand identify the proteases responsible for their cleavage. Herewe examine the expression patterns of Mhp385 and Mhp384,P97 and P102 paralogues, respectively, whose genes reside in atwo-gene operon. We show that Mhp384 undergoes one majorcleavage event at an S/T-X-F↓X-D/E-like motif while Mhp385undergoes inefficient endoproteolytic cleavage at a unique site.We precisely mapped cleavage sites in Mhp384 and Mhp385and examined the association between the location of pro-teolytic cleavage sites and protein disorder in these and other

P97 and P102 paralogues. Finally, we determined whetherthese proteins play a role in adherence to porcine cilia andGAGs.

■ EXPERIMENTAL PROCEDURESBacterial Strains

M. hyopneumoniae strain 232 cells were grown and harvested asdescribed previously.24 The Escherichia coli strains TOP10(Invitrogen, USA) and XL1-Blue (Stratagene, USA) were usedfor plasmid maintenance, while E. coli BL21 Star (DE3)(Invitrogen) was used for protein expression. E. coli wereroutinely grown at 37 °C on Luria−Bertani (LB) agar platesor cultured in LB broth with shaking at 200 rpm, supple-mented with antibiotics where appropriate.

Cloning of mhp385 and mhp384 Gene Fragments

Gene fragments were amplified from M. hyopneumoniae strain232 DNA, using the oligonucleotide primers (Sigma-Aldrich,Australia) listed in Supporting Information Table S1. TGAcodons encode tryptophan in M. hyopneumoniae; therefore, allin-frame TGA codons were changed to TGG by PCR usingmutated primers or site-directed mutagenesis.The mhp385 and mhp384 PCR amplified gene fragments

were cloned into the pET100/D-TOPO and pET160/GW/D-TOPO vectors (Invitrogen), respectively, and transformed intoE. coli TOP10 cells according to the manufacturer’s in-structions. For binding studies (see below), mhp384 genefragments were subcloned into pET100/D-TOPO. All clonedgene fragments were sequenced prior to performing expressionstudies.

Expression and Purification of Recombinant Fragments ofMhp385 and Mhp384

For protein expression, each plasmid construct was transformedinto E. coli BL21 Star (DE3) cells according to the manu-facturer’s instructions. Expression and purification of the re-combinant proteins was as described by Deutscher et al.,16 exceptE. coli BL21 Star (DE3) cells transformed with plasmid constructpET100/D-TOPO:f1384 or pET100/D-TOPO:f2385 were in-duced with 1 mM isopropyl-β-galactopyranoside (IPTG) over-night at 22 °C or for 8 h at 30 °C, respectively. Antiserum toeach of the F1384-F3384 (pET160/GW/D-TOPO) and F1385-F4385 (pET100/D-TOPO) fusion proteins was generated inNew Zealand White rabbits as detailed previously.20

Bioinformatics

To identify transmembrane domains, the Tied Mixture HiddenMarkov Model (TMHMM) Server v. 2.025,26 and Phobius,27 acombined transmembrane topology and signal peptide pre-dictor, were used. The ProtParam algorithm28 was used tocalculate the theoretical mass and pI of a protein. Basic LocalAlignment Search Tool (BLAST) searches, performed usingthe nonredundant protein database at the National Center forBiotechnology Information (NCBI), were used to identify pro-teins that share regions of identity or similarity.Protein disorder was predicted using the Predictor of Naturally

Disordered Regions (PONDR) algorithms VSL129,30 andPONDR-FIT,31 and the Intrinsically Unstructured Proteins Pre-dictor (IUPred) algorithm32,33 [Access to PONDR was providedby Molecular Kinetics (6201 La Pas Trail, Ste 160, Indianapolis,IN 46268; 317-280-8737; E-mail: main@molecularkinetics.com).PONDR is copyright 1999 by the Washington State UniversityResearch Foundation, all rights reserved]. Amino acid scoresabove 0.5 are considered to be disordered.

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361925

The search for coiled-coils was conducted by means of theCOILS2 algorithm34 with window widths of 14, 21, and 28 andthe MTIDK matrix. The outputs from both weighted andunweighted scans were compared. In addition, the Paircoil2algorithm,35 with a minimum window size of 21 and a P-scorecutoff of 0.025, was used.

Preparation of Whole Cell Lysates from M. hyopneumoniaeCells

Whole cell lysates from M. hyopneumoniae strain 232 cells wereprepared using two different methods. For the analysis ofM. hyopneumoniae proteome by “slice and dice” of a one-di-mensional (1D) sodium dodecyl sulfate (SDS)-polyacrylamidegel, or by two-dimensional (2D) SDS-polyacrylamide gel elec-trophoresis (SDS-PAGE), a 0.1 g pellet of M. hyopneumoniaestrain 232 cells was prepared as described by Seymour et al.14

For all other experiments, a 0.1 g pellet of M. hyopneumoniaestrain 232 cells was resuspended in 5 mL of 0.5% (v/v) TritonX-100 in phosphate-buffered saline (PBS).

Trypsin Digest of M. hyopneumoniae Cells

Trypsin digestion of M. hyopneumoniae cells and Western blot-ting were performed as described previously.16

Protein Separation by Polyacrylamide Gel Electrophoresisand Extraction

One-dimensional SDS-PAGE was performed as described byDeutscher et al.16 or carried out in 4−12% Bis-Tris gels using aCriterion XT gel system (Bio-Rad Laboratories, USA). For“slice and dice” experiments, whole cell lysates from M. hyo-pneumoniae strain 232 were separated by 1D SDS-PAGE, andthen either they were sliced equally into 15 pieces across themolecular mass range or a single slice was cut from the gel asindicated in the Results section. They were then destained andtrypsin digested as described by Seymour et al.17 PageRulerPrestained Protein Ladder (Thermo Fisher Scientific Inc., USA)or Unstained Precision Plus Protein Standards (Bio-RadLaboratories) were run on all polyacrylamide gels. Gels werestained with either GelCode Blue (Thermo Fisher Scientific Inc.),Coomassie Blue G-250, or Flamingo Fluorescent Gel Stain (Bio-Rad Laboratories).

2D SDS-PAGE and Western Blotting

Two-dimensional gel electrophoresis (2DGE) combined withWestern blotting was employed to identify the C-terminalcleavage fragment of Mhp385. M. hyopneumoniae strain232 cells whole cell extracts were passed through a Nanosep300K omega centrifugal filter (Pall, USA) prior to isoelectricfocusing (IEF). Bio-Lyte 3/10 ampholytes (Bio-Rad Laboratories)were added to a final concentration of 0.2% (v/v). The proteins(250 μg in 75 μL) were cup loaded onto an 11 cm pH 6−11Immobiline DryStrip (GE Healthcare, USA) according to manu-facturer’s instructions. IEF was carried out using a Protean IEFcell (Bio-Rad Laboratories) at a constant 20 °C and 50 μAcurrent limit per strip with a 3-step program: slow ramp to4000 V for 4 h, linear ramp to 10,000 V for 4 h, then 10,000 Vuntil 115 kVh was reached. The strips were equilibrated with5 mL of equilibration solution [6 M urea, 2% (w/v) SDS,250 mM Tris-HCl (pH 8.5), 0.0025% (w/v) bromophenolblue] for 20 min before 2DGE using the Bio-Rad LaboratoriesCriterion gel system.Two-dimensional polyacrylamide gels were visualized by

staining with Coomassie Blue G-250, or alternatively, proteinswere transferred to PVDF using a semidry transfer method.36

Membranes were blocked, probed, and developed as described

by Seymour et al.,14 using a 1:100 dilution of anti-F4385 serumand antirabbit peroxidase conjugate (Sigma-Aldrich) diluted1:3000.

1D Liquid Chromatography−Tandem Mass Spectrometry(LC−MS/MS) Using Q-TOF

LC−MS/MS was performed as described previously37 withminor modifications. The peptides were washed off the trap at300 nL/min onto a PicoFrit column (75 μm × 100 mm) (NewObjective, USA) packed with Magic C18AQ resin (MichromBioresources, USA), and the following program was used toelute peptides from the column and into the source of aQSTAR Elite hybrid Q-TOF mass spectrometer (AB Sciex,USA): 5−30% MS buffer B (98% acetonitrile−0.2% formicacid) over 30 min, 30−80% MS buffer B over 3 min, 80% MSbuffer B for 2 min, 80−85% MS buffer B for 3 min. The mass ofthe precursor peptide was then excluded for 120 s. Peak list fileswere generated by MSConvert from ProteoWizard38 release2.1.2674.

MS/MS Data Analysis

The MS/MS data files were searched using Mascot v.2.2.0739

through the UNified user InTErface (UNITE) interfaceprovided by the Australian Proteomics Computational Facility(APCF, http://www.apcf.edu.au/) against the LudwigNRdatabase (comprised of the UniProt, plasmoDB, and Ensembldatabases; http://www.ludwig.edu.au/archive/LudwigNR/LudwigNR.pdf). The MS/MS results from the M. hyopneumo-niae proteome “slice and dice” experiment, and thepolyacrylamide gel bands containing the Mhp385 preproteinand its N-terminal cleavage fragment, were searched againstversion Q111 with 14359729 sequences and 4875314780residues. The MS/MS results for the protein gel spots con-taining the C-terminal cleavage fragment of Mhp385 weresearched against version Q311 with 16981479 sequences and5747323621 residues. The MS/MS results for the polyacryla-mide gel bands containing the Mhp384 cleavage fragments weresearched against version Q211 with 12931687 sequences and4483500103 residues. The following parameter settings wereused in all searches: fixed modifications, none; variable modi-fications, propionamide, oxidized methionine, deamidatedasparagine; enzyme, semiTrypsin; number of allowed missedcleavages, 3; precursor ions mass tolerance, 100 ppm; fragmentions mass tolerance, 0.2 Da; charge state, 2+ and 3+. The resultsof the search were then filtered by including only protein hitswith at least one unique peptide and excluding peptide hits withP values of >0.05. Peptides identified by Mascot were furthervalidated by manual inspection of the MS/MS spectra for thepeptide to ensure that the b and y ion series were sufficientlyextensive for an accurate identification.

Cilia Adherence Assay

Cilia adherence assays and statistical analysis were performed asdescribed previously.16

Heparin-Binding Assays

Heparin microtiter plate binding assays and dot blots wereperformed and analyzed as described by Deutscher et al.,16 withminor modifications. Flat bottom Linbro/Titertek 96-wellmicrotiter plates (MP Biomedicals Inc., USA) were used formicrotiter plate binding assays, and dot blot membranes werewashed with TBS [10 mM Tris-HCl, 0.9% (w/v) NaCl, pH7.4]containing 0.05% (v/v) Tween 20 twice for 5 min, followed byone 5 min wash in TBS only.

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361926

■ RESULTS

Molecular Analysis of Mhp385 and Mhp384

mhp385 shares a two-gene operon with a gene encoding the P102paralogue Mhp384 (GenBank accession number AAV27855)(Figure 1A). A possible ribosome binding sequence (AGGAG) islocated within the twelve nucleotides that separate the genes.Mhp385 (GenBank accession number AAV27856) consists of988 amino acid residues and has a theoretical molecular weightand pI of 114.8 kDa and 8.87, respectively. A putativetransmembrane region resides between residues 22−44 (score0.99993, TMHMM 2.0). No signal peptide is predicted.A feature of the mhp385 gene locus is the presence of ahomopolymeric tract of 22 thymine residues located 100 basepairs upstream of the start codon of mhp385. This repeat regionis potentially placed between a Pribnow box (TATAAT) and apossible −35 region (TTGAGA). In addition, a homopoly-meric tract of 19 adenines is located 25 base pairs downstreamof the start codon for mhp385. Such regions are known hotspots for slipped-strand mispairing,40 which could result inframeshift mutations affecting the expression of mhp385. Thepresence of thymine and adenine tracts in very close proximityof each other may also lead to hairpin formation, which wouldinterfere with gene transcription.Mhp385 is one of three M. hyopneumoniae proteins that

contain tandem R1 pentapeptide repeats. Cilium-bindingoccurs in the presence of eight of these repeats.41 Mhp385contains four tandem R1 repeats (AAKV-AAKPS-AAKPV-AAKPE) within the C-terminus of the protein. The number of

tandem R1 repeats in both P97 and Mhp271 can vary betweenstrains of M. hyopneumoniae;16,41−44 however, alignment of theR1 region in Mhp385 with homologues in M. hyopneumoniaestrains J (GenBank accession number AAZ44457), 7448(AAZ53744), and 168 (ADQ90625) suggests that this regionis not highly variable (data not shown). On the basis of thenumber of repeats, this region is expected to be recognized by thecilia adherence-blocking monoclonal antibody (mAb) F2G5.41

The Mhp385 amino acid sequence shares greater than 96%identity with its respective homologues in M. hyopneumoniaestrains J, 7448, and 168, and it possesses a hexapeptide“NNNLRL” that is repeated twice in strain 232 but not in anyof its homologues. Outside the M. hyopneumoniae proteome,Mhp385 shows greatest similarity to LppS, a putative adhesin ofMycoplasma conjunctivae45 (GenBank accession numberCAC87274; the Mhp385 amino acids 39−399 share 30%identity and 50% positives; and the amino acids 360−735 share25% identity and 42% positives). The majority of the homologythat Mhp385 shares with P97 and its paralogues is within itsfirst 780 amino acids. Apart from the similarity of the R1 repeatregion with P97 and Mhp271, the C-terminal sequence ofMhp385 only shares some similarity with collagen alpha-3(VI)chain isoform 1 (Sus scrofa) (NCBI accession numberXP_001928122; 25% identity, 42% positives with Mhp385amino acids 723−930). The last 58 amino acids of Mhp385 donot share any significant similarity to proteins in the NCBIprotein database.

Figure 1. Cloning and expression of Mhp385 and Mhp384 recombinant proteins. (A) Schematic representation of mhp385 and mhp384 in thechromosome ofM. hyopneumoniae strain 232. The arrows show the direction of translation, and sizes are in base pairs. (B) Schematic of recombinantconstructs relative to the Mhp385 and Mhp384 proteins. Numbers refer to amino acid positions. * indicates the presence of a TGA codon. (C)GelCode Blue stained SDS−12% polyacrylamide gel of purified recombinant Mhp385 and Mhp384 6xHis-tagged fragments. Molecular weightmarkers are indicated on the left.

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361927

Mhp384 (GenBank accession number AAV27855) consistsof 957 amino acids and has a theoretical molecular weight of109.2 kDa and a pI of 8.97. Between residues 9−31 lies a putativetransmembrane region (score 0.99876, TMHMM 2.0), and nosignal peptide is predicted. The Mhp384 amino acid sequenceshares greater than 96% identity with its respective homologuesin M. hyopneumoniae strains J, 7448, and 168. Outside theM. hyopneumoniae proteome, Mhp384 shows greatest similarityto the M. conjunctivae putative adhesin LppT46 (GenBankaccession number CAC87275; the Mhp384 amino acid residues13−501 share 23% identity and 44% positives). The genes of theputative adhesins LppS and LppT reside in a two-geneoperon.45,46 The similarity Mhp384 shares with P102 and itsparalogues is within the first 324 amino acids, while the lastapproximately 450 amino acids of the protein do not sharesignificant sequence similarity to proteins in the NCBI proteindatabase. Both Mhp385 and Mhp384 protein sequences do notcontain any cysteine residues.

Cloning and Expression of Mhp385 and Mhp384Recombinant Protein Fragments

Lack of genet ic tools to mutate and transformM. hyopneumoniae and reliable growth on agar medium aremajor impediments to generating knockout mutants in thisspecies. To characterize the functions of Mhp385 and Mhp384,nonoverlapping fragments of the mhp385 and mhp384 geneswere cloned and expressed in E. coli. Due to the high adenineand thymine (A+T) content of mhp385 (72%) and mhp384(71%) and the potential difficulty in expressing A+T richgenes in E. coli,21,47 the entire genes were not cloned. Fragmentsthat span mhp385 and mhp384, apart from the transmem-brane domains (Figure 1B), were PCR amplified, cloned, andexpressed as N-terminal hexahistidine (6xHis)-tagged fusionproteins (F1385-F4385, F1384-F3384). One-dimensional SDS-PAGEanalysis of the 6xHis-tagged proteins (Figure 1C) confirmedpredicted molecular masses for F1385 (41.1 kDa); F2385 (32.4 kDa);F3385 (25.9 kDa); F4385 (26.9 kDa); F1384 (40.4 kDa); and F3384(38.7 kDa). F2384 traveled slightly more slowly than its theoreticalmass of 39.2 kDa, but this aberrant migration may be due to acluster of acidic and charged amino acids within its sequence(NEEVKQQESQVKDQAKQEKSSKDSQSKQ), since abnormalmigration occurs with proteins enriched in acidic residues.21,48

Expression of Mhp385 in M. hyopneumoniae

To detect the expression of Mhp385 byM. hyopneumoniae cells,whole cell lysates of M. hyopneumoniae strain 232 cells wereprobed with rabbit immune sera raised against each of theMhp385 recombinant fragments (Figure 2A). All anti-Mhp385sera detected a protein with an approximate molecular weightof 115 kDa (P115385), which corresponds to the theoreticalmolecular weight of the Mhp385 preprotein. Anti-F1385, anti-F2385, and anti-F3385 sera also detected an 88 kDa protein(P88385), while anti-F3385 and anti-F4385 sera also detected a27 kDa protein (P27385). P88385 and P27385 are most likelyN-terminal and C-terminal cleavage fragments of P115385, re-spectively. Since anti-F3385 serum detected all three bands, thecleavage site of Mhp385 is likely to fall within the sequence ofF3385. To confirm the identities of the 115 and 88 kDa proteins,the respective protein bands were cut from a 1D SDS-7% poly-acrylamide gel, digested with trypsin, and analyzed by LC−MS/MS(Figure 2B). Peptides unique to Mhp385 were found in boththe 115 kDa and 88 kDa bands (Figure 2D and SupportingInformation Tables S2, S3, and S4). Blots of a 2D gel (pI range6−11) of whole cell lysates from M. hyopneumoniae strain

232 cells probed with anti-F4385 serum identified two spots withmass of approximately 27 kDa (Figure 2C). The identities ofthese proteins were confirmed to be the C-terminal fragment ofMhp385 by LC−MS/MS (Figure 2D and SupportingInformation Tables S2, S5, and S6).Currently all cleavage sites in the P97 and P102 paralogue

families, except for the cleavage event that generates P28 fromthe C-terminus of the P97 homologue in strain J,23 adhere to themotif S/T-X-F↓X-D/E.15 Analysis ofM. hyopneumoniae strain 232proteins separated by 1D SDS-PAGE, extracted and charac-terized by LC−MS/MS, identified a semitryptic peptide withthe sequence 752IQFELEPISLNV763, matching to Mhp385(Figure 2D and Supporting Information Figure S1 and Table S7).This peptide falls within the 38 amino acids 752IQFELEPISLN-VAVSQEKTNPNNNLRLNNNLRLKYWYK789 that separatethe peptide matches that map to P88385 and P27385. Evidenceof an S/T-X-F↓X-D/E-like motif was not observed in thisstretch of amino acids. The semitryptic cleavage at the C-terminusof the peptide indicates that the cleavage site that separates P88385from P27385 in Mhp385 is 761L-N-V↓A-V-S766. This cleavage sitedoes not adhere to the previously established cleavage motif,suggesting that a different protease is responsible for this event.Consistent with this view, cleavage at this site is comparativelyinefficient because we identified the Mhp385 preprotein.

Expression of Mhp384 in M. hyopneumoniae

Whole cell lysates from M. hyopneumoniae strain 232 cellsprobed with anti-F1384 serum identified a 60 kDa protein(P60384), while anti-F3384 serum detected a 50 kDa protein(P50384) (Figure 3A). No evidence of the 109 kDa Mhp384preprotein was observed by probing blots with F1384-F3384antisera (data not shown), indicating that the single endo-proteolytic cleavage event that generates P60384 and P50384 isefficient. Since anti-F2384 serum detected both P60384 and P50384,the cleavage site of Mhp384 is likely to occur within the se-quence of this fragment. LC−MS/MS was performed on trypsindigests of proteins found within gel slices containing P60384 andP50384 (Figure 3B). Tryptic peptides from P60384 matched to theN-terminal sequence of Mhp384, while peptides from P50384matched to the C-terminal sequence of Mhp384 (Figure 3C andSupporting Information Tables S2, S8, and S9).Two semitryptic peptides, with sequences 530NEEVKQ-

QESQVK541 and 530NEEVKQQESQVKDQAK545, were identi-fied in a global analysis of the M. hyopneumoniae strain 232proteome by 1D SDS-PAGE followed by LC−MS/MS(Supporting Information Figure S2 and Table S7). These semi-tryptic peptides define the N-terminus of P50384 (Figure 3C).These data indicate that the cleavage site for Mhp384 spans thesequence 527I-L-F↓N-E-E532 with a phenylalanine residue (F) inthe −1 position of the cleavage site and a glutamic acid residue(E) in the position +2. This cleavage site is similar to thecleavage motif S/T-X-F↓X-D/E suggested by Bogema et al.15

Mapping the Cleavage Site in the P97 Paralogue Mhp107

To further refine the cleavage sites in the P97 and P102paralogue families, we also looked for the presence of asemitryptic peptide within another P97 paralogue, Mhp107.Previously we were unsuccessful in attempts to determine if themultifunctional adhesin Mhp107 undergoes endoproteolyticprocessing.18 Immunoblotting studies showed that Mhp107and homologues in different strains ofM. hyopneumoniae migratewith a mass of approximately 100 kDa, although the predictedmolecular weight of Mhp107 is approximately 120 kDa.18 LC−MS/MS analysis of M. hyopneumoniae strain 232 proteins

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361928

separated by 1D SDS-PAGE that migrated within the mass rangeof approximately 25−30 kDa identified tryptic peptides thatmatch to the N-terminal 199 amino acids of Mhp107 (Figure S3

and Supporting Information Tables S2 and S10). In this massrange a semitryptic peptide with the sequence 184QINENIL-NIGNFTTNF199 matched to Mhp107 (Supporting Information

Figure 2. Expression of Mhp385 byM. hyopneumoniae. (A) Whole cell lysates from M. hyopneumoniae strain 232 cells were separated on a 1D SDS−12% polyacrylamide gel and stained with GelCode Blue or transferred onto PVDF membrane. The membranes were incubated with preimmune (PI)or immune (I) antisera raised against each Mhp385 recombinant protein (F1385−F4385), followed by sheep antirabbit serum. Identified proteins areindicated on the right by arrowheads, and molecular weight markers are indicated on the left. (B) A 1D SDS−7% polyacrylamide gel of M.hyopneumoniae strain 232 cells whole cell lysates stained or transferred onto PVDF membrane and reacted with anti-F1385 serum. The arrowsindicate proteins recognized by the antiserum, which were excised, digested, with trypsin and subjected to LC−MS/MS. (C) Section of a 2DGE gel(Bio-Safe Coomassie stained) of whole cell lysates from M. hyopneumoniae strain 232 cells. A corresponding gel was transferred onto PVDF andprobed with anti-F4385 serum. Two prominent protein spots were detected (circled), excised from the gel, and subjected to trypsin digestionfollowed by LC−MS/MS. (D) Protein sequence of Mhp385. Peptides identified by LC−MS/MS of the 115 kDa band shown in panel B areunderlined. Highlighted in bold are the peptides identified by LC−MS/MS of the 88 kDa band in panel B. Peptides identified by LC−MS/MS ofboth protein spots boxed in panel C are in bold and italicized. Highlighted in bold and dash underlined is the Mhp385 semitryptic peptide identifiedin a global analysis of theM. hyopneumoniae strain 232 proteome by 1D SDS-PAGE followed by LC−MS/MS. The transmembrane domain is boxed.

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361929

Figure S4 and Table S7). The semitryptic cleavage at theC-terminus of this peptide indicates that the cleavage site is 197

T-N-F↓S-D-Q202. The cleavage site strongly resembles the cleavagemotif suggested by Bogema et al.15 Cleavage at this site wouldremove 199 amino acids from the N-terminus of Mhp107,generating a mature C-terminal cleavage fragment with a theoreticalmolecular weight of 96.4 kDa and a 23.4 kDa N-terminal cleavagefragment. The C-terminal cleavage protein has a theoreticalmolecular weight similar to that of the protein detected by Westernblotting with anti-Mhp107 sera in ref 18. The Mhp107 preproteinwas not identified in the whole proteome analysis, suggesting thatcleavage of Mhp107 is efficient.

Prediction of Regions of Disorder among the P97 and P102Paralogue Families

It was recently shown that the two TTKF↓QE cleavage sites inthe P102 paralogue Mhp683 reside within long (greater than 40amino acids) disordered regions.15 To determine if proteindisorder plays a role in the endoproteolytic cleavage of otherP97 and P102 paralogues, we applied several protein disorderprediction algorithms, as recommended by Ferron et al.,49 toMhp385, Mhp384, and Mhp107, and to MHJ_0194 andMHJ_0195, respectively the P97 and P102 homologues in

M. hyopneumoniae strain J, whose cleavage sites were previouslyidentified.13,23 Since several disorder predicting algorithmsinterpret coiled-coil regions as disordered,49 coiled-coil analyseswere also performed.The endoproteolytic cleavage sites in Mhp384 and in

MHJ_0195 occur within a long region of disorder (greater than50 amino acids long) (Supporting Information Figure S5B and E).However, Mhp107 is predicted not to be disordered at the site ofcleavage (Supporting Information Figure S5C). The endoproteo-lytic cleavage site in Mhp385 occurs prior to the start of a short,five amino acid long disordered region (Supporting InformationFigure S5A). Although PONDR VSL1 and PONDR-FITpredicted that the endoproteolytic cleavage site of MHJ_0194occurs within a disordered region (Supporting InformationFigure S5D), both Paircoil2 (amino acid residues 173−203; P-scoreless than 0.025) and COILS2 (residues 175−203; probability greaterthan 0.9 using weights and a 21-long window) identified a coiled-coildomain at the cleavage site, suggesting that this region is not likely tobe disordered.

Cellular Location of Mhp385 and Mhp384

Trypsin digestion of viable Mycoplasma cells has been widelyused to demonstrate the surface location of proteins.19,21,42

Figure 3. Expression of Mhp384 byM. hyopneumoniae. (A) Whole cell lysates from M. hyopneumoniae strain 232 cells were separated on a 1D SDS−12% polyacrylamide gel and stained with GelCode Blue or transferred onto a PVDF membrane. Membranes were treated with preimmune (PI) orimmune (I) antisera raised against each Mhp384 recombinant protein (F1384−F3384), followed by sheep antirabbit serum. Arrows point to cleavagefragments of Mhp384. Molecular weight markers are indicated on the left. (B) A 1D SDS−10% polyacrylamide gel of whole cell lysate ofM. hyopneumoniae strain 232 cells, GelCode Blue stained or transferred onto PVDF membrane and reacted with anti-F1384 and anti-F3384 sera. Thearrows indicate proteins recognized by the antisera that were excised, digested with trypsin, and subjected to LC−MS/MS. (C) Amino acid sequenceof Mhp384. Peptides identified by LC−MS/MS of the 60 kDa band shown in panel B are in bold. Highlighted in bold and underlined are thepeptides identified by LC−MS/MS of the 50 kDa band in panel B. Highlighted in bold and dash underlined is one of the two Mhp384 semitrypticpeptides identified in a global analysis of the M. hyopneumoniae strain 232 proteome by 1D SDS-PAGE followed by LC−MS/MS. Thetransmembrane domain is boxed.

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361930

Electrophoretic profiles of whole cell lysates fromM. hyopneumoniaestrain 232 cells, which were exposed to concentrations of trypsinranging from 0 to 300 μg/mL prior to 1D SDS-PAGE, showedgradual loss of proteins across the entire mass range, but particularlywith proteins with masses greater than 80 kDa (Figure 4A).

Immunoblots of lysates from trypsin digested M. hyopneumoniaecells showed that the Mhp385 preprotein (P115385) and theMhp385 and Mhp384 cleavage fragments (P88385, P27385, P60384,and P50384) were readily digested by trypsin (Figures 4B−E).P115385 and the C-terminal cleavage fragments P27385 and P50384were more susceptible to trypsin than the N-terminal cleavagefragments P88385 and P60384. Antiserum against the ribo-somal protein L7/L12 was used to control for cell lysis.21

L7/L12 protein was identified at trypsin concentrations up to

300 μg/mL, indicating that the cells remained intact during thedigestion procedure (Figure 4F).

Mhp385 and Mhp384 Bind Porcine Cilia

To investigate whether Mhp385 and Mhp384 have a role inadherence, the recombinant fragments were assessed for theirability to bind to purified swine tracheal cilia coated on a micro-titer plate. F1384, F2384, F3384, F1385, F2385, F3385, and positivecontrol F2P97, which includes the R1 cilium-binding domainand R2 repeat region from P97,20 reproducibly bound to cilia(P < 0.05) (Figure 5). F4385 did not bind cilia.

Mhp385 and Mhp384 Bind Heparin

P9720 and its paralogues Mhp493,19 Mhp271,16 and Mhp107,18

and the P102 paralogue Mhp68315 bind heparin. Since GAGssensitive to heparinase are found along the surface of porcinerespiratory cilia,50 our prevailing hypothesis is that members ofthe P97 and P102 paralogue families bind cilia during the earlystages of colonization by targeting proteoglycans decoratedwith highly sulfated heparin-like GAGs. To determine if re-combinant proteins spanning Mhp385 and Mhp384 bindheparin, an analogue for the highly sulfated iduronic acid-(1→4)-D-glucosamine disaccharide domains of heparan sul-fate,51 dot blots and microtiter plate binding studies were per-formed using biotinylated porcine heparin (Figure 6). F2385,F3385, F2384, and F3384 bound heparin in dot blot and the micro-titer plate binding assays (Figure 6A and B). F1384 displayedfaint reactivity in heparin-binding dot blot assays and lowabsorbance values in microtiter plate heparin-binding assays,indicating that it is at best a weak heparin-binding protein. F4385showed evidence of weak heparin-binding in the dot blot assaysand low absorbance values in the microtiter plate heparin-binding assays. In contrast, F1385 bound heparin in microtiterplate binding assays, but only faint reactivity was observed inheparin dot blot assays. Since the absorbances of F1384 and F4385in the heparin microtiter plate assays were low, indicating weakheparin-binding potential, they were not investigated further.

Figure 4. Trypsin sensitivity of Mhp385 and Mhp384 inM. hyopneumoniae cells. Whole cell lysates from freshly culturedM. hyopneumoniae strain 232 cells treated with a range of trypsinconcentrations (μg/mL), as listed at the top of the gel, were subjectedto SDS-PAGE and stained with GelCode Blue (A) or transferred ontoPVDF for immunoblot analyses. (B−E) Immunoblots probed withantiserum raised against F1385 (1:500), F4385 (1:1000), F1384 (1:500),and F3384 (1:500), respectively. (F) Control immunoblot probed withantiserum (1:5000) against cytosolic ribosomal protein L7/L12.21

Molecular weight markers are indicated on the left.

Figure 5. Adherence of Mhp385 and Mhp384 recombinant proteins toswine cilia. Mhp385 and Mhp384 recombinant protein fragments andthe positive control F2P97 protein

20 were incubated with purified swinetracheal cilia coated on a microtiter plate. Bound protein was detectedusing polyclonal antiserum raised against the respective recombinantprotein. Each bar represents the mean response above backgroundplus the upper standard error from triplicate wells from twoindependent experiments (n = 6). Bars marked with an asteriskrepresent results that are significantly different (P < 0.05) from thebackground.

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361931

Heparin-binding to F1385, F2385, F3385, F2384, and F3384 wasdose-dependent and saturating (Figure 6C). Unlabeled heparineffectively inhibited biotinylated heparin binding to F1385, F2385,F3385, F2384, and F3384, indicating that the interaction was specific(Figure 6C). From the specific heparin-binding curves, the

equilibrium binding dissociation constant (KD) for heparin toeach recombinant protein was calculated: F1385, 19 ± 2.6 nM;F2385, 2.7 ± 0.15 nM; F3385, 16 ± 1.7 nM; F2384, 26 ± 2.2 nM;and F3384, 8.9 ± 2.6 nM. To provide some indication of thetype of interaction occurring between the recombinants and

Figure 6. Recombinant protein fragments of Mhp385 and Mhp384 bind heparin. (A) Dot blot of F1−F4385 and F1−F3384 recombinant proteinsstained with amido black or probed with biotinylated heparin or streptavidin peroxidase only. (B) Binding curves of biotinylated heparin to F1−F4385and F1−F3384 recombinant proteins immobilized on a microtiter plate. (C) Specific biotinylated heparin-binding curves to F1385, F2385, F3385, F2384,and F3384 recombinant proteins. (D−H) Competitive-binding assays for F1385, F2385, F3385, F2384, and F3384, respectively, with biotinylated heparinand various unlabeled sulfated polysaccharides: heparin (■), fucoidan (▲), porcine mucin type II (●), chondroitin sulfate A (▼), and chondroitinsulfate B (◆). All data represent the mean and standard error of triplicate readings from two independent experiments (n = 6); except for thecompetition assays, a representative experiment of two independent experiments is shown (n = 3).

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361932

biotinylated heparin, competitive-binding assays were performedusing unlabeled heparin, fucoidan, chondroitin sulfates A and B,and porcine intestinal mucin (Figure 6D-H). Consistent withprevious studies,15,16 fucoidan, a highly sulfated polysaccharideextracted from seaweed, was most efficient at inhibiting theF2385, F2384, and F3384 interaction with heparin. At 600 μg/mL(30 times the concentration of heparin), porcine intestinalmucin reduced the adherence of heparin to F1385, F3385, F2384,and F3384 by at least 30%. Chondroitin sulfates A and B did notinhibit adherence of F2385, F2384, and F3384 to heparin, andchondroitin sulfate A was a minor inhibitor of heparin bindingto F1385 and F3385.

■ DISCUSSIONAdherence to host tissues is an early and essential step inpathogenesis and a complex molecular interplay betweenbacterial adhesins and host receptor molecules. Here weshow that the P97 and P102 paralogues, Mhp385 and Mhp384,respectively, are expressed during broth culture and undergoendoproteolytic cleavage, generating P88385 and P27385, andP60384 and P50384 on the surface of M. hyopneumoniae cells. Onthe basis of the binding abilities of the recombinant Mhp385and Mhp384 fragments expressed in this study, P88385 andP50384 bind heparin and porcine cilia. The sequences ofrecombinant F1384 and half of F2384 make up P60384; therefore,the location of the heparin-binding site(s) within F2384 willdetermine the affinity P60384 has for heparin. However, thiscleavage fragment has at least one cilium-binding site. P27385was only shown to be a weak heparin-binder. The affinity of therecombinant fragments F1385, F2385, F3385, F2384, and F3384 toheparin is similar to that reported for other M. hyopneumoniaeheparin-binding P97 and P102 paralogues and the Mycoplasmagallisepticum protein MG1142.15,16,18,19,52 The fact thatMhp385 and Mhp384 bind to cilia and heparin, combinedwith the presence of heparinase sensitive GAGs along the sur-face of porcine respiratory cilia,50 means that they are bothlikely to be important to the ability of M. hyopneumoniae tocolonize the porcine respiratory tract.The P97 and P102 paralogue families are targets of

endoproteolytic processing events.13−15,17−19,21,23 Prior to thisstudy, a P97 or P102 paralogue at a mass consistent with itsdeduced amino acid sequence had not been found by MS,indicating that, at least in vitro, cleavage is efficient and critical forprotein function. Recent studies comparing the endoproteolyticcleavage sites in MHJ_0194 (P97 homologue), MHJ_0195 (P102homologue), and the P97 and P102 paralogues Mhp493 (P216)and Mhp683 (P135), respectively, showed that most endoproteo-lytic cleavage events occur at a phenylalanine residue (−1 position)with an alcohol-containing residue in the −3 position and an acidic(D/E) residue in the +2 position.15 Cleavage at these S/T-X-F↓X-D/E sites is invariably efficient. Here, we show that cleavage sites inMhp384 and Mhp107 conform to the S/T-X-F↓X-D/E motif.LC−MS/MS revealed that both cleavage sites have a phenylalanineresidue at the −1 position and an acidic residue in the +2 positionand possess the sequences 527ILF↓NEE532 and 197TNF↓SDQ202,respectively. The location of the endoproteolytic cleavage site inMhp384 and MHJ_0195, like Mhp683 and Mhp493,15 waslocalized within a disordered region spanning more than 40 aminoacids. Our inability to find evidence of the Mhp384 preprotein byMS or Western blotting indicates that cleavage is efficient. Ourstudies show that the primary amino acid sequence and the degreeof protein disorder encompassing, or in close proximity to, the siteof endoproteolysis appear to influence protease cleavage behavior.

Disordered regions are flexible and therefore more accessible toproteases.53 Notably, however, the cleavage site in Mhp107 doesnot fall within a predicted region of disorder. The in-efficient MHJ_0194 cleavage site that removes the C-terminal 28kDa R2-containing region of the adhesin occurs within a coiled-coilregion. As further cleavage sites are mapped in sur-face proteins of M. hyopneumoniae, we will examine the associa-tion between protein disorder and endoproteolytic cleavage.A P97 paralogue that migrates at a mass equivalent to its

predicted mass (of 115 kDa) was identified here for the firsttime. Western blots of whole cell lysates from M. hyo-pneumoniae strain 232 cells probed with F1385−F4385 antiseraidentified a 115 kDa protein, and LC−MS/MS analysesgenerated peptide coverage spanning the length of P115385.These data confirm that a proportion of the Mhp385 pre-protein remains intact and that the cleavage event thatgenerates P88385 and P27385 is inefficient. Trypsin digestionstudies show P115385 is accessible on the cell surface ofM. hyopneumoniae. The cleavage site in Mhp385 that generatesP88385 and P27385 displays the sequence

761LNV↓AVS766, whichbears no sequence resemblance to the S/T-X-F↓X-D/Ecleavage motif within Mhp384 and other paralogues of P97and P102.15 This suggests that Mhp385 is processed by adifferent protease. The cleavage site 859ATNT↓NTNTGFS869

in the C-terminal region of MHJ_019423 also bears no se-quence identity to the S/T-X-F↓X-D/E consensus motif or tothe Mhp385 cleavage site sequence. Therefore, cleavage at thissite is likely to be performed by a different protease again. Thiscleavage event in MHJ_0194 is likely to be significant in abiological context because it results in the separation of the R1cilium-binding domain from the R2 repeat region and ef-fectively destroys a heparin-binding site.20 Collectively, our datasuggest that endoproteolytic processing of the P97 and P102paralogues involves at least three proteases.Both Mhp385 and Mhp384 contain several more S/T-X-F↓

X-D/E-like motifs. Further research is required to ascertain ifcleavage is occurring at these sites. The cleavage productsidentified in the current study were the strongly reactive proteinbands, other than Mhp385 preprotein (P115385), identified byimmunoblotting, whose identities were confirmed by LC−MS/MS.Immunoblots of whole cell lysates of M. hyopneumoniae probedwith anti-F4385 serum display a faint ladder suggesting furthercleavage of Mhp385, cross reactivity with other R1-containingM. hyopneumoniae proteins, or a combination of both. Futurestudies examining enriched fractions of the M. hyopneumoniaeproteome will assist in the identification and characterization offurther cleavage fragments. It is also possible that these cleavagesites are not accessible to the protease(s) due to protein folding orthe presence (or absence) of post-translational modificationssurrounding cleavage motifs. It is clear from data presented hereand with other P97 and P102 paralogue family mem-bers13−15,17−19,21,23 that processing is complex and strain-specific.Further research is needed to refine our understanding of surfaceadhesin processing and the precise role it plays in pathogenesis.F4385 does not bind cilia and only showed evidence of weak

heparin-binding. Apart from 20 amino acids missing at theN-terminus of the fragment, F4385 is identical to P27385.Therefore, in this study, we identify only weak heparin-bindingas a function of P27385; however, the weak binding suggests thatthis is not the main function of this cleavage fragment. Therelease of P27385 from the C-terminus of Mhp385 is likely to beimportant because the molecule contains a short R1-like motifcomprised of four tandem pentapeptide repeats. The cilium

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361933

adhesins Mhp183 (P97) and Mhp271 are the only othermolecules in the four sequenced strains of M. hyopneumoniae tocontain a R1 repeat region. Although the R1 repeat region inMhp385 does not contain a sufficient number of tandem R1repeats for cilium-binding, the P97 R1 region is an im-munogenic domain,54 and R1 repeat regions may also be asource of surface variation with the number of tandem repeatsin P97 and Mhp271 varying between field strains.16,42−44 Thefunction of the R1-like domain in Mhp385 is currently un-known, although it contains the minimum number of tandempentapeptide repeats sufficient to bind the mAb F2G5.41 Thisrepeat region also shares high similarity with porcine collagentype VI (α3 chain), which implicates that molecular mimicrymay be utilized as a means of avoiding host immune systemrecognition. P27385 remains associated with the surface ofM. hyopneumoniae cells despite the absence of anchorage motifsand transmembrane domains.Members of the P97 and P102 paralogue families have been

shown to be multifunctional proteins.14−21 We add Mhp385and Mhp384 to that list with their ability to bind cilia andGAGs. Gene families are common among the mycoplasmas,and many encode surface exposed lipoproteins that arepurported to play a role in immune evasion.55−57 Most lipo-protein family members are transcriptionally silent, with a singlelipoprotein family member dominating the cell surface architec-ture.58 Phase and antigenic size variation of M. hyopneumoniaelipoproteins have not been reported, and we have not identifiedevidence for such a mechanism from our proteome analyses. Onlya few genes in the M. hyopneumoniae genome display tandemrepeat sequences.1,59 Clearly, M. hyopneumoniae does not appearto employ the same evasion strategies as reported for many otherMycoplasma species. The surface architecture ofM. hyopneumoniaeis potentially quite variable, with P97, P102, and their paraloguesbeing the specific target of several proteases. Genes encodingmembers of the P97 and P102 paralogue families are derivedfrom gene fusions with unrelated sequences,1 and the C-terminalregions of most P97 and P102 paralogues are novel. Our cumulativedata suggest that processing of P97, P102, and their paraloguesgenerates N-terminal endoproteolytic fragments, that retain atransmembrane domain and remain tethered to the membrane,and C-terminal cleavage fragments, often novel in sequence, thatretain a presence on the surface of M. hyopneumoniae despitelacking known anchorage motifs. All members of the P97 and P102paralogue families characterized to date are targets of endoproteo-lytic cleavage generating fragments that play essential roles inbinding cilia, ECM molecules, and circulatory host proteins.Consequently, processing may represent a novel mechanism togenerate surface diversity. Further studies are needed to gain anunderstanding of the proteases involved in adhesin processing andthe circumstances that govern the efficiency of cleavage.

■ ASSOCIATED CONTENT

*S Supporting Information

Supplementary tables and figures. This material is available freeof charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATION

Corresponding Author

*Phone: +61 2 9514 4127. Fax: +61 2 9514 4143. E-mail:steven.djordjevic@uts.edu.au.

■ ACKNOWLEDGMENTS

A.T.D. is the recipient of an Australian Postgraduate Award.This work was partly funded by ARC-Linkage Grant LP776711and a grant from the McGarvie Smith Trust to S.P.D.

■ REFERENCES(1) Minion, F. C.; Lefkowitz, E. J.; Madsen, M. L.; Cleary, B. J.;Swartzell, S. M.; Mahairas, G. G. The genome sequence ofMycoplasmahyopneumoniae strain 232, the agent of swine mycoplasmosis.J. Bacteriol. 2004, 186 (21), 7123−33.(2) Dybvig, K.; Voelker, L. L. Molecular Biology of Mycoplasmas.Annu. Rev. Microbiol. 1996, 50, 25−57.(3) Pollack, J. D.; Williams, M. V.; McElhaney, R. N. Thecomparative metabolism of the Mollicutes (Mycoplasmas): the utilityfor taxonomic classification and the relationship of putative geneannotation and phylogeny to enzymatic function in the smallest free-living cells. Crit. Rev. Microbiol. 1997, 23 (4), 269−354.(4) Matic, J. N.; Wilton, J. L.; Towers, R. J.; Scarman, A. L.; Minion,F. C.; Walker, M. J.; Djordjevic, S. P. The pyruvate dehydrogenasecomplex of Mycoplasma hyopneumoniae contains a novel lipoyl domainarrangement. Gene 2003, 319, 99−106.(5) Underdahl, N. R.; Kennedy, G. A.; Ramos, A. S. Duration ofMycoplasma hyopneumoniae infection in gnotobiotic pigs. Can. Vet. J.1980, 21 (9), 258−261.(6) Livingston, C. W.; Stair, E. L.; Underdahl, N. R.; Mebus, A.Pathogenesis of mycoplasmal pneumonia of swine. Am. J. Vet. Res.1972, 33 (11), 2249−2258.(7) Mebus, A.; Underdahl, N. R. Scanning electron microscopy oftrachea and bronchi from gnotobiotic pigs inoculated with Mycoplasmahyopneumoniae. Am. J. Vet. Res. 1977, 38 (8), 1249−1254.(8) Tajima, M.; Yagihashi, T. Interaction of Mycoplasmahyopneumoniae with the porcine respiratory epithelium as observedby electron microscopy. Infect. Immunol. 1982, 37 (3), 1162−1169.(9) Duensing, T. D.; Wing, J. S.; van Putten, J. P. Sulfatedpolysaccharide-directed recruitment of mammalian host proteins: anovel strategy in microbial pathogenesis. Infect. Immunol. 1999, 67 (9),4463−4468.(10) Zhang, Q.; Young, T. F.; Ross, R. F. Microtiter plate adherenceassay and receptor analogs for Mycoplasma hyopneumoniae. Infect.Immunol. 1994, 62 (5), 1616−1622.(11) Zhang, Q.; Young, T. F.; Ross, R. F. Identification andcharacterization of a Mycoplasma hyopneumoniae adhesin. Infect.Immunol. 1995, 63 (3), 1013−1019.(12) DeBey, M. C.; Ross, R. F. Ciliostasis and loss of cilia induced byMycoplasma hyopneumoniae in porcine tracheal organ cultures. Infect.Immunol. 1994, 62 (11), 5312−5318.(13) Hsu, T.; Artiushin, S.; Minion, F. C. Cloning and functionalanalysis of the P97 swine cilium adhesin gene of Mycoplasmahyopneumoniae. J. Bacteriol. 1997, 179 (4), 1317−1323.(14) Seymour, L. M.; Jenkins, C.; Deutscher, A. T.; Raymond, B. B. A.;Padula, M. P.; Tacchi, J. L.; Bogema, D. R.; Eamens, G. J.; Woolley, K. L.;Dixon, N. E.; Walker, M. J.; Djordjevic, S. P. Mhp182 (P102) bindsfibronectin and contributes to the recruitment of plasmin(ogen) to theMycoplasma hyopneumoniae cell surface. Cell. Microbiol. 2012, 14 (1),81−94.(15) Bogema, D. R.; Scott, N. E.; Padula, M. P.; Tacchi, J. L.; Jenkins,C.; Cordwell, S. J.; Minion, F. C.; Walker, M. J.; Djordjevic, S. P. Thesequence TTKF↓QE defines the site of proteolytic cleavage inMhp683, a novel glycosaminoglycan and cilium adhesin ofMycoplasmahyopneumoniae. J. Biol. Chem. 2011, 286 (48), 41217−41229.(16) Deutscher, A. T.; Jenkins, C.; Minion, F. C.; Seymour, L. M.;Padula, M. P.; Dixon, N. E.; Walker, M. J.; Djordjevic, S. P. Repeatregions R1 and R2 in the P97 paralogue Mhp271 of Mycoplasmahyopneumoniae bind heparin, fibronectin and porcine cilia. Mol.Microbiol. 2010, 78 (2), 444−458.(17) Seymour, L. M.; Deutscher, A. T.; Jenkins, C.; Kuit, T. A.;Falconer, L.; Minion, F. C.; Crossett, B.; Padula, M.; Dixon, N. E.;Djordjevic, S. P.; Walker, M. J. A processed multidomain Mycoplasma

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361934

hyopneumoniae adhesin binds fibronectin, plasminogen, and swinerespiratory cilia. J. Biol. Chem. 2010, 285 (44), 33971−33978.(18) Seymour, L. M.; Falconer, L.; Deutscher, A. T.; Minion, F. C.;Padula, M. P.; Dixon, N. E.; Djordjevic, S. P.; Walker, M. J. Mhp107 isa member of the multifunctional adhesin family of Mycoplasmahyopneumoniae. J. Biol. Chem. 2011, 286 (12), 10097−10104.(19) Wilton, J. L.; Jenkins, C.; Cordwell, S. J.; Falconer, L.; Minion,F. C.; Oneal, D. C.; Djordjevic, M. A.; Connolly, A.; Barchia, I.;Walker, M. J.; Djordjevic, S. P. Mhp493 (P216) is a proteolyticallyprocessed, cilium and heparin binding protein of Mycoplasmahyopneumoniae. Mol. Microbiol. 2009, 71 (3), 566−582.(20) Jenkins, C.; Wilton, J. L.; Minion, F. C.; Falconer, L.; Walker,M. J.; Djordjevic, S. P. Two domains within Mycoplasmahyopneumoniae cilium adhesin bind heparin. Infect. Immunol. 2006,74 (1), 481−487.(21) Burnett, T. A.; Dinkla, K.; Rohde, M.; Chhatwal, G. S.; Uphoff,C.; Srivastava, M.; Cordwell, S. J.; Geary, S.; Liao, X.; Minion, F. C.;Walker, M. J.; Djordjevic, S. P. P159 is a proteolytically processed,surface adhesin of Mycoplasma hyopneumoniae: defined domains ofP159 bind heparin and promote adherence to eukaryote cells. Mol.Microbiol. 2006, 60 (3), 669−686.(22) Adams, C.; Pitzer, J.; Minion, F. C. In vivo expression analysis ofthe P97 and P102 paralog families of Mycoplasma hyopneumoniae.Infect. Immunol. 2005, 73 (11), 7784−7787.(23) Djordjevic, S. P.; Cordwell, S. J.; Djordjevic, M. A.; Wilton, J. L.;Minion, F. C. Proteolytic processing of the Mycoplasma hyopneumoniaecilium adhesin. Infect. Immunol. 2004, 72 (5), 2791−2802.(24) Scarman, A. L.; Chin, J. C.; Eamens, G. J.; Delaney, S. F.;Djordjevic, S. P. Identification of novel species-specific antigens ofMycoplasma hyopneumoniae by preparative SDS-PAGE ELISAprofiling. Microbiology 1997, 143 (2), 663−673.(25) Krogh, A.; Larsson, B.; von Heijne, G.; Sonnhammer, E. L. L.Predicting transmembrane protein topology with a hidden Markovmodel: Application to complete genomes. J. Mol. Biol. 2001, 305 (3),567−580.(26) Sonnhammer, E. L. L.; von Heijne, G.; Krogh, A. A hiddenMarkov model for predicting transmembrane helices in proteinsequences. In Proceedings of the Sixth International Conference onIntelligent Systems for Molecular Biology; Glasgow, J., Littlejohn, T.,Major, F., Lathrop, R., Sankoff, D., Sensen, C., Eds.; AAAI Press:Menlo Park, CA, 1998; Vol. 6, pp 175−182.(27) Kal̈l, L.; Krogh, A.; Sonnhammer, E. L. L. A combinedtransmembrane topology and signal peptide prediction method. J. Mol.Biol. 2004, 338 (5), 1027−1036.(28) Gasteiger, E.; Hoogland, C.; Gattiker, A.; Duvaud, S.; Wilkins,M. R.; Appel, R. D.; Bairoch, A. Protein identification and analysistools on the ExPASy Server. In The Proteomics Protocols Handbook;Walker, J. M., Ed.; Humana Press: Totowa, NJ, 2005; pp 571−607.(29) Peng, K.; Radivojac, P.; Vucetic, S.; Dunker, A. K.; Obradovic, Z.Length-dependent prediction of protein intrinsic disorder. BMC Bioinf.2006, 7, DOI: 10.1186/1471-2105-7-208.(30) Obradovic, Z.; Peng, K.; Vucetic, S.; Radivojac, P.; Dunker, A. K.Exploiting heterogeneous sequence properties improves prediction ofprotein disorder. Proteins: Struct., Funct., Bioinf. 2005, 61 (Suppl 7),176−182.(31) Xue, B.; Dunbrack, R. L.; Williams, R. W.; Dunker, A. K.; Uversky,V. N. PONDR-FIT: A meta-predictor of intrinsically disordered aminoacids. Biochim. Biophys. Acta, Proteins Proteomics 2010, 1804 (4), 996−1010.(32) Dosztanyi, Z.; Csizmok, V.; Tompa, P.; Simon, I. IUPred: webserver for the prediction of intrinsically unstructured regions ofproteins based on estimated energy content. Bioinformatics 2005, 21(16), 3433−3434.(33) Dosztanyi, Z.; Csizmok, V.; Tompa, P.; Simon, I. The pairwiseenergy content estimated from amino acid composition discriminatesbetween folded and intrinsically unstructured proteins. J. Mol. Biol.2005, 347 (4), 827−839.(34) Lupas, A.; Vandyke, M.; Stock, J. Predicting coiled coils fromprotein sequences. Science 1991, 252 (5009), 1162−1164.

(35) McDonnell, A. V.; Jiang, T.; Keating, A. E.; Berger, B. Paircoil2:improved prediction of coiled coils from sequence. Bioinformatics2006, 22 (3), 356−358.(36) Kyhse-Andersen, J. Electroblotting of multiple gels: a simpleapparatus without buffer tank for rapid transfer of proteins frompolyacrylamide to nitrocellulose. J. Biochem. Biophys. Methods 1984, 10(3−4), 203−209.(37) Szczepanek, S. M.; Frasca, S. Jr.; Schumacher, V. L.; Liao, X.;Padula, M.; Djordjevic, S. P.; Geary, S. J. Identification of lipoproteinMslA as a neoteric virulence factor of Mycoplasma gallisepticum. Infect.Immunol. 2010, 78 (8), 3475−3483.(38) Kessner, D.; Chambers, M.; Burke, R.; Agusand, D.; Mallick, P.ProteoWizard: open source software for rapid proteomics toolsdevelopment. Bioinformatics 2008, 24 (21), 2534−2536.(39) Perkins, D. N.; Pappin, D. J. C.; Creasy, D. M.; Cottrell, J. S.Probability-based protein identification by searching sequence data-bases using mass spectrometry data. Electrophoresis 1999, 20 (18),3551−3567.(40) Citti, C.; Browning, G. F.; Rosengarten, R. Phenotypic diversityand cell invasion in host subversion by pathogenic mycoplasmas. InMycoplasmasMolecular Biology, Pathogenicity and Strategies forControl; Blanchard, A., Browning, G. F., Eds.; Horizon Bioscience:Norfolk, 2005; pp 439−484.(41) Minion, F. C.; Adams, C.; Hsu, T. R1 region of P97 mediatesadherence of Mycoplasma hyopneumoniae to swine cilia. Infect.Immunol. 2000, 68 (5), 3056−3060.(42) Wilton, J. L.; Scarman, A. L.; Walker, M. J.; Djordjevic, S. P.Reiterated repeat region variability in the ciliary adhesin gene ofMycoplasma hyopneumoniae. Microbiology 1998, 144 (7), 1931−1943.(43) Stakenborg, T.; Vicca, J.; Maes, D.; Peeters, J.; de Kruif, A.;Haesebrouck, F.; Butaye, P. Comparison of molecular techniques forthe typing of Mycoplasma hyopneumoniae isolates. J. Microbiol. Methods2006, 66 (2), 263−275.(44) de Castro, L. A.; Pedroso, T. R.; Kuchiishi, S. S.; Ramenzoni,M.; Kich, J. D.; Zaha, A.; Vainstein, M. H.; Ferreira, H. B. Variablenumber of tandem amino acid repeats in adhesion-related CDSproducts in Mycoplasma hyopneumoniae strains. Vet. Microbiol. 2006,116 (4), 258−269.(45) Belloy, L.; Vilei, E. M.; Giacometti, M.; Frey, J. Characterizationof LppS, an adhesin of Mycoplasma conjunctivae. Microbiology 2003,149, 185−193.(46) Zimmermann, L.; Peterhans, E.; Frey, J. RGD motif oflipoprotein T, involved in adhesion ofMycoplasma conjunctivae to lambsynovial tissue cells. J. Bacteriol. 2010, 192 (14), 3773−3779.(47) Notarnicola, S. M.; McIntosh, M. A.; Wise, K. S. Multipletranslational products from a Mycoplasma hyorhinis gene expressed inEscherichia coli. J. Bacteriol. 1990, 172 (6), 2986−2995.(48) Li, Y. Y.; Martinez, G.; Gottschalk, M.; Lacouture, S.; Willson,P.; Dubreuil, J. D.; Jacques, M.; Harel, J. Identification of a surfaceprotein of Streptococcus suis and evaluation of its immunogenic andprotective capacity in pigs. Infect. Immunol. 2006, 74 (1), 305−312.(49) Ferron, F.; Longhi, S.; Canard, B.; Karlin, D. A practicaloverview of protein disorder prediction methods. Proteins: Struct.,Funct., Bioinf. 2006, 65 (1), 1−14.(50) Erlinger, R. Glycosaminoglycans in porcine lungan ultra-structural study using cupromeronic blue. Cell Tissue Res. 1995, 281(3), 473−483.(51) Rabenstein, D. L. Heparin and heparan sulfate: structure andfunction. Nat. Prod. Rep. 2002, 19 (3), 312−331.(52) Jenkins, C.; Geary, S. J.; Gladd, M.; Djordjevic, S. P. TheMycoplasma gallisepticum OsmC-like protein MG1142 resides on thecell surface and binds heparin. Microbiology 2007, 153, 1455−1463.(53) Uversky, V. N.; Dunker, A. K. Understanding protein non-folding. Biochim. Biophys. Acta, Proteins Proteomics 2010, 1804 (6),1231−1264.(54) Chen, J. R.; Liao, C. W.; Mao, S. J. T.; Weng, C. N. Arecombinant chimera composed of repeat region RR1 of Mycoplasmahyopneumoniae adhesin with Pseudomonas exotoxin: in vivo evaluation

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361935

of specific IgG response in mice and pigs. Vet. Microbiol. 2001, 80 (4),347−357.(55) Citti, C.; Watson-McKown, R.; Droesse, M.; Wise, K. S. Genefamilies encoding phase- and size-variable surface lipoproteins ofMycoplasma hyorhinis. J. Bacteriol. 2000, 182 (5), 1356−1363.(56) Markham, P. F.; Duffy, M. F.; Glew, M. D.; Browning, G. F.A gene family in Mycoplasma imitans closely related to the pMGAfamily of Mycoplasma gallisepticum. Microbiology 1999, 145, 2095−2103.(57) Rosengarten, R.; Wise, K. S. Phenotypic switching inmycoplasmasphase variation of diverse surface lipoproteins. Science1990, 247 (4940), 315−318.(58) Glew, M. D.; Markham, P. F.; Browning, G. F.; Walker, I. D.Expression studies on four members of the pMGA multigene family inMycoplasma gallisepticum S6. Microbiology 1995, 141, 3005−3014.(59) Vasconcelos, A. T. R.; Ferreira, H. B.; Bizarro, C. V.; Bonatto,S. L.; Carvalho, M. O.; Pinto, P. M.; Almeida, D. F.; Almeida, L. G. P.;Almeida, R.; Alves, L.; Assuncao, E. N.; Azevedo, V. A. C.; Bogo, M. R.;Brigido, M. M.; Brocchi, M.; Burity, H. A.; Camargo, A. A.; Camargo,S. S.; Carepo, M. S.; Carraro, D. M.; Cascardo, J. C. D.; Castro, L. A.;Cavalcanti, G.; Chemale, G.; Collevatti, R. G.; Cunha, C. W.;Dallagiovanna, B.; Dambros, B. P.; Dellagostin, O. A.; Falcao, C.;Fantinatti-Garboggini, F.; Felipe, M. S. S.; Fiorentin, L.; Franco, G. R.;Freitas, N. S. A.; Frias, D.; Grangeiro, T. B.; Grisard, E. C.; Guimares,C. T.; Hungria, M.; Jardim, S. N.; Krieger, M. A.; Laurino, J. P.; Lima,L. F. A.; Lopes, M. I.; Loreto, E. L. S.; Madeira, H. M. F.; Manfio, G. P.;Maranhao, A. Q.; Martinkovics, C. T.; Medeiros, S. R. B.; Moreira,M. A. M.; Neiva, M.; Ramalho-Neto, C. E.; Nicolas, M. F.; Oliveira, S. C.;Paixao, R. F. C.; Pedrosa, F. O.; Pena, S. D. J.; Pereira, M.; Pereira-Ferrari,L.; Piffer, I.; Pinto, L. S.; Potrich, D. P.; Salim, A. C. M.; Santos, F. R.;Schmitt, R.; Schneider, M. P. C; Schrank, A.; Schrank, I. S.; Schuck, A. F.;Seuanez, H. N.; Silva, D. W.; Silva, R.; Silva, S. C.; Soares, C. M. A.; Souza,K. R. L.; Souza, R. C.; Staats, C. C.; Steffens, M. B. R.; Teixeira, S. M. R.;Urmenyi, T. P.; Vainstein, M. H.; Zuccherato, L. W.; Simpson, A. J. G.;Zaha, A. Swine and poultry pathogens: the complete genome sequences oftwo strains of Mycoplasma hyopneumoniae and a strain of Mycoplasmasynoviae. J. Bacteriol. 2005, 187 (16), 5568−5577.

Journal of Proteome Research Article

dx.doi.org/10.1021/pr201115v | J. Proteome Res. 2012, 11, 1924−19361936