Cell surface processing of the P1 adhesin of Mycoplasma pneumoniae identifies novel domains that bind host molecules

Mycoplasma pneumoniae is a genome reduced pathogen and causative agent of community acquired pneumonia. The major cellular adhesin, P1, localises to the tip of the attachment organelle forming a complex with P40 and P90, two cleavage fragments derived by processing Mpn142, and other molecules with adhesive and mobility functions. LC-MS/MS analysis of M. pneumoniae M129 proteins derived from whole cell lysates and eluents from affinity matrices coupled with chemically diverse host molecules identified 22 proteoforms of P1. Terminomics was used to characterise 17 cleavage events many of which were independently verified by the identification of semi-tryptic peptides in our proteome studies and by immunoblotting. One cleavage event released 1597TSAAKPGAPRPPVPPKPGAPKPPVQPPKKPA1627 from the C-terminus of P1 and this peptide was shown to bind to a range of host molecules. A smaller synthetic peptide comprising the C-terminal 15 amino acids, 1613PGAPKPPVQPPKKPA1627, selectively bound cytoskeletal intermediate filament proteins cytokeratin 7, cytokeratin 8, cytokeratin 18, and vimentin from a native A549 cell lysate. Collectively, our data suggests that ectodomain shedding occurs on the surface of M. pneumoniae where it may alter the functional diversity of P1, Mpn142 and other surface proteins such as elongation factor Tu via a mechanism similar to that described in Mycoplasma hyopneumoniae.

The gene mpn141 encoding the major adhesin P1 is located in the same operon along with mpn140 and mpn142 and these three genes constitute a polycistronic transcriptional unit 22,23 . mpn140 encodes for a 28 kDa putative phosphoesterase 24 and while it has been shown to degrade nanoRNA and dephosphorylate 3′-phosphoadenosine 5′-phosphate to AMP 25 , no role in adherence has been assigned for this protein. mpn142 generates a 130 kDa product (Mpn142) that is cleaved into two fragments of 40 kDa (P40) and 90 kDa (P90) immediately after or concurrent with translation 26,27 . The cleavage event in Mpn142, first described over 25 years ago, was the first in what is now known to be a highly processed molecule on the surface of M. pneumoniae 28 . P1 is a remarkably versatile molecule and the subject of numerous studies over the past 30 years. The only cleavage event that has been accurately assigned to P1 is the removal of the N-terminal 59 amino acids as a leader peptide 29 . Molecular cross-linking and immunogold-labelling studies indicated that P1 forms a complex with P30, P40, and P90 30,31 that colocalise to the tip of the attachment organelle to act in concert to effect different functions 5,6,23,32 . Cross-linking studies with paraformaldehyde identified P1 complexes containing Mpn309 (P65), Mpn272 (DnaK), C-terminal truncated forms of DnaK and P1, pyruvate dehydrogenase α subunit (Pdh-A), and ication in 7 M urea, 2 M thiourea, 40 mM Tris-HCl, and 1% (w/v) C7BzO detergent (Sigma) after washing with PBS. Proteins were reduced and alkylated with 5 mM tributylphosphine and 20 mM acrylamide monomers before precipitation with acetone. Protein was resuspended in 7 M urea, 2 M thiourea, and 1% (w/v) C7BzO for 1D-and 2D-SDS PAGE.
Gel electrophoresis was performed as described previously 60,61 . Approximately 80 µg and 250 µg of protein was used for 1D-and 2D-SDS PAGE, respectively. Gels were fixed and stained by either Flamingo fluorescent gel stain (Bio-Rad) or Coomassie Blue G-250 (Sigma).
In-gel trypsin digestion was performed as described previously 62 for mass spectrometry analysis. Gel pieces were excised, destained, dehydrated, and then incubated with trypsin Gold MS grade (Promega) in 100 mM NH 4 HCO 3 . Tryptic peptides were extracted by sonication and stored in 4 °C until needed for mass spectrometry.

Liquid chromatography tandem mass spectrometry (LC-MS/MS) and data analysis. LC-MS/
MS was performed as described previously 61 . In brief, 5 μg of peptides in 15 μl was loaded into an Eksigent AS-1 autosampler connected to a Tempo nanoLC system (Eksigent, Livermore, CA, USA) and washed onto a PicoFrit column (75 μm × 150 mm) packed with Magic C18AQ resin (Michrom Biosciences, CA). Peptides were eluted from the column into the source of a QSTAR Elite hybrid quadrupole-time-of-flight mass spectrometer (Sciex, Redwood, CA, USA).
Surface proteome analysis of M. pneumoniae. Biotinylation of the M. pneumoniae cells was performed as described previously 28 . The biotinylation reaction was allowed to proceed for 30 seconds on ice. Biotinylated surface proteins were confirmed with western blots using ExtrAvidin-HRP (Sigma).
Trypsin shaving of M. pneumoniae cells was carried out as described previously 12 . Shaving was for 5 minutes at 37 °C and released peptides were trypsin digested a second time before analysis by LC-MS/MS. Immunoblot of M. pneumoniae cell lysates using Anti-P1 serum. 60 µg of M. pneumoniae cell lysate proteins were separated on 1D-SDS PAGE as described above. Proteins were transferred to PVDF (polyvinylidene fluoride) membranes using a semidry method 71 . Membranes were blocked with 5% (w/v) skim milk powder in PBS, and 0.1% (v/v) Tween 20 (PBS-Tween) for 1 hour at 25 °C. Membranes were cut in to individual lanes and then separately probed with guinea pig sera raised against different regions of the P1 adhesin (guinea pig sera was generated in a previous study 55 ) for 1.5 hours at 25 °C in PBS-Tween. Membranes were washed three times over 30 minutes before being probed a second time in peroxidase-conjugated anti-guinea pig antibodies (1:3000, Sigma) for 1 hour at 25 °C in PBS-Tween. Membranes were washed again three times over 30 minutes and developed with DAB tablets (3,3′-Diaminobenzidine, Sigma).
Binding affinity measured by ELISA was performed as described previously 17 . Recombinant protein RP15 was produced as described 55 and both C-terminal peptides were synthesised by Chempeptide Limited (China). P1-30 ( 1597 TSAAKPGAPRPPVPPKPGAPKPPVQPPKKPA 1627 ) without any tags, but P1-15 ( 1613 PGAPKPPVQPPKKPA 1627 ) was sequenced with an N-terminal biotin tag.
Binding of the P1 C-terminus to A549 human lung cells. Freshly grown A549 cells were immobilised in 96-well microtitre plates as described in 17 . Immobilised A549 cells were incubated with 10 µg/ml of either RP15, P1-30, or P1-15 and binding affinity was measured with antiserum raised against RP15 (1:100) as described above. Absorbance detection at 450 nm is the same as described above.
Affinity chromatography of complexes that bind the P1 C-terminus. The C-terminal sequence of P1 (P1-15) was synthesised with an N-terminal biotin tag by Chempeptide Limited (China). Affinity chromatography was performed similar to the section above. In brief, 1 mg of the peptide was added to Avidin Agarose beads for 16 h at 4 °C. The beads were washed four times (5 ml per wash) with PBS before being incubated with native A549 cell lysates (harvested in 1% w/v C7BzO in PBS) for 16 h at 4 °C. Non-binding proteins were washed from  55 are shown. Predicted disordered regions appear as purple boxes in the grey bar. Acidic and basic regions within P1 are identified as yellow and blue bars, respectively. Peptides released from surface shaving experiments and identified by mass spectrometry are shown in the light green boxes within the grey bar. Grey bars represent fragments of P1 identified during SDS-PAGE of whole cell lysates. Red bars represent fragments of P1 recovered from lysates of M. pneumoniae that have their surface proteins labelled with biotin (surface exposed fragments of P1). Peptides identified by mass spectrometry of P1 Ethical approval. Guinea pig sera used in this study was generated in a previous study 55 . The animal experiments in that previous study were proved by the ethical board of Landesdirektion Sachsen, Dresden, Germany (no. 24-9168.25-1).

Results
Bioinformatic analysis of the P1 adhesin. The P1 adhesin has a predicted mass of 176.3 kDa and a pI of 8.53 and contains six predicted transmembrane regions and nine putative glycosaminoglycan binding sites (Fig. 1). The first transmembrane region (spanning the N-terminus), and the last transmembrane region (spanning the C-terminus) have been identified in previous studies of P1 32,36,66 , and a P1 paralog of Mycoplasma genitalium 73 . The glycosaminoglycan binding sites consist of reiterated copies of positively charged amino acids that are likely to be important in interactions with sulphated derivatives of heparin and heparan sulfate. Analysis of P1 using PONDR ® identified seven putative disordered regions that span at least 30 amino acids (Fig. 1). Modules in P1 enriched in acidic (E, K) and basic (K, R, H) amino acids were identified. Disordered region and protein modules enriched in acidic and basic amino acids have been described in adhesin families in the respiratory pathogen M. hyopneumoniae and these were influential in the location of a subset of important cleavage sites [60][61][62]74 . We confirmed the precise location of 17 cleavage sites in P1 (shown below), 11 of which reside in predicted regions of disorder ( Fig. 1). Cleavage sites did not seem to be over-represented in acidic or basic domains.
The P1 adhesin is processed extensively on the M. pneumoniae cell surface. P1 peptides identified by LC-MS/MS analyses of size fractionated M. pneumoniae lysates identified 23 proteoforms ranging in size from 17 to 176 kDa including the full length proteoform without the N-terminal signal sequence (Fig. 1). The full length and an additional 16 smaller proteoforms of P1 were identified by LC-MS/MS of size fractionated cell lysates separated by SDS-PAGE (grey bars; Fig. 1). The migration behaviour of these 17 proteoforms of P1 was consistent with masses predicted by ProtParam 65 . Trypsin shaving of the M. pneumoniae cell surface released trypsin accessible peptides (green boxes within a grey bar in Fig. 1) that span most of the adhesin indicating that P1 is exposed on the cell surface. This was consistent with LC-MS/MS analysis of size-fractionated biotinylated proteins that were first enriched using avidin chromatography which identified 14 proteoforms (full and fragments 2, 3, 5, 7, 10, 11, 13, 14, 16, 17, 20, 21, and 22) of P1 (red bars in Fig. 1). These data suggest that cleaved P1 proteoforms are surface accessible.
A global M. pneumoniae dimethyl labelling approach was used to identify internal neo-N termini. Ten cleavage sites were identified in P1 using this approach (Table 1, blue arrows in Fig. 1). Semi-tryptic peptides, defined as peptides with only one tryptic end (Table 1, red arrows in Fig. 1) were also identified, implying seven additional cleavage sites in P1. Four distinct sites in P1 showed evidence that surface accessible amino-peptidases may alter neo-N-termini ( Fig. 1 Table 1; sequence: 1341 STS↓D↓G↓N↓T↓S↓S↓T↓N↓N↓L↓A↓P↓N↓T↓N ↓T↓G↓NDV 1363 ).

Functional analysis of the C-terminal tail of P1.
Dimethyl labelling data indicated that the carboxy-terminal 30 residues of P1 is released by a cleavage event at serine 1598 (cleavage site 17 in Table 1, sequence: 1595 KQT↓SAA 1600 ). The C-terminal peptide has a composition comprising five alanine, five lysine, and thirteen proline residues. This C-terminal region also shares sequence identity (53.1%) with the carboxy-terminal 31 residues of Mpn142. Furthermore, the final 15 residues of P1 shares 73.3% sequence identity with the last 14 residues of Mpn142 (11 identical positions). The C-terminal 30 amino acids (named P1-30: 1597 TSAAKPGAPRPPVPPKPGAPKPPVQPPKKPA 1627 ), and the C-terminal 15 amino acids (named P1-15 1613 PGAPKPPVQPPKKPA 1627 ) were synthesised chemically (Table 2; Chempeptide Limited, China) and an N-terminal biotin tag was added to the P1-15 peptide. Microtitre binding assays revealed that P1-15, P1-30, and the recombinant protein, RP15 55 , bind a range of host molecules in a dose dependent manner (Fig. 3). M. pneumoniae cells and RP15 bound lactoferrin, vitronectin, plasminogen, fibronectin, and fibrinogen. Only M. pneumoniae cells bound laminin. P1-30 bound fibronectin, fibrinogen and plasminogen in a dose dependent manner but failed to bind laminin. P1-15 only bound plasminogen in a dose dependent manner but also bound to vitronectin but failed to bind laminin, lactoferrin, fibronectin, and fibrinogen (Fig. 3). Compared with P1-30 and P1-15, the C-terminal 106 amino acids of P1 represented by RP15 consistently showed the most consistent and most diverse binding capabilities for the panel of host proteins tested here suggesting that multiple binding domains increase the binding capabilities of P1 proteoforms. Consistent with this hypothesis, RP15 spans two putative glycosaminoglycan binding motifs (underlined motifs in Table 2) that are absent in P1-30 and P1-15.
To investigate whether binding was due to the specific amino acid sequence or to amino acid composition, microscale thermophoresis was performed on P1-30 and a scrambled version of P1-30 (PKPPRAAPPKAPTPVPPGPASPVKKPKQAPG). P1-30 had a medium binding affinity for plasminogen (K D = 554 ± 2.1 nM) and a medium/low binding affinity for fetuin (K D = 2.4 ± 0.7 μM). No binding affinity could be detected for the scrambled peptide (Fig. 4).
Microtitre binding assays were also employed to determine the binding capabilities of regions spanning the C-terminus of P1 to A549 human epithelial cells (Fig. 5). Recombinant pyruvate dehydrogenase subunit B of M. pneumoniae (rPdhB; positive control 16 ) and RP15 bound immobilised A549 cells, but not P1-30. We were not able to determine if P1-15 bound using this assay because we lacked reagents that could detect this peptide.
To overcome this experimental limitation and to attempt to identify potential binding partners for P1-15, we designed an affinity bait-prey experiment. The biotinylated P1-15 was coupled to avidin agarose and, in parallel with uncoupled avidin-agarose (negative control), were exposed to a native A549 cell lysate as described in Methods, washed and eluants were characterised by SDS-PAGE and LC-MS/MS (Fig. S1). Three protein bands identified in eluents from avidin-agarose coupled with biotinylated P1-15 that were absent in the control were analysed by LC-MS/MS (Fig. S1) (Fig. S1). Tryptic peptides to these filament proteins were not identified in the control experiment. Tryptic peptides identified in slice 3 identified glyceraldehyde-3-phosphate dehydrogenase, however, this protein was also identified in the eluents from the control and was not considered further as a potential binding partner with P1-15.
The P1 adhesin and proteins it associates with at the tip of the attachment organelle are central to binding interactions that enable M. pneumoniae to target host cell receptors and is likely to contain binding domains  www.nature.com/scientificreports www.nature.com/scientificreports/ for some or all of these host molecules. Here we show that Mpn141 is processed extensively generating 23 proteoforms and that many proteoforms are retained on affinity matrices loaded with different host molecules and mimics of regions of host proteins including fetuin, fibronectin, actin, heparin, and plasminogen. Microtitre plate binding assays and microscale thermophoresis assays confirmed several of these preliminary findings and showed that the C-terminal region of P1 binds vitronectin, fibrinogen and fibronectin. Apart from removal of a 59 amino acid N-terminal leader peptide, only a ~40 kDa carboxyl terminal truncated fragment of P1 (potentially representing fragment 18 from this study), that forms a complex with full length P1 protein, and other accessory proteins has been reported previously 33 but earlier immunoblotting studies with anti-P1 monospecific antisera identified numerous smaller proteoforms of P1 that were not characterised 2 . Dimethyl labelling experiments enabled us to map the precise location of cleavage events in P1 (Table 1). P1 proteoforms are likely generated by proteases on the cell surface of M. pneumoniae or associated with the protein translocation machinery but their identities have not been confirmed. Biotinylation studies identified 13 proteoforms of P1 that were accessible on the surface of M. pneumoniae and our surface labelling and trypsin shaving experiments indicate that the www.nature.com/scientificreports www.nature.com/scientificreports/ proteoforms remain attached to the extracellular side of M. pneumoniae cell membranes. Our data is consistent with electron micrographs of M. pneumoniae immunostained with ferretin-labelled anti-P1 antibodies that depict gold particles at: i) the tip of the attachment organelle; ii) along the shaft of this structure; iii) at sites along the cell body; and iv) at sites distant from the M. pneumoniae membrane 2 . It is not known if some proteoforms are excreted into the extracellular milieu but it is conceivable that processing of P1 occurs after translocation and the fragments may remain anchored to the surface via the predicted C-terminal transmembrane domain similarly seen in P40 and P90 of M. pneumoniae 27 . Consistent with this view, we were unable to find tryptic peptides that mapped to the putative leader peptide residing in the N-terminus of P1 or in the bioinformatically predicted transmembrane domains, or the well characterised C-terminal transmembrane domain. However, we did find tryptic peptides in the bioinformatically predicted transmembrane domain located around residue 1294.
Regions in P1 have been extensively characterised in an earlier study 55 . Highly immunogenic regions and adherence mediating regions were found distributed throughout P1 particularly in the carboxy-terminal half of the molecule 55 . Sera from patients infected with M. pneumoniae bound to regions in P1 that were not responsible for adherence 55 . It is conceivable that P1-derived proteoforms divert the binding of host antibodies away from regions in P1 required for adherence. We hypothesise that post-translational processing events release a proportion of P1-derived proteoforms into the extracellular milieu, a process that may represent an immune decoy mechanism that seeks to bind and direct host antibodies away from M. pneumoniae. A similar scenario has been hypothesized for Protein M of Mycoplasma genitalium; a close relative of M. pneumoniae 79 .  www.nature.com/scientificreports www.nature.com/scientificreports/ Our affinity studies suggest that the different proteoforms retain the ability to bind to different host proteins, glycosaminoglycans and sialoglyconjugates. RP15 was observed to bind immobilised A549 cells in microtitre plate assays (Fig. 4). This was surprising as no adherence regions have been previously identified within RP15. Anti-RP15 antibodies were reported to be unable to inhibit M. pneumoniae adherence to primary human bronchial epithelial (HBEC) cells, human fetal lung fibroblasts (MRC-5), and human cervical carcinoma cells (HeLa) 55 suggesting that RP15 may bind to specific receptors only present on the A549 cell surface. We were unable to determine binding activity to A549 cells for P1-30 or P1-15 (Fig. 4) because anti-RP-15 antibodies did not detect these peptides. To investigate the binding capabilities of the C-terminal peptide P1-15, it was bound to avidin agarose and incubated with A549 cell lysates. This strategy selectively recovered cytoskeletal proteins, vimentin, cytokeratin 7, cytokeratin 8, and cytokeratin 18 (Fig. S1) from P1-15-avidin agarose but not from avidin agarose control experiments. Although preliminary, these observations are worthy of further study. Cytokeratin 7 is found in epithelia of lungs and other tissues 80 , and has been shown to be involved in stabilising cytokeratin 18 81 . Both cytokeratin 8 and 18 are major structural proteins of epithelial cells 82 and are found in the intermediate filaments of A549 cells 83 . Cytokeratin 8 has been identified to reside on the cellular surface of carcinogenic keratinocyte cells (HaCat) 84 , carcinogenic mammary cells 85 , and carcinogenic hepatocytes 86 suggesting they may be surface accessible on many cells. Cytokeratin 8 and 18 are co-expressed and frequently found associated together 87,88 . Vimentin forms filaments and is primarily expressed when epithelial cells transition into mesenchymal cells and function to induce changes in cell shape, motility and adhesin during this transition 89,90 . Vimentin has also been observed to be secreted to the extracellular matrix and on the surface of activated macrophages 91 . Cytokeratin 8, 18, and vimentin are suggested to be targeted by different pathogens after successfully invading host cells 84,[92][93][94][95] or after inducing cytoskeletal rearrangement [96][97][98][99][100] . Pathogenic bacteria are known to interact with these cytoskeletal proteins during infection 95,101,102 . Although mycoplasma have long been considered to be cell surface-associated parasitic bacteria, this dogma has been challenged with numerous reports citing phylogenetically-divergent mycoplasmas residing within eukaryote cells and possessing the molecular machinery for selective uptake into, survival within, and release from phagosomes [103][104][105][106][107][108][109] .
We recently showed that Mpn142, a member of the same operon that houses the P1 gene (mpn141), and the surface accessible moonlighting adhesin, elongation factor Tu (Ef-Tu), are cleaved extensively 12,28 . Post-translational processing of adhesins has been well characterised in M. hyopneumoniae where cleavage fragments have been shown to adhere to porcine cilia, porcine kidney epithelial cells, and a range of host molecules such as the glycosaminoglycan mimic heparin [59][60][61][62]64,72,74,[110][111][112][113][114][115] , plasminogen 60,[112][113][114] , actin 116 , and fibronectin 59,72,[112][113][114] . Processing of adhesin molecules is not confined to M. hyopneumoniae but has been described in Mycoplasma gallisepticum 117 , Mycoplasma fermentans [118][119][120] , M. genitalium 121 , and Spiroplasma citri 122 . Here we show that major adhesion molecules in M. pneumoniae, a phylogenetically distinct human pathogen, are processed 12,28 . All these studies suggest that the processing of surface accessible proteins is widespread in Mollicutes. It is notable that all the P1 fragments that were recovered during heparin affinity chromatography contained putative glycosaminoglycan binding motifs except an N-terminal and a central fragment (Fig. 1, fragments 14 and 16). These motifs consist of clustered, positively charged amino acids that have been shown to have a role in binding to glycosaminoglycans 69,72 , actin 123 , and plasminogen 123 . Heparin mimics the glycosaminoglycans found in the extracellular matrix and on the surface of host cells 124 . M. hyopneumoniae, and M. gallisepticum have been shown to bind heparin to aid in host adherence 110,125 . Pathogens such as Staphylococcus and Neisseria spp., Helicobacter pylori, and Streptococcus pyogenes are able to recruit heparin to the bacterial cell surface and employ bound heparin to bind other host molecules 126 . Finally, heparin has also been implicated in biofilm formation by increasing cell-cell interactions in the Gram-positive pathogens, S. aureus 127 and Lactobacillus rhamnosus 128 . M. pneumoniae forms large, complex biofilms on abiotic surfaces 34 . Heparin affinity chromatography of M. pneumoniae has been performed previously 129 identifying only nine proteins, none of which was P1. Recently, we showed that Ef-Tu in M. pneumoniae displays a strong affinity to heparin 12 . Collectively, our studies suggest that the ability to bind heparin is a universal strategy in microbial pathogenesis.
In several instances, we observed multiple cleavage sites within P1 that clustered within a defined region of P1. For example, 18 cleavage sites clustered between amino acids 1343-1361 in the C-terminus of P1 (Table 1). Sequential cleavage patterns similar to this was also reported in Mpn142 28 and in Mhp493, a paralog of the major adhesin P97 (Mhp183) in M. hyopneumoniae 74 . Surfaceome studies of M. pneumoniae (data not shown) revealed the presence of surface accessible aminopeptidases that may target a neo-N-terminal cleavage event and sequentially clip amino acids subsequent to the initial cleavage event. The function of these clipping events remains unknown but could be a mechanism to alter function and localisation of cleavage fragments, or represent a mechanism to recycle amino acids 74 . Cleavage site 14 in P1 (Fig. 1) occurs within a large predicted disordered region (amino acid range 1187-1382). The inherent flexibility of disordered regions make them accessible to protease activity 130 . Many major cleavage events identified in M. hyopneumoniae adhesin molecules reside with large disordered regions [60][61][62]74,114,115 .
The C-terminus of the P1 tail is homologous to the C-terminus of Mpn142 and the C-terminal 15 amino acids of P1 ( 1613 PGAPKPPVQPPKKPA 1627 ) has 73.3% sequence identity with the same region in Mpn142. Almost half of this sequence consists of proline residues while lysine is also heavily represented in this region. Proline-rich regions in proteins have been implicated in protein:protein interactions [131][132][133] and it has been suggested that proline residues could anchor the C-terminus of P1 in the cell membrane 49 . Lysine-rich regions are associated with binding plasminogen 60,64,123,134,135 , heparin 59,61,69,72,115,136 , actin 116,123 , and DNA 75,137 . While P1-15 and P1-30 bound plasminogen in a dose-responsive manner, it was notable that RP-15 bound it more strongly. RP-15 also bound fibronectin and fibrinogen more strongly than P1-30 (Fig. 3). These data suggest that extra binding sites for these host molecules are located upstream of the C-terminal 30 amino acids of P1. Previous work suggests that sialic acid is the dominant host receptor for the P1 adhesin [18][19][20][21] . Consistent with these earlier studies the P1 tail has a strong affinity to the sialic acid rich protein, fetuin. Our data indicates that the mature P1 proteoform and a Scientific RepoRtS | (2020) 10:6384 | https://doi.org/10.1038/s41598-020-63136-y www.nature.com/scientificreports www.nature.com/scientificreports/ further nine smaller proteoforms of P1 bind fetuin. The ability to bind fetuin has been linked with biofilm formation in M. pneumoniae 34 . conclusion In summary, this study reports that the P1 adhesin is subject to extensive post-translational processing forming twenty-two proteoforms from seventeen cleavage sites. Each of the proteoforms retain the ability to bind to host molecules or their structural mimics and are surface accessible. Processing has been described in M. hyopneumoniae, M. gallisepticum, and S. citri and is likely to be a widespread mechanism to generate surface protein diversity and promote protein:protein interactions. Specifically we show that the C-terminus of P1 plays a role in adhering to a range of host molecules including cytoskeletal proteins. This study expands on our knowledge of the role that the P1 adhesin plays in interactions between M. pneumoniae and host cells.

Data availability
Data for this study is available on request from the corresponding author.