Protein Kinase/Phosphatase Function Correlates with Gliding Motility in Mycoplasma pneumoniae

Abstract

Mycoplasma pneumoniae exhibits a novel form of gliding motility that is mediated by the terminal organelle, a differentiated polar structure. Given that genes known to be involved in gliding in other organisms are absent in M. pneumoniae, random transposon mutagenesis was employed to generate mutants with gliding-deficient phenotypes. Transposon insertions in the only annotated Ser/Thr protein kinase gene (prkC; MPN248) and its cognate phosphatase gene (prpC; MPN247) in M. pneumoniae resulted in significant and contrasting effects on gliding frequencies. prkC mutant cells glided at approximately half the frequency of wild-type cells, while prpC mutant cells glided more than twice as frequently as wild-type cells. Phosphoprotein staining confirmed the association between phosphorylation of the cytoskeletal proteins HMW1 and HMW2 and membrane protein P1 and the gliding phenotype. When the prpC mutant was complemented by transposon delivery of a wild-type copy of the prpC allele, gliding frequencies and phosphorylation levels returned to the wild-type standard. Surprisingly, delivery of the recombinant wild-type prkC allele dramatically increased gliding frequency to a level approximately 3-fold greater than that of wild-type in the prkC mutant. Collectively, these data suggest that PrkC and PrpC work in opposition in M. pneumoniae to influence gliding frequency.

INTRODUCTION

Mycoplasma pneumoniae is a cell wall-less bacterial pathogen of the human respiratory tract causing primary atypical pneumonia and tracheobronchitis (1). Mycoplasmas lack major biosynthetic pathways, classical transcriptional regulators, chemotactic and other two-component systems, and the prototypical prokaryotic cell division apparatus (2, 3). The limited biosynthetic capabilities of mycoplasmas are generally explained by their evolution as obligate parasites of diverse eukaryotic hosts (4). Colonization of the host respiratory epithelium by M. pneumoniae requires gliding motility (5), which might facilitate access to receptors on the host cell surface and subsequent lateral spread.

The gliding apparatus of M. pneumoniae is a polar terminal structure (6) that also functions in cell division (7) and adhesion to host receptors (2, 5). While the cytoskeletal protein HMW1 (MPN447) and membrane proteins P1, B/C, and P30 (MPN141, MPN142, and MPN453, respectively) localize to the terminal organelle and are required for gliding (5, 8–10), these proteins are also essential for cytadherence and attachment to surfaces. Accordingly, their distinct functions in gliding motility are difficult to define by mutagenesis alone. Given that the M. pneumoniae genome exhibits no homology to elements of defined gliding mechanisms (2, 3), including those of other gliding mycoplasmas (11–14), it was necessary to perform saturating transposon mutagenesis in order to identify the components specific to gliding (15). Transposon insertions in the genes encoding the cytoskeletal proteins P41 and P65 (MPN311 and MPN309, respectively) (16), which are known to localize to the terminal structure of M. pneumoniae (7), produce gliding-deficient phenotypes (15, 17), but the defect in each of them is not clearly associated with an actual gliding motor and suggests little about structure and function of the motor. Gliding-deficient phenotypes were also found to result from transposon interruption of genes encoding putative lipoproteins and ABC transporters, as well as genes with metabolic functions and several hypothetical genes (15). However, one of the most striking gliding phenotypes observed was the lawn-like satellite growth of a mutant resulting from transposon interruption of MPN247, which encodes PrpC, the sole annotated protein phosphatase in M. pneumoniae (Fig. 1) and a component of its phosphotransferase system (6).

Fig 1.

Schematic of the prpC-prkC gene cluster in M. pneumoniae. Arrowheads indicate the insertion sites of the transposons Tn4001:2065 (11) and pMT85 in MPN247 and MPN248, respectively. Insertions generated the prpC mutant 247-100 and the prkC mutant 248-377. Arrows indicate the positions of likely promoters based on RT-PCR analysis (data not shown).

PrpC is a member of a family termed “eukaryotic-like” Ser/Thr phosphatases (eSTPs) despite a growing recognition of their ubiquity in eubacteria. PrpC and its homologs are often partnered with the cognate kinase PrkC, a highly conserved eukaryotic-like Ser/Thr kinase (eSTK). Mutant analysis in Bacillus subtilis has shown that homologs to these enzymes regulate sporulation and cell wall development by reversible phosphorylation (18–20). Similar reversible phosphorylation might occur in M. pneumoniae, where terminal organelle proteins HMW1 and HMW2 (MPN310) are phosphorylated in an ATP-dependent manner (21).

Here, we explored further the relationship between PrkC/PrpC function and gliding motility. The prpC mutant exhibited a higher gliding velocity and a higher frequency than the wild type. We compared this prpC mutant to a mutant with a transposon insertion near the 3′ end of the prkC gene (Fig. 1). The catalytic region of PrkC in this mutant is predicted to be intact, yet gliding occurred at roughly half the frequency of that of the wild type. Considered together, these data suggest that PrpC and PrkC work in opposition in M. pneumoniae, where gliding frequency might be subject to regulation by Ser/Thr phosphorylation. Significantly, complementation of the prpC mutant restored gliding speed and frequency to the wild-type standard, while complementation of the prkC mutant increased the gliding frequency to a level approximately 3-fold higher than that of the wild type, which was somewhat surprising yet consistent with a correlation between gliding frequency and PrkC/PrpC function. Phosphorylation levels of several terminal organelle proteins, including HMW1 and P1, were elevated in the prpC mutant, as expected (22), and reduced in the prkC mutant. However, in contrast to the findings of Schmidl et al. (22), reduced phosphorylation of HMW1, HMW2, HMW3, and P1 did not impact their stability. Finally, P1 phosphorylation was significantly elevated in mutant M6, which lacks HMW1 and produces a truncated P30, raising the possibility of a phosphorelay among terminal organelle components.

MATERIALS AND METHODS

Mycoplasma strains and culture conditions.

Wild-type M. pneumoniae strain M129, cytadherence mutants II-3 (23), M6 (24, 25), H9 (26, 27), HMW3 (28), and IV-22 (29), and the prpC gliding mutant 247-100 (15) were described previously. The gliding-deficient prkC mutant 248-377 was generated serendipitously by transformation of wild-type M129 with a derivative of the transposon vector pMT85 (7). The nomenclature indicates the site of transposon insertion by open reading frame (ORF) (MPN247 or MPN248), followed by the last amino acid residue encoded by that truncated ORF (15). Cultures were grown to mid-log phase in SP-4 medium (30), with the addition of 18 μg/ml gentamicin and 24 μg/ml chloramphenicol where appropriate for antibiotic selection. Hemadsorption and satellite growth were characterized as described previously (8).

Complementation of prpC and prkC mutants.

MPN247 was amplified from wild-type M129 genomic DNA, along with the upstream ORF MPN246 and its promoter, to construct a recombinant wild-type prpC allele for complementation of the prpC mutant. Forward (5′-CGTGGCGATATCCATAACCCTGGTGC-3′) and reverse (5′-CAGGTAAGAGGATATCCCGCCTGA-3′) primers were designed with EcoRV sites (indicated in bold) providing blunt ends for ligation into the SmaI site of pKV104. To complement the prkC mutant, we implemented a comparable strategy using forward and reverse primers flanking MPN248 (5′-CGTGGCGATATCCTGTACAACTTCTTGG-3′ and 5′-GCGACAATGACAGATATCTCAAGGG-3′, respectively). Transformation of competent mutant and wild-type cells by electroporation was as described previously (31). Transformants were incubated on PPLO agar (32) with gentamicin and chloramphenicol, and colonies were carefully picked and expanded in SP-4 medium. Individual transformants were screened for satellite growth by time-lapse analysis (15).

Western immunoblotting.

Samples were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western immunoblotting (33, 34) using a monoclonal P30-specific antibody (at a 1:1,000 dilution) (35) and rabbit antisera to HMW1 (1:10,000) (36), HMW2/P28 (1:2,000) (23), HMW3 (1:10,000) (37), P1 (1:2,000) (38), P200 (MPN567) (1:2,000) (39), TopJ (MPN119) (1:2,000) (40), P65 (1:3,000) (41), B (1:5,000) and C (1:1,000) (10), P41 (1:1,000), and P24 (MPN312) (1:250) (23).

Time-lapse analysis of satellite growth, microcolony development, and cell gliding.

Satellite growth of M. pneumoniae mutants was evaluated as described previously (15) but at ×100 magnification. We measured microcolony diameter microscopically to quantify the differences in colony area between the wild-type, mutant, and complemented strains.

We previously described a procedure to measure M. pneumoniae gliding frequency and speed (8), but a modification to this protocol improved precision with the more fragile mutant strains. Overnight cultures were incubated to a cell density of 75 to 150 cells per field prior to image capture, eliminating the need for the destructive thawing and needle passage steps used in the previous protocol. In addition, 10 min prior to image capture, we replaced the spent medium in which the cultures were inoculated with a defined gliding medium (20 mM HEPES, 150 mM NaCl, 1.0 mM sodium phosphate monobasic, 27.5 mM glucose, 3% [wt/vol] gelatin [pH 7.2]). This was necessary because gliding measurements in SP-4 medium plus gelatin, as described previously (8), varied with the batch of some medium components, especially fetal bovine serum (data not shown). Wild-type gliding speed and frequency as measured here in the defined gliding medium were comparable to published values (8). Finally, for some studies, staurosporine (1 μm in methanol) was added to overnight cultures of wild-type M. pneumoniae for 2 h prior to replacement with the defined gliding medium.

Phosphoprotein staining.

Pro Q Diamond stain (Life Technologies, Carlsbad, CA) was used to compare phosphorylation in wild-type, mutant, and complemented M. pneumoniae. SDS-PAGE gels (7.5% polyacrylamide) were stained with Pro Q Diamond according to the manufacturer's protocol and imaged on a Typhoon Trio laser scanner (GE Healthcare, Little Chalfont, United Kingdom) at wavelengths of 532 nm (emission) and 580 nm (excitation). To confirm equal protein levels per sample, gels were also stained with Sypro Ruby and imaged at wavelengths of 488 nm (emission) and 610 nm (excitation). ImageJ (Wayne Rasband, National Institutes of Health, Bethesda, MD [http://rsb.info.nih.gov/ij/]) was used to perform background fluorescence subtraction and densitometric analysis of fluorescence-imaged bands corresponding to HMW1, HMW2, and P1. The final intensity measurements were normalized to those of the wild type.

P1 immunoprecipitation.

M. pneumoniae cells (300 μg protein) were dissolved in TDSET (10 mM Tris-hydrochloride [pH 7.8], 0.2% [wt/vol] sodium deoxycholate, 0.1% [wt/vol] sodium dodecyl sulfate, 10 mM tetrasodium EDTA, 1% [vol/vol] Triton X-100) (42), incubated for 30 min at 37°C, and then centrifuged at 20,000 × g for 20 min at 4°C to remove insoluble material. Supernatant (500 μl) was combined with 100 μl of P1 antibody and incubated for 1 h at 4°C, after which immune complexes were collected with 25 μl Dynabeads protein G (Invitrogen) at 4°C for 1 h. Dynabeads were recovered from the lysate-antibody mixture by magnetic separation and washed twice with TDSET. The immune complexes were eluted at 68°C for 10 min in 3× SDS-PAGE sample buffer for a final magnetic separation and analyzed by SDS-PAGE and Pro Q Diamond staining.

RESULTS

Gliding phenotype of prpC and prkC mutants.

The prpC mutant was described previously (15) as exhibiting lawn-like satellite growth without the large microcolonies that are characteristic of wild-type cultures over time. The pattern observed visually (Fig. 2) was confirmed by comparative analysis of the average microcolony areas (Table 1). A similar growth pattern was noted for the prkC mutant both visually (Fig. 2) and by average microcolony area (Table 1), although this mutant achieved confluence more slowly. The observed differences in growth patterns could not be accounted for strictly on the basis of growth rate, although both mutants grew slightly slower than the wild type.

Fig 2.

Microcolony satellite growth for the prpC and prkC mutants and a complemented transformant (C1) of the prpC mutant. Cultures were incubated in chamber slides for 144 h, and images were captured at 24-h intervals. Scale bars, 15 μm.

Table 1.

Mean microcolony areas for wild-type M. pneumoniae, the prpC and prkC mutants, and a complemented prpC mutant (C1), at the indicated time points

Strain	Mean areaa (μm2 ± 95% confidence interval) at:
	72 h	96 h
Wild type	80.65 ± 11.1	116.1 ± 15.6
prpC mutant	42.56 ± 7.48	49.34 ± 6.54
prpC-C1 strain	82.78 ± 22.3	134.2 ± 41.8
prkC mutant	32.26 ± 6.54	38.69 ± 7.21

^an ≥ 100 microcolonies measured.

Our original analysis of gliding by prpC mutant cells indicated that gliding speed and frequency were severely reduced (15). However, modification of the protocol for gliding quantification of more fragile mutant strains revealed that prpC mutant cells were, in fact, faster and glided at a frequency nearly 3-fold higher than wild-type cells (Fig. 3). In contrast, prkC mutant cells had a gliding frequency approximately half that of wild-type cells but had a comparable gliding speed (Fig. 3).

Fig 3.

Cell gliding frequency (A) and velocity (B) for wild-type M. pneumoniae (WT), the prpC and prkC mutants, the complemented transformants of each mutant (C1 and C2), and wild-type M. pneumoniae with an extra copy of the prkC allele. Gliding frequency is shown as the percentage of cells gliding during the observation period ± the 95% confidence interval (n > 4,400 cells observed for WT, >2,800 for each mutant, and >800 for all transformants). Velocities (μm/s) were recorded for ≥30 cells per strain, with one 95% confidence interval indicated.

Spontaneous mutations affecting M. pneumoniae terminal organelle function arise at a high frequency (29). Therefore, to explore whether the prpC and prkC mutants had potential secondary mutations, we employed Western immunoblotting to assess the steady-state levels of proteins known to contribute to cytadherence and/or motility, including those most prone to loss by spontaneous mutation (29). All of these proteins were present at wild-type levels in both the prpC and prkC mutants (Fig. 4). Hemadsorption screening confirmed that transposon disruption of MPN247 or MPN248 had no qualitative impact on attachment to erythrocytes (Fig. 5A), although quantitative analysis revealed an approximately 30% reduction in erythrocyte binding for the prpC mutant relative to that for the wild type (Fig. 5B).

Fig 4.

Western immunoblot analysis of wild-type M. pneumoniae (WT) and the prpC and prkC mutants. Twenty micrograms of protein was loaded per lane on a 3 to 10% polyacrylamide gradient gel. Antisera specificities are indicated on the left.

Fig 5.

(A) Qualitative hemadsorption analysis of prpC and prkC mutants, with wild-type M. pneumoniae and P1 mutant IV-22 (29) as positive and negative controls, respectively. Scale bars, 65 μm. (B) Quantitative hemadsorption analysis of the prpC and prkC mutants relative to wild-type M. pneumoniae (WT) and mutant II-3 (23) as positive and negative controls, respectively. Adherence is shown as the percentage of wild-type adherence, with one 95% confidence interval indicated. Insets, approximately 3-fold magnifications.

Protein phosphorylation in prkC and prpC mutants.

The contrasting gliding phenotypes of the prpC and prkC mutants suggested a role for reversible phosphorylation in M. pneumoniae gliding. We previously demonstrated that the terminal organelle proteins HMW1 and HMW2 contain phosphoserine and phosphothreonine (8), and recent studies reported that phosphorylation of HMW1 and adhesin protein P1 is impacted by mutations in prkC and prpC (22, 43). We employed Pro Q Diamond staining to confirm that disruption of prpC and prkC here was accompanied by changes in the protein phosphorylation profiles of these mutants (Fig. 6). Pro Q Diamond staining detected hyperphosphorylated bands corresponding to HMW1 and P1 in the prpC mutant. By comparison, the HMW1 and HMW2 bands in the prkC mutant were consistently less intense than those in the wild type, whereas P1 appeared to be phosphorylated at wild-type levels (Fig. 6A). Subsequent imaging by Sypro Ruby staining for total protein revealed that the profiles of the prpC and prkC mutants were indistinguishable from those of the wild type (Fig. 6B), consistent with the Western immunoblotting data (Fig. 4). Band intensities were analyzed in triplicate for Pro Q Diamond fluorescence intensity (Fig. 7), confirming the results noted visually for HMW1 and P1 and allowing the calculation of 95% confidence intervals.

Fig 6.

Analysis of protein phosphorylation by Pro Q Diamond staining for wild-type M. pneumoniae (WT) and the indicated mutants and complemented transformants. (A) Thirty micrograms of protein was loaded per lane on a 7.5% polyacrylamide gel and stained with Pro Q Diamond stain following electrophoresis. Bar, 200-kDa protein size standard. Ovals, absence of HMW1 and HMW2 in the mutants M6 and H9, respectively; dashed oval, absence of P1 in the mutant IV-22; open arrowhead, likely P1 band based on its absence in mutant IV-22; solid arrowhead, likely truncated HMW2 protein in mutant H9. (B) Sypro Ruby stain of the same gel shown in panel A.

Fig 7.

Relative fluorescence intensity with Pro Q Diamond phosphoprotein staining of bands representing HMW2, HMW1, and P1 from wild-type M. pneumoniae (WT), the prpC and prkC mutants, the complemented transformants (C1 and C2) of the prpC and prkC mutants, and the M6 mutant. Staining was calculated by densitometric analysis in ImageJ. Error bars, one 95% confidence interval based on three independent experiments.

We also examined the phosphoprotein profiles of the terminal organelle mutants M6, H9, HMW3, and IV-22 by Pro Q Diamond staining. The absence of phosphorylated HMW1 and HMW2 in mutants M6 and H9, respectively (Fig. 6A, ovals), was consistent with our previous findings (8). Interestingly, a band corresponding in size to the truncated HMW2 in mutant H9 (23) was also apparent by Pro Q Diamond staining (Fig. 6A, solid arrowhead). The absence of a phosphorylated band corresponding to P1 in mutant IV-22 (Fig. 6A, dashed oval) was likewise consistent with its identification as a phosphoprotein. A band migrating at the size of P1 (Fig. 6A, open arrowheads) was more intense in the prpC mutant, consistent with recent findings of others (22), and unexpectedly in the mutant M6 (Fig. 6A), which lacks HMW1 and has a truncated P30 (25). Densitometric analysis indicated an approximately 75% increase in P1 band intensity in the M6 mutant relative to that in the wild type (Fig. 7). We confirmed the identity of this band as P1 in the mutant M6 by immunoprecipitation (Fig. 8) and its hyperphosphorylation by ³²P incorporation (data not shown).

Fig 8.

Pro Q Diamond staining following immunoprecipitation of protein P1 from wild-type M. pneumoniae (WT) and the mutant M6.

Complementation of prpC and prkC mutants.

Complemented transformants of the prpC and prkC mutants were evaluated for gliding and protein phosphorylation phenotypes. The gliding speeds and frequencies of complemented transformants of the prpC mutant returned to the wild-type standards (Fig. 3). Additionally, the increased phosphorylation of P1 observed in the prpC mutant was ameliorated by complementation with the wild-type prpC allele (Fig. 6A and 7), although the phosphorylation of P1 was somewhat reduced relative to that of the wild type (Fig. 7). Complementation of the prpC mutant also resulted in wild-type satellite growth (Fig. 2) and an increase in the average microcolony size (Table 1). Complementation of the prkC mutant, on the other hand, yielded somewhat unexpected results, where delivery of the recombinant wild-type allele resulted in hyperphosphorylation of P1 (Fig. 6A and 7) and a gliding frequency greater than that of wild-type M. pneumoniae or the prpC mutant (Fig. 3). A similar phenotype was observed when a second copy of the wild-type prkC allele was introduced into wild-type M. pneumoniae (data not shown and Fig. 3).

Gliding motility in staurosporine-treated cells.

Staurosporine is an inhibitor of protein kinase C-type kinases and serine/threonine kinases in general (44). We treated wild-type M. pneumoniae cultures with 1 μm staurosporine in methanol (1% final concentration) for 2 h prior to standard motility analysis. Treated cells glided at approximately half the frequency of untreated cells (Fig. 9), while gliding speeds were not affected (data not shown). The impact on gliding frequency of treatment with methanol alone was significantly less than that of staurosporine plus methanol (Fig. 9).

Fig 9.

Cell gliding frequency for wild-type M. pneumoniae treated with the protein kinase inhibitor staurosporine (1 μM in 1% methanol) or methanol only. The gliding frequency is shown as the percentage of cells gliding during the observation period. Error bars, one 95% confidence interval based on 3 independent experiments.

DISCUSSION

The terminal organelle proteins HMW1 and HMW2 are phosphorylated, probably in an ATP-dependent manner by a Ser/Thr kinase (8, 21). Mutant analyses by others identified the likely kinase as the ORF MPN248 product, a homolog of the highly conserved PrkC (22, 43). Additionally, two other phosphoprotein bands were identified: the adhesin protein P1 and the predicted cell surface protein encoded by MPN474 (22). The neighboring ORF MPN247 encodes PrpC, a phosphatase that acts in concert with PrkC (22, 43). Our findings here suggest that the PrkC/PrpC pair functions in the control of gliding frequency of M. pneumoniae.

The prpC and prkC mutants examined here were originally identified by anomalous satellite growth morphology featuring lawn-like spreading with microcolonies smaller than those of wild-type M. pneumoniae (Fig. 2; Table 1). The increased gliding frequency of prpC mutant cells (Fig. 3) might contribute to a reduced accumulation of cells in microcolonies and more uniform lawn-like spreading compared to that of the prkC mutant, which displayed a sparser spread pattern. While satellite growth morphology might not necessarily correlate specifically with single-cell behavior, as noted with other gliding bacteria (45), complementation of each mutant with the corresponding recombinant wild-type allele supported the correlation between reversible protein phosphorylation, PrkC/PrpC function, and control of gliding frequency. Furthermore, the contrasting gliding phenotypes of these mutants strongly suggest that PrkC and PrpC work in opposition to up- and downregulate gliding, respectively.

The effect of phosphorylation on gliding was most clearly seen in the frequency at which cells glide, although an increase in velocity was also observed with the prpC mutant. While complementation returned gliding velocity to wild-type levels, the prkC mutant was not subject to an attendant decrease in speed along with its decrease in gliding frequency. In an interesting corollary, we also detected reduced gliding frequency, but not velocity, for wild-type cells treated with the protein kinase inhibitor staurosporine (44). We are inclined to conclude that protein phosphorylation clearly impacts the activation of gliding, but a potential role in regulating gliding velocity might be more complex.

Surprisingly, P1 phosphorylation, as assessed by Pro-Q Diamond staining, remained at wild-type levels in the prkC mutant, in contrast to the findings of Schmidl et al. (22). It is possible that P1 is phosphorylated by a different, unidentified protein kinase, but it is more likely that the PrkC truncation in this mutant affects the function beyond enzymatic activity. For example, the function of eSTKs such as PrkC can involve the formation of homodimers via extracellular ligand binding (19), and truncation of PrkC here might impact that aspect of activity. The comparable hyperphosphorylation and gliding phenotypes of the complemented prkC mutant and wild-type M. pneumoniae carrying a second prkC allele suggest that the truncated PrkC might form fully functional dimers when paired with full-length wild-type PrkC.

An additional unexpected finding here was P1 hyperphosphorylation in the HMW1 mutant M6 (Fig. 6A and 7). While this raises the possibility that P1 is an intermediate in PrkC-dependent phosphorylation of HMW1, it should be noted that the absence of P1 in the mutant IV-22 did not affect HMW1 phosphorylation levels. Moreover, if P1 were an intermediate in HMW1 phosphorylation, then one might also expect P1 hyperphosphorylation in mutant H9, where HMW1 is present at greatly reduced levels (23, 27), but this was not the case (Fig. 6A and 7). P1 hyperphosphorylation in M6 might reflect a requirement for HMW1 in PrkC localization, as noted previously for P1 (24).

In their analysis of similar mutants, Schmidl et al. (22) described a major variance in the prkC mutant phenotype from that observed here. In their study, mutant prkC::Tn exhibits reduced levels of several proteins involved in cytadherence, including HMW1 and HMW2, with HMW2 completely absent when assessed by Western immunoblotting. The authors attributed this to a role for PrkC phosphorylation in the stability of these proteins, HMW2 in particular, but the wild-type levels of these proteins in the PrkC mutant described here (Fig. 4) appear to contradict this finding. This discrepancy might be explained by the fact that unlike prkC::Tn, the PrkC mutant described here retains the predicted PrkC catalytic domain and might therefore be capable of phosphorylating HMW2 and other target phosphoproteins to a degree sufficient to maintain their stability. However, the “haystack mutagenesis” method of Schmidl et al. (22) is more likely than simple transposon mutagenesis to yield spontaneous secondary mutations, given its reliance on multiple passages to select for specific mutants, and the altered protein profile of the prkC::Tn mutant is similar to that of the spontaneous HMW2 mutant I-2 (9). Complementation of prkC::Tn with recombinant prkC should confirm whether its phenotype is truly the result of MPN248 disruption alone. Schmidl et al. (46) concluded from phosphoproteome analysis that unidentified protein kinases and/or autophosphorylation were likely responsible for phosphorylation of phosphoproteins unaffected by PrkC truncation. Analysis of the phosphoproteome here was limited, but detection of comparable band intensities between the wild-type and prkC mutant strains for some protein bands is consistent with that hypothesis (data not shown).

In conclusion, our study establishes a clear correlation between reversible protein phosphorylation by the PrkC/PrpC pair and gliding motility in M. pneumoniae but does not assert that protein phosphorylation by PrkC drives the gliding motor. Rather, we suggest that PrpC and PrkC constitute part of a regulatory circuit comparable to that in B. subtilis wherein the phosphorylation of elongation factor Tu and other phosphoproteins influences aspects of sporulation and cell wall biosynthesis (18, 20). Homologous eSTP/eSTK-associated regulation of swarming in Myxococcus xanthus, cell division in Mycobacterium tuberculosis, and virulence in Yersinia pseudotuberculosis have been previously reported (19), indicating that reversible protein phosphorylation is coupled to a diverse slate of functions in bacteria. We submit that gliding motility in M. pneumoniae is a strong candidate for inclusion in the growing category of cellular processes in prokaryotes controlled in part by reversible Ser/Thr phosphorylation.