Members of the mycoplasmas are known to have some the smallest free-living prokaryotes. They are characterized by their small genomes (0.58 – 2.20 Mb) and lack of a cell wall [1]. The availability of complete genome sequences of several mycoplasmas provides a great opportunity to understand the minimal genome biology of these small bacteria [2]. Considerable advances have been made towards understanding the pathogenesis of mycoplasmas; of particular interest is the interaction of these pathogens with the immune system, macrophage activation, cytokine induction, superantigens, autoimmune manifestations, and the evasion of host immune responses. However, despite the wealth of new genetic information emerging from bacterial genome sequencing, many newly discovered genes remain functionally unclassified [3]. It is suggested that these mollicutes are phylogenetically related to Gram-positive bacteria whereby their genomes were reduced through multiple events of mutations. This implies that these simple bacteria are not found at the root of the phylogenetic tree but are more likely late evolutionary products [4]. Owing to their reduced genomes mycoplasmas have limited biosynthetic capabilities and therefore tend to take on a parasitic lifestyle in which they obtain nutrients and other metabolic building blocks such as amino acids, nucleotides, fatty acids, and sterols from their hosts [1]. Generally Mycoplasmas enter their host and multiply over long periods of time, slowly wearing down the host [1]. This is one reason perhaps why mycoplasmas have been linked to AIDS pathogenesis and to malignant transformation, chromosomal aberrations, the Gulf War Syndrome, and other unexplained and complex illnesses, including chronic fatigue syndrome, Crohn’s disease, and various arthritis’ [5]. These microorganisms have developed a variety of mechanisms in order to deal with host immune responses including; mimicry of host antigens, survival within phagocytic and non-phagocytic host cells and the generation of phenotypic plasticity [1]. One of the most important questions about the pathogenesis of mycoplasmas is how they cause damage to the host cells and to what extent is the damage clinically apparent in the host.

M. capricolum is the causative agent of caprine arthritis, mastitis, and respiratory disease [6]. The sophisticated antigenic variation of M. capricolum makes it difficult for the host immune system to produce effective antibodies to suppress the infection [6]. The accumulation of hydrolytic enzymes produced by this species also contributes to tissue damage during infection [6]. A better understanding of the pathogenesis of these bacterial infections may be the result of exposing the nature of virulence factors elaborated by these pathogens during growth. The secretome is of particular interest as it shows the potential for elaborating such virulence factors.

We have undertaken this study to characterize the secretome of M. capricolum using two-dimensional electrophoresis and mass spectrometry of tryptic digests. The results were compared to bioinformatics analysis of the secretome and supported with enzymatic characterization. This study provides new information and adds to the somewhat limited data on the biology of M. capricolum. Response of this pathogen to acidic pH provides a hint as to the mechanism of adaptation during infection and following phagocytic uptake in the host.

Bioinformatic analysis

The sequenced genome of M. capricolum (Accession # NC_007633) was retrieved from the World Wide Web NCBI site (www.ncbi.nlm. nih.gov/). The genome was first analyzed using the in-house program ExProt [7]. The program was used to identify the secreted / exported proteins of gram positive bacteria. The proteins that were predicted as being secreted or exported were further analyzed using the program SignalP 3.0 server (www.cbs.dtu.dk/services/SignalP/) [8]. The amino acid sequences of the Signal peptidase I proteins were processed through the SignalP 3.0 server, specified to analyze gram positive bacteria.

Growth

Complete growth media (Modified Eaton’s Medium, MEM) was composed of 8 g Mycoplasma Broth Base (Viragen, Ontario), 3.2 g of dextrose phenol red combination, and 100 ml of donor horse serum (The Cell Culture Company, PAA, Ontario), 100 μg/ml ampicillin (Sigma-Aldrich, St. Louis, MO) and 400 ml of distilled water, for a total final volume of 500 ml. The growth medium was modified in order to exclusively study the secretome of M. capricolum. The protein content was removed from the growth media by filtration through an Amicon filter (Millipore Stirred Cell 8400) with a 10,000 kDa cut off limit [9]. The culture medium was inoculated with Mycoplasma capricolum subsp. capricolum (California Kid type strain; kindly provided by Dr. Pascal Sirand-Pugnet, INRA, France) incubated at 37°C for 48 hours under 5% CO₂. This culture was then spun down at 13,000 rpm for 30 minutes at 4°C (Beckman, GS-15R Centrifuge). The cell pellet was then resuspended and washed in phosphate buffered saline (PBS) and centrifuged at 13,000 rpm for 30 minutes at 4°C. This pellet was split and resuspended in either 150 ml of Amicon Modified culture medium (AM) buffered with 2-(N-morpholino)ethanesulfonic acid (MES) at pH 5.5 (acidic) or 150 ml of AM buffered with MES at pH 7.5 (neutral). Culture supernatants were collected by centrifugation at 13,000 rpm for 30 minutes at 4°C followed by filtration through a 0.22 μm filter.

API ZYM enzyme assay

API ZYM (bioMerieux, Durham, NC, USA), a system designed for the detection of enzyme activities with certain chromogenic substrates, was used to determine the presence or absence of certain enzymes in the culture supernatants. Approximately 10 μg (65 μl) of proteins form the neutral and acidic protein sample was dispensed into each of the 20 cupules of the individual strips. The strips were then placed into their respective incubation boxes with 5 ml of distilled water. After inoculation, the plastic lid was placed on the incubation box and the box was incubated for 4 hours at 37ºC. After the incubation period 20 μl of Fast Garnet GBC solution (5mls of 500 mM acetate buffer pH 6.2, and 2.5 mg/ml of Fast Garnet GBC salt) was added to each cupule on each strip and left exposed to daylight at room temperature for approximately 30 minutes until sufficient colour development had occurred.

Two-dimension SDS-PAGE

The volume of CFP used for isoelectric focussing was measured according to the protein concentration determined by the BCA protein assay. The immobilized pH gradient (IPG) strips (linear pH 3-10, 17- cm length, Bio-Rad) can accommodate up to 3 mg of protein while still maintaining proper protein resolution. Therefore volumes of CFP containing 3 mg of the Acidic CFP and 3mg of the neutral CFP were dried in the SpeedVac (DNA 120, Savant, Thermo Electron Corporation) for approximately 45 minutes at low heat. The dried CFPs were resuspended for 30 minutes in 350 μl of IPG rehydration buffer at room temperature (7M urea, 50mM dithiothreitol (DTT), 4% Chaps, 0.2% ampholytes, 0.001% bromophenol blue (BPB). The IPG strips were then rehydrated and isoelectric focussing (8-10,000V-hr at 20°C) was performed overnight (Protean IEF® Cell, Bio-Rad). IEF focusing conditions were the same for both the acidic and neutral CFPs. Once the first dimension was completed, the electrophoresed IPG strips were incubated for 20 minutes in equilibration buffer (6M urea, 0.375 M Tris HCl pH8.8, 2% sodium dodecyl sulphate (SDS), 20% glycerol) containing 2% DTT followed by and additional 20 minutes in equilibration buffer containing 2.5% iodoacetamide. Samples were subsequently subjected to the second dimension on a 10% separating gel (4% stacking) at a constant voltage of 200 V, and at 16 mA/gel for 30 minutes and 25 mA/gel for 5 hours using a vertical gel electrophoresis system by Whatman, Inc. Model V16-2 (Florham, NJ).

Silver staining

Proteins spots on the 17 cm 2D gels were visualized by silver staining methods (compatible with Mass Spectrometry analysis) based on the work of Blum et al. [10]. The gels underwent fixing in 40% methanol and 10% acetic acid. Two washes of 30% ethanol followed for 20 minutes each. A single wash of distilled H₂O followed for 20 minutes. The sensitizer containing 0.02% DTT was then applied to the gels for 10 minutes followed by three washes of distilled H₂O for 20 seconds each. The gels were placed in 0.2% silver nitrate for 20 minutes. Three washes of distilled H₂O followed for 20 seconds each. The gels were developed with 2% sodium carbonate and 0.04% Formaldehyde until the spots became visible. Three washes of distilled H₂O followed for 20 seconds each. The gels were placed in stop reagent containing 0.5% glycine overnight.

Mass spectrometry and protein identification

Protein spots of interest were manually excised from the 2D-gels and placed in siliconized sterile 1.5 ml microfuge tubes. In gel digestion and peptide extraction was based on a protocol provided by the manufacturer of the silver staining kit (Promega Inc.) with a modification of the protocol of Shevchenko et al. [11]. The protein spots were first destained by mixing in 50 μl of 50 mM ammonium bicarbonate, leaving them to sit for 5 minutes, and the wash solution was discarded as waste. The spots were then shrunk by the addition of 50 μl of 50% acetonitrile/25 mM ammonium bicarbonate, again the spots were vortexed and left to sit for 10 minutes; the liquid was discarded as waste. The spots were then reduced by the addition of 30 μl of 10 mM DTT solution for 30 minutes and incubation at 56°C, the supernatant liquid was discarded as waste. The spots were alkylated next by the addition of 30 μl of iodoacetamide for 15 minutes in the dark at room temperature. The supernatant liquid was again removed and discarded as waste. The protein spots were then shrunk again by the addition of 50 μl of 50% acetonitrile/25 mM ammonium bicarbonate. The mix was vortexed and left to stand for 15 minutes, the supernatant was discarded as waste. The protein spots were then digested with trypsin. Each spot was covered with 13 ng/μl of trypsin, and the gel pieces were incubated overnight at 37°C. The tubes were removed from incubation and the liquid surrounding the gel pieces was transferred to a corresponding fresh sterile 500 μl microfuge tube. The tryptic peptides were sequentially extracted with 20 μl of each of the following: 25 mM ammonium bicarbonate, 5% trifluoroacetic acid (TFA), and 20 μl of 100% acetonitrile. The extracted peptides were dried in the SpeedVac for approximately 2.5 hours to decrease the volume from approximately 100 μl to 10 μl. Samples were subjected to MALDI-TOF Mass spectrometry (North Eastern Ontario Regional Cancer Center, Sudbury, Ontario). Proteins were identified using the Profound [12] server (http://prowl.rockefeller.edu/prowl-cgi/profound.exe).

Bioinformatic analysis of the putative secretome of M. capricolum

Using the in-house program ExProt the entire predicted proteome of M. capricolum was treated for identifying proteins with signal sequences similar to Type I (secretory) or Type II (lipoproteins) signal sequences of Gram positive bacteria. Out of the 812 proteins from the M. capricolum genome (Accession # NC_007633) 247 of the proteins were identified as having either of these two signal sequences. Members of this set were then individually processed through the SignalP server which confirmed 60 lipoproteins (signal peptidase II) and 55 nonlipoproteins (signal peptidase I) within this set (Tables 1 and 2). This “filtered” set of proteins constitutes the core secretome and makes up about 14% of the total theoretical proteome of this pathogen.

Table 1: Secreted / exported signal peptidase I proteins, predicted through the use of ExProt (ExProt) and SignalP 3.0 server.

Table 2: Secreted / exported Lipoproteins (signal peptidase II proteins) from Mycoplasma capricolum, predicted through the use of ExProt (ExProt).

API ZYM enzyme profiling

Culture supernatants were profiled for enzyme activities to compare with the findings of the bioinformatics and experimental protein identifications. This enzyme assay system provides qualitative evaluation of 20 enzyme activities. Culture supernatants of M. capricolum displayed similar enzyme activity profiles with 7 positive out of a total of 20 activities screened in this test (Table 3). These were the acid and alkaline phosphatases, C4 and C8 esterases/lipases, phosphohydrolase, β-galactosidase, and β-glucuronidase. There were distinct differences between the two culture supernatants however where the CFP from the acidic media showed higher activity for 4 of those enzymes: the acid and alkaline phosphatase, the C8 esterase/lipase, and the phosphohydrolase. This CFP also showed activity for 2 additional enzymes not detected in the neutral CFP: leucine arylamidase and α-chymotrypsin.

Table 3: Enzymatic profiling of the culture filtrates of M. capricolum.

Proteomic analysis of culture supernatants

It was generally observed that the growth of M. capricolum in the acidic medium was retarded compared to that at neutral pH. It was also noted that the protein content of the acidic culture filtrate was consistently less compared to that at neutral pH. Representative maps of the 2D gels are presented in Figures 1 and 2. A total of 82 gel spots can be resolved from the 2D map of the neutral CFP (Figure 1) and 111 spots were recorded for the acidic CFP (Figure 2) using silver staining. Superimposition of both maps (not shown) revealed that 18 spots were unique to the neutral CFP which are numbered 4, 15, 23, 37, 58, 59, 60, 61, 62, 63, 70, 72, 74, 77, 78, and 82. The composite image also showed 26 spots that were unique to the acidic CFP and included spot numbers 83-108. It is important to note here however that there is no information on whether these represent unique proteins or simply proteolytic fragments of others. A significant number of protein spots failed to provide usable tryptic peptide fingerprint and were thus not identified. From the neutral CFP gel, a total of 27 proteins were identified (Table 4) compared to 12 proteins for the acidic CFP gel (Table 5).

Figure 1: 2D-PAGE reference map of the culture filtrate proteins of neutral culture of M. capricolum.

Figure 2: 2D-PAGE reference map of the culture filtrate proteins of acidic culture of M. capricolum.

Table 4: List of identified proteins from the culture filtrate of neutral culture of M. capricolum.

Table 5: List of identified proteins from the culture filtrate of acidic culture of M. capricolum.

Treatment of the theoretical proteome of M. capricolum with our in-house program ExProt resulted in a predicted secretome size of 247 proteins out of a total proteome size of 812 proteins. This fraction of the proteome (30.4%) is comparable to predicted secretomes of other mycoplasmas including M. pneumonia (37%) and M. genitalium (24%) predicted with the same program [7]. A value of 32% for the M. pneumonia secretome was reported using the LocateProtein program [7]. These secretome sizes fall in the high end of the spectrum for secretome sizes and inspection of the predicted secretome of M. capricolum reveals that this high size possibly due to the large number of predicted lipoproteins as part of the secretome. Over 7% of the proteome of M. capricolum could potentially represent lipoproteins (Table 2). Treatment of the proteome using the SignalP program reduced the secretome size from 247 proteins to 115, including the 60 predicted lipoproteins (Table 2). This reduces the secretome size from 30% to 14%. SignalP predicted 55 proteins with Type I signal sequence with a composition made up mostly of hypothetical proteins, transport proteins, and proteins involved in protein synthesis (Figure 3).

Figure 3: Functional grouping of the predicted secretome of M. capricolum.

Enzymatic profiling of the culture supernatants of M. capricolum identified 7 enzyme activities in the neutral and acidic culture supernatants of M. capricolum. These were the acid and alkaline phosphatases, C4 and C8 esterases/lipases, phosphohydrolase, β-galactosidase, and β-glucuronidase. There were distinct differences between the two culture supernatants however where the CFP from the acidic media showed higher activity for 4 of those enzymes: the acid and alkaline phosphatase, the C8 esterase/lipase, and the phosphohydrolase. This CFP also showed activity for 2 additional enzymes not detected in the neutral CFP: leucine arylamidase and α-chymotrypsin. What was more curious however is the observation that the predicted secretome did not contain any proteins that could account for the observed enzymatic activities. This result was unusual and it’s not clear why at this time. The ExProt and SignalP programs have been tested previously on many proteomes and their comparative prediction power was established many years ago [7]. If it is not the quality of the predictive programs then what the possible reasons for this observation. Cell lysis contribution as a possibility was looked at by measuring the levels of the cytoplasmic enzyme lactate dehydrogenase. Typical values for this activity in the CFP’s used in this study were typically 10% that of the cytoplasmic extracts (data not shown). This activity is regularly used as part of the quality control of the CFP’s used in our study. Other possible reasons may include for instance the dependence of M. capricolum on other secretory pathways to excrete proteins and that the general secretory pathway may only play a minor role in protein secretion in this bacterium. We should also not dismiss the fact that a large number of the predicted proteins in the secretome of M. capricolum have unknown functions (Tables 1 and 2). It is therefore possible that the observed enzymatic activities in the CFP’s are in fact due to correctly predicted components of the secretome but have been annotated as having unknown functions due to lack of homologous proteins for sequence comparison.

The genomes of several mycoplasma species include distinctive sets of genes that encode different surface lipoproteins [13]. For the most part, it seems that these lipoproteins remain functionally unknown. However, some research suggests that surface lipoproteins confer both protective and opportunistic capabilities to M. capricolum based on analogous proteins. Lipoproteins of M. hyorhinis or M. pulmonis protect these pathogens from immunological targeting by host antibodies [14] and complement [15]. Additional functions for these surface proteins include the phase variation that is produced by mutations in certain gene clusters such as the variable M. capricolum subsp. capricolum (Vmc). The predicted secretome of M. capricolum contains the six Vmc (A-F) lipoproteins (Table 2). These genes and their products have been well characterized [13]. From the CFP of the M. capricolum grown at neutral pH a total of 28 proteins were identified by trypsin digest and MALDI-TOF mass spectrometry (Table 4). A major component of this fraction contained putative lipoproteins (8 out of 28, or 29%). The second major component of the CFP was made up of hypothetical proteins (5 out of 28, or 18%). The abundance of these two groups of proteins identified in this CFP matches that obtained through the bioinformatics analysis (Tables 1 and 2, Figure 3) where the lipoproteins and the hypothetical proteins represented the largest two fractions of the predicted secretome. A recent report by Couto et al. [16] outlined the identification of 19 proteins from the CFP fraction of M. synoviae cultured in modified Frey’s medium. From the 19 identified proteins 3 were listed as hypothetical proteins. This is similar to our results were the hypothetical proteins made up 18% of the extracellular fraction. The authors did not report the presence of any lipoproteins in their findings. There have been several other proteomic studies of various mycoplasmas in the past few years but none of them focussed on the exported/secreted proteins [17-21]. Other notable findings in our study include the identification of several peptidases. A total of 4 different proteases were identified in 5 spots from the 2D gels (Table 4) to include oligoendopeptidase F (spots 4 and 9), dipeptidase, cytosol aminopeptidase, and zinc metalloprotease. Gelatinases are metalloproteases and in our study gelatin zymography confirmed the presence of gelatinase activity in the CFP of M. capricolum (results not shown). The zinc metalloprotease identified in the CFP from the neutral culture medium (MCAP_0804) may be responsible for this activity. Aminopeptidases are exopeptidases that release N-terminal amino acid residues from their substrates; see review by Gonzales and Robert-Baudouy [22]. Aminopeptidase activity is well established in mycoplasma and is used in classification and identification of these microorganisms [23-25], including M. capricolum [26]. Although none of those studies looked at the secretome, aminopeptidases can be secreted and not only found in the cytosol or associated with the membrane or envelope in bacteria [22].

The 2D gel profile and the number of identified proteins of the CFP from the acidic media of M. capricolum were generally less complex than that from the neutral media. The bacterium did show reduced growth potential in acidic media and the reduced complexity of the CFP profile may be a reflection of this observation. A total of 12 proteins identified in this acidic CFP and its composition showed marked differences in comparison to that from neutral CFP. Of all the identified neutral and acidic proteins there were 3 that were seen in the secretome of M. capricolum at both a neutral and acidic pH. These included a putative lipoprotein (LppA), a putative dipeptidase, and a cytosol aminopeptidase. Only one lipoprotein in comparison was identified and the number of hypothetical proteins was 2. The acidic CFP also contained a putative hemolysin, a protein not identified in the neutral CFP. A previous study reported on the hemolytic activity of a number of mycoplasmas but not in M. capricolum [27]. One other protein of interest identified in the CFP from acidic growth media was the peptide methionine sulfoxide reductase (MCAP_0722; Table 5). This type of enzyme is typically involved in protecting bacteria from oxidative stress but a deletion mutant from M. genitalium was shown to exhibit reduced cytadherence and virulence [28].

Enzymatic profiling of both CFP’s was evaluated to compare with the identification of the proteins in these preparations. The CFP from the neutral culture of M. capricolum contained a number of proteins that could explain some of the enzymatic activities observed in this protein fraction and includes the phosphatase, aryl amidase, and peptidase activities. No proteins were identified to explain the other enzymatic activities identified by the API ZYME system. Similarly, no proteins were identified in the acidic CFP that would explain those enzyme activities identified by the API ZYME system. The 2D gels however show a number of proteins for which the trypsin digestion and mass spectrometry failed to identify. What could possibly aid in the functional classification of the large number of functionally unknown proteins are the recent developments in the field of genome transplantation [29]. Gene knockout experiments are limited by the fact that gene function can be redundant and phenotypic changes associated with certain gene knockouts would thus not be observed to hint to a possible gene function. An additional limitation to these experiments is functional compensation where loss of a specific gene function is compensated for by up-regulation of other gene products. Genome transplantation on the other hand allows for deletion of large segments or gene clusters without causing fatal consequences to the recipient cell. The recent engineering of Mycoplasma mycoides JCVI [30] attests to the promise that this technique can be attainable and can provide a wealth of information on genome evolution and function.

The authors declared that there are no competing interests.

We would like to thank Dr. Pascal Sirand-Pugnet (Institut national de la recherche agronomique, INRA, France) for providing us with the culture of mycoplasma used in this study. This work was supported in part by a research grant from the Laurentian University Research Fund (LURF).