Nucleases are easy to detect in vitro by visualization of DNA hydrolysis. For instance, DNA can be embedded into a solid matrix, either agar growth medium or a polyacrylamide gel, and its hydrolysis results in a clear halo around colonies or a clear band after electrophoretic migration of the nuclease (Minion et al., 1993; Sharma et al., 2015; Zhang et al., 2016; Yamamoto et al., 2017). Extracellular nucleases have been suspected in mycoplasmas since 1993, when Minion et al. demonstrated degradation of linear DNA by mycoplasma-free supernatants obtained after incubation of several mycoplasma species in a nuclease assay buffer (Minion et al., 1993). More recently, DNA degradation around M. pneumoniae colonies indicated the secretion of an extracellular nuclease, which was found by mutagenesis to be the MPN491 nuclease (Yamamoto et al., 2017). Other extracellular proteins with nuclease activity in vitro have been reported, including the P40 protein of M. penetrans, as demonstrated using zymography (Bendjennat et al., 1997), and the MAM superantigen of M. arthritidis, as demonstrated by degradation of genomic DNA (Diedershagen et al., 2007). Homologs of these nucleases have also been detected in the cytoplasmic membrane of other species, but their extracellular localization has yet not been investigated (Sharma et al., 2015).

Specific antibodies can be used to detect and localize nucleases. An example is the Mhp597 nuclease of M. hyopneumoniae, an ortholog of M. pneumoniae MPN491, which was detected in the growth medium supernatant by western blotting using specific antibodies (Li et al., 2019).

Although less straightforward, bulk proteomics data can be examined to detect potential nucleases in the cell environment. For instance, several nucleases have been identified by MS in the exoproteome purified from cell-free culture supernatants [e.g., Mhp7448_0580 (homolog of Mhp597) and MBOV_RS02825] or in the supernatant of a swine tracheal cell line infected with M. hyopneumoniae (Zhang et al., 2016; Paes et al., 2017; Leal Zimmer et al., 2019; Table 1).

Once secreted into the mycoplasmal environment, extracellular nucleases could participate in (i) nutrient scavenging (through degradation of the DNA released by cell lysis during the infection process) or in (ii) host evasion through hydrolysis of the neutrophil extracellular traps (NET), which are DNA networks produced by neutrophils that can trap and kill pathogens. NET evasion has been shown in vitro for two important lung pathogens, M. bovis in cattle (by nuclease MBOV_RS02825) and M. pneumoniae in humans (with nuclease MPN491) (Zhang et al., 2016; Yamamoto et al., 2017). The inactivation/deletion of the gene coding MPN491 resulted in a reduced survival of M. pneumoniae in the presence of neutrophils in vitro as well as in vivo in a mouse nasal infection model. In the most recent example, poor survival of a mutant was associated with its inability to degrade NETs induced by Escherichia coli lipopolysaccharides (Yamamoto et al., 2017).

Extracellular proteases produced by lung-colonizing ruminant mycoplasmas have been demonstrated by measuring the degradation of fluorescent casein in culture supernatants and observing clearer digestion areas around colonies on casein-enriched solid medium (Allam et al., 2010; Ganter et al., 2019). Zymography and mutagenesis experiments further showed that M. mycoides subsp. capri MMCAP2_0241 and M. capricolum subsp. capricolum MCAP_0240, which both belong to the S41 peptidase family, are the main extracellular peptidases of these two species (Allam et al., 2010, 2012; Ganter et al., 2019). By zymography, the apparent molecular mass of these polypeptidases was estimated to be 55 kDa, in contrast with the 75 kDa molecular mass predicted from the sequences of the corresponding genes. Two predicted transmembrane domains at the N- and C-terminal ends suggest a potential membrane localization of the native proteins that may be cleaved, by endoproteolysis for example, to be released into the medium (Ganter et al., 2019).

Sequence analysis and function prediction have identified several peptidases in the exoproteome of M. capricolum subsp. capricolum (Voros et al., 2015) and M. bovis (Zubair et al., 2020b), as well as in the supernatant of cells infected with M. hyopneumoniae and M. flocculare (Leal Zimmer et al., 2019; Table 1).

In other bacterial models, such as Staphylococcus (S.) aureus, extracellular proteases are involved in nutrient acquisition, bacterial dissemination and immune evasion (Tam and Torres, 2019). In mycoplasmas, their precise roles have yet to be defined, with a few exceptions. For instance, in M. hyopneumoniae the endopeptidase F and the XAA-pro aminopeptidase degrade immunologically active peptides (see below), and the cell-free form of aminopeptidase AAZ44217.1 is involved in adhesion to plasminogen and heparin (Robinson et al., 2013). Extracellular peptidases of M. capricolum subsp. capricolum MCAP_0240 may have a direct or indirect role in cell surface shaving and thus modulate adhesion and immune invasion, as deletion mutants had a modified surface proteome (Ganter et al., 2019).

Chromogenic substrates, such as 2-naphthyl caprylate and 2-naphthyl butyrate, have made it possible to detect lipase activity in the supernatant of M. capricolum subsp. capricolum cultures (Voros et al., 2015), but the corresponding protein has not yet been identified. Lipase activity has also been detected in two immunodominant surface proteins, the P65 lipoprotein of M. hyopneumoniae (Schmidt et al., 2004) and the M. bovis MilA protein (Wawegama et al., 2014), using assays based on hydrolysis of lipid substrates such as O-dilauryl-rac-glycero-3-glutaric acid resorufin ester, p-nitrophenyl caproate or p-nitrophenyl palmitate. Western blotting further demonstrated that one of them, the M. bovis MilA lipase, is released into the culture supernatant (Adamu et al., 2020) and it has also been found in the exoproteome of M. bovis strain HB0801 (Zubair et al., 2020b). In contrast, the P65 lipase of M. hyopneumoniae was not directly shown to be extracellular, but has been found in the exoproteomes of M. hyopneumoniae strains J and 7448 obtained from culture supernatants (Paes et al., 2017) and from the supernatant of infected cells (Leal Zimmer et al., 2019; Table 1).

In other bacteria, such as S. aureus, secreted extracellular lipases are involved in a broad spectrum of functions, including acquisition of lipids from the host (Delekta et al., 2018), immune evasion (Chen and Alonzo, 2019), biofilm formation, and host cell invasion (Nguyen et al., 2018). The dual potential localization—i.e., extracellular as well as cell-attached—of MilA in mycoplasmas complicates research into its function(s) (Adamu et al., 2020). MilA seems to be essential, as M. bovis growth is inhibited by anti-MilA antibodies, and screening of M. bovis transposon libraries has failed to detect a MilA mutant (Sharma et al., 2014; Josi et al., 2019; Adamu et al., 2020). Cell-free recombinant MilA has been shown to bind lipid and heparin, suggesting a putative role in the processing and transport of lipids and in adhesion to the extracellular matrix (Adamu et al., 2020).

Dozens of mycoplasma proteins have the capacity to bind to host cells or to components of the host extracellular matrix such as actin, fibronectin and glycosaminoglycans. Several of these adhesins or adhesion-related proteins, although classically described as cell surface-associated (and belonging to the surfaceome), are regularly identified in the exoproteome of several mycoplasma species (Table 1).

For instance, the P102 adhesin of M. hyopneumoniae has been shown, by immunoelectron microscopy after experimental infection of swine, to be localized either within mycoplasma cells or distant from them and directly attached to respiratory cilia of the experimentally infected swine (Djordjevic et al., 2004; Adams et al., 2005). P102-homologs have been identified in the exoproteomes of M. hyopneumoniae and M. flocculare (Paes et al., 2017; Leal Zimmer et al., 2019; Table 1). Other adhesins have been found in the exoproteome of M. hyopneumoniae, M. flocculare and M. synoviae, but no role has yet been firmly defined for them as released proteins (Table 1). A reasonable hypothesis could be that the release of adhesins from the mycoplasma cell could promote mycoplasma dispersion by limiting their adhesion to host cells and tissues, as described for other bacteria (Coutte et al., 2003). They could also be blocking anti-adhesin antibody binding to the cell, but this has yet to be demonstrated.

However, it is becoming increasingly evident that adhesins and other proteins could also play a role in degradation of the host extracellular matrix (ECM) by binding to plasminogen. This is the case for M. hyopneumoniae adhesin P102 (Seymour et al., 2012; Leal Zimmer et al., 2020). Binding of P102 to plasminogen resulted in increased conversion of plasminogen into plasmin, a serine protease able to degrade the ECM either directly or through activation of other enzymes, such as metalloproteases (Seymour et al., 2012). This strategy of ECM degradation by subversion of the host plasmin system is used by many other bacteria to invade and spread (Lahteenmaki et al., 2005). Non-adhesin proteins that could also contribute to degradation of the ECM include the glycolytic enzyme glyceraldehyde-3-P-dehydrogenase (GAPDH) in M. hyorhinis (Wang et al., 2021), the chaperone protein DNAK, GAPDH and subunit E1α of the pyruvate dehydrogenase complex (PDHB), another enzyme involved in carbon metabolism, in M. pneumoniae (Grundel et al., 2016; Hagemann et al., 2017), the enzyme enolase, which catalyzes the interconversion of phosphoenolpyruvate into 2-phosphoglycerate in M. bovis (Song et al., 2012), and the elongation factor EF-Tu, which orchestrates transport of aminoacylated tRNA to the ribosome, in M. pneumoniae and M. hyopneumoniae (Widjaja et al., 2017). These proteins have been found in the exoproteome of mycoplasmas (Table 1).

The cell-free form of the chaperone protein DnaK from M. fermentans was also shown to be taken up by the host cells and localize in the cytoplasm, the perinuclear space and the nucleus (Benedetti et al., 2020). Once in the cell, DnaK interacts with several host proteins, including those involved in DNA repair, like PARP1 (poly-ADP ribose polymerase-1) and USP10 (ubiquitin carboxyl-terminal hydrolase protein-10). The interaction with USP10 in turn leads to a reduction in P53 activity, which is known to have an anti-oncogenic effect (Zella et al., 2018; Benedetti et al., 2020). For bacteria, this is a way to redirect host-cell metabolism to facilitate bacterial growth (Siegl and Rudel, 2015). Using immunoprecipitation techniques with an anti-P53 monoclonal antibody, Zella et al. identified other mycoplasmal proteins able to interact with P53, one of which was the glycolytic enzyme enolase (Zella et al., 2018), which has been found in the exoproteome of five mycoplasma species (Table 1).

A western blotting study recently confirmed that the MbovP280 protein, predicted in silico to be secreted, was effectively released by M. bovis cells cultured for 36 h (Zhao et al., 2021). The secreted MbovP280 protein was further shown to bind to and induce the apoptosis of bovine macrophages through a complex signaling pathway (Zhao et al., 2021). However, the effect of a recombinant protein was greater than that of whole cells expressing MbovP280. Apoptotic activity has also been described for mycoplasmal extracellular proteins with other main roles. For instance, the nucleases P40 in M. penetrans, Mhp597 in M. hyopneumoniae and RS_02825 in M. bovis induce apoptosis in human lymphocytes (Bendjennat et al., 1999), in swine kidney epithelial cells (Li et al., 2019), and in bovine macrophages via the NFkB p65 pathway (Zhang et al., 2016), respectively. Extracellular nucleases can also play a role in modulation of cytokine production. M. hyopneumoniae Mhp597 nuclease has been shown to suppress INFγ secretion and stimulate IL1, IL8 and TNF secretion in macrophages (Li et al., 2019). This is an important finding, as INFγ is a first-line protection against viral infection. The MAM superantigen of M. arthritidis has also been shown to modulate cytokine release (Mu et al., 2000). In addition to nucleases, some proteases may also have immunomodulatory effects. M. hyopneumoniae endopeptidase F and XAA pro-aminopeptidase, which are both found in the exoproteome (Table 1), have recently been shown to be involved in the degradation of peptides that play a role in innate immunity (Jarocki et al., 2019). These data have started to demonstrate a role of the mycoplasma releasome in evasion of the immune system and modulation of its activity, which are two important features of mycoplasma virulence (Leal Zimmer et al., 2020; Askar et al., 2021; Jiang et al., 2021; Yiwen et al., 2021).

The combination of (i) numerous potential experimental biases (failure to pellet some mycoplasma cells, cell lysis during the purification protocol, and so on), (ii) the high proportion (circa 30%) of hypothetical proteins with no associated function found in exoproteomes, (iii) the extent of moonlighting activity in mycoplasma proteins, and (iv) underpowered in silico prediction capacity (Zhao et al., 2021) means that further studies complementary to studies characterizing the core exoproteome are necessary to confirm the extracellular localization of the exoproteins and decipher their role. With the growing interest in mycoplasma exoproteins, the mycoplasmology community would welcome a consensus methodology guideline to improve the quality of results and enable valuable comparisons of exoproteomes between species. For instance, specific labeling of newly synthesized proteins using new methodologies such as bioorthogonal non-canonical amino acid tagging and proximity labeling (Shin et al., 2019; Sukumaran et al., 2021) could be helpful to distinguish the exoproteome of mycoplasmas from proteins contained in a complex environment. This includes complex growth medium, the cytoplasmic compartment in case of intracellular mycoplasmas, or different host sites occupied by mycoplasmas in the course of infection [for example M. hyopneumoniae has been detected in the heart, kidneys, liver and spleen of pigs (Le Carrou et al., 2006; Marois et al., 2007; Woolley et al., 2012)].

Studies addressing the role of exoproteins in the interplay with the host are also essential. As the generation and use of mycoplasma mutants is still limited, approaches based on recombinant proteins are increasingly being used (Bendjennat et al., 1999; Zhang et al., 2016; Yamamoto et al., 2017; Zhao et al., 2021), but they can sometimes exaggerate the actual role of a protein because of the high concentrations tested, which do not correspond to the actual quantities secreted by the cell (Zhao et al., 2021).

The mycoplasmology community is starting to gain a global picture of the role of exoproteome interactions with host cells and with components of the host extracellular matrix and how it shapes immune evasion or modulation. However, the exoproteome composition is likely to vary over time, depending on the host context and the time since infection. Variability in exoproteome composition between species, and potentially between strains, might also explain some of the variability in virulence.

Last but not least, although some Sec genes have been identified in silico (Staats et al., 2007), further investigation is needed into the mechanisms involved in protein release by Mycoplasma spp. For instance, surface shaving via proteases could contribute to the release of exoproteins from the cell-surface, highlighting the tight connection between the surfaceome and the releasome (Raymond et al., 2013; Jarocki et al., 2015, 2019; Tacchi et al., 2016; Berry et al., 2017; Ganter et al., 2019; Machado et al., 2020). Other non-classical protein release mechanisms have been suggested such as explosive cell events or ghost cells formation (Raymond et al., 2018a). Those could contribute to free into the extracellular medium membrane or cytoplasmic proteins (Wang et al., 2013).

Mycoplasma exopolysaccharides were first reported in the 1930s, when Kurotchkin described a carbohydrate released by M. mycoides subsp. mycoides into the culture medium and the blood of animals with acute CBPP (Kurotchkin, 1937; Kurotchkin and Benaradsky, 1938). It was then of unknown composition and structure, but cross-reacted serologically with the M. mycoides subsp. mycoides capsular polysaccharide (Plackett et al., 1963), which is composed of galactose and named galactan (Plackett and Buttery, 1958). By the 1960s, it was clear that M. mycoides subsp. mycoides produced both a CPS and an EPS, with a shared antigenic signature recognized by the same antibodies. Despite several efforts, purification of the EPS from culture supernatants remained thwarted by difficulties in eliminating contamination from polysaccharides, such as glycogen, contained in the growth medium (Plackett et al., 1963; Hudson et al., 1967). It was only in 2013 that Bertin et al. eliminated contamination with medium-associated polysaccharides by transferring PPLO-grown M. mycoides subsp. mycoides cells into CMRL, a defined cell culture medium with no polysaccharides and only glucose as a carbon source (Bertin et al., 2013). CMRL was shown to sustain mycoplasmal metabolism but not growth (Bertin et al., 2013). An EPS, in free form in the spent CMRL, was purified and its composition and structure were shown by NMR and HPLC to be identical to the polysaccharide moiety of the capsular galactan described 50 years earlier (Bertin et al., 2013). This identity explained the immunological cross-reactivity between the EPS and CPS, but was limited to the polysaccharide moiety, as galactan CPS contains a lipid anchor (Buttery and Plackett, 1960) that is not present when the galactan is released from the cells (Bertin et al., 2013).

Other EPS were identified among the members of the M. mycoides cluster using the same purification method: a β-(1→2)-glucopyranose was detected in the culture supernatant of M. capricolum subsp. capricolum, M. capricolum subsp. capripneumoniae and M. leachii (Bertin et al., 2015; Table 2), and weak exopolysaccharide release was detected in vitro from the two serovars of M. mycoides subsp. capri (Bertin et al., 2015; Gaurivaud et al., 2016). Interestingly, the two serovars of the goat pathogen M. mycoides subsp. capri secreted two different polysaccharides: serovar LC (large colony) produced galactan and serovar capri produced β-(1→6)-glucopyranose. Both EPS were homopolysaccharides with no ramification and no chemical modifications. However, not all mycoplasmal polysaccharides are as simple. The EPS isolated from M. pneumoniae biofilms, for instance, is composed of galactose and N-acetyl glucosamine (Simmons et al., 2013; Table 2).

All these EPS were also detected as CPS (Table 2). However, the presence of a CPS does not guarantee secretion of the corresponding EPS, as shown by the β-(1→6)-glucopyranose of M. agalactiae, which was only detected as a CPS (Gaurivaud et al., 2016; Table 2). Other CPS have been detected in M. feriruminatoris, which is able to produce a galactan and a β-(1→6)-glucopyranose CPS (Ambroset et al., 2017), and in M. genitalium (Daubenspeck et al., 2020) and M. pulmonis (Daubenspeck et al., 2009; Table 2). However, their release as EPS has not been demonstrated to date.

In M. mycoides subsp. mycoides, the galactan is not secreted concomitantly as CPS and EPS in vitro, but rather alternatively by phenotypic variants undergoing phase variation (Bertin et al., 2013). One variant secreted a galactan CPS but no EPS, whereas the other one was not capsulated but produced a galactan EPS. This phenomenon was reversible and was shown to be related to the expression of a permease of the glucose-phosphoenolpyruvate phosphotransferase system (PTS) undergoing on/off phase variation (Gaurivaud et al., 2004; Bertin et al., 2013). The PTS system is a well-established sugar transport system, but also regulates carbon metabolism in bacteria (Deutscher et al., 2014) and has already been shown to regulate polysaccharide synthesis in Vibrio cholerae (Houot et al., 2010). One group posited glucose PTS system-mediated regulation of a glycosyltransferase possibly involved in attachment of the polysaccharide moiety to the membrane lipid anchor (Bertin et al., 2013), but the regulatory and attachment mechanisms have yet to be demonstrated.

A complete biosynthesis pathway involving a membrane glycosyltransferase belonging to the synthase family (Whitney and Howell, 2013), and catalyzing both polymerization and transfer of the galactan of M. mycoides subsp. mycoides in its CPS or EPS across the cytoplasmic membrane, has been proposed based on in silico data (Bertin et al., 2013). The role of synthases in mycoplasma polysaccharide synthesis was also demonstrated using a functional genomics approach for the synthesis of the M. agalactiae β-(1→6)-glucopyranose (Gaurivaud et al., 2016). Among the many glycosyltransferases identified in mycoplasmal genomes, synthases are easily identified as they have 4 or 7 transmembrane domains and a cytoplasmic loop bearing the glycosyltransferase-active sites (Gaurivaud et al., 2016). The number of transmembrane domains may be linked to substrate specificity and the structure of the resulting polymer. Synthases have been predicted in silico for 14 Mycoplasma species, which suggests that several other polysaccharides have yet to be identified (Gaurivaud et al., 2016). In contrast, no synthesis pathway has been identified for the polysaccharides produced by M. pulmonis, M. genitalium or M. pneumoniae (Daubenspeck et al., 2009, 2020). However, an ABC transporter pathway (Schmid, 2018) has been suggested for M. pulmonis, as the mutation of two ABC permease genes was associated with a loss of polysaccharide production (Daubenspeck et al., 2009).

In mycoplasmas, as in other bacterial models, CPS are known to be involved in the modulation of cytoadherence (Bolland and Dybvig, 2012) and in protection against phagocytosis (Shaw et al., 2013) and against the bactericidal activity of the complement system (Bolland et al., 2012; Gaurivaud et al., 2014). In contrast, with the exception of circulating galactan in the body fluids of animals experiencing CBPP, the role of mycoplasmal EPS has been under-researched. Intravenous injection of galactan purified from M. mycoides subsp. mycoides culture supernatant into calves resulted in an increase of pulmonary arterial blood pressure and a transient apnea (Buttery et al., 1976). However, in its EPS form, the galactan did not induce any fever and did not affect the susceptibility of calves to subsequent subcutaneous infection with M. mycoides subsp. mycoides (Hudson et al., 1967; Buttery et al., 1976). However, these results need to be interpreted with caution, as the purified galactan was shown to contain peptides (Hudson et al., 1967; Buttery et al., 1976). This ambiguity was overcome recently by Totte et al. (2015) who used a high-purity-grade galactan obtained from M. mycoides subsp. mycoides supernatant, checked by SDS PAGE and NMR, to examine its effect on immune cells. Pure galactan failed to activate naive lymphocytes but induced IL-10 release by bovine macrophages, thus echoing the observation of peak IL-10 1–2 weeks after experimental infection of cattle with M. mycoides subsp. mycoides (Sacchini et al., 2012). Moreover, the purified free galactan was able to reduce the release of pro-inflammatory cytokines by macrophages in response to Escherichia coli lipopolysaccharide (Totte et al., 2015). These observations indicated that, overall, the galactan EPS acts as an immunosuppressor by inhibiting pro-inflammatory cytokines and increasing the IL-10 secretion that depresses T-cell responses (Totte et al., 2015).

Polysaccharides are a major constituent of biofilm matrices, where they play roles in binding different components and in physical resistance to stress or stratification of the biofilm structure (Limoli et al., 2015). Both CPS and EPS could contribute to biofilm formation in mycoplasmas. For instance, in M. pulmonis, different types of expressed dominant polysaccharides changed the propensity of strains to form a biofilm (Daubenspeck et al., 2009). Similarly, in M. pneumoniae, the volume, texture, robustness and internal structure of the biofilm was shown to differ between two strains (of different type), depending on how loosely or tightly the polysaccharides were attached to the mycoplasma cell (Simmons et al., 2013). This could ultimately completely modify the role of the resulting biofilm in virulence and chronicity of infection.

Although extracellular polysaccharides have been detected in several Mycoplasma species, only three polymers have had their structure elucidated thus far. This relatively low rate of structural characterization may be attributable to the necessity of purifying a high quantity (approximately 1 mg) of material for structure analyses by HPLC and NMR, which could be technically challenging depending on the species, culture conditions and overall complexity of the exopolysaccharide structure (e.g., biofilms) (Table 2).

Clear identification of the pathways involved in the synthesis of galactan, β-(1→6)-glucopyranose and β-(1→2)-glucopyranose has been helpful for in silico screening of other species for their potential to produce polysaccharides. However, for M. pulmonis, M. genitalium and M. pneumoniae, the EPS biosynthesis pathways have yet to be deciphered. In M. pulmonis, random transposon mutagenesis identified the role of two genes coding for ABC transporters, but failed to find the glycosyltransferase catalyzing the polymerization. This suggests that polymerization of the polysaccharides could be an essential function for mycoplasmas. Other mutations in genes involved in polysaccharide biosynthesis were shown to have an adverse effect on cell viability: for example, a mutation in the UDP galactofuranose mutase gene of M. mycoides subsp. capri reduced the membrane integrity of the cells (Schieck et al., 2016). The genes involved in polysaccharide attachment to the cell surface and the membrane anchor have not yet been identified.

A final crucial point is that we are still a long way from identifying the roles of EPS and their level of secretion in vivo. The example of galactan illustrates that, despite having the same polysaccharide moiety, CPS and EPS have two different roles: protection of the cell (CPS) and suppression of inflammation (EPS). This has some similarity with the moonlighting proteins of mycoplasmas that play different roles depending on their localization. Phase variation between capsulated variants and uncapsulated variants secreting EPS is thought to be involved in the adaptation of M. mycoides subsp. mycoides to changing environments during host colonization (Gaurivaud et al., 2014), but this hypothesis has yet to be validated in vivo.

Extracellular vesicles are the least investigated elements of the mycoplasmal releasome. Membranous particles, with a diameter of 75 to 210 nm, were first observed in the 1960s by electron microscopy in mycoplasma cultures during studies of cell ultrastructure and cellular division (Domermuth et al., 1964; Hummeler et al., 1965). More recently, using a classical method for EV purification, similar nanosized particles were observed by TEM in Acholeplasma laidlawii and several Mycoplasma species (Chernov et al., 2011; Gaurivaud et al., 2018). The diameters of these particles ranged from 30 to 170 nm, although a few larger particles of around 200 nm were also observed for M. mycoides subsp. mycoides (Gaurivaud et al., 2018). Although the shapes and sizes of these particles were typical of classical prokaryotic EV, further evidence was needed to definitively confirm that they were EV: they had to be (i) non-replicative, (ii) surrounded by a lipid bilayer, (iii) produced by viable cells, and (iv) contain cytoplasmic proteins in order to prove that the particle was not just a circularization/reassembly of membrane fragments (Thery et al., 2018). They were shown to meet the first three criteria (Gaurivaud et al., 2018), but the fourth criterion—internal protein contents of the EV—has not yet been assessed. However, TEM images of mycoplasma cultures clearly showed EV budding from cells, suggesting that they are not the result of reassembling membrane fragments nor of an aberrant division of small mycoplasma cells (Hummeler et al., 1965; Gaurivaud et al., 2018).

Production of vesicles by exploding bacterial cells has been described recently for Bacillus spp. (Toyofuku et al., 2019). The “explosion” results from the expression of an autolysin, i.e., a peptidoglycan hydrolyzing enzyme. After the degradation of the cell-wall, the cell is lysed because of the high intracellular pressure [10 atm for Bacillus (Rojas and Huang, 2018)]. Subsequent reassembly of bacterial membrane debris can generate vesicles. However, these vesicles differ from surface-budding EV with respect to their composition (Toyofuku et al., 2019). Explosive cell-lysis events leading to membrane vesicles formation were once observed for M. hyopneumoniae cells embedded in a biofilm produced in vitro (Raymond et al., 2018a). Comparison of size, structure and cargo composition (expected to be random in case of explosive cells) between EV and vesicles produced from explosive cell events would certainly shed new light onto their respective roles.

Bacterial EV are involved in an array of processes, including stress responses, cell communication, and protein secretion (Kim et al., 2015; Coelho and Casadevall, 2019). Several proteins involved in host interactions have been identified in the proteome of EV membranes in three species, including pathogens of both domestic animals (M. mycoides subsp. mycoides and M. agalactiae) and humans (M. fermentans) (Gaurivaud et al., 2018). Of all the putative virulence factors found in EV membranes, DnaK is one of the most abundant proteins (Gaurivaud et al., 2018) and this protein has also been detected in host cells (Zella et al., 2018) (see earlier). Secretion through EV and delivery into the host cells could be used by mycoplasmas that are devoid of other more classical secretion systems. Secretion of proteins through EV confers both protection of the cargo molecules against degradation and an efficient transport system able to deliver high concentrations of cargo molecules to the cells (Coelho and Casadevall, 2019; McMillan and Kuehn, 2021).

Bacterial EV could be involved in nutrient acquisition, as recently described for iron acquisition by EV of Mycobacterium tuberculosis and Pseudomonas (P.) aeruginosa (Prados-Rosales et al., 2014; Lin et al., 2017). These bacteria are able to scavenge the iron sequestered by EV, and because of their dissemination over long distances, EV are assumed to supply nutrients to bacteria localized at different infection sites. Given that nutrient binding proteins, such as those belonging to ABC transporter systems, have been identified in the EV membranes of mycoplasmas (Gaurivaud et al., 2018), EV could participate in nutrient acquisition, which is a crucial feature for these biosynthetically limited bacteria.

Data on the biogenesis, composition and role of mycoplasmal EV are scarce. Despite the considerable efforts made to date, there is still a need to further optimize the purification of EV (Thery et al., 2018) and completely eliminate the small-sized mycoplasma cells observed in the purified material from some species (Gaurivaud et al., 2018). A second gap to address is to complete the characterization of the EV composition and, crucially, their cytoplasmic content. The presence of polypeptides, nucleic acids, DNA and RNA within EV has not been yet assessed. A method that can directly assess the presence of DNA or RNA inside EV has been proposed recently (Bitto et al., 2017).

Extracellular vesicles secretion is energetically expensive for bacteria (McMillan and Kuehn, 2021) and may be even more expensive for the small-sized (300–800 nm diameter) mycoplasmas (Razin et al., 1998), which produce EV of 30–170 nm diameter. This raises the question of the fitness burden that EV release might represent for mycoplasmas, which needs to be compensated for by beneficial roles of EV release. Remodeling mycoplasmas membrane by vesicle production could be a rapid mechanism to maintain membrane integrity and its adaptation to changing environments, as it is described for other bacteria (McMillan and Kuehn, 2021). Deciphering the cargo composition of EV, and an in-depth comparison of virulence factors in the vesicle membranes (Gaurivaud et al., 2018) with those of parental cells, would certainly help define these beneficial roles. Selective cargo packaging, i.e., preferential exclusion or selection of vesicle cargo, has already been demonstrated in enterotoxigenic Escherichia coli during stress (Orench-Rivera and Kuehn, 2021). EV produced by explosive-cell events [observed for M. hyopneumoniae biofilm in vitro (Raymond et al., 2018a)] carry different cargo than classical EV and may have different roles (Toyofuku et al., 2019; McMillan and Kuehn, 2021) such as release of “public goods” useful for the bacterial community inside the biofilm (Turnbull et al., 2016).

The in vitro use of cellular models to study EV release and cytotoxicity or their effect on immunity compared to mycoplasmas alone would bring further insights into their role. Similarly, experimental infection of animals could provide information about the potential long-distance dissemination of mycoplasmal EV (Stentz et al., 2018).

Extracellular vesicles-based vaccines could also hold promise for use in mycoplasmosis control as it has been shown for other bacterial diseases (Jiang et al., 2019; Behrens et al., 2021). However, the presence of proinflammatory lipoproteins in the mycoplasmal EV membrane and the yield rate and cost of in vitro EV production remain technical and economic bottlenecks, although they could be solved by expressing extracellular virulence factors within heterologous EV systems (Carvalho et al., 2019; Stentz et al., 2022).