Development of replicative plasmids for M. feriruminatoris
The genetic tools currently available for manipulation of Mferi are limited and only the introduction and modification of the bacterial genome in yeast has been reported, which is technically challenging and time consuming. To accelerate the shuttle in of genes into Mferi, we decided to generate replicative plasmids based on the origin of replication (oriC) sequence of the chromosome (Fig. 1A), as previously reported for other Mollicutes species 35,42–44. We adapted the oriC-plasmid pMYCO1 for use in Mferi by exchanging the oriC region of Mmc by the oriC region of Mferi strain G5847. This first oriC-plasmid, named pIVB03, could be successfully transformed in Mferi and promoted resistance to tetracycline at 15 µg mL− 1. However, as Mferi strains engineered in yeast already have a tetracycline resistance cassette, we developed a derivative plasmid of pIVB03 named pIVB04 carrying the puromycin acetyl transferase (pac) gene, which was reported to confer resistance to puromycin in other closely related Mollicutes species 45. This puromycin resistance-conferring plasmid was transformed but did not produce puromycin-resistant colonies carrying pIVB04 (Fig. 1B). This negative outcome was changed by switching the direction of the pac gene (Fig. 1A), which became plasmid pIVB06, and allowed the recovery of resistant colonies at 16 µg mL− 1 puromycin. In other Mycoplasma species, oriC-plasmids have been shown to be more stable when reducing the length of the oriC region or by deleting the dnaA gene 19,46. Therefore, we modified the oriC-plasmids pIVB03 and pIVB06 by replacing the dnaA gene of Mferi by the antibiotic resistance cassette, generating streamlined oriC-plasmids with resistance to tetracycline (pIVB08) or to puromycin (pIVB09) (Fig. 1A). These plasmid versions transformed with an efficiency up to four times higher than their parental plasmids with dnaA counterparts (Fig. 1B). They can be used for subsequent mutant phenotype complementation studies and as versatile backbones for heterologous expression systems reported in this study.
MIB-MIP tandem gene copies are widespread, divergent, and horizontally transferred among the Mollicutes
The MIB-MIP system is widespread between different members of the Mollicutes 5, but little is known about their diversity. We aimed at studying the occurrence and diversity of the different existing MIB-MIP cleavage systems in different phylogenetic clades of the Mollicutes. Therefore, we selected Mollicutes genomes including the ones reported in this study and analyzed the presence of the MIB-MIP system. Altogether, we included a total of 34 genomes from 23 different species (Fig. 2A). We analyzed the sequences of a total of 70 genes coding for putative MIB proteins and 71 genes encoding putative MIPs. In most genomes analyzed, both MIB and MIP counterparts were found adjacent in the same genetic locus, most likely forming a single transcriptomic unit. Occasionally we observed “orphan” MIB-encoding genes without a MIP-encoding gene in its vicinity. However, in those genomes with orphan MIBs, there was always another MIB-MIP pair present elsewhere in the genome, suggesting that Ig cleavage activity can potentially still occur. To facilitate the analysis, only the paired MIB-MIP proteins were included in the construction of the phylogenetic trees and the analyses (Fig. 2A and Figure S1, Supp. Table S1).
An initial scrutiny of the phylogenetic relations between MIB-MIP proteins of different species revealed three major distinctive branches, (i) encompassing Mycoplasma and Mycoplasmopsis species infecting ruminants; (ii) Ureaplasma, Metamycoplasma and some MIB-MIP pairs of porcine Mesomycoplasma species; and (iii) a more distant branch containing the MIB-MIP pairs of Mycoplasmoides and Mesomycoplasma species. Interestingly, we detected several examples of closely related MIB-MIP systems between Mollicutes species that are phylogenetically very distant but share the same hosts, such as Ureaplasma species and Metamycoplasma hominis, Mycoplasmoides gallisepticum and Mycoplasmopsis synoviae, or Mycoplasmopsis pulmonis and Metamycoplasma arthritidis, infecting humans, poultry, or rodents, respectively (Fig. 2A and 2B). This data strongly suggests that these genes have been acquired through horizontal gene transfer between distantly related Mollicutes infecting the same host. Species belonging to the ‘M. mycoides cluster’ and their close relatives have multiple pairs located adjacent to each other in the same chromosomal locus. Here we show that there is a strong conservation between each pair in the same position, suggesting that multiple copies of this tandem system were present in a common ancestor of these mycoplasmas. Furthermore, there is a significant difference between the two strains of Mferi analyzed, as the MIB-MIP pairs of the type-strain G5847T are similar to the ones present in Mccp, while the MIB-MIP pairs of the IVB14/OD_0535 strain cluster closely to the ones of the other members of the ‘M. mycoides cluster’. Remarkably, the two paired MIB-MIP copies of M. hyopneumoniae, located on different chromosomal loci, are highly divergent, with one MIB-MIP being more similar to the only system present in M. hyorhinis and the other more similar to MIB-MIP present in Mesomycoplasma ovipneumoniae (Fig. 2A). Overall, our analysis shows how widespread these tandem systems are within the Mollicutes, with high sequence divergence and multiple possible gene acquisitions, particularly between species sharing the same hosts.
Analysis of MIB-MIP expression in M. feriruminatoris
We studied the expression of each MIB-MIP gene pair of this species by transcriptomics and proteomics analyses to see whether the gene pairs are organized as operons and to identify different promotors that can be used to express MIB-MIPs from other Mollicutes species. Exploration of the transcriptomic data showed that each MIB-MIP pair constitutes an individual transcriptional unit, with the first and last pairs (MM1mfe and MM4mfe) being expressed at higher levels than the other two pairs (MM2mfe and MM3mfe) (Fig. 3A). Moreover, all MIB-MIP gene pairs are among the mid-to-high expressed genes of Mferi, with no apparent differences between MIB and MIP expression (Fig. 3B). However, when analyzing the proteomics data, MIB-MIP pairs are not among the highly expressed proteins of the cell highlighting the lack of correlation between RNA and protein levels (Fig. 3C). Besides, the expression levels of all MIP copies of each pair are significantly higher expressed compared to their MIB counterparts. Given the similar disposition and number of the different MIB-MIP gene pairs between Mferi IVB14/OD_0535 and Mmc GM12, we also analyzed transcriptomics and proteomics data of GM12 obtained in a previous study 23. Our results showed a similar mRNA expression trend for each MIB-MIP pair, with the first tandem of genes being expressed higher than the rest (Supp. Fig. S2). Moreover, proteomics data showed that most MIP proteins are expressed at higher levels than their respective MIB counterparts, with most of them being not reliably detected (Supp. Fig. S2). Overall, these results suggest that each MIB-MIP system constitutes an individual expression unit and that expression of the different tandem of genes is species-specific, despite having similar genetic organization or distribution in the different chromosomes.
Identification of MIB-MIP promoter and terminator regions
To further characterize the transcriptional profile shown by the RNAseq analysis of the MIB-MIP locus of Mferi, we decided to experimentally determine the transcriptional start sites (TSSs) present in that genomic area by primer extension. We could only reliably detect the TSS of the last MIB-MIP gene tandem encoded in Mferi, while no clear signal could be seen for the other gene pairs (Fig. 4A). The determination of the TSS of the elongation factor Tu gene (tufB) was used to establish the experimental conditions. Moreover, genetic analysis of the downstream regions of MIB-MIP gene clusters in a number of members of the 'Mycoplasma mycoides cluster' revealed the presence of inverted repeat elements with a structure reminiscent of rho-independent terminators (Fig. 4B). These elements are 32-34bp long and are located very close to the stop codon of the MIP genes. The last MIB-MIP gene pair is always devoid of this downstream element in all studied species, suggesting that it is part of a larger operon unit that also comprises putative ATPase-coding genes located downstream of the MIB-MIP cluster. The consensus sequence of these elements was determined as 5’TA(A/C)NATCCTTT(A/G)G-NT(A/T2)T(A/T2)-CTAAAGGATTTTT using all available sequences (Fig. 4C). Employing RNAfold 47, the RNA structure of these downstream elements was predicted to be a small hairpin with a minimum free energy of approximately − 12.70 kcal mol− 1 (Fig. 4C).
Generation and characterization of a M. feriruminatoris ΔMIB-MIP strain
To examine the effectiveness and versatility of the new oriC-plasmids developed for Mferi we decided to test expression of the recently identified MIB-MIP cleavage systems. Mferi strain IVB14/OD_0535 has several genes coding for a total of four MIB-MIP tandem systems clustered in a single chromosomal locus 14, in a similar disposition as in Mmc GM12 5 (Fig. 5A). To assess activity of each unique MIB-MIP gene pair of Mferi using oriC-plasmids in cellulo, we aimed at the generation of a ΔMIB-MIP knock-out mutant by cloning and modifying the genome of Mferi in yeast prior transplantation of the modified genome to a new mycoplasma cell. The genome of Mferi IVB14/OD_0535 was transformed into S. cerevisiae carrying the necessary plasmids to replace the chromosomal locus coding for the four MIB-MIP gene pairs (MF5583_00301 to MF5583_00308, ~ 20Kb) by a recombination template using the CReasPy-Cloning method 37. YACs containing the modified genome were transplanted into an Mcap ΔRE recipient cell to obtain the desired Mferi ΔMIB-MIP strain (Figure S3). This knock-out mutant could grow similarly to the WT strain in the absence of antibiotics (Fig. 5B), with a doubling time of 50 ± 2min compared to 45 ± 3min of the WT in SP-5 (Fig. 5C). Besides, the ΔMIB-MIP phenotype was confirmed by an IgG cleavage assay in which this mutant lost the ability to cleave the heavy chain of purified goat IgGs (Fig. 5D, lane 4), compared to the wild-type strain (Fig. 5D, lane 3).
Expression of native MIB-MIP gene pairs of M. feriruminatoris in trans
Each unique MIB-MIP gene pair of Mferi was cloned in a pIVB09 backbone under the control of their own natural promoters previously determined by RNAseq or primer extension. The newly constructed Mferi ΔMIB-MIP strain was transformed with each of these plasmids individually. Positive clones were exposed to goat IgGs to assess cleavage activity (Fig. 5D). Despite analyzing multiple clones harboring each MIB-MIP gene pair (data not shown), we could only detect IgG heavy chain cleavage in clones expressing the first and last gene tandems of the cluster (MM1mfe and MM4mfe), while other clones expressing the other MIB-MIP pairs (MM2mfe and MM3mfe) showed marginal cleavage or activity below the detection limit. As the transcriptomics data suggested that the MM1mfe and MM4mfe gene tandems were expressed at higher levels than MM2mfe and MM3mfe, and that expression from the promoter region of MM1mfe (PMM1mfe) had no apparent interplay with any rho-independent terminator or upstream regulatory sequences, we decided to reintroduce the MM2mfe and MM3mfe gene pairs under the control of PMM1mfe in a ΔMIB-MIP genetic background. Under these conditions, all MIB-MIP gene pairs could be expressed, and mutants showed clear IgG cleavage activity (Fig. 5E), suggesting that all MIB-MIP gene pairs are functional in cellulo, and that the PMM1mfe region was sufficient to drive expression of two relatively large membrane-associated proteins organized in an operon.
In vivo IgG cleavage by M. hyopneumoniae and M. hyorhinis
Cleavage of immunoglobulins by the MIB-MIP system has been reported in Mollicutes of the formerly known 'Spiroplasma phylogenetic group', i.e Mmc 9 or Mferi 12, but never in Mollicutes species like Mesomycoplasma spp. To determine if important porcine pathogens such as M. hyopneumoniae or M. hyorhinis can target and cleave host IgGs, we analyzed immunoglobulin cleavage activity of two strains isolated in Switzerland. Genome analysis revealed that M. hyopneumoniae Ue273 contains two complete MIB-MIP gene pairs and a single orphan MIB gene, while M. hyorhinis JF5820 only contains a single MIB-MIP gene pair 17 (Fig. 6A). This contrasts with many Mollicutes species of the ‘Spiroplasma phylogenetic group’, where all MIB-MIP gene copies are clustered in a single chromosomal locus containing 3–4 complete MIB-MIP gene pairs. Interestingly, neither of the different MIB-MIP copies in M. hyopneumoniae were downstream followed by an ATPase gene cluster, as it is the case in most Mollicutes species. This ATPase gene cluster in M. hyopneumoniae is found in a different chromosomal location instead, seemingly controlled by a DNA slippage mechanism in a similar fashion as antigenic or phase variation switches present in Mollicutes (Figure S4) (Citti 2010). Incubation with purified commercial IgGs isolated from naïve pig serum showed cleavage activity by both pathogens M. hyorhinis and M. hyopneumoniae (Fig. 6B), with the heavy chain of the immunoglobulins being targeted.
Heterologous expression of MIB-MIP gene pairs of other Mollicutes
The availability of a ΔMIB-MIP strain together with a vector capable of expression of MIB-MIP pairs in trans in Mferi prompted us to complement this strain with MIB-MIP systems from other Mollicutes species. First, we cloned the last gene tandem of the MIB-MIP operon of Mmc GM12 (MM4Mmc) under the control of the promoter region of the first gene tandem of the same species (PMM1Mmc), mimicking a similar disposition performed in situ at the chromosomal location in a previous work 9. This construction was transformed in the ΔMIB-MIP strain and positive clones exhibit restored capacity to cleave goat IgGs (Fig. 6C, lane 6). Hereafter, we cloned the two complete MIB-MIP gene pairs and the single MIB-MIP gene tandem from M. hyopneumoniae Ue273 (MM1Mhp and MM2Mhp) and M. hyorhinis JF5820 (MMMhr) in a pIVB09 backbone. In a first attempt, we introduced each MIB-MIP set under the control of their natural promoters, as previously done with the MIB-MIP genes of Mferi and Mmc. However, despite obtaining similar number of transformants carrying the different oriC-plasmids, no IgG cleavage was detected (Fig. 6C). To facilitate recombinant expression, we adapted all the MIB-MIP coding sequences to the codon usage of Mmc GM12, the closest species of the ‘Mycoplasma mycoides cluster’ to Mferi with an available characterized codon usage table (kazusa.or.jp), and replaced the natural promoters with the PMM1Mfe, which proved capable of generating mRNA of similar length as previously shown in this study. However, transformants carrying these new constructs could neither cleave goat nor porcine IgGs (Fig. 6C), suggesting that the system was not active or could not be correctly exported, folded or displayed at the membrane of Mferi cells.
To further investigate this, we decided to clone in pIVB09 tagged-versions of the MIB-MIP systems of Mferi (MM4Mfe), Mmc (MM4Mmc) and the single MIB-MIP system of M. hyorhinis (MMMhr) to track protein expression by immunoblotting. All MIB genes were fused with a C-terminal 6xHis tag, while their MIP counterparts were tagged with a C-terminal FLAG tag. Analysis of Mferi ΔMIB-MIP strains carrying these plasmids showed that neither of the proteins forming the MIB-MIB system of M. hyrorhinis was expressed in these conditions (Fig. 7A), which explained the lack of IgG cleavage showed previously. Sequence analysis of the MIB-MIP systems of Mferi, Mmc and M. hyorhinis showed significant differences in the N-terminal residues (Supp. Fig. S5), which could prevent export of these proteins to the cell surface. To test this, we replaced the predicted signal peptides of the tagged MIB-MIP system of M. hyorhinis with the ones present in the MIB-MIP pair 1 of Mferi. Strains carrying this construction could correctly express the protease component of the MMMhr, but not the binding protein (Fig. 7B).
Goats infected intranasally and transtracheally with Mycoplasma feriruminatoris did not develop disease
Mferi is considered a promising candidate for the development of a vaccine chassis 12. However, Mferi has only been isolated from wild caprinae 10, data regarding its pathogenic potential in closely related domestic animals are absent. Therefore, we decided to assess the pathogenicity of the type-strain G5847T of Mferi as the representative member of the species. We used a challenge model established for the phylogenetically related species Mmc 8 and modified for Mccp 41, which is robust and reproducible 48. Positive control was the highly virulent Mccp ILRI181 49. Clinical evaluation was assessed daily and was carried out 10 days pre-infection up to 25 days post-infection (dpi). Goats infected with Mferi showed no clinical signs in contrast to animals infected with Mccp, which showed onset of clinical disease including elevated body temperature at 6–8 dpi (Fig. 8). This was followed by high fever (> 40.5°C for all animals), associated with respiratory distress, coughing and wheezing (8–10 dpi), less movement and reduced intake of food. All criteria considered, this clinical evaluation led to a severity grade of 3 at 10 dpi; consequently, the three animals infected with Mccp were euthanized (Supp. Fig S6). During the course of infection, we monitored the hematological parameters. All animals infected with either species did not show a clear difference in the hematological parameters compared to their baseline levels prior infection (Supp. Fig S6). Postmortem analysis did reveal CCPP-typical pathomorphological changes including the detection of Mccp, while the animals infected with Mferi did not have any lesions pointing towards Mferi-related disease and Mferi could not be isolated from the animals.