A Polymorphic Gene within the Mycobacterium smegmatis esx1 Locus Determines Mycobacterial Self-Identity and Conjugal Compatibility

Abstract

Mycobacteria mediate horizontal gene transfer (HGT) by a process called distributive conjugal transfer (DCT) that is mechanistically distinct from oriT-mediated plasmid transfer. The transfer of multiple, independent donor chromosome segments generates transconjugants with genomes that are mosaic blends of their parents. Previously, we had characterized contact-dependent conjugation between two independent isolates of Mycobacterium smegmatis. Here, we expand our analyses to include five independent isolates of M. smegmatis and establish that DCT is both active and prevalent among natural isolates of M. smegmatis. Two of these five strains were recipients but exhibited distinct conjugal compatibilities with donor strains, suggesting an ability to distinguish between potential donor partners. We determined that a single gene, Msmeg0070, was responsible for conferring mating compatibility using a combination of comparative DNA sequence analysis, bacterial genome-wide association studies (GWAS), and targeted mutagenesis. Msmeg0070 maps within the esx1 secretion locus, and we establish that it confers mycobacterial self-identity with parallels to kin recognition. Similar to other kin model systems, orthologs of Msmeg0070 are highly polymorphic. The identification of a kin recognition system in M. smegmatis reinforces the concept that communication between cells is an important checkpoint prior to DCT commitment and implies that there are likely to be other, unanticipated forms of social behaviors in mycobacteria. IMPORTANCE Conjugation, unlike other forms of HGT, requires direct interaction between two viable bacteria, which must be capable of distinguishing between mating types to allow successful DNA transfer from donor to recipient. We show that the conjugal compatibility of Mycobacterium smegmatis isolates is determined by a single, polymorphic gene located within the conserved esx1 secretion locus. This gene confers self-identity; the expression of identical Msmeg0070 proteins in both donor-recipient partners prevents DNA transfer. The presence of this polymorphic locus in many environmental mycobacteria suggests that kin identification is important in promoting beneficial gene flow between nonkin mycobacteria. Cell-cell communication, mediated by kin recognition and ESX secretion, is a key checkpoint in mycobacterial conjugation and likely plays a more global role in mycobacterial biology.

Keywords: conjugation; esx1; horizontal gene transfer; kin recognition; mycobacteria.

FIG 1

Genetic map of the mid locus within esx1 of the donor strain mc²155. The mid region of the donor strain M. smegmatis mc²155 was defined by linkage mapping to Msmeg0069–0078 (8). Conserved core esx genes, homologs of which are typically found in esx loci, are shown above the locus. The encoded core and other Esx proteins are highly conserved (>96.6% amino acid identity), in contrast to the Mid proteins; percentages below the encoding mid genes indicate amino acid sequence identities between donor (mc²155) and recipient (MKD8) orthologs. Msmeg0071 is poorly conserved at its N terminus (27%), but its C terminus is highly conserved (93% amino acid identity).

FIG 2

Summary of interstrain mating that occurs between independent environmental isolates of M. smegmatis ( and 13). Arrows indicate the directionality of transfer, from donor to recipient. Jucho and MKD8 act as recipients in cocultures with donor-only strains, but they can exchange DNA bidirectionally when cocultured together, indicating that they can also act as donors.

FIG 3

Genome comparisons of the five independent environmental isolates of M. smegmatis. (A) Pangenome analysis identifies conserved core genes among the M. smegmatis isolates. Roary v.3.13.0 (42) was used to perform a pangenome analysis with default settings. Orthologs were required to share at least 95% global amino acid identity and are indicated by a vertical blue line in each genome along the x axis (the pangenome); the white (no blue) line indicates that an ortholog of >95% identity is absent in that genome. Clustered are conserved core genes (left), accessory genes (middle right), or lineage-specific genes (right) found in only one of the five strains. Results were visualized with Phandango v.1.3.0 (43) and Figtree v.1.4.4 (https://github.com/rambaut/Figtree). (B) A multiple alignment of the five mycobacterial genomes was performed with Parsnp and visualized using Gingr (37) to show collinearity and SNVs in a 79-kb region spanning nt coordinates 130000 to 209000 in the mc²155 genome. In the upper panels, vertical purple bars indicate SNVs present in each genome, using mc²155 as the reference sequence, which is depicted in the middle row as a white bar. Consecutive panels zoom in on the indicated regions. The bottom panel is at single-nucleotide resolution, and the colors correspond to GCAT (yellow, blue, green, and red, respectively). Bold colors are SNVs, and faintly colored vertical bars match the mc²155 sequence.

FIG 4

Circos plots showing the distribution of parental DNA in sequenced transconjugants. (A) Crosses between defined donor and recipient strains in which transfer is unidirectional. (B) Transconjugant progeny of MKD8 and Jucho strains in which DNA transfer is bidirectional. Each parent genome is color-coded (as in Fig. 2) to indicate the source and visualize the distribution of transferred DNA in the transconjugant genomes. All crosses are between a Km-resistant donor and an Str-resistant recipient. The Km gene is inserted at the attL5 locus at ∼8 o’clock on the chromosome. Transfer of this DNA segment can be seen against the recipient background in most crosses. In a few examples, the transferred segment was too small to be resolved in the Circos plot.

FIG 5

Comparative gene analysis of the esx1 regions of the five M. smegmatis genomes. The schematic shows the overall collinearity and conservation of esx1 while highlighting the diversity of the mid region. The non-mid esx1 genes (Msmeg0055–0068 and Msmeg0076–0083) encode highly conserved proteins (>96.6% amino acid identity). However, proteins encoded by genes between these regions are poorly conserved (<30% amino acid identity) and include gene rearrangements, duplications, and multiple insertion sequence (IS) elements (mid genes within this region are boxed [Msmeg0069–0071]). Remarkably, mc²155 and Jucho are identical throughout the region, with 2 nucleotide differences from mc²155 in Msmeg0067 (resulting in an Arg-to-Pro amino acid change) and eccE (a silent C-to-T nucleotide substitution). Comparisons of esx1-encoded proteins were generated by Clinker using a best-BLAST-hit approach to identify orthologous genes at the 5′ and 3′ ends of each region (44). Orthologous genes are color-coded, and ISs and remnants of ISs are shaded in gray. Vertical lines drawn in the same color connect homologs. Note that the low amino acid conservation in the N terminus of Msmeg0071 prevented Clinker from identifying the complete gene, which we indicate here with a green striped box. Similarly, the low conservation of Msmeg0069 resulted in two classifications, identical in mc²155 and Jucho (olive arrow) and related but depicted in blue in Rabinowitchi, Nishi, and MKD8, immediately upstream of Msmeg0070–0071 orthologs. The names of conserved esx genes are indicated at the top of the alignment, and M. smegmatis numerical gene identifiers are shown at the bottom for reference.

FIG 6

mid confers self-identity and determines mating compatibility. (A) Map of esx1 and mid indicating the relevant genes used in the study. (B) Mating pairs boxed inside the shaded rectangle generated transconjugants at high frequencies (>1 × 10⁻⁴ per donor) (DCT “+”). Unboxed mating pairs produced no or very few transconjugants (<1 × 10⁻⁹ donors). In all crosses, parent donor and recipient counts were similar.

FIG 7

Msmeg0070 confers self-identity. (A) Mating pairs used in the analysis and their transfer proficiency. Pairs boxed inside a shaded rectangle were transfer proficient (DCT +). Unboxed mating pairs produced no or very few transconjugants (<1 × 10⁻⁹ donors). In all crosses, viable donor and recipient counts were similar. (B) Image of transconjugant-selective plates at a 10⁻¹ dilution for the donor-recipient pairs indicated. Crosses between mc²155 and Jucho are nonproductive (bottom left plate), but disruption of Msmeg0070 in mc²155 results in high transfer frequencies, which can be suppressed by the ectopic expression of Msmeg0070. Note the hyperconjugative phenotype of the Msmeg0070 mutant with MKD8 (top, compare the left and middle plates), which is also partly suppressed by expressing Msmeg0070 (right plate).