T6SS is a bacteria strategy to compete for a niche. The gene clusters of this system are often located in genomic islands, which have the potential to be transferred, as a unit, to other cells. To date, T6SSs have been identified in several genera of six phyla of Gram-negative bacteria, Acidobacteria, Bacteroidetes, Gemmatimonadetes, Nitrospirae, Planctomycetes, and Proteobacteria, being abundant in the latter [6, 9, 11, 12]. This wide (but not ubiquitous) distribution and diversity of T6SS in different genera of these phyla suggest an association of T6SS with horizontal gene transfer. However, so far, only twenty-nine plasmids with T6SS had been reported [2]. Even so, despite being in mobile elements, plasmid T6SSs can be functional [13–15].
Here, we mined hundreds of T6SS-harboring plasmids, and in dozens of cases, the bacteria did not have chromosomal T6SS, indicating and reinforcing the role of carrier elements, such as plasmids, in the dispersion of this secretion system. Furthermore, although present in the chromosome of several phyla and genera [11], indeed, the distribution of T6SS in plasmids is limited, as only ~1% of them encoded this secretion system, mainly Proteobacteria. Several factors may contribute to this phenomenon: (i) the dissemination of T6SS via plasmids, at least in Proteobacteria, seems to have barriers, since bacteria with chromosomal T6SS (abundant in Proteobacteria) may present a defense mechanism via T6SS against the acquisition of new plasmids [2]; (ii) carrying an extra copy of T6SS does not seem advantageous if the bacterium already has a chromosomal copy, as it is a niche-specific system and different T6SSs do not confer different functions, depending more on the effectors that are secreted [4]; (iii) since T6SS-harboring plasmids have a large median size (1.6 Mb), this would likely impose a high fitness cost. Previously, Abby et al., (2016) showed that chromosomal T6SS was more prevalent in γ-Proteobacteria than in α- and β-Proteobacteria, and curiously, here, we observed that the plasmid T6SS prevail in α- and β-Proteobacteria. On the other hand, plasmids with T6SS were prevalent in γ-Proteobacteria when considering those carrying less than nine T6SS genes, which suggests that smaller clusters of T6SS would be more common in γ-Proteobacteria plasmids and that once these elements are stabilized in the chromosome, the T6SS plasmid copy would undergo a process of degradation and eventually lost, which could result in a low prevalence of T6SS in such mobile elements. Furthermore, in the loose analysis, several non-Proteobacteria plasmids harbored a T6SS gene, TssE, which is homologous to the bacteriophage T4 gp25 baseplate [16]. Thus, it is likely that these non-Proteobacteria plasmids are not related to T6SS. Here we also observed T6SS in a few plasmids from other phyla, in addition to Proteobacteria, Acidobacteria, and Gemmatimonadetes, which may indicate their acquisition from other bacteria in the environment.
Of the four phylogenetic T6SS types, type i was the only one found in plasmids. Furthermore, this type is the most common in Proteobacteria [11]. This scenario is evidence that plasmids have irradiated T6SSi, at least, in Proteobacteria. Looking at our dataset and the T6SS subtypes, types are prevalent i and ii on chromosomes, while types i4b and i3 prevail on plasmids. Of note, there was a subcluster in the subtype iii cluster with dozens of plasmids sequences that were not closely related to any reference chromosomal sequence. Most of these sequences belonged to environmental bacteria and could be evolving independently of the others of subtype iii.
Considering the gene cargo of the analyzed plasmids, we did not observe in most of them a prevalence of resistance or virulence genes (disregarding the T6SS). Thus, unless these T6SSs play a virulence role in their host niche, these plasmids would be more related to some ecological role. Even because some of them also encode secondary metabolites related to survival and protection. Among the T6SS effectors identified in the plasmids, most of them would be related to virulence, however, modA is associated with nitrate metabolism and anaerobiosis, being an important element for plant-associated bacteria [17]. Interestingly, this effector was found only in Azospirillum plasmids (9/13). These ecological gene cargos contrast with virulent T6SS-harboring plasmids from clinical bacteria, such as Cronobacter spp. and Campylobacter jejuni [18, 19]. The few plasmids identified carrying resistance and virulence genes were mainly associated with bacteria isolated from human or animal hosts. Indeed, clinical T6SS-positive bacteria were observed to have a higher resistance and frequency of virulence genes [20]. Although 92/303 of these T6SS-carrying plasmids have been characterized as conjugative or mobilizable [21], their median size (~601 kb) would represent a natural restriction to transmission. Thus, these mobility genes may be part of integrative elements, such as integrative conjugative elements (ICEs) [22]. Indeed, ICEs have been predicted in some plasmids characterized as conjugative (e.g., NZ_CP064859.1, NZ_CP057783.1, NZ_CP017887, and NZ_CP053574.1). Furthermore, recently, it was shown in Bacteroidales that T6SS presents an extensive intra-ecosystem transfer and multi-species spread due to its association with ICEs [23]. Therefore, in addition to plasmids, other mobile platforms, such as ICEs, may be involved in the spread of T6SS. Finally, in dozens of these T6SS-harboring plasmids we identified genes associated with chromosomes, such as rRNA, and this, added to the fact that most of them are megabases in size, raised the question of whether they were in fact plasmids or another type of replicon. In fact, some of the genera identified here were associated with secondary essential replicons (secondary chromosomes or forming chromosomes), such as Burkholderia, Cupriavidus, Ensifer/Sinorhizobium, Pantoea, Ralstonia, Rhizobium, Vibrio [24].
Therefore, our findings do not fully support the hypothesis that T6SS irradiation within bacteria was plasmid-mediated, as occurred with T7SS in Mycobacteriaceae [10]. Even so, the evidence gathered here points to the involvement of mobile platforms in the spread of T6SS within bacteria.