FIC domain toxins are the origin of intra- and inter-kingdom effectors of Bartonella

Bacterial secretion systems are machineries that allow one cell to take direct control of molecular processes in another one so that they are ubiquitously involved in the host interaction of mutualistic, commensal, and pathogenic bacteria. However, the evolutionary origin of these systems lies in genuine bacterial machineries. One of the most well-studied host-interacting secretion systems is the type IV secretion system (T4SS) that originally evolved to mediate bacterial conjugation, i.e., the interbacterial transfer of DNA. During conjugation, the DNA is transferred covalently linked to a protein called relaxase that is the actual substrate of the type IV secretion machinery.
It is the common focus of most studies that come together in this thesis to unravel the adaptive path by which the host-interacting VirB/D4 T4SS of Bartonella and its secreted effectors evolved from conjugative ancestors. In short, these studies allowed to propose a two-step model in which an ancestral conjugation machinery first acquired an interbacterial effector protein and was later exapted for host interaction which resulted in the evolution of the VirB/D4 T4SS.
In Review article I we comprehensively reviewed available literature regarding the molecular pathogenesis of the α-proteobacterial pathogen Bartonella and concluded that different phylogenetic lineages of this genus use partially divergent sets of virulence factors for essentially the same stealth infection strategy. Though the overall course of infection in their respective mammal reservoir hosts is similar, the bartonellae showed major differences in their host adaptability that correlated with the presence of the VirB/D4 T4SS in the promiscuous lineages. This machinery contributes to Bartonella virulence by secreting a cocktail of effector proteins (Beps) into mammalian host cells where they manipulate cellular signaling processes in favor of the pathogen. Previous studies had shown that all effectors of the VirB/D4 T4SS evolved from a single common ancestor that consisted of a FIC protein domain fused to a conserved type IV secretion signal which is also the most prevalent domain architecture among extant Beps. FIC domains are enzymatic domains that typically mediate adenylylation, the transfer of an adenosine 5’-monophosphate onto target proteins, suggesting that a prototypic effector of the VirB/D4 T4SS may have targeted host proteins by post-translational modification. Interestingly, a bacterial conjugation system called Vbh T4SS had been discovered in Bartonella and is encoded on a plasmid together with a Bep-like protein consisting of FIC domain and type IV secretion signal that had been named VbhT.
In Research article I we investigated the phylogeny and regulation of Fic proteins (i.e., proteins containing FIC domains) and found that their adenylylation activity is controlled by a conserved mechanism of active site obstruction via an inhibitory α-helix (αinh). Depending on the positioning of αinh either in a separate polypeptide or N- or C-terminally of the FIC domain in the same protein we classified FIC domain proteins as class I, class II, and class III, respectively. Fic proteins have repeatedly evolved into bacterial virulence factors, but the far majority of them are genuine bacterial proteins of unknown function. We showed that the mutational activation of bacterial Fic proteins of all three classes by abrogation of active site obstruction results in bacterial growth inhibition, suggesting that most Fic proteins may be some kind of toxins. Interestingly, the elusive Bep-like VbhT protein of the Vbh T4SS was found to inhibit bacterial growth via target adenylylation unless it is inhibited by a small protein VbhA that contains the αinh.as a hallmark of class I Fic proteins. This arrangement is reminiscent of toxin-antitoxin (TA) modules that are genetic elements consisting of a toxin that inhibits bacterial growth and an antitoxin that suppresses the toxin’s activity but can unleash it in order to induce a phenotypic switch to the dormant persister state.
In Review article II we firmly established that class I Fic proteins form a new toxin-antitoxin (TA) module FicTA with FicT toxins being homologous to the VbhT FIC domain and FicA antitoxins being homologous to VbhA. A follow-up study described in Research article III characterized the FicTA toxin-antitoxin module and discovered that FicT toxins inhibit bacterial growth via the adenylylation and concomitant inactivation of DNA gyrase and topoisomerase IV (topo IV), the two bacterial type IIA topoisomerases. The resulting disruption of cellular DNA topology induced a phenotypic switch to the persister state. Although an important role of TA modules in persister formation is well established, previously known TA systems invariably induced dormancy by inhibiting translation or unsetting the proton-motive force. Our results therefore uncovered a new physiological path to the persister state that is likely involved in the inherent physiological heterogeneity of persisters as a main obstacle to their eradication. Furthermore, we found that the Vbh T4SS in Bartonella is sometimes associated to a functional FicTA module that is encoded at the locus where VbhTA would be expected but shows no trace of a type IV secretion signal.
The concluding Perspective section of this work combines the results of the aforementioned articles with a considerable amount of unpublished data and proposes a model that can explain the evolution of the VirB/D4 T4SS and its effectors in Bartonella from conjugative ancestors. In short, it is evident that an ancestral conjugative T4SS as well as a FicTA module (supplying the FIC domain for effector evolution) entered the genus Bartonella via horizontal gene transfer, though likely not together. After sequence reshuffling events had resulted in a ficAT locus encoded next to a conjugative Vbh-like machinery (like in some extant bartonellae), one of the two type IV secretion signals of the relaxase was transferred onto the FicT toxin via terminal reassortment and created a situation identical to the Vbh T4SS with VbhT. Based on additional evidence, we propose that extant Vbh T4SS and VbhT represent such an ancestral evolutionary state as a “living fossil”. Furthermore, we suggest that VbhT is an interbacterial effector that is secreted via the Vbh T4SS during bacterial conjugation in order to promote complementary strand synthesis in the recipient. Like this, the Vbh T4SS with VbhT constitutes a “missing link” in the evolution of the host-interacting, effector secreting VirB/D4 T4SS from purely conjugative ancestors. The remarkable sequence similarity between Vbh T4SS and VbhT on one side and the VirB/D4 T4SS and Beps on the other side suggests that such a conjugative machinery with interbacterial effector may have been exapted for host interaction in a second step. During that process, a primordial vbhT like effector gene served as the template for a series of gene duplication and diversification events that created the extant effector repertoires by gradual evolution. The different paralogous effectors created by this process were generally suspected to have different roles in host interaction. Among other evidence, I could show earlier that one effector (Bep1) targets the host protein Rac1 by adenylylation. In Research article II we followed up on these results and identified the adenylylated target of another effector, called Bep2, as vimentin. For this purpose we had developed a new technique that identifies adenylylated target proteins by mass spectrometry using a characteristic pattern of peaks upon adenylylation with a mixture of heavy-isotope labeled ATP. Together with additional data of others and myself, these results suggest that the diversified effector repertoires of the VirB/D4 T4SS evolved to promote the stealth infection strategy of Bartonella by fine-tuned manipulations of host cell signaling. In particular, the modular architecture of both the type IV machinery and its effectors conferred an inherent evolvability that is likely involved in the remarkable host adaptability of those bartonellae that encode VirB/D4 T4SS and effectors.
The most important open point in this model is the biological function of VbhT as the first bona fide interbacterial type IV secretion effector. Future studies should address a possible role of VbhT in bacterial conjugation in order to strengthen the claim of a “missing link”.
In Research article IV we expanded the understanding of FIC domain proteins without obvious connection to Bartonella. This study was focused on NmFicT of Neisseria meningitidis as a model for class III Fic proteins that encode the αinh at their C-terminus, but contain no additional domains or modules for their regulation. We reported strong evidence for a model in which NmFicT is controlled by a double-lock mechanism where the first level of control is tetramerization that generates an inactive storage form of the protein. Upon an unknown trigger, NmFicT monomers would be freed from oligomerization and then remove the second lock via intermolecular auto-adenylylation which abolishes αinh –mediated active site obstruction and unleashes the full catalytic activity of NmFicT. Furthermore, we showed that activated NmFicT adenylylates DNA gyrase at the same residue as FicT toxins, though these also adenylylate topo IV and primarily inhibit the latter target in vivo. The biological function of NmFicT is unclear, but may be related to DNA repair and protection as described for other proteic gyrase inhibitors.