- Research article
- Open Access
Fratricide activity of MafB protein of N. meningitidis strain B16B6
BMC Microbiology volume 15, Article number: 156 (2015)
Neisseria meningitidis is an inhabitant of the mucosal surfaces of the human nasopharynx. We recently demonstrated that the secreted meningococcal Two-partner secretion protein A (TpsA) is involved in interbacterial competition. The C-terminal end of the large TpsA protein contains a small toxic domain that inhibits the growth of target bacteria. The producing cells are protected from this toxic activity by a small immunity protein that is encoded by the gene immediately downstream of the tpsA gene. Further downstream on the chromosome, a repertoire of toxic modules, designated tpsC cassettes, is encoded that could replace the toxic module of TpsA by recombination. Each tpsC cassette is associated with a gene encoding a cognate immunity protein.
Blast searchers using the toxic domains of TpsA and TpsC proteins as queries identified homologies with the C-terminal part of neisserial MafB proteins, which, for the rest, showed no sequence similarity to TpsA proteins. On the chromosome, mafB genes are part of genomic islands, which include cassettes for additional toxic modules as well as genes putatively encoding immunity proteins. We demonstrate that a MafB protein of strain B16B6 inhibits the growth of a strain that does not produce the corresponding immunity protein. Assays in E. coli confirmed that the C-terminal region of MafB is responsible for toxicity, which is inhibited by the cognate immunity protein. Pull-down assays revealed direct interaction between MafB toxic domains and the cognate immunity proteins.
The meningococcal MafB proteins are novel toxic proteins involved in interbacterial competition.
The cell envelope of Gram-negative bacteria consists of two membranes, the inner and outer membrane, which are separated by the periplasm containing a peptidoglycan layer. Different transport machineries have evolved for delivering proteins from their site of synthesis in the cytoplasm to the bacterial surface and beyond. One of them is the two-partner secretion (TPS) system, which is broadly used in many Gram negatives for the specific secretion of usually very large proteins, generically called TpsA . After transport across the inner membrane via the Sec system, the TpsA protein is translocated across the outer membrane via a dedicated transporter, called TpsB. Recognition between the proteins is mediated by a specific N-terminal domain of ~300 amino acid residues in TpsA, the TPS domain, which interacts with a periplasmically exposed domain of the outer-membrane-embedded TpsB protein .
In several microorganisms, including Escherichia coli and Neisseria meningitidis, TpsA functions in inhibiting the growth of related bacteria in competition for the same niche in a process called contact-dependent growth inhibition (CDI) [2–4]. In the proposed model , the surface-exposed TpsA interacts with a conserved receptor, BamA, on a target cell, after which a small C-terminal part of TpsA is proteolytically released and transported into the target cell in a process that also requires the inner-membrane protein AcrB . An alternative import pathway requires the interaction of the toxin with the F pilus and pilus retraction . In the target cell, the toxin, can exert different toxic activities, e.g. as a DNase or an RNase. The gene immediately downstream of tpsA, called tpsI, encodes a small immunity protein that protects the producing cell against the toxic activity of the C-terminal domain of TpsA. On the chromosome, the tpsB, tpsA and tpsI genes are present on genetic islands often additionally containing a number of tpsC cassettes . These tpsC cassettes potentially encode N-terminally truncated TpsA proteins, which, however, present an entirely different toxic module at the C terminus. Each tpsC cassette is associated with a cognate tpsI gene. Because of the N-terminal truncation, these putative TpsC proteins lack the sequences necessary for secretion and it is not sure whether they are expressed. However, the tpsC cassettes can recombine with the tpsA locus, thereby replacing the toxic module present at the C terminus of TpsA . Thus, TpsA constitutes an interbacterial competition system that can use a broad repertoire of toxic modules.
Other secretion systems found in Gram-negative bacteria are also meant for inhibiting competing bacteria or even eukaryotic cells. Examples include RhsA (rearrangement hot spot) of E. coli  or the broadly distributed Type VI secretion system . These growth inhibition systems present similar toxic modules at the C terminus of the exported proteins as found in the TpsA proteins, but show no further sequence similarity with TpsA, consistent with a different secretion mechanism. In the present study, we demonstrate that the MafB proteins of Neisseria spp., previously thought to function as adhesins , present similar toxic modules at their C terminus as the TpsA proteins, but show no further sequence similarity to TpsA. We demonstrate that these MafB proteins represent a novel growth inhibition system in the meningococcal strain B16B6 that functions in interbacterial competition. Whilst this manuscript was in preparation, another study of the MafB proteins of Neisseria spp. was published . For clarity, we have adopted the nomenclature for the Maf proteins of that study.
Structural organization of meningococcal Maf islands
BLAST searches using different toxic domains of various meningococcal TpsA and TpsC sequences as queries yielded hits with various TpsAs and TpsCs of different bacterial species. Additional hits were also retrieved with the C termini from a large variety of other proteins, including neisserial MafB proteins. MafB proteins are present in different Neisseria spp., including N. meningitidis, N. lactamica and N. gonorrhoeae, and they were earlier assigned as members of a multiple adhesin family of proteins thought to be involved in adhesion to host cells . The sequence similarity of MafB with TpsAs or TpsCs is restricted to the C-terminal toxic module, indicating that MafB is not secreted via a TPS mechanism. Inspection of its genetic context in available genome sequences indicated that the mafB genes are components of genetic islands. The genes in the islands may form an operon composed, from 5’ to 3’ end, of mafA, mafB and a variable number of mafB-related genes, potentially encoding alternative C-terminal domains and designated mafB-CT. These mafB and mafB-CT genes are interspersed with one or more intervenient ORFs, which may encode immunity proteins (designated mafI) (see Fig. 1 for examples). Up to three independent maf genetic islands, present on different chromosomal locations, can be recognized in meningococcal genomes (Fig. 1), designated MGI-1, 2, and 3 according to a recent proposal . The predicted MafA proteins contain a lipoprotein signal sequence, and phylogenetic analysis of MafA proteins from different strains of various Neisseria spp. revealed clustering of the sequences in two phylogenetic groups (Fig. 2) with > 95 % of identity within each group and < 70 % identity between groups. MafB proteins contain a predicted N-terminal signal sequence, and the mature part is further structured into three regions: an N-terminal DUF1020 domain of ~ 260–320 aa in length, a central region of ~140 aa containing a Hint domain, which is usually involved in autocatalytic protein splicing, and a variable C-terminal toxic domain of ~80-100 aa. The N-terminal DUF1020 domains of MafB proteins can be clustered in three variants, I, II and III, where variants I and II are phylogenetically more related (Fig. 2). The central region is conserved in sequence, but its presence is variable as some MafBs lack the Hint domain. The C-terminal region is polymorphic and constitutes an independent toxic domain. The mafB-CTs are variable in length (260–893 bp). At their 3’ end, they contain sequences that putatively encode a toxic domain, whilst upstream thereof they share sequence similarity with sequences adjacent to the toxic region of mafB genes, including sometimes an intact Hint domain. Some mafB-CTs, for instance NMBG2136_0586 of strain G2136 (Fig. 1), encode a VENN domain previously recognized as being adjacent to the toxin domain in TpsA proteins . The mafB-CT genes generally lack sequences corresponding to the 5´ end of mafB genes, but occasionally, a complete mafB is found among the mafB-CT genes, for example NMO_0056 or NMC2090 in MGI-3 of strains Alpha14 and FAM18, respectively (Fig. 1). The shared sequences between mafB and mafB-CT genes may be used for genetic rearrangements, resulting in the substitution of the toxic domain at the C terminus of MafB. Examination of various neisserial genome sequences suggests that such recombination events have indeed occurred in several cases, since the C-terminal domains of some MafB proteins have > 95 % of sequence similarity with predicted MafB-CT proteins from other genomes. For example, the 3’ end of mafB-CT NMBG2136_0598 in MGI-2 of G2136 is placed at the 3’ end of the mafB NMC1790 in MGI-1 of strain FAM18 (Fig. 1). In general, such rearrangements also affect the position of the immediately downstream located mafI gene, which remains associated with the 3’ end of the upstream mafB or mafB-CT gene, supporting the notion that these modules function as specific toxin-antitoxin pairs. However, in some islands, several mafIs are associated with a specific mafB or mafB-CT (Fig. 1); it is not clear whether they are all required to confer immunity against the toxic domain of the corresponding MafB or MafB-CT.
Examination of the genomic islands revealed that mafAs of phylogenetic group I are always immediately followed by a mafB of group I, whilst islands containing a mafA of group II always contain an adjacent mafB of group III (Fig. 1). Remarkably, the mafB genes of group II (NMBG2136_0582, NMC0586 and NMO_0529 of strains G2136, FAM18, and Alpha14, respectively, in Fig. 1) are not directly adjacent to a mafA, but they are located on an island (MGI-2) usually also containing a mafA of group I and a mafB of group I. Furthermore, the mafB-CTs only share sequences with the upstream located mafB or mafBs of the same phylogenetic group located on a different island (see Fig. 1 for examples), suggesting the existence of independent mafB/mafB-CT systems with genetic exchange between mafB and mafB-CTs being restricted to each system.
Overall, this analysis revealed that Maf systems share several genetic features with the TPS systems involved in interbacterial competition. Accordingly, it is logical to hypothesize that (i) MafB is a secreted protein with growth-inhibitory properties, (ii) MafI proteins confer specific immunity against the toxic MafB, and (iii) mafB-CTs represents a source of switchable toxic modules.
Characterization of MGIs in strain B16B6
To investigate our hypotheses, we used N. meningitidis strain BB-1, an unencapsulated derivative of B16B6 of the clonal complex cc11 , in which TpsA-mediated growth-inhibitory activity was previously demonstrated [4, 12]. Extensive PCR analysis and subsequent sequencing of PCR fragments revealed that this strain contains three MGI corresponding with islands MGI-1 (GenBank accession number KR006907), MGI-2 (GenBank accession number KR012999), and MGI-3 (GenBank accession number KR006906) in other meningococcal genomes (Fig. 3). Comparison of the genetic organization of the MGIs of strains FAM18 and B16B6, which are both of the same clonal complex (cc11), revealed considerable differences (Fig. 3). The genetic composition of MGI-3 is similar in both strains, with > 99 % of sequence identity. However, comparison of islands MGI-1 and MGI-2 of both strains revealed complex rearrangements. In MGI-2, the toxic domain encoded by the first mafB gene (mafB1 MGI-2 ) and the downstream mafI and mafC genes differ between the two strains (Fig. 3). In contrast, the second mafB gene in this island (mafB2 MGI-2) contains a similar 3’ end in the two strains. However, downstream of mafB2 MGI-2 and its cognate mafI, there are large differences in the mafB-CT and mafI content of the two strains (Fig. 3), and the organization in this part of island MGI-2 of strain B16B6 is actually very similar to that of MGI-2 of cc8 strain G2136 (Fig. 1), except that the IS4 transposase gene is lacking. In island MGI-1, the 3’ end of the mafB of B16B6 is different from that of FAM18 (Fig. 3) but similar to that of the second mafB in MGI-2 of strain G2136 (locus tag NMBG2136_0596; Fig. 1). After the cognate mafI, the organization of mafB-CT and mafI genes in MGI-1 of B16B6 is also deviant from that in FAM18 (Fig. 3) but similar to that in the corresponding island of strain MC58 (compare Figs 1 and 3).
The very different organization of the MGI-1 and MGI-2 islands in cc11 strains B16B6 and FAM18 is striking, since comparative analysis of the TPS island in these strains and a large collection of other cc11 strains evidenced only a very low frequency of genetic rearrangements . To confirm the genetic variability of the maf loci, we extended our analysis to available genome sequences of nine other cc11 strains, i.e. ES14902, M6190, M7124, NM3682, NM3683, NM126, M10208, NM174 and NM82. In all these strains, the C terminal toxic domains of MafB2MGI-2 and MafBMGI-3 are similar to those of FAM18 and B16B6, whilst that of MafBMGI-1 is similar to that in B16B6 but different from that in FAM18. In two strains, i.e. NM3682 and NM3683, the toxic domain of MafB1MGI-2 is similar to that of B16B6, while it is similar to that of FAM18 in the other seven strains. Thus, these results suggest that the C-terminal toxic domain of MafB of phylogenetic cluster III, which is encoded by MGI-3, is rather stable in strains of the same cc11 lineage, as previously found for TpsA , but the MafBs of clusters I and II, which are encoded by MGI-1 and MGI-2, seems to be more prone to alteration.
MafB functions in meningococcal growth inhibition
To investigate the putative role of MafB proteins in mediating growth inhibition, we first constructed several mutants of strain BB-1, lacking diverse genes in different maf islands. These mutants were mixed with a rifampicin-resistant derivative of the parent strain in different ratios and the numbers of wild-type and mutant bacteria were determined by plating on selective media at time 0 and after different periods (3 –5 h) of co-incubation at 37 °C while shaking. However, under these conditions, we did not observe inhibition of the growth of the mutants relative to the parent. Therefore, we next performed these assays on GC plates in a candle jar, conditions that were previously found to be optimal to detect TpsA-mediated growth inhibition . Under these conditions, growth of a mutant lacking the complete MGI-3 island (ΔMGI-3; indicated in blue in Fig. 4) and of a mutant lacking the entire MGI-3 island except for mafA (data not shown) were significantly reduced in the presence of the parent, whilst that of a mutant lacking only mafA of MGI-3 was not inhibited (ΔmafA MGI-3 ; indicated in black in Fig. 4). When a mutant lacking mafB2 and all downstream genes of island MGI-2 (ΔmafB2IB-CTs MGi-2 ; green in Fig. 4), or a mutant lacking all genes except for mafA of island MGI-1 (ΔmafBIB-CTs MGI-1 ; red in Fig. 4), were tested in similar conditions, no significant growth inhibition by the parent was detected. These results demonstrate that at least island MGI-3 mediates growth inhibitory activities.
mafB and mafI genes encode toxins and cognate immunity proteins
To investigate whether indeed the C-terminal part of MafBMGI-3 is responsible for the growth-inhibitory effect observed in the in vivo assays, we cloned the DNA for this fragment from strain B16B6 with a FLAG-epitope tag into a pAYC184-derived vector behind an arabinose-inducible promoter. Although we did not detect growth inhibitory activity in vivo of the MafBs of phylogenetic cluster I encoded by islands MGI-1 and MGI-2, we decided to clone also, as a representative, the DNA for the C-terminal fragment of MafB2MGI-2 to investigate whether also this protein is a toxin. It is noteworthy that this domain is recognized in the Conserved Domain Database (CDD) as a toxin64 domain with a possible RNase function. In contrast, the C-terminal domain of MafBMGI-3 of B16B6 was not recognized in CDD as a putative toxic domain, but it was predicted to be a toxin_56 (putative RNase) with 95 % probability using the HHpred server. We also cloned the mafIs located immediately downstream of these mafBs on the chromosome into a pET26b vector behind an isopropyl-β-D-1-thiogalactopyranoside (IPTG)-inducible T7 promoter; the resulting recombinant proteins contain a His6 tag on their C-terminal end. The plasmids were introduced in E. coli strain BL21 (DE3). The expression of the recombinant proteins was independently controlled with specific inducers and tested by Western blotting, and cell viability after overnight induction was tested by plating the bacterial suspensions on LB plates with appropriate antibiotics. The expression of the C-terminal fragment of MafB2MGI-2 in E. coli resulted in a drastic reduction of cell viability compared to the non-induced control (Fig. 5a). Viability was restored when the cognate MafI2MGI-2 was co-expressed, but not by co-expression of MafIMGI-3. Similarly, the expression of the C-terminal part of MafBMGI-3 resulted in a drastic reduction of cell viability, which could be restored by co-expression of the cognate MafIMGI-3 but not of the non-cognate MafI2MGI-2 (Fig. 5a). These assays confirmed that indeed the C-terminal parts of both MafBs are toxic and that their activity is specifically inhibited by the cognate MafI.
Usually, the immunity proteins of toxin-antitoxin systems confer protection through their direct interaction with the cognate toxins. To investigate whether the MafI proteins interact with the C-terminal fragments of the corresponding MafBs, pull-down assays were performed. The His-tagged immunity proteins were purified on Ni2+ beads from E. coli cells also expressing the C-terminal MafB fragments, and the co-purification of the MafB fragments was examined on Western blots. As anticipated, both MafB2MGI-2 and MafBMGI-3 C-terminal fragments were purified together with their cognate immunity proteins but not with non-cognate MafI (Fig. 5b). Thus, these assays confirmed the role of MafI in conferring protection by specifically interacting with C-terminal domain of the cognate MafB.
Toxins play an important role in the competition between bacteria for limited resources. They are produced to inhibit the growth of competitors for the same niche and, in addition, they are used for self recognition in a bacterial community. Bacteria have developed several independent competition systems, including bacteriocins, TPS systems, type 6 secretion systems, Rhs, or the ESX systems in Gram-positive bacteria [13–15]. The neisserial Maf systems appear to share features with several of these previously characterized competition systems. Growth inhibition in these systems is mediated by a cell-surface-exposed or secreted protein, i.e. MafB in the case of the Maf system. Like in many other systems, the C-terminal end of MafB contains a polymorphic toxic module, whilst the rest of the protein might be required for its secretion and/or for its interaction with receptors on the target cells. Producing cells are protected against the toxic activity by a small protein, i.e. MafI in the Maf system, which interacts with the toxin and is usually encoded by the gene immediately adjacent to the toxin gene. Like in several other systems, the mafB and mafI genes are located on genetic islands together with a variable repertoire of genes putatively encoding N-terminally truncated toxins, called mafB-CTs in the Maf system , with the cognate immunity genes. By genetic rearrangements, these mafB-CTs may substitute the 3’ end of the mafB gene, thereby replacing its toxic activity, whilst the array of encoded immunity proteins may confer protection against a variety of toxins that may be encountered in the natural habitat. Indeed, a variety of toxic domains may be encoded at the 3’ end of the same mafB gene in different genomes, and the same domains may be found at the 3’ end of mafB-CT genes in other genomes (Fig. 1), clearly supporting the idea that such rearrangements occur in vivo. The substitution of the toxic module may offer a selective advantage in a given bacterial community [4, 16] but may also lead to the loss of immunity genes . Indeed, an analysis of the frequency of substitution of toxic domains in the tpsA loci of N. meningitidis evidenced that this event occurs but is very rare in isolates of the same genetic lineage and that such alteration may generate a competitive disadvantage relative to other members of the community . Similarly, our analysis of a limited set of strains of the same genetic lineage in the current study evidenced no alterations at the C terminus of MafB proteins of phylogenetic group III encoded on MGI-3. However, rearrangements were evident in MafB proteins of phylogenetic groups I and II encoded on MGI-1 and MGI-2. Note that the mutual similarity between MafB proteins of groups I and II is higher than their similarity to group 3 (Fig. 2). It is conceivable that MafBMGI-3 plays a role in the maintenance of the bacterial community, while MafB proteins encoded on MGI-1 and MGI-2 may offer a selective advantage to members of a community only under specific conditions. Perhaps, this may explain why we detected growth inhibitory activity only of MafBMGI-3 and not of MafBMGI-1 and MafB2MGI-2. Possibly, the latter MafB proteins or their cognate receptors on the target cells are not expressed under the experimental conditions used. Alternatively, these proteins may target other bacterial species or eukaryotic cells. This question remains to be elucidated in future work.
The secretion mechanism of MafB was not elucidated in this work. MafB proteins were detected in the culture supernatant of N. meningitidis strain when overexpressed from a plasmid but the secretion mechanism was not solved . MafB proteins are synthesized with an N-terminal signal sequence for transport across the inner membrane via the Sec system. Furthermore, all mature MafB proteins contain a conserved DUF2010 domain at their N-terminal end, which may constitute a recognition signal for the machinery that mediates its subsequent transport across the outer membrane. The MafA proteins, previously reported to function as glycolipid-binding adhesins , are likely candidates to constitute this machinery. Interestingly, our phylogenetic analysis revealed two main clusters of MafA variants, which paralleled the clustering of the MafB proteins into two main clusters based on their DUF1020 domains, suggesting specificity in the recognition of MafB proteins by the cognate MafA. However, MafA is a predicted lipoprotein and secondary structure predictions were inconsistent with a β-barrel conformation, which is common among integral outer membrane proteins, including, for example, the TpsB proteins, which mediate the transport of TpsA proteins across the outer membrane in the TPS systems . Perhaps, the quaternary structure of MafA generates a channel in the outer membrane as reported, for example, for the CsgG protein, which mediates the translocation of curli subunits across the outer membrane .
Some MafB proteins contain a Hint domain in between the DUF1020 domain and the toxic domain. Such domains are responsible and sufficient for protein splicing and could have a role in toxin processing and delivery into the target cells. There are two types of Hint domains, designated type A and B, based on sequence motifs . The Hint domains, if present in MafB proteins, are of type A. Proteins containing this domain undergo splicing and release of the C-terminal part . However, most MafB proteins do not contain such a Hint domain; how the C-terminal toxic moiety is released and transported into the target cells in those cases is not clear and remains to be investigated.
Whilst this manuscript was in preparation, an independent study reported the toxic activity of the C-terminal domains of some of the MafB proteins of another meningococcal strain, 8013, and the competitive advantage of overexpression of one of these MafB proteins , consistent with the observations reported in this paper.
In this work, we identified the neisserial MafB proteins as a novel family of toxic proteins involved in interbacterial competition. We demonstrated that the toxic activity of MafB of strain B16B6 resides in the C-terminal region, while MafI specifically interacts with this toxic domain to confer protection to the producing cells. Further studies will be directed to elucidate the secretion mechanism of MafB and to identify possible applications of this system to fight pathogens.
The C-terminal 150 aa of several meningococcal TpsA and TpsC proteins present in available genome sequences were used as input queries in BLAST searches. Analysis and comparisons of DNA and protein sequences were performed with Clone Manager software Suite 7. Alignment of protein sequences was performed in MAFFT version 7 (http://mafft.cbrc.jp/alignment/server/). Aligned sequences were then used for phylogenetic analyses using the Neighbour-joining method with the available MEGA software version 6 (http://www.megasoftware.net/). The matrices and phylogenetic trees derived from this study were deposited in Treebase repository (http://treebase.org/treebase-web/search/study/anyObjectAsRDF.rdf?namespacedGUID=TB1:S17890). The subcellular localization of proteins was predicted using the servers SignalP (http://www.cbs.dtu.dk/services/SignalP/) and LipoP (http://www.cbs.dtu.dk/services/LipoP/). Secondary and tertiary structure of the mature proteins were predicted using the public web-based programs Phyre2 (http://www.sbg.bio.ic.ac.uk/phyre2/html/page.cgi?id=index) and PSIPRED v3.3 (http://bioinf.cs.ucl.ac.uk/psipred/). For homology detection, the HHpred server was used at http://toolkit.tuebingen.mpg.de/hhpred.
Bacterial strains and growth conditions
N. meningitidis strain BB-1 is a non-capsulated derivative of strain B16B6 of clonal complex cc11 . A spontaneous rifampicin-resistant mutant of BB-1 was also described . The bacteria were grown at 37 °C on GC medium base (Difco) supplemented with IsovitaleX (Becton Dickinson) and appropriate antibiotics (100 μg ml−1 of kanamycin, 5 μg ml−1 of chloramphenicol, or 50 μg ml−1 of rifampicin) at 37 °C in a candle jar. For liquid cultures, bacteria grown on plates were diluted in tryptic soy broth (TSB; Beckton Dickinson) to an OD550 of 0.1 and incubated in 25-cm2 polystyrene cell-culture flasks with constant shaking at 110 rpm.
E. coli strains DH5α and BL21 (DE3) (Invitrogen) were grown in lysogeny broth (LB) or LB agar at 37 °C. When required, the media were supplemented with antibiotics (100 μg ml−1 of kanamycin, 25 μg ml−1 of chloramphenicol, or 10 μg ml−1 of tetracycline).
PCR amplification and sequencing
Chromosomal DNA was extracted from BB-1 and its mutant derivatives as described  and used as template DNA for PCRs. PCRs were performed using 1–2 μl of extracted DNA, 200 μM dNTPs (Fermentas), 0.25 μM of different primer combinations (see in Additional file 1: Table S1), 0.5 U of Expand High Fidelity Enzyme Mix, and PCR buffer of the Expand High Fidelity PCR System (Roche). Thermal cycling conditions were 30 cycles of 1 min at 95 °C, 0.5 min at 58 °C and elongation at 72 °C during 1 min per kbp of expected amplicon size. Reactions were preceded by incubation at 95 °C for 10 min and terminated with an extended elongation step for 10 min at 72 °C. The PCR products were separated by electrophoresis in 1 % agarose gels and stained with ethidium bromide for visualization. Some PCR products were cloned into the pCR2.1-TOPO vector (Invitrogen), according to the manufacturer’s instructions. Plasmids and PCR products were purified using the Plasmid Extraction kit and Clean-Up System, respectively, (Promega). Purified plasmids and PCR products of variable length were sequenced at the Macrogen sequencing service (Seoul, Korea). For sequencing the maf islands, several overlapping DNA fragments were amplified by PCR, and then sequenced either directly or after cloning in the TOPO vector using internal primers. Amplification reactions were carried out independently and repeated at least twice for each fragment. Assembly of sequences was performed with SeqMan II software (DNAstart Inc.).
DNA manipulation techniques
Cloning was performed in E. coli DH5α. All primers used for PCR amplifications with the restriction sites they contain are listed in Additional file 1: Table S1. To generate knockout constructs, DNA fragments upstream and downstream of the gene (s) of interest were amplified from chromosomal DNA of strain B16B6 as indicated in Fig. 3. PCR products were purified, digested with proper restriction enzymes, and sequentially cloned into pnhbA::kan  thereby replacing DNA fragments flanking the kanamycin-resistance cassette in the original plasmid. The knockout constructs were purified and used to transform strain B16B6. Appropriate PCRs were used to verify the presence of the correct mutations in kanamycin-resistant transformants.
For expression of recombinant proteins in E. coli, DNA fragments corresponding to the toxic regions (the C-terminal 100 aa residues) of the MafB proteins MafB2MGI-2 and MafBMGI-3 and the cognate MafI proteins of BB-1 were amplified by PCR. The mafI genes were cloned into the expression vector pET26b behind a T7 promoter, which is inducible with IPTG and repressed with glucose. Cloning in this vector resulted in a 6xHis-tag at the C terminus of the recombinant protein. The mafB gene fragments were cloned behind the arabinose-inducible araBAD promoter (P BAD ) on a pACYC184-derived vector. This vector was prepared by amplifying P BAD together with the downstream yciM-phoA chimeric gene from plasmid pBADyciM-phoA  and cloning the resulting PCR product into the chloramphenicol-resistance gene on pACYC184 , resulting in plasmid pACYC184-P BAD -yciM-phoA. The mafB fragments were substituted for the yciM-phoA chimeric gene on pACYC184-P BAD -yciM-phoA. The cloning procedure resulted in the presence of an N-terminal FLAG tag in the recombinant proteins. To avoid the lethal effect of the production of the toxic MafB fragments, the plasmids encoding these fragments were introduced in DH5α and BL21 (DE3) under repression conditions (i.e. in the absence of arabinose and the presence of 2 % glucose), and when necessary, the expression of the cognate MafI from a plasmid in trans was induced. The correct expression of recombinant proteins was determined by Western blotting.
Growth inhibition assays
For growth inhibition assays, the rifampicin-resistant derivative of BB-1 was used as the killer and various maf mutant derivatives of BB-1 containing a kanamycin-resistance marker as target cells. Growth inhibition was performed as described [4, 12]. Briefly, bacteria from overnight cultures were grown independently in TSB for 4–5 h. Strains were then mixed 1:1 in TSB and drops were spotted on GC medium plates and incubated for 48 h at 37 °C in candle jars in a humid atmosphere. Subsequently, bacteria were scraped from the plates, resuspended in TSB, and the ratio of the different bacteria at time 0 and after 48 h of co-incubation was determined by plating on selective GC plates containing either rifampicin or kanamycin and counting colonies after overnight incubation. Data of three independent experiments performed in duplicate were considered for statistical comparisons using an unpaired statistical t-test.
Toxicity in E. coli and pull-down assays
E. coli BL21 (DE3) harboring plasmids carrying the 3’ end of mafB and mafI genes were used for toxicity assays. Bacteria from overnight cultures were diluted to an OD600 of 0.1 in 10 ml of LB with appropriate antibiotics for plasmid selection. The concentrations of glucose (0.5 or 2 %), arabinose (2 %) and IPTG (0.1 or 1 mM) used for repression or induction of protein synthesis were optimized for each strain. After overnight incubation, bacteria were plated on LB plates containing proper antibiotics for determination of cell viability. The assays were repeated at least three times. Plates of a representative experiment were photographed with a conventional camera.
For pull-down assays, recombinant MafB and MafI proteins were simultaneously overproduced and the His-tagged MafI was purified by affinity chromatography using Ni2+-NTA beads (QIAGEN GmbH, Hilden) according to the manufacturer’s instructions. Briefly, bacteria were lysed by sonication in 50 mM sodium phosphate, pH 8.0, 300 mM sodium chloride, and 10 mM imidazole, and cell debris was removed by centrifugation (6000 g, 30 min). The supernatant was mixed with Ni2+ beads, and the beads were washed with 50 mM sodium phosphate, pH 8.0, 300 mM sodium chloride, 20 mM imidazole, and eluted with the same buffer containing 250 mM imidazole. The correct purification of the His-tagged MafI and the co-purification of the FLAG-tagged MafB were examined by SDS-PAGE and Western blotting.
Sample preparation, SDS-PAGE and Western blotting
For sample preparation, cultures were adjusted to an OD600 of 1.0, and cells were collected by centrifugation (8000 rpm during 3 min, Eppendorf Centrifuge 5424) and suspended in sample buffer. Purified proteins eluted from Ni2+ beads were mixed with double-strength sample buffer. SDS-PAGE was performed at 200 V during ∼ 45 min in a discontinuous buffer system using 12 % or 15 % polyacrylamide gels. Gels were stained with Coomassie brilliant blue G250 or the proteins were transferred to nitrocellulose membranes. Membranes were subsequently blocked with PBS containing 0.1 % (v/v) Tween 20 (PBST) and 0.5 % (w/v) non-fat dried milk (PBST-M) for 1 h and incubated with anti-His (Thermo Fisher Scientific) or anti-FLAG monoclonal antibodies in PBST-M for 1 h at room temperature. After washes with PBST, the membrane was incubated for 1 h with horseradish peroxidase-conjugated goat anti-mouse immunoglobulins diluted in PBST-M. After extensive washing in PBST, blots were developed with the Pierce ECL Western blotting substrate.
PBS containing 0.1 % (v/v) Tween 20
PBST with 0.5 % (w/v) non-fat dried milk
Polymerase chain reaction
Two-partner secretion protein A
Tryptic soy broth
Rearrangement hot spot
Jacob-Dubuisson F, Guérin J, Baelen S, Clantin B. Two-partner secretion: as simple as it sounds? Res Microbiol. 2013;164:583–95.
Aoki SK, Diner EJ, t’Kint De Roodenbeke C, Burgess BR, Poole SJ, Braaten BA, et al. A widespread family of polymorphic contact-dependent toxin delivery systems in bacteria. Nature. 2010;468:439–42.
Aoki SK, Pamma R, Hernday AD, Bickham JE, Braaten BA, Low DA. Contact-dependent inhibition of growth in Escherichia coli. Science. 2005;309:1245–8.
Arenas J, Schipper K, van Ulsen P, van der Ende A, Tommassen J. Domain exchange at the 3’ end of the gene encoding the fratricide meningococcal two-partner secretion protein A. BMC Genomics. 2013;14:622.
Aoki SK, Malinverni JC, Jacoby K, Thomas B, Pamma R, Trinh BN, et al. Contact-dependent growth inhibition requires the essential outer membrane protein BamA (YaeT) as the receptor and the inner membrane transport protein AcrB. Mol Microbiol. 2008;70:323–40.
Beck CM, Diner EJ, Kim JJ, Low DA, Hayes CS. The F pilus mediates a novel pathway of CDI toxin import. Mol Microbiol. 2014;93:276–90.
Poole SJ, Diner EJ, Aoki SK, Braaten BA, t’Kint De Roodenbeke C, Low DA, et al. Identification of functional toxin/immunity genes linked to contact-dependent growth inhibition (CDI) and rearrangement hotspot (Rhs) systems. PLoS Genet. 2011;7: e1002217.
Kapitein N, Mogk A. Deadly syringes: Type VI secretion system activities in pathogenicity and interbacterial competition. Curr Opin Microbiol. 2013;16:52–8.
Parkhill J, Achtman M, James KD, Bentley SD, Churcher C, Klee SR, et al. Complete DNA sequence of a serogroup A strain of Neisseria meningitidis Z2491. Nature. 2010;404:502–6.
Jamet A, Jousset AB, Euphrasie D, Mukorako P, Boucharlat A, Ducousse A, et al. A new family of secreted toxins in pathogenic Neisseria species. PLoS Pathog. 2015;11: e1004592.
Arenas J, Nijland R, Rodriguez FJ, Bosma TNP, Tommassen J. Involvement of three meningococcal surface-exposed proteins, the heparin-binding protein NhbA, the α-peptide of IgA protease and the autotransporter protease NalP, in initiation of biofilm formation. Mol Microbiol. 2013;87:254–68.
ur Rahman S, Arenas J, Öztürk H, Dekker N, van Ulsen P. The polypeptide transport-associated (POTRA) domains of TpsB transporters determine the system specificity of two-partner secretion systems. J Biol Chem. 2014;289:19799–809.
Benz J, Meinhart A. Antibacterial effector/immunity systems: it's just the tip of the iceberg. Curr Opin Microbiol. 2014;17:1–10.
Braun V, Patzer SI. Intercellular communication by related bacterial protein toxins: colicins, contact-dependent inhibitors, and proteins exported by the type VI secretion system. FEMS Microbiol Lett. 2013;345:13–21.
Ruhe ZC, Low DA, Hayes CS. Bacterial contact-dependent growth inhibition. Trends Microbiol. 2013;21:230–7.
Koskiniemi S, Garza-Sánchez F, Sandegren L, Webb JS, Braaten BA, Poole SJ, et al. Selection of orphan Rhs toxin expression in evolved Salmonella enterica serovar Typhimurium. PLoS Genet. 2014;10: e1004255.
Paruchuri DK, Seifert HS, Ajioka RS, Karlsson KA, So M. Identification and characterization of a Neisseria gonorrhoeae gene encoding a glycolipid-binding adhesin. Proc Natl Acad Sci U S A. 1990;87:333–7.
Goyal P, Krasteva PV, Van Gerven N, Gubellini F, Van den Broeck I, Troupiotis-Tsaïlaki A, et al. Structural and mechanistic insights into the bacterial amyloid secretion channel CsgG. Nature. 2014;516:250–3.
Amitai G, Belenkiy O, Dassa B, Shainskaya A, Pietrokovski S. Distribution and function of new bacterial intein-like protein domains. Mol Microbiol. 2003;47:61–73.
Putker F, Grutsch A, Tommassen J, Bos MP. Ght protein of Neisseria meningitidis is involved in the regulation of lipopolysaccharide biosynthesis. J Bacteriol. 2014;196:780–9.
Chang ACY, Cohen SN. Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the P15A cryptic miniplasmid. J Bacteriol. 1978;134:1141–56.
Laura Catón and Juan Cruz Herrero were supported by personal fellowships for student mobility (Spanish government) and Flavio Ferrera by an Erasmus fellowship. We thank Tom van den Hoeven for technical support and discussions.
The authors declare they have no competing interests.
VdM, LC, MK, JCH, FF, JT, and JA conceived and designed the experiments; LC, JT, and JA performed bioinformatics analysis; VdM, LC, MK, JCH, FF, and JA performed the experiments; JA and JT analyzed the data and wrote the paper. All authors have read and approved the manuscript.