Information Excavation Of Novel Toxins And Their Immunity Proteins In Large Genomes Of Myxobacteria

The manufacture and use of toxins are the most common survival strategies for life.Bacteria have been found to secrete protein toxins into the environment or directly into other cells that are in contact in a variety of ways.Most of these protein toxins have enzymatic activities,which can destroy important cellular components of target cells,such as DNA,RNA,peptidoglycans in cell wall,cell membranes,and so on.Bacteria also produce immunity proteins that are antagonistic to toxin proteins to protect cells from toxins from themselves or other cells.Genes encoding toxin proteins and their corresponding immunity proteins are usually adjacent in the bacterial genomes,thus comprising a toxin-immunity protein system.Many studies have shown that the bacterial toxin-immunity protein system plays a key role in some cellular events.For example,the pathogenic mechanism of many bacterial pathogens is achieved through the secretion of toxin proteins that kill or affect the eukaryotic target cells.However,the toxin-immunity protein system can be used to obtain competitive advantage at various levels of conflict between bacteria.There are still many toxin-immunity protein systems in bacteria those have not yet been identified,and a large number of sequenced genomes provide the opportunity to identify new types of toxins,associated immunity proteins,and export mechanisms through computational analysis.Myxobacterium is a type of gram-negative sliding bacteria with complex multicellular behaviors.It is a type of dominant bacteria in soil both in quantity and species.Most of the aerobic myxobacterial genomes are more than 9 Mb,in particular,Sorangiineae have been found to possess the largest genome in prokaryotes.The survival advantage of myxobacteria probably benefits from the diverse toxin-immunity protein systems encoded in their large genomes.However,only two toxin-immunity protein systems were reported in myxobacteria recently.In this paper,we attempted to obtain more toxin-immunity protein systems in the large genomes of myxobacteria through extensive comparative genomics and structural bioinformatics analysis based on a series of phenotypic mutant strains found in the previous work inour laboratory.In our laboratory,random mutations of Myxococcus xanthus DK1622,a myxobacterial model strain,were carried out by the random insertion of transposons into the genome.The eleven mutant strains each formed a colony boundary with the wild-type strain,and were screened from a large number of mutants.The insertion sites of these mutants were scattered in the genome,but the related genes were all unknown in function.Starting from these genes,we identified seven gene clusters related to the colony expansion boundary phenotype in the M.xanthus DK1622 genome through a series of bioinformatics analysis.These involved 50 genes of 9 homologous families?family A-family I?.And we found that the homologous genes of these families were widely distributed in various myxobacteria and constitute at least 68 similar gene clusters.Further protein structural modeling and conserved motif analysis revealed that family E is the AHH nuclease toxin and family D is its corresponding immunity protein.In the genome,the family E gene is always adjacent with family D gene in downstream,and both constitute a novel type of toxin-immunity protein system.However,the family B protein is highly structurally similar to the PAAR protein,a structural component of the bacterial type VI secretion system that is able to export toxin proteins.Therefore,we speculated that the colony expansion boundary phenotype involves a novel nuclease toxin-immunity protein system,whose output is mediated by the PAAR protein and type VI secretion system,and several other families may be the related auxiliary proteins.In addition,we also found that two PAAR proteins possessed a novel type of toxic domain at the C-terminal,and the SMI1_KNR4 protein in the family H is its immunity protein.Several collaborators carried out a large number of relevant experimental works on the basis of these bioinformatics analyses.The results showed that the destroying of immunity protein-encoding genes in those mutant strains resulted in their inability to defend against toxins from the wild-type strain,thus forming the boundary line constituting by a large numbers of dead cells on the junctions of both colonies.Our large-scale scanning to bacterial genomes revealed that the nuclease toxin-immunity protein system consisting of family E and family D is widely distributed in various types of bacteria.Moreover,this type of conserved gene pair is multiple copies in many bacterial genomes,especially the most copies in Myxobacteria.Further,we identified all the known 2,610 non-redundant homologous proteins of family D through large-scale sequence iteration search.These proteins neither belong to any known family,nor contain almost any known domain.Thus,we categorized them into a new MIP protein family.To ascertain the structural characteristics of the MIP family proteins,we determined the three-dimensional structure of MXAN₀049 in M.xanthus DK1622 and found that its fold is different from all known structures.It has a large acidic loop on the protein surface.We divided the MIP family into seven subfamilies with various characteristics based on the phylogenetic analysis.The MIP genes in five of seven subfamilies are in downstream of AHH-encoding gene;the rest two subfamilies' upstream is DUF2380 and pentapeptide-encoding genes,respectively.According to the conserved motif assay,we suggest that AHH and DUF2380 proteins probably had nuclease activity,while pentapeptide could be a structural analog of nucleic acid.Thus,all three kinds of proteins may be nucleic acid-targeting toxins in bacteria.In addition to the AHH nuclease toxin that has been experimentally confirmed by us,the collaborators also have demonstrated that the DUF2380 protein is a nuclease toxin and it is able to be immunized by the corresponding MIP protein.Further,we found out that these toxins were able to be exported by multiple secretion ways through the analysis of genomic contexts and domain architectures,including T6SS?PAAR?,lipoprotein signal peptide?SP??,VgrG,RHS,T7SS,and so on.Taken together,the MIP family first reported has abilities to defend against multiple secreted nucleic acid-targeting toxins in bacteria.Several hundred pairs of unreported toxins-immunity protein genes were accordingly identified from bacterial genomes,many of them are from the pathogenic bacteria.Further,we conducted a comprehensive systematic bioinformatics analysis of the PAAR protein?family B?.Both the PAAR protein and the VgrG protein form a needle-like complex at the top of the bacterial type VI secretion system and carry the toxin output.In addition to the known PAAR domains,we found that the three-dimensional structures of DUF4150 and DUF4280 that are two unknown functional domains are highly similar to PAAR,inferring that they should also be the members of the PAAR superfamily.Through large-scale sequence searches,we found as many as 33,901 proteins containing PAAR superfamily domains.These PAAR proteins come from almost all life,but more than 99%of them come from bacteria,especially Proteobacteria.The PAAR gene was found in approximately 25%of the completely sequenced bacterial genomes,and more than 60%of these strains have multi-copy PAAR genes,especially the number of copies is highest in the myxobacteria.The architecture analysis of domain revealed that more than 55%of the PAAR proteins had multiple domains,of which the PAAR domain was almost always located on the N-terminal,and at least 20%of the multi-domain PAAR proteins possessed unknown domains.We focused on the C-terminal domain of the PAAR protein.In addition to the most common Rhs elements-related domains,there are at least 58 kinds of domains with high frequencies,of which 37 are known toxins,the remaining 21 are functionally unknown,and we speculated that most might also be toxins.At least 33 of these domains have nucleic acid-related activity,PAAR proteins prefer to carry domains with nuclease activity or other nucleic acid-related functions.Furthermore,we also analyzed the upstream and downstream regions of the PAAR gene in the genome,and its peripheral genes were mostly related to the type VI secretion system,including family C?PKS_KS?and family G?DUF2169?.Some conserved gene pairs appearing with high frequency near the PAAR gene were also found,including Ntox15 and GAD-like,VRR_NUC and DUF3396,Tox-REase-5 and Imm52,PAAR+DUF2235+unknown domains and DUF2931,DUF3274 and DUF2875,and so on.They are probably potential new toxin-immunity protein systems.To seek for more toxin-immunity protein systems,we investigated all the conserved gene pairs in the large genomes of the myxobacteria.We predicted 174,992 domains from 4,773 superfamilies in all the encoding genes of 23 completely sequenced myxobacterial genomes.These domains combine more than 10,000 different architectures in proteins.Among them,we have identified 135 and 81 copies of the two nuclease toxins AHH and DUF2380,respectively,which are usually adjacent downstream to the genes encoding the immunity proteins of the MIP family.Forty-four genes encoding the nuclease toxin Tox-REase-5 were also found,all of which are adjacent downstream to the genes encoding immunity protein Imm52.The three types of nuclease toxin-immunity protein systems are mainly distributed in the Cystobacterineae and the maximum number of their gene copies in the single strain is more than 10.We also found 181 PAAR genes and 250 VgrG genes in the myxobacterial genome.The number of these two gene copies was positively correlated in the bacterial strain,and the strains with high gene copy number were all present in the Sorangiineae.With the exception of VRR_NUC,no other known toxic domains appear at high frequencies in the myxobacterial VgrG protein.However,5%of PAAR proteins in myxobacteria have a protease Peptidase_S41 domain at the C-terminal,and these PAAR genes may form a gene cluster with three conserved genes in the upstream.Three percent of the PAAR protein may have a novel and unknown toxic domain at the C-terminal because their encoding genes are adjacent downstream to the gene encoding the immunity protein SMI1_KNR4.In addition,pentapeptide proteins have also been found to occur at high frequencies in the Sorangiineae,and more than 20%of them have a DUF2169 domain at the N-terminal.Seven gene pairs encoding HNHc nuclease and SMI1_KNR4 immunity protein,and four gene pairs encoding nuclease toxin Ntox16 and ANK with unknown function were also identified.Further,we also found that the conserved gene pairs in the myxobacterial genome encoded a large number of signal transduction systems,including more than 300 pairs of sigma and anti-sigma factors and more than 200 pairs of two-component systems.In conclusion,a large number of novel toxin-immunity protein systems were identified in the large genome of myxobacteria by bioinformatics analysis,and some of them have been confirmed by the collaborators.These toxin-immunity protein systems and other functionally important conserved gene pairs in the myxobacterial genome may help them to gain widespread environmental adaptability and the advantages of competition and survival against other bacteria.These findings help increase the understanding to the bacterial toxin systems,and those new nuclease toxins and immunity proteins may also have potential biotechnological applications.