A toxin-antitoxin system consists of a "toxin" and a corresponding "antitoxin", usually encoded by closely linked genes. The toxin is usually a protein, while the antitoxin can be a protein or an RNA. Toxin-antitoxin systems are widely distributed in prokaryotes, and organisms often have them in multiple copies. When these systems are contained on plasmids – transferable genetic elements – they ensure that only the daughter cells that inherit the plasmid survive after cell division. If the plasmid is absent in a daughter cell, the unstable antitoxin is degraded and the stable toxic protein kills the new cell; this is known as 'post-segregational killing' (PSK).
Toxin-antitoxin systems are typically classified according to how the antitoxin neutralises the toxin. In a type I toxin-antitoxin system, the translation of messenger RNA (mRNA) that encodes the toxin is inhibited by the binding of a small non-coding RNA antitoxin that binds the toxin mRNA. The toxic protein in a type II system is inhibited post-translationally by the binding of an antitoxin protein. Type III toxin-antitoxin systems consist of a small RNA that binds directly to the toxin protein and inhibits its activity. There are also types IV, V and VI, which are less common. Toxin-antitoxin genes are often inherited through horizontal gene transfer and are associated with pathogenic bacteria, having been found on plasmids conferring antibiotic resistance and virulence.
Chromosomal toxin-antitoxin systems also exist, some of which are thought to perform cell functions such as responding to stresses, causing cell cycle arrest and bringing about programmed cell death. In evolutionary terms, toxin-antitoxin systems can be considered selfish DNA in that the purpose of the systems are to replicate, regardless of whether they benefit the host organism or not. Some have proposed adaptive theories to explain the evolution of toxin-antitoxin systems; for example, chromosomal toxin-antitoxin systems could have evolved to prevent the inheritance of large deletions of the host genome. Toxin-antitoxin systems have several biotechnological applications, such as maintaining plasmids in cell lines, targets for antibiotics, and as positive selection vectors.
Biological functions
Stabilization and fitness of mobile DNA As stated above, toxin-antitoxin systems are well characterized as plasmid addiction modules. It was also proposed that toxin-antitoxin systems have evolved as plasmid exclusion modules. A cell that carries two plasmids from the same incompatibility group will eventually generate two daughter cells carrying either plasmid. Should one of these plasmids encode for a TA system, its "displacement" by another TA-free plasmid system will prevent its inheritance and thus induce post-segregational killing. This theory was corroborated through computer modelling. Toxin-antitoxin systems can also be found on other mobile genetic elements such as conjugative transposons and temperate bacteriophages and could be implicated in the maintenance and competition of these elements.
Genome stabilization
Toxin-antitoxin systems could prevent harmful large deletions in a bacterial genome, though arguably deletions of large coding regions are fatal to a daughter cell regardless. In Vibrio cholerae, multiple type II toxin-antitoxin systems located in a super-integron were shown to prevent the loss of gene cassettes.
Altruistic cell death mazEF, a toxin-antitoxin locus found in E. coli and other bacteria, was proposed to induce programmed cell death in response to starvation, specifically a lack of amino acids. This would release the cell's contents for absorption by neighbouring cells, potentially preventing the death of close relatives, and thereby increasing the inclusive fitness of the cell that perished. This would be an example of altruism and how bacterial colonies could resemble multicellular organisms. However, the "mazEF-mediated PCD" has largely been refuted by several studies.
Stress tolerance Another theory states that chromosomal toxin-antitoxin systems are designed to be bacteriostatic rather than bactericidal. RelE, for example, is a global inhibitor of translation, is induced during nutrient stress. By shutting down translation under stress, it could reduce the chance of starvation by lowering the cell's nutrient requirements. However, it was shown that several toxin-antitoxin systems, including relBE, do not give any competitive advantage under any stress condition.
Anti-addiction It has been proposed that chromosomal homologues of plasmid toxin-antitoxin systems may serve as anti-addiction modules, which would allow progeny to lose a plasmid without suffering the effects of the toxin it encodes. For example, a chromosomal copy of the ccdA antitoxin encoded in the chromosome of Erwinia chrysanthemi is able to neutralize the ccdB toxin encoded on the F plasmid and thus, prevent toxin activation when such a plasmid is lost. Similarly, the ataR antitoxin encoded on the chromosome of E. coli O157:H7 is able to neutralize the ataTP toxin encoded on plasmids found in other enterohemorrhagic E. coli.
Phage protection Type III toxin-antitoxin (AbiQ) systems have been shown to protect bacteria from bacteriophages altruistically. During an infection, bacteriophages hijack transcription and translation, which could prevent antitoxin replenishment and release toxin, triggering what is called an "abortive infection". Similar protective effects have been observed with type I, type II, and type IV (AbiE) toxin-antitoxin systems.
Abortive initiation (Abi) can also happen without toxin-antitoxin systems, and many Abi proteins of other types exist. This mechanism serves to halt the replication of phages, protecting the overall population from harm.
Antimicrobial persistence When bacteria are challenged with antibiotics, a small and distinct subpopulation of cells is able to withstand the treatment by a phenomenon dubbed as "persistence" (not to be confused with resistance). Due to their bacteriostatic properties, type II toxin-antitoxin systems have previously been thought to be responsible for persistence, by switching a fraction of the bacterial population to a dormant state. However, this hypothesis has been widely invalidated.
Selfish DNA Toxin-antitoxin systems have been used as examples of selfish DNA as part of the gene centered view of evolution. It has been theorised that toxin-antitoxin loci serve only to maintain their own DNA, at the expense of the host organism. Thus, chromosomal toxin-antitoxin systems would serve no purpose and could be treated as "junk DNA". For example, the ccdAB system encoded in the chromosome of E. coli O157:H7 has been shown to be under negative selection, albeit at a slow rate due to its addictive properties.
System types
Type I
Type I toxin-antitoxin systems rely on the base-pairing of complementary antitoxin RNA with the toxin mRNA. Translation of the mRNA is then inhibited either by degradation via RNase III or by occluding the Shine-Dalgarno sequence or ribosome binding site of the toxin mRNA. Often the toxin and antitoxin are encoded on opposite strands of DNA. The 5' or 3' overlapping region between the two genes is the area involved in complementary base-pairing, usually with between 19–23 contiguous base pairs.
Toxins of type I systems are small, hydrophobic proteins that confer toxicity by damaging cell membranes. Few intracellular targets of type I toxins have been identified, possibly due to the difficult nature of analysing proteins that are poisonous to their bacterial hosts. Also, the detection of small proteins has been challenging due to technical issues, a problem that remains to be solved with large-scale analysis.
Type I systems sometimes include a third component. In the case of the well-characterised hok/sok system, in addition to the hok toxin and sok antitoxin, there is a third gene, called mok. This open reading frame almost entirely overlaps that of the toxin, and the translation of the toxin is dependent on the translation of this third component. Thus the binding of antitoxin to toxin is sometimes a simplification, and the antitoxin in fact binds a third RNA, which then affects toxin translation.
Example systems Toxin AntitoxinNotes Ref.hoksok The original and best-understood type I toxin-antitoxin system (pictured), which stabilises plasmids in a number of gram-negative bacteriafstRNAII The first type I system to be identified in gram-positive bacteria tisBistR A chromosomal system induced in the SOS responsedinQagrB A chromosomal system induced in the SOS responseldrDrdlD A chromosomal system in Enterobacteriaceae flmAflmBA hok/sok homologue, which also stabilises the F plasmidibssibDiscovered in E. coli intergenic regions, the antitoxin was originally named QUAD RNA txpA/brnTratAEnsures the inheritance of the skin element during sporulation in Bacillus subtilissymEsymRA chromosomal system induced in the SOS responseXCV2162ptaRNA1A system identified in Xanthomonas campestris with erratic phylogenetic distribution.timPtimRA chromosomal system identified in SalmonellaaapA1isoA1A type 1 TA module in Helicobacter pylori sprA1sprA1asLocated within S. aureus small Pathogenicity island (SaPI). SprA1 encodes for a small cytotoxic peptide, PepA1, which disrupts both S. aureus membranes and host erythrocytes.
Type II
Type II toxin-antitoxin systems are generally better-understood than type I. In this system a labile proteic antitoxin tightly binds and inhibits the activity of a stable toxin. The largest family of type II toxin-antitoxin systems is vapBC, which has been found through bioinformatics searches to represent between 37 and 42% of all predicted type II loci. Type II systems are organised in operons with the antitoxin protein typically being located upstream of the toxin, which helps to prevent expression of the toxin without the antitoxin. The proteins are typically around 100 amino acids in length, and exhibit toxicity in a number of ways: CcdB, for example, affects DNA replication by poisoning DNA gyrase whereas toxins from the MazF family are endoribonucleases that cleave cellular mRNAs, tRNAs or rRNAs at specific sequence motifs. The most common toxic activity is the protein acting as an endonuclease, also known as an interferase.
One of the key features of the TAs is the autoregulation. The antitoxin and toxin protein complex bind to the operator that is present upstream of the TA genes. This results in repression of the TA operon. The key to the regulation are (i) the differential translation of the TA proteins and (ii) differential proteolysis of the TA proteins. As explained by the "Translation-responsive model", the degree of expression is inversely proportional to the concentration of the repressive TA complex. The TA complex concentration is directly proportional to the global translation rate. The higher the rate of translation more TA complex and less transcription of TA mRNA. Lower the rate of translation, lesser the TA complex and higher the expression. Hence, the transcriptional expression of TA operon is inversely proportional to translation rate.
A third protein can sometimes be involved in type II toxin-antitoxin systems. In the case of the ω-ε-ζ (omega-epsilon-zeta) system, the omega protein is a DNA binding protein that negatively regulates the transcription of the whole system. Similarly, the paaR2 protein regulates the expression of the paaR2-paaA2-parE2 toxin-antitoxin system. Other toxin-antitoxin systems can be found with a chaperone as a third component. This chaperone is essential for proper folding of the antitoxin, thus making the antitoxin addicted to its cognate chaperone.
Example systems