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.[2][3] 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).[4][5]
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.[6] There are also types IV-VI, which are less common.[7] Toxin-antitoxin
genes are often inherited through
horizontal gene transfer[8][9] and are associated with
pathogenic bacteria, having been found on plasmids conferring
antibiotic resistance and
virulence.[1]
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.[1][10] 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.[11] Toxin-antitoxin systems have several
biotechnological applications, such as maintaining plasmids in
cell lines, targets for
antibiotics, and as positive selection vectors.[12]
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 would carry two plasmids from the same incompatibility group will eventually generate two daughters 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.[13] This theory was corroborated through
computer modelling.[14] Toxin-antitoxin systems can also be found on other
mobile genetic elements such as conjugative
transposons and
temperatebacteriophages and could be implicated in the maintenance and competition of these elements.[15]
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.[11] In Vibrio cholerae, multiple type II toxin-antitoxin systems located in a
super-integron were shown to prevent the loss of gene cassettes.[18]
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.[19] 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.[14] However, the "mazEF-mediated PCD" has largely been refuted by several studies.[20][21][22]
Stress tolerance
Another theory states that chromosomal toxin-antitoxin systems are designed to be
bacteriostatic rather than
bactericidal.[23] 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.[24] However, it was shown that several toxin-antitoxin systems, including relBE, do not give any competitive advantage under any stress condition.[21]
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.[9] 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.[25] Similarly, the ataR antitoxin encoded on the chromosome of
E. coli O157:H7 is able neutralize the ataTP toxin encoded on plasmids found in other
enterohemorragic E. coli.[26]
Phage protection
Type III toxin-antitoxin (AbiQ) systems have been shown to protect bacteria from
bacteriophages altruistically.[27][28] During an infection, bacteriophages hijack transcription and translation, which could prevent antitoxin replenishment and release toxin, triggering what is called an "abortive infection".[27][28] Similar protective effects have been observed with type I,[29] type II,[30] and type IV (AbiE)[31] 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.[32]
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).[33] 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.[34] However, this hypothesis has been widely invalidated.[35][36][37]
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.[1][38] 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.[8]
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.[39]
Toxins of type I systems are small,
hydrophobic proteins that confer toxicity by damaging
cell membranes.[1] 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.[10] Also, the detection of small proteins has been challenging due to technical issues, a problem that remains to be solved with large-scale analysis.[40]
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.[5] 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.[39]
Located 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 toxin-antitoxin systems are generally better-understood than type I.[39] In this system a
labile proteic antitoxin tightly binds and inhibits the activity of a stable toxin.[10] The largest family of type II toxin-antitoxin systems is vapBC,[53] which has been found through
bioinformatics searches to represent between 37 and 42% of all predicted type II loci.[16][17] 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.[54] The proteins are typically around 100
amino acids in length,[39] and exhibit toxicity in a number of ways:
CcdB, for example, affects
DNA replication by poisoning
DNA gyrase[55] whereas toxins from the MazF family are endoribonucleases that cleave cellular mRNAs,[56][57] tRNAs [58][59] or rRNAs [60] at specific
sequence motifs. The most common toxic activity is the protein acting as an
endonuclease, also known as an
interferase.[61][62]
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-reponsive model",[63] 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.[64] Similarly, the paaR2 protein regulates the expression of the paaR2-paaA2-parE2 toxin-antitoxin system.[65] Other toxin-antitoxin systems can be found with a
chaperone as a third component.[66] This chaperone is essential for proper
folding of the antitoxin, thus making the antitoxin addicted to its cognate chaperone.
Type III toxin-antitoxin systems rely on direct interaction between a toxic protein and an RNA antitoxin. The toxic effects of the protein are neutralised by the RNA gene.[6] One example is the ToxIN system from the bacterial plant pathogen Erwinia carotovora. The toxic ToxN protein is approximately 170 amino acids long and has been shown to be toxic to E. coli. The toxic activity of ToxN is inhibited by ToxI RNA, an RNA with 5.5 direct
repeats of a 36 nucleotide motif (AGGTGATTTGCTACCTTTAAGTGCAGCTAGAAATTC).[27][71]Crystallographic analysis of ToxIN has found that ToxN inhibition requires the formation of a trimeric ToxIN complex, whereby three ToxI monomers bind three ToxN monomers; the complex is held together by extensive protein-RNA interactions.[72]
Type IV
Type IV toxin-antitoxin systems are similar to type II systems, because they consist of two proteins. Unlike type II systems, the antitoxin in type IV toxin-antitoxin systems counteracts the activity of the toxin, and the two proteins do not necessarily interact directly. DarTG1 and DarTG2 are type IV toxin-antitoxin systems that modify DNA. Their toxins add ADP-ribose to guanosine bases (DarT1 toxin) or thymidine bases (DarT2 toxin), and their antitoxins remove the toxic modifications (NADAR antitoxin from guanosine and DarG antitoxin from thymidine).[73][74][75][76]
Type V
ghoST is a type V toxin-antitoxin system, in which the antitoxin (GhoS) cleaves the ghoT mRNA. This system is regulated by a type II system, mqsRA.[77]
Type VI
socAB is a type VI toxin-antitoxin system that was discovered in Caulobacter crescentus. The antitoxin, SocA, promotes degradation of the toxin, SocB, by the
protease ClpXP.[78]
Type VII
Type VII has been proposed to include systems hha/tomB, tglT/takA and hepT/mntA, all of which neutralise toxin activity by post-translational chemical modification of amino acid residues.[79]
Type VIII
Type VIII includes the system creTA. In this system, the antitoxin creA serves as a guide RNA for a CRISPR-Cas system. Due to incomplete complementarity between the creA guide and the creAT promoter, the Cas complex does not cleave the DNA, but instead remains at the site, where it blocks access by RNA polymerase, preventing expression of the creT toxin (a natural instance of
CRISPRi). When expressed, the creT RNA will sequester the rare arginine codon tRNAUCU, stalling translation and halting cell metabolism.[80]
Biotechnological applications
The
biotechnological applications of toxin-antitoxin systems have begun to be realised by several biotechnology organisations.[12][23] A primary usage is in maintaining plasmids in a large bacterial
cell culture. In an experiment examining the effectiveness of the hok/sok locus, it was found that segregational stability of an
inserted plasmid expressing
beta-galactosidase was increased by between 8 and 22 times compared to a
control culture lacking a toxin-antitoxin system.[81][82] In large-scale
microorganism processes such as
fermentation, progeny cells lacking the plasmid insert often have a higher
fitness than those who inherit the plasmid and can outcompete the desirable microorganisms. A toxin-antitoxin system maintains the plasmid thereby maintaining the efficiency of the industrial process.[12]
Additionally, toxin-antitoxin systems may be a future target for
antibiotics. Inducing suicide modules against pathogens could help combat the growing problem of
multi-drug resistance.[83]
Ensuring a plasmid accepts an insert is a common problem of DNA
cloning. Toxin-antitoxin systems can be used to positively select for only those cells that have taken up a plasmid containing the inserted gene of interest, screening out those that lack the inserted gene. An example of this application comes from the ccdB-encoded toxin, which has been incorporated into
plasmid vectors.[84] The gene of interest is then targeted to recombine into the ccdB locus, inactivating the transcription of the toxic protein. Thus, cells containing the plasmid but not the insert perish due to the toxic effects of CcdB protein, and only those that incorporate the insert survive.[12]
Another example application involves both the CcdB toxin and CcdA antitoxin. CcdB is found in recombinant bacterial genomes and an inactivated version of CcdA is inserted into a
linearised plasmid vector. A short extra sequence is added to the gene of interest that activates the antitoxin when the insertion occurs. This method ensures
orientation-specific gene insertion.[84]
^Page R, Peti W (April 2016). "Toxin-antitoxin systems in bacterial growth arrest and persistence". Nature Chemical Biology. 12 (4): 208–14.
doi:
10.1038/nchembio.2044.
PMID26991085.
^Szekeres S, Dauti M, Wilde C, Mazel D, Rowe-Magnus DA (March 2007). "Chromosomal toxin-antitoxin loci can diminish large-scale genome reductions in the absence of selection". Molecular Microbiology. 63 (6): 1588–605.
doi:
10.1111/j.1365-2958.2007.05613.x.
PMID17367382.
S2CID28191383.
^
abHazan R, Engelberg-Kulka H (September 2004). "Escherichia coli mazEF-mediated cell death as a defense mechanism that inhibits the spread of phage P1". Molecular Genetics and Genomics. 272 (2): 227–34.
doi:
10.1007/s00438-004-1048-y.
PMID15316771.
S2CID28840747.
^Loh SM, Cram DS, Skurray RA (June 1988). "Nucleotide sequence and transcriptional analysis of a third function (Flm) involved in F-plasmid maintenance". Gene. 66 (2): 259–68.
doi:
10.1016/0378-1119(88)90362-9.
PMID3049248.
^Robson J, McKenzie JL, Cursons R, Cook GM, Arcus VL (July 2009). "The vapBC operon from Mycobacterium smegmatis is an autoregulated toxin-antitoxin module that controls growth via inhibition of translation". Journal of Molecular Biology. 390 (3): 353–67.
doi:
10.1016/j.jmb.2009.05.006.
PMID19445953.
^
abBernard P, Couturier M (August 1992). "Cell killing by the F plasmid CcdB protein involves poisoning of DNA-topoisomerase II complexes". Journal of Molecular Biology. 226 (3): 735–45.
doi:
10.1016/0022-2836(92)90629-X.
PMID1324324.
^Christensen-Dalsgaard M, Overgaard M, Winther KS, Gerdes K (2008). "Chapter 25 RNA Decay by Messenger RNA Interferases". RNA Turnover in Bacteria, Archaea and Organelles. Methods in Enzymology. Vol. 447. pp. 521–35.
doi:
10.1016/S0076-6879(08)02225-8.
ISBN978-0-12-374377-0.
PMID19161859.
^Yamaguchi Y, Inouye M (2009). "Chapter 12 mRNA Interferases, Sequence-Specific Endoribonucleases from the Toxin–Antitoxin Systems". mRNA interferases, sequence-specific endoribonucleases from the toxin-antitoxin systems. Progress in Molecular Biology and Translational Science. Vol. 85. pp. 467–500.
doi:
10.1016/S0079-6603(08)00812-X.
ISBN978-0-12-374761-7.
PMID19215780.
^Wang X, Yao J, Sun YC, Wood TK (May 2021). "Type VII Toxin/Antitoxin Classification System for Antitoxins that Enzymatically Neutralize Toxins". Trends in Microbiology. 29 (5): 388–393.
doi:
10.1016/j.tim.2020.12.001.
PMID33342606.
S2CID229341165.