Insights into the mRNA cleavage mechanism by MazF, an mRNA interferase

MazF is an Escherichia coli toxin that is highly conserved among the prokaryotes and plays an important role in growth regulation. When MazF is induced, protein synthesis is effectively inhibited. However, the mechanism of MazF action has been controversial. Here we unequivocally demonstrate that MazF is an endoribonuclease that specifically cleaves mRNAs at ACA sequences. We then demonstrate its enzymatic specificity using short RNA substrates. MazF cleaves RNA at the 5'-end of ACA sequences, yielding a 2',3'-cyclic phosphate at one side and a free 5'-OH group at the other. Using DNA-RNA chimeric substrates containing XACA, the 2'-OH group of residue X was found absolutely essential for MazF cleavage, whereas all the other residues may be deoxyriboses. Therefore, MazF exhibits exquisite site specificity and has utility as an RNA-restriction enzyme for RNA structural studies or as an mRNA interferase to regulate cell growth in prokaryotic and eukaryotic cells.     

Insights into the specificity of RNA cleavage by the Escherichia coli MazF toxin

The mazEF (chpA) toxin-antitoxin system of Escherichia coli is involved in the cell response to nutritional and antibiotic stresses as well as in bacterial-programmed cell death. Valuable information on the MazF toxin was derived from the determination of the crystal structure of the MazE/MazF complex and from in vivo data, suggesting that MazF promoted ribosome-dependent cleavage of messenger RNA. However, it was concluded from recent in vitro analyses using a MazF-(His6) fusion protein that MazF was an endoribonuclease that cleaved messenger RNA specifically at 5'-ACA-3' sites situated in single-stranded regions. In contrast, our work reported here shows that native MazF protein cleaves RNA at the 5' side of residue A in 5'-NAC-3' sequences (where N is preferentially U or A). MazF-dependent cleavage occurred at target sequences situated either in single- or double-stranded RNA regions. These activities were neutralized by a His6-MazE antitoxin. Although essentially consistent with previous in vivo reports on the substrate specificity of MazF, our results strongly suggest that the endoribonuclease activity of MazF may be modulated by additional factors to cleave messenger and other cellular RNAs.     

Characterization of ChpBK, an mRNA interferase from Escherichia coli

Escherichia coli contains a number of antitoxin-toxin modules on its chromosome, which are responsible for cell growth arrest and possible cell death. ChpBK is a toxin encoded by the ChpBIK antitoxin-toxin module. This module consists of a pair of genes, chpBI and chpBK encoding antitoxin ChpBI and toxin ChpBK, respectively. ChpBK consists of 116 amino acid residues, and its sequence shows 35% identity and 52% similarity to MazF, another E. coli toxin. MazF has been shown to be a sequence-specific (ACA) endoribonuclease that cleaves cellular mRNAs and effectively blocks protein synthesis and is thus termed as an mRNA interferase. Here we demonstrate that ChpBK is another mRNA interferase in E. coli whose induction effectively blocks cell growth in a manner similar to that of MazF. The protein synthesis as judged by incorporation of [35S]methionine was, however, reduced by only 60% upon ChpBK induction. We demonstrate that ChpBK is a new sequence-specific endoribonuclease that cleaves mRNAs both in vivo and in vitro at the 5'-or3'-side of the A residue in ACY sequences (Y is U, A, or G). The ChpBK cleavage of a synthetic RNA substrate generated a 2',3'-cyclic phosphate group at the 3'-end of the 5'-end product and a 5'-OH group at the 5'-end of the 3'-end product in a manner identical to that of MazF.     

Crystal structure of the MazE/MazF complex: molecular bases of antidote-toxin recognition

A structure of the Escherichia coli chromosomal MazE/MazF addiction module has been determined at 1.7 A resolution. Addiction modules consist of stable toxin and unstable antidote proteins that govern bacterial cell death. MazE (antidote) and MazF (toxin) form a linear heterohexamer composed of alternating toxin and antidote homodimers (MazF(2)-MazE(2)-MazF(2)). The MazE homodimer contains a beta barrel from which two extended C termini project, making interactions with flanking MazF homodimers that resemble the plasmid-encoded toxins CcdB and Kid. The MazE/MazF heterohexamer structure documents that the mechanism of antidote-toxin recognition is common to both chromosomal and plasmid-borne addiction modules, and provides general molecular insights into toxin function, antidote degradation in the absence of toxin, and promoter DNA binding by antidote/toxin complexes.     

ACA‐specific RNA sequence recognition is acquired via the loop 2 region of MazF mRNA interferase

MazF is an mRNA interferase that cleaves mRNAs at a specific RNA sequence. MazF from E. coli (MazF‐ec) cleaves RNA at A and CA. To date, a large number of MazF homologs that cleave RNA at specific three‐ to seven‐base sequences have been identified from bacteria to archaea. MazF‐ec forms a dimer, in which the interface between the two subunits is known to be the RNA substrate‐binding site. Here, we investigated the role of the two loops in MazF‐ec, which are closely associated with the interface of the MazF‐ec dimer. We examined whether exchanging the loop regions of MazF‐ec with those from other MazF homologs, such as MazF from Myxococcus xanthus (MazF‐mx) and MazF from Mycobacterium tuberculosis (MazF‐mt3), affects RNA cleavage specificity. We found that exchanging loop 2 of MazF‐ec with loop 2 regions from either MazF‐mx or MazF‐mt3 created a new cleavage sequence at (A/U)(A/U)AA and C in addition to the original cleavage site, A and CA, whereas exchanging loop 1 did not alter cleavage specificity. Intriguingly, exchange of loop 2 with 8 or 12 consecutive Gly residues also resulted in a new RNA cleavage site at (A/U)(A/U)AA and C. The present study suggests a method for expanding the RNA cleavage repertoire of mRNA interferases, which is crucial for potential use in the regulation of specific gene expression and for biotechnological applications. Proteins 2013. © 2012 Wiley Periodicals, Inc.     

Characterization of the interactions within the mazEF addiction module of Escherichia coli

In bacteria, programmed cell death is mediated through the unique genetic system called "addiction module," which consists of a pair of genes encoding a stable toxin and an unstable antitoxin. The mazEF system is known as an addiction module located on the Escherichia coli chromosome. MazF is a stable toxin, and MazE is a labile antitoxin interacting with MazF to form a complex. MazE and the MazE-MazF complex can bind to the mazEF promoter region to regulate the mazEF expression. Here we show that the binding of purified (His)6MazE to the mazEF promoter DNA was enhanced by MazF. The site-directed mutations at the conserved amino acid residues in MazE N-terminal region (K7A, R8A, S12A, and R16A) disrupted the DNA binding ability of both (His)6MazE and the MazE-MazF-(His)6 complex, suggesting that MazE binds to the mazEF promoter DNA through the N-terminal domain. The ratio of MazE to MazF(His)6 in the MazE-MazF(His)6 complex is about 1:2. Because both MazE and MazF-(His)6 exist as dimers by themselves, the MazE-MazF-(His)6 complex (76.9 kDa) is predicted to consist of one MazE dimer and two MazF(His)6 dimers. The interaction between MazE and MazF was also characterized with the yeast two-hybrid system. It was found that the region from residues 38 to 75 of MazE was required for its binding to MazF. Site-directed mutagenesis at this region revealed that Leu55 and Leu58 play an important role in the MazE-MazF complex formation but not in MazE binding to the mazEF promoter DNA. The present results demonstrate that MazE is composed of two domains, the N-terminal DNA-binding domain and the C-terminal domain interacting with MazF.     

An ADP‐ribosyltransferase Alt of bacteriophage T4 negatively regulates the Escherichia coli MazF toxin of a toxin–antitoxin module

Prokaryotic toxin–antitoxin (TA) systems are linked to many roles in cell physiology, such as plasmid maintenance, stress response, persistence and protection from phage infection, and the activities of toxins are tightly regulated. Here, we describe a novel regulatory mechanism for a toxin of Escherichia coli TA systems. The MazF toxin of MazE‐MazF, which is one of the best characterized type II TA systems, was modified immediately after infection with bacteriophage T4. Mass spectrometry demonstrated that the molecular weight of this modification was 542 Da, corresponding to a mono‐ADP‐ribosylation. This modification disappeared in cells infected with T4 phage lacking Alt, which is one of three ADP‐ribosyltransferases encoded by T4 phage and is injected together with phage DNA upon infection. In vivo and in vitro analyses confirmed that T4 Alt ADP‐ribosylated MazF at an arginine residue at position 4. Finally, the ADP‐ribosylation of MazF by Alt resulted in the reduction of MazF RNA cleavage activity in vitro, suggesting that it may function to inactivate MazF during T4 infection. This is the first example of the chemical modification of an E. coli toxin in TA systems to regulate activity.     

Stimulation of HIV-1 minus strand strong stop DNA transfer by genomic sequences 3' of the primer binding site

The mechanism of human immunodeficiency virus 1 (HIV-1) minus strand transfer was examined using a genomic RNA sequence-based donor-acceptor template system. The donor RNA, D199, was a 199-nucleotide sequence from the 5'-end of the genome to the primer binding site (PBS) and shared 97 nucleotides of homology with the acceptor RNA. To investigate the influence of RNA structure on transfer, a second donor RNA, D520, was generated by extending the 3'-end of D199 to include an additional 321 nucleotides of the genome. The position of priming, length of homology with the acceptor, and length of cDNA synthesized were identical with the two donors. Interestingly, at 200% NC coating, donor D520 yielded a transfer efficiency of about 75% compared with about 35% with D199. A large proportion of the D520 promoted transfers occurred after the donor RNA was copied to the end. Analysis of donor RNA cleavage, the acceptor invasion site and R homology requirements indicated that transfers with D520 involved a similar but more efficient acceptor invasion mechanism compared with D199. RNA structure probing by RNase T1 and the RT pause profile during synthesis indicated conformational differences between D199 and D520 in the starting structure, and in dynamic structures formed during synthesis within the R region. Overall observations suggest that regions 3' of the primer binding site influence the conformation of the R region of D520 to facilitate steps that promote strand transfer.     

A novel endonucleolytic mechanism to generate the CCA 3' termini of tRNA molecules in Thermotoga maritima

The tRNA 3'-terminal CCA sequence is essential for aminoacylation of the tRNAs and for translation on the ribosome. The tRNAs are transcribed as larger precursor molecules containing 5' and 3' extra sequences. In the tRNAs that do not have the encoded CCA, the 3' extra sequence after the discriminator nucleotide is usually cleaved off by the tRNA 3' processing endoribonuclease (3' tRNase, or RNase Z), and the 3'-terminal CCA residues are added thereto. Here we analyzed Thermotoga maritima 3' tRNase for enzymatic properties using various pre-tRNAs from T. maritima, in which all 46 tRNA genes encode CCA with only one exception. We found that the enzyme has the unprecedented activity that cleaves CCA-containing pre-tRNAs precisely after the CCA sequence, not after the discriminator. The assays for pre-tRNA variants suggest that the CA residues at nucleotides 75 and 76 are required for the enzyme to cleave pre-tRNAs after A at nucleotide 76 and that the cleavage occurs after nucleotide 75 if the sequence is not CA. Intriguingly, the pre-tRNA(Met) that is the only T. maritima pre-tRNA without the encoded CCA was cleaved after the discriminator. The kinetics data imply the existence of a CCA binding domain in T. maritima 3' tRNase. We also identified two amino acid residues critical for the cleavage site selection and several residues essential for the catalysis. Analysis of cleavage sites by 3' tRNases from another eubacteria Escherichia coli and two archaea Thermoplasma acidophilum and Pyrobaculum aerophilum corroborates the importance of the two amino acid residues for the cleavage site selection.     

Cleavage of oligoribonucleotides by a ribozyme derived from the hepatitis delta virus RNA sequence.

A self-cleaving RNA sequence from hepatitis delta virus was modified to produce a ribozyme capable of catalyzing the cleavage of RNA in an intermolecular (trans) reaction. The delta-derived ribozyme cleaved substrate RNA at a specific site, and the sequence specificity could be altered with mutations in the region of the ribozyme proposed to base pair with the substrate. A substrate target size of approximately 8 nucleotides in length was identified. Octanucleotides containing a single ribonucleotide immediately 5' to the cleavage site were substrates for cleavage, and cleavage activity was significantly reduced only with a guanine base at that position. A deoxyribose 5' to the cleavage site blocked the reaction. These data are consistent with a proposed secondary structure for the self-cleaving form of the hepatitis delta virus ribozyme in which a duplex forms with sequences 3' to the cleavage site, and they support a proposed mechanism in which cleavage involves attack on the phosphorus at the cleavage site by the adjacent 2'-hydroxyl group.     

RNase/anti-RNase activities of the bacterial parD toxin-antitoxin system

The bacterial parD toxin-antitoxin system of plasmid R1 encodes two proteins, the Kid toxin and its cognate antitoxin, Kis. Kid cleaves RNA and inhibits protein synthesis and cell growth in Escherichia coli. Here, we show that Kid promotes RNA degradation and inhibition of protein synthesis in rabbit reticulocyte lysates. These new activities of the Kid toxin were counteracted by the Kis antitoxin and were not displayed by the KidR85W variant, which is nontoxic in E. coli. Moreover, while Kid cleaved single- and double-stranded RNA with a preference for UAA or UAC triplets, KidR85W maintained this sequence preference but hardly cleaved double-stranded RNA. Kid was formerly shown to inhibit DNA replication of the ColE1 plasmid. Here we provide in vitro evidence that Kid cleaves the ColE1 RNA II primer, which is required for the initiation of ColE1 replication. In contrast, KidR85W did not affect the stability of RNA II, nor did it inhibit the in vitro replication of ColE1. Thus, the endoribonuclease and the cytotoxic and DNA replication-inhibitory activities of Kid seem tightly correlated. We propose that the spectrum of action of this toxin extends beyond the sole inhibition of protein synthesis to control a broad range of RNA-regulated cellular processes.     

Ribonucleases in bacterial toxin–antitoxin systems

Toxin–antitoxin (TA) systems are widespread in bacteria and archaea and play important roles in a diverse range of cellular activities. TA systems have been broadly classified into 5 types and the targets of the toxins are diverse, but the most frequently used cellular target is mRNA. Toxins that target mRNA to inhibit translation can be classified as ribosome-dependent or ribosome-independent RNA interferases. These RNA interferases are sequence-specific endoribonucleases that cleave RNA at specific sequences. Despite limited sequence similarity, ribosome-independent RNA interferases belong to a limited number of structural classes. The MazF structural family includes MazF, Kid, ParE and CcdB toxins. MazF members cleave mRNA at 3-, 5- or 7-base recognition sequences in different bacteria and have been implicated in controlling cell death (programmed) and cell growth, and cellular responses to nutrient starvation, antibiotics, heat and oxidative stress. VapC endoribonucleases belong to the PIN-domain family and inhibit translation by either cleaving tRNA^fMet in the anticodon stem loop, cleaving mRNA at -AUA(U/A)-hairpin-G- sequences or by sequence-specific RNA binding. VapC has been implicated in controlling bacterial growth in the intracellular environment and in microbial adaptation to nutrient limitation (nitrogen, carbon) and heat shock. ToxN shows structural homology to MazF and is also a sequence-specific endoribonuclease. ToxN confers phage resistance by causing cell death upon phage infection by cleaving cellular and phage RNAs, thereby interfering with bacterial and phage growth. Notwithstanding our recent progress in understanding ribonuclease action and function in TA systems, the environmental triggers that cause release of the toxin from its cognate antitoxin and the precise cellular function of these systems in many bacteria remain to be discovered. This article is part of a Special Issue entitled: RNA Decay mechanisms.     

Kid cleaves specific mRNAs at UUACU sites to rescue the copy number of plasmid R1

Stability and copy number of extra-chromosomal elements are tightly regulated in prokaryotes and eukaryotes. Toxin Kid and antitoxin Kis are the components of the parD stability system of prokaryotic plasmid R1 and they can also function in eukaryotes. In bacteria, Kid was thought to become active only in cells that lose plasmid R1 and to cleave exclusively host mRNAs at UA(A/C/U) trinucleotide sites to eliminate plasmid-free cells. Instead, we demonstrate here that Kid becomes active in plasmid-containing cells when plasmid copy number decreases, cleaving not only host- but also a specific plasmid-encoded mRNA at the longer and more specific target sequence UUACU. This specific cleavage by Kid inhibits bacterial growth and, at the same time, helps to restore the plasmid copy number. Kid targets a plasmid RNA that encodes a repressor of the synthesis of an R1 replication protein, resulting in increased plasmid DNA replication. This mechanism resembles that employed by some human herpesviruses to regulate viral amplification during infection.     

Nucleases of the metallo-beta-lactamase family and their role in DNA and RNA metabolism

Proteins of the metallo-beta-lactamase family with either demonstrated or predicted nuclease activity have been identified in a number of organisms ranging from bacteria to humans and has been shown to be important constituents of cellular metabolism. Nucleases of this family are believed to utilize a zinc-dependent mechanism in catalysis and function as 5' to 3' exonucleases and or endonucleases in such processes as 3' end processing of RNA precursors, DNA repair, V(D)J recombination, and telomere maintenance. Examples of metallo-beta-lactamase nucleases include CPSF-73, a known component of the cleavage/polyadenylation machinery, which functions as the endonuclease in 3' end formation of both polyadenylated and histone mRNAs, and Artemis that opens DNA hairpins during V(D)J recombination. Mutations in two metallo-beta-lactamase nucleases have been implicated in human diseases: tRNase Z required for 3' processing of tRNA precursors has been linked to the familial form of prostate cancer, whereas inactivation of Artemis causes severe combined immunodeficiency (SCID). There is also a group of as yet uncharacterized proteins of this family in bacteria and archaea that based on sequence similarity to CPSF-73 are predicted to function as nucleases in RNA metabolism. This article reviews the cellular roles of nucleases of the metallo-beta-lactamase family and the recent advances in studying these proteins.     

Importance of the +73/294 interaction in Escherichia coli RNase P RNA substrate complexes for cleavage and metal ion coordination

We have studied an interaction, the "73/294-interaction", between residues 294 in M1 RNA (the catalytic subunit of Escherichia coli RNase P) and +73 in the tRNA precursor substrate. The 73/294-interaction is part of the "RCCA-RNase P RNA interaction", which anchors the 3' R(+73)CCA-motif of the substrate to M1 RNA (interacting residues underlined). Considering that in a large fraction of tRNA precursors residue +73 is base-paired to nucleotide -1 immediately 5' of the cleavage site, formation of the 73/294-interaction results in exposure of the cleavage site. We show that the nature/orientation of the 73/294-interaction is important for cleavage site recognition and cleavage efficiency. Our data further suggest that this interaction is part of a metal ion-binding site and that specific chemical groups are likely to act as ligands in binding of Mg(2+) or other divalent cations important for function. We argue that this Mg(2+) is involved in metal ion cooperativity in M1 RNA-mediated cleavage. Moreover, we suggest that the 73/294-interaction operates in concert with displacement of residue -1 in the substrate to ensure efficient and correct cleavage. The possibility that the residue at -1 binds to a specific binding surface/pocket in M1 RNA is discussed. Our data finally rationalize why the preferred residue at position 294 in M1 RNA is U.     

RNA heptamers that direct RNA cleavage by mammalian tRNA 3' processing endoribonuclease.

Mammalian tRNA 3' processing endoribonuclease (3' tRNase) can recognize and cleave any target RNA that forms a precursor tRNA-like complex with another RNA. Various sets of RNA molecules were tested to identify the smallest RNA that can direct target RNA cleavage by 3' tRNase. A 3' half tRNAArgwas cleaved efficiently by 3' tRNase in the presence of small 5' half tRNAArgvariants, the D stem-loop region of which was partially deleted. Remarkably, 3' tRNase also cleaved the 3' half tRNAArgin the presence of a 7 nt 5' tRNAArg composed only of the acceptor stem region with a catalytic efficiency comparable with that of cleavage directed by an intact 5' half tRNAArg. The catalytic efficiency of cleavage directed by the heptamer decreased as the stability of the T stem-loop structures of 3' half tRNAArg variants decreased. No heptamer-directed cleavage of a 3' half tRNAArg without T stem base pairs was detected. A heptamer also directed cleavage of an HIV-1 RNA containing a stable hairpin structure. These findings suggest that in the presence of an RNA heptamer, 3' tRNase can discriminate and eliminate target RNAs that possess a stable hairpin adjacent to the heptamer binding sequence from a large complex RNA pool.     

MoxT toxin of Bacillus anthracis exhibits sequence specific ribonuclease activity

MoxXT module of Bacillus anthracis encodes MoxX, a labile protein and MoxT, a ribonuclease. However, mechanism of cleavage of RNA by MoxT has not been explored till date. In the present study, we have demonstrated that MoxT is a sequence specific ribonuclease which recognizes UACAU sequence in ss RNA and cleaves between U and A. Moreover, cleavage of RNA requires 2′ OH group of first residue, i.e. U of UACAU RNA sequence. An interesting finding which makes it distinct from the other MazF family toxins was also observed, i.e. its ability to cleave RNA in DNA–RNA hybrid.     

Proton NMR studies on the covalently linked RNA-DNA hybrid r(GCG)d(TATACGC). Assignment of proton resonances by application of the nuclear Overhauser effect.

Proton NMR spectra of a covalently linked self-complementary RNA X DNA hybrid, r(GCG)-d(TATACGC), are recorded in H2O and D2O. Imino proton resonances as well as the non-exchangeable base and H-1' resonances are unambiguously assigned by means of nuclear. Overhauser effect measurements. Additional information was obtained by 31P NMR and circular dichroism spectra. The RNA parts in the duplex attain full conformational purity and adopt the usual A-RNA conformation. The DNA residues opposite the RNA tract do not adopt an A-type structure completely. Their respective sugar rings still appear to possess a certain conformational freedom. The same holds true for the central d(-TATA-) sequence which forms a DNA X DNA duplex. There appears to be a structural break in this part: the first two residues, T(4) and A(5), are clearly influenced by the adjacent RNA structure, whereas residues T(6) and A(7) behave quite similar to what usually is found in DNA duplexes in aqueous solution.     

Sequence and translation of the murine coronavirus 5''-end genomic RNA reveals the N-terminal structure of the putative RNA polymerase.

A 28-kilodalton protein has been suggested to be the amino-terminal protein cleavage product of the putative coronavirus RNA polymerase (gene A) (M.R. Denison and S. Perlman, Virology 157:565-568, 1987). To elucidate the structure and mechanism of synthesis of this protein, the nucleotide sequence of the 5' 2.0 kilobases of the coronavirus mouse hepatitis virus strain JHM genome was determined. This sequence contains a single, long open reading frame and predicts a highly basic amino-terminal region. Cell-free translation of RNAs transcribed in vitro from DNAs containing gene A sequences in pT7 vectors yielded proteins initiated from the 5'-most optimal initiation codon at position 215 from the 5' end of the genome. The sequence preceding this initiation codon predicts the presence of a stable hairpin loop structure. The presence of an RNA secondary structure at the 5' end of the RNA genome is supported by the observation that gene A sequences were more efficiently translated in vitro when upstream noncoding sequences were removed. By comparing the translation products of virion genomic RNA and in vitro transcribed RNAs, we established that our clones encompassing the 5'-end mouse hepatitis virus genomic RNA encode the 28-kilodalton N-terminal cleavage product of the gene A protein. Possible cleavage sites for this protein are proposed.