To access the private area of this site, please log in.

Username: 

Password:  

 

Chapter 51

The DNases, Topoisomerases, and Helicases of Escherichia coli

STUART LINN

 

INTRODUCTION

This chapter provides an overview of enzymes of Escherichia coli known to alter the primary, secondary, or tertiary structure of DNA. A minimal number of references is provided so as to allow entry into the literature. Other references and additional information can be found in chapters of this volume focusing upon DNA replication (chapter 50), DNA restriction and modification (chapter 52), DNA methylation (chapter 53), DNA recombination (chapters 119 and 125), and DNA repair (chapter 121), in a recent review by Kowalczykowski et al. (37), and in a recent monograph on nucleases (48).

Table 1 lists the genes, map positions, and subunit molecular weights of each enzyme. Map positions are according to Berlyn et al. (chapter 109), and molecular weights are based upon the translated gene sequence where available.

It is particularly striking in perusing the collection of enzymes to note that each function of each enzyme appears to be duplicated by one or more other enzymes. Such duplication emphasizes the importance of these enzymes and provides fascinating puzzles to solve for the researcher who is studying these E. coli enzymes. It also is important for those using E. coli as an expression system for homologous enzymes from other sources to be aware of this complex background. Finally, it is impressive, given this exhaustive battery of enzymes, that phage and plasmids still often further supplement this array for their DNA metabolism needs.

 

DNASES

All DNases of E. coli produce 3'-hydroxyl termini and 5'-phosphomonoester termini except for the type I restriction enzymes, which produce unknown 5' termini, and the apurinic/apyrimidinic (AP) endonucleases acting by a β-lyase mechanism, which produce an α,β-unsaturated aldehyde at the 3' terminus.

 

Exonucleases

Exonuclease I (28, 62).

Exonuclease I is specific for single-stranded DNA or 3' single-stranded tails. It degrades processively in a 3'→5' direction to yield 5'-mononucleotides and a dinucleotide which originates from the 5' terminus of the single-stranded DNA substrate.

Certain sbcB mutants suppress the recB and recC mutant phenotypes, apparently by allowing recombination to proceed more efficiently via the RecF pathway. Exonuclease I also takes part in methylation-directed mismatch repair (Fig. 1) and, along with RecBCD, degrades restricted DNA in vivo. We expect, therefore, that the role of exonuclease I is to aid in DNA repair, DNA restriction, and the control of undesirable levels or forms of DNA recombination. It is used as a reagent to remove linear single-stranded DNA from single-stranded circles, duplex DNA, polymerase chain reactions, etc., as well as to remove 3' single-stranded tails.

DNA Polymerase I Exonucleases (36).

Originally called exonucleases II and VI, respectively, the DNA polymerase I exonuclease activities consist of two independent catalytic entities. The first is a 3'→5' exonuclease activity which forms mononucleotides and acts on mismatched termini on duplex DNA; it also degrades single-stranded DNA and oligonucleotides. It serves as the "editor" which acts in conjunction with the polymerizing activity to increase fidelity.

The second activity is a 5'→3' exonuclease present on the C-terminal domain of the protein and absent from the Klenow fragment. It acts upon nicked duplex DNA or RNA:DNA hybrids, from which it removes oligodeoxyribonucleotides or ribonucleotides, respectively. In the former case, it functions in a nick-translation mechanism to remove DNA damages; in the latter case, it functions to remove the RNA primers associated with DNA replication.

DNA Polymerase II Exonuclease (36).

DNA polymerase II exonuclease has only a 3'→5' exonuclease, presumably serving in an editor capacity.

DNA Polymerase III Exonuclease (36).

DNA polymerase III exonuclease also has only the 3'→5' editing exonuclease, which is associated with a separate subunit, ε. The activity also degrades single-stranded DNA. It forms mononucleotides, in a nonprocessive reaction, but leaves the 5'-terminal dinucleotide intact. It is stimulated by the presence of the catalytic (α) subunit, but does not require it. Mutants in the ε gene, dnaQ, have a mutator phenotype. It is noteworthy that in this case the editor exonuclease activity is on an accessory subunit, unlike E. coli DNA polymerases I and II or eukaryotic DNA polymerases γ, δ, or ε, but like eukaryotic DNA polymerases α and β.

Exonuclease III (36, 65).

Also referred to as endonuclease II or endonuclease VI in the older literature, exonuclease III is a remarkable DNA repair powerhouse. It acts on double-stranded (but not single-stranded) DNA as a 3'→5' exonuclease, releasing mononucleotides and leaving the 5' portions of the duplex (averaging half the length of the original duplex) intact. It acts from ends or nicks and has been applied as a reagent for the formation of "nested deletions" and for widening gaps. The enzyme demonstrates similar exonuclease activity on both the DNA and the RNA strands of RNA:DNA hybrids.

In repair, the enzyme removes 3'-terminal phosphomonoester groups, 3'-sugar phosphate groups such as the α,β-unsaturated deoxyribose phosphate formed by β-elimination at AP sites, or 3'-sugar phosphate fragments such as 3'-phosphoglycerate which are formed by oxyradical damage. In short, the enzyme "cleans up" DNA 3' termini.

Finally, exonuclease III is the major AP endonuclease of E. coli, acting by a hydrolysis mechanism so as to leave 3'-nucleoside termini and 5'-deoxyribose phosphate termini. Homologs of exonuclease III exist in eukaryotes.

Exonucleases IVA and IVB (34).

Exonucleases IVA and IVB release mononucleotides from deoxyoligonucleotides. They have not been characterized in recent years.

RecBCD Enzyme (Exonuclease V) (23, 37, 53).

The trimeric enzyme RecBCD (exonuclease V) consists of one each of the RecB, RecC, and RecD subunits. It processively degrades both single-stranded and duplex DNAs from 3' and 5' termini in an ATP-dependent, exonucleolytic manner to form oligonucleotides; it also has endonuclease activity upon single-stranded DNA which is stimulated by ATP. The enzyme unwinds duplexes from their ends by virtue of an ATP-dependent helicase activity. Upon encountering a chi site in duplex DNA in the proper orientation (5'-GCTGGTGG-3'), it nicks the DNA 4, 5, or 6 nucleotides 3' to that site. Figure 2 describes a model of how these reactions might be integrated to help to catalyze homologous recombination.

RecBCD is implicated in the majority of homologous recombination events in E. coli and also is responsible, with exonuclease I, for degrading the products of restriction. In vitro the enzyme has been used to efficiently remove single-stranded DNA and duplex linear molecules from duplex DNA circle preparations. With RNA:DNA hybrid molecules, it degrades DNA tails, but leaves flush RNA:DNA ends or RNA tails intact.

Exonuclease VII (15, 28).

The heterodimeric exonuclease VII acts upon single-stranded DNA, in particular single-stranded tails or displaced single strands on DNA duplexes. It processively degrades this material to oligonucleotides which can include damaged nucleotides. It acts from 5' ends in methylationdirected mismatch repair (Fig. 1). Exonuclease VII appears to be responsible for cleaning up loose ends from DNA, presumably an important process since all of its activities appear to be duplicated by other enzymes in E. coli. In vitro the enzyme is useful for removing tails from DNA duplexes.

Exonuclease VIII (30).

Exonuclease VIII is a fusion of the products of the adjacent recE and recT genes. The RecE (N-terminal) portion catalyzes the DNase activity, whereas the RecT portion promotes DNA renaturation. The two activities show functional, but not sequence, homology to the α and β gene products of the λ red system (λ exonuclease and beta protein, respectively). Indeed, they seem to derive from the presence of the cryptic Rac prophage.

Like λ exonuclease, exonuclease VIII processively removes mononucleotides from the 5' ends of duplex, but not single-stranded DNA. Certain mutations in recE (sbcA mutants) are able phenotypically to suppress recB and recC mutants by substituting RecA-independent recombination for the defective RecBCD-mediated recombination. Hence, in a recB sbcA double mutant, the situation is analogous to λ red-mediated recombination in which the λ gam gene product inactivates RecBCD enzyme, allowing a 5'→3' exonuclease/renaturase mechanism to substitute.

RecJ (28, 50, 51).

The RecJ activity is to some degree the complement of exonuclease I: it acts on single-stranded DNA, displaced single strands, or single-stranded tails on duplex DNA, but in a processive 5'→3' manner to give mononucleotides. It functions in methylation-directed mismatch repair (Fig. 1). As opposed to sbcB mutants, which increase recombination in recB or recC strains by inactivating exonuclease I, recJ mutants decrease recombination which is independent of RecBCD but dependent upon RecE or RecF. The differing phenotypes as a consequence of the loss of exonuclease I or RecJ emphasize the importance of single-strand tail polarity in governing DNA transactions.

dRPase(s) (11, 26).

The DNA deoxyribophosphodiesterase (dRPase) activities specifically remove sugar phosphate residues from DNA termini. A novel activity was reported (26) that removes deoxyribose 5-phosphate when present at 5' termini, a substrate formed by exonuclease III acting at AP sites, for example. Another novel activity removed α,β-unsaturated deoxyribose 5-phosphate when present at 3' termini (11), a substrate formed by β-elimination at AP sites by endonuclease III or Fpg, for example (see below). The relation between these two activities with regard to their protein association is currently unclear, but they presumably both function in base excision DNA repair.

 

Endonucleases

Endonuclease I (24, 45).

The very potent enzyme endonuclease I makes double-strand breaks in duplex DNA, releasing roughly 400 nucleotides as oligonucleotides during each such cleavage. It is a periplasmic enzyme of unknown function. It is strongly inhibited by double-stranded RNA, and in the presence of tRNA it makes only widely scattered nicks in DNA duplexes. Although endonuclease I is the dominant DNase activity in E. coli, crude extracts from the organism do not demonstrate such activity due to the presence of tRNA. It is generally wise to add tRNA or use endA strains when studying DNA enzymes in nonpurified fractions from E. coli extracts.

Endonuclease III(8, 39).

The ubiquitous endonuclease III is responsible for excising oxidized pyrimidine residues from DNA via base excision repair. The enzyme acts as a DNA glycosylase, removing such damaged pyrimidines as thymine glycol and dihydrothymine, or fragments of pyrimidine bases. The resulting baseless sites (or preexisting baseless sites) are also cleaved by the enzyme via a β-lyase mechanism to leave α,β-unsaturated deoxyribose at the 3' terminus. The enzyme acts in EDTA and is inhibited by tRNA. It has been the prototype for a ubiquitous activity found in most organisms.

Endonuclease IV (46, 64).

Endonuclease IV catalyzes many of the same reactions as exonuclease III. It acts on duplex DNA, removing 3'-phosphomonoester or 3'-sugar phosphate groups or their fragments. It hydrolyzes AP sites to yield 5'-deoxyribose residues. It is active in EDTA. While it generally accounts for a small part of these activities in unstressed E. coli, it is inducible by superoxide- and oxygen radical-generating agents such as paraquat. It is similar in activity and specificity to the major mammalian AP endonucleases.

Endonuclease V (19).

Endonuclease V hydrolyzes single-stranded DNA and cleaves duplex DNA containing certain bulky adducts or uracil, forming one double-strand break per roughly eight nicks. It also hydrolyzes undamaged DNA at alkaline pH. It processively degrades DNA molecules and requires Mg²⁺. Identification of its function and gene remains elusive.

Endonuclease VII(13).

Endonuclease VII has the unique property of cleaving single-stranded DNA with AP sites in the presence of EDTA. Its activity and genetic basis warrant further study.

Endonuclease VIII (58, 77).

The monomeric enzyme endonuclease VIII seems to duplicate the activities of endonuclease III. It removes oxidized pyrimidines from DNA via a DNA glycosylase mechanism, and it cleaves AP sites in DNA duplexes via β-elimination. The presence of endonuclease VIII probably explains the very weak phenotype of nth mutants that lack endonuclease III activity, but its exact niche in the cell’s DNA repair functions is unclear.

Fpg (MutM) (12, 14, 16).

The ubiquitous enzyme formamidopyrimidine DNA glycosylase (Fpg, MutM), which acts upon purine residues which were subject to damage by oxyradicals, is the purine counterpart of endonuclease III. It catalyzes the removal of oxidized or ring-opened purine bases (most notably, perhaps, formamidopyrimidines and 8-hydroxyguanine). It also cleaves AP sites by a β-elimination mechanism. However, it appears to then go on to catalyze δ-elimination so as to release α,β,γ,δ-unsaturated deoxyribose and leave a 3'-phosphomonoester residue on the DNA. Why it continues on to the δ-elimination reaction whereas endonuclease III does not is curious.

UvrABC(29, 47, 74).

By the elaborate mechanism outlined in Fig. 3, the UvrABC complex excises bulky DNA adducts as part of a 12- to 13-mer with the damage 7 to 8 nucleotides from the 5' end of the oligonucleotide. ATP is hydrolyzed to drive translocation and enzyme reactivation. The nick 3' to the damage is catalyzed by UvrB; that 5' to the damage is catalyzed by UvrC. This system is responsible for a large part of nucleotide excision repair in E. coli.

MutH (17, 27).

As outlined in Fig. 1, MutH nicks DNA adjacent to the unmodified strand of a hemimethylated GATC modification site during methylation-directed mismatch repair. In response to a replication error resulting in a mismatch, MutS and MutL recruit MutH to its substrate site. The system operates most efficiently with G:T, A:C, A:A, and G:G mismatches; it does not respond to C:C mismatches.

Vsr Endonuclease (31, 68).

Vsr endonuclease is the endonucleolytic component of Vsp (very short patch) repair, the responsibility of which is to correct mismatches brought about by deamination of 5-methylcytosine to form thymine at DNA cytosine methyltransferase (DCM) modification sites. As with MutH, Vsr is recruited by MutL and MutS, but since the methylation and mismatch sites overlap in this case, no helicase is needed. The nick is made 5' to the dTMP:

5' . . . C↓T(A/T)G G . . .

. . . G G(T/A)C*C . . . 5'

RuvC (10, 33, 66, 67).

RuvC cleaves Holliday junction recombination intermediates during postreplication repair and RecA-dependent homologous recombination. It introduces symmetrically related nicks into two strands of the same polarity, each 3' to a TT sequence within the consensus 5' . . . (A/T) T T ↓ (G/C) . . . 3' (Fig. 4). It probably acts in conjunction with RuvA and RuvB.

EcoK, EcoB (25, 49; chapter 52, this volume).

EcoK (EcoB) acts as a multimeric complex of at least one each of three peptides in catalyzing the surprisingly complex reaction illustrated in Fig. 5. The enzyme requires S-adenosylmethionine (AdoMet) and ATP. It does not turn over, and it translocates from an unmodified recognition site to an undefined cleavage site several kilobase pairs away by hydrolyzing ATP. It forms loops in the process so that the cleavage and recognition sites are juxtaposed. One enzyme molecule makes a gap of some 100 nucleotides releasing oligonucleotides; then a second cleaves the opposite strand to leave a 3' overhang of roughly 100 nucleotides and a 5' terminus blocked in an unknown manner. (No part of the ATP or AdoMet is found at the block.) The recognition sites for EcoB and EcoK are 5'-TGA(N₈)TGCT and 5'-AAC(N₆)GTCG, respectively. It remains a mystery as to why this system is so elaborate compared to type II restriction enzymes, though some speculations are mentioned in the legend to Fig. 5.

Whereas the HsdR, HsdM, and HsdS subunits are required for restriction, HsdM and HsdS suffice for methylation. No methyl group transfer seems to occur during restriction, however. Bickle (chapter 52, this volume) elaborates further on these and other type I restriction enzymes.

McrBC (22, 69).

Mcr (modified cytosine restriction) degrades DNA with fully or hemimethylated sites separated by 40 to 80 nucleotides of the type RmC N_40–80 RmC, where mC is 5-methyl-, 5-hydroxymethyl-, or N₄-methylcytosine. The N_40–80 spacer region is subject to multiple cleavages in the process. McrB is activated by GTP and then recruits McrC to form the active complex. The system restricts nonglycosylated T-even phage in vivo.

McrA (63).

McrA also restricts T-even phage. Its gene is associated with an unstable element found at map position 25, and its reaction has not been studied in vitro.

Mrr (38, 76).

Mrr recognizes DNA containing A or C modifications including 6-methyladenine and 5-methylcytosine. It degrades DNA carrying a number of type II modification methylations. To date, no successful demonstration of its in vitro activity has been reported, and it is likely to act as part of a complex. The adjacent location of the genes, mcrB-hsdS-hsdM-hsdR-mrr, at 99 min is intriguing in this regard.

Topoisomerase I (52).

The first topoisomerase discovered, also known as ω protein, topoisomerase I (a type I topoisomerase) relaxes negatively coiled, supercoiled DNA by forming intermediates with DNA whose nicks have 5' termini bound to tyrosine 319 of the enzyme. Like other topoisomerases, these nicks can be made permanent if the intermediates are disrupted by chemicals or heat.

Topoisomerase II (32).

Topoisomerase II (a type II topoisomerase), also known as DNA gyrase, is unique in its ability to catalyze negative supercoiling of DNA in the presence of ATP. The enzyme, acting as an A₂B₂ tetramer, forms intermediates with double-strand breaks containing 4-nucleotide-long 5' tails bound to tyrosine 122 of the GyrA subunits. These double-strand cleavages become permanent in the presence of a number of drugs or heat.

Topoisomerase III (20, 21).

Topoisomerase III (a type I topoisomerase) relaxes supercoils in a reaction which, when disrupted, puts nicks into DNA and RNA at like sequences. It can act as a backup in topA mutants.

Topoisomerase IV (4, 35, 61).

Topoisomerase IV, a type II topoisomerase, seems to function in the resolution of daughter molecules after circular chromosomal replication. Its detailed in vitro characterization is in progress.

Helicase I (1, 3, 40).

Helicase I was the first helicase to be identified by its DNA-dependent ATPase activity. This large enzyme acts as a multimer in an ATP-dependent, processive DNA unwinding in the 5' → 3' direction. It was found to map to the traI gene on the F plasmid. It does not seem to turn over from one DNA molecule to another. It is probably involved in the transfer of the F plasmid to recipient cells and is not utilized for host chromosomal metabolism.

Helicase II (2, 54, 55, 59).

Helicase II unwinds DNA in a stoichiometric reaction that depends upon the ratio of DNA to protein. The enzyme is usually associated with DNA repair, and its particular roles in excision repair and mismatch repair have been noted above (see also Fig. 1 and 3). Helicase II appears to regulate recombination in a negative sense in that certain uvrD mutants have enhanced recombination. It has also been speculated that helicase II might be associated with replication. Helicase II also efficiently unwinds RNA:DNA hybrids. Clearly the level of helicase II in the cell is important in determining the extent of unwinding by the enzyme due to its stoichiometric nature.

Helicase III (18, 57, 81).

Helicase III still requires detailed analysis of its activity and its gene. The 20-kDa peptide appears to act as a dimer and appears to act in a 5'→3' direction.

Helicase IV (77, 79, 80).

Helicase IV seems to unwind efficiently only short duplexes in the 3'→5' direction. It is not efficient with large duplexes, and its role is unknown.

Rep (6, 7, 41, 82). Rep protein, the first helicase discovered in E. coli, was identified by its requirement for replication of the DNAs of small phage (φX174, fd, etc.). The enzyme unwinds DNA in the 3' → 5' direction and prefers short duplexes. The Rep protein may interact with the CisA protein of φX174 during replication of that phage. The role of Rep in host cell metabolism is not established, though rep mutants have slower than normal replication fork movement.

PriA (42, 43, 44, 60).

Previously known as factor Y and n' protein, PriA was discovered and characterized by several laboratories studying the replication of φX174 DNA. The protein binds a specific sequence on the φX174 phage DNA, directs the primosome assembly, and then functions as a helicase. The enzyme prefers small duplexes, does not appear to turn over to a new substrate, and acts in the 3' → 5' direction.

DnaB (5, 9, 70, 75).

Isolated as a φX174 replication factor, the 52-kDa DnaB protein appears to act as a hexamer. The enzyme unwinds DNA in the 5' → 3' direction and appears to require a 3' tail to effectively unwind the DNA. DnaB is required for initiating unwinding of DNA from the replication origin. It also is a component of the primosome complex for lagging-strand synthesis. While DnaB is said to be the helicase for driving DNA replication, PriA is certainly also a component of that process.

RecQ (72, 73).

RecQ unwinds DNA in the 3' → 5' direction and is especially efficient in the presence of single-strand binding protein. It appears to be involved in the RecE and RecF recombination pathways, in which it may aid in RecA-mediated strand exchange.

RuvAB (71).

The RuvA and RuvB proteins act in a complex to catalyze branch migration of Holliday junctions and are required for normal levels of genetic recombination and DNA repair.

The helicase also promotes ATP-dependent strand displacement in the 5' → 3' direction, most efficiently with small duplexes. The DNA helicase sequence motifs are on the RuvB protein. RuvAB appears to act in concert with RuvC to resolve Holliday junctions.

RecG (78).

The RecG protein has the Holliday junction translocase and DNA-dependent ATPase activities expected for a protein with the same function as RuvAB, a prediction also made on genetic grounds. Although DNA helicase activity has not been reported with a strand displacement assay, the protein contains the DNA helicase sequence motifs. RecG appears not to act in conjunction with RuvC, but to resolve Holliday junctions by reverse branch migration (78). It will be interesting to learn the precise roles of the RuvAB versus RecG systems.

 

TOPOISOMERASES

Like many (or all?) organisms, E. coli has two each of type I and type II topoisomerases, probably reflecting their roles in virtually all biochemical processes utilizing DNA. The discovery and elucidation of the functions and mechanisms of the topoisomerases are among the true milestones of DNA enzymology.

 

DNA HELICASES

As for DNases and topoisomerases, the duplication of helicases often obscures their individual roles in vivo. Besides the helicases noted below, the RecBCD and UvrAB helicase activities are noted in preceding sections of this chapter. A recent review (56) also elaborates on the E. coli DNA helicases.

 

PERSPECTIVES

The most evident characteristic of the list of enzymes discussed in this chapter is its length; the number of these enzymes was certainly not anticipated, yet it will certainly increase. There are probably several reasons for this multiplicity besides the simple "back-up" explanation. Clearly replication, repair, recombination, and restriction are each much more multifaceted than our original simplistic models anticipated. Some of these enzymes probably act only under specific circumstances and when complexed specifically to particular accessory proteins. Others probably are induced or activated only under certain stress or growth conditions. Undoubtedly molecular geneticists, biochemists, and structural biologists will sort out many of these questions as E. coli continues to be the paradigm organism for understanding DNA metabolism.