Transcription infidelity and genome integrity: the parallax view

ABSTRACT

It was recently shown that removal of GreA, a transcription fidelity factor, enhances DNA break repair. This counterintuitive result, arising from unresolved backtracked RNA polymerase impeding DNA resection and thereby facilitating RecA-loading, leads to an interesting corollary: error-free full-length transcripts and broken chromosomes. Therefore, transcription fidelity may compromise genomic integrity.

KEWORDS:

• Transcription fidelity
• DNA break repair
• GreAB
• DksA
• RecBCD
• Epimutation
• Genomic integrity
• Backtracked RNA polymerase
• DNA resection
• Genetic noise

Gene expression and genetic replication are parallel and linear readouts of a template and therein lies complication and conflict: RNA transcription and DNA replication use the same template often at the same time but at different speeds and with different fidelities and often in opposing directions.

The agents of gene expression and genetic replication, nucleic acid polymerases, once engaged in the act, can be relentless when afforded 3ʹ complementarity with the last few nucleotides of the growing chain and an available template: DNA polymerase can misalign, realign and continue chain synthesis leading to complex de novo inversion, deletion and duplication mutational events1 RNA polymerase (RNAP) can also slip and skip along the template leading to insertions and deletions in the nascent transcript[2]. However, a polymerase may also be stopped dead in its tracks.

RNAP can backtrack, effectively stopping and creating an effective barrier to other enzymes and proteins travelling along the DNA double helix. Backtracking can occur as a result of ribonucleotide misincorporation[3], or transcription encounters with helix-distorting DNA lesions[4]. During backtracking, the nascent mRNA is displaced from the RNAP active site and extruded into the secondary channel preventing transcription elongation[5]. The Escherichia coli GreA and GreB factors interact with the RNAP secondary channel and restart backtracked RNAP by stimulating RNA cleavage[6], thereby realigning the 3ʹ end of the transcript with the active site, allowing transcription to resume[5]. Therefore, these Gre factors are considered transcription fidelity factors [7,8], and in their absence errors accumulate in nascent mRNA [9,10]. However, GreAB have also been shown to reduce conflicts between transcription and replication during nutrient stress[11], and now a new complexity has recently been shown between GreA, transcription and DNA break repair[12].

Pregnant pauses and futile repair

Surprisingly, deletion of greA increased cell survival to phleomycin (PHL), a radiomimetic drug (Figure 1(a)). The greA*(D41N) allele abolishes the transcript cleavage[13] and transcription restart function of GreA, and greA* cells exhibited similar survival to ΔgreA cells, implicating that loss of these functions is sufficient for increased DNA break survival. Interestingly, survival of ΔgreB cells did not look like that found in ΔgreA cells, but survival of the ΔgreBΔgreA double mutant was similar to the ΔgreA mutant. GreA and GreB have similar functions but differ in transcript cleavage specificity. While GreA cleaves di – and tri-nucleotides arising from misincorporation, GreB stimulates the cleavage of larger fragments[6] and this may explain the differential survival patterns observed and suggests that short RNAP backtracking due to misincorporation may enhance DNA break survival.

DksA, a transcription factor that also interacts with the RNAP secondary channel, does not perform transcript cleavage but can compete with the Gre factors for RNAP binding[14]. DksA is also involved in transcription fidelity [15,16] and can resolve conflicts between transcription and replication during nutrient stress[11]. Deletion of dksA increased sensitivity to PHL (Figure 1(a)) but, again, the increased survival of the ΔdksAΔgreA double mutant approached the ΔgreA level. Moreover, DksA overexpression increased survival of wild-type cells to the level of ΔgreA, suggesting that DksA competition for RNAP effectively excludes GreA rendering wild-type cells GreA-less (Figure 1(a)).

Figure 1. The absence of GreA increases survival to PHL treatment by the RecBCD-RecA pathway. (a) Average survival data of indicated mutant on 2 μg ml⁻¹ PHL (compared to non-PHL) for two days at 37°; for detailed results see Sivaramakrishnan et al. [12]. Red bars indicate strains with deleted or non-functional greA alleles; striped red bar indicates over-expression of DksA, which we suggest leads to increased survival by preventing GreA function through DksA successfully competing for RNAP binding to the exclusion of GreA. (b) DSB repair is initiated by RecBCD which resects DNA ends and then loads RecA to promote strand exchange. During transcription RNAP can become backtracked, creating a barrier to RecBCD resection. In the presence of GreA, transcription restarts and the barrier to RecBCD is resolved allowing continued resection leading to inefficient DNA repair. In the absence of GreA, the backtracked RNAP barrier remains causing RecBCD to pause and resection to cease allowing RecA loading and increased DNA repair[12].

Deletion of greA has little bearing on survival in a recombination defective ΔrecA or ΔrecB background, suggesting that increased backtracked RNAP promotes DNA double-strand break (DSB) repair by the canonical RecBCD–RecA pathway. In E. coli, DSB repair is initiated predominantly by the RecBCD exonuclease enzyme[17]. RecBCD resects DNA ends, degrading it into single-stranded DNA onto which it loads the RecA recombinase to promote strand exchange. RecBCD switches from DNA degradation to RecA loading at Chi (crossover hotspot instigator) sites in the DNA. By using a novel deep-sequencing method, XO-Seq, it was found that the build-up of backtracked RNAP that occurs in the absence of GreA reduces RecBCD resection[12], as well as the requirement for specific RecA loading activity of RecB. Therefore, this DNA transaction conundrum, that a transcription fidelity factor can compromise genomic integrity, has a resolution.

During transcription RNAP can become backtracked[5], posing a barrier to RecBCD resection (see Figure 1(b)). In the presence of GreA, an anti-backtracking factor, transcription restarts and the barrier to RecBCD is resolved allowing continued resection leading to inefficient DNA repair. Therefore, an actively elongating RNAP is a less efficient barrier to RecBCD resection than is an arrested backtracked RNAP[12]. In a sense this is another case of futile repair: it is the presence of a wild-type function causing subsequent problems for DNA repair, so while the operation is successful, the cell dies. Here, the presence of GreA restarts backtracked RNAP allowing resection to continue, and degraded chromosomes result; in alkylation-treated human cell lines, it is the presence of active mismatch repair that causes the cells problems since rounds of interfering misinsertion and correction occur, while in mismatch repair deficient lines, the misinsertions remain and the cells are now tolerant to the treatment and grow[18]. In the absence of GreA, the backtracked RNAP barrier remains causing RecBCD to pause and resection to cease allowing RecA loading and increased DNA repair (Figure 1(b)). Therefore, we propose that backtracked RNAPs can instigate recombination in a manner analogous to Chi site function[12]. Indeed, wild-type cells have tamed backtracked RNAP to better repair DNA through the use of the UvrD helicase [4,12], which when associated with RNAP stimulates backtracking in the presence of ppGpp [12,19].

Gre factors maintain epigenetic inheritance

The trade-off between transcription fidelity and DNA break repair highlights the importance of maintaining quality mRNA. Our previous studies demonstrate that the Gre factors play a role in maintaining epigenetic inheritance in a gene network [20,21]. Our epimutation system is based on the classic 1957 study of Novick and Weiner[22], who described how the lac operon epigenetic switch can produce alternative heritable phenotypes from genetically identical cells grown in the same environment.

The lac operon comprises an autocatalytic positive feedback loop Figure 2(a)) allowing a heritable all-or-none epigenetic switch at a maintenance concentration of inducer (that concentration of inducer which does not activate transcription of the operon but allows an already induced cell to remain induced [22]; Figure 2(b)). The lac repressor is rare (~ 5 tetramers per cell) and the lacI gene is infrequently transcribed and therefore this system is sensitive to fluctuations in lac repressor numbers. A transient depletion of repressor within a cell will lead to a transient derepression of the operon, producing a burst of lacY permease gene expression.[23]. At the maintenance concentration of inducer, the presence of permease will activate the positive feedback loop, so that the new induced state will be heritably maintained through cell division in a clonal cell population [21,24]. We harnessed this lac memory-module to capture and monitor the consequences of transient transcription errors in E. coli, providing a forward epimutation approach to study the factors and processes involved in modulating RNA fidelity[20].

Figure 2. Heritable stochastic switching in the E. coli lac operon: a forward epimutation system. (a) The lac operon comprises an autocatalytic positive feedback loop (the presence of LacY permease will produce more permease) allowing a heritable all-or-none epigenetic switch at a maintenance concentration of inducer. LacY is presented in green because the lacA gene is replaced by gfp, and the lac operon is now lacZYA::gfp, so when permease is made GFP will also be made and the cell will fluoresce green. Stochastic events that lead to a transient depletion of repressor (in red) within a cell will result in a burst of LacY permease expression and the presence of permease will activate the positive feedback loop, so that the new induced state will be heritably maintained, mimicking mutation[22]. While it is the presence of LacY permease that allows the positive feedback loop to initiate and continue, it is the faithful transcription of the lacI gene that dictates cell fate. (b) Wild-type cells that were originally ON (open circles) or OFF (closed circles) were sub-cultured and grown in media containing various concentrations of inducer thio-methylgalactoside (TMG). The shaded area highlights the maintenance concentration of inducer for these strains; hysteresis and bistability in this system is observed [21,24]. (c) When OFF cells are grown in the maintenance concentration of inducer, the absence of GreA and GreB (blue flow cytometry histograms) increases the proportion of ON cells with respect to wild-type cell levels (red flow cytometry histograms) [20,21]. Each line represents an independent histogram of 10,000 cells. The increase in stochastic switch frequency is 38-fold over wild-type level[21].

Using single-cell analysis, it was shown that the frequency of epigenetic switching from the OFF state to the ON state of the lac operon is increased when the fidelity of RNA transcription is decreased due to an error-prone RNA polymerase[21], an error-prone transcription sequence[24] or in the absence of RNA fidelity factors GreA and GreB [21,24] (see Figure 2(c)), or DksA[16]. This system was also used to show that the absence of MutT function does not create an epimutator phenotype[25], which had previously been suggested to be a RNA mutator[26]. Therefore, transcription infidelity contributes to molecular noise (stochastic fluctuations in the level of functional proteins produced) and can effect heritable phenotypic change in genetically identical cells in the same environment. The steps in other microbial epigenetic phenotypic switches that are susceptible to molecular noise have already been highlighted[27], and epimutation may play a role in stochastic switching in these systems.

Gre factor redundancy and revival

A recent reverse epigenetic/genetic approach has been described to monitor riboA misinsertions opposite template C at one specific site in the E. coli genome: a catalytically inactive Cre recombinase can be restored to function by such a transcriptional misinsertion and this event monitored by Cre-dependent DNA recombination that converts a mutant gal gene to Gal⁺[28]. Therefore this system requires a full-length revertant mRNA to be translated into functional Cre tetramers. These authors found that the absence of GreA alone increases this phenotype ~ 100-fold over wild-type or a ΔgreB strain, and that the absence of GreB alone had no affect. It was noted though, that overexpression of GreB in ΔgreA cells reduced the misincorporation event to wild-type levels, therefore, GreB can correct G to A transcription errors in vivo. Similarly, and in equally dramatic fashion, GreB overexpression in ΔgreA cells also returns PHL sensitivity back to wild-type levels (greater than ~ 100-fold increase)[12], presumably by resolving those backtracked RNAP that would normally be processed by GreA in vivo. Therefore, under these very different conditions, misincorporations leading to full-length mutant mRNA (where backtracked RNAP would not be relevant)[28] and misincorporations leading to backtracked recalcitrant RNAP (where full-length mutant mRNA would not be relevant)[12], GreB lingers in the shadow of GreA, bordering on functional redundancy, but GreB can emerge and act under certain circumstances, and has been likened to an ‘emergency’ proofreading factor[28].

In our epigenetic switch system, we do not observe an increase in stochastic switching in the absence of GreA or GreB alone, but only observe a significant increase in switching when both Gre factors are missing (Figure 2(c)) [21]. Our forward epimutation system (any epimutation event that precludes a full-length functional mRNA will have consequences) is based on (i) infrequent transcription of a repressor gene that is (ii) oriented toward the origin of replication, which would result in head-on collision with the replication fork when encountered, and (iii) performed under slow growth conditions (minimal salts and succinate; generation time ~ 90 min). All these aspects may come into play, to allow the innate character of GreB to manifest itself and its loss contribute to a phenotype. Interestingly, ΔgreB mutants are more sensitive to sublethal amounts of the translation inhibitor chloramphenicol than wild-type cells implicating backtracking in chromosomal instability and the importance of translation in the maintenance of chromosome integrity[29]. Therefore, GreB may aid a solitary RNAP moving down a template and possibly encountering another solitary DNA polymerase arriving head-on. Indeed, it has been shown that an array of closely spaced elongating RNAP prevents excessive backtracking [30,31]. It would be of particular interest to know the sequence nature of any transcript indels produced in the ΔgreAB background and how they compare to transcript indels found in the presence of GreA or GreB since these factors revive different forms of backtracked RNAP.

The medium is the message?

A pause assumes a resumption, if not it is a termination. Misincorporations lead to backtracked RNAP that can be restarted by Gre factors and that prompts the question is Gre more a processivity factor or more a fidelity factor? James et al. [10] cogently distinguish between RNA misincorporations during transcription leading to the production of full-length mutant mRNA (the message) and RNA misincorporations during transcription leading to backtracked RNAP elongation complexes that may physically block transcription of a gene or interfere with the progression of replication forks and subsequently alter gene expression patterns (the medium). They emphasize that it is not the correctness of the final RNA product per se that is paramount but rather it is the absence of protein product due to the physical block of a misincorporated backtracked RNAP or other deleterious effects of stalled misincorporated complexes that can contribute to molecular noise and effect phenotypic heterogeneity and possibly human (epi)genetic disease [32,33].

We view transcriptional infidelity as any event (misincorporation, frameshift, stalled, aborted or premature transcription termination) that precludes the eventual production of a wild-type functional mRNA once a transcript has been initiated20 we consider such events as epimutations, a breakdown in the accurate flow of genetic information from DNA to RNA to protein due to transcription errors. Therefore, we find no dichotomy between misincorporations leading to full-length mutant mRNA and misincorporations leading to backtracked RNAP blockages preventing gene expression, since such infidelity would contribute to molecular noise and epimutagenesis.

Transcription allows constant monitoring of the genome and is harnessed to aid DNA repair[34]. Here we describe how GreA, an auxiliary transcription factor, can impede DNA break repair[12] and yet GreA, GreB and DksA can also mediate DNA/RNA conflicts and promote genome stability [11,29]. Above these DNA events, such factors also function in transcription fidelity and in their absence epimutations increase [9,10,15,16,21,28]. Therefore, these auxiliary factors modulate RNAP capability to influence the fidelity and integrity of the parallel processes of DNA replication/repair and RNA transcription from a singular template, and the relative importance of their functional consequences depends on the process in question.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by a grant from the National Institutes of Health USA (RO1-GM088653) awarded to C.H;National Institute of General Medical Sciences [GM088653].