Understanding DNA Repair in Hyperthermophilic Archaea: Persistent Gaps and Other Reasons to ...

13 Pages·2015·1.43 MB·English

Hindawi Publishing Corporation
Archaea
Volume 2015, Article ID 942605, 12 pages
http://dx.doi.org/10.1155/2015/942605
Review Article
Understanding DNA Repair in Hyperthermophilic Archaea:
Persistent Gaps and Other Reasons to Focus on the Fork
DennisW.Grogan
DepartmentofBiologicalSciences,UniversityofCincinnati,614RieveschlHall,CliftonCourt,Cincinnati,OH45221-0006,USA
CorrespondenceshouldbeaddressedtoDennisW.Grogan;[email protected]
Received26February2015;Accepted21May2015
AcademicEditor:YoshizumiIshino
Copyright©2015DennisW.Grogan.ThisisanopenaccessarticledistributedundertheCreativeCommonsAttributionLicense,
whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
AlthoughhyperthermophilicarchaeaarguablyhaveagreatneedforefficientDNArepair,theylackmembersofseveralDNArepair
proteinfamiliesbroadlyconservedamongbacteriaandeukaryotes.Conversely,theputativeDNArepairgenesthatdooccurin
thesearchaeaoftendonotgeneratetheexpectedphenotypewhendeleted.Theprospectthathyperthermophilicarchaeahavesome
uniquestrategiesforcopingwithDNAdamageandreplicationerrorshasintellectualandtechnologicalappeal,butresolvingthis
questionwillrequirealternativecopingmechanismstobeproposedandtestedexperimentally.Thisreviewevaluatesacombination
offourenigmaticpropertiesthatdistinguishesthehyperthermophilicarchaeafromallotherorganisms:DNApolymerasestallingat
dU,apparentlackofconventionalNER,lackofMutSLhomologs,andapparentessentialityofhomologousrecombinationproteins.
Hypotheticaldamage-copingstrategiesthatcouldexplainthissetofpropertiesmayprovidenewstartingpointsforeffortstodefine
howarchaeadifferfromconventionalmodelsofDNArepairandreplicationfidelity.
1.GenomeIntegrityandArchaea archaea employ “eukaryotic” DNA replication proteins to
propagate small, circular chromosomes in the context of
Theimportanceofmaintainingtheintegrityofcellulargen-
a small “prokaryotic” cell [3–7]. In addition, many groups
omes, and the diversity of processes that threaten it, can
of archaea are adapted to environmental extremes which
be seen in the network of sophisticated mechanisms that
impose molecular constraints not accommodated by the
cope with DNA damage and replication errors in even the
microorganisms used historically as models of molecu-
simplest microorganisms. Consistent with their deep inte-
lar biology. To the extent that archaea may have genetic
grationintocellularbiology,damage-copingstrategiesshow
mechanismsnotrepresentedinbacteriaoreukaryotes,they
both conservation and divergence; eukaryotes and bacteria
employsimilarsetsofbasiccopingstrategies,forexample,but representapotentialsourceofnewinformationconcerning
thesedifferintheirmechanisticdetails.Thesimilaritiesmay earlycellularevolution,aswellasprocessesthatmaybeused
reflectthecommonbiologicalthreatsposedbyDNAlesions for biotechnology, especially if these processes occur in the
and mutations, whereas the differences point to the deep archaeathatareadaptedtoextremeconditions.
evolutionary divergence between these two fundamentally Molecular mechanisms not known from other systems
differenttypesofcells[1]. are by nature, however, difficult to define, and confirming
There are legitimate reasons to expect that archaea, such mechanisms will require genetics, including the con-
especially the hyperthermophiles, differ from both bacteria struction and analysis of mutant strains. Because hyper-
andeukaryoteswithrespecttomechanismsofgenomemain- thermophilic archaea (HA) are challenging to manipulate
tenance. As predicted by its distinct evolutionary history as living cells, their genetic processes have been addressed
[2], the third domain of life shares certain fundamental overwhelmingly via biochemical and structural analyses
molecular features only with eukaryotic cells, others only of purified proteins, with little input from the comple-
with bacterial cells, and has yet other features that are mentary and essential perspective of functional genetics.
uniquely archaeal. With respect to genome maintenance, Paraphrasing a sentiment once expressed by the author’s
2 Archaea
doctoral research advisor, “When you do biochemistry, at least one TLS polymerase, Holliday-junction resolvases,
󸀠
you tell the organism what you think is important; when a 3 -flap endonuclease, a recombinase homolog, and two
you do genetics, the organism tells you what it thinks is putativeuracilDNAglycosylases(UDG).Inmostcases,the
important.” documented phenotype is increased sensitivity to one or
more DNA-damaging agents, although the UDG mutants
showed impaired growth [10] and the Y-family polymerase
2.TheFunctional-GeneticErafor
mutantshowedanalteredspectrumofspontaneousmutation
HyperthermophilicArchaea
[11].
MutantsinGroupIIidentifygeneswithminimalimpact
ForHA,basicgeneticanalyseshavebecomefeasibleonlyin
onDNArepaircapacity,asevidencedbylittleornodetectable
thepastdecade,andoneofthefirstofthesestudiestargeted
increaseinsensitivitytoradiationorDNA-damagingchem-
the reverse DNA gyrase of Thermococcus kodakaraensis
icals. In some cases, notably HJ resolvases, Hjc and Hje,
[8]. Reverse gyrases (Rgy) are single-strand-nicking DNA
singlemutantsexhibitedonlymarginalphenotypes,butthe
topoisomerasesthatintroducepositivesupercoilsintoDNA
double mutant was apparently nonviable [12]. The latter
and appear only in extreme- or hyperthermophilic archaea
case thus suggests that some cellular function of these two
andbacteria[9].Althoughitsnaturaldistributionsuggested
enzymesisimportantandisfulfilledbytwoatleastpartially
tomanythatreversegyraseactivityisessentialforlifeathigh
redundant enzymes. An important feature of the Group II
temperature,Atomietal.[8]successfullyisolatedadeletion
cases (Table1) is that most of them encode homologs of
mutant,demonstratingthatRgyisnotessentialforThermo-
eukaryoticNERproteins,whichwereexpectedtoparticipate
coccusviability.TheThermococcusmutantneverthelessgrew
ingeneralizedrepairofbulky,helix-distortingDNAlesions.
atareducedrateandwithadecreasedmaximaltemperature.
Althoughredundancyoffunctioncould,inprinciple,account
Thus, whereas Rgy appears not to be absolutely required
forthisresult,thecorrespondingeukaryoticgenestypically
forcellularfunctioninThermococcus,themutantphenotype
donotexhibitsuchredundancywithrespecttoDNArepair
arguedthatitmustplayaroleofsomeimportance.Thelatter
[13].
conclusion has been reinforced by failure of later efforts to
Group III of Table1 represents genes which could not
delete the corresponding genes of other hyperthermophilic
be deleted and thus may encode an essential function.
archaea[10].
Although the genetic tools for HA remain limited, in
In the ensuing decade, progress in the genetic analysis
some cases lethality could be further supported by specific
ofDNAmetabolicenzymesofHAhasbeensubstantialbut
genetic tests. Some of the proteins are predicted to play
hard-won. As in the Rgy study, the archaeal mutants are
important roles in basic replication, and thus their appear-
typically constructed by replacing the target gene with a
ance in this group is fully plausible; examples include the
selected marker and therefore represent unambiguous loss
structure-specific endonuclease Fen1, which is implicated
of gene function. However, the resulting genotype often
in maturation of Okazaki fragments. Most of the genes in
doesnotmatchthatpredictedbythebacterialoreukaryotic
this category are associated specifically with homologous
counterpart. On one hand, this reinforces the evidence
recombination,whichishowevernotanessentialprocessin
of deep divergence separating archaea from bacteria and
normalcellsofbacteria,unicellulareukaryotes,ormesophilic
eukaryotes and the incentive for their analysis, but on the
archaea.
other hand, it intensifies the need to formulate and test the
As noted in several of these studies, the number of
next generation of hypotheses. This review and commen-
putativeDNA-repairgenesthatfallintothelattertwocate-
tary selectively interprets functional genetic data relating
gories(GroupsIIandIII)addstotheaccumulatingevidence
to genome stability in archaea, with an emphasis on the
that HA do not follow the bacterial or eukaryotic models
hyperthermophiles.Itproposesthatcertainpatternsofresults
of DNA repair in any strict sense. The available evidence
do not fit well with the schemes of genome-maintenance
(discussedbelow)neverthelessarguesthattheydoreplicate
mechanisms identified in eukaryotic or bacterial models
their genomes accurately and cope effectively with DNA
andsuggestswaysinwhichatypicalorprimordialdamage-
damage.Thisposestheinterestingbutchallengingopportu-
coping strategies could, in principle, support fundamen-
nity of identifying precisely how the genome-maintenance
tally different strategies of genome maintenance in the
mechanismsofHAdifferfromthosedefinedhistoricallyin
HA.
bacteria and eukaryotic cells. This review and commentary
highlights four enigmata of genome-stability mechanisms
3.MutantPhenotypes in HA, some of which have persisted for two decades: (i)
the functional role of archaeal DNA polymerase stalling at
Table1 lists functional studies of HA genes that have been dU,(ii)theapparentlackofrepair-specificfunctionsforthe
implicated in DNA repair or related processes (usually by HA homologs of eukaryotic NER proteins, (iii) the specific
sequence similarity), grouped according to the phenotypic absenceofMutSLhomologsinHA,and(iv)theessentiality
outcome that was observed. Mutants placed in Group I ofHRproteinsinHA.Althoughtheinterpretationsprovided
seemtoberelativelystraightforwardtointerpret,sincetheir here are broad and speculative, attention to unresolved
phenotypes indicate a role of the deleted gene in DNA questions such as these represents a logical prerequisite to
repair or mutagenesis roughly in line with that predicted uncovering genetic processes that distinguish archaea, and
from its sequence. Examples include DNA photolyases, especiallyHA,fromotherformsoflife.
Archaea 3
Table1:Gene-deletionstudiesinhyperthermophilicarchaea.
Gene ORFID Organism Predictedfunction Phenotype Reference
GroupI(relevantphenotype)
udg2 SiRe 0084 S.islandicus UracilDNAglycosylase Impairedgrowth [10]
udg4 SiRe 1884 S.islandicus UracilDNAglycosylase Impairedgrowth [10]
hef/xpf TK1021 Tc.kodakarensis Structure-specificendonuclease Generaldamagesensitivity [34]
hjm/hel308 TK1332 Tc.kodakarensis DNAhelicase SensitivetomitomycinC [34]
phr Saci 1227 S.acidocaldarius DNAphotolyase Lossofphotoreactivation [59]
dbh Saci 0554 S.acidocaldarius Y-familyDNApolymerase IncreasedG:CtoT:A [11]
radC1 SiRe 0240 S.islandicus RadAparalog Increaseddamagesensitivity [60]
phrB SiRe 0261 S.islandicus DNAphotolyase Photorepairdeficiency [10]
󸀠
GroupI (possiblephenotype)
󸀠
xpf SiRe 1280 S.islandicus 3-flapendonuclease SlowDNAreplication? [10]
dpo4 S.solfataricus Y-familyDNApolymerase Possiblesensitivitytocisplatin [61]
xpb TK0928 Tc.kodakarensis NER-relatedhelicase Marginaldamagesensitivity [36]
xpd TK0784 Tc.kodakarensis NER-relatedhelicase Marginaldamagesensitivity [36]
GroupII(noobservedimpact)
xpd SiRe 1685 S.islandicus NER-relatedhelicase Noneobserved [10]
xpb1 SiRe 1128 S.islandicus NER-relatedhelicase Noneobserved [10]
xpb2 SiRe 1526 S.islandicus NER-relatedhelicase Noneobserved [10]
bax1 SiRe 1524 S.islandicus Nuclease Noneobserved [10]
uvde Saci 1096 S.acidocaldarius PutativeUVendonuclease NoeffectonUVsurvival [59]
hjc TK1175 Tc.kodakarensis Holliday-junctionresolvase Noneobserved [36]
hjc SiRe 1431 S.islandicus Holliday-junctionresolvase Noneobserved [10]
hje SiRe 930 S.islandicus Holliday-junctionresolvase Noneobserved [10]
󸀠
GroupII (syntheticlethality)
hje+hjc SiRe 930&1431 S.islandicus Holliday-junctionresolvases Doublemutantislethal [62]
GroupIII(presumedlethal)
radA TK1899 Tc.kodakarensis Recombinase Lethal(notrecovered) [36]
rad50 TK2211 Tc.kodakarensis DSBend-processing Lethal(notrecovered) [36]
mre11 TK2212 Tc.kodakarensis DSBend-processing Lethal(notrecovered) [36]
herA TK2213 Tc.kodakarensis DSBend-processing Lethal(notrecovered) [36]
nurA TK2210 Tc.kodakarensis DSBend-processing Lethal(notrecovered) [36]
󸀠
xpg/fen1 TK1281 Tc.kodakarensis 5-flapendonuclease Lethal(notrecovered) [36]
radA SiRe 1747 S.islandicus Recombinase Lethal [10]
herA SiRe 0064 S.islandicus Bipolarhelicase Confirmedlethal [62,63]
󸀠
xpg/fen1 SiRe 1830 S.islandicus 5-flapendonuclease Lethal [10]
nurA SiRe 0061 S.islandicus DSBend-processing Confirmedlethal [10,62]
rad50 SiRe 0062 S.islandicus DSBend-processing Confirmedlethal [10,62]
mre11 SiRe 0063 S.islandicus DSBend-processing Confirmedlethal [10,62]
hjm SiRe 0250 S.islandicus DSBend-processing Lethal [62,64]
4.TheEnigmaofDNAPolymerases polymerases,thesetechniquesillustratethepotentialbiolog-
Blocked byTemplatedU ical challenges posed by this uniquely archaeal property of
thesepolymerases.
Although early characterization of HA DNA polymerases One technique which promotes PCR by archaeal poly-
showedthattheyhavecertainadvantagesforPCRrelativeto merases involves supplementing the reaction mixture with
TaqDNApolymerase,itwasalsofoundthatHApolymerases athermostableuracilDNAglycosylase(UDG)[15].Because
bind tightly to any dU they encounter in the template UDGdestroysthedU-containingstrandasausabletemplate
strand [14], which impedes PCR. Driven by the practical in any subsequent cycle, its beneficial impact implies that
incentives to use HA DNA polymerases in PCR, various dU binding does not merely slow the HA polymerase but
biochemical techniques have been developed to overcome somehowremovesasignificantfractionofitfromcontribut-
this problem. In addition to enabling PCR by HA DNA ingtoDNAsynthesis.Thissuggeststhatthepolymerase:dU
4 Archaea
dU in
leading-strand Polymerase uncoupling
template
Fork regression U
U
(b) (subject to breakage?)
(a)
U (c)
(substrate for BER)
U U
dUin
Polymerase uncoupling
lagging-strand
template
󸀠
Figure1:PossiblefatesofareplicativepolymerasestalledattemplatedU.(a)Archaealpolymerasesstallwiththenascent3 end4ntahead
ofthetemplatedU[23].(b)Ifthereplicationforkrespondswithimmediateforkreversal,thedUisrestoredtoitsoriginalcontext,andBER
isrequiredforresumptionofforkprogress.(c)Iftightcouplingisnotpreserved,theMcmhelicaseandnonstalledpolymerase(notdrawn)
wouldcontinue,creatingssDNAregiononthestalledtemplatevulnerabletostructure-orsingle-strand-specificnucleases.Arrowheadson
󸀠
DNAstrandsrepresent3 ends.
complexisbothunproductiveandstableandraisesquestions archaeathathavebeenexamined,butnoneoftheeukaryotic
aboutwhatimpactsuchpropertieswouldhaveinvivo. or bacterial enzymes, stall at dU [20]. This natural distri-
Another technique involves supplementing the reaction bution thus seems to undermine the argument that HA
mixture with a thermostable dUTPase [16]. This enzyme need polymerase stalling specifically to cope with a high
removes dUTP from the cellular pool of dNTPs, thus pre- rateofspontaneouscytosinedeaminationinDNA,ashyper-
venting its incorporation into nascent DNA in vivo. The thermophilic bacteria should have a similar need, whereas
impactofthistechniquesuggeststhatmuchofthetemplate mesophilic archaea should not. It also seems striking that
dU that the polymerases of HA encounter during PCR D-family DNA polymerases, which replicate genomes and
results from cytosine deaminated before, not after, it was occuronlyinarchaea,arealsoinhibitedbydU,butthrough
incorporated into DNA. This is consistent with in vitro adifferentmechanism[21].Moreover,thereseemstobe(at
studiesdemonstratinglimiteddiscriminationagainstdUTP leastsuperficially)noevidencethatHAareanylessefficient
by DNA polymerases of HA [17] and raises the biologically in the conventional base-excision repair (BER) of dU than
relevantquestionastohowwellHApurgedUTPfromtheir thermophilic bacteria are. Corresponding N-glycosylases,
nucleotide pools during chromosome replication. A third supportedbyAPlyases,APendonucleases,andrelatedBER
approach to improved PCR by these polymerases has been enzymes,arewellrepresentedinHA[22].
to mutate the specific dU-binding pocket which mediates Other,moremechanistic,questionsconcernthefactthat
the stalling. This has yielded fully active and thermostable thestalledDNApolymeraseprecludesBER,bysequestering
archaealDNApolymeraseswhichareneverthelessinsensitive the dU within the binding pocket of the polymerase [23].
totheoccurrenceofdUinthetemplatestrand[18,19]. Repairthusrequiressomeformofreplication-forkreversalor
The only biological rationale offered consistently in the disassembly.Howthisprocessmightworkremainsdifficultto
literature for the specific stalling of the replicative poly- test.Theapparentlymostfavorableandcommonlyproposed
merasesofHAattemplatedUistoavoidG:CtoA:Ttransi- outcome is reversal of the fork, allowing BER of the dU in
tionmutations.Thelogicisadmittedlysimpleandplausible: its original context (Figure1). Recently, an archaeal DNA
(i) hydrolytic deamination of dC occurs spontaneously in primase has been found to bypass dU and other damaged
DNAandisgreatlyacceleratedbyhightemperatureand(ii) bases in vitro [24], and this may allow bypass with mini-
most DNA polymerases (including those of HA) insert dA mal reversal or rearrangement of the fork. Other plausible
oppositetheresultingdU,thuscreatingatransitionmutation. scenarios can be proposed which have not been discussed
Although avoiding such mutation is clearly a benefit, the in the literature; these involve intermediate structures that
magnitude of this benefit has not been weighed against the are vulnerable to single-strand or structure-specific nucle-
risks and costs which polymerase stalling imposes on the ases, leading potentially to breakage (or “collapse”) of the
archaealcellnoragainstthebenefitsofconventionalrepair, replication fork (Figure1). As argued below, fork reversal,
nor has it been evaluated critically in the wider context of disassembly, or even breakage may be a routine process
DNAmetabolism. in HA chromosome replication and may underlie uniquely
For example, why does the property of polymerase archaealstrategiestorepairDNAlesions.Eveninthiscontext,
stallingatdUcorrelatewithphylogeneticdomain,notwith however, the antimutagenesis argument fails to identify the
growthtemperature?AllthereplicativeDNApolymerasesof net advantage for archaea to invest deeply in a radical
Archaea 5
alternativetotherelativelysimple,low-riskstrategyofBER, and of individual UDGs through cloning in appropriate
whichthesearchaeaalsoemploy. expression vectors [29]. It also may be possible to express
More evidence of deeper complexity in the dU-stalling physiologicallevelsoftherecentlyengineereddU-insensitive
phenomenon comes from recent gene-deletion studies. DNApolymerases[18,19]inHAcellsusingsimilarvectors.
Genome annotations depict S. islandicus as having ensured
an adequate level of dU repair, as it encodes no fewer than
6.TheEnigmaofUnemployed Eukaryotic
fourputativeUDGproteins.However,effortstodeleteeach
of the genes individually were successful in only two cases, NERProteins
andthesuccessfulconstructsshowedimpairedgrowthunder
Historically, the molecular dissection of DNA repair path-
laboratoryconditions[10].Sinceneitherlethalitynorgrowth
ways, including that of nucleotide excision repair (NER),
impairment is a characteristic of UDG mutants, even of
depended heavily on the nonessential nature of these func-
similarlythermophilicbacteria[25],theseresultsarguethat
tions in microorganisms, which allowed completely repair-
theencodedSulfolobusproteinshaverolesbeyondthebasic,
defective mutants to be isolated and characterized. NER
redundantBERofdUpredictedbytheirannotationasfour
is one of the most generalized systems of DNA repair;
UDGs.
its substrates represent a broad spectrum of lesions that
are bulky, helix-distorting, or both. Despite having broadly
5.DoArchaeaNeeddUforaSpecificAspectof
similarstrategies,bacterialandeukaryoticNERaremediated
NormalChromosome Replication?
by proteins that share little homology. Furthermore, major
groups of eukaryotes diverge with respect to the damage
As noted above, deamination of dC is not the only source
recognitionproteinswhichinitiateNER;yeasthasonesetof
of dU; it can also be inserted opposite dA via dUTP
damage-recognitionproteins,plantcellshaveadifferentset,
incorporationduringpolymerization.ThedUformedbythis
andmammaliancellshaveseveraladditionalproteins[13,30].
routeisnotintrinsicallymutagenic,anditsabundancewould
Inadditiontobeingdiverse,thedamage-recognitionproteins
be determined by the discriminative accuracy of the DNA
are also the only NER-specific proteins of eukaryotes, as
polymeraseagainstdUTP[17]andthelevelofdUTPinthe
all “downstream” NER proteins play other roles in the cell.
cellularpools.Totheauthor’sknowledge,neitherdUTPlevels
The overall pattern thus suggests an evolutionary scenario
nor rates of dUTP incorporation have been measured in
in which NER arose rather recently and multiple times, by
archaealcells.
recruitment of suitable proteins to perform the damage-
RemovalofdUfromarchaealDNAisexpectedtooccur
recognitionstep.
atasteadyratedictatedlargelybyintracellularlevelsofUDG.
In contrast, dU incorporation occurs only at sites of DNA Superficially,genomeannotationsofmesophilicarchaea
synthesis;dUwouldthusbemostabundantimmediatelyafter suggestaredundancyofNERsystemsinwhichanincomplete
passageofthereplicationforkandwoulddecaysteadilyover set of eukaryotic NER genes, commonly annotated as “XP”
time thereafter. The contrasting kinetics of incorporation (xeroderma pigmentosum) homologs, is superimposed on a
versusremovalcould,inprinciple,makedUa“biochemical complete set of bacterial genes, represented by homologs
signatureofyouth,”markingstrandsthatremainuncorrected of E. coli uvrABC. None of the archaeal genomes encodes
by BER because they have been synthesized recently. It is a clear homolog of an eukaryotic NER damage-recognition
temptingtoquestionwhetherthisorsomeothermolecular protein,andtheavailableexperimentalevidencesuggeststhat
“time stamp” could control important features of genome only the Uvr proteins mediate NER in these archaea. For
replicationinarchaea.E.coliusesmethylationofthetetranu- example,thedamagedoligonucleotideexcisedbythemoder-
cleotide GATC in an analogous way to distinguish old and ately thermophilic Methanobacterium thermoautotrophicum
new strands after the replication fork has passed, and this (i.e., Methanothermobacter thermoautotrophicus) strain ΔH
signalisnecessaryfortheaccuracyofpostreplicativeMMR (which has a complete set of uvr genes) has the length
[26–28]. In HA, dU could, in principle, play an analogous characteristicofbacterialNER,noteukaryoticNER[31].Sim-
role in a uniquely archaeal error-correction system, or it ilarly,inthemesophilichalophileHalobacteriumsalinarum,
couldservetolimittheextentortimingofDNAsynthesisin inactivation of any uvr homolog is sufficient to render the
recentlyreplicatedregions.Alternatively,optimalreplication cells UV-sensitive [32]. Thus, the redundancy of NER in
in HA may simply require periodic fork reversal in the mesophilic archaea suggested by genome analysis is more
absenceofreplication-blockinglesions,forwhichdUinthe apparentthanreal,asthesearchaeaseemtorelyfunctionally
templatestrandmayprovideaspecifictrigger. onthesimplerNERsystemtheysharewithbacteria[33].
Whatever cellular processes dU abundance may con- ThisconclusionraisesquestionsfortheHA,whichuni-
trol, the “biochemical marker” hypothesis predicts that a formlylacktheUvrABChomologsfoundinthemesophilic
particular balance of the relevant biochemical activities archaea. Three possibilities can be considered: (i) in HA,
(dUTPase,UDG,andpolymerasestalling)maybeimportant lack of the UvrABC system is compensated by recruitment
fornormalcellularfunction.Thus,artificialmanipulationof of unidentified, HA-specific, damage-recognition proteins
these activities may perturb normal replication or growth, which, in combination with the archaeal XP homologs,
which may explain the unexpected results of UDG gene reconstructaneukaryotic-styleNER,(ii)HAusefundamen-
deletion in S. islandicus [10]. It may be feasible to test this tally different “alternative excision” pathways [1] to repair
moredirectlybymanipulatingintracellularlevelsofdUTPase duplexDNAbetweenroundsofreplication,or(iii)HAsimply
6 Archaea
Lesion in
leading-strand Nicked or gapped chromosome
template plus ds end with lesion
X F∨en1/Xpg X
∧
X Hef, Xpf X
Lesion in
Nicked or gapped chromosome lagging-strand
plus ds end with lesion template
(a) (b)
Figure2:Forkbreakage“collapse”inresponsetoreplication-blockinglesions.Adductsorotherlarge,helix-distortinglesions(symbolized
byX)blockapolymerasemolecule,inducingacertaindegreeofuncouplingandssDNAexposureatthebaseofthefork.Cleavagebyan
appropriatesingle-strand-specificorstructure-specificendonucleasecan,foreachcase,generatedsDNAendthatretainsthelesion.Thetop
󸀠 󸀠
strandineachDNAduplexisorientedwiththe5 endonleft;free3 endsarerepresentedbyarrowheads,whereasbrokenlinesindicatethe
restofthechromosome.Carets(∧)locatethesinglecutthatwouldbemadebytheindicatednuclease.Forthecleavagepreferencesofarchaeal
Hef,Xpf,andFen1/Xpg,see[36,65,66].
forgopreemptiverepairanddealwithbulky,helix-distorting the resulting end. Conversely, a lesion encountered on the
DNAlesionsaftertheystallthereplicationfork. lagging-strand template would yield a similar result if the
󸀠
Theaccumulatingexperimentalevidencearguesincreas- Y-structure were cleaved by a 5 -flap endonuclease, such
inglyagainst(i),asdeletionsofmostarchaealXPhomologs as Fen1/Xpg (Figure2). The logic of the proposed strand
have negligible effects on the survival of UV and other specificity is that in both cases it positions the lesion for
relevant DNA damage (Table1). While it could be argued removalbydsend-processing,whichiscentraltoreplication
that the genes which showed no apparent role in repair forkreassemblyviaHR(discussedinmoredetailbelow).In
duplicate the function of another gene, such redundancy contrast,modelsofreplication-fork“collapse”andreassem-
is not seen typically in bacteria and eukaryotes. It should bly in bacteria have historically proposed steps that leave
also be noted that HA can remove such lesions and do thelesioninacontinuous(albeitpotentiallygap-containing)
not merely bypass them, as UV photoproducts are steadily DNA; this feature may reflect the rationale that providing
lostfromthegenomicDNAofUV-irradiatedculturesupon additionalchancesforconventionalNERrepresentsthebest
resumptionofincubation[34,35].Inaddition,theremovalof strategyforsuccessfulreplication[36,38–40].
UVphotoproductsisnotaffectedbythetranscriptionofthe Although inducing fork disassembly and breakage in
damagedDNA,whicharguesfurtherthatHAdonotemploy order to expose and remove DNA lesions seems drastic
thearchaealRNApolymeraseextensivelyasaDNAdamage (as argued above for dU), the lesions in question are, by
detector[34,35]. definition, intrinsically incompatible with replicative poly-
merasesand(unlikedU)mayhavenoalternativepathwaysof
removalinHA.Astrategylikethismayrepresentoneofthe
7.DoHAUseDNAReplication
simplestandearliestresponsestoDNAdamage,anditsbasic
ItselftoTargetBulky,Helix-Distorting
feasibility is demonstrated by the relatively robust growth
LesionsforRemoval?
of bacterial and eukaryotic NER mutants under laboratory
cultivation. In these cells, the low level of bulky lesions
Otherexperimentalevidencecanbeinterpretedassupport-
that form endogenously (e.g., protein:DNA cross-links) is
ing (iii) above. Among various HA deletion mutants, for
usually handled at the replication fork by HR-dependent
example, oneofthefew that fits thephenotype expected of
processes,suchasforkreassemblyinitiatedbyforkbreakage
ageneralizedexcisionrepairdeficiencyistheΔhef mutantof
(discussed below), template-switching, or lesion-skipping
Thermococcuskodakarensis[36].TheencodedHefproteinis
processes[41].
relatedtoeukaryoticFANCMproteinsand,likethem,binds
tobranchedDNAsthatmimicstalledreplicationforks,where
itcleavestheduplexnearthejoint[36].Thisspecificityseems 8.TheEnigmaofAccurateReplication
tailored to the expected structures of archaeal replication withoutMutSLHomologs
forksstalledbyunrepairedbulkyadducts,particularlyifthe
replicative helicase, archaeal Mcm, mimics the eukaryotic Researchers analyzing the first Pyrobaculum genome
McminitsabilitytoprogresspastmostDNAlesions,leaving sequencesnotedthattheseHAgenomesdidnotencodeany
them to stall the polymerase [37]. If the leading-strand homologs of the E. coli MutS or MutL proteins [42]. These
polymeraseencountersthelesion,itsblockagewouldcreate two proteins cooperate to remove replication errors from
a partially single-stranded region near the junction of the dsDNA soon after the bacterial replication fork has passed
fork (Figure2). Cleavage at the base of the fork by Hef or (sometimes termed the “spellchecker” function); they also
󸀠
a 3 -flap endonuclease such as archaeal Xpf, or cleavage of trigger the abortion of HR between sequences that differ
the large single-stranded region by ssDNA endonuclease, at multiple sites (termed the “antirecombination” function)
would break the fork in a way that leaves the lesion on [26, 27]. Among archaea, the absence of MutSL homologs
Archaea 7
applies specifically to hyperthermophiles and thus parallels base pair substitutions (BPSs) as genetic markers [48]. The
the absence of Uvr proteins. Also similar to Uvr, absence complex pattern of markers recovered after transformation
of MutSL homologs in HA does not correlate with any withssDNA(Figure3(a))indicatedthatashort-patchmode
obvious defect in maintaining genome integrity, although of gene conversion (nonreciprocal HR), representing local-
it must be noted that few HA have been analyzed in this ized strand removal and resynthesis, had created multiple
respect. Sulfolobus spp. have been found to replicate their discontinuous tracts during recombination. Even greater
chromosomesaccurately,althoughinsomespeciesthismay complexity was observed in cells transformed with dsDNA
be masked by high levels of transposable element activity that was itself a preformed heteroduplex in which both
[43–45].TheerrorratepergenomereplicationislowerinS. strandsweredistinguishedfromeachotheraswellasfromthe
acidocaldarius than in nearly all other microbial genomes recipientchromosome(Figure3(b)).Thepatternsfromthese
similarly analyzed; interestingly, however, single-nucleotide trialleliccrosses suggested that thepreformedheteroduplex
frameshifts in short mononucleotide runs are prominent wasprocessedviasimilarlocalizedstrandremovalandresyn-
in the spectrum, which is a qualitative characteristic of thesisbeforesynapsiswiththerecipientchromosome[48].
MMR-deficientcells[44].
These results demonstrate that Sulfolobus spp. imitate 10.CouldForkRegressionTargetReplication
theresultsofconventionalMMRquantitatively,butperhaps ErrorsforRemoval?
notqualitatively.TheabsenceofanyMutSLpairarguesthat
Sulfolobus spp. and other HA either (i) lack any form of Transientreversalofthereplicationforktoformafour-arm
postreplicationalmismatchrepairor(ii)employsomealter- structure,commonlytermeda“chickenfoot,”isthoughttobe
native to the conventional(i.e., MutSL-dependent) strategy. relativelyfrequentinDNAreplicationandisproposedtohave
In order to explain the low error rate, hypothesis (i) would functionalroles,suchasprovidingerror-freealternativesto
require HA to have a much higher accuracy of polymer- TLS when the leading-strand template has an unrepaired
ization in vivo than that exhibited by known replicative lesion [49, 50]. One property of regressed forks that has
DNApolymerasesinvitro.Similarly,(ii)wouldrequiresome not received much attention is the distinctive nature of the
systemthatrecognizesmismatchesaftertheduplexleavesthe fourth(i.e.,extruded)arm,which(i)representstheonlyDNA
polymerase and removes the newly synthesized strand but endofthestructure(albeitnotnecessarilyabluntone)and
doesnotinvolveaMutSorMutLprotein. (ii) includes only the newly synthesized daughter strands
ExplainingtheaccuratechromosomereplicationinSul- (Figure4). Thus, a “chicken foot” contains, in the config-
folobusspp.mustalsoaccountfortheirmultipletranslesion uration of its strands, the information needed to find and
synthesis (TLS) DNA polymerases, which are nonproces- remove a recent replication error (Figure4). Although this
sive but error-prone. The Sulfolobus Y-family polymerases hypothetical “spellchecker” would remove a correct strand
Dpo4 (S. solfataricus) and Dbh (S. acidocaldarius) have alongwiththereplicationerror,thispricemaybeconsidered
been cocrystallized with many DNA substrates, providing small in light of the potential simplicity of the mechanism
unprecedentedresolutionofbiochemicalandstructuralanal- (whichdoesnotneedtotrackstranddiscontinuities)andthe
ysisofTLSinvitro[46].Bothpolymerasesarehighlyerror- benefits of the resulting replication accuracy. However, this
prone when replicating intact template in vitro, with Dbh hypothetical mode of strand discrimination would impose
making 1-nt deletions at frequencies over 50% opposite some unusual and perhaps unrealistic requirements. The
󸀠
common sequence motifs, such as 5-GYY- [47]. However, replication fork should presumably reverse only when a
the genetic consequences of inactivating Dbh indicate that replicationerrorhaslefteitheroftheDNApolymerases,for
thispolymeraseplaysatleastoneantimutagenicroleinvivo, example;also,destructionofthefourtharmwouldneedtobe
namely,thesuppressionofspontaneousG:CtoT:Atransver- preventedinforksreversedinthecourseofotherprocesses,
sions [11]. The hypothesis that this suppression reflects suchaserror-freelesionbypass[50].
accuratebypassofoxidizedguanine(oxoG)issupportedby
analysis of individual TLS events past this damaged base 11.TheEnigmaofEssentialRecombination
(Sakofsky&Grogan,unpublished).However,themostcom-
monspontaneousmutationsobservedintheS.acidocaldarius Recombination proteins play diverse roles across biology.
chromosomearenotaffectedbyinactivatingDbh(Sakofsky Historically, analysis of homologous recombination (HR)
& Grogan, unpublished). Given the extreme inaccuracy of emphasizedgeneticdiversification,thatis,thereassortment
thisDNApolymeraseinvitro,theseresultssuggestthatDbh of alleles in eukaryotic meiosis and the acquisition of new
isnormallyexcludedfromtheSulfolobusreplicationforkby traits in bacteria following chromosomal DNA transfer.
anunidentified,yetapparentlyeffective,mechanism. Increasingly,however,HRisconsideredtosupportsuccessful
genome replication by providing repair of double-strand
9.MismatchProcessinginSulfolobus breaks(DSBs).Inparticular,thecentralprocess,invasionof
󸀠
a DNA duplex by a single-stranded 3 end with a matching
ProcessingofmismatchedDNAhasbeendetectedgenetically sequence,providesaroutetoreassemblingbrokenreplication
in S. acidocaldarius, but in the context of genetic recombi- forks[37,38].
nation,notthatofreplicationaccuracy.Theevidencecomes HR in bacteria can be eliminated nearly quantitatively
from transforming auxotrophic mutants with linear DNAs by inactivating the homolog of the E. coli recA gene or in
that contain several closely spaced, phenotypically silent eukaryotes by inactivating the Rad52 protein, which loads
8 Archaea
Preformed
ssDNA heteroduplex
+ Recipient chromosome Recipient chromosome +
Heteroduplex Localized strand removal Partial resynthesis
formation
+
Localized
strand removal
Heteroduplex
formation
Resynthesis
Localized strand removal Partial resynthesis
Transformant chromosome Transformant chromosome
(a) (b)
Figure3:PatchyrecombinationinSulfolobus.CellstransformedwithlinearDNAsmarkedatmultiplesites(ormatedwithmultiplymarked
donor cells) produce recombinants that indicate erratic, localized strand loss from heteroduplex intermediates [48, 67]. (a) Markers are
acquiredfromssDNAasmultipleshorttracts,someofwhichconsistofasinglemarker.(b)Apreformedheteroduplex,containingtwo
distinctdonorallelesateachmarkedposition,generatessimilarshortpatchesrepresentingallthreepossiblealleles.Semicircularsymbols
depict “silent” genetic markers (synonymous mutations), and the colors (black, white, and gray) depict different alleles. Different colors
oppositetoeachotherindicateamismatchthatiseventuallyresolvedbystrandremovalandresynthesis.
disabled.Thissupportstheargumentthattheimmediate,and
(fork (ds end perhapstheoriginal,functionofHRistoenablereplication
regression?) degraded?) forksblocked by unrepairedlesionstodisassemble, reform,
Replication andresumeDNAreplication[39,52].
error
The remaining proteins required for HR prepare the
DNAend(s)forstrandinvasionorresolvethestructuresthat
result.Inbacteria,homologsoftheE.coliRecBCDproteins
Figure 4: Repositioning of replication errors by fork reversal. If
prepare the ds end for invasion of the intact duplex. The
areplicationforkreversesshortlyaftermakingareplicationerror,
theresultingmismatchwouldbelocalizedtotheshort(extruded) RecBCD complex is a powerful helicase/exonuclease that
arm.Inprinciple,thiscouldprovideabasisforremovingreplication initially unwinds and degrades both strands of dsDNA end
󸀠
errorswithoutinvolvingaconventional(MutSL-dependent)MMR but eventually converts to a form that leaves 3 extensions,
system.Thewhitesemicirclerepresentsapolymeraseerror;heavy which are needed for strand invasion [37, 38]. Another
linesdepictparental(template)strands. complex in bacteria, consisting of the SbcCD proteins, can
alsoperformthisfunction,andeukaryoticcellshavecorre-
sponding“MRX”complexes,composedofrelatedeukaryotic
proteinsMre11andRad50(thethirdprotein,X,differsamong
the eukaryotic equivalent of RecA (Rad51) onto ssDNA. different eukaryotes). Archaea have their own version of
Archaea have their own recombinases, designated “RadA,” MRX, composed of archaeal Mre11 and Rad50 homologs,
related to both the bacterial and eukaryotic proteins [51]. and a uniquely archaeal complex of bipolar helicase and
All these proteins mediate a central step in HR, in which endo/exonuclease,HerA-NurA[53].
ssDNA finds its complement in dsDNA and pairs with it. Among archaea, genetic dissection of HR functions has
Inadditiontoabolishingtheprocessesofcrossing-overand madethemostprogressintheextremehalophiles.TheradA,
gene conversion between two homologous duplex DNAs, rad50,andmre11geneshavebeendeletedindividuallyandin
recAorrad52mutationsmakecellsdramaticallysensitiveto combinationsfromHalobacteriumsalinarumandHaloferax
DNAdamage,andthissensitivityisexacerbatedifNERisalso volcanii,forexample,[54–56].Mostofthemutantsshowed
Archaea 9
Fork breakage products
X
X }
Resection of both strands,
X formation of a3󳰀extension
+ +
Gaps filled
}
Strand invasion, D-loop extension, Holliday-junction
D-loop formation opposite-strand priming resolution, gap filling
(a) (b)
Figure5:RegenerationofbrokenreplicationforksbyHRfunctions.(a)Theproductsofarchaealreplicationforkbreakageproposedin
Figure2areexpectedtoretainvariouslesions,gaps,andoverhangs;possiblestructuresareillustratedtotheleftofeachbracket.Thesevarious
󸀠
structuresmustbeconvertedtoa“clean”3 extensiononthedsendandanintactcontinuousduplexwithwhichitcanrecombine.(b)Proteins
ofdouble-strand-breakrepair(homologousrecombination)promotesubsequentreassemblyofthereplicationfork.Thestepsdepictedhere
arethosecommonlyproposedforbacteriaandeukaryotes[37,38,52].Asinpreviousfigures,brokenlinesindicatetheremainderofthe
󸀠
chromosome,andarrowheadsmark3 ends.
a modest increase in radiation sensitivity, but in Haloferax, the essentiality of basic HR functions for viability (Table1)
radA could not be deleted from a mutant lacking all four combine to suggest the possibility that breakage and HR-
of the normal replication origins [57]. In contrast to the mediated reassembly of replication forks blocked at DNA
halophiles,HAhavenotpermitteddeletionoftheradAgene, lesions may serve as the major pathway for removing such
despite multiple attempts by multiple groups. Furthermore, lesions.
four other proteins of HA share this property of apparent Mostschemesoffork“collapse”(breakage)andreassem-
essentiality: the archaeal homologs of Mre11 and Rad50, as blythathavebeenproposedforbacteriaoreukaryoteseither
wellastheHerAandNurAproteins.Alloftheseproteinsare restore the context of the original lesion, so that it can be
repaired by conventional NER, or leave it behind in an
implicatedinthecentralstepsofHR,thatis,theprocessing
associateddaughter-strandgap[39,52].InthecontextofHA,
of dsDNA and for RadA loading and subsequent strand
itwouldseemtobenefitthecellmoreiftheendonucleolytic
invasion,andtotheauthor’sknowledgetheyhavenotbeen
cleavage that breaks the replication fork would leave the
identified with any other, nonrecombination, function. The
lesion near the ds end (Figure4). The proposed advantage
mre11,rad50,herA,andnurAgenesaretypicallyclusteredin
ofthisfeatureisthattheunwindingandbidirectionalnucle-
HA genomes, and theproteinshave been observed to form
foci in 𝛾-irradiated cells [58]. The apparent essentiality of olytic activities typical of DSB end-processing complexes
seem well suited to remove a broad spectrum of DNA
these proteins therefore seems to coincide with the other
lesions,includingclusteredandopposedones,withminimal
geneticfeaturesdiscussedabovethatdistinguishHAfromall
requirementsfordamagerecognition(Figure5).Similarly,by
othercellularorganisms,includingmesophilicarchaea.
analogytotheRecBCDcomplex,onceabidirectionalstrand
󸀠
removalhasdiscardedthelesion,limitingorinhibitingthe3-
12.DoHANeedHRtoSurviveFork-Centered specificactivitywould,inprinciple,besufficienttogenerate
󸀠
DNARepair? the3 tailrequiredtoinitiateforkreconstruction.
Regardlessofitsmoleculardetails,thebasicprinciplethat
In principle, the essentiality of RadA, Mre11, Rad50, HerA, allreplication-blockingDNAlesionsmaybefunneleddown
and NurA specifically in HA could reflect biochemical a pathway culminating in fork breakage can be expected to
functions unique to the HA versions of these proteins. makesuccessfulreplicationofthechromosomedependenton
Although this possibility remains intriguing, it seems more HRtoadegreenotrepresentedinanyothermajorgroupof
parsimonioustoaskwhetherthebroadlyconservedfunction organisms.AlthoughNERmutantsofbacteriaoreukaryotes
oftheseproteins(specifically,HRbetweenadsDNAendand managetoreproduceunderlaboratoryconditionsonabasis
acorrespondingduplex)isnotsimplyingreaterdemandor similartothatproposedhereforwild-typeHA,theenvisaged
playsamorecriticalroleinHAthaninotherorganisms.If archaealstrategyseemsunlikelytosucceedinnaturewithout
HAindeedlackadedicated,preventive-maintenancesystem molecularadaptationsthatmaketherequisitefork-breakage
like NER to repair bulky, helix-distorting lesions before and reassembly processes unusually efficient and reliable.
replication,theyshouldexperiencefrequentlesion-induced Genes or biochemical functions found to be unexpectedly
fork stalling. Bacterial and eukaryotic cells experience such important for normal chromosomal replication in HA may
arrest, and HR proteins are required for many of their includethosethatembodysuchadaptations.
responsestothischallenge[39,52].InHA,therefore(i)the Finally,whetherornotanyoftheprocessesproposedhere
ability to remove UV lesions from genomic DNA in the ultimatelyprovetooperateinHA,theprospectremainsthat
absence of conventional NER [34], (ii) the importance of these organisms may integrate repair, recombination, and
Hef/Xpf activity for full DNA repair capacity [36], and (iii) replication of DNA much more tightly than any other type
10 Archaea
of cell that has been examined previously. Such integration Sulfolobusgenomeathightemperature,”DNARepair,vol.11,no.
is expected to make functional studies challenging and 4,pp.391–400,2012.
anticipatesaneedforgenetictechniquessuchasconstruction [12] Q.Huang,Y.Li,C.Zengetal.,“GeneticanalysisoftheHoll-
of single-gene knockout libraries and conditional mutants, idayjunctionresolvasesHjeandHjcinSulfolobusislandicus,”
epistasis testing, and analysis of mutants with biophysical Extremophiles,vol.19,no.2,pp.505–514,2015.
methods and the full spectrum of molecular biology tech- [13] O.D.Scharer,“Nucleotideexcisionrepairineukaryotes,”Cold
niques. Although activities and molecular interactions of SpringHarborPerspectivesinBiology,vol.5,no.10,ArticleID
individualenzymeswillcontinuetobeidentifiedinHAthat a012609,2013.
haveaplausibleconnectionwithDNArepairandreplication [14] R.S.Lasken,D.M.Schuster,andA.Rashtchian,“Archaebacte-
fidelity,themostmeaningfultestofprogressinthisareawill rialDNApolymerasestightlybinduracil-containingDNA,”The
be whether the proposed contribution to genome mainte- JournalofBiologicalChemistry,vol.271,no.30,pp.17692–17696,
nancecanbeconfirmedinlivingarchaealcells. 1996.
[15] X.-P. Liu and J.-H. Liu, “Characterization of family IV UDG
fromAeropyrumpernixanditsapplicationinhot-startPCRby
ConflictofInterests
familyBDNApolymerase,”PLoSONE,vol.6,no.11,ArticleID
e27248,2011.
Theauthordeclaresthatnoconflictofinterestsexistsregard-
[16] H. H. Hogrefe, C. J. Hansen, B. R. Scott, and K. B. Nielson,
ingthepublicationofthispaper.
“ArchaealdUTPaseenhancesPCRamplificationswitharchaeal
DNA polymerases by preventing dUTP incorporation,” Pro-
Acknowledgment ceedingsoftheNationalAcademyofSciencesoftheUnitedStates
ofAmerica,vol.99,no.2,pp.596–601,2002.
TheauthorthanksC.Sakofskyforhelpfuldiscussionregard- [17] P.Gru´z,M.Shimizu,F.M.Pisani,M.deFelice,Y.Kanke,and
ingthepaper. T.Nohmi,“ProcessingofDNAlesionsbyarchaealDNApoly-
merases from Sulfolobus solfataricus,” Nucleic Acids Research,
References vol.31,no.14,pp.4024–4030,2003.
[18] E.Gaidamaviciute,D.Tauraite,J.Gagilas,andA.Lagunavicius,
[1] E. C. Friedberg, G. C. Walker, W. Siede, R. D. Wood, R. A. “Site-directedchemicalmodificationofarchaealThermococcus
Schultz,andT.Ellenberger,DNARepairandMutagenesis,ASM litoralisSh1BDNApolymerase:acquiredabilitytoreadthrough
Press(AmericanSocietyforMicrobiology),Washington,DC, template-strand uracils,” Biochimica et Biophysica Acta, vol.
USA,2006. 1804,no.6,pp.1385–1393,2010.
[2] C. R. Woese, O. Kandler, and M. L. Wheelis, “Towards a [19] S.K.Jozwiakowski,B.J.Keith,L.Gilroy,A.J.Doherty,andB.A.
naturalsystemoforganisms:proposalforthedomainsArchaea, Connolly,“Anarchaealfamily-BDNApolymerasevariantable
Bacteria,andEucarya,”ProceedingsoftheNationalAcademyof to replicate past DNA damage: occurrence of replicative and
SciencesoftheUnitedStatesofAmerica,vol.87,no.12,pp.4576– translesionsynthesispolymeraseswithintheBfamily,”Nucleic
4579,1990. AcidsResearch,vol.42,no.15,pp.9949–9963,2014.
[3] E. R. Barry and S. D. Bell, “DNA replication in the archaea,” [20] J. Wardle, P. M. J. Burgers, I. K. O. Cann et al., “Uracil
MicrobiologyandMolecularBiologyReviews,vol.70,no.4,pp. recognition byreplicative DNA polymerases islimitedtothe
876–887,2006. archaea,notoccurringwithbacteriaandeukarya,”NucleicAcids
[4] S.Ishino,L.M.Kelman,Z.Kelman,andY.Ishino,“Thearchaeal Research,vol.36,no.3,pp.705–711,2008.
DNA replication machinery: past, present and future,” Genes [21] T. T. Richardson, L. Gilroy, Y. Ishino, B. A. Connolly, and
andGeneticSystems,vol.88,no.6,pp.315–319,2014. G. Henneke, “Novel inhibition of archaeal family-D DNA
[5] L. M. Kelman and Z. Kelman, “Archaeal DNA replication,” polymerasebyuracil,”NucleicAcidsResearch,vol.41,no.7,pp.
AnnualReviewofGenetics,vol.48,no.1,pp.71–97,2014. 4207–4218,2013.
[6] M. O’Donnell, L. Langston, and B. Stillman, “Principles and [22] S.GrassoandG.Tell,“BaseexcisionrepairinArchaea:backto
conceptsofDNAreplicationinbacteria,archaea,andeukarya,” thefutureinDNArepair,”DNARepair,vol.21,pp.148–157,2014.
ColdSpringHarborPerspectivesinBiology,vol.5,no.7,2013. [23] M. J. Fogg, L. H. Pearl, and B. A. Connolly, “Structural basis
[7] F.Sarmiento,F.Long,I.Cann,andW.B.Whitman,“Diversity foruracilrecognitionbyarchaealfamilyBDNApolymerases,”
oftheDNAreplicationsysteminthearchaeadomain,”Archaea, NatureStructuralBiology,vol.9,no.12,pp.922–927,2002.
vol.2014,ArticleID675946,15pages,2014. [24] S. K. Jozwiakowski, F. Borazjani Gholami, and A. J. Doherty,
[8] H.Atomi,R.Matsumi,andT.Imanaka,“Reversegyraseisnota “Archaeal replicative primases can perform translesion DNA
prerequisiteforhyperthermophiliclife,”JournalofBacteriology, synthesis,”ProceedingsoftheNationalAcademyofSciencesofthe
vol.186,no.14,pp.4829–4833,2004. UnitedStatesofAmerica,vol.112,no.7,pp.E633–E638,2015.
[9] P. Forterre, “A hot story from comparative genomics: reverse [25] T. Sakai, S.-I. Tokishita, K. Mochizuki, A. Motomiya, H.
gyraseistheonlyhyperthermophile-specificprotein,”Trendsin Yamagata,andT.Ohta,“Mutagenesisofuracil-DNAglycosylase
Genetics,vol.18,no.5,pp.236–237,2002. deficient mutants of the extremely thermophilic eubacterium
[10] C. Zhang, B. Tian, S. Li et al., “Genetic manipulation in Sul- Thermusthermophilus,”DNARepair,vol.7,no.4,pp.663–669,
folobusislandicusandfunctionalanalysisofDNArepairgenes 2008.
,”BiochemicalSocietyTransactions,vol.41,no.1,pp.405–410, [26] S. Acharya, P. L. Foster, P. Brooks, and R. Fishel, “The coor-
2013. dinated functions of the E. coli MutS and MutL proteins in
[11] C. J. Sakofsky, P. L. Foster, and D. W. Grogan, “Roles of the mismatch repair,” Molecular Cell, vol. 12, no. 1, pp. 233–246,
Y-familyDNApolymeraseDbhinaccuratereplicationofthe 2003.