Bacterial type 1A topoisomerases maintain the stability of the genome by preventing and
dealing with R-loop-and nucleotide excision repair-dependent topological stress.
Julien Brochu, Emilie Vlachos-Breton and Marc Drolet*.
Département de microbiologie, infectiologie et immunologie, Université de Montréal, Montréal,
P. Québec, Canada, H3C 3J7.
*To whom correspondence should be addressed. Tel: 1-514-343-5796. Email:
marc.drolet@umontreal.ca
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 11, 2021. ; https://doi.org/10.1101/2021.07.10.451908doi: bioRxiv preprint
ABSTRACT
E. coli type 1A topoisomerases (topos), topo I (topA) and topo III (topB) have both relaxation and
decatenation activities. B. subtilis and E. coli topA topB null cells can survive owing to DNA
amplifications allowing overproduction of topo IV, the main cellular decatenase that can also relax
supercoiling. We show that overproducing human topo IB, a relaxase but not a decatenase, can
substitute for topo IV in allowing E. coli topA null but not topA topB null cells to survive. Deleting
topB exacerbates phenotypes of topA null mutants including hypernegative supercoiling, R-loop
formation, and RNase HI-sensitive replication, phenotypes that are not corrected by topo IV
overproduction. These phenotypes lead to Ter DNA amplification causing a chromosome
segregation defect that is corrected by topo IV overproduction. Furthermore, topA topB null
mutants not overproducing topo IV acquire uvrB or uvrC mutations, revealing a nucleotide
excision repair (NER)-dependent problem with replication fork progression. Thus, type IA topos
maintain the stability of the genome by providing essential relaxase and decatenase activities to
prevent and solve topological stress related to R-loops and NER. Moreover, excess R-loop
formation is well tolerated in cells that have enough topoisomerase activity to support the
subsequent replication-related topological stress.
INTRODUCTION
Because of the double helical structure of DNA, each time the two strands are separated during
replication, transcription, or repair, underwinding and overwinding (supercoiling) occur. In turn,
such supercoiling interferes with normal gene expression and replication. Furthermore, tangling
of the DNA occurs during replication and repair. If not properly resolved, such entanglements
inhibit chromosome segregation and may lead to DNA breaks and genomic instability. To solve
these topological problems, cells possess DNA topoisomerases (topos), which are nicking-closing
enzymes (1,2). These enzymes cut one (type I) or two (type II) DNA strands. Type I and II are
further divided respectively in class IA and IB and IIA and IIB. To solve topological problems,
type IA and type II enzymes use a strand-passage mechanism whereas enzymes of the class IB
family, also named swivelases, use a rotation mechanism.
Class A enzymes are the only ubiquitous topos (3,4). Furthermore, unlike other topos, they use
ssDNA as substrates and many of them possess RNA topo activity (5,6). Type IA topos are
classified into three subfamilies, the two main ones being topo I and topo III with E. coli topA and
topB respectively encoding the prototype enzyme of these two subfamilies. The third one, reverse
gyrase, is only found in hyperthermophilic and some thermophilic organisms. Topo I is present in
all bacteria but not in archaea and eukaryotes, whereas topo III is found in most bacteria and in all
archaea and eukaryotes. Topo III has a higher requirement for ssDNA than topo I and mostly acts
as a decatenase (7), whereas topo I mostly acts to relax transcription-induced negative supercoiling
(8). Notwithstanding these preferences, it is quite clear from in vitro experiments that topo I and
III can perform both relaxation and decatenation reactions (9-11). The results of recent single-
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 11, 2021. ; https://doi.org/10.1101/2021.07.10.451908doi: bioRxiv preprint
molecule experiments suggest that a dynamic fast gate for topo I may promote efficient relaxation
of negatively supercoiled DNA, whereas a slower gate-closing for topo III may facilitate capture
of dsDNA and, as a result, efficient decatenation (12). Additional factors that must be considered
for their in vivo activity are, among others, their relative abundance which is reported to be about
200 and 20 ppm respectively for topo I and topo III (https://pax-db.org/species/511145), and their
interacting partner, e.g. RNAP for topo I (13,14) and the replication fork for topo III (likely via
SSB and DnaX; ref. (15,16)).
Null mutations in the topA gene of E. coli inhibit cell growth. Few topA null cells manage to
generate visible colonies owing to the acquisition of compensatory mutations (reviewed in (4)).
Such mutations are either substitutions in gyrA or gyrB genes that reduce the negative supercoiling
activity of gyrase (17,18), or amplification of a chromosomal region allowing topo IV
overproduction (4,19-22). The occurrence of gyrase mutations is easy to interpret since the in vivo
function of gyrase is the introduction of negative supercoiling (via the relaxation of positive
supercoiling generated during replication and transcription). However, the main function of topo
IV is to act as a decatenase behind the replication fork to remove pre-catenanes and, once
chromosomal replication is completed, to remove catenanes to allow chromosome segregation
(23,24). Topo IV can also relax negative supercoiling, albeit much less efficiently than positive
supercoiling relaxation and decatenation (25-27). Because of the established main function of topo
I, it is believed that topo IV overproduction complement the growth defect of topA null mutants
by relaxing negative supercoiling. In fact, topo IV plays a role in the regulation of global negative
supercoiling (28).
Topo I relaxes negative supercoiling generated behind moving RNAPs during transcription (8).
The failure to relax transcription-induced negative supercoiling in topA null mutants leads to
hypernegative supercoiling and the formation of R-loops that can inhibit gene expression and
activate DnaA-independent unregulated replication (29-35). In fact, overproduction of RNase HI,
an enzyme degrading the RNA portion of a DNA:RNA hybrid in the R-loop, can compensate for
the growth defect of topA null mutants (36). Moreover, the supercoiling activity of gyrase acting
in front of replication and transcription complexes promotes R-loop formation both in vitro and in
vivo (36,37). Thus, one major function of topo I in E. coli is the relaxation of transcription-induced
negative supercoiling to inhibit R-loop formation.
Unlike topA null mutants, topB null cells grow as well as wild-type cells, but show a delayed and
disorganized nucleoid segregation phenotype, as compared to wild-type cells (16). Considering
that topo III is found associated with replication forks in vivo where its substrate, ssDNA, is present
and that in vitro it has a strong decatenation activity that can fully support replication, it is likely
that topo III plays a role in the removal of pre-catenanes. However, at least under normal growth
conditions and when topo I is present, this role appears to be minor as compared to topo IV, the
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 11, 2021. ; https://doi.org/10.1101/2021.07.10.451908doi: bioRxiv preprint
absence of which inhibits growth and causes severe chromosome segregation defects (the par
phenotype (21)).
Initially, topA topB null mutants were shown to generate a RecA-dependent chromosome
segregation and extreme cellular filamentation phenotype and were reported to be non-viable (38).
Nevertheless, it was later found upon prolonged incubation that topA topB null transductants could
be obtained (39). More recently, a correlation between R-loop-triggered RecA-dependent
amplification in the chromosomal Ter region, cellular filamentation, chromosome segregation
defects and growth inhibition phenotypes was demonstrated in topA topB null mutants (22). RNase
HI overproduction was shown to correct these phenotypes and R-loop formation was detected in
a topA null as well as in a topA topB null mutant, but at a higher level in the latter. Importantly,
chromosomal DNA amplifications carrying parC and parE genes for topo IV were detected in
every topA topB null mutant studied, irrespective of the topA or topB null alleles, and whether
RNase HI was overproduced ((22) and see this work). In parallel to this study, a similar observation
was reported for B. subtilis i.e., amplification of a DNA region carrying the parEC operon (for
topo IV) allowing topA topB null mutants to survive (40).
Here, to better understand the relative contribution of the relaxase and decatenase activity in the
essential function(s) of type 1A topos in bacteria, we sought to characterize the effect of deleting
topB on supercoiling in topA null mutants and the mechanism by which topo IV overproduction
allows topA topB null cells to survive. Our main findings are: 1- there is a substantial effect of the
absence of topo III on hypernegative supercoiling in topA null mutants. This illustrates the
significant backup roles that topo I and topo III can play for each other when one of them is absent.
2- the inhibition of R-loop formation requires a topo that can act locally (topo I or topo III, but not
topo IV) to relax transcription-induced negative supercoiling. 3- topo IV overproduction is mostly
required to solve R-loop- and NER-dependent topological problems related to replication in the
absence of type IA topos. 4- excess R-loop formation is well tolerated in cells that have enough
topoisomerase activity to support the ensuing replication-related topological stress.
MATERIALS AND METHODS
Bacterial strains and plasmids
The list of bacterial strains used in this study as well as the details on their constructions can be
found in Table S1. This table also give the list of plasmids used in this work. Transductions with
phage P1vir were done as described previously (41). PCR with appropriate oligonucleotides were
performed to confirm the transfer of the expected alleles in the selected transductants. Strains
JB136, JB335 and JB336 also carry a ΔyncE mutation that was introduced in the gyrB(Ts) strain
before the ΔtopB and topA20::Tn10 alleles. This mutation in the Ter DNA region was introduced
in the frame of our study aiming at the identification of loci that could affect Ter DNA
amplification. It was found that this mutation had no effects on topA topB null cells growth, cell
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 11, 2021. ; https://doi.org/10.1101/2021.07.10.451908doi: bioRxiv preprint
morphology or MFA profiles including Ter DNA amplification (Fig. S1, compare the MFA profile
of JB137, topA topB and JB335, topA topB ΔyncE). The inversion in JB136 involves
recombination between short homologous sequences (FRT, FLP recognition target) in the yncE
and the topB loci that are the result of FLP recombinase activity to remove the Km
R
cassette (42),
that was introduced to substitute for these two genes during the process of JB121 strain
construction. The inversion in JB136 was therefore named IN(1.52-1.84). The mutations uvrB24,
uvrBC25 and gyrA21 all appeared spontaneously in topA topB null mutants, as we confirmed by
sequencing that they were not present in the topB null strains in which the topA null allele was
introduced.
Bacterial cells growth
Cells were grown overnight at 37
o
C on LB plates added with the appropriate antibiotics and
supplements. Cell suspension were prepared, and aliquots were used to obtain an OD
600
of 0.01 in
fresh LB media. The cells were grown at the specified temperature to log phase at the indicated
OD
600
.
Plasmid extraction for supercoiling analysis.
pACYC184∆tet5’ DNA extraction for supercoiling analysis was performed by using the
Monarch
®
Plasmid Miniprep Kit (NEB). Cells were grown at the indicated temperature to an
OD
600
of 0.4. We found that after the first centrifugation step, keeping the cell pellet on ice for 30
minutes instead of resuspending it immediately in the plasmid resuspension buffer did not affect
the results. Chloroquine gel electrophoresis was done as previously reported (43). The gels were
stained with SYBR™ Gold Nucleic Acid Gel Stain (ThermoFisher Scientific). The one-
dimensional gels (7.5 µg/ml chloroquine) were photographed under UV light by using
the Alphaimager HP. The two-dimensional gels (7.5 and 30 µg/ml chloroquine
respectively in the first and second dimension) were photographed by using the Blue
light transilluminator (ThermoFisher scientific).
Spot tests
Cells were grown at 30
o
C to an OD
600
of 0.6. Five µl of 10-fold serial dilutions were spotted on
LB plates that were then incubated at 30
o
C.
Detection of DnaA-independent replication
For the detection of DnaA-independent replication, EdU (ethynyl deoxyuridine) incorporation and
click-labeling using the “Click-It
®
Alexa Fluor 448
®
Imaging kit” (Life Technologies, Molecular
Probes) were done as previously described (34,44). Briefly, the cells were grown in LB medium
at 30°C to an OD
600
of 0.3. An aliquot of cells was used for EdU incorporation for 60 min to detect
ongoing replication in log phase cells. To detect DnaA-independent replication the log phase cells
were first treated with spectinomycin for two hours, to allow the termination of replication rounds
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 11, 2021. ; https://doi.org/10.1101/2021.07.10.451908doi: bioRxiv preprint
