Bacterial Argonaute Proteins Aid Cell Division in the Presence of
Topoisomerase Inhibitors in
Escherichia coli
Anna Olina
,
a
*
Aleksei Agapov
,
a
§
Denis Yudin
,
a
^
Dmitry Sutormin
,
b
Alina Galivondzhyan
,
b
Anton Kuzmenko
,
a,c
Konstantin Severinov
,
d
Alexei A. Aravin
,
c
Andrey Kulbachinskiy
a,e
a
Institute of Molecular Genetics, National Research Center
“
Kurchatov Institute
”
, Moscow, Russia
b
Skolkovo Institute of Science and Technology, Moscow, Russia
c
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, California, USA
d
Waksman University for Microbiology, Rutgers, New Jersey, USA
e
Institute of Gene Biology, Russian Academy of Sciences, Moscow, Russia
ABSTRACT
Prokaryotic Argonaute (pAgo) proteins are guide-dependent nucleases that
function in host defense against invaders. Recently, it was shown that TtAgo from
Thermus
thermophilus
also participates in the completion of DNA replication by decatenating chro-
mosomal DNA. Here, we show that two pAgos from cyanobacteria
Synechococcus elongatus
(SeAgo) and
Limnothrix rosea
(LrAgo) are active in heterologous
Escherichia coli
and aid cell
division in the presence of the gyrase inhibitor cipro
fl
oxacin, depending on the host dou-
ble-strand break repair machinery. Both pAgos are preferentially loaded with small guide
DNAs (smDNAs) derived from the sites of replication termination. Cipro
fl
oxacin increases
the amounts of smDNAs from the termination region and from the sites of genomic DNA
cleavage by gyrase, suggesting that smDNA biogenesis depends on DNA replication and is
stimulated by gyrase inhibition. Cipro
fl
oxacin enhances asymmetry in the distribution of
smDNAs around Chi sites, indicating that it induces double-strand breaks that serve as a
source of smDNA during their processing by RecBCD. While active in
E. coli
, SeAgo does
not protect its native host
S. elongatus
from cipro
fl
oxacin. These results suggest that pAgo
nucleases may help to complete replication of chromosomal DNA by promoting chromo-
some decatenation or participating in the processing of gyrase cleavage sites, and may
switch their functional activities depending on the host species.
IMPORTANCE
Prokaryotic Argonautes (pAgos) are programmable nucleases with incom-
pletely understood functions
in vivo
. In contrast to eukaryotic Argonautes, most studied
pAgos recognize DNA targets. Recent studies suggested that pAgos can protect bacteria
from invader DNA and counteract phage infection and may also have other functions
including possible roles in DNA replication, repair, and gene regulation. Here, we have
demonstrated that two cyanobacterial pAgos, SeAgo and LrAgo, can assist DNA replica-
tion and facilitate cell division in the presence of topoisomerase inhibitors in
Escherichia
coli
.Theyarespeci
fi
cally loaded with small guide DNAs from the region of replication
termination and protect the cells from the action of the gyrase inhibitor cipro
fl
oxacin,
suggesting that they help to complete DNA replication and/or repair gyrase-induced
breaks. The results show that pAgo proteins may serve as a backup to topoisomerases
under conditions unfavorable for DNA replication and may modulate the resistance of
host bacterial strains to antibiotics.
KEYWORDS
prokaryotic Argonautes, gyrase, DNA replication,
ter
sites, cell division,
DNA gyrase, antibiotic resistance
A
rgonaute (Ago) proteins are an evolutionary conserved family of programmable
nucleases that are found in all three domains of life (1
–
4). Eukaryotic Argonautes
(eAgos) participate in RNA interference and use small RNA guides to recognize RNA
Editor
Silvia T. Cardona, University of Manitoba
Copyright
© 2023 Olina et al. This is an open-
access article distributed under the terms of
the
Creative Commons Attribution 4.0
International license
.
Address correspondence to Andrey
Kulbachinskiy, avkulb@yandex.ru.
*
Present address: Anna Olina, Environment
and Sustainability Institute, Biosciences,
University of Exeter, Penryn, United Kingdom.
§Present address: Aleksei Agapov, Environment
and Sustainability Institute, Biosciences,
University of Exeter, Penryn, United Kingdom.
^
Present address: Denis Yudin, Department of
Biology, Institute of Molecular Biology and
Biophysics, ETH Zurich, Zurich, Switzerland.
The authors declare no con
fl
ict of interest.
Received
12 October 2022
Accepted
29 March 2023
Published
27 April 2023
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
1
RESEARCH ARTICLE
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
targets (5
–
8). This is followed by target RNA cleavage through an intrinsic endonucleolytic
activity of eAgo or by recruitment of accessory factors, resulting in posttranscriptional or
transcriptional gene silencing (9
–
12). When
fi
rst identi
fi
ed in bacteria and archaea, prokary-
otic Argonautes (pAgos) served as models to study the structure and biochemical proper-
ties of Ago proteins (3, 13
–
19), but their functional activities in host species remained
unknown. Full-length eAgos and pAgos contain four main domains, N-terminal, PAZ (Piwi/
Argonaute/Zwille), MID (middle), and PIWI (P-element Induced WImpy testis), responsible
for interactions with the guide and target molecules and for target cleavage, with a cata-
lytic tetrad located in the PIWI domain. However, phylogenetic analysis demonstrated that
pAgos are much more diverse than eAgos and only a smaller part of them is catalytically
active, while others contain substitutions of the catalytic residues and may include only
MID and PIWI domains (1, 2, 4, 20, 21). This analysis also revealed widespread horizontal
transfer of pAgos among bacterial and archaeal species, suggesting that they can function
in various genetic contexts and environments.
In vitro
studies of pAgo proteins demonstrated that their primary target is DNA, except
for a small group of pAgos that can target RNA, discovered recently (22, 23). Recognition
of DNA targets by pAgos can be guided by small DNAs, as observed for most studied
pAgos, or by small RNAs (24
–
33). These
fi
ndings were corroborated by
in vivo
analysis of
several DNA-targeting pAgo proteins in bacterial cells, which revealed that pAgos preferen-
tially recognize foreign DNA, such as plasmids, mobile elements, and phages. In particular,
it was shown that RsAgo from
Rhodobacter sphaeroides
,TtAgofrom
Thermus thermophilus
,
PfAgo from
Pyrococcus furiosus
,MjAgofrom
Methanocaldococcus jannaschii
,andCbAgo
from
Clostridium butyricum
decrease plasmid DNA content and transformation ef
fi
ciency,
and CbAgo counteracts phage infection (25, 29, 31
–
34).Accordingly,thesepAgosare
actively loaded with guide molecules corresponding to plasmid or phage sequences dur-
ing their expression in
Escherichia coli
(29, 32, 34). As demonstrated for CbAgo, generation
of small guide DNAs from foreign genetic elements depends on both the catalytic activity
of pAgo and the action of cellular nucleases (34). These studies suggested that catalytically
active pAgos perform targeted degradation of foreign DNA, even if they are expressed in
heterologous cells.
Recent studies suggested that elimination of invader DNA may not be the sole func-
tion of pAgos. In particular, TtAgo was shown to increase the resistance of its host ther-
mophilic bacterium
T. thermophilus
to cipro
fl
oxacin (Cfx), an inhibitor of DNA gyrase
that impairs DNA replication and prevents normal cell division (35). TtAgo was shown
to target the region of replication termination and participate in decatenation of chro-
mosomal DNA, thus helping to complete DNA replication when the gyrase function is
inhibited (35). However, it remained unknown whether this function in cell division
might be conserved among other DNA-targeting pAgos.
Here, we have analyzed two pAgo proteins from mesophilic cyanobacteria, SeAgo
from
Synechococcus elongatus
and LrAgo from
Limnothrix rosea
. SeAgo is more closely
related to TtAgo (35.7% identity in the MID and PIWI domains), while LrAgo is more
distant from it on the phylogenetic tree (25.8% identity) (20). Both SeAgo and LrAgo
are DNA-guided DNA nucleases that can perform precise cleavage of target DNA
in
vitro
at moderate temperatures, suggesting that they may be active in other meso-
philic bacteria (27, 28). SeAgo was previously shown to interact with small guide DNAs
in its native host
S. elongatus
but without obvious target speci
fi
city (28). We have
found that, when expressed in a heterologous
E. coli
host, both SeAgo and LrAgo are
loaded with small DNAs corresponding to the sites of replication termination.
Furthermore, both SeAgo and LrAgo increase
E. coli
resistance to the gyrase inhibitor
cipro
fl
oxacin, suggesting that targeting of termination sites by pAgos may aid DNA
replication in various prokaryotic species.
RESULTS
Cyanobacterial pAgos rescue
E. coli
growth in the presence of topoisomerase
inhibitors.
To explore whether SeAgo and LrAgo can affect DNA replication and cell
division, we expressed them in the heterologous
E. coli
system. The SeAgo and LrAgo
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
2
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
genes were cloned under the control of an arabinose-inducible promoter in pBAD-
based vectors (Fig. 1). Western blots con
fi
rmed that both proteins were expressed after
the addition of arabinose (Ara) (see Fig. S1A in the supplemental material). We then
studied their effects on cell growth and resistance to gyrase inhibition and further ana-
lyzed their DNA speci
fi
city
in vivo
(Fig. 1).
We
fi
rst determined the range of sublethal concentrations of cipro
fl
oxacin (Cfx) in the
absence of pAgo proteins by measuring the kinetics of cell growth (optical density at 600
nm [OD
600
]curves)ofan
E. coli
strain containing an empty expression plasmid, either in the
absence or in the presence of Ara. In both cases, cell growth was partially inhibited starting
from 0.2 to 0.5 ng/mL of Cfx and was completely inhibited at concentrations higher than
2ng/mL,whichwasde
fi
nedastheMICofCfx(Fig.2,toptworows).Wethentestedthe
effects of the pAgo proteins on cell growth. LrAgo and SeAgo did not affect the growth
kinetics in the absence of Cfx; however, both proteins protected the cells from the sublethal
and lethal concentrations of Cfx (0.5 to 15.6 ng/mL or 0.25 to 8 MIC) (Fig. 2, bottom two
rows). Comparable effects were observed at different concentrations of the Ara inducer
(0.01% and 0.1%) (Fig. S1B). We therefore used milder expression conditions (with lower
Ara) in all subsequent experiments. Even at high Cfx concentrations (31 to 62 ng/mL or 16
to 32 MIC), pAgos partially restored cell growth, which was completely absent in the control
conditions (Fig. 2). These experiments suggested that both pAgos can help the cells over-
come the inhibitory effects of Cfx on DNA replication.
Previous experiments with TtAgo suggested that it helps to decatenate chromo-
somes by directly introducing double-strand breaks in the genomic DNA of
T. thermo-
philus
or by assisting their processing by the DNA repair machinery (35). If SeAgo and
LrAgo acted by a similar mechanism, double-strand breaks generated during this pro-
cess should subsequently be repaired by homologous recombination, depending on
the RecA protein and the RecBCD helicase-nuclease involved in double-strand break
processing (36
–
38). To test whether this was the case, we compared the effects of Cfx
and pAgos in wild-type and mutant
E. coli
strains with deletions of
recA
as well as
recB
and
recD
. As expected, Cfx had stronger effects on the growth of the mutant strains in
comparison with wild-type cells (Fig. 3A). Notably, SeAgo and LrAgo did not stimulate
the growth of
recA
-minus and
recBD
-minus strains in the presence of Cfx (Fig. 3A), sug-
gesting that the function of the pAgo proteins depends on the homologous recombi-
nation machinery.
FIG 1
Analysis of pAgo functions in
E. coli
.
E. coli
strains expressing plasmid-encoded SeAgo or LrAgo or containing
a control empty plasmid were grown in the absence or in the presence of cipro
fl
oxacin (Cfx), followed by CFU
counting, cell microscopy, and analysis of pAgo-associated smDNAs.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
3
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
The observed effects of Cfx on the density of cell cultures may not directly corre-
spond to changes in the number of viable bacteria because Cfx induces formation of
multinucleate cell
fi
laments that increase culture density but do not divide properly
(39) (see Discussion). We therefore directly compared the number of CFU and analyzed
cell morphology in bacterial cultures grown in the absence and in the presence of Cfx.
After 4.5 h of growth, when the effects of Cfx just became visible on the growth curves
obtained by optical density measurements (Fig. 3A), Cfx dramatically decreased CFU
numbers in the wild-type
E. coli
strain (20- to 360-fold in three biological replicates)
(Fig. 3B; see also Fig. S2 in the supplemental material). Similar differences in CFU num-
bers were observed for the
rec
-minus strains (with lower numbers of viable cells in the
recBD
strain even in the absence of Cfx) (Fig. 3B). Microscopy analysis con
fi
rmed that
inhibition of cell division by Cfx caused formation of long
fi
lamentous cells containing
unsegregated DNA (Fig. 4A, top; see also Fig. S3 in the supplemental material).
We then analyzed the effects of pAgo expression on the number of viable bacteria. In
the absence of Cfx, CFU numbers were similar in control and pAgo-expressing strains. In
contrast, in the presence of Cfx, CFU numbers were strongly increased in the strains
expressing pAgos (Fig. 3B; see also Fig. S2). This effect was especially prominent in the case
of SeAgo; for this strain, Cfx decreased cell numbers only 3- to 16-fold in comparison with
an up to 360-fold reduction observed in the control strain. Microscopy analysis showed
that expression of both SeAgo and LrAgo resulted in disappearance of long
fi
laments
induced by Cfx and increased the number of normal cells and shorter
fi
laments. This effect
was especially pronounced for SeAgo, while elongated cells were still observed in the case
of LrAgo (Fig. 4A, middle and bottom; see also Fig. S3).
Since Cfx primarily targets gyrase, we analyzed the effects of gyrase knockdown on cell
morphology in control cells and upon pAgos expression. To silence gyrase expression, cat-
alytically inactive dCas9 and single guide RNA (sgRNA) corresponding to the beginning of
the coding region of the
gyrA
gene were expressed in the
E. coli
strains lacking or contain-
ing pAgos. Control experiments demonstrated that only a low level of gyrase knockdown
couldbeachievedbythisapproach(10to20%decreaseinthemRNAlevelsmeasuredby
quantitative PCR). However, this was suf
fi
cient to observe formation of short cell
fi
laments
by microscopy (Fig. 4B, top). Similarly to the experiments with Cfx, these
fi
laments disap-
peared in the presence of pAgos (Fig. 4B, middle and bottom). Together, these results
FIG 2
Growth of
E. coli
strains lacking or expressing pAgos at different concentrations of Cfx.
E. coli
transformed with an empty pBAD plasmid or pBAD
encoding SeAgo or LrAgo were grown at 30°C in a plate reader with indicated concentrations of Cfx (0 to 62.5 ng/mL) and Ara to induce expression of
pAgos, and OD
600
was monitored over time.
E. coli
transformed with empty pBAD and grown without Ara was used as a control. Means from four
biological replicates are shown; shades represent 0.95 con
fi
dence intervals of the mean. Gray rectangle highlights the range of Cfx concentrations (0.5 to
15.6 ng/mL), the inhibitory effects of which are suppressed by pAgos expression.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
4
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
FIG 3
Comparison of the effects of pAgos on the viability of wild-type and
rec
-minus
E. coli
strains grown in the absence and in the presence of Cfx. (A)
Growth kinetics of the wild-type and
recA
-minus and
recBD
-minus strains, measured in a plate reader. Cfx was added to 0.5 ng/mL when indicated. Means
and standard deviations from three biological replicates are shown. (B) Comparison of the number of viable cells (CFU) in the wild-type and mutant
E. coli
strains lacking or expressing pAgos in the absence and in the presence of Cfx. The samples were taken from
E. coli
cultures grown for 4.5 h in the absence
or in the presence of Cfx (indicated with dashed lines in panel A), and CFU numbers were determined by their serial plating without Cfx. Representative
LB
plates from one of biological replicates are shown for
E. coli
strains grown in the absence (top) or in the presence (bottom) of Cfx (see Fig. S2 in the
supplemental material for CFU numbers from the three biological replicates).
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
5
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
suggestthatpAgoscanaidcelldivisionin
E. coli
cells when gyrase is inhibited by either
Cfx or transcriptional knockdown, indicating that they may help to complete DNA replica-
tion impaired by topoisomerase de
fi
ciency.
pAgo-bound smDNAs are enriched in the termination region of the chromo-
some and are generated with participation of RecBCD.
Previous analysis of small
guide DNAs (smDNAs) associated with TtAgo in
T. thermophilus
demonstrated that they
are highly enriched around the region of replication termination. This led to the suggestion
that TtAgo may facilitate decatenation of chromosomal DNA by targeting the
ter
region of
the chromosome (35). To explore whether SeAgo and LrAgo have preference for speci
fi
c
genomic regions or sequence motifs, we checked whether they were loaded with smDNAs
during their expression in the heterologous
E. coli
host. Electrophoretic analysis of nucleic
acids copuri
fi
ed with SeAgo and LrAgo revealed that both pAgos were associated with
;
14- to 19-nucleotide (nt) smDNAs (see Fig. S4 in the supplemental material). We
sequenced libraries of pAgo-associated smDNAs obtained from late logarithmic or station-
ary bacterial cultures (for 5.5- and 12.5-h time points) (Fig. S4) in the absence or in the pres-
ence of Cfx and analyzed the distribution of smDNAs along the chromosomal and plasmid
DNAs.
Sequence analysis of smDNAs con
fi
rmed that the majority of smDNA associated
with SeAgo and LrAgo had a length of 15 to 19 nt and 14 to 19 nt, respectively
(Fig. 5A). Except for a slight preference for G at the
fi
rst guide position in the case of
SeAgo, no strong nucleotide biases were found along the guide length (Fig. 5B). The
mean GC content of smDNAs corresponded to genomic DNA of
E. coli
(51%), and it
was only slightly increased at the
fi
rst guide position for SeAgo and slightly decreased
upstream of the guide 5
9
-end and around 10 to 15 guide nucleotides for LrAgo
FIG 4
Effects of pAgo expression and gyrase inhibition on
E. coli
cell morphology. (A)
E. coli
cells lacking or containing pAgos were
grown in the absence (left) or in the presence (right) of Cfx. The samples were taken at 4.5 h from the cultures shown in Fig. 3. (B)
Effects of dCas9 gyrase (
gyrA
) knockdown on cell morphology. Fluorescence microscopy after acridine orange staining. The scale bar
is 10
m
m. See Fig. S3 in the supplemental material for additional
fi
elds of view.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
6
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
(Fig. 5C). Overall, this analysis suggests that both SeAgo and LrAgo have no speci
fi
c
motif preferences and can likely interact with guide DNAs of any sequence.
Similar to several previously studied pAgos, both SeAgo and LrAgo were enriched
with smDNAs derived from plasmid DNA (11- to 14-fold enrichment over chromosomal
DNA for SeAgo and 3- to 9-fold preference for LrAgo after accounting for the relative
replicon lengths and the plasmid copy number, 12 for pBAD). A similar bias for plasmid
DNA was observed in the presence of Cfx (see Table S1 in the supplemental material).
Nevertheless, the majority of smDNA guides (78 to 95% in various experiments) (Table
S1) were derived from the
E. coli
chromosome, indicating that genomic DNA can be
ef
fi
ciently targeted by both pAgos.
The distribution of smDNA guides along the chromosome was highly uneven, with two
large peaks at the
terA
and
terC
sites of replication termination observed for both pAgos
(Fig. 6A and B). In addition, two smaller peaks were present at the next pair of
ter
sites,
terB
and
terD
(Fig. 6B). For SeAgo, similar targeting of the
ter
sites was observed at different
stages of cell growth. For LrAgo, targeting of the
ter
sites was less ef
fi
cientatthelogarith-
mic stage but was increased in the stationary culture (Fig. 6B).
The outer edges of the smDNA peaks precisely correspond to the
ter
motifs in chro-
mosomal DNA bound by the Tus protein (Fig. 6B). The inner borders of the peaks coin-
cide with the closest Chi-sites co-oriented with the direction of replication for each
replichore (forward site in the plus strand for
terC
and reverse site in the minus strand
for
terA
). Chi-sites (5
9
-GCTGGTGG-3
9
in
E. coli
) serve as stop-signals for the RecBCD heli-
case-nuclease during processing of double-strand DNA breaks (DSBs), located down-
stream (in the 3
9
-direction) of the Chi sequence (36
–
38). This suggests that smDNAs in
the
ter
region are produced with participation of RecBCD from double-strand DNA
ends that are formed after stalling of the replication forks at Tus-bound
ter
sites.
In a further support for this notion, the ratio of smDNAs loaded from the plus and
minus genomic strands in the
ter
region is asymmetric depending on the orientation
FIG 5
Characteristics of smDNAs associated with pAgos in
E. coli
in the absence or in the presence of Cfx. The smDNA samples were isolated at the
exponential phase of growth (5.5 h) (see Fig. S4 in the supplemental material for the growth curves). (A) Distribution of smDNA lengths for each smDNA
library. (B) Nucleotide logos for different smDNA positions starting from the 5
9
-end of guide DNA. (C) GC content of the smDNA sequences and of the
surrounding genomic regions for each condition. The mean GC content of the
E. coli
genome is indicated.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
7
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
of the
ter
sites. For both
terA
and
terC
(and
terB
), more smDNAs correspond to the DNA
strand with the 3
9
-end oriented toward the
ter
site (the minus genomic strand for
terA
and the plus strand for
terC
; 2- to 3-fold more than for complementary 5
9
-terminated
DNA strands). This corresponds to the asymmetry in smDNA processing by RecBCD
previously reported for another pAgo protein, CbAgo (34).
FIG 6
Whole-genome mapping of smDNAs associated with pAgos in
E. coli
. (A) Genomic distribution of smDNAs isolated from SeAgo and LrAgo in the
absence or in the presence of Cfx at the logarithmic or stationary phases of growth (5.5 and 12.5 h) (see Fig. S4 in the supplemental material). The numbe
rs
of smDNAs along the genomic coordinate are shown in reads per kilobase per million aligned reads in the smDNA library (RPKM), individually for the plus
(green) and minus (red) genomic strands. The
ori
,
ter
sites, and the direction of replication are indicated. (B) Targeting of the
ter
region by pAgos. SmDNA
densities in each genomic strand (plus strand, green; minus strand, red) in the
ter
region in
E. coli
cultures grown in the absence (top) and in the presence
(bottom) of Cfx are shown. Positions of the
terA
,
terC
,
terD
,and
terB
sites are shown with dashed lines and the directions of replichores are shown with
arrows. Chi sites in the plus (green) and minus (red) strands are shown above the plots. The closest Chi sites oriented toward
terA
(
“
Chi
2
”
), and
terC
(
“
Chi
1
”
)
are indicated. SmDNA numbers are shown in RPKM. The numbers of smDNAs from each DNA strand from each
ter
site, calculated for genomic regions
between the
ter
site and the closest Chi site in the correct orientation, are shown as a percentage of the total number of smDNAs mapped to both strands of
the whole genomic sequence in each smDNA library. (C) Effects of Sfx on the distribution of pAgo-associated smDNA between genomic DNA strands during
replication. The ratio of pAgo-associated smDNAs corresponding to the plus and minus genomic strands was calculated independently for samples isol
ated
from strains grown in the presence and in the absence of Cfx, and the obtained pro
fi
les were then divided by each other.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
8
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
Targeting of the
ter
region by both SeAgo and LrAgo was increased in the presence of
Cfx (Fig. 6A and B, compare upper and lower rows). Speci
fi
cally, the fraction of smDNAs
corresponding to the peak at
terA
was increased
;
2.6-fold and 3.9-fold for SeAgo and
LrAgo, respectively, at the logarithmic stage of growth, and 2.4-fold and 2.9-fold at the sta-
tionarystageofgrowth(Fig.6B).Furthermore,Cfxchangedtherelativesizesofthe
smDNA peaks at
terA
and
terC
. In the absence of Cfx, the peak at
terC
was larger than at
terA
for both pAgos at both stages of growth (Fig. 6B). This corresponds to the shorter
length of the rightward replichore that terminates at
terC
and indicates that smDNAs at
ter
sites are likely produced during replication termination, which is more frequent at
terC
(34,
40). In contrast, the peaks at
terA
and
terC
became comparable in the presence of Cfx
because smDNA loading was stimulated to a higher extent at
terA
than at
terC
(Fig. 6B). For
both pAgos, this effect became especially prominent in the stationary cultures. This indi-
cates that the relative frequencies of replication termination at
terA
and
terC
may be lev-
eled in the presence of Cfx or that the ef
fi
ciency of smDNA processing in these regions
may become less dependent on replication in these conditions.
Effects of cipro
fl
oxacin-induced double-strand breaks on generation of pAgo-
associated smDNAs.
To explore a possible connection between smDNA processing
and replication on the whole-genome level, we compared the ratio of pAgo-bound
smDNAs produced from the plus and minus genomic strands in the presence or in the
absence of Cfx (Fig. 6C). In the case of SeAgo (but not LrAgo), Cfx increased generation
of smDNAs from the leading DNA strand in both replichores, resulting in an increased
ratio (
.
1) of smDNAs corresponding to the plus and minus strands for the rightward
replichore and a decreased ratio (
,
1) for the leftward replichore (Fig. 6C). The only
exception was the
ter
region, in which this ratio was changed in favor of the 3
9
-termi-
nated strand at each
ter
site for both pAgos (
,
1 for
terA
and
.
1 for
terC
) as discussed
in the previous section. Preferential generation of smDNAs from the leading DNA
strand observed for SeAgo correlates with preferential co-orientation of Chi sites with
the direction of replication and indicates that smDNAs may be produced during proc-
essing of double-strand breaks after gyrase inhibition and replisome stalling.
To further explore the role of double-strand DNA breaks (DSBs) and the RecBCD ma-
chinery in the biogenesis of smDNA guides bound by pAgos, we calculated the distri-
bution of smDNAs around Chi sites throughout the whole chromosome, excluding the
ter
region (Fig. 7). This metaplot analysis revealed that distribution of pAgo-bound
smDNA guides was asymmetric and dependent on the orientation of genomic DNA
strands relative to the Chi sites. For the DNA strand co-oriented with Chi (forward ori-
entation of Chi [F]), the amounts of smDNAs derived from the 3
9
-side of Chi were
much higher than those from the 5
9
-side, with an abrupt drop immediately at the Chi
sequence (Fig. 7A, B, green; see also Fig. S5A in the supplemental material). In particu-
lar, for SeAgo, this drop corresponded to 23% and 22% changes in the amounts of
smDNAs at Chi sites relative to background levels in the exponential and stationary
cultures, respectively. For the DNA strand that was oriented in the opposite direction
relative to Chi (reverse orientation [R]), the changes in the amounts of smDNAs around
Chi were much less pronounced (7% and 8% for SeAgo) (Fig. 7B, light gray; see also
Fig. S5A). This indicates that recognition of Chi sites by RecBCD during processing of
DSBs inhibits generation of smDNAs from the 3
9
-terminated DNA strand.
Inhibition of gyrase with Cfx stabilizes DSBs by covalently trapping cleaved DNA
strands with the gyrase. This should increase RecBCD-dependent processing of gyrase-
dependent DSBs and further stimulate biogenesis of smDNAs at Chi sites. Indeed, the
asymmetry of smDNA loading around Chi sites was increased in the presence of Cfx for
both pAgos at both exponential and stationary stages. In particular, the drop in the
amounts of smDNAs at the properly oriented Chi sites was changed from 23% to 40%
for SeAgo and from 14% to 18% for LrAgo in the logarithmic phase of growth (Fig. 7B
and C) and from 22% to 40% for SeAgo and from 13% to 27% for LrAgo in the station-
ary phase (Fig. S5A and B). Thus, Cfx stimulates RecBCD-dependent processing of
smDNAs, likely by increasing the number of DSBs formed in the chromosomal DNA
due to inhibition of gyrase.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
9
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
To directly test the effects from Cfx-induced DSBs on smDNA abundance, we ana-
lyzed the distribution of pAgo-associated smDNAs relative to Cfx-stabilized gyrase
cleavage sites (GCSs), which were previously mapped in the
E. coli
genome (41). No
enrichment of smDNAs was seen immediately at GCSs in Cfx-treated versus untreated
cells (see Fig. S6 in the supplemental material). However, when we compared Chi sites
with an adjacent GCS located downstream of the Chi sequence (Fig. 8A) and all other
Chi sites with no adjacent downstream GCSs, a highly signi
fi
cant increase in smDNA
FIG 7
Asymmetry of smDNA distribution around Chi sites. (A) Scheme of DNA processing by RecBCD between
a DSB and an upstream Chi site, illustrating the observed polarity of smDNA loading into pAgos. DNA
unwinding by RecBCD is followed by asymmetric processing of the two DNA strands by RecBCD and other
nucleases; RecBCD loads RecA onto the 3
9
-terminated strand at the Chi site, thus shielding this strand from
further degradation. (B) Metaplots of the densities of smDNAs around Chi sites analyzed for smDNA libraries
isolated from logarithmic cultures (5.5 h of growth) of
E. coli
strains grown in the absence and in the presence
of Cfx (averages from two replicate experiments). SmDNA numbers were independently calculated for the DNA
strands co-oriented (green, F) and oppositely oriented (gray, R) with the Chi sequence (5
9
-GCTGGTGG-3
9
) for all
Chi sites in both genomic strands (833 sites in total) and smoothened with a 400-bp sliding window. (C)
Comparison of metaplots of normalized densities of smDNAs around co-oriented Chi sites for smDNA libraries
isolated from
E. coli
grown in the absence and in the presence of Cfx (the data correspond to green smDNA
pro
fi
les in panel A). For normalization, averaged and smoothened RPKM values around Chi sites were divided
by the background RPKM value calculated for regions remote from Chi sites (
2
50 to
2
35 kb and
1
35 to
1
50 kb from the Chi sequence). Arrows indicate relative differences between background and minimal
densities of smDNAs at the Chi sites.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
10
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
FIG 8
Effects of Cfx on the enrichment of pAgo-associated smDNAs around Chi sites with an adjacent GCS at the 3
9
-
side. (A) Scheme of DNA processing between a GCS and the closest upstream Chi site, illustrating the observed
polarity of smDNA loading into pAgos. (B) Metaplots of the relative densities of smDNAs around Chi sites adjacent to
a downstream GCS from the 3
9
-side of the Chi sequence (carmine curves, 188 sites in total) and all other Chi sites,
lacking adjacent downstream GCSs (lilas curves, 645 sites in total). A Chi site and a GCS were considered adjacent if
there were no co-oriented Chi sites in between. The data were averaged for two replicate experiments for the
exponential (5.5 h, top) and early stationary (12.5 h, bottom) phases of growth. SmDNA density was calculated for
DNA strands co-oriented with the Chi sites for
1
Cfx and
2
Cfx conditions independently, and then
1
Cfx density was
divided by Cfx density. The resultant relative density was smoothened with a 1-kb sliding window. Blue rectangles
mark the region (from the Chi sequence to
1
5 kb) used to quantify the relative enrichments of smDNAs in panel B.
(C) Quanti
fi
cation of relative enrichments of smDNAs at the 3
9
-sides of Chi sites for Chi sites adjacent to a
downstream GCS (DS, carmine) and all other Chi sites (Oth, violet). The error bars represent mean values
6
standard
deviation (SD) for the same sets of Chi sites as in panel A. The enrichments were compared by a two-sided
t
test.
P
values of
,
0.0005 are indicated with three asterisks.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
11
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
amounts was observed for the
fi
rst group of Chi sites (Fig. 8B). Speci
fi
cally, the
amounts of smDNAs were increased from the 3
9
-side of the Chi sequence, directly sug-
gesting that gyrase-induced DNA cleavage stimulates smDNA processing between the
GCS and the upstream Chi site. This increase was highly statistically signi
fi
cant for both
pAgos and both stages of growth and was especially pronounced for SeAgo (Fig. 8C)
(
P
values of 1.4e
2
5 and 2.9e
2
5 for 5.5 and 12.5 h, respectively). To con
fi
rm that the
position of the GCS relative to the Chi sequence was important for smDNA processing,
we also compared Chi sites with an adjacent downstream GCS (Chi sites in the active
orientation) versus Chi sites with an adjacent upstream GCS (Chi sites in the inactive
orientation). As expected, the enrichment of smDNAs at the 3
9
-side of Chi sites from
the
fi
rst group was signi
fi
cantly higher than in the case of Chi sites with upstream
GCSs (see Fig. S7 in the supplemental material). Taken together, our data indicate that
pAgos preferentially load smDNAs generated with participation of RecBCD at the 3
9
-
sides of Chi sites during processing of downstream DSBs generated by Cfx.
Distribution of pAgo-associated smDNAs relative to coding regions.
In vitro
experiments with several pAgo proteins, including TtAgo, CbAgo, and LrAgo, demon-
strated that their ability to process double-stranded DNA depends on DNA supercoil-
ing (25, 27, 32). Our results indicate that gyrase inhibition, which induces DSBs and
changes the supercoiling state of the chromosome, also changes the distribution of
smDNAs. Another factor that could potentially affect smDNA processing by changing
DNA supercoiling is transcription, which induces positive supercoils in front of the
moving RNA polymerase and negative supercoils behind it (41
–
44). To explore the role
of transcription in the targeting of chromosomal DNA by pAgos, we
fi
rst compared the
abundances of pAgo-associated smDNAs derived from each genomic strand for genes
codirected or oppositely directed relative to replication. No differences between these
groups of genes were found for either SeAgo or LrAgo (see Fig. S8A in the supplemen-
tal material). Divergent and convergent gene orientation is associated with negative
and positive DNA supercoiling in intergenic regions, respectively, which could affect
smDNA production (42). We analyzed smDNA abundance in intergenic DNA regions
for divergent, convergent, and co-oriented gene pairs. No signi
fi
cant differences in the
amounts of smDNAs were detected for the different types of gene pairs for both
pAgos (Fig. S8B). Together, our data suggest that transcription-induced supercoiling is
unlikely to have major effects on smDNA biogenesis.
Analysis of the effects of cipro
fl
oxacin and SeAgo on cell division in
S. elongatus
.
The experiments presented above demonstrated that pAgo proteins from mesophilic cya-
nobacteria can suppress defects in DNA replication in
E. coli
caused by gyrase inhibition,
with stronger effects observed for SeAgo. To test whether SeAgo may have similar func-
tions in its native species, we compared
S. elongatus
strains with the natural level of
expression of SeAgo (wild-type strain), without SeAgo (
D
SeAgo), and with an increased
level of SeAgo, expressed from a strong constitutive promoter (
:
SeAgo). Titration of Cfx
demonstrated that growth of all three strains in liquid culture was fully inhibited at
$
15 ng/mL of the antibiotic, indicating that
S. elongatus
is slightly less sensitive to Cfx
than
E. coli
(see Fig. S9A in the supplemental material). We then compared cell growth at
sublethal Cfx concentrations. Wild-type and
D
SeAgo strains had identical growth kinetics
in the absence and in the presence of 10 ng/mL of Cfx (Fig. S9A). In contrast, the growth
of the strain with overexpression of SeAgo was strongly inhibited in the presence of the
same Cfx concentration. Microscopy analysis revealed no changes in the cell number or
morphology in either wild-type or
D
SeAgo strains in these conditions (Fig. S9B). Cell mor-
phology also remained unchanged in the case of the overexpressor strain, despite the
lower number of cells observed in the presence of Cfx (Fig. S9B). Thus, SeAgo does not
increase the resistance of the wild-type strain of
S. elongatus
to Cfx in comparison with
the deletion strain, while its overexpression is toxic in the presence of the antibiotic.
Furthermore, Cfx inhibits
S. elongatus
growth without formation of
fi
lamentouscells,sug-
gesting that its mechanism of action in cyanobacterial cells may be different from
E. coli
.
To get further insight into possible functions of SeAgo in
S. elongatus
, we compared
sensitivity of
S. elongatus
strains with various levels of expression of SeAgo to other
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
12
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
genotoxic agents, 4-nitroquinoline 1-oxide (4-NQO) and nitrofurazone (NFZ). Both
compounds produce DNA guanosine adducts that are mainly removed by nucleotide
excision repair but may also induce secondary DNA damage and DSBs during replica-
tion (45
–
47). It was found that strains with wild-type SeAgo, deletion of SeAgo, overex-
pression of wild-type SeAgo, and overexpression of a catalytically dead variant of
SeAgo all had similar sensitivities to these treatments (see Fig. S10 in the supplemental
material). Therefore, the primary function of SeAgo in
S. elongatus
may not be related
to DNA replication and cell division (see Discussion).
DISCUSSION
In contrast to eAgos that recognize RNA targets, the majority of studied pAgos preferen-
tially target DNA
in vitro
and
in vivo
, suggesting that their mechanism of action is different
from eAgos. However, similarly to eAgos, studied pAgos can target foreign genetic ele-
ments (25, 29, 31
–
34), demonstrating that the defensive function of Ago proteins is con-
served between prokaryotes and eukaryotes (48). At the same time, pAgos might play other
roles, including the regulation of gene expression, participation in cellular suicide systems,
and DNA repair (21, 29, 49
–
52). In particular, TtAgo from
T. thermophilus
was recently shown
to participate in separation of chromosomal DNA during replication (35). This activity of
TtAgo became crucial for cell division when the DNA gyrase, the sole type II topoisomerase
in
T. thermophilus
, was inhibited by Cfx. TtAgo associated with small guide DNAs corre-
sponding to the termination region of replication and coprecipitated with several proteins
involved in DNA processing, including gyrase and factors involved in DNA recombination
(AddAB, a RecBCD homolog in
T. thermophilus
). It was therefore proposed that TtAgo might
help to decatenate daughter chromosomes by direct DNA cleavage and/or by recruiting
accessory factors to the termination region (35). However, suppressive effects of TtAgo
were not observed for another DNA gyrase inhibitor, novobiocin, which prevents DNA
cleavage by gyrase without producing DSBs in chromosomal DNA (35). This suggests that,
alternatively, TtAgo may be involved in the repair of DSBs at GCSs induced by Cfx.
Here, we have demonstrated that two pAgo proteins from mesophilic cyanobacteria,
SeAgo and LrAgo, facilitate cell division and prevent cell
fi
lamentation in the presence of
Cfx in
E. coli
. The primary target of Cfx in
E. coli
is gyrase (53
–
56), and both pAgos also
suppress a milder phenotype caused by dCas9 knockdown of gyrase expression in
E. coli
.
Inhibition of gyrase by Cfx strongly affects replication by changing DNA supercoiling and
integrity and by introducing direct roadblocks to the moving replisomes (56
–
59).
Additionally, Cfx can target topoisomerase IV in
E. coli
, which is also involved in chromo-
some decatenation (55). Thus, formation of cell
fi
laments with unsegregated DNA in the
presence of Cfx can be explained by failure to separate daughter chromosomes due to
their incomplete replication and decatenation and by the induction of the SOS response
after DNA damage, preventing cell division (59
–
64).
Available data suggest several possible mechanisms explaining the observed effects of
pAgos on DNA replication and cell division. The patterns of smDNA loading into both
SeAgo and LrAgo suggest that genomic smDNAs are primarily generated during nuclease-
dependent processing of DSBs and replication intermediates (Fig. 9). When loaded with
smDNAs, SeAgo, and LrAgo may help the cells to complete DNA replication by direct target-
ing of chromosomal DNA through their nuclease activity. By doing this, they may help to (i)
relax excessive supercoils accumulated in the absence of gyrase by their DNA nicking activ-
ity, and (ii) decatenate sister chromosomes by introducing DSBs in the
ter
region. Cleavage
of both DNA strands may be performed by pAgos loaded with complementary smDNAs
produced from the
ter
sites or by their binding to pre-existing nicks and cutting the second
DNA strand as previously proposed for TtAgo (35). Alternatively, (iii) pAgos may assist the
cellular DNA repair machinery in the removal of covalent gyrase-DNA intermediates stabi-
lized by Cfx or in the processing of the resulting DSBs, thus accelerating the repair process.
This mechanism is supported by the observed enrichment of smDNAs from the 3
9
-sides of
Chi sites adjacent to GCSs in Cfx-treated cells. Finally, (iv) TtAgo was proposed to recruit DSB
repair factors to the
ter
region in
T. thermophilus
, possibly facilitating DNA repair even
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
13
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
without pAgo-dependent DNA cleavage (35). However, SeAgo and LrAgo are not expected
to speci
fi
cally interact with host-speci
fi
c factors in heterologous
E. coli
and are more likely to
directly participate in chromosomal DNA processing. Indeed, the catalytic activity of pAgos
contributes to smDNA biogenesis since a catalytically dead mutant of SeAgo does not bind
smDNAs in
E. coli
(28), and much lower amounts of smDNAs are associated with a catalyti-
cally dead mutant of TtAgo in
T. thermophilus
. The catalytic activity of pAgos is also impor-
tant for their protective function since mutant TtAgo provides only low levels of protection
against Cfx in
T. thermophilus
(35).
Both SeAgo and LrAgo preferentially target the replication termination region of the
E.
coli
chromosome and are loaded with smDNAs corresponding to the
ter
sites. Similar
enrichment of pAgo-associated smDNAs at the
ter
sites was previously observed for TtAgo
in
T. thermophilus
(35) and for CbAgo expressed in
E. coli
(34). SmDNAs loaded into pAgos
in the
ter
region are preferentially generated from 3
9
-ends of DNA strands oriented toward
ter
sites and are con
fi
ned to the areas between
ter
sites and the closest Chi sites. SmDNA
loading into pAgos also depends on Chi sites on the genomic scale and is not restricted to
the
ter
region. This pattern indicates that smDNAs are produced during asymmetric proc-
essing of the two DNA strands at replicated DNA ends or DSBs by RecBCD and/or other cel-
lular nucleases that cooperate with RecBCD during DNA unwinding (34, 36
–
38).
The proposed role of pAgos in DNA decatenation requires introduction and repair
of DSBs in chromosomal DNA. Indeed, the anti-inhibitory function of pAgos depends
on the homologous recombination machinery and is not observed in
E. coli
strains
lacking the key components of double-strand break repair, RecBCD or RecA. As previ-
ously shown for CbAgo, RecBCD is not required for biogenesis of smDNAs and their
loading into pAgos; however, inactivation of RecBCD changes the pattern of chromo-
somal smDNA distribution and removes smDNA peaks around
ter
sites (34). Thus, it
FIG 9
Possible activities of pAgos explaining their observed suppressor effects on bacterial cell growth. pAgos capture
guide smDNAs generated during chromosomal DNA replication and repair with the participation of RecBCD. Guide-
loaded pAgo may then (i) relax supercoils in the replicating chromosome through its DNA-nicking activity, (ii) help to
decatenate chromosomes under conditions of gyrase inhibition by attacking both DNA strands with guide DNAs
corresponding to the
ter
region or by processing of nicked DNA strands, (iii) participate in the processing of DSBs
generated after gyrase inhibition by Cfx and possibly facilitate removal of covalently bound gyrase, or (iv) recruit
additional factors involved in DNA processing and repair (as proposed for TtAgo [35]).
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
14
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
remains to be established whether the absence of protective activity of SeAgo and
LrAgo in
rec
-minus strains may result from changes in smDNA processing or from
defects in DSB repair caused by inactivation of RecBCD.
The asymmetry of smDNA processing in the
ter
region and around Chi sites on
the chromosomal scale are both increased in the presence of Cfx, likely as a result of
increased levels of gyrase-mediated DNA fragmentation induced by the antibiotic.
Indeed, this increase is primarily explained by stimulation of DSB formation down-
stream of Chi sites at GCSs by Cfx. This may result in increased targeting of the
ter
region and GCSs by guide-loaded pAgos and help to overcome problems with chro-
mosomal DNA processing during replication. In support of this, the stronger effects
of SeAgo on cell growth and morphology in the presence of Cfx correlate with its
higher loading with smDNAs from the
ter
sites and a stronger asymmetry of smDNA
processing around Chi sites in comparison with LrAgo. At the same time, the ability
of SeAgo and LrAgo to suppress the effects of mild gyrase knockdown, which is not
associated with Cfx-induced DNA cleavage, suggests that they may have pleiotropic
effects on DNA processing. Further research is needed to decipher the interplay
between pAgos and the recombination machinery in various species and to eluci-
date the possible roles of pAgos in DNA replication, decatenation, or repair (Fig. 9).
Strikingly, while SeAgo protects
E. coli
cells from the toxic effects of Cfx, it does not pro-
tect its host strain of
S. elongatus
. Moreover, overexpression of SeAgo makes
S. elongatus
more susceptible to this antibiotic. This may be explained by differences in chromosomal
DNA targeting by SeAgo and/or by differences in the mechanism of inhibition of cell division
by Cfx in cyanobacteria and in
E. coli
. Indeed, a different pattern of SeAgo-associated
smDNAs was observed for SeAgo in its native host
S. elongatus
,withnomajorpeaksinthe
predicted termination region (28). While
S. elongatus
encodes several topoisomerases includ-
ing Topo I, Topo II, and gyrase, it lacks recognizable
ter
sites in the chromosome and its
mechanisms of replication termination remain poorly understood (65). Similarly to
T. thermo-
philus
,
S. elongatus
encodes the RecBCD homolog AddAB, but its potential Chi sites have not
been identi
fi
ed in the genome of
S. elongatus
or other cyanobacteria (66, 67). Thus, the spec-
i
fi
city of chromosomal DNA targeting by different pAgos may be different in various species
depending on the mechanisms of DSB processing and replication termination. Furthermore,
cyanobacteria do not have classical SOS response characterized in
E. coli
,and
S. elongatus
apparently lacks its master regulator, the LexA repressor (68, 69). Accordingly, cell
fi
lamenta-
tion was not observed in cyanobacteria (
S. elongatus
or
Synechocystis
)treatedwithCfxor
other DNA-damaging agents (70
–
72).Atthesametime,changesinthesupercoilingstateof
the chromosome caused by gyrase inhibition in cyanobacteria were shown to induce major
changes in the expression of genes involved in adaptation and to impair stress response (70,
73). Thus, by acting as a DNA endonuclease, SeAgo may additionally increase DNA damage
rather than aid DNA processing in
S. elongatus
cells treated with Cfx.
The natural functions of SeAgo in cyanobacterial cells remain to be investigated.
We have previously shown that its deletion or overexpression do not affect the kinetics
of cell growth under laboratory conditions (28). Furthermore, SeAgo does not change
the sensitivity of
S. elongatus
to other DNA-damaging agents tested here. At the same
time, loss of function of SeAgo was shown to increase the ef
fi
ciency of plasmid DNA
transfer in
S. elongatus
(74, 75). SeAgo may therefore modulate natural transformation
in its host bacterium by targeting and processing of foreign DNA. At the same time,
CbAgo was recently shown to stimulate homologous recombination in
E. coli
by induc-
ing DSBs in chromosomal DNA, suggesting that pAgos may promote insertion of ho-
mologous sequences into the chromosome (76, 77). Further research is needed to
understand possible roles of SeAgo in its host species, including natural competence
and homologous recombination. Importantly, our
fi
ndings suggest that the functions
of pAgo proteins may change upon their transfer between prokaryotic species (1, 2, 4)
and may switch between cell protection, regulation of horizontal gene transfer, and
processing and repair of chromosomal DNA.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
15
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
MATERIALS AND METHODS
Plasmids and strains.
All
E. coli
strains were isogenic to
E. coli
BL21(DE3).
recA
-minus and
recBD
-
minus strains were obtained previously (34).
E. coli
cells were routinely cultivated in LB Miller broth (2%
tryptone, 0.5% yeast extract, 1% NaCl, pH 7.0) with the addition of ampicillin (Amp) (100
m
g/mL) if
needed. The gene of SeAgo (Protein database accession number
WP_011378069.1
) was codon-opti-
mized using IDT Codon Optimization Tool for expression in
E. coli
, synthesized by the IDT core facility,
and cloned into pBAD-HisB in frame with an N-terminal His
6
tag. The pBAD-HisB plasmid encoding
LrAgo was obtained previously (27
).
E. coli
strains were transformed with pBAD plasmids encoding each
of the two pAgos or a control plasmid without pAgo, grown overnight in LB with Amp, diluted twice
with 50% glycerol, aliquoted, frozen in liquid nitrogen, and stored at
2
80°C.
Cipro
fl
oxacin treatment and analysis of growth kinetics in
E. coli
.
To determine the range of in-
hibitory concentrations of Cfx and analyze the effects of pAgos on the growth kinetics, overnight bacte-
rial cultures of
E. coli
BL21(DE3) carrying the control pBAD vector or pBAD-encoding pAgo genes were
obtained from the frozen aliquots and inoculated into 200
m
L of fresh LB medium supplemented with a
range of Cfx concentrations (from 62.5 ng/mL to 0.2 ng/mL obtained by 2-fold serial dilutions; 0.5 ng/mL
for comparison of wild-type and mutant strains) or without Cfx, without or with 0.01%
L
-arabinose (Ara),
and with Amp in 96-well plates (TPP;
fl
at bottom). The plates were incubated at 300 rpm at 30°C in a
SPECTROstar Nanomicroplate reader, and cell density was monitored by measuring OD
600
every 10 min.
Three independent biological replicates were performed in each case. Additionally, after 4.5 h, cells were
harvested for Western blot analysis, microscopy, and determination of CFU numbers. To calculate CFU,
serial dilutions of the culture were plated on selective agar medium containing Amp.
Western blotting.
The levels of SeAgo and LrAgo expression were determined by Western blotting.
E. coli
cells expressing pAgos were harvested by centrifugation, the pellet was mixed with 1
Laemmli
sample buffer (120 mM Tris-HCl, 4% SDS, 4%
b
-mercaptoethanol, 10% glycerol, pH 6.8) and heated at
95°C for 5 min, and the samples were resolved by electrophoresis in a 4 to 20% Tris-glycine gel (Bio-
Rad). Proteins were transferred onto a nitrocellulose membrane in Towbin transfer buffer (25 mM Tris,
192 mM glycine, 20% methanol) using semidry procedure at 25 V, 1 A for 30 min (Bio-Rad; Trans-Blot
Turbo). The transfer membrane was washed in phosphate-buffered saline (PBS) (10 mM phosphate
buffer, 137 mM NaCl, 2.7 mM KCl) for 5 min. The membrane was blocked with blocking buffer (PBS,
Tween 20 0.1% [vol/vol], nonfat milk 5% [wt/vol]) for 30 min at room temperature, and then incubated
with anti-His
6
monoclonal antibodies (1:1,000; Sigma) for 1 h at room temperature. The membrane was
washed four times with PBST buffer (PBS, Tween 20 0.1% [vol/vol]) and incubated with horseradish per-
oxidase (HRP)-conjugated anti-mouse secondary antibodies (1:10,000, Sigma) for 1 h at room tempera-
ture and washed again as described above. Antigen-antibody complexes were detected with Immobilon
ECL Ultra Western HRP substrate (Millipore) on a ChemiDoc XRS
1
imager (Bio-Rad).
Preparative growth of
E. coli
and puri
fi
cation and sequencing of pAgo-associated smDNAs.
Overnight bacterial cultures were obtained from frozen aliquoted cultures and inoculated into 500 mL
of fresh LB medium supplemented with 0 or 0.3 ng/mL Cfx, 0.01% Ara, and Amp in 2-L
fl
asks. The
fl
asks
were incubated at 190 rpm at 30°C in an orbital shaker, and cell density was monitored by measuring
OD
600
every 30 min. Two independent biological replicates were performed. The cells were harvested by
centrifugation at 7,000
g
, 4°C for 15 min after 5.5 and 12.5 h of growth for protein pulldown. The cells
were disrupted with a high-pressure homogenizer (EmulsiFlex-C5; Avestin) at 18,000 lb/in
2
. pAgos were
pulled down using Co
2
1
-Talon metal af
fi
nity resin (TaKaRa) as described previously (29). Eluted proteins
were treated with Proteinase K for 30 min at 37°C, and small nucleic acids were extracted with phenol-
chloroform, ethanol precipitated, dissolved in water, and analyzed by PAGE as described (29).
Libraries for high-throughput sequencing of smDNAs were prepared according to the previously
published splinted ligation protocol (29). Brie
fl
y, nucleic acids extracted from pAgos were treated with
RNase A (Thermo Fisher), puri
fi
ed by PAGE, and small DNAs (14 to 20 nt) were eluted from the gel in 0.4
M NaCl overnight at 21°C, ethanol precipitated, dissolved in water, phosphorylated with polynucleotide
kinase (New England Biolabs), and ligated with adaptor oligonucleotides using bridge oligonucleotides
as described in reference 29. The ligated DNA fragments were puri
fi
ed by denaturing PAGE, ampli
fi
ed,
and indexed by the standard protocol for small RNA sequencing (New England Biolabs). Small DNA libra-
ries were sequenced using the HiSeq 2500 platform (Illumina) in the rapid run mode (50-nucleotide sin-
gle-end reads). The list of all sequenced smDNA libraries is presented in Table S1.
Analysis of chromosomal distribution of smDNAs.
Analysis of smDNA sequences was performed
as described previously (34). After trimming the adaptors, reads shorter than 14 nt were removed with
CutAdapt (v. 2.8). Bowtie (v. 1.2.3) was used to align the reads to the reference genomic DNA (RefSeq
accession number
NC_012971.2
) with no mismatches allowed. Nucleotide logos were calculated using
reads longer than 16 nt, and reads longer than 17 nt were truncated to 17 nt from the 3
9
-end. Genome
coverage was calculated using BEDTools (v. 2.27.1) and custom Python scripts. Whole-genome coverage
of each DNA strand was calculated in 1,000-nt windows and normalized by the total number of mapped
reads in the library, expressed as RPKM (reads per kilobase per million aligned reads). To calculate the
percent of reads mapped to the regions around
ter
sites, the number of smDNAs for each DNA strand
from each region (from 1,328,000 to 1,350,000 for
terA
, from 1,524,000 to 1,557,000 for
terC
, and from
1,626,000 to 1,629,000 for
terB
) was divided by the total number of reads mapped to both strand of the
genome. To determine the asymmetry of smDNA distribution between the two genomic strands
depending on Cfx treatment, the ratio between the coverage of the plus and minus strands from the
Cfx-treated cell culture was divided by the same ratio obtained for the control library and plotted as a
rolling mean (50-kb window, 10-kb step).
To calculate smDNA distribution around Chi sites, GCSs, in genes, and intergenic intervals, the region
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
16
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
of replication termination (1.2 to 1.7 Mb chromosomal coordinates) and several other highly enriched
regions, including rRNA operons, were removed from the analysis (see the full list of masked regions in
Table S2 in the supplemental material). Gene borders were extracted from the RefSeq annotation
fi
le.
Intergenic intervals with lengths of
,
500 nt were classi
fi
ed into four groups depending on the relative
orientation of the surrounding gene, and 100 nucleotides were additionally included from each side to
each interval. SmDNA coverage of genes and intergenic intervals was calculated in RPKM. GCSs induced
by Cfx were mapped in the reference genome essentially as described previously using Topo-Seq data
obtained for
E. coli
with a 0.01
P
value threshold for the Audic-Claverie test (41). In total, 1,931 GCSs
were identi
fi
ed in the reference genome, 1,833 of which were considered for further analysis and 98 of
which were mapped in excluded regions. To prepare metaplots with GCSs as anchors, smDNA densities
were extracted for both DNA strands around GCSs (
6
15-kb region) for
1
Cfx and
2
Cfx conditions.
SmDNA densities were averaged position-wise for the
1
Cfx and
2
Cfx data sets independently. The aver-
aged
1
Cfx density was divided by the averaged
2
Cfx density, and the resultant relative density was
smoothed with a 2-kb sliding window. Plots were generated in R (v. 3.6.3) using ggplot2 (v. 3.3.3) and
ggseqlogo (v. 0.1) (78) libraries and in python (v. 3.7) using matplotlib (v. 3.5.1).
Strains of
S. elongatus
, growth conditions, and cipro
fl
oxacin titration.
Wild-type strain of
S. elon-
gatus
PCC 7942 was obtained from Invitrogen. Strains with deletion and overexpression of the
ago
gene
were obtained previously using standard protocols (28). Cyanobacterial strains were maintained in liquid
BG11 with shaking or on solid BG11 plates (79) under constant light conditions (
;
250
m
Em
2
2
s
2
1
)at
30°C with 10
m
g/mL spectinomycin if needed.
Dense cultures of
S. elongatus
wild-type and mutant strains were inoculated into 800
m
L of fresh
BG11 medium supplemented with 0 to 35 ng/mL Cfx and spectinomycin (in the case of the mutant
strains) in 48-well plates (Eppendorf). The plates were incubated at 300 rpm at 30°C under constant light
conditions, and cell density was monitored by measuring OD
750
every 12 h. One biological replicate was
performed for titration and three independent biological replicates were performed for the growth
experiment with 0 and 10 ng/mL Cfx. Cells were harvested for microscopy 48 h after inoculation.
To analyze the effects of 4-NQO and NFZ on cell growth, serial dilutions of the wild-type strain of
S.
elongatus
and strains with deletion of SeAgo and with overexpression of wild-type or catalytically dead
SeAgo were plated on LB agar containing increasing concentrations of 4-NQO and NFZ (from 1 to
10
m
M) and grown for 12 days at 30°C under constant light conditions.
Cell microscopy.
E. coli
cells were visualized using acridine orange staining. Sterilized slides were
fi
xed with 95% ethanol for 2 min, excess ethanol was drained, and the slides were allowed to air dry.
Bacterial culture was placed onto a slide, dried, and brie
fl
y
fi
xed in the
fl
ame of a burner. The slide was
fl
ooded with acridine orange stain for 2 min, rinsed thoroughly with tap water, and allowed to air dry.
Cell pictures were taken on a ZEISS LSM 900 confocal laser scanning microscope (Carl Zeiss) with a 63
oil immersion objective and with 100
m
m confocal pinhole aperture. Cell pictures of
S. elongatus
were
taken similarly but using chlorophyll auto
fl
uorescence, which was excited at 488 nm and recorded using
a 650-nm long-pass
fi
lter. The obtained pictures were processed using the ZEN Microscopy software
(Carl Zeiss).
Gyrase knockdown with dCas9.
To knockdown the
gyrA
gene, a plasmid carrying the p15A origin,
chloramphenicol resistance and dCas9 genes, and sgRNA was obtained by Gibson assembly. The 20-nt
guide sequence in sgRNA corresponding to the
gyrA
gene (5
9
-AGCTCTTCCTCAATGTTGAC-3
9
) was chosen
based on the algorithm developed in reference 80.
E. coli
Bl21(DE3) strains containing pBAD plasmids
encoding or lacking pAgos were transformed with the dCas9 plasmid and grown under standard condi-
tions with the addition of 0.01% Ara, Amp, and chloramphenicol. The drop of the
gyrA
expression level
was measured by quantitative reverse transcription-PCR using primers corresponding to the
gyrA
gene
(qPCR_gyrA_for 5
9
-TTATGACACGATCGTCCGTATG and qPCR_gyrA_rev 5
9
-TTCCGTGCCGTCATAGTTATC)
and the
rpoB
gene for normalization (rpoB_Fw1 5
9
-ATGGTTTACTCCTATACCGAGAAAAAAC and rpoB_Rv1
5
9
-TATTGCAGCTCGGAATTACCG).
Data availability.
The smDNA sequencing data sets generated in this study are available from the
Sequence Read Archive (SRA) database under BioProject number
PRJNA878808
. The code used for data
analysis is available at the GitHub repository at
https://github.com/AlekseiAgapov/SeAgo_LrAgo
. All pri-
mary data are available from the corresponding author upon request.
SUPPLEMENTAL MATERIAL
Supplemental material is available online only.
SUPPLEMENTAL FILE 1
, PDF
fi
le, 6 MB.
ACKNOWLEDGMENTS
We thank Daria Esyunina for continued advice on this study and Phillip Zamore and
Samson M. Jolly for helpful discussions.
This work was supported by the Russian Science Foundation (grant 19-14-00359 to
Daria Esyunina, analysis of smDNA biogenesis and the effects of Cfx on DNA damage;
grant 20-74-10127 to A.A., analysis of the effects of pAgos on replication termination).
D.S. and A.G. were supported by a grant from the Ministry of Science and Higher
Education of Russian Federation (agreement 075-10-2021-114; analysis of gyrase-
induced DNA cleavage).
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
17
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.
We declare no competing interests.
Andrey Kulbachinskiy and Alexei A. Aravin conceived the study; Anna Olina performed
experiments; Alina Galivondzhyan analyzed cell growth in the presence of Cfx; Aleksei
Agapov and Dmitry Sutormin performed bioinformatic analysis of small DNA libraries; Denis
Yudin made the original discovery of the chromosomal speci
fi
city of pAgos under
supervision of Anton Kuzmenko and performed initial analysis of small DNAs; Anna Olina,
Aleksei Agapov, and Dmitry Sutormin prepared the
fi
gures; and Andrey Kulbachinskiy and
Dmitry Sutormin wrote the manuscript with contributions from all the authors.
REFERENCES
1. Makarova KS, Wolf YI, van der Oost J, Koonin EV. 2009. Prokaryotic homo-
logs of Argonaute proteins are predicted to function as key components
of a novel system of defense against mobile genetic elements. Biol Direct
4:29.
https://doi.org/10.1186/1745-6150-4-29
.
2. Swarts DC, Makarova K, Wang Y, Nakanishi K, Ketting RF, Koonin EV, Patel
DJ, van der Oost J. 2014. The evolutionary journey of Argonaute proteins.
Nat Struct Mol Biol 21:743
–
753.
https://doi.org/10.1038/nsmb.2879
.
3. Hall TM. 2005. Structure and function of argonaute proteins. Structure 13:
1403
–
1408.
https://doi.org/10.1016/j.str.2005.08.005
.
4. Burroughs AM, Ando Y, Aravind L. 2014. New perspectives on the diversi
fi
ca-
tion of the RNA interference system: insights from comparative genomics
and small RNA sequencing. Wiley Interdiscip Rev RNA 5:141
–
181.
https://doi
.org/10.1002/wrna.1210
.
5. Ghildiyal M, Zamore PD. 2009. Small silencing RNAs: an expanding uni-
verse. Nat Rev Genet 10:94
–
108.
https://doi.org/10.1038/nrg2504
.
6. Hammond SM, Boettcher S, Caudy AA, Kobayashi R, Hannon GJ. 2001.
Argonaute2, a link between genetic and biochemical analyses of RNAi.
Science 293:1146
–
1150.
https://doi.org/10.1126/science.1064023
.
7. Martinez J, Patkaniowska A, Urlaub H, Luhrmann R, Tuschl T. 2002. Single-
stranded antisense siRNAs guide target RNA cleavage in RNAi. Cell 110:
563
–
574.
https://doi.org/10.1016/s0092-8674(02)00908-x
.
8. Meister G, Landthaler M, Patkaniowska A, Dorsett Y, Teng G, Tuschl T. 2004.
Human Argonaute2 mediates RNA cleavage targeted by miRNAs and siRNAs.
Mol Cell 15:185
–
197.
https://doi.org/10.1016/j.molcel.2004.07.007
.
9. Joshua-Tor L. 2006. The Argonautes. Cold Spring Harbor Symp Quant Biol
71:67
–
72.
https://doi.org/10.1101/sqb.2006.71.048
.
10. Peters L, Meister G. 2007. Argonaute proteins: mediators of RNA silencing.
Mol Cell 26:611
–
623.
https://doi.org/10.1016/j.molcel.2007.05.001
.
11. Pratt AJ, MacRae IJ. 2009. The RNA-induced silencing complex: a versatile
gene-silencing machine. J Biol Chem 284:17897
–
17901.
https://doi.org/
10.1074/jbc.R900012200
.
12. Olina AV, Kulbachinskiy AV, Aravin AA, Esyunina DM. 2018. Argonaute pro-
teins and mechanisms of RNA interference in eukaryotes and prokaryotes. Bio-
chemistry (Mosc) 83:483
–
497.
https://doi.org/10.1134/S0006297918050024
.
13. Parker JS, Roe SM, Barford D. 2004. Crystal structure of a PIWI protein sug-
gests mechanisms for siRNA recognition and slicer activity. EMBO J 23:
4727
–
4737.
https://doi.org/10.1038/sj.emboj.7600488
.
14. Song JJ, Smith SK, Hannon GJ, Joshua-Tor L. 2004. Crystal structure of
Argonaute and its implications for RISC slicer activity. Science 305:
1434
–
1437.
https://doi.org/10.1126/science.1102514
.
15. Ma JB, Yuan YR, Meister G, Pei Y, Tuschl T, Patel DJ. 2005. Structural basis
for 5'-end-speci
fi
c recognition of guide RNA by the A. fulgidus Piwi pro-
tein. Nature 434:666
–
670.
https://doi.org/10.1038/nature03514
.
16. Parker JS, Roe SM, Barford D. 2005. Structural insights into mRNA recogni-
tion from a PIWI domain-siRNA guide complex. Nature 434:663
–
666.
https://doi.org/10.1038/nature03462
.
17. Yuan YR, Pei Y, Ma JB, Kuryavyi V, Zhadina M, Meister G, Chen HY, Dauter
Z, Tuschl T, Patel DJ. 2005. Crystal structure of A. aeolicus argonaute, a
site-speci
fi
c DNA-guided endoribonuclease, provides insights into RISC-
mediated mRNA cleavage. Mol Cell 19:405
–
419.
https://doi.org/10.1016/j
.molcel.2005.07.011
.
18. Wang Y, Juranek S, Li H, Sheng G, Tuschl T, Patel DJ. 2008. Structure of an
argonaute silencing complex with a seed-containing guide DNA and target
RNA duplex. Nature 456:921
–
926.
https://doi.org/10.1038/nature07666
.
19. Wang Y, Juranek S, Li H, Sheng G, Wardle GS, Tuschl T, Patel DJ. 2009.
Nucleation, propagation and cleavage of target RNAs in Ago silencing
complexes. Nature 461:754
–
761.
https://doi.org/10.1038/nature08434
.
20. Ryazansky S, Kulbachinskiy A, Aravin AA. 2018. The expanded universe of
prokaryotic Argonaute proteins. mBio 9:e01935-18.
https://doi.org/10
.1128/mBio.01935-18
.
21. Willkomm S, Makarova K, Grohmann D. 2018. DNA-silencing by prokary-
otic Argonaute proteins adds a new layer of defence against invading
nucleic acids. FEMS Microbiol Rev 42:376
–
387.
https://doi.org/10.1093/
femsre/fuy010
.
22. Li W, Liu Y, He R, Wang L, Wang Y, Zeng W, Zhang Z, Wang F, Ma L. 2022.
A programmable pAgo nuclease with RNA target preference from the
psychrotolerant bacterium Mucilaginibacter paludis. Nucleic Acids Res
50:5226
–
5238.
https://doi.org/10.1093/nar/gkac315
.
23. Lisitskaya L, Shin Y, Agapov A, Olina A, Kropocheva E, Ryazansky S, Aravin
AA, Esyunina D, Murakami KS, Kulbachinskiy A. 2022. Programmable RNA
targeting by bacterial Argonaute nucleases with unconventional guide
binding and cleavage speci
fi
city. Nat Commun 13:4624.
https://doi.org/
10.1038/s41467-022-32079-5
.
24. Cao Y, Sun W, Wang J, Sheng G, Xiang G, Zhang T, Shi W, Li C, Wang Y, Zhao
F, Wang H. 2019. Argonaute proteins from human gastrointestinal bacteria
catalyze DNA-guided cleavage of single- and double-stranded DNA at 37°C.
Cell Discov 5:38.
https://doi.org/10.1038/s41421-019-0105-y
.
25. Hegge JW, Swarts DC, Chandradoss SD, Cui TJ, Kneppers J, Jinek M, Joo C, van
der Oost J. 2019. DNA-guided DNA cleavage at moderate temperatures by
Clostridium butyricum Argonaute. Nucleic Acids Res 47:5809
–
5821.
https://
doi.org/10.1093/nar/gkz306
.
26. Kaya E, Doxzen KW, Knoll KR, Wilson RC, Strutt SC, Kranzusch PJ, Doudna
JA. 2016. A bacterial Argonaute with noncanonical guide RNA speci
fi
city.
Proc Natl Acad Sci U S A 113:4057
–
4062.
https://doi.org/10.1073/pnas
.1524385113
.
27. Kuzmenko A, Yudin D, Ryazansky S, Kulbachinskiy A, Aravin AA. 2019. Pro-
grammable DNA cleavage by Ago nucleases from mesophilic bacteria Clos-
tridium butyricum and Limnothrix rosea. Nucleic Acids Res 47:5822
–
5836.
https://doi.org/10.1093/nar/gkz379
.
28. Olina A, Kuzmenko A, Ninova M, Aravin AA, Kulbachinskiy A, Esyunina D.
2020. Genome-wide DNA sampling by Ago nuclease from the cyanobac-
terium Synechococcus elongatus. RNA Biol 17:677
–
688.
https://doi.org/
10.1080/15476286.2020.1724716
.
29. Olovnikov I, Chan K, Sachidanandam R, Newman DK, Aravin AA. 2013.
Bacterial argonaute samples the transcriptome to identify foreign DNA.
Mol Cell 51:594
–
605.
https://doi.org/10.1016/j.molcel.2013.08.014
.
30. Sheng G, Zhao H, Wang J, Rao Y, Tian W, Swarts DC, van der Oost J, Patel DJ,
Wang Y. 2014. Structure-based cleavage mechanism of Thermus thermophi-
lus Argonaute DNA guide strand-mediated DNA target cleavage. Proc Natl
Acad Sci U S A 111:652
–
657.
https://doi.org/10.1073/pnas.1321032111
.
31. SwartsDC,HeggeJW,HinojoI,ShiimoriM,EllisMA,DumrongkulraksaJ,Terns
RM, Terns MP, van der Oost J. 2015. Argonaute of the archaeon Pyrococcus
furiosus is a DNA-guided nuclease that targets cognate DNA. Nucleic Acids
Res 43:5120
–
5129.
https://doi.org/10.1093/nar/gkv415
.
32. Swarts DC, Jore MM, Westra ER, Zhu Y, Janssen JH, Snijders AP, Wang Y,
Patel DJ, Berenguer J, Brouns SJJ, van der Oost J. 2014. DNA-guided DNA
interference by a prokaryotic Argonaute. Nature 507:258
–
261.
https://doi
.org/10.1038/nature12971
.
33. Zander A, Willkomm S, Ofer S, van Wolferen M, Egert L, Buchmeier S, Stockl S,
Tinnefeld P, Schneider S, Klingl A, Albers SV, Werner F, Grohmann D. 2017.
Guide-independent DNA cleavage by archaeal Argonaute from Methanocaldo-
coccus jannaschii. Nat Microbiol 2:17034.
https://doi.org/10.1038/nmicrobiol
.2017.34
.
34. Kuzmenko A, Oguienko A, Esyunina D, Yudin D, Petrova M, Kudinova A,
Maslova O, Ninova M, Ryazansky S, Leach D, Aravin AA, Kulbachinskiy A.
Chromosome Targeting by Argonautes Aids Cell Division
Microbiology Spectrum
May/June 2023 Volume 11 Issue 3
10.1128/spectrum.04146-22
18
Downloaded from https://journals.asm.org/journal/spectrum on 20 June 2023 by 131.215.220.163.