The telomerase reverse transcriptase elongates reversed replication forks at telomeric repeats

Abstract

The telomerase reverse transcriptase elongates telomeres to prevent replicative senescence. This process requires exposure of the 3'-end, which is thought to occur when two sister telomeres are generated at replication completion. Using two-dimensional agarose gel electrophoresis (2D-gels) and electron microscopy, we found that telomeric repeats are hotspots for replication fork reversal. Fork reversal generates 3' telomeric ends before replication completion. To verify whether these ends are elongated by telomerase, we probed de novo telomeric synthesis in situ and at replication intermediates by reconstituting mutant telomerase that adds a variant telomere sequence. We found variant telomeric repeats overlapping with telomeric reversed forks in 2D-gels, but not with normal forks, nontelomeric reversed forks, or telomeric reversed forks with a C-rich 3'-end. Our results define reversed telomeric forks as a substrate of telomerase during replication.

Full text

GENETICS
The telomerase reverse transcriptase elongates
reversed replication forks at telomeric repeats
Armela Huda
1
, Hiroshi Arakawa
1
, Giulia Mazzucco
1
, Martina Galli
1
, Valentina Petrocelli
2
,
Stefano Casola
1
, Lu Chen
3,4
, Ylli Doksani
1
*
The telomerase reverse transcriptase elongates telomeres to prevent replicative senescence. This process re-
quires exposure of the 3′-end, which is thought to occur when two sister telomeres are generated at replication
completion. Using two-dimensional agarose gel electrophoresis (2D-gels) and electron microscopy, we found
that telomeric repeats are hotspots for replication fork reversal. Fork reversal generates 3′telomeric ends before
replication completion. Toverify whether these ends are elongated by telomerase, we probed de novo telomeric
synthesis in situ and at replication intermediates by reconstituting mutant telomerase that adds a variant telo-
mere sequence. We found variant telomeric repeats overlapping with telomeric reversed forks in 2D-gels, but
not with normal forks, nontelomeric reversed forks, or telomeric reversed forks with a C-rich 3′-end. Our results
define reversed telomeric forks as a substrate of telomerase during replication.
Copyright © 2023 The
Authors, some rights
reserved; exclusive
licensee American
Association for the
Advancement of Science.
No claim to original U.S.
Government Works.
Distributed under a
Creative Commons
Attribution
NonCommercial License
4.0 (CC BY-NC).
INTRODUCTION
Linear chromosomes end in several kilobases of telomeric repeats,
bound by the shelterin complex, which prevents activation of a
DNA damage response at chromosome ends (1). Telomere function
is essential for cellular proliferation and chromosome stability.
Gradual erosion of telomeres is counteracted by the telomerase
reverse transcriptase (TERT), which uses an internal RNA template
(TR) to add telomeric repeats to the 3′-ends of chromosomes (2).
Telomere maintenance, however, depends also on faithful duplica-
tion of the bulk of repeats by the replication machinery every S
phase. Several lines of evidence suggest that this task is challenging.
Telomeres behave like replication fragile sites and constitutively
require the action of specialized helicases to assist replication (3–
7). In their absence, large fragments of telomeric repeats can be
lost in a single S phase, likely due to replication failure (3,6). The
severe telomere replication defects observed in shelterin mutants,
from yeast to mammals, suggest that shelterin plays an essential
function in assisting semiconservative replication of telomeric
repeats (4,8,9). Several telomeric features have been invoked as po-
tential obstacles to fork progression, including the presence of
tightly bound proteins, ongoing transcription, or secondary struc-
tures like G4-DNA, t-loops, R-loops, and damage-induced i-loops
(6,10–15). However, the relative contribution of each of these
factors to telomere replication stress is not clear. Electron microsco-
py (EM) analysis of model replication forks has shown that the pres-
ence of TTAGGG repeats induces spontaneous fork reversal in vitro
(16). Two-dimensional agarose gel electrophoresis (2D-gels) has re-
vealed that replication fork pausing at yeast telomeres and fork col-
lapse events have been observed in mutants with telomere
dysfunction (8,17–19). In taz1 mutants in Schizosaccharomyces
pombe, fork collapse at telomeric repeats is associated with an in-
crease recruitment and activity of telomerase, suggesting that col-
lapsed telomeric forks might represent a robust substrate for
telomerase (20). In these settings, it was found that telomerase
played an important role in restoring telomere length after replica-
tion failures by synthesizing long stretches of telomeres or, possibly,
by acting directly on collapsed replication forks (8,19,20). While in
yeast telomerase activity seems to alleviate the consequences of telo-
mere replication stress, genetic analysis of rapid telomere deletion
events in mouse cells has suggested a different scenario, where tel-
omerase loading on reversed telomeric forks would lead to replica-
tion failures, a pathological activity that is kept at bay by the Rtel1
helicase (21). Replication stress at mammalian telomeres is often
monitored through the scars it leaves on telomere fluorescence in
situ hybridization (FISH) signals on chromosome spreads. Pheno-
types like fragile (decondensed) telomeres, signal heterogeneity, or
telomere loss are indicative of replication failures but provide no
molecular detail on the fate of telomeric forks. Differently from
yeast, 2D-gel analysis of mammalian telomeric forks has not been
possible so far and, overall, there is a lack of structural information
on the fate of telomeric forks in mammalian cells.
To overcome these limitations, we have monitored by 2D-gels
and EM the structure of replication forks traveling through long
stretches of telomeric repeats, introduced into a Simian virus 40
(SV40) mini-chromosome. We found that replication of telomeric
repeats is characterized by an increased frequency of replication fork
reversal, which occurred in a length-dependent but orientation-in-
dependent fashion. In agreement with this result, EM analysis of
enriched mouse telomeres revealed accumulation of reversed
forks also at endogenous telomeric repeats, when compared with
the bulk DNA of the same sample. To probe the activity of telome-
rase at stalled telomeric forks, we expressed a mutated telomerase
TR, hTR-TSQ1 (tolerated sequence 1), which introduces
GTTGCG repeats, in cells with the SV40 mini-chromosome con-
taining telomeric repeats (22). 2D-gel analysis, combined with spe-
cific probes, revealed the addition of TSQ1 repeats specifically on
1
IFOM ETS-The AIRC Institute of Molecular Oncology, Milan, Italy.
2
Institute for
Tumor Biology and Experimental Therapy, Frankfurt Cancer Institute, Goethe Uni-
versity Frankfurt, Frankfurt am Main, Germany.
3
Nuclear Dynamics and Cancer
Program, Cancer Epigenetics Institute, Fox Chase Cancer Center, Philadelphia,
PA, USA
4
Department of Cancer and Cellular Biology, Lewis Katz School of Medi-
cine, Temple University, Philadelphia, PA, USA.
*Corresponding author. Email: ylli.doksani@ifom.eu
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 1 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

molecules that migrate as reversed forks, in the cone signal. No
signal was detected at normal replication forks, at nontelomeric re-
versed forks, or at telomeric reversed forks with a C-rich 3′-end.
Our results estimate a nearly twofold increase in the probability
of replication fork reversal at telomeric repeats compared to the
rest of the genome. We show that reversed telomeric forks generate
a 3′-end that is elongated by the TERT in vivo. These results reveal
another substrate for telomerase, which is generated before the
completion of chromosome replication. Our results also suggest
the existence of replisome-associated factors that control telomerase
activity during telomere replication, similar to what has been pro-
posed for Rtel1 (21). Frequent fork reversal would result in the gen-
eration of free telomeric ends during replication that, if left
unresolved or unprotected, could be engaged in aberrant DNA
repair, resulting in telomere dysfunction and genome instability.
RESULTS
Replication fork reversal at telomeric repeats in an SV40
mini-chromosome
To monitor the structure of replication forks traveling through te-
lomeric repeats, we used an SV40 mini-chromosome–based system,
previously used to study other repetitive elements in human cells
(23). A stretch of 115 telomeric repeats was inserted next to the
SV40 origin of replication in both orientations (Fig. 1A). The con-
structs were transfected in human embryonic kidney (HEK)–293T
cells where they undergo replication, and 40 hours after transfec-
tion, the SV40 mini-chromosomes were extracted following precip-
itation of intact genomic DNA in high salt, according to the Hirt
procedure (24). These constructs are chromatinized in vivo (23),
and consistent with previous work showing binding of shelterin
components on telomeric repeats in episomal constructs (25), we
found that TRF1 is recruited at the telomeric repeats on the SV40
mini-chromosome (Fig. 1B and fig. S1A). When equal amounts of a
construct with 115 telomeric repeats and a nontelomeric control of
equal length were cotransfected and allowed to replicate, we recov-
ered around 30% less of the former, suggesting an increased inci-
dence of replication failure at telomeric repeats (fig. S1, B and C).
We monitored by neutral 2D-gel electrophoresis the structure of
replication forks arising from the SV40 origin and moving
through the telomeric repeats. To prevent branch migration of rep-
lication intermediates, the DNA was psoralen–cross-linked in vivo,
before extraction (26). A 5.26-kb Bam HI–Sac I restriction fragment,
containing the SV40 origin and the telomeric insert, was separated
Fig. 1. Replication fork reversal at telomeric repeats on the SV40 mini-chromosome. (A) Schematic representation of the SV40 constructs used in this study. The Bam
HI–Sac I restriction sites are shown. (B) Chromatin immunoprecipitation (ChIP) analysis showing loading of the shelterin component TRF1 to ectopic telomeric repeats on
the pTelN construct (see also fig. S1A). HEK-293T cells were transfected with the indicated constructs, allowed to replicate, collected 40 hours after transfection, and
processed for ChIP with an anti-hTRF1 antibody. Bars show mean with SD from three independent experiments. The Pvalue was derived from an unpaired, two-tailed
Student’sttest. (C) 2D-gels showing the accumulation of the cone signal in the constructs with telomeric repeats (see also fig. S1D). HEK-293T cells were transfected with
the constructs indicated in (A), allowed to replicate, collected 40 hours after transfection, and psoralen–cross-linked in vivo. Then, extra-chromosomal DNA was extracted
with the Hirt procedure; digested with Bam HI, Sac I, and Dpn I (to remove unreplicated plasmids); and separated in 2D-gels. The gels were blotted into a membrane and
hybridized with a
32
P-labeled Bam HI–Sac I 4.53-kb fragment from the pML113 construct (backbone) (top); then, the membranewas stripped and hybridized with a probe
recognizing the telomeric repeats (telomeric) (bottom). (D) Schematic representation of the 2D-gel migration profile of the indicated DNA replication intermediates. Note
that the distinct pattern, generated by each class of intermediates, is color-coded. (E) Quantification of the cone signal expressed as a percentage of all replication
intermediates from three to five independent experiments as the one shown in (C). Bars show mean with SD. Pvalues derived from unpaired, two-tailed Student’sttest.
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 2 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

in 2D-gels (Fig. 1, A, C, and D). All constructs accumulated bubbles
and large Y-s, indicating that the nearby telomeric repeats do not
affect SV40 origin activity (Fig. 1, C and D). No obvious pausing
spots were visible in 2D-gels in the fragment with the telomeric
repeats (see below). The telomeric sample showed a more intense
cone signal (Fig. 1, C to E). This signal is generated by X-shaped
DNA structures associated with replication fork reversal (27–29).
While the cone signal accumulated in the construct with 115 telo-
meric repeats, constructs with 34 or 54 repeats did not show a con-
sistent increase in cone signal intensity (fig. S1, D and E). The cone
signal accumulated both in the pTelN construct (where the G-rich
parental strand is replicated as the lagging strand) and in the pTelR
construct with reversed orientation (where the C-rich parental
strand is replicated as the lagging strand), arguing against the for-
mation of G4-DNA as the main cause of fork reversal in this setting
(Fig. 1, C to E). Consistently, treatment with the G4 ligand pyridos-
tatin did not result in a furtheraccumulation of reversed forks in the
sample with the telomeric repeats (fig. S1, F and G). These results
suggest an increased incidence of replication fork reversal at telo-
meric repeats in the SV40 mini-chromosome, which occurs in a
length-dependent and orientation-independent manner.
To corroborate these findings, replication forks within the same
5.26-kb Bam HI–Sac I restriction fragment analyzed above were
spread and visualized by EM. For this type of analysis, replicating
Fig. 2. Fork reversal at the ectopic telomeric repeats, visualized by electron microscopy. (A) Representative EM images of normal replication bubbles (top), rep-
lication bubbles with fork reversal, and an X-shaped intermediate compatible with fork reversal, in the pTelN construct. The DNA, derived from an identical experimental
procedure as the one in Fig. 1C, was digested with Bam HI and Sac I, and the replication intermediates were purified from an agarose gel (see also fig. S2). The samples
were spread and visualized in EM. A scheme of the molecule visualized is shown above each image. Inset shows a magnification of the area inside the red square. Scale
bars, 180 nm (0.5 kb). (B) Quantification of the percentage of molecules compatible with a reversed fork structure, over all replication forks from three independent
experiments as the one shown in (A) for pTelN and pCtrl and one experiment for pTelR. N= 157, 60, and 320 and 124, 56, and 166 replication forks for pTelN and
pCtrl, respectively, and 76 forks for pTelR. Bars represent mean with SD. The Pvalue was derived from unpaired, two-tailed Student’sttest.
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 3 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

molecules were enriched by separating them from the linear form
on a preparative agarose gel (fig. S2A) and only molecules with a
size compatible with the Bam HI–Sac I restriction fragment
(>60% of the replication intermediates isolated from the gel) were
analyzed (fig. S2B). Normal forks were visualized either within a
replication bubble or as three-way junctions. Events of replication
fork reversal were visible either within a replication bubble or as
X-shaped (four-way) DNA structures with two arms of equal
length (Fig. 2A). Reversed forks were observed in both the pTelN,
pTelR, and pCtrl fragment of matching length; however, the pres-
ence of telomeric repeats was associated with an over twofold in-
crease in reversed replication forks (Fig. 2B). This result is
consistent with the increase in cone signal intensity observed in
2D-gels and with previous EM analysis of model telomeric
forks (16).
In the circular SV40 constructs, the two diverging forks originat-
ing from the SV40 origin typically travel a similar distance and con-
verge on the opposite site of the origin (Fig. 3A). 2D-gel analysis of
the other restriction fragment generated by the digestion—the 3.71-
kb Sac I–Bam HI fragment—showed a strong termination signal, in
both the control and pTelN construct (Fig. 3B). As expected, in EM
analysis, the 5.26-kb fragment containing the SV40 origin showed
very rare double Y-shaped termination intermediates; however, the
presence of the telomeric repeats leads to a significant increase of
termination intermediates detected in this fragment (Fig. 3, C and
D). This result is consistent with an increased stalling frequency of
the clockwise fork traveling through the telomeric repeats and con-
sequent entry of the counterclockwise fork in the restriction frag-
ment (Fig. 3A).
Fig. 3. Fork stalling and termination at the telomeric repeats in the SV40 construct. (A) Schematic representation of two possible scenarios of replicationtermination
in the SV40 construct. The blue tract indicates the 5.26-kb Bam HI–Sac I restriction fragment containing the SV40 origin, and the yellow tract indicates the 3.71-kb Sac
I–Bam HI fragment, opposite to the origin. The red tract indicates the telomeric repeats. Normally, the two replication forks travel at similar distances to merge in the
yellow region (pCtrl). If the clockwise fork stalls in the telomeric repeats, termination might occur also in the blue tract (pTelN). (B) 2D-gel analysis showing that the bulk of
termination in the SV40 constructs (in both pCtrl and pTelN) occurs in the 3.71-kb Sac I–Bam HI fragment as expected [the yellow tract in the scheme in (A)]. DNA from the
same experimental procedure, described in Fig. 1C, was separated in 2D-gels and hybridized with a
32
P-labeled Sac I–Bam HI 3.71-kb fragment from the pML113 con-
struct. (C) Example of a double Y-shaped termination intermediate, visualized in the 5.26-kb fragment, containing the replication origin and the telomeric repeats from
the pTelN construct. A scheme of the molecule visualized is shownabove the image. The inset represents a magnification of the area inside the red square. Scale bar, 180
nm (0.5 kb). (D) Quantification of the fraction of termination intermediates visualized in the 5.26-kb Bam HI–Sac I fragment, from the three independent experiments
described in Fig. 2B. Bars represent mean with SD. The Pvalue was derived from unpaired, two-tailed Student’sttest.
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 4 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

Fig. 4. Increased incidence of reversed forks at mouse telomeres. (A) Dot blot analysis showing the enrichment of telomeric repeats. Telomeres were enriched
through two restriction digestion and size exclusion steps as described in (31). After each enrichment step, the amounts of DNA indicated in the figure were spotted
on a membrane and hybridized with a
32
P-labeled probe recognizing telomeric repeats. The amount of TTAGGG repeat signal per nanogram was quantified and reported
relative to the signal per nanogram value in the initial, nonenriched, bulk DNA. (B) Quantification of the percentage of molecules compatible with a reversed fork struc-
ture, over all replication forks from three independent experiments, where either telomere-enriched or Kpn I–digested bulk DNA was spread and visualized by EM. (A)
N= 81, 74, and 144 and 64, 72, and 130 replication forks for telomeric and bulk samples, respectively. Bars represent mean with SD. The Pvalue was derived from unpaired,
two-tailed Student’sttest. (C) Examples of normal replication forks from atelomere-enriched sample. (D) Examples of molecules compatible with a reversed fork structure
from a telomere-enriched sample. Insets represent magnifications of the area inside the red square. Scale bars, 360 nm (1 kb).
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 5 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

Increased incidence of replication fork reversal at mouse
telomeres
Prompted by the results obtained at the ectopic telomeric repeats in
the SV40 system, we asked if the increased frequency of replication
fork reversal is observed also at endogenous telomeres. To address
this point, we decided to analyze mouse telomeres by EM. Because
this type of analysis is limited by the paucity of replication interme-
diates in the final sample, we exploited a rapidly dividing mouse B
cell lymphoma line established from the lambda-MYC transgenic
model with long telomeres (30). Telomeres from this cell line
were enriched through two successive rounds of restriction diges-
tion with frequent cutters, followed by size fractionation (Fig. 4A)
(31). For EM analysis, Y-shaped molecules (replication forks) and
X-shaped intermediates with two equal arms (compatible with fork
reversal) were quantified in parallel in the nonenriched (bulk) DNA
and after telomere enrichment (telomeric) (Fig. 4, B to D). In these
experiments, telomere-enriched samples showed a twofold increase
in the fraction of reversed forks, compared to the respective bulk
DNA (Fig. 4B). These results were observed in the absence of any
exogenously induced replication stress and show that, consistently
with the replication of ectopic telomeric repeats, replication of en-
dogenous telomeres is also associated with an increased probability
of replication fork reversal.
Extension of reversed telomeric forks by the TERT in vivo
Several studies have suggested that reversed telomeric forks could
become a substrate for the TERT (19–21). However, direct evidence
for this activity is missing, also due to the inability to differentiate it
from postreplicative elongation of telomeric double-strand breaks.
We sought to take advantage of the SV40 mini-chromosome,
bearing a long stretch of telomeric repeats, to address this question
directly. This system offers two main advantages: First, it does not
contain a natural telomeric end that would provide a substrate for
telomerase, and second, its replication intermediates are readily vi-
sualized in 2D-gels, which could allow detection of telomerase ac-
tivity on specific telomeric structures (see below). To label
specifically telomerase extension events, we took advantage of a
mutated version of the telomerase TR, hTR-TSQ1 (tolerated se-
quence 1), which introduces a GTTGCG motif at telomeres (22).
As expected, ectopic expression of the TERT (hTERT) together
with a mutated hTR-TSQ1 template in 293T cells leads to the addi-
tion of the GTTGCG motif at telomeres, detected by peptide nucleic
acid (PNA)–FISH and, in telomere blots, hybridized with a TSQ1-
specific probe (fig. S3, A and B). We then monitored, under the
same conditions, telomerase activity at the telomeric replication in-
termediates in the SV40 mini-chromosome. The pCtrl or pTelN
constructs were transfected together with hTERT and hTR-TSQ1
and allowed to replicate, and then the 5.26-kb Bam HI–Sac I frag-
ment was separated in 2D-gels as described above. The membranes
were hybridized sequentially, first with a probe recognizing the
TSQ1 repeats and then with another against the entire Bam
HI–Sac I fragment (backbone). No TSQ1 signal was detected in
the hTR-WT controls, or when hTR-TSQ1 was expressed in cells
replicating the pCtrl construct (Fig. 5A). However, when hTR-
TSQ1 was expressed in cells replicating the pTelN construct, a
clear TSQ1 signal was visible in 2D-gels and the TSQ1 signal (de-
riving from telomerase activity) overlapped in 2D-gels with the cone
signal that is generated by telomeric reversed forks (Fig. 5A and fig.
S4A). Consistent with this observation, pTelN constructs with 34 or
Fig. 5. Extension of reversed telomeric forks by telomerase. (A) 2D-gel exper-
iments showing detection of the telomerase-dependent addition of TSQ1 repeats
at the cone signal, where reversed forks migrate in 2D-gels. HEK-293T cells were
transfected with the indicated constructs, and 40 hours after transfection, cells
were collected and psoralen–cross-linked in vivo. Extra-chromosomal DNA was ex-
tracted with the Hirt procedure; digested with Bam HI, Sac I, and Dpn I; and sep-
arated in 2D-gels. The gels were blotted onto a membrane and hybridized with a
32
P-labeled (GCAACC)4 (TSQ1) oligo to detect telomerase extension events. Then,
the membrane was stripped and hybridized with a
32
P-labeled Bam HI–Sac I 4.53-
kb fragment from the pML113 construct (Backbone). To facilitate visualization, the
images were false-colored and merged (bottom). (B) Schematic representation
showing that the sequence of the 3′-end of the reversed arm will depend on
the orientation of the telomeric repeats with respect to the clockwise moving
fork. (C) 2D-gels showing that the telomerase-mediated extension of reversed
forks depends on the orientation of the telomeric repeats. The same experimental
procedure as the one described in (A) was performed, including the construct
pTelR with the inverted orientation.
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 6 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

54 telomeric repeats (that have lower levels of fork reversal) showed
the same pattern, but a fainter TSQ1 probe signal (fig. S4B). This
result shows that telomerase can extend reversed telomeric forks
in vivo.
We considered the possibilities that the TSQ1 sequence at the
cone signal might derive from extension of telomeric double
strand breaks (DSBs), which are then engaged in homologous re-
combination, and/or that recruitment of telomerase on ectopic te-
lomeric repeats might lead to unspecific or spurious activities on
telomeric reversed forks. To address these points, we repeated the
experiment including a pTelR construct, which is identical to
pTelN, except that the orientation of the telomeric repeats, with
respect to the SV40 origin, is reversed. While this construct
should be identical to pTelN in terms of telomerase recruitment
and eventual extension of DSBs formed during the replication of
telomeric repeats, reversal of the clockwise fork in the pTelR con-
struct will generate a C-rich 3′-end that should not be targeted by
telomerase (Fig. 5B). While a strong cone signal was detected by the
TSQ1 probe in the pTelN construct, no cone signal was detected by
the TSQ1 probe in the pTelR construct (Fig. 5C and fig. S4C).
Therefore, the results obtained in 2D-gels cannot be attributed to
an unspecific activity of the telomerase recruited at the ectopic telo-
meric repeats or extension of telomeric DSBs, but rather represent a
bona fide telomerase activity extending reversed telomeric forks.
DISCUSSION
Fork reversal is thought to occur when the replication machinery
encounters a roadblock on the template. This phenomenon is typ-
ically observed genome-wide after treatment with genotoxic agents
that induce DNA strand damage, cross-links, inhibit replicative
polymerases, or deplete the deoxynucleoside triphosphate (dNTP)
pool (32). Tightly bound proteins and replication-transcription col-
lisions have also been invoked as potential sources of fork reversal,
while accumulation of positive supercoiling ahead of the fork has
been shown to drive fork regression in reconstituted systems (33,
34). We observe a nearly twofold increase in the incidence of re-
versed forks at telomeric repeats compared to a nontelomeric se-
quence in the SV40 system or to the bulk genomic DNA in
mouse cells. While the absolute levels of fork reversal that we
observe may be influenced by the experimental tools used here,
such as the deregulated expression of the MYC oncogene in the
mouse B cell lymphoma line, or the use of the large T-antigen heli-
case in the SV40 constructs, comparison with the nontelomeric,
control DNA replicated exactly under the same conditions indicates
that mammalian telomeric repeats are hotspots for replication fork
reversal. Our results are consistent with long-standing observations
of telomeres being hard-to-replicate regions from yeast to mammals
(4,8,17,18). Reversed forks at ectopic telomeric repeats in the SV40
system occurred in an orientation-independent manner and were
not affected by treatment with the G4-stabilizing ligand pyridosta-
tin. These results argue against G4-DNA being a major driver of
fork reversal at telomeric repeats, at least in the episomal, SV40
setting. The experiments with ectopic repeats also indicate that re-
versed forks can occur in the absence of a t-loop structure. Other
potential obstacles to telomere replication include the presence of
damage-induced i-loops, or tightly bound shelterin subcomplexes
that need to be removed at the passage of the replication fork
(12–15,18). Other intrinsic features of stretches of repetitive
DNA, like the abundance of local homology, could also contribute
to replication fork reversal at telomeres (16,23). Formation of re-
versed forks during telomere replication could have important con-
sequences for telomere metabolism. The observation that the
SMARCAL1 helicase, which is thought to promote fork reversal,
is also required for telomere integrity suggests that fork reversal
might be a mechanism that deals with replication stress at telomeres
(35,36). The reversed arm would generate an unprotected end,
which could, in principle, lead to the activation of a double-
strand break response at telomeres (37). Furthermore, the reversed
arm could be engaged in interchromosomal break-induced replica-
tion (BIR), favored by the abundance of homology at telomeric
repeats. This would be consistent with reports of BIR in conditions
associated with telomere replication stress (38–40).
Our 2D-gel experiments reveal telomerase-mediated extension
events on DNA structures migrating in the cone signal. This obser-
vation implies generation of the reversed arm in vivo, providing or-
thogonal evidence to the visualization of reversed forks in EM.
Furthermore, they imply that telomerase can be recruited at a re-
versed telomeric fork, where it can add telomeric repeats, extending
the reversed arm. The consequences of this process on telomere
maintenance will depend on at least two factors: (i) the ability of
telomeric reversed forks to restart and (ii) the length of the unrepli-
cated telomeric repeats ahead of the reversed fork. If a telomeric re-
versed fork is unable to restart, the telomeric tract downstream the
fork cannot be completed by semiconservative replication. If this
tract exceeds the repeats that telomerase can add in one S phase,
then this process will result in the loss of telomeric repeats
(Fig. 6). This scenario would apply to mammalian cells with long
telomeres and is consistent with the phenotype of RTEL1-deleted
mouse cells, characterized by telomerase-dependent telomere loss.
Fig. 6. Possible consequences of telomerase engagement at reversed forks. If
reversed forks elongated by telomerase can be reactivated in a timely fashion, then
telomere replication can be completed. In this scenario, telomerase elongation of
the 3′-end may even assist the bypass of certain lesions on the leading strand. If
reversed forks elongated by telomerase cannot restart, then telomeric repeats
ahead of the fork cannot be copied by semiconservative DNA replication. In
cells where the size of telomerase-extension events is comparable to the
average telomere length (i.e., yeast), incomplete replication can be compensated
for by telomerase. However, in cells with kilobase-long telomeres (i.e., mouse or
human), this process would lead to the loss of the telomeric tract ahead of the
reversed fork, which could then only be recovered by recombination-mediated
mechanisms like BIR.
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 7 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

In that setting, it was suggested that abnormal telomerase loading at
telomeres prevented the restart of telomeric reversed forks (21).
However, if telomerase activity can fully compensate for the
length of telomeric repeats ahead of the reversed forks, then the
process will not result in the loss of telomeric repeats, but rather
it might have the advantage of elongating telomeres during replica-
tion, without generating two newly replicated and unprotected
sister ends that would be at risk of fusion (Fig. 6). This scenario is
consistent with the proposed mechanisms of telomere elongation in
yeast, where telomerase extension events are comparable to the
average telomere length (19,20,41). A similar scenario could be en-
visioned also for mammalian cells, with long telomeres, if telome-
rase recruitment and/or fork reversal occurs near the 3′-end of the
chromosome, similarly to what have been proposed for replication
forks encountering a DSB (42). Last, unscheduled telomerase activ-
ity at reversed forks might contribute to the instability of interstitial
telomeric repeats, especially in telomerase-positive cells experienc-
ing replication stress, a recurrent condition in cancer cells.
MATERIALS AND METHODS
Plasmids and cloning
All the SV40 constructs derive from pML113 (23). A Bsm BI site,
leaving Eco RI compatible ends, was introduced next to the Kpn I
site of pML113 to generate pSV40. For the constructs with 34 telo-
meric repeats, an insert containing 34 TTAGGG repeats was ampli-
fied by polymerase chain reaction (PCR) from pTH5, a gift from
T. de Lange, with the following oligos: forward: 5′-TACGGGTAC-
CAATTCGGCCTAATTCGGCCT; reverse: 5′-TCAGGGTAC-
CAATTCGGCCAAATTCGGTAG. The PCR product was
digested with Kpn I and introduced into Kpn I–linearized
pML113. To generate the constructs with 54 and 115 repeats, an
Eco RI fragment containing TTAGGG repeats from pTH9 (a gift
from T. de Lange) was introduced into Bsm BI–linearized pSV40,
and products with one or two insertions in tandem, in either orien-
tation, were selected. For matching size controls (pCtrl), fragments
of 200, 400, and 800 base pairs (bp) (for inserts with 34, 54, and 115
telomeric repeats, respectively) were amplified from the hygromy-
cin resistance marker of pHEBO (a gift from T. de Lange) (43) with
Eco RI–compatible ends and were introduced into Bsm BI–digested
pSV40. For hTERT and hTR-TSQ1 expression, cells were transfect-
ed with pcDNA-3F-hTERT-WT, pBlueScript-U1-hTR-WT, or
pBlueScript-U1-hTR-TSQ1.
Cell culture
HEK-293T Interlab Cell Line Collection (ICLC) cells were grown in
Dulbecco’s modified Eagle’s medium (DMEM) (Lonza, BE12-
614F), supplemented with 10% fetal bovine serum (EuroClone,
ECS0180L) and 2 mM L-glutamine (EuroClone, LOBE17605F).
The mouse B cell lymphoma line #126800 was established from a
lambda-MYC transgenic animal, and early-passage cells were main-
tained in DMEM (Lonza, BE12-614F), supplemented with 10% fetal
bovine serum (EuroClone, ECS0180L), 2 mM L-glutamine (Euro-
Clone, LOBE17605F), 0.1 mM nonessential amino acids (Micro-
tech, X-0557), 1 mM sodium pyruvate (Microtech, L0642), and
40 μM 2-mercaptoethanol (Life Technologies, 31350-010). Cells
were diluted to 1 × 10
5
cells/ml in new plate every 2 days. Transfec-
tion of HEK-293T cells was performed with the calcium phosphate
method. Cells (6 × 10
6
to 7 × 10
6
) were plated in a 15-cm dish (~50%
confluency) 1 day before the transfection. For each plate, 190 μl of
CaCl
2
(2 M) together with 15 μg of plasmid DNA and 1 μg of
pAcGFP1-C1 (Clontech, 632470), in 1500-μl final volume of
ddH
2
O, were mixed with an equal amount of HEPES-buffered
saline (HBS) (2×). During the process of mixing, the solution was
aerated by blowing air through a 2-ml pipette with Pipet-Aid; then,
the solution was added dropwise to the cells. The medium was
changed 16 to 18 hours after transfection. Alternative cells were
transfected with polyethylenimine (PEI Max 40,000, PEI 2500,
Histo-Line Laboratories, catalog no. 24765-1). Cells were transfect-
ed at 50 to 70% confluence. The medium was changed 30 min
before transfection. For each 15-cm plate containing 20 ml of
medium, the DNA (10 to 15 μg per construct) is diluted in
DMEM to a final volume of 150 μl. PEI (stock 1 mg/ml) is
diluted 1:10 to a final volume of 150 μl for each 15-cm plate. The
PEI and DNA dilutions were mixed 1:1, vortexed briefly, and incu-
bated for 10 min at room temperature (RT). The mix was added
dropwise to the cells and gently distributed by rocking the plate
back and forth. The medium was changed again 1 day after the
transfection. When indicated, cells were treated with 50 μM pyri-
dostatin (Sigma-Aldrich, SML0678) 1 hour before the harvest. Pso-
ralen cross-linking was performed as described in (31). Briefly, cells
resuspended in 3 ml of ice-cold Dulbecco's phosphate-buffered
saline (DPBS) with Ca
++
and Mg
++
(Life Technologies, 14080-
089) were poured in a 6-cm dish and kept on ice, in the dark,
while stirring, throughout the procedure. The suspension was first
incubated with 4,5′,8-trimethylpsoralen (~7 μg/ml) (Sigma-
Aldrich, T6137; stock 2 mg/ml in dimethyl sulfoxide, stored at
−20°C) for 5 min in the dark and then exposed to 365-nm ultravi-
olet (UV) light for 2.5 min in UV Stratalinker 1800 (Stratagene),
with 365-nm UV bulbs (model UVL-56, UVP) at 2 to 3 cm from
the light source.
Hirt extraction of extrachromosomal DNA (ecDNA)
The plasmid DNA was recovered using a modified Hirt protocol
(24,44). Cells were washed with 4 ml of ice-cold tris-buffered
saline [50 mM tris-HCl (pH 7.0) and 150 mM NaCl] and lysed by
incubation for 20 min at RT in 0.7 ml of lysis solution [50 mM tris-
HCl (pH 7.0), 20 mM EDTA, 10 mM NaCl, 2% SDS, and proteinase
K (0.4 mg/ml)]. Then, chromosomal DNA was precipitated by the
addition of 0.3 ml of 5 M NaCl, followed by overnight incubation at
4°C. The samples were centrifuged in a SW41-Ti rotor (Beckman) at
16,000 rpm (31,611g) for 50 min at 4°C, and the supernatant (~1 ml)
was collected, supplemented with 2 μl of proteinase K (Roche,
31158870015, stock 50 mg/ml), and incubated at 55°C for 2
hours. DNA was extracted with 1 volume of phenol/chloroform/
isoamyl alcohol (25:24:1; Sigma-Aldrich, P2069) and chloroform
extraction, followed by precipitation with isopropanol.
Two-dimensional agarose gel electrophoresis
Around 3 to 4 μg of plasmid DNA, extracted from 293T cells with
the procedure described above, were digested as follows: The plas-
mids with 115 repeats were digested overnight with 90 U of Sac I and
Bam HI. The day after, 40 U of Dpn I (NEB) was added and the
digestion was incubated again at 37°C for 1 hour. The plasmids
with 34 and 54 telomeric repeats were digested with 90 U of Eco
RI and Not I, respectively, overnight at 37°C and 1 additional
hour with 40 U of Dpn I. The digestion was precipitated with iso-
propanol and resuspended in 10 mM tris-HCl (pH 8.0). The first
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 8 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

dimension was run on 0.4% agarose (US Biological, A1015) in 0.5×
tris-borate-EDTA (TBE) without ethidium bromide (EtBr) for 42
hours at 0.65 V/cm. Subsequently, the first dimension was stained
in 0.5× TBE with EtBr for 45 min. For each lane, a slice of 6.5 cm
from the 5-kb size and up was cleaved. The second dimension was
run on 0.7% agarose gel with EtBr for 23 hours at 2 to 3 V/cm at 4°C.
Before blotting, psoralen cross-linking was reversed by irradiating
the gel for 10 min under 254-nm UV light (UVP CL1000 Ultraviolet
crosslinker). For Southern blotting, the gel was first incubated for 30
min with the depurination solution (0.25 N HCl), 2 × 40 min with
denaturing solution (0.5 M NaOH and 1.5 M NaCl), and 2 × 40 min
with neutralizing solution [0.5 M tris (pH 7.5) and 3 M NaCl]. The
DNA was then transferred by capillarity in 10× SSC onto an Amer-
sham Hybond-X membrane (GE Healthcare, RPN203). Probes were
labeled with the Prime-a-Gene Labeling system (Promega, U1100)
using [α-
32
P]deoxycytidine triphosphate (dCTP) (3000 Ci/μmol;
PerkinElmer, 1300000327). For the TSQ1 repeat detection, the oli-
gonucleotides (complementary of TSQ1 sequence or forward TSQ1
sequence) were end-labeled with T4-PNK (NEB, M0201) and
[γ-
32
P]adenosine triphosphate (ATP) (6000 Ci/μmol; PerkinElmer,
1300000372).
Alternatively, the TSQ1-hi-a (5′-AAACCGCAACCGCAACCG-
CAACCGC) and Template-TSQ1-hi-a (3′-TTTGGCGUUGGC-
GUUGGCGUUGGCG) oligos were used to generate a probe as
described in (45). Briefly, the oligos were annealed by mixing 3.4
μl of 10 μM Template-TSQ1-hi-a and 15.6 μl of 100 μM TSQ1-
high activity with 1 μl of 1 M NaCl and incubating at 99°C for
1 min, 37°C for 15 min, then at RT for 15 min. The labeling reaction
was set up mixing pre-annealed template oligos with 50 μM deox-
yadenosine triphosphate (dATP) and deoxyguanosine triphosphate
(dGTP) (Invitrogen, catalog no. 15612-01), 5 μl of [α-
32
P]dCTP,
and 5 U of DNA polymerase I Klenow fragment (Promega,
catalog no. M220A) in a final volume of 25 μl in buffer M [10
mM tris (pH 7.6), 10 mM MgCl
2
, 50 mM NaCl, 0.5 mM dithiothrei-
tol (DTT), and bovine serum albumin (400 ng/μl)]. The reaction
was incubated for 20 min at RT and then at 95°C for 5 min. After
cooling down to RT, 5 U of uracil deglycosylase (NEB, BM0280S)
was added and the reaction was incubated at 37°C for 10 min and
then at 98°C for 10 min. The volume was brought to 50 μl with TES
(tris-EDTA-SDS), passed through a microspin G25 column (GE
Healthcare, GE275325), and diluted in the hybridization mix
(Church mix). Hybridization was carried out overnight at 50°C.
The membranes were washed three times with 4× SSC for 30 min
each and with 4× SSC, 0.1% SDS for another 30 min at 50°C. For the
detection of telomeres, the pSty11 plasmid was digested with Eco
RI, and the 800-bp fragment was purified from the gel and
diluted at 20 ng/μl in ddH
2
O. Ten microliters (200 ng) of the
800-bp telomeric insert was mixed with 5 μl of telomeric oligo (1
ng/μl) and 24 μl of ddH
2
O, incubated at 95°C for 5 min, and
quickly chilled on ice. For probe labeling, 5 μl of the 10× reaction
buffer, 5 μl of [α-
32
P]dCTP (3000 Ci/μmol; PerkinElmer,
1300000370), and 5 U of DNA polymerase I Klenow fragment
were added, incubated at RT for 90 min. The probe was passed
through a microspin G50 column (GE Healthcare, GE27533001),
incubated at 95°C for 5 min, and chilled on ice before being
diluted in the hybridization mix. Hybridization was carried out at
65°C for at least 4 hours to overnight. The membrane was washed
twice for 15 min with 2× SSC at 65°C and exposed on a phos-
phoscreen for 2 hours. Radioactive signal was captured on phosphor
screens (FUJIFILM Storage Phosphor screen MS3543 E), read on
Typhoon Trio (GE), and analyzed on ImageJ.
Quantitative PCR
Two 10-cm plates of HEK-293T were transfected with a 1:1 mix of
pTelN and pCtrl. Forty hours after transfection, plasmid DNA was
extracted using the Hirt protocol. Quantitative PCR (qPCR) was
performed using LightCycler 480 (Roche), and the amplifications
were done using SsoFast EvaGreen Supermix (Bio-Rad, 1725201),
according to the manufacturer’s indications. The following
primers recognizing the SV40 backbone were used as an internal
control: SV40_qPCR_Fw: 5′-GATAATGCTTATCCAGTGGA;
SV40_qPCR_Rev: 5′-GTAGGTTCCAAAATATCTAG. The primer
couple for the pTelN plasmid was as follows: SV40_Fw: 5′-
CCCTAACCCTCCGAATTGGA; Telo-specific_Rev: 5′-
CCCTAACTGACACACATTCC. The primer couple for the pCtrl
plasmid was as follows: SV40_Fw: 5′-CCCTAACTGACACA-
CATTCC; Random-specific_Rev: 5′-GTCAGGCTCTCGCAATTG-
GA. Reactions were carried out in triplicate for each set of data. The
relative quantification in plasmid replication was determined using
the 2
−ΔΔCt
method (46), obtaining the fold changes in plasmid rep-
lication, normalized with the Ct of the internal control.
Chromatin immunoprecipitation
Four- to 15-cm plates for each condition, containing 20 ml of
medium each, were cross-linked for 10 min at RT by adding 540
μl of 37% formaldehyde (Merck, 104002) and incubating with
gentle shaking. Formaldehyde was quenched for 5 min with 125
mM glycine. Cells were scraped with 1× PBS, keeping the plates
on ice. Chromatin extraction was prepared by lysing in four differ-
ent buffers: Buffer A [100 mM tris-HCl (pH 8.0) and 10 mM DTT]
for 15 min at RT and 15 min at 30°C, Buffer B [10 mM Hepes (pH
7.5), 10 mM EDTA, 0.5 mM EGTA, and 0.25% Triton X-100] for
5 min on ice, Buffer C [10 mM Hepes (pH 7.5), 10 mM EDTA,
0.5 mM EGTA, and 200 mM NaCl] for 5 min on ice, and Buffer
D [50 mM tris-HCl (pH 8.0), 10 mM EDTA, 1% SDS, and 1× pro-
teinase inhibitor cocktail]. Then, the chromatin was sonicated for 10
cycles, 30 s on and 30 s off on Bioruptor Plus (Diagenode). The
chromatin was resuspended in 10 volumes of dilution buffer [16.7
mM tris-HCl (pH 8.0), 167 mM NaCl, 1.2 mM EDTA, and 1.1%
Triton X-100], on a concentration of 100 μg for each condition,
and precleared with 100 μl of blocking beads, prepared for 2
hours at 4°C with 40 μl of Dynabeads Protein G beads (Thermo
Fisher Scientific, 10003D), 10 μl of anti-TRF1 antibody (Santa
Cruz, GO715), and IP dilution buffer [1.1% Triton X-100, 1.2
mM EDTA, 16.7 mM tris-HCl (pH8), and 150 mM NaCl]. After
overnight incubation at 4°C, beads were washed three times for 5
min with a low-salt washing buffer [0.5% Triton X-100, 5 mM
EDTA, 50 mM tris-HCl (pH 8.0), and 150 mM NaCl]. Beads were
then resuspended in 500 μl of elution buffer (0.1 M NaHCO
3
and
1% SDS) for 40 min at RT and subsequently decross-linked by
adding 36 µl tris-HCl 1M (pH 8.0), 182 mM NaCl, and proteinase
K (360 μg/ml; Roche) and incubated for 2 hours at 52°C and then
overnight at 65°C. The DNA was purified using the phenol/chloro-
form/isoamyl alcohol and eluted in 30 μl of 10 mM tris-HCl buffer.
qPCR was performed with LightCycler 480 (Roche) combining 1 μl
of DNA with 18 μl of SsoFast EvaGreen Supermix (Bio-Rad,
1725201) and the following primers: forward: 5′-
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 9 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

GGACTTTCCACACCTGGTTG; reverse: 5′-GGAGACGGG-
TACCTTCTGAG, according to the manufacturer’s indications.
Purication of SV40 replication intermediates for EM
Digested DNA was separated on a preparative gel [0.6% (w/v) low-
melting agarose (Lonza, catalog no. 50100) gel in 1× tris-acetate-
EDTA (TAE)]. The gel was run at 0.8 V/cm for 1 hour and at 2
V/cm for 2 hours. The lane portion above the 5.2-kb fragment
was excised, and the DNA was recovered with the silica bead
DNA gel extraction kit (Thermo Fisher Scientific, K0513) following
the manufacturer’s instructions, except that, once the DNA was
bound, the beads were not resuspended to avoid mechanical shear-
ing of the DNA. The DNA was eluted in 1× TE and quantified using
a Qubit dsDNA HS assay kit (Invitrogen, Q32854).
Enrichment of telomeric repeats
Cells (500 × 10
6
) were used for the telomere enrichment procedure
as described in detail in (31).
EM spreading, acquisition, and analysis
EM analysis was performed according to (26). Typically, 5 μl of telo-
mere-enriched DNA corresponding to 20 to 100 ng was used for
each spread. The DNA was added to premixed 5 μl of formamide
(Thermo Fisher Scientific, 17899) and 0.4 μl of 0.02% benzalko-
nium chloride (BAC; Sigma-Aldrich, B6285) in TE (0.2% BAC
stock solution in formamide was diluted 1:10 in 1× TE before
use). After mixing, the drop was immediately spread on a water
surface in a 15-cm dish containing 50 ml of distilled water, using
a freshly cleaved mica sheet (Ted Pella Inc., product no. 52-6) as a
ramp. For nonenriched controls, 30 ng of Kpn I–digested genomic
DNA was spread using the same method. The monomolecular layer
was gently touched with an EM grid prepared as follows. A thin (4 to
8 nm), homogeneous, and low-grain carbon layer was deposited on
EM grids (Nanovision, PEG400). The carbon layer was created on a
mica glass surface (2 cm × 2 cm) using the MED020 e-beam evap-
orator (Leica), equipped with the QSG monitor, two EK030 electron
guns controlled by the EVM030 control unit. The e-beam evapora-
tion parameters described in the instruction manual were used. The
carbon layer deposited on the mica surface was floated on the
surface of the water and transferred on the 400-mesh copper
grids. Before use, carbon-coated EM grids were placed in contact
with an ethidium bromide solution (33.3 μg/ml in H
2
O) for 30 to
45 min at RT. Carbongrids with absorbed DNA molecules were im-
mediately stained with a solution of uranyl acetate (0.2 μg/μl) in
ethanol and coated with 8 nm of platinum using the MED020 evap-
orator modified with the low-angle grid shadowing kit (Leica
16770525) so that the sample holder was placed at an angle of
280.5° and made an angle of around 3° with the platinum gun
fixed on the head of the instrument. For platinum e-beam evapora-
tion, we used the parameters indicated in the MED020 instruction
manual. TEM pictures were taken using a FEI Tecnai12 BioTwin
microscope operated at 120 kV and equipped with a side-
mounted GATAN Orius SC-1000 camera controlled by the
Digital Micrograph software. Images in DM3 format were analyzed
using the ImageJ software. In these conditions, 0.36 μm corresponds
to 1 kb of double-stranded DNA. Molecules with reversed arm
inside a replication bubble or X-shaped molecules with two equal
arms were scored as reversed forks.
Telomere blots
Genomic DNA was extracted from 10
7
cells with phenol/chloro-
form/isoamyl alcohol, as described in (31). Thirty micrograms of
genomic DNA was digested overnight with 150 U of Hin fI and
Rsa I and separated on a 0.8% agarose gel in 1× TAE at 1.8 V/cm
for 16 hours. The gel was blotted onto a membrane and hybridized
with
32
P-labeled probes as described above.
Fluorescence in situ hybridization
HEK-293T cells were plated onto glass coverslips 2 days after the
transfection. The next day, the cells were fixed with formaldehyde
4% for 10 min at RT and then washed with PBS. The cells were de-
hydrated in ethanol series 70%-90%-100%, 5 min each, and then
air-dried. The coverslips were then incubated with the hybridizing
solution [70% formamide, Thermo Fisher Scientific, 17899; 10 mM
tris (pH 7.4), 0.5% blocking reagent, Roche, 11096176001] with a
mix of Cy5-labeled (TTAGGG)3 probe (PNA Bio) and CY3-
labeled (CGCAAC)3 PNA probe (Panagene) (1:500; stock 54 μM,
used 100 nm each). The coverslips were incubated for 5 min at
80°C and then overnight at RT. The coverslips were washed twice
with wash buffer I [70% formamide and 10 mM tris (pH 7.4)] at
RT, once with PBS-Tween (0.1% final) with 4′,6-diamidino-2-phe-
nylindole (DAPI) (0.5 mg/ml, 3000×), and once with PBS-Tween,
dehydrated in ethanol series 70%-90%-100%, 5 min each, air-dried,
and mounted with ProLong Gold antifade reagent (Invitrogen,
catalog no. P36930). The images were acquired on a DV Elite
system (GE Healthcare) equipped with an IX71 microscope
(Olympus) and a scientific complementary metal-oxide semicon-
ductor (sCMOS) camera and driven by SoftWoRx version 7.0.0
and analyzed in ImageJ.
Supplementary Materials
This PDF le includes:
Figs. S1 to S4
REFERENCES AND NOTES
1. T. de Lange, Shelterin-mediated telomere protection. Annu. Rev. Genet. 52,
223–247 (2018).
2. R. A. Wu, H. E. Upton, J. M. Vogan, K. Collins, Telomerase mechanism of telomere synthesis.
Annu. Rev. Biochem. 86, 439–460 (2017).
3. L. Crabbe, R. E. Verdun, C. I. Haggblom, J. Karlseder, Defective telomere lagging strand
synthesis in cells lacking WRN helicase activity. Science 306, 1951–1953 (2004).
4. A. Sfeir, S. T. Kosiyatrakul, D. Hockemeyer, S. L. MacRae, J. Karlseder, C. L. Schildkraut, T. de
Lange, Mammalian telomeres resemble fragile sites and require TRF1 for efficient repli-
cation. Cell 138, 90–103 (2009).
5. P. Martinez, M. Thanasoula, P. Muñoz, C. Liao, A. Tejera, C. M. Nees, J. M. Flores,
O. Fernández-Capetillo, M. Tarsounas, M. A. Blasco, Increased telomere fragility and fusions
resulting from TRF1 deficiency lead to degenerative pathologies and increased cancer in
mice. Genes Dev. 23, 2060–2075 (2009).
6. J.-B. Vannier, V. Pavicic-Kaltenbrunner, M. I. Petalcorin, H. Ding, S. J. Boulton, RTEL1 dis-
mantles T loops and counteracts telomeric G4-DNA to maintain telomere integrity. Cell
149, 795–806 (2012).
7. M. Zimmermann, T. Kibe, S. Kabir, T. de Lange, TRF1 negotiates TTAGGG repeat-associated
replication problems by recruiting the BLM helicase and the TPP1/POT1 repressor of ATR
signaling. Genes Dev. 28, 2477–2491 (2014).
8. K. M. Miller, O. Rog, J. P. Cooper, Semi-conservative DNA replication through telomeres
requires Taz1. Nature 440, 824–828 (2006).
9. G. Sarek, J.-B. Vannier, S. Panier, J. H. J. Petrini, S. J. Boulton, TRF2 recruits RTEL1 to telo-
meres in S phase to promote t-loop unwinding. Mol. Cell 57, 622–635 (2015).
10. E. Gilson, V. Géli, How telomeres are replicated. Nat. Rev. Mol. Cell Biol. 8, 825–838 (2007).
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 10 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE

11. B. Silva, R. Arora, S. Bione, C. M. Azzalin, TERRA transcription destabilizes telomere integrity
to initiate break-induced replication in human ALT cells. Nat. Commun. 12, 3760 (2021).
12. G. Mazzucco, A. Huda, M. Galli, D. Piccini, M. Giannattasio, F. Pessina, Y. Doksani, Telomere
damage induces internal loops that generate telomeric circles. Nat. Commun. 11,
5297 (2020).
13. R. Ohki, F. Ishikawa, Telomere-bound TRF1 and TRF2 stall the replication fork at telomeric
repeats. Nucleic Acids Res. 32, 1627–1637 (2004).
14. M. E. Douglas, J. F. X. Diffley, Budding yeast Rap1, but not telomeric DNA, is inhibitory for
multiple stages of DNA replication in vitro. Nucleic Acids Res. 49, 5671–5683 (2021).
15. E. A. Radchenko, A. Y. Aksenova, K. V. Volkov, A. A. Shishkin, Y. I. Pavlov, S. M. Mirkin,
Partners in crime: Tbf1 and Vid22 promote expansions of long human telomeric repeats at
an interstitial chromosome position in yeast. PNAS Nexus 1, pgac080 (2022).
16. N. Fouché, S. Ozgür, D. Roy, J. D. Griffith, Replication fork regression in repetitive DNAs.
Nucleic Acids Res. 34, 6044–6050 (2006).
17. S. Makovets, I. Herskowitz, E. H. Blackburn, Anatomy and dynamics of DNA replication fork
movement in yeast telomeric regions. Mol. Cell. Biol. 24, 4019–4031 (2004).
18. A. S. Ivessa, J. Q. Zhou, V. P.Schulz, E. K. Monson, V. A. Zakian, Saccharomyces Rrm3p, a 5′to
3’DNA helicase that promotes replication fork progression through telomeric and sub-
telomeric DNA. Genes Dev. 16, 1383–1396 (2002).
19. S. Matmati, S. Lambert, V. Géli, S. Coulon, Telomerase repairs collapsed replication forks at
telomeres. Cell Rep. 30, 3312–3322.e3 (2020).
20. P.-M. Dehe, O. Rog, M. G. Ferreira, J. Greenwood, J. P. Cooper, Taz1 enforces cell-cycle
regulation of telomere synthesis. Mol. Cell 46, 797–808 (2012).
21. P. Margalef, P. Kotsantis, V.Borel, R. Bellelli, S. Panier, S. J. Boulton, Stabilization of reversed
replication forks by telomerase drives telomere catastrophe. Cell 172, 439–453.e14 (2018).
22. M. E. Diolaiti, B. A. Cimini, R. Kageyama, F. A. Charles, B. A. Stohr, In situ visualization of
telomere elongation patterns in human cells. Nucleic Acids Res. 41, e176 (2013).
23. C. Follonier, J. Oehler, R. Herrador, M. Lopes, Friedreich’s ataxia-associated GAA repeats
induce replication-fork reversal and unusual molecular junctions. Nat. Struct. Mol. Biol. 20,
486–494 (2013).
24. B. Hirt, Selective extraction of polyoma DNA from infected mouse cell cultures. J. Mol. Biol.
26, 365–369 (1967).
25. Z. Deng, C. Atanasiu, J. S. Burg, D. Broccoli, P. M. Lieberman, Telomere repeat binding
factors TRF1, TRF2, and hRAP1 modulate replication of Epstein-Barr virus OriP. J. Virol. 77,
11992–12001 (2003).
26. M. Lopes, Electron microscopy methods for studying in vivo DNA replication intermedi-
ates. Methods Mol. Biol. 521, 605–631 (2009).
27. M. Lopes, C. Cotta-Ramusino, A. Pellicioli, G. Liberi, P. Plevani, M. Muzi-Falconi, C. S. Newlon,
M. Foiani, The DNA replication checkpoint response stabilizes stalled replication forks.
Nature 412, 557–561 (2001).
28. J. Courcelle, J. R. Donaldson, K. H. Chow, C. T. Courcelle, DNA damage-induced replication
fork regression and processing in Escherichia coli.Science 299, 1064–1067 (2003).
29. D. T. Long, K. N. Kreuzer, Regression supports two mechanisms of fork processing in phage
T4. Proc. Natl. Acad. Sci. U.S.A. 105, 6852–6857 (2008).
30. A. L. Kovalchuk, C.-F. Qi, T. A. Torrey, L. Taddesse-Heath, L. Feigenbaum, S. S. Park, A. Gerbitz,
G. Klobeck, K. Hoertnagel, A. Polack, G. W. Bornkamm, S. Janz, H. C. Morse III, Burkitt
lymphoma in the mouse. J. Exp. Med. 192, 1183–1190 (2000).
31. G. Mazzucco, A. Huda, M. Galli, E. Zanella, Y. Doksani, Purification of mammalian telomeric
DNA for single-molecule analysis. Nat. Protoc. 17, 1444–1467 (2022).
32. K. J. Neelsen, M. Lopes, Replication fork reversal in eukaryotes: From dead end to dynamic
response. Nat. Rev. Mol. Cell Biol. 16, 207–220 (2015).
33. J. Atkinson, P. McGlynn, Replication fork reversal and the maintenance of genome stability.
Nucleic Acids Res. 37, 3475–3492 (2009).
34. L. Postow, C. Ullsperger, R. W. Keller, C. Bustamante, A. V. Vologodskii, N. R. Cozzarelli,
Positive torsional strain causes the formation of a four-way junction at replication forks. J.
Biol. Chem. 276, 2790–2796 (2001).
35. L. A. Poole, R. Zhao, G. G. Glick, C. A. Lovejoy, C. M. Eischen, D. Cortez, SMARCAL1 maintains
telomere integrity during DNA replication. Proc. Natl. Acad. Sci. U.S.A. 112,
14864–14869 (2015).
36. K. P. Bhat, D. Cortez, RPA and RAD51: Fork reversal, fork protection, and genome stability.
Nat. Struct. Mol. Biol. 25, 446–453 (2018).
37. R. E. Verdun, L. Crabbe, C. Haggblom, J. Karlseder, Functional human telomeres are rec-
ognized as DNA damage in G2 of the cell cycle. Mol. Cell 20, 551–561 (2005).
38. R. L. Dilley, P. Verma, N. W. Cho, H. D. Winters, A. R. Wondisford, R. A. Greenberg, Break-
induced telomere synthesis underlies alternative telomere maintenance. Nature 539,
54–58 (2016).
39. Z. Yang, K. K. Takai, C. A. Lovejoy, T. de Lange, Break-induced replication promotes fragile
telomere formation. Genes Dev. 34, 1392–1405 (2020).
40. R. M. Porreca, E. Herrera-Moyano, E. Skourti, P. P. Law, R. Gonzalez Franco, A. Montoya,
P. Faull, H. Kramer, J.-B. Vannier, TRF1 averts chromatin remodelling, recombination and
replication dependent-break induced replication at mouse telomeres. eLife 9,
e49817 (2020).
41. E. Bonnell, E. Pasquier, R. J. Wellinger, Telomere replication: Solving multiple end repli-
cation problems. Front. Cell Dev. Biol. 9, 668171 (2021).
42. Y. Doksani, R. Bermejo, S. Fiorani, J. E. Haber, M. Foiani, Replicon dynamics, dormant origin
firing, and terminal fork integrity after double-strand break formation. Cell 137,
247–258 (2009).
43. J. L. Yates, N. Warren, B. Sugden, Stable replication of plasmids derived from Epstein-Barr
virus in various mammalian cells. Nature 313, 812–815 (1985).
44. G. S. Chandok, K. K. Kapoor, R. M. Brick, J. M. Sidorova, M. M. Krasilnikova, A distinct first
replication cycle of DNA introduced in mammalian cells. Nucleic Acids Res. 39,
2103–2115 (2011).
45. Y. Zhao, J. W. Shay,W. E. Wright, Telomere G-overhang length measurement method 1: The
DSN method. Methods Mol. Biol. 735, 47–54 (2011).
46. K. J. Livak, T. D. Schmittgen, Analysis of relative gene expression data using real-time
quantitative PCR and the 2
−ΔΔCT
method. Methods 25, 402–408 (2001).
Acknowledgments: We are grateful to the IFOM cell biology unit, IFOM imaging facility for
support, and M. Giannattasio for assistance with EM. We thank M. Lopes for providing the
pML113 plasmid and technical advice. We thank T. de Lange for providing the plasmids with
telomeric repeats. We are grateful to F. Pessina for assistance with the ChIP and qPCR
experiments and E. Zanella for critical reading. Funding: This work was supported by the
Associazione Italiana per la Ricerca sul Cancro, AIRC, IG 19901 (to A.H., G.M., M.G., and Y.D.);
Concern Foundation (to Y.D.); Associazione Italiana per la Ricerca sul Cancro, AIRC, IG 23747 (to
S.C.); Umberto Veronesi Foundation (to V.P.); National Cancer Institute (NCI) grant P30CA006927
(to L.C.); and National Institute of Allergy and Infectious Diseases (NIAID) grant 1R21AI164333
(to L.C.). Author contributions: A.H. performed all the 2D-gel experiments and ChIP and qPCR
experiments and prepared samples for the EM experiments with the SV40 constructs. H.A.
cloned the telomeric repeats in the pML113 vector. G.M. and M.G. performed the EM spreading
and acquisition with the pTelN and pCtrl samples. G.M. performed the experiments with the
mouse B cell lymphoma line. V.P.and S.C. generated the mouse B cell lymphoma cell line used
for telomere enrichment and assisted with the experimental design in using this cell line. L.C.
provided the hTERT and hTR-TSQ1 constructs and assisted with the experimental design and
analysis involving those constructs. Y.D. conceived the study, designed the experiments, and
wrote the manuscript. Competing interests: The authors declare thatthey have no competing
interests. Data and materials availability: All data needed to evaluate the conclusions in the
paper are present in the paper and/or the Supplementary Materials.
Submitted 5 October 2022
Accepted 23 February 2023
Published 22 March 2023
10.1126/sciadv.adf2011
Huda et al.,Sci. Adv. 9, eadf2011 (2023) 22 March 2023 11 of 11
SCIENCE ADVANCES |RESEARCH ARTICLE