Ataxia telangiectasia mutated (ATM) modulates long
interspersed element-1 (L1) retrotransposition in
human neural stem cells
Nicole G. Coufala, Josè Luis Garcia-Perezb,c, Grace E. Penga,d, Maria C. N. Marchettoa, Alysson R. Muotrie,
Yangling Mua, Christian T. Carsonf, Angela Maciab,c, John V. Morang,h, and Fred H. Gagea,1
aLaboratory of Genetics, Salk Institute, La Jolla, CA 92037;bAndalusian Stem Cell Bank, Centro de Investigación Biomédica, Consejería Salud Junta de Andalucia-
Universidad de Granada, 18100 Granada, Spain;cDepartment of Human DNA Variability, Pfizer-University of Granada and Andalusian Government Center for
Genomics and Oncology (GENYO), 18007 Granada, Spain;dCardiovascular Research Institute, University of California, San Francisco, CA 94158;eDepartment of
Pediatrics/Rady Children’s Hospital San Diego, School of Medicine, University of California San Diego, La Jolla, CA 92093;fBD Biosciences, San Diego, CA 92121;
andgDepartment of Human Genetics and Internal Medicine,hHoward Hughes Medical Institute, University of Michigan Medical School, Ann Arbor, MI 48109
Edited by Marlene Belfort, University at Albany, Albany, NY, and approved November 3, 2011 (received for review May 10, 2011)
Long interspersed element-1 (L1) retrotransposons compose ∼20%
of the mammalian genome, and ongoing L1 retrotransposition
events can impact genetic diversity by various mechanisms. Pre-
vious studies have demonstrated that endogenous L1 retrotrans-
position can occur in the germ line and during early embryonic
development. In addition, recent data indicate that engineered
human L1s can undergo somatic retrotransposition in human neu-
ral progenitor cells and that an increase in human-specific L1 DNA
content can be detected in the brains of normal controls, as well as
in Rett syndrome patients. Here, we demonstrate an increase in
the retrotransposition efficiency of engineered human L1s in cells
that lack or contain severely reduced levels of ataxia telangiectasia
mutated, a serine/threonine kinase involved in DNA damage sig-
naling and neurodegenerative disease. We demonstrate that the
increase in L1 retrotransposition in ataxia telangiectasia mutated-
deficient cells most likely occurs by conventional target-site
primed reverse transcription and generate either longer, or per-
haps more, L1 retrotransposition events per cell. Finally, we pro-
vide evidence suggesting an increase in human-specific L1 DNA
copy number in postmortem brain tissue derived from ataxia tel-
angiectasia patients compared with healthy controls. Together,
these data suggest that cellular proteins involved in the DNA dam-
age response may modulate L1 retrotransposition.
and they mobilize (i.e., retrotranspose) by a “copy-and-paste”
mechanism termed target-site primed reverse transcription (TPRT)
(1, 2). Although the vast majority of human L1 sequences are ret-
rotransposition defective, ∼80–100 full-length retrotransposition-
competent L1s (RC-L1s) persist in the genome (3, 4). RC-L1s are
6 kb in length and contain two ORFs that encode proteins required
for their mobility (5). ORF1 encodes a protein (ORF1p) with RNA
binding and nucleic acid chaperone activity (6, 7), whereas ORF2
encodes a protein (ORF2p) with endonuclease (8) and reverse
transcriptase (9) activities. L1 retrotransposition occasionally can
lead to disease and can impact human genome structural variation
by various mechanisms (1, 10, 11). Heritable L1 insertions must
occur in the germ line or during early embryonic development (11).
However, engineered human L1s can undergo somatic retro-
transposition in the mammalian nervous system, and previous
studies have demonstrated an increase in the DNA copy number of
human-specific L1s in the brains of normal individuals compared
with other somatic tissues (12, 13).
Host DNA repair processes may also impact L1 retro-
transposition. For example, DNA repair pathways may either
insertions (14, 15). Studies of cultured cells and comparative
genomics analyses have further demonstrated that L1 retro-
transposition events are associated with various genomic struc-
tural DNA rearrangements, which include intrachromosomal
deletions, intrachromosomal duplication/inversions, and perhaps
ong interspersed element-1 (L1) retrotransposons are the only
autonomously active retrotransposons in the human genome,
interchromosomal translocations (11, 16–20). Finally, mutations
in genes required for the nonhomologous end-joining (NHEJ)
pathway of DNA repair allow for an alternate, endonuclease-in-
dependent pathway of L1 retrotransposition (ENi) in select p53-
deficient Chinese hamster ovary (CHO) cell lines (14, 21). ENi
retrotransposition may occur at areas of DNA disrepair or at
dysfunctional telomeres, and the resultant retrotransposition
events generally lack canonical L1 structural hallmarks (14, 21).
The ataxia telangiectasia mutated (ATM) gene encodes a 350-
kDa serine/threonine kinase that is a sensor of cellular DNA
damage. ATM is activated by double-strand DNA breaks and
subsequently phosphorylates downstream substrates, such as
CHK2, p53, BRCA1, and the MRN complex (MRE11, Rad50,
and NBS1), leading to the activation of a DNA damage check-
point and cell cycle arrest (22). The damaged DNA subsequently
is repaired, or the cell may undergo p53-mediated apoptosis (22).
In humans, autosomal recessive mutations that inactivate ATM
result in ataxia telangiectasia (AT), and AT patients exhibit
progressive neurodegeneration and eventual death in the second
or third decade of life (22). Gene knockout studies indicate that
ATM-deficient mice are viable (23) and model human AT, ex-
hibiting evidence of neurodegeneration, T-cell deficits, growth
retardation, infertility, and sensitivity to gamma radiation (23).
deficiency. Using a cell culture-based retrotransposition assay
(5, 24) and a transgenic mouse model, we consistently observed
a modest two- to fourfold increase in L1 retrotransposition
in ATM-deficient cells compared with ATM-proficient controls.
This increase in L1 retrotransposition likely occurs by conven-
tional TPRT and may result in longer, or perhaps more, L1 ret-
rotransposition events in cells deficient for ATM. We also present
evidence for an increase in L1 DNA content in postmortem
hippocampal brain tissues derived from human AT patients
when compared to healthy age-matched controls. These findings
strongly suggest that the ATM-signaling pathway plays a role in
regulating L1 retrotransposition.
This paper results from the Arthur M. Sackler Colloquium of the National Academy of
Sciences, “Telomerase and Retrotransposons: Reverse Transcriptases That Shaped Ge-
nomes,” held September 29–30, 2010, at the Arnold and Mabel Beckman Center of the
National Academies of Sciences and Engineering in Irvine, CA. The complete program and
audio files of most presentations are available on the NAS Web site at www.nasonline.
Author contributions: N.G.C., J.L.G.-P., J.V.M., and F.H.G. designed research; N.G.C., J.L.G.-P.,
G.E.P., M.C.N.M., A.R.M., Y.M., C.T.C., and A.M. performed research; N.G.C. J.L.G.-P.,
and J.V.M. contributed new reagents/analytic tools; N.G.C., G.E.P., M.C.N.M., A.R.M.,
Y.M., C.T.C., A.M., J.V.M., and F.H.G. analyzed data; and N.G.C., J.L.G.-P., J.V.M., and
F.H.G. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| December 20, 2011
| vol. 108
| no. 51 www.pnas.org/cgi/doi/10.1073/pnas.1100273108
We previously demonstrated that an engineered human RC-L1
(L1RP) (25) containing a retrotransposition indicator cassette
could retrotranspose in the brain of transgenic mice (13, 25).
L1RP–enhanced green fluorescent protein (EGFP) transgenic
mice harbor a retrotransposition-competent human L1 driven by
its native promoter (13, 24). The 3′ UTR of the L1 also contains
a retrotransposition indicator cassette (mEGFPI) that can be
activated only upon L1 retrotransposition (13, 24). To analyze the
role of ATM in L1 retrotransposition in vivo, we generated ATM
knockout (KO) mice containing the L1RP–EGFP transgene (Fig.
contained the L1RP–EGFP transgene but lacked detectable de
novo germ-line retrotransposition events (Fig. S1A). Compar-
isons of the brain samples derived from 3-mo-old wild-type (WT)
and ATM KO mice revealed a statistically significant increase in
the number of EGFP-positive cells in ATM KO mice compared
with WT animals (Fig. 1 A and B). The most marked increase was
observed in the hippocampus (Fig. 1A), and EGFP-positive cells
were not apparent in other somatic tissues (Fig. S1B). In-
terestingly, EGFP-positive cells were not seen in the testes of
ATM-deficient animals; however, as described previously, rare
EGFP-positive cells were noted in WT testes (Fig. S1C) (13, 27).
These data suggest an increased rate of L1 retrotransposition in
ATM KO animals. The L1 insertions may occur during embryonic
development or during adult hippocampal neurogenesis.
We next examined L1 retrotransposition in ATM-deficient
human neural progenitor cells (NPCs). Briefly, we transduced
human embryonic stem cells (HUES6-hESCs) with three differ-
ent lentiviral vectors harboring small hairpin RNAs (shRNAs)
designed to knock down ATM expression (Fig. 2A). Each shRNA
suppressed ATM expression, whereas a scrambled control
shRNA did not (Fig. 2B and Fig. S2 C and D). The derived hESC
lines were karyotypically stable (Fig. S3 A–D) and exhibited
marked reductions in both ATM expression (Fig. 2C) and protein
levels (Fig. 2B). Notably, RT-PCR experiments revealed minor
alterations in the expression of p53 and ataxia telangiectasia and
Rad3-related protein (ATR) in ATM-deficient cells (Fig. 2C).
NPC lines expressing the shRNA vectors were transfected
with an LRE3–EGFP construct, and L1 retrotransposition was
monitored by flow cytometry (EGFP-positive cells) 14 d post
transfection (Fig. 2A; Methods). An L1 construct containing a
pair of missense mutations in ORF1p (JM111/L1RP) served as a
negative control (5, 24). We reliably detected an approximately
twofold increase in L1 retrotransposition in ATM-deficient
NPCs compared with controls (Fig. 2E; n = 3). Notably, the
number of EGFP-positive cells was more pronounced upon the
retrotransposition occurs in the brains of wild type (WT) and ATM KO ani-
mals. (Left) Images of maximal numbers of EGFP-positive cells (indicative of
L1 retrotransposition) in ATM KO mice brain. (Right) An image of EGFP-
positive cells in an ATM-proficient control mouse brain; the most marked
increase was found in the hippocampus. (Scale bar, 50 μm.) (B) Quantifica-
tion of brain sections in an ATM KO transgenic mouse background revealed
more EGFP-positive cells compared with a WT control (six age-matched
animals per group). Error bars indicate SEM. *P < 0.05.
ATM modulates neuronal L1 retrotransposition in vivo. (A) L1–EGFP
Diagram was adapted from ref. 12. HUES6-hESCs were infected with shRNA-
expressing lentiviruses and selected with blasticidin. HESC colonies with
a strong knockdown of ATM were differentiated to embryoid bodies, then
neural rosettes, and were next manually dissected. The resulting NPCs were
propagated with fibroblast growth factor 2 (FGF2). NPCs were transfected
with an engineered human L1 and were assayed for retrotransposition and/
or their ability to differentiate. (B) Western blot analysis indicates robust
depletion of ATM expression with ATM 1-3 shRNAs lentiviruses in HUES6-
derived NPCs. SOX2 is a loading control. (C) RT-PCR of ATM-deficient HUES6-
derived NPCs for genes involved in DNA damage signaling, exhibiting an
effect of ATM knockdown on ATR and, to a lesser extent, on p53. (D) ATM-
deficient NPCs accommodate L1 retrotransposition; cells exhibiting EGFP
expression continued to express neural progenitor markers SOX2 and nestin
(arrows indicate colabeled cells). (E) ATM-deficient HUES6-derived NPCs ex-
hibit increased L1 retrotransposition when transfected with an engineered
human L1. The addition of trichostatin A (TsA) revealed epigenetic silencing
of some engineered L1 insertions. Error bars indicate SEM. *P < 0.05 and
**P < 0.01 are a result of a one-way ANOVA with Bonferroni correction. (F and
G) LRE3–EGFP-positive ATM-deficient HUES6-derived neurons exhibit supra-
threshold responses to somatic current injections from a current clamped
potential of −70 mV (F) and can exhibit spontaneous action potentials (G).
ATM modulates L1 retrotransposition in HUES6-derived NPCs. (A)
Coufal et al.PNAS
| December 20, 2011
| vol. 108
| no. 51
addition of the histone deacetylase inhibitor trichostatin A (500
nM) (Fig. 2E), suggesting that some engineered L1 retro-
transposition events in NPCs are subject to epigenetic silencing
(12, 13, 28).
Experiments were next conducted to examine the integrity of
ATM-deficient NPCs. EGFP-positive, ATM-deficient NPCs ex-
pressed the neural progenitor markers Nestin and Sox2 (Fig. 2D)
and could be differentiated into both neuronal (Map2a+b, βIII
tubulin, and synapsin) and glial (glial fibrillary acidic protein-pos-
itive) cell fates (Fig. S4 A–E). Characterization of differentiated
EGFP-positive, ATM-deficient neurons revealed that some ex-
pressed subtype-specific neuronal markers such as the dopami-
nergic marker tyrosine hydroxylase (Fig. S5A), the cholinergic
marker choline acetyltransferase (Fig. S5A), and the inhibitory
neuronal marker GABA (γ-aminobutyric acid) (Fig. S4D). Finally,
whole-cell perforated patch-clamp recording demonstrated that
certain EGFP-positive, ATM-deficient neurons exhibited both
responses to somatic current injection and spontaneous action
potentials that were similar to ATM-proficient HUES6-derived
and G and Fig. S6 A–E). Thus, WT and ATM-deficient NPCs ac-
commodate engineered human L1 retrotransposition and can be
differentiated into functional, mature neurons.
We next investigated why L1 retrotransposition is increased in
ATM-deficient NPCs. A recent study reported that loss of MeCP2
well as L1 ribonucleoprotein particle (RNP) levels, in ATM-pro-
ficient and ATM-deficient hESC-derived NPCs. We did not ob-
serve significant differences in L1 promoter activity upon the
differentiation of ATM-proficient or ATM-deficient NPCs (Fig.
3A); similarly, the synapsin promoter showed that the differentia-
tionpotentialandactivity ofa cellularpromoterweresimilarinthe
two cell types (Fig. 3A). Moreover, we did not observe marked
differences in the level of endogenous ORF1p present in RNPs
isolated from ATM-proficient or ATM-deficient hESC-derived
NPCs or hESCs (Fig. 3B and Fig. S7D). L1 expression has been
reported to induce cellular toxicity (29). Thus,ATM-deficient cells
may have an increased tolerance for L1-induced toxicity, allowing
for increased survival of cells containing L1 events. Finally, we
observed similar cell cycle profiles and cell division rates, as well as
similar 8-d growth and survival rates between ATM-proficient and
ATM-deficient hESCs and NPCs (Fig. S2 A, B, and E).
Previous studies revealed an alternative pathway of L1 retro-
transposition (ENi), which occurs in select CHO cell lines that
were defective NHEJ (14, 21). Therefore, we next analyzed
whether ENi was responsible for the increase in L1 retro-
transposition observed in ATM-deficient cells. As above, we ob-
in ATM-deficient NPCs compared with ATM-proficient controls
(Fig. 3D). By comparison, we did not observe retrotransposition
from constructs containing missense mutations in the ORF1p
RNA-binding domain (RR261-262AA; JM111/L1RP) (5, 24), the
ORF2p RT domain (L1.3/D702A) (30), or the ORF2p endonu-
clease domain (L1.3/D205A and L1.3/H230A) (5, 8, 14) (Fig. 3D
and Fig. S7A). Similar results were obtained from experiments
conducted with ATM-deficient and ATM-proficient human
fetal brain stem cells (hCNS-SCns) (31) and with HUES6 hESCs
(32) (Fig. S7 B and C). Finally, PCR amplification experiments
conducted on genomic DNA derived from ATM-deficient,
HUES6-derived NPCs revealed splicing of the intron from the
mEGFPI reporter cassette in cells transfected with a WT L1
(LRE3–EGFP) but not from cells transfected with ORF1p or
retrotransposition does not appear to be responsible for the in-
crease in engineered L1 retrotransposition observed in ATM-
Using inverse PCR, we characterized engineered L1 retro-
transposition events from either ATM-proficient (3 events) or
ATM-deficient (13 events) NPCs (12–14, 28) (Table S1). Com-
parisons of the integration sequences revealed that each
retrotransposition event occurred in an actual or inferred L1
endonuclease consensus cleavage site (5′-TTTT/A and deriva-
tives) (14). Three L1 insertions were fully characterized and the
remaining 13 insertions were characterized only at their 3′ ends,
each ending in a poly(A) tail that ranged in size from ∼22 to >
∼130 bp. Thus, the majority of L1 retrotransposition events in
ATM-deficient cells likely occur by conventional TPRT.
Given the discrepancy between our findings and those of a
previous study (33), we sought to independently analyze the role
of ATM in L1 mobilization in a nonembryonic/neuronal cell type.
We also investigated possible interactions between NHEJ and the
ATM DNA repair pathways. We obtained a parental human
Relative Luciferase Activity
ATM deficient HUES6-derived NPCs
D205A + TsA
H230A + TsA
D702A + TsA
derived NPCs. (A) The L1 5′ UTR is rapidly induced upon HUES6-derived
NPC differentiation in both ATM-deficient and ATM-proficient shRNA-
infected cells, as is synapsin, a marker of differentiation. The x axis indicates
days postdifferentiation; the y axis indicates luciferase fold activity. (B) RNP
particles were isolated from control, and ATM-deficient NPCs and were
subjected to Western blotting using anti-ORF1p and antiribosomal S6 anti-
bodies. SOX1 is a whole-cell lysate control. (C) PCR of genomic DNA con-
firmed L1 retrotransposition. The 1,243-bp product corresponds to the EGFP
retrotransposition indicator cassette (lane 1, L1 + intron); the 342-bp product
is diagnostic for intron loss (lane 2, control EGFP no intron). (D) L1 retro-
transposition (EGFP-positive cells) is detected in HUES6-derived NPCs trans-
fected with a WT L1 (L1.3–EGFP), but not in cells transfected with mutant
L1 constructs [L1.3/JM111–EGFP (ORF1 mutant), L1.3/D205A–EGFP, or L1.3/
H230A–EGFP (endonuclease mutants) or L1.3/D702A–EGFP (reverse tran-
scriptase mutant)]. Notably, L1.3–EGFP retrotransposition is modestly in-
creased in ATM-deficient cells compared with controls. Also, the addition of
the histone deacetylase inhibitor trichostatin A (TsA) enhanced the ability to
detect retrotransposition. Error bars indicate SEM. *P < 0.05 and **P < 0.01
are a result of a one-way ANOVA with Bonferroni correction.
Examination of L1 retrotransposition in ATM-deficient HUES6-
| www.pnas.org/cgi/doi/10.1073/pnas.1100273108Coufal et al.
HCT116 colorectal cancer cell line (34) as well as isogenic mutant
HCT116 derivatives that lacked the p53 gene or lacked genes
important forNHEJ(DNA-PKcs,XRCC4-likefactor (XLF),orthe
DNALigase IVgenes, respectively) (Fig.4A).Notably, each ofthe
NHEJ-deficient cell lines assayed supported LRE3–EGFP ret-
rotransposition but did not accommodate ENi retrotransposition
(Endo-, LRE3–H230A–EGFP) (Fig. 4B).
Because human HCT116 cells possess WT p53 and CHO cells
lack p53 expression (35), we investigated whether the loss of both
NHEJ and p53 signaling is required for efficient ENi L1 retro-
transposition in HCT116 cells. To accomplish this goal, we used
stable transfection to introduce a dominant-negative p53 con-
struct (DNp53) into the NHEJ-deficient HCT116 lines (36).
Remarkably, the resultant cell lines accommodated readily de-
tectable levels of ENi retrotransposition when transfected with
an endonuclease-deficient L1 construct (LRE3–H230A–EGFP)
(Fig. 4C). Thus, consistent with previous reports, efficient ENi
retrotransposition in mammalian cells seems to require muta-
tions that inactivate both NHEJ and p53 functions (14, 21).
Notably, in contrast to a previous report (37), we observed effi-
cient L1 retrotransposition in both p53-proficient and p53-de-
ficient HCT116 cells (Fig. 4 B and C).
We next used lentiviral shRNAs to knock down ATM expres-
sion in both parental and isogenic p53-deficient HCT116 cells
(Fig. 4D). As above, ATM-deficient HCT116 cells transfected
L1 retrotransposition (Fig. 4E). However, we did not observe ENi
retrotransposition in ATM/p53-deficient cells (Fig. 4F). Finally,
we demonstrated that shRNAs directed against the ATM-sig-
naling partners MRE11 and BRCA1 led to slight increases in
LRE3–EGFP retrotransposition (Fig. 4 G and H). These data
confirm that ATM deficiency leads to a modest increase in engi-
neered L1 retrotransposition.
We next sought to explain why L1 retrotransposition is en-
hanced in ATM-deficient cells. Notably, most genomic L1 inser-
shown to be highly processive in vitro (39), we hypothesized that
cellular DNA repair and damage sensing proteins may impact L1
perhaps more, L1 insertions in ATM-deficient cells.
As an initial test of the above hypothesis, we generated con-
structs that contained either one or two 500-bp spacer regions
(ColE1) between the LRE3 polyadenylation site and the
mEGFPI retrotransposition indicator cassette (Fig. 5A). A ret-
rotransposition of 1.6 kb (no spacer; LRE3*–EGFP), 2.1 kb (one
spacer; LRE3*–B1–EGFP), or 2.6 kb (two spacers; LRE3*–B2-
EGFP) would be required to activate EGFP expression (24).
Notably, the ColE1 spacer sequence does not prohibit engineered
L1 retrotransposition but, when located 3′ to the mneoI indicator,
reduces the detection of L1 retrotransposition (16). The apparent
decrease in L1 retrotransposition likely reflects the increased
length of retrotransposed products that are needed to allow the
expression of the retrotransposed mneoI indicator cassette.
We first tested the spacer constructs for retrotransposition in
retrotransposition by ∼3-fold in HCT116 cells and by ∼2.5-fold in
HeLa cells compared with the parental construct (Fig. S8 A and
B). Similarly, the addition of two spacers resulted in an ∼5-fold
and an ∼3.5-fold reduction of L1 retrotransposition in HCT116
cells and HeLa cells, respectively (Fig. S8 A and B) (16). The
presence of two spacers resulted in an incremental decrease in
EGFP-positive cells compared with one spacer; however, this
decrease was not statistically significant in a direct comparison
between one and two spacers (Fig. S8 A and B). Control PCR
reactions with primers flanking the inserted cassette (Fig. S8E,
primers cF, cR) confirmed the increased insertion size of the
vector and retrotransposed products containing the additional
spacers (Fig. S8C). Finally, PCR reactions with primers flanking
the intron in the EGFP indicator cassette (Fig. S8E, primers dF,
dR) confirmed the presence of the retrotransposed products in
each sample (Fig. S8D).
Given the above results, we hypothesized that the spacer con-
structs could be used as a proxy to determine the overall length of
L1 retrotransposition events in ATM-proficient and ATM-de-
ficient NPCs. As expected, the addition of one or two spacers
resulted in decreased L1 retrotransposition (∼1.5- and ∼2- to 3-
fold, respectively) in ATM-proficient NPCs (Fig. 5B, white bars).
In contrast, we did not observe a significant decrease in the ret-
rotransposition efficiency in ATM-deficient NPCs transfected
with the construct that contained one spacer (1.6% LRE3* vs.
1.7% LRE3*–B1 EGFP-positive cells) (Fig. 5B). Notably, we
HCT116-derived cell lines. (A) RT-PCR analysis confirmed the knockout of
NHEJ genes in each of the isogenic HCT116 cell lines. ATM shRNA-infected
cells showed a decrease in ATM expression compared with controls. (B)
Isogenic HCT116 lines defective for NHEJ, but harboring a WT p53, exhibited
robust LRE3–EGFP retrotransposition; however, they did not accommodate
the retrotransposition of mutant L1 constructs [ORF1 mutant (JM111–LRE3)
and endonuclease mutant (LRE3–H230A–EGFP)]. (C) LRE3–EGFP retro-
transposition in HCT116-derived cell lines expressing a dominant-negative
p53 (DNp53). Control DNp53 transfection shows robust L1 retrotransposition
similar to isogenic p53−/−lines (F), but no ENi retrotransposition was ob-
served (white bar). However, the combined loss of NHEJ (DNA-PKcs, DNA
Ligase IV, XLF) and p53 function led to detectable ENi retrotransposition.
(D) Western blot analyses in HCT116 (Left) and HCT116 p53−/−(Right) cells
infected with control and ATM shRNA lentiviruses. (E) HCT116 cells infected
with ATM shRNAs exhibit increased LRE3–EGFP retrotransposition compared
with cells infected with control shRNA. (F) In the p53−/−background,
knockdown of ATM does not result in ENi retrotransposition activity. How-
ever, a modest increase in LRE3–EGFP retrotransposition activity is observed
in the ATM shRNA p53−/−cell lines. (G) HCT116 cells were infected with
shRNAs against BRCA1 and MRE11, which led to slight increases in engi-
neered L1 retrotransposition. Error bars indicate SEM. *P < 0.05 and **P <
0.01 are a result of a one-way ANOVA with Bonferroni correction. (H) RT-PCR
analysis of HCT116 cells transfected with shRNA vectors targeting MRE11
and BRCA1. The expression of other NHEJ genes was not severely affected.
The pursuit of endonuclease-independent L1 retrotransposition in
Coufal et al.PNAS
| December 20, 2011
| vol. 108
| no. 51
observed a similar trend in data using the construct with two
spacers (LRE3*–B2–EGFP), although both control and ATM-
deficient NPCs also showedstatistically significant drops in EGFP
expression (Fig. 5B). These data are consistent with the hypoth-
esis that the resultant L1 retrotransposition insertions are longer
in ATM-deficient NPCs compared with ATM-proficient controls.
Notably, a similar finding indicated that deficiencies in NHEJ
resulted in longer insertions when assaying a zebrafish LINE
element in chicken DT40 cells (15). Thus, we speculate that ATM
may recognize intermediates that are generated during the pro-
cess of L1 integration as DNA damage and thereby reduce the
length of the resultant retrotransposition events (Fig. 5C).
Finally, we analyzed the copy number of endogenous, human-
specific L1s using a previously described quantitative (Taqman-
based) PCR (qPCR) methodology (12, 27). Briefly, we designed
primers that amplify ∼4,500 genomic L1s (12, 27). The primers
primarily target L1PA1/L1Hs elements (∼1,800 genomic ele-
ments); however, because of sequence homology, they also can
amplify L1s from slightly older subfamilies (e.g., L1PA2 and
L1PA3). Because hippocampal sections from AT patients exhib-
ited distorted ultrastructures, we used laser capture for neuronal
nuclei (NeuN)-stained nuclei. We then used qPCR to compare
changes in L1 copy numbers (12, 27) in postmortem hippocampal
samples from AT and age- and sex-matched control patients (n =
L1 ORF2 sequences normalized to nonmobile repetitive DNA
sequences (SATA, HERVH, 5sRNA gene). In all comparisons,
we observed a statistically significant increase in ORF2 copy
number in AT neurons compared with controls (Fig. 5D). Nota-
bly, we did not observe a marked difference between two other
repeat sequences, HERVH and SATA (12, 27), between AT
patients and controls (Fig. 5E).
It is worth pointing out caveats regarding the above analysis.
First, we are measuring changes only in L1 copy number, and the
characterization of endogenous retrotransposition events is needed
to unequivocally prove de novo, somatic L1 retrotransposition.
Second, and unfortunately, we could not obtain other somatic
tissues from AT patients, which limited our ability to estimate
changes in L1 copy number. With these caveats clearly stated, our
findings suggest that L1 retrotransposition might be exacerbated
in AT patients. Clearly, recent advances in DNA sequencing
technology should help definitively show if L1 retrotransposition
is indeed elevated in certain somatic cells in the brain.
Using various strategies, we have shown that ATM can impact L1
retrotransposition. First, we observed an increase in the retro-
transposition efficiency of engineered human L1s, likely by con-
ventional TPRT, in ATM-deficient hESCs, hESC-derived NPCs,
hCNS-SCns, and HCT116 cells. Second, we observed an increase
in the retrotransposition efficiency of an engineered human L1 in
ATM knockout transgenic mice. Third, we observed an increase
in L1 DNA copy number in brain samples from AT patients.
Our initial experiments suggested that engineered L1 retro-
transposition events in ATM-deficient cells may be of a longer
length compared with ATM-proficient controls. However, we
cannot exclude that ATM deficiency allows cells to accommodate
more L1 retrotransposition events per cell. Finally, although we
did not observe major changes in the cell cycle, growth, pro-
liferation, or survival rates between ATM-proficient and ATM-
deficient cells, it is formally possible that ATM deficiency could
render cells more tolerant to L1-induced toxicity (29). Clearly,
further experiments are required to determine why ATM de-
ficiency results in higher levels of retrotransposition.
During the course of our studies, we also analyzed L1 retro-
transposition in NHEJ- and p53-defective human cell lines. Al-
though previous studies suggested that p53 limits L1 retro-
transposition by inducing apoptosis (37), we observed similar L1
retrotransposition efficiencies in p53-proficient and p53-deficient
HCT116 cells; these data compel a re-examination of the im-
portance of p53 in L1 retrotransposition.
Consistent with previous studies in chicken DT40 cells (15),
we did not observe an increase in ENi retrotransposition in
DNA-PKcs, XRCC4-like factor (XLF), or DNA Ligase IV-deficient
HCT116 cell lines. However, each of these HCT116 cell lines
when p53 function was abrogated by the expression of a domi-
nant-negative allele of p53. The latter findings are consistent
with previous experiments conducted in XRCC4 and DNA-PKcs
JM111 LRE3* LRE3*-B1 LRE3*-B2
% EGFP Positive Cells
Relative ORF2 DNA Content
Relative ORF2 DNA Content
Relative DNA Content
Relative ORF2 DNA Content
HUES6-derived NPCs and multiplex qPCR analysis of L1 DNA copy number. (A)
A derivative of LRE3–EGFP was created where the BamH1 site in ORF2 is si-
lently mutated (*BamH1). A second BamH1 site immediately upstream of the
polyadenylation sequence (pA) was used to insert one or two copies of
two ColE1 spacers resulted in decreasing L1–EGFP expression. There was
a statistically significant decrease in EGFP expression between constructs with
zero and one copy of ColE1 in control shRNA-infected cells; however, there
was no change in the amount of EGFP expression in ATM-deficient cells re-
gardless of the ATM shRNA used. Using a one-way t test, we compared the
each sample (n = 6). (C) A schematic model for the role of ATM in L1 retro-
transposition. L1 retrotransposition through TPRT involves first-strand
nicking by the ORF2p endonuclease and priming by the exposed 3′ hydroxyl
group for reverse transcription and likely is followed by second-strand
cleavage and second-strand L1 cDNA synthesis. In our model, ATM is involved
in recognizing the DNA breakage intermediate created during L1 integration
and is implicated in its resolution and repair. Loss of ATM leads to either more
or longer insertions as a result. (D) Results from the multiplex qPCR analysis of
L1 sequences from the hippocampus. The ratio of ORF2 to internal control
represents the amount of L1 ORF2 DNA sequence in each sample relative to
the amount of another multicopy DNA control, such as 5S RNA sequence,
HERVH, and α-satellite. Under these conditions, the copy number of L1 ORF2
sequences was higher in the AT hippocampal neurons (NeuN-positive) com-
pared with those in control hippocampal neurons. (E) Multiplexing of control
HERVHprimerswithsatellite α(SATA)primers indicatednosignificantchange
in copy number; P ≤ 0.15. Error bars in all panels indicate SEM. *P < 0.05 and
**P < 0.005 were obtained using a one-way t test.
An assay to detect changes in L1 insertion sizes in ATM-deficient,
| www.pnas.org/cgi/doi/10.1073/pnas.1100273108Coufal et al.
deficient CHO cells that lack p53 function (14, 21). Thus, it is Download full-text
L1 to use endogenous DNA lesions as substrates for the alter-
native ENi retrotransposition pathway in mammalian cells.
Notably, Gasior and colleagues (33) have previously reported
that ATM is required for L1 retrotransposition in immortalized
ATM-deficient human fibroblasts. Those data are in stark con-
trast to what we observed in hESCs, NPCs, hCNS-SCns, HCT116
cells, and ATM-KO mice; these discrepancies warrant further
investigation. It is worth highlighting differences between our
studies. First, Gasior and colleagues used an L1 retrotrans-
position assay based on the generation of G418-resistant foci;
they noted a decrease in G418-resistant foci formation with the
control neoR vector in ATM-deficient cells compared with
a control cell line (33). Thus, ATM-deficient fibroblasts either
may be more sensitive to experimental manipulation, or may re-
spond differently to L1 retrotransposition compared with other
cell types. Previous reports have indicated that L1 retro-
transposition occurs at very low levels in fibroblasts (40),
and we did not observe L1 retrotransposition in skin in our trans-
genic mouse model. Finally, it is possible that our use of shRNAs
may provoke cellular perturbations that result in compensatory
mechanisms; however, the finding of increased L1 retro-
transposition in ATM-KO mice leads us to conclude that ATM is
not strictly required for L1 retrotransposition.
Our data provide evidence that ATM has a role in L1 retro-
transposition in neuronal and nonneuronal cell types. It is
tempting to speculate that L1 retrotransposition may be involved
in diverse neurological disease processes. Whether increased L1
retrotransposition contributes to the neurodegenerative disease
process of AT remains to be determined, and may merely be
a consequence, rather than an actor, in the course of disease.
Detailed methods and data collection are provided in SI Methods.
Animals. The L1RP–EGFP transgenic mouse (13) was crossed into an ATM-
deficient background (23).
Constructs and PCR. ATM lentiviral shRNA constructs were commercially
available. Cells were transfected with L1s containing an EGFP retro-
transposition cassette in a modified version of pCEP4 (Invitrogen) (24). Adult
human tissues from AT patients and control samples were obtained from the
National Institute of Child Health and Human Development Brain and Tissue
Bank for Developmental Disorders (University of Maryland). Inverse PCR and
PCR cycling conditions were described previously (12, 27, 32).
Cell Culture and Analysis. HUES6 hESCs and NPCs were derived as described
(12). Lentivirus was produced as previously described (13).
ACKNOWLEDGMENTS. We thank J. Simon for schematics, M. L. Gage and
N. Leff for editorial comments, Prof. M. Weitzman for the kind gifts
of cell lines, K. Stecker for laser capture, E. Mejia for tissue assistance, and
L. Randolph-Moore for molecular advice. F.H.G.’s laboratory is supported by
the Mather’s Foundation, California Institute for Regenerative Medicine,
and National Institutes of Health/National Institute of Neurological Disorders
and Stroke (MH088485) J.L.G.-P’s laboratory is supported by Instituto Salud
Carlos III-Consejeria de Salud Junta de Andalucia- Fondo Europeo Desarrollo
Regional (EMER07/056); by Marie Curie International Reintegration Grant
Action FP7-PEOPLE-2007-4-3-IRG; by Consejeria de Innovacion Ciencia y
Economia- Fondo Europeo Desarrollo Regional (P09-CTS-4980); by Proyectos
en Salud-FEDER PI-002 from Junta de Andalucia (Spain); and by the Spanish
Ministry of Health (FIS-FEDER PI08171). A.R.M. is supported by the National
Institutes of Health through the National Institutes of Health Director’s New
Innovator Award Program (1-DP2-OD006495-01) and by the Emerald Foun-
dation. J.V.M. is supported by National Institutes of Health Grants GM06518
and GM082970. J.V.M. is an Investigator of the Howard Hughes Medical
1. Ostertag EM, Kazazian HH, Jr. (2001) Biology of mammalian L1 retrotransposons.
Annu Rev Genet 35:501–538.
2. Luan DD, Korman MH, Jakubczak JL, Eickbush TH (1993) Reverse transcription of
R2Bm RNA is primed by a nick at the chromosomal target site: A mechanism for non-
LTR retrotransposition. Cell 72:595–605.
3. Beck CR, et al. (2010) LINE-1 retrotransposition activity in human genomes. Cell 141:
4. Brouha B, et al. (2003) Hot L1s account for the bulk of retrotransposition in the hu-
man population. Proc Natl Acad Sci USA 100:5280–5285.
5. Moran JV, et al. (1996) High frequency retrotransposition in cultured mammalian
cells. Cell 87:917–927.
6. Hohjoh H, Singer MF (1997) Sequence-specific single-strand RNA binding protein
encoded by the human LINE-1 retrotransposon. EMBO J 16:6034–6043.
7. Martin SL, Bushman FD (2001) Nucleic acid chaperone activity of the ORF1 protein
from the mouse LINE-1 retrotransposon. Mol Cell Biol 21:467–475.
8. Feng Q, Moran JV, Kazazian HH, Jr., Boeke JD (1996) Human L1 retrotransposon
encodes a conserved endonuclease required for retrotransposition. Cell 87:905–916.
9. Mathias SL, Scott AF, Kazazian HH, Jr., Boeke JD, Gabriel A (1991) Reverse tran-
scriptase encoded by a human transposable element. Science 254:1808–1810.
10. Kazazian HH, Jr. (1998) Mobile elements and disease. Curr Opin Genet Dev 8:343–350.
11. Beck CR, Garcia-Perez JL, Badge RM, Moran JV (2011) LINE-1 elements in structural
variation and disease. Annu Rev Genomics Hum Genet 12:187–215.
12. Coufal NG, et al. (2009) L1 retrotransposition in human neural progenitor cells. Na-
13. Muotri AR, et al. (2005) Somatic mosaicism in neuronal precursor cells mediated by L1
retrotransposition. Nature 435:903–910.
14. Morrish TA, et al. (2002) DNA repair mediated by endonuclease-independent LINE-1
retrotransposition. Nat Genet 31(2):159–165.
15. Suzuki J, et al. (2009) Genetic evidence that the non-homologous end-joining repair
pathway is involved in LINE retrotransposition. PLoS Genet 5:e1000461.
16. Gilbert N, Lutz-Prigge S, Moran JV (2002) Genomic deletions created upon LINE-1
retrotransposition. Cell 110:315–325.
17. Gilbert N, Lutz S, Morrish TA, Moran JV (2005) Multiple fates of L1 retrotransposition
intermediates in cultured human cells. Mol Cell Biol 25:7780–7795.
18. Symer DE, et al. (2002) Human l1 retrotransposition is associated with genetic in-
stability in vivo. Cell 110:327–338.
19. Han K, et al. (2005) Genomic rearrangements by LINE-1 insertion-mediated deletion
in the human and chimpanzee lineages. Nucleic Acids Res 33(13):4040–4052.
20. Lin C, et al. (2009) Nuclear receptor-induced chromosomal proximity and DNA breaks
underlie specific translocations in cancer. Cell 139:1069–1083.
21. Morrish TA, et al. (2007) Endonuclease-independent LINE-1 retrotransposition at
mammalian telomeres. Nature 446:208–212.
22. Shiloh Y (2001) ATM (ataxia telangiectasia mutated): Expanding roles in the DNA
damage response and cellular homeostasis. Biochem Soc Trans 29:661–666.
23. Barlow C, et al. (1996) Atm-deficient mice: A paradigm of ataxia telangiectasia. Cell
24. Ostertag EM, Prak ET, DeBerardinis RJ, Moran JV, Kazazian HH, Jr. (2000) De-
termination of L1 retrotransposition kinetics in cultured cells. Nucleic Acids Res 28:
25. Ostertag EM, et al. (2002) A mouse model of human L1 retrotransposition. Nat Genet
26. Muotri AR, Zhao C, Marchetto MC, Gage FH (2009) Environmental influence on L1
retrotransposons in the adult hippocampus. Hippocampus 19:1002–1007.
27. Muotri AR, et al. (2010) L1 retrotransposition in neurons is modulated by MeCP2.
28. Garcia-Perez JL, et al. (2010) Epigenetic silencing of engineered L1 retrotransposition
events in human embryonic carcinoma cells. Nature 466:769–773.
29. Wallace NA, Belancio VP, Deininger PL (2008) L1 mobile element expression causes
multiple types of toxicity. Gene 419(1–2):75–81.
30. Wei W, et al. (2001) Human L1 retrotransposition: Cis preference versus trans com-
plementation. Mol Cell Biol 21:1429–1439.
31. Uchida N, et al. (2000) Direct isolation of human central nervous system stem cells.
Proc Natl Acad Sci USA 97:14720–14725.
32. Garcia-Perez JL, et al. (2007) LINE-1 retrotransposition in human embryonic stem cells.
Hum Mol Genet 16:1569–1577.
33. Gasior SL, Wakeman TP, Xu B, Deininger PL (2006) The human LINE-1 retrotransposon
creates DNA double-strand breaks. J Mol Biol 357:1383–1393.
34. Topaloglu O, Hurley PJ, Yildirim O, Civin CI, Bunz F (2005) Improved methods for the
generation of human gene knockoutand knockin cell lines. Nucleic AcidsRes 33(18):e158.
35. Moro F, et al. (1995) p53 expression in normal versus transformed mammalian cells.
36. Irwin M, et al. (2000) Role for the p53 homologue p73 in E2F-1-induced apoptosis.
37. Haoudi A, Semmes OJ, Mason JM, Cannon RE (2004) Retrotransposition-competent
human LINE-1 induces apoptosis in cancer cells with intact p53. J Biomed Biotechnol
38. Pavlícek A, Paces J, Zíka R, Hejnar J (2002) Length distribution of long interspersed
nucleotide elements (LINEs) and processed pseudogenes of human endogenous retro-
viruses: Implications for retrotransposition and pseudogene detection. Gene 300(1–2):
39. Piskareva O, Schmatchenko V (2006) DNA polymerization by the reverse transcriptase
of the human L1 retrotransposon on its own template in vitro. FEBS Lett 580:661–668.
40. Kubo S, et al. (2006) L1 retrotransposition in nondividing and primary human somatic
cells. Proc Natl Acad Sci USA 103:8036–8041.
Coufal et al.PNAS
| December 20, 2011
| vol. 108
| no. 51