Cernunnos, a Novel Nonhomologous End-Joining Factor, Is Mutated in Human Immunodeficiency with Microcephaly
DNA double-strand breaks (DSBs) occur at random upon genotoxic stresses and represent obligatory intermediates during physiological DNA rearrangement events such as the V(D)J recombination in the immune system. DSBs, which are among the most toxic DNA lesions, are preferentially repaired by the nonhomologous end-joining (NHEJ) pathway in higher eukaryotes. Failure to properly repair DSBs results in genetic instability, developmental delay, and various forms of immunodeficiency. Here we describe five patients with growth retardation, microcephaly, and immunodeficiency characterized by a profound T+B lymphocytopenia. An increased cellular sensitivity to ionizing radiation, a defective V(D)J recombination, and an impaired DNA-end ligation process both in vivo and in vitro are indicative of a general DNA repair defect in these patients. All five patients carry mutations in the Cernunnos gene, which was identified through cDNA functional complementation cloning. Cernunnos/XLF represents a novel DNA repair factor essential for the NHEJ pathway.
Cernunnos, a Novel Nonhomologous
End-Joining Factor, Is Mutated in Human
Immunodeﬁciency with Microcephaly
gina de Chasseval,
Franc¸ oise le Deist,
Jean-Pierre de Villartay,
and Patrick Revy
pital Necker-Enfants Malades, U768 Unite
veloppement Normal et Pathologique du Syste
me Immunitaire, Paris,
F-75015 France; Universite
Paris Descartes, Faculte
Descartes, Paris, F-75005 France
Immunology Division, Hacettepe University, Ankara, Turkey
Clinica Pediatrica and Istituto di Medicina Moleculare ‘‘Angelo Nocivelli,’’ Universita
di Brescia, Italy
matologie et Oncologie Pe
diatrique, CHU Saint Etienne, France
Children’s Hospital, Department of General Pediatrics, University Hospital Schleswig-Holstein, Campus Kiel, Germany
pital Necker-Enfants Malades, Unite
d’Immunologie et d’He
matologie, Paris, F-75015 France
DNA double-strand breaks (DSBs) occur at
random upon genotoxic stresses and repre-
sent obligatory intermediates during physio-
logical DNA rearrangement events such as
the V(D)J recombination in the immune sys-
tem. DSBs, which are among the most toxic
DNA lesions, are preferentially repaired by
the nonhomologous end-joining (NHEJ) path-
way in higher eukaryotes. Failure to properly
repair DSBs results in genetic instability,
developmental delay, and various forms of
immunodeﬁciency. Here we describe ﬁve pa-
tients with growth retardation, microcephaly,
and immunodeﬁciency characterized by
a profound T+B lymphocytopenia. An in-
creased cellular sensitivity to ionizing radia-
tion, a defective V(D)J recombination, and
an impaired DNA-end ligation process both
in vivo and in vitro are indicative of a general
DNA repair defect in these patients. All ﬁve
patients carry mutations in the Cernunnos
gene, which was identiﬁed through cDNA
functional complementation cloning. Cer-
nunnos/XLF represents a novel DNA repair
factor essential for the NHEJ pathway.
DNA double-strand breaks (DSBs) are among the most toxic
DNA lesions caused by cell-intrinsic sources such as replica-
tion errors or other DNA-damaging agents naturally present
in cells, including reactive oxygen species. DSBs can also re-
sult from exposure to a variety of extrinsic factors, including
ionizing radiation (IR). Lastly, DSBs represent obligatory inter-
mediates of physiological DNA rearrangement processes
taking place during the development and maturation of the
adaptive immune system (V[D]J recombination and immuno-
globulin [Ig] heavy chain class switch recombination [CSR])
(Sancar et al., 2004). The V(D)J recombination is a somatic
DNA rearrangement of Variable, Diversity, and Joining gene
segments encoding the T and B cell antigen receptor loci.
The lymphoid-speciﬁc recombination activating gene (RAG)
1 and 2 encoded endonuclease initiates the reaction by intro-
ducing a DSB at recombination-speciﬁc sequences (RSS)
that ﬂank all V, D, and J gene units (Bassing et al., 2002).
The efﬁcient repair of DNA DSBs introduced during these re-
actions is required to maintain genome integrity, thus pre-
venting the development of cancer and other ‘‘DNA instabil-
ity’’ disorders. DSBs are primarily repaired by the accurate
homologous recombination (HR) in yeast and by the error-
prone nonhomologous end-joining (NHEJ) pathway in higher
eukaryotes (Sancar et al., 2004). To date, six mammalian fac-
tors have been identiﬁed as constituting the core NHEJ appa-
ratus involved in V(D)J recombination. The Ku70/Ku80 heter-
odimer binds to DNA ends, present at RAG1/2-generated
DSBs, and recruits the ATM-related kinase DNA-PKcs (Got-
tlieb and Jackson, 1993; Dynan and Yoo, 1998). DNA-PKcs
then phosphorylates and activates the Artemis endonucle-
ase, the enzyme required to resolve DNA hairpin structures
created upon RAG1/2-speciﬁc DNA cleavage (Ma et al.,
2002). The XRCC4/DNA-LigaseIV (Lig4) complex is respon-
sible for the ﬁnal ligation step (Grawunder et al., 1997).
The fundamental role of NHEJ factors in the immune system
has been recognized through genetic deletion in mice and
studies of speciﬁc human conditions (de Villartay et al., 2003).
In all animal models, NHEJ defects result in the arrest of B
and T lymphocyte maturation, accompanied by an embryonic
lethality in the case of XRCC4 and Lig4 deﬁciencies (Dudley
et al., 2005; Revy et al., 2005). In humans, null mutations in
Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc. 287
Artemis result in RS-SCID, characterized by a complete ab-
sence of peripheral T and B lymphocytes and an increased
cellular sensitivity to IR (Moshous et al., 2001). Variable immu-
nodeﬁciency, developmental delay, chromosome alterations,
and microcephaly have also been attributed to hypomorphic
mutations in the Lig4 gene in humans (O’Driscoll et al., 2001;
Buck et al., 2006). Finally, a radiosensitive SCID condition
(patient 2BN) without mutation in any of the known NHEJ
factors was identiﬁed, supporting the existence of additional
uncharacterized DNA repair factors (Dai et al., 2003).
With the aim of identifying new factors involved in general
DNA repair, we undertook a systematic survey of human con-
ditions characterized by developmental anomalies, such as
microcephaly, associated with various degrees of immune
deﬁciency and/or the onset of lymphopoietic malignancy. In
this article, we describe a new syndrome of human combined
immunodeﬁciency (CID) associated with microcephaly and
increased cellular sensitivity to IR. This condition is caused
by a general DNA repair defect owing to mutations in a novel
NHEJ factor encoding gene, Cernunnos.
A New Syndrome Associated with T and B Cell
Combined Immunodeﬁciency, Growth Retardation,
The main clinical features characterizing the patients in-
cluded in this study were growth retardation, microcephaly,
and immunodeﬁciency (for detailed case reports, see Exper-
imental Procedures and Tables 1 and 2). Severe growth retar-
dation and dystrophy were observed in four of the ﬁve pa-
tients, and microcephaly was present in all ﬁve patients at
birth. Dysmorphic features and various malformations were
observed in four of the patients (Table 1). Recurrent infections
of bacterial, viral, and/or parasitic origin occurred in all pa-
tients and were lethal in two. Laboratory analysis indicated
a mild to severe B (CD19+) and T (CD3+) lymphocytopenia
in all patients, whereas the NK cell subset was not affected
(Table 2). Circulating B cell counts declined with age from
close to normal values to undetectable levels, and most pa-
tients displayed hypogammaglobulinemia affecting IgG and
IgA, accompanied by ﬂuctuating levels of IgM (Table 2). The
T lymphocyte subsets were composed only of memory T
cells (CD45RO positive) with impaired functions determined
by low in vitro PHA mitogen-induced T cell proliferation.
In association with microcephaly, the absence of circulat-
ve T cells and the progressive disappearance of B
cells were indicative of a molecular defect affecting the mat-
uration of the immune system as observed in Lig4 deﬁcient
or Nijmegen breakage syndrome (NBS) patients.
Increased Radiosensitivity of Patients’ Fibroblasts
Lig4 and NBS conditions are two DNA repair deﬁciencies
characterized by an increased cellular sensitivity to IR (Varon
et al., 1998; O’Driscoll et al., 2001; Buck et al., 2006). Pri-
mary skin ﬁbroblasts from four of the tested patients (P1-
P4) exhibited an increased sensitivity to g ray exposure as
Table 1. Clinical Features of Patients
Patients Origin Consanguinity Microcephaly
Features Infections Autoimmunity Outcome
P1 French no yes yes chromosomal
P2 Turkish +, 1st degree yes yes no bacterial
no died at
P3* Turkish +, 1st degree yes yes birdlike face,
P4* Turkish +, 1st degree yes yes birdlike face bacterial
P5 Italian +, 3rd degree yes yes birdlike face,
* P3 and P4 are siblings.
288 Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc.
compared to control cells in a clonogenic assay (Figure 1A).
Although the level of IR sensitivity was variable among the
patients, it was equivalent or sometimes more pronounced
than what was observed in other radiosensitive conditions,
such as Ataxia telangiectasia (A-T), NBS, Lig4, or Artemis
deﬁciencies (Figure 1A and data not shown).
IR-Induced Foci Formation Is Normal
in Patients’ Fibroblasts
The phosphorylation of the histone variant H2AX (then called
gH2AX) occurs rapidly in response to a DNA-DSB (Rogakou
et al., 1998) and is essential to keeping DNA ends in close
proximity and to stabilizing the association of DNA-repair
factors such as the MRE11-RAD50-NBS1 (MRN) complex
and 53BP1 at the site of the damage (Bassing and Alt,
2004), forming ionizing radiation-induced foci (IRIF). gH2AX,
MRE11, 53BP1, and RAD51 IRIFs were equally well formed
in P1-P4 and control ﬁbroblasts 2 hr after 10 Gy irradiation
(Figures 1B and S1 and data not shown). This result sug-
gests that the increased sensitivity to IR does not result
from a defect in the initial DNA damage sensing. In contrast
to normal control cells, gH2AX signal persisted in patients’
cells 24 hr following DSB induction, as observed by both mi-
croscopy and ﬂuorocytometry, signifying a major DNA repair
defect in these cells (Figure 1B and data not shown).
Normal Cell-Cycle Checkpoints in Patients’
Fibroblasts Following IR
IR induce cell-cycle checkpoints, which can be analyzed at
three critical steps: the inhibition of entry into the S phase
(G1/S checkpoint), the S phase progression, and the inhibi-
tion of entry into mitosis (G2/M checkpoint). The G1/S
checkpoint following a 5 Gy IR was comparable in all pa-
tients’ and controls’ ﬁbroblasts with a 75% mean inhibition
of S phase entry (Figures 1C and 1D). This contrasted with
the defect of this checkpoint in ﬁbroblasts from A-T or
NBS patients (Figures 1C and 1D). The S phase progression,
analyzed through radioresistant DNA synthesis (RDS) assay,
was also similarly inhibited in patients’ and controls’ ﬁbro-
blasts following IR (Figure S2). Lastly, we examined the IR-
induced G2/M checkpoint by analyzing the phosphorylation
of the Histone 3 (H3) (Figure 1E). H3 phosphorylation (P-H3)
Table 2. Immunological Characteristics of Patients
Patients P1 P2 P3 P4 P5
Age at diagnosis
14 2 13 2 7
750 (1400–3300) 1220 (2300–5400) 1610 (1400–3300) 2570 (2300–5400) 1101 (1900–3700)
630 (1000–2200) 730 (1400–3700) 870 (1000–2200) 591 (1400–3700) 693 (1200–2600)
428 (530–1300) 150 (650–1500) 490 (530–1300) 334 (650–1500) 319 (530–1300)
248 (330–920) 559 (370–1100) 370 (330–920)77(370–1100) 330 (330–920)
92 (4–23)54(4–16)65(4–23) n.i. 83 (4–21)
0(110–570)75(390–1400)0(110–570) 154 (390–1400)44(110–570)
n.d. 281 (92–918) 402 (42–726) 1336 (92–918) 195 (76–629)
T cell proliferation
30 (>40) 5.8 (>40) 119 (>40)84(>40)16(>40)
Ig values before
2.4 (6.4–14.2) <0.33 (5.2–10.8) <0.5 (2.8–6.8)* 2.37 (maternal;
<0.06 (0.52–2.2) <0.06 (0.36–1.65) <0.5 (0.1–0.58)* <0.05 (0.02–0.22)** 0.06 (0.65–2.4)
6.8 (0.4–1.8) 1.39 (0.72–1.60) 0.15 (0.4–0.84)* 0.27 (0.36–0.56)** 1.05 (0.6–1.75)
n.d.: not done.
n.i.: not interpretable.
Age-matched control values are from Shearer et al. (Shearer et al., 2003) and are shown in italics in parentheses. Tested at *5 or
**3 months of age.
in counts/ml; presence of maternal T cells was excluded by HLA typing of sorted T cells.
Phytohemagglutinin (PHA)-induced proliferations are expressed as counts per minute 10
Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc. 289
Figure 1. Cellular Response to DNA Damage
(A) Survival of primary ﬁbroblasts after IR (up to 3 Gy). Results are expressed as the fraction of colony-forming cells in relation to unirradiated cells. Each point
represents the mean value and standard deviation of three separate determinations. Control radiosensitive ﬁbroblasts were from Lig4 (L4), A-T, and NBS
(B) Left panel: Primary ﬁbroblasts from the control and P1 were seeded onto glass slides and either left untreated or else irradiated with 10 Gy and ﬁxed 2or
24 hr postirradiation before staining with anti-gH2AX. Nuclei were stained with DAPI. Right panel: gH2AX detection by ﬂuorocytometry was performed in
untreated ﬁbroblasts (black curve) and 30 min (red curve) or 24 hr (green curve) post-DSB induction. Persistence of gH2AX signal at 24 hr posttreatment in
patients’ cells in both experiments is indicative of a general DNA-repair defect.
290 Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc.
correlates with chromosome condensation occurring imme-
diately before mitosis (Wei et al., 1999). The 80% decrease
in P-H3-positive cells in patients’ and controls’ ﬁbroblasts
following a 5 Gy IR (Figures 1E and 1F) indicated a normal
G2/M checkpoint and contrasted with the impaired G2/M
checkpoint in A-T cells. Altogether, these results demon-
strated that the IR-induced biochemical events leading to
the various cell-cycle checkpoints were not altered in this
group of patients.
These observations suggested that these patients, al-
though presenting clinical manifestations similar to those
seen in A-T or NBS conditions, clearly differ from these two
settings. In addition, the normal DNA-damage sensing, IRIF
detection, and IR-induced cell-cycle control indicate that
the increased cellular radiosensitivity and gH2AX persistence
result from an intrinsic defect in DNA repair.
Defect in the NHEJ Process
In Vivo DNA-Ligation Assay
To analyze the ability of patients’ cells to join double-
stranded DNA ends by the NHEJ pathway, ﬁbroblasts
were transfected with restriction-enzyme-digested, linear-
ized plasmids containing either blunt-blunt or incompatible
overhang ends. Recircularized plasmids were recov-
ered 48 hr after transfection, and their junctions were ana-
lyzed by DNA sequencing. The junctions in plasmids recov-
ered from two of the control cell lines (C-1 and C-2) were
relatively accurate (71% to 95%), independently of the nature
of the DNA-ends (Figures 2A, 2B, and 2C). In striking con-
trast, in plasmids recovered from the three of the tested pa-
tients’ cell lines (P1, P2, and P5), the junctions were rarely ac-
curate (0% to 16%), irrespective of the nature of the DNA
ends (Figures 2A, 2B, and 2C). These junctions involved var-
ious degrees of nucleotide deletion. The mean number of nu-
cleotide loss in inaccurate junctions was signiﬁcantly higher
(p < 0.001) in plasmids recovered from patients’ cells than
in those recovered from controls (Figure 2D).
Because of its critical dependency on a functional NHEJ ma-
chinery, we next analyzed V(D)J recombination in patients
ﬁbroblasts upon transfection of the pRecCS extrachromo-
somal V(D)J recombination substrate and the human
full-length lymphoid-speciﬁc RAG1- and RAG2-expressing
constructs (Nicolas et al., 1998). Coding and signal-join for-
mation, identiﬁed by speciﬁc PCR ampliﬁcation of recovered
rearranged pRecCS plasmids, was similar or reduced in pa-
tients’ ﬁbroblasts as compared to control (Figure 3A). Re-
duced V(D)J recombination activity was further documented
in P1 and P2 using an in-chromosome V(D)J assay (Fig-
ure S4). Sequence analysis of recovered, rearranged sub-
strates revealed a slight increase in nucleotide loss in coding
joins (Figure S3A) compare to controls but not the major al-
terations typically observed in other NHEJ deﬁciency situa-
tions (Bogue et al., 1997; Rooney et al., 2002; Rooney
et al., 2003). In particular, these junctions were not charac-
terized by extensive nucleotide loss or long P nucleotides.
In contrast, the ﬁdelity of signal joins was severely impaired
in patients’ cells. This was ﬁrst demonstrated by the resis-
tance to ApalI digestion, a restriction-enzyme site created
by the perfect fusion of RSS in normal signal joins (Figure 3B).
Sequence analysis revealed that 45% of signal joins were im-
precise, with nucleotide loss ranging from 6 to 17 bp (Fig-
ure S3B). Several sequences suggested the possible usage
of microhomology during joining. These V(D)J recombination
anomalies are a characteristic feature previously recognized
in NHEJ-deﬁcient ﬁbroblasts from Lig4 patients (Badie et al.,
1997; Riballo et al., 2001; Buck et al., 2006) and in 2BN cells
(Dai et al., 2003).
The V(D)J deﬁciency in these patients is less severe than in
Artemis-deﬁcient RS-SCIDs and probably accounts for the
presence of low counts of B and T cells in these patients.
In Vitro NHEJ Assay
Lastly, we analyzed the in vitro end-joining of linearized plas-
mid DNA by using cell-free extracts (Baumann and West,
1998; Diggle et al., 2003). This DNA-end ligation assay,
which results in the formation of DNA concatemers when us-
ing control extracts (Figure 3C, lane C), requires a functional
NHEJ apparatus, as demonstrated by the absence of DNA
oligomers when using extract from a Lig4 deﬁcient patient
(L4). Defective DNA multimerization when using extracts
from P1-P5 (Figure 3C) strongly argued for a NHEJ defect
in these patients. Importantly, the patients’ extracts cross-
complemented the L4 extract in this assay (Figure 3D),
thus eliminating a DNA Lig4 deﬁciency in P1-P5. Finally,
pair-wise mixing of P1-P5 extracts failed to complement
the absence of DNA end-joining (Figure 4E), suggesting
that the molecular defect may be identical in the ﬁve patients.
In conclusion, these data indicated that the ﬁve patients
present with an identical general DNA-repair defect affecting
the NHEJ machinery. This molecular defect translates into
an increased cellular sensitivity to DNA-damaging agents
in vitro and subsequently into a major failure in the develop-
ment of both B and T lymphocytes in vivo, thus resulting in
a combined immunodeﬁciency. Although these features
are known to be shared by Lig4-deﬁcient patients, a Lig4
gene defect was excluded in all ﬁve of our patients.
(C) BrdU incorporation in proliferating cells and DNA content were analyzed either in untreated or in 5 Gy irradiated cells from a control, an A-T patient, and
P1 as a means of analyzing the G1/S cell-cycle checkpoint. While A-T cells did not arrest cycling following IR, P1 cells did.
(D) The fraction of cells in early S phase was determined by FACS analysis with and without 5 Gy IR, as in (C), and was used to calculate the inhibition of S
phase entry. A-T and NBS cells were defective in G1/S checkpoints, whereas P1-P5 cells behaved like normal controls.
(E) The phosphorylation of histone H3 and DNA content of either untreated or of 5 Gy-irradiated P1 cells were analyzed by FACS. While A-T cells retained
phospho-H3-positive cells following IR, indicating their impaired G2/M checkpoint, P1 cells did not.
(F) The fraction of phospho-H3 positive cells in IR and untreated cells was used to calculate the % inhibition of entry into mitosis in A-T and control cells as
well as into cells from P1, P2, and P5. While A-T cells demonstrate a defect in the G2/M checkpoint following IR, patients’ cells behaved like normal controls.
Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc. 291
Cloning of Cernunnos cDNA
Genetic analysis using highly polymorphic microsatellite
markers allowed us to formally exclude a role for the six
known NHEJ factor-encoding genes Ku70, Ku80, DNA-
PKcs, Lig4, XRCC4, and Artemis in the four patients (P2-
P5) who were born to three consanguineous families since
transmission of the disease was independent from the seg-
regation of markers ﬂanking the above-mentioned genes
(not shown). We designed a functional complementation
cloning strategy based on the rescue of the increased cel-
lular sensitivity to DNA-damaging agents to identify the de-
fective gene. A human thymic cDNA library was introduced
into P2 ﬁbroblasts by means of retroviral transduction. Eight
separate pools of transduced cells were then treated nine
times with the radiomimetic drug bleomycin (0.5 mg/ml)
over a period of 88 days. Transduced cDNAs were recov-
ered by PCR ampliﬁcation of genomic DNA with vector-spe-
ciﬁc ﬂanking primers. This resulted in discrete bands (from 1
Figure 2. DNA-End Ligation of Linearized Plasmids Is Defective in Patients’ Fibroblasts
(A) DNA sequence of junctions was obtained after ligation of blunt-blunt linearized plasmids in ﬁbroblasts from two controls (C-1 and C-2) and patients P1,
P2, and P5. The occurrence of a given sequence is indicated on the right.
(B) DNA sequence of junctions was obtained after ligation of 3
linearized plasmid in ﬁbroblasts from two controls (C-1 and C-2) and patients P1 and P2.
The occurrence of a particular sequence is indicated on the right.
(C) Percentages (mean values and standard deviations) of accurate junctions in clones from each set of experiments depicted in (A) and (B). Accurate junc-
tions are deﬁned as the sequences preponderantly shared by the two controls.
(D) Chart representing the mean nucleotide loss in inaccurate junctions observed under the two experimental conditions.
292 Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc.
to 6) in all eight pools of transduced cells (data not shown),
which suggested a speciﬁc outgrowth of cells that had re-
ceived a cDNA complementing their endogenous gene de-
fect. DNA sequencing identiﬁed an identical cDNA present
in the PCR products from seven of the pools.
This cDNA (EMBL:AJ972687) covers 2063 nucleotides
and encodes a hypothetical protein of 299 amino acids
(aa), which we named Cernunnos (Figure 4A). The Cernun-
nos gene (GeneID: 79840), located on human chromosome
2q35, is composed of eight exons with sizes ranging from 59
bp to 1130 bp (Figure 4B). The Cernunnos gene has never
been described previously (except for the in silico-derived
hypothetical protein FLJ12610), and database searching
did not disclose any obvious homology with other known
proteins or functional domains. cDNA cloning of the murine
Cernunnos counterpart revealed a 74% protein identity
over the whole sequence, suggesting a high degree of con-
servation among higher eukaryotes (not shown). RT-PCR
analysis of a panel of 15 cDNAs representing a wide range
of human tissues demonstrated that Cernunnos is ubiqui-
tously expressed (Figure 4C). Finally, immunoﬂuorescence
analysis of ectopically expressed, V5 epitope-tagged Cer-
nunnos in ﬁbroblasts revealed a predominant nuclear locali-
zation (Figure 4D). These last two ﬁndings are compatible
with Cernunnos being a general DNA repair factor.
Mutations affecting the Cernunnos encoding gene were
identiﬁed in all ﬁve patients (Figure 4B). P1 carries two mis-
sense heterozygous mutations, C259G (cDNA) in exon 2
and T457C in exon 3, leading to R57G and C123R aa sub-
stitutions respectively. P2 displays a homozygous C622T
nonsense mutation in exon 5, changing an arginine codon
at position 178 to a stop codon (R178X), thus resulting in
a putative protein lacking about one-third of the C terminus
region. P3 and P4 carry a homozygous deletion of G267,
the last nucleotide of exon 2, as well as a homozygous A
to T change three nucleotides downstream in intron 2 (Fig-
ure 4B). These two associated mutations result in a mixed
RNA population. Normally spliced RNAs carry the G267 de-
letion, resulting in a frameshift at K69 followed by a premature
stop codon, while numerous aberrant splicing events lead to
severely truncated putative proteins. Of note, one of these
mutated forms may retain some activity as it presents an in-
ternal in-frame deletion covering A25-R57 (data not shown).
Lastly, P5 carries a homozygous C259G substitution in exon
Figure 3. V(D)J Recombination in Fibroblasts and In Vitro NHEJ Are Defective in Patients
(A) SV40-transformed ﬁbroblasts were transfected with RAG1- and RAG2-expressing vectors and the pRecCS extrachromosomal V(D)J recombination
substrate. The substrate was recovered 48–72 hr posttransfection, and the formation of signal and coding joins were analyzed by PCR.
(B) The signal-join ﬁdelity was analyzed by ApalI digestion of the PCR products. Signal joins detected in P1 and P3 were imprecise.
(C) Protein extracts from control (C), a Lig4 patient (L4), and P1-P5 cells were incubated with a linearized plasmid. The reaction products were run on aga-
rose gel and stained with SYBR-Green. Multimerized DNA molecules (2 and 3) were formed with control but not with P1-P5 extracts.
(D) While mixing protein extracts from Lig4 (L4) together with P1-P5 cells eliminated the NHEJ defect, using either one independently resulted in the emer-
gence of the defect, indicating that P1-P5 defect is not Lig4.
(E) Pair-wise mixingof P1-P5 cellular extractsfailed to complement the patient’s NHEJdefect, suggestingthat P1-P5belong to the same complementationgroup.
Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc. 293
294 Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc.
2, resulting in R57G aa modiﬁcation. We conﬁrmed the auto-
somal recessive inheritance of the identiﬁed mutations by
sequencing parents’ DNA in all cases (data not shown).
Complementation of the NHEJ Defect
by wt Cernunnos
We next evaluated the complementation of the NHEJ defect
by ectopic expression of wt Cernunnos. Primary ﬁbroblasts
from P1, P2, and P3 were transduced with a bicistronic
Cernunnos-ires-GFP retrovirus, resulting in 86%, 75%, and
80% GFP-positive cells respectively. The survival curves of
these transduced cells after ionizing radiation (0, 1, and 3
Gy) indicated a return to an overall normal radiosensitivity
as compared to control cells (Figure 5A). Likewise, the
V(D)J recombination was fully restored upon transfection of
wt Cernunnos, as demonstrated by the complete ApalI di-
gestion of PCR-ampliﬁed signal joins (Figure 5B) and in-
chromosome V(D)J assay (Figure S4). Lastly, cell-free ex-
tracts from Cernunnos-transfected ﬁbroblasts were able to
oligomerize linear DNA molecules in the in vitro NHEJ assay
(Figure 5C). Taken together, these results conﬁrm that the
identiﬁed Cernunnos mutations are responsible for the de-
fective DNA repair in this series of patients.
T and B Cell Combined Immunodeﬁciency,
Microcephaly, and Growth Retardation
Caused by Cernunnos Mutations
We herein describe a T and B cell combined immunodeﬁ-
ciency characterized by a progressive lymphocytopenia as-
sociated with developmental defects including microcephaly
as well as growth retardation and an increased cellular sen-
sitivity to IR. This new rare human autosomal recessive pri-
mary immunodeﬁciency is caused by deleterious mutations
in a novel general DNA repair factor we have named Cernun-
nos. The independent identiﬁcation of the same factor,
named XLF, is reported by Ahnesorg et al. in another article
in this issue of Cell (Ahnesorg et al., 2006). Strikingly, the clin-
ical phenotype of Cernunnos-deﬁcient patients shares sev-
eral characteristics with Nijmegen breakage syndrome and
Lig4 deﬁciency. However, Cernunnos deﬁciency does not
lead to impaired cell-cycle checkpoints, as observed in NBS
condition, but leads rather to an NHEJ defect as observed in
Lig4 deﬁciency as well as in the 2BN patient (Carney et al.,
1998; Matsuura et al., 1998; Varon et al., 1998; Dai et al.,
2003; Buck et al., 2006).
DSB are introduced in DNA during the V(D)J recombination
process. A complete failure to resolve these breaks leads to
an arrest in B and T cell development, as observed in Arte-
mis-deﬁcient RS-SCID condition and in the respective animal
models (Revy et al., 2005). A residual DNA end-joining activ-
ity, as observed in the case of hypomorphic Lig4 and Artemis
mutations, allows for limited development of B and T cells
(Moshous et al., 2003; Smith et al., 2003). The partially com-
promised (around 10% of normal value as shown in Figure
S4C) V(D)J recombination identiﬁed in Cernunnos-deﬁcient
ﬁbroblasts is comparable to that of hypomorphic Lig4-deﬁ-
cient cells and may account for the severe T and B cell lym-
phopenia observed in Cernunnos deﬁciency. The low counts
of circulating T and B cells may in fact reﬂect a residual Cer-
nunnos activity caused by hypomorphic mutations. The na-
ture of the different mutations identiﬁed in the patients is com-
patible with expression of a Cernunnos protein, albeit
truncated in some cases (P2-P4). In Cernunnos patients,
the presence of memory-only T cells in combination with
the progressive loss of B lymphocytes suggest that while
a wave of lymphocyte production did occur at some point,
the sustained renewal of the immune system is profoundly
impaired in these patients. These observations, taken to-
gether with our observation of the occurrence of bone mar-
row aplasia in P5, suggest that the Cernunnos defect may
be associated with a compromised haemo-/lymphopoiesis.
T cell lymphopenia appears relatively milder as compared
to B cell, at least in the eldest patients. Expansion of long-
lived memory T cells as observed in other combined T+B im-
munodeﬁciencies caused by other hypomorphic gene muta-
tions (DiSanto et al., 1994; Moshous et al., 2003; De Villartay
et al., 2005) can account for this observation.
Although patients displayed a hypogamaglobulinemia, the
levels of serum IgM (Table 1) were occasionally as high as
observed in classical CSR-deﬁcient hyper-IgM syndromes
(Durandy et al., 2005). This suggests a possible role for Cer-
nunnos in CSR. Autoimmunity is another consequence of
the immunodeﬁciency caused by Cernunnos deﬁciency, as
also observed in other combined T+B cell immunodeﬁ-
ciencies caused by hypomorphic mutation of RAG1 (
lartay et al., 2005). It may result from a skewed T cell reper-
toire subsequent to partially defective T+B development or
to defective peripheral control of response to self antigens.
Beside the patent immunological consequences of Cer-
nunnos deﬁciency, developmental defects, notably micro-
cephaly, are also present. This is reminiscent of the micro-
cephaly that has been noted in some Lig4-deﬁcient
patients (Riballo et al., 1999; O’Driscoll et al., 2001; Buck
et al., 2006), as well as of the embryonic lethality of Lig4
(and XRCC4) KO mice due to massive apoptosis of postmi-
totic neurons (Barnes et al., 1998; Frank et al., 1998; Gao
et al., 1998). This suggests that Cernunnos plays a role in
the development of the central nervous system. Indeed, pro-
ﬁcient DNA-damage responses, including NHEJ and HR,
have been shown to be essential during the development
Figure 4. Cernunnos: cDNA Sequence, Genomic Organization, Patients’ Mutations, cDNA Expression, and Subcellular Localiza-
(A) Cernunnos cDNA sequence and protein (1-letter code).
(B) Cernunnos gene organization with the mutations identiﬁed in patients. Gray regions represent the coding sequence.
(C) RT-PCR analysis using a panel of cDNA from 15 different human tissues. G3PDH-speciﬁc primers were used as control.
(D) Nuclear localization of transfected V5-epitope-tagged Cernunnos in ﬁbroblasts determined by immunoﬂuorescence.
Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc. 295
of the brain (Abner and McKinnon, 2004). Whether Cernun-
nos, comparably to Lig4/XRCC4, is an essential factor for vi-
ability is currently unknown. The development of a Cer-
nunnos KO mouse is likely to assist in clarifying this issue.
NHEJ factors are considered genomic caretakers since
they guarantee genomic integrity through the proper repair
of DNA lesions (Ferguson and Alt, 2001; Dudley et al.,
2005; Revy et al., 2005). Indeed, analysis of several animal
models has indicated that NHEJ deﬁciency may result in
chromosomal aberrations such as translocations or telo-
meric fusions, the consequence of which is a high propensity
to develop cancer. In the immune system, the introduction of
an NHEJ defect into a P53 mutated background invariably
leads to the development of pro-B cell lymphomas in mice
(Ferguson and Alt, 2001). Likewise, hypomorphic mutations
in Lig4 and Artemis in humans have been found to be asso-
ciated with the onset of lymphopoietic malignancies (Riballo
et al., 1999; Moshous et al., 2003). The question of the ge-
nomic caretaker status of Cernunnos is of importance be-
cause it may turn out that Cernunnos deﬁciencies are linked
to a higher risk of developing cancer. The implication of Cer-
nunnos in NHEJ in combination with the fact that two Cer-
nunnos deﬁcient patients (P1 and P5) exhibited several chro-
mosome alterations support this assumption. The analysis of
Cernunnos KO mice will help to clarify this important issue.
What Is the Role for Cernunnos in NHEJ?
The in vitro NHEJ assay clearly identiﬁed Cernunnos as a key
constituent of the DNA end-joining reaction. The remarkable
parallel in the clinical/biological presentations of Cernunnos
with Lig4 deﬁciencies strongly suggests that Cernunnos
could be a third partner in the XRCC4/Lig4 complex. Indeed,
the accompanying article from Stephen Jackson’s labora-
tory reports on the identiﬁcation of the same factor (named
XLF) through two-hybrid screen using XRCC4. Moreover,
the interaction of Cernunnos/XLF with the XRCC4:DNA-Li-
gaseIV complex in vivo was further documented through
coimmunoprecipitation experiments (Ahnesorg et al., 2006
and unpublished data). Several biochemical and genetic
studies in yeast and mammals have previously identiﬁed
XRCC4 and/or Lig4 interacting factors. Among those,
Nej1/Lif2p has been described only in S. cerevisae (Frank-
Vaillant and Marcand, 2001; Kegel et al., 2001; Ooi et al.,
2001; Valencia et al., 2001). Nej1 interacts with the yeast
XRCC4 homolog (Lif1) and is critically required for NHEJ.
Moreover, Nej1 is speciﬁcally repressed during mating
type-switching in yeast, a DNA recombination process that
relies exclusively on homologous recombination. No homo-
logs for Nej1 have yet been identiﬁed in higher eukaryotes.
Further studies are now required to determine whether or
not Cernunnos represents a functional, although highly di-
vergent, homolog of Nej1 in mammals.
The patients’ clinical phenotypes are shown in Table 1.
Five patients from four families (including three consanguineous fami-
lies) were included. P3 and P4 were siblings. Family history revealed
that three older sisters of P2 died from severe infections during their ﬁrst
year of life. All patients presented with growth retardation (-3SD), already
existent at birth in two of them, or during the ﬁrst year of life in the other
three. Microcephaly (-3SD) at birth was a constant feature in all patients.
Facial dysmorphia was mild in P1 but was more prominent in P3-P4 and
P5, who presented with a birdlike face. Bone malformations were present
in P1 (i.e., low implantation of the thumb) and in P3-P4 (i.e., hypoplasia of
Figure 5. Complementation of the DNA-Repair Defect by wt Cernunnos
(A) Primary ﬁbroblasts from P1 (squares), P2 (circles), and P3 (triangles) were transduced with Cernunnos-iresGFP retrovirus and exposed to ionizing ra-
diation (solid lines), and their survival compared to untransduced cells (dash lines) and normal control cells (gray envelope).
(B) V(D)J recombination coding join-formation and signal join-ﬁdelity (complete ApalI digestion) were restored by wt Cernunnos in P1-P5 ﬁbroblasts.
(C) Wt Cernunnos complemented the in vitro NHEJ defect in cellular extracts from transfected P1-P5 cells.
296 Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc.
the middle phalanx of the ﬁfth ﬁnger). Malformations of the kidney (i.e.,
nephroptosis) and of the genital organs were also present in P1. Mental
retardation was evident only in P1. P5 suffered from bone-marrow apla-
sia, which was ﬁrst diagnosed at the age of one year. Chromosomal anal-
yses revealed a normal 46XY karyotype but several spontaneous, nonre-
current, chromosomal translocations in patients P1 and P5 lymphocytes.
All patients experienced severe infections; P1 developed recurrent
bacterial infections of the respiratory tract beginning at the age of three
years and invasive warts and a life-threatening cholangitis which had
been diagnosed at eight years of age. P2 suffered from recurrent bacterial
infections of the respiratory and digestive tracts from the ﬁrst months of life
onward. P3 presented with a Pneumocystis carinii pneumonia, chronic
Giardia lamblia enteritis, and Salmonella and Campylobacter enteritis as
well as molluscum contagiosum and warts from three years of age on-
ward. His sister, P4, had neonatal pneumonia and suffered from mild re-
current infections of the upper respiratory tract. P5 developed bacterial in-
fections of the respiratory tract from the ﬁrst year of life onward. Two
patients (P1 and P3) also developed autoimmune manifestations—hemo-
lytic anemia at four years of age and thrombocytopenia at eight years of
age, respectively. P1 received immunosuppressive therapy over a ﬁve-
year period and was ﬁnally splenectomized. As soon as lymphopenia
and hypogammaglobulinemia were diagnosed, patients were started
on immunoglobulin substitution. Although under treatment, P1 and P2
both died from a septic shock. P3, P4, and P5 currently remain alive, sup-
ported by immunoglobulin substitution and antibiotic prophylaxis (in P3
and P4). In accordance with the Helsinki Declaration, informed consent
for our study was obtained from the families. This study was also ap-
proved by the (INSERM) Institutional Review Board. Primary skin ﬁbro-
blasts were obtained from skin biopsies of all ﬁve patients as well as
RS-SCID, Lig4, Ataxia telangiectasia, Nijmegen breakage syndrome pa-
tients in addition to normal controls. SV40-transformed and telomerase-
immortalized cell lines were obtained as described (Nicolas et al., 1998;
Poinsignon et al., 2004).
Rabbit polyclonal anti-phospho-histone H3 and anti-phospho-H2A-X
(Ser139) came from Upstate (Lake Placid, NY). Mouse monoclonal anti-
bodies directed against MRE11 (clone 12D7) and RAD51 (clone 14B4)
originated from Abcam (Cambridge, UK). Alexa-568 and Alexa-488
secondary antibodies were from Molecular probes (Eugen,
OR). FITC-conjugated anti-BrdU antibody and goat F(Ab
munoglobulin G were purchased from Becton Dickinson (Mountain View,
CA). Mouse monoclonal anti-V5 antibody came from Invitrogen.
IR Sensitivity Assay
IR sensitivity was analyzed by clonogenic assays upon g rays exposure of
primary ﬁbroblasts, as previously described (Nicolas et al., 1998). For IR
sensitivity complementation analysis, primary ﬁbroblasts were trans-
duced using a bicistronic Cernunnos-iresGFP-expressing retrovirus and
subjected to various doses of g rays (0, 1, 3 Gy).
Immunoﬂuorescence Detection of IRIFs
Primary ﬁbroblasts were seeded on cover slips and X-irradiated (Faxitron-
160FW, EDIMEX, Wheeling, IL; 10 Gy; 166 rad/min). Two hours after irra-
diation, cells were washed with PBS and ﬁxed with 4% paraformaldehyde
in PBS for 15 min. After each step, the cover slips were rinsed three times
with PBS. Cells were incubated for 20 min with PBS containing 0.1M gly-
cine and permeabilized in 0.5% Triton X-100 in PBS for 15 min. Cells were
incubated for 30 min with the PBS-BSA solution. Cells were then labeled
with the appropriate primary antibodies, followed by washes with PBS-
BSA solution and incubation with secondary antibodies. Slides were
counterstained with 0.1 mg/ml of DAPI (4
and mounted in Fluorsave (Calbiochem). Slides were viewed by epiﬂuor-
escence microscopy (Axioplan ZEISS), and images were taken by
a cooled charge-coupled device (CCD) camera. Grayscale images were
processed using Adobe Photoshop 5.5. (San Jose, CA).
gH2AX Detection by Fluorocytometry
Fibroblasts were incubated at 37ºC with medium alone or containing
10 mg/ml of Bleomycin (Aventis) for 1 hr and then washed and incubated
in complete medium. Thirty minutes or 24 hr later, cells were harvested
and ﬁxed in ethanol 70%. Permeabilization was performed with 0.1% tri-
ton X-100 and 4% FBS. Cells were resuspended with anti-gH2AX anti-
body (dilution 1:500) and incubated for 2 hr. Cells were washed and re-
suspended with a 1:1000 dilution of goat anti-mouse IgG (Alexa 488)
for 1 hr. Samples were washed and analyzed by a Becton Dickinson
G1/S checkpoint analysis was performed as previously described
(Moshous et al., 2003). For G2/M checkpoint analysis, 3 10
ized SV40-transformed ﬁbroblasts were harvested 1 hr after 5 Gy X-irra-
diation, washed with PBS, ﬁxed with cold 70% ethanol, and kept at –20ºC
for 16 hr. After washing, the cells were resuspended in PBS 0.25% with
Triton X-100 and incubated on ice for 15 min. Cells were washed in
PBS and incubated with 1 mg of anti-phospho-H3 in PBS-1% BSA for
3 hr. Anti-P-H3 was detected with FITC-conjugated F(Ab
bit IgG antibody. The cells were washed and resuspended in PBS con-
taining 25 mg/ml PI. Fluorescence was measured by using a Becton Dick-
inson FACScan ﬂowcytometer. For RDS assay, primary ﬁbroblasts were
pulse-labeled for 2 hr with 10 mlof[
H]thymidine (1.5 MBq/ml) 45 min fol-
lowing 0 to 40 Gy IR exposure.
In Vivo Plasmid Religation Assay
Five mg of a modiﬁed version of EGFP-N2 plasmid (Clontech) were linear-
ized by either SacI and KpnI in order to generate 3
ends or by Ecl136II
and SmaI in order to generate blunt endsand introduced into SV40-trans-
formed ﬁbroblasts by electroporation. Recircularized plasmids were re-
covered by hirt lysis 48 hr after transfection, and the junctions were
PCR-ampliﬁed using primers CMV 5
and GFPr 5
. PCR products were
cloned in pGemT (Promega) and sequenced.
V(D)J Recombination Assay
V(D)J recombination on extra chromosomal substrates was performed as
previously described using full-length human Rag1 and Rag2 (Nicolas
et al., 1998; Moshous et al., 2001). The pRecCS V(D)J recombination re-
porter substrate has RSS in an orientation which drives V(D)J recombina-
tion by inversion, thus retaining both coding and signal joins, these then
being PCR-ampliﬁed using speciﬁc combinations of primers. For comple-
mentation analysis in ﬁbroblasts, 2.5 mgofwtCernunnos cloned in
pCDNA3.1D (Invitrogen) was cotransfected. V(D)J recombination assay
on chromosome integrated substrate was performed as previously de-
scribed (Poinsignon et al., 2004).
In Vitro NHEJ Assay
Whole-cell extracts (WCE) preparation and in vitro NHEJ assay were per-
formed using a procedure adapted from Baumann et al. and Diggle et al.
(Baumann and West, 1998; Diggle et al., 2003). Brieﬂy, after washing in
1 PBS, cells were lysed through three freeze/thaw cycles in LB buffer
(25 mM Tris [pH 7.5], 333 mM KCl, 1.3 mM EDTA, 4 mM DTT, protease
and phosphatase inhibitors). Lysates were incubated for 20 min at 4ºC
and cleared by centrifugation. Supernatants were dialyzed against 1 E
buffer (20 mM Tris [pH 8.0], 20% glycerol, 0.1 M K(OAc), 0.5 mM
EDTA, 1 mM DTT). WCE were adjusted to 5 mg/ml and kept frozen
(80ºC) until use. For NHEJ assay, 15 mg of WCE was incubated (10 ml
reaction) with 25 ng of linear DNA (EcoRI digested pEGFPN2) in 1
LigB (250 mM Tris [pH 8.0], 300 mM K(OAc), 2.5 mM Mg(OAc)2, 5 mM
ATP, 5 mM DTT, 0.5 mg/ml BSA, 1 mg/ml IP6) for 2 hr at 37ºC. Reactions
were then treated with 1 m l RNase (1 mg/ml) for 5 min at RT and with 2 mlof
5 deproteination solution (10 mg/ml Proteinase K, 2.5% SDS, 50 mM
EDTA, 100 mM Tris [pH 7.5]) for 30 min at 55ºC. After migration of the
samples in 0.7% agarose, the gels were stained with SYBR-Green
(30 min), and ﬂuorescence was detected via a FluorImager.
Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc. 297
cDNA Complementation Cloning
SV40-transformed, telomerase-immortalized ﬁbroblasts from P2 were
transduced with a commercially available human thymic cDNA library
cloned into the pFB retroviral vector (ViraPort XR Plasmid Human Thymus
cDNA library, Stratagene) as described (Moshous et al., 2003). Trans-
duced cells were kept in culture for 88 days. During this time period,
they were treated nine times with bleomycin (0.5 mg/ml in RPMI for 1 hr).
cDNA inserts were recovered from genomic DNA by PCR-ampliﬁcation
using vector-speciﬁc primers FBP1 (5
) and FBP2 (5
). PCR bands were gel-puriﬁed and directly sequenced using FBP1
and subsequent internal primers. The Cernunnos ORF was PCR-ampliﬁed
(F1 CACCTCGCCACATGGAAGAACTGGAGCAA, R1 ACTGAAGAGACC
CCTTGCCCTTCTTCC) and cloned with a C terminus V5-epitope/6xHis
tag in pCDNA3.1D (Invitrogen) and a C terminus myc-epitope/6xHis tag
in the retroviral vector pMND-MFG (Robbins et al., 1998). The genomic or-
ganization of Cernunnos gene was determined in silico through Blast anal-
ysis of the cDNA against human genomic sequences.
Cernunnos Gene Sequencing and Expression Analysis
The eight exons of Cernunnos were PCR-ampliﬁed by using genomic DNA
and Taq High Fidelity polymerase (Roche Diagnostics, Mannheim, Ger-
many) in addition to the following oligonucleotide primer pairs: 1F 5
. RT-PCRs were performed with primers Cernunnos-F 5
and Cernunnos-R 5
. All PCR products were directly sequenced. PCR-ready cDNAs
from several tissues (Clontech) were ampliﬁed with primers Cernu-F 5
and Cernunnos-R or G3PDH primers.
They were then run onto 1% agarose gels, and bands were stained with
Supplemental Data include four ﬁgures and can be found with this article
online at http://www.cell.com/cgi/content/full/124/2/287/DC1/.
We are indebted to Dr. Irene Ward for the kind gift of the mouse monoclo-
nal anti-53BP1 antibody and to Dr. Tobias Ankermann for blood and ﬁ-
broblast samples from P3 and P4. We acknowledge Monique Forveille
for her technical assistance. We thank Dr. Sylvain Latour for discussions
and critical reading of the manuscript. We acknowledge Drs. I. Tezcan
and F. Ersoy for their contribution to the followup of the patient P2. This
work was supported by institutional grants from INSERM as well as grants
from the Association de Recherche sur le Cancer (ARC), the Ligue
National contre le Cancer (Equipe labelise
e LA LIGUE 2005,
EL2005.LNCC/JPDV1), the Commissariat a
l’Energie Atomique (LRC-
CEA Nº40V), the French Rare Disease Program (GIS), the INCa/Cance
le IdF, and the Euro-Policy-PID (nºPL 006411). P.R. is a scientist from
the Centre National de la Recherche Scientiﬁque (CNRS). D.B. received
fellowships from the European Academy of Allergy and Clinical Immunol-
ogy (EAACI), the Association pour la Recherche sur le Cancer (ARC), and
the City of Paris.
Received: October 3, 2005
Revised: November 22, 2005
Accepted: December 14, 2005
Published: January 26, 2006
Abner, C.W., and McKinnon, P.J. (2004). The DNA double-strand break
response in the nervous system. DNA Repair (Amst.) 3, 1141–1147.
Ahnesorg, P., Smith, P., and Jackson, S.P. (2006). XLF interacts with the
XRCC4-DNA ligase IV complex to promote DNA nonhomologous end-
joining. Cell 124, this issue, 301–313.
Badie, C., Goodhardt, M., Waugh, A., Doyen, N., Foray, N., Calsou, P.,
Singleton, B., Gell, D., Salles, B., Jeggo, P., et al. (1997). A DNA dou-
ble-strand break defective ﬁbroblast cell line (180BR) derived from a radio-
sensitive patient represents a new mutant phenotype. Cancer Res. 57,
Barnes, D.E., Stamp, G., Rosewell, I., Denzel, A., and Lindahl, T. (1998).
Targeted disruption of the gene encoding DNA ligase IV leads to lethality
in embryonic mice. Curr. Biol. 8, 1395–1398.
Bassing, C.H., and Alt, F.W. (2004). H2AX may function as an anchor to
hold broken chromosomal DNA ends in close proximity. Cell Cycle 3,
Bassing, C.H., Swat, W., and Alt, F.W. (2002). The mechanism and reg-
ulation of chromosomal V(D)J recombination. Cell 109 Suppl, S45–S55.
Baumann, P., and West, S.C. (1998). DNA end-joining catalyzed by hu-
man cell-free extracts. Proc. Natl. Acad. Sci. USA 95, 14066–14070.
Bogue, M.A., Wang, C., Zhu, C., and Roth, D.B. (1997). V(D)J recombi-
nation in Ku86-deﬁcient mice: distinct effects on coding, signal, and hy-
brid joint formation. Immunity 7, 37–47.
Buck, D., Moshous, D., de Chasseval, R., Ma, Y., Le Deist, F., Cavaz-
zana-Calvo, M., Fischer, A., Casanova, J.L., Lieber, M.R., and De Villar-
tay, J.P. (2006). Severe combined immunodeﬁciency and microcephaly in
siblings with hypomorphic mutations in DNA ligase IV. Eur. J. Immunol.
Carney, J.P., Maser, R.S., Olivares, H., Davis, E.M., Le Beau, M., Yates,
J.R., 3rd, Hays, L., Morgan, W.F., and Petrini, J.H. (1998). The hMre11/
hRad50 protein complex and Nijmegen breakage syndrome: linkage of
double-strand break repair to the cellular DNA damage response. Cell
Dai, Y., Kysela, B., Hanakahi, L.A., Manolis, K., Riballo, E., Stumm, M.,
Harville, T.O., West, S.C., Oettinger, M.A., and Jeggo, P.A. (2003). Non-
homologous end joining and V(D)J recombination require an additional
factor. Proc. Natl. Acad. Sci. USA 100, 2462–2467.
de Villartay, J.P., Fischer, A., and Durandy, A. (2003). The mechanisms of
immune diversiﬁcation and their disorders. Nat. Rev. Immunol. 3, 962–
De Villartay, J.P., Lim, A., Al-Mousa, H., Dupont, S., De
J., Caumau-Gatbois, E., Gougeon, M.L., Lemainque, A., Eidenschenk,
C., Jouanguy, E., et al. (2005). A novel immunodeﬁciency associated
with hypomorphic RAG1 mutations and CMV infection. J. Clin. Invest.
Diggle, C.P., Bentley, J., and Kiltie, A.E. (2003). Development of a rapid,
small-scale DNA repair assay for use on clinical samples. Nucleic Acids
DiSanto, J.P., Rieux-Laucat, F., Dautry-Varsat, A., Fischer, A., and de
Saint Basile, G. (1994). Defective human interleukin 2 receptor gamma
chain in an atypical X chromosome-linked severe combined immunodeﬁ-
ciency with peripheral T cells. Proc. Natl. Acad. Sci. USA 91 , 9466–9470.
Dudley, D.D., Chaudhuri, J., Bassing, C.H., and Alt, F.W. (2005). Mech-
anism and control of V(D)J recombination versus class switch recombina-
tion: similarities and differences. Adv. Immunol. 86, 43–112.
Durandy, A., Revy, P., Imai, K., and Fischer, A. (2005). Hyper-immuno-
globulin M syndromes caused by intrinsic B-lymphocyte defects. Immu-
nol. Rev. 203, 67–79.
Dynan, W.S., and Yoo, S. (1998). Interaction of Ku protein and DNA-
dependent protein kinase catalytic subunit with nucleic acids. Nucleic
Acids Res. 26, 1551–1559.
298 Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc.
Ferguson, D.O., and Alt, F.W. (2001). DNA double strand break repair and
chromosomal translocation: lessons from animal models. Oncogene 20,
Frank, K.M., Sekiguchi, J.M., Seidl, K.J., Swat, W., Rathbun, G.A.,
Cheng, H.L., Davidson, L., Kangaloo, L., and Alt, F.W. (1998). Late em-
bryonic lethality and impaired V(D)J recombination in mice lacking DNA
ligase IV. Nature 396, 173–177.
Frank-Vaillant, M., and Marcand, S. (2001). NHEJ regulation by mating
type is exercised through a novel protein, Lif2p, essential to the ligase
IV pathway. Genes Dev. 15, 3005–3012.
Gao, Y., Sun, Y., Frank, K.M., Dikkes, P., Fujiwara, Y., Seidl, K.J., Seki-
guchi, J.M., Rathbun, G.A., Swat, W., Wang, J., et al. (1998). A critical
role for DNA end-joining proteins in both lymphogenesis and neurogene-
sis. Cell 95, 891–902.
Gottlieb, T.M., and Jackson, S.P. (1993). The DNA-dependent protein ki-
nase: requirement for DNA ends and association with Ku antigen. Cell 72,
Grawunder, U., Wilm, M., Wu, X., Kulesza, P., Wilson, T.E., Mann, M.,
and Lieber, M.R. (1997). Activity of DNA ligase IV stimulated by complex
formation with XRCC4 protein in mammalian cells. Nature 388, 492–495.
Kegel, A., Sjostrand, J.O., and Astrom, S.U. (2001). Nej1p, a cell type-
speciﬁc regulator of nonhomologous end joining in yeast. Curr. Biol. 11 ,
Ma, Y., Pannicke, U., Schwarz, K., and Lieber, M.R. (2002). Hairpin open-
ing and overhang processing by an Artemis/DNA-dependent protein ki-
nase complex in nonhomologous end joining and V(D)J recombination.
Cell 108, 781–794.
Matsuura, S., Tauchi, H., Nakamura, A., Kondo, N., Sakamoto, S., Endo,
S., Smeets, D., Solder, B., Belohradsky, B.H., Der Kaloustian, V.M., et al.
(1998). Positional cloning of the gene for Nijmegen breakage syndrome.
Nat. Genet. 19, 179–181.
Moshous, D., Callebaut, I., de Chasseval, R., Corneo, B., Cavazzana-
Calvo, M., Le Deist, F., Tezcan, I., Sanal, O., Bertrand, Y., Philippe, N.,
et al. (2001). Artemis, a novel DNA double-strand break repair/V(D)J
recombination protein, is mutated in human severe combined immune
deﬁciency. Cell 105, 177–186.
Moshous, D., Pannetier, C., Chasseval Rd, R., Deist Fl, F., Cavazzana-
Calvo, M., Romana, S., Macintyre, E., Canioni, D., Brousse, N., Fischer,
A., et al. (2003). Partial T and B lymphocyte immunodeﬁciency and pre-
disposition to lymphoma in patients with hypomorphic mutations in Arte-
mis. J. Clin. Invest. 111, 381–387.
Nicolas, N., Moshous, D., Cavazzana-Calvo, M., Papadopoulo, D., de
Chasseval, R., Le Deist, F., Fischer, A., and de Villartay, J.P. (1998). A hu-
man severe combined immunodeﬁciency (SCID) condition with increased
sensitivity to ionizing radiations and impaired V(D)J rearrangements de-
ﬁnes a new DNA recombination/repair deﬁciency. J. Exp. Med. 188,
O’Driscoll, M., Cerosaletti, K.M., Girard, P.M., Dai, Y., Stumm, M., Kysela,
B., Hirsch, B., Gennery, A., Palmer, S.E., Seidel, J., et al. (2001). DNA li-
gase IV mutations identiﬁed in patients exhibiting developmental delay
and immunodeﬁciency. Mol. Cell 8, 1175–1185.
Ooi, S.L., Shoemaker, D.D., and Boeke, J.D. (2001). A DNA microarray-
based genetic screen for nonhomologous end-joining mutants in Sac-
charomyces cerevisiae. Science 294, 2552–2556.
Poinsignon, C., Moshous, D., Callebaut, I., de Chasseval, R., Villey, I., and
de Villartay, J.P. (2004). The metallo-beta-lactamase/beta-CASP domain
of Artemis constitutes the catalytic core for V(D)J recombination. J. Exp.
Med. 199, 315–321.
Revy, P., Buck, D., Le Deist, F., and De Villartay, J.P. (2005). The repair of
DNA damages/modiﬁcations during the maturation of the immune sys-
tem: lessons from human primary immunodeﬁciency disorders and ani-
mal models. Adv. Immunol. 87, 237–295.
Riballo, E., Critchlow, S.E., Teo, S.H., Doherty, A.J., Priestley, A.,
Broughton, B., Kysela, B., Beamish, H., Plowman, N., Arlett, C.F., et al.
(1999). Identiﬁcation of a defect in DNA ligase IV in a radiosensitive leukae-
mia patient. Curr. Biol. 9, 699–702.
Riballo, E., Doherty, A.J., Dai, Y., Stiff, T., Oettinger, M.A., Jeggo, P.A.,
and Kysela, B. (2001). Cellular and biochemical impact of a mutation in
DNA ligase IV conferring clinical radiosensitivity. J. Biol. Chem. 276,
Robbins, P.B., Skelton, D.C., Yu, X.J., Halene, S., Leonard, E.H., and
Kohn, D.B. (1998). Consistent, persistent expression from modiﬁed retro-
viral vectors in murine hematopoietic stem cells. Proc. Natl. Acad. Sci.
USA 95, 10182–10187.
Rogakou, E.P., Pilch, D.R., Orr, A.H., Ivanova, V.S., and Bonner, W.M.
(1998). DNA double-stranded breaks induce histone H2AX phosphoryla-
tion on serine 139. J. Biol. Chem. 273, 5858–5868.
Rooney, S., Sekiguchi, J., Zhu, C., Cheng, H.L., Manis, J., Whitlow, S.,
DeVido, J., Foy, D., Chaudhuri, J., Lombard, D., and Alt, F.W. (2002).
Leaky Scid phenotype associated with defective V(D)J coding end pro-
cessing in Artemis-deﬁcient mice. Mol. Cell 10, 1379–1390.
Rooney, S., Alt, F.W., Lombard, D., Whitlow, S., Eckersdorff, M., Fleming,
J., Fugmann, S., Ferguson, D.O., Schatz, D.G., and Sekiguchi, J. (2003).
Defective DNA repair and increased genomic instability in Artemis-deﬁ-
cient murine cells. J. Exp. Med. 197, 553–565.
Sancar, A., Lindsey-Boltz, L.A., Unsal-Kacmaz, K., and Linn, S. (2004).
Molecular mechanisms of mammalian DNA repair and the DNA damage
checkpoints. Annu. Rev. Biochem. 73
Shearer, W.T., Rosenblatt, H.M., Gelman, R.S., Oyomopito, R., Plaeger,
S., Stiehm, E.R., Wara, D.W., Douglas, S.D., Luzuriaga, K., McFarland,
E.J., et al. (2003). Lymphocyte subsets in healthy children from birth
through 18 years of age: the Pediatric AIDS Clinical Trials Group P1009
study. J. Allergy Clin. Immunol. 112, 973–980.
Smith, J., Riballo, E., Kysela, B., Baldeyron, C., Manolis, K., Masson, C.,
Lieber, M.R., Papadopoulo, D., and Jeggo, P. (2003). Impact of DNA li-
gase IV on the ﬁdelity of end joining in human cells. Nucleic Acids Res.
Valencia, M., Bentele, M., Vaze, M.B., Herrmann, G., Kraus, E., Lee, S.E.,
Schar, P., and Haber, J.E. (2001). NEJ1 controls non-homologous end
joining in Saccharomyces cerevisiae. Nature 414, 666–669.
Varon, R., Vissinga, C., Platzer, M., Cerosaletti, K.M., Chrzanowska,
K.H., Saar, K., Beckmann, G., Seemanova, E., Cooper, P.R., Nowak,
N.J., et al. (1998). Nibrin, a novel DNA double-strand break repair protein,
is mutated in Nijmegen breakage syndrome. Cell 93, 467–476.
Wei, Y., Yu, L., Bowen, J., Gorovsky, M.A., and Allis, C.D. (1999). Phos-
phorylation of histone H3 is required for proper chromosome condensa-
tion and segregation. Cell 97, 99–109.
Human Cernunnos cDNA sequence was deposited into EMBL under ac-
cession number AJ972687.
Cell 124, 287–299, January 27, 2006 ª2006 Elsevier Inc. 299