Content uploaded by Georges Snounou
Author content
All content in this area was uploaded by Georges Snounou
Content may be subject to copyright.
References and Notes
1. E. A. Bach, M. Aguet, R. D. Schreiber, Annu. Rev.
Immunol. 15, 563 (1997).
2. U. Boehm, T. Klamp, M. Groot, J. C. Howard, Annu.
Rev. Immunol. 15, 749 (1997).
3. J. E. Darnell, I. M. Kerr, G. R. Stark, Science 264, 1415
(1994).
4. M. A. Meraz et al.,Cell 84, 431 (1996).
5. S. Huang et al.,Science 259, 1742 (1993).
6. D. K. Dalton et al.,Science 259, 1739 (1993).
7. J. E. Durbin, R. Hackenmiller, M. C. Simon, D. E. Levy,
Cell 84, 443 (1996).
8. E. Jouanguy et al.,N. Engl. J. Med. 335, 1956 (1996).
9. M. J. Newport et al.,N. Engl. J. Med. 335, 1941
(1996).
10. C. Pierre-Audigier et al.,Clin. Infect. Dis. 24, 982
(1997).
11. S. Dupuis et al.,Science 293, 300 (2001).
12. V. E. Papaioannou, Trends Genet. 13, 212 (1997).
13. J. Smith, Curr. Opin. Genet. Dev. 7, 474 (1997).
14. S. J. Szabo et al.,Cell 100, 655 (2000).
15. J. L. Grogan et al.,Immunity 14, 205 (2001).
16. A. C. Mullen et al.,Science 292, 1907 (2001).
17. Supplementary Web material is available on Science
Online at www.sciencemag.org/cgi/content/full/295/
5553/338/DC1.
18. A. O’Garra, N. Arai, Trends Cell Biol. 10, 542 (2000).
19. H. Asnagli, K. M. Murphy, Curr. Opin. Immunol. 13,
242 (2001).
20. M. Bamshad et al.,Nature Genet. 16, 311 (1997).
21. C. T. Basson et al.,Nature Genet. 15, 30 (1997).
22. I. Riviere, M. J. Sunshine, D. R. Littman, Immunity 9,
217 (1998).
23. G. A. Hollander et al.,Science 279, 2118 (1998).
24. S. J. Szabo et al., unpublished observations.
25. S. L. Reiner, R. M. Locksley, Annu. Rev. Immunol. 13,
151 (1995).
26. E. Tomasello et al.,Immunity 13, 355 (2000).
27. C. K. Lee et al.,J. Immunol. 165, 3571 (2000).
28. S. Finotto et al.,Science 295, 336 (2002).
29. S. L. Peng, A. J. Gerth, A. M. Ranger, L. H. Glimcher,
Immunity 14, 13 (2001).
30. G. M. Monteforte et al.,J. Immunol. 164, 5890
(2000).
31. R. Coico, Protocols in Immunology (Wiley, New York,
1994), pp. 3.11.13–3.11.20.
32. We thank A. Dighe, A. Wurster, N. Iwakoshi, and J.
Rengarajan for thoughtful review of the manuscript;
C. McCall for manuscript preparation; and L. (Nacho)
Terrazas for assistance with the Leishmania experi-
ment. Supported by grants from the National Insti-
tutes of Health (B.P.S. and L.H.G) and a gift from The
G. Harold and Leila Y. Mathers Charitable Foundation
(L.H.G), and by an Ohio State University Seed Grant
(A.R.S), a Leukemia Society Special Fellowship (S.J.S.),
and the Cancer Research Institute Investigator Award
(B.P.S.). S.J.S and B.P.S. are recipients of the Burroughs
Wellcome Foundation Career Development Award.
20 August 2001; accepted 14 November 2001
Stage-Specific Transcription of
Distinct Repertoires of a
Multigene Family During
Plasmodium Life Cycle
P. R. Preiser,
1
* S. Khan,
1
F. T. M. Costa,
2
W. Jarra,
1
E. Belnoue,
2
S. Ogun,
1
A. A. Holder,
1
T. Voza,
3
I. Landau,
3
G. Snounou,
4
L. Re´nia
2
Members of a multigene family in the rodent malaria parasite Plasmodium yoelii
yoelii code for 235-kilodalton proteins (Py235) that are located in the merozoite
apical complex, are implicated in virulence, and may determine red blood cell
specificity. We show that distinct subsets of py235 genes are expressed in
sporozoites and hepatic and erythrocytic stages. Antibodies to Py235 inhibited
sporozoite invasion of hepatocytes. The switch in expression profile occurred
immediately after transition from one stage to another. The results suggest that
this differential expression is driven by strong biological requirements and
provide evidence that hepatic and erythrocytic merozoites differ.
Invasive stages (oo¨kinete, sporozoite, and
merozoite) of the malaria parasite penetrate
specific host cell types at different stages of
the life cycle. The 235-kD rhoptry proteins
(Py235) of the rodent parasite Plasmodium
yoelii yoelii are implicated in the type of
erythrocyte (normocyte or reticulocyte) in-
vaded by merozoites and in parasite virulence
(1–3). There are ⬃35 copies of py235 genes
in the parasite genome. Analysis of the tran-
scription pattern of py235 in blood stages has
revealed a mechanism of clonal phenotypic
variation (4): Merozoites from a single in-
fected erythrocyte differ with respect to
Py235 in their rhoptries, suggesting a unique
survival strategy (4,5). Homologs of Py235
are found in other malaria species (6–13), and
antibodies to both Py235 and a P. falciparum
homolog inhibit merozoite invasion (1,2,
13). We investigated the transcription pattern
of py235 during the different stages of the
parasite’s life cycle and the effect of specific
antibodies on cell invasion.
We used a panel of antibodies specific to
Py235 to establish that Py235 proteins are
found in sporozoites and infected hepato-
cytes. A 235-kD protein was detected in ex-
tracts of sporozoites (Fig. 1A). By immuno-
fluorescence, staining was only obtained with
the pAb-S6 and pAb-F sera, indicating that
Py235 proteins in pre-erythrocytic stages dif-
fer from those in erythrocytic parasites. All
sporozoites were labeled, with diffuse stain-
ing outlining each cell and regions of more
intense label at both ends (Fig. 1B). Infected
hepatocytes were labeled with a patchy pat-
tern that may correspond to developing mero-
zoites (Fig. 1C). Evidence that sporozoite
Py235 proteins have a functional role was
obtained from antibody inhibition of sporo-
zoite invasion of cultured primary hepato-
cytes (14). A Py235-specific antibody reac-
tive with the proteins in sporozoites inhibited
invasion, but an antibody to Py235 expressed
only during blood stages did not inhibit inva-
sion (Fig. 1D).
We analyzed py235 transcripts using nest-
ed reverse transcription–polymerase chain re-
action (RT-PCR). The size-polymorphic 3⬘-
end of py235 (4,15) (Fig. 2A) was amplified
with RNA purified from 10 to 100 oocysts
(found on a single midgut) or from 10,000 to
100,000 salivary gland sporozoites (265BY
line) (Fig. 2B). A single-sized fragment was
consistently amplified from early (5-day) and
mature (10-day) oocysts (Fig. 2B, ⫹lanes)
and from different batches of salivary gland
sporozoites (Fig. 2B, ⫹lanes); in contrast,
multiple-sized products were obtained with
RNA purified from an equivalent number of
erythrocytic parasites (Fig. 2C). Sequencing
of about 200 different cloned fragments de-
rived from at least three independent RT-
PCRs showed that these single products all
had the same sequence (Fig. 2D, type IIb). A
single band was also consistently amplified
from RNA extracted from liver-stage para-
sites grown in vitro or in vivo (Fig. 2C), and
sequence analysis of about 100 cloned prod-
ucts also showed that they had the identical
sequence. No transcript was detected in very
early hepatic trophozoites (in liver biopsies 3
hours after sporozoite inoculation), indicating
that the sporozoite py235 mRNA is degraded
very soon after hepatocyte invasion.
Although multiple py235 genes are tran-
scribed in the erythrocytic stages (4,15), it is
not known how soon this pattern is estab-
lished after initiation of blood infection by
hepatic merozoites. Therefore, we prepared
RNA from blood samples of sporozoite-in-
1
Division of Parasitology, National Institute for Med-
ical Research, The Ridgeway, London, NW7 1AA, UK.
2
INSERM U445, De´partement d’Immunologie, Institut
Cochin de Ge´ne´tique Mole´culaire, Hoˆpital Cochin,
Universite´ Rene´ Descartes, Baˆtiment G. Roussy, 27
Rue du Faubourg Saint-Jacques, 75014 Paris, France.
3
Laboratoire de Biologie Parasitaire, Muse´um National
d’Histoire Naturelle, 61 Rue Buffon, 75231 Paris Ce-
dex 05, France.
4
Unite´ de Parasitologie Biome´dicale,
Institut Pasteur, 25 and 28 Rue du Dr. Roux, 75724
Paris Cedex 15, France.
*To whom correspondence should be addressed. E-
mail: ppreise@nimr.mrc.ac.uk
REPORTS
11 JANUARY 2002 VOL 295 SCIENCE www.sciencemag.org342
fected mice, collected at closely spaced inter-
vals before and after maturation of the liver
schizonts. As expected, no RT-PCR products
were obtained from the samples obtained 24
and 44 hours after sporozoite inoculation,
because no merozoites would be released into
the blood before 45 hours. Multiple tran-
scripts were detected at 66 hours (Fig. 2C),
corresponding to the first erythrocytic
schizogony.
The analysis was extended to two addi-
tional cloned lines, YM and 1.1, derived in-
dependently from the 17X isolate. For clone
1.1, the pattern of py235 transcription was
identical to that in the 265BY line (16). For
the YM clone, transcription of two different
single genes in the pre-erythrocytic parasites
was suggested by the size of the RT-PCR
products (Fig. 3A). Sequencing established
that a single type of 3⬘-end variant is present
in each of these stages; types IIb and I were
expressed in the sporozoite and the hepatic
parasite, respectively (Fig. 2B). Multiple
transcripts were detected during the first
erythrocytic schizogony of YM parasites
(16). Although no differences could be de-
tected in the size of the RT-PCR products
from the 3⬘ends of transcribed py235 in the
1.1 and 265BY parasites, different py235
genes can share the same 3⬘-end repeat se-
quence (17,18), and the repertoire of py235
genes differs among parasite lines (15).
Therefore, to refine the RT-PCR analysis, we
used a variable region, vr (Fig. 2A), found at
the 5⬘end of the py235 genes (17). The
products were of the expected size [264 or
267 base pairs (bp)] (Fig. 3B) from the dif-
ferent stages of YM parasites, and restriction
fragment length polymorphism (RFLP)
Southern blot analysis (Fig. 3C) of the indi-
vidual products demonstrated a specific sub-
set of the py235 family expressed at each
stage. Sequencing of approximately 100
cloned vr regions derived from at least three
independent RT-PCR reactions per stage
showed that py235 genes transcribed in the
sporozoite and hepatic forms differed from
those transcribed in the erythrocytic parasite
(Fig. 3D). For the sporozoite, two vr types
(vr1 and vr3) were detected equally, whereas
for hepatic schizonts the vr3 type predomi-
nated and the vr2 type was observed once.
Whether or not Py235 expression at these
stages undergoes phenotypic variation (4)
could not be determined, because it could not
be established that sporozoites or hepatic
merozoite progeny were derived from a sin-
gle oocyst or hepatic schizont, respectively.
Two vr types (vr4 and vr5) were detected in
blood-stage parasites. All the vr sequences
detected by RT-PCR, except vr2 and vr3,
were also found in fragments amplified from
genomic DNA, and two other (vr6 and vr7)
were detected by direct amplification of
genomic DNA but not in any of the RT-PCR
products. It is therefore unlikely that the dif-
ferent vr sequences detected in the RNA
arose as a result of RNA editing.
Our results showed that a distinct subset
of py235 genes was expressed at each of the
following stages: sporozoite, hepatic schi-
zont, and erythrocytic schizont. The expres-
sion of the py235 family was reset for each of
the three invasive forms of the parasite. In-
variably, the same py235 repertoire was de-
tected in samples from independent experi-
mental infections with a cloned parasite line.
Additional preliminary analysis of the vr re-
gion from RNA obtained from the uncloned
265BY parasite was consistent with the con-
clusion that nonoverlapping sets of py235
genes are expressed during the three devel-
opmental stages studied. All together, these
observations demonstrate that differential
transcription of py235 is a general feature in
P. y. yoelii.
Little is known about the mechanisms reg-
ulating the transcription of multigene families
in Plasmodium. The two or three different
ribosomal RNA genes of malaria parasites
(depending on species) are differentially ex-
pressed in the insect and vertebrate stages
(19); however, expression of the insect stage–
specific type begins in the vertebrate host,
and this rRNA can still be detected in the
newly invaded hepatocyte of the next cycle.
Of the extensive multigene families coding
for antigenically variant proteins, only the
var gene family of P. falciparum has been
investigated in detail (20–22). PfEMP1, the
var gene product, is only detected in eryth-
rocytic-stage asexual and sexual parasites,
and the pattern of var transcription differs
from that of py235 (4). The expression of the
py235 family differs in two major respects
from that of the trypanosome variant surface
glycoprotein genes (23,24), because the
py235 repertoire expressed at a given stage is
immediately switched off at transit to the next
stage; and, in blood-stage parasites, expres-
sion from the py235 repertoire does not ap-
pear to be sequential, with several transcripts
observed in individual multinucleate para-
sites (4). The tightly regulated stage-specific
expression of different subsets of py235 may
thus be a newly discovered type of transcrip-
tional regulation in protozoa.
The demonstration of a distinct set of
merozoite rhoptry protein genes expressed
only in hepatic schizonts is molecular evi-
dence for a difference between hepatic and
erythrocytic merozoites. The activation or re-
pression of py235 expression appears to be
mediated by the cellular environment of the
parasite. However, the biological require-
Fig. 1. Py235 proteins are ex-
pressed in P. y. yoelii sporozoites
and hepatic schizonts. (A) Western
blot of P. y. yoelii proteins ob-
tained from infected erythrocyte
schizonts (PE) and sporozoites (SP)
using the serum pAb-S6 raised
against the protein sequence en-
coded by the 3⬘terminal region of
the E8 gene (25). The arrow indi-
cates the 235-kD protein band.
Immunofluorescence assays were
performed with an antiserum
(pAb-F) against the recombinant
fragment F derived from the gene
E8 (26). Immunoreactivity with (B)
air-dried and methanol-fixed sporozoites (⫻1000 magnification) and (C) methanol-fixed 48-hour
liver-stage schizonts (⫻750 magnification) is shown (27). Identical results were obtained with the
serum pAb-S6. The arrowhead indicates a liver schizont. (D) Sporozoite invasion inhibition assays
(14) on two occasions (Exp. 1 and Exp. 2) with two different preparations of polyclonal antisera
pAb-S6 (with different levels of Py235 antibody titers), pAb-D (26,28), or a monoclonal antibody
(␣CS) specific to circumsporozoite protein. The full range of Py235-specific antibodies tested is
presented in (29).
REPORTS
www.sciencemag.org SCIENCE VOL 295 11 JANUARY 2002 343
5’ UTR 3’ UTR
E8S6-5’ E8S6-3’
p235all-5’ p235all-3’
5’ F VRU 3’ R VRU
3’ R VRU(a)
5’ R VRU(a)
Fig. 2. Expression of a
single 3⬘-end repeat
type in the mosquito
and hepatic stages of P.
y. yoelii and multiple
types in blood stages.
(A) Schematic structure
of py235. The gray area
indicates a region that
has significant homolo-
gy with the reticulo-
cyte-binding protein of
P. vivax (30). The black
box indicates the trans-
membrane domain, and
the cross-hatched area
shows the repeat re-
gion, which was ampli-
fied with the nested
primers E8S6-5⬘/E8S6-
3⬘and p235all-5⬘/
p235all-3⬘.vr(17) (stip-
pled area), was ampli-
fied with the nested
primers 5⬘F-VRU/3⬘R-
VRU and 5⬘F-VRU(a)/
3⬘R-VRU(a). UTR indi-
cates 5⬘and 3⬘untrans-
lated regions. (B) Nested RT-PCR 3⬘-end repeat region products obtained from
5-day (5d) and 10-day (10d) oocysts and from salivary gland sporozoites (SP) (29).
RT reaction (4) was performed in the presence (⫹) or absence (⫺) of reverse
transcriptase (29). Products were cloned and sequenced as previously described
(15). (C) In vitro infected hepatocytes panel: RT-PCR products from in vitro
infected hepatocytes 48 (48h) and 72 (72h) hours after sporozoite infection (29).
Complete maturation of the parasite in this in vitro system takes 72 hours (31),
whereas this phase is completed in 45 hours in vivo (32). In vivo infected hepatocytes panel: RT-PCR products from in vivo infected hepatocytes 3 (3h), 24
(24h), and 44 (44h) hours after sporozoite infection. Five- to 8-week-old female BALB/c mice (Harlan Laboratories, Orle´ans, France) were injected
intravenously with 20,000 sporozoites. At different times after inoculation, liver biopsies were removed from three infected mice and immediately immersed
in lysis buffer for further RNA extraction. In vivo erythrocytes panel: RT-PCR products from circulating erythrocytes taken at 24 to 84 (24h to 84h)
hours after sporozoite infection. Blood was obtained from groups of three similarly infected mice at different times after sporozoite inoculation and
processed for RNA extraction. Blood-stage parasitemia was ascertained by examination of Giemsa-stained blood smears. We also indicate the RT
reaction in the presence (⫹) or absence (⫺) of reverse transcriptase. (D) A list of the translations of the 3⬘-end repeats identified at the time of
manuscript preparation, indicating those detected in the cloned parasite line YM. PE, SP, and LS refer to transcripts detected in erythrocytic parasites,
sporozoites, and hepatic parasites, respectively. (⫹) indicates the presence of py235 in the genome but not its detection as a transcript, and (⫺)
indicates py235 not being detected at either the DNA or RNA levels in this parasite line.
Fig. 3. Unique expression of multiple py235 genes
at different stages of parasite development. (A)
RT-PCR analysis of the 3⬘-end repeat region from
the different stages of the cloned parasite line P.
y. yoelii ( YM). Salivary gland sporozoites (SP) as
well as uninfected 0- (0h), 24- (24h) and 72-
(72h) hour-old in vitro infected hepatocytes (LS)
are shown. The location of the 300-bp marker is
indicated. (B) PCR and RT-PCR of the vr region
(29) using genomic DNA (33) (D) or RNA ob-
tained from hepatic stages (LS), sporozoites (SP),
or erythrocytic parasites (PE). (C) RFLP Southern
blot analysis of genomic parasite DNA probed
with radiolabeled PCR or RT-PCR product ob-
tained from D, PE, LS, or SP was carried out as
previously described (34). A maximum of four
bands was detected for Xmn I– (9.0, 4.1, 3.4, and
2.7 kb), or Hpa I– (10.2, 7.4, 6.0, and 2.25 kb)
digested DNA. (D) vr amino acid sequences. Se-
quence analysis of products was performed as
previously described (15). Because PCR is known
to introduce mutations, a vr allele was only con-
sidered as a variant when it was obtained from
two independent PCR amplifications or when it
differed at several sites from other variants. Ami-
no acids differing from the consensus sequence
(Majority) are boxed, and differences among
closely related sequences that are expressed in
different stages are shaded.
REPORTS
11 JANUARY 2002 VOL 295 SCIENCE www.sciencemag.org344
ments for the complex pattern of transcrip-
tional regulation of the py235 genes remain to
be elucidated. Py235 proteins have previous-
ly been shown to be involved in red blood
cell invasion. Because a subset of these pro-
teins is expressed in the sporozoite and is the
target of antibodies that inhibit hepatocyte
invasion, these proteins may be important in
the recognition and/or invasion of the mos-
quito salivary glands and the liver. Merozo-
ites released from both the liver and the
infected erythrocyte invade red blood cells,
so the need to express a distinct set of py235
genes in the infected hepatocyte is puzzling.
This differential expression of py235 in the
hepatic schizont reinforces the idea that the
obligatory passage of the parasite through the
liver not only amplifies the number of para-
sites injected by the mosquito but also pre-
adapts the parasite to invade red blood cells.
The presence of distinct rhoptry proteins in
the sporozoite and the liver-stage malaria par-
asite may form the basis of an efficient vac-
cination strategy to target these pre-erythro-
cytic–stage parasites, which are present in
small numbers and are at their most vulner-
able. Conserved regions of the rhoptry pro-
teins that are the target of protective immune
responses may also form the basis of a vac-
cine against both pre-erythrocytic– and eryth-
rocytic-stage parasites.
References and Notes
1. R. R. Freeman, A. J. Trejdosiewicz, G. A. Cross, Nature
284, 366 (1980).
2. A. A. Holder, R. R. Freeman, Nature 294, 361 (1981).
3. S. A. Ogun, A. A. Holder, Mol. Biochem. Parasitol. 76,
321 (1996).
4. P. R. Preiser, W. Jarra, T. Capiod, G. Snounou, Nature
398, 618 (1999).
5. G. Snounou, W. Jarra, P. R. Preiser, Parasitol. Today
16, 28 (2000).
6. A. A. Holder, R. R. Freeman, Philos. Trans. R. Soc.
London Ser. B 307, 171 (1984).
7. M. R. Galinski, C. C. Medina, P. Ingravallo, J. W.
Barnwell, Cell 69, 1213 (1992).
8. M. R. Galinski, J. W. Barnwell, Parasitol. Today 12,20
(1996).
9. C. A. Owen, K. A. Sinha, J. K. Keen, S. A. Ogun, A. A.
Holder, Mol. Biochem. Parasitol. 99, 183 (1999).
10. M. R. Galinski, M. Xu, J. W. Barnwell, Mol. Biochem.
Parasitol. 108, 257 (2000).
11. J. C. Rayner, M. R. Galinski, P. Ingravallo, J. W. Barn-
well, Proc. Natl. Acad. Sci. U.S.A. 97, 9648 (2000).
12. H. M. Taylor et al.,Infect. Immun. 69, 3635 (2001).
13. T. Triglia et al.,Infect. Immun. 69, 1084 (2001).
14. M. Marussig et al.,Int. Immunol. 9, 1817 (1997).
15. P. R. Preiser, W. Jarra, Exp. Parasitol. 89, 50 (1998).
16. P. R. Preiser et al., data not shown.
17. S. Khan, W. Jarra, H. Bayele, P. R. Preiser, Mol. Bio-
chem. Parasitol. 114, 197 (2001).
18. D. L. Narum, J. L. Green, S. A. Ogun, A. A. Holder, Mol.
Biochem. Parasitol. 112, 193 (2001).
19. J. Thompson et al.,Mol. Microbiol. 31, 253 (1999).
20. Q. Chen et al.,Nature 394, 392 (1998).
21. A. Scherf et al.,EMBO J. 17, 5418 (1998).
22. R. E. Hayward, B. Tiwari, K. P. Piper, D. I. Baruch, K. P.
Day, Proc. Natl. Acad. Sci. U.S.A. 96, 4563 (1999).
23. J. D. Barry, R. McCulloch, Adv. Parasitol. 49, 1 (2001).
24. P. Borst, S. Ulbert, Mol. Biochem. Parasitol. 114,17
(2001).
25. J. K. Keen, A. A. Holder, J. H. L. Playfair, M. J. Lockyer,
A. P. Lewis, Mol. Biochem. Parasitol. 42, 241 (1990).
26. S. A. Ogun, A. A. Holder, Exp. Parasitol. 79, 270
(1994).
27. L. Re´nia et al.,Eur. J. Immunol. 20, 1445 (1990).
28. K. A. Sinha, J. K. Keen, S. A. Ogun, A. A. Holder, Mol.
Biochem. Parasitol. 76, 329 (1996).
29. Supplementary data are available on Science Online
at www.sciencemag.org/cgi/content/full/295/5553/
342/DC1.
30. J. Keen, K. A. Sinha, K. N. Brown, A. A. Holder, Mol.
Biochem. Parasitol. 65, 171 (1994).
31. L. Re´nia, unpublished observations.
32. I. Landau, P. Gautret, in Malaria: Parasite Biology,
Pathogenesis,and Protection, I. W. Sherman, Ed.
(American Society for Microbiology, Washington, DC,
1998), pp. 401–417.
33. G. Snounou, S. Viriyakosol, W. Jarra, S. Thaithong,
K. N. Brown, Mol. Biochem. Parasitol. 58, 283 (1993).
34. P. R. Preiser et al.,EMBO J. 15, 684 (1996).
35. We thank R. J. M. Wilson, A. Scherf, P. Druilhe, and
A. C. Gru¨ ner for a critical review of the manuscript
and we thank A. M. Vigario for help. F.T.M.C. is
supported by a Brazilian fellowship from the CAPES
Foundation. Partly funded by the Royal Society’s
Joint Project under European Science Exchange Pro-
gramme, the British Council’s Alliance: Franco-British
Partnership Programme, and European Union con-
tract IC18 CT98 0369.
31 July 2001; accepted 25 October 2001
CTCF, a Candidate Trans-Acting
Factor for X-Inactivation Choice
Wendy Chao, Khanh D. Huynh, Rebecca J. Spencer,
Lance S. Davidow, Jeannie T. Lee*
In mammals, X-inactivation silences one of two female X chromosomes. Si-
lencing depends on the noncoding gene, Xist (inactive X-specific transcript), and
is blocked by the antisense gene, Tsix. Deleting the choice/imprinting center in
Tsix affects X-chromosome selection. Here, we identify the insulator and tran-
scription factor, CTCF, as a candidate trans-acting factor for X-chromosome
selection. The choice/imprinting center contains tandem CTCF binding sites
that function in an enhancer-blocking assay. In vitro binding is reduced by CpG
methylation and abolished by including non-CpG methylation. We postulate
that Tsix and CTCF together establish a regulatable epigenetic switch for
X-inactivation.
Dosage compensation ensures equal expres-
sion of X-linked genes in XX females and
XY males. In mammals, this process results
in inactivation of one female X chromosome
(XCI) (1) in a random or imprinted manner.
In the random form (eutherian), a zygotic
counting mechanism initiates dosage com-
pensation and enables a choice mechanism to
randomly designate one active (Xa) and one
inactive (Xi) X [reviewed in (2)]. In the
imprinted form, zygotic counting and choice
are superseded by parental imprints that di-
rect exclusive paternal X-silencing (3, 4).
Imprinted XCI is found in ancestral marsupi-
als (3) but vestiges remain in the extraembry-
onic tissues of eutherians such as mice (4).
An epigenetic mark for random and imprint-
ed XCI has long been postulated (2). The marks
are placed at the X-inactivation center (Xic)(5),
which includes the cis-acting noncoding gene,
Xist (6, 7), and its antisense counterpart, Tsix
(8). Xist RNA accumulation along the Xi ini-
tiates the silencing step (9, 10), whereas Tsix
represses silencing by blocking Xist RNA ac-
cumulation (11, 12).Acis-acting center for
choice and imprinting lies at the 5⬘end of Tsix,
as its deletion abolishes random choice in epi-
blast-derived cells to favor inactivation of the
mutated X (11, 13) and disrupts maternal Xist
imprinting in extraembryonic tissues (14, 15).
Thus, while imprinted XCI is parentally direct-
ed and random XCI is zygotically controlled,
both work through Tsix to regulate Xist.
To date, only X-linked cis-elements have
been identified as XCI regulators. Yet, virtually
all models invoke trans-acting factors which
interact with the X-linked sites. In one model
for imprinted XCI, a maternal-specific trans-
factor confers resistance to XCI (16). In models
for random XCI, an autosomally expressed
“blocking factor” protects a single X from si-
lencing (2). We have proposed that Tsix is the
cis-target of both trans-factors (11, 14).
To isolate candidate trans-factors, we now
used computational analysis (Fig. 1) to identify
mouse-to-human conserved elements within the
2- to 4-kilobase (kb) sequence implicated in
choice and imprinting (11, 13–15), a region
including DXPas34 (17). We found that the
region is composed almost entirely of 60 – to
70 – base pair (bp) repeats with striking resem-
blance to known binding sites for CTCF, a
transcription factor with a 60-bp footprint and
11 zinc fingers that work in various combina-
tions to generate a wide range of DNA-binding
activities (18). CTCF functions as a boundary
element at the globin locus (19), regulates en-
hancer access to the H19-Igf2 imprinted genes
(20–23), and associates with CTG/CAG repeats
Howard Hughes Medical Institute, Department of Mo-
lecular Biology, Massachusetts General Hospital, De-
partment of Genetics, Harvard Medical School, Bos-
ton, MA 02114, USA.
*To whom correspondence should be addressed. E-
mail: lee@frodo.mgh.harvard.edu
REPORTS
www.sciencemag.org SCIENCE VOL 295 11 JANUARY 2002 345