© 2001 Oxford University Press Human Molecular Genetics, 2001, Vol. 10, No. 26 3017–3024
SALL1, the gene mutated in Townes–Brocks syndrome,
encodes a transcriptional repressor which interacts
with TRF1/PIN2 and localizes to pericentromeric
Christian Netzer*, Leonie Rieger, Alessandro Brero1, Chang-Dong Zhang, Markus Hinzke,
Jürgen Kohlhase and Stefan K. Bohlander
Institute of Human Genetics, University of Göttingen, Heinrich-Düker-Weg 12, 37073 Göttingen, Germany and
1Institute of Anthropology and Human Genetics, University of Munich, Richard-Wagnerstraße 10/I, 80333 München,
Received August 20, 2001; Revised and Accepted October 26, 2001
The Townes–Brocks syndrome (TBS) is an auto-
somal dominantly inherited malformation syndrome
presenting as an association of imperforate anus,
triphalangeal and supernumerary thumbs, malformed
ears and sensorineural hearing loss. Mutations in
SALL1, a gene mapping to 16q12.1, were identified as
a cause for TBS. To elucidate how SALL1 mutations
lead to TBS, we have performed a series of functional
studies with the SALL1 protein. Using epifluorescence
and confocal microscopy it could be shown that a
GFP–SALL1 fusion protein localizes to chromo-
centers and smaller heterochromatin foci in tran-
siently transfected NIH-3T3 cells. Chromocenters
consist of clustered pericentromeric heterochromatin
and contain telomere sequences. Indirect immuno-
fluorescence revealed a partial colocalization of
GFP–SALL1 with M31, the mouse homolog of the
Drosophila heterochromatic protein HP1. It was
further demonstrated that SALL1 acts as a strong
transcriptional repressor in mammalian cells. Tran-
scriptional repression could not be relieved by the
addition of the histone deacetylase inhibitor Tricho-
statin-A. In a yeast two-hybrid screen we identified
PIN2, an isoform of telomere-repeat-binding factor 1
(TRF1), as an interaction partner of SALL1, and showed
that the N-terminus of SALL1 is not necessary for the
interaction with PIN2/TRF1. The interaction was
confirmed in vitro in a GST-pulldown assay. The
association of the developmental regulator SALL1
with heterochromatin is striking and unexpected.
Our results propose an involvement of SALL1 in the
regulation of higher order chromatin structures and
indicate that the protein might be a component of a
distinct heterochromatin-dependent silencing process.
We have also provided new evidence that there is a
close functional link between the centromeric and
telomeric heterochromatin domains not only in
Drosophila and yeast, but also in mammalian cells.
Townes–Brocks syndrome (TBS; OMIM 107480) is an autosomal
dominantly inherited malformation syndrome. Mutations in
SALL1, a gene mapping to 16q12.1, were identified as the
cause for TBS (1). The clinical presentation of TBS is highly
variable within and between affected families (2). Characteristic
features of TBS are anorectal abnormalities (imperforate anus,
anal stenosis), abnormalities of the hands (preaxial poly-
dactyly, triphalangeal thumbs), abnormalities of the feet
(syndactyly, pes planus, fused metatarsals), deformities of the
outer ear (‘lop ears’, microtia), preauricular tags and hearing
loss, which can be sensorineural, conductive or mixed. Renal
malformations have also been reported from several cases and
can lead to renal failure in TBS patients. Cardiac malformations as
well as mental retardation are rarely reported. In addition,
some families are known in which affected subjects show
features typical for both TBS and the Goldenhar syndrome/
oculo-auriculo-vertebral spectrum (OMIM 164210). Penetrance
seems to be complete in TBS (2).
SALL1 has homologies to the essential developmental
regulator gene spalt (sal) of Drosophila (3). In Drosophila, sal
is required for the specification of posterior head and anterior
tail segment identity (4) as well as for larval tracheal system
development (5) and adult wing development (6). In the wing
imaginal discs, sal is activated in response to hedgehog signalling
mediated by the TGFβ-like protein DPP (7–9).
Prior to the cloning of SALL1, sal-related genes had been
isolated from mouse (Msal, now Sall3) (10) and from Xenopus
*To whom correspondence should be addressed. Tel: +49 551 397592; Fax: +49 551 399303; Email: firstname.lastname@example.org.
Stefan K.Bohlander and Chang-Dong Zhang, Department of Medicine III, University of Munich, and GSF, Clinical Cooperative Group ‘Leukemia’,
Marchioninistrasse 25, 81377 München, Germany
3018 Human Molecular Genetics, 2001, Vol. 10, No. 26
laevis (Xsal-1) (11). Both genes are expressed in the developing
limbs, heart, kidney, inner ear and central nervous system,
i.e. in tissues/organs affected in TBS. Human SALL1 is very
similar in structure to both Msal and Xsal-1, and shows a
similar expression pattern in adult tissues (12). The observation
that the sal gene in the fish Medaka is activated in response to
Sonic hedgehog suggests that not only the SAL-like protein
structure was conserved in evolution but that SAL-like
proteins might also act as essential developmental regulators in
The ORF of SALL1 encodes a protein of 1325 amino acids
with four characteristically arranged SAL-like C2H2 double
zinc finger domains (Fig. 1). A single zinc finger is attached to
the second double zinc finger domain (12). All known vertebrate
SAL-like proteins contain an additional C2HC zinc finger
close to the N-terminus (11–14). The majority of the SALL1
mutations detected in TBS patients to date result in preterminal
stop codons (12). It is very likely that TBS is caused by a
haploinsufficiency of SALL1. Among typical TBS patients, we
reached a detection rate of SALL1 mutations of 64.3% (15).
However, there are a number of families with typical TBS-like
phenotypes in which SALL1 mutations were not found.
To elucidate how SALL1 mutations lead to TBS, we
performed a series of functional studies with the SALL1
protein. First, we determined the intracellular localization of a
GFP–SALL1 fusion protein in NIH-3T3 cells by epifluorescence
and confocal microscopy. Secondly, we tested SALL1 for its
ability to repress or activate gene expression in transient
transfection assays. Thirdly, we performed a yeast two-hybrid-
screen to identify proteins interacting with SALL1.
SALL1 colocalizes with heterochromatin
NIH-3T3 cells (mouse fibroblasts) were transfected transiently
with pEGFP–SALL1 to determine the intracellular localization
of the fusion protein in interphase cells. By epifluorescence
microscopy, we found an exclusive nuclear localization of
GFP–SALL1 in almost all cells. However, in some cells GFP–
SALL1 was also present in the cytoplasma (data not shown).
The cytoplasmatic fraction often looked like inclusion bodies
and possibly represented membrane-bound vesicles that were
formed due to the relative overexpression in certain cells. In all
transfected cells (even in those few cells with a cytoplasmatic
GFP–SALL1 fraction) GFP–SALL1 was distributed in the
nucleus as distinct aggregates (Fig. 2A). These aggregates
strikingly corresponded to the 4′,6-diamidin-2-phenylindol
(DAPI)-bright regions of the nucleus (Fig. 2B). The DAPI-
bright regions in the nuclei of NIH-3T3 cells represent the
chromocenters and consist of clustered pericentromeric
Confocal microscopy was used in order to analyse the
pattern of the GFP–SALL1 distribution at a higher resolution.
Heterochromatin (and hence chromocenters) were stained with
TO-PRO®-3. This technique revealed distinct distribution
patterns of GFP–SALL1: in the nucleus depicted in Figure 2C
and D GFP–SALL1 shows the previously described
colocalization with the chromocenters, which were located at
the nuclear membrane or at the periphery of the nucleoli. In the
nucleus shown in Figure 2E and F GFP–SALL1 not only
colocalizes with the chromocenters, but is also distributed in
smaller foci which seem to cover the inner surface of the
nuclear membrane and the outer surface of the nucleoli. A
tangential section of such a cell (Fig. 2G) reveals that SALL1
covers the inner surface of the nuclear membrane in a mesh-like
fashion. The nuclei depicted in Figure 2A–F show a homo-
genous distribution of GFP–SALL1 within the chromocenters.
However, in other nuclei a pattern was observed which looked
like coating of the outer surface of the chromocenters (Fig. 2H–J).
In these cells, the coating did not cover the complete surface of the
chromocenters, but contained openings.
SALL1 partially colocalizes with M31/HP1
Since the mammalian homologs of Drosophila hetero-
chromatin protein 1 (HP1) are known to localize to constitutive
heterochromatin, we used indirect immunofluorescence with a
monoclonal antibody (raised against M31, the murine HP1
homolog) to stain the pericentromeric heterochromatin.
Epifluorescence microscopy (Fig. 2K–M) and confocal microscopy
(Fig. 2N–P) revealed that GFP–SALL1 and M31/HP1 partially
colocalize at the chromocenters and showed a M31/HP1 distri-
bution in the nucleus as described by Minc et al. (16) and
Wreggett et al. (17). Comparing the distribution patterns of the
two proteins in detail showed that most subregions of the
chromocenters are either exclusively covered by GPP–SALL1
or by M31/HP1.
SALL1 acts as a transcriptional repressor in mammalian
Since one of the hallmarks of heterochromatin is that it
constitutes a transcriptionally repressive environment, we next
wished to examine whether SALL1 acts as a transcriptional
repressor. To this end, SALL1 was tested for its transcriptional
repression properties in a reporter gene assay. A series of
transient transfections of NIH-3T3 cells was performed with a
full-length SALL1 fusion protein (Fig. 1) with the GAL4-DNA-
binding domain (GAL4-DB). As a reporter plasmid we used
pGAL45tkLUK, which contains the luciferase gene under the
control of a thymidine kinase promoter with a GAL4 binding
site. The results of these assays are shown in Figure 3. GAL4-
DB–SALL1 very strongly represses the activity of this
promoter. Luciferase expression was repressed ∼20-fold
(Fig. 3, lane 3). We next tested if transcription repression of
Figure 1. Diagram of full-length SALL1 and of the three SALL1 deletion
mutants (SALL1∆1, SALL1∆2 and SALL1∆3) used in this study. The ORF of
SALL1 is shown as a box. The positions of the double zinc fingers (black) and
of the single zinc fingers (white) are depicted as ovals. The names of the
mutants are indicated on the left side.
Human Molecular Genetics, 2001, Vol. 10, No. 26 3019
SALL1 is relieved by the addition of the histone deacetylase
inhibitor Trichostatin-A (TSA). As a control the central
domain (amino acids 218–345) of the ets transcription factor
ETV6 (Fig. 3, lane 5) was used. Transcription repression of the
Figure 2. Localization of GFP–SALL1 (A–J) and colocalization of GFP–SALL1 with M31/HP1 (K–P) in NIH-3T3 cells. (A, B, K–M) Epifluorescence
microscopy images. (C–F, H–J, N–P) Confocal laser microscopy images, mid-sections and (G) tangential section. (A and B) Epifluorescence images showing a
nucleus expressing GFP–SALL1 (A) and the corresponding DAPI staining (B). Note the distribution of GFP–SALL1 in distinct aggregates which correspond to
the DAPI-bright regions representing chromocenters (arrow). (C and D) Confocal images showing the colocalization of GFP–SALL1 (C) with chromocenters (D;
TO-PRO®-3 staining), which are usually located at the nuclear membrane (arrowhead) or at nucleoli (arrow). Note the homogenous distribution of GFP–SALL1
within chromocenters. (E and F) Confocal images illustrating that GFP–SALL1 (E) is additionally distributed in smaller foci which seem to cover the inner surface
of the nuclear membrane (arrow) and the outer surface of the nucleoli (arrowhead). (F) Corresponding TO-PRO®-3 image. (G) A tangential section of a nucleus
reveals that GFP–SALL1 covers the inner surface of the nuclear membrane in a mesh-like fashion. (H–J) Confocal image showing an inhomogenous distribution
of GFP–SALL1 (H) at heterochromatin masses which resembles a coating of the outer surface of the chromocenters (I; TO-PRO®-3 staining). The arrow points at
one of the incomplete ring structures with openings. (J) Merged picture from (H) (green) and (I) (red). (K–M) Epifluorescence images showing that GFP–SALL1
(K) is distributed in the nucleus in distinct aggregates. The arrow points at one of these aggregates to demonstrate the colocalization of the GFP–SALL1 signal (K)
with the M31/HP1 signal (L) at DAPI bright regions representing the chromocenters (M; DAPI staining). (N–P) Confocal images showing that the colocalization
of the GFP–SALL1 signals (N) with the corresponding M31/HP1 signals (O) is incomplete. In (P) the images from (N) (green) and (O) (red) are merged. Bar = 5 µm
in (B, F, G, I and M).
3020 Human Molecular Genetics, 2001, Vol. 10, No. 26
ETV6 central domain is known to be mediated through histone
deacetylases (18) (S.K.Bohlander, J.Putnik, S.Bartels and
M.Kickstein, manuscript in preparation). The results, shown in
Figure 3, lane 4, clearly indicate that TSA had no effect on the
transcription repression of SALL1. However, the repression by
the central domain is relieved in the presence of TSA (Fig. 3,
SALL1 interacts with PIN2 in yeast
The yeast two-hybrid system was applied to identify proteins
that interact with SALL1 using pGBT9–SALL1∆2 (encoding
amino acids 87–1058 of SALL1; Fig. 1) as a bait. Among 2 ×
106 transformants we identified one clone that was positive for
the expression of the selection markers (β-galactosidase and
histidine). Subsequent sequence analysis showed that the clone
had a 792 bp insert containing an open reading frame (ORF) of
264 amino acids. A database search revealed 100% homology
with the human PIN2 protein (amino acids 67–331; Fig. 4), an
isoform of telomere repeat binding factor 1 (TRF1).
To determine the domains of SALL1 necessary for the inter-
action with PIN2, pGAD-GH–∆PIN2 was cotransformed with
pGBT9–SALL∆1, –SALL1∆2, –SALL1∆3 and –SALL1
(Fig. 1) into yeast strain CG1945 and assayed for β-galacto-
sidase and histidine activity. Full length SALL1, SALL1∆2
and SALL1∆3, but not SALL1∆1 interacted with ∆PIN2
(Fig. 5; data for complete SALL1 not shown). This indicates
that the N-terminus of SALL1 is not necessary for interaction
SALL1 interacts with PIN2 in vitro
To confirm that SALL1 and PIN2 interact physically, we performed
a GST-pulldown assay. The fusion protein GST–∆PIN2 (amino
acids 67–331 of complete PIN2; Fig. 4) was expressed,
purified and incubated with in vitro translated 35S-Met-labeled
SALL1∆2. After extensive washing, the proteins bound to
glutathione sepharose beads were separated by SDS–PAGE
and detected by autoradiography. The result of the assay is
shown in Figure 6: GST–∆PIN2 (lane 2), but not GST alone
(lane 3) is able to efficiently retain SALL1∆2.
Using epifluorescence microscopy we have shown that in
interphase NIH-3T3 mouse fibroblasts the GFP–SALL1 fusion
protein is concentrated in DAPI-bright regions of the nucleus,
which are known to represent chromocenters. The chromocenters
of mouse nuclei consist of pericentromeric heterochromatin
(19) and contain some telomeres, since mouse chromosomes
are acro- or telocentric (20–22). Confocal microscopy with
TO-PRO®-3 heterochromatin staining confirmed the observed
SALL1 localization and revealed that GFP–SALL1 exhibits
different distribution patterns within the nuclei: in some cells
the protein was not only localized at the chromocenters, but
was also distributed in smaller foci which seemed to cover the
inner surface of the nuclear membrane and the outer surface of
Figure 3. SALL1 acts as a TSA-independent repressor. NIH-3T3 cells were
transiently transfected with the empty pM1 vector (lanes 1 and 2), pM1–SALL1
(lanes 3 and 4) or pM1–ETV6 (containing a TSA-sensitive repression domain;
lanes 5 and 6). The luciferase activity was normalized for transfection
efficiency. Gray bars indicate repression activity after addition of TSA (lanes 2,
4 and 6). Full-length SALL1 (lane 3) represses luciferase activity ∼20-fold
(compared to the luciferase activity of NIH-3T3 cells transfected with the
reporter plasmid plus empty pM vector). Addition of TSA had no effect on
SALL1 repression activity (lane 4), but relieved the repression induced by the
ETV6 central domain (lanes 5 and 6).
Figure 4. Schematic presentation of the domain structure of TRF1, PIN2 and of
the PIN2 fragment (∆PIN2) isolated from the HeLa cDNA library. Shaded areas
indicate from left to right an acid (D/E)-rich domain, a D-like box, a bipartite
nuclear localization signal and a Myb-type HTH DNA-binding domain. PIN2
(and ∆PIN2) carries an internal 20 amino acid deletion. ∆PIN2 consists of
amino acids 67–331 of full-length PIN2.
Figure 5. Interaction of SALL1 with PIN2 in yeast. Yeast strain CG-1945 was
cotransformed with pGAD-GH–∆PIN2 and pGBT9–SALL1∆1/∆2/∆3 as
indicated in (B). Transformed CG-1945 cells containing both plasmids were
streaked out on plates lacking tryptophan, leucine and histidine in the presence
of 10 mM 3-AT (A). Yeast cells were tested for β-galactosidase activity by a
filter assay (C).
Human Molecular Genetics, 2001, Vol. 10, No. 26 3021
the nucleoli. These differences might reflect cell-cycle
dependent changes in heterochromatin organization and hence
GFP–SALL1 localization. Interestingly, we never observed a
GFP–SALL1-transfected cell in mitosis or arrested in metaphase
after colchicin treatment. Therefore, overexpression of SALL1
might either be toxic or lead to cell-cycle arrest. Further experi-
ments are needed to interpret this observation and to examine
the localization of native SALL1 at different stages of the cell
cycle. Nevertheless, at least three independent lines of
evidence strongly argue against the possibility that the
observed association of SALL1 with heterochromatin is an
artefact of the overexpression of the fusion protein. First, GFP
alone shows a diffuse cytoplasmic and nuclear localization
without any accumulation at DAPI-bright regions in the nucleus
(data not shown). Secondly, the localization of GFP–SALL1 is
highly specific; the protein is (in the nucleus) exclusively
found at heterochromatic sites. Thirdly, we identified a SALL1
interaction partner (PIN2/TRF1) that binds telomeric hetero-
chromatin and also localizes to chromocenters (22,23).
Taking into account that mutations of SALL1 lead to a well
defined and specific phenotype, the accumulation of SALL1 at
sites of highly repetitive DNA is striking and unexpected. Our
results suggest that SALL1 is not a ‘classical’ transcription
factor regulating the expression of a few target genes, but that
the protein is involved in the regulation of higher order
chromatin structures such as pericentromeric heterochromatin.
There are only two other examples of human malformation
syndromes that are caused by mutations in a gene coding for a
heterochromatin-associated protein. (i) The ATR-X syndrome,
a severe form of syndromal mental retardation characterized by
the presence of α-thalassemia with urogenital abnormalities and
facial dysmorphism, is caused by mutations of the X-chromosomal
gene hATRX. The ATRX protein is a putative transcriptional
regulator which is localized at the pericentromeric hetero-
chromatin and at the short arms of acrocentric chromosomes (24).
(ii) The ICF syndrome (an acronym for immunodeficiency,
centromeric instability and facial anomalies), an autosomal
recessive disorder, is caused by mutations in the DNA methyl-
transferase DNMT3B and involves extensive loss of methylation
from pericentromeric regions (25). Just recently it was shown
that DNMT3B is a methylation-independent transcriptional
repressor and colocalizes with HP1α to pericentromeric
heterochromatin regions in murine embryonic stem cells (26).
We have shown that SALL1 is a strong transcriptional
repressor when it is tethered to a promoter. This effect is
apparently not mediated by histone-deacetylases. The exact
repression mechanism remains to be determined. Since the
SALL1 mutations detected in Townes–Brocks syndrome are
believed to result in haploinsufficiency, the Townes–Brocks
phenotype might be caused by the up-regulation of SALL1 target
genes. The fact that SALL1 is located at heterochromatin regions
suggests that the protein is involved in the establishment or
stabilization of a transcriptionally repressive environment.
Changes in chromatin organization and the directed formation
of heterochromatin-like complexes are thought to be important
for regulating gene expression during development (27). For
example, Ikaros, another zinc finger protein that binds centro-
meric heterochromatin, is required for normal T, B and natural
killer cell development and is thought to repress genes by
selectively recruiting them to centromeric heterochromatin
(28). Taking into account the fact that SALL1 exhibits a
spatiotemporally restricted expression profile, our findings
indicate that SALL1 might be a component of such a distinct
heterochromatin-dependent silencing process.
A similar clustering at chromocenters as observed for
SALL1 has been reported for M31, the murine homolog of the
Drosophila heterochromatic protein HP1. Therefore, we
visualized HP1/M31 in GFP–SALL1 transfected cells by
indirect immunofluorescence. As would be expected from the
published data on the intracellular localization of HP1/M31
(16,17), the signals detected from both proteins partially over-
lapped at DAPI-bright regions of the nucleus. In addition,
HP1/M31 staining revealed smaller spots that were not over-
lapping with GFP–SALL1. These signals most likely represent
unspecific background, as they were also detectable in the
cytoplasma (data not shown). Interestingly, most subregions of
the chromocenters were either covered exclusivly by SALL1
or by HP1. We did not perform any experiments to test whether
the two proteins directly interact with each other. HP1 was first
described in Drosophila (29) as a heterochromatin-associated
protein with dosage-dependent effects on heterochromatin-
induced gene silencing known as position effect variegation
(30). In Drosophila, HP1 is also associated with telomeres and
is thought to be involved in preventing telomere fusions (31).
In Schizosaccharomyces pombe, the swi6 gene encodes a HP1-like
chromodomain protein that localizes to heterochromatin domains,
including the centromeres and telomeres, and is involved in
silencing at these loci (32). A number of interactions with other
proteins have been reported for the mammalian homologs of
HP1, among which the interaction of HP1α with human Ku70
is especially intriguing (33). Ku, which is a heterodimer of 70
and 80 kDa subunits, is involved in DNA repair and the main-
tenance of telomeres (34–36). Ku70 also interacts with TRF2,
a mammalian telomere-binding protein (37). It has been
suggested that HP1α is a mammalian counterpart of the yeast
Sir4 protein and represents a telomere protein of the silencing
There seem to be some striking parallels between HP1 and
SALL1 not only concerning their intracellular localization, but
also regarding functional aspects, since our yeast two-hybrid
screen identified a SALL1 interaction partner which also acts at
telomeres. PIN2 is a splice variant of the telomere-repeat-binding
protein TRF1, carrying an internal 20 amino acid deletion (23).
TRF1/PIN2 is involved in regulating telomere length. Long-term
overexpression of TRF1 in a telomerase-positive tumor cell
line results in progressive telomere shortening, whereas inhibition
of TRF1 induces telomere elongation (38). PIN2 is more abun-
dant in cells than TRF1, but functional differences between
PIN2 and TRF1 have not been reported. So far, TRF1 protein
interactions with Tankyrase, TIN2 and NBSI have been
reported. Tankyrase, a protein with homology to ankyrins and
the catalytic domain of poly(ADP-ribose) polymerase, colocal-
izes with telomeres and can ribosylate both itself and TRF1 in
vitro (39). Ribosylation of TRF1 inhibits its binding to telo-
meric DNA. TIN2 is thought to mediate TRF1 function and
negatively regulates telomere length (40). NBS1 colocalizes with
TRF1 at PML bodies during late S/G2 phases in immortalized
telomerase-negative cells (41) and is encoded by the gene
mutated in Nijmegen breakage syndrome, a chromosomal
instability disorder. It has been suggested that NBS1 may be
involved in alternative lengthening of telomeres in telomerase-
negative immortalized cells (41).
3022 Human Molecular Genetics, 2001, Vol. 10, No. 26
TRF1/PIN2 binds telomeres in interphase and will therefore
partially colocalize with the chromocenters of mouse nuclei
and with SALL1 (22,23). The colocalization between these
two proteins in interphase can be expected to be incomplete,
since not all telomeres of mouse chromosomes cluster at the
chromocenters and the telomeres only occupy a small portion
of the chromocenters. At the present time, many questions
concerning the exact nature of the interaction of the two
proteins and their role for telomere function and/or hetero-
chromatin formation and maintenance remain to be answered.
These matters have just become more complicated by the first
report of an extra-telomere function and localization of TRF1/PIN2:
the protein was localized at the mitotic spindle, suggesting a
new role for TRF1/PIN2 in modulating the function of
microtubules during mitosis (42). Futhermore, it is tempting to
speculate that the interaction between TRF1/PIN2 and SALL1
is involved in the recently reported telomere position effect in
human cells, which results in reversible silencing of genes near
telomeres depending on the length of the adjacent telomere
(43). We have provided strong evidence that besides HP1α
(which interacts with Ku70) a second mammalian hetero-
chromatin-associated protein is interacting with a protein
involved in telomere function. With these data a picture of a
close functional link between the centromeric and telomeric
heterochromatin domains not only in Drosophila and yeast,
but also in mammalian cells is emerging.
MATERIALS AND METHODS
Various cDNA fragments of SALL1 were cloned into pBluescript
as described previously (44). We subcloned SALL1∆1 (bp 1–549
of the complete coding cDNA of SALL1; Fig. 1), SALL1∆2
(bp 261–3175; Fig. 1) and SALL1∆3 (bp 2067–3975; Fig. 1)
into the yeast expression vector pGBT9 (Clontech, Palo Alto, CA)
by using standard techniques to produce fusion proteins with the
GAL4-DNA-binding domain. The construct pGBT9–SALL1
(containing the complete ORF of SALL1; Fig.1) was generated
by a multistep subcloning strategy using appropriate restriction
enzymes. The full-length SALL1 cDNA was inserted (i) into
pEGFP-C1 (CLONTECH) to generate a GFP fusion protein,
and (ii) into pM1 (CLONTECH), a mammalian vector used for
expression of hybrid proteins with the DNA-binding domain of
GAL4. By further subcloning, we generated pM2-SALL1∆2.
As a reporter in the transcription repression assays, the
plasmid pGAL45tkLUC was used. This plasmid was constructed
by replacing the chloramphenicol acetyltransferase gene (CAT)
in pGAL45tkCAT (45) with the firefly luciferase gene. This
was accomplished by excising the CAT gene from
pGAL45tkCAT using BglII and SMI and replacing it by a
BglII–SalI fragment containing the luciferase gene from the
pGL3-Basic plasmid (Promega, Madison, WI). The resulting
construct expresses the luciferase gene under the control of the
Herpes simplex virus thymidine kinase promoter.
Cell transfection and fixation
NIH-3T3 cells were grown on 18 × 18 mm coverslips in 5 ml
DMEM with 10% FCS and 1% antibiotics at 37°C in a humidified
5% CO2 atmosphere. Transfection with 10 µg pEGFP–SALL1
was carried out with Roti®-Fect (ROTH, Karlsruhe, Germany)
according to the manufacturer’s instructions. Forty-eight hours
after transfection cells were washed in 1× PBS and fixed by a
10 min incubation in 3.7% formaldehyde (Sigma, Taufkirchen,
Germany)/1× PBS at room temperature.
For indirect immunofluorescence cells were permeabilized using
0.5% Triton X-100 (Calbiochem, Bad Soden, Germany)/1× PBS
for 20 min at room temperature. To avoid unspecific antibody
binding, cells were incubated for 15 min at 37°C in a blocking
solution containing 4% BSA in PBT [(1× PBS, 0.02% Tween
20 (Calbiochem)]. HP1/M31 detection was accomplished
using a commercially available monoclonal rat antibody
(Serotec; 10–50 µg/ml) as primary antibody (undiluted) and a
Cy3-conjugated goat anti-rat antibody (1 mg/ml; Amersham
Pharmacia Biotech, Freiburg, Germany) as secondary antibody
(diluted 1:500 in blocking solution). Antibody incubation was
performed at 37°C for 30 min. Between the two detection
layers, cells were washed in PBT (3 × 3 min at 37°C). Nuclei
were counterstained with DAPI (0.05 µg/ml) for epifluorescence
microscopy or TO-PRO®-3 (Molecular Probes, Leiden, The
Netherlands; 1 µM) for confocal microscopy (DAPI could not
be used for confocal imaging as the confocal microscope used
was not equipped with a UV-laser. TO-PRO®-3 reveals the
same nuclear staining pattern as DAPI). Coverslips were
mounted on slides using Vectashield antifade (Vector Laboratories,
Burlingame, CA) and sealed with nail polish.
Epifluorescence images were obtained using a Zeiss Axiophot
II microscope (Zeiss, Jena, Germany) equipped with a Zeiss
Plan-Apochromat 40×/1.4 NA objective lens, and with single-
band pass filter sets for visualization of green (GFP), red (Cy3)
and infrared (TO-PRO®-3) fluorescence. For image acquisition
a cooled charged couple device camera (Photo Science Ltd,
Millham, UK) was used; 8 bit grayscale images were recorded
and merged to RGB images via Cytovision 2.7 software
(Applied Imaging International Ltd, Newcastle-Upon-Tyne,
UK). Pictures were processed with Adobe Photoshop 6.0
(Adobe Systems, Mountain View, CA).
Figure 6. GST-pulldown with SALL1 and PIN2. In vitro translated (ivt)
SALL1∆2 (lane 1) was incubated with GST–∆PIN2 (lane 2) or with GST alone
(lane 3), washed and separated by SDS–PAGE as described in Materials and
Methods. SALL1∆2 was only retained by GST–∆PIN2, not by GST alone.
Human Molecular Genetics, 2001, Vol. 10, No. 26 3023
Optical sections of 324 or 203 nm were acquired with a Leica
TCS SP confocal laser scanning microscope (Leica Micro-
systems, Heidelberg, Germany) equipped with an oil immersion
Plan-Apochromat 100×/1.4 NA objective lens. Fluorochromes
were visualized either using an argon laser with excitation
wavelengths of 488 nm (for GFP) and 568 nm (for Cy3), or
using a helium–neon laser with an excitation wavelength of
633 nm (for TO-PRO®-3). Image resolution was 512 × 512
pixels with a pixel size ranging from 25 to 55 nm depending on
the selected zoom factor. RGB pictures were generated by
merging confocal images in Adobe Photoshop 6.0.
Transcription repression assay
NIH-3T3 cells (1.4 × 105 cells/35 mm plate) were transfected
with Roti®-Fect (ROTH). Each transfection assay included
(i) 0.5 µg of the pM1–SALL1 construct, (ii) 0.5 µg of the
pGAL45tkLUC reporter plasmid and (iii) 0.1 µg of pCMV-β-
Gal (Clontech), expresssing the β-galactosidase enzyme.
Forty-eight hours after transfection, cells were harvested, lysed
in 100 µl of lysis buffer [100 mM KH2PO4 pH 7.8, 0.2% (v/v)
Triton X-100, 0.5 mM DTT] and assayed for luciferase and
β-galactosidase activity with an Autolumat LB953 (Berthold,
Wildbad, Germany) as described by Schlüter et al. (46).
Trichostatin A (TSA) was added to the culture medium after
transfection at a final concentration of 500 ng/ml. The levels of
β-galactosidase expression were used to normalize the
efficiency of transfection. All assays were repeated at least five
times. The median of repression activity and the SD were
Yeast two-hybrid screen
We used the Matchmaker Two-Hybrid Kit (Clontech) with
pGBT9–SALL1-∆2 as a bait. The assay was performed as
recommended by the manufacturer. In brief, yeast cells (strain
CG-1945) containing the bait plasmid pGBT9–SALL1∆2 were
transformed with a HeLa cDNA library (Clontech) using the
lithium acetate method. We screened approximately 2 × 106
transformants for growth on synthetic dropout (SD) plates
lacking histidine, leucine, and tryptophan in the presence of
10 mM 3-amino-1,2,4-triazol (3-AT). Colonies growing on the
SD plates were analyzed for β-galactosidase activity by filter
assay. The pGAD-GH plasmids containing the cDNA
sequence of the putative SALL1 interaction partners were
recovered from the yeast cells, transformed into Escherichia
coli DH5α and sequenced. The plasmids were then retrans-
formed into the CG-1945 yeast and assayed for growth on SD
plates lacking histidine and leucine to exclude DNA-binding
activity. In a second round of testing, the pGAD-GH plasmids
and pGBT9–SALL1∆2 were cotransformed into CG-1945, and
transformants were assayed again for growth on SD plates
lacking histidine and for β-galactosidase activity in a filter
Mapping of interacting domains
To map the SALL1 domains responsible for the interaction
with ∆PIN2, yeast strain CG-1945 was cotransformed with
SALL1 and its deletion mutants (SALL1∆1, SALL1∆2 and
SALL1∆3; Fig. 1) cloned into pGBT9 (Clontech) and ∆PIN2
(Fig. 4) cloned into pGAD-GH (Clontech). Growth was
assayed on SD plates lacking leucine and tryptophane (trans-
formation control) or SD plates supplemented with 10 mM 3-
AT lacking leucine, tryptophane and histidine (interaction
assay). Colonies growing on SD plates lacking leucine, tryp-
tophane and histidine were tested for β-galactosidase activity
in a filter assay.
Glutathion S-transferase (GST)-pulldown and in vitro
The ∆PIN2-cDNA was cut out of pGAD-GH and cloned into
the pGEX-KT vector (Amersham Pharmacia) to produce the
fusion protein GST–∆PIN2 (amino acids 67–331 of PIN2;
Fig. 4). Expression and purification of GST–∆PIN2 and GST
alone (as a control) using glutathione Sepharose 4B (Amersham
Pharmacia) were performed as recommended by the manufacturer.
In vitro translation of SALL1∆2 (amino acids 87–1058 of
SALL1; Fig. 1) was performed with rabbit reticulocyte lysate
(Promega) and 35S-methionine according to the manufacturer’s
instruction. The glutathione Sepharose beads with bound GST
fusion proteins (40 µl) and 22 µl of in vitro-translated
SALL1∆2 were incubated in 200 µl of bead-binding buffer
[50 mM K-phosphate pH 7.5, 100 mM KCl, 10% glycerol (v/v),
0.1% Triton X-100] at 4°C for 2 h. The beads were washed
four times with bead-binding buffer devoid of glycerol and
Triton X-100. Beads were then heated for 5 min at 95°C in
SDS-sample buffer and analyzed by SDS–PAGE. Radio-
labeled bands were visualized by autoradiography.
The authors would like to thank Wolfgang Engel for his
support and critically reading the manuscript, Nicole Richter
for technical assistance and Folker Garbe for practical help.
We are also grateful to Harry Scherthan for helpful discussions
and Thomas Cremer for support of the project. This study was
funded by the Fritz-Thyssen-Stiftung (grant no. 2000 1071).
1. Kohlhase, J., Wischermann, A., Reichenbach, H., Froster, U. and
Engel, W. (1998) Mutations in the SALL1 putative transcription factor
gene cause Townes–Brocks syndrome. Nat. Genet., 18, 81–83.
2. Powell, C.M. and Michaelis, R.C. (1999) Townes–Brocks syndrome.
J. Med. Genet., 36, 89–93.
3. Kühnlein, R.P., Frommer, G., Friedrich, M., Gonzalez-Gaitan, M.,
Weber, A., Wagner-Bernholz, J.F., Gehring, W.J., Jäckle, H. and
Schuh, R. (1994) spalt encodes an evolutionarily conserved zinc finger
protein of novel structure which provides homeotic gene function in the
head and tail region of the Drosophila embryo. EMBO J., 13, 168–179.
4. Jürgens, G. and Weigel, D. (1988) Terminal versus segmental development in
the Drosophila embryo: the role of the homeotic gene fork head. Roux’s
Arch. Dev. Biol., 197, 345–354.
5. Kühnlein, R.P. and Schuh, R. (1996) Dual function of the region-specific
homeotic gene spalt during Drosophila tracheal system development.
Development, 122, 2215–2223.
6. Sturtevant, M.A., Biehs, B., Marin, E. and Bier, E. (1997) The spalt gene
links the A/P compartment boundary to a linear adult structure in the
Drosophila wing. Development, 124, 21–32.
7. de Celis, J.F., Barrio, R. and Kafatos, F.C. (1996) A gene complex acting
downstream of dpp in Drosophila wing morphogenesis. Nature, 381,
8. Lecuit, T., Brook, W.J., Ng, M., Calleja, M., Sun, H. and Cohen, S.M.
(1996) Two distinct mechanisms for long-range patterning by Decapentaplegic
in the Drosophila wing. Nature, 381, 387–393.
3024 Human Molecular Genetics, 2001, Vol. 10, No. 26
9. Nellen, D., Burke, R., Struhl, G. and Basler, K. (1996) Direct and
long-range action of a DPP morphogen gradient. Cell, 85, 357–368.
10. Ott, T., Kaestner, K.H., Monaghan, A.P. and Schutz, G. (1996) The mouse
homolog of the region specific homeotic gene spalt of Drosophila is
expressed in the developing nervous system and in mesoderm-derived
structures. Mech. Dev., 56, 117–128.
11. Hollemann, T., Schuh, R., Pieler, T. and Stick, R. (1996) Xenopus Xsal-1,
a vertebrate homolog of the region specific homeotic gene spalt of
Drosophila. Mech. Dev., 55, 19–32.
12. Kohlhase, J., Schuh, R., Dowe, G., Kühnlein, R.P., Jäckle, H., Schroeder, B.,
Schulz-Schaeffer, W., Kretzschmar, H.A., Köhler, A., Müller, U. et al.
(1996) Isolation, characterization, and organ-specific expression of two
novel human zinc finger genes related to the Drosophila gene spalt.
Genomics, 38, 291–298.
13. Köster, R., Stick, R., Loosli, F. and Wittbrodt, J. (1997) Medaka spalt acts
as a target gene of hedgehog signaling. Development, 124, 3147–3156.
14. Ott, T., Kaestner, K.H., Monaghan, A.P. and Schütz, G. (1996) The mouse
homolog of the region specific homeotic gene spalt of Drosophila is
expressed in the developing nervous system and in mesoderm-derived
structures. Mech. Dev., 56, 117–128.
15. Kohlhase, J. (2000) SALL1 mutations in Townes–Brocks syndrome and
related disorders. Hum. Mutat., 16, 460–466.
16. Minc, E., Allory, Y., Worman, H.J., Courvalin, J.C. and Buendia, B.
(1999) Localization and phosphorylation of HP1 proteins during the cell
cycle in mammalian cells. Chromosoma, 108, 220–234.
17. Wreggett, K.A., Hill, F., James, P.S., Hutchings, A., Butcher, G.W. and
Singh, P.B. (1994) A mammalian homologue of Drosophila heterochromatin
protein 1 (HP1) is a component of constitutive heterochromatin.
Cytogenet. Cell Genet., 66, 99–103.
18. Chakrabarti, S.R. and Nucifora, G. (1999) The leukemia-associated gene
TEL encodes a transcription repressor which associates with SMRT and
mSin3A. Biochem. Biophys. Res. Commun., 264, 871–877.
19. Brown, S.W. (1966) Heterochromatin. Science, 151, 417–425.
20. Vourc’h, C., Taruscio, D., Boyle, A.L. and Ward, D.C. (1993) Cell
cycle-dependent distribution of telomeres, centromeres, and
chromosome-specific subsatellite domains in the interphase nucleus of
mouse lymphocytes. Exp. Cell Res., 205, 142–151.
21. Scherthan, H., Weich, S., Schwegler, H., Heyting, C., Harle, M. and
Cremer, T. (1996) Centromere and telomere movements during early
meiotic prophase of mouse and man are associated with the onset of
chromosome pairing. J. Cell Biol., 134, 1109–1125.
22. Scherthan, H., Jerratsch, M., Li, B., Smith, S., Hulten, M., Lock, T. and
deLange, T. (2000) Mammalian meiotic telomeres: protein composition and
redistribution in relation to nuclear pores. Mol. Biol. Cell, 11, 4189–4203.
23. Shen, M., Haggblom, C., Vogt, M., Hunter, T. and Lu, K.P. (1997)
Characterization and cell cycle regulation of the related human telomeric
proteins Pin2 and TRF1 suggest a role in mitosis. Proc. Natl Acad. Sci.
USA, 94, 13618–13623.
24. McDowell, T.L., Gibbons, R.J., Sutherland, H., O’Rourke, D.M.,
Bickmore, W.A., Pombo, A., Turley, H., Gatter, K., Picketts, D.J.,
Buckle, V.J. et al. (1999) Localization of a putative transcriptional
regulator (ATRX) at pericentromeric heterochromatin and the short arms of
acrocentric chromosomes. Proc. Natl Acad. Sci. USA, 96, 13983–13988.
25. Hansen, R.S., Wijmenga, C., Luo, P., Stanek, A.M., Canfield, T.K.,
Weemaes, C.M. and Gartler, S.M. (1999) The DNMT3B DNA
methyltransferase gene is mutated in the ICF immunodeficiency
syndrome. Proc. Natl Acad. Sci. USA, 96, 14412–14417.
26. Bachman, K.E., Rountree, M.R. and Baylin, S.B. (2001) Dnmt3a and
Dnmt3b are transcriptional repressors that exhibit unique localization
properties to heterochromatin. J. Biol. Chem., 26, 26.
27. Jones, D.O., Cowell, I.G. and Singh, P.B. (2000) Mammalian chromodomain
proteins: their role in genome organisation and expression. Bioessays, 22,
28. Brown, K.E., Guest, S.S., Smale, S.T., Hahm, K., Merkenschlager, M. and
Fisher, A.G. (1997) Association of transcriptionally silent genes with
Ikaros complexes at centromeric heterochromatin. Cell, 91, 845–854.
29. James, T.C. and Elgin, S.C. (1986) Identification of a nonhistone chromosomal
protein associated with heterochromatin in Drosophila melanogaster and
its gene. Mol. Cell. Biol., 6, 3862–3872.
30. Eissenberg, J.C., James, T.C., Foster-Hartnett, D.M., Hartnett, T.,
Ngan, V. and Elgin, S.C. (1990) Mutation in a heterochromatin-specific
chromosomal protein is associated with suppression of position-effect
variegation in Drosophila melanogaster. Proc. Natl Acad. Sci. USA, 87,
31. Fanti, L., Giovinazzo, G., Berloco, M. and Pimpinelli, S. (1998) The
heterochromatin protein 1 prevents telomere fusions in Drosophila.
Mol. Cell, 2, 527–538.
32. Wang, G., Ma, A., Chow, C.M., Horsley, D., Brown, N.R., Cowell, I.G.
and Singh, P.B. (2000) Conservation of heterochromatin protein 1
function. Mol. Cell. Biol., 20, 6970–6983.
33. Song, K., Jung, Y., Jung, D. and Lee, I. (2001) Human Ku70 interacts with
heterochromatin protein 1α. J. Biol. Chem., 276, 8321–8327.
34. Rathmell, W.K. and Chu, G. (1994) Involvement of the Ku autoantigen in
the cellular response to DNA double-strand breaks. Proc. Natl Acad. Sci.
USA, 91, 7623–7627.
35. Mimori, T. and Hardin, J.A. (1986) Mechanism of interaction between
Ku protein and DNA. J. Biol. Chem., 261, 10375–10379.
36. Bianchi, A. and de Lange, T. (1999) Ku binds telomeric DNA in vitro.
J. Biol. Chem., 274, 21223–21227.
37. Song, K., Jung, D., Jung, Y., Lee, S.G. and Lee, I. (2000) Interaction of
human Ku70 with TRF2. FEBS Lett., 481, 81–85.
38. van Steensel, B. and de Lange, T. (1997) Control of telomere length by the
human telomeric protein TRF1. Nature, 385, 740–743.
39. Smith, S., Giriat, I., Schmitt, A. and de Lange, T. (1998) Tankyrase, a
poly(ADP-ribose) polymerase at human telomeres. Science, 282, 1484–1487.
40. Kim, S.H., Kaminker, P. and Campisi, J. (1999) TIN2, a new regulator of
telomere length in human cells. Nat. Genet., 23, 405–412.
41. Wu, G., Lee, W.H. and Chen, P.L. (2000) NBS1 and TRF1 colocalize at
promyelocytic leukemia bodies during late S/G2 phases in immortalized
telomerase-negative cells. Implication of NBS1 in alternative lengthening
of telomeres. J. Biol. Chem., 275, 30618–30622.
42. Nakamura, M., Zhou, X.Z., Kishi, S., Kosugi, I., Tsutsui, Y. and Lu, K.P.
(2001) A specific interaction between the telomeric protein Pin2/TRF1
and the mitotic spindle. Curr. Biol., 11, 1512–1516.
43. Baur, J.A., Zou, Y., Shay, J.W. and Wright, W.E. (2001) Telomere
position effect in human cells. Science, 292, 2075–2077.
44. Kohlhase, J. (1996) Isolierung und Charakterisierung einer humanen
Genfamilie mit Ähnlichkeit zu spalt, einem regionsspezifischen
homöotischen Gen von Drosophila melanogaster. Doctoral Thesis,
University of Göttingen, Göttingen.
45. Shi, Y., Seto, E., Chang, L.S. and Shenk, T. (1991) Transcriptional
repression by YY1, a human GLI-Kruppel-related protein, and relief of
repression by adenovirus E1A protein. Cell, 67, 377–388.
46. Schlüter, G., Boinska, D. and Nieman-Seyde, S.C. (2000) Evidence for
translational repression of the SOCS-1 major open reading frame by an
upstream open reading frame. Biochem. Biophys. Res. Commun., 268,