Báo cáo khoa học: The C-terminal region of CHD3/ZFH interacts with the CIDD region of the Ets transcription factor ERM and represses transcription of the human presenilin 1 gene pot

The C-terminal region of CHD3/ZFH interacts with the
CIDD region of the Ets transcription factor ERM and
represses transcription of the human presenilin 1 gene
Martine Pastorcic
1
and Hriday K. Das
1,2
1 Department of Pharmacology & Neuroscience, University of North Texas Health Science Center at Fort Worth, TX, USA
2 Department of Molecular Biology & Immunology, and Institute of Cancer Research, University of North Texas Health Science Center at
Fort Worth, TX, USA
Keywords
CHD3; ERM; presenilin-1; transcription;
yeast-two-hybrid
Correspondence
H.K. Das, Department of Pharmacology &
Neuroscience, and Institute of Cancer
Research, University of North Texas Health
Science Center at Fort Worth, 3500 Camp
Bowie Boulevard, Fort Worth, TX 76107,
USA
Fax: +1 817 735 2091
Tel: +1 817 735 5448
E-mail:
(Received 8 August 2006, revised 2 January
2007, accepted 9 January 2007)
doi:10.1111/j.1742-4658.2007.05684.x
Presenilins are required for the function of c-secretase: a multiprotein
complex implicated in the development of Alzheimer’s disease (AD). We
analyzed expression of the presenilin 1 (PS1) gene. We show that ERM
recognizes avian erythroblastosis virus E26 oncogene homolog (Ets) motifs
on the PS1 promoter located at )10, +90, +129 and +165, and activates

PS1 transcription with promoter fragments containing or not the )10 Ets
site. Using yeast two-hybrid selection we identified interactions between the
chromatin remodeling factor CHD3 ⁄ ZFH and the C-terminal 415 amino
acids of ERM used as bait. Clones contained the C-terminal region of
CHD3 starting from amino acid 1676. This C-terminal fragment (amino
acids 1676–2000) repressed transcription of the PS1 gene in transfection
assays and PS1 protein expression from the endogenous gene in SH-SY5Y
cells. In cells transfected with both CHD3 and ERM, activation of PS1
transcription by ERM was eliminated with increasing levels of CHD3. Pro-
gressive N-terminal deletions of CHD3 fragment (amino acids 1676–2000)
indicated that sequences crucial for repression of PS1 and interactions with
ERM in yeast two-hybrid assays are located between amino acids 1862 and
1877. This was correlated by the effect of progressive C-terminal deletions
of CHD3, which indicated that sequences required for repression of PS1 lie
between amino acids 1955 and 1877. Similarly, deletion to amino acid 1889
eliminated binding in yeast two-hybrid assays. Testing various shorter frag-
ments of ERM as bait indicated that the region essential for binding
CHD3 ⁄ ZFH is within the amino acid region 96–349, which contains the
central inhibitory DNA-binding domain (CIDD) of ERM. N-Terminal
deletions of ERM showed that residues between amino acids 200 and 343
are required for binding to CHD3 (1676–2000) and C-terminal deletions of
ERM indicated that amino acids 279–299 are also required. Furthermore,
data from chromatin immunoprecipitation (ChIP) indicate that
CHD3 ⁄ ZFH interacts with the PS1 promoter in vivo.
Abbreviations
3-AT, 3-amino-1,2,4-triazole; AD, Alzheimer’s disease; APP, amyloid precursor protein; CAT, chloramphenicol acetyl transferase; CHD3,
chromodomain helicase DNA-binding protein 3; ChIP, chromatin immunoprecipitation; CIDD, central inhibitory DNA binding domain; ER81
(ETV1), Ets-related protein 81 or Ets translocation variant 1; ERM, Ets-related molecule 5 or ETV5-Ets translocation variant 5; Ets, avian
erythroblastosis virus E26 oncogene homolog; HDAC, histone deacetylase complex; PEA3 (or E1AF, ETV4), polyoma enhancer A3;
PS1, presenilin 1; ZFH, zinc finger helicase.

1434 FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS
Presenilins (PS1 and PS2) are highly homologous
multipass transmembrane proteins [1,2]. PS1 mutations
have been linked to early-onset familial Alzheimer’s
disease (AD) [3,4]. Presenilins are required for the
function of c-secretase, a multiprotein complex that
has also been implicated in the development of AD
[5–8]. They may act as a catalyst or be involved in the
structure and metabolism of the complex itself.
c-Secretase has been implicated in the development of
AD because of its role in cleavage of the amyloid pre-
cursor protein (APP) and the production of Ab pep-
tide, which is central to the pathogenesis of AD [9].
Similarly, processing of the Notch receptor protein,
which controls signaling and cell–cell communication
has indicated a role for presenilin in development [10].
Presenilin and c-secretase also appear to cleave a vari-
ety of type 1 transmembrane proteins which all release
intracellular fragments with the ability to interact with
transcription coactivators [11,12]. They include CD44,
a ubiquitous cell-adhesion protein [13], and neuronal
cadherin (N-cadherin) [14]. Hence it appears that pres-
enilins may affect the expression of many genes
through intramembrane proteolysis [12]. Control of the
level of presenilins and its coordination with other
components of the c-secretase complex are likely to be
tightly regulated and we studied the transcriptional
control of the PS1 gene.
We identified DNA sequences required for expres-
sion of the human PS1 gene. A promoter region has

been mapped in SK-N-SH cells and includes sequences
from )118 to +178 flanking the major initiation site
(+1). However, we have shown that the promoter is
utilized in alternative modes in SK-N-SH cells and its
SH-SY5Y subclone [15]. The )10 Ets site controls
80% of transcription in SK-N-SH cells, whereas by
itself it plays only a minor role in SH-SY5Y cells.
Conversely, the Ets element at +90 controls 70% of
transcription in SH-SY5Y cells, whereas it affects tran-
scription by < 50% in SK-N-SH cells [15]. However,
in both cell types, mutations at both the )10 and +90
Ets sites substantially eliminate transcription activity,
indicating the crucial importance of these two Ets
motifs [15]. In addition to controlling the level of gene
expression, Ets factors may direct the choice of the
promoter elements in play, and therefore, determine
the selective combination of transcription factors
involved and the regulatory pathways modulating tran-
scription. We have identified several Ets factors that
specifically recognize the ) 10 Ets motif using yeast
one-hybrid selection including avian erythroblastosis
virus E26 oncogene homolog 2 (Ets2), Ets-like gene 1
(Elk1), Ets translocation variant 1 (ER81) and Ets-
related molecule (ERM) [15–17]. The ets genes encode
a family of transcription factors and most are tran-
scriptional activators [18,19]. They share a conserved
85-amino acid motif, the ETS domain, which recog-
nizes a nine-nucleotide DNA sequence with the central
consensus 5¢-GGAA ⁄ T-3 (Fig. 1A) [18,20]. Based on
the sequence homology of the ETS domain, and the

conservation of other functional domains, a phylogenic
tree of the ETS gene family has been derived [20] iden-
tifying 13 subgroups. ERM belongs to the PEA3 sub-
group, which also includes ER81 (or ETV1) and PEA3
(or ETV4) [20–23]. They share three conserved
domains. The most highly conserved ETS sequence,
the Ets domain, is required for specific DNA binding
to the consensus motif [18,24,25]. Two transactivating
domains, at the N-terminus and the C-terminus are
less conserved within the Ets family [26–28]. However
the N-terminus is highly conserved among ERM,
ER81 and PEA3, and has been shown to interact with
TAFII60 [27]. The C-terminal transactivation domain
functions in synergy with the N-terminal activation
domain but is not functionally equivalent [26,28]. The
central inhibitory domain (CIDD) of ERM shows sig-
nificantly less conservation with the other two PEA3
members [26]. Both the C-terminal and central domain
modulate DNA binding by the Ets domain and con-
tain an inhibitory function [18].
We chose to analyze the role of ERM because little
is known about its mode of action and particularly the
transcription factors with which it interacts. ERM
recognizes specifically Ets motifs located at )10 as well
as downstream at +90, +129 and +165 on the PS1
promoter and it activates PS1 transcription with pro-
moter fragments containing or not the Ets motif at
)10. In this report we have identified a new interaction
between ERM and CHD3 ⁄ ZFH using yeast two-
hybrid selection, and we show that this interaction

occurs between the C-terminal amino acid residues
1862–1877 of CHD3 and the CIDD domain of ERM.
Results
ERM interacts with CHD3
To identify proteins interacting with ERM we used the
C-terminal region (415 amino acid) of ERM, excluding
the first 95 amino acids, as a bait (Fig. 1A) to screen a
human brain cDNA library in pACT2 [17,26] using a
yeast two-hybrid selection assay. The excluded N-ter-
minus includes a transcription activation domain
highly conserved among ERM, ER81 and PEA3,
which has been shown to interact with TAFII60 [27].
The bait included the CIDD, the Ets domain and the
C-terminal domain (Fig. 1A) [18]. The Ets domain is
M. Pastorcic and H.K. Das Regulation of the presenilin 1 gene
FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS 1435
highly conserved among the Ets family and is required
for specific DNA binding to the consensus motif
[18,24,25]. The C-terminal domain includes a transacti-
vation domain functioning in synergy with the N-ter-
minal activation domain but it is not functionally
equivalent [26,28]. Both the C-terminal and central
domain modulate DNA binding by the Ets domain
and contain an inhibitory function. The ERM
A
B
C
Fig. 1. Interactions of ERM protein domains
with CHD3. (A) The domains conserved
within the PEA3 family are boxed, including

the N-terminal a-helical acidic domain con-
tained within the first 72 amino acids, the
Ets domain, the CIDD and the C-terminal
domain. The fragment of cDNA included in
the bait used for two-hybrid screening of
the brain cDNA library included amino acids
96–510 at the C-terminus. Shorter ERM
fragments were also tested as bait in yeast
two-hybrid assays and are indicated by
boxes below. Growth was scored at 6 days.
Black boxes indicate activity similar to the
larger bait fragment. White boxes indicate
fragments with no or little binding activity at
5m
M 3-AT. Striped and gray boxes show an
intermediate level of binding activity. Gray
box had 50% growth at 30 and 60 m
M 3-AT
compared with the larger construct. Striped
box showed no growth at 60 m
M 3-AT and
25% growth on 30 m
M 3-AT. (B) The
growth of yeast patches from AH109 trans-
formed with the CHD3 C-terminal fragment
including amino acids 1676–2000 together
with the various ERM bait fragments indica-
ted on the left was scored at 6 days.
Growth on medium excluding leucine, tryp-
tophan and histidine and including increasing

concentration of 3-AT from 0 to 60 m
M was
compared with growth on medium lacking
leucine and tryptophan (C). (C) Similarly the
growth of AH109 cotransformed with CHD3
and either N-terminal deletions (left) or
C-terminal deletions (right) of the ERM frag-
ment spanning amino acids 96–349 was tes-
ted in the presence of increasing amounts
of 3-AT (0–60 m
M; upper). The end point of
each deletion is indicated alongside.
Regulation of the presenilin 1 gene M. Pastorcic and H.K. Das
1436 FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS
fragment was cloned inframe with the GAL4
BD
(amino
acids 1–147 of GAL4 protein). Expression of the
GAL4–HIS3 reporter gene is leaky in AH109: a low
level of expression occurs in the absence of GAL4 acti-
vation. 3-Amino-1,2,4-triazole (3-AT; an inhibitor of
histidine biogenesis) is used to quench background
growth on His minus medium and the minimum level
of 3-AT required varies with the bait. Library screen-
ing identified two clones encoding the C-terminal por-
tion of the chromodomain helicase–DNA-binding
protein 3 (CHD3) or zinc finger helicase (ZFH), con-
ferring growth on 3-AT at concentrations as high as
60 mm by 6 days after plating. CHD3 is expressed in
several forms derived by alternative splicing [29–31]

(Figs 2 and 3) and ZFH is one such form of CHD3.
The two clones selected using yeast two-hybrid screen-
ing contained sequences downstream of amino acid
1676, and were identical to the CHD3 variants 1, 2
and 3 over their C-terminal protein sequence (Fig. 3).
Identification of the ERM region interacting with
CHD3 using yeast two-hybrid analysis
We constructed a series of shorter bait fragments to
identify the region required for interaction of ERM
with CHD3 (Fig. 1A,B). We tested CIDD + Ets,
Ets + C-terminus, Ets, CIDD, or C-terminus only.
3-AT (5 mm) was used to screen the library and was
sufficient to quench background growth on His-minus
medium for all the baits in the 6-day timeframe
observed. CIDD binding activity was identical to the
larger fragment including amino acids C-terminal to
residue 96. No growth was observed at 6 days on His-
medium containing as little as 5 mm 3-AT with baits
including C-terminal domain alone or a fragment
including Ets and C-terminal domains together. The
Ets domain by itself conferred growth at 5 and 30 mm
but showed no colonies at 60 mm. Hence, the Ets
domain is able to bind CHD3 independently but the
C-terminal domain appears to modulate this interac-
tion. A more precise location of ERM sequences
required to bind CHD3 (1676–2000) was derived from
a set of N- and C-terminal deletions of the ERM frag-
ment spanning amino acids 96–349 with CHD3
(Fig. 1C). N-Terminal deletions to residue 304 totally
eliminated growth at 60 mm 3-AT, but allowed resid-

ual growth at 30 mm at later time points (6–8 days).
Deletions to residue 343 eliminated all growth on
30 mm 3-AT. C-Terminal deletion to residue 279 elim-
inated growth at 30 mm 3-AT, indicating that amino
acids 299–279 are also required. Both series indicate
the importance of the interval 279–299. However,
sequences from 304 to 343 are also important. Hence
mutating both regions may be required to eliminate
binding to CHD3.
CHD3 represses the transcription of PS1
in SH-SY5Y neuronal cells
We tested the effects of pC1.CHD3 on expression of
the PS1 gene in SH-SY5Y cells. We compared the
action of the C-terminal fragment (amino acids 1676–
2000) with larger fragments including amino acids
295–1717, 1005–2000 and 1327–2000. The most signifi-
cant activity was observed with the C-terminal frag-
ment (1676–2000), which represses transcription of PS1
in transient infection assays by nearly 10-fold (Fig. 3).
These results suggest that in a particular cellular con-
text the interactions of full-length CHD3 with specific
proteins may result in conformation changes that
enable the same protein interactions that occur more
readily with the isolated C-terminal domain. Because
ERM acts as a transactivator of PS1, we asked whe-
ther CHD3 would alter the activation of transcription
of PS1 by ERM (Fig. 4). In cotransfections of
PS1CAT reporter with pC1.ERM, increasing amounts
of pC1.CHD3 appeared to eliminate the activation of
PS1 by ERM. We have previously shown that ERM

activates PS1 transcription through sequences
upstream as well as downstream of the major trans-
cription start site in SH-SY5Y cells [15]. Hence, we
compared the effect of CHD3 on the two promoter
fragments: )118,+178, which contains sequences flank-
ing the transcription start site, and +6,+178, which
contains only downstream sequences (Fig. 5A). Both
promoter sequences conferred a significant inhibition
by pC1.CHD3 (Fig. 5A), which is consistent with the
presence of ERM-binding sites upstream as well as
downstream of the transcription start site. We also
examined the effects of point mutations within several
of the Ets sites present within the PS1 promoter
(Fig. 5B). None of the single mutants at +20,+90 or
double mutants ()10,+90) ()10,+65) (+65,+129)
(+90,+129) appears to affect repression by CHD3,
suggesting a redundancy between Ets sites.
Delineation of the CHD3 domain(s) required
for the repression of PS1 transcription and
interaction with ERM by deletion mapping
Different N-terminal deletions of CHD3 were cloned
into pCMV-Tag2 vector to generate pCMV-Tag2ÆCHD3
expression constructs. These pCMV-Tag2ÆCHD3 con-
structs were transiently cotransfected into SH-SY5Y
cells. We examined the effects of N-terminal deletions of
CHD3 on the transcription of the ()118,+178) PS1CAT
M. Pastorcic and H.K. Das Regulation of the presenilin 1 gene
FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS 1437
reporter (Fig. 6A). Progressive N-terminal deletions
from amino acid 1676 towards amino acid 2000 indica-

ted that crucial amino acid residues for repression activ-
ity of CHD3 are present between 1810 and 1818 as well
as 1862 and 1877, because deletion of each region
reduced the repression activity by  50%. Finally, the
repression activity is completely eliminated with deletion
to 1877. Similar to the effects on transcription, the inter-
actions with ERM tested in yeast two-hybrid assays
(Fig. 6B) were unaffected by N-terminal deletions from
1676 to 1810 (end points at 1801 and 1810 not shown).
Unlike the effects on PS1 transcription (Fig. 6A), further
deletions with end points at amino acids 1810, 1818,
1830, 839, 1851 and 1861 did not affect interactions with
ERM (data not shown). Deletion from amino acids
1862–1871 (N-terminal end point at amino acid 1872)
drastically reduced 3-AT resistance: < 10% growth was
observed even at 5 mm 3-AT in deletions to 1872
(Fig. 6B). No further reduction was observed with dele-
tions to 1877, 1891, 1904, 1914 and 1924 (data not
shown). Only deletion reaching position 1943 totally
eliminated any growth down to the level observed in
Fig. 2. CHD3 proteins. The structure of
CHD3 and ZFH forms of CHD3 proteins
derived by alternative splicing are summar-
ized. Black shadows with white letters indi-
cate the sequences present in ZFH, but
absent in CHD3: a 34 amino acid insertion
at 1642, and a 34 amino acid terminal region
substituted by 12 heterologous amino acids
in CHD3 (boxed). In the functional assays
reported here we have considered the ZFH

form of the gene. The position of helicase
domains (I–VI) is underlined [29]. The histi-
dine and cysteine residues involved in puta-
tive zinc fingers are marked by stars. An
acidic region at amino acid 431 and the nuc-
lear localization signals at 691 and 954 are
shadowed in gray. The end-points of the
N-terminal deletions tested in binding or
transcription assays are marked by arrow-
heads and gray letters. The two clones
selected by yeast two-hybrid screening
contained the 325 amino acid C-terminal
fragment of CHD3: amino acids 1676–2000.
Regulation of the presenilin 1 gene M. Pastorcic and H.K. Das
1438 FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS
transformants containing the empty vector (Fig. 6B).
Hence the binding interface appears to be located in the
C-terminal of CHD3, after amino acid 1861 and crucial
sequences for CHD3–ERM interaction are located
between 1862 and 1877. However, in both in transcrip-
tion and in yeast two-hybrid assays, none of the single
amino acid mutations in this interval had any effect (data
not shown). Although they are not identical, the results
from the transfection assays in neuroblastoma cells and
from the yeast assays both indicate the presence of a
crucial ERM-binding domain of CHD3 located at the
C-terminal from amino acid residue 1862.
Further delineation was obtained with C-terminal
deletions (Fig. 7A). Deletion to 1877 eliminated repres-
sion activity (Fig. 7A), which is consistent with the

N-terminal deletions. Deletions to 1902 had only a
minor effect (Fig. 7A). Hence essential sequences
appear to be located between residues 1902 and 1877.
Similarly, binding measured by yeast two-hybrid assay
delineated important sequences between 1955 and 1902
(Fig. 7B).
Fig. 3. Inhibition of PS1 transcription by CHD3. SH-SY5Y cells were transiently transfected using the calcium phosphate precipitation method
with 6 lgof()118, +178) PS1CAT reporter together with 3 lg of either pC1 vector or pC1.CHD3. Various fragments of CHD3 protein
expressed from pC1.CHD3 include amino acids 295–1717, 1005–2000, 1327–2000 and 1676–2000. The relationship of these CHD3 protein
fragments to the different variants reported for CHD3 is summarized on top. Base pair positions are according to gi#2645432 for CHD3 and
gi#3298561 for ZFH. Promoter activity in the presence of each construct is indicated laterally, with the activity in the presence of the empty
pC1 vector arbitrarily set as 100%. CAT activity in different samples was standardized using the amount of protein present in the cellular
extracts as an internal control. Each experiment was repeated three times, with at least triplicate tests of each construct combination.
Fig. 4. CHD3 antagonizes PS1 activation by ERM. SH-SY5Y cells
were transiently transfected using the calcium phosphate precipita-
tion method with 6 lgof()118, +178) PS1CAT reporter together
with 3 lg of pC1 vector or pC1.ERM in the presence of 0, 1 or
3 lg pC1.CHD3. pC1.ERM expresses the full-length ERM protein.
pC1.CHD3 contains the CHD3 fragment expressing amino acids
1676–2000. Promoter activity in different samples was standardized
using the amount of protein present in the cellular extracts as an
internal control. Each experiment was repeated three times, with
at the minimum triplicate tests of each construct combination.
M. Pastorcic and H.K. Das Regulation of the presenilin 1 gene
FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS 1439
Interaction of CHD3 with the PS1 promoter
in vivo
We sought to document more directly the role of
CHD3 in the regulation of PS1 gene in vivo. We first
A

B
Fig. 6. Effect of N-terminal deletions of CHD3 on the inhibition of
transcription by CHD3 and on the binding of CHD3 with ERM. (A)
Effect of N-terminal deletions on the inhibition of transcription by
CHD3. SH-SY5Y cells were transiently transfected using the cal-
cium phosphate precipitation method with 6 lg of the ()118, +178)
PS1CAT reporter together with 3 lg of either pCMV-Tag2 or
pCMV-Tag2ÆCHD3. pCMV-Tag2ÆCHD3 expresses various N-terminal
fragments of CHD3 with the indicated N-terminal end-point (from
position 1676–1943). The C-terminal end point of the above N-ter-
minal deletions of CHD3 expressed by pCMV-Tag2. CHD3 is at
amino acid 2000. Promoter activity in different samples was stan-
dardized using the amount of protein present in the cellular extracts
as an internal control. Each experiment was repeated three times,
with a minimum of n ¼ 4 for each data point. Values that differ sig-
nificantly from the level of inhibition observed in cotransfections
with the C-terminal amino acids 1676–2000 of CHD3 (1676) with
P < 0.05 by t-test ⁄
ANOVA are indicated (*). (B) Identification of
CHD3 domains interacting with ERM by yeast two-hybrid assay.
The same N-terminal deletions of CHD3 mentioned in (A) were
introduced into pACT2 to generate various pACT2.CHD3 construct
which were introduced into AH109 pretransformed with the
Gal4
BD
–ERM fusion bait (amino acids 96–510). The ability of the
mutants to promote growth on SD medium lacking tryptophan, leu-
cine or histidine and including 15 m
M of 3-AT was scored after
4 days. Patch L represents transformants with pACT2.CHD3 which

expresses CHD3 fragment containing amino acids 1676–2000.
Fig. 5. Inhibition of PS1 transcription by CHD3. (A) SH-SY5Y cells
were transiently transfected using the calcium phosphate precipita-
tion method with 6 lgof()118, +178) PS1CAT (circles) or (+ 6,
+178) PS1CAT (triangles) reporter together with either pC1 vector
or increasing amounts of pC1.CHD3. pC1.CHD3 contains the CHD3
fragment expressing amino acids between 1676 and 2000. The
total amount of (pC1.CHD3 + pC1) was kept constant at 2 lg. Pro-
moter activity in different samples was standardized using the
amount of protein present in the cellular extracts as an internal con-
trol. (B) Effect of PS1 promoter mutations at Ets elements on the
inhibition of transcription by CHD3. SH-SY5Y cells were transfected
with 5 lgof()118, +178) PS1CAT wild-type (wt) or containing
point mutations at the sites indicated in the presence of 4 lgof
pC1.CHD3 (1676–200) or the empty pC1 expression vector. Promo-
ter activity in different samples was standardized with the amount
of protein present in cell extracts. Each experiment was repeated
three times, with a minimum of three tests for each construct
combination.
Regulation of the presenilin 1 gene M. Pastorcic and H.K. Das
1440 FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS
examined the interactions of the endogenous CHD3
and ERM produced in SH-SY5Y cells with the cellular
chromatin (Fig. 8A). We tested whether interactions of
CHD3 or ERM and the PS1 promoter area around the
main transcription start site (+1) could be detected
using chromatin immunoprecipitation assays (ChIPs).
Cross-linked DNA–protein complexes were immuno-
precipitated with antibodies to CHD3 or ERM, and the
DNA was analyzed for the presence of PS1 promoter

sequences. Although we detected interactions with
ERM (lane 2) we did not detect any interaction of the
PS1 promoter with endogenous CHD3 (lane 1). An
alternative ChIP assay was carried out using fusion
proteins including the Flag epitope and CHD3 inserted
downstream in the pCMV-Tag2 vector transfected with
high frequency into SH-SY5Y cells using lipofectamine.
Cross-linked DNA–protein complexes were immuno-
precipitated with anti-Flag2 serum. The DNA in the
complexes was then analyzed by PCR for the presence
of the PS1 promoter (Fig. 8B). Promoter sequences
from a gene unrelated to PS1 and not containing Ets
elements (IRL) do not appear to be enriched in cells
transfected with CHD3 (lanes 1 and 2) as compared to
cells transfected with pCMV-Tag2 vector alone (lane
3). However the PS1 promoter sequences flanking the
PS1 transcription initiation site are more enriched in
cells transfected with the C-terminal fragment of CHD3
spanning residues 1676–2000 (lane 2) than CHD3 frag-
ment spanning residues 1676–1877 (lane 1), a poor
repressor of the PS1 gene (Fig. 6A). Hence it appears
that CHD3 (1676–2000) interacts somehow with the
area around the PS1 promoter in vivo.
CHD3 inhibits PS1 protein levels in SH-SY5Y cells
Western blot analysis of total protein from SH-SY5Y
cells transiently transfected with pCMV-Tag2 or
pCMV-Tag2ÆCHD3 showed that the CHD3 gene frag-
ment encoding amino acids 1676–2000 decreased the
amount of C-teminal fragment of PS1 protein
( 20 kDa PS1CTF) by  60%, whereas transfection

of the CHD3 gene fragment encoding amino acids
1741–2000 decreased PS1 protein level by  75%.
Hence increasing the level of CHD3 in neuroblastoma
cells reduced the amount of PS1 protein produced by
the endogenous gene. These results suggest that CHD3
may indeed reduce the level of PS1 in vivo and may
affect the level of PS1 ⁄ c-secretase activity.
Discussion
We have identified interaction(s) of the CIDD domain
of ERM with the C-terminal domain of CHD3. The
A
B
Fig. 7. Effect of C-terminal deletions of CHD3 on the inhibition of
transcription by CHD3 and its binding to ERM. (A) Effects of C-ter-
minal deletions on the inhibition transcription by CHD3. SH-SY5Y
cells were transiently transfected by calcium phosphate precipita-
tion method with 6 lg of the () 118, +178) PS1CAT reporter
together with 3 lg of either pCMV-Tag2 vector or pCMV-
Tag2ÆCHD3. pCMV-Tag2ÆCHD3 expresses various C-terminal frag-
ments of CHD3 with the indicated C-terminal end-point (from posi-
tion 2000–1790). The N-terminal end point of the above C-terminal
deletions of CHD3 expressed by pCMV-Tag2ÆCHD3 is at amino acid
1676. Promoter activity in different samples was standardized using
the amount of protein present in the cellular extracts as an internal
control. Each experiment was repeated three times, with a mini-
mum of n ¼ 4 for each data point. Values that differ significantly
from the level of inhibition observed in cotransfectoins with the
C-terminal amino acids 1676–2000 of CHD3 (2000) with P < 0.05
by t-test ⁄
ANOVA are indicated (*). (B) Effect of C-terminal deletions

of CHD3 on its interaction with ERM in yeast two-hybrid assays. A
subset of the same C-terminal deletions of CHD3 were introduced
into pACT2 and transformed into AH109 pretransformed with the
Gal4
BD
–ERM fusion bait (amino acids 96–510). The deletion end
points to 1955, 1902 and 1889 are indicated on the right. The ability
of the mutants to promote growth on SD medium lacking trypto-
phan, leucine or histidine and including 0, 5, 30 or 60 m
M of 3-AT
was scored after 4 days.
M. Pastorcic and H.K. Das Regulation of the presenilin 1 gene
FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS 1441
C-terminal amino acid residues between 1862 and 1877
(Fig. 6B) of CHD3 appears to be crucial for inter-
action with the CIDD domain of ERM. We have also
detected interactions of CHD3 with the Ets domain of
ERM. Both the interactions with CIDD and Ets
appear to be modulated by adjacent domains. Binding
to CIDD is stronger than CIDD + Ets, and binding
to Ets alone is stronger than Ets + C-terminal. The
effect of C-terminal deletions of ERM-delineated
sequences required for interaction with CHD3 between
residues 279 and 299, and this is consistent with the
effects of N-terminal deletions of ERM. However
N-terminal deletions of ERM indicate that the residues
304–343 region may also be important in the interac-
tions with CHD3.
A
B

Fig. 8. CHD3 interaction with the PS1 promoter. (A) Interaction of
the endogenous CHD3 produced in SH-SY5Y cells with the cellular
chromatin. Exponentially growing SH-SY5Y were cross-linked with
1% formaldehyde, lyzed and chromatin was sheared. The nuclear
protein–DNA complexes were immunoprecipitated by incubation
with antibodies to CHD3 (aCHD3: sc-11378X; Santa Cruz Biotech-
nology) and ERM (aERM1: sc-1955X or aERM2: sc22807X). Control
serum was added in (C). The DNA precipitated in the complexes
was analyzed by PCR to detect PS1 promoter sequences from )25
to +66 as well as +45 to +100. No DNA was added to the PCR
reaction in lane 5. Unrelated control DNA sequences (IRL:
monoamine oxydase B gene) were tested as internal standard. (B)
Interaction of the C-terminal fragment of CHD3 (amino acids 1676–
2000) with the promoter of the endogenous PS1 gene. SH-SY5Y
cells were transiently transfected by lipofectamine with 8 lgof
either pCMV-Tag2 vector [C] or pCMV-Tag2–CHD3 expressing the
Flag Tag2–CHD3 fusion protein including residues 1676–1877
(1877) or 1676–2000 (1676). Forty hours after transfection cells
were cross-linked with 1% formaldehyde, lyzed and chromatin was
sheared. Nuclear protein–DNA complexes were immunoprecipitat-
ed by incubation with anti-Flag M2 agarose beads. DNA in the
cross-linked complexes was analyzed by PCR to detect PS1 promo-
ter sequences (91 bp) from )25 to +66 and (55 bp) from +45 to
+100 (Table 1). Unrelated control DNA sequences (IRL) without
Ets-binding site were tested as internal standard.
Fig. 9. Expression of CHD3 inhibits PS1 protein expression.
SH-SY5Y cells were transiently transfected by lipofectamine with
8 lg of either pCMV-Tag2 vector [C] or pCMV-Tag2-CHD3 expres-
sing the Flag Tag2-CHD3 fusion protein. A, a typical Western blot
shows the expression of PS1, FLAG-tagged-CHD3, and GAPDH

protein levels. (C) 1676, and 1741 represent western blot analysis
with protein extracts from cells transfected with pCMV-Tag2,
pCMV-Tag2–CHD3 (amino acids 1676–2000), and pCMV-Tag2–
CHD3 (amino acids 1741–2000), respectively. Arrows indicate the
position of the FLAG-CHD3 fusion protein (amino acids 1676–2000)
on the left, and FLAG-CHD3 (amino acids 1741–2000) on the right.
A protein band unrelated to CHD3 appears in all the samples. The
nature of this nonspecific protein band is unknown. Blots were
developed by chemiluminescence and protein gel bands were
quantified using
SCION IMAGE software (n ¼ 4). (B) Bar graphs show
relative expression of PS1 protein ( 20 kDa PS1CTF) normalized
to the expression of GAPDH. Data was analyzed by paired
t-test ⁄
ANOVA and (*) indicates that protein level in1676 and 1741
samples were different from the control (C) with P < 0.05.
Regulation of the presenilin 1 gene M. Pastorcic and H.K. Das
1442 FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS
The tertiary structure of ERM is not entirely known
[32]. The DNA-binding Ets domain adopts a winged
helix–turn–helix form [24,25]. DNA binding is inhib-
ited by the CIDD at amino acids 203–290 and the C-
terminus at positions 468–510, which appear to act in
synergy [26–28,33,34]. A shorter domain at 291–355
also plays an inhibitory role. The mode of inhibition
of DNA binding is unknown. Furthermore, both
CIDD and the C-terminus contain transactivation
activity and are able to synergize, but are not func-
tionally equivalent [22,27,28,33,34]. The CIDD and
C-terminus appear mostly devoid of identifiable ter-

tiary structure except for a short a helix in CIDD at
position 216 [32]. Post-translational modifications
have been identified in the CIDD that may poten-
tially affect the structure and function of ERM
through its ability to inhibit DNA binding or to
interact with other proteins. Phosphorylation sites at
amino acids 223 and 248 are conserved with ER81
where they appear to affect transactivation [35,36].
Sumoylation sites are present in ERM at residues 89,
263, 293 and 350 [37]. Modification at these sites does
not appear to affect DNA binding but does decrease
transactivation by ERM [37]. Transactivation of the
human Ets-responsive ICAM-1 promoter by ERM
was increased by mutants eliminating sumoylation at
all three sites 263–293–350 together in COS-7 cells
[37]. It is interesting to note that although the activa-
tion of the synthetic reporter containing three E74
Ets-binding sites inserted upstream from the thymi-
dine kinase promoter by ERM requires modification
such as phosphorylation at the serine residue 367 [38]
that simultaneously decreases DNA binding, it can be
activated by the triple mutant eliminating sumoylation
at sites 263–293–350 together [37]. The mechanism of
action of sumoylation of ERM is unknown and could
potentially involve the regulation of interactions with
cofactors.
We have identified an interaction between CHD3 and
the CIDD region of ERM. CHD3 (ZFH) is a member
of the chromodomain family of proteins and includes
chromo (chromatin organization modifier) domains and

helicase ⁄ ATPase domains (Fig. 2). It is a component of
a histone deacetylase complex which participates in the
remodeling of chromatin by deacetylating histones.
Chromatin remodeling and the unwinding activity of
helicases are required for many aspects of DNA meta-
bolism: replication, recombination, chromatin pack-
aging and transcription. Chromatin-remodeling factors
have been implicated in the repression of transcription
[29,30]. CHD3 is widely expressed as three different iso-
forms (variants 1, 2 and 3) derived by alternative spli-
cing [29,30]. It is interesting to note that CHD3 has been
found to interact with SUMO-1 [39], indeed a potential
SUMO-1 motif ‘VKKE’ is located at position 1970
within the C-terminal region required for repression of
PS1 (Fig. 6A). However, C-terminal deletion mapping
indicates that it is not required for repression in our sys-
tem (Fig. 7A). Interactions of CHD3 with SUMO-1
were detected using yeast two-hybrid screening for pro-
teins interacting with the p73 protein, a p53-related fac-
tor often mutated in neuroblastoma [40]. However,
yeast two-hybrid assays also showed interactions
between SUMO-1 and p53 and p73 [39]. Hence it is
possible that CHD3 is implicated in complexes with p73
and p53 through interaction with SUMO-1 (and in yeast
the SUMO-1 equivalent Smt3p) and that such interac-
tions participate in the repression of transcription by
p53 and p73. It is possible that we are observing a
similar case and that our yeast two-hybrid selection
implicated a SUMO-1 yeast protein intermediate.
An increasing number of reports have implicated

CHD3 in the repression of transcription by its participa-
tion in histone deacetylase complexes (HDAC) [41–43],
which have been implicated in the repression of gene
activity [41,44]. In particular, repression by p53 protein
appears to involve HDAC complexes in which p53 inter-
acts with HDAC indirectly via the corepressor mSin3a.
It may be worth noting that the transcription of PS1 is
repressed by p53 and the cofactor p300 [45]. It is poss-
ible that CHD3 participates in the repression of PS1
in vivo via mechanisms including some of the aspects
outlined above. Such mechanisms may function to ant-
agonize the induction of apoptosis by p53.
Several transcriptional repressors have implicated
CHD3 repression [42,43,46–50]. Interestingly, several
of these corepressors also interact with the C-terminal
regions of CHD3 downstream of residue 1676 [46] or
the conserved regions of CHD4 and dMi2 [50]. Models
proposed to explain the repression of transcription
involve a transcriptional repressor functioning as a
hinge between the HDAC complex and the factors
binding to the gene promoter DNA [50]. In the case
described here we may be observing a direct inter-
action between a component of HDAC and the tran-
scription factor ERM. However, we were not able to
show this interaction in vitro using a pull-down assay,
although CHD3 was found to be associated with the
PS1 promoter region using chromatin immunoprecipi-
tation. It may be due to lack of stability of the com-
plex because ChIP assays involve cross-linked proteins.
We cannot rule out that the interaction was indirect

and implicated a protein such as the yeast homolog of
SUMO-1.
Two homologous proteins, Ki-1 ⁄ 57 and CGI-55,
have recently been shown to interact with the
M. Pastorcic and H.K. Das Regulation of the presenilin 1 gene
FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS 1443
C-terminal region of CHD3 [51]. Ki-1 ⁄ 57 was identi-
fied as the phosphoprotein antigen recognized by the
first antibody to specifically detect malignant Hodgkin
and Sternberg–Reed cells in Hodgkin lymphoma. CGI-
55 is a homologous protein. Sequencing of the CHD3
clones selected by yeast two-hybrid indicates that all
the CHD3 clones selected contained sequences C-ter-
minal from residue 1551. The shortest clone indicated
that the region of interaction C-terminal from amino
acid 1839 still contains binding activity [51]. Hence
there is growing evidence indicating that the C-ter-
minal region of CHD3 plays a crucial physiological
role.
Experimental procedures
Yeast two-hybrid assays
The two-hybrid system first described by Fields and Song
[52] enables the capture of cellular proteins interacting with
a specific protein used as a bait. We have used the Match-
maker system from Clontech, a version of the system based
on the GAL4 transcription factor. A hybrid of GAL4
DNA-binding domain and the protein of interest is used as
the bait. A library of cellular proteins expressed as fusions
with the GAL4 transcription activation domain is screened
for those that interact with the bait, thereby bridging

GAL4 BD and AD and activating transcription of a repor-
ter gene containing a GAL upstream activating element.
ERM-containing baits were generated by PCR amplifica-
tion with Pfu polymerase (Stratagene, La Jolla, CA) and
inserted between the EcoRI and BamHI sites of the plasmid
pGBKT7 (Clontech, Palo Alto, CA) in order to generate in
frame fusions with the Gal4
BD
(amino acids 1–147). The
bait used originally to screen the brain cDNA library in
ACT2 (Cat #HL4004AH; Clontech) contained the 984 bp
3¢-terminal portion of ERM or C-terminal residues 96–510
[26]. Baits containing subfragments of ERM were generated
similarly with the following oligonucleotides listed in
Table 1: residues 96–510, primers 96FW and 510R where
EcoRI and BamHI sites were added to ERM sequences to
direct insertion; residues 96–454, primers 96FW and 454R;
residues 342–510, 342FW and 510R; residues 342–454,
342FW and 454R; residues 447–510, 447FW and 510R; res-
idues 96–349, 96FW and 349R. In C-terminal deletions
(Fig. 1C) the common sense primer was 96FW. In N-ter-
minal deletions the common antisense primer was 349Rb.
PCR conditions were 2 min at 94 °C, 30 cycles with (30 s
at 94 °C, 30 s at annealing temperature, 90 s at 72 °C), and
10 min at 72 °C. Annealing temperature was usually 5 °C
below the T
m
.
AH109 competent cells from Clontech were transformed
with each bait. In the original screening of the library a

total of 1.5 · 10
6
independent transformants were screened
and two independent CHD3 clones were obtained. Sequen-
cing revealed that they were identical and contained CHD3
sequences from amino acids 1676–2000.
To examine the activity of subfragments from the ori-
ginal bait, AH109 cells were transfected with each bait
together with pACT2.CHD3 (amino acids 1676–2000).
Analysis of the activity of CHD3 domains
Various CHD3 subclones spanning most of the protein
sequence were derived by PCR amplification using Pfu
(Stratagene) from a set of clones generously provided by
F. Aubry [29]. CHD3 fragment from amino acids 295–1717
was derived from clone 37-1 [29] with primers 37-1S and 37-
1A including the NheI and BstZI sites, respectively (Table 1)
and inserted into the corresponding sites of pC1 (Promega,
Madison, WI). Primers also included an extra 3 amino acid
5¢-sequence providing and ATG translation start site in the
context of a sequence obeying the Kozak consensus. CHD3
fragment from amino acids 1005–2000 was derived from
clone 37-9 [29] with the primers 37-9S and 37-9A. CHD3
fragment from amino acids 1327–2000 was derived from
clone 37-8 with the primers 37-8S and 37-9A. CHD3 frag-
ment from amino acids 1676–2000 was derived from the
clone obtained in our yeast two-hybrid selection with the
sense primer 1676S and the reverse primer 37-9A.
Constructs into pCMV-Tag2 (Stratagene) were generated
in order to analyze the CHD3 domains required for interac-
tions with the PS1 promoter. The maximum construct with

amino acids 1676–2000 was derived by PCR amplification
with Pfu polymerase and the primers 1676ST (sense) and
2000AT (antisense) including the EcoRI and SalI sites,
respectively, and inserted between the EcoRI and XhoI sites
of pCMV-Tag 2 form C. A series of N-terminal deletions
were derived using sense primers including an EcoRI site:
1741S, 1800S, 1810S, 1818S, 1830S, 1839S, 1851S, 1861S,
1872S, 1877S, 1891S, 1904S, 1914S, 1924S and 1943S. The
common antisense primer was 2000AX and included an
XhoI site.
C-Terminal deletions were also constructed with the fol-
lowing primers including a SalI site: 1955A, 1902A, 1889A,
1877A, 1870A, 1862A, 1851A, 1843A, 1832A, 1821A,
1810A, 1801A and 1790A. The common sense primer was
1676ST and included and EcoRI site. The same series of
inserts was inserted into the EcoRI and XhoI sites of
pACT2 in order to also test the CHD3 domains in yeast
two-hybrid binding assays.
Transfection assays in SH-SY5Y neuroblastoma
cells
Transfection by calcium phosphate precipitation and gly-
cerol shock has been described previously [16]. Cells were
seeded at a density of 10
4
Æcm
)2
, 1 day before transfection.
Regulation of the presenilin 1 gene M. Pastorcic and H.K. Das
1444 FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS
Table 1. Primers.

ERM
96FW: GATCGAATTCTCCTCTGAGCTGTCGTCTTGTA
342FW: GATCGAATTCCCTGAGAGACTGGAAGGCAAAG
349R: GATCGGATCCACTTTGCCTTCCAGTCTCTCAG
447FW: GATCGAATTCTTTGTCTGTGACCCAGATGC
454R: GATCGGATCCTTAGAGGGCATCTGGGTCACAG
510R: GATCGGATCCACTCCGCCACTCAGAAACTTAG
ERM N-terminal deletions
140S: GATCGAATTCTCACCCACCCATCAGAATC
200S: GATCGAATTCCCCCTGCAGATGCCAAAGATG
304S: GATCGAATTCGAAGTGCCTAACTGC
343S: GATCGAATTCGAGAGACTGGAAGGCAAAG
349Rb: GATCGGATCCTCATTTGCCTTCCAGTCTCTCAG
ERM C-terminal deletions
320A: GATCGGATCCTCAGCTGGAGAAATAAC
299A: GATCGGATCCTCAGTAATCCCGAGGCTCCTG
279A: GATCGGATCCTCATGGCATGCCCGGGAC
256A: GATCGGATCCTCAGGGAGCTGCAGGGAC
239A: GATCGGATCCTCAATTATCTCCAGGAAC
219A: GATCGGATCCTCACTGAAATCTCTGTTC
199A: GATCGGATCCTCACTGATGTGGTGGTC
CHD3 subclones
37-1S: GATCGCTAGCGTGTGATGAAGGCGCTAGGGCTTCTGGGT
37-1A: GATCCGGCCGTCATGTGAAGCCACCAT
37-9S: GATCGCTAGCGTGTGATGAAGGCGGAGGCCTTGAATTCACG
37-9A GATCCGGCCGATCCAGTCAGTCGTCTATAC
37-8S: GATCGCTAGCGTGTGATGAAGGCGGAGCAACAGCAGGAAGAC
1676S: GATCGCTAGCGTGTGATGAAGGCGGAAGATGTAAAAGGTG
CHD3 inserts in pCMV-Tag2
1676ST: GATCGAATTCTGGAAGATGTAAAAGGTG

2000AT: GATCGTCGACATCCAGTCAGTCGTCTATAC
2000AX: GATCCTCGAGATCCAGTCAGTCGTCTATAC
N-terminal deletions
1741S: GATCGAATTCACAGAAGACATGA,
1800S: GATCGAATTCTGGAGCAGGCGCTG
1810S: GATCGAATTCTGCGGCGGGCGGCCTACTCG,
1818S: GATCGAATTCTGTCGCAGGAGCCGGCGCAG,
1830S: GATCGAATTCACGCCCGCTTCGCCGAG,
1839S: GATCGAATTCTGGCCGAGAGCCACCAGCAC,
1851S: GATCGAATTCTGGCGGGGAACAAGCCG,
1861S: GATCGAATTCTGCACAAGGTTCTGAACCA,
1872S: GATCGAATTCTGAGCGACATGAAGGCGGAC,
1877S: GATCGAATTCTAGCGGACGTGACCCGCCTG,
1891S: GATCGAATTCCCATCGCAGCCCGCCTTCAGATG,
1904S: GATCGAATTCTCAGCCGGCTGGCCAGCA,
1914S: GATCGAATTCCTCACCCCACACCGGCCTAC,
1924S: GATCGAATTCCCTACGCTACACCTCCG,
1943S: GATCGAATTCTGGCCGCCGCAGGC,
C-terminal deletions
1955A: GATCGTCGACTCATGCAGGCATCTGGCTGTAATTG
1902A: GATCGTCGACTCAGCTGCGCTCGGACATCTGAAGGC
1889A: GATCGTCGACTCATATTCGGGACAGCGTGGCTG
1877A: GATCGTCGACTCACGCCTTCATGTCGCTCAGCAAC
1870A:
GATCGTCGACTCA CTC CTC CAG CTG GTT CAG AAC CTT G
1862A: GATCGTCGACTCA GTG CAG GAC GGC GTT GGC
1851A: GATCGTCGACTCA CAG CGA CTC CTT GGA GAG G
1843A: GATCGTCGACTCA GTG GCT CTC GGC CAG GCA C
1832A: GATCGTCGACTCA GCG GGC GTG GAG GGC CAT GG
M. Pastorcic and H.K. Das Regulation of the presenilin 1 gene

FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS 1445
Promoter activity was determined by chloramphenicol ace-
tyl transferase (CAT) assay in different samples and was
standardized using the amount of protein present in the cel-
lular extracts as an internal control as described previously
[16]. Each experiment was repeated three times, with at
least triplicate tests of each construct and treatment. The
ERM expression construct pC1.ERM used in Fig. 4 has
been described previously [15] and included the entire
cDNA (full-length ERM protein). Sequences from position
132–2066 according to gi: 33873571 were inserted between
the MluI and SalI of pC1. Lipofectamine 2000 transfections
(Invitrogen, Carlsbad, CA) were according to the manufac-
turer’s instructions with the following modifications. On
transfection day, medium in 6 cm plates was replaced by
2.5 mL serum-free modified Eagle’s medium. For each
plate: 0.25 mL of serum free medium containing 10 lLof
lipofectamine were combined with 0.25 mL medium con-
taining 8 lg DNA, after 20 min the 0.5 mL mixture was
added to the plates and cells were returned to the incubator
for 5 h. The lipofectamine solution was then removed and
replaced by culture medium including 10% fetal bovine
serum (HyClone, Logan, UT). Cells were refed after 24 h
and harvested at 40 h after transfection.
Chromatin immunoprecipitation
SH-SY5Y cells were transiently transfected with 8 lgof
pCMV-Tag2–CHD3 expressing the Flag tag–CHD3 fusion
protein (amino acids 1676–1877 or 1676–2000) or the empty
control pCMV-Tag2 vector by lipofectamine. Forty hours
after transfection cells were cross-linked with 1% formalde-

hyde in serum-free modified Eagle’s medium for 10 min at
room temperature. Cells were then washed with NaCl ⁄ Pi
and treated with 0.125 m glycine for 5 min to stop the reac-
tion. Cells were washed again and harvested in NaCl ⁄ P
i
containing protease inhibitors at 4 °C. Cells were lyzed and
enzymatic chromatin shearing was performed as directed by
the kit from Active Motif (Carlsbad CA). The nuclear pro-
tein–DNA complexes were immunoprecipitated by incuba-
tion with anti-Flag M2 agarose beads from Sigma
(St. Louis, MO) overnight at 4 °C with rotation in IP buf-
fer (20 mm Tris pH 8.0, 150 mm NaCl, 1 mm EDTA, 1%
Triton X-100, 10% glycerol, and protease inhibitors). The
resin was collected and washed successively with IP buffer,
high salt buffer (20 mm Tris pH 8.0, 1% Triton X-100,
2mm EDTA, 500 mm NaCl) and LiCl buffer (20 mm Tris
pH 8.0, 0.1% NP40, 2 mm EDTA, and 250 mm LiCl), TE
buffer (10 mm Tris pH 8.0 and 1 mm EDTA) and eluted
with 1% SDS, 0.1 m NaHCO
3
. Cross-links were reversed by
incubation overnight at 65 °C in the presence of 200 mm
NaCl. Samples were treated with RNase and DNA was
purified with minielute spin columns (Qiagen, Valencia,
CA). Samples were analyzed by PCR with the primers
PS1()25 to +66) and PS1()45 to +100) as well as IRL (the
amplified fragment corresponds to position 35 821–35 980
of the sequence gi:2440066 for the monoamine oxydase B
gene) primers as a control (Table 1). DNA was fractionated
by electrophoresis on 12% polyacrylamide gels.

For ChIP assay of the endogenous CHD3 expressed in
SH-SY5Y cells, complexes were precipitated with antibodies
to CHD3 (sc-11378X, Santa Cruz Biotechnology, Santa
Cruz, CA) and ERM (aERM1: sc-1955X or aERM2:
sc22807X). Antibody complexes were captured with protein
A ⁄ G (Santa Cruz Biotechnology).
Western blot analysis
SH-SY5Y cells were transiently transfected with 8 lgof
pCMV-Tag2 or pCMV-Tag2ÆCHD3 using lipofectamine.
Total cellular proteins (50 lg) were fractionated by electro-
phoresis on denaturing 12% polyacrylamide gels and ana-
lyzed by western blotting using successively anti-PS1
(MAB5232; Chemicon, Temecula, CA), anti-Flag (F1804;
Sigma), and anti-GAPDH (MAB374; Chemicon) as des-
cribed previously [16].
Acknowledgements
This research was supported by a grant from the
National Institutes of Health (AG 18452). We wish to
thank Dr Florence Aubry (CNRS, France) for
CHD3 ⁄ ZFH clones.
References
1 Dewji NN, Valdez D & Singer SJ (2004) The preseni-
lins turned inside out: implications for their structures
and functions. Proc Natl Acad Sci USA 101, 1057–
1062.
Table 1. Continued
1821A: GATCGTCGACTCA CTC CTG CGA CAG GTT CAG GTA G
1810A: GATCGTCGACTCA CAGCTGCTCCTCAAT C
1801A: GATCGTCGACTCA CTCCAGGAGCTTGAACCTCCG
1790A: GATCGTCGACTCA ATTTTTCATCTCCAGAAAGTTC

ChiP
PS1()25 to +66): CGACGCCAGAGCCGGAAATGAC (sense), TTCCGATGTGAAACCGCGGACC (antisense)
PS1(+ 45 to +100): GGTCCGCGGTTTCACATCG (sense), GCTCAGGTTCCTTCCAGAC (antisense)
IRL: TTTGCTGTCTCAGGCCCTTTATA (sense), ATGAATGGAGAGGATCTGCTACG (antisense)
Regulation of the presenilin 1 gene M. Pastorcic and H.K. Das
1446 FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS
2 Li X & Greenwald I (1998) Additional evidence for an
eight-transmembrane-domain topology for Caenorhabdi-
tis elegans and human presenilins. Proc Natl Acad Sci
USA 95, 7109–7114.
3 Sherrington R, Rogaev EI, Liang Y, Rogaeva EA, Lev-
esque G, Ikeda M, Chi H, Lin C, Li G, Holman K
et al. (1995) Cloning of a gene bearing missense muta-
tions in early-onset familial Alzheimer’s disease. Nature
375, 754–760.
4 Tanzi RE, Kowacs DM, Kim TW, Moir RD, Guenette
SY & Wasco W (1996) The gene defects responsible for
familial Alzheimer’s disease. Neurobiol Dis 3, 159–168.
5 Chyung JH, Raper DM & Selkoe DJ (2005) Gamma-
secretase exists on the plasma membrane as an intact
complex that accepts substrates and effects intramem-
brane cleavage. J Biol Chem 280, 4383–4392.
6 De Strooper B (2003) Aph-1, Pen-2, and nicastrin with
presenilin generate an active-secretase complex. Neuron
38, 9–12.
7 Kimberly WT, LaVoie MJ, Ostaszewski BL, YeW,
Wolfe MS & Selkoe DJ (2003) Gamma-secretase is a
membrane protein complex comprised of presenilin,
nicastrin, Aph-1, and Pen-2. Proc Natl Acad Sci USA
100, 6382–6387.

8 Takasugi N, Tomita T, Hayashi I, Tsuruoka M,
Niimura M, Takahashi Y, Thinakaran G & Iwatsubo T
(2003) The role of presenilin cofactors in the c-secretase
complex. Nature 422, 438–441.
9 De Strooper B, Saftig P, Craessaerts K, Vanderstichele
H, Guhde G, Annaert W, Von Figura K & Van Leuven
F (1998) Deficiency of presenilin-1 inhibits the normal
cleavage of amyloid precursor protein. Nature 391,
387390.
10 Kopan R & Goate A (2000) A common enzyme con-
nects Notch signaling and Alzheimer’s disease. Genes
Dev 14, 2799–2806.
11 Koo EH & Kopan R (2004) Potential role of presenilin-
regulated signaling pathways in sporadic neurodegenera-
tion. Nat Med 10, S26–S33.
12 Thinakaran G & Parent AT (2004) Identification of the
role of presenilins beyond Alzheimer’s disease. Pharma-
col Res 50, 411–418.
13 Okamoto I, Kawano Y, Murakami D, Sasayama T,
Araki N, Miki T, Wong AJ & Saya H (2001) Proteo-
lytic release of CD44 intracellular domain and its role
in the CD44 signaling pathway. J Cell Biol 155, 755–
761.
14 Marambaud P, Wen PH, Dutt A, Shioi J, Takashima
A, Siman A & Robakis NK (2003) A CBP binding tran-
scriptional repressor produced by the PS1 ⁄ -cleavage of
N-cadherin is inhibited by PS1 FAD mutations. Cell
114, 635–645.
15 Pastorcic M & Das HK (2004) Alternative initiation of
transcription of the human presenilin 1 gene in SH-

SY5Y and SK-N-SH cells. The role of Ets factors in the
regulation of presenilin 1. Eur J Biochem 271, 4485–
4494.
16 Pastorcic M & Das HK (1999) An upstream element
containing an ETS binding site is crucial for transcrip-
tion of the human presenilin-1 gene. J Biol Chem 274,
24297–24307.
17 Pastorcic M & Das HK (2003) Ets transcription factors
ER81 and Elk1 regulate the transcription of the human
presenilin 1 gene promoter. Brain Res Mol Brain Res
113, 57–66.
18 Sharrocks AD (2001) The ETS-domain transcription
factor family. Nat Rev, Mol Cell Biol 2, 827–837.
19 Oikawa T & Yamada T (2003) Molecular biology of the
Ets family of transcription factors. Gene (Amst) 303,
11–34.
20 Laudet V, Hanni C, Stehelin D & Duterque-Coquillaud
M (1999) Molecular phylogeny of the ETS gene family.
Oncogene 118, 1351–1359.
21 Monte D, Baert JL, Defossez PA, de Launoit Y & Ste-
helin D (1994) Molecular cloning and characterization
of human ERM, a new member of the Ets family clo-
sely related to mouse PEA3 and ER81 transcription
factors. Oncogene 9, 1397–1406.
22 de Launoit Y, Chotteau-Lelievre A, Beaudoin C, Coutte
L, Netzer S, Brenner C, Huvent L & Baert JL (2000)
The PEA3 group of ETS-related transcription factors.
Role in breast cancer metastasis. Adv Exp Med Biol
480, 107–116.
23 de Launoit Y, Baert JL, Chotteau-Lelievre A, Monte D,

Coutte L, Mauen S, Firlej V, Degerny C & Verreman K
(2006) The Ets transcription factors of the PEA3 group:
transcriptional regulators in metastasis. Biochim Biophys
Acta 1766(1), 79–87.
24 Werner MH, Clore M, Fisher CL, Fisher RJ, Trinh L,
Shiloach J & Gronenborn AM (1995) The solution
structure of the human ETS1–DNA complex reveals a
novel mode of binding and true side chain intercalation.
Cell 83, 761–771.
25 Donaldson LW, Petersen JM, Graves BJ & McIntosh
LP (1996) Solution structure of the ETS domain from
murine Ets-1: a winged helix–turn–helix DNA binding
motif. EMBO J 15, 125–134.
26 de Launoit Y, Baert JL, Chotteau A, Monte D,
Defossez PA, Coutte L, Pelczar H & Leenders F (1997)
Structure–function relationships of the PEA3 group of
Ets-related transcription factors. Biochem Mol Med 61,
127–135.
27 Defossez PA, Baert JL, Monnot M & de Launoit Y
(1997) The ETS family member ERM contains an
alpha-helical acidic activation domain that contacts
TAFII60. Nucleic Acids Res 25, 4455–4463.
28 Laget MP, Defossez PA, Albagli O, Baert JL, Dewitte
F, Stehelin D & de Launoit Y (1996) Two functionally
distinct domains responsible for transactivation by the
Ets family member ERM. Oncogene 12, 1325–1336.
M. Pastorcic and H.K. Das Regulation of the presenilin 1 gene
FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS 1447
29 Aubry F, Mattei MG & Galibert F (1998) Identification
of a human 17p-located cDNA encoding a protein of

the Snf2-like helicase family. Eur J Biochem 254, 558–
564.
30 Woodage T, Basrai MA, Baxevanis AD, Hieter P &
Collins FS (1997) Characterization of the CHD family
of proteins. Proc Natl Acad Sci USA 94, 11472–11477.
31 Ge Q, Nilasena DS, O’Brien CA, Frank MB & Targoff
IN (1995) Molecular analysis of a major antigenic
region of the 240 kD protein of Mi-2 autoantigen.
J Clin Invest 96, 1730–1737.
32 Mauen S, Huvent I, Raussens V, Demonte D, Baert JL,
Tricot C, Ruysschaert JM, Van Lint C, Moguilevsky N
& de Launoit Y (2006) Expression, purification, and
structural prediction of the Ets transcription factor
ERM. Biochim Biophys Acta 1760, 1192–1201.
33 Greenall A, Willingham N, Cheung E, Boam DS &
Sharrocks AD (2001) DNA binding by the ETS-domain
transcription factor PEA3 is regulated by intramolecular
and intermolecular protein–protein interactions. J Biol
Chem 276, 16207–16215.
34 Bojovic BB & Hassell JA (2001) The PEA3 Ets tran-
scription factor comprises multiple domains that regu-
late transactivation and DNA binding. J Biol Chem
276, 4509–45021.
35 Janknecht R (2001) Cell type-specific inhibition of the
ETS transcription factor ER81 by mitogen-activated
protein kinase-activated protein kinase 2. J Biol Chem
276, 41856–41861.
36 Wu J & Janknecht R (2002) Regulation of the ETS
transcription factor ER81 by the 90-kDa ribosomal S6
kinase 1 and protein kinase A. J Biol Chem 277, 42669–

42679.
37 Degerny C, Monte
´
D, Beaudoin C, Jaffray E, Portois
L, Hay RT, de Launoit Y & Baert JL (2005) SUMO
modification of the Ets-related transcription factor
ERM inhibits its transcriptional activity. J Biol Chem
280, 24330–24338.
38 Baert JL, Beaudoin C, Coutte L & de Launoit Y (2002)
ERM transactivation is up-regulated by the repression
of DNA binding after the PKA phosphorylation of a
consensus site at the edge of the ETS domain. J Biol
Chem 277, 1002–1012.
39 Minty A, Dumont X, Kaghad M & Caput D (2000)
Covalent modification of p73alpha by SUMO-1. Two-
hybrid screening with p73 identifies novel SUMO-1-
interacting proteins and a SUMO-1 interaction motif.
J Biol Chem 275, 36316–36323.
40 Romani M, Tonini GP, Banelli B, Allemanni G, Maz-
zocco K, Scaruffi P, Boni L, Ponzoni M, Pagnan G,
Raffaghello L et al. (2003) Biological and clinical role
of p73 in neuroblastoma. Cancer Lett 197, 111–117.
41 Tong JK, Hassig CA, Schnitzler GR, Kingston RE &
Schreiber SL (1998) Chromatin deacetylation by an
ATP-dependent nucleosome remodelling complex. Nat-
ure 395, 917–921.
42 Koipally J, Renold A, Kim J & Georgopoulos K (1999)
Repression by Ikaros and Aiolos is mediated through
histone deacetylase complexes. EMBO J 18, 3090–3100.
43 Kim J, Sif S, Jones B, Jackson A, Koipally J, Heller E,

Winandy S, Viel A, Sawyer A, Ikeda T et al. (1999) Ika-
ros DNA-binding proteins direct formation of chroma-
tin remodeling complexes in lymphocytes. Immunity 10,
345–355.
44 Gill G (2005) Something about SUMO inhibits tran-
scription. Curr Opin Genet Dev 15, 536–541.
45 Pastorcic M & Das HK (2000) Regulation of transcrip-
tion of the human presenilin-1 gene by ets transcription
factors and the p53 protooncogene. J Biol Chem 275,
34938–34945.
46 Schultz DC, Friedman JR & Rauscher FJ III (2001)
Targeting histone deacetylase complexes via KRAB-zinc
finger proteins: the PHD and bromodomains of KAP-1
form a cooperative unit that recruits a novel isoform of
the Mi-2alpha subunit of NuRD. Genes Dev 15, 428–
443.
47 Hong W, Nakazawa M, Chen YY, Kori R, Vakoc CR,
Rakowski C & Blobel GA (2005) FOG-1 recruits the
NuRD repressor complex to mediate transcriptional
repression by GATA-1. EMBO J 24, 2367–2378.
48 Kehle J, Beuchle D, Treuheit S, Christen B, Kennison
JA, Bienz M & Muller J (1998) dMi-2, a Hunchback-
interacting protein that functions in polycomb repres-
sion. Science 282, 1897–1900.
49 Murawsky CM, Brehm A, Badenhorst P, Lowe N,
Becker PB & Travers AA (2001) Tramtrack69 interacts
with the dMi-2 subunit of the Drosophila NuRD chro-
matin remodelling complex. EMBO Rep 2, 1089–1094.
50 Srinivasan R, Mager GM, Ward RM, Mayer J & Sva-
ren J (2006) NAB2 represses transcription by interacting

with the CHD4 subunit of the nucleosome remodeling
and deacetylase (NuRD) complex. J Biol Chem 281,
15129–15137.
51 Lemos TA, Passos DO, Nery FC & Kobarg J (2003)
Characterization of a new family of proteins that inter-
act with the C-terminal region of the chromatin-remo-
deling factor CHD-3. FEBS Lett 533, 14–20.
52 Fields S & Song O (1989) A novel genetic system to
detect protein–protein interactions. Nature 340, 245–
246.
Regulation of the presenilin 1 gene M. Pastorcic and H.K. Das
1448 FEBS Journal 274 (2007) 1434–1448 ª 2007 The Authors Journal compilation ª 2007 FEBS