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.. Pro-g
Trang 1CIDD region of the Ets transcription factor ERM and
represses transcription of the human presenilin 1 gene
Martine Pastorcic1and Hriday K Das1,2
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: hdas@hsc.unt.edu
(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.
Trang 2Presenilins (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
Trang 3highly 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
5 m 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.
Trang 4fragment was cloned inframe with the GAL4BD(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 encodscreen-ing 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
Trang 5reporter (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.
Trang 6transformants 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 lg of ( )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 lg of ( )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.
Trang 7Interaction 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 Gal4BD–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 lg of ( )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 lg of ( )118, +178) PS1CAT wild-type (wt) or containing
point mutations at the sites indicated in the presence of 4 lg of
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.
Trang 8examined 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 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.
Trang 9C-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 lg of 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.
Trang 10The 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