© 1990 Oxford University Press
Nucleic Acids Research, Vol. 18, No. 24 7331
Phosphorylation influences the binding of the yeast RAP1
protein to the upstream activating sequence of the PGK
gene
Jimmy S.H.Tsang, Yves A.L.Henry, Alistair Chambers, Alan J.Kingsman1 and Susan
M.Kingsman*
Department of Biochemistry, University of Oxford, South Parks Road, Oxford, 0X1 3QU and
department of Molecular Biology, British Biotechnology Ltd, Watlington Road, Cowley, Oxford,
0X4 5LY, UK
Received September 6, 1990; Revised and Accepted November 16, 1990
ABSTRACT
Yeast repressor activator protein 1 (RAP1) binds in vitro
to specific DNA sequences that are found in diverse
genetic elements. Expression of the yeast
phosphoglycerate kinase gene (PGK) requires the
binding of RAP1 to the activator core sequence within
the upstream activating sequence (UAS) of PGK. A DNA
fragment Z + which contains the activator core
sequence of the PGKUAS has been shown to bind
RAP1. Here we report that phosphatase treatment of
RAP1 affected its binding to the PGKUAS but that this
depended on the nature of the sequence flanking the
5' end of the activator core sequence. When the
sequence flanking the 5' end of the activator core
sequence was different from the PGK RAP1-binding
site, phosphatase treatment of RAP1 decreased its
binding to the DNA. When the 5' end of the binding site
was a match to the PGK RAP1-binding site
dephosphorytation of RAP1 increased RAP1 binding to
the DNA. These observations were reproduced when
the minimal functional DNA-binding domain of the
RAP1 protein was used, Implicating a phosphorylationdependent binding of RAP1. This is the first evidence
for phosphorylation-dependent binding of RAP1.
INTRODUCTION
The protein RAP1 was originally identified as a silencer-binding
factor at HMR (1) and was recently shown to mediate DNA-loop
formation in vitro at HML (2). RAP1 was also proposed to
function as both a repressor and activator protein (3). Two
independently identified proteins, TUF, which binds to the RPGbox in genes coding for components of the translation apparatus
(4), and GRF1 (5), have similar properties to RAP1 (6) and may
be identical factors (3, 6, 7). Potential and proven RAPl-binding
sites have now been identified in numerous genes encoding
unrelated products (3, 6, 8-13). The protein RAP1 therefore
appears to possess multiple activities: it appears to bind to many
• To whom correspondence should be addressed
unrelated promoters, it can activate or repress expression and
it can bind differentially to the same promoter to affect regulation.
The promoter-specific properties of RAP1 could be conferred
either by differences in binding affinity due to the nature or the
context of the binding sites, or to the influence of flanking DNA,
or to interactions with adjacent proteins that may themselves be
regulated or to modifications to the RAP1 protein. There is
already evidence that binding affinity is influenced by the
recognition site and by flanking sequences (3, 6). The question
of modifications to the RAP1 protein has not been addressed.
In the yeast Saccharomyces cerevisiae the phosphoglycerate
kinase (PGK) gene encodes one of the highly expressed glycolytic
enzymes accounting for 1 - 5 % of total mRNA and protein (14).
The PGK gene, like most glycolytic genes, is subject to carbonsource regulation. RNA levels are low on non-fermentable carbon
sources such as pyruvate and high on glucose (8, 14). The
promoter region of PGK contains an upstream activating sequence
(PGKUAS, 15) that comprises at least three different sequence
elements (16). These are 1) an activator core (16), 2) three repeats
of the motif 5'CTTCC3' (16) and 3) a strong protein-binding
site called Ypp (17). The activator core, located between bases
- 4 5 8 and - 4 7 3 , contains a RAPl-binding site. Deletion of the
activator core reduces PGK expression to less than 20% (16).
We have recently proven that the RAPl-binding site in the
activator core is bound by RAP1 present in crude nuclear protein
extracts and by RAP1 produced in vitro (8). The CTTCC blocks
are necessary for maximal expression but not essential for
transcriptional activation (16). Each CTTCC block contributes
differendy to PGK expression and the cognate binding factors,
if any, have not been identified (16). The Ypp, located between
bases - 4 % to -523, binds the ARS-binding factor ABF1 (5,
18, 19). The role of Ypp is at present unclear, as its deletion
does not affect the activity of the authentic PGK promoter (8).
It does, however, augment the function of the RAPl-binding site
in an artificial assay promoter (17).
We have recently shown that the binding of RAP1 to the
activator core sequence of PGK is carbon-source dependent. Little
7332 Nucleic Acids Research, Vol. 18, No. 24
or no DNA-protein complex was observed when nuclear protein
extracts from pyruvate-grown cells were used in gel retardation
assays (8). On the other hand, the expression of RAP1 is not
regulated. Similar amounts of RAP 1-specific RNA (8) and
RAPl-/3-galactosidase fusion protein (unpublished data) were
detected in cells grown on glucose or on pyruvate. The carbonsource- dependent binding of RAP1 to PGKUAS may therefore
be explained by post-translational modification such as
phosphorylation.
Phosphorylation is a common post-translational modification
that affects the properties of a wide range of proteins with
different functions. The binding of at least two transcription
factors, the c-fos serum response factor (20) and the nuclear factor
CREB (21), is influenced by phosphorylation. In addition several
viral proteins that bind DNA are phosphoproteins (22-24). In
this paper we have addressed the role of phosphorylation of RAP 1
in determining binding to the PGK promoter.
MATERIALS AND METHODS
Strains and growth conditions
The Escherichia coli strain used was MC1061 (F~ araD139,
(ara, leu)1696, lacY74, gal U~, gal K~, hsr~, hsm+, str A).
The yeast strain used was Saccharomyces cerevisiae DBY745
(a adel-100 leu2-3 Ieu2-112 ura3-52). E. coli cultures were
grown at 37°C in L-broth (1% tryptone, 0.5% yeast extract,
0.5% NaCl). Yeast cultures were grown at 30°C in YEPD (1 %
yeast extract, 2% peptone, 2% glucose) or minimal medium
containing 0.67% yeast nitrogen base without amino acid (Difco)
supplemented with the necessary amino acids and 2% glucose.
DNA probes
The UAS of PGK is schematically represented in Figure 1. The
DNA fragments used in this study and their corresponding
locations in the UAS are also depicted. The DNA fragments Z +
(-409 to -473) was subcloned from PGK into pSP46 (16, 25).
DNA fragments Z + 6 (-406 to -479), Z + 2 3 (-406 to - 4 % )
and Z + 4 5 ( - 4 0 6 to -518) were isolated from plasmids
pMA731, pMA732 and pKV530, respectively, by
BamHUEcoOl09l digests (15). The isolated fragments were endlabelled using cr^P dTTP (400Ci/mmol) or a 32 P dCTP
(3000Ci/mmol) and Klenow polymerase (Amersham).
Retardation Gel analysis
Preparation of yeast nuclear protein extracts by homogenization
and production of in vitro translated RAP1 using a rabbit
reticulocyte lysate system (Promega) were as described previously
(8, 25). The gel retardation assay was based on the methods of
Fried and Crothers and Gamer and Revzin (26, 27). Binding
reactions using yeast nuclear protein extracts contained 1/tg of
nuclear protein. This was incubated with the labelled DNA
fragment in a buffer containing 5% glycerol, lmM Na2EDTA,
lOmM /3-mercaptoethanol, 25mM Tris-HCl pH 7.5, 25mM NaCl
and 20mM KC1. 200ng of calf-thymus DNA was added to each
binding reaction as a non-specific competitor. After a thirtyminute incubation at 22°C, DNA-protein complexes were
resolved by electrophoresis on a 5% polyacrylamide gel
containing 0.5xTBE. Binding reactions using the in vitro
translated RAP1 were similar to those using the yeast protein
extracts. 0.2-0.5/tl of RAP1 or mock lysate was incubated with
the labelled probe in the presence of 400ng of calf-thymus DNA.
DNA-protein complexes were resolved as described earlier.
Phosphatase treatment
Calf intestine alkaline phosphatase and potato acid phosphatase
were obtained from Boehringer Corporation London (products
713023 and 108197, respectively). Insoluble alkaline phosphatase
was obtained from Sigma (P-0762). These enzymes were added
to binding reactions without the labelled DNA fragment. The
enzyme was allowed to react with the extract or in vitro translated
RAP1 for fifteen minutes at 22°C before the labelled probe and
the non-specific competitor DNA were added.
RESULTS
Phosphatase treatment of a crude nuclear protein extract
abolishes RAP1 binding to the PGK Z + fragment
The PGKUAS, located between bases -402 to -538 upstream
of the initiation codon, has been subcloned into smaller fragments
(16, 25). The Z + fragment contains sequences from -409 to
-473. This region contains the RAPl-binding site and the three
5'CTTCC3' blocks. The Z + fragment has been shown to
produce a single major complex when retarded with a crude yeast
nuclear protein extract prepared from glucose-grown cells (16).
This complex was caused by the binding of RAP1 to the
consensus binding site located between —458 and —473 and its
occurrence was carbon-source dependent (8). To examine the
role of phosphorylation on RAP1 binding activity we treated the
crude yeast nuclear protein extract with calf intestine alkaline
phosphatase (Fig. 2 lanes 4 and 5) prior to the normal retardation
gel assay. The results show that the specific RAP 1-DNA complex
(Fig. 2 complex R) diminished with phosphatase treatment (Fig.
2, compare lanes 3 and 5). Moreover, a new faster migrating
complex (complex C) was identified.
To further support the notion that dephosphorylation modifies
RAP1 binding to the Z + fragment the experiment was repeated
using another source of phosphatase. Potato acid phosphatase was
used and the results (Fig. 2 lanes 9—13) confirmed the
observations obtained using alkaline phosphatase, namely that
the RAP 1-DNA complex (R) was replaced by a faster-migrating
complex (C). This faster-migrating complex was not a modified
• Ml
z"
I
1
f""
M«.
Figure 1. Structural organization of the P G K ^ j and the DNA fragments used
in this study. The numbers represent the location of the nucleotides upstream
of the initiation codon.
Nucleic Acids Research, Vol. 18, No. 24 7333
form of RAPl as it was not recognised by a specific anti-RAPl
antibody. We used a concentration of acid phosphatase that
allowed both the normal RAP1-Z+ complex (R) and the fastermigrating complex (C) to form (Fig. 3 lanes 3 and 6). The
addition of specific anti-RAPl antibody (a generous gift from
D. Shore, Columbia University) would abolish any complex
formation related to RAPl binding (Fig. 3, compare lanes <t and
5; 8). The use of anti-RAPl antibody abolished the formation
of the normal slow-migrating complex R but had no effect on
the formation of the faster-migrating complex C (Fig. 3, compare
lanes 6 and 7). In order to eliminate the presence of other yeast
proteins which might interact with the Z + fragment specifically
or non-specifically after phosphatase treatment and complicate
the result, in vitro translated (TVT) RAPl protein was used (8).
Figure 4A shows that there was no binding of the mock lysate
proteins (an IVT reaction without any RNA), with or without
phosphatase treatment, to the Z + fragment (Fig. 4A lanes 3 and
4). Moreover, the binding of the IVT RAPl to the Z + fragment
was abolished after phosphatase treatment (Fig. 4A, compare
lanes 5 and 6). This abolition of RAPl binding was not due to
protein degradation as incubation of IVT RAPl with phosphatase
does not cause degradation (data not shown). This confirmed our
results using the crude extracts, which showed that
dephosphorylation of RAPl abolished its binding to the Z +
fragment. These changes could be inhibited by the addition of
the phosphatase inhibitor 30mM ammonium molybdate (Fig. 4B,
compare lanes 3 to 4). Furthermore, the DNA-binding activity
of dephosphorylated RAPl can be restored by phosphorylating
the protein. Insoluble phosphatase was used to dephosphorylate
IVT RAPl and the phosphatase was removed by centrifugation
through a column of glass beads. Half of this phosphatase-treated
lysate was then treated with catalytic unit of cAMP-dependent
protein kinase (ca. 40 picomole unit, Sigma P-2645). Both lysates
:alff
intestine
alkaline
phosphatase
potato
acid
phosphata so
Extract
Phosphatas*
were then used for retardation gel assays. Figure 4C shows that
insoluble phosphatase decreased the DNA binding activity of
RAPl while protein kinase restored its binding to the Z +
fragment (Fig. 4C, compare lanes 3 and 4).
DNA sequences 5' to the activator core sequence determine
the modifiction of RAPl binding
The Z + fragment, which contains sequences -409 to -473
upstream of the initiation codon of PGK, is a subfragment of
the PGKUAS located between bases -402 to -538. The
PGKUAS, denoted as the X fragment, has been shown to give
more than one protein-DNA complex when retarded with a crude
nuclear protein extract (25). When the X fragment was used to
test the effect of phosphatase on RAPl binding a contrasting result
to that for the Z + fragment was obtained. Figure 5A shows that
when IVT RAPl was pretreated with alkaline phosphatase an
increase in binding to the X fragment was detected (Fig. 5A,
compare lanes 5 and 6). This result was very surprising because
the two DNA fragments, Z + and X, contain the same RAPl
core recognition sequence. This prompted us to investigate
whether the nature of the DNA fragment affects RAPl binding.
In the X fragment the RAPl-binding site is located roughly
in the centre, flanked by 65bp from one end and 61bp from the
other end, of the DNA fragment. In the Z + fragment the RAP1binding site is located on one end adjacent to the polylinker
cloning site where the fragment was isolated. In order to place
the RAPl binding site in the central region of the DNA fragment
plasmid pSP46-Z+ (16) was digested with SalUBanl and a
208bp fragment (Z +B ) was isolated and used for retardation gel
analysis. This fragment, Z + B , contains additional pSP46 vector
DNA sequence upstream of the Z + fragment. In this fragment
the RAPl-binding site is flanked by 138bp and 60bp, thus
imitating the X fragment. The result indicated that the binding
of in vitro produced RAPl to this fragment was also decreased
by phosphatase treatment (Fig. 5B, compare lanes 5 and 6). This
implied that the position of the RAPl-binding site in the fragments
Extract
Phosphatase
Pre-immune
anti-RAP1
0
t
fctftfc
: I- ¥
3
4
5
6
1
8
9
10 11 12 13
Figure 2. Effect of phosphalase treatment of crude nuclear protein extract on
DNA binding. The positions of the free fragment (F) and the retardation complexes
(R, O are indicated. Fragments with (lanes 2 and 7) and without (lanes 1 and
6) phosphatase are shown as controls. Binding reactions on the Z + fragment with
protein extract prepared from cells grown on glucose were pretreated with calf
intestine alkaline phosphatase (lanes 3-5) or potato acid phosphatase (lanes 9-13).
The amounts of phosphatase used (in units) are indicated.
Figure 3. Identification of RAPl-related complexes. Binding reactions on the
Z* fragment containing protein extract prepared from cells grown on glucose
were pretreated with (lanes 3, 6 and 7) or without (lanes 2, 4 and 5) 0.2 units
of acid phosphatase. Some of the reactions were also reacted with pre-immune
(lanes 4 and 6) or anti-RAPl (lanes 5 and 7) antiserum. The position of the free
fragment (F) and the retardation complexes (R, C) are indicated.
7334 Nucleic Acids Research, Vol. 18, No. 24
B
Lyaata
Phoaphatasa
Mock
R*Pl
RAP1
phosphatase
molybdate
+
1 2
3
+ +
+
•#-
-
+ +
-
4
RAF1
phosphatase
kinase
Figure 4. Binding of in vitro translated RAP1 to the Z + fragment. The positions of the free fragment (F) and the specific complex (R) are indicated. (A) Effect
of phosphatase on binding. Binding reactions were pretrcated with (lanes 2, 4 and 6) or without (lanes 1, 3 and 5) 2 units of alkaline phosphatase. Lanes 3 and
4 contain mock lysate and lanes 5 and 6 contain FVT RAP1. (B) Inhibition of phosphatase effect on binding. Binding reactions on the Z + fragment with IVT RAP1
were treated with (lanes 3 and 4) or without (lanes 2 and 5) 2 units of alkaline phosphatase in the presence (lanes 4 and 5) or absence (lanes 2 and 3) of 30mM
ammonium molybdate. ( Q Removal of insoluble phosphatase and post-treatment of the lysate with protein kinase re-establish RAP 1-binding activity. Insoluble alkaline
phosphatase (Sigma) was used to pretrcat IVT RAP1 and was removed by centrifugation. The lysate was then divided into two portions (lanes 3 and 4) and one
of which (lane 4) was then treated with catalytic unit of cAMP-dependent protein kinase (ca. 40pMU) prior to retardation gel analysis. Lane 7 is a control to show
that the protein kinase does not contain DNA-binding activity.
being tested cannot account for the differences in the effects of
phosphatase treatment on RAP1 binding to the Z + fragment and
to the X fragment.
A previous study using the methylation interference footprinting
technique indicated that RAP1 interacted with sequences upstream
of the activator core sequence of the Z + fragment (8). This
suggested that RAP1 might contact sequences upstream of -473,
the end point of the Z + fragment. Furthermore, a close
comparison between the RAP 1-binding consensus and the
activator core sequence of DNA fragments Z + and Z + B
revealed that these fragments do not contain matches to the first
two bases of the consensus. DNA fragments containing various
lengths of PGK sequences upstream of the 5' end point of the
Z + fragment were therefore tested. DNA fragments Z + 6 , Z + 2 3
and Z + 4 5 , containing 6, 23 and 45 additional authentic bases,
respectively, upstream of the 5' end point of the Z + fragment,
were isolated from plasmids pMA731, pMA732 and pKV530,
respectively (15). Retardation gel analysis showed that the binding
of RAP1 to these fragments was increased by phosphatase
treatment (Fig. 5C, compare lanes 2 and 3, 5 and 6, and 8 and
Nucleic Acids Research, Vol. 18, No. 24 7335
Lysate
Mock
RAP1
phosphatase
Lysate
Mock
RAP1
3
5
phosphatase
1 2
3
4
5
1 2
4
6
RAP1
phosphatase
Figure 5. Effect of phosphatase on binding to the PGK^^ subfragments. The position of the free fragment (F) and the RAPI-DNA specific complex (R) are indicated.
(A) Binding reactions on the x fragment with mock lysate (lanes 3 and 4), or IVT RAP1 (lanes 5 and 6). The reactions were pretreated with (lanes 2, 4 and 6)
or without (lanes I, 3 and 5) 1 unit of alkaline phosphatase. (B) As in (A) except the DNA fragment Z + B is used. (C) Binding reactions with IVT RAPI on DNA
fragments Z + 6 (lanes 1-3), Z + 2 3 (lanes 4 - 6 ) , and Z + 4 5 (lanes 7 - 9 ) . Lanes 3, 6 and 9 were pretreated with 1 unit of alkaline phosphatase.
9). This suggested that DNA sequences 5' to the activator core
sequence of PGKUAS determine the outcome of the
phosphorylation-dependent binding of RAPI.
The response of a RAPI DNA binding domain (28) to
phosphatase treatment is similar to that of full-length RAPI.
Figure 6 shows that the RAPI DNA-binding domain decreased
its binding to the Z + and Z + B fragments but increased its
binding to the X, Z + 6 , Z + 2 3 and the Z + 4 5 fragments, when
treated with phosphatase. This implied that the observed responses
of RAPI to phosphatase were most likely achieved via the DNAbinding domain.
DISCUSSION
Phosphorylation has been shown to affect the DNA-binding ability
and biological activity of numerous cellular components such as
c-fos serum response factor (20), the heat shock element binding
protein (29) and cAMP response element binding protein (21).
Here we show that the DNA-binding ability of the yeast RAPI
protein is also affected by its phosphorylation state.
Treatment of crude yeast protein extracts and reticulocyte
lysates with phosphatase could result in a non-specific interference
with RAPI binding. This is however unlikely, as both an increase
or decrease of binding with the same extracts have been observed
dependent upon the target DNA fragment. The modification of
RAPI binding is most likely phosphatase-specific. First, similar
results on DNA binding modification were obtained by using calf
intestine alkaline phosphatase or potato acid phosphatase (Fig.
2), suggesting that the effect was not due to the storage buffers
of the enzymes. The use of insoluble phosphatase (Fig. 4C)
removes the doubt of any buffer effect as no other buffer was
supplemented. Second, the modification was inhibited by
phosphatase inhibitor, in this case, ammonium molybdate, further
indicating that the effect was due to phosphatase (Fig. 4B). Third,
7336 Nucleic Acids Research, Vol. 18, No. 24
KSSPDDFETA PAEYVDALDP SMWVDSGSA AVTAPSDSAA EVKANQNEEN
t
TGATAAETSE KVDQTEVEKK DDDDTTEVGV TTTTPSIADT AATANIASTS
Phosphatasa
GASVTEPTTD DTAADEKKEQ VSGPPLSNMK FYLNRDADAH DSLNDIDQLA
RIIRANGGEV LDSKPRESKE NVFIVSPYNB TNLPTVTPTY IKACCQSNSL
LNHENYLVPY DHFREWDSR LQEESHSNGV DNSNSHSDNK DSIRPKTEII
0
STOTNGATED STSEKVHVDA EQQARLQEQA QLLRQHVSST ASITSGGHND
0
LVQJEQPQKD TSNWHNSNVM DEDNDLLTQD NNPQTADEGN ASFQAQRSMI
1 2
3
4
5
+ 23
Z
+45
Z
Phosphatase
SRGALPSHNK ASFTDEEDEF ILDWRKNPT RRTTHTLYDE ISHYVPNHTG
o •
tt
AA
A
NSIRHRFRVY LSKRLEYVYE
DDGNLIKTKV LPPSIKRKFS
A
• • VDKFGKLVRD
A
At
ADEDYTLAIA VKKQFYRDLF
ITDEDTPTAI ARRNMTMDPN
QIDPDTGRSL t
HVPGSEPNFA AYRTQSRRGP
EEHAAHTENA HRDRFRKFLL
A
IAREFFKHFA
AYGIDDYISY YEAEKAQNRE
PKRPGVPTPG NYNSAAKRAR
PEPMKHLTNR
NKSSQRNVQP TAHAASANAA
YAIPENELLD EDTMNFISSL
AAAAAAASNS
KNDLSNISNS LPFEYPHEIA
DIYDNIDPDT ISFPPKIATT
ERIRSDFSNE
DLFLPLEFHF GSTRQFMDKL
SQAEKLVQDL CDETGIRKNF
HEVISGDYEP
STSILTCLSG HLHVFPRYFL
PNVPGIWTHD DDESLKSNDQ
0
HMFKDNVNPP
t
EQIR1CLVKKH GTGRMEMRKR
FFEKDLL
Protein kinase C
casein kinase I
tyrosine kinase
casein kinase II
Figure 7. Potential phosphorylation sites of RAP1. The arrows demarcate the
limits of the minimal DNA binding domain of RAP1.
Figure 6. Effect of phosphatase treatment of RAP1 DNA binding domain on
binding to various DNA fragments. Binding reactions with an IVT RAP1 DNA
binding domain on Z + (lanes 1 and 2), Z + B (lanes 3 - 5 ) , X (lanes 6 and 7),
Z + 6 (lanes 8-10), Z + 2 3 (lanes 11-13), and Z + 4 5 Canes 14-16) fragments
were treated with (lanes 2, 5, 7, 10, 13 and 16) or without (lanes 1, 4, 6, 9,
12 and 15) 1 unit of alkaline phosphatase. Lanes 3, 8, 11 and 14 contain free
fragments only. The arrows indicate the position of the specific retardation
complexes.
the loss of DNA-binding activity on the Z + fragment caused by
phosphatase can be reversed by phosphorylating the protein (Fig.
4C).
A previous study showed that mutations at the 5' end of the
RAP1 recognition site did not affect RAP1 binding (30). This
agrees with our observations that DNA fragments Z + and Z + B
bind RAP1 efficiently. However, we show that the mere addition
of six authentic base pairs to the Z + fragment, which is fully
capable of binding to RAP1, altered its response to RAP1 binding
with respect to phosphatase treatment. DNA fragments Z + and
derivative Z + B both decreased their binding (Figs. 4A and 5B)
while DNA fragments X, Z + 6 , Z+ 23 and Z + 4 5 all showed
increases in binding when RAP1 was treated with phosphatase
(Figs. 5A and 5C). This probably reflects the involvement of
the 5' end of the RAP 1-recognition site in determining the stability
of RAP1 binding, especially in response to phosphatase
treatment. DNA sequences 5' to the RAPl-binding site have been
shown to be the DNA-bending site induced by RAP1 (30). It
may be possible that phosphorylation affects DNA binding and
bending abilities of RAP1 depending on the nature of the
sequences 5' to the core recognition site. The involvement of
additional sequences may help the protein to recognise different
binding sites and to act selectively in response to different stimuli.
It should be noted that the decrease in binding to the Z +
fragment in this in vitro assay may not necessarily reflect the
in vivo situation. The Z + fragment has been shown to display
activation and regulation functions when inserted upstream of
an artifical assay promoter (8, 17). It is likely that the vector
sequences adjacent to the RAPl-binding site of the Z + and the
Z + B fragments disrupt or de-stabilise the structural integrity of
the binding complex during the analysis in vitro (31) while in
in vivo situation RAP1 interacts with other promoter elements
(32, 33), thus stabilising its binding to the binding site.
Regulation of transcription involves DNA-protein interactions
as well as protein-protein interactions. Using a minimal promoter
assay system derived from PGK, we have demonstrated that
RAP1 requires additional interactions to form an active
transcription complex (17). Either the Y-binding protein ABF1
which binds -496 to -523 or the CTTCC blocks which probably
bind a protein factor, can supply the necessary additional
interactions. When we treated a nuclear protein extract with
phosphatase we detected a faster-migrating complex (Fig. 2
complex C) in retardation gel analysis. This faster- migrating
complex is unlikely to be a modified form of RAP1 because an
incubation with specific anti-RAPl antibody prior to DNA
binding did not inhibit the formation of this complex (Fig. 3).
Furthermore, this phosphatase-induced complex was also detected
when a DNA fragment Z, containing sequences —402 to - 4 6 0
(25), which lacks the RAPl-binding site, was used for analysis
(unpublished data). This newly detected complex may be a
specific DNA-protein complex caused by the interaction of other
specific factors with the sequence shared by the Z + and the Z
fragments, namely the CTTCC blocks or the yATF (34) binding
site located between bases -416 and —425, or it may be a nonspecific complex induced by phosphatase. Further investigation
is required to prove that this complex is caused by a specific DNA
binding protein.
We have recently identified a minimal functional DNA-binding
domain of the protein RAP1 to residues 361 to 596 (28). An in
vitro produced protein containing this minimal DNA-binding
domain possesses DNA-binding ability similar to that of fulllength RAP1 (28). When this protein was treated with
phosphatase and retarded with various PGK subfragments a
Nucleic Acids Research, Vol. 18, No. 24 7337
similar result to that of full-length RAPl was observed (Fig. 6).
This implied that the phosphorylation-dependent binding of
RAPl is most likely achieved via the binding domain. Located
in this minimal-binding domain is a cluster of potential
phosphorylation sites for protein kinase C (Fig. 7; 28). Moreover,
18 out of a total of 28 potential phosphorylation sites, as
determined by a computer search using the PIP program
developed by Staden using the kinase recognition sequences
described by Leader and Katan (35), were located in this domain.
However, we have no direct evidence, yet, to show that RAPl
is a phosphoprotein. The actual phosphorylation site(s) and
enzyme(s) involved in modifying RAPl DNA-binding activity,
are still waiting for identification.
ACKNOWLEDGEMENTS
We thank D. Shore for the generous gift of the anti-RAPl and
pre-immune antisera.
YH acknowledges the financial support of United Kingdom
M.R.C. This work was supported by grants from the United
Kingdom A.F.R.C.
REFERENCES
1. Shore, D., Stillman, D.J., Brand, A.H. and Nasmyth, K.A. (1987) EMBO
J. 6, 461-467.
2. Hofmann, J.F.-X., Laroche, T., Brand, A.H. and Gasser, S.M. (1989) Cell
57, 725-737.
3. Shore, D. and Nasmyth, K. (1987) Cell 51, 721-732.
4. Huet, J., Cottrelle, P., Cool, M., Vignais, M.L., Thiele, D., Marck, C ,
Buhler, J.M., Sentenac, A. and Fromageot, P. (1985) EMBO J. 4,
3539-3547.
5. Buchman, A., Kimmerley, W.J., Rine, J. and Komberg, R. (1988) Mol.
Cell. Biol. 8, 210-225.
6. Buchman, A.R., Lue, N.F. and Komberg, R.D. (1988) Mol. Cell. Biol.
8, 5086-5099.
7. Huet, J. and Sentenac, A. (1987) Proc. Natl. Acad. Sci. USA 84, 3648-3652.
8. Chambers, A., Tsang, J.S.H., Stanway, C , Kingsman, AJ. and Kingsman,
S.M. (1989) Mol. Cell. Biol. 9, 5516-5524.
9. Capieaux, E., Vignais, M.-L., Sentenac, A. and Goffeau, A. (1989) J. Biol.
Chem. 264, 7437-7446.
10. Longtine, M.S., Wilson, N.M., Petracek, M.E. and Berman, J. (1989) Curr.
Genet. 16, 225-239.
11. Hurd, H.K. and Roberts, J.W. (1989) Mol. Cell. Biol. 9, 5359-5372.
12. Nishizawa, M., Araki, R. and Teramishi, Y. (1989) Mol. Cell. Biol. 9,
442-451.
13. Vignais, M.L., Woudt, L.P., Wassenaar, G.M., Mager, W.H., Sentenac,
A. and Planta, R.J. (1987) EMBO J. 6, 1451-1457.
14. Holland, M.J. and Holland, J.P. (1978) Biochem.17, 4900-4907.
15. Odgen, J.E., Stanway, C , Kim, S., Mellor, J., Kingsman, A.J. and
Kingsman, S.M. (1986) Mol. Cell. Biol. 6, 4335-4343.
16. Chambers, A., Stanway, C , Kingsman, A.J. and Kingsman, S.M. (1988)
Nucleic Acids Res., 16, 8245-8260.
17. Stanway, C.A., Chambers, A., Kingsman, AJ. and Kingsman, S.M. (1989)
Nucleic Acids Res., 17,9205-9218.
18. Diffley, J,F-X. and Stillman, B. (1989) Science 246, 1034-1038.
19. Chambers, A., Stanway, C , Tsang, J.S.H., Henry, Y.A.L., Kingsman,
A.J. and Kingsman, S.M. (1990) Nucleic Acids Res. 18, 5393-5399.
20. Prywes, R., Dana, A., Cromlish, J.A. and Roeder, R.G. (1988) Proc. Natl.
Acad. Sci. USA 85, 7206-7210.
21. Yamamoto, K.K., Gonzalez, G. A., Biggs III, W.H. and Montminy, M.R.
(1988) Nature 334, 494-498.
22. Klein, H., Maltzman, W. and Lcvine, A.J. (1979) J. fliol. Chem. 254,
11051-11060.
23. Raychaudhuri, P., Bagchi, S. and Nevins, J.R. (1989) Genes Dev. 3,
620-627.
24. WiJcox, K.W., Kohn, A., Skylanskaya, E. and Roizman, B. (1980) J. Virol.
33, 167-182.
25. Stanway, C , Mellor, J., Odgen, J. E., Kingsman A.J. and Kingsman, S.M.
(1987) Nucleic Acids Res. 15, 6855-6873.
26. Fried, M. and Crother, D.M. (1981) Nucleic Acids Res. 9, 6505-6525.
27. Garner, M.M. and Revzin, A. (1981) Nucleic Acids Res. 9, 3037-3060.
28. Henry, Y.A.L., Chambers, A., Tsang, J.S.H., Kingsman, A.J. and
Kingsman, S.M. (1990) Nucleic Acids Res. 18,2617-2623.
29. Sorger, P.K. and Pelham, H.R.B. (1988) Cell 54, 855-864.
30. Vignais, M-L. and Sentenac, A. (1989) J. Biol. Chem. 264,8463-8466.
31. Baumruker, T., Sturm, R. and Herr, W. (1988) Genes Dev. 2, 1400-1413.
32. Brand, A.H., Micklem, G. and Nasmyth, K. (1987) Cell 51,709-719.
33. Sousa, R. and Arcangioli, B. (1989) EMBO J. 8, 1801-1808.
34. Lin, Y-S. and Green, M.R. (1989) Proc. Natl. Acad. Sci. USA 86, 109-113.
35. Leader, D.P. and Katan, M. (1988) J. gen. Virol. 69, 1441-1464.