ABSTRACT
The yeast zinc cluster protein HAP1, a member of the GAL4 family, is a
transcriptional activator that binds as a homodimer to target DNA sequences.
These targets include the upstream activating sequences of the
CYC1
and
CYC7
genes, which have no obvious sequence similarity. Even though both sites have
the same affinity for HAP1, activation differs at these two sites, even when
the sequences are placed in an identical promoter context. In addition, mutants
of HAP1 that can bind to both sites but are specifically transcriptionally
inactive at
CYC7
have been previously isolated. In order to identify nucleotides that are
responsible for this differential activity, we have performed random and site-directed mutagenesis of these target sites and assayed their binding to
HAP1
in vitro
and their activity
in vivo
in reporter plasmids. Our results show that HAP1 binding sites are degenerate
forms of the direct repeat CGG N
3
TA N CGG N
3
TA. Moreover, we show that activity of HAP1 mutants defective for activation of
the
CYC7
gene is restored by specific mutations in the
CYC7
binding site. Conversely, other mutations of the target sites prevent
activation by HAP1, without interfering with DNA binding. The results suggest
that the sequence of the target sites influences the conformation and, hence,
the activity of DNA-bound HAP1.
Many fungal transcriptional regulators contain a zinc finger known as a
binuclear zinc cluster. The consensus sequence of the cluster is CX
2
CX
6
CX
5-9
CX
2
CX
6-8
C, where the cysteines bind two zinc atoms which coordinate folding of the
domain. A well-characterized member of this family is GAL4, which binds as a homodimer to
palindromic DNA sequences (CGG N
11
CCG), with each zinc cluster recognizing a CGG triplet, as shown by X-ray crystallography (
1
and references therein). The zinc cluster domain is followed by a short linker
region and a dimerization domain. While the spacing between the CGG triplets is
11 base pairs (bp) for GAL4, other zinc cluster proteins also recognize the
same palindromic triplets, but with a different spacing. For example, the PUT3
and PPR1 activators bind CGG triplets spaced by 10 and 6 bp respectively (
2
-
3
). Moreover, the geometry of the zinc cluster controls binding specificity, as
changing the spacing of the GAL4 site from 11 to 10 bp greatly reduces GAL4
binding
in vitro
(
4
). Construction of GAL4, PUT3 and PPR1 chimeric proteins has shown that the
linker region is responsible for binding to a site of a given spacing (
5
-
7
).
HAP1, another member of this family of zinc cluster proteins (
8
-
10
), possesses an acidic activation domain at its C-terminus (residues 1307-1483) and an N-terminal DNA binding domain (residues 1-174) (
10
,
11
). The zinc cluster of HAP1 (residues 64-93) is homologous to that of GAL4 (43% identity) and the coiled coil
sequence of the dimerization domain is similar to the domain identified in GAL4
(
11
). The middle region of HAP1 is required for regulation of DNA binding through
interaction with heme. In the absence of heme HAP1 is present in a high
molecular weight complex (
12
,
13
). Binding of heme to HAP1 apparently stimulates dissociation of HAP1 from the
complex and subsequent homodimerization, which is a prerequisite for specific
DNA binding.
Upstream activating sequences (UASs) for HAP1 (see Fig.
2
for DNA sequences) have been identified in the
CYC1
and
CYC7
genes, which encode isoforms of cytochrome c (
14
-
16
). In addition, other binding sites for HAP1 have been identified in the
CTT1
(catalase T), the
CYB2
(cytochrome b
2
) and the
CYT1
(cytochrome c
1
) genes (
17
-
19
). These sites share little homology. This is particularly evident when
comparing the UAS of
CYC1
with that of
CYC7
, which has led to the suggestion that HAP1 binds to unrelated DNA sequences (
14
).
Dissimilarity between these sites results in different levels of HAP1-dependent activation. Activity at the
CYC1
promoter is higher than at
CYC7
, even if the UASs are placed in a similar promoter context. This different
transcriptional activity is not a result of different affinity of HAP1 for
CYC1
and
CYC7
(
20
). Furthermore, a single amino acid change in HAP1 of Ser63 to Arg, immediately
N-terminal of the first cysteine of the zinc cluster, prevents binding to
the UAS of
CYC1
and results in greatly increased activation (10- to 100-fold) at
CYC7
, even though the affinity of the mutant HAP1-18 for the UAS of
CYC7
is similar to wild-type HAP1 (
20
). Other HAP1 mutants, termed positive control (PC) mutants, with mutations that
flank the zinc cluster show wild-type binding and activation at
CYC1
but are transcriptionally inactive at
CYC7
, even though they show wild-type or increased binding at that site (
21
).
We wished to learn more about the nucleotides responsible for this differential
activity of HAP1 (and its mutants) at
CYC1
and
CYC7
. We first performed saturation mutagenesis of the
CYC1
and
CYC7
UASs (as well as site-directed mutagenesis at other UASs of HAP1) to identify key nucleotides
responsible for binding of HAP1. Our results show that all UASs of HAP1 are
imperfect versions of the direct repeat CGG N
3
TA N CGG N
3
TA. Moreover, at a similar position the
CYC7
site has two CGC triplets instead of the two CGGs. Changing these triplets to
CGG restores the activity of HAP1 PC mutants, while the activity of HAP1-18 is greatly decreased.
Strain THl
(Mat
a
ura3-52 his4-519 ade1-100 leu2-2
[Delta]
hap1
::
hisG
) is a derivative of LPY22 (
21
). Most of the HAP1 coding region has been deleted in that strain. Rich (YPD)
and synthetic (SD) media were prepared as described (
25
).
Cells were grown to saturation in YEP containing 2% raffinose. They were then
diluted into minimal medium containing 1% glucose and 1% galactose and grown
for ~12-18 h (OD
600
0.6-1.0) before assaying for [beta]-galactosidase activity. [beta]-Galactosidase assays were performed with
permeabilized cells (
27
).
Extracts were prepared and DNA binding assays were carried out as described (
21
,
28
) using a HAP1 expression vector under the control of UAS
GAL
(SD5-HAP1;
20
). Probes were generated by PCR amplification using the reporter plasmids as
templates and purified with G50 spin columns. Quantitative data were obtained
by measuring the amount of radioactivity present in the retarded bands using a
phosphorimager (Fuji).
Reporter plasmids with mutated
CYC1
and
CYC7
UASs were generated using `spiked' oligonucleotides (level of contamination 3%
of each of the three other nucleotides), flanked by appropriate restriction
sites (see Materials and Methods). Mutagenized UASs were then subcloned into a
reporter plasmid containing a minimal
CYC1
promoter driving
lacZ
transcription (Fig.
1
). Mutations in the UASs were sequenced and the activity of the reporters was
determined
in vivo
by transforming them, along with a HAP1 expression vector, into a hap1
-
strain. In the presence of HAP1, activity at
CYC1
and
CYC7
was 6 and 2.7 U respectively (Tables
0
and
0
). Moreover, when a HAP1 mutant deleted of its activation domain was used (HAP1[Delta]Kp) only background activity was measured (0.1-0.3 U [beta]-galactosidase activity; data not shown), even though
this HAP1 mutant showed wild-type DNA binding
in vitro
(
21
). This indicates that activity of these reporters was dependent on the
activation domain of HAP1. Mutants were also tested for their
in vitro
DNA binding by the electrophoretic mobility shift assay (EMSA) using an extract
prepared from a strain that contained a HAP1 overexpression vector under the
control of UAS
GAL4
.
Many mutants generated by random mutagenesis of the
CYC1
UAS displayed reduced binding and
in
vivo activity, like the single point mutants 1-9 and 1-14 (Table
1
). Activation from these sites is 10- to 20-fold lower than from the wild-type UAS of
CYC1
. Strikingly, these mutations are located in CGG triplets, motifs recognized by
the GAL4, PUT3 and PPR1 activators. Moreover, all other mutants with
alterations in either triplet showed reduced
in vitro
binding and
in vivo
activation. For example, mutant 1-2, with the first CGG triplet changed to CAC, and mutant 1-28, with a change to GGG, showed reduced
in vitro
binding and
in vivo
activation. Similar results were observed with mutants 1-1, 1-2, 1-6, 1-7, 1-18, 1-22, 1-41 and 1-42, containing nucleotide
alterations in either CGG triplet (Table
1
). These results suggest that two CGG triplets are important for binding of HAP1
to the UAS of
CYC1
.
Interestingly, another class of mutants was isolated with mutations outside the
CGG triplets. In the case of mutant 1-3, which carries five mutations, binding was not affected, but activation
was 15 times lower than wild-type
CYC1
. Similar results were seen with mutant 1-26. The properties observed at this UAS are analogous to those seen for
HAP1 positive control mutants at the wild-type UAS of
CYC7
, i.e. wild-type
in vitro
binding but no transcriptional activation (
21
).
Other mutants (1-10, 1-32, 1-40 and 1-43) with changes outside the triplets showed wild-type binding properties, as well as significant [beta]-galactosidase activity, suggesting
that many nucleotides outside the CGG triplets are not important for binding of
HAP1.
Two mutants (1-8 and 1-16; Table
1
) of the UAS of
CYC1
showed increased transcriptional activity (3-fold). Mutant 1-8 carries a single nucleotide change 5 nt downstream of the second
CGG triplet and mutant 1-16 has 4 nt alterations, including that found in 1-8. In order to rule out the possibility that these UASs could be
bound by an activator other than HAP1 we measured their activity with a HAP1
mutant that is transcriptionally inactive (SD5-HAP1[Delta]Kp-HIS; see Materials and Methods). Activity of the mutants was
reduced to background levels (0.2 U [beta]-galactosidase activity; data not shown), indicating that their
activity was dependent on HAP1.
The UAS of
CYC7
has two CGC triplets instead of the two CGGs found at a similar position in
CYC1
. Deletion of one G in one triplet (mutant 7-25) had drastic effects on HAP1 activity (8-fold reduction in activity; Table
2
). Some other mutants with changes in these triplets also had decreased binding
in vitro
and activation
in vivo
. For example, mutant 7-20, with a change of the first CGC triplet to CTC, had 6-fold less [beta]-galactosidase activity (Table
2
). Similarly, 7-30, which had reduced activity, carries a mutation in the second CGC
triplet (along with two other changes). Interestingly, changing one CGC to CGG
(7-24) increased the
in vivo
activity ~4-fold. A single nucleotide change between the two CGC triplets (7-5) reduced
in vitro
binding ~10-fold. Many mutations only minimally affected binding and activation
at
CYC7
. For instance, mutant 7-1 with 3 nt changes before the first CGC triplet, had wild-type activity. Similar results were obtained with mutants 7-2, 7-3, 7-8 and 7-34. All of these alterations fall outside
the CGC repeats.
We then tested the affinity of various key mutants by competition assays, as
shown in Figure
3
. As expected from previous results (
20
), the UASs of
CYC1
and
CYC7
had a similar affinity for HAP1. Higher
in vivo
activity of mutants 1-8 and 7-24 was correlated with an increased affinity of HAP1 for 7-24 and, to a lesser extent, for 1-8, as compared with the UAS of
CYC1
. Conversely, 1-9, which had reduced
in vitro
binding, did not compete with the wild-type probe
CYC1
, even at a 200 molar excess (Fig.
3
). Similarly, 7-5 showed reduced activity and affinity. On the other hand, mutants 1-3, 1-26 and 7I (Table
3
) showed increased affinity for HAP1 as compared with the wild-type UAS of
CYC1
, but were transcriptionally inactive.
Based on the above results and comparison with other UASs of HAP1 (see Fig.
2
), we concentrated on specific bases within the CGG/C motifs, as well as the
downstream TA motifs, and made other alterations using site-directed mutagenesis, as shown in Table
3
. First, we changed either (or both) CGG triplet(s) in
CYC1
to the sequence found at similar positions in
CYC7
: CGC. This resulted in a reduction in binding and activation (10-fold) (mutants 1A, 1B and 1C; Table
3
). Conversely, changing CGC to CGG in
CYC7
resulted in increased activity (mutants 7D, 7E and 7F; Table
3
). In addition, any alteration of the TA nucleotides in
CYC7
reduced binding and activation. For example, changing 1 nt between the two CGC
triplets (mutant 7K; Table
3
) had drastic effects on binding of HAP1
in vitro
and activation. Similar results were obtained with mutants 7H, 7I and 7J. At
similar positions the
CYC1
site has TT and AT sequences (Table
3
). Mutant 1G (TT -> TA) showed increased activity, as did 1F (AT -> AA). However, an AT -> TT change had a negative effect (mutant 1E). Finally, mutating
the
CYC1
site to match the CGC triplets and the TA repeats found in
CYC7
(mutant 1D; Table
3
) gave levels of activation that were 30% of the UAS of
CYC7
, indicating a secondary role for flanking sequences. From these data it appears
that HAP1 can recognize a CGG repeat and also a CGC repeat, although less
efficiently. In addition, TA repeats are also important for HAP1 binding. Their
presence (and probably that of some other nucleotides, as suggested by mutant
1D) seems to be more important when HAP1 is bound to the weaker CGC motif of
CYC7
. We propose that the `optimal' UAS for HAP1 is the direct repeat CGG N
3
TA N CGG N
3
TA.
Table 3
Since a major difference between the
CYC1
and
CYC7
sites is the sequence of the two triplets, we tested the possibility that they
are be responsible for the hyperactivity of HAP1-18 and the PC phenotype of HAP1 mutants, i.e. a defect in activation but
not binding at
CYC7
. As expected (
21
), HAP1-PC1 had wild-type activity at
CYC1
and gave background activity at
CYC7
(Table
3
), while HAP1-18 shows the opposite pattern of activity (Table
3
). Changing either CGC triplet (or both) of
CYC7
to CGG resulted in increased activity of HAP1-PC1 and reduced activity of HAP1-18 (mutants 7D, 7E and 7F; Table
3
). Similar results were observed with another PC mutant, HAP1-PC2 (data not shown). In addition, activation by HAP1-PC1 was greater when the second CGC triplet was changed to CGG as
compared with the first one. The opposite pattern was seen with HAP1-18, where some activity was retained with a mutation in the first CGC
triplet (mutant 7D), while HAP1-18 was transcriptionally inactive with a mutation in the second CGG
triplet (7F). The increased activity (>50-fold) of HAP1-PC1 at 7E and 7F (relative to wild-type
CYC7
) cannot be explained by greater binding, since only a modest increase in
binding (<2-fold) was observed with HAP1-PC1 tested with 7E and 7F probes in EMSA (data not shown). These
results show that the effect of HAP1 PC mutations can be suppressed by mutating
1 or 2 nt of the
CYC7
site. Conversely, activity of HAP1-PC1 was abolished when the CGG triplets of
CYC1
were mutated to CGC (mutants 1B and 1D). It is possible that mutants of HAP1
adopt a different conformation according to the site they are bound to; this
would result in different transcriptional activity (see Discussion).
Comparisons with other known UASs of HAP1 are shown in Figure
2
(see also Table
4
). All these sites have, in the middle of the UAS, the sequence `TA N CGG',
which matches the `optimal' sequence. However, instead of CGG as the first
triplet, the sequences TGG and AGG are found in
CTT1
and
CYB2
respectively. In addition, the sequences TT and GC are found downstream of the
second CGG triplet for
CTT1
and
CYB2
respectively, as compared with TA for the `optimal' sequence. If HAP1 binds
similarly to these sites, then changes that would allow a better fit with the
`optimal' sequence should result in higher activity. Mutating the first triplet
of the UAS of
CTT1
from TGG to CGG resulted in 7-fold higher activity (Table
4
). Similarly, changing AGG of
CYB2
to CGG also increased the [beta]-galactosidase activity >5-fold. Increases were also observed when nucleotides were
mutated at a position equivalent to the second TA of
CYC7
. However, these changes resulted in a more modest increase (2-fold) in activity (Table
4
). Comparison of the UAS of
CYT1
with the optimal sequence shows that this site has the consensus sequence
except for the second TA, which is CC. The second TA appears to be less
important if other nucleotides match the `optimal sequence'. Taken together the
data show that all HAP1 binding sites are related sequences, i.e. they are
imperfect versions of an `optimal site' that is a direct repeat.
We have performed random mutagenesis of the HAP1 binding sites found in the
CYC1
and
CYC7
genes. The results show that HAP1 binds to a direct repeat with the `optimal'
sequence CGG N
3
TA N CGG N
3
TA. This is in contrast to other zinc binuclear cluster proteins, such as GAL4,
PPR1 and PUT3, which recognize palindromic sequences containing inverted CGG
triplets (
2
,
3
,
29
). HAP1 can accommodate some changes at these triplets. For instance, the UASs
of
CTT1
and that of
CYB2
have the sequence TGG N
6
CGG and AGG N
6
CGG, respectively (Fig.
2
). However, activation at these sites is increased 5- to 7-fold when TGG or AGG is mutated to CGG. Variation of the CGG
triplets is also seen for other zinc cluster proteins. For instance, one UAS
GAL4
has an AGG triplet instead of CGG (
30
).
Divergence from the `optimal' site is even greater in
CYC7
, where two CGC triplets are found instead of the two CGGs. Again, mutating
these triplets to CGG results in greater activation. This is in agreement with
studies performed on the native
CYC7
promoter (
31
). In addition, our mutational analysis shows that the 2 nt (TA) located 4 bp
downstream of the two CGG/C triplets are important for binding of HAP1.
Increased activity at the UAS of
CYC1
was observed when some nucleotides were mutated at positions equivalent to the
first or the second TA of
CYC
7 (Table
3
). However, similar changes resulted in a more modest increase in activity of
CYB2
and
CTT1
(Table
4
). Comparison of the UAS of
CYT1
(Fig.
2
) shows that this site has the consensus sequence except for the second TA,
which is CC. The second TA appears to be less critical if other nucleotides
match the `optimal' sequence. More drastic effects are seen when mutating the
second TA in
CYC7
. Since two CGC triplets instead of CGG are present at that site, it is likely
that the second TA sequence helps to stabilize interaction of HAP1 with that
UAS. Moreover, the
CYC7
site, as opposed to other HAP1 UASs, is an almost perfect direct repeat, but is
the only site that does not have at least one CGG triplet. Other nucleotides
must also be important for binding of HAP1 to that site, as suggested by a
mutant of the UAS of
CYC1
(mutant 1D; Table
3
). This mutant has the repeat CGC N
3
TA N CGC N
3
TA found in
CYC7
, but shows reduced binding and activation as compared with the wild-type UAS of
CYC7
. Therefore, the presence of two unfavorable CGC triplets in
CYC7
is compensated for by other nucleotides that form an almost perfect direct
repeat. Moreover, mutant 7E, which matches the `optimal' sequence, shows a
lower activity than mutant CTT1A, in agreement with a secondary role for other
nucleotides located outside the `optimal' sequence.
Taken together these results show that HAP1 binds to related DNA sequences. Our
mutational analysis is in agreement with recently published experiments (
32
) where random site selection was used to identify HAP1 binding sites. However,
in that study the second TA repeat was not identified as being important for
HAP1 binding. Our results show that this second repeat is important for
increased activity at
CYC1
,
CTT1
and
CYB2
, and essential for activity at
CYC7
.
All of the known target sites for GAL4 or the
Kluyveromyces lactis
homolog LAC9 are palindromic (see
4
,
29
,
33
for a compilation of the UASs). Conversely, all the known binding sites for
HAP1 are imperfect direct repeats (
14
-
19
). In addition, no binding of HAP1 could be detected with various palindromic
sequences derived from the consensus UAS
GAL4
with spacing between the two triplets varying from 1 to 14 bp (unpublished
results). This is in agreement with the random site selection, where only
sequences with direct repeats were recovered (
32
). In addition, changing the orientation of the second CGG triplet to generate a
palindromic sequence prevents binding of HAP1 (
32
). These observations suggest that there are constraints that prevent HAP1 from
binding to a palindromic sequence and GAL4 from binding to a direct repeat. It
has been suggested that HAP1 binds to a direct repeat through swiveling of one
DNA binding domain relative to the dimerization domain (
32
). However, if the DNA binding domain of HAP1 shows such flexibility, HAP1
should also be able to bind to a palindromic sequence. One possibility is that
the relatively long N-terminal segment of HAP1 that precedes its zinc finger (63 amino acids as
compared with 10 for GAL4) prevents HAP1 from binding to a palindromic
sequence. Orientation and spacing of the CGG triplets appear to be major
determinants for the binding specificity of a given zinc cluster protein. This
contrasts with nuclear receptors, which show more flexibility for their binding
sites (
34
-
38
). Thus our results show that all HAP1 sites are related and are imperfect
direct repeats with the optimal sequence CGG N
3
TA N CGG N
3
TA.
Some mutants do not show a correlation between
in vitro
binding and
in vivo
activation. For instance, in the case of mutant 1-3 (Table
1
), which carries five mutations, binding is not affected, but activation is 15
times lower than wild-type
CYC1
. Similarly, mutant 1-26, with one alteration 3 nt downstream of the first CGG triplet, shows
greatly reduced activation. Affinity of these UASs for HAP1 was shown to be
equivalent to that of
CYC1
by competition assay (Fig.
3
). Similar results were observed for nuclear receptors, where a heterodimer
formed of LXR/RXR binds to various sites but activates only a certain subset. (
39
).
In addition, the properties observed at these UASs are analogous to those seen
for HAP1 PC mutants (which carry amino acid changes in the DNA binding domain),
i.e. wild-type
in vitro
binding but no transcriptional activation (
21
; Table
3
). For instance, HAP1-PC1 shows background activity at
CYC7
, while binding of HAP1-PC1 to
CYC7
was shown to be stronger than to wild-type
CYC7
(
21
). However, when either CGC triplet is changed to CGG increased activity was
observed (Table
3
). Binding of HAP1-PC1 to
CYC7
was shown to be stronger than wild-type
CYC7
(
21
). In contrast to PC mutants, HAP1-18 shows a dramatic increase in activation at the UAS of
CYC7
(~16 times more than wild-type HAP1), even though it has a similar affinity for the UAS of
CYC7
as compared with wild-type HAP1 (
20
). Introduction of CGG triplets decreases the activity of HAP1-18 as opposed to the effect seen for PC mutants, as suggested from studies
with the intact
CYC7
promoter (
31
).
We proposed (
21
) that the phenotype of the PC mutants could be explained by the fact that these
mutants would prevent interaction of the DNA binding domain of HAP1 with a
cofactor protein that would normally act in synergy with the activation domain
of HAP1. The model was based on genetic evidence that suggested that a mutant
of GAL11, GAL11P, increases the activity of a GAL4 mutant by interacting with
its DNA binding domain (
40
), which contains a zinc finger homologous to that of HAP1. However, more recent
data showed that wild-type GAL11, unlike GAL11P, does not interact with GAL4 and is, rather, a
component of the RNA polymerase II holoenzyme (
41
-
42
). Therefore, HAP1-PC mutants may simply have an altered conformation that may inhibit the
activation domain. Changing the CGC triplets to CGG would alter the
conformation of PC mutants, enabling them to be active, as seen when bound to
the
CYC1
site, which contains two CGG triplets. Conversely, HAP1-18 could have an alternate conformation that would allow very efficient
interaction with some components of the basic transcriptional machinery. A
similar model has been suggested to explain the phenotype of PC mutants of the
glucocorticoid receptor (
43
), where it is proposed that DNA can act as an allosteric effector.
In conclusion, two different types of mutations modulate the activity of HAP1
without changing its affinity for target sites. Firstly, mutations in the DNA
binding domain of HAP1 can lead to increased (HAP1-18) or decreased (PC mutants) activity at
CYC7
. Secondly, mutations in the target sites for HAP1 also affect activation of
HAP1 (or mutants). It will be interesting to determine if these two types of
mutations have an allosteric effect on the activation domain of HAP1.
We are grateful to Andrew Nice for help in sequencing mutants. We thank Caterina
Russo for synthesis of the spiked oligonucleotides. We are grateful to Drs John
White and Hans Zingg for critical reading of the manuscript. We also thank
members of the Laboratory of Molecular Endocrinology for helpful suggestions.
This work was supported by grants from the Medical Research Council of Canada
and the Fonds de la Recherche en Santé du Québec. NH is the recipient of a FCAR (Quebec government)
studentship. BT is a scholar of the Medical Research Council of Canada.
HAP1 expression vectors.
A
HIS4
version of SD5-HAP1 (
20
) was constructed by deleting the
URA3
gene of SD5-HAP1 by cutting with
Stu
I, adding
Not
I linkers and ligating a
Not
I fragment (containing the
HIS4
gene) from CYC7-5lacZ-HIS4 (
21
). SD5-HAP1-PC1-HIS, SD5-HAP1-PC2-HIS and SD5-HAP1-18- HIS were constructed by
introducing
Dra
III fragments (containing the HAP1 mutations) from
URA3
marked expression vectors (
21
) into
Dra
III-cut SD5-HAP1-HIS4. HAP1 expression vector deleted of its activation domain
(SD5-HAP1[Delta]Kp-HIS) was constructed by linearizing SD5-HAP1-HIS4 with
Kpn
I, treating with T4 DNA polymerase and inserting a
Xba
I linker (New England Biolabs) containing nonsense codons in the three reading
frames.
Reporter plasmid.
Plasmid pLG178-M was constructed by destroying the unique
Mun
I site in the 2[mu] origin of replication of pSLF[Delta]178K (
26
) by cutting with
Mun
I, filling in with Klenow fragment and ligating. pLG178-M was linearized with
Xho
I and an oligonucleotide (5'-TCGAGAGATCTAAAAAACAATTGC-3') and its complement were inserted at that site to
give p178MB. This plasmid has unique
Mun
I and
Bgl
II sites (flanked by
Xho
I sites) in front of a minimal
CYC1
promoter driving
lacZ
transcription (see Fig.
1
).
REFERENCES
Return