ABSTRACT
The
Saccharomyces cerevisiae IRE1
gene, encoding a putative receptor-type protein kinase, is known to be required for inositol prototrophy and
for the induction of a chaperon molecule, BiP, encoded by
KAR2
, under stress conditions such as tunicamycin addition. We have characterized a
yeast gene,
IRE2
, which was isolated as a suppressor gene that complements the inositol
auxotrophic phenotype of the
ire1
mutation. Sequencing analysis revealed that
IRE2
is identical to
HAC1
, which encodes a transcription factor having a basic-leucine zipper motif. Introduction of
IRE2/HAC1
into the
ire1
mutant clearly restored the expression of
KAR2
upon tunicamycin treatment.
ire2/hac1
-disrupted yeast cells showed not only the inositol auxotrophic phenotype but also
the tunicamycin sensitivity, and failed to induce the expression of
KAR2
. These results clearly indicate that the
IRE2/HAC1
gene product plays a critical role in the induction of
KAR2
expression and in the inositol prototrophy mediated by
IRE1
.
In our previous study,
IRE1
was identified as the gene required for the
myo
-inositol (inositol) prototrophy of
Saccharomyces
cerevisiae
(
1
). The predicted gene product shows sequence similarity to receptor kinases,
such as the epidermal growth factor receptor, having a putative signal
sequence, an N-terminal ligand binding domain, a transmembrane domain and a C-terminal kinase domain.
IRE1
has also been identified as the gene required for the transcriptional induction
of
KAR2
(
2
,
3
).
KAR2
encodes a protein chaperon, BiP, which is required for protein folding in the
endoplasmic reticulum (ER), as in higher eukaryotic cells. The expression of
KAR2
is induced by a variety of treatments, such as the addition of tunicamycin,
that causes the accumulation of unfolded proteins in the ER (
4
). Mutants having a defect in
IRE1
are unable to induce the transcription of
KAR2
, resulting in an inability of growth on tunicamycin-containing media. It has also been shown that a mammalian cell line
harboring multi-copies of the yeast
IRE1
gene exhibits modest increases in the basal and drug-induced levels of mammalian BiP (
5
). However, the exact function of
IRE1
in
KAR2
expression and the cause of the inositol auxotrophy of
ire1
mutants are unclear. Furthermore, no transcription factor involved in
KAR2
expression is known.
In our previous study, we also isolated a gene which could suppress the
ire1
mutation (
1
). However, the nature of this gene, designated as
IRE2
, was not studied in detail. Since the
IRE2
gene product could be involved in
IRE1
-mediated signal transduction, we attempted to characterize it. Here we
show that
IRE2
is identical to
HAC1
. The
HAC1
gene was originally isolated by Nojima
et al.
as a cDNA of
S.cerevisiae
which can suppress the
Schizosaccharomyces
pombe
cdc10
mutation (
6
). They reported that the gene product of
HAC1
had a basic leucine zipper (bZIP) motif, showing local sequence similarity to
the mammalian cAMP response element binding protein (CREB) and its yeast
homologue, the
SKO1/ACR1
gene product. They also showed that the
HAC1
protein expressed in
Escherichia coli
could bind to a cAMP response element (CRE)
in vitro
, and the disruption of
HAC1
rendered yeast cells caffeine sensitive. However, the physiological role of
HAC1
in
S.cerevisiae
has not been elucidated yet. We present evidence that
HAC1
is involved in
IRE1
-mediated
KAR2
gene expression in
S.cerevisiae
.
Saccharomyces cerevisiae
strain D452-2 (
MAT
[alpha]
leu2 ura3 his3
) was used as the wild-type strain (
7
). Strain SFY526, for the two-hybrid experiments, was obtained from Clonetech Laboratory Inc. (California). Yeast cells were grown aerobically with shaking at 30oC. The composition of inositol-free minimum medium M-i was as described previously (
8
). Inositol, L-leucine, L-histidine and uracil were each added to the culture medium at a concentration of 20 [mu]g/ml. Tunicamycin was used at a concentration of 1 [mu]g/ml.
Escherichia coli
JM109 was used for the amplification of all plasmids and M13 recombinant
phages. Bacteria were cultured in Luria broth at 37oC (
9
). Ampicillin was used at the concentration of 50 [mu]g/ml.
Escherichia coli
plasmids were prepared by the alkaline lysis method (
9
). Yeast transformation was carried out by the lithium acetate method (
10
). The transformation of
E.coli
was performed by the standard method (
9
).
Plasmids pIR2, pIR3 and pIR4 (formerly pIRE2, pIRE3 and pIRE4, respectively)
were described previously (
1
). YEpM4 is a 2[mu]m DNA-based shuttle vector containing a multicloning site and the
LEU2
gene as an yeast selectable marker (
11
). Plasmid pIR3 was digested with
Xho
I plus
Sal
I and
Stu
I plus
Sma
I, separated by electrophoresis to remove the small fragment, and then self-ligated to yield pIR3[Delta]X and pIR3[Delta]St, respectively. Plasmid pIR3[Delta]Sp was constructed by inserting the 5 kb
Sph
I fragment of pIR3 into the
Sph
I site of the YEpM4 vector. pIR3[Delta]S was constructed as follows. Plasmid pIR3 was digested with
Sal
I and the resulting fragments were separated by electrophoresis. Fragments of ~0.4 and 14 kb were ligated, and the plasmid, pIR3[Delta]S, having the same orientation as the original pIR3, was selected by
restriction enzyme analysis.
Plasmids pBT-HAC1 and pAD-HAC1, for the two-hybrid experiments, were constructed as follows. The 0.8 kb
Spe
I-
Hin
dIII fragment of plasmid pIR3 was ligated between the
Sma
I and
Hin
dIII sites of pUC18. The plasmid, pUC-IRE2SH, thus obtained was digested with
Hin
dIII, treated with Klenow large fragment, and then digested with
Eco
RI. An ~0.8 kb fragment was isolated by electrophoresis, and inserted between the
Sma
I and
Eco
RI sites of plasmids pGBT9 and pGAD424 to yield plasmids pBT-HAC1 and pAD-HAC1, respectively. Plasmids pGBT9 and pGAD424 were purchased from
Clonetech Laboratory Inc.
pUC-ire1::URA3, for construction of the
ire1
disruptant, was prepared by the same method as described previously except that
the
URA3
gene was used instead of the
HIS3
gene (
1
). The 2 kb
Eco
RI-
Hin
dIII fragment of plasmid pUC-ire1::URA3 was used for replacement of the chromosomal
IRE1
gene of strain D452-2.
For construction of the
ire2/hac1
disruptant, a 1.6 kb
Eco
RV-
Hin
dIII DNA fragment harboring the
IRE2/HAC1
gene was inserted between the
Sma
I and
Hin
dIII sites of pUC19. The plasmid, pUC-IRE2, thus obtained was digested with
Pst
I and treated with Klenow fragment, the small fragment was removed by
electrophoresis, and then the plasmid was ligated with the
Hin
dIII fragment of the
URA3
gene, which had been treated with the Klenow fragment. The resulting plasmid,
pUC-ire2::URA3, was digested with
Xho
I and
Hin
dIII, and used for replacement of the chromosomal
IRE2/HAC1
gene of yeast haploid strain D452-2. Gene replacement in the
ire1
and
ire2/hac1
disruptants thus obtained (YF4 and HU1, respectively) was confirmed by genomic
Southern hybridization.
DNA sequencing was carried out by the dideoxy termination method of Sanger
et al
. (
12
) using a universal primer or synthetic oligonucleotides for part of the
sequence of
IRE2/HAC1
.
For Northern blot analysis, total RNA was isolated from yeast cells as described
previously (
13
). Samples were subjected to electrophoresis on a 1% agarose gel containing
formaldehyde, blotted onto a Biodyne A membrane (Nihon Pole, Tokyo), and then
hybridized with a
32
P-labelled probe. A 2 kb
Xba
I fragment containing the
KAR2
coding region and a part of the 5' flanking region was excised from plasmid pSVYB-1 and used as a probe. Hybridization and detection were carried out
according to the manufacturer's manual.
[beta]-Galactosidase activity was determined according to the
manufacturer's manual (Clonetech Laboratory Inc.) and expressed in Miller units
(
14
).
[[alpha]-
32
P]dCTP was purchased from Amersham International plc (Buckinghamshire).
Restriction endonucleases, T4 DNA ligase and the Klenow fragment of DNA
polymerase were obtained from Takara Shuzo (Kyoto). Tunicamycin was purchased
from Wako Chemicals (Osaka).
Plasmid pIR3 harboring
IRE2
was originally obtained as a suppressor for the inositol auxotrophic phenotype
of
ire1
mutant D437-1B, which was isolated by mutagenesis with ethyl methanesulfonate (
1
). Therefore, we first examined whether or not the
IRE2
gene could suppress the
ire1
null mutation. Strain YF4 is a null mutant, in which three-quarters of the
IRE1
coding region was deleted and replaced with the yeast
URA3
gene. Strain YF4 was transformed with pIR3 as well as the vector plasmid and
pIR2 harboring the
IRE1
gene. The growth phenotype of the transformants was determined on inositol-free medium. As shown in Figure
1
, introduction of
IRE2
, as well as
IRE1
, into the
ire1
disruptant clearly reversed the growth defect of the mutant. The growth defect
of the mutant was not suppressed by introducing a single-copy of
IRE2
on a centromere-containing vector (data not shown). These results indicate that multi-copies of
IRE2
can bypass the defect in the
IRE1
function.
ORF YFL031w is also known as
HAC1
(accession number D26506), which was originally reported by Nojima
et al.
(
6
). They isolated the cDNA of this gene as a suppressor of the
S.pombe
cdc10
mutant. The entire sequence of cDNA as well as that of its 5'-regulatory region were reported. However, we found that the DNA
sequences of
HAC1
and YFL031w (same as our sequence) differ by 6 nt (insertion, deletion and
alteration), resulting in a difference in their predicted amino acid sequences.
All the differences are localized in the C-terminal region of the predicted protein. Therefore, we determined whether
or not the sequence difference might be important for the ability of
IRE2/HAC1
to suppression of the
ire1
mutation. We constructed plasmid pIR3[Delta]S, which has a truncated form of
IRE2/HAC1
and thus lacks 26 amino acids of the C-terminus in its gene product. Plasmid pIR3[Delta]S was introduced into strain YF4 and the growth phenotype of the
transformants on inositol-free and inositol-containing media was determined. As shown in Figure
3
, the truncated form of the
IRE2/HAC1
gene product clearly suppressed the inositol auxotrophic phenotype of the
ire1
null mutant. Therefore, the C-terminus of the
IRE2/HAC1
gene product was found not to be essential for suppression of the
ire1
mutation.
Yeast mutants having the
ire1
mutation are not only inositol auxotrophic but also defective in
KAR2
induction, showing the growth defect under stress conditions such as
tunicamycin addition. Since we isolated the
IRE2/HAC1
gene as a suppressor for the inositol auxotrophic phenotype of the
ire1
mutant, we wondered whether or not
IRE2/HAC1
could suppress the tunicamycin sensitivity of
ire1
mutants. We introduced the
IRE2/HAC1
gene in multi-copies into the
ire1
-disrupted strain, YF4, and then its growth on tunicamycin-containing medium was determined. As shown in Figure
3
, plasmid pIR3 clearly restored the tunicamycin sensitivity of the
ire1
-disrupted strain. Plasmid pIR3[Delta]S also suppressed the tunicamycin sensitivity, though the
suppression was slightly weaker than that of the full-size
IRE2/HAC1
. These results strongly suggested that
IRE2/HAC1
might be involved in the
IRE1
-mediated
KAR2
induction.
To determine the effect of
IRE2/HAC1
on the induction of
KAR2
more directly, we next determined the mRNA level of
KAR2
by Northern blot analysis. The
ire1
-disrupted strain harboring either
IRE2/HAC1
or vector plasmid YEpM4 was cultured in the presence of tunicamycin for 3 h.
Total RNA was isolated and used for the determination of
KAR2
mRNA. As shown in Figure
4
A, the
ire1
-disrupted strain could not induce
KAR2
mRNA expression (lane 2), compared with the wild-type strain (lane 1). On the other hand, the introduction of
IRE2/HAC1
(lane 4), as well as
IRE1
itself (lane 3), clearly restored the induction of
KAR2
mRNA, although the extent was slightly lower for
IRE2/HAC1
.
Nojima
et al.
reported that a
HAC1
-disrupted strain is not lethal but shows caffeine sensitivity. They also
showed that this caffeine sensitivity is overcome on introduction of the gene
for high affinity cAMP phosphodiesterase, suggesting that the growth inhibition
of the mutant is caused by a high level of cAMP. However, the detailed
mechanism is not known yet. Since we showed above that
IRE2/HAC1
could suppress the inositol auxotrophic phenotype and the tunicamycin
sensitivity of the
ire1
-disrupted strain, we next determined the growth phenotype of the
ire2/hac1
-disrupted strain on inositol-free medium and tunicamycin-containing medium. For this purpose, we constructed an
ire2/hac1
-disrupted strain, of which two-thirds of the coding region was replaced with the yeast
URA3
gene. As expected, the disrupted strain, HU1, thus obtained showed the growth
defect on inositol-free medium and on tunicamycin-containing medium (Fig.
5
). Furthermore, the growth defect of the
ire2/hac1
-disrupted strain was not suppressed by introducing the multi-copies of
IRE1
into the mutant (data not shown). These results strongly suggest that
IRE2/HAC1
acts downstream of
IRE1
and is necessary for the induction of
KAR2
under stress conditions. To clarify this point, we next determined the changes
in the mRNA level of
KAR2
in the
ire2/hac1
-disrupted strain upon the addition of tunicamycin. The
ire2/hac1
-disrupted strain, HU1, as well as the wild-type strain, D452-2, were cultured in the presence of tunicamycin, and then the
KAR2
mRNA levels were determined by Northern blot analysis. As shown in Figure
4
B, the expression of
KAR2
was rapidly induced on the addition of tunicamycin in the wild-type cells (lanes 1-3), whereas no induction of
KAR2
mRNA was observed in the disruptant cells (lanes 4-6). These results together with those described above clearly indicate
that the
IRE2/HAC1
gene product is essential for
KAR2
induction mediated by
IRE1
under stress conditions.
It has been shown that the
HAC1
gene product exhibits sequence similarity to CREB and has a bZIP motif (
6
). Since CREB, like other proteins having a bZIP motif, are known to form a
homodimer to bind CRE (
15
), and the
HAC1
gene product expressed in
E.coli
could bind to a DNA fragment having a CRE motif (
6
), we speculated that the
IRE2/HAC1
gene product could form a homodimer. To examine this possibility we used the
yeast two-hybrid system (
16
). We constructed chimeric plasmids in which the
IRE2/HAC1
gene was fused to the gene for the
GAL4
DNA binding domain or activating domain. The resultant plasmids, pBT-HAC1 and pAD-HAC1, were introduced into the tester strain, SFY526. Figure
6
shows the [beta]-galactosidase activities of the transformants which contained pBT-HAC1 or pAD-HAC1, or both. The transformant harboring pBT-HAC1 or pAD-HAC1 alone did not exhibit [beta]-galactosidase activity. On the
other hand, the transformant harboring the combination of pBT-HAC1 and pAD-HAC1 clearly showed a significant level of [beta]-galactosidase activity, indicating that the
IRE2/HAC1
gene product can form a homodimer
in vivo
.
At present we do not know the targets of the
IRE2/HAC1
gene product, but
KAR2
might be one of the target genes. The
KAR2
promoter region has been analyzed and shown to contain a GC-rich region, heat shock element and unfolded protein response element (
17
,
18
). One of these elements might be a target region for the binding of the
IRE2/HAC1
gene product. Otherwise, the gene product of
IRE2/HAC1
might be a transcription factor for the protein which regulates the expression
of
KAR2
under stress conditions. To clarify this point we are now investigating the
correlation between
IRE2/HAC1
and
KAR2
. In any case, our present study showed that the
IRE2/HAC1
gene product plays a critical role in the induction of
KAR2
expression mediated by
IRE1
.
We would like to thank Drs Kenji Kohno and Masao Tokunaga for providing plasmid
pSVYB-1, and Dr Hiroshi Nojima for the helpful discussions. This work was
supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science
and Culture of Japan.
REFERENCES
Return