ABSTRACT
Recessive mutations in the
SSU71
,
SSU72
and
SSU73
genes of
Saccharomyces cerevisiae
were identified as either suppressors or enhancers of a TFIIB defect (
sua7-1
) that confers both a cold-sensitive growth phenotype and a downstream shift in transcription start
site selection. The
SSU71
(
TFG1
) gene encodes the largest subunit of TFIIF and
SSU72
encodes a novel protein that is essential for cell viability. Here we report
that
SSU73
is identical to
RPB9
, the gene encoding the 14.2 kDa subunit of RNA polymerase II. The
ssu73-1
suppressor compensates for both the growth defect and the downstream shift in
start site selection associated with
sua7-1
. These effects are similar to those of the
ssu71-1
suppressor and distinct from the
ssu72-1
enhancer. The
ssu73-1
allele was retrieved and sequenced, revealing a nonsense mutation at codon 107. Consequently,
ssu73-1
encodes a truncated form of Rpb9 lacking the C-terminal 16 amino acids. This Rpb9 derivative retains at least partial
function since the
ssu73-1
mutant exhibits none of the growth defects associated with
rpb9
null mutants. However, in a
SUA7
+
background,
ssu73-1
confers the same upstream shift at
ADH1
as an
rpb9
null allele. This suggests that the C-terminus of Rpb9 functions in start site selection and demonstrates that
the previously observed effects of
rpb9
mutations on start site selection are not necessarily due to complete loss of
function. These results establish a functional interaction between TFIIB and
the Rpb9 subunit of RNA polymerase II and suggest that these two components of
the preinitiation complex interact during transcription start site selection.
Accurate initiation of transcription by RNA polymerase II (RNAP II) requires
several factors in addition to RNAP II (
1
). These general transcription factors (GTFs) include TBP (TATA binding
protein), TFIIB, TFIIE, TFIIF and TFIIH (reviewed in
2
,
3
). The GTFs assemble with RNAP II in a defined order on a DNA template to
generate a preinitiation complex (PIC) that is sufficient for accurate, basal
level initiation
in vitro
(
4
). Although step by step assembly of the PIC has been demonstrated by several
laboratories, the recent identification of RNAP II holoenzyme complexes that
include GTFs from both yeast and mammalian cells suggests that RNAP II binds
promoters
in vivo
as a pre-assembled or partially assembled complex (
5
-
7
). Several of the GTFs have been shown to be dispensable for accurate initiation
in vitro
, depending upon the structure of the promoter (
8
-
12
). The minimum requirements for accurate initiation from a negatively
supercoiled template are TBP, TFIIB and RNAP II, implying that TBP and TFIIB
are sufficient to position RNAP II at the transcription start site (
8
). TFIIB interacts directly with RNAP II (
13
,
14
), as well as with the TBP-TATA complex, binding DNA both upstream and downstream of TATA (
15
,
16
). These results are consistent with the suggestion that TFIIB forms a bridge
between the TATA element and the initiator region (
4
).
Despite identification of the factors required for accurate initiation, the
mechanism by which RNAP II selects start sites is unclear. In higher eukaryotes initiation generally occurs at a single start site
located ~30 bp from the TATA element, suggesting that start site selection might be defined by a fixed distance from TATA. However, in the yeast
Saccharomyces cerevisiae
transcription often occurs at multiple sites within a broad window located 40-120 bp from TATA (
17
,
18
). Nonetheless, the TATA element in
Saccharomyces cerevisiae
defines the window within which initiation can occur (
19
-
23
).
Genetic analysis of transcription initiation has identified components of the
transcriptional machinery that are important for accurate start site selection
in
S.cerevisiae
. The gene encoding yeast TFIIB (
SUA7
) was first identified based on the ability of
sua7
mutations to shift start site selection downstream of normal (
24
). Similar effects on start site selection were conferred by
sua8
mutations, which are allelic to
RPB1
and encode altered forms of the largest subunit (Rpb1) of RNAP II (
25
). Mutations in the
RPB9
gene, which encodes the 14.2 kDa subunit (Rpb9) of RNAP II, can also affect
start site selection, in these cases by shifting transcription initiation
upstream of normal (
26
,
27
). Mutations in
RPB1
can also shift initiation upstream of normal (
28
), although the reported effects are more subtle than either the downstream
shifts associated with
sua7
and
sua8
, or the upstream shifts associated with
rpb9
. These results demonstrate that TFIIB and the Rpb1 and Rpb9 subunits of RNAP II
are important determinants of transcription start site selection
in vivo
and suggest that accurate initiation involves specific interactions between
TFIIB, Rpb1 and Rpb9. Furthermore, these conclusions are consistent with results demonstrating that TFIIB and RNAP II are both necessary and sufficient for accurate start site selection in a yeast
in vitro
transcription system (
29
).
In an effort to identify factors that functionally interact with TFIIB, we are
isolating and characterizing suppressors of specific TFIIB defects. The
sua7-1
mutation encodes a Glu-62 -> Lys (E62K) replacement that is responsible for both aberrant start
site selection and a cold-sensitive (Csm
-
) growth defect (
30
). By selecting for revertants of the
sua7-1
Csm
-
phenotype, we have identified three genes designated
ssu71
,
ssu72
and
ssu73
. The
ssu71
(
TFG1
) gene encodes the largest subunit of TFIIF and the
ssu71-1
suppressor compensates not only for the Csm
-
phenotype of
sua7-1
, but restores the normal initiation pattern in the
sua7-1
background (
31
).
SSU72
is an essential gene encoding a novel factor that functionally interacts with
TFIIB (
32
). In contrast with
ssu71
and
ssu73
, the
ssu72-1
allele does not suppress
sua7-1
. Rather,
ssu72-1
, in combination with
sua7-1
, confers a heat-sensitive (Tsm
-
) growth defect and shifts start site selection further downstream of normal (
32
).
Here we define
ssu73-1
. This suppressor is similar to the
ssu71
suppressors in that it restores growth in the cold and compensates for the
downstream start site shift associated with
sua7-1
. Molecular and genetic experiments established that
ssu73-1
is allelic to
RPB9
. These results confirm and extend previous conclusions that the Rpb9 subunit of RNAP II plays an important role in start site
selection. Moreover, the genetic relationship between
sua7
and
ssu73
suggests that interaction between TFIIB and Rpb9 is an important determinant of accurate initiation
in vivo
.
The strains used in this study are listed in Table
1
. Strain YZS19 (Csm
+
Tsm
-
) was isolated as a spontaneous Csm
+
revertant of YMH71-9C (Csm
-
Tsm
+
). YZS19-3B (Csm
+
Tsm
+
) is a segregant of a diploid strain derived from a cross between YZS19 and YDW546 (Csm
-
Tsm
+
). YAT30 was constructed by introducing plasmid pM706 into strain YDW546. To stimulate integration of
RPB9
:
URA3
at the
RPB9
locus, pM706 was first linearized at the unique
Sna
BI site located within the
Pst
I-
Bam
HI fragment encompassing
RPB9
. YZS19 and YZS19-3B derivatives were constructed by introducing the indicated plasmids
(YCplac33, pDW11, pM586 or pM681) into the respective strains and selecting for
uracil prototrophy.
Standard procedures were used for crossing yeast strains, selecting diploids,
inducing sporulation and dissecting tetrads (
33
). Yeast transformations were done by the lithium acetate procedure (
34
). The symbols Csm
-
(cold-sensitive), Tsm
-
(heat-sensitive) and Slg
-
(slow growth) refer to distinctly impaired growth (or no growth) on rich medium (YPD) at 16, 37 and 30oC respectively. Ura
-
denotes failure to grow on -Ura omission medium; Flo
-
refers to a flocculant growth phenotype in liquid culture (YPD, 30oC).
Plasmids used in this study were constructed by standard recombinant DNA methods
(
35
). YCplac33 is a centromere-based
URA3
vector (
36
). pDW11 is a centromere-based
URA3
plasmid carrying the
SUA7
gene (
24
). pM586 carries the 1.05 kb
Eco
RI-
Hin
dIII DNA fragment encompassing
SSU72
in YCplac33 (
32
). pM681 was constructed by transferring the 3.2 kb
Pst
I-
Bam
HI fragment encompassing
RPB9
from plasmid Ro406 (
26
) into YCplac33. pM706 is an integrating plasmid derived by transferring the 3.2 kb
Pst
I-
Bam
HI
RPB9
fragment into the
URA3
integrating vector YIplac211 (
36
). Plasmids pM714 and pM715 carry the
Eco
RI-
Spe
I DNA fragment encompassing the
ssu73-1
open reading frame (amplified by the polymerase chain reaction) in the vectors pRS426 (
37
) and pRS314 (
38
) respectively.
Table 1
Primer extension was performed as described previously, using total RNA and the
ADH1
-specific primer oIP-87 (
24
). Primer extension products were resolved in a 6% polyacrylamide DNA sequencing
gel and visualized by autoradiography. Sequenced
SSU72
DNA was used as a molecular size marker.
DNA encompassing the
ssu73
allele was amplified from genomic DNA (strain YZS19) by the polymerase chain
reaction using primers that flank the
SSU73
open reading frame. Primers oAT-202 (5'-GG
The
sua7-1
allele encodes an E62K replacement in TFIIB that confers both altered start
site selection and a severe Csm
-
phenotype (
30
). 120 spontaneous Csm
+
revertants of YMH71-9C (
sua7-1
) were isolated on rich medium at 16oC. When scored for pleiotropic phenotypes, three Csm
+
revertants were found to be heat-sensitive (Tsm
-
), failing to grow at 37oC. Two of these strains, YZS14 and YZS45, were described previously and are
the result of mutations in the
SSU71
(
TFG1
) gene, which encodes the largest subunit of TFIIF (
31
). The third strain was designated YZS19 and is depicted in Figure
1
.
Although the
ssu73-1
allele confers multiple phenotypes, we were unable to identify a phenotype,
either associated with
ssu73
alone or with an
sua7
ssu73
double mutant, that could be used to select for the
SSU73
wild-type gene from a yeast genomic library. In an effort to avoid cloning
SSU73
by scoring a library of transformants for restoration of the Csm
-
phenotype, we first asked if
ssu73
might be allelic to previously cloned yeast genes encoding either general transcription factors or subunits of RNAP II. Since
ssu71
is allelic to
TFG1
, obvious candidates were
TFG2
and
TFG3
, the genes encoding the other two subunits of TFIIF (
39
). However, neither of these genes complemented
ssu73-1.
We also tested
RPB9
, not only because it encodes a subunit of RNAP II, but because alleles of
RPB9
have been reported to affect transcription start site selection (
26
,
27
).
To determine whether
RPB9
would complement
ssu73-1
, a low-copy-number plasmid carrying
RPB9
(pM681) was introduced into strain YZS19 (
sua7 ssu72 ssu73
). A Ura
+
transformant (YZS19/pM681) was subsequently scored for restoration of Csm
-
. Strain YZS19/pM681 exhibited the same Csm
-
phenotype associated with the
sua7-1
mutation, whereas control strains that had been transformed with plasmids
carrying
SUA7
(YZS19/pDW11),
SSU72
(YZS19/pM586) or vector only (YZS19/YCplac33) remained Csm
+
(Fig.
2
). Plasmid pM681 (
RPB9
) was also introduced into strain YZS19-3B (
sua7 ssu73
). The presence of
RPB9
restored the Csm
-
phenotype, elevated iso-1-cytochrome
c
levels to ~30% of normal, eliminated the flocculant phenotype associated with
ssu73-1
, and rendered the strain Slg
-
at 30oC (Table
2
). Therefore,
RPB9
fully complements all phenotypes associated with the
ssu73-1
mutation in the
sua7-1
background, suggesting that
ssu73-1
is allelic to
RPB9
.
In addition to its effect on growth at reduced temperature, the
sua7-1
mutation shifted transcription start site selection downstream of normal at the
ADH1
,
CYC1
and other genes (
24
,
25
). We therefore asked if
ssu73-1
affected start site selection. The
ADH1
gene was chosen for this study because both
sua7
and
rpb9
mutations are known to affect start site selection at
ADH1
. Transcription start sites at
ADH1
were mapped by primer extension and results are shown in Figure
3
. In a wild-type strain transcription initiates with equal efficiency at two principal
sites located 37 and 27 bp upstream of the ATG start codon (lane 1). Consistent
with previous results (
24
,
25
), the
sua7-1
mutation shifted initiation downstream of normal, rendering position -37 a minor site relative to -27 and enhancing initiation at sites downstream of -27 (lane 2). The
ssu73-1
mutation partially compensated for the
sua7-1
shift (lane 3), resulting in an initiation pattern intermediate between that of
the
sua7
mutant and the wild-type strain; this effect is most pronounced at position -37 (cf. lane 3 with lanes 2 and 5). In an
SUA7
+
background,
ssu73-1
diminished initiation at the minor sites downstream of -27, enhanced initiation at position -37 relative to -27, and, importantly, generated new sites upstream of -37 (lane 4). The start sites located upstream of -37 are identical to those reported previously
for an
rpb9
null mutant (
27
). These effects can be attributed specifically to
ssu73-1
because the downstream shift conferred by
sua7-1
(lane 2) is fully restored by plasmid-borne
RPB9
(
SSU73
) in the
sua7
ssu73
mutant (lane 5). Therefore, the
ssu73-1
mutation exerts an upstream shift on start site selection in both the
sua7-1
and
SUA7
+
backgrounds.
The
ssu73-1
allele of
RPB9
was cloned by amplification of
ssu73
genomic DNA from strain YZS19 by the polymerase chain reaction. Amplified DNA
was inserted into the vectors pRS426 and pRS314 and single-stranded DNA was isolated. DNA sequence analysis of the entire
ssu73-1
open reading frame revealed a single base-pair substitution encoding a premature stop codon (T
The
SSU71
,
SSU72
and
SSU73
genes were identified based on the ability of recessive mutations at these loci
to either suppress (
ssu71
,
ssu73
) or enhance (
ssu72
) a conditional growth defect associated with the altered form of TFIIB encoded
by the
sua7-1
allele.
SSU71
is identical to
TFG1
, the gene encoding the largest subunit of TFIIF (
31
).
SSU72
is a new gene encoding a novel protein that is essential for cell viability (
32
). Here we report that
SSU73
is identical to
RPB9
, the gene encoding the 14.2 kDa subunit of RNAP II. In addition to either
suppressing or enhancing the
sua7-1
conditional growth defects, all
ssu71
,
ssu72
and
ssu73
alleles affect transcription start site selection. Both
ssu71-1
and
ssu73-1
compensate for the downstream shift associated with
sua7-1
by at least partially restoring the normal initiation pattern at
CYC1
or
ADH1
(Fig.
3
; ref.
31
). In contrast,
ssu72-1
shifts initiation further downstream of normal, a result consistent with
identification of
ssu72-1
as an enhancer of
sua7-1
(
32
). These results imply functional interactions between TFIIB/Sua7 and
Ssu71/Tfg1, Ssu72 and Ssu73/Rpb9, and that these interactions are important for
accurate start site selection
in vivo
.
Although the
ssu71-1
and
ssu72-1
mutations exert clear effects on start site selection in the
sua7-1
background, neither allele appears to affect initiation in an
SUA7+
background (Z.-W.S. and M.H., unpublished results). This is in contrast with
ssu73-1
, which shifts initiation at
ADH1
upstream of normal in an
SUA7+
strain (Fig.
3
). This upstream shift is consistent with previously reported effects of
shi
/
rpb9
mutations on start site selection (
26
,
27
,
41
). Furthermore, the initiation pattern at
ADH1
reported here (Fig.
3
, lane 4) appears to be identical to that reported previously for an
rpb9
null mutant (
27
). Although
RPB9
is not essential for cell viability,
rpb9
null mutants are severely impaired, exhibiting Csm
-
, Tsm
-
and Slg
-
phenotypes (
40
). Since
ssu73-1
mutants remain Csm
+
, Tsm
+
and Slg
+
, the
ssu73-1
-encoded Rpb9 protein must be functional. Therefore, complete loss of Rpb9
function is not required for altered start site selection; rather an impaired
but functional form of Rpb9 is sufficient to shift initiation upstream of
normal.
The Rpb9 protein contains two Cys
4
zinc binding motifs, one near the N-terminus, the other near the C-terminus (Fig.
4
). The C-terminal Cys
4
motif is predicted to form a zinc ribbon, defined as a zinc binding domain
comprised of three antiparallel [beta]-sheets, based on sequence similarity to the Cys
4
domain of elongation factor TFIIS (
42
). The
ssu73-1
allele is the result of a nonsense mutation that encodes a truncated form of
Rpb9 lacking the C-terminal 16 amino acid residues (Fig.
4
). Although this protein retains the four C-terminal cysteine residues, the Cys
4
domain terminates at the fourth cysteine. In the case of TFIIS, the third [beta]-sheet of the zinc ribbon is composed entirely of residues downstream
of the fourth cysteine (
42
). If this region is critical for formation of the zinc ribbon, then the altered
form of Rpb9 encoded by
ssu73-1
is unlikely to include an intact zinc ribbon. Thus, the C-terminal 16 amino acids of Rpb9 and, presumably, the C-terminal zinc ribbon are dispensable for normal cell growth. The
effect on transcription initiation by the
shi
allele of
RPB9
is due to a C7F replacement in the N-terminal Cys
4
motif (
26
), and a C7A replacement has a similar effect on initiation (
27
). However, these replacements apparently abolish Rpb9 function since plasmids
carrying either allele fail to complement the Tsm
-
phenotype of an
rpb9
null mutant (
26
,
27
). Therefore, the N-terminal Cys
4
motif appears to be critical for Rpb9 function but might not be directly
involved in transcription start site selection, whereas the C-terminal Cys
4
motif is apparently not essential for Rpb9 function but plays an important role
in initiation.
In a yeast
in vitro
transcription system, replacement of both RNAP II and TFIIB from
S.cerevisiae
by their counterparts from
Schizosaccharomyces pombe
is sufficient to shift start sites from the pattern characteristic of
S.cerevisiae
to that of
S.pombe
. This result defines TFIIB and RNAP II as the sole determinants of accurate
start site selection
in vitro
(
29
). TFIIB is required to recruit RNAP II/TFIIF to the PIC (
4
; reviewed in
43
) and TFIIB directly binds RNAP II
in vitro
(
13
). Mutational analyses have demonstrated that RNAP II interacts with the N-terminal region of TFIIB (
14
,
44
-
46
). Previous genetic studies demonstrated that interaction between RNAP II and
TFIIB involves the Rpb1 subunit. Double
sua7
(TFIIB)
sua8
(Rpb1) mutants are inviable (synthetic lethality) and an
sua7
/
SUA7
sua8
/
SUA8
heterozygous diploid strain exhibits mutant phenotypes despite the presence of
the dominant, wild-type alleles of both genes (non-allelic non-complementation) (
25
). In addition,
sua7
and
sua8
mutations confer nearly identical effects on start site selection
in vivo
(
25
). These results establish a functional interaction between Rpb1 and TFIIB, and
suggest that these two proteins directly interact.
The results presented here extend the functional interaction between RNAP II and
TFIIB to include the Rpb9 subunit. However, our genetic data do not rigorously
exclude the possibility that the TFIIB and Rpb9 contributions to start site
selection are independent of one another. For example, the downstream shift
associated with altered TFIIB might be offset by an independent upstream shift
associated with altered Rpb9. However, this seems unlikely for several reasons.
First, the
ADH1
start sites observed upstream of position -37 in the
SUA7
+
ssu73-1
mutant (Fig.
3
, lane 4) are not seen in the
sua7-1 ssu73-1
mutant (lane 3), even upon prolonged exposure of the autoradiogram. If the upstream and downstream shifts were independent of one another, the
upstream shift associated with
ssu73-1
should be observed in both strains. Secondly, the
ssu73-1
effect is not limited to its role in start site selection; rather
ssu73-1
is an effective suppressor of the
sua7-1
Csm
-
growth defect (Fig.
1
). The Csm
-
phenotype of
sua7-1
mutants is presumably a consequence of the altered form of TFIIB on assembly or
stability of the PIC (
32
), rather than an effect of altered start site selection on the expression of a
gene(s) required for growth in the cold. In this case, suppression of the
sua7-1
Csm
-
phenotype by
ssu73-1
provides further support for a functional interaction between TFIIB and Rpb9.
Finally, given the physical interaction between TFIIB and RNAP II (
13
), and the dependence of accurate initiation on this interaction (
29
), it is difficult to imagine that the role of Rpb9 in start site selection is
independent of TFIIB.
We have failed to detect a direct interaction between Rpb9 and TFIIB and
therefore do not know if the functional interaction between these two proteins
involves a physical interaction. Moreover, we have failed to detect direct
interactions between TFIIB and all three proteins (Ssu71/Tfg1, Ssu72 and
Ssu73/Rpb9) identified in our genetic screen for factors that functionally
interact with TFIIB (Z.-W.S. and M.H., unpublished results). In this regard it is interesting to
note that the zinc binding motifs of both TFIIB and TFIIS are predicted to
engage in intramolecular interactions that mask functional domains. The zinc
ribbon of TFIIS has been proposed to form a cryptic nucleic acid binding site
that is exposed when TFIIS binds an elongation complex (
42
).
In vitro
studies indicate that TFIIB is comprised of a protease-susceptible N-terminus, which includes the zinc binding motif, and a protease-resistant C-terminal core domain (
46
). These two domains appear to engage in intramolecular interactions that are disrupted by an activator-induced conformational change, thereby exposing cryptic binding sites for other general
transcription factors (
47
). Perhaps Ssu71/Tfg1, Ssu72 and Ssu73/Rpb9 interact directly with TFIIB
in vivo
, but a conformational change in TFIIB is required to expose the interacting
domains.
We are grateful to David Gross for valuable discussions and critical comments on
the manuscript. We thank Rolf Furter, Lynn Henry, Roger Kornberg and David
Drubin for plasmids carrying the
RPB9
,
TFG2
and
TFG3
/
ANC1
genes. This work was supported by research grants from the American Cancer
Society (NP-842) and the National Institutes of Health (GM-39484).
*To whom correspondence should be addressed at present address: Department of
Biochemistry, Robert Wood Johnson Medical School, UMDNJ,
675 Hoes Lane, Piscataway, NJ 08854, USA
Strain
Genotype
T15
MAT
[alpha]
CYC1
+
cyc7-67 leu2-3,112 ura3-52 cyh2
T16
MAT
[alpha]
cyc1-5000 cyc7-67 leu2-3,112 ura3-52 cyh2
YDW546
MAT
[alpha]
cyc1-5000 cyc7-67 leu2-3,112 ura3-52 cyh2 sua7-1
YMH71-9C
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1
YZS19
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu72-1 ssu73-1
YZS19-3B
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu73-1
YZS19/YCplac33
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu72-1 ssu73-1
[
URA3
]
YZS19/pDW11
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu72-1 ssu73-1
[
URA3 SUA7
]
YZS19/pM586
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu72-1 ssu73-1
[
URA3 SSU72
]
YZS19/pM681
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu72-1 ssu73-1
[
URA3 RPB9
]
YZS19-3B/YCplac33
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu73-1
[
URA3
]
YZS19-3B/pDW11
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu73-1
[
URA3
SUA7
]
YZS19-3B/pM586
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu73-1
[
URA3
SSU72
]
YZS19-3B/pM681
MAT
a
cyc1-5000 cyc7-67 trp5-48 his5-2 ura3-52 sua7-1 ssu73-1
[
URA3 RPB9
]
YAT30
MAT
[alpha]
cyc1-5000 cyc7-67 leu2-3,112 ura3-52 cyh2 sua7-1 RPB9:URA3
REFERENCES
Return