(Received for publication, December 9, 1996, and in revised form, February 24, 1997)
From the Department of Biochemistry, University of Missouri, Columbia, Missouri 65211
Two subunits of about 36-44 kDa and 13-19 kDa
in the eukaryotic nuclear RNA polymerases share limited amino acid
sequence similarity to the subunit in Escherichia coli
RNA polymerase. The
subunit in the prokaryotic enzyme has a
stoichiometry of 2, but the stoichiometry of the
-like subunits in
the eukaryotic enzymes is not entirely clear. To gain insight into the
subunit stoichiometry and assembly pathway for eukaryotic RNA
polymerases, in vitro reconstitution experiments have been
carried out with recombinant
-like subunits from yeast and plant RNA
polymerase II. The large and small
-like subunits from each species
formed stable heterodimers in vitro, but neither the large
or small
-like subunits formed stable homodimers. Furthermore, mixed
heterodimers were formed between corresponding subunits of yeast and
plants, but were not formed between corresponding subunits in different RNA polymerases from the same species. Our results suggest that RNA
polymerase II
-like heterodimers may be the equivalent of
homodimers found in E. coli RNA polymerase.
Escherichia coli core RNA polymerase contains three
different subunits of 155 kDa (), 151 kDa (
), and 37 kDa (
).
Two copies of the
subunit are present in the core enzyme, resulting
in a
(
)2 subunit structure (1). Subunits related
to the E. coli
,
, and
subunits are found in most
multimeric RNA polymerases, including eukaryotic RNA polymerases I, II,
and III. The
and
subunits share at least eight and nine blocks
of amino acid sequence similarity with their counterparts in archeal
and eukaryotic nuclear RNA polymerases (2). The
subunit shares two
blocks of limited amino acid sequence similarity with two different
subunits found in archeal RNA polymerase and eukaryotic RNA polymerases I, II, and III. The eukaryotic
-like subunits consist of a large
-like subunit, 36-44 kDa, and a small
-like subunit, 13-19 kDa (2-12). Blocks of amino acid sequence similarity between the E. coli
subunit and the eukaryotic
-like subunits extend over 25-40 amino acids (2-12). Yeast (Saccharomyces cerevisiae)
nuclear RNA polymerases are the most thoroughly studied of the
eukaryotic nuclear enzymes (2), and yeast RNA polymerases I and III or A and C contain common
-like subunits of 40 kDa (AC40) and 19 kDa
(AC19), while RNA polymerase II or B contains
-like subunits of 44 kDa (B44 or RPB3) and 12.5 kDa (B12.5 or RPB11). The
-like subunits
of the plant Arabidopsis thaliana are comparable to those found in yeast (5-7). Arabidopsis contains AC
-like
subunits of 42/43 and 14 kDa and B subunits of 36 and 13.6 kDa (5-7). While the yeast
-like subunits are encoded by single-copy genes (3,
10-12), the large AC and B
-like subunits are each encoded by two
genes (i.e. two genes encode AC42 and AC43, and two genes encode B36a and B36b) in Arabidopsis (5, 6), but the
functional significance, if any, of these redundant genes in
Arabidopsis has not been determined.
The stoichiometries of the -like subunits in eukaryotic RNA
polymerases have not been totally resolved. Yeast AC40 and AC19 have
been reported to have stoichiometries of one in RNA polymerases I and
III (13), while yeast B44 has been reported to have a stoichiometry of
2 in RNA polymerase II (14). The stoichiometry of yeast B12.5 has not
been determined. Lalo et al. (15) showed that yeast AC40 and
AC19 subunits interact with one another, but that neither AC40 nor AC19
interacted with themselves in a yeast two-hybrid system. In a related
study, Ulmasov et al. (6) used the yeast two-hybrid system
and association studies with in vitro translated subunits to
demonstrate that Arabidopsis B36 and B13.6 subunits interact
with one another but not with themselves in vivo and
in vitro. The results of Lalo et al. (15) and
Ulmasov et al. (6) suggested that the large
-like subunit
and the small
-like subunit might form heterodimers in yeast RNA
polymerase I and III and in Arabidopsis RNA polymerase II
and that these heterodimers might be the equivalent of the
2 homodimer found in E. coli RNA polymerase.
In contrast to the above studies, in vivo labeling studies
with yeast RNA polymerase II have indicated that the large
-like
subunit, B44, is present at a stoichiometry of 2 (14, 16). On the other
hand, in vivo labeling studies with Arabidopsis
RNA polymerase II are consistent with a stoichiometry of 1 for the
large
-like subunit, B36 (6). A number of other studies in a variety
of eukaryotes have suggested that the stoichiometry of the third
largest subunit in RNA polymerase II is 1 (17-26).
We have used a renaturation protocol with recombinant -like subunits
to determine which combinations of
-like subunits can form stable
complexes in vitro. Here, we report on renaturation studies
with yeast and Arabidopsis RNA polymerase II large and small
-like subunits that provide information on RNA polymerase class
specificity, species specificity, and stoichiometry of subunit interactions.
Plasmids encoding
nuclear -like RNA polymerase subunits lacking polyhistidine tags
(His tags) were constructed by inserting ORFs1 into the vector pET-16b (Novagen,
Madison, WI) between NcoI and BamHI sites. The
ORF encoding Arabidopsis AC14 was inserted into pET-16b in
two steps. First, cDNA encoding amino acids 1-117 of AC14 was
excised from a plasmid in which the first codon of the AC14 ORF was
converted to a NcoI site2 and
the AC14 cDNA fragment was ligated into the NcoI site of pET-16b. Second, cDNA encoding amino acids 107-122 and the
3
-untranslated region of the Arabidopsis AC14 was excised
from an EST clone (GenBankTM accession number Z25617[GenBank]) as an
AflII-BamHI fragment. The
AflII-BamHI fragment was ligated into the partial
Arabidopsis AC14 in pET-16b construct between the
AflII and BamHI sites. The ORFs encoding four
different nuclear RNA polymerase subunits were amplified from cDNA
clones using Pfu DNA polymerase according to the
manufacturer's recommendations (Stratagene, La Jolla CA) with the
following oligonucleotides: CATGTCATGAATGCTCCAGACAGATTCG and
TCAAAATGCGTCGTCGG for yeast B12.5 (3), GCCTCATGAATGCTCCCGAACGAT and
TTAAAACTGATTCGAAAACTTGG for Arabidopsis B13.6 (6),
CATGCCATGGACGGTGCCACATACC and TCATCCTCCACGCATATGG for
Arabidopsis B36a (6), and GTCGTCTCATGACTGAAGAAGGTCCTCAAG and
CTACCAAGCATTATCATACCCTC for yeast B44 (11). Using appropriate restriction endonucleases, the ORFs encoding the yeast B12.5, Arabidopsis B13.6, and Arabidopsis B36a subunits
were inserted into pET-16b between the NcoI and
BamHI sites. To obtain a plasmid encoding a full-length
yeast B44 subunit, the ORF encoding this subunit was isolated using
oligonucleotides GTCGTCTCATGAGTGAAGAAGGTCCTCAAG and
CTACCAAGCATTATCATACCCTC with PCR-mediated amplification as described above. The PCR product was digested with
BspHI, and the BspHI product encoding amino acids
1-37 of the yeast B44 subunit was inserted into pET-16b at the
NcoI site. The pET-16b derivative containing the
partial yeast B44 subunit ORF was digested with BamHI,
filled in with Klenow, digested with SalI, and ligated to a
PCR product containing the yeast B44 ORF that had been digested with
SalI. To obtain a construct encoding the a
polyhistidine-tagged (His-tagged) Arabidopsis AC42, the AC42
ORF (5) was amplified by PCR using the oligonucleotides
GGAATTCCATATGGTGACTAAAGCAGAAAAACA and TCAAAAGTCGGTTATAGCTTCAC with
Pfu DNA polymerase as described above. PCR products
containing the Arabidopsis AC42 subunit ORF were
ligated into pET-16b between NdeI and BamHI.
Plasmids encoding the His-tagged Arabidopsis B36a, B36b, and
AC43 and yeast B44 subunits were constructed using a similar strategy
to that described above. NdeI sites were introduced at
translational start codons using specific oligonucleotides and PCR. The
large
-like subunit ORFs were inserted into pET-16b between
NdeI and BamHI sites. Each plasmid was
introduced into E. coli strain BL21(DE3) and expressed using
the procedure of Studier and Moffatt (27).
Each
recombinant RNA polymerase subunit accumulated as an insoluble product
in E. coli. The insoluble products were partially purified
using the methods of Borukhov and Goldfarb (28) and dissolved in buffer
B (6 M guanidine hydrochloride, 50 mM Tris-HCl, pH 7.9, 10 mM MgCl2, 10 µM
ZnCl2, 1 mM EDTA, 10 mM
dithiothreitol, 10% glycerol) (29). In most experiments (Figs. 2, 3, 4, 5), 500 µg of denatured, His-tagged large -like subunit
(Arabidopsis B36a, B36b, AC42, or AC43 or yeast B44) was
mixed with a 5-fold molar excess of denatured small
-like subunit
(Arabidopsis B13.6 or AC14, or yeast B12.5) and diluted to
0.5 mg/ml in buffer B. As controls to rule out nonspecific interaction
of subunits with the Ni-NTA resin, the small
-like subunits were
diluted to 0.5 mg/ml in buffer B without other
-like subunits and
tested for binding on Ni-NTA columns (see below). In one set of
experiments (Fig. 7, A and B), 250 µg of large
-like subunits that contained or lacked a His tag were mixed with a
5-fold molar excess of small
-like subunits and diluted to 0.5 mg/ml
in buffer B. In another experiment (Fig. 7C), approximately
equal amounts of yeast B12.5 and His-tagged yeast B12.5 were mixed with
1 molar eq of yeast B44 and diluted to 0.5 mg/ml in buffer B. Renaturation was carried out by dialysis against 500 ml of buffer C (50 mM Tris-HCl, pH 7.9, 200 mM KCl, 10 mM MgCl2, 10 µM
ZnCl2, 1 mM EDTA, 5 mM
-mercaptoethanol, 20% glycerol) (29), followed by dialysis against
two 500-ml changes of buffer D (buffer C lacking EDTA). Renatured
subunits were clarified by centrifugation at 14,000 rpm at 4 °C in
an Eppendorf centrifuge. Clarified mixtures were applied to 0.5-ml
columns containing Ni-NTA resin (QIAGEN, Chatsworth, CA) that were
equilibrated in buffer D. Columns were washed with 10 volumes of buffer
D, followed with 10 volumes of buffer D containing 40 mM
imidazole. Columns were eluted with buffer D containing 500 mM imidazole, and four drop fractions (0.2 ml) were
collected. Eluted fractions were analyzed in 10% SDS gels (30) or by
gel filtration chromatography. Gel filtration chromatography was
carried out by applying 200 µl of Ni-NTA eluted fractions to a 24-ml
Superose 12 column (Pharmacia Biotech Inc.) that was equilibrated in
buffer C containing 10% glycerol instead of 20% glycerol. Columns
were run at a flow rate of 0.25 ml/min, and 0.5-ml fractions were
collected and analyzed in 10% SDS gels. Molecular mass was
calculated as recommended by Stellwagen (31).
Both the large and small -like subunits in eukaryotic
RNA polymerases contain two motifs related to the motifs in the
E. coli
subunit (2-12). These motifs consist of a
25-residue amino-terminal block of similarity and a 40-residue block of
similarity that are separated by about 130 amino acids in the E. coli
subunit and by 140-200 amino acids in the eukaryotic
large
-like subunits, but are juxtaposed in the eukaryotic small
-like subunits (Fig. 1). To determine if large and
small
-like subunits interacted with themselves or with one another,
we carried out renaturation experiments with recombinant large
-like
subunits (36-44 kDa) and small
-like subunits (13-14 kDa) from
Arabidopsis and yeast RNA polymerases using the methods
described by Tang et al. (29). The large RNA polymerase II
-like subunits that were tested included Arabidopsis B36a
and B36b and yeast B44. The Arabidopsis B36a and B36b
subunits are encoded by two different genes, and the B36a subunit is
the predominant form of the large
-like subunit associated with
purified RNA polymerase II (6). The small RNA polymerase II
-like
subunits that were analyzed consisted of Arabidopsis B13.6
and yeast B12.5. Recombinant
-like subunits expressed in E. coli were recovered as insoluble polypeptides in inclusion body
fractions and solubilized in a denaturing buffer containing guanidine
hydrochloride. Various combinations of denatured large and small
-like subunits were mixed together and renatured by dialysis against
buffers that lacked guanidine hydrochloride. In these experiments, the
large
-like subunit contained a His tag on the amino terminus, which
facilitated purification of soluble, renatured subunits on
Ni-NTA-agarose columns. The small
-like subunits lacked His tags and
were retained on the Ni-NTA columns only if they formed stable
complexes with the His-tagged large
-like subunits.
When the Arabidopsis B13.6 subunit was subjected to the
renaturation protocol, the subunit precipitated and little if any B13.6
subunit was retained in a soluble form (Fig.
2A). In contrast, the bulk of the His-tagged
Arabidopsis B36a (or B36b; data not shown) subunit remained
soluble during the renaturation protocol and could be purified by
Ni-NTA affinity chromatography (Fig. 2B). Renaturation of
the B13.6 subunit in the presence of either His-tagged B36a or B36b
subunit resulted in a soluble form of B13.6 that bound to the Ni-NTA
column along with the His-tagged B36a or B36b subunit (Fig. 2,
C and D), suggesting that the B13.6 subunit
formed a soluble, stable complex with the B36a or B36b subunit. Similar
experiments were carried out with the yeast B44 and B12.5 subunits. In
this case, however, the majority of the B12.5 subunit was recovered in
the soluble fraction whether renatured in the absence or presence of
the His-tagged B44 subunit (Fig. 2, E and F). The
yeast His-tagged B44 subunit, like the Arabidopsis His-tagged B36a or B36b subunit, remained soluble during the
renaturation protocol and could be recovered by chromatography on the
Ni-NTA resin (data not shown). The B12.5 subunit was not retained on the Ni-NTA column if renatured in the absence of the His-tagged B44
subunit (Fig. 2E), but was retained on the nickel resin if renatured in the presence of the His-tagged B44 subunit (Fig. 2F). The results with yeast -like subunits are consistent
with the results obtained with the Arabidopsis
-like
subunits and indicate a small
-like RNA polymerase II subunit forms
a stable complex with a large
-like RNA polymerase II subunit
in vitro.
Another set of renaturation experiments was
carried out to determine if the yeast small -like subunit could
stably associate with the Arabidopsis large
-like subunit
and if the Arabidopsis small
-like subunit could stably
associate with the yeast large
-like subunit. Results based upon the
retention of small
-like subunits and His-tagged large
-like
subunits on Ni-NTA columns presented in Fig. 3 indicate
that yeast B12.5 forms stable complexes with Arabidopsis
B36a (Fig. 3A) and B36b (Fig. 3B) and that
Arabidopsis B13.6 forms stable complexes with yeast B44
(Fig. 3C). The renaturation of soluble
Arabidopsis B13.6 appears to be less effective when carried
out in the presence of the yeast B44 subunit, however, than
renaturation in the presence of Arabidopsis B36a or B36b (compare Fig. 3C to Fig. 2, C and D).
This less effective renaturation with the Arabidopsis B13.6
and yeast B44 subunit appears to result from precipitation of these two
subunits when renatured together, while the yeast B44 subunit remains
soluble when renatured alone or in combination with the yeast B12.5
subunit (compare Fig. 3C to Fig. 2F). Because
only small amounts of B13.6 and B44 complexes remained soluble, little
of this complex was applied and bound to the affinity resin (Fig.
3C).
To determine if the large and small -like subunit interactions are
specific for RNA polymerase II
-like subunits, renaturation experiments were conducted with mixtures of small
-like RNA
polymerase II subunits and large
-like RNA polymerase I and III
subunits. The large RNA polymerase I and III
-like subunits
subjected to analysis were Arabidopsis AC42 (42 kDa) and
AC43 (43 kDa). The Arabidopsis AC42 and AC43 subunits are
encoded by two different genes, and the AC42 subunit is the predominant
form associated with RNA polymerase III, but the form associated with
RNA polymerase I has not been determined (5). Arabidopsis
B13.6 or yeast B12.5 subunits were subjected to the renaturationation
protocol as mixtures with His-tagged Arabidopsis AC42 or
AC43. When His-tagged AC42 or AC43 subunits were renatured alone, they
behaved similar to the Arabidopsis B13.6 subunit in that the
bulk of these subunits precipitated during the renaturation protocol
(data not shown). Only small amounts of soluble His-tagged AC42 or AC43
could be recovered on the Ni-NTA resin. Likewise, when His-tagged AC42 or AC43 were renatured with Arabidopsis B13.6, most of the
AC42, AC43, and B13.6 precipitated, and only small amounts of
His-tagged AC42 or AC43 could be recovered on Ni-NTA columns (Fig.
4, A and B). No B13.6 subunit was
retained on the nickel column. Similar results were obtained when
renaturation experiments between yeast B12.5 and Arabidopsis
AC42 or AC43 were conducted (Fig. 4, C and D). In
this case, however, the yeast B12.5 subunit remained soluble during the
renaturation protocol, but failed to bind to the Ni-NTA column even
though small amounts of His-tagged AC42 or AC43 were retained on the
nickel resin.
To assess whether Arabidopsis
AC42 and AC43 subunits could form stable complexes with a RNA
polymerase I and III small -like subunit, the Arabidopsis
small
-like subunit, AC14 (7), was renatured in the absence or
presence of a His-tagged AC42 or AC43 subunit. Most of the AC14 subunit
remained soluble whether or not an AC42 or AC43 subunit was present
during the renaturation protocol, but the AC14 subunit was only
retained on the Ni-NTA column if renatured in the presence of a
His-tagged AC42 or AC43 subunit (Fig. 5,
A-C). Most of the AC42 and AC43 subunits precipitated out
of solution if renatured in the absence of AC14, but both subunits
remained soluble if renatured in the presence of AC14. The AC42 and
AC43 subunits failed to stay in solution if renatured in the presence
of small
-like Arabidopsis B13.6 or yeast B12.5 subunit
(Fig. 4, A-D), indicating that the large RNA polymerase I/III
-like subunits only remain soluble if they form a stable complex with a small
-like subunit. These results suggest that Arabidopsis RNA polymerase I and III large
-like
subunits, AC42 and AC43, form soluble stable complexes with the
Arabidopsis RNA polymerase I and III small
-like subunit,
AC14. On the other hand, RNA polymerase I and III large
-like
subunits do not form stable complexes with RNA polymerase II small
-like subunits.
To further
characterize -like subunit complexes, renatured
-like subunits
were fractionated on gel filtration columns. When renatured (but not
affinity-purified) Arabidopsis His-tagged B36a was
fractionated on Superose 12, a major portion of this subunit eluted as
a symmetrical peak with an estimated size of 54 kDa (Fig.
6A). A minor fraction eluted as a broad peak
with an average size of about 400 kDa, suggesting that some aggregation
of B36a subunits occurs during the renaturation or fractionation
process. The Arabidopsis B13.6 subunit could not be analyzed
by gel filtration because this subunit precipitated during the
renaturation protocol. When the His-tagged B36a subunit was renatured
(but not affinity-purified) with a molar excess of B13.6 subunit and
fractionated on Superose 12, most of the B36a and B13.6 subunits eluted
as a symmetrical peak with an estimated size of 67 kDa (Fig.
6B). The peak fraction containing the B36a and B13.6
subunits eluted one fraction earlier than the peak fraction containing
only the B36a subunit. A minor fraction of the B36a and B13.6 subunits
eluted in fractions with an estimated size of 200-500 kDa,
suggesting that some aggregation occurs similar to that found with the
B36a subunit alone. His-tagged B36A and B13.6 complexes that had been
renatured together and purified by affinity chromatography on an Ni-NTA
column were also analyzed by gel filtration on Superose 12. This
purified complex displayed a fractionation profile similar to that
observed with the unpurified B36a·B13.6 mixture (Fig. 6C);
however, high molecular weight aggregates were not observed with the
affinity-purified B36·B13.6 complex. Taken together, the gel
fractionation results are consistent with B36a and B13.6 associating
together as a stable heterodimer.
Affinity-purified complexes containing yeast His-tagged B44·yeast
B12.5 and Arabidopsis B36a·yeast B12.5 were also analyzed by gel filtration on Superose 12 (Fig. 6, D and
E). In each of these cases, the large -like subunit and
small
-like subunit fractionated as a stable complex similar
to that observed with affinity-purified Arabidopsis
His-tagged B36a·B13.6 complexes. The yeast His-tagged B44·B12.5
complex eluted in fractions with an average molecular mass of greater
than 67 kDa, however. The yeast His-tagged B44 subunit shows this same
fractionation pattern (i.e. eluting in fractions with an
estimated molecular mass in excess of 67 kDa) when renatured by itself
and analyzed by Superose 12 gel filtration (data not shown). Since the
large
-like subunit may not form a globular structure, it is
possible that some fractionation abnormalities are observed on Superose
12. Support for the fractionation abnormalities of the yeast large
-like subunit as opposed to association between these subunits was
obtained from the next set of experiments.
Gel filtration experiments did not allow us to
definitively determine if large and small -like subunits formed
heterodimers or some type of higher order multimers. For example,
complexes observed in affinity-purified fractions or gel filtration
fractions might consist of two large
-like subunits associated with
one small
-like subunit, one large subunit associated with two small subunits, or two large subunits associated with two small subunits. To
further verify that one large and one small
-like subunit associate
with one another to form heterodimers and not higher order multimers,
renaturation experiments were carried out with one His-tagged large
-like subunit, a second untagged large
-like subunit, and an
untagged small
-like subunit. Complexes were purified by affinity
chromatography on Ni-NTA columns, and purified complexes were analyzed
by SDS-gel electrophoresis. Because His-tagged and untagged large
-like subunits displayed different mobilities on SDS gels, these
subunits could be distinguished from one another. Fig.
7A shows that with affinity-purified
renatured mixtures of Arabidopsis His-tagged B36a, untagged
B36a, and untagged B13.6, only the His-tagged B36a and B13.6 subunits
bound to the affinity resin. No untagged B36a subunit was found in the
affinity-purified complex. Likewise, Fig. 7B shows that with
affinity-purified mixtures of yeast His-tagged B44, untagged B44, and
untagged B12.5, only the His-tagged B44 and B12.5 subunits bound to the
affinity resin. These results indicate that only one molecule of the
large
-like subunit associates in a complex with the small
-like
subunit.
To determine if only one molecule of the small -like subunit is
associated with this complex, renaturation experiments were carried out
with mixtures of yeast His-tagged B12.5, untagged B12.5, and untagged
B44. Complexes were purified by affinity chromatography on Ni-NTA
columns, and purified complexes were analyzed by SDS gel
electrophoresis. His-tagged and untagged B12.5 subunits displayed different mobilities on SDS gels and could be distinguished from one
another. Fig. 7C shows that affinity-purified complexes
contained the His-tagged B12.5 subunit and B44 subunit, but no untagged B12.5 subunit. These results indicate that only one molecule of the
small
-like subunit associates in a complex with the large
-like
subunit. Taken together, our results indicate that large and small
-like subunits from Arabidopsis and yeast form stable heterodimers in vitro.
Previous results obtained by using the yeast two-hybrid system
have indicated that a large -like subunit and small
-like subunit
from yeast RNA polymerase I and III and Arabidopsis RNA polymerase II could interact stably with one another in vivo
(6, 15). These studies also indicated that large
-like subunits do
not interact with themselves and small
-like subunits do not interact with themselves. Extragenic suppression studies in yeast support the interaction between a large and small
-like subunit in
yeast RNA polymerases I and III (15). Using an antibody pull-down experimental approach, Ulmasov et al. (6) further
demonstrated that an in vitro translated epitope-tagged
Arabidopsis RNA polymerase II large
-like subunit
associated with an untagged RNA polymerase II small
-like subunit.
While these results are consistent with one large
-like subunit and
one small
-like subunit associating as a heterodimer in RNA
polymerases I, II, and III, the experimental approaches described above
do not rule out additional interactions with other molecules that might
be required for stable association of
-like subunits and do not
establish stoichiometries for these
-like subunit interactions. For
example, previous studies do not rule out the possibility that
B36(B13.6)2 or (B36)2B13.6 complexes were
formed. We have taken an alternative approach using in vitro association studies with recombinant
-like subunits from yeast and
Arabidopsis RNA polymerases that allows us to rule out
accessory protein requirements for
-like subunit interactions, to
establish stoichiometries for
-like subunits in purified stable
complexes, and to provide a framework for the reconstitution of
eukaryotic RNA polymerases. Our results indicate that one large
-like subunit associates with one small
-like subunit to form
stable heterodimers in vitro and that heterodimerization of
-like subunits in vitro is not species-specific but is
specific for the class of RNA polymerase (i.e. RNA
polymerase II versus RNA polymerases I and III).
E. coli and other prokaryotic RNA polymerases contain an
2 homodimer that is associated with
and
in core
enzymes. The
2 homodimer is formed as the initial event
in the E. coli RNA polymerase assembly pathway (32). The
assembly pathway for eukaryotic nuclear RNA polymerases has not been
established, but studies in yeast have suggested that the third largest
subunit in yeast RNA polymerase II, which is related to the E. coli
subunit, may homodimerize and subsequently interact with
the second largest and largest subunits in enzyme assembly (33). Our
results suggest an alternative to the
2 homodimer enzyme
assembly pathway in eukaryotic RNA polymerases might involve the
initial heterodimerization of a large
-like subunit and a small
-like subunit. Future experiments will be required to determine if
2 heterodimers can stably associate with second largest
and/or the largest subunit of eukaryotic RNA polymerases to form a core
enzyme analogous to
(
)2.
Mutation studies with the E. coli subunit have shown
that the two motifs conserved in eukaryotic
-like subunits are
important in enzyme assembly of prokaryotic RNA polymerase (34-38).
Mutations in the NH2-terminal
-like motif in yeast AC40
subunit can be suppressed by elevated expression of the AC19 subunit
(15), and mutations in the yeast B44 subunit
-like motifs prevent
assembly of the B44 subunit with the second largest and largest
subunits of yeast RNA polymerase II (33). NH2-terminal and
COOH-terminal mutations in the Arabidopsis B36b subunit that
disrupt either
-like motif prevent interaction between the B36 and
B13.6 subunits in in vitro pull-down experiments (6). Taken
together, preliminary results with the eukaryotic
-like subunits
suggest that
-motifs conserved in prokaryotic
subunits and
eukaryotic
-like subunits may be important for subunit association
and enzyme assembly.
There is evidence that the subunits in E. coli RNA
polymerase are not structurally (i.e. tertiary structure
within the core enzyme) or functionally equivalent (39-45). Based on
the nonequivalence of the
subunits in prokaryotic RNA polymerases,
we propose that eukaryotes may have evolved
-like subunits that are
nonequivalent in primary and tertiary structure as well as in function.
-Like subunits in eukaryotes are clearly not equivalent in function to prokaryotic
subunits because eukaryotic subunits lack the COOH-terminal motif of prokaryotic
subunits that plays a role in
transcriptional activation and repression and in binding to promoter
DNA elements (45-47). Eukaryotic enzymes may have evolved distinct
-like subunits that serve nonequivalent roles in enzyme assembly, in
contacting other subunits, and possibly in enzymatic function.
Recent immunoelectron microscopy results with antibodies directed
against the largest (-like), second largest (
-like), and
-like subunits of yeast RNA polymerase I have provided information about the location of these subunits within the enzyme and have indicated that the large and small
-like subunits are located near
each other within the apical region of the enzyme (48). These results
have not, however, provided definitive information about the
stoichiometries of the
-like subunits within the yeast enzyme. Our
results, based upon in vitro subunit interaction studies, are consistent with there being one large
-like subunit and one small
-like subunit in eukaryotic RNA polymerases. Nonetheless, there remains the possibility that additional
-like subunits might
associate with eukaryotic RNA polymerases by interacting with a higher
order subunit complex (i.e. after additional subunits have
associated with the
-like subunit heterodimer).