ABSTRACT
Human NC2 utilizes a unique mechanism of repression of transcription by
associating with TBP and inhibition of preinitiation complex formation. Here we
have cloned two genes from
Saccharomyces cerevisiae
and functionally characterized them as yeast NC2. We show that yeast NC2 binds
to TBP as a heterodimer and represses RNA polymerase II transcription during
assembly of the preinitiation complex. Yeast NC2 is highly homologous to its
human counterpart within histone fold domains. C-Terminal regions previously discussed to be important for repression in
man are in part not conserved. The human
[alpha]
but not the
[beta]
subunit efficiently heterodimerizes and represses transcription in combination with the corresponding yeast subunit. Yeast and human NC2 inhibit transcription in the presence of yeast and human TBP. However,
repression is optimal within one species. The N-terminus of human TBP supports repression of transcription by human but
not by yeast NC2.
Transcription by RNA polymerase II in the eukaryotic cell is regulated by the
complex interplay of positive and negative regulators. Transcriptional
inhibitors have evolved to utilize a variety of repression mechanisms (reviewed
in
1
). Many of the repressors are site-specific DNA binding proteins that control the activity of specific genes.
Some examples have been reported that regulate basal promoter activity by
directly targeting the basal transcription machinery. For example, ADI/Mot1
releases the TATA binding protein (TBP) from the promoter (
2
). Another example is provided by negative cofactor 2 (NC2) (
3
). Human NC2 was recently purified and cloned (
4
,
5
). NC2 is a heterodimer consisting of NC2[alpha]/DRAP1 and NC2[beta]/Dr1 (
6
). NC2 binds to TBP and inhibits binding of both TFIIA and TFIIB. NC2 contact points on TBP have been mapped to a basic region within the core of TBP (
7
) that overlaps with the binding site of TFIIA (
8
,
9
), arguing for a steric exclusion of TFIIA and NC2. The precise mechanism of
TFIIB exclusion is not yet understood. Surprisingly, the TBP-TFIIB interface lies opposite the interaction surface of TBP with NC2 (
10
,
11
). However, it is possible that inhibition of TFIIB assembly also depends on
conformational changes in the promoter structure induced by NC2.
Here, we reasoned that NC2 function should be conserved in different species,
since the region of TBP that binds to human NC2 is highly conserved among
species in its primary sequence and structure (
12
). We report cloning of the NC2 subunits from
Saccharomyces cerevisiae
and characterize the recombinant proteins in an
in vitro
system reconstituted with human components that allowed the interchange of human and yeast subunits.
Restriction enzymes
Nde
I and
Bam
HI and Vent DNA polymerase were obtained from New England Biolabs (Schwalbach, Germany). Bovine serum albumin solution (20 mg/ml) was from Boehringer Mannheim
(Mannheim, Germany). All chemicals were purchased either from Sigma
(Deisenhofen, Germany) or Merck (Darmstadt, Germany). Oligonucleotides were
synthesized by Eurogentec (Belgium). Ni-NTA-agarose was purchased from Diagen (Hilden, Germany), heparin-Sepharose from Pharmacia (Freiburg, Germany) and cyanogen
bromide-activated Sepharose 4B from Sigma (Deisenhofen, Germany).
Aliquots of 55 ng yeast cDNA library (a gift of J.Regenbogen, Ludwig-Maximilians-Universität München, Germany) or 10 ng genomic yeast DNA were employed
in a standard PCR containing 200 pmol forward primer and 200 pmol reverse primer in a final volume of 0.1
ml. Oligonucleotide primers for amplification of yNC2[alpha] were the forward primer yNC2[alpha]UP, 5'-CCATGGAACAT
Electrophoretic mobility shift assays (EMSAs) were carried out as described (
4
). Reactions contained 4 ng hNC2[alpha], 30 ng hNC2[beta]/Dr1, 30 ng yNC2[alpha], 30 ng yNC2[alpha][Delta]31, 30 ng yNC2[beta] and 10 ng recombinant yeast TBP.
Aliquots of 50 fmol of a 90 bp end-labeled DNA fragment containing the HIV promoter TATA box were used as
probe. EMSAs were analyzed in 5% (50:1) 0.5* Tris-borate-EDTA polyacrylamide gels containing 5% (v/v) glycerol and 2
mM MgCl
2
.
Human and yeast NC2[beta] and BSA as a control were covalently immobilized on BrCN Sepharose.
Briefly, the beads were washed with 1 mM HCl before coupling and equilibrated
in coupling buffer (0.1 M NaHCO
3
, pH 8.3, containing 0.5 M NaCl). The ligands were dialyzed to coupling buffer
and incubated with the beads at a final concentration of 1 mg/ml. Excess ligand
was washed away with coupling buffer and the remaining active groups were
blocked with 0.1 M Tris-HCl, pH 8.0. Samples of 10 [mu]g human or yeast NC2[alpha] were incubated with 10 [mu]l affinity beads in a final volume of 100 [mu]l under transcription conditions for 1 h at room
temperature. After extensive washing the beads were boiled in Laemmli buffer
and subjected to SDS-PAGE.
In vitro
transcription reactions were carried out with the
Sma
I-linearized HIV core promoter pMRG5 as described (
20
). Reactions contained the full set of general transcription factors, including
recombinant TFIIA (co-renatured and purified [alpha][beta] and [gamma] subunits), TFIIB, TBP, TFIIE, TFIIF, partially
purified TFIIH and RNA polymerase II. In preincubation experiments NC2 subunits
were added either before or after 30 min incubation of the general factors with
the template at 30oC. Quantitation of the transcripts was performed on an Instant Imager
(Canberra Packard).
Database searches were performed with the Wisconsin Sequence Analysis Package
and included the programs FASTA, BLAST and BESTFIT.
In order to identify yeast homologs of human NC2 we performed a computer search
of the non-redundant NCBI protein database with the human NC2 sequences. Two open
reading frames (ORFs) shared significant homology with hNC2 and were designated
yNC2[alpha] and yNC2[beta] respectively (
4
). The gene for the yNC2[beta] subunit contained an intron of 92 bp (Fig.
1
B). Here, we isolated the corresponding cDNAs encoding the putative NC2 subunits
(Fig.
1
A and B). The ORF of yNC2[alpha] encodes a protein of 142 amino acids with a calculated molecular mass of
15.5 kDa, while yNC2[beta] is comprised of 146 amino acids with a calculated molecular mass of 16.6
kDa. Both yeast NC2 subunits contain multiple potential target sites for
protein kinases (Fig.
1
A and B). These could be important for yeast NC2 activity, as it has been
demonstrated that phosphorylation of human NC2 by casein kinase II (CKII) increases the specificity of NC2 for TBP-promoter complexes (
10
).
Alignment of the putative yeast NC2 subunits with their human counterparts
revealed striking conservation of the histone fold, a region which is conserved
in all four histones and reconstitutes the structured region of the nucleosome
(
14
). The histone fold motif consists of a long central [alpha]-helix flanked on each side by a short [alpha]-helix set apart from the central helix by [beta]-strands.
The human and the yeast [alpha] subunit share 41.9% identity and 66.7% similarity in amino acids 7-100 of hNC2[alpha], which encompasses the entire histone fold. Yeast NC2[alpha] is much smaller than hNC2[alpha], although it contains an additional N-terminal region of 54 amino acids.
However, it lacks essentially the entire C-terminal region of hNC2[alpha]. The homologous region to hNC2[alpha] includes residues 48-56 of yNC2[alpha], which are predicted to be part of helix I of
the histone fold, but which are not conserved in DRAP1 (
5
). Homologies in the [alpha] subunits also include a short hydrophobic motif located in the extended
helix III, which we term the FDFL box (Fig.
1
C). Both human and yeast NC2 are also highly related to components of the CCAAT
box binding factor CBF (also termed NF-Y or CP1). Both the histone fold and the FDFL box are conserved in two
subunits of CBF, the CBF-C protein and the corresponding yeast homolog, the HAP5 protein.
The [beta] subunits share 41.3% identity and 62.7% similarity in a region spanning
amino acids 5-130 of hNC2[beta]/Dr1 (Fig.
1
D). This region again includes the histone fold motif and extends to a conserved
acidic stretch between amino acids 115 and 126 of hNC2[beta]/Dr1 with the sequence EEELLRQQEELF. This acidic stretch is also partly
conserved in a putative NC2[beta]/Dr1 protein from
Arabidopsis thaliana
(
15
) and in one subunit of CBF, the CBF-A protein. No function, however, has been assigned to this region so far.
The glutamine/alanine-rich region which has been characterized as the repression domain in hNC2[beta]/Dr1 (
16
) is not present in yeast NC2[beta] (Fig.
1
E).
In addition to conservation of the histone fold domain and the FDFL box, the [alpha] subunits display a similar overall charge and charge distribution. The
histone fold domains of hNC2[alpha] and yNC2[alpha] are basic, whereas the entire polypeptides are acidic. Positive
charges in the histone fold may provide an explanation for the observation that
the [alpha] subunit non-specifically contacts DNA (
4
). In the case of the [beta] subunits the histone fold domains and the full-length proteins possess an acidic isoelectric point.
Escherichia coli
-expressed and purified NC2 polypeptides were analyzed for their capability
to form complexes with DNA and TBP. As in the case of human NC2, each yeast NC2
subunit alone failed to form stable complexes with promoter DNA and TBP (Fig.
2
, lane 3, and data not shown). In contrast, a combination of yNC2[alpha] and yNC2[beta] efficiently formed yNC2-yTBP-DNA complexes (Fig.
2
, lane 4). Thus, both hNC2 and yNC2 are heterodimers that form quarternary
complexes with DNA and TBP. The heterodimer of yeast NC2[beta] and human NC2[alpha] (lane 2) but not yeast NC2[alpha] and human NC2[beta]/Dr1 (lane 3) bound to TBP. Yeast NC2-yTBP complexes multimerize on promoter-bound TBP when longer DNA fragments are
used as a probe, as has been reported for human NC2-TBP complexes (lanes 1 and 4 and data not shown).
The evolutionary conservation of the repressor NC2 was further underscored
through analysis of recombinant proteins in a highly purified
in vitro
transcription system. Yeast NC2 efficiently inhibited transcription of a given
model promoter (Fig.
4
B, lane 3 versus 6), demonstrating that the yeast polypeptides are a potent
repressor of class II transcription. The individual subunits alone did not
affect transcription (lane 3 versus lanes 1 and 2), which shows that
heterodimerization of yNC2 subunits is required for efficient repression of
transcription. When individual subunits were exchanged between yeast and man,
yNC2[beta] could fully substitute for its human counterpart (lane 4 versus lane 5).
In contrast, yNC2[alpha] and hNC2[beta]/Dr1 repressed transcription far less efficiently (lane 3 versus
lane 7), which is consistent with the DNA and the protein binding analysis.
Figure
Repression of transcription by combinations of human and yeast NC2 subunits
correlated with the ability to prevent TFIIB from binding to TBP-promoter complexes (data not shown), as described earlier for hNC2 (
4
). In agreement with the model that NC2 primarily acts on preinitiation complex
formation, preincubation of the general factors with the template abrogated
repression by NC2 (Fig.
4
A, lane 2 versus lane 3).
Both human and yeast NC2 were capable of inhibiting transcription in the
presence of human full-length TBP, human core TBP and yeast TBP, although repression of
transcription is most effective within the same species (Fig.
4
C, lanes 2 and 3 versus 5 and 6). As the origin of TBP clearly influenced
function of NC2, we reasoned that the differential behaviour might be due to
the N-terminal domain of TBP, which is not conserved between species. Human NC2
inhibited transcription 6-fold more efficiently in the presence of full-length hTBP than the core of hTBP (Fig.
4
C, lane 5 versus lane 8), indicating that the N-terminus of hTBP enhances the function of hNC2. On the other hand,
repression of transciption by yeast NC2 was more pronounced when the N-terminus of hTBP was absent (lane 6 versus 9). In summary, the N-terminus of TBP supports the function of NC2 provided that both are
compatible (e.g. originate from the same species).
In this study we have identified the homolog of a human repressor of class II
transcription, NC2, in the yeast
S.cerevisiae.
Yeast NC2 exhibits all the characteristics attributed to the human factor.
Yeast NC2 is a basal repressor of RNA polymerase II transcription
in vitro
and consists of two subunits which form quarternary complexes with TBP and DNA.
The yeast NC2 sequences are highly conserved between yeast and man, pointing to
an evolutionarily important role of NC2. This is also reflected in that both
NC2 subunits are essential for yeast viability (R.A.Young, personal
communication).
The comparison of the human and yeast NC2 proteins is an evolutionary saturating
mutagenesis study, which reveals important regions and amino acids in detail. This study contains several new findings
regarding the repressor NC2. NC2 contains a highly conserved core, which
consists of the histone fold domains and comprises two regions (the FDFL motif
of NC2[alpha] and the acidic region of NC2[beta]) which have not yet been characterized.
The highest degree of homology is found in the histone fold domains. In the
histone proteins it serves as an architectural motif utilized in protein
dimerization and DNA compaction (
17
-
19
,
20
). The homology beween the NC2 [alpha] subunits is restricted almost exclusively to the histone fold. Deletion
of N-terminal sequences of yNC2[alpha], leaving the histone fold intact, did not affect its ability to
interact with TBP-promoter complexes. The homology between the [beta] subunits extends beyond the histone fold and includes the acidic
region. The region between amino acids 80 and 100 of hNC2[beta]/Dr1, however, which has previously been characterized as the TBP binding
domain (
5
,
16
), is not conserved between yeast and man. Also, our experiments do not support
a role in TBP binding, as heterodimers consisting of hNC2[alpha] and hNC2[beta]/Dr1(1-83), which lack this region, efficiently bound to hTBP or
yTBP and DNA
in vitro
(
4
; data not shown). The glutamine/alanine-rich region of hNC2[beta]/Dr1 is also not present in yeast. We have previously demonstrated
that repression of transcription requires the glutamine/alanine-rich region of hNC2[beta]/Dr1 in the presence of the human TFIID complex but not in the
absence of TAFs (
4
). This may indicate that the interplay of NC2 and TAFs in yeast differs from
that in man. In addition to the TAFs, components of the RNA polymerase II (Pol
II) holoenzyme (
21
,
22
) which are associated with the C-terminal domain of the largest Pol II subunit counteract NC2. Young and
colleagues isolated yeast NC2[alpha] as a functional antagonist of SRB4
in vivo
, which is a constituent of the Pol II holoenzyme and characterized yeast NC2 as a global repressor of class II transcription (R.A.Young, personal communication).
Yeast and human NC2 are interchangeable as entities in
in vitro
transcription. Interestingly, the N-terminus of TBP modulates the function of NC2. Repression of transcription
was most pronounced when NC2 and TBP originated from the same species. The
presence of the N-terminal domain of human TBP enhanced the function of human NC2, which
provides the first positive function for this domain. Initially it was proposed
that the N-terminus of human TBP was essential for Sp1-dependent transcription
in vitro
(
28
). Berk and colleagues, however, showed that functional holo-TFIID, which mediated transcriptional activation
in vitro
(e.g. by Sp1), could be purified from a cell line stably expressing only the
core domain of human TBP (
23
) or the core domain of yeast TBP (
24
).
Mixing experiments of human and yeast NC2 subunits revealed that at least the [beta] subunits can efficiently replace each other in function, whereas a
combination of yeast NC2[alpha] and human NC2[beta]/Dr1 failed to bind to TBP-DNA complexes and had little effect on transcription, which
is at least in part due to inefficient subunit association. Another example for
a sucessful interchange of a transcription factor between these species is
provided by TBP (see Fig.
4
C; reviewed in
12
). Functional exchange of TBP with regard to response to acidic activators
in vitro
has been reported, although human TBP failed to support yeast cell growth (
25
). Examples of other transcription factors that can be exchanged between yeast
and man include TFIIA (
26
-
28
), subunits of the human CCAAT binding factor CP1 and the yeast HAP2-HAP3 complex (
7
). The latter example is intriguing, as CCAAT binding factors are also thought
to contain heterodimeric histone folds (
29
,
3
0) and show the highest degree of homology to NC2.
Other reports describe exchanges between species that are more evolutionarily
linked, as exemplified by the yeasts
S.cerevisiae
and
Schizosaccharomyces pombe
or the fruit fly
Drosophila melanogaster
and man. Exchange of general transcription factors of
S.cerevisiae
and
S.pombe
revealed that three components, counterparts of human TFIIB, TFIIE and Pol II,
could not be exchanged individually but could be swapped in the pairs TFIIE-TFIIH and TFIIB-Pol II, demonstrating functional interactions between the
components of these pairs (
22
). The general transcription factor TFIIB can be exchanged from
D.melanogaster
to man (35). In the case of the yeast activator SWI2/SNF2, the
D.melanogaster
homolog could functionally replace the
S.cerevisiae
but not the
Homo sapiens
protein. Chimeric proteins composed of the ATPase domains of
D.melanogaster
brahma and SWI2 rescued the slow growth phenotype of swi
-
cells, whereas the fusion of SWI2 with the corresponding domain of the highly
homologous human ISWI protein failed to do so (
31
).
The low incidence of cases where interchange of factors from man to the yeast
S.cerevisiae
, representing two of the most divergent species of the eukaryotic kingdom, was
possible indicates a stringently conserved function for NC2 in regulating
transcription. Further experiments will have to show how specificity within one
system is maintained, given the rising number of histone fold proteins in the
cell, ranging from general transcription factors (
32
,
3
3
) to activators of transcription (
29
,
3
0).
We thank Ellen L.Gadbois and Richard A.Young for sharing results prior to
publication, Johannes Regenbogen for providing the yeast cDNA library, Gabi
Stumpf and Lutz Zeitlmann for critical reading of the manuscript and E.-L.Winnacker for general support. This work was supported by grants from
the BMBF, the DFG and the Genzentrum to M.M.

REFERENCES
Return

