1883-1888. 9. Ingraham, H.A., Flynn, S.E., Voss, J.W., Albert, V.R., Kapiloff, M.S., ... Finney, M., Ruvkun, G., and Horvitz, H.R. (1988) Cell, 55, 757-769. 14. Finney ...
Nucleic Acids Research, 1992, Vol. 20, No. 23 6369-6375
%.j 1992 Oxford University Press
POU domain transcription factors from different subclasses stimulate adenovirus DNA replication C.Peter Verrijzer+, Marijke Strating, Yvonne M.Mul and Peter C.van der Vliet* Laboratory for Physiological Chemistry, University of Utrecht, Vondellaan 24a, 3521 GG Utrecht, The Netherlands Received August 4, 1992; Revised and Accepted October 29, 1992
ABSTRACT POU domain proteins constitute a family of eukaryotic transcription factors that exert critical functions during development. They contain a conserved 160 amino acids DNA binding domain, the POU domain. Genetic data have demonstrated that some POU domain proteins are essential for the proliferation of specific cell types, suggesting a possible role in DNA replication. In addition, the ubiquitous POU transcription factor Oct-1 or its isolated POU domain enhances adenovirus DNA replication. Here we compared the DNA binding specificities of POU domain proteins from different subclasses. They exhibit overlapping, yet distinct binding site preferences. Furthermore, purified Pit-1, Oct-i, Oct-2, Oct-6, Oct-4 and zebrafish POU[C] could all stimulate adenovirus DNA replication in a reconstituted in vitro system. Thus, activation appears to depend on a property common to most POU domain proteins. Adenovirus DNA replication is also stimulated by the transcription factor NFI/CTF. In contrast to NFI, the POU domain did not enhance binding of precursor terminal protein-DNA polymerase to the origin nor did it stabilize the preinitiation complex. These results suggest that the POU domain acts on a rate limiting step after formation of the preinitiation complex. INTRODUCTION DNA replication and transcription are essential cellular processes that are regulated by sequence-specific DNA binding proteins. Many of these proteins can be classified by the presence of a conserved DNA binding motif (1, 2, 3). The POU domain is a 160 amino acids domain that characterizes a family of transcription factors that exert crucial developmental functions (reviewed in: 4, 5, 6). The POU domain is composed of a 60 amino acids POU-type homeodomain (POUHD) and a 74-82 amino acids POU-specific (POUs) domain, connected by a 15-27 amino acids variable spacer region. Whereas the POUHD is distantly related to the classic homeodomain, the POUs domain is unique to the POU protein family. In contrast to the To whom correspondence should be addressed + Present address: Department of Molecular and Cell
classic homeodomain proteins, DNA binding requires the entire POU domain. Both the POUS domain and POUHD directly contact the DNA and determine the sequence specificity and binding affinity (7, 8, 9, 10, 11, 12). The POU domain sequence is about 40% conserved between POU domain proteins. Based upon sequence comparison, POU proteins have been classified in five different groups (4). However, the recent cloning of a zebrafish POU domain cDNA, POU[C], that is difficult to classify in the existing groups suggests that there are six classes (T. Johansen pers. comm.). POU domain transcription factors are required for the determination and proliferation of specific cell types. Unc-86 is a Caenorhabditis elegans developmental control gene (13) that is required for the development of three neuroblast lineages (14). In these lineages unc-86 is critical for daughter cells to become different from their mothers. Moreover, unc-86 is also required for the determination of specific neural identities. Direct evidence for a role of POU proteins in the specification and proliferation of specific cell-types in mammals is provided by mutations in the Pit-I (GHF-1) (15, 16) locus in dwarf mice. Two allelic null mutations in the Pit-i gene abort specification of three of the five cell types in the anterior pituitary and thus prevents normal development of the pituitary gland resulting in dwarfism (17). A role for SCIP (Oct-6, Tst-1) (18, 19, 20, 21) in cellular proliferation has been suggested since its expression is high in rapidly dividing Schwann cell progenitors but is down regulated during cell differentiation (22). Similarly, Oct-4 (Oct-3) and Oct-6 expression is high in proliferating, undifferentiated embryonal carcinoma and embryonic stem cells, but expression is extinguished when the cells are induced to differentiate by retinoic acid (19, 23, 24, 25, 26, 27). All these data suggest an, either direct or indirect role for POU proteins in cellular DNA replication. However, direct proof for such a function is still lacking. Adenovirus (Ad) DNA replication is an attractive model system to study molecular mechanisms of eukaryotic DNA replication as this process can be reconstituted with highly purified components (28, 29, 30, 31). Initiation of DNA replication takes place at either terminus of the 36 kbp linear DNA genome by a protein priming mechanism. The first nucleotide, a dCMP
*
Biology, University of California, Berkeley, USA
6370 Nucleic Acids Research, 1992, Vol. 20, No. 23 residue, is covalently coupled to a precursor of the terminal protein (pTP). The 3'-OH group of this dCMP serves as a primer for further elongation by a strand displacement mechanism. In vitro replication requires three viral proteins, pTP, DNA polymerase (pol) and DNA binding protein (DBP) as well as two cellular transcription factors NFI/CTF (32, 33, 34, 35) and NFIII/Oct-I (36, 37, 38). In the Ad origin of DNA replication two functional regions can be recognized, a core origin that is absolutely required and an auxiliary region that is recognized by NFI/CTF and Oct-1. Both give a strong stimulation of the initiation reaction. Deletion analysis of Oct-I demonstrated that the POU domain is sufficient for activation of DNA replication (8). Since additional domains of Oct-I are required for transcriptional activation (39, 40), these results indicate that although Oct-I functions in DNA replication as well as transcription, the mechanisms underlying these processes are distinct. Similar observations have been made for NFI (41, 42). Binding of Oct-I to the origin is mediated by the conserved octamer motif. Although the Oct-I POUHD can bind the Ad2 octamer site it does not activate DNA replication, implying that the POUs domain is essential in this process (8). In order to determine a direct role in DNA replication of other POU domain proteins we have tested their potential to enhance Ad DNA replication. In addition, we have studied the mechanism of stimulation by the POU domain.
MATERIALS AND METHODS Expression and purification of recombinant POU domain proteins Most of the proteins were expressed in bacteria as fusions with glutathione S-transferase (GST) (43). The Oct-I POU domain encoding HincII-PflMI fragment from plasmid pl-440 (8) was cloned in pGEX-2T (43). The GST-Oct-6 POU domain fusion expression plasmid was a generous gift from Dr Dies Meijer. The HindIfl-XbaI fragment from CMV-Oct-4(+) (25), kindly provided by Dr Hans Scholer was cloned in the pGEX-2T derivative pRP265 (kindly provided by Drs Cees Vink and Ronald Plasterk). The zebrafish POU[C] expression plasmid was a generous gift from Dr Tjerje Johansen. Pit-l was expressed in a pET expression system (44). A plasmid expressing the Pit-I DNA binding domain was kindly provided by Dr Holly Ingraham (9). Oct-2 was expressed in recombinant vaccinia virus infected HeLa cells (8) and purified essentially as described (45). For bacterial expression the strain BL21 (DE3) was used (44). The GST fusion proteins were purified as follows. Overnight cultures were diluted 1:20 in LB medium to a volume of 21 and grown at 37°C until an OD6W of 1.0 was reached. Next isopropylthiogalactopyranoside (IPTG) was added to 0.4 mM. Three hrs. after induction, cells were harvested by centrifugation at 4'C and resuspended in 40 ml cold buffer A (50 mM Tris-HCl, pH 8.0; 250 mM NaCl; 1 mM EDTA; 5 mM dithiotreitol (DTT); 10mM metabisulfiet (MBS); 0.5 mM phenyl-methylsulfonyl fluoride (PMSF); and 0.5 mM L- 1-chloro-3-(4-fosylamino)-4-phenyl-2butanone (TPCK)). After a freeze/thaw step, 2 ml of a 7 mg/ml fresh lysozyme stock was added. After a 30 min. incubation at room temperature (RT), the extract was frozen for the second time at -80°C and thawed at RT. DNaseI and RNaseA were added to final concentrations of 10 jg/ml and 1 yig/ml, respectively. After incubation for 30 min. on a rotating wheel, Triton X-100 was added to a final concentration of 1 %. Unless indicated otherwise all subsequent steps were at 4°C or on ice.
Insoluble material was removed by centrifugation (100,000 x g). Extracts were diluted with buffer B (25 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES)-KOH, pH 7.6; 10% glycerol; 10 mM MBS; 5 mM DTT; 0.5 mM PMSF, 0.5 mM TPCK 0.1 S% Nonidet P40 (NP40)) to 75 mM NaCl and loaded onto a 12 ml glutathione-agarose column (Sigma). The column was washed with 50 ml buffer B/75 mM NaCl/ 1 % NP-40 and eluted with buffer B containing 50 mM NaCl, 0.1 % NP-40 and 5mM glutathione. The elution of GST-POU domain fusion proteins was monitored by bandshift assays. Peak fractions were pooled, diluted to 50 mM NaCl and brought to pH 6.5 by addition of 2-[N-morpholino]ethanesulfonic acid (MES) buffer, pH 6.2 and applied to a 5 ml fast flow S column. The column was washed with buffer C (25 mM MES, pH 6.5; 10% glycerol; 5 mM DTT; 50 mM NaCl; 0.01 % NP-40; 2 mM MBS; 0.2 mM PMSF; and 0.2 mM TPCK) and eluted with a linear gradient of 50 to 500 mM NaCl. Octamer binding activity eluted at 100 mM NaCl (POU[C]), 175 mM NaCl (Oct-l and Oct-6) or 250 mM NaCl (Oct-4). By addition of Tris-HCl pH 8.8, the pH of the samples was raised to 8.0 and the GST part was cleaved of by thrombin (Sigma) at RT. For Oct-4 this step was omitted since it gave rise to several additional breakdown products that obscured the interpretation of the results. The POU domain was separated from
0 :
*.,.,
4
-
AA.A
G,,
..
Figure 1. Comparison of the binding site preference of different POU domain proteins. Binding of Pit- 1, Oct- 1, Oct-2, Oct-6, and Oct-4 to various recognition sequences was tested in a bandshift experiment. Sequences of the labelled oligonucleotides were derived from the prolactin promoter (PIP), the canonical octamer motif in the Ad4 origin, the degenerated octamer in the Ad2 origin, the HSV ICP4 TAATGARAT motif, the heptamer-octamer element in the IgH promoter (H+O+) and the isolated heptamer site (Hept). Monomeric or dimeric protein-DNA complexes are indicated with Ci or C2, respectively. Oct-2 was purified from recombinant vaccinia infected HeLa cells. The other proteins were expressed in bacteria and purified as described (Materials and Methods). Oct-2 was a full-length protein, Oct-4 was an almost full-length protein fused to GST, for Oct- I and Oct-6 the isolated POU domains were used and for Pit- 1 the POU domain containing a small C-terminal extension (see Materials and Methods for details). (B) Sequences of the POU domain protein binding sites and summary of the results. The core sequences of the binding elements are shown in bold. Gels were quantified as described (Materials and Methods) and binding affinities are indicated.
Nucleic Acids Research, 1992, Vol. 20, No. 23 6371 the GST moiety and thrombin on a 2 ml fast flow S column equilibrated with buffer D (buffer C with MES pH 6.8 instead of 6.5). The column was washed with buffer D/50 mM NaCl and eluted with a linear gradient of 50 to 500 mM NaCl. Peak fractions were used for DNA binding and replication assays. Pit-I was purified from a 41 culture. After induction and lysis as described above, the extract was applied to a 30 ml DEAE column equilibrated with buffer E/300 mM NaCl (buffer C with 50 mM Tris, pH 8.0 instead of MES). The flowthrough plus 30 ml wash were pooled, diluted with buffer F (buffer C with 25 mM HEPES-KOH, pH 7.6 instead of MES) to 50 mM NaCl and applied to a 12 ml fast flow S column. The column was washed with buffer F/50 mM NaCl and eluted with a linear gradient of 50 to 1000 mM NaCl. Pit-I eluted at 480 mM NaCl. Fractions were pooled, diluted to 50 mM NaCl and loaded onto a 8 ml heparin-sepharose column. The column was washed with buffer F/50 mM NaCl and eluted with a linear gradient of 50 to 1000 mM Nacl. The Pit-I peak fraction eluted at 340 mM NaCl and was used in the replication and bandshift assays. All the protein preparations were nearly homogeneous as determined by SDS-PAGE followed by staining with coomassie brilliant blue (data not shown). DNA binding studies The probes used for bandshift assays were oligonucleotides, endlabelled with T4 polynucleotide kinase and purified by preparative polyacrylamide gel electrophoresis (46). The binding sites are shown in Figure 1B and the oligonucleotides have been described before: PIP (9) Ad4, Ad2 (10), ICP4 (47), Hept (= H+O-) and H+O+ (48). Binding reactions were carried out for 30 min. at RT in 20 Al of binding buffer (20 mM HEPES-KOH (pH 7.5); 100 mM NaCl; 1mM EDTA; 1mM DTT; 0.025% NP-40; 4% Ficoll and I Ag poly (dI-dC)). Free DNA and protein-DNA complexes were analyzed on a 6% polyacrylamide gel run in 0.5 x TBE (45 mM Tris, 45 mM boric acid, 1 mM EDTA) containing 0.01 % NP-40. Gels were dried and exposed. For quantitative analysis, 3 fmol of Ad2 probe was added to the reaction. One binding unit is defined as the amount of protein that could bind 50% of the probe in a bandshift assay. Quantification was by Cerenkov counting of excised gel slices. Ad DNA replication in vitro TP-DNA complex, pTP-pol, DBP and NFI-BD were purified as described (49, 50). Reaction mixtures (15 Al) contained 30 ng of XhoI-digested AdS DNA-TP (2.1 fmol each of NFI and Oct-I binding sites), 0.9 jig DBP, 20 mU pTP-pol that equals 7 ng pTP-pol (49) and, when added to the reaction, 3 ng NFI-BD in a buffer containing 25 mM HEPES-KOH (pH 7.5); 0.4 mM DTT; 2 mM ATP; 4 mM Mg C12 40 yM of dGTP, dATP and dTTP and 2.5 liM [ai-32P]dCTP (5.5 Ci/mmol.). Incubations were for lhr at 37 C and reactions were stopped by addition of 1.5 Al stopmix (40% sucrose, 1 % SDS, 0.1 % bromophenol blue). Template commitment experiment were performed as described before (50). Products were analyzed on a 1 % agarose gel run in 0.5 xTBE containing 0.1 % SDS. After electrophoresis the gels were dried and exposed. Quantification was by Cerenkov counting of gel slices corresponding to the double-stranded and single-stranded replication products. Stimulation was plotted against the amount of each protein present as described before (56), The value of activation at 2.0 binding units was used to calculate the relative activity shown in figure 2B. When 2.0 binding units of POU protein are added, approximately 80% of
the maximal level of stimulation is reached. The values in figure 2B are the result of three independent experiments.
RESULTS Distinct POU domain proteins have overlapping but dissimilar DNA binding specificities On the basis of sequence similarity, the POU domain proteins have been classified into five groups (4). The class II proteins Oct-1 (51) and Oct-2 (52, 53, 54, 55) show 88% identity in their POU domains. Pit-I (class I), Oct-6 (class III), and Oct4 (class V) are 66%, 68% and 56% identical to the Oct-I POU domain, respectively. The total identity between the POU domains of these proteins is about 43 %. We expressed the intact proteins or their DNA binding domains in bacteria (Materials and Methods) and, after purification, compared their recognition sequence preferences on several natural response elements in a bandshift assay (Figure 1A). Binding is clearly specific and complexes formed could be competed away with an excess of cold oligonucleotides (data not shown). As described before (9), Pit-I binds very efficiently as a dimer (complex C2) to the prolactin promoter PIP element. The canonical octamer (Ad4), albeit with lower afflnity. Surprisingly,
A Oct-l binding units
B/C
ss-B
ss-C
Pit-I
Oct-2
Oct-4
Oct-6
POU[C]
-
I
I
I
,' ou oob -
;
°e4' °6°
:-
:2 Xs
.-
*
..
so
"
40
-
4d
c=
Itz sx .11;: " n
-
.w A-M
a*
..
.0 mm
so
W so
so a* so, ow
M
de
-
0 lo
en
: .: =
-
-
-
"
O.-
MAO
am
B activity
Oct-i
Pit-i
100% 80%
Oct-2 Oct-6
110%
Oct4
90%
POUiC]
55%
115%
Figure 2. Stimulation of Ad DNA replication by POU domain proteins. (A) In vitro replication of AdS DNA-TP digested with Xhol. The origin containing fragments, B and C as well as displaced single stranded fragments, ss-B and ssC, are indicated. Reaction mixtures contained, in addition to the DNA-TP template (30 ng), NFI-BD (3 ng), DBP (0.9 Mg) and pTP-pol (20 mU). Purified recombinant POU domain proteins were tested for activation of replication. Varying amounts of each POU factor were added to the reaction mixture, expressed as binding units. One binding unit is defined as the amount of protein that can bind 50% (1.5 fmol) of an Ad2 octamer probe in a bandshift assay. (B) Activities of the various POU domain proteins were calculated as the ratio a[32P]dCMP incorporation relative to the amount of protein. This allows a comparison of the stimulation by other POU domain proteins with that by Oct-1. The stimulation in the presence of 2.0 binding units was used to calculate the relative activities of the POU proteins. The activity of Oct-i was assigned a value of 100%. No stimulation of DNA replication was set at 0% activity. Data are from the experiment shown in (A) and at least two additional independent assays. The variation between the different experiments was never more than 10%.
6372 Nucleic Acids Research, 1992, Vol. 20, No. 23 Pit-I has a considerable affinity for degenerated Ad2 site and the ICP4 TAATGARAT motif, the latter is a weak Oct-I binding site. The mobility of the complexes indicates that Pit-I binds to these motifs as a monomer (complex CI). The affinity for the isolated heptamer site (Hept) is very low. However, Pit-I dimerizes very efficientdy on the intact heptamer-octamer element (H+O+) from the immunoglobulin heavy chain promoter. Apparently, this element allows the Pit-I POU domain to bind in such an orientation as to allow dimerization and cooperative DNA binding. This supports the notion that the sequence relationship between the PIP site and the heptamer-ctamer element is functional. Consistent with their closely related DNA binding domains, Oct-I and Oct-2 demonstrate an identical preference for binding sequences. Under the conditions used, binding to the PIP element or the heptamer sequence is hardly detectable. Also the affinity for the TAATGARAT motif is low compared to the canonical octamer. In contrast, Oct-6, like Pit-i, has a high affinity for the ICP4 TAATGARAT motif (see also ref. 19). Oct-6 binds very weakly to the PIP element and the isolated heptamer. The relative sequence preference of Oct-4 is most similar to that of Oct-I and Oct-2 but there are some differences. Oct-4 dimerizes more efficiently on the heptameroctamer element as well as on the PiP element. Like Oct-I and Oct-2 and in contrast to Pit-I and Oct-6, Oct-4 has only a weak affinity for the TAATGARAT motif. Pit-i binds the prolactin PiP element much more efficient than the other POU domain proteins and Oct-4 has an intermediate affinity for this element. However, at higher protein concentrations, the other POU proteins also bind to this motif but dimerization is very poor (data not shown). All POU factors that we tested can bind to the heptamer-octamer element as a dimer. Comparison of binding to the H+O+ and Hept probes indicates that binding is cooperative in all cases. These data are summarized in Figure 1B and demonstrate that the POU domain proteins Pit-i, Oct-I/Oct-2, Oct-6 and Oct-4 share DNA recognition motifs but with differing preferences.
POU transcription factors stimulate DNA replication We assayed purified Oct-i, Pit-i, Oct-2, Oct-6, Oct-4 and zebrafish POU[C] or their POU domains (see Materials and Methods) for their potential to activate Ad DNA replication in
B/c
I
=
_:
5s--B I
-,
Flgure 3. POU domain proteins function independently of NFI. The various POU domain transcription factors were tested for stimulation of DNA replication in the absence of NFI-BD. Conditions were similar to those described above only NFI-BD was omitted. The amount of POU domain protein added corresponds to the maximal amounts used in Figure 2A.
a reconstituted system. To allow a comparison of the stimulation by each POU protein, binding activity was determined in bandshift assays, using the Ad2 octamer as a probe (data not shown). Ad5 DNA-TP complex digested with XhoI was used as a template in the replication assay. Furthermore, pTP-pol complex, DBP and NFI were added to the reaction. Figure 2A shows the effects of various POU transcription factors on DNA replication. The origin containing fragments B and C are
indicated. When replication is very efficient, labelled singlestranded B and C fragments are displaced (ss-B and ss-C) during second and further rounds of replication. Increasing amounts of each POU domain protein were added to the replication reaction. The Oct-I POU domain stimulated Ad replication in a dosedependent manner up to 10-fold (lanes 1 to 4). All other POU domain proteins tested also stimulated DNA replication albeit at different levels. In Figure 2B, the level of replication activity of these proteins is compared with that of Oct-i. Activity was calculated relative to the amount of POU domain protein added (2.0 binding units). Stimulation by Pit-I is slightly reduced compared to Oct-i. Oct-2, Oct-6 and Oct-4 enhance the level of DNA replication to the same extent as Oct-1. POU(C] is the only POU domain protein with a significantly lower level of stimulation, 55% ofthe Oct-I activity. POU[C] reaches a plateau when more than 2.0 binding units are added to the assay. This is most likely due to binding to a second low affinity site that is present in the core origin which causes inhibition of replication. This site is hardly occupied at 2.0 binding units and when lower amounts of POU polypeptides were added, stimulation by POU[C] remained at about 55% of the stimulation by Oct-1. Control experiments showed that an intact octamer element was critical for stimulation (data not shown). Oct-1 and NFI function independently in the replication reaction (56). In agreement with this, other POU domain transcription factors could also support Ad DNA replication in the absence of NFI (Figure 3). The level of stimulation by POU domain proteins in the absence or presence of NFI was similar. POU domain does not enhance binding of pTP-pol to the origin Both for NFI and for Oct-i, the domains required for DNA binding and replication coincide (8, 41, 42). The DNA binding domain of NFI (NFI-BD) interacts directly with the pTP-pol complex (56, 57, 58). NFI-BD tethers pTP-pol to the origin, increases the specific binding of pTP-pol and mediates the formation of a stable preinitiation complex, thereby enhancing the initiation reaction (50). Analogously, we have studied the effect of the Oct-I POU domain on the binding of pTP-pol to the origin (Figure 4A). In agreement with previous results (50, 59), pTP-pol binds weakly to the Ad2 origin (lanes 2 and 6). Addition of a saturating amount of POU domain, that is optimal in the DNA replication reaction, does not affect binding of pTPpol (compare lanes 2 and 3). In contrast, NFI-BD strongly increases binding of pTP-pol to the origin (compare lanes 6 and 7). We also studied the effects of the POU domain on the stability of the preinitiation complex in a functional assay (Figure 4B). Two closely related virion DNA templates were used, Ad2 and AdS. Their origins are identical but it is possible to generate origin containing fragments of different lengths (AdS-C: 5505 bp; Ad2-C: 4140 bp). Ad5 was preincubated with pTP-pol and either NFI-BD or POU domain. The Ad2 template and other replication components were added together with the nucleotides to start
Nucleic Acids Research, 1992, Vol. 20, No. 23 6373 replication. When a stable preinitiation complex is formed and when the amount of pTP-pol is limiting, replication of the competing Ad2 template should be reduced compared to the preincubated AdS template. Without preincubation Ad2 and AdS were replicated equally well (lanes 1 and 4). Preincubation in the presence of NFI leads to a higher level of replication of Ad5 compared to Ad2 (lanes 5 and 6). If Ad2 was preincubated and AdS used as a challenging template, replication of Ad2 was much more productive than AdS (data not shown). The total level of replication with or without preincubation is equal in all reactions,
A pTP-pol - + + NFI-BD - - + +
pTP-pol - + + Pou - - + + POU-pTP-polI pTP-pol
NFI-pTP-pol > pTP-polo.
-
*_l
POU3
free DNA> _Miu _
1 2 3 4
DISCUSSION POU proteins have related but differing preferences of
a
NFI-BD>
freeDNAp -
5 6 7 8
B POU
START REPLICATION
t=
Ad5 * Ad2 *
0
10 20
-----
1
2 3
NFI-BD 0 10
20
- '
4
d
4
demonstrating that preincubation does not affect the replication components. These results show that NFI mediates the formation of a template committed complex (50). In contrast, the POU domain had no effect on the stability of the preinitiation complex (lanes 1 to 3). In agreement with these functional experiments we could never detect an interaction between the POU domain and the pTP-pol complex in coimmunoprecipitation and crosslinking experiments, whereas the NFI/pTP-pol interaction was readily detectable (data not shown, 49, 50). Moreover, no direct protein-protein interactions nor functional cooperation between the POU domain and the other replication proteins, NFI and DBP, have been detected (50, 60, 61, data not shown). We conclude that the POU domain activates DNA replication after formation of a preinitiation complex and does not interact with the pTPpol complex.
5
o
6
Figure 4. The POU domain does not enhance binding of pTP-pol to the origin. (A) Influence of the POU domain on binding of pTP-pol to the origin DNA was determined by a bandshift assay, using the labelled Ad2 origin as a probe. The positions of free DNA as well as the various DNA-protein complexes are indicated.
sequence recognition Within families of eukaryotic transcription factors, different members often demonstrate similarities in their DNA binding specificities. This has been observed for the AP-1 family, the hormone receptors and for the Drosophila homeodomain proteins (1, 2). We have compared the DNA binding specificities of several POU transcription factors, Pit-i, Oct-i, Oct-2, Oct-6 and Oct-4. Our results revealed that they recognize common sequences but with a different order of preference. Moreover, POU proteins can associate cooperatively on dimeric target sequences as homodimers (9, 48, 62, 63) but also as heterodimers (61, 64). Thus, it is probable that many target sequences do not have one unique cognate POU transcription factor. Rather, depending on their expression pattern and sequence preferences, competition and cooperation among different POU proteins would determine the transcriptional effects of natural response elements. This provides a strategy for differential gene regulation by a network of related transcription factors. In addition to diversification at the level of DNA binding, specific proteinprotein interactions provide an alternative mechanism of regulation. The POU domain protein family demonstrates several examples of this principle. First, the Oct-I POUHD can specifically interact with the potent viral transactivator VP-16/Vmw65 (65, 66). Second, the POUHD of the non DNA binding I-POU protein has a strong interaction with the POUHD of transcription factor Cfl-a that aborts DNA binding by this factor (67, 68). Finally, the distinct types of transcriptional activation domains in different POU domain proteins can result in promoter-specific transactivation (40). POU domain proteins are DNA replication factors DNA replication and transcription are regulated through the function of sequence-specific DNA binding proteins. Evidence is accumulating that several of these proteins function in both processes
which is reflected by the common elements shared by
transcriptional control regions and origins of DNA replication.
(B) The POU domain does not contribute to the formation of a template-committed Man Xbaldigested Ad5 DNA-TP was preincubated in the preseey vral but also cellular eukaryotic origins are organized in two functional regions: a core region that is essential for origin of pTP-pol and POU domain for 0 (lane 1), 10 (lane 2) or 20 minutes (lane 3) function and determines the site of initiation, and an adjacent prior to the addition of the competing template (Xbal digested Ad2 DNA-TP), the other replication proteins, NFI and DBP, and dNTPs to start replication. The auxiliary region that is important for efficient replication (69, positions of the replicated origin containing fragments, Ad5-C (5508 bp) and Ad2-C ..c. r 70, 71, 72). Within (4140 bp) are indicated. In the second panel AdS DNA-TP was preincubated appropriate context these auxliary regions in the presence of pTP-pol and NFI for the same time periods, followed by addition can also function as transcriptional enhancers. There are several of Ad2 DNA-TP, dNTPs and DBP and POU domain (lanes 4 to 6). examples of transcription factors that function in the control of preinitiated complex.
an
6374 Nucleic Acids Research, 1992, Vol. 20, No. 23 DNA replication. In the yeast Saccharomyces cerevisiae, ABF1 is important for autonomous replicating sequence (ARS) function but it also directs gene expression via silencers and upstream activating sequences (72, 73, 74). The yeast protein MCM1 regulates mating-type-specific genes and is also required for DNA replication (75, 76). Replication of simian virus 40 (SV40) and polyoma virus origins is stimulated by specific transcription factors (71, 77, 78). Further examples of multifunctional DNA binding proteins are provided by the transcription factors NF/CTF and Oct-I that activate Ad DNA replication. The SV40 enhancer contains two octamer motifs that have also been implicated in viral DNA replication (79). In this study we show that, in addition to Oct-i, divergent POU domain transcription factors stimulate Ad DNA replication in vitro. These data suggest the possibility that POU proteins have a dual function in both cellular DNA replication and transcription. Consistent with this notion, POU domain proteins are highly expressed during periods of rapid cellular divisions of several cell-types. Moreover, genetic evidence demonstrated that Pit-i and unc-86 are critical for proliferation of specific cell-types (14, 17). It must be noted, however, that it also possible that adenovirus simply utilizes hostcell proteins for its replication but that this does not reflect cellular processes. Direct proof for a function of POU domain proteins (and other transcription factors) in cellular DNA replication could come from depletion studies, similar to those performed for p34cdc2 in Xenopus Laevis egg extracts (80) or from a further analysis of cellular origins of DNA replication. What is the mechanism by which the POU domain activates DNA replication? Several not mutually exclusive mechanisms by which transcription factors stimulate the initiation of replication can be envisioned. These factors could have a specific interaction with other components of the replication machinery and, for example, stabilize the preinitiation complex, as demonstrated for NFI (50). The bovine papilloma virus (BPV) transcription factor E2 brings the replication protein El to the origin by a direct interaction (81, 82) Alternatively, activation could result from displacement of inhibitory components, like nucleosomes (83). Finally, DNA binding per se could be sufficient and act via alteration of the DNA structure, as proposed for the SV40 large T antigen (84, 85, 86). Nucleosomal structures are not involved in the in vitro Ad DNA replication system. The possibility that Oct-I acts via direct protein-protein interactions with other replication proteins is highly unlikely. No such interactions were observed in various assays, whereas the specific association of NFI and pTP-pol was readily detected. Finally, in contrast to NFI-BD, the POU domain does not stabilize binding of pTP-pol to the origin DNA. Thus, the POU domain probably activates a rate limiting step in the initiation reaction after formation of the preinitiation complex. In agreement with the absence of specific interactions, divergent POU domain proteins are capable of stimulation of DNA replication. This is in contrast to the interaction between the POUHD and VP-16 that is aborted by a limited number of residue changes between the highly related Oct-I and Oct-2 POU domains (65). Previously we showed that the POUs domain is essential for stimulation of Ad DNA replication (8). Furthermore, binding of the POU domain, but not binding of the POUHD, induces bending of the DNA implying that the POUs domain is critical (87). Although there is no direct evidence that these two functions of the POUs domain are related, all data are consistent with a model in which the POU domain stimulates DNA replication through a
conformational change in the DNA structure. We are currently investigating whether DNA bending plays functions in the activation of Ad DNA replication.
ACKNOWLEDGEMENTS We are very grateful to Drs Holly Ingraham, Dies Meijer, Tjerje Johanson, Hans Sch6ler, Winship Herr, Patrick Mathias, Walter Schaffner, Cees Vink and Ronald Plasterk for making the various plasmids available to us. This work was supported in part by the Netherlands Foundation for Chemical Research (SON), with financial support from the Netherlands Organization for Scientific research (NWO).
REFERENCES 1. Mitchell, P.J., and Tjian, R. (1989) Science, 245, 371-378. 2. Johnson, P.F., and McKnight, S.L. (1989) Annu. Rev. Biochem., 58, 799-839. 3. Harrison, S.C. (1991) Nature, 353, 715-719. 4. Rosenfeld, M.G. (1991) Genes Dev., 5, 897-907. 5. Ruvkun, G., and Finney, M. (1991) Cell, 64, 475-478. 6. Scholer, H.R. (1991) Trends in Gen., 7, 323-329. 7. Sturm, R.A., and Herr, W. (1988) Nature, 336, 601-604. 8. Verrijzer, C.P., Kal, A.J., and Van der Vliet, P.C. (1990a) EMBO J., 9, 1883-1888. 9. Ingraham, H.A., Flynn, S.E., Voss, J.W., Albert, V.R., Kapiloff, M.S., Wilson, L., and Rosenfeld, M.G. (1990) Cell, 61, 1021-1033. 10. Verrijzer, C.P., Kal, A.J., and Van der Vliet, P.C. (1990b) Genes Dev., 4, 1964-1974. 11. Kristie, T.M., and Sharp, P.A. (1990) Genes Dev., 4, 2383-2396. 12. Aurora, R., and Herr, W. (1992) Mol. Cell. Biol., 12, 455-467. 13. Finney, M., Ruvkun, G., and Horvitz, H.R. (1988) Cell, 55, 757-769. 14. Finney, M., and Ruvkun, G. (1990) Cell, 63, 895-905. 15. Bodner, M., Castrillo, J.L., Theill, L.E., Deerinck, T., Ellisman, M., and Karin, M. (1988) Cell, 55, 505-518. 16. Ingraham, H.A., Chen, R., Mangalam, H.J., Elsholtz, H.P., Flyn, S.E., Lin, C.R., Simmons, D.M., Swanson, L., and Rosenfeld, M.G. (1988) Cell, 55, 519-529. 17. Li., S., Crenshaw, E.B., Rawson, E.J., Simmons, D.M., Swanson, L.W., and Rosenfeld, M.G. (1990) Nature, 347, 528-533. 18. Monuki, E.S., Weinmaster, G., Kuhn, R., and Lemke, G. (1989) Neuron3, 783-793. 19. Meijer, D., Graus, A., Kraay, R., Langeveld, A., Mulder, M.P., and Grosveld, G. (1990) Nucleic Acids Res., 18, 7357-7365. 20. Suzuki, N., Rohdewohld, H., Neuman, T., Gruss, P., and Sch6ler H.R. (1990) EMBO J., 9, 3723-3732. 21. He, X., Gerrero, R., Simmons, D., Park, R. E., Lin C.R., Swanson, L.W. and Rosenfeld, M.G. (1991) Mol. Cell. Biol., 11, 1739-1744. 22. Monuki, E.S., Kuhn, R., Weinmaster, Trapp, B., Lemke, G. (1990) Science, 249, 1300-1303. 23. Okamoto, K., Okazawa, H., Okuda, A., Sakai, M., Muramatsu, M., and Hamada, H. (1990) Cell, 60, 461-472. 24. Scholer, H.R., Ruppert, S., Suzuki, N., Chowdhury, K., and Gruss, P. (1990a) Nature, 344, 435-439. 25. Scholer, H.R., Dressler, G.R., Balling, R., Rohdenwohld, H., and Gruss, P. (1990b) EMBO J., 9, 2185-2195. 26. Suzuki, N., Rohdewohld, H., Neuman, T., Gruss, P., and Scholer H.R. (1990) EMBO J., 9, 3723-3732. 27. Rosner, M.H., Vigano, M.A., Ozato, K., Timmons, P.M., Poirier, F., Rigby, P.W.J., and Staudt, L.M. (1990) Nature, 345, 686-692. 28. Stillman, B.M. (1989) Annu. Rev. Cell. Biol., 5, 197-245. 29. Challberg, M.D. and Kelly, T.J. (1989) Annu. Rev. Biochem., 58, 671-717. 30. Hay, R.T., and Russell, W.C. (1989) Biochem. J., 238, 3-16. 31. Van der Vliet, P.C. (1991) Semin. Virol., 2, 271-280. 32. Nagata, K., Guggenheimer, R.A., Enomoto, T., Lichy, J.H., and Hurwitz, J. (1982) Proc. Natl. Acad. Sci. USA, 79, 6438-6442. 33. Jones, K.A., Kadonaga, J.T., Rosenfeld, P.J., Kelly, J.T., and Tjian, R.T. (1987) Cell, 48, 79-89. 34. Santoro, C., Mermod, N., Andrews, P.C., and Tjian, R.T. (1988) Nature, 334, 218-224.
Nucleic Acids Research, 1992, Vol. 20, No. 23 6375 35. Meisteremst, M., Rogge, L., Foechler, L., Karaghiosoff, M., and Winnacker, E.L. (1989) Biochemistry, 28, 8191-8200. 36. Pruijn, G.J.M., van Driel, W., and van der Vliet, P.C. (1986) Nature, 322, 656-659. 37. Pruijn, G.J.M., Van der Vliet, P.C., Dathan, N.A., and Mattaj, I.W. (1989) Nucleic Acids Res., 17, 1845-1863. 38. O'Neill, E.A., Fletcher, C.,Burrow, C.R., Heintz, N., Roeder, R.G., and Kelly, T.J. (1988) Science, 241, 1210-1213. 39. Tanaka, M., and Herr, W. (1990) Cell, 60, 375-389. 40. Tanaka, M., Lai, J.-S., and Herr W. (1992) Cell, 68, 755-768. 41. Mermod, N., O'Neill, E.A., Kelly, T.J., and Tjian, R.T. (1989) Cell, 58, 741-753. 42. Gounari, F., De Francesco, R., Schmidt, J., Van der Vliet, P.C., Cortese, R., and Stunnenberg, H.G. (1990) EMBO J., 9, 559-566. 43. Smith, D.B., and Johnson, K.S. (1988) Gene 67: 31-40. 44. Studier, F.W., Rosenberg, A.H., Dunn, J.J., and Dubendorf, J.W. (1990) Meth. in Enzym., 185, 60-89. 45. Poellinger, L., and Roeder, R.G. (1989) Mol. Cell. Biol., 9, 747-756. 46. Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory. Cold Spring Harbor, N.Y. 47. O'Hare, P., Goding, C.R. (1988) Cell, 52, 435-445. 48. Kemier, I., Schreiber, E., Muller, M.M., Matthias, P. and Schaffner, W. (1989) EMBO J., 8, 2001-2008. 49. Mul, Y.M., Van Miltenburg, R.T., De Clercq, E., and Van der Vliet, P.C. (1989) Nucleic Acids Res., 17, 8917-8929. 50. Mul, Y.M., and Van der Vliet, P.C. (1992) EMBO J., 11, 751-760. 51. Sturm, R.A., Das, G., and Herr, W. (1988) Genes Dev., 2, 1582-1599. 52. Clerc, R.G., Corcoran, L.M., LeBowitz, J.H., Baltimore, D., and Sharp, P.A. (1988) Genes Dev., 2, 1570-1581. 53. Muiller-Immergluck, M.M., Ruppert, S., Schaffner, W., and Matthias, P. (1988) Nature 336, 544-551. 54. Scheidereit, C., Cromilich, J.A., Gerster, T., Kawakami, K., Balmaceda, R.A., Currie, R.A., and Roeder, R.G.(1988) Nature, 336, 551-557. 55. Ko, H.-S., Fast, P., McBride, W., and Staudt, L.M. (1988) Cell, 55, 135-144. 56. Mul, Y.M., Verrijzer, C.P., and Van der Vliet, P.C. (1990) J. Virol., 64, 5510-5518. 57. Bosher, J., Robinson, E.C., and Hay, R.T. (1990) New Biol., 2, 1083-1090. 58. Chen, M., Mermod, N., and Horwitz, M. (1990) J. Biol. Chem., 265, 18634-18642. 59. Temperley, S. and Hay, R.T. (1992) EMBO J. 11, 761-768. 60. Stuiver, M. and Van der Vliet, P.C. (1990) J. Virology, 64, 379-386. 61. Verrijzer, C.P., van Oosterhout, J.A.W.M., and van der Vliet, P.C. (1992) Mol. Cell. Biol., 12, 542-551. 62. LeBowitz, J.H., Clerc, R.G., Brenowitz, M., Sharp, P.A. (1989) Genes Dev., 3, 1625-1638. 63. Poellinger, L., Yoza, B.K., and Roeder, R.G. (1989) Nature, 337, 573-576. 64. Voss. J.W., Wilson L., and Rosenfeld, M.G. (1991) Genes dev., 5, 1309-1320. 65. Stem, S., Tanaka, M., and Herr, W. (1989) Nature, 341, 624-630. 66. Goding, C.R., and O'Hare, P. (1989) Virology, 173, 363-367. 67. Treacy, M.N., He, X., and Rosenfeld, M.G. (1991) Nature, 350, 577-584. 68. Treacy, M.N., Nelson, L.I., Turner, E.E., He, X., and Rosenfeld, M.G. (1992) Cell, 68, 491-505. 69. DePamphilis, M.L. (1988) Cell, 52, 635-638. 70. Diffley, J.F., and Stillman, B.M. (1990) Trends Genet., 6, 426-432. 71. Guo, Z.-S., and DePamphilis, M.L. (1992) Mol. Cell. Biol., 12, 2514-2524. 72. Marahrens, Y., and Stillman, B.M. (1992) Science, 255, 817-823. 73. Buchman, A.R., Kimmerly, W.J., Rine, J. and Komberg, R.D. (1988) Mol. Cell. Biol., 8, 210-225. 74. Diffley, J.F.X., and Stillman, B.M. (1988) Proc. Natl. Acad. Sci. USA, 85, 2120-2124. 75. Passmore, S., Elbe, R., and Tye, B.K. (1989) Genes Dev., 3, 921-935. 76. Christ, C., and Tye, B.K. (1991) Genes Dev., 5, 751-764. 77. Bennet-Cook, E.R., and Hassell, J.A. (1991) EMBO J., 10, 959-969. 78. Murakami, Y, Satake, M., Yamaguchi-Iwai, Y., Sakai, M., Muramatsu, M., and Ito, Y. (1991) Proc. Natl. Acad. Sci. USA, 88, 3947-3951. 79. Haas, M.W., Ramanujam, P., Chandrasekharappa, S.C., and Subramania, K.N. (1991) Virology, 180, 41-48. 80. Blow, J.J., and Nurse, P. (1990) Cell, 62, 855-862. 81. Mohr, I.J., Clark, R., Sun, S., Androphy, E.J., Macphearson, P., and Botchan, M.R. (1990) Science, 250, 1694-1699. 82. Yang, L., Li. R., Mohr, I.J., Clark, R., and Botchan, M.R. (1991) Nature, 353, 628-632.
83. Cheng, L., and Kelly, T.J. (1989) Cell, 59, 541-551. 84. Dean, F.B., Bullock, D., Murakami, Y, Wobbe, C.R., Weissbach, L., and Hurvitz, J. (1987) Proc. Natl. Acad. Sci. USA, 84, 16-20. 85. Wold, M.S., Li, J.J., and Kelly, T.J. (1987) Proc. Natl. Acad. Sci. USA, 84, 3643-3647. 86. Sengupta, D.J., and Borowiec, J.A. (1992) Science, 256, 1656-1661. 87. Verrijzer, C.P., van Oosterhout, J.A.W.M., van Weperen W.W., and van der Vliet, P.C. (1991) EMBO J., 10, 3007-3014.
