Gene, 111 (1992) 249-254 © 1992 Elsevier Science Publishers B.V. All rights reserved. 0378-1119/92/$05.00
249
GENE0~86
Protein binding sites within the human thymidine kinase promoter (DNase I footprinting; CCAAT and GC elements; insertion mutants; promoter strength; recombinant DNA)
Santosh S. Areot a'* and Prescott L. Deininger a'b a Department of Biochemistry and Molecular Biology, Louisiana State University Medical Center, New Orleans, LA 70112 (U.S.A.) and b Laboratory of Molecular Genetics, Ochsner Medical Foundation, 1516 Jefferson Highway, New Orleans, LA 70121 (U.S.A.) Received by J.L. Slightom: 14 August 1991 Revised/Accepted: 28 September 1991 Received at publishers: 22 November 1991
SUMMARY
DNase I footprint analysis, using total HeLa cell nuclear extract and purified transcription factor Spl, was carried out to determine the various protein binding sites within the human thymidine kinase promoter. The promoter has two separate CCAAT elements and multiple Spl-binding sites, as well as at least one undefined protein binding site. Detailed analysis of protein binding to the two CCAAT elements showed that changing the spacing between the two CCAAT elements altered both protein binding to the distal CCAAT element as well as promoter activity. Both CCAAT elements can act as functional transcription dements, but are not oriented for optimal promoter strength in the human tk promoter. Our studies show that a promoter fragment that has been previously shown to be the minimal region to maintain a serum responsive promoter regulation apparently contains only a single Spl-binding domain and the more distal of the CCAAT elements.
INTRODUCTION
Thymidine kinase (TK, EC 2.7.1.21) is a crucial enzyme in the salvage pathway of pyrimidine nucleotide biosynthesis. Since this enzyme activity is required only for DNA synthesis, the levels of the enzyme show a high degree of correlation with the proliferative state of the cells (Belle, 1974). Regulation of the gene that encodes this enzyme has
Correspondenceto: Dr. P.L. Deininger, Biochemistry and Molecular Biology, LSUMC, 1901 Perdido St., New Orleans LA 70112 (U.S.A.) Tel. (504)568-4735; Fax (504)568-6158. * Present address: College of Pharmacy, University of Kentucky, Lexington, KY 40536 (U.S.A.) Tel. (606)257-2573. Abbreviations: bp, base pair(s); CAT, chloramphenicoi acetyl transferase; cat, gene (DNA) encoding CAT; nt, nucleotide(s); TK, thymidine kinase; tk. gone (DNA) encoding TK; poIIk, Klenow (large) fragment of the E. coil DNA polymerase I; tsp, transcription start point; USI, unknown site 1, wt, wild type.
been shown to occur at multiple levels which include transcriptional and post-transcriptional mechanisms (Coppock and Pardee, 1987; Ire and Conrad, 1990; Kasid et al., 1986; Sherley and Kelly, 1988; Stewart et al., 1987). More recently, studies on the role of the promoter in the regulation of the gone have shown that sequences within the promoter arc capable of conferring cell cycle dependent regulation onto a heterologous gene (Kim etal., 1988; Lipson etal., 1989; Roehl and Conrad, 1990; Travali et al., 1988). From the nt sequence analysis, we (Flemington etal., 1987), and others (Sauve etal., 1990) have previously identified a number of putative transcriptional elements within the human tk promoter. These include two CCAAT elements at -40 and -71 and several GC elements farther upstream. Subsequent studies using a deletion analysis confirmed these putative elements as functional promoter elements (Areot et al., 1989). Tile specific binding of proteins to the two CCAAT elements using gel retardation assays had previously demonstrated that both CCAAT elements could independently bind the same factor with high
250
A t')
¢'3t"~
I-! i-1
A[ -~
o..o.°o.°" "........
..........tk325 . . . . . . . . . . . . . . . . .
B
tk325
.
tk167 ........
tk167
+
+
MabbaM
MabbaM
GC" GC
US]
dCCAAT
GC" GO" pCCAAT
TATA
affinity (Arcot et al., 1989). Although the deletion analysis suggested the importance of a number of putative GC elements within the tk promoter based upon their ability to contribute to overall promoter strength, the actual role of these sequences to bind Spl (Dynan and Tjian, 1983) had not been demonstrated. The deletion and protein-binding analyses were also not comprehensive, in that they may very well have overlooked binding to other promoter elements which had not been identified from the sequence analysis (Flemington et al., 1987). Finally, although both the CCAAT elements were independently capable of binding a nuclear factor with similar affinities (Arcot et al., 1989), the potential roles for these CCAAT elements in the promoter appeared to be complex. First, it is clear that their contributions to promoter strength were not additive (Arcot et al., 1989), with two CCAAT elements contributing only a minor increase in promoter strength relative to either CCAAT element alone. Second, the hamster tk promoter contains only a single CCAAT element (Lewis, 1986). Third, this element has been implicated by several studies to have a potential role in the cell cycle regulation oftk gene expression (Knight et al., 1987; Lipson et al., 1989; Roehl and Conrad, 1990) with only a region encompassing one of the CCAAT sequences and one of the GC elements being required to maintain a serum responsive promoter (Kim and Lee, 1991). In order to resolve these questions and map out the various protein binding sites within the tk promoter, we have utilized the technique of DNase I footprint analysis (Galaz and Schmitz, i978) using both crude nuclear extracts as well as purified transcription factor, Sp 1 (Briggs et al., 1986).
EXPERIMENTAL AND DISCUSSION
Fig. I. DNase I footprint analysis of the human tk promoter. (A)The structure of the human tk promoter is shown schematically with the triangle representing the TATA element, the two CCAAT elements indicated as squares, and the GC elements as ovals. The tsp and the direction of transcription are indicated by the bent arrow. Horizontal arrows in the tk325 region indicate a 27-bp inverted repeat. The position of the Hint site used to generate the two fragments, tk325 and tk167 is shown. (B) DNasc I footpri,t analysi~ on tk325 and tk167 fragments using total HeLa cell nuclear extract. The lanes marked M represent a G+A chemical sequencing reaction of the fragment. Reactions with no protein are indicated as a, while those with protein added (100/~g of total HoLa cell nuclear extract) are indicated as b. The + sign indicates that the footprint reactions were carded out on the coding strand. The locations of the various protein binding elements are as indicated. Methods. The 492-bp HindIII-RsaI fragment ofthe human tk promoter (Flemington et at., 1987) was cleaved using the unique Hint site. The resulting fragments were end-repaired and cloned into the SmaI site o f p U C l l 9 (Vieira and Messing, 1987) to obtain the plasmids ptk167 and ptk325, The orientation of the inserts was determined by sequence analysis. The insert fragments from the plasmids ptk167 and ptk325 (designated as tk167 and tk325)
(a) Identification of the various protein binding sites within the tk promoter using crude nuclear extracts From previous analyses of the tk promoter, it had been determined that the 492-bp HindIII-RsaI fragment of 5' were Y-end labeled by digestion of the plasmids with either EcoRl or BamHI (which cut in the polylinker on the opposite sides of the insert) followed by the repair of the overhang using a polIk fill-in step (Maniatis et al,, 1982) with [~¢-32p]dATP (3000 Ci/mmol). The labeling reactions were terminated by heating at 68°C for 10 rain and the plasmids were redigested with either BmnHI or EcoRI, respectively, to excise the insert. All the fragments were purified using a 6% polyacrylamide gel (nondenaturing) and soak-eluted (Maniatis et al., 1982). Preparation of nuclear extracts from logarithmically growing HeLa $3 cells (ATCC CCL2.2) and DNase I footprinting were carried out essentially as described (Flemington and Speck, 1990). The DNase I concentrations for reactions with no protein and reactions with 100 ~g of nuclear extract were empirically adjusted to produce an even digestion pattern. 100 pg of protein was used as the maximum amount of extract compatible with the footprinting protocol utilized.
251
A Del6
I-1
Del6/i Del6/I
~ r'l ~
B Del6
ebbs
dC
I-I
/~
100
[7
/~
146+34
["]
/k
271+57
Del6/i
Del6/l e bbe
8bb8
~~dC ~
INt
i,-i! o 9" "e' I
$8 o. o, |6
i ~
JQ - 0
:T ~ :T :T t: i l' I'-i' ill ~ ~ "l 8
g ~
.9999
9V,9
Fig. 2. Effect of insertions between the two CCAAT elements on promoter strength and protein binding. (A) Schematic of the Del6 insertion mutants with the squares representing the two CCAAT elements and the triangle representing the TATA element. The 5- and 10-bp insertions between the two CCAAT elements arc shown as hatched boxes. The numbers on the right indicate the average values and standard deviation of the CAT assays which were repeated six times. All values were normalized to the CAT activity resulting from the Del6 construct, which was arbitrarily taken as 100%. (B) DNase I footprint analysis of the insertion mutants using total HeLa cell nuclear extract. The lanes marked a are reactions with no protein while those marked b are reactions with 100 #g of total H eLa cell nuclear extract. The location of the TATA element, the proximal CCAAT and the distal CCAAT element are indicated as T, pC and dC, respectively. Methods. Plasmid Del6 [previously defined as -88 (Arcot et al., 1989)] is a deletion mutant of the human tk promoter linked to the cat gene and has only the TATA and the two CCAAT elements (Fig. 2A). This plasmid was used to create the insertion mutants Del6/i and Del6/I, which have 5 and 10 T nt inserted between the two CCAAT elements, respectively. Site-directed insertions of 5 and 10 T nt between the CCAAT elements were carried out using the Muta-GeneT M system (BioRad) as per the manufacturer's instructions. Transfection of the Del6 plasmids (20/~g) into mouse L929 (ATCC CCL1) and CAT assays
non-coding sequences was sufficient for maximal promoter activity (Arcot et el., 1989; Flemington et el., 1987). In order to facilitate the footprint analysis, this fragment was cleaved into two fragments, tk167 and tk325, by making use of the H i n t site (Fig. 1A). The larger of the two fragments, tk325, has at least three consensus GC elements, two of which are a part era large 27-bp inverted repeat. The smaller fragment, tk167 has one consensus GC element, two CCAAT elements (in the inverted orientation) and a TATA element. Using total HeLa cell nuclear extract, a number of protein binding sites could be identified within the tk promoter (Fig. 1B). In the case of the tk325 fragment, besides the three previously identified GC elements, a few other protein binding regions, mostly in and around the 27-bp inverted repeat could be identified. A closer examination of the sequences of these new binding sites showed that all, except one region (designated as US1, Fig. 1B), were potential GC elements, with a single mismatch to the GC core consensus sequence (see Fig. 4). The U S 1 region shares no identity with any of the known transcription factor binding sites (see Fig. 4; Sauve et al., 1990), and so the potential role for this binding region is unknown. In our previous functional study, the US 1 region did not appear to have a major contribution to overall promoter strength (Arcot et al., 1989). Studies on the role of the promoter in the cell-cycle regulation of tk gene expression have given no indication of this sequence being involved (Kim et ai., 1988; Lipson et al., 1989; Roehl and Conrad, 1990; Travail et el., 1988). Furthermore, a similar sequence is not found in the promoters of the chicken or rodent tk genes (Kwoh and Engler, 1984; Lewis, 1986; Seiser et al., 1989), and so we did not investigate this binding site any further. There is some chance that other protein binding sites were not detected in these studies. For instance, we should note that, although tk has been shown to be transcriptionally regulated in several cell types under different stimuli (Kasid et el., 1986; Stewart et al., 1987), this has not been definitively demonstrated in HeLa cells, from which our protein extracts originated. Thus, it is possible that in different cells, or under different footprinting conditions, other protein binding sites might also be detected. We have utilized
were carried out using standard procedures (Lopata et al., 1984) as described (Arcot et al., 1989). Quantitations of the CAT activities were performed by cutting out the regions of the silica gel corresponding to acetylated and unacetylated derivatives of [ t4C]chloramphenicol followed by liquid scintillation counting. Inserts from the Del6 plasmids for the footprint analysis (designated by the name of the respective plasmids) were T-end labeled by first digesting the plasmids with Nhel (which cleaves at the downstream end of the promoter) followed by the pollk fill-in step. Isolation of the fragments was achieved by redigestion of the plasmids with HindIII (which cleaves upstream from the promoter fragment). Purification of the inserts and footprint analysis were performed exactly as described in legend to Fig. 1.
252 extracts from a human Burkitt's lymphoma cell line, Ramos, with identical results, however (data not shown). Similar nuclear extract binding studies on the tk167 fragment showed a distinct protection around the single GC element present within this fragment along with a very broad region of protein binding encompassing the proximal CCAAT element, the TATA element and the mRNA tsp. The distal CCAAT element showed significantly less prorein binding when compared to the proximal CCAAT element (Fig. 1B). A detailed analysis of this anomaly is presented below. Our studies do not support the presence of other putative transcription elements predicted from the sequence to be within this region of the promoter (Sauve et al., 1990). The cell cycle regulatory unit defined by Kim and Lee (1991) extends from the Hinfl site to between the two CCAAT elements (see Fig. 1). This fragment apparently contains only one GC and one CCAAT element, which are the minimal requirement for the serum induction seen for this promoter.
(b) Protein binding to both CCAAT elements and promoter strength In our previous analysis of the human tk promoter, both the CCAAT elements were shown to be capable of binding a nuclear factor with the proximal CCAAT element having a slightly higher affinity, as judged by competition experiments in a gel-retardation assay (Afoot et al., 1989). However, the slight difference in affinity did not seem sufficient to explain the much poorer binding to the distal CCAAT element, as judged by the DNase I footprint analysis of this region. To determine if this difference was due to close proximity of the two related sites, or merely a slightly lower DNA-protein affinity at the distal site, the spacing between the two CCAAT elements in a deletion mutant of the promoter (DeI6, previously defined as -88 by Arcot et al., 1989) was increased by 5 and 10 bp (Del6/i and DeI6/I, Fig. 2A) and assayed for promoter strength by transfection into mouse L929 cells. The results of the CAT assays showed that promoter strength increased by approx. 1.5- and 2.7-fold respectively (Fig. 2A). Furthermore, this increase in promoter strength was consistent with increased protein binding to the distal CCAAT element, as seen in the footprint analysis of the insertion mutants (Fig. 2B). These data suggest that the proximity of the two CCAAT elements limits the binding to the distal site in the wt fragment. Coupled with our previous observations which showed a requirement for extensive flanking sequence in the binding of this CCAAT factor (Arcot et al., 1989), these data suggest that the binding of this factor (or complex) to the proximal CCAAT element at least partially excludes simultaneous binding to the distal CCAAT element. Thus, it appears that the placement of/he CCAAT elements in the human tk promoter is not optL~.zl for pro-
tk325
tk167
Mabc
Mabc
+
+
~~~q]Spl
| :il, I
|I dCCAAT
I
I
tE
I I TATA
|' :
!,)
Fig. 3. Interactionof purified Spl withthe variousGC elementswithin the tk325 and tk167 fragmentsdeterminedby DNase I footprintanalysis. The Spl bindingsites are bracketedwith arrowheadsindicatingthe cryptic GC elements (see Fig.4). The 27.bp inverted repeat in tk325 fragmentis indicatedby the invertedarrows.The positionsof the other protein bindingelementsare as indicated.The + signindicatesthat the reactions were performedon the coding strand. The lanes marked M indicatea G+Achemicalsequencingreaction(Maniatiset al., 1982)ofthe fragment. Lanesmarkeda, b and c represent reactionswithno protein. 100/~gof total HeLa cell nuclearextract and 10 ng of purified Spl, respectively.Reactionswith purified Spl (lane¢) were treated under the same conditionsas DNA sampleswith no protein added (lanea), since lessthan 10ng ofpurifiedproteinwas requiredto saturatethe Spl binding sites. meter strength. It may represent a redundancy built into the promoter, with the presence of multiple GC and CCAAT elements increasing the reliability, rather than the strength of the promoter. This redundancy of CCAAT elements is not essential to tk promoters, however, as the hamster tk promoter has only a single CCAAT dement (Lewis, 1986)
253
AAGCTTCCTTCTTGGAATTCCAAAC -432 TAATAAATGAGCTAACTCCGCCCCAGCCCCTTAGTCCCTCCCTGCAATCCACCTACCTCTGCAGACATCTTCTTC -357 CAAGGAACCTTGCTTGGGAAACCCACACCAG~CACATCCATCATGGCGTCT~CAGCCG~ATGGGCGtGC~TCCCT -282 (
)
CTGTTTATATGGCCAGAGCCCCGCCtCGCTCCC,CCC~TTTAAAC~TGGTGGGC~'CCGaGGCGG~GCTCAGACCA -208 ~GCCCCaCCCCGA~'CAGCCACGTCCATCGCCCTGATTTCCAGGCCCTCCCAGTCCCTGGGCGCACGTCCCGGAT -132 TCCTCCC,ACGAG~GdTGCGGCCAAATCTCCCGCCAGGTCA~CGGCCGGGCGCTGATTGGCCCC~TGGCG -57 GCGGG~CCGGCTCGTGKTTGGCCAGCACGCCGT~GTTTAKkGCGGTCGGCGCGGGAACCAG~GGCTTACTG A
Fig. 4. Summaryof the various footprintswithin the human tk promoter.The sequenceof the human tk promoteris presentedfrom -456 to position + 15 (numberingas describedin Flemingtonet al., 1987).Position + 1 is arbitrarilydefinedin the middle of the major transcription initiation site and is markedwith an arrowhead.Varioustranscription elementcore consensus sequences are in bold type. These include a TATA sequence(TI'rAAA), two CCAATelements in the inverted orientation (ATFGG) and eight sequences with close matches to the GC elementcore (either GGGCGG or CCGCCC). The nt in GC elementsthat do not match the abovecore consensus sequencesare shownin lower-caseletters.The undefinedproteinbinding site, USI, is underlinedin its entiretywith no consensus sequencehighlightedin bold. The brackets abovethe sequencerepresentthe maximalfootprints observedusing total nuclearextractwhilebrackets underneaththe sequenceindicatethe footprintsusingpurifiedSpl. The largedivergentarrows above the sequenceindicate the 27-bp invertedrepeats. and mouse apparently has none (Seiser et al., 1989). Recently, it has been demonstrated that the region of the tk promoter spanning from the distal CCAAT element to the first GC element was sufficient to confer cell-cycle regulation (Kim and Lee, 1991). The same study also showed that the region of the promoter comprising just the proximal CCAAT and the TATA element was unable to confer cellcycle regulation. Taken together with our footprint analysis, it would suggest that either one of the CCAAT elements in conjunction with the first GC element could function as the cell-cycle regulatory unit (Kim and Lee, 1991).
(c) Identification of the various Spl binding sites within the tk promoter In order to confirm that the predicted GC elements could interact with Spl (Briggs et al., 1986), footprint reactions on both tk325 and tk167 fragments were performed using purified Spl. The results showed that Spl does bind to all the previously identified GC elements (Fig. 3) as well as several unpredicted, cryptic GC elements (Fig. 3, regions indicated by arrowheads, also see Fig. 4). The physiological significance of such cryptic sites is not clear from these experiments. In our previous deletion analysis, we were unable to detect these cryptic GC elements, presumably due to the lack of sensitivity of the assay used. The arrangement of the G C elements in the 27-bp inverted repeat is reminiscent of the three direct repeats consisting of palmadly GC elements present in the SV40 early promoter (Dynan and Tjian, 1983). Using purified Spl, the single G C element in tk167 showed an altered pattern of protein ~nding (Fig. 3, lane c) when compared to the reaction using total nuclear extract (Fig. 3, lane b, summarized in Fig. 4). A possible
explanation could be that another nuclear protein in the total nuclear extract interacts with Spl forming a larger complex that results in a modified DNA footprint. It has previously been demonstrated that Spl activity requires an interaction with coactivator molecules (Pugh and Tjian, 1990). It is clear that Spl binding occurs at many sites in the promote1, and almost certainly is a major contributor to promoter function as judged by genetic (Arcot et at., 1989) as well as this footprint analysis. From the previous functional studies, deleting the regions of the promoter harboring the GC elements decreased the promoter strength and therefore, we conclude that the binding of Spl to the several upstream GC elements results in a moderate additive increase in the tk promoter strength over the basal promoter activity, which is mainly conferred by the CCAAT and the first GC element.
ACKNOWLEDGEMENTS We wish to thank Drs. R. Tjian and S. Jackson for kindly providing us with purified Spl protein. We also thank Drs. E. Flemington and S. Speck for technical assistance with footprint analysis. This work was supported by USPHS grant CA37673.
REFERENCES Arcot, S.S., Flemington,E.K. and Deininger,P.L.:The humanthymidine kinase genepromoter.J. Biol. Chem. 264 (1989)2343-2349. Bdio, LJ.: Regulationof thymidinekinasein human cells. Exp. CellRes. 89 (1974)263-274.
254 Briggs, M.R., Kadonaga, J.T., Bel!, S.P. and Tjian, R.: Purification and biochemical characterization of the promoter specific transcription factor Spl. Science 234 (1986) 47-52. Coppock, D.L. and Pardee, A.B.: Control of thymidine kinase mRNA during the cell cycle. Mol. Cell. Biol. 7 0987) 2925-2932. Dynan, W.S. and Tjian, R.: The promoter-specific transcription factor Spl binds to upstream sequences in the SV40 early promoter. Cell 235 (1983) 79-87. Flemington, E. and Speck, S.H.: Identification of phorbol ester response elements in the promoter of Epstein-Barr virus putative lyric switch gone BZLFI. J. Viral. 64 (1990) 1217-1226. Flemington, E., Bradshaw, H.D., Traina-Dorge, V., Slagel, V. and Deiningcr, P.L.: Sequence, structure and promoter characterization of the human thymidine kinase gone. Gone 52 (1987) 267-277. Gaiaz, D. and Schmitz, A.: DNase footprinting: a simple method for detection of protein-DNA binding specificity. Nucleic Acids Res. 5 (1978) 3157-3170. Ito, M. and Conrad, S.E.: Independent regulation of human thymidine kinase mRNA and enzyme levels in serum stimulated cells. J. Biol. Chem. 265 (1990) 6954-6960. Kasid, A., Davidson, N.E., Gelmann, E.P. and Lippman, M.E.: Transcriptional control of thymidine kinase gone expression by estrogen and antiestrogens in MCF-7 human breast cancer cells. J. Biol. Chem. 261 (1986) 5o'~62-5567. Kim, Y.K. and Lee, A.S.: Identification of a 70-base-pair cell cycle regulatory unit within the promoter of the human thymidine kinase gone and its interaction with cellular factors. Mol. Cell. Biol. 6 (1991) 2296-2302. Kim, Y.K., Wells, S., Lau, Y.F.C. and Lee, A.S.: Sequences contained within the promoter of the human thymidine kinasv genc can direct cell-cycle regulation of hcterologous fusion genes. Prec. Natl. Acad. Sci. USA 85 (1988) 5894-5898. Knight, G.B., Gudas, J.M. and Pardee, A.B.: Cell-cycle specific interaction of nuclear DNA-binding proteins with a CCAAT sequence from the human thymidine kinase gone. Prec. Natl. Acad. Sci. USA 84 (1987) 8350-8354. Kwoh, T.J. and Engier, J.A,: The nuc[eotide sequence of the chicken
thymidine kinase gene and the relationship of its predicted polypeptide to that of the vaccinia virus thymidine kinase. Nucleic Acids Res. 12 (1984) 3959-3971. Lewis, J.: Structure and expression of the chinese hamster thymidine kinase gone. Mol. Cell. Biol. 6 (1986) 1998-2009. Lipson, K.L., Chen, S.T., Koniecki, J., Ku, D.H. and Baserga, R.: Sphase specific regulation by deletion mutants of the human thymidine kinase promoter. Prec. Natl. Acad. Sci. USA 86 (1989) 6848-6852. Lopata, M.A., Cleveland, D.W. and Sollner-Webb, B.: High level expression of a chloramphenicol acetyl transferase gone by DEAE-dextran mediated DNA transfection coupled with a dimethyl sulfoxide or glycerol shock treatment. Nucleic Acids Res. 12 (1984) 5707-5717. Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982. Pugh, B.F. and Tjian, R.: Mechanism of transcriptional activation by Spl: Evidence for coactivators. Cell 61 (1990) 1185-1197. Roehl, H.H. and Conrad, S.E.: Identification of a G~-S-phase regulated region in the human thymidine kinase gone promoter. Mol. Cell. Biol. 10 (1990) 3834-3837. Sauce, GJ., Lipson, K.E., Chen, S.-T. and Baserga, R.: Sequence anal. ysis of the human thymidine kinase promoter: Comparison with the human PCNA promoter. DNA Seq. 1 (1990) 13-24. Seiser, C, Knofler, M., Rudelstorfer, I., Haas, R. and Wintersberger, E.: Mouse thymidine kinase: The promoter sequence and the gone and pseudogene structures in normal cells and in thymidine kinase deficient mutants. Nucleic Acids Res. 17 (1989) 185-195. Shcrley, J.L. and Kelly, TJ.: Regulation of the human thymidine kinase during the cell cycle. J. Biol. Chem. 263 0988) 8350-8358. Stewart, C J., Ire, M. ~.nd Conrad, S.E.: Evidence for transcriptional and post-transcriptional control of cellular thymidine kinase gone. Mol. Cell. Biol. 7 (1987) 1156-1163. Travail, S., Lipson, K.E., Jaskulski, D., Lanret, E. and Baserga, R.: Role of the promoter in the regulation of the thymidine kinase gone. Mol. Cell. Biol. 8 (1988) 1551-1557. Vieira, J. and Messing, .I.: Production of single-stranded plasmid DNA. Methods Enzymol. 153 (1987)3-11.
249
GENE0~86
Protein binding sites within the human thymidine kinase promoter (DNase I footprinting; CCAAT and GC elements; insertion mutants; promoter strength; recombinant DNA)
Santosh S. Areot a'* and Prescott L. Deininger a'b a Department of Biochemistry and Molecular Biology, Louisiana State University Medical Center, New Orleans, LA 70112 (U.S.A.) and b Laboratory of Molecular Genetics, Ochsner Medical Foundation, 1516 Jefferson Highway, New Orleans, LA 70121 (U.S.A.) Received by J.L. Slightom: 14 August 1991 Revised/Accepted: 28 September 1991 Received at publishers: 22 November 1991
SUMMARY
DNase I footprint analysis, using total HeLa cell nuclear extract and purified transcription factor Spl, was carried out to determine the various protein binding sites within the human thymidine kinase promoter. The promoter has two separate CCAAT elements and multiple Spl-binding sites, as well as at least one undefined protein binding site. Detailed analysis of protein binding to the two CCAAT elements showed that changing the spacing between the two CCAAT elements altered both protein binding to the distal CCAAT element as well as promoter activity. Both CCAAT elements can act as functional transcription dements, but are not oriented for optimal promoter strength in the human tk promoter. Our studies show that a promoter fragment that has been previously shown to be the minimal region to maintain a serum responsive promoter regulation apparently contains only a single Spl-binding domain and the more distal of the CCAAT elements.
INTRODUCTION
Thymidine kinase (TK, EC 2.7.1.21) is a crucial enzyme in the salvage pathway of pyrimidine nucleotide biosynthesis. Since this enzyme activity is required only for DNA synthesis, the levels of the enzyme show a high degree of correlation with the proliferative state of the cells (Belle, 1974). Regulation of the gene that encodes this enzyme has
Correspondenceto: Dr. P.L. Deininger, Biochemistry and Molecular Biology, LSUMC, 1901 Perdido St., New Orleans LA 70112 (U.S.A.) Tel. (504)568-4735; Fax (504)568-6158. * Present address: College of Pharmacy, University of Kentucky, Lexington, KY 40536 (U.S.A.) Tel. (606)257-2573. Abbreviations: bp, base pair(s); CAT, chloramphenicoi acetyl transferase; cat, gene (DNA) encoding CAT; nt, nucleotide(s); TK, thymidine kinase; tk. gone (DNA) encoding TK; poIIk, Klenow (large) fragment of the E. coil DNA polymerase I; tsp, transcription start point; USI, unknown site 1, wt, wild type.
been shown to occur at multiple levels which include transcriptional and post-transcriptional mechanisms (Coppock and Pardee, 1987; Ire and Conrad, 1990; Kasid et al., 1986; Sherley and Kelly, 1988; Stewart et al., 1987). More recently, studies on the role of the promoter in the regulation of the gone have shown that sequences within the promoter arc capable of conferring cell cycle dependent regulation onto a heterologous gene (Kim etal., 1988; Lipson etal., 1989; Roehl and Conrad, 1990; Travali et al., 1988). From the nt sequence analysis, we (Flemington etal., 1987), and others (Sauve etal., 1990) have previously identified a number of putative transcriptional elements within the human tk promoter. These include two CCAAT elements at -40 and -71 and several GC elements farther upstream. Subsequent studies using a deletion analysis confirmed these putative elements as functional promoter elements (Areot et al., 1989). Tile specific binding of proteins to the two CCAAT elements using gel retardation assays had previously demonstrated that both CCAAT elements could independently bind the same factor with high
250
A t')
¢'3t"~
I-! i-1
A[ -~
o..o.°o.°" "........
..........tk325 . . . . . . . . . . . . . . . . .
B
tk325
.
tk167 ........
tk167
+
+
MabbaM
MabbaM
GC" GC
US]
dCCAAT
GC" GO" pCCAAT
TATA
affinity (Arcot et al., 1989). Although the deletion analysis suggested the importance of a number of putative GC elements within the tk promoter based upon their ability to contribute to overall promoter strength, the actual role of these sequences to bind Spl (Dynan and Tjian, 1983) had not been demonstrated. The deletion and protein-binding analyses were also not comprehensive, in that they may very well have overlooked binding to other promoter elements which had not been identified from the sequence analysis (Flemington et al., 1987). Finally, although both the CCAAT elements were independently capable of binding a nuclear factor with similar affinities (Arcot et al., 1989), the potential roles for these CCAAT elements in the promoter appeared to be complex. First, it is clear that their contributions to promoter strength were not additive (Arcot et al., 1989), with two CCAAT elements contributing only a minor increase in promoter strength relative to either CCAAT element alone. Second, the hamster tk promoter contains only a single CCAAT element (Lewis, 1986). Third, this element has been implicated by several studies to have a potential role in the cell cycle regulation oftk gene expression (Knight et al., 1987; Lipson et al., 1989; Roehl and Conrad, 1990) with only a region encompassing one of the CCAAT sequences and one of the GC elements being required to maintain a serum responsive promoter (Kim and Lee, 1991). In order to resolve these questions and map out the various protein binding sites within the tk promoter, we have utilized the technique of DNase I footprint analysis (Galaz and Schmitz, i978) using both crude nuclear extracts as well as purified transcription factor, Sp 1 (Briggs et al., 1986).
EXPERIMENTAL AND DISCUSSION
Fig. I. DNase I footprint analysis of the human tk promoter. (A)The structure of the human tk promoter is shown schematically with the triangle representing the TATA element, the two CCAAT elements indicated as squares, and the GC elements as ovals. The tsp and the direction of transcription are indicated by the bent arrow. Horizontal arrows in the tk325 region indicate a 27-bp inverted repeat. The position of the Hint site used to generate the two fragments, tk325 and tk167 is shown. (B) DNasc I footpri,t analysi~ on tk325 and tk167 fragments using total HeLa cell nuclear extract. The lanes marked M represent a G+A chemical sequencing reaction of the fragment. Reactions with no protein are indicated as a, while those with protein added (100/~g of total HoLa cell nuclear extract) are indicated as b. The + sign indicates that the footprint reactions were carded out on the coding strand. The locations of the various protein binding elements are as indicated. Methods. The 492-bp HindIII-RsaI fragment ofthe human tk promoter (Flemington et at., 1987) was cleaved using the unique Hint site. The resulting fragments were end-repaired and cloned into the SmaI site o f p U C l l 9 (Vieira and Messing, 1987) to obtain the plasmids ptk167 and ptk325, The orientation of the inserts was determined by sequence analysis. The insert fragments from the plasmids ptk167 and ptk325 (designated as tk167 and tk325)
(a) Identification of the various protein binding sites within the tk promoter using crude nuclear extracts From previous analyses of the tk promoter, it had been determined that the 492-bp HindIII-RsaI fragment of 5' were Y-end labeled by digestion of the plasmids with either EcoRl or BamHI (which cut in the polylinker on the opposite sides of the insert) followed by the repair of the overhang using a polIk fill-in step (Maniatis et al,, 1982) with [~¢-32p]dATP (3000 Ci/mmol). The labeling reactions were terminated by heating at 68°C for 10 rain and the plasmids were redigested with either BmnHI or EcoRI, respectively, to excise the insert. All the fragments were purified using a 6% polyacrylamide gel (nondenaturing) and soak-eluted (Maniatis et al., 1982). Preparation of nuclear extracts from logarithmically growing HeLa $3 cells (ATCC CCL2.2) and DNase I footprinting were carried out essentially as described (Flemington and Speck, 1990). The DNase I concentrations for reactions with no protein and reactions with 100 ~g of nuclear extract were empirically adjusted to produce an even digestion pattern. 100 pg of protein was used as the maximum amount of extract compatible with the footprinting protocol utilized.
251
A Del6
I-1
Del6/i Del6/I
~ r'l ~
B Del6
ebbs
dC
I-I
/~
100
[7
/~
146+34
["]
/k
271+57
Del6/i
Del6/l e bbe
8bb8
~~dC ~
INt
i,-i! o 9" "e' I
$8 o. o, |6
i ~
JQ - 0
:T ~ :T :T t: i l' I'-i' ill ~ ~ "l 8
g ~
.9999
9V,9
Fig. 2. Effect of insertions between the two CCAAT elements on promoter strength and protein binding. (A) Schematic of the Del6 insertion mutants with the squares representing the two CCAAT elements and the triangle representing the TATA element. The 5- and 10-bp insertions between the two CCAAT elements arc shown as hatched boxes. The numbers on the right indicate the average values and standard deviation of the CAT assays which were repeated six times. All values were normalized to the CAT activity resulting from the Del6 construct, which was arbitrarily taken as 100%. (B) DNase I footprint analysis of the insertion mutants using total HeLa cell nuclear extract. The lanes marked a are reactions with no protein while those marked b are reactions with 100 #g of total H eLa cell nuclear extract. The location of the TATA element, the proximal CCAAT and the distal CCAAT element are indicated as T, pC and dC, respectively. Methods. Plasmid Del6 [previously defined as -88 (Arcot et al., 1989)] is a deletion mutant of the human tk promoter linked to the cat gene and has only the TATA and the two CCAAT elements (Fig. 2A). This plasmid was used to create the insertion mutants Del6/i and Del6/I, which have 5 and 10 T nt inserted between the two CCAAT elements, respectively. Site-directed insertions of 5 and 10 T nt between the CCAAT elements were carried out using the Muta-GeneT M system (BioRad) as per the manufacturer's instructions. Transfection of the Del6 plasmids (20/~g) into mouse L929 (ATCC CCL1) and CAT assays
non-coding sequences was sufficient for maximal promoter activity (Arcot et el., 1989; Flemington et el., 1987). In order to facilitate the footprint analysis, this fragment was cleaved into two fragments, tk167 and tk325, by making use of the H i n t site (Fig. 1A). The larger of the two fragments, tk325, has at least three consensus GC elements, two of which are a part era large 27-bp inverted repeat. The smaller fragment, tk167 has one consensus GC element, two CCAAT elements (in the inverted orientation) and a TATA element. Using total HeLa cell nuclear extract, a number of protein binding sites could be identified within the tk promoter (Fig. 1B). In the case of the tk325 fragment, besides the three previously identified GC elements, a few other protein binding regions, mostly in and around the 27-bp inverted repeat could be identified. A closer examination of the sequences of these new binding sites showed that all, except one region (designated as US1, Fig. 1B), were potential GC elements, with a single mismatch to the GC core consensus sequence (see Fig. 4). The U S 1 region shares no identity with any of the known transcription factor binding sites (see Fig. 4; Sauve et al., 1990), and so the potential role for this binding region is unknown. In our previous functional study, the US 1 region did not appear to have a major contribution to overall promoter strength (Arcot et al., 1989). Studies on the role of the promoter in the cell-cycle regulation of tk gene expression have given no indication of this sequence being involved (Kim et ai., 1988; Lipson et al., 1989; Roehl and Conrad, 1990; Travail et el., 1988). Furthermore, a similar sequence is not found in the promoters of the chicken or rodent tk genes (Kwoh and Engler, 1984; Lewis, 1986; Seiser et al., 1989), and so we did not investigate this binding site any further. There is some chance that other protein binding sites were not detected in these studies. For instance, we should note that, although tk has been shown to be transcriptionally regulated in several cell types under different stimuli (Kasid et el., 1986; Stewart et al., 1987), this has not been definitively demonstrated in HeLa cells, from which our protein extracts originated. Thus, it is possible that in different cells, or under different footprinting conditions, other protein binding sites might also be detected. We have utilized
were carried out using standard procedures (Lopata et al., 1984) as described (Arcot et al., 1989). Quantitations of the CAT activities were performed by cutting out the regions of the silica gel corresponding to acetylated and unacetylated derivatives of [ t4C]chloramphenicol followed by liquid scintillation counting. Inserts from the Del6 plasmids for the footprint analysis (designated by the name of the respective plasmids) were T-end labeled by first digesting the plasmids with Nhel (which cleaves at the downstream end of the promoter) followed by the pollk fill-in step. Isolation of the fragments was achieved by redigestion of the plasmids with HindIII (which cleaves upstream from the promoter fragment). Purification of the inserts and footprint analysis were performed exactly as described in legend to Fig. 1.
252 extracts from a human Burkitt's lymphoma cell line, Ramos, with identical results, however (data not shown). Similar nuclear extract binding studies on the tk167 fragment showed a distinct protection around the single GC element present within this fragment along with a very broad region of protein binding encompassing the proximal CCAAT element, the TATA element and the mRNA tsp. The distal CCAAT element showed significantly less prorein binding when compared to the proximal CCAAT element (Fig. 1B). A detailed analysis of this anomaly is presented below. Our studies do not support the presence of other putative transcription elements predicted from the sequence to be within this region of the promoter (Sauve et al., 1990). The cell cycle regulatory unit defined by Kim and Lee (1991) extends from the Hinfl site to between the two CCAAT elements (see Fig. 1). This fragment apparently contains only one GC and one CCAAT element, which are the minimal requirement for the serum induction seen for this promoter.
(b) Protein binding to both CCAAT elements and promoter strength In our previous analysis of the human tk promoter, both the CCAAT elements were shown to be capable of binding a nuclear factor with the proximal CCAAT element having a slightly higher affinity, as judged by competition experiments in a gel-retardation assay (Afoot et al., 1989). However, the slight difference in affinity did not seem sufficient to explain the much poorer binding to the distal CCAAT element, as judged by the DNase I footprint analysis of this region. To determine if this difference was due to close proximity of the two related sites, or merely a slightly lower DNA-protein affinity at the distal site, the spacing between the two CCAAT elements in a deletion mutant of the promoter (DeI6, previously defined as -88 by Arcot et al., 1989) was increased by 5 and 10 bp (Del6/i and DeI6/I, Fig. 2A) and assayed for promoter strength by transfection into mouse L929 cells. The results of the CAT assays showed that promoter strength increased by approx. 1.5- and 2.7-fold respectively (Fig. 2A). Furthermore, this increase in promoter strength was consistent with increased protein binding to the distal CCAAT element, as seen in the footprint analysis of the insertion mutants (Fig. 2B). These data suggest that the proximity of the two CCAAT elements limits the binding to the distal site in the wt fragment. Coupled with our previous observations which showed a requirement for extensive flanking sequence in the binding of this CCAAT factor (Arcot et al., 1989), these data suggest that the binding of this factor (or complex) to the proximal CCAAT element at least partially excludes simultaneous binding to the distal CCAAT element. Thus, it appears that the placement of/he CCAAT elements in the human tk promoter is not optL~.zl for pro-
tk325
tk167
Mabc
Mabc
+
+
~~~q]Spl
| :il, I
|I dCCAAT
I
I
tE
I I TATA
|' :
!,)
Fig. 3. Interactionof purified Spl withthe variousGC elementswithin the tk325 and tk167 fragmentsdeterminedby DNase I footprintanalysis. The Spl bindingsites are bracketedwith arrowheadsindicatingthe cryptic GC elements (see Fig.4). The 27.bp inverted repeat in tk325 fragmentis indicatedby the invertedarrows.The positionsof the other protein bindingelementsare as indicated.The + signindicatesthat the reactions were performedon the coding strand. The lanes marked M indicatea G+Achemicalsequencingreaction(Maniatiset al., 1982)ofthe fragment. Lanesmarkeda, b and c represent reactionswithno protein. 100/~gof total HeLa cell nuclearextract and 10 ng of purified Spl, respectively.Reactionswith purified Spl (lane¢) were treated under the same conditionsas DNA sampleswith no protein added (lanea), since lessthan 10ng ofpurifiedproteinwas requiredto saturatethe Spl binding sites. meter strength. It may represent a redundancy built into the promoter, with the presence of multiple GC and CCAAT elements increasing the reliability, rather than the strength of the promoter. This redundancy of CCAAT elements is not essential to tk promoters, however, as the hamster tk promoter has only a single CCAAT dement (Lewis, 1986)
253
AAGCTTCCTTCTTGGAATTCCAAAC -432 TAATAAATGAGCTAACTCCGCCCCAGCCCCTTAGTCCCTCCCTGCAATCCACCTACCTCTGCAGACATCTTCTTC -357 CAAGGAACCTTGCTTGGGAAACCCACACCAG~CACATCCATCATGGCGTCT~CAGCCG~ATGGGCGtGC~TCCCT -282 (
)
CTGTTTATATGGCCAGAGCCCCGCCtCGCTCCC,CCC~TTTAAAC~TGGTGGGC~'CCGaGGCGG~GCTCAGACCA -208 ~GCCCCaCCCCGA~'CAGCCACGTCCATCGCCCTGATTTCCAGGCCCTCCCAGTCCCTGGGCGCACGTCCCGGAT -132 TCCTCCC,ACGAG~GdTGCGGCCAAATCTCCCGCCAGGTCA~CGGCCGGGCGCTGATTGGCCCC~TGGCG -57 GCGGG~CCGGCTCGTGKTTGGCCAGCACGCCGT~GTTTAKkGCGGTCGGCGCGGGAACCAG~GGCTTACTG A
Fig. 4. Summaryof the various footprintswithin the human tk promoter.The sequenceof the human tk promoteris presentedfrom -456 to position + 15 (numberingas describedin Flemingtonet al., 1987).Position + 1 is arbitrarilydefinedin the middle of the major transcription initiation site and is markedwith an arrowhead.Varioustranscription elementcore consensus sequences are in bold type. These include a TATA sequence(TI'rAAA), two CCAATelements in the inverted orientation (ATFGG) and eight sequences with close matches to the GC elementcore (either GGGCGG or CCGCCC). The nt in GC elementsthat do not match the abovecore consensus sequencesare shownin lower-caseletters.The undefinedproteinbinding site, USI, is underlinedin its entiretywith no consensus sequencehighlightedin bold. The brackets abovethe sequencerepresentthe maximalfootprints observedusing total nuclearextractwhilebrackets underneaththe sequenceindicatethe footprintsusingpurifiedSpl. The largedivergentarrows above the sequenceindicate the 27-bp invertedrepeats. and mouse apparently has none (Seiser et al., 1989). Recently, it has been demonstrated that the region of the tk promoter spanning from the distal CCAAT element to the first GC element was sufficient to confer cell-cycle regulation (Kim and Lee, 1991). The same study also showed that the region of the promoter comprising just the proximal CCAAT and the TATA element was unable to confer cellcycle regulation. Taken together with our footprint analysis, it would suggest that either one of the CCAAT elements in conjunction with the first GC element could function as the cell-cycle regulatory unit (Kim and Lee, 1991).
(c) Identification of the various Spl binding sites within the tk promoter In order to confirm that the predicted GC elements could interact with Spl (Briggs et al., 1986), footprint reactions on both tk325 and tk167 fragments were performed using purified Spl. The results showed that Spl does bind to all the previously identified GC elements (Fig. 3) as well as several unpredicted, cryptic GC elements (Fig. 3, regions indicated by arrowheads, also see Fig. 4). The physiological significance of such cryptic sites is not clear from these experiments. In our previous deletion analysis, we were unable to detect these cryptic GC elements, presumably due to the lack of sensitivity of the assay used. The arrangement of the G C elements in the 27-bp inverted repeat is reminiscent of the three direct repeats consisting of palmadly GC elements present in the SV40 early promoter (Dynan and Tjian, 1983). Using purified Spl, the single G C element in tk167 showed an altered pattern of protein ~nding (Fig. 3, lane c) when compared to the reaction using total nuclear extract (Fig. 3, lane b, summarized in Fig. 4). A possible
explanation could be that another nuclear protein in the total nuclear extract interacts with Spl forming a larger complex that results in a modified DNA footprint. It has previously been demonstrated that Spl activity requires an interaction with coactivator molecules (Pugh and Tjian, 1990). It is clear that Spl binding occurs at many sites in the promote1, and almost certainly is a major contributor to promoter function as judged by genetic (Arcot et at., 1989) as well as this footprint analysis. From the previous functional studies, deleting the regions of the promoter harboring the GC elements decreased the promoter strength and therefore, we conclude that the binding of Spl to the several upstream GC elements results in a moderate additive increase in the tk promoter strength over the basal promoter activity, which is mainly conferred by the CCAAT and the first GC element.
ACKNOWLEDGEMENTS We wish to thank Drs. R. Tjian and S. Jackson for kindly providing us with purified Spl protein. We also thank Drs. E. Flemington and S. Speck for technical assistance with footprint analysis. This work was supported by USPHS grant CA37673.
REFERENCES Arcot, S.S., Flemington,E.K. and Deininger,P.L.:The humanthymidine kinase genepromoter.J. Biol. Chem. 264 (1989)2343-2349. Bdio, LJ.: Regulationof thymidinekinasein human cells. Exp. CellRes. 89 (1974)263-274.
254 Briggs, M.R., Kadonaga, J.T., Bel!, S.P. and Tjian, R.: Purification and biochemical characterization of the promoter specific transcription factor Spl. Science 234 (1986) 47-52. Coppock, D.L. and Pardee, A.B.: Control of thymidine kinase mRNA during the cell cycle. Mol. Cell. Biol. 7 0987) 2925-2932. Dynan, W.S. and Tjian, R.: The promoter-specific transcription factor Spl binds to upstream sequences in the SV40 early promoter. Cell 235 (1983) 79-87. Flemington, E. and Speck, S.H.: Identification of phorbol ester response elements in the promoter of Epstein-Barr virus putative lyric switch gone BZLFI. J. Viral. 64 (1990) 1217-1226. Flemington, E., Bradshaw, H.D., Traina-Dorge, V., Slagel, V. and Deiningcr, P.L.: Sequence, structure and promoter characterization of the human thymidine kinase gone. Gone 52 (1987) 267-277. Gaiaz, D. and Schmitz, A.: DNase footprinting: a simple method for detection of protein-DNA binding specificity. Nucleic Acids Res. 5 (1978) 3157-3170. Ito, M. and Conrad, S.E.: Independent regulation of human thymidine kinase mRNA and enzyme levels in serum stimulated cells. J. Biol. Chem. 265 (1990) 6954-6960. Kasid, A., Davidson, N.E., Gelmann, E.P. and Lippman, M.E.: Transcriptional control of thymidine kinase gone expression by estrogen and antiestrogens in MCF-7 human breast cancer cells. J. Biol. Chem. 261 (1986) 5o'~62-5567. Kim, Y.K. and Lee, A.S.: Identification of a 70-base-pair cell cycle regulatory unit within the promoter of the human thymidine kinase gone and its interaction with cellular factors. Mol. Cell. Biol. 6 (1991) 2296-2302. Kim, Y.K., Wells, S., Lau, Y.F.C. and Lee, A.S.: Sequences contained within the promoter of the human thymidine kinasv genc can direct cell-cycle regulation of hcterologous fusion genes. Prec. Natl. Acad. Sci. USA 85 (1988) 5894-5898. Knight, G.B., Gudas, J.M. and Pardee, A.B.: Cell-cycle specific interaction of nuclear DNA-binding proteins with a CCAAT sequence from the human thymidine kinase gone. Prec. Natl. Acad. Sci. USA 84 (1987) 8350-8354. Kwoh, T.J. and Engier, J.A,: The nuc[eotide sequence of the chicken
thymidine kinase gene and the relationship of its predicted polypeptide to that of the vaccinia virus thymidine kinase. Nucleic Acids Res. 12 (1984) 3959-3971. Lewis, J.: Structure and expression of the chinese hamster thymidine kinase gone. Mol. Cell. Biol. 6 (1986) 1998-2009. Lipson, K.L., Chen, S.T., Koniecki, J., Ku, D.H. and Baserga, R.: Sphase specific regulation by deletion mutants of the human thymidine kinase promoter. Prec. Natl. Acad. Sci. USA 86 (1989) 6848-6852. Lopata, M.A., Cleveland, D.W. and Sollner-Webb, B.: High level expression of a chloramphenicol acetyl transferase gone by DEAE-dextran mediated DNA transfection coupled with a dimethyl sulfoxide or glycerol shock treatment. Nucleic Acids Res. 12 (1984) 5707-5717. Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982. Pugh, B.F. and Tjian, R.: Mechanism of transcriptional activation by Spl: Evidence for coactivators. Cell 61 (1990) 1185-1197. Roehl, H.H. and Conrad, S.E.: Identification of a G~-S-phase regulated region in the human thymidine kinase gone promoter. Mol. Cell. Biol. 10 (1990) 3834-3837. Sauce, GJ., Lipson, K.E., Chen, S.-T. and Baserga, R.: Sequence anal. ysis of the human thymidine kinase promoter: Comparison with the human PCNA promoter. DNA Seq. 1 (1990) 13-24. Seiser, C, Knofler, M., Rudelstorfer, I., Haas, R. and Wintersberger, E.: Mouse thymidine kinase: The promoter sequence and the gone and pseudogene structures in normal cells and in thymidine kinase deficient mutants. Nucleic Acids Res. 17 (1989) 185-195. Shcrley, J.L. and Kelly, TJ.: Regulation of the human thymidine kinase during the cell cycle. J. Biol. Chem. 263 0988) 8350-8358. Stewart, C J., Ire, M. ~.nd Conrad, S.E.: Evidence for transcriptional and post-transcriptional control of cellular thymidine kinase gone. Mol. Cell. Biol. 7 (1987) 1156-1163. Travail, S., Lipson, K.E., Jaskulski, D., Lanret, E. and Baserga, R.: Role of the promoter in the regulation of the thymidine kinase gone. Mol. Cell. Biol. 8 (1988) 1551-1557. Vieira, J. and Messing, .I.: Production of single-stranded plasmid DNA. Methods Enzymol. 153 (1987)3-11.
