Human cytomegalovirus infection induces cellular ... - CiteSeerX

Journal of General Virology (2002), 83, 2983–2993. Printed in Great Britain ...................................................................................................................................................................................................................................................................................

Human cytomegalovirus infection induces cellular thymidylate synthase gene expression in quiescent fibroblasts Giorgio Gribaudo,1 Ludovica Riera,1 Thomas L. Rudge,2 Patrizia Caposio,1 Lee F. Johnson2 and Santo Landolfo1 1 2

Department of Public Health and Microbiology, University of Torino, Via Santena, 9 – 10126 Torino, Italy Department of Molecular Genetics, The Ohio State University, Columbus, OH 43210, USA

Productive infection of non-proliferating cells by cytomegalovirus (CMV) requires the coordinated stimulation of host biochemical pathways that prepare cells to synthesize DNA. Here we illustrate the ability of human CMV (HCMV) to stimulate cellular thymidylate synthase (TS) gene expression in quiescent human embryonic lung fibroblasts. TS mRNA and protein levels are nearly undetectable in quiescent cells, but are greatly increased following HCMV infection. Inhibition of TS activity was shown to impair HCMV DNA synthesis, demonstrating that TS upregulation is required for efficient HCMV replication in quiescent cells. The increase in TS gene expression was due to an increase in gene transcription, since the expression of a reporter gene driven by the human TS promoter was strongly induced by HCMV infection. Deletion analysis of the human TS promoter identified two positive elements that are important for this increased transcription. We have previously shown that murine CMV (MCMV) stimulates the mouse TS promoter by a mechanism that depends on the presence of an E2F element in the promoter region. However, deletion of the two potential E2F binding sites in the human TS promoter did not prevent the virus-induced increase in TS promoter activity. Our data suggest that HCMV activates human TS gene transcription by mechanisms that are independent of E2F and different from those used by MCMV to stimulate the mouse TS promoter.

Introduction Human cytomegalovirus (HCMV) is a ubiquitous betaherpesvirus that generally causes benign or asymptomatic infections. However, in immunocompromised hosts, such as transplant recipients and AIDS patients, and in immunologically immature individuals, it can cause life-threatening diseases. Furthermore, HCMV is the major viral cause of birth defects in humans (Fortunato & Spector, 1999 ; Griffith, 2000 ; Mocarski & Courcelle, 2001). Its success as an opportunistic pathogen depends largely on its ability to establish latent lifelong infections, counteract host defence mechanisms and replicate in a wide variety of cells and tissues, including differentiated, post-mitotic cells (Hengel et al., 1998 ; Hengel & Weber, 2000 ; Lalani et al., 2000). The absence of functional virus-encoded dNTP-synthesizing enzymes such as thymidine kinase, dihydrofolate reductase (DHFR), thymidylate synthase (TS) and ribonucleotide reductase (Chee et al., 1990 ; Rawlinson Author for correspondence : Giorgio Gribaudo. Fax j39 011 6636436. e-mail giorgio.gribaudo!unito.it

0001-8383 # 2002 SGM

et al., 1996) compels HCMV to depend mainly on host-cell metabolism to ensure a sufficient supply of dNTPs for its DNA replication. However, these enzymes are expressed at very low levels in post-mitotic cells. To explain this apparent paradox, it has been hypothesized that HCMV infection of quiescent cells induces biochemical pathways that are required for DNA replication, including enzymes involved in nucleotide metabolism. Subsequently, to gain selective access to the newly synthesized dNTPs, the virus prevents replication of the host cell DNA by blocking cell-cycle progression at the beginning of the S phase (Fortunato et al., 2000 ; Wiebusch & Hagemeier, 2001). TS is an essential enzyme that catalyses the de novo biosynthesis of thymidylic acid (dTMP) by the reductive transfer of the methylene group from 5,10-methylenetetrahydrofolate to the 5 position of the substrate, deoxyuridylic acid, to form dTMP and dihydrofolate. TS activity is associated with cell proliferation and its inhibition in rapidly proliferating cells leads to inhibition of DNA synthesis and cell death. For this reason, TS is an important target enzyme in cancer chemotherapy. Both substrate and cofactor analogues have CJID

G. Gribaudo and others

been widely used as anti-neoplastic agents. TS mRNA and enzyme levels are very low in quiescent cells, but increase at the G –S border during a serum-induced transition from the " resting (G ) phase to the S phase (Jenh et al., 1985 ; Ayusawa et ! al., 1986). TS mRNA content is primarily controlled at the post-transcriptional level in growth-stimulated cells, since nuclear run-on transcription assays have revealed that TS gene transcription in human and mouse cells does not change during the G –S transition (Ayusawa et al., 1986 ; Ash et al., 1995). " Studies with transfected TS minigenes have shown that both the TS essential promoter region and an intron in the transcribed region are required for proper S-phase regulation, suggesting that some form of communication between the TS promoter and the RNA processing machinery may be important for regulation of TS mRNA production in growthstimulated cells (Kaneda et al., 1992 ; Takayanagi et al., 1992 ; Ash et al., 1993, 1995 ; Johnson, 1994 ; Ke et al., 1996). We have been studying the effects of CMV infection on the expression of enzymes involved in dTMP biosynthesis, since elucidation of the molecular mechanisms of virus-mediated regulation could lead to the design of molecules with antiviral activity. In the murine system, we have observed that murine CMV (MCMV) replication in quiescent NIH 3T3 cells increases the expression of enzymes involved in dTMP biosynthesis, namely folylpolyglutamate synthetase (FPGS), DHFR and TS (Lembo et al., 1998 ; Gribaudo et al., 2000 ; Cavallo et al., 2001). Here we report that HCMV infection of quiescent human embryonic lung fibroblasts induces the expression of TS mRNA and protein, and that TS activity is required for efficient HCMV DNA synthesis. Regulation of TS gene expression takes place at the transcriptional level, and several DNA elements within the TS promoter are necessary for its increase. Lastly, HCMV and MCMV transactivate the corresponding cellular TS promoters by different mechanisms.

Methods

Cells and culture conditions. Low-passage human embryonic lung fibroblasts (HELF) were grown in minimal essential medium (MEM) (Gibco\BRL) supplemented with 10 % foetal calf serum (Gibco\BRL). NIH 3T3 murine fibroblasts were grown in Dulbecco’s modified Eagle’s medium supplemented with 10 % calf serum. Quiescent cells (arrested in G \G phase) were obtained by incubating confluent monolayers for ! " 48 h in medium containing 0n5 % serum (low-serum medium). Flow cytometry demonstrated that more than 90 % were growth-arrested.

Virus preparation and infections. HCMV (strain AD169 ; ATCC VR-538) was passaged in HELF cells. MCMV (strain Smith ; ATCC VR194) was propagated in NIH 3T3 cells. Both viruses were prepared in low-serum medium. Quiescent HELF or NIH 3T3 cells were infected with HCMV or MCMV, respectively, at an m.o.i. of 5, unless otherwise stated. Mockinfected control cultures were exposed to an equal volume of mockinfecting fluid. At the end of the adsorption period, which was defined as 0 h post-infection (p.i.), the low-serum medium removed before infection was restored to the cultures to avoid stimulation due to the addition of fresh serum growth factors. UV-inactivated HCMV was prepared with one pulse of 1n2 J\cm# UV light. The UV-inactivated HCMV did not CJIE

replicate or produce detectable levels of immediate-early (IE) gene products. For infection in the presence of cycloheximide (CHX), quiescent HELF cells were pretreated for 1 h with 100 µg\ml CHX before infection and then maintained in the medium throughout the infection. For infection in the presence of phosphonoformic acid (PFA), PFA was added to the medium at a final concentration of 200 µg\ml at the time of infection. RNA was then harvested at 24 (CHX) or 48 (PFA) h p.i.

Northern blot analysis. Total RNA or poly(A)+ mRNA was fractionated on a 1 % agarose, 2n2 M formaldehyde gel and then blotted on to a nitrocellulose membrane (Hybond C-extra ; Amersham). The filters were hybridized to random-primed radiolabelled probes corresponding to the human TS cDNA (Ayusawa et al., 1984) and human glyceraldehyde-3-phosphate dehydrogenase (G3PDH) cDNA as a control.

Plasmids. Plasmids containing the indicated regions of the human TS promoter (numbered relative to the A of the translational start codon ; see Fig. 3) driving expression of the luciferase indicator gene were constructed in the following manner. Plasmids phTS k161\k98 Luc, phTS k161\k141 Luc and phTS k140\k98 Luc were constructed by inserting the HindIII–NheI fragment of the corresponding TS minigene (Dong et al., 2000) into the HindIII and NheI sites of pGL3 basic-EP, which is the same as pGL3 basic (Promega) except that the multiple cloning sequence was modified by the insertion of a synthetic DNA fragment to give a new multiple cloning sequence (KpnI, HindIII, SacI, XbaI, NheI, SmaI, BglII). Plasmid phTSk161\k141 Sp1 mut. Luc was constructed by inserting synthetic oligonucleotides with the desired nucleotide changes into the HindIII and NheI sites of pGL3 basic-EP. Plasmid phTS k243\j30 Luc was constructed by PCR-amplifying the indicated region of the human TS promoter and inserting it into the NheI site of pGL3-basic that had been modified by deleting 13 nucleotides between the BglII and HindIII sites (Dong et al., 2000). The promoter regions of all constructs were sequenced to verify the presence of the desired promoter regions and\or mutations and the absence of undesired changes. Plasmids phTS k243\k70 Luc, phTS k243\k98 Luc and phTS k243\k141 Luc were as described previously (Dong et al., 2000). pTSWTGL3 was constructed by inserting the mouse TS promoter and 5h flanking region from the XbaI site at k985 to an engineered BglII site at k11 into the NheI and BglII sites of pGL3-basic. pTSWTGL3(k110) was the same as pTSWTGL3, except that the potential E2F element just upstream from the mouse TS essential promoter region (−""&GATTCTGGCGGCC−"!$) was mutated to −""&GATTCGCTAGCCC−"!$ (Geng & Johnson, 1993). pSGIE72 and pSGIE86 contained cDNAs corresponding to the HCMV IE1 72 kDa and IE2 86 kDa proteins, respectively. Their expression was driven by the simian virus 40 (SV40) early promoter (Klucher et al., 1993). p729CAT contained the CAT reporter gene driven by the HCMV UL112\113 early promoter (Staprans et al., 1988).

Transient transfection and reporter gene assays. Cells were transfected by the calcium phosphate procedure, as previously described (Gribaudo et al., 2000). The transfected cells were washed twice with medium and incubated in MEM supplemented with 0n5 % foetal calf serum (low-serum medium) for 48 h. Luciferase and CAT activity were assayed as previously described (Gribaudo et al., 1995). Reporter gene activity was normalized to the amount of plasmid DNA introduced into recipient cells, as previously described (Abken & Reifenrath, 1992).

Immunoblotting. Whole-cell protein extracts were prepared as previously described (Gribaudo et al., 2000). Proteins were separated by SDS–PAGE and then transferred to Immobilon-P membranes (Millipore). Filters were immunostained with the mouse anti-TS human mAb (Johnston et al., 1991) (clone TS106 ; Labvision Neomarkers), the mouse

HCMV stimulates thymidylate synthase expression anti-HCMV IEA mAb (E13 clone ; Argene Bio-Soft) recognizing both the IE1 72 kDa and IE2 86 kDa proteins, or the mouse anti-actin mAb (Boehringer) at room temperature for 1 h. Immune complexes were detected with sheep anti-mouse IgG antibody conjugated to horseradish peroxidase (Amersham) and visualized by enhanced chemiluminescence (Super Signal ; Pierce).

Inhibition of viral DNA synthesis. To evaluate the inhibition of HCMV DNA synthesis, cells were grown to subconfluence, incubated in low-serum medium for 48 h and infected with HCMV at an m.o.i. of 1. The infected cultures were treated in low-serum medium with different concentrations of N-(5-[N-(3,4-dihydro-2-methyl-4-oxoquinazolin6-ylmethyl)-N-methylamino]-2-thenoyl)--glutamic acid (raltitrexed, Tomudex, formerly ZD1694 ; Zeneca). At 96 h p.i. cells were harvested and total DNA was isolated. Dot-blot hybridization was then performed using $#P-labelled probes prepared from the PstI–BamHI DNA fragment of the HCMV IE1 gene (exon 4) and mouse G3PDH cDNA. The membranes were autoradiographed and the hybridization signals were quantified using a phosphoimager.

Results Inhibition of cellular TS activity blocks HCMV DNA synthesis

To determine whether cellular TS activity is required for HCMV replication, we analysed the effects of raltitrexed, a TS inhibitor, on HCMV DNA synthesis. Quiescent HELF cells were infected at an m.o.i. of 1. After virus adsorption, raltitrexed was added at concentrations ranging from 10−$ to 10 µM. After 96 h p.i., viral DNA levels were measured by dot-blot analysis using a radiolabelled viral DNA probe. As shown in Fig. 1, addition of raltitrexed greatly reduced HCMV DNA levels. The drug EC was 0n03 µM, while its 50 % &! cytotoxic concentration was 10 µM. These results demonstrated that inhibition of TS activity impairs HCMV DNA synthesis in quiescent cells and that this inhibition is not due to generalized cellular toxicity. HCMV infection upregulates TS gene expression in quiescent HELF cells

Having established that TS activity is required for efficient HCMV DNA synthesis in quiescent fibroblasts, we analysed TS gene expression in serum-arrested HELF cells during productive HCMV replication. Poly(A)+ RNA was prepared at different times p.i., and TS mRNA content was measured by Northern blot analysis. As shown in Fig. 2(A), HCMV infection increased TS mRNA levels by more than eightfold by 24 h p.i. compared with mock-infected cells. To identify which phase of viral gene expression was temporally associated with the upregulation of TS mRNA, we infected quiescent fibroblasts in the presence of CHX to block protein synthesis so that we could monitor the effect of the binding and entry of HCMV on TS expression. Alternatively, the infection was carried out in the presence of PFA to inhibit viral DNA replication and restrict viral gene expression to IE and E genes. Thereafter, total RNA was prepared at 24 (CHX) or 48 (PFA) h p.i., and TS mRNA content was measured by

Fig. 1. Inhibition of TS activity abrogates HCMV DNA synthesis in quiescent HELF cells. Quiescent HELF cells were infected with HCMV at an m.o.i. of 1 and then incubated with the indicated concentrations of raltitrexed. At 96 h p.i., total genomic DNA was purified and twofold dilutions were immobilized on hybridization membrane. The same filter was sequentially hybridized with 32P-labelled HCMV IE1 and G3PDH probes. Hybridization signals were quantified by a phosphoimager. To determine cell viability, HELF cells were growth-arrested in low-serum medium and then exposed to increasing concentrations of raltitrexed. After 4 days of incubation, the number of viable cells was determined by the 3(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) method, as previously described (Pauwels et al., 1988).

Northern blot analysis. As shown in Fig. 2(B), the increase in TS mRNA content occurred only in the presence of PFA treatment, indicating that TS gene expression was regulated by newly expressed IE and\or E proteins, rather than by the binding and entry of the virus particle or by the action of virion proteins within the infected cell. To determine whether viral infection would also lead to an increase in TS protein, cell extracts were prepared at different times p.i. and analysed by immunoblotting with an anti-TS mAb. As shown in Fig. 2(C), TS protein was undetectable in mock-infected cells, started to increase at 24 h p.i. in infected cells and peaked at 72 h p.i. In contrast, it was undetectable in cells infected for 48 h with UV-inactivated HCMV, further demonstrating that newly synthesized viral proteins are required to induce cellular TS expression. The efficacy of UV treatment was confirmed by the lack of viral IE gene expression. These results confirmed that the increases in TS mRNA and protein levels during HCMV replication in quiescent HELF cells depend on viral gene expression. HCMV infection transactivates the TS promoter in quiescent HELF cells

To determine whether the increase in TS mRNA was a consequence of stimulation of TS promoter activity, we analysed the effects of HCMV infection on the expression of a transiently transfected luciferase reporter gene driven by the human TS promoter. HELF cells were transiently transfected CJIF


expression of an indicator gene (Dong et al., 2000). Transfected cells were serum-starved and then infected with HCMV or UV-inactivated virus, mock-infected or stimulated with serum. At different times p.i., cells extracts were prepared and assayed for luciferase activity. As shown in Fig. 4, HCMV infection led to a sixfold increase in luciferase activity by 12 h, a 26-fold increase by 24 h and an 18-fold increase by 48 h p.i. UVinactivated virus did not affect luciferase activity, demonstrating that HCMV-mediated transactivation requires de novo viral protein expression. As expected, serum stimulation did not increase the reporter activity since, as previously observed, the S-phase-specific human TS expression requires the presence of both the promoter region and a spliceable intron (Takayanagi et al., 1992). Taken together, these results demonstrate that HCMV regulates TS gene expression primarily at the transcriptional level using mechanisms that are different from those employed by growth-stimulated cells. Transactivation of the human TS promoter by HCMV requires the essential promoter region and sequences between N98 and N70

Fig. 2. HCMV infection stimulates TS gene expression in quiescent fibroblasts. (A) HCMV infection increases TS mRNA content. HELF cells were serum-starved for 48 h and then infected with HCMV (m.o.i. of 5) or mock-infected. Poly(A)+ RNA was isolated and purified at the indicated times p.i. and 2 µg was analysed by Northern blot as described in Methods. The filter was sequentially hybridized with human TS and G3PDH radiolabelled probes. Hybridization signals were quantified by a phosphoimager. (B) HCMV immediate-early and early gene expression are required to increase TS mRNA levels. Quiescent HELF cells were infected with HCMV (m.o.i. of 5) in the presence of CHX or PFA or were left untreated. Total RNA was isolated and purified at 24 (CHX-treated) or 48 (PFA-treated) h p.i., and 20 µg of total RNA was analysed by Northern blot. The filter was sequentially hybridized with human TS and G3PDH radiolabelled probes. Hybridization signals were quantified by a phosphoimager. (C) HCMV increases TS protein levels. Quiescent HELF cells were infected with HCMV or UV-inactivated HCMV, or mock-infected. Total cell extracts were prepared at the indicated times p.i. and fractionated by SDS–PAGE (50 µg protein per lane). Proteins were analysed by immunoblotting with the anti-TS mAb or with the anti-IEA mAb, as described in Methods. Actin immunodetection with a mAb was used as an internal control. Cell extracts were prepared from mockinfected cells (Mock), from cells infected with HCMV for 24, 48 or 72 h, and from cells infected with UV-treated HCMV for 48 h (48 h UV).

with the indicator plasmid phTS k243\j30 Luc, which corresponds to an intronless luciferase gene driven by the promoter and adjacent sequences of the human TS gene (Fig. 3). These promoter sequences are sufficient to drive high-level CJIG

To determine the locations of nucleotide sequences that are important for the HCMV-mediated regulation of TS promoter activity, we first performed a promoter deletion analysis. We have previously shown that the sequences responsible for high-level promoter activity of the human TS gene are located between k243 and j30 relative to the AUG start codon (Dong et al., 2000). Our first goal was to establish the 3h boundary of the sequences that are important for virus induction of the TS promoter. To accomplish this, the 5h end of the promoter region was kept at position k243 and the 3h boundary was varied from j30 to k141 (Fig. 3). HELF cells were transfected with the different constructs, serum-starved and then infected with active or UV-inactivated HCMV. At 24 h p.i., cell extracts were prepared and assayed for luciferase activity. The results (Fig. 5) show that deletion of the sequences between j30 and k70 had little effect on HCMV-mediated transactivation. In contrast, deletion of sequences downstream from k98 led to a threefold reduction in reporter activity, suggesting that a positive-acting promoter element regulating TS gene expression in response to HCMV infection is located between k98 and k70. Further deletion of sequences downstream of k141 (phTS k243\k141 Luc) did not significantly reduce the luciferase activity relative to that of phTS k243\k98 Luc. To establish the 5h boundary of the regulatory elements necessary for virus stimulation of TS promoter activity, we constructed two more plasmids, phTS k161\k98 Luc and phTS k140\k98 Luc, which retained the region of the TS promoter containing most of the transcription initiation sites (Fig. 3). As shown in Fig. 5, deletion of the sequences between k243 and k161 had no effect on the virus induction of promoter activity since both plasmids were transactivated eightfold by virus infection. However, deletion of sequences

HCMV stimulates thymidylate synthase expression

Fig. 3. Sequence of the human TS promoter region. The sequence of the human TS gene (Takeishi et al., 1989) between k250 and j51 relative to the A of the translational start codon (***) is shown. The sequence of the TS promoter region analysed in the present study is identical to the published sequence except that the boxed region (one of the repeated sequences) is missing and there is a G A change at k39. The differences are presumably the result of sequence polymorphisms and do not appear to affect TS promoter activity (Dong et al., 2000). Direct and inverted repeat sequences are indicated by arrows. Potential binding sites for several transcription factors are shown below the sequence. Sp1 mutation indicates the nucleotide changes (underlined) that inactivated the Sp1 element.

Fig. 4. HCMV infection transactivates the human TS promoter in quiescent cells. HELF cells were transiently transfected with 2 µg of the indicator construct phTS k243/j30 Luc (Fig. 3) and 10 µg carrier DNA (pBluescript SK). After 18 h, cells were washed and growth-arrested in low-serum medium for 48 h. Thereafter, transfectants were infected with HCMV, UV-inactivated HCMV, mock-infected or serum-stimulated. Total cytoplasmic extracts were isolated at the indicated times p.i. and assayed for luciferase activity. Reporter gene activity was normalized to the amount of plasmid DNA introduced into recipient cells as determined by dot blot analysis. The normalized luciferase activity is expressed as fold induction relative to basal levels measured in cells transfected with phTS k243/j30 Luc and then mock-infected, which was set at 1. The experiment was repeated twice and representative results are shown.

between k161 and k140 (phTS k140\k98 Luc) almost completely prevented virus induction of promoter activity. These results suggested that the sequences between k161 and k140 are important for stimulation of TS promoter activity in response to HCMV infection. To confirm this conclusion, the 20-nucleotide region was inserted upstream of the luciferase gene to obtain the phTS k161\k141 Luc plasmid. As shown in Fig. 5, HCMV infection led to a more than fivefold increase in luciferase activity with this plasmid. Previous studies have demonstrated that the sequence between

k161 and k141 is highly conserved between the human and mouse TS promoter (Horie & Takeishi, 1997) and is the minimum essential region for TS promoter activity in proliferating cells (Dong et al., 2000). We have recently observed that MCMV infection transactivates the mouse TS promoter and that this transactivation was abolished by inactivating the functional E2F element immediately upstream of the mouse TS essential promoter region. These observations indicated an E2F-dependent mechanism for induction of the mouse TS promoter in response to CJIH


Fig. 5. Human TS promoter deletion analysis. HELF cells were transiently transfected with 2 µg of the indicated phTS constructs (Fig. 3) and 10 µg carrier DNA (pBluescript SK). After 18 h, cells were washed and growth-arrested in low-serum medium for 48 h. Thereafter, transfectants were infected with HCMV or UV-irradiated HCMV or were mock-infected. Total cytoplasmic extracts were isolated at 24 h p.i. and assayed for luciferase activity. Reporter gene activity was normalized to the amount of plasmid DNA introduced into recipient cells as determined by dot-blot analysis. The normalized luciferase activity is expressed as fold induction relative to basal levels measured in cells transfected with the different phTS constructs and then mock-infected, which was set at 1. The data shown are the averages of three experimentspSEM (error bars).

Fig. 6. HCMV IE1 protein transactivates the human TS promoter. HELF cells were transiently cotransfected with 2 µg of the indicated human TS plasmid and 4 µg of the expression vector for HCMV IE1 72 kDa protein, HCMV IE2 86 kDa protein, or the empty pSG5 vector. At 18 h after transfection, cells were washed and then maintained in low-serum medium for 48 h. Thereafter, total cytoplasmic extracts were isolated and assayed for luciferase activity. Reporter gene activity was normalized to the amount of plasmid DNA introduced into recipient cells by dot-blot analysis. The normalized luciferase activity is expressed as fold induction relative to basal levels measured in cells transfected with each phTS construct and the pSG5, which was set at 1. The data shown are the averages of three experimentspSEM (error bars).

MCMV infection (Gribaudo et al., 2000). The presence of two potential E2F binding sites within the inverted repeat element between k128 and k98 of the human TS promoter (Fig. 1) raised the possibility that these elements may contribute to the regulation of human TS promoter activity in response to HCMV infection. We recently found that recombinant E2F is able to bind to the human TS promoter region between k134 and k105 in EMSA assays (data not shown). Unexpectedly, we found that the E2F elements were not necessary for the CJII

HCMV-mediated increase in TS promoter activity since a similar increase in luciferase activity was observed with phTS k161\k141 Luc, which lacks the E2F elements, as with phTS k161\k98 Luc, which retains the E2F elements (Fig. 5). Taken together, these results indicate that the k98\k70 element and the minimal essential TS promoter segment play important positive roles in the response of the human TS promoter to HCMV infection. Furthermore, in contrast to our earlier observations in the murine system, stimulation of the


human TS promoter by HCMV does not appear to involve an E2F-dependent mechanism. HCMV IE1 protein transactivates the human TS promoter

We have previously observed that the expression of MCMV IE1 protein activates the mouse TS promoter and that this activation is dependent on the integrity of the E2F element (Gribaudo et al., 2000). To examine the potential role of the HCMV major immediate-early proteins in the regulation of the human TS promoter, we co-transfected an expression vector for the IE1 72 kDa or the IE2 86 kDa protein with different human TS constructs into HELF cells. The major HCMV IE proteins were expressed under the control of the early SV40 promoter to avoid the potential complication of negative autoregulation of the HCMV IE promoters by the IE2 protein (Klucher et al., 1993). Fig. 6 demonstrates that only the IE1 product transactivated the human TS promoter, although to a lesser extent than that observed with virus infection. The magnitude of activation was 4n5-fold for phTS k243\j30 Luc, 3n5-fold for phTS k243\k70 Luc, threefold for phTS k243\k98 Luc and 2n5-fold for phTS k243\k141 Luc relative to the control. None of the other constructs analysed was significantly transactivated by IE1 (data not shown). The ability of the IE1 or the IE2 constructs to express functional proteins was verified by co-transfection with the p729CAT indicator plasmid containing the HCMV UL112\ 113 early promoter. As previously reported (Klucher et al., 1993), the cDNA construct expressing the IE2 protein transactivated the early promoter, whereas when both the IE1 and IE2 constructs were transfected together, a small but reproducible synergistic effect was observed (data not shown). Differential regulation of mouse and human TS promoters by CMV infection

As indicated above, activation of the mouse TS promoter by MCMV is mediated through the E2F site in the mouse TS promoter (Gribaudo et al., 2000), whereas activation of the human TS promoter by HCMV does not depend on the presence of the E2F sites. This discrepancy prompted us to see whether HCMV regulates the mouse TS promoter by a different mechanism than that used by MCMV. We reasoned that if HCMV were able to control the mouse TS promoter through the E2F element, differences in the architecture of the regulatory elements between the human and mouse TS promoter might account for the different observations. However, if the mouse TS promoter were not regulated in the same manner, this would indicate that the two viruses exploit different molecular mechanisms. To explore these possibilities, HELF cells were transiently transfected with the pTSWTGL3 and pTSWTGL3(k110) constructs, serum-starved and then infected with HCMV for

Fig. 7. HCMV and MCMV transactivate the TS promoter by different mechanisms. (A) HELF cells were transiently transfected with 2 µg of the promoter-less pGL3, pTSWTGL3 (which contains the wild-type mouse TS promoter region), or pTSWTGL3(k110) (which contains the mouse TS promoter with the E2F mutation) and 10 µg carrier DNA (pBluescript SK). After 18 h, cells were washed and growth-arrested in low-serum medium for 48 h. Thereafter, transfectants were infected with HCMV or mockinfected. Total cytoplasmic extracts were isolated at 24 h p.i. and assayed for luciferase activity. Reporter gene activity was normalized to the amount of plasmid DNA introduced into recipient cells as determined by dot-blot analysis. The normalized luciferase activity is expressed as fold induction relative to basal levels measured in cells transfected with pTSWTGL3 or pTSWTGL3(k110) and then mock-infected, which was set at 1. The data shown are the averages of two experimentspSEM (error bars). (B) NIH 3T3 cells were transiently transfected with 2 µg of the promoter-less pGL3, phTS k161/k98 Luc, phTS k161/k141 Luc and 10 µg carrier DNA (pBluescript SK). After 18 h, cells were washed and then maintained for 48 h in medium containing 0n5 % serum. Thereafter, transfectants were infected with MCMV or mock-infected. Total cytoplasmic extracts were isolated at 24 h p.i. and assayed for luciferase activity. The resulting luciferase activity is expressed as fold induction relative to basal levels measured in cells transfected with phTS k161/k98 Luc or phTS k161/k141 Luc and mock infected, which was set as 1. The data shown are the averages of two experimentspSEM (error bars).

24 h. The pTSWTGL3 contains the luciferase gene driven by the wild-type promoter and 5h flanking region of the mouse TS gene. pTSWTGL3(k110) is the same as pTSWTGL3 except that the E2F binding site at k110 (TCTGGCGG) has been mutated to TCGCTAGC (Geng & Johnson, 1993). Fig. 7(A) CJIJ


shows that pTSWTGL3 and pTSWTGL3(k110) were transactivated by HCMV infection by fivefold and by more than 11-fold respectively. Surprisingly, inactivation of the E2F binding site led to a more pronounced response to HCMV infection than that measured with the wild-type construct. We then performed the reciprocal experiment. Murine NIH 3T3 cells were transiently transfected with the human TS promoter constructs phTS k161\k98 Luc or phTS k161\ k141 Luc, serum-starved and then infected with MCMV for 24h. As shown in Fig. 7(B), MCMV infection increased the luciferase activity by about twofold in cells transfected with phTS k161\k98 Luc, which retains the E2F elements. However, the increase was almost completely abolished in cells transfected with phTS k161\k141 Luc, which lacks the E2F elements. This indicates that MCMV is able to regulate (at least in part) the human TS promoter through an E2Fdependent mechanism. Taken together these results confirm that MCMV and HCMV activate the TS promoter by different mechanisms.

Discussion During natural infection, HCMV productively replicates in terminally differentiated epithelial, endothelial and smooth muscle cells despite the absence of virus-encoded enzymes involved in the de novo biosynthesis of dTMP. In addition, infection of fibroblasts with HCMV leads to an expansion of cellular dTTP pools (Suzuki et al., 1985 ; Biron et al., 1986). These observations suggest that HCMV is able to stimulate the expression of cellular enzymes involved in the synthesis of dTMP. In the present study, we show that HCMV infection of quiescent fibroblasts stimulates the transcription of the cellular TS gene and that TS activity is required for efficient viral DNA synthesis. The dependence of HCMV replication on TS activity is supported by the results obtained with the folate analogue raltitrexed, a powerful TS inhibitor that prevents DNA synthesis and repair by blocking the de novo synthesis of dTMP. We have shown that raltitrexed inhibited HCMV DNA synthesis in quiescent cells, indicating that induction of TS activity is required for viral DNA replication. In line with this, we have previously observed that raltitrexed had little effect on MCMV and HCMV IE gene expression, although it strongly inhibited late gene expression as well as the replication of the virus in quiescent cells (Lembo et al., 2000). Taken together, these observations suggest that raltitrexed inhibits HCMV replication by blocking viral DNA synthesis. We also found that raltitrexed is far more detrimental to CMV DNA replication than to the survival of uninfected quiescent cells. This is probably due to the fact that TS is present at very low levels and is irrelevant for the survival of quiescent cells (Jenh et al., 1985 ; Johnson, 1994), but is required for viral DNA replication. These observations suggest that drugs that target TS may be highly effective in treating CMV infections. CJJA

Several laboratories have recently shown that HCMV infection blocks cell-cycle progression and cellular DNA replication while activating the cellular DNA synthetic machinery. These alterations create a favourable environment for high-level viral DNA replication, since they provide the viral DNA polymerase with dNTPs while preventing competition by cellular DNA synthesis (Fortunato et al., 2000 ; Flemington, 2001). Activation of host genes important for DNA synthesis has been reported to depend on binding of CMV to the cell surface or expression of viral IE proteins (Fortunato et al., 2000). Here we have shown that : (i) TS mRNA upregulation was prevented in the absence of protein synthesis ; and (ii) inactivation of HCMV by UV exposure abolished both the induction of TS protein and the activation of the TS promoter. These results indicate that viral IE and E gene expression, rather than interaction of viral particles with the cell surface, is required to stimulate TS gene expression. Furthermore, we have shown that the product of the IE1 gene partially transactivated the TS promoter and that the virus-dependent activation of TS transcription coincided with expression of the IE1 protein. However, the difference between the levels of transactivation of the human TS promoter in HCMV-infected and in IE1-transfected cells suggests that viral IE and\or E gene products other than IE1 contribute to the overall response of the human TS promoter to virus infection. To map the sequences important for activation of the TS promoter by HCMV, we assessed its ability to stimulate TS promoters modified by deletions or point mutations. Two positive-acting elements were identified. The first was located between k70 and k98 (relative to the AUG start codon) and corresponded to one of the direct repeats of a 28-nucleotide G\C-rich sequence. The human TS promoter is polymorphic, containing either two or three tandem direct repeats in addition to a single inverted copy of the repeat (Horie et al., 1995 ; Marsh et al., 1999). The promoter examined in this study had two copies of the direct repeat sequence. We showed that deletion of the downstream direct repeat had little effect on TS promoter activity following HCMV infection, but elimination of both direct repeats significantly reduces expression of the reporter gene in virus-infected cells. Earlier studies did not identify potential binding sites for transcription factors within the repeated sequences (Horie et al., 1992). In addition, we have shown that deletion of all of the direct repeats has little effect on expression of reporter genes driven by the TS promoter in proliferating cells (Dong et al., 2000). Therefore, the direct repeat may contribute to the human TS gene expression only under certain conditions, such as virus infection. The repeated sequences have also been reported to affect the efficiency of translation of TS mRNA (Kaneda et al., 1987 ; Horie et al., 1995 ; Kawakami et al., 2001). For this reason, it is not clear if the repeated sequence has a positive effect on TS promoter activity or on mRNA translation following HCMV infection. The second postive-acting element was located between k161 and k141. This segment has been defined as the TS


essential promoter region, since it is both necessary and sufficient for efficient promoter activity in proliferating cells. This region contains binding sites for Ets, Sp1 and LSF transcription factors, and inactivation of any of these motifs leads to a significant decrease in promoter activity (Dong et al., 2000 ; Powell et al., 2000). Previously, it was shown that Sp1 binds to its cognate motif within the essential region of the human TS promoter (Horie & Takeishi, 1997) and that an increase in Sp1 DNA binding activity occurs in HELF cells during HCMV infection (Yurochko et al., 1995, 1997). However, mutation of the Sp1 site (Fig. 3) in the phTS k161\k141 Luc construct had no significant effect on stimulation of TS promoter activity following virus infection (data not shown), suggesting that the Sp1 site is not important for the response to HCMV infection. It remains to be established whether the Ets and\or the LSF elements may be required for the stimulation of the TS essential promoter region by HCMV. Relevant to this, the DNA-binding activity of an Ets family member, Elk-1, was found to increase in HCMV-infected fibroblasts (Chen & Stinski, 2000). The human TS promoter has two E2F-binding sites within the inverted repeat sequence downstream from the essential promoter region (Dong et al., 2000). However, deletion of both of these E2F motifs did not significantly diminish the expression of the reporter gene following HCMV infection, indicating that these sequences do not have a significant role in the response of the TS promoter to the virus. This is surprising since it has been reported that HCMV infection results in an E2F-dependent activation of the DHFR promoter (Wade et al., 1992 ; Margolis et al., 1995) and since an E2F element is necessary for the activation of the mouse TS promoter by murine CMV (Gribaudo et al., 2000). These observations raised the possibility that MCMV and HCMV may activate the TS promoter by different mechanisms. To explore this possibility, we examined the effect of MCMV and HCMV on the expression of the human or mouse TS promoters, respectively. We found that the murine TS promoter with a mutated E2F site was activated to a higher extent than the wild-type promoter when introduced into human cells that were then infected with HCMV. In contrast, deletion of both E2F sites of the human TS promoter abolished its response to MCMV when the construct was introduced into NIH 3T3 cells. These results suggest that MCMV activates the TS promoter through an E2F-dependent pathway, whereas HCMV relies on a different mechanism. TS catalyses the de novo synthesis of dTMP by the reductive methylation of dUMP, and availability of the substrate may thus be rate-limiting for dTMP biosynthesis. There are several pathways by which dUMP can be produced in the cell, although deamination of dCMP by deoxycytidylate deaminase (dCMP deaminase) appears to be the most important route. dCMP deaminase activity is much higher in rapidly dividing cells than in quiescent cells and is expressed at the highest levels during the S phase of the cell cycle (Maley

& Maley, 1990). We have recently observed that HCMV infection of quiescent HELF cells also leads to an increase in dCMP deaminase gene expression (data not shown). These findings, along with previous studies showing increased DHFR activity during HCMV infection (Lembo et al., 1999), demonstrate that HCMV is able to coordinately activate the expression of multiple cellular enzymes involved in the synthesis of dTMP, thereby releasing the virus from the requirement of an S-phase environment for its replication. This work was supported by grants from MURST-CNR Biotechnology program L. 95\95, MURST (40 % and 60 %) to G. G. and S. L, from the AIDS Research Project to S. L., and by grants from the National Institute for General Medical Sciences (GM29356) and a Comprehensive Cancer Center Core Grant from the National Cancer Institute (CA16058) to L. F. J. T. L. R. was supported by a training grant (T32 CA09498) from the National Cancer Institute. This work is dedicated to Franco Tato' .

References Abken, H. & Reifenrath, B. (1992). A procedure to standardize CAT

reporter gene assay. Nucleic Acids Research 20, 3527. Ash, J., Ke, Y., Korb, M. & Johnson, L. F. (1993). Introns are essential for growth-regulated expression of the mouse thymidylate synthase gene. Molecular and Cellular Biology 13, 1565–1571. Ash, J., Liao, W.-C., Ke, Y. & Johnson, L. F. (1995). Regulation of mouse thymidylate synthase gene expression in growth-stimulated cells : upstream S phase control elements are indistinguishable from the essential promoter elements. Nucleic Acids Research 23, 4649–4656. Ayusawa, D., Takeishi, K., Kaneda, S., Shimizu, K., Koyama, H. & Seno, T. (1984). Isolation of functional cDNA clones for human thymidylate

synthase. Journal of Biological Chemistry 259, 14361–14364. Ayusawa, D., Shimizu, K., Koyama, H., Kaneda, S., Takeishi, K. & Seno, T. (1986). Cell-cycle-directed regulation of thymidylate synthase

messenger RNA in human diploid fibroblasts stimulated to proliferate. Journal of Molecular Biology 190, 559–567. Biron, K. K., Fyfe, J. A., Stanat, S. C., Leslie, L. K., Sorrell, J. A., Lambe, C. U. & Coen, D. M. (1986). A human cytomegalovirus mutant resistant

to the nucleoside analog 9-[2-hydroxy-1-(hydroxymethyl)ethoxy]methylguanine (BW B759U) induces reduced levels of BW B759U triphosphate. Proceedings of the National Academy of Sciences, USA 83, 8769–8773. Cavallo, R., Lembo, D., Gribaudo, G. & Landolfo, S. (2001). Murine cytomegalovirus infection induces cellular folylpolyglutamate synthetase activity in quiescent cells. Intervirology 44, 224–226. Chee, M. S., Bankier, A. T., Beck, S., Bohni, R., Brown, C. M., Cerny, R., Horsnell, T., Hutchison, III, C. A., Kouzarides, T., Martignetti, J. A., Preddie, E., Satchwell, S. C., Tomlinson, P., Weston, K. M. & Barrell, B. G. (1990). Analysis of the protein-coding content of human cytome-

galovirus strain AD169. Current Topics in Microbiology and Immunology 154, 125–169. Cheng, J. & Stinski, M. F. (2000). Activation of the human cytomegalovirus early UL4 promoter by the Ets transcription factor binding element. Journal of Virology 74, 9845–9857. Dong, S., Lester, L. & Johnson, L. F. (2000). Transcriptional control elements and complex initiation pattern of the TATA-less bidirectional human thymidylate synthase promoter. Journal of Cellular Biochemistry 77, 50–64. CJJB

G. Gribaudo and others Flemington, E. K. (2001). Herpesvirus lytic replication and the cell

cycle : arresting new developments. Journal of Virology 75, 4475–4481. Fortunato, E. A. & Spector, D. H. (1999). Regulation of human cytomegalovirus gene expression. Advances in Virus Research 54, 61–128. Fortunato, E. A., McElroy, A. K., Sanchez, V. & Spector, D. H. (2000).

Exploitation of cellular signaling and regulatory pathways by human cytomegalovirus. Trends in Microbiology 8, 111–119. Geng, Y. & Johnson, L. F. (1993). Lack of an initiator element is responsible for multiple transcriptional initiation sites of the TATA-less mouse thymidylate synthase promoter. Molecular and Cellular Biology 13, 4894–4903. Gribaudo, G., Ravaglia, S., Gaboli, M., Gariglio, M., Cavallo, R. & Landolfo, S. (1995). Interferon-α inhibits the murine cytomegalovirus

immediate-early gene expression by down-regulating NF-κB activity. Virology 211, 251–260. Gribaudo, G., Riera, L., Lembo, D., De Andrea, M., Gariglio, M., Rudge, T. L., Johnson, L. F. & Landolfo, S. (2000). Murine cytomegalovirus

stimulates cellular thymidylate synthase gene expression in quiescent cells and requires the enzyme for replication. Journal of Virology 74, 4979–4987. Griffith, P. D. (2000). Cytomegalovirus. In Principles and Practice of Clinical Virology, 4th edn, pp. 79–115. Edited by A. J. Zuckerman, J. E. Banatvala & J. R. Pattison. New York : Wiley. Hengel, H. & Weber, C. (2000). Driving cells into atherosclerotic lesions – a deleterious role for viral chemokine receptors? Trends in Microbiology 8, 294–296. Hengel, H., Brune, W. & Koszinowski, U. H. (1998). Immune evasion by cytomegalovirus – survival strategies of a highly adapted opportunist. Trends in Microbiology 6, 190–197. Horie, N. & Takeishi, K. (1997). Identification of functional elements in the promoter region of the human gene for thymidylate synthase and nuclear factors that regulate the expression of the gene. Journal of Biological Chemistry 272, 18375–18381. Horie, N., Nozawa, R. & Takeishi, K. (1992). Identification of cellular differentiation-dependent nuclear factors that bind to a human gene for thymidylate synthase. Biochemical and Biophysical Research Communications 185, 127–133. Horie, N., Aiba, H., Oguro, K., Hojo, H. & Takeishi, K. (1995).

Functional analysis and DNA polymorphism of the tandemly repeated sequences in the 5h-terminal regulatory region of the human gene for thymidylate synthase. Cell Structure and Function 20, 191–197. Jenh, C. H., Rao, L. G. & Johnson, L. F. (1985). Regulation of thymidylate synthase enzyme synthesis in 5-fluorodeoxyuridine-resistant mouse fibroblasts during the transition from the resting to the growing state. Journal of Cellular Physiology 122, 149–154. Johnson, L. F. (1994). Posttranscriptional regulation of thymidylate synthase gene expression. Journal of Cellular Biochemistry 54, 387–392. Johnston, P. G., Liang, C.-M., Henry, S., Chabner, B. A. & Allegra, C. J. (1991). Production and characterization of monoclonal antibodies

Kawakami, K., Salonga, D., Park, J. M., Danenberg, K. D., Uetake, H., Brabender, J., Omura, K., Watanabe, G. & Danenberg, P. V. (2001).

Different lengths of a polymorphic repeat sequence in the thymidylate synthase gene affect translational efficiency but not its gene expression. Clinical Cancer Research 7, 4096–4101. Ke, Y., Ash, J. & Johnson, L. F. (1996). Splicing signals are required for S-phase regulation of the mouse thymidylate synthase gene. Molecular and Cellular Biology 16, 376–383. Klucher, K. M., Sommer, M., Kadonaga, J. T. & Spector, D. H. (1993).

In vivo and in vitro analysis of transcription activation mediated by the human cytomegalovirus major immediate-early proteins. Molecular and Cellular Biology 13, 1238–1250. Lalani, A. S., Barrett, J. W. & McFadden, G. (2000). Modulating chemokines : more lessons from viruses. Immunology Today 21, 100–106. Lembo, D., Angeretti, A., Gariglio, M. & Landolfo, S. (1998). Murine cytomegalovirus induces expression and enzyme activity of dihydrofolate reductase in quiescent cells. Journal of General Virology 78, 2803–2808. Lembo, D., Gribaudo, G., Cavallo, R., Riera, L., Angeretti, A., Hertel, L. & Landolfo, S. (1999). Human cytomegalovirus stimulates cellular

dihydrofolate reductase activity in quiescent cells. Intervirology 42, 30–36. Lembo, D., Gribaudo, G., Riera, L., Mondo, A., Cavallo, R., Angeretti, A. & Landolfo, S. (2000). The thymidylate synthase inhibitor ZD1694

potently inhibits murine and human cytomegalovirus replication in quiescent cells. Antiviral Research 47, 111–120. Maley, F. & Maley, G. F. (1990). A tale of two enzymes, deoxycytidylate deaminase and thymidylate synthase. Progress in Nucleic Acid Research and Molecular Biology 39, 49–80. Margolis, M. J., Pajovic, S., Wong, E. L., Wade, M., Jupp, R., Nelson, J. A. & Clifford Azizkhan, J. (1995). Interaction of the 72-kilodalton

human cytomegalovirus IE1 gene product with E2F1 coincides with E2Fdependent activation of dihydrofolate reductase transcription. Journal of Virology 69, 7759–7767. Marsh, S., Collie-Diguid, E. S. R., Li, T., Liu, X. & McLeod, H. L. (1999).

Ethnic variation in the thymidylate synthase enhancer region polymorphism among Caucasian and Asian populations. Genomics 58, 310–312. Mocarski, E. S. & Courcelle, C. T. (2001). Cytomegalovirus and their replication. In Fields Virology, 4th edn, pp. 2629–2673. Edited by D. M. Knipe & P. M. Howley. New York : Lippincott–Raven. Pauwels, R., Balzarini, J., Baba, M., Snoeck, R., Schols, D., Hederwijin, P., Desmyter, J. & De Clerq, E. (1988). Rapid and automated

tetrazolium-based colorimetric assay for the detection of anti HIV compounds. Journal of Virological Methods 20, 309–321. Powell, C. M. H., Rudge, T. L., Zhu, Q., Johnson, L. F. & Hansen, U. (2000). Inhibition of the mammalian transcription factor LSF induces S-

Kaneda, S., Horie, N., Takeishi, K., Takayanagi, A., Seno, T. & Ayusawa, D. (1992). Regulatory sequences clustered at the 5h end of the first intron

phase-dependent apoptosis by downregulating thymidylate synthase expression. EMBO Journal 19, 4665–4675. Rawlinson, W. D., Farrell, H. E. & Barrell, B. G. (1996). Analysis of the complete DNA sequence of murine cytomegalovirus. Journal of Virology 70, 8833–8849. Staprans, S. I., Rabert, D. K. & Spector, D. H. (1988). Identification of sequence requirement and trans-acting functions necessary for regulated expression of a human cytomegalovirus early gene. Journal of Virology 62, 3463–3473.

of the human thymidylate synthase gene function in cooperation with the promoter region. Somatic Cell and Molecular Genetics 18, 409–415.

Suzuki, S., Saneyoshi, M., Nakayama, C., Nishiyama, Y. & Yoshida, S. (1985). Mechanism of selective inhibition of human cytomegalovirus

that localize human thymidylate synthase in the cytoplasm of human cells and tissue. Cancer Research 51, 6668–6676. Kaneda, S., Takeishi, K., Ayusawa, D., Shimizu, K. A., Seno, T. & Altman, S. (1987). Role in translation of a triple tandemly repeated

sequence in the 5h-untranslated region of human thymidylate synthase mRNA. Nucleic Acids Research 15, 1259–1270.

CJJC

HCMV stimulates thymidylate synthase expression replication by 1-β--arabinofuranosyl-5-fluorouracil. Antimicrobial Agents and Chemotherapy 28, 326–330. Takayanagi, A., Kaneda, S., Ayusawa, D. & Seno, T. (1992). Intron 1 and the 5h-flanking region of the human thymidylate synthase gene as a regulatory determinant of growth-dependent expression. Nucleic Acids Research 20, 4021–4025. Takeishi, K., Kaneda, S., Ayusawa, D., Shimizu, K., Gotoh, O. & Seno, T. (1989). Human thymidylate synthase gene : isolation of phage clones

which cover a functionally active gene and structural analysis of the region upstream from the translation initiation codon. Journal of Biochemistry 106, 575–583.

Wiebusch, L. & Hagemeier, C. (2001). The human cytomegalovirus

immediate early 2 protein dissociates cellular DNA synthesis from cyclindependent kinase activation. EMBO Journal 20, 1086–1098. Yurochko, A. D., Kowalik, T. F., Huong, S. M. & Huang, E.-S. (1995).

Human cytomegalovirus upregulates NF-kappaB activity by transactivating the NF-κB p105\p50 and p65 promoter. Journal of Virology 69, 5391–5400. Yurochko, A. D., Mayo, M. W., Poma, E. E., Baldwin, A. S., Jr & Huang, E.-S. (1997). Induction of the transcription factor SP1 during human

cytomegalovirus infection mediates upregulation of the p65 and p105\p50 NF-κB promoters. Journal of Virology 71, 4638–4648.

Wade, M., Kowalik, T. F., Mudryj, M., Huang, E.-S. & Clifford Azizkhan, J. (1992). E2F mediates dihydrofolate reductase promoter activation and

multiprotein complex formation in human cytomegalovirus infection. Molecular and Cellular Biology 12, 4364–4374.

Received 7 February 2002 ; Accepted 22 August 2002

CJJD