Journal of Clinical Virology 35 (2006) 478–484

Human cytomegalovirus temporally regulated gene expression indifferentiated, immortalized retinal pigment epithelial cells

Richard Adaira,1, Gregory W. Liebischa,Bruce J. Lermana,c, Anamaris M. Colberg-Poleya,b,∗

a Center for Cancer and Immunology Research, Children’s Research Institute, Children’s National Medical Center, Washington, DC 20010, USAb Department of Pediatrics, George Washington University School of Medicine and Health Sciences, Washington, DC 20037, USA

c Pharmacology Program of the Institute for Biomedical Sciences at the George Washington University, Washington, DC 20037, USA

Received 13 May 2005; received in revised form 21 October 2005; accepted 25 October 2005

Abstract

Background: Human cytomegalovirus (HCMV) replication in epithelial cells is crucial for its pathogenesis. To date, HCMV gene expressionhas been primarily studied in human foreskin fibroblasts (HFFs), although their importance for HCMV pathogenesis remains unclear. Primary

UL37

ome keyteins

UL37roduced72 hpi.ced within

r, HCMVn suggests

alto

sedimar-;-

piteto

retinal pigment epithelial (RPE) cells are permissive for HCMV.Objectives: Our objectives were to determine the production of alternatively processed HCMV major immediate-early (MIE) andRNAs and their essential products in infected, terminally differentiated immortalized RPE (hTERT-RPE) cells.Study design: hTERT-RPE cells were studied because of their notable similarities with primary RPE cells, and because they overclimitations of primary cells. hTERT-RPE cells were terminally differentiated in vitro and infected with HCMV. Total RNA or cell prowere analyzed at various times post-infection.Results: We show for the first time that HCMV-infected, differentiated hTERT-RPE cells produce IE1, IE2, UL37 exon 1 (UL37x1) andalternatively spliced RNAs, albeit with abundances and kinetics distinct from those observed in HCMV-infected HFFs. IE1-72 was pin HCMV-infected, differentiated hTERT-RPEs within 24 h post-infection (hpi); whereas, IE2-86 and pUL37x1 were produced withinIE2-86 was detected after IE1-72 even though its transcript appeared first. Early/late (pp65) and late (pp28) proteins were produ96–120 hpi.Conclusions: The temporal cascade of HCMV gene expression was observed in infected, differentiated hTERT-RPE cells. MoreoveIE RNAs are alternatively and accurately processed in differentiated hTERT-RPE cells. However, the delayed temporal expressiofurther regulation of HCMV gene expression at post-transcriptional/translational levels in differentiated hTERT-RPE cells.© 2005 Elsevier B.V. All rights reserved.

Keywords: HCMV MIE; UL37 RNAs; Alternative processing; RPE cells; hTERT-RPE cells

Abbreviations: FCS, fetal calf serum; HFFs, human foreskin fibroblasts;hTERT-RPE, human telomerase reverse transcriptase immortalized retinalpigment epithelial cells; RT-PCR, reverse transcription-polymerase chainreaction; IE, immediate-early

∗ Corresponding author at: Center for Cancer and Immunology Research,Children’s Research Institute, Children’s National Medical Center, Room5720, 111 Michigan Ave., NW, Washington, DC 20010, USA. Tel.: +1 202884 3984; fax: +1 202 884 3929.

E-mail address: acolberg-poley@cnmcresearch.org (A.M. Colberg-Poley).

1 Present address: MRC Virology Unit, Institute of Virology, ChurchStreet, Glasgow, G11 5JR Scotland, UK.

1. Introduction

Human cytomegalovirus (HCMV) is the leading vircause of congenital infections and a major contributormorbidity in immunosuppressed patients (Pass, 2001). Incongenitally infected newborns and in immunosupprespatients, HCMV causes diseases in tissues composed prily of epithelial cells including the retina (Coats et al., 2000Holland, 1994). HCMV infections are a well-described complication of allogeneic transplantation and one that, dessignificant advances in prophylactic therapy, continues

1386-6532/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.jcv.2005.10.015

R. Adair et al. / Journal of Clinical Virology 35 (2006) 478–484 479

Fig. 1. HCMV genes examined in differentiated hTERT-RPE cells. The HCMV genome (strain AD169, EMBL accession #17403) is depicted with the terminalrepeat long (TRL), internal repeat long (IRL), internal repeat short (IRS) and terminal repeat short (TRS) regions marked with grey boxes. The unique long(UL) and unique short (US) sequences are shown by a thin line. The approximate physical location and direction of transcription of the pp65 early/late andthe pp28 late genes are indicated above the genome (Chee et al., 1990). Depicted in more detail below the genome are the UL36-38 (Adair et al., 2003) andUL122-123 (Awasthi et al., 2004) IE loci. Alternatively processed UL37 and IE1 and IE2 transcripts are shown. Throughout, white boxes represent non-codingexons, black boxes represent exonic sequences that are translated, thin bent lines represent splicing of introns, thin straight lines represent untranslated regionsand arrow heads represent polyadenylation sites.

produce significant morbidity and mortality following bonemarrow transplantation (Boeckh and Nichols, 2004). HCMVinfections may contribute to the development of atheroscle-rosis, cardiac allograft rejection, and restenosis (Blum et al.,1998; Epstein et al., 1996; Speir et al., 1994; Zhou et al., 1996)and of malignant glioma (Cobbs et al., 2002). Despite reduc-tions due to HAART therapy, retinitis is still a significantcomplication of HCMV infection in AIDS patients (Salmon-Ceron, 2001; Scholz et al., 2003, 2004). Moreover, HCMVDNA load is a more reliable negative survival predictor thanHIV RNA load for patients with advanced AIDS (Spector etal., 1999).

HCMV genes are transcribed in a tightly regulatedcascade during its life cycle: the first genes transcribedare immediate-early (IE) genes. The most abundant IERNAs, products of the major IE (MIE, UL122-123) locus,encode IE1-72 and IE2-86 (Mocarski and Courcelle, 2001),which are detectable within a few hours in HCMV-infectedfibroblasts (Stinski et al., 1983; Xu et al., 2001). Otherkey HCMV IE products are expressed from the UL36-38locus (Colberg-Poley, 1996) (Fig. 1). The essential UL37exon 1 protein (pUL37x1) is produced within 8 h post-infection (hpi), traffics through the secretory apparatusand into mitochondria, and has anti-apoptotic activity(Colberg-Poley et al., 2000; Goldmacher et al., 1999;Hayajneh et al., 2001; Mavinakere and Colberg-Poley,2 04T theU es

pUL37x1, and a lower abundance UL37 spliced RNA, whichencodes the UL37 glycoprotein, gpUL37 (Kouzarides etal., 1988; Tenney and Colberg-Poley, 1991a,b). In addi-tion, there are multiple alternatively spliced UL37 RNAsproduced at low abundance in infected human foreskinfibroblasts (HFFs) (Adair et al., 2003; Goldmacher et al.,1999).

Primary HFFs are most commonly used for studying theHCMV life cycle, as they are permissive for its replication.However, HFFs are difficult to transfect efficiently, are slowgrowing, and have a limited lifespan. Moreover, their rele-vance for HCMV pathogenesis remains unclear (Hendrix etal., 1997; Sinzger et al., 1995). Nevertheless, due to theiravailability and permissivity, HCMV gene expression hasbeen mostly studied in HFFs.

HCMV replicates in cultured primary retinal pigmentepithelial (RPE) cells, although more slowly than in HFFs(Detrick et al., 1996). HCMV enters RPE cells by endocyto-sis, rather than membrane fusion as with HFFs (Bodaghi etal., 1999). IE1 was first detected by immunofluorescence by24 hpi in HCMV-infected RPE cells (Bodaghi et al., 1999).Since HCMV particles enter RPE cells within minutes, slowproduction of HCMV IE1 protein could not be accountedfor by rate of virus uptake. In addition, we have recentlyshown that HCMV infection rapidly alters the abundanceof polyadenylation and splicing factors in actively dividingh ,w sioni

004a,b; McCormick et al., 2003; Reboredo et al., 20).he two major alternatively processed UL37 RNAs areL37 exon 1 (UL37x1) unspliced RNA, which encod

TERT-RPE cells (Adair et al., 2004). In this investigatione set out to examine the kinetics of HCMV gene expres

n differentiated hTERT-RPE cells.

480 R. Adair et al. / Journal of Clinical Virology 35 (2006) 478–484

Table 1RT-PCR primers used in this study

Primer Gene Sequence (nt)

140 UL37x1 52259–52236a,b

196 UL37x3 50542–50521a,b

264 UL37x1 52541–52520a,b

298 UL83 (pp65) 120833–120817a,b

299 UL83 (pp65) 120272–120256a,b

300 UL123 (IE1) exon 2 172752–172726a,b

301 UL123 (IE1) exon 4 172219–172194a,b

321 UL122 (IE2) exon 5 170582–170561a,b

322 UL122 (IE2) exon 5 170319–170300a,b

395 �-Actin 832–851c,d

396 �-Actin 1147–1165c,d

a HCMV genome (EMBL accession no. X17403).b Primers 140, 196, 264, 298-301, 321 and 322 were published inAdair

et al. (2004).c �-Actin (GenBank accession no. BC012854).d The sequence used to design primers 395 and 396 was published in

Strausberg et al. (2002).

2. Methods

2.1. Cells and viruses

hTERT-RPEs (Clontech) were grown as previouslydescribed (Adair et al., 2004). To differentiate the cells,hTERT-RPEs were grown to confluency and then serumstarved (0% FCS) for 4 weeks. Cellular differentiationresulted in a confluent monolayer of polygonal cells asdescribed (Rambhatla et al., 2002). Differentiated hTERT-RPEs were infected with HCMV strain AD169 variant ATCC(varATCC, Manassas, VA) or variant HB5 (BAC-AD169,Borst et al., 1999) at multiplicities of infection (moi) of3 PFU/cell, 10 PFU/cell, or 30 PFU/cell. Infections were per-formed without any drug treatment, using high titer stocks,which were diluted prior to use. Following infection, the cellswere overlaid with serum free medium.

2.2. RT-PCR

Total RNA was harvested at various hpi using the Ribop-ure Kit (Ambion) as previously described (Su et al., 2003a,b).PCR products were not observed in the absence of reversetranscriptase, confirming the use of RNA as templates.cDNAs (50 ng, 250 ng or 500 ng) were amplified by PCRwith HotStar Taq polymerase (Qiagen) using primers tods

2

reae arya 00)o p28( L.

Pereira, Pers. Commun.) or actin (sc8432, 1:100, SantaCruz). The membranes were then incubated with horseradishperoxidase-conjugated goat anti-mouse or rabbit IgG (Bio-Rad) and reactive proteins were detected by ECL chemilu-minescent reagents (Amersham).

3. Results

3.1. HCMV MIE and UL37x1 IE RNAs are produced indifferentiated hTERT-RPE cells

The kinetics of MIE RNA or alternative processing ofUL37 RNAs in HCMV-infected, differentiated hTERT-RPEcells has not been examined. To that end, hTERT-RPEs werefully differentiated and infected with HCMV (Fig. 2). IE2RNA was detected by RT-PCR from 1 hpi. Surprisingly, IE1RNA was detected later than IE2 RNA since IE1 was firstdetected weakly at 3 and 2 hpi at moi of 3 and of 30, respec-tively. This experiment was independently repeated with thesame result.

UL37x1 unspliced IE RNA, which is the predomi-nant UL37 RNA in HCMV-infected HFFs, was detectedin HCMV-infected, differentiated hTERT-RPEs at 8 hpi

F ofH PEc C,3 atedu g totT -gen) and cDNAs amplified by PCR to detect UL37x1 (250 ng, primers264/140), UL37 (500 ng, primers 264/196), IE1 (250 ng, primers 300/301),IE2 (250 ng, primers 321/322) or pp65 (250 ng, primers 298/299) products.PCR reactions were incubated (i) at 94◦C for 15 min, (ii) then at 95◦C for45 s, 60◦C for 1 min, 72◦C for 2 min (35 cycles), and (iii) finally, at 72◦Cfor 10 min. The PCR products were resolved by agarose gel electrophore-sis. The lanes include size markers (m), water control, uninfected cell andHCMV-infected cell RNA at 1, 2, 3, 4, and 8 hpi.

etect UL37x1, UL37, IE1, IE2, pp65, and control�-actinequences (Table 1).

.3. Western analysis

HCMV-infected differentiated hTERT-RPE cells wenalyzed by Western analysis as previously described (Adairt al., 2003). The membranes were reacted with primntibodies including rabbit anti-pUL37x1 (Ab1064, 1:10r mouse anti-IE1/IE2 (MAb810, 1:500, Chemicon), pAb6502-100, 1:1000, Abcam), pp65 (CH65-1, 1:50,

ig. 2. Production of UL37 and MIE transcripts during initial timesCMV infection in differentiated hTERT-RPEs. Differentiated hTERT-Rells were uninfected or infected with HCMV strain AD169 (variant ATCPFU/cell or 30 PFU/cell) for the indicated times. Total RNA was isolsing a commercially available kit (Ribopure kit, Ambion) accordin

he manufacturer’s instructions as previously described (Su et al., 2003a,b).otal RNA was reverse transcribed using HotStarTaq polymerase (Qia

R. Adair et al. / Journal of Clinical Virology 35 (2006) 478–484 481

(moi = 3) and weakly at 3 hpi with a higher multiplicity(moi = 30). Conversely, UL37 spliced RNA was detected at8 hpi (moi = 3 and 30) but its detection required a largeramount of input cDNA (∼500 ng). Thus, the abundanceof UL37 spliced RNA is extremely low, as in HCMV-infected HFFs (Adair et al., 2003, 2004; Tenney and Colberg-Poley, 1991a). Nonetheless, it is noteworthy that expressionis delayed in hTERT-RPEs, since UL37x1 RNA has beendetected from infected HFFs by Northern blots (Tenney andColberg-Poley, 1991a) and RT-PCR at 4 hpi (Adair et al.,2004). Thus, HCMV IE RNAs can be detected in differenti-ated RPEs within 8 hpi, but their expression was later than ininfected HFFs.

pp65 early/late RNA, carried within HCMV virions(Terhune et al., 2004), was detected from initial times ofinfection but its abundance markedly increased within 8 hpi.Thus, the transition to early/late gene transcription wasobserved in HCMV-infected hTERT-RPE cells within 8 hpi.

3.2. Complexity of alternative processing of UL37pre-mRNA in HCMV-infected differentiated hTERT-RPEs

Multiple, alternatively spliced UL37 RNAs are producedin HCMV-infected HFFs and hTERT-HFFs (Adair et al.,2003, 2004). To determine whether this is also the case indifferentiated hTERT-RPE cells, we examined the complex-i n( Aw ast timepa s ins Fs,w andpC ned,a crip-t hpii

3d

ont C)i lysis( sedi s firstd wasm Aw sd 2-86i nsla-t

inedb As

Fig. 3. Alternative processing of UL37 spliced RNAs at early and late timesof HCMV infection in differentiated hTERT-RPEs. hTERT-RPEs were dif-ferentiated by serum starvation as inFig. 2and infected with HCMV (strainAD169, varATCC) at an moi of 3 PFU/cell. Total RNA was isolated at theindicated times, reverse transcribed and the resulting cDNAs were PCRamplified for one round of PCR to detect IE1 (200 ng, primers 300/301),pp65 (200 ng, primers 298/299), or�-actin (200 ng, primers 395/396). Thealternatively spliced UL37 RNAs were detected by RT-PCR (50 ng, primers264/196) as inFig. 2but for 40 cycles and subsequent amplification of thePCR product (2�l) for a second round (primers 264/196). The lanes includesize markers (m), water control, uninfected cell, and HCMV-infected cellRNA at 8, 24, 48, 72, 96, and 120 hpi.

Fig. 4. Delayed production of IE2 86 kDa in HCMV-infected differentiatedhTERT-RPE cells. hTERT-RPEs were differentiated as inFig. 2and infectedwith HCMV strain AD169 (varATCC) at an moi of 10 PFU/cell. Total proteinwas harvested by addition of lysis buffer (1 ml) to each flask of confluent dif-ferentiated hTERT-RPE cells (T-175) at the indicated times. Protein extracts(20�l/lane) were separated by electrophoresis in SDS-PAGE gels, blottedand reacted with antibodies against IE1/IE2 (MAb810, 1:500, Chemicon),pp65 (CH65-1, 1:50), and actin (sc8432, 1:100, Santa Cruz). The reactedmembranes were incubated with horseradish peroxidase-conjugated goatanti-mouse IgG (Bio-Rad), reacted with ECL chemiluminescent reagents(Amersham), and exposed to film. The migration of size markers are indi-cated on the left of the Western blots.

ty of UL37 pre-mRNA splicing during HCMV infectioFig. 3). In this experiment (moi = 3), UL37 spliced RNas not detected until 24 hpi. Notably, the UL37 RNA w

he most abundant UL37 spliced RNA detected at alloints after its initial production, although UL37di, UL37Mnd UL37Sdi, were detected in lower abundance. This itriking contrast to HCMV-infected HFFs and hTERT-HFhere UL37 splice variants differ in relative abundanceredominance throughout infection (Adair et al., 2003, 2004).ontrol reactions detected IE1 RNA at all times examis expected. Moreover, transition to early/late gene trans

ion was detected by the production of pp65 RNA within 8n HCMV-infected RPEs.

.3. HCMV IE1 and IE2 proteins are expressed withelayed kinetics in differentiated hTERT-RPE cells

To determine if MIE RNAs are rapidly translated upheir synthesis, we examined HCMV (AD169, varATCnfected differentiated hTERT-RPEs by Western anaFig. 4). IE1-72 was detected weakly at 24 hpi and increan abundance at subsequent times; whereas, IE2-86 waetected at 96 hpi. Equivalent loading of protein samplesonitored by reactivity with anti-actin antibody. As IE2 RNas detected prior to IE1 RNA (Fig. 2) and, yet, IE1-72 ietected first, these results suggest that production of IE

s regulated at a post-transcriptional and, possibly, traional level in HCMV-infected hTERT-RPE cells.

Transition to early/late gene expression was examy the production of pp65 in HCMV-infected RPEs.

482 R. Adair et al. / Journal of Clinical Virology 35 (2006) 478–484

Fig. 5. Temporal expression of HCMV IE (pUL37x1 and MIE), early/late(pp65), and late (pp28) proteins in HCMV-infected differentiated RPE cells.hTERT-RPEs were differentiated as inFig. 2 and infected with HCMVstrain AD169 variant HB5 (BAC-AD169,Borst et al., 1999) at an moi of3 PFU/cell. Total protein was harvested by addition of lysis buffer (1 ml) toeach flask of cells (T-175) at the indicated times. Protein extracts (20�l/lane)were separated by electrophoresis in SDS-PAGE gels, blotted and reactedwith antibodies against pUL37x1 (rabbit anti-pUL37x1 aa27-40, Ab1064,1:1000), IE1/IE2 (MAb810, 1:500, Chemicon), pp65 (CH65-1, 1:50), pp28(Ab6502-100, 1:1000, Abcam), and actin (sc8432, 1:100, Santa Cruz). Thereacted membranes were incubated with horseradish peroxidase-conjugatedgoat anti-mouse or rabbit IgG (Bio-Rad), reacted with ECL chemilumi-nescent reagents (Amersham), and exposed to film. The migration of sizemarkers are indicated on the left of the Western blots.

expected, tegument pp65 was detected immediately follow-ing HCMV infection. However, at 96 hpi, following theincreased synthesis of IE1 and IE2 proteins, a markedincrease in pp65 levels was observed indicating a transitionto early protein synthesis.

3.4. Production of pUL37x1 in HCMV-infecteddifferentiated RPEs

To determine whether expression of other IE proteins isalso delayed, we examined the production of pUL37x1 inHCMV-infected differentiated hTERT-RPE cells (Fig. 5). Forthese studies, we used HCMV (AD169 variant HB5,Borst etal., 1999) to determine whether other AD169 variants tempo-rally express their genes with similar patterns and kinetics.pUL37x1 was detected weakly within 72 hpi. The doublet,corresponding to phosphorylated pUL37x1 species previ-ously observed in HCMV-infected HFFs (Mavinakere andColberg-Poley, 2004a), was detected in infected differenti-ated hTERT-RPE cells. IE1-72 was produced within 24 hpi;whereas, IE2-86 was observed within 72 hpi. Transition toearly/late protein synthesis was documented by a markedincrease in pp65 production within 96 hpi (Fig. 5), as above(Fig. 4). This was followed by the production of pp28, within120 hpi, indicating transition to late protein synthesis (Fig. 5).T wasd , thek ig-n ost-

transcriptional and/or translational levels in HCMV-infectedhTERT-RPE cells.

4. Discussion

For these studies, we used hTERT-RPE cells that fol-lowing differentiation are similar to primary RPE cells incell growth, morphology and physiology (Rambhatla et al.,2002) and, yet, overcome key limitations of primary cells,including senescence and lot variations. In all cases whereexamined, differentiated hTERT-RPE cells behave compara-bly to primary RPE cells (Rambhatla et al., 2002). We setout to determine whether HCMV alternatively processed IEtranscripts and their products were produced and whethertransition to late stages of the HCMV life cycle would beobserved in differentiated hTERT-RPE cells. We now docu-ment that differentiated hTERT-RPE cells behave similarlyto primary RPE cells during HCMV infection.

Surprisingly, IE1 RNA was detected after IE2 RNA inHCMV-infected differentiated hTERT-RPEs; whereas, theirencoded proteins have the opposite kinetics of production.This is in contrast to HCMV-infected HFFs, where expres-sion of IE1 RNA typically precedes IE2 RNA (Stinski etal., 1983). The reason for this delay in hTERT-RPE cells isunknown. It is possible that failure to detect IE1 RNA priort thes fol-l IE2R uslyt res-c ,1

A inH Ap tiono ot locki pe-c unda alter-n tedH -t verp tions ;S tt IEp ono

-s nti-a tions gousa o IE1R in

hus, the temporal cascade of HCMV gene expressionocumented in differentiated hTERT-RPE cells. Howeverinetics of HCMV IE2-86 and pUL37x1 production are sificantly delayed, suggesting additional regulation at p

o 3 hpi in HCMV-infected hTERT-RPE cells is due toubstantial amount of virion degradation that is observedowing endocytosis. However, this seems unlikely sinceNA is readily detected and it has been reported previo

hat pp65 positive nuclei are observed using immunofluoence 1–2 h after viral entry into RPE cells (Bodaghi et al.999).

Our results suggest that processing of MIE pre-mRNCMV-infected hTERT-RPE cells initially favors IE2 RNroduction. This may result from transcriptional regulaf MIE gene.Sinclair et al. (1992)previously showed in vitr

hat monocytes are non-permissive for HCMV due to a bn expression from the MIE promoter via a differentiation sific cellular factor. Further, Spector and her coworkers forole of cyclin dependent kinases (1, 2, 5, 7, and 9) in theative processing of IE1 and IE2 RNAs in HCMV-infecFFs (Sanchez et al., 2004). Alternatively, IE2 RNA produc

ion may result from preferential splicing to IE2 acceptor oolyadenylation at the IE1 site. As the IE1 polyadenylaignal is close to the IE2 exon 5 acceptor (Chee et al., 1990anchez et al., 2004; Stinski et al., 1983), it is possible tha

he balance between cellular RNA factors which bind to Mre-mRNAcis-elements favor splicing and initial productif IE2 RNA.

In spite of similarities, HCMV UL37 and MIE trancripts are differentially expressed in infected, differeted hTERT-RPE cells. The UL37x1 RNA polyadenylaite is juxtaposed to the UL37x2 acceptor, in an analorrangement to the MIE locus. Nonetheless, in contrast tNA, UL37x1 RNA appears before UL37 spliced RNAs

R. Adair et al. / Journal of Clinical Virology 35 (2006) 478–484 483

HCMV-infected hTERT-RPE cells. Moreover, UL37 splicedRNAs are produced later than IE2 spliced RNAs in HCMV-infected RPEs. Finally, the production of alternatively splicedUL37 RNAs in HCMV-infected hTERT-RPE cells differs intiming and relative abundance from infected HFFs (Adairet al., 2003, 2004). UL37 RNA is predominately detectedat IE times (Tenney and Colberg-Poley, 1991a). At earlytimes, UL37M RNA is predominant and, at later times, otherUL37 splice RNAs increase in abundance (Adair et al., 2003;Goldmacher et al., 1999). In HCMV-infected, differentiatedhTERT-RPE cells, UL37 RNA was consistently the mostabundant UL37 splice variant throughout, although UL37di,UL37M, and UL37Sdi were detected. Identity of the UL37splice variants was confirmed by DNA sequencing the PCRproducts (unpublished results).

The role of the minor UL37 species during HCMV infec-tion is unknown. It is possible that the splice variants arenon-functional and result from aberrant splicing. However,the finding that the alternatively spliced UL37 RNAs areproduced in all HCMV-infected cells tested (HFFs, hTERT-HFFs, and differentiated hTERT-RPE cells), argues that theylikely represent products that perform important functionsduring HCMV infection (Adair et al., 2003). Splice vari-ants of IE1 have recently been described, although theircomplexity is less than UL37 transcripts (Adair et al.,2003; Awasthi et al., 2004). UL37 alternatively splicedR , ina tionb ,2

-86a PEc con-s da ellsw ctedw hpi.S -t Fs,I enicdT 37x1i u-l h asn ggestt tedh rtedi i etat andl e ofp usHe

RT-R for

studying the HCMV life cycle in a more biologically relevantcell line than previously available.

Acknowledgements

The authors thank Dr. Derrick Dargan for his critical com-ments. We are grateful to Dr. Martin Messerle for providingHCMV strain AD169 variant HB5 (BAC-AD169) and Dr.Lenore Pereira for the monoclonal antibody against pp65(CH65-1). This work was supported by Public Health ServiceGrants AI-46459 and AI-57906 from the National Institute ofAllergy and Infectious Diseases to ACP and an award fromthe Board of Visitors at Children’s National Medical Centerto RA.

References

Adair R, Liebisch GW, Colberg-Poley AM. Complex alternative pro-cessing of human cytomegalovirus UL37 pre-mRNA. J Gen Virol2003;84:3353–8.

Adair R, Liebisch GW, Su Y, Colberg-Poley AM. Alteration of cellu-lar RNA splicing and polyadenylation machineries during productivehuman cytomegalovirus infection. J Gen Virol 2004;85:3541–53.

Awasthi S, Isler JA, Alwine JC. Analysis of splice variants ofthe immediate-early 1 region of human cytomegalovirus. J Virol

B , ettedlloon

B RC,intovest

B s oftationlood

B ancialof

C R,e ofnol

C mo-ion.

C H, etalig-

C odedand

C anughVirol

D virusres-

NAs may alter the translation of other UL37 productsnalogy to repression of IE1 and IE2 protein producy the IE1-19 and IE1-17.5 kDa products (Awasthi et al.004).

We noted a significant delay in the production of IE2nd pUL37x1 in HCMV-infected, differentiated hTERT-Rells. Our results in differentiated hTERT-RPE cells areistent with the findings ofDetrick et al. (1996)who observedelay in HCMV protein expression in primary RPE chen compared to fibroblasts. Herein, IE2 RNA was deteithin 2 hpi and its product was first detected at 72imilarly, pUL37x1 is delayed >2 days (∼64 h) after the ini

ial detection of its encoding transcript at 8 hpi. In HFE1-72 localizes to nuclear promyelocytic protein oncogomains within 1–2 hpi just prior to IE2 (Xu et al., 2001).hese results suggest that production of IE2-86 and pUL

n HCMV-infected differentiated hTERT-RPE cells is regated at post-transcriptional or translational levels, sucuclear export or translation exclusion. These results su

hat the HCMV gene regulation and life cycle in differentiaTERT-RPEs are significantly different from those repo

n HFFs (Mavinakere and Colberg-Poley, 2004a; Stinskl., 1983; Tenney and Colberg-Poley, 1991a,b). This is fur-

her evidenced by the slow appearance of the early/lateate proteins in infected hTERT-RPEs. The appearancp28 correlated well with the initial production of infectioCMV progeny at 6 days pi in primary RPE cells (Detrickt al., 1996).

We conclude from these studies that differentiated hTEPEs provide a powerful yet tractable cell culture system

2004;78:8191–200.lum A, Giladi M, Weinberg M, Kaplan G, Pasternack H, Laniado S

al. High anti-cytomegalovirus (CMV) IgG antibody titer is associawith coronary artery disease and may predict post-coronary baangioplasty restenosis. Am J Cardiol 1998;81:866–8.

odaghi B, Slobbe-van Drunen ME, Topilko A, Perret E, Vossenvan Dam-Mieras MC, et al. Entry of human cytomegalovirusretinal pigment epithelial and endothelial cells by endocytosis. InOphthalmol Vis Sci 1999;40:2598–607.

oeckh M, Nichols WG. The impact of cytomegalovirus serostatudonor and recipient before hematopoietic stem cell transplanin the era of antiviral prophylaxis and preemptive therapy. B2004;103:2003–8.

orst EM, Hahn G, Koszinowski UH, Messerle M. Cloning of the humcytomegalovirus (HCMV) genome as an infectious bacterial artifichromosome inEscherichia coli: a new approach for constructionHCMV mutants. J Virol 1999;73:8320–9.

hee MS, Bankier AT, Beck S, Bohni R, Brown CM, Cernyet al. Analysis of the protein-coding content of the sequenchuman cytomegalovirus strain AD169. Curr Top Microbiol Immu1990;154:125–69.

oats DK, Demmler GJ, Paysse EA, Du LT, Libby C. Ophthallogic findings in children with congenital cytomegalovirus infectJ AAPOS 2000;4:110–6.

obbs CS, Harkins L, Samanta M, Gillespie GY, Bharara S, King Pal. Human cytomegalovirus infection and expression in human mnant glioma. Cancer Res 2002;62:3347–50.

olberg-Poley AM. Functional roles of immediate early proteins encby the human cytomegalovirus UL36-38, UL115-119 TRS1/IRS1US3 loci. Intervirology 1996;39:350–60.

olberg-Poley AM, Patel MB, Erezo DPP, Slater JE. Humcytomegalovirus immediate early regulatory proteins traffic throthe secretory apparatus and to mitochondria. J Gen2000;81:1779–89.

etrick B, Rhame J, Wang Y, Nagineni CN, Hooks JJ. Cytomegaloreplication in human retinal pigment epithelial cells altered exp

484 R. Adair et al. / Journal of Clinical Virology 35 (2006) 478–484

sion of viral early proteins. Invest Ophthalmol Vis Sci 1996;37:814–25.

Epstein SE, Speir E, Zhou YF, Guetta E, Leon M, Finkel T. The role ofinfection in restenosis and atherosclerosis: focus on cytomegalovirus.Lancet 1996;348(Suppl 1):s13–7.

Goldmacher VS, Bartle LM, Skaletskaya A, Dionne CA, Kedersha NL,Vater CA, et al. A cytomegalovirus-encoded mitochondria-localizedinhibitor of apoptosis structurally unrelated to Bcl-2. Proc Natl AcadSci USA 1999;96:12536–41.

Hayajneh WA, Colberg-Poley AM, Skaletskaya A, Bartle LM, LesperanceMM, Contopoulos-Ioannidis DG, et al. The sequence and antiapop-totic functional domains of the human cytomegalovirus UL37 exon1 immediate early protein are conserved in multiple primary strains.Virology 2001;279:233–40.

Hendrix RM, Wagenaar M, Slobbe RL, Bruggeman CA. Widespread pres-ence of cytomegalovirus DNA in tissues of healthy trauma victims. JClin Pathol 1997;50:59–63.

Holland GN. AIDS: retinal and choroidal infections. In: Lewis H,Ryan SJ, editors. Medical and surgical retina: advances, contro-versies, and management. St. Louis, MO: Mosby; 1994. p. 415–33.

Kouzarides T, Bankier AT, Satchwell SC, Preddy E, Barrell BG. Animmediate early gene of human cytomegalovirus encodes a potentialmembrane glycoprotein. Virology 1988;165:151–64.

Mavinakere MS, Colberg-Poley AM. Dual targeting of the humancytomegalovirus UL37 exon 1 protein during permissive infection.J Gen Virol 2004a;85:323–9.

Mavinakere MS, Colberg-Poley AM. Internal cleavage of the humancytomegalovirus UL37 immediate-early glycoprotein and divergenttrafficking of its proteolytic fragments. J Gen Virol 2004b;85:1989–94.

M ofene

irol

M . In:hia:

P ields1.

R dif-cells.

R teinsirus-n. J

S ed

S RA,s forovirusimes

Scholz M, Doerr HW, Cinatl J. Human cytomegalovirus retinitis:pathogenicity, immune evasion and persistence. Trends Microbiol2003;11:171–8.

Scholz M, Vogel JU, Hover G, Prosch S, Kotchetkov R, Cinatl J, et al.Thrombin induces Sp1-mediated antiviral effects in cytomegalovirus-infected human retinal pigment epithelial cells. Med MicrobiolImmunol (Berl) 2004;193:195–203.

Sinclair JH, Baillie J, Bryant LA, Taylor-Wiedeman JA, Sissons JG.Repression of human cytomegalovirus major immediate early geneexpression in a monocytic cell line. J Gen Virol 1992;73:433–5.

Sinzger C, Grefte A, Plachter B, Gouw AS, The TH, Jahn G. Fibroblasts,epithelial cells, endothelial cells and smooth muscle cells are majortargets of human cytomegalovirus infection in lung and gastrointesti-nal tissues. J Gen Virol 1995;76:741–50.

Spector SA, Hsia K, Crager M, Pilcher M, Cabral S, StempienMJ. Cytomegalovirus (CMV) DNA load is an independent pre-dictor of CMV disease and survival in advanced AIDS. J Virol1999;73:7027–30.

Speir E, Modali R, Huang ES, Leon MB, Shawl F, Finkel T, et al. Poten-tial role of human cytomegalovirus and p53 interaction in coronaryrestenosis. Science 1994;265:391–4.

Stinski MF, Thomsen DR, Stenberg RM, Goldstein LC. Organization andexpression of the immediate early genes of human cytomegalovirus.J Virol 1983;46:1–14.

Strausberg RL, Feingold EA, Grouse LH, Derge JG, Klausner RD, CollinsFS, et al. Generation and initial analysis of more than 15,000 full-length human and mouse cDNA sequences. Proc Natl Acad Sci USA2002;99:16899–903.

Su Y, Adair R, Davis CN, DiFronzo NL, Colberg-Poley AM. Conver-gence of RNAcis elements and cellular polyadenylation factors in theregulation of human cytomegalovirus UL37 exon 1 unspliced RNA

S oleyini-Virol

T virusogy

T andn ofuring

T anen-

X etainsfor

ML-

Z T,andMed

cCormick AL, Smith VL, Chow D, Mocarski ES. Disruptionmitochondrial networks by the human cytomegalovirus UL37 gproduct viral mitochondrion-localized inhibitor of apoptosis. J V2003;77:631–41.

ocarski ES, Courcelle CT. Cytomegaloviruses and their replicationKnipe DM, Howley PM, editors. Fields virology. 4th ed. PhiladelpLippincott, Williams & Wilkins; 2001. p. 2629–73.

ass RF. Cytomegaloviruses. In: Knipe DM, Howley PM, editors. Fvirology. 4th ed. Philadelphia: Lippincott, Williams & Wilkins; 200p. 2675–705.

ambhatla L, Chiu CP, Glickman RD, Rowe-Rendleman C. In vitroferentiation capacity of telomerase immortalized human RPEInvest Ophthalmol Vis Sci 2002;43:1622–30.

eboredo M, Greaves RF, Hahn G. Human cytomegalovirus proencoded by UL37 exon 1 protect infected fibroblasts against vinduced apoptosis and are required for efficient virus replicatioGen Virol 2004;85:3555–67.

almon-Ceron D. Cytomegalovirus infection: the point in 2001. HIV M2001;2:255–9.

anchez V, McElroy AK, Yen J, Tamrakar S, Clark CL, Schwartzet al. Cyclin-dependent kinase activity is required at early timeaccurate processing and accumulation of the human cytomegalUL122-123 and UL37 immediate-early transcripts and at later tfor virus production. J Virol 2004;78:11219–32.

production. J Virol 2003a;77:12729–41.u Y, Testaverde JR, Davis CN, Hayajneh WA, Adair R, Colberg-P

AM. Human cytomegalovirus UL37 immediate early target mgene RNAs are accurately spliced and polyadenylated. J Gen2003b;84:29–39.

enney DJ, Colberg-Poley AM. Expression of the human cytomegaloUL36-38 immediate early region during permissive infection. Virol1991a;182:199–210.

enney DJ, Colberg-Poley AM. Human cytomegalovirus UL36-38US3 immediate-early genes: temporally regulated expressionuclear, cytoplasmic, and polysome-associated transcripts dinfection. J Virol 1991b;65:6724–34.

erhune SS, Schroer J, Shenk T. RNAs are packaged into humcytomegalovirus virions in proportion to their intracellular conctration. J Virol 2004;78:10390–8.

u Y, Ahn JH, Cheng M, apRhys CM, Chiou CJ, Zong J,al. Proteasome-independent disruption of PML oncogenic dom(PODs), but not covalent modification by SUMO-1, is requiredhuman cytomegalovirus immediate-early protein IE1 to inhibit Pmediated transcriptional repression. J Virol 2001;75:10683–95.

hou YF, Leon MB, Waclawiw MA, Popma JJ, Yu ZX, Finkelet al. Association between prior cytomegalovirus infectionthe risk of restenosis after coronary atherectomy. N Engl J1996;335:624–30.