1

Differential Properties of Cytomegalovirus pUL97 Kinase Isoforms Affect Viral 1 Replication and Maribavir Susceptibility 2

3 Rike Webel,a Morgan Hakki,b Mark Prichard,c William D. Rawlinson,d 4

Manfred Marschall,a Sunwen Choub, e * 5 6

Institute for Clinical and Molecular Virology, University of Erlangen-Nuremberg, 7 Erlangen, Germanya; Division of Infectious Diseases, Oregon Health and Science 8 University, Portland, Oregon, USAb; Department of Pediatrics, University of Alabama at 9 Birmingham, Birmingham, AL, USAc; Microbiology Department SEALS, Virology 10 Division, Prince of Wales Hospital, Randwick, NSW, Australiad; Department of Veterans 11 Affairs Medical Center, Portland, Oregon, USAe 12

13 R.W. and M.H. contributed equally to this article. 14 * Address correspondence to Sunwen Chou, chous@ohsu.edu 15 Running title: Functional properties of HCMV pUL97 isoforms 16 17

JVI Accepts, published online ahead of print on 12 February 2014J. Virol. doi:10.1128/JVI.00192-14Copyright © 2014, American Society for Microbiology. All Rights Reserved.

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

2

ABSTRACT 18 The human cytomegalovirus (HCMV)-encoded kinase pUL97 is required for 19

efficient viral replication. Previous studies described two isoforms of pUL97, full length 20 (M1) and a smaller isoform likely resulting from translation initiation at codon 74 (M74). 21 Here, we report the detection of a third pUL97 isoform during viral infection resulting 22 from translation initiation at codon 157 (isoform M157). The consistent expression of 23 isoform M157 as a minor component of pUL97 during infection with clinical and 24 laboratory-adapted HCMV strains was suppressed when codon 157 was mutagenized. 25 Viral mutants expressing specific isoforms were generated to compare their growth and 26 drug susceptibility phenotypes, as well as pUL97 intracellular localization patterns and 27 kinase activity. The exclusive expression of isoform M157 resulted in substantially 28 reduced viral growth and resistance to the pUL97 inhibitor maribavir while retaining 29 susceptibility to ganciclovir. Confocal imaging demonstrated reduced nuclear import of 30 amino-terminal deletion isoforms compared to isoform M1. Isoform M157 showed 31 reduced efficiency of various substrate protein interactions and autophosphorylation 32 whereas Rb phosphorylation was preserved. These results reveal differential properties of 33 pUL97 isoforms that affect viral replication, with implications for the antiviral efficacy of 34 maribavir. 35 36

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

3

37 IMPORTANCE 38

The HCMV UL97 kinase performs important functions in viral replication that 39 are targeted by the antiviral drug maribavir. Here, we describe a naturally occurring short 40 isoform of the kinase that when expressed by itself in a recombinant virus results in 41 altered intracellular localization, impaired growth and high level resistance to maribavir 42 when compared to the predominant full length counterpart. This is another factor to 43 consider in explaining why maribavir appears to have variable antiviral activity in cell 44 culture and in vivo. 45 46

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

4

INTRODUCTION 47 The product of the HCMV UL97 gene (pUL97) is a serine/threonine protein 48

kinase that phosphorylates itself and multiple viral and host proteins (1). It is expressed 49 early in infection (2, 3), localizes predominantly to the nucleus using amino-terminal 50 nuclear localization signals (4, 5), and is also observed in the cytoplasm later in the 51 infection cycle (2, 5, 6). It is also present in the HCMV virion (7). Genetic deletion of 52 UL97 results in a severe (10-1000 fold) replication defect in vitro (8-10), as does 53 mutation of the K355 lysine residue critical for kinase activity (6). UL97 is therefore an 54 attractive target for antiviral drug development (1, 11-14) as an alternative to the UL54 55 DNA polymerase target of all currently licensed HCMV antivirals including ganciclovir 56 (GCV). 57

Maribavir (MBV) is a benzimidazole riboside pUL97 inhibitor that inhibits 58 HCMV replication potently (15) but variably depending on cell culture conditions (16). 59 While MBV showed promise in early clinical trials, two recently-published phase III 60 trials demonstrated that MBV was no more effective than placebo at preventing HCMV 61 viremia after allogeneic stem cell transplantation or liver transplantation (17, 18). Since 62 resistance to MBV was not observed in these trials (19), potential explanations for their 63 outcome may include an insufficient drug dosage or understanding of how in vivo host 64 cell conditions affect the function of pUL97 and consequences of its inhibition. 65

It is difficult to determine which of the multiple functions of pUL97 are most 66 important in facilitating viral replication. Early in the infection cycle it may activate the 67 major immediate-early promoter (20), phosphorylate retinoblastoma (Rb) protein (21) to 68 influence the cell cycle as a CDK ortholog (22-24), interact with cyclins (25), and 69

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

5

phosphorylate the viral polymerase accessory protein pUL44 (26, 27) and the nuclear 70 mRNA export factor pUL69 (28). Later in the infection cycle, pUL97 may facilitate 71 nuclear egress by phosphorylating the viral tegument protein pp65 (29), nuclear lamins 72 (30, 31), and may also function through kinase-independent mechanisms (32). 73

The UL97 open reading frame (ORF) contains several potential alternative in-74 frame ATG translation initiation sites downstream of the first ATG codon (M1) and 75 upstream of K355, located at codons 38, 74, 111, 157, and 330. Previous studies 76 documented two pUL97 isoforms, and results using pUL97 expressed by transient 77 transfection implicated M74 as the start site for the second isoform (4, 27, 33). Both 78 isoforms exhibited autophosphorylation activity and were incorporated in HCMV virions 79 (4) . Moreover, two nuclear localization sequences, NLS1 and NLS2, were identified in 80 the N-terminus of pUL97 indicating a differential, isoform-specific efficiency of nuclear 81 import (5). 82

The present objective was to investigate the use of alternate translation initiation 83 sites in the UL97 ORF in the context of HCMV replication. We found that during normal 84 HCMV infection, a majority component of full-length pUL97 is consistently 85 accompanied by a shorter isoform translated beginning with M157. Recombinant 86 phenotyping of HCMV strains showed significant differences in substrate 87 phosphorylation, subcellular localization, and growth and drug susceptibility profiles of 88 the smaller isoform that may affect the antiviral efficacy of MBV. 89 90

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

6

MATERIALS AND METHODS 91 Cells. Human foreskin fibroblast (HFF) and human embryonic lung (HEL) fibroblast cell 92 lines were maintained at 37°C in a 5% CO2 atmosphere in Eagle’s minimal essential 93 media supplemented with 7.5 - 10% fetal bovine serum (FBS, Hyclone) during the 94 growth phase and 3% FBS after reaching confluence. HFFs and HELs were maintained 95 by passaging (app. 1:3) once or twice per week. Human embryonic kidney 293T cells 96 were cultivated in Dulbecco's modified Eagle media supplemented with 10% FBS. For 97 HCMV infections, cells were selectively seeded in 6-well, 12-well, or 24-well plates and 98 infected at MOIs between 0.01 to 3 under standard conditions (12, 34) . 99 Viral clones and strains. The recombinant viruses UL97-HA ( expressing wild type 100 UL97 tagged with HA epitope), UL97(Mx4)-F (expressing Flag tagged UL97 containing 101 the mutations M38A/M74L/M111L/M157A) and UL97(M1L)-HA were constructed 102 using traceless bacterial artificial chromosome (BAC) mutagenesis (35). To generate the 103 internal kanamycin resistance cassette bounded by a repeated UL97 segment, the 104 oligonucleotides 5-aphAI/UL97-PstI and 3-aphAI-PstI (Table 1) were used to amplify a 105 fragment containing PstI restriction sites (unique in UL97) using the vector pEP-S/aphAI 106 (kindly provided by B. K. Tischer, Free Univ. Berlin, Germany (35)). This PCR product 107 was then inserted into the cloning constructs for UL97-HA, UL97(Mx4)-F and 108 UL97(M1L)-HA, and tagged UL97 mutant versions were generated. Recombination was 109 achieved by the use of these transfer vectors, inserted at homologous sites into the 110 previously constructed AD169-derived BAC pHB15-ΔUL97 using primers 5-ΔUL97 and 111 3-ΔUL97 (Table 1). After cleavage at a specific I-SceI site, the kanamycin resistance 112 marker was removed by a second recombination step. The resulting recombinant BACs 113

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

7

were verified via restriction fragment length polymorphism, PCR amplification of the 114 inserted gene region and nucleotide sequencing. For virus reconstitution, BAC DNA was 115 transfected into HFF cells, before infectious virus released to the supernatants could be 116 passaged and grown to viral stocks. 117

Additional strains were generated for secreted alkaline phosphatase (SEAP) 118 reporter-based analysis of replication and drug susceptibility. These are based on the 119 BAC BA1 that was cloned from the AD169-derived wild type HCMV strain T2211 and 120 contains a SEAP reporter gene under control of an ectopic HCMV major immediate early 121 (MIE) promoter adjacent to gene US3 (36, 37). The BAC BA9 was derived from BA1 by 122 replacing codons 134-654 of UL97 with a galK selection marker (36) and transfected into 123 HFF to yield HCMV strain T3110 (ΔUL97). Strains expressing kinase-inactive pUL97 124 (K355M and ΔK355) have been previously described (38). Additional HCMV 125 recombinant strains harboring the desired UL97 mutation(s) (Table 2) were created by 126 recombination of BAC BA9 with a transfer vector containing the UL97 mutation 127 (generated by PCR mutagenesis) in addition to an Frt-flanked kanamycin resistance 128 marker, followed by removal of the kanamycin marker with Flp recombinase, as 129 previously described (36). BAC DNA was then transfected into HFF cultures to 130 reconstitute live HCMV. The absence of parental BAC galK sequences in the 131 recombinant virus was confirmed by PCR, and the entire UL97 sequence of each 132 recombinant virus was assessed by standard automated fluorescent dideoxy sequencing to 133 ensure the presence of the desired mutation(s) and the absence of unintended ones. 134

Historical clinical HCMV isolates C1-C3 and C076 (39) were obtained from 135 transplant recipients not receiving anti-CMV therapy. A series of additional clinical 136

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

8

isolates collected between 1990 and 2012 was obtained from the diagnostic repositories 137 for clinical specimens of the virological institutes at Erlangen (11 isolates) and Sydney 138 (31 isolates). The UL97 gene region of primary or low- passage virus isolates was 139 amplified using a nested-PCR reaction. First round amplification was performed with 140 primers UL97.140304T and UL97.142644B, and second round with primers 141 UL97.140341T and UL97.142418B (Table 1). All PCR products were analyzed by 142 nucleotide sequencing. 143 Comparative growth of UL97 mutant strains. SEAP activity in the supernatants of 24-144 well cultures of HEL or HFF cells was assayed using a chemiluminescent substrate as 145 relative light units (RLU) at days 1, 4, 5, 6, 7, and 8 after the inoculation of a set of 146 comparison strains at a multiplicity of infection (MOI) of 0.01 to 0.02 (36, 37). The mean 147 SEAP activities and standard deviations from 4 replicate wells were used to construct the 148 growth curves. Similar SEAP activities at 24h were used to calibrate an equivalent MOI. 149

Viral genome equivalents were measured by quantitative real-time PCR (qPCR, 150 7500 RealTime PCR System and SDS Software, Applied Biosystems) of supernatants 151 collected from HFFs infected with recombinant HCMVs (MOI 0.1) at days 0, 2, 4, 6, 8, 152 10 and 12 post-infection. Parental AD169 was taken as a control for normal viral 153 replication in vitro. Following digestion of supernatant samples with proteinase K, an IE1 154 gene region within exon 4 was amplified using primers CMV 3' and CMV 5' together 155 with the TaqMan probe CMV MIE FAM/TAMRA (Table 1). Control plasmid containing 156 IE1 cDNA served as an internal amplification standard. All samples were analyzed in 157 triplicate and the standard deviation was calculated. 158

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

9

Phenotypic assay of antiviral susceptibility. SEAP yield reduction assays for GCV and 159 MBV susceptibility of HCMV strains were performed as previously described (37, 40) . 160 Briefly, cell-free virus stock was inoculated onto 6-day-old HEL (MBV) or HFF (GCV) 161 monolayers in 24-well culture plates at an MOI of 0.01 to 0.02 and cultured under a range 162 of drug concentrations (5 per assay) and a no-drug control. The drug concentration 163 required to reduce the supernatant SEAP activity by 50% (EC50) was determined by 164 curve fitting. A minimum of seven assay replicates performed on at least four separate 165 dates were used to calculate the mean EC50 and standard deviation for each strain. 166 Susceptible and resistant control strains were included for quality control and to enable 167 comparison with previous publications. 168 Western blot analysis. HFF or HEL cells infected with HCMV recombinant strains were 169 washed in ice-cold phosphate-buffered saline and lysed using 2% sodium dodecyl sulfate 170 (SDS), standard lysis buffer (50mM Tris/HCl (pH 8.0), 150mM NaCl, 5mM EDTA, 171 0.5% NP-40, 1mM PMSF, 2 μg/ml aprotinin, leupeptin and pepstatin) or NSRB (62.5 172 mM Tris-HCl pH 6.8, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 2% SDS, 173 10% Glycerol, 5% beta-mercaptoethanol). Lysate was separated on SDS-polyacrylamide 174 gels followed by transfer to a polyvinylidene difluoride (PVDF) membrane (Millipore). 175 Immunoblot analysis was performed for pUL97 using a carboxy-terminal mouse 176 monoclonal antibody (Alabama) (6), a central region monoclonal antibody (Rijeka) (41) 177 or rabbit polyclonal antibodies (2, 42). Antibodies against pp65 (Virusys), UL44 178 (Virusys), actin (Sigma AC-15 and 2066), and Rb (4H1; Cell Signaling Technology) 179 were used according to manufacturers’ instructions. Horseradish peroxidase-conjugated 180 goat anti-mouse (Santa Cruz Biotechnology; sc-2005) or anti-rabbit (Santa Cruz 181

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

10

Biotechnology; sc-2004) IgG secondary antibody, and luminol were used according to 182 manufacturers’ instructions (Sigma-Aldrich; A8511). Digital images were acquired using 183 LAS-1000 (Fujifilm Europe GmbH), ChemiDoc XRS (BioRad), or 4000 MM (Kodak) 184 luminescence imagers. 185 Confocal laser-scanning microscopy. Immunofluorescence analysis and confocal 186 microscopy was performed as described elsewhere (43). In brief, HCMV-infected cells 187 (MOI 0.1) were grown on coverslips, before fixation and immunostaining was performed 188 by the use of 4% paraformaldehyde and a polyclonal antibody against pUL97 (42). 189 Immunoprecipitation and in vitro kinase assays. pUL97 immunoprecipitation and in 190 vitro kinase assays using anti-pUL97 rabbit polyclonal antibody was performed as 191 previously described (29). 192

RESULTS 193 Isoforms of pUL97 result from translation initiation at M1 and M157 during 194 HCMV infection. The presence of two pUL97 isoforms during infection of human HFFs 195 with HCMV strain AD169 has been described (4, 5, 27, 33). Potential translation start 196 sites for UL97 coding sequence include M1 for full-length pUL97 and others at M38, 197 M74, M111, M157, and M330. Prior work indicated that M74 may serve as an initiation 198 site when pUL97 is expressed by transient transfection (4). We therefore focused on 199 M74, but also evaluated a role for M111 or M157 during viral infection. Immunoblot 200 analysis was performed on protein extracts from HFFs infected with parental HCMV 201 (AD169) or recombinant strains containing stop codons at positions 74, 111 or 157. 202 Using a rabbit polyclonal antibody (Ulm) to pUL97 (Fig. 1A), infection with wt HCMV 203 resulted in a prominent pUL97 isoform migrating at ~ 100kDa (band 1) and a shorter, 204

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

11

less prominent isoform at ~80 kD (band 3). Mutation of M1 to a leucine residue (M1L) 205 resulted in loss of band 1, confirming loss of the full-length isoform and appearance of an 206 intermediate-size isoform (band 2) likely resulting from translation initiation at M74. The 207 isoform represented by band 3 was also observed during infection with strains containing 208 stop codons at 74 and 111 but not 157, indicating that it may result from initiation at 209 codon M157. 210 To confirm these results, a mouse monoclonal antibody was used to detect pUL97 during 211 infection with the strains described above. Also included were strains lacking intact 212 pUL97, including one with stop codons replacing both M74 and M330, and another with 213 codons 134-654 deleted (HCMVΔUL97). Two isoforms were observed during infection 214 with BAC-cloned strain AD169 derivatives (wt HCMV, Figure 1B and C) and an 215 intermediate-size isoform (M74) was again detected during infection with HCMV 216 UL97(M1L) (band 2). The smallest isoform (band 3) was not detected in strains with stop 217 codons at 157 or 330. No signal was observed during infection with strains lacking intact 218 pUL97, demonstrating the specificity of the pUL97 monoclonal antibody. These results 219 are consistent with isoforms of pUL97 translated from M1 and M157 during HCMV 220 infection. 221 In order to verify M157 as a site of isoform initiation, and to control for the possibility 222 that this isoform is simply a degradation product of full-length pUL97, a recombinant 223 virus with M157 replaced by leucine (HCMV UL97(M157L)) was generated. The 224 recombinant viruses HCMV UL97(M74L) and HCMV UL97(M74L/M157L) were made 225 to address the previously mentioned role for M74 in pUL97 isoform translation (4) and 226 for the finding that translation appears to initiate at M74 when M1 is mutated (Figures 1A 227

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

12

and B). To assess whether smaller isoforms of pUL97 resulted from autophosphorylation, 228 a recombinant virus lacking the critical kinase residue (HCMV UL97(ΔK355)) was 229 utilized. Finally, a virus in which amino acids 1-156 were deleted was made to provide a 230 size marker for the pUL97 polypeptide comprising amino acids 157-707. Two isoforms 231 were clearly detected during infection with wt HCMV (Figure 1C, arrows), with the 232 smaller isoform again being present in a minority amount relative to full-length pUL97 233 and co-migrating with pUL97(157-707). Mutation of M157, but not M74, markedly 234 reduced the intensity of the smaller isoform despite equal amounts of full-length pUL97, 235 leaving a faint background signal that was observed in this blot but not consistently in 236 others (Fig 2A). The presence of the smaller isoform was not affected by the absence of 237 pUL97 kinase activity (HCMV UL97(ΔK355)). 238 Taken together, these results indicate that multiple isoforms of pUL97 are made 239 during viral infection, with full-length pUL97 (isoform M1) predominant over the 240 isoform originating at M157 (isoform M157). 241 Translation of the M157 isoform occurs during infection with clinical isolates of 242 HCMV. To determine whether translation of the M157 isoform occurred during infection 243 with clinical isolates, lysate was prepared from HFFs infected with one of four clinical 244 isolates of HCMV (CS1-3, C076) and analyzed by immunoblot using anti-pUL97 mAb. 245 Lysate from cells infected with HCMV UL97(M74L/M157L) and HCMV UL97(157-246 707) served as negative and positive controls, respectively, for detection of isoform 247 M157. Both isoforms M1 and M157 were observed during infection with CS1, 2, and 3 248 (Figure 2A, arrows). For CS076, the M157 isoform band was barely visible probably 249 because of the reduced total amount of pUL97 in this sample. 250

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

13

Additional clinical isolates, R3 and R4 (12, 44) were analyzed for the presence of 251 pUL97 isoforms using anti-pUL97 polyclonal antibody (Ulm). Three polypeptides were 252 detected, consistent with the presence of isoforms (Fig. 2B), but in different relative 253 amounts than findings using the anti-pUL97 monoclonal antibody (Fig 1B and C and Fig. 254 2A). To determine whether this difference in detection depended on the pUL97 antibody 255 reagents, the same lysates from cells infected with strain AD169 or the clinical isolates 256 R3 and R4 were subjected to immunoblot analysis using anti-pUL97 polyclonal antibody 257 (Ulm) as in Fig 2B, and two anti-pUL97 monoclonal antibodies – “Alabama” (as used in 258 Figs 1B and C, 2A), and “Rijeka”. The relative detection of isoforms varied among the 259 strains and pUL97 antibody reagents (Figure 2C). In contrast to findings with the 260 polyclonal antibody, both monoclonal antibodies showed the smaller isoforms to be 261 always present in lesser quantities compared to the full-length isoform. 262 Thus, the translation of multiple pUL97 isoforms in HCMV-infected cultured cells 263 could be demonstrated for several strains of HCMV and may therefore also have 264 relevance during infection in vivo. Indeed, the amino terminal portion of UL97 is known 265 to be well conserved among >60 published clinical isolate sequences, including codons 266 M1, M38, M74, M111 and M157 (45, 46). Sequence analysis of 5 laboratory HCMV 267 strains and an additional 42 clinical isolates also showed complete conservation of these 268 ATG codons (Table 2). Therefore it is anticipated that all clinical isolates have the 269 potential to express multiple pUL97 isoforms. 270 Isoform M1 but not M157 is sufficient for normal viral replication. The role of each 271 pUL97 isoform in HCMV replication was assessed by multi-step growth curve analysis 272 of HCMV recombinants. Replication of HCMV UL97(M1L) and HCMV UL97(M74L) 273

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

14

in HEL cells was similar to that of a simultaneously inoculated control, parental AD169-274 derived (wt) strain (Figure 3A). Additional multi-step growth curve analysis after 275 infection of HEL cells indicated that while HCMV UL97(M74L/M157L) grew with wild-276 type kinetics, HCMV UL97(157-707) was growth attenuated but not to the degree of a 277 recombinant strain in which the kinase activity of pUL97 was genetically abrogated by 278 deletion of the critical catalytic residue K355 (HCMV UL97(ΔK355) (Figure 3B). 279 HCMV UL97(157-707) also demonstrated growth impairment in HFFs (Figure 3C), 280 although to a lesser extent than in HELs. An alternate growth comparison of this mutant 281 in HFF culture against wild type AD169 by DNA qPCR showed a modest reduction of 282 viral DNA yield over time, while strains producing the other isoforms appeared to grow 283 as well as AD169 (Fig. 3D). Therefore, it appears that no single isoform, including full-284 length pUL97 (isoform M1), is required for efficient viral replication. 285 An HCMV mutant lacking isoform M1 results in altered subcellular localization of 286 pUL97. NLS1 is located between codons 6-35 and is therefore missing in smaller 287 isoforms, whereas NLS2, located between codons 190-213, is retained (4, 5). To assess 288 for changes in localization of pUL97 resulting from loss of NLS1 during infection, 289 confocal microscopy experiments were performed after infection of HFFs (MOI=0.01) 290 with recombinant viruses representing full length (UL97(Mx4)-F) or short isoforms 291 (UL97(M1L)-HA) (Fig. 4) . Virus expressing only full length pUL97 (UL97(Mx4)-F, 292 panels 10-18) showed dominant nuclear localization as expected, with concentration of 293 signals at the periphery of viral replication compartment (panel 17b). In contrast, during 294 infection with HCMV UL97(M1L)-HA, pUL97 demonstrated reduced nuclear 295 localization, and the pUL97 that was imported into the nucleus was concentrated in 296

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

15

replication compartments (panels 25-27). Thus, the absence of NLS1 impairs but does not 297 abrogate nuclear import of pUL97 during infection. 298 Isoform 3 exhibits a reduced level of substrate protein interaction. pUL97 has been 299 shown to directly interact with several proteins, including the HCMV-encoded UL83 300 (pp65) and UL44 proteins (27, 42). To determine the roles of the various isoforms in 301 protein-protein interaction, pUL97 from cells infected with wt HCMV, HCMV 302 UL97(Mx4)-F, HCMV UL97(M1L), and HCMV UL97(157-707) was 303 immunoprecipitated using rabbit polyclonal anti-pUL97 antibody. The relative amounts 304 of input pUL97 isoforms and pUL44 were comparable across infections, whereas input 305 pp65 was slightly more variable but was similar in HCMV AD169 and UL97(157-707) 306 (Figure 5, bottom panels). The pUL97 polypeptides immunoprecipitated to a similar 307 degree with the exception of the isoforms made during infection with HCMV 308 UL97(M1L)-HA, which were present in greater abundance (Figure 5, top panel). Both 309 pp65 and pUL44 co-immunoprecipated with pUL97 during infection with recombinant 310 viruses expressing all pUL97 isoforms (AD169 and UL97-HA) as well as solely isoform 311 M1 (UL97(Mx4)-F). The efficiency of pUL44 co-immunoprecipitation was slightly 312 reduced based on the amount relative to immunoprecipitated pUL97 during infection 313 with HCMV UL97(M1L)-HA, whereas pp65 co-immunoprecipitated proportionately. 314 The interaction of pUL97 isoform M157 (HCMV UL97(157-707)) with both pUL44 and 315 pp65 appeared less robust, with both being detectable only after prolonged exposure. 316 Similar results pertaining to the interaction of pUL97(157-707) with pp65 were observed 317 after immunoprecipitation of pUL97 from HCMV-infected cells using the mouse 318 monoclonal antibody (data not shown). These results indicate that the amino-terminus of 319

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

16

pUL97 is necessary for normal pUL97 protein-protein interactions during infection. 320 Phosphorylation of pUL97 substrates by the M157 isoform. pUL97 321 autophosphorylation is often used as a surrogate assay of biologically relevant kinase 322 activity in its wild type and mutant forms (1, 41, 47). To determine isoform-specific 323 properties of autophosphorylation, an in vitro kinase assay was performed using HCMV-324 infected HFFs (MOI of 0.1). Five dpi, cell lysates were prepared in CoIP buffer (Fig. 6A) 325 or RIPA buffer (Fig. 6B), and pUL97 was immunoprecipitated and analyzed in an in vitro 326 kinase assay (29). Analysis of a recombinant virus expressing only isoform M1 327 (UL97(Mx4)-F) demonstrated autophosphorylation comparable to wild-type controls 328 (AD169, UL97-HA) (Figure 6A, top panel). During infection with HCMV UL97(M1L)-329 HA, autophosphorylation was readily detectable, whereas when expressed in isolation, 330 isoform M157 exhibited relatively reduced autophosphorylation. The arrow indicates a 331 phosphorylated product that likely represents pp65. While phosphorylation of pp65 by 332 pUL97(157-707) appeared to be relatively reduced, the interaction of pUL97(157-707) 333 with pp65 was, as previously observed (Figure 5), less robust, thus pp65 phosphorylation 334 appears to be limited at the level of substrate protein binding. To exclude the presence of 335 additional co-immunoprecipitating kinases, we routinely performed control reactions in 336 which we applied the pUL97 inhibitor Gö6976 so that pUL97 specificity of the 337 phosphorylation signals could be assured (data not shown). The reduced phosphorylation 338 capacity of isoform M157 was also reflected by a limited degree of standard substrate 339 phosphorylation, i.e. exogenously added recombinant human histones 1-4 (Fig. 6B). 340 Based on these results, it appears that isoform M157 is capable of autophosphorylation 341 but to a lesser degree than isoform M1, and substrate phosphorylation may be affected as 342

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

17

well, possibly due to loss of direct interaction. 343 Phosphorylation of Rb during HCMV infection is dependent on pUL97 kinase 344

activity (21). pUL97 contains three putative Rb binding domains, of which one located 345 between codons 149-153 is absent from isoform M157 (1). To assess whether infection 346 with HCMV UL97(157-707) resulted in pUL97-kinase dependent Rb phosphorylation, 347 HFFs that were cultured in conditions of serum starvation were mock-infected or infected 348 with wt HCMV or HCMV UL97(157-707) at an MOI of ~3 under various concentrations 349 of MBV and lysate was prepared at 24 hpi. The phosphorylation status of Rb was 350 determined by assessing the electrophoretic mobility shift observed after immunoblot 351 analysis of total Rb (21). Infection with HCMV UL97(157-707) resulted in Rb 352 phosphorylation comparable to infection with AD169-derived parental wt HCMV, and 353 this was effectively inhibited by MBV (Figure 6C). Therefore, normal phosphorylation of 354 Rb occurs during infection in a pUL97-kinase dependent manner when isoform M157 355 alone is expressed. 356 Isolated expression of isoform M157 is associated with MBV resistance. In order to 357 assess whether pUL97 isoforms have a role in determining susceptibility to the antiviral 358 agents GCV and MBV, EC50 values for each drug against recombinant strains were 359 determined (Table 3). While HCMV UL97(M157L) was susceptible to both GCV and 360 MBV, HCMV UL97(157-707) was significantly resistant to MBV but susceptible to 361 GCV. Therefore, isoform M157 is dispensable for the antiviral efficacy of GCV and 362 MBV, but the exclusive expression of this pUL97 isoform results in resistance to MBV 363 and not to GCV. 364 365

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

18

DISCUSSION 366 Expanding on previous reports of pUL97 isoforms, we used specific viral mutants 367

to demonstrate that an additional isoform results from translation initiation at codon 368 M157 in HCMV-infected cell cultures. Phenotypic characterization of a HCMV 369 recombinant strain expressing only isoform M157 indicated that the amino terminus is 370 required for the normal nuclear localization of pUL97, efficient HCMV replication, and 371 maribavir susceptibility. 372

The alternate ATG start sites for translation of pUL97 are all highly conserved 373 among clinical isolates on the basis of published (45, 46) and new data (Table 2), 374 supporting the presumption that isoform M157 of pUL97 is routinely expressed during 375 HCMV infection in vivo in similar relative quantities and kinetics as observed in the 376 laboratory and clinical strains examined here. 377

Not all publications have detected multiple pUL97 isoforms, instead reporting just 378 one isoform that corresponds in size to full-length pUL97 (48). Indeed, at early times 379 during infection, or when DNA synthesis was inhibited with foscarnet (data not shown), 380 this isoform was not readily observed, likely due to the low overall amount of total 381 pUL97. However, as total pUL97 accumulated at late times during infection, we 382 consistently detected the isoform translated from M157 and confirmed its identity by 383 mutagenizing that residue, although it represented a minor component relative to full-384 length pUL97. 385

The isoform of pUL97 described in transient transfection systems as resulting 386 from translation initiation at M74 (4, 5, 33) was not readily detected in virus-infected 387 cells in this study using a monoclonal antibody (Fig. 1 and 2). These discordant results 388

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

19

may reflect the differences in the expression of pUL97 during infection versus transient 389 transfection, as well as the performance characteristics of different anti-pUL97 antibodies 390 (Fig 2C). 391

The exclusive expression of isoform M157 of pUL97 by an in-frame deletion 392 mutant resulted in a growth defect that was more apparent in HEL fibroblasts than HFFs, 393 consistent with the known greater sensitivity of HCMV growth in HEL cells to loss of 394 pUL97 kinase activity (16). However, in contrast to UL97 knockout mutants, infection 395 with HCMV UL97(157-707) did not show the typical UL97-defective cytopathic 396 appearance (49, 50). In addition, isoform M157 maintained sufficient kinase activity to 397 result in full ganciclovir susceptibility, which requires initial pUL97-mediated 398 phosphorylation, (51, 52). Rb phosphorylation was also maintained, despite loss of a Rb-399 binding LxCxE motif in pUL97 centered on C151 (53). Isoform M157 also exhibited 400 partial autophosphorylation activity, adding to prior data showing that a wide range of 401 pUL97 residues affect autophosphorylation in vitro (41, 54). In this case, decreased 402 autophosphorylation may result from the absence of known serine and threonine targets 403 that are clustered in the missing amino terminal part of pUL97, such as S3 (54). 404 However, it is unclear if autophosphorylation is required for the normal function of 405 pUL97. 406

The partial growth defect of HCMV UL97(157-707) may result from loss of 407 pUL97 function due to the observed disruption of normal intracellular localization and/or 408 protein-protein interactions. Isoform M157 exhibited reduced interaction with at least two 409 known pUL97 interacting partners, pp65 and pUL44. While the pp65 interaction domain 410 has not yet been defined, the pUL44 interaction domain has been mapped to codons 366-411

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

20

459 (27). Since this domain is present in isoform M157, the impaired interaction with 412 pUL44 most likely results from abnormal subcellular localization during infection 413 (Figure 4). As pp65 is nonessential for efficient HCMV replication in HFFs (55), it is 414 unlikely that loss of phosphorylation of this substrate explains the observed growth 415 defect. It remains to be determined whether phosphorylation of pUL44 (26) and other 416 pUL97 substrates such as pUL69 (28) are affected. 417

MBV resistance has emerged in HCMV isolates from treated individuals, 418 involving mutations mapping to the ATP binding region of pUL97 (codons 409 and 411) 419 (56, 57), consistent with the action of MBV as an ATP-competitive kinase inhibitor (58). 420 Additional mutations in this vicinity that confer MBV resistance have been observed in 421 cell culture (15, 19, 38, 59). All of these mutations confer significant MBV resistance 422 while retaining near-normal viral growth and kinase activity sufficient for GCV 423 phosphorylation and susceptibility. Deletion of the amino terminal residues of pUL97 424 would not appear to involve the ATP-competitive MBV binding site. The suppression of 425 Rb phosphorylation by MBV during infection with HCMV UL97(157-707) at 426 concentrations well below the EC50 for this virus, and comparable to the EC50 for a 427 strain expressing full length pUL97 (Figures 6C), provides indirect evidence that MBV 428 inhibits the kinase activity of isoform M157. Rather, it is likely that isoform M157 lacks 429 some critical function of full-length pUL97 that results in MBV resistance because a 430 greater fraction of the remaining viral growth would then be independent of the pUL97 431 kinase activity that is inhibited by MBV. This isoform could have nonkinase biological 432 activities as previously described for pUL97 (32). 433

The finding of multiple pUL97 isoforms with differential kinase and growth 434

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

21

properties adds to the biological complexity of this antiviral drug target. A variable 435 degree of HCMV growth that can occur without normal pUL97 kinase activity may have 436 influenced the results of the two Phase III MBV prophylaxis trials. As additional Phase II 437 trials of MBV are ongoing (ClinicalTrials.gov Identifier NCT01611974 and EudraCT 438 number 2010-024247-32), a more detailed understanding of the pUL97 isoform-specific 439 interactions, substrates, and functional overlap with cellular enzymes and pathways is 440 needed to guide the development of antiviral compounds targeting pUL97. 441

442 ACKNOWLEDGMENTS 443

We gratefully acknowledge Donald Coen (Harvard Medical School, Boston, 444 USA), Tihana Lenac and Stipan Jonjić (Univ. Rijeka, Croatia), Detlef Michel (Univ. 445 Ulm, Germany), Bodo Plachter (Univ. Mainz, Germany), B. K. Tischer (Free Univ. 446 Berlin, Germany) and William J. Britt (Univ. of Alabama, Birmingham, USA) for the gift 447 of reagents. This study was supported by the Deutsche Forschungsgemeinschaft SFB796 448 (C3), DFG MA 1289/6-1, DAAD/Go8 grant 54390135, U.S. National Institutes of Health 449 grant R01-AI39938, and Department of Veterans Affairs grant I01-BX00925. We thank 450 Ronald Ercolani, Gail Marousek, Coyne Drummond and Ben Houser for technical 451 assistance. Further contributions by Klaus Korn, Thomas Stamminger, Sabrina Wagner, 452 members of the M.M. laboratory (Inst. Virology, Univ. Erlangen-Nuremberg) and Zin 453 Naing (Virology Div., POW Hospital, Randwick, Australia) are gratefully acknowledged. 454 We are grateful to Hanife Bahsi for excellent technical assistance. 455 456

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

22

REFERENCES 457 1. Prichard MN. 2009. Function of human cytomegalovirus UL97 kinase in viral 458

infection and its inhibition by maribavir. Rev Med Virol 19:215-229. 459 2. Michel D, Pavic I, Zimmermann A, Haupt E, Wunderlich K, Heuschmid M, 460

Mertens T. 1996. The UL97 gene product of human cytomegalovirus is an early-461 late protein with a nuclear localization but is not a nucleoside kinase. J Virol 462 70:6340-6346. 463

3. Wing BA, Huang ES. 1995. Analysis and mapping of a family of 3'-coterminal 464 transcripts containing coding sequences for human cytomegalovirus open reading 465 frames UL93 through UL99. J Virol 69:1521-1531. 466

4. Webel R, Milbradt J, Auerochs S, Schregel V, Held C, Nobauer K, Razzazi-467 Fazeli E, Jardin C, Wittenberg T, Sticht H, Marschall M. 2011. Two isoforms 468 of the protein kinase pUL97 of human cytomegalovirus are differentially 469 regulated in their nuclear translocation. J Gen Virol 92:638-649. 470

5. Webel R, Solbak SM, Held C, Milbradt J, Gross A, Eichler J, Wittenberg T, 471 Jardin C, Sticht H, Fossen T, Marschall M. 2012. Nuclear import of isoforms 472 of the cytomegalovirus kinase pUL97 is mediated by differential activity of NLS1 473 and NLS2 both acting through classical importin-alpha binding. J Gen Virol 474 93:1756-1768. 475

6. Gill RB, James SH, Prichard MN. 2012. Human cytomegalovirus UL97 kinase 476 alters the accumulation of CDK1. J Gen Virol 93:1743-1755. 477

7. Varnum SM, Streblow DN, Monroe ME, Smith P, Auberry KJ, Pasa-Tolic L, 478 Wang D, Camp DG, 2nd, Rodland K, Wiley S, Britt W, Shenk T, Smith RD, 479

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

23

Nelson JA. 2004. Identification of proteins in human cytomegalovirus (HCMV) 480 particles: the HCMV proteome. J Virol 78:10960-10966. 481

8. Krosky PM, Baek MC, Coen DM. 2003. The human cytomegalovirus UL97 482 protein kinase, an antiviral drug target, is required at the stage of nuclear egress. J 483 Virol 77:905-914. 484

9. Prichard MN, Gao N, Jairath S, Mulamba G, Krosky P, Coen DM, Parker 485 BO, Pari GS. 1999. A recombinant human cytomegalovirus with a large deletion 486 in UL97 has a severe replication deficiency. J Virol 73:5663-5670. 487

10. Wolf DG, Courcelle CT, Prichard MN, Mocarski ES. 2001. Distinct and 488 separate roles for herpesvirus-conserved UL97 kinase in cytomegalovirus DNA 489 synthesis and encapsidation. Proc Natl Acad Sci U S A 98:1895-1900. 490

11. Chou S. 2008. Cytomegalovirus UL97 mutations in the era of ganciclovir and 491 maribavir. Rev Med Virol 18:233-246. 492

12. Herget T, Freitag M, Morbitzer M, Kupfer R, Stamminger T, Marschall M. 493 2004. Novel chemical class of pUL97 protein kinase-specific inhibitors with 494 strong anticytomegaloviral activity. Antimicrob Agents Chemother 48:4154-495 4162. 496

13. Marschall M, Stamminger T. 2009. Molecular targets for antiviral therapy of 497 cytomegalovirus infections. Future Microbiol 4:731-742. 498

14. Schleiss M, Eickhoff J, Auerochs S, Leis M, Abele S, Rechter S, Choi Y, 499 Anderson J, Scott G, Rawlinson W, Michel D, Ensminger S, Klebl B, 500 Stamminger T, Marschall M. 2008. Protein kinase inhibitors of the quinazoline 501

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

24

class exert anti-cytomegaloviral activity in vitro and in vivo. Antiviral Res 79:49-502 61. 503

15. Biron KK, Harvey RJ, Chamberlain SC, Good SS, Smith AA, 3rd, Davis 504 MG, Talarico CL, Miller WH, Ferris R, Dornsife RE, Stanat SC, Drach JC, 505 Townsend LB, Koszalka GW. 2002. Potent and selective inhibition of human 506 cytomegalovirus replication by 1263W94, a benzimidazole L-riboside with a 507 unique mode of action. Antimicrob Agents Chemother 46:2365-2372. 508

16. Chou S, Van Wechel LC, Marousek GI. 2006. Effect of cell culture conditions 509 on the anticytomegalovirus activity of maribavir. Antimicrob Agents Chemother 510 50:2557-2559. 511

17. Marty FM, Ljungman P, Papanicolaou GA, Winston DJ, Chemaly RF, 512 Strasfeld L, Young JA, Rodriguez T, Maertens J, Schmitt M, Einsele H, 513 Ferrant A, Lipton JH, Villano SA, Chen H, Boeckh M. 2011. Maribavir 514 prophylaxis for prevention of cytomegalovirus disease in recipients of allogeneic 515 stem-cell transplants: a phase 3, double-blind, placebo-controlled, randomised 516 trial. Lancet Infect Dis 11:284-292. 517

18. Winston DJ, Saliba F, Blumberg E, Abouljoud M, Garcia-Diaz JB, Goss JA, 518 Clough L, Avery R, Limaye AP, Ericzon BG, Navasa M, Troisi RI, Chen H, 519 Villano SA, Uknis ME. 2012. Efficacy and safety of maribavir dosed at 100 mg 520 orally twice daily for the prevention of cytomegalovirus disease in liver transplant 521 recipients: A randomized, double-blind, multicenter controlled trial. Am J 522 Transplant 12:3021-3030. 523

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

25

19. Chou S, Hakki M, Villano S. 2012. Effects on maribavir susceptibility of 524 cytomegalovirus UL97 kinase ATP binding region mutations detected after drug 525 exposure in vitro and in vivo. Antiviral Res 95:88-92. 526

20. Bigley TM, Reitsma JM, Mirza SP, Terhune SS. 2013. Human 527 cytomegalovirus pUL97 regulates the viral MIE promoter by phosphorylation-528 mediated disruption of HDAC1 binding. J Virol 87:7393-7408. 529

21. Hume AJ, Finkel JS, Kamil JP, Coen DM, Culbertson MR, Kalejta RF. 2008. 530 Phosphorylation of retinoblastoma protein by viral protein with cyclin-dependent 531 kinase function. Science 320:797-799. 532

22. Hertel L, Chou S, Mocarski ES. 2007. Viral and cell cycle-regulated kinases in 533 cytomegalovirus-induced pseudomitosis and replication. PLoS Pathogens 3:e6. 534

23. Kuny CV, Chinchilla K, Culbertson MR, Kalejta RF. 2010. Cyclin-dependent 535 kinase-like function is shared by the beta- and gamma- subset of the conserved 536 herpesvirus protein kinases. PLoS Pathog 6:e1001092. 537

24. Marschall M, Feichtinger S, Milbradt J. 2011. Regulatory roles of protein 538 kinases in cytomegalovirus replication. Adv Virus Res 80:69-101. 539

25. Graf L, Webel R, Wagner S, Hamilton ST, Rawlinson WD, Sticht H, 540 Marschall M. 2013. The Cyclin-Dependent Kinase Ortholog pUL97 of Human 541 Cytomegalovirus Interacts with Cyclins. Viruses 5:3213-3230. 542

26. Krosky PM, Baek MC, Jahng WJ, Barrera I, Harvey RJ, Biron KK, Coen 543 DM, Sethna PB. 2003. The human cytomegalovirus UL44 protein is a substrate 544 for the UL97 protein kinase. J Virol 77:7720-7727. 545

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

26

27. Marschall M, Freitag M, Suchy P, Romaker D, Kupfer R, Hanke M, 546 Stamminger T. 2003. The protein kinase pUL97 of human cytomegalovirus 547 interacts with and phosphorylates the DNA polymerase processivity factor 548 pUL44. Virology 311:60-71. 549

28. Thomas M, Rechter S, Milbradt J, Auerochs S, Muller R, Stamminger T, 550 Marschall M. 2009. Cytomegaloviral protein kinase pUL97 interacts with the 551 nuclear mRNA export factor pUL69 to modulate its intranuclear localization and 552 activity. J Gen Virol 90:567-578. 553

29. Becke S, Fabre-Mersseman V, Aue S, Auerochs S, Sedmak T, Wolfrum U, 554 Strand D, Marschall M, Plachter B, Reyda S. 2010. Modification of the major 555 tegument protein pp65 of human cytomegalovirus inhibits virus growth and leads 556 to the enhancement of a protein complex with pUL69 and pUL97 in infected 557 cells. J Gen Virol 91:2531-2541. 558

30. Hamirally S, Kamil JP, Ndassa-Colday YM, Lin AJ, Jahng WJ, Baek MC, 559 Noton S, Silva LA, Simpson-Holley M, Knipe DM, Golan DE, Marto JA, 560 Coen DM. 2009. Viral mimicry of Cdc2/cyclin-dependent kinase 1 mediates 561 disruption of nuclear lamina during human cytomegalovirus nuclear egress. PLoS 562 Pathog 5:e1000275. 563

31. Marschall M, Marzi A, aus dem Siepen P, Jochmann R, Kalmer M, 564 Auerochs S, Lischka P, Leis M, Stamminger T. 2005. Cellular p32 recruits 565 cytomegalovirus kinase pUL97 to redistribute the nuclear lamina. The Journal of 566 biological chemistry 280:33357-33367. 567

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

27

32. Goldberg MD, Honigman A, Weinstein J, Chou S, Taraboulos A, Rouvinski 568 A, Shinder V, Wolf DG. 2011. Human cytomegalovirus UL97 kinase and 569 nonkinase functions mediate viral cytoplasmic secondary envelopment. J Virol 570 85:3375-3384. 571

33. Schregel V, Auerochs S, Jochmann R, Maurer K, Stamminger T, Marschall 572 M. 2007. Mapping of a self-interaction domain of the cytomegalovirus protein 573 kinase pUL97. J Gen Virol 88:395-404. 574

34. Marschall M, Freitag M, Weiler S, Sorg G, Stamminger T. 2000. 575 Recombinant green fluorescent protein-expressing human cytomegalovirus as a 576 tool for screening antiviral agents. Antimicrob Agents Chemother 44:1588-1597. 577

35. Tischer BK, von Einem J, Kaufer B, Osterrieder N. 2006. Two-step red-578 mediated recombination for versatile high-efficiency markerless DNA 579 manipulation in Escherichia coli. Biotechniques 40:191-197. 580

36. Chou S. 2010. Recombinant phenotyping of cytomegalovirus UL97 kinase 581 sequence variants for ganciclovir resistance. Antimicrob Agents Chemother 582 54:2371-2378. 583

37. Chou S, Van Wechel LC, Lichy HM, Marousek GI. 2005. Phenotyping of 584 cytomegalovirus drug resistance mutations by using recombinant viruses 585 incorporating a reporter gene. Antimicrob Agents Chemother 49:2710-2715. 586

38. Chou S, Ercolani RJ, Marousek G, Bowlin TL. 2013. Cytomegalovirus UL97 587 kinase catalytic domain mutations that confer multi-drug resistance. Antimicrob 588 Agents Chemother 57:3375-3379. 589

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

28

39. Chou S, Erice A, Jordan MC, Vercellotti GM, Michels KR, Talarico CL, 590 Stanat SC, Biron KK. 1995. Analysis of the UL97 phosphotransferase coding 591 sequence in clinical cytomegalovirus isolates and identification of mutations 592 conferring ganciclovir resistance. J Infect Dis 171:576-583. 593

40. Chou S. 2009. Diverse cytomegalovirus UL27 mutations adapt to loss of viral 594 UL97 kinase activity under maribavir. Antimicrob Agents Chemother 53:81-85. 595

41. Michel D, Schaarschmidt P, Wunderlich K, Heuschmid M, Simoncini L, 596 Muhlberger D, Zimmermann A, Pavic I, Mertens T. 1998. Functional regions 597 of the human cytomegalovirus protein pUL97 involved in nuclear localization and 598 phosphorylation of ganciclovir and pUL97 itself. J Gen Virol 79:2105-2112. 599

42. Kamil JP, Coen DM. 2007. Human cytomegalovirus protein kinase UL97 forms 600 a complex with the tegument phosphoprotein pp65. J Virol 81:10659-10668. 601

43. Milbradt J, Auerochs S, Sevvana M, Muller YA, Sticht H, Marschall M. 602 2012. Specific residues of a conserved domain in the N terminus of the human 603 cytomegalovirus pUL50 protein determine its intranuclear interaction with 604 pUL53. J Biol Chem 287:24004-24016. 605

44. Efferth T, Marschall M, Wang X, Huong SM, Hauber I, Olbrich A, 606 Kronschnabl M, Stamminger T, Huang ES. 2002. Antiviral activity of 607 artesunate towards wild-type, recombinant, and ganciclovir-resistant human 608 cytomegaloviruses. J Mol Med 80:233-242. 609

45. Boutolleau D, Burrel S, Agut H. 2011. Genotypic characterization of human 610 cytomegalovirus UL97 phosphotransferase natural polymorphism in the era of 611 ganciclovir and maribavir. Antiviral Res 91:32-35. 612

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

29

46. Lurain NS, Weinberg A, Crumpacker CS, Chou S. 2001. Sequencing of 613 cytomegalovirus UL97 gene for genotypic antiviral resistance testing. Antimicrob 614 Agents Chemother 45:2775-2780. 615

47. Marschall M, Stein-Gerlach M, Freitag M, Kupfer R, van Den Bogaard M, 616 Stamminger T. 2001. Inhibitors of human cytomegalovirus replication drastically 617 reduce the activity of the viral protein kinase pUL97. J Gen Virol 82:1439-1450. 618

48. Gentry BG, Vollmer LE, Hall ED, Borysko KZ, Zemlicka J, Kamil JP, 619 Drach JC. 2013. Resistance of human cytomegalovirus to cyclopropavir maps to 620 a base pair deletion in the open reading frame of UL97. Antimicrob Agents 621 Chemother 57:4343-4348. 622

49. Chou S, Marousek GI. 2008. Accelerated evolution of maribavir resistance in a 623 cytomegalovirus exonuclease domain II mutant. J Virol 82:246-253. 624

50. Prichard MN, Britt WJ, Daily SL, Hartline CB, Kern ER. 2005. Human 625 cytomegalovirus UL97 kinase is required for the normal intranuclear distribution 626 of pp65 and virion morphogenesis. J Virol 79:15494-15502. 627

51. Littler E, Stuart AD, Chee MS. 1992. Human cytomegalovirus UL97 open 628 reading frame encodes a protein that phosphorylates the antiviral nucleoside 629 analogue ganciclovir. Nature 358:160-162. 630

52. Sullivan V, Talarico CL, Stanat SC, Davis M, Coen DM, Biron KK. 1992. A 631 protein kinase homologue controls phosphorylation of ganciclovir in human 632 cytomegalovirus-infected cells. Nature 358:162-164. 633

53. Prichard MN, Sztul E, Daily SL, Perry AL, Frederick SL, Gill RB, Hartline 634 CB, Streblow DN, Varnum SM, Smith RD, Kern ER. 2008. Human 635

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

30

cytomegalovirus UL97 kinase activity is required for the hyperphosphorylation of 636 retinoblastoma protein and inhibits the formation of nuclear aggresomes. J Virol 637 82:5054-5067. 638

54. Baek MC, Krosky PM, Coen DM. 2002. Relationship between 639 autophosphorylation and phosphorylation of exogenous substrates by the human 640 cytomegalovirus UL97 protein kinase. J Virol 76:11943-11952. 641

55. Schmolke S, Kern HF, Drescher P, Jahn G, Plachter B. 1995. The dominant 642 phosphoprotein pp65 (UL83) of human cytomegalovirus is dispensable for 643 growth in cell culture. J Virol 69:5959-5968. 644

56. Strasfeld L, Lee I, Villano S, Chou S. 2010. Virologic characterization of multi-645 drug-resistant cytomegalovirus infection in two transplant recipients treated with 646 maribavir. J Infect Dis 202:104-108. 647

57. Schubert A, Ehlert K, Schuler-Luettmann S, Gentner E, Mertens T, Michel 648 D. 2013. Fast selection of maribavir resistant cytomegalovirus in a bone marrow 649 transplant recipient. BMC Infect Dis 13:330. 650

58. Shannon-Lowe CD, Emery VC. 2010. The effects of maribavir on the 651 autophosphorylation of ganciclovir resistant mutants of the cytomegalovirus 652 UL97 protein. Herpesviridae 1:4. 653

59. Chou S, Van Wechel LC, Marousek GI. 2007. Cytomegalovirus UL97 kinase 654 mutations that confer maribavir resistance. J Infect Dis 196:91-94. 655

656 657 658

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

31

FIGURE LEGENDS 659 660 FIG 1. pUL97 isoform translation during HCMV infection. 661 HFFs were infected with wt AD169-derived HCMV or recombinant strains in which the 662 indicated ATG codons were replaced with ones encoding leucine (L) or a stop codon (*). 663 pUL97 isoforms were detected by immunoblots of infected cell lysates using (A) a rabbit 664 polyclonal antibody (pAb-Ulm) or (B and C) a pUL97 mouse monoclonal antibody 665 (mAb) (Alabama). Arrows 1, 2 and 3 indicate bands corresponding to isoforms M1, M74 666 and M157. 667 668 FIG 2. pUL97 isoform translation during infection with clinical isolates. 669 Lysates of cells infected with clinical isolates CS1-3, C076, R3 and R4 were analyzed by 670 immunoblotting with indicated antibodies. (A) Sizes of isoform M1 and M157 bands are 671 confirmed by control UL97 mutant laboratory strains. (B) Comparison of lysates 672 harvested at various time points after infection at MOI=0.5. (C) Comparison of isoform 673 detection in the same infected cell lysates by different antibody reagents. Arrows 1, 2 and 674 3 indicate the locations of isoforms M1, M74 and M157 as in Figure 1. 675 676 FIG 3. Growth in cell culture of HCMV strains containing mutations in pUL97 677 isoform translation start sites. 678 (A and B) HEL cells or (C) HFFs were infected with wt or recombinant HCMV strains 679 with indicated genotypes. All infections were at calibrated MOI of 0.01-0.02. Supernatant 680 SEAP activity at days 1 and 4-7 or 8 post-infection was used to assess relative growth. 681

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

32

Data points are the mean and standard deviation of four simultaneous replicates and are 682 representative of at least three independent experiments. (D) HFFs were infected at MOI 683 0.1 with the indicated recombinant HCMV strains. Virus growth was measured by 684 quantitative PCR of IE-1 sequences in culture supernatant samples at the time-points 685 indicated. 686 687 FIG 4. Intracellular localization of pUL97 isoforms. HFFs were infected with the 688 recombinant HCMV strains UL97(Mx4)-F and UL97(M1L)-HA, followed by fixation 689 and immunostaining at the time points indicated. Specific localization phenotypes of 690 pUL97 were illustrated in an enlarged presentation by the use of image insets taken from 691 mostly nuclear (inset nuc) or cytoplasmic areas (inset cyt). 692 693 FIG 5. Interaction of pUL97 isoforms with pUL44 and pp65. HFFs were infected with 694 HFFs were infected with AD169 or recombinant HCMVs expressing individual pUL97 695 isoforms. Using anti-pUL97 polyclonal antiserum (pAb-Boston), pUL97 was 696 precipitated from cell lysates followed by immunoblot analysis for pUL97 (pAb-Ulm), 697 pUL44, and pp65. Expression controls consisted of immunoblots of cell lysates before 698 immunoprecipitation. 699 700 FIG 6. Phosphorylation of pUL97 substrates in vitro and during viral infection. 701 (A) An in vitro kinase assay on proteins immunoprecipitated from infected cell lysates by 702 pUL97 antibody. Autophosphorylation of pUL97 (arrows) and phosphorylation of co-703 immunoprecipitated pp65 was assessed by autoradiography (upper panel). Expression 704

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

33

and precipitation control immunoblot panels show pUL97 and pp65 before and after 705 immunoprecipitation. 706 (B) In vitro phosphorylation of human histone mixture H1–H4 (Roche) by pUL97 707 immunoprecipitated from HCMV-infected cells, as shown by autoradiography. 708 (C) HFFs were cultured in media containing 0.1% FBS for 24 hours prior to infection 709 with HCMV strains (MOI of 1-3) expressing wt pUL97 or pUL97(157-707) in the 710 presence or absence of MBV at the indicated concentrations. After infection, cells were 711 maintained under serum starvation until protein lysates were obtained at 24 hpi. Rb 712 phosphorylation was assessed by differential mobility on SDS-PAGE. 713

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

Table 1. Oligonucleotide sequences

Oligonucleotide name Sequence (5'-3')CMV 3' GAGCAGACTCTCAGAGGATCGG

CMV 5' AAGCGGCCTCTGATAACCAAG

CMV MIE FAM/TAMRA CATGCAGATCTCCTCAATGCGGCG

5-aphAI/UL97-PstI TAGCTGCAGAAGCTGCTCATCTGCGACCCGCACGCGCGTTTCCCCGTAGCCGGCCTACGTAGGGATAACAGGGTAATCGATTT

3-aphAI-PstI TAGCTGCAGGCCAGTGTTACAACCAATTAACC

5-ΔUL97 CGGTGTGGTAGCTAGTGCAGCCTTAGGAACAGGGAAGACTGTCGCCACTTAGGGATAACAGGGTAATCGATTT

3-ΔUL97 ACCTTCTCTGTTGCCTTTCCCCTCAGCAACCGTCACGTTCCGCGTCCCGGAGTGGCGACAGTCTTCCCTGTTCCTAAGGCTGCACTAGCTACCACACCGAGCCAGTGTTACAACCAATTAACC

UL97.140304T TTGGTGGACTCGGTTTCGGCG

UL97.142644B CCTTTCCCCTCAGCAACCGTC

UL97.140341T CTGGAGGTCGACGACGCCGTC

UL97.142418B GCGACACGAGGACATCTTGG

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

Table 2. Analysis of the N-terminal sequence of UL97 from clinical HCMV isolates

HCMV source Specimen isolation year

ATG (M1/38/74/111/157)

conservation

NLS1 / NLS2 conservation*

Laboratory strains AD169 - / Merlin - / TB40E - / Towne - Q19E /

FIX - /

Clinical isolates from Australia

27 isolates 1991-1996 / 4 isolates 1990-1994 Q19E /

Clinical isolates from Europe

10 isolates 2012 / 1 isolate 2012 Q19E /

*Amino acid substitution indicated where noted, otherwise no changes

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

Table 3. Genotypes and Phenotypes of Recombinant Viruses

UL97Genotype2

EC503 SD4 N5 Ratio6 EC503 SD4 N5 Ratio6

Controls

BA29 T3261 none 1.17 0.29 45 0.12 0.03 44

BA27 T3259 C592G 3.07 0.49 36 2.6

BA202 T3679 V353A 1.35 0.37 12 11.3

BA109 T3418 ΔK355 19.0 7.80 13 16 38.5 8.20 14 321

UL97 isoform mutants

BA235 T3820 M1L 1.26 0.34 11 1.1 0.08 0.02 12 0.7

BA234 T3819 M74L 1.17 0.29 16 1.0 0.07 0.03 8 0.6

BA270 T3947 M157L 1.42 0.43 13 1.2 0.10 0.02 9 0.8

BA299 T4045 157-707 0.88 0.23 8 0.8 18.8 6.35 8 157

BA296 T4047 M74L/M157L 1.62 0.52 9 1.4 0.12 0.04 10 1

Table footnotes bold Items associated with decreased drug susceptibility BAC HCMV strain AD169-derived bacterial artificial chromosome clone name

1 Recombinant HCMV strain derived from BAC clone (serial number)2 Amino acid substitution or codon range in BAC and recombinant virus3 Mean drug concentration (µM) required to reduce SEAP growth by 50% 4 Standard deviation of the EC50 values5 Number of assays (performed over at least 4 separate dates)6 Ratio of EC50 to baseline strain

MaribavirBAC Virus1

Ganciclovir

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

FIG 1. pUL97 isoform translation during HCMV infection. HFFs were infected with wt AD169-derived HCMV or recombinant strains in which the indicated ATG codons were replaced with ones encoding leucine (L) or a stop codon (*). pUL97 isoforms were detected by immunoblots of infected cell lysates using (A) a rabbit polyclonal antibody (pAb-Ulm) or (B and C) a pUL97 mouse monoclonal antibody (mAb) (Alabama). Arrows 1, 2 and 3 indicate bands corresponding to isoforms M1, M74 and M157.

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

FIG 2. pUL97 isoform translation during infection with clinical isolates.

Lysates of cells infected with clinical isolates CS1-3, C076, R3 and R4 were analyzed by immunoblotting with indicated antibodies. (A) Sizes of isoform M1 and M157 bands are confirmed by control UL97 mutant laboratory strains. (B) Comparison of lysates harvested at various time points after infection at MOI=0.5. (C) Comparison of isoform detection in the same infected cell lysates by different antibody reagents. Arrows 1, 2 and 3 indicate the locations of isoforms M1, M74 and M157 as in Figure 1.

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

FIG 3. Growth in cell culture of HCMV strains containing mutations in pUL97 isoform translation start sites. (A and B) HEL cells or (C) HFFs were infected with wt or recombinant HCMV strains with indicated genotypes. All infections were at calibrated MOI of 0.01-0.02. Supernatant SEAP activity at days 1 and 4-7 or 8 post-infection was used to assess relative growth. Data points are the mean and standard deviation of four simultaneous replicates and are representative of at least three independent experiments. (D) HFFs were infected at MOI 0.1 with the indicated recombinant HCMV strains. Virus growth was measured by quantitative PCR of IE-1 sequences in culture supernatant samples at the time-points indicated.

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

FIG 4. Intracellular localization of pUL97 isoforms.

HFFs were infected with the recombinant HCMV strains UL97(Mx4)-F and UL97(M1L)-HA, followed by fixation and immunostaining at the time points indicated. Specific localization phenotypes of pUL97 were illustrated in an enlarged presentation by the use of image insets taken from mostly nuclear (inset nuc) or cytoplasmic areas (inset cyt).

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

FIG 5. Interaction of pUL97 isoforms with pUL44 and pp65.

HFFs were infected with AD169 or recombinant HCMVs expressing individual pUL97 isoforms. Using anti-pUL97 polyclonal antiserum (pAb-Boston), pUL97 was precipitated from cell lysates followed by immunoblot analysis for pUL97 (pAb-Ulm), pUL44, and pp65. Expression controls consisted of immunoblots of cell lysates before immunoprecipitation.

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from

FIG 6. Phosphorylation of pUL97 substrates in vitro and during viral infection. (A) An in vitro kinase assay on proteins immunoprecipitated from infected cell lysates by pUL97 antibody. Autophosphorylation of pUL97 (arrows) and phosphorylation of co-immunoprecipitated pp65 was assessed by autoradiography (upper panel). Expression and precipitation control immunoblot panels show pUL97 and pp65 before and after immunoprecipitation. (B) In vitro phosphorylation of human histone mixture H1–H4 (Roche) by pUL97 immunoprecipitated from HCMV-infected cells, as shown by autoradiography. (C) HFFs were cultured in media containing 0.1% FBS for 24 hours prior to infection with HCMV strains (MOI of 1-3) expressing wt pUL97 or pUL97(157-707) in the presence or absence of MBV at the indicated concentrations. After infection, cells were maintained under serum starvation until protein lysates were obtained at 24 hpi. Rb phosphorylation was assessed by differential mobility on SDS-PAGE

on May 14, 2018 by guest

http://jvi.asm.org/

Dow

nloaded from