Proteomic and Functional Analyses of the Virion Transmembrane Proteome of Cyprinid Herpesvirus 3

ABSTRACT Virion transmembrane proteins (VTPs) mediate key functions in the herpesvirus infectious cycle. Cyprinid herpesvirus 3 (CyHV-3) is the archetype of fish alloherpesviruses. The present study was devoted to CyHV-3 VTPs. Using mass spectrometry approaches, we identified 16 VTPs of the CyHV-3 FL strain. Mutagenesis experiments demonstrated that eight of these proteins are essential for viral growth in vitro (open reading frame 32 [ORF32], ORF59, ORF81, ORF83, ORF99, ORF106, ORF115, and ORF131), and eight are nonessential (ORF25, ORF64, ORF65, ORF108, ORF132, ORF136, ORF148, and ORF149). Among the nonessential proteins, deletion of ORF25, ORF132, ORF136, ORF148, or ORF149 affects viral replication in vitro, and deletion of ORF25, ORF64, ORF108, ORF132, or ORF149 impacts plaque size. Lack of ORF148 or ORF25 causes attenuation in vivo to a minor or major extent, respectively. The safety and efficacy of a virus lacking ORF25 were compared to those of a previously described vaccine candidate deleted for ORF56 and ORF57 (Δ56-57). Using quantitative PCR, we demonstrated that the ORF25 deleted virus infects fish through skin infection and then spreads to internal organs as reported previously for the wild-type parental virus and the Δ56-57 virus. However, compared to the parental wild-type virus, the replication of the ORF25-deleted virus was reduced in intensity and duration to levels similar to those observed for the Δ56-57 virus. Vaccination of fish with a virus lacking ORF25 was safe but had low efficacy at the doses tested. This characterization of the virion transmembrane proteome of CyHV-3 provides a firm basis for further research on alloherpesvirus VTPs. IMPORTANCE Virion transmembrane proteins play key roles in the biology of herpesviruses. Cyprinid herpesvirus 3 (CyHV-3) is the archetype of fish alloherpesviruses and the causative agent of major economic losses in common and koi carp worldwide. In this study of the virion transmembrane proteome of CyHV-3, the major findings were: (i) the FL strain encodes 16 virion transmembrane proteins; (ii) eight of these proteins are essential for viral growth in vitro; (iii) seven of the nonessential proteins affect viral growth in vitro, and two affect virulence in vivo; and (iv) a mutant lacking ORF25 is highly attenuated but induces moderate immune protection. This study represents a major breakthrough in understanding the biology of CyHV-3 and will contribute to the development of prophylactic methods. It also provides a firm basis for the further research on alloherpesvirus virion transmembrane proteins.


C yprinid herpesvirus 3 (CyHV-3; species Cyprinid herpesvirus 3, genus Cyprinivirus,
family Alloherpesviridae, order Herpesvirales), also known as koi herpesvirus (KHV), is the etiological agent of a lethal disease in common and koi carp. The common carp (Cyprinus carpio) is one of the main fish grown for human consumption (1). Moreover, its colorful ornamental varieties (koi carp) represent one of the most lucrative markets for individual freshwater fish (2,3). As a result, CyHV-3 is considered to be the archetypal fish herpesvirus and is the subject of a growing number of studies (4).
The CyHV-3 genome is 295 kbp in size and thus the largest described among all herpesviruses (5). It encodes 155 potential protein-coding open reading frames (ORFs), a large number of which lack similarity to other herpesvirus genes (5)(6)(7). Nonetheless, CyHV-3 virions present the characteristic herpesvirus morphology (8), consisting of an icosahedral capsid containing the genome surrounded by an amorphous layer of proteins termed the tegument and enveloped in a lipid membrane bearing virion transmembrane proteins (VTPs) (8).
The protein composition of CyHV-3 virions has been investigated by mass spectrometrybased proteomic approaches for one European and two Chinese strains (9,10). These analyses led to the identification of 46 virion proteins, of which 34 were detected in all three strains. Among these 34 proteins, 13 were classified as VTPs (9,10). Since none of these proteins are related to viral proteins with known roles, it is not possible to predict their functions.
VTPs mediate key functions in the herpesvirus infectious cycle, such as binding virions to the cell and mediating entry, providing immune evasion mechanisms, facilitating virion morphogenesis, and promoting virion egress from the host cell. The expression of VTPs on the virion surface may also make them targets for neutralizing antibodies. As a result, it is important to study VTPs in order to understand the biology of infection and to aid in the design of candidate vaccines (attenuated, nonreplicative, or subunit). However, current knowledge of CyHV-3 VTPs is very limited. For example, it is not known whether any of the VTPs identified to date are required for viral replication in vitro.
Here, we performed proteomic and functional analyses of CyHV-3 VTPs. The proteome of CyHV-3 virions was revisited. Based on this new information and that published previously (9,10), 16 VTPs were identified and selected for further study. Using recombination technologies, VTPs that are essential to viral growth in vitro were identified. Furthermore, the effects of deleting nonessential VTPs on viral growth in vitro and virulence in vivo were investigated. The knowledge gained represents a major breakthrough in understanding the biology of CyHV-3 and provides a firm basis for further research on alloherpesvirus VTPs.

RESULTS AND DISCUSSION
Viral protein composition of CyHV-3 virions. We took advantage of the experience gained during our previous studies of the virion proteomes of herpesviruses (9,(11)(12)(13) to revisit the protein composition of extracellular CyHV-3 virions reported by us and others (9,10). We selected a strategy based on SDS-PAGE separation of proteins, followed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) of tryptic peptides, because this approach has been shown to enhance the recovery of peptides derived from proteins that are prone to aggregation or contain hydrophobic domains (9,(11)(12)(13). The data listed in Table 1 represent independent analyses of three independent preparations of CyHV-3 virions. A total of 43 viral proteins were identified (based on detection in at least two of the preparations; Table 1 and Fig. 1). This number is identical to that determined by Yi et al. (10). Comparison of the proteomes derived from both studies revealed 39 proteins in common, with four proteins in each study that were not identified in the other (Fig. 1).
Yi et al. (10) identified ORF11, ORF27, ORF91, and ORF116 as potential virion proteins, whereas we did not. The lack of detection of ORF27 in our study was anticipated because this coding region is disrupted in the FL strain, as has been reported for other CyHV-3 strains (5). ORF91 and ORF116 were not detected in any of   No transmembrane domain was detected by software prediction. This sample was classified as a putative type 1 transmembrane protein on the basis of significant similarity to previously studied viral membrane proteins.
Vancsok et al. Journal of Virology the preparations in the present study, and were identified by Yi et al. (10) for only one of the two Chinese strains, with detection being based on a single peptide that was supported by a low Mascot score (10). Similarly, ORF11 was detected on the basis of only one or two peptides (depending on the isolate) with relatively low Mascot scores. Among the four proteins that were detected in the present study and not by Yi et al. (10), two were detected in all three replicates (ORF4 and ORF64), and the two other were detected in only two replicates (ORF119 and ORF125). Finally, three proteins identified in our previous study were detected neither by Yi et al. (10) nor in the present study. Taking into account that these proteins were detected initially at relatively low abundances (9), it is likely that they represented contamination by nonstructural proteins.
CyHV-3 VTPs essential to viral growth in vitro. The roles of CyHV-3 VTPs in viral replication in vitro were investigated by using bacterial artificial chromosome (BAC) cloning and prokaryotic recombination technologies to generate viral mutants (Fig. 2). To facilitate the reconstitution of infectious viruses from recombinant plasmids, the BAC cassette was left in the viral genome leading to viruses expressing a truncated form of thymidine kinase (TK; encoded by ORF55) and enhanced green fluorescent protein (EGFP) (the BAC cassette is inserted at the 3= end of ORF55 and encodes an EGFP expression cassette). The effect of TK truncation on viral replication in vitro and virulence in vivo has been documented previously. TK truncation was shown to have no effect on viral growth in vitro and to reduce virulence slightly in vivo (14). For each ORF predicted to encode a VTP, a single-gene-deletion recombinant plasmid was produced by replacing the ORF by a galK expression cassette. The molecular structures of all recombinant BAC plasmids produced were confirmed by a combined SacI digestion and Southern blotting approach and by sequencing the regions used to target recombination (data not shown). We then investigated the ability of the recombinant plasmids to reconstitute infectious virus after transfection into permissive CCB cells (Fig.  3). Transfection was monitored by detecting EGFP-expressing cells (since the BAC lower left circle, analyses of the FL strain performed in a former study (9); and lower right circle, analyses of two Chinese strains (GZ10 and GZ11) (10). Numbers represent CyHV-3 ORFs. Asterisks indicate viral proteins that were detected in only one of the two Chinese isolates. Predicted transmembrane proteins are underlined.

CyHV-3 Virion Transmembrane Proteins
Journal of Virology cassette encodes an EGFP reporter gene) at 2 days postinfection (dpi). Examination of cell cultures at 4 dpi revealed the formation of viral plaques for the FL BAC plasmid (used as a positive control) and the BAC plasmids with ORF25, ORF64, ORF65, ORF108, ORF132, ORF136, ORF148, or ORF149 deleted, thereby demonstrating that these ORFs are nonessential for viral growth in vitro. Reconstituted viruses (i.e., viruses mutated in nonessential genes) were amplified, and their genomes were validated by full-length genome sequencing (data not shown) prior to further investigations (see below). Transfection of BAC plasmids with ORF32, ORF59, ORF81, ORF83, ORF99, ORF106, ORF115, or ORF131 deleted did not induce the formation of CyHV-3 plaques (Fig. 3), suggesting that these ORFs encode VTPs essential for viral growth in vitro. To test this hypothesis further, additional recombinant BAC plasmids were produced (Fig. 2). The molecular structures of all recombinant BAC plasmids described below were confirmed by a combined SacI digestion and Southern blotting approach and by sequencing the regions used to target recombination (data not shown). To exclude the possibility that the lack of infectivity of the deleted recombinant BAC plasmids resulted from unexpected mutations generated during BAC manipulation, plasmids encoding revertant wild-type (WT) VTP ORF were derived (FL BAC ORFX Rev1 plasmids, Fig. 2) from the deleted BAC plasmids. Transfection of each of the revertant plasmids into CCB cells led to viral growth (Fig. 4, second column). To exclude the possibility that the lack of FIG 2 Production of CyHV-3 recombinants to identify essential and nonessential VTPs. Deleted recombinant plasmids were produced for each ORF predicted to encode a VTP (FL BAC ORFX Del galK plasmids, with "X" standing for the number of the ORFs tested; these included ORF25, ORF32, ORF59, ORF64, ORF65, ORF81, ORF83, ORF99, ORF106, ORF108, ORF115, ORF131, ORF132, ORF136, ORF148, and ORF149). The effect of the deletion on the ability of the BAC plasmid to reconstitute infectious virus was tested by transfection into permissive CCB cells. Subsequently, for each gene identified as essential, two additional recombinant plasmids were produced to revert to wild-type ORFX sequence (FL BAC ORFX Rev1) or to insert a nonsense mutation (FL BAC ORFX NS). Plasmids were transfected into CCB cells to determine their ability to induce reconstitution of infectious virus. As an additional revertant control, FL BAC ORFX NS plasmids were cotransfected into CCB cells together with a fragment encoding the WT sequence of ORFX and flanking regions, in order to facilitate reversion to wild-type ORFX sequence by recombination in eukaryotic cells (FL BAC ORFX Rev2). To simplify the reading of the manuscript, some recombinants were given a short name (in red). The right part of the Fig. summarizes  infectivity observed for the deleted BAC plasmids resulted from a polar effect of the deletion on the expression of a nearby essential gene, a recombinant plasmid in which the ORF was disrupted by a nonsense (NS) mutation was derived for each putative essential VTP gene (FL BAC ORFX NS plasmids, Fig. 2). Transfection of these plasmids into CCB cells did not induce plaque formation (Fig. 4, third column). To exclude the possibility that the lack of infectivity of the NS recombinant BAC plasmids resulted from unexpected mutations generated during BAC manipulation at the stage of NS mutagenesis, the FL BAC ORFX NS plasmids were transfected into CCB cells, together with a WT DNA fragment containing the relevant ORF and flanking regions, in order to induce reversion to a WT sequence after homologous recombination in eukaryotic cells. All plasmids led to viral growth when cotransfected with WT DNA fragments (Fig. 4, fourth column). In summary, the results demonstrated that ORF32, ORF59, ORF81, ORF83, ORF99, ORF106, ORF115, and ORF131 encode VTPs essential to viral growth in vitro.
These experiments identified essential versus nonessential VTPs of the FL strain of CyHV-3, as summarized in Table 2. To test whether the findings extend across the species Cyprinid herpesvirus 3, the ORFs encoding VTPs were sequenced from nine CyHV-3 strains with different geographical origins and passage histories ( Table 2). All of the ORFs encoding essential VTPs were intact and exhibited a limited number of polymorphisms. In contrast, various ORFs encoding nonessential VTPs were truncated by disruptive mutations (highlighted in red in Table 2), the precise pattern depending on the strain.

Effects of deleting genes encoding nonessential VTPs on CyHV-3 growth in vitro.
Considerations of the CyHV-3 gene arrangement suggested that deletion of ORF25, ORF64, ORF65, ORF108, ORF132, ORF136, or ORF148 would be unlikely to affect the expression of neighboring genes. In contrast, deletion of ORF149 might affect the expression of ORF148, resulting in the phenotype of TK BAC Δ149 galK representing the absence of expression of both proteins. To test this hypothesis, TK BAC Δ148 galK and TK BAC Δ149 galK plaques were stained with monospecific polyclonal antibodies (pAbs) against ORF148 or ORF149. Staining of TK BAC Δ149 galK plaques with the anti-ORF148 antibodies demonstrated the expression of the ORF148 protein (data not shown), implying that TK BAC Δ149 galK is adequate for investigating the effects of deleting ORF149.
The effect of deleting genes encoding nonessential VTPs on viral growth in vitro was investigated by multistep growth assay (Fig. 5A) and plaque size assay (Fig. 5B). Deletion of ORF25, ORF132, ORF136, ORF148, or ORF149 impaired CyHV-3 replication in the multistep growth assays to various degrees (based on observation of a significant effect for at least two consecutive time points). The more pronounced effect was observed for deletion of ORF25, which was associated with a 10-fold reduction in viral titer. In contrast, deleting ORF64, ORF65, or ORF108 had no major effect on CyHV-3 growth. The plaque size assays (Fig. 5B), based on the results collected at 8 dpi, revealed that deleting ORF25, ORF108, ORF132, or ORF149 resulted in reduced plaque size, whereas deleting ORF64 led to an increased plaque size. Deleting ORF65, ORF136, or ORF148 did not significantly affect plaque size.
Effect of deleting genes encoding nonessential VTPs on CyHV-3 virulence in vivo. All recombinant viruses deleted for a nonessential ORF (TK BAC ΔX galK , where X is the number of the ORF) were tested in triplicate for their virulence in carp (Fig. 6). Fish infected with the TK BAC virus (used as a positive control) exhibited all the clinical signs associated with CyHV-3 disease, including apathy, folding of the dorsal fin, hyperemia, increased mucus secretion, skin lesions, suffocation, erratic swimming, and loss of equilibrium. At 30 dpi, the mean survival rate of TK BAC-infected fish was 41%. Fish inoculated with TK BAC Δ64 galK , TK BAC Δ65 galK , TK BAC Δ108 galK , TK BAC Δ132 galK , TK BAC Δ136 galK , or TK BAC Δ149 galK exhibited comparable effects, leading to similar survival rates at 30 dpi (ranging between 33 and 56%). Conversely, fish infected with TK BAC Δ148 galK exhibited mild disease and a higher survival rate (77%, P Ͻ 0.001). Fish inoculated with TK BAC Δ25 galK expressed no symptoms, and all survived the infection.
To ensure that each group of fish was infected with the correct virus and to exclude any possibility of viruses spreading among the tanks, PCR assays were performed on three randomly selected dead fish from each infected group (for TK BAC Δ25 galKinfected fish, three fish were selected randomly from the living fish at 30 dpi), one fish infected with the TK BAC virus, and one mock-infected fish selected randomly at the end of the experiment (Fig. 7). PCR performed with the ORF112intfw/ORF112intrev primer pair confirmed that samples from all but one group contained the CyHV-3 genome (the exception being the TK BAC Δ25 galK -infected group). PCR performed with primers specific for the deleted ORFs confirmed that each group had been infected  CyHV-3 Virion Transmembrane Proteins Journal of Virology with the correct virus and that cross-contamination among tanks had not occurred (Fig. 7).
Testing of an ORF25-deleted recombinant as an attenuated vaccine candidate against CyHV-3. The attenuation described above for TK BAC Δ25 galK , from which ORF25 was deleted and encoding a truncated form of TK, suggested that a virus deleted only for ORF25 would have potential as an attenuated recombinant vaccine against CyHV-3. The requisite recombinant (Δ25 galK , with ORF25 deleted and encoding a wild-type TK) was produced by following the procedure described in Fig. 8 and in Materials and Methods. The failure to detect virus by PCR in the gills of fish infected with the TK BAC Δ25 galK virus (Fig. 7) suggests that deletion of ORF25 could affect the ability of deleted virions to infect fish or to spread within their host. To test this hypothesis and to compare the tropism of a virus lacking ORF25 (Δ25 galK ) to the tropism of a previously described recombinant vaccine candidate (deleted for ORF56 and ORF57, Δ56-57), the following experiment was performed. Fish were infected with the FL BAC revertant virus (WT, wild-type for TK and for ORF25), FL BAC revertant ORF25 Del galK virus (Δ25 galK , wild-type for TK but with ORF25 deleted), or the FL BAC revertant ORF56-57 Del virus (Δ56-57, wild-type for TK but with ORF56-57 deleted) (Fig.  9). The viral charge was measured by qPCR over time in the skin (the portal of entry of CyHV-3) and the heart (selected as an internal organ). The results demonstrated that the ORF25 deleted virus infects fish through skin infection and then spreads to internal organs as reported previously for the WT parental virus and the Δ56-57 virus (15). However, compared to the parental WT virus, the replication of the ORF25 deleted virus was reduced in intensity and duration to levels similar to those observed for the Δ56-57 virus.
Next, the virulence of the Δ25 galK virus was compared to that of the WT strain and the previously described Δ56-57 attenuated virus (Fig. 10) (15). A high degree of attenuation of Δ25 galK was indicated by the finding that infected fish exhibited no mortality or symptoms (Fig. 10). The level of immune protection conferred by primary infection with Δ25 galK or Δ56-57 virus was tested by challenging infected fish with the M3 or FL strains of CyHV-3 (Fig. 10, middle and lower lines of the graph, respectively). The fish vaccinated with Δ25 galK exhibited a poor level of protection, independent of the dose used for primary infection and the challenge strain used, with the relative percentage survival at a maximum of 34 when the fish were vaccinated at a dose of 400 PFU/ml and challenged with the M3 strain. In comparison, vaccination with the Δ56-57 attenuated virus resulted in a relative percentage survival of 84 under the same conditions (Fig. 10).
The data above demonstrated that the VTP encoded by ORF25 is nonessential for viral replication in vitro but acts as a key virulence factor in vivo. ORF25 is the prototypical gene of the ORF25 family, which contains six paralogs (ORF25, ORF26, ORF27, ORF65, ORF148, and ORF149) encoding type I transmembrane proteins (5). Although ORF26 is a pseudogene, all other paralogs encode CyHV-3 VTPs. The FL strain used in this study as the parental strain encodes a disrupted ORF27, as has been reported for other strains of CyHV-3 (5). Individual deletions of each member of the ORF25 family in this ORF27-negative background revealed that none of them is essential for viral replication in vitro, although it is possible that this is the consequence of redundant functions expressed by the paralogs. The attenuation observed for a mutant lacking ORF25 suggested that this locus might serve as a target for generating FIG 8 Flow chart of the production of CyHV-3 recombinants deleted for ORF25 or ORF56-57. Recombinants were given a short name (in red) for use in the text. Viral genotypes for ORF55 (TK), ORF25, and ORF56-57 are given on the right. Del, deleted; WT, wild type.

FIG 9
Effect of ORF25 deletion on viral tropism according to qPCR analysis. At time zero, carp (average weight, 6.10 g Ϯ 1.39 g; average age, 6 months old) were infected for 2 h by immersion in water containing the virus indicated at 800 PFU/ml and then returned to larger tanks. At the indicated time points, fish were sampled and submitted to qPCR analysis. The data obtained for each fish (according to viral genotype and time postinfection) are represented by the same symbol to allow correlation of the data obtained for the two tested organs. The number of viral genome copies is expressed as the log 10 per 10 6 carp glucokinase gene copies. Individual values represent the mean of duplicate measurements. Mock-infected fish were used as a negative control and no viral genome copies were detected in these fish. The number of positive fish among six analyzed fish is represented by gray bars. Viral charges [viral genome copies/10 6 carp glucokinase gene copies (log 10 )] were compared using one-way ANOVA taking the viral genotype as a variable.

CyHV-3 Virion Transmembrane Proteins
Journal of Virology recombinant attenuated vaccines. However, there are two points of caution in relation to this hypothesis: the ORF25 deletion virus induced weak immune protection, and it is possible that reversion to virulence could occur by mutation of the paralogous genes. Concluding remarks. Herpesvirus VTPs play crucial roles in the viral life cycle. Although the functions of VTPs in members of the family Herpesviridae have been studied extensively, very little information is available on VTPs in members of the family Alloherpesviridae. The very distant evolutionary origin of these families has resulted in a lack of detectable genetic similarity between VTPs in the two groups. This situation makes it impossible to advance hypotheses on the roles of alloherpesvirus VTPs on the basis of the much more extensive knowledge available for other herpesviruses. The present study of the archetypal fish alloherpesvirus has enabled major steps to be taken toward characterizing alloherpesvirus VTPs: (i) the number of CyHV-3 VTPs was expanded from 14 to 16; (ii) these were characterized as consisting of eight essential and nine nonessential proteins; (iii) individual deletion of seven nonessential VTPs affected viral growth in vitro, and individual deletion of two affected virulence in vivo; and (iv) a mutant lacking ORF25 proved to be highly attenuated in vivo but induced a poor immune protection against a lethal challenge. This study provides a firm basis for the further study of alloherpesvirus VTPs.

Cells and viruses.
Cyprinus carpio brain (CCB) cells (16) were cultured as described previously (14). The CyHV-3 FL strain was isolated in Belgium from a fish that had died from CyHV-3 infection and had been used to produce the FL BAC plasmid (14). Other CyHV-3 strains from different geographical origins were also used: the M3 strain from Belgium, the U strain from the United States (5); the J strain from Japan (5), the E strain from England (kindly provided by K. Way), the T strain from Taiwan (17), the GZ11 strain from China (18), and the KHV-I and I strains from Israel (5). The attenuated Cavoy strain from Israel was produced by passaging in cell culture (19).
Production and purification of CyHV-3 virions. CCB cells were infected with the CyHV-3 FL strain at a multiplicity of infection (MOI) of 0.02 PFU/cell (20). The supernatant was harvested at the beginning of viral release, when cell lysis was minimal (4 dpi). Virions were purified from the cell supernatant and monitored for purity as described previously (9).

SDS-PAGE and LC-MS/MS proteomic analysis.
Virion proteins were separated by SDS-PAGE and detected by Coomassie blue staining. The entire lanes of the gel were cut into 30 equally sized slices, which were subjected to in-gel trypsin digestion (21). Peptides were separated by reversed-phase chromatography using a 40-min CAN gradient (4 to 45%) and analyzed by using an HCT ultra-ion trap (Bruker) in data-dependent AutoMS (2) mode with a scan range of 100 to 2,800 m/z and three averages, and five precursor ions at 300 to 1,500 m/z were selected from the MS scan. Precursors were actively excluded within a 0.5 min window, and all singly charged ions were excluded. The MS data were processed by using Mascot distiller, and all single-slice analyses were merged into a single data set. The data were searched for matches to CyHV-3 protein sequences (160 entries) by using an in-house Mascot 2.2 server (Matrix Science). The default search parameters used were as follows: enzyme ϭ trypsin, maximum missed cleavages ϭ 1, fixed modifications ϭ carbamidomethyl (C), variable modifications ϭ oxidation (M), peptide tolerance Ϯ 1.3 Da; MS/MS tolerance ϭ Ϯ 0.5 Da, and instrument ϭ ESI-TRAP. Only sequences identified by a Mascot score of Ͼ30 were considered, and single peptide identification was systematically evaluated manually. In addition, the emPAI (22) value was calculated to estimate relative protein abundance in the GC extract only.
Cloning and site-directed mutagenesis of CyHV-3 ORFs. DNA fragments containing CyHV-3 ORFs of interest plus at least 200 bp upstream and downstream were amplified by PCR using CyHV-3 FL BAC DNA as a template (see Table 3 for primer details). The amplification products were TA cloned into the pGEM-T-Easy vector (Promega), and the resulting clones (pGEM-T-ORFX WT, X standing for the number of the ORF tested) with the correct sequence were used as a recombination cassette or as a template for mutagenesis. For ORFs encoding VTPs, site-directed mutagenesis was used to create 1-or 2-bp substitutions changing a tyrosine or a tryptophan codon into a stop codon, resulting in pGEM-T-ORFX NS. Oligonucleotide primers were designed by using the QuikChange primer design tool (http://www .genomics.agilent.com/primerDesignProgram.jsp) ( Table 3).

Production of CyHV-3 FL BAC recombinant plasmids and CyHV-3 recombinant viruses using BAC cloning and prokaryotic recombination technologies. CyHV-3 FL BAC recombinant plasmids and
CyHV-3 recombinant viruses were produced on the basis of the strategy outlined in Fig. 2 and 8, by using a two-step galactokinase gene (galK) positive/negative selection in bacteria (15,23). Single-gene deleted recombinant plasmids were produced for various ORFs by replacing the sequence (initiation codon to stop codon) by a galK expression cassette (galK positive selection). Recombination cassettes encoding galK were produced by PCR using the primers listed in Table 3 and the pgalK vector as the template. The derived ORFX Del galK cassette consisted of the galK gene flanked by 75-bp sequences homologous to the regions of the CyHV-3 genome immediately upstream and downstream of ORFX. The flanking sequences served as targets to induce homologous recombination with the FL BAC parental plasmid in competent E. coli cells, resulting in replacement of the entire ORF by the galK expression cassette (FL BAC ORFX Del galK plasmid).
The recombinant FL BAC plasmids were transfected into permissive CCB cells in the presence of polyethylenimine (PEI). Transfection of the BAC plasmids and eventual reconstitution of infectious virus were monitored by epifluorescence microscopy, since the FL BAC plasmids and derived viruses contained an EGFP expression cassette. For mutations that did not prevent reconstitution of infectious virus (i.e., in nonessential genes), cell supernatants were collected and the viruses were amplified, leading to FL BAC recovered ORFX Del galK viruses. For mutations that prevented reconstitution of infectious virus (i.e., in essential genes), further investigations were performed. For each ORF identified as being essential for viral growth, WT revertant (Rev1) and nonsense (NS) mutation clones were produced by using a second recombination process (galK negative selection). ORFX Rev1 and ORFX NS recombination cassettes were produced by PCR using WT FL BAC DNA and pGEMT-ORFX NS as the templates, respectively (see   Table 3 for primers). The resulting FL BAC ORFX Rev1 and FL BAC ORF X NS plasmids were transfected into CCB cells as described above. To exclude the possibility that the lack of infectivity of the NS-null recombinant BAC plasmids was due to unexpected mutations during BAC manipulation, FL BAC ORFX NS plasmids were also transfected into CCB cells together with a WT DNA fragment containing the appropriate ORF and upstream and downstream flanking regions, in order to induce reversion to a WT sequence after homologous recombination in eukaryotic cells. Genetic characterization of CyHV-3 recombinant BAC plasmids and derived recombinant viruses. CyHV-3 FL BAC recombinant plasmids and derived recombinant viruses were characterized by sequencing regions of interest and by assessing the sizes of SacI fragments by agarose gel electrophoresis and Southern blotting (24). Probes for Southern blotting were produced by PCR using the primers listed in Table 3. CyHV-3 recombinant viruses deleted for nonessential genes (the FL BAC recovered ORFX Del galK viruses and the FL BAC revertant ORF25 Del galK virus) were characterized further by full-length genome sequencing as described previously (15).
Multistep growth curves. Triplicate cultures of CCB cells were infected with CyHV-3 viruses at an MOI of 0.05 PFU/cell. After incubation for 2 h, the cells were washed with PBS and overlaid with Dulbecco modified Eagle medium (DMEM) containing 4.5 g/liter glucose and 10% (vol/vol) FCS. The supernatant was removed from the infected cultures at successive intervals (2, 4, and 6 dpi) and stored at Ϫ80°C. Titers of infectious viral particles were determined by duplicate plaque assays in CCB cells (24).
Plaque size assay. CCB cells grown in 6-well plates were infected with viruses at an MOI of 200 PFU/well. After incubation for 2 h, the cells were overlaid with DMEM containing 10% (vol/vol) FCS and 0.6% (wt/vol) carboxymethyl cellulose (medium viscosity; Sigma) in order to obtain isolated plaques (25).