Array Analysis of Viral Gene Transcription during Lytic Infection of Cells in Tissue Culture with Varicella-Zoster Virus

Abstract

Varicella-zoster virus (VZV), a neurotropic alphaherpesvirus, causes childhood chickenpox (varicella), becomes latent in dorsal root and autonomic ganglia, and reactivates decades later to cause shingles (zoster) and other neurologic complications. Although the sequence and configuration of VZV DNA have been determined, relatively little is known about viral gene expression in productively infected cells. This is in part because VZV is highly cell associated, and sufficient titers of cell-free virus for use in synchronizing infection do not develop. PCR-based transcriptional arrays were constructed to simultaneously determine the relative abundance of the ≈70 predicted VZV open reading frames (ORFs). Fragments (250 to 600 bp) from the 5′ and 3′ end of each ORF were PCR amplified and inserted into plasmid vectors. The virus DNA inserts were amplified, quantitated, and spotted onto nylon membranes. Probing the arrays with radiolabeled cDNA synthesized from VZV-infected cells revealed an increase in the magnitude of the expressed VZV genes from days 1 to 3 after low-multiplicity virus infection but little change in their relative abundance. The most abundant VZV transcripts mapped to ORFs 9/9A, 64, 33/33A, and 49, of which only ORF 9 corresponded to a previously identified structural gene. Array analysis also mapped transcripts to three large intergenic regions previously thought to be transcriptionally silent, results subsequently confirmed by Northern blot and reverse transcription-PCR analysis. Array analysis provides a formidable tool to analyze transcription of an important ubiquitous human pathogen.

Varicella-zoster virus (VZV) is propagated by cocultivating infected cells with uninfected cells. Because VZV is cell associated and high titers of cell-free virus that can be used to synchronize infection do not develop, analysis of VZV gene transcription in infected cells in tissue culture has been limited. Three reports described global analysis of VZV transcription in cultures harvested at advanced stages of infection; Northern blot analysis of RNA extracted from late-stage VZV-infected cultures identified 58 to 67 discrete transcripts mapping throughout the virus genome (26, 32), while single-stranded DNA probes used in Northern blots revealed the direction of transcription of 57 of the 58 previously identified VZV transcripts and also identified 20 novel virus transcripts (35). Because the RNA in those studies was sampled at the height of cytopathic effect, the transcripts were presumed to reflect VZV structural genes.

To better understand the pattern of virus gene transcription during lytic infection, we constructed PCR-based VZV transcriptional arrays to analyze the magnitude of expression and relative abundance of each predicted VZV gene as well as detect transcription for three large intergenic regions of the virus genome.

MATERIALS AND METHODS

Virus and cells.

VZV (strain Ellen) was propagated as described previously (13) by cocultivation of infected and uninfected BSC-1 cells, a continuous cell line derived from African green monkey kidney cells. Since VZV is highly cell associated, and infection in vitro is achieved by cocultivation of infected cells with uninfected cells (unlike herpes simplex virus, where high titers of cell-free virus can be used to infect cells), accurate assessment of multiplicity of infection is not feasible. Instead, virus-infected cells were mixed with uninfected cells, resulting in a typical multiplicity of 8.6 × 10³ infected cells and 1.2 × 10⁶ uninfected cells (multiplicity of infection = 0.007).

VZV DNA was extracted from isolated nucleocapsids (44) by digestion with 50 μg of pronase K (Roche, Indianapolis, Ind.) per ml in 2% sodium dodecyl sulfate (SDS) and spooling on glass rods upon the addition of ethanol. The integrity of virus DNA was determined by restriction endonuclease digestion and agarose gel electrophoresis.

Array construction.

VZV transcription arrays were modeled after similar arrays developed to analyze human cytomegalovirus (2) and herpes simplex virus type 1 (39, 36) transcription. PCR primers (Integrated DNA Technologies, Coralville, Iowa) were selected to amplify 250- to 600-bp DNA segments from both the 5′ start and 3′ terminus of each predicted unique VZV open reading frame (ORF) (GenBank accession number NC 001348). For ORFs < 350 bp, one set of primers was selected to amplify the entire predicted ORF. Table 1 lists the oligonucleotide primers, the ORF to which the primers map, the 5′ location of each 20- to 25-bp primer, and the size and molar G+C content of each amplified fragment.

TABLE 1.

VZV oligonucleotide primers

Clone	ORF	Forward	5′ location	Reverse	5′ location	Size (bp)	G + C (%)
1	1	1FA	915	1RA	588	327	50.0
2	2	2FA	1134	2RA	1503	369	48.9
3	2	2FB	1483	2RB	1850	367	47.1
4	3	3FA	2452	3RA	2146	306	50.4
5	3	3FB	2198	3RB	1907	291	53.0
6	4	4FA	4142	4RA	3836	306	47.0
7	4	4FE	3088	4RE	2781	307	43.3
8	5	5FA	5275	5RA	4972	303	39.8
9	5	5FD	4553	5RD	4252	301	39.6
10	6	6FA	8577	6RA	8271	306	41.2
11	6	6FL	5640	6RL	5325	315	43.9
12	7	7FA	8605	7RA	8940	335	50.7
13	7	7FB	9078	7RB	9387	309	40.3
14	8	8FA	10667	8RA	10324	343	41.9
15	8	8FE	9783	8RE	9476	307	39.2
16	9	9FA	11009	9RA	11484	475	49.5
17	9	9FB	11464	9RB	11917	453	48.5
18	10	10FA	12160	10RA	12600	440	41.4
19	10	10FC	12995	10RC	13392	397	48.2
20	11	11FA	13590	11RA	13919	329	45.4
21	11	11FH	15752	11RH	16049	297	43.4
22	12	12FA	16214	12RA	16551	337	47.9
23	12	12FF	17854	12RF	18199	345	42.7
24	13	13FA	18440	13RA	18743	303	42.3
25	13	13FC	19043	13RC	19346	303	40.1
26	14	14FA	21113	14RA	20668	445	65.0
27	14	14FF	19731	14RF	19431	300	42.8
28	15	15FA	22479	15RA	22177	302	45.8
29	15	15FE	21554	15RE	21258	296	35.9
30	16	16FA	23794	16RA	23467	327	43.4
31	16	16FD	22885	16RD	22568	317	39.7
32	17	17FA	24149	17RA	24443	294	45.5
33	17	17FE	25233	17RE	25518	285	39.0
34	18	18FA	26493	18RA	26153	340	38.1
35	18	18FD	25876	18RD	25573	303	39.6
36	19	19FA	28845	19RA	28547	298	46.0
37	19	19FH	26826	19RH	26517	309	40.6
38	20	20FA	30475	20RA	30160	315	40.4
39	20	20FE	29371	20RE	29023	348	44.4
40	21	21FA	30755	21RA	31083	328	38.6
41	21	21FJ	33567	21RJ	33875	308	43.0
42	22	22FA	34082	22RA	34402	320	45.7
43	22	22FZ	42050	22RZ	42374	324	44.6
44	23	23FA	43119	23RA	42763	356	52.1
45	23	23FB	42774	23RB	42430	344	56.5
46	24	24FA	44021	24RA	43678	343	45.2
47	24	24FB	43560	24RB	43211	349	44.8
48	25	25FA	44618	25RA	44303	315	41.0
49	25	25FB	44448	25RB	44148	300	39.4
50	26	26FA	44506	26RA	44816	310	41.4
51	26	26FF	45966	26RF	46263	297	46.6
52	27	27FA	46127	27RA	46483	356	49.0
53	27	27FC	46792	27RC	47128	336	42.8
54	28	28FA	50636	28RA	50318	318	44.6
55	28	28FL	47356	28RL	47051	305	42.2
56	29	29FA	50857	29RA	51165	308	47.7
57	29	29FM	54168	29RM	54471	303	42.9
58	30	30FA	54650	30RA	54973	323	39.9
59	30	30FG	56631	30RG	56963	332	47.9
60	31	31FA	57008	31RA	57278	270	47.6
61	31	31FH	59315	31RH	59615	300	46.8
62	32	32FA	59766	32RA	59993	227	48.9
63	32	32FB	59974	32RB	60198	224	48.4
64	33	33FA	62138	33RA	61799	339	42.4
65	33	33FG	60674	33RG	60321	353	45.6
66	34	34FA	63911	34RA	63597	314	47.2
67	34	34FF	62496	34RF	62170	326	45.3
68	35	35FA	64753	35RA	64440	313	46.9
69	35	35FB	64283	35RB	63977	306	44.4
70	36	36FA	64806	36RA	65082	276	48.5
71	36	36FD	65566	36RD	65833	267	47.7
72	37	37FA	66074	37RA	66382	308	41.1
73	37	37FH	68317	37RH	68601	284	47.1
74	38	38FA	70293	38RA	69951	342	50.8
75	38	38FF	68979	38RF	68668	311	44.0
76	39	39FA	70633	39RA	70894	261	46.1
77	39	39FC	71055	39RC	71355	300	42.5
78	40	40FA	71540	40RA	71835	295	40.1
79	40	40FO	75421	40RO	75730	309	50.4
80	41	41FA	75846	41RA	76145	299	48.6
81	41	41FC	76478	41RC	76798	320	43.6
82	42	42FA	78038	42RA	77729	309	35.5
83	42	42FD	77182	42RD	76849	333	39.3
84	43	43FA	78170	43RA	78468	298	43.0
85	43	43FF	79883	43RF	80178	295	41.8
86	44	44FA	80337	44RA	80612	275	47.4
87	44	44FD	81129	44RD	81429	300	47.5
88	45	45FA	82574	45RA	82271	303	49.6
89	45	45FC	81854	45RC	81522	332	39.6
90	46	46FA	82717	46RA	83028	311	47.2
91	46	46FB	83008	46RB	83319	311	47.4
92	47	47FA	83168	47RA	83456	288	48.6
93	47	47FG	84405	47RG	84702	297	47.8
94	48	48FA	84667	48RA	84967	300	44.7
95	48	48FG	86029	48RG	86322	293	49.8
96	49	49FA	86226	49RA	86472	246	46.5
97	50	50FA	87883	50RA	87556	327	48.0
98	50	50FE	86886	50RE	86575	311	49.2
99	51	51FA	87880	51RA	88182	302	52.1
100	51	51FI	90090	51RI	90388	298	46.6
101	52	52FA	90493	52RA	90854	361	51.5
102	52	52FI	92512	52RI	92808	296	40.2
103	53	53FA	93850	53RA	93526	324	48.4
104	53	53FC	93153	53RC	92854	299	42.8
105	54	54FA	95984	54RA	95662	322	42.2
106	54	54FI	93983	54RI	93675	308	46.1
107	55	55FA	95996	55RA	96295	299	43.1
108	55	55FI	98343	55RI	98642	299	40.2
109	56	56FA	98568	56RA	98871	303	45.2
110	56	56FC	99006	56RC	99303	297	49.3
111	57	57FA	99628	57RA	99411	217	46.5
112	58	58FA	100272	58RA	99953	319	46.6
113	58	58FB	99925	58RB	99607	318	43.7
114	59	59FA	101219	59RA	100909	310	48.8
115	59	59FC	100605	59RC	100301	304	51.8
116	60	60FA	101651	60RA	101324	327	44.9
117	60	60FB	101466	60RB	101168	298	47.4
118	Int	6F	101650	6R	102019	369	38.8
119	Int	7F	102000	7R	102232	232	45.3
120	Int	8F	102221	8R	102680	459	40.3
121	Int	9F	102661	9R	103082	421	42.9
122	61	61FA	104486	61RA	104156	330	50.4
123	61	61FF	103381	61RF	103082	299	48.1
124	Int	10F	104486	10R	104924	438	42.6
125	Int	11F	104905	11R	105013	108	65.1
126	Int	12F	104996	12R	105182	186	64.7
127	62/71	71FA	109133	71RA	108834	299	58.8
128	62/71	71FB	108834	71RB	108426	408	60.1
129	62/71	71FC	108409	71RC	108014	395	58.2
130	62/71	71FD	108015	71RD	107657	358	59.8
131	62/71	71FE	107656	71RE	107177	479	64.3
132	62/71	71FF	107178	71RF	106759	419	65.8
133	62/71	71FG	106762	71RG	106242	520	67.2
134	62/71	71FH	123654	71RH	124086	432	66.8
135	62/71	71FI	105815	71RI	105204	611	70.8
136	62/71	71FK	105476	71RK	105200	276	69.3
137	Int	1F	109134	1R	109625	491	45.0
138	Int	2F	109626	2R	109895	269	68.2
139	Int	3F	109894	3R	110077	183	46.4
140	Int	4F	110085	4R	110328	243	40.8
141	Int	5F	110347	5R	110580	233	47.4
142	63/70	63FA	110580	63RA	110880	300	57.4
143	63/70	63FC	111118	63RC	111417	299	59.7
144	64/69	64FA	118332	64RA	118034	298	62.3
145	64/69	64FB	111809	64RB	112107	298	55.6
146	65	65FA	112640	65RA	112332	308	41.9
147	66	66FA	113037	66RA	113344	307	42.3
148	66	66FC	113911	66RC	114220	309	40.7
149	67	67FA	114496	67RA	114796	300	40.2
150	67	67FC	115336	67RC	115562	226	43.3
151	68	68FA	115808	68RA	116105	297	44.8
152	69	68FG	117371	68RG	117679	308	48.3

Initially, PCR analysis of VZV-infected and control BSC-1 cell DNA showed that each primer pair was specific for VZV DNA (data not shown). Each primer pair was then used to amplify 10⁶ copies of purified VZV DNA. To facilitate cloning, the primers contained restriction endonuclease sites (EcoRI, PstI, SstII, KpnI, SphI, or HindIII) 5′ proximal to the virus DNA sequence. The amplified product was digested at the engineered restriction endonuclease site, gel purified, and inserted into the multiple cloning site of pGEM3zf⁻ (Promega, Madison, Wis.). Plasmid construction was confirmed by PCR, restriction endonuclease digestion, agarose gel electrophoresis, and, when required, DNA sequencing (Macrogene, Inc., Seoul, Korea). Recombinant clones were propagated, plasmid DNA was extracted, and the VZV DNA insert was PCR amplified with universal plasmid primers. The integrity and quantity of each PCR product were determined by agarose gel electrophoresis, and all DNA targets samples were diluted to 10 ng/μl.

With the aid of multichannel pipettes, each target DNA (4 μl) was spotted onto nylon-based filters (Sigma, St. Louis, Mo.), air dried, denatured in 0.5 M NaOH for 12 min, neutralized in 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) for 4 min, again air dried, and UV cross-linked. Control targets on each array consisted of a 284-bp fragment of human actin cDNA, an 84-bp fragment of human glyceraldehyde phosphate dehydrogenase (GAPdH) cDNA, a 284-bp fragment from the chloramphenicol acetyl transferase (CAT) gene PCR amplified from the pCAT (Gibco-BRL, Carlsbad, Calif.) plasmid (plasmid DNA), and the PCR mixture consisting of vector primers and lacking template plasmid DNA (no DNA). Each set of four control spots was spotted in quadruplet on each array. Each membrane (15.5 by 11 cm) contained one spot each for the VZV targets and four spots for each control target. Each experiment consisted of duplicate arrays, and all experiments were repeated at least thrice with independently obtained probes. Computer analysis of DNA sequences was performed with DNAMax (MiraiBio, Inc., Alameda, Calif.).

RNA extraction and cDNA synthesis.

VZV-infected and control BSC-1 cells were scraped, pelleted, and frozen in liquid nitrogen. Cell pellets were thawed in Tri-Reagent (Molecular Research Center, Cincinnati, Ohio) and briefly sonicated to reduce viscosity, and after addition of 1-bromo-3-chloropropane for phase separation, RNA was precipitated, washed, and dissolved in water. Residual DNA was digested with RNase-free DNase (DNA-free; Ambion, Austin, Tex.) at 37°C for 20 min and poly(A)-containing RNA was extracted by batch-chromatography on oligo(dT)-cellulose (NucleoTrap, Clontech, Palo Alto, Calif.). Poly(A)⁺ RNA (1 μg) was primed with oligo(dT), and first-strand cDNA was synthesized with modified Moloney murine leukemia virus reverse transcriptase (SuperScript II; Invitrogen, Carlsbad, Calif.). After incubation at 37°C for 90 min, the enzyme was inactivated at 65°C for 5 min and cDNA was cleaned by filtration (Microcon-PCR; Millipore, Bedford, Mass.).

Array probe synthesis.

Array probes were synthesized either by directly incorporating [³²P]dCTP during cDNA synthesis or by labeling the cDNA with [³²P]dCTP by nick translation. To directly label cDNA, 50 μCi of [³²P]dCTP was lyophilized and added to the cDNA reaction modified to contain 500 ng of poly(A)⁺ RNA and a 10% concentration of unlabeled dCTP. For post-cDNA synthesis, nick translation probe labeling, five reactions (Prediprime II; Amersham Pharmacia Biotech, Piscataway, N.J.), each containing 100 ng of cDNA and 10 μCi of [³²P]dCTP, were incubated at 37°C for 60 min. Unincorporated isotope was removed from all labeling reactions by size exclusion gel filtration (Micro Bio-Spin P30; Bio-Rad, Hercules, Calif.). The array probes were heat denatured, quenched on ice, and applied to prehybridized arrays in 20 ml of Perfecthyb (Sigma-Aldrich, St. Louis, Mo.). Hybridization was for 48 h at 65°C, after which the blots were washed three times at 65°C at low stringency (2× SSC-0.1% SDS), twice at high stringency (0.5× SSC-0.1% SDS) and once at ultrahigh stringency (0.1× SSC-0.1% SDS). Arrays were air dried and exposed for uniform times to phosphor-storage screens (Molecular Dynamics, Piscataway, N.J.).

Data analysis.

Arrays were scanned (Storm; Molecular Dynamics) and target spot intensities were determined (Image-Qaunt; Molecular Dynamics). All experiments consisted of replicate arrays and were repeated at least three times. Each experiment also contained arrays hybridized to probes synthesized from uninfected control cells. To enable comparison of results from independent arrays, data were normalized according to the formula:

where X̄_VZVi is average VZV array spot intensity (i), X̄_coni is average control array spot intensity, X̄_VZVnd is average VZV array no-DNA spot intensity, and X̄_connd is average control array no-DNA spot intensity.

Northern blot analysis.

RNA was denatured in formaldehyde at 65°C for 15 min and resolved in 1.8% formamide-denaturing agarose gels, transferred to nylon-based membranes (Zeta Probe; Bio-Rad), and probed with either PCR-generated fragments labeled by nick translation or ³²P-end-labeled oligonucleotides (11).

3′-RACE.

First-strand cDNA was synthesized by priming the RNA with a modified oligonucleotide containing a PCR primer pad at the 5′ end of the oligo(dT) (CDS-primer; Clontech, Palo Alto, Calif.). Amplification of the cleaned cDNA with a VZV gene-specific primer located ≈500 nucleotides from the 3′ end of the selected ORF and the 3′ rapid amplification of cDNA ends (RACE) PCR primer resulted in the 3′-terminal sequences of the selected VZV transcript, including the poly(A) tract. Products were cleaned and the DNA sequence wasdetermined.

PCR.

Each 100-μl reaction mixture consisted of 0.3 μg of plasmid DNA or 10⁶ VZV DNA molecules in 1× PCR buffer, 1.2 pmol of each primer, 1.5 μmol of MgCl₂ and 0.5 to 1.0 units of Taq (PerkinElmer Life Sciences, Boston, Mass.). Reaction cycles consisted of denaturation for 1 min at 95°C, 40 cycles of amplification at 95°C for 30 s, 55°C for 30 s, and 72°C for 1.5 min, and one extension cycle (94°C for 30 s, 52°C for 30 s, and 72°C for 7 min).

RESULTS

Specificity of VZV arrays.

Figure 1 illustrates the array specificity. Each segment of cloned VZV DNA was flanked by ≈50 bp of common plasmid DNA. ³²P-end-labeled oligonucleotide probes mapping to this common plasmid sequence detected all amplified targets on the array (panel A), while probes made from cell DNA detected only the actin and GAPdH targets (panel B). Probes made from VZV DNA enriched from virus-infected cells revealed all cloned VZV targets and the cellular actin target (panel C). VZV DNA extracted from enriched nucleocapsids contained small amounts of cellular DNA resulting in detectable levels of cellular actin sequences in the array analysis. The intensity of the actin spots was greater than the intensity of GAPdH spots when the array probe consisted of cellular DNA (panel B), explaining the lack of GAPdH DNA hybridization signal in panel C.

FIG. 1.

Specificity of VZV arrays. VZV arrays were incubated with ³²P-end-labeled vector primer (A), ³²P-nick-translated control cell DNA (B), or VZV DNA (C) probes. All arrays were configured the same; Table 2 lists the location (column and row) of the 152 VZV DNA ORF targets. Each array also contained a quadruplet set of four control targets. Control targets in each array consisted of actin (row 10, columns 1 and 5; row 11, columns 9 and 13), GAPdH (row 10, columns 2 and 6; row 11, columns 10 and 14), plasmid DNA (row 10, columns 3 and 7; row 11, columns 11 and 15), and no DNA (row 10, columns 4 and 8; row 11, columns 12 and 16). After hybridization, the intensity of each spot was quantitated by phosphoimaging. The graph to the right of each array shows the spot intensity (arbitrary units) associated with each of the 152 VZV target clones, along with the average spot intensity associated with the array controls. Error bars indicate the standard error of the mean for each data point. The red shaded regions in the graphs show the mean ± 1 standard deviation for the VZV target DNAs. The three-dimensional Lorentzian transformation of spot intensity versus insert molar G+C and insert size (in base pairs) revealed no statistical correlation when the array was probed with the plasmid oligonucleotide (D), but a slight correlation between spot intensity and insert G+C content when the array was probed with VZV DNA (E).

Quantitation of spot intensity from the array in panel A revealed no clustering of targets with intensities of >1 standard deviation from the mean (red-shaded region in the plasmid graph), indicating the absence of geometric bias in the array. No statistical correlation between spot intensity and insert G+C molar content (R² = 0.07) or target size (R² < 0.01) was detected in arrays probed with the plasmid-specific oligonucleotide (panel D). Similarly, quantitation of spot intensity from the array probed with VZV DNA (panel C) showed that most hybridization signals were within 1 standard deviation from the mean (red shaded region). However, a cluster of targets corresponding to clones 129 to 144 showed ≈2-fold higher than average signal intensity. These targets mapped to VZV ORFs 62, 63, and 64 and the intergenic region separating VZV ORFs 62 and 63, and all of these targets mapped within the inverted repeat of the unique short segment (IR_S) of the VZV genome. Thus, a randomly labeled, nick-translated probe synthesized from VZV DNA would be expected to overrepresent (by twofold) the amount of probe accessible to targets within the IR_S. Alternatively, the IR_S contain areas of high G+C molar content, which could result in a higher hybridization signal. Panel E illustrates this minimal association of spot intensity with (G+C)% (R² = 0.57), and the absence of any significant association between spot volume and VZV DNA insert size (R² = 0.08).

Sensitivity of VZV arrays.

The relative abundance of individual transcripts within an RNA preparation is most accurately determined by array analysis when the probe is labeled during first-strand cDNA synthesis (17, 40). Whereas direct cDNA labeling of VZV-infected cell mRNA yielded probe with activity sufficient to detect numerous VZV transcripts, labeling the synthesized cDNA by nick translation yielded probes with higher activity. With the single exception of clone 136, which contains a 276-bp insert mapping to the 3′ terminus of ORF 62, the relative abundance of the virus transcripts was maintained independent of the labeling method (data not shown).

Transcription of VZV genes during lytic infection.

VZV is highly cell associated, which precludes the ability to obtain high titers of cell-free virus, a requirement for synchronized high-multiplicity infections and one-step growth analysis. Instead, virus infection is initiated by cocultivation of infected cells with uninfected cells. Such pseudo-one-step analysis has been used to compare the growth kinetics of wild-type and mutant VZVs containing site-directed mutations (4, 9, 30, 34). Thus, sequential array analysis reveals virus gene transcription as the infection progresses through the culture (Fig. 2). Nick-translated cDNA array probes were synthesized from VZV-infected cell RNA harvested at 1 to 4 days postinfection and compared to probes synthesized from control cells harvested after 2 days in culture (Fig. 2). Probe synthesized from control cells detected only actin and GAPdH transcripts (panel A), whereas probe synthesized from VZV-infected cells harvested at 1 to 4 days postinfection detected, along with actin and GAPdH, numerous VZV targets (panels B to E). Quantitative analysis of individual spot intensities revealed increased expression of VZV genes as a function of time but no significant variation in the pattern of virus gene expression (Fig. 2, graphs).

FIG. 2.

Array analysis of VZV transcripts during lytic infection. Arrays were hybridized to probes synthesized from control cell RNA (A) or RNA extracted from VZV-infected cells harvested 1 day postinfection (dpi) (B), 2 days postinfection (C), 3 days postinfection (D), or 4 days postinfection (E). VZV DNA target and control spots are configured on each array as described in the Fig. 1 legend. After hybridization, spot intensities were determined and normalized to the values obtained from the array hybridized with the control cell probe. Graphs show the relative expression (rel. exp.) associated with each VZV DNA target clone and the control spots in the arrays. The overall abundance of VZV transcription peaked at 3 days postinfection, when targets mapping to all possible virus genes were detected.

Table 2 presents the quantitative analysis of VZV gene transcription. Columns 1 to 5 correspond to the array images and list the clone number, predicted VZV ORF, and region to which the cloned VZV DNA insert maps and the physical coordinates that locate each clone on the array by column and row. Other columns list the relative expression and standard deviation of each array target generated from probe synthesized from VZV-infected cells harvested at 1 to 4 days postinfection and the order (in descending relative expression) of each VZV target resulting from the probe at 1 to 4 days postinfection. For example, the highest relative expression obtained with the 1-day postinfection probe was associated with clone 17, the VZV insert mapping to the 3′ terminus of ORF 9. Since DNA corresponding to clones 26, 27, 126, 135, and 139 (indicated in italics) was not applied to the arrays, the order number of these spots was used to determine the threshold for detection of VZV gene expression. Thus, transcripts mapping to a particular target VZV clone are listed as present (indicated in bold) if the order number was less than the minimum order number corresponding to clones 26, 27, 126, 135, or 139.

TABLE 2.

Quantitative analysis of VZV gene transcription^a

A1 clone	A2 ORF	A3 region	A4 column	A5 row	1 dpi			2 dpi			3 dpi			4 dpi
					B1 rel exp	B2 std	B3 order	C1 rel exp	C2 std	C3 order	D1 rel exp	D2 std	D3 order	E1 rel exp	E2 std	E3 order
1	1	all	1	1	3.6	0.7	36	19.3	3.8	39	62.5	8.9	24	36.0	1.5	19
2	2	5′	1	2	3.7	0.7	25	28.2	0.6	22	71.9	6.6	19	38.0	0.7	17
3	2	3′	1	3	3.4	0.5	56	13.2	0.3	55	26.5	1.0	47	13.1	0.4	43
4	3	5′	1	4	3.5	0.2	49	12.8	0.4	59	19.3	0.6	59	15.0	0.6	41
5	3	3′	1	5	4.1	0.7	8	25.5	1.2	28	51.3	4.9	29	39.8	2.4	15
6	4	5′	1	6	2.7	0.1	117	5.7	0.9	119	7.8	0.4	118	5.3	0.1	116
7	4	3′	1	7	2.7	0.1	122	21.8	0.3	34	42.7	3.8	36	19.2	0.8	33
8	5	5′	1	8	1.5	0.7	147	2.5	0.5	148	4.8	0.0	141	1.9	0.1	150
9	5	3′	1	9	0.3	0.1	152	0.4	0.1	152	0.5	0.0	152	0.5	0.0	152
10	6	5′	2	1	3.3	0.2	59	10.8	0.3	68	11.4	0.1	91	5.5	0.1	114
11	6	3′	2	2	3.6	0.3	35	9.4	0.7	77	12.1	0.2	85	6.8	0.3	90
12	7	5′	2	3	3.5	0.3	46	9.7	1.2	74	17.3	0.1	69	8.8	0.1	70
13	7	3′	2	4	4.5	0.3	5	14.0	3.2	51	26.6	1.2	46	14.8	1.7	42
14	8	5′	2	5	3.2	0.6	81	6.1	0.3	111	8.6	0.0	106	5.8	0.0	110
15	8	3′	2	6	3.1	0.2	96	9.3	1.3	79	8.5	0.0	107	6.7	0.1	91
16	9	5′	2	7	2.9	0.7	109	46.5	1.0	11	164.5	35.3	8	114.3	2.3	1
17	9	3′	2	8	10.8	0.7	1	179.5	20.7	1	321.0	2.7	1	93.9	9.0	2
18	10	5′	2	9	1.7	0.3	145	2.4	0.4	149	3.8	0.5	145	2.5	0.2	146
19	10	3′	3	1	4.0	1.5	11	28.8	1.9	21	63.3	0.7	22	26.1	0.3	24
20	11	5′	3	2	4.4	1.0	6	24.6	4.7	30	82.3	2.2	17	25.3	0.7	25
21	11	3′	3	3	3.7	0.5	30	31.4	4.9	17	93.8	13.9	13	29.8	0.2	22
22	12	5′	3	4	3.3	0.2	63	6.0	2.0	112	8.2	0.3	113	6.0	0.2	102
23	12	3′	3	5	3.2	0.1	87	5.4	1.5	126	5.8	0.3	137	4.8	0.0	124
24	13	5′	3	6	3.2	0.9	89	9.3	1.3	78	14.3	0.0	78	10.0	0.3	56
25	13	3′	3	7	3.2	0.6	82	14.6	2.4	49	27.8	0.5	44	15.3	0.1	40
26	14	5′	3	8	2.4	0.3	130	2.7	0.0	145	3.2	0.3	148	2.5	0.2	147
27	14	3′	3	9	2.4	0.2	131	2.6	0.0	146	2.4	0.2	149	2.4	0.2	148
28	15	5′	4	1	3.8	0.8	23	17.5	1.9	44	36.1	3.9	39	18.7	0.5	36
29	15	3′	4	2	3.3	0.8	61	12.1	3.1	64	27.4	3.6	45	11.5	0.2	48
30	16	5′	4	3	3.3	0.3	67	6.4	1.5	106	6.9	0.2	126	3.9	0.0	137
31	16	3′	4	4	3.3	0.1	74	5.5	1.5	122	6.7	0.0	132	5.1	0.0	119
32	17	5′	4	5	3.3	0.0	66	7.0	1.5	94	7.3	0.0	124	5.3	0.0	117
33	17	3′	4	6	3.3	0.2	69	8.9	2.9	82	11.2	0.3	94	6.9	0.1	89
34	18	5′	4	7	2.8	0.1	114	8.9	1.0	84	18.0	2.4	64	7.9	0.2	76
35	18	3′	4	8	3.7	0.3	28	15.4	0.5	48	24.8	2.7	52	9.4	0.4	60
36	19	5′	4	9	2.4	0.3	128	3.8	0.8	138	5.7	0.4	138	3.9	0.4	136
37	19	3′	5	1	3.7	0.0	26	9.6	1.1	76	15.2	0.5	76	7.5	0.1	80
38	20	5′	5	2	3.4	0.1	52	21.6	7.2	35	48.7	2.3	31	15.9	0.1	39
39	20	3′	5	3	3.7	0.4	29	30.7	6.1	19	69.2	8.2	20	21.5	0.6	29
40	21	5′	5	4	3.2	0.1	90	5.8	2.5	117	6.6	0.2	133	4.3	0.1	133
41	21	3′	5	5	3.6	0.2	39	13.8	2.7	52	21.1	0.0	56	8.7	0.0	71
42	22	5′	5	6	3.0	0.2	103	13.3	4.4	54	16.9	0.9	72	8.8	0.1	69
43	22	3′	5	7	2.5	0.0	126	10.4	1.4	70	18.5	1.5	61	7.1	0.0	85
44	23	5′	5	8	0.9	0.2	149	6.0	0.0	114	8.3	0.6	111	6.6	0.3	92
45	23	3′	5	9	0.4	0.1	150	5.7	0.8	118	12.2	1.6	83	3.3	0.2	143
46	24	5′	6	1	3.6	0.2	40	25.0	3.9	29	61.0	5.6	25	33.4	1.4	20
47	24	3′	6	2	3.8	0.1	20	26.1	3.0	26	87.2	2.0	15	41.2	2.4	14
48	25	5′	6	3	3.7	0.3	32	12.3	2.6	63	19.4	0.9	58	9.2	0.1	65
49	25	3′	6	4	3.0	0.4	101	5.3	0.7	128	6.8	0.4	129	4.3	0.1	132
50	26	5′	6	5	3.4	0.7	58	13.8	2.6	53	17.0	0.8	71	7.8	0.1	78
51	26	3′	6	6	3.1	0.2	94	12.9	2.2	58	19.4	0.0	57	8.9	0.1	68
52	27	5′	6	7	2.3	0.1	133	9.0	3.0	81	16.0	0.4	74	7.4	0.4	82
53	27	3′	6	8	1.9	0.1	143	9.7	1.6	75	25.5	0.5	49	9.4	0.4	61
54	28	5′	6	9	1.8	0.2	144	7.3	0.6	93	17.5	0.1	68	6.0	0.7	101
55	28	3′	7	1	3.8	0.3	17	27.6	2.2	23	56.1	7.1	27	20.8	0.8	32
56	29	5′	7	2	3.5	0.7	45	17.4	0.5	45	25.3	0.7	51	12.7	0.0	44
57	29	3′	7	3	3.4	0.4	50	18.5	0.1	41	34.4	2.0	40	11.2	0.1	51
58	30	5′	7	4	3.3	0.1	75	5.9	1.6	115	7.3	0.0	123	4.7	0.0	126
59	30	3′	7	5	3.2	0.0	80	10.8	2.1	69	12.3	0.1	82	9.9	0.2	57
60	31	5′	7	6	2.8	0.4	111	6.7	1.9	100	11.7	0.1	88	5.9	0.2	105
61	31	3′	7	7	2.6	0.2	124	12.5	3.5	61	18.3	1.1	62	9.4	0.5	64
62	32	5′	7	8	1.9	0.1	142	8.9	2.9	83	17.6	0.4	66	9.4	0.4	63
63	32	3′	7	9	3.1	0.3	92	13.0	1.6	57	38.4	0.3	38	12.3	0.2	45
64	33	5′	8	1	4.0	0.8	14	53.7	0.5	10	192.9	22.6	5	55.6	2.2	9
65	33	3′	8	2	4.7	1.5	4	78.7	11.2	3	222.9	39.4	3	68.3	2.9	7
66	34	5′	8	3	3.4	0.2	51	10.0	0.6	72	11.8	0.9	87	5.8	0.0	108
67	34	3′	8	4	3.3	0.1	71	8.1	0.2	90	11.9	0.6	86	5.8	0.3	109
68	35	5′	8	5	2.8	0.2	112	5.4	1.0	127	8.2	0.2	112	4.5	0.2	130
69	35	3′	8	6	4.0	1.8	13	5.9	1.6	116	9.9	0.1	100	6.0	0.2	103
70	36	5′	8	7	2.7	0.7	119	4.8	2.2	135	7.4	0.4	122	4.6	0.2	128
71	36	3′	8	8	2.6	0.3	123	12.1	2.4	65	22.6	1.5	54	8.1	0.8	74
72	37	5′	8	9	2.4	0.3	129	3.8	0.9	140	5.0	0.0	140	3.9	0.3	135
73	37	3′	9	1	4.1	0.3	10	31.0	5.0	18	55.4	8.9	28	16.9	0.9	38
74	38	5′	9	2	3.2	0.5	76	17.9	3.6	42	50.3	2.7	30	27.1	1.7	23
75	38	3′	9	3	3.2	0.2	85	38.8	0.8	13	87.0	10.0	16	36.5	1.7	18
76	39	5′	9	4	3.0	0.1	102	6.2	1.6	110	9.2	0.4	102	5.9	0.2	106
77	39	3′	9	5	3.2	0.5	78	11.3	2.8	66	11.6	1.2	90	7.4	0.3	81
78	40	5′	9	6	3.1	0.4	97	6.2	1.5	109	8.3	0.2	110	5.7	0.3	111
79	40	3′	9	7	2.9	0.4	108	5.0	1.3	134	8.5	0.1	108	5.6	0.5	113
80	41	5′	9	8	2.1	0.3	138	6.7	1.4	102	17.8	0.1	65	11.3	0.7	50
81	41	3′	9	9	1.6	0.3	146	17.3	4.2	46	28.6	5.9	43	10.2	0.9	55
82	42	5′	9	10	0.3	0.0	151	0.5	0.1	151	1.0	0.0	151	0.5	0.0	151
83	42	3′	10	1	3.2	0.6	84	20.8	1.2	37	39.1	3.8	37	10.8	0.1	54
84	43	5′	10	2	3.3	0.2	62	15.5	2.2	47	17.3	1.6	70	7.0	0.1	88
85	43	3′	10	3	2.7	0.2	120	12.4	2.2	62	8.0	0.8	115	4.2	0.2	134
86	44	5′	10	4	3.0	0.2	104	14.1	2.6	50	11.2	1.0	95	5.9	0.3	107
87	44	3′	10	5	3.5	0.3	47	28.8	11.1	20	47.9	4.8	32	23.5	1.8	27
88	45	5′	10	6	3.1	0.6	93	6.4	3.0	104	8.4	0.6	109	5.6	0.3	112
89	45	3′	10	7	2.8	0.4	113	5.0	2.3	131	7.6	0.6	120	4.8	0.3	125
90	46	5′	10	8	2.9	0.7	105	5.5	2.9	124	7.9	0.6	116	5.1	0.2	122
91	46	3′	10	9	2.6	0.5	125	3.8	1.6	139	4.5	0.2	143	3.8	0.3	141
92	47	5′	10	10	2.1	0.6	136	2.6	1.1	147	3.7	0.1	146	3.0	0.2	144
93	47	3′	11	1	3.2	1.1	83	19.8	3.2	38	6.7	0.4	131	5.0	0.1	123
94	48	5′	11	2	3.4	1.3	57	18.7	7.0	40	18.2	1.5	63	12.2	0.4	46
95	48	3′	11	3	3.9	0.6	15	54.4	7.0	9	130.8	36.3	10	54.3	2.6	10
96	49	all	11	4	5.9	0.7	3	69.5	25.1	5	152.1	31.9	9	70.5	13.8	6
97	50	5′	11	5	3.6	0.9	37	25.9	11.8	27	56.2	10.4	26	25.0	1.2	26
98	50	3′	11	6	3.3	1.1	60	57.4	26.6	8	112.0	6.7	11	43.2	1.9	13
99	51	5′	11	7	3.7	0.5	31	24.1	14.6	31	44.1	4.5	35	21.5	0.9	30
100	51	3′	11	8	2.9	0.6	107	5.6	3.4	121	6.7	0.6	130	5.1	0.2	121
101	52	5′	11	9	2.7	0.9	116	6.3	3.4	107	7.0	0.8	125	5.9	0.0	104
102	52	3′	11	10	2.3	0.6	132	5.2	2.6	129	6.8	0.4	127	4.4	0.4	131
103	53	5′	12	1	3.4	1.2	54	21.9	3.1	33	21.3	0.6	55	11.4	0.2	49
104	53	3′	12	2	3.4	1.2	55	33.5	4.7	16	30.1	2.9	42	17.8	0.5	37
105	54	5′	12	3	3.8	1.3	19	20.9	5.5	36	45.6	4.2	34	23.4	0.1	28
106	54	3′	12	4	3.1	0.8	91	6.9	2.4	96	14.1	0.4	79	9.1	0.1	67
107	55	5′	12	5	2.9	0.4	106	6.7	3.3	98	8.9	0.6	105	5.1	0.3	120
108	55	3′	12	6	3.6	0.8	41	8.5	4.4	85	11.3	1.2	93	6.5	0.2	93
109	56	5′	12	7	3.3	0.4	68	6.7	3.7	99	8.9	0.4	104	6.2	0.3	97
110	56	3′	12	8	3.1	0.2	100	8.4	5.8	87	11.0	0.7	96	9.9	0.0	58
111	57	all	12	9	4.0	1.5	12	40.0	25.4	12	90.4	8.2	14	52.0	2.6	11
112	58	5′	12	10	2.1	0.6	139	11.2	5.5	67	25.6	1.4	48	9.4	0.3	62
113	58	3′	13	1	3.8	1.5	18	38.5	2.8	14	107.0	0.6	12	39.0	0.2	16
114	59	5′	13	2	2.8	0.9	115	12.7	6.7	60	11.7	1.1	89	6.1	0.1	99
115	59	3′	13	3	3.6	1.4	34	13.1	3.2	56	25.5	2.3	50	11.0	0.1	53
116	60	5′	13	4	3.3	1.4	65	9.9	3.4	73	14.7	0.5	77	8.3	0.2	72
117	60	3′	13	5	3.3	1.4	72	10.1	3.6	71	17.6	0.8	67	7.9	0.5	75
118	Int 6	na	13	6	3.2	0.0	86	7.3	2.9	92	10.8	0.8	97	7.1	0.3	84
119	Int 7	na	13	7	3.5	0.1	48	6.9	3.4	97	8.1	1.2	114	6.3	0.3	96
120	Int 8	na	13	8	2.5	0.2	127	3.5	1.9	142	4.6	0.3	142	3.8	0.2	139
121	Int 9	na	13	9	3.2	0.8	79	8.4	5.2	86	12.6	0.7	81	9.8	0.6	59
122	61	5′	13	10	1.0	0.5	148	2.9	1.5	144	6.8	0.0	128	2.8	0.0	145
123	61	3′	14	1	3.3	1.4	73	26.5	1.5	25	64.4	4.4	21	45.1	0.7	12
124	Int 10	na	14	2	3.5	1.0	42	17.7	6.5	43	23.4	1.1	53	11.1	0.3	52
125	Int 11	na	14	3	3.9	0.9	16	5.5	2.5	123	7.8	0.1	117	4.6	0.0	127
126	Int 12	na	14	4	3.5	0.5	44	3.7	0.7	141	3.5	0.1	147	3.5	0.1	142
127	62	5′-a	14	5	2.7	0.5	121	4.6	1.9	136	6.4	0.3	134	4.5	0.0	129
128	62	5′-b	14	6	3.2	0.7	77	6.0	2.1	113	10.4	0.0	98	7.1	0.2	86
129	62	5′-c	14	7	3.1	0.1	95	5.0	1.7	132	5.9	0.2	136	6.3	0.1	95
130	62	5′-d	14	8	2.7	0.0	118	5.0	1.5	133	5.6	0.1	139	5.2	0.2	118
131	62	5′-e	14	9	2.8	0.5	110	9.0	3.6	80	16.4	0.2	73	12.1	0.4	47
132	62	5′-f	14	10	2.0	0.8	141	6.4	1.6	105	12.1	0.8	84	8.2	0.5	73
133	62	5′-g	15	1	2.3	0.9	135	8.3	0.7	88	12.7	0.6	80	6.1	0.3	100
134	62	5′-h	15	2	3.6	1.1	38	8.0	0.0	91	10.3	0.2	99	6.2	1.0	98
135	62	5′-i	15	3	3.8	0.2	22	4.0	0.8	137	4.0	0.2	144	3.9	0.1	138
136	62	3′	15	4	3.1	0.7	99	71.6	14.4	4	174.9	5.7	6	60.4	1.0	8
137	Int 1	na	15	5	4.1	0.3	9	23.6	12.9	32	47.0	1.0	33	21.4	0.7	31
138	Int 2	na	15	6	3.3	0.3	70	6.7	3.3	101	11.4	0.1	92	7.8	0.0	79
139	Int 3	na	15	7	2.3	0.2	134	2.0	0.2	150	1.9	0.0	150	2.4	0.0	149
140	Int 4	na	15	8	2.1	0.4	137	3.4	1.2	143	6.3	0.9	135	3.8	0.3	140
141	Int 5	na	15	9	3.1	0.5	98	5.2	1.6	130	9.3	0.4	101	7.9	0.0	77
142	63	5′	15	10	2.0	0.4	140	6.6	4.2	103	30.4	0.1	41	19.0	0.4	34
143	63	3′	16	1	3.7	1.7	33	64.4	15.7	7	194.3	3.7	4	91.2	6.4	3
144	64	5′	16	2	3.8	1.2	24	69.0	11.2	6	165.0	8.2	7	74.3	4.0	5
145	64	3′	16	3	4.3	1.4	7	100.4	31.8	2	236.0	31.8	2	83.7	2.3	4
146	65	all	16	4	3.8	0.5	21	8.1	3.0	89	15.2	1.9	75	9.2	0.0	66
147	66	5′	16	5	3.7	0.2	27	6.3	2.8	108	7.7	0.7	119	5.5	0.2	115
148	66	3′	16	6	3.5	0.1	43	5.6	3.0	120	8.9	0.3	103	6.4	0.1	94
149	67	5′	16	7	3.4	0.4	53	7.0	4.7	95	19.0	0.2	60	7.3	0.2	83
150	67	3′	16	8	6.4	2.4	2	26.6	19.7	24	62.5	8.4	23	18.8	0.2	35
151	68	5′	16	9	3.2	0.4	88	5.5	3.3	125	7.5	0.4	121	7.0	0.3	87
152	68	3′	16	10	3.3	0.3	64	34.4	23.7	15	76.8	6.4	18	33.0	0.8	21
Array controls
Avg plasmid DNA					2.3	0.2	132	3.3	0.3	144	4.1	0.2	144	4.7	0.1	126
Avg no DNA					1.0	0.1	149	1.0	0.0	152	1.0	0.1	152	1.0	0.1	152
Avg actin					0.1	0.0	155	0.6	0.0	153	0.2	0.0	155	0.2	0.0	155
Avg GAPdH					0.1	0.0	156	0.0	0.0	156	0.0	0.0	156	0.1	0.0	156

^aVZV open reading frame region [5′, 3′, or entire (all)] was PCR amplified and inserted into pGEM. Target DNA array coordinates (column and row) are given. rel exp, relative expression of the individual target region. std, standard deviation of mean relative expression. order, order (in decreasing magnitude) of the individual target relative expression. dpi, day postinfection.

RNA samples harvested at 1 day postinfection detected 21 (14%) of the total 147 VZV-specific array targets. Similarly, VZV-infected samples harvested at 2 days postinfection detected 136 (92%), samples harvested 3 days postinfection detected 143 (97%), and samples harvested 4 days postinfection detected 137 (93%) of the available VZV-specific array targets. Considering that 64 of the 68 predicted unique VZV ORFs contained multiple targets and that targets corresponding to ORF 14 were not present on the arrays, samples harvested 1 day postinfection detected 18 of the possible 67 VZV ORFs. Similarly, samples harvested 2 and 3 days postinfection detect 100% of the 67 VZV ORFs. Samples from day postinfection 4, which lacked detectable ORF 5 transcripts (clones 8 and 9), detected 66 of the possible 67 VZV ORFs.

A bias was detected when multiple targets were associated with a specific VZV gene, especially evident in the 1 day postinfection sample, 12 of 14 times towards the 3′ end compared to 2 of 14 times towards the 5′ end, most likely reflecting incomplete oligo(dT)-primed, first-strand cDNA synthesis resulting in underrepresentation of the 5′ start of the ORF.

Table 3 summarizes the relative expression of VZV ORFs, given in decreasing order of abundance and as the average relative expression for both the 3′ and 5′ targets for each ORF from 2 to 4 days postinfection. The function assigned to the individual virus genes (21, 42) is also listed. Figure 3, . visually representing the transcription expression data in Table 3, shows that the four most abundant (relative expression ≈100 and greater) VZV transcripts detected during virus growth mapping to ORFs 9, 64, 33, and 49 are located throughout the virus genome.

TABLE 3.

Relative expression of VZV ORFs during lytic infection

ORF	Relative expression	Gene function
9	153.3	Abundant tegument phosphoprotein
64	121.4
33	112.0	Protease
49	97.4
63	67.7	Tegument (transactivator/transrepressor)
57	60.8
62	53.7	Transactivator, tegument protein
50	53.3
48	48.1
11	47.9
24	45.6
38	42.9
1	39.3	Membrane protein
58	38.5
20	34.6
2	31.8
10	30.1	Transactivator, tegument protein
61	29.1	Transactivator, transrepressor
68	27.4	Glycoprotein (gE)
3	27.3
42	23.6
67	23.5	Glycoprotein (gl)
53	22.7
28	22.6	DNA polymerase
37	22.4	Glycoprotein (gH)
44	21.9
15	20.6
54	20.0
29	19.9	Single-stranded DNA-binding protein
51	17.9	Origin-binding protein
4	17.1	Transactivator, tegument protein
32	16.6	Phosphoprotein
41	15.3
13	15.2	Thymidylate synthetase
7	15.2
18	14.1	Ribonucleotide reductase, small subunit
59	13.4	Uracil-DNA glycosylase
26	13.3
27	12.8
22	12.5
60	11.4	Glycoprotein (gL)
65	10.8
31	10.7	Glycoprotein (gB)
43	10.7
21	10.1
36	9.9	Thymidine kinase
25	9.6
6	9.3
34	8.9
39	8.6
56	8.5
47	8.5	Protein kinase
30	8.5
19	8.4	Ribonucleotide reductase, large subunit
55	7.8
17	7.8
23	7.8
8	7.5	dUTPase
66	6.7	Protein kinase
35	6.6
40	6.6	Major nucleocapsid protein
45	6.3
12	6.0
52	5.9
16	5.7
46	5.7
5	4.8	Glycoprotein (gK)
14	nca	Glycoprotein (gC)

^aORF 14 was not cloned (nc).

FIG. 3.

Relative expression of VZV genes during lytic infection. The VZV genome consists of unique long (U_L) and short (U_S) segments, each bounded by terminal and internal inverted repeated DNA sequences (TR_L/IR_L and IR_s/TR_s). Transcription of VZV genes was determined during lytic virus growth in cell culture (Fig. 2), and relative expression of virus genes was assessed (Table 3). The relative size, direction of transcription, and expression of each of the 71 predicted virus ORFs show that the most abundant transcripts (ORFs 9, 33, 49, 63/70, and 64/69) mapped across the virus genome.

Confirmation of array results (DNase digestion).

Because array cDNA probes generated by nick translation are also expected to radiolabel residual VZV DNA, PCR was used to demonstrate the removal of residual DNA from the RNA preparations. RNA extracted at 1 to 4 days postinfection was incubated with or without reverse transcriptase (RT), and the resulting material was amplified with actin or ORF 9 primers. Resolution of RT-dependent bands at 242 bp (actin) and 452 bp (VZV ORF 9) indicated that residual DNA was eliminated from the RNA preparations, and the band intensities reflected the relative expression levels determined by array analysis (data not shown).

Confirmation of array results (Northern blot and 3′-RACE).

RNA from control and VZV-infected BSC-1 cells was resolved on Northern filters and hybridized with various ORF and intergenic VZV DNA probes (Fig. 4). . Both the ORF 9 probes detected two predominant transcripts at 1.68 and 1.25 kb; the ORF 64 probes detected transcripts at 1.67 and 1.01 kb; ORF 63 probes detected transcripts at 1.67 and 1.22 kb; the ORF 62 probe detected a single 3.35-kb transcript; ORF 61 probes detected a 1.54-kb transcript; and the ORF 40 probe detected a major transcript at 5.19 kb.

FIG. 4.

Northern blot analysis of VZV transcripts. Control (C) and VZV-infected (V) cell RNA was resolved on denaturing gels and hybridized to PCR-generated probes specific for the 5′ or 3′ end of ORFs 9, 64, 63, 62, 61. and 40, the antisense oligonucleotide 9RB (Table 1), or the PCR-generated intergenic (Int) fragments 1, 9, and 10. Asterisks indicate major bands detected in each lane. The relative expression (rel. exp.) of each ORF determined by array analysis (Table 3) is listed below the respective Northern blot. The DNA sequences of the ORF 9 and ORF 61 termini determined by 3′-RACE are also listed. The termination codon (TAG) is underlined, and the putative poly(A) addition signal is both italicized and underlined. The proposed ORF 9A, 9, and 61 mRNA structures are shown at the bottom of the figure.

To determine whether both the transcripts detected with the ORF 9 probes mapped to the same DNA strand, the Northern blot was probed with oligonucleotide 9FB (Table 1). This oligonucleotide maps antisense to the 3′ terminus of ORF 9 and detected both the 1.68- and 1.25-kb transcripts. The 3′ end of the ORF 9 transcripts was PCR amplified, and the DNA sequence located the poly(A) tract to a unique site 52 nucleotides distal to the ORF 9 termination codon, 12 nucleotides downstream from the predicted poly(A) addition signal. ORF 9A is a newly identified transcript mapping 107 nucleotides upstream of the ORF 9 initiation codon (38). Thus, Northern blot analysis with PCR-generated and strand-specific oligonucleotide probes, coupled with the 3′-RACE and DNA sequence results, are consistent with readthrough from the ORF 9A promoter to the ORF 9 termination/poly(A) addition signal, thereby generating two 3′-coterminal transcripts (1.69-kb ORF 9A and 1.25-kb ORF 9).

The major 1.01-kb ORF 64 transcript, detected with both 5′ and 3′ probes, corresponds to the predicted 539-bp ORF 64, indicating ≈470 nucleotides of untranslated RNA. The major 1.22-kb transcript detected by both ORF 63 probes corresponds to the predicted 833-nucleotide ORF 63, indicating 387 nucleotides of untranslated RNA. Comigrating bands at 1.67 kb were detected with all ORF 63 and ORF 64 probes, indicating that this transcript is the result of gene 64 readthrough to the gene 63 termination/poly(A) addition signal.

VZV DNA analysis indicates that gene 62 includes a 3,930-nucleotides ORF. Northern blots analyzed with ORF 62, 5′, and Int 1 probes detected a single VZV-specific transcript at ≈3.4 kb. This transcript most likely corresponds to the gene 62 transcript, with the Int-1 probe mapping to the 5′ untranslated region of the full-length transcript. The size discrepancy between the expected gene 62 transcript and that seen on the Northern blot rests in the inherent difficulty in determining the size of large transcripts on denaturing agarose gels.

A single 1.54-kb transcript was detected in VZV-infected BSC-1 RNA with probes generated from both ORF 61 termini, along with Int-9 and Int-10 DNA sequences. 3′-RACE used to amplify the end of VZV gene 61 from virus-infected cell RNA showed that the DNA sequence contains a poly(A) tract 99 nucleotides distal to the ORF 61 termination codon, 17 nucleotides distal to the predicted poly(A) addition signal. Primer extension has mapped the 5′ start of gene 61 transcription 69 nucleotides upstream of the initiation codon (31), and our present results map the 3′ terminus 99 nucleotides downstream of the termination codon. Int-9 and Int-10 probes detected the 5′ and 3′ untranslated regions, respectively, associated with the 1.54-kb ORF 61 transcripts.

ORF 40 encodes the major VZV nucleocapsid protein. This predicted 4,187-nucleotide ORF is contained within the ≈4.9-kb transcript detected with the 3′ ORF 40 probe. The low relative abundance of this transcript determined by array analysis was reflected in the Northern analysis, which required longer exposure times to detect ORF 40 transcripts. With the single exception of the ORF 61 transcript detected with the 5′ probe, the relative expression of each gene determined by array analysis was reflected in the band intensity on the Northern blots. The discrepancy noticed in ORF 61 between the relative abundance determined by array analysis and that inferred by Northern blot analysis is especially noteworthy (Fig. 4). For example, array analysis ranks ORF 61 expression below ORF 62 expression (29.1 and 53.7, respectively). However, Northern blot analysis suggests that ORF 61 expression is greater than ORF 62 expression. Some of the major factors which may interfere with both analyses are RNA stability, probe activity, and means of detection. It is possible that the ≈4-kb ORF 62 transcript is less stable than the ≈1.5-kb ORF 61 transcript, resulting in under representation of ORF 62 by Northern analysis. Alternatively, probe synthesis and hybridization kinetics for ORF 61, 5′ probe (G+C = 50.4%) may be more efficient that that of ORF 62, 3′ probe (G+C = 69.3%).

Confirmation of array results (RT-PCR analysis of intergenic regions).

Array analysis indicated the presence of transcripts mapping to the three major intergenic regions of the VZV genome in virus-infected tissue culture cells. While Northern blot analysis mapped Int-1 and Int-9/10 probes to gene 62 and 61 transcripts, respectively, results for transcripts mapping to the remaining Int regions were not conclusive. Thus, PCR was used to confirm transcription within the intergenic regions (Fig. 5). . First-strand cDNA was synthesized from control and VZV-infected cell RNA either with or without reverse transcriptase, and PCR-amplified with primers mapping to the three major intergenic regions of the virus genome. Comparison of the PCR product obtained by amplification of VZV-infected cell cDNA with that obtained from the VZV DNA amplification showed that, with the exception of Int-3(F/R), all Int primer pairs tested yielded product transcription across these “silent” regions of the virus genome. The lack of product from control cell RNA and VZV-infected cell RNA demonstrated the specificity of amplification and removal of residual virus DNA the reactions.

FIG. 5.

RT-PCR analysis of VZV intergenic DNA. Control (C) and VZV-infected (V) cell RNA was incubated with (+) or without (−) reverse transcriptase and PCR amplified with selected intergenic (Int)-specific primers. Amplified products were compared with those after PCR amplification of VZV DNA. The predicted size of the amplification product (amplicon size) for each Int region primer set is listed. Int regions 6 to 8 map between ORFs 60 and 61; Int regions 10 to 12 map between ORFs 61 and 62 and contain the IR_L/IR_S junction; and Int regions 1 to 5 map between ORFs 62 and 63 and contain the VZV DNA origin replication (ori).

Int-8 primers yielded two PCR products in the amplification of VZV cDNA, possibly indicating posttranscriptional processing of the full-length message. Int-11 primers resulted in two bands when VZV DNA or cDNA was amplified. The presence of the top band in all samples is the result of nonspecific amplification; however, the ≈100-bp band, present only in the VZV cDNA and in VZV DNA, approximates the 109-bp theoretical size indicating virus transcription across this region.

DISCUSSION

This study is the first to use PCR-based arrays for the simultaneous detection and quantitation of VZV transcripts. The array targets contained PCR fragments from all originally predicted ORFs (10) with the single exception of ORF 14, which contains multiple R2 reiterations that render the clones unstable. ORF S/L is the first demonstrated spliced VZV transcript and maps to the leftward end of the virus genome (19). The 21-kDa cytoplasmic ORF S/L protein is expressed during lytic virus growth (in vitro and in vivo), and null mutations of this gene yield virus with altered cell adhesion characteristics. ORF S/L-specific targets were not placed on this first-generation VZV array.

The four most abundant (relative expression > ≈100) VZV transcripts mapped to ORFs 9, 64, 33, and 49. Northern blot and DNA sequencing indicated that the ORF 9 products detected by array analysis were composed of two 3′-coterminal overlapping transcripts. ORF 9A is located 367 bp upstream from the ORF 9 initiation codon. The ORF 9-associated bands detected by Northern analysis differed by 0.34 kb. Together, the data indicate that ORF 9 and 9A are coterminal, consistent with the ORF 9 doublet band detected by Ross et al. (38), and map the 3′ terminus of these transcripts. Even though quantitation of the Northern blots indicated a ≈4-fold higher abundance of ORF 9 than ORF 9A transcripts (3.8 ± 0.4, n = 5), ORF 9 expression was greater than that of ORF 64, the second most abundant transcript detected by array analysis.

Recently, OFR 9 protein has been shown to be an abundant tegument protein phosphorylated by ORF 47-encoded protein kinase (42). Phosphorylated ORF 9 protein associates with phosphorylated IE62, and the complex is liberated during virus uncoating. Along with stabilizing the virion tegument, ORF 9 protein may also help direct IE62 to appropriate promoter sites. The VZV ORF 9 homologue in herpes simplex virus type 1 is UL49, a posttranslationally modified tegument protein which associates with cellular chromatin (1, 22). Thus, the structural ORF 9 protein may also help orchestrate virus gene expression. ORF 9A encodes a nonessential 7-kDa membrane-associated protein which is associated with syncytium formation (38).

VZV ORF 64, which encodes a predicted 19.8-kDa protein, is located within the repeat region of the unique short segment of VZV DNA and is therefore also present as ORF 69. Site-directed gene knockout studies identified a phenotype only when both ORF 64 and 69 were deleted (41). In VZV containing both ORF 64 and 69 mutations, glycoprotein E (gE) expression and syncytium formation are increased. While VZV ORF 64 and 69 proteins are dispensable for both in vitro and in vivo virus growth, they may play a role in the establishment or maintenance of VZV latency (41). Herpes simplex virus type 1 US10, the homologue to VZV ORF 64, is a capsid/tegument-associated phosphoprotein that associates with the nuclear matrix of infected cells (46).

Our Northern blot analysis suggests low-level ORF 63 readthrough to the ORF 64 termination signal. Multiple termination signals may be more common than previously thought, particularly since ORFs 18 and 19 have been shown to be coterminal (unpublished data) in cDNA libraries constructed from VZV-infected cells (8). An additional example of readthrough may be seen in the relative abundance of ORFs 42 and 45, which encode putative exons from a single processed transcript. The relative expression of ORF 42 is greater than that of ORF 45. A possible explanation is that transcription originating from upstream genes such as ORFs 40 and 41 with the termination signal supplied by ORF 43 or 44 would result in greater expression of ORF 42 than of 45. Further studies are needed to confirm such a hypothesis. Meanwhile, even the prediction that both ORFs (45 and 42) are spliced has not been confirmed.

ORFs 33 and 33.5 encode highly immunogenic proteins associated with the assembly of virus nucleocapsids (12, 16, 27, 33). The ORF 33 transcript encodes the full-length precursor protein, which is posttranslationally cleaved to yield the assembly protein protease (ORF 33) and the structural protein substrate (ORF 33.5). Thus, the nested genes are in-frame, 3′-coterminal encoding proteins with identical C termini.

The protein encoded by VZV ORF 49 has not been identified. The herpes simplex virus type 1 homologue UL11 is a posttranslationally modified, Golgi-associated, viral tegument protein that functions in nucleocapsid envelopment and egress. UL11 protein may be involved in membrane structure, since constitutive cellular expression of UL11 blocks herpes simplex virus type 1 gD-mediated virus attachment (23-25, 36).

The most abundant VZV transcripts expressed during virus propagation are associated with membrane structure (ORF 9/9A, ORF 64, and ORF 49) or capsid assembly (ORF 33/33.5). Contrary to previous assumptions that ORF 40 transcripts, encoding the major VZV nucleocapsid protein, were highly expressed during lytic virus infection (3, 6-8, 20), our array analysis suggests and Northern blot analysis confirms that ORF 40 transcripts in productively infected cells are low abundance (23-fold less abundant than ORF 9/9A).

Unlike experiments involving high-titer cell-free virus or metabolic inhibitors of protein and DNA synthesis (5), this experiment was designed to determine the relative expression of VZV genes as virus infection progresses through the culture and was not designed to determine the traditional one-step growth kinetics. As such, it might be expected that the virus genes expressed are predominantly late (i.e., virus structural genes); however, immediate-early VZV genes are also detected. Of the 67 VZV genes detected by array analysis, immediate-early genes 63 and 62 are ranked 5 and 7, immediate-early genes 61 is ranked 18, and immediate-early gene 4 is ranked 31 (Table 3). Thus, propagation of the virus by cocultivation resulting in large multinucleated syncytia is accompanied by high levels of expression of membrane-associated proteins, immediate-early proteins 63 and 62, and the expression of numerous genes whose function has not been identified (for example, ORF 57).

Oligonucleotide-based arrays have been used to determine the time-course of herpes simplex virus type 1 transcription (43). Comparison of the relative abundance of herpes simplex virus type 1 versus VZV transcripts reveals striking differences in the pattern of gene expression. For example, only the homologue of VZV ORF 64 is included among the five most abundant herpes simplex virus type 1 genes transcribed at early times after infection. Transcription of this herpes simplex virus type 1 gene diminishes late in infection and is superseded by transcription of the gene encoding the herpes simplex virus major capsid protein. Both early and late in herpes simplex virus infection, transcription of the VZV ORF 9A/9 homologue approximates the mean herpes simplex virus transcriptional level. Just as DNA sequence analysis demonstrates genetic differences between the two human neurotropic alphaherpesviruses (28), transcriptional array analysis indicates how these genetic differences are reflected functionally.

A caveat in interpreting transcriptional data arises from the ability of array and Northern blot analysis to provide only snapshot of the cell at time of harvest. Thus, high abundance may result from low-level transcription of stable RNA, while low abundance may rest in rapid degradation of an actively transcribed gene. Our studies addressed relative abundance of VZV gene transcripts, not RNA stability. On the other hand, the global approach of array analysis, besides providing a temporal description of virus gene expression, enabled discovery of novel transcripts, such as those mapping to the three major intergenic regions. Northern blot analysis confirmed three of the Int-target results and demonstrated an association with untranslated regions outside ORFs 61 and 62. PCR evidence confirmed the array analysis by demonstrating the RT-dependent amplification of cDNA within the intergenic regions. Future research is focused on mapping these transcripts, but preliminary evidence suggests posttranscriptional processing of transcripts within these regions (unpublished data).

Global transcriptional analysis is potentially valuable in several areas. For example, wild-type and VZV gene knock-out constructs can be compared to determine how the virus compensates for the lack of specific genes. Comparison of virus gene transcription in various cell lines or in human biopsy and explanted tissue in SCID-hu mice (29) might elucidate cellular effects on virus gene expression. Combining VZV transcriptional array data with wild-type and specific gene knockout mutant virus with the global cellular transcription analysis afforded by cellular gene chip experiments (18) will demonstrate the cellular response to specific virus genes. Transcriptional array analysis of parental and vaccine OKA VZV along with the recently available complete virus genetic information (14) will help to determine the molecular effects of virus attenuation.

Various methods to obtain cell-free VZV, albeit at low titers (14, 15), together with methods of probe signal enhancement (37, 45) will allow synchronized virus infection and temporal classification of VZV genes, information required for a better understanding of the virus's growth cycle and especially important when comparing VZV to the prototype neurotropic alphaherpesvirus herpes simplex virus type 1. Herpes simplex virus type 1 infection is characterized by the ordered sequential expression of immediate-early, early, and late genes (43). In this report, at the earliest time examined (1 day postinfection), little transcription (cellular or virus) was detected. The differences in gene expression between VZV (this report) and herpes simplex virus type 1 (43), especially early in virus infection, most likely reflect the mode of infection (cocultivation and unsynchronized versus synchronized high-multiplicity infection). Nevertheless, even with unsynchronized infection, we detected some VZV-specific transcripts within 24 h after infection, followed by increasing numbers and abundance of VZV-specific transcripts corresponding to all putative kinetic classes during the first 3 days (Fig. 2).

Finally, application of high-throughput technology with global VZV transcriptional array analysis to the study of human ganglia will increase our understanding of virus latency to an unprecedented level. Whereas these studies were previously limited to PCR or in situ results for selected virus genes, the technology is now available to search the entire VZV genome for transcription in a single experiment. Thus, a most vexing and basic question in the field of VZV research, i.e., which VZV genes are transcribed in latently infected human ganglia, can now be addressed.