Kaposi's Sarcoma-Associated Herpesvirus ORF66 Is Essential for Late Gene Expression and Virus Production via Interaction with ORF34

KSHV ORF66 is expressed during the early stages of lytic infection, and ORF66 and vPIC are thought to contribute significantly to late gene expression. However, the physiological importance of ORF66 in terms of vPIC formation remains poorly understood. Therefore, we generated an ORF66-deficient BAC clone and evaluated its viral replication. The results showed that ORF66 plays a critical role in virus production and the transcription of L genes. To our knowledge, this is the first report showing the function of ORF66 in virus replication using ORF66-deficient KSHV. We also clarified that ORF66 interacts with the transcription start site of the K8.1 gene, a late gene. Furthermore, we identified the ORF34-binding motifs in the ORF66 C terminus: three C-X-X-C sequences and a leucine-repeat sequence, which are highly conserved among beta- and gammaherpesviruses. Our study provides insights into the regulatory mechanisms of not only the late gene expression of KSHV but also those of other herpesviruses.

ABSTRACT

Kaposi’s sarcoma-associated herpesvirus (KSHV) is closely associated with B-cell and endothelial cell malignancies. After the initial infection, KSHV retains its viral genome in the nucleus of the host cell and establishes a lifelong latency. During lytic infection, KSHV-encoded lytic-related proteins are expressed in a sequential manner and are classified as immediate early, early, and late (L) gene transcripts. The transcriptional initiation of KSHV late genes is thought to require the complex formation of the viral preinitiation complex (vPIC), which may consist of at least 6 transcription factors (ORF18, -24, -30, -31, -34, and -66). However, the functional role of ORF66 in vPIC during KSHV replication remains largely unclear. Here, we generated ORF66-deficient KSHV using a bacterial artificial chromosome (BAC) system to evaluate its role during viral replication. While ORF66-deficient KSHV demonstrated mainly attenuated late gene expression and decreased virus production, viral DNA replication was unaffected. Chromatin immunoprecipitation analysis showed that ORF66 bound to the promoters of a late gene (K8.1) but did not bind to those of a latent gene (ORF72), an immediate early gene (ORF16), or an early gene (ORF46/47). Furthermore, we found that three highly conserved C-X-X-C sequences and a conserved leucine repeat in the C-terminal region of ORF66 were essential for the interaction with ORF34, the transcription of K8.1, and virus production. The interaction between ORF66 and ORF34 occurred in a zinc-dependent manner. Our data support a model in which ORF66 serves as a critical vPIC component to promote late viral gene expression and virus production.

IMPORTANCE KSHV ORF66 is expressed during the early stages of lytic infection, and ORF66 and vPIC are thought to contribute significantly to late gene expression. However, the physiological importance of ORF66 in terms of vPIC formation remains poorly understood. Therefore, we generated an ORF66-deficient BAC clone and evaluated its viral replication. The results showed that ORF66 plays a critical role in virus production and the transcription of L genes. To our knowledge, this is the first report showing the function of ORF66 in virus replication using ORF66-deficient KSHV. We also clarified that ORF66 interacts with the transcription start site of the K8.1 gene, a late gene. Furthermore, we identified the ORF34-binding motifs in the ORF66 C terminus: three C-X-X-C sequences and a leucine-repeat sequence, which are highly conserved among beta- and gammaherpesviruses. Our study provides insights into the regulatory mechanisms of not only the late gene expression of KSHV but also those of other herpesviruses.

INTRODUCTION

Kaposi’s sarcoma-associated herpesvirus (KSHV; also known as human herpesvirus 8 [HHV8]) causes Kaposi’s sarcoma, primary effusion lymphoma (PEL), and multicentric Castleman’s disease (1–3). Over a quarter of a century has passed since the first discovery of KSHV DNA fragments in a Kaposi’s sarcoma lesion of an AIDS patient in 1994 (4), and several aspects of KSHV pathogenesis and the KSHV life cycle and viral protein function have been elucidated. Compared with other viruses, herpesvirus has a large number of viral genes in its genome. KSHV encodes not only viral proteins but also microRNAs (miRNAs) and long noncoding RNAs (lncRNAs). These viral molecules are thought to be essential for KSHV replication and KSHV pathogenesis. One characteristic of the KSHV life cycle is the establishment of a lifelong latency in the infected individual, leading to KSHV-associated malignancy in patients with severe immunosuppression caused by drugs after organ transplantation or AIDS (1–5). The development of KSHV-associated neoplasms occurs due to infected cells which express few latent KSHV genes (including LANA, v-FLIP, kaposin, and miRNAs) (6). These genes modulate cell proliferation and apoptosis pathways (7–11). Another characteristic of the KSHV life cycle is active virus production, known as lytic infection. The genes related to lytic infection have been categorized into three groups: (i) primary lytic/immediate early (IE) genes, (ii) secondary lytic/delayed early/early (E) genes, and (iii) tertiary lytic/late (L) genes (12–14). The IE gene product RTA/ORF50 is a transcription factor inducing the transcriptional activation of other lytic genes and initiates the shift from latency to lytic infection (13, 15). The transcribed products of E genes start viral genome DNA replication. Finally, the transcriptional products of L genes contribute to viral particle formation by encoding viral structural proteins (14).

The viral preinitiation complex (vPIC) has recently been proposed to regulate L gene expression (16). vPIC is composed of several viral proteins conserved among beta- and gammaherpesviruses (16) and has functional homology with the host preinitiation complex, which consists of the TATA-binding protein (TBP) and general transcriptional factors (GTFs). Whereas the host preinitiation complex accumulates on the TATA box of the transcription start site (TSS) and initiates cellular RNA polymerase II (Pol II)-mediated transcription, vPIC accumulates on the TATT box of the viral gene TSS and initiates RNA Pol II-mediated transcription (16).

The function of the vPIC machinery and its components has been extensively studied in the context of Epstein-Barr virus (EBV), murine gammaherpesvirus 68 (MHV68), cytomegalovirus (CMV), and KSHV (16). In KSHV, at least 6 viral proteins contribute to vPIC formation. The viral TBP homolog KSHV ORF24 directly binds to the TATT box on the promoter sequences of the KSHV genome and is essential for the recruitment of host RNA Pol II (17, 18). We and other groups revealed that KSHV ORF34 acts as a hub for the interaction between ORF24 and other vPIC components, such as ORF18, ORF30, ORF31, and ORF66 (17, 19, 20). Split luciferase assays and coimmunoprecipitation experiments revealed that ORF34 directly or indirectly interacts with ORF18, -30, -31, and 66 (17, 19–21). ORF24 binds to the promoter of the L genes with RNA Pol II, and ORF34 serves as a bridge between ORF24 and a complex of ORF18, -30, -31, and -66 (17, 19, 20). Furthermore, ORF18 (22), ORF30 (20), ORF31 (22), and ORF34 (19) are essential for virus replication and gene expression. Although ORF66 appears to be a vPIC component (17, 19, 21), its importance and function during KSHV replication remain unknown. Therefore, we established ORF66-deficient KSHV and evaluated its physiological role during viral replication. Here, we show that ORF66 is essential for virus production and L gene expression via its interaction with ORF34.

(This article was submitted to an online preprint archive [23].)

RESULTS

Construction of ORF66-deficient KSHV BAC.

We constructed an ORF66-deficient recombinant KSHV bacterial artificial chromosome 16 (ΔORF66-BAC16) clone to study the impact of ORF66 during KSHV replication. Three stop codons (3-stop element) were inserted into the ORF66 coding region of KSHV BAC16 using a two-step markerless Red recombination system (24, 25) (Fig. 1a). Because ORF66 overlaps ORF67 (Fig. 1a), the three stop codons were inserted within the ORF66 gene to avoid interference with the coding frame of ORF67. The insertion and deletion of a kanamycin resistance gene (Kan^r) were analyzed by EcoRV digestion (Fig. 1b). The mutations and the insertion of the three stop codons into ΔORF66-BAC16 were confirmed by Sanger sequencing (Fig. 1c).

FIG 1.

Construction of ORF66-deficient recombinant KSHV BAC16. (a) Schematic illustration of the KSHV genome, including the ORF66-coding region. Using a two-step Red recombination system, three stop codons were inserted into the ORF66 coding region of KSHV BAC16 (nucleotides [nt] 113417 to 113416; GenBank accession number GQ994935) to construct an ORF66-deficient BAC clone (ΔORF66-BAC16). a.a., amino acids. (b) Agarose gel electrophoresis of the recombinant KSHV BACmids, which were digested with EcoRV. The asterisks indicate insertion and deletion of a kanamycin resistance cassette in each BAC clone. K, thousand. (c) DNA sequencing results for ORF66 mutagenesis sites in ΔORF66-BAC16.

ORF66 deficiency abrogates virus production and late gene expression.

To efficiently induce recombinant KSHV, tetracycline-inducible (Tet-on) RTA/ORF50-expressing SLK cells (iSLK) and Vero cells (iVero) were used as virus producer cells (19). The wild-type (WT) recombinant KSHV BAC16 (WT-BAC16) or ΔORF66-BAC16 clone was transfected into iVero or iSLK cells and then selected with hygromycin to generate the recombinant KSHV-inducible stable cell lines iVero-WT, iVero-ΔORF66, iSLK-WT, and iSLK-ΔORF66. To evaluate whether ORF66 is critical for KSHV replication, virus production and virus genome replication in iSLK-ΔORF66 and iVero-ΔORF66 cells were analyzed. iSLK-WT, iSLK-ΔORF66, iVero-WT, and iVero-ΔORF66 cells were treated with doxycycline (Dox) and sodium butyrate (NaB), and culture supernatants were harvested. The amount of WT KSHV and ΔORF66-KSHV in iSLK or iVero cell culture supernatants was measured by real-time PCR (Fig. 2a and e). As a result, the amount of ΔORF66-KSHV produced in iSLK and iVero cells was about 1,000-fold and 100-fold lower than the amount of WT KSHV produced. In contrast, there was no significant difference in cell-associated KSHV DNA levels between WT KSHV- and ΔORF66-KSHV-producing cells (Fig. 2b and f). Next, to clarify that the reduction of virus production in iSLK-ΔORF66 and iVero-ΔORF66 cells was caused by ORF66 deficiency, we tested whether exogenous ORF66 expression could rescue virus production in iSLK-ΔORF66 and iVero-ΔORF66 cells. iSLK-ΔORF66 and iVero-ΔORF66 cells stably transfected with the empty plasmid or the 3xFLAG-tagged ORF66 expression plasmid were cloned by neomycin (G418) or blasticidin S. The protein expression levels of exogenous 3xFLAG-tagged ORF66 were confirmed by Western blotting (Fig. 2c and g). Virus production from the iSLK-ΔORF66 and iVero-ΔORF66 cell lines was partially but significantly recovered when 3xFLAG-ORF66 was exogenously expressed (Fig. 2d and h). Furthermore, we asked whether exogenous ORF66 could rescue the transcription of L genes in iSLK-ΔORF66 cells. The iSLK-ΔORF66 cells stably expressing 3xFLAG-tagged ORF66 were treated with Dox and NaB, and expression of the L gene K8.1 and the E gene K-bZIP was analyzed (Fig. 2i). Lytic induction in the presence of exogenous ORF66 induced high levels of K8.1 expression in iSLK-ΔORF66 cells, whereas lytic induction in the absence of exogenous ORF66 did not. K-bZIP expression was unchanged in iSLK-Δ66 KSHV cells that underwent lytic induction with and without exogenous ORF66. These results indicate that ORF66 is crucial for KSHV replication and may function in steps following viral DNA replication.

FIG 2.

ORF66 is essential for virus production but not DNA replication of KSHV. Virus production in iSLK-WT and iSLK-ΔORF66 cells (a) and iVero-WT and iVero-ΔORF66cells (e). Each cell line was cultured for 72 h with medium containing NaB and Dox. KSHV DNA was purified from capsidated KSHV virions in culture supernatants, and KSHV genome copies were determined by real-time PCR. Virus production in iSLK-WT and iSLK-ΔORF66 cells (b) and iVero-WT and iVero-ΔORF66 cells (f). Each cell line was cultured for 48 h with medium containing NaB and Dox to induce a lytic state. Cellular DNA containing KSHV genomic DNA was purified from each cell line. KSHV genome copies were determined by real-time PCR and normalized by the total DNA amount. (c, g) Establishment of iSLK-ΔORF66 (c) and iVero-ΔORF66 (g) stable cell lines stably expressing 3xFLAG-tagged ORF66. The Western blots show exogenous ORF66. (d, h) Rescue of virus production in iSLK-ΔORF66 cells (d) and iVero-ΔORF66 cells (h) by exogenous ORF66 expression. Each stable cell line was cultured with NaB- and Dox-containing medium for 3 days, and the culture supernatant containing virus was harvested and quantified. (a, b, d to f, h) Three or four independent samples were evaluated by real-time PCR. The error bars indicate standard deviations. (i) Effect of exogenous ORF66 expression on the transcription of the K8.1 gene (a late gene) and the K-bZIP gene (an early gene) in iSLK-Δ66 KSHV cells. iSLK-ΔORF66 cells stably expressing 3xFLAG-tagged ORF66 were cultured for 72 h in medium with or without NaB and Dox. The expression of K8.1 and K-bZIP was analyzed by Western blotting. N.S., not significant.

To evaluate the contribution of ORF66 to KSHV gene expression, we performed a reverse transcription (RT)-quantitative real-time PCR (qPCR) array on the viral gene. Total RNA was extracted from iSLK-WT and iSLK-ΔORF66 cells that had been stimulated with NaB and Dox for 72 h. RNA was subjected to the RT-qPCR array as previously reported (26). Primers were designed based on the work of Fakhari and Dittmer (26) and modified to fit a unique coding DNA sequence (CDS) described by Bruce et al. (27), as shown in Table 1. Our data showed a broad reduction in the amount of KSHV mRNA in iSLK-ΔORF66 cells compared to iSLK-WT cells (Fig. 3), where 5 out of the top 10 downregulated genes were late transcripts (K8.1; ORF26 and ORF25; ORF27, ORF26, and ORF25; K9/vIRF1; ORF25). In particular, K8.1, ORF26, ORF27, and ORF25 have been previously reported by Nandakumar and Glaunsinger (28) to be direct vPIC targets. In contrast to L genes, latent and IE transcripts were mildly downregulated.

TABLE 1.

Primers for RT-qPCR of KSHV transcripts

Description	Unique CDSa	Unique CDS start-stop (nt)a	Strand	Timingb	Classificationc	Primer name	Primer sequence (5′ → 3′)	Primer sequence start-stop (nt)d
K1	K1	105–747	+	Latent	Latent	K1-F	GCAATCGTCTCCAAATCTCTGC	167–188
						K1-R	GCAATACCAGGATGTTGGCAAG	239–260
ORF4	ORF4	2592–2764	+	24 h	E	ORF4-F	GCCTCAGAGACCGCGAGA	2608–2625
						ORF4-R	AGCGATTTTTAGACGCCGG	2654–2672
ORF6	ORF6	3179–6577	+	24–48 h	E	ORF6-F	CTGCCATAGGAGGGATGTTTG	5271–5291
						ORF6-F	CCATGAGCATTGCTCTGGCT	5317–5336
ORF7	ORF7	6980–8618	+	NA		ORF7-F	TTTATTTCCCAGTCCTCCAAATG	8225–8247
						ORF7-R	GGGAAGCATGCCCGC	8274–8288
ORF7 + 8	ORF8	8681–11202	+	48–72 h	L	ORF8-F	CCCGACGTAGATCGCAGG	10969–10986
						ORF8-R	GTTTTTGATTTCCTCCCGTGTT	11010–11031
ORF7 + 8+9	ORF9	11329–14367	+	48–72 h	L	ORF9-F	TAGGCGCTTCGTGCTGG	12036–12052
						ORF9-R	CCGGATTGCTGCACTCGTA	12079–12097
ORF7 + 8+9 + 10	ORF10	14485–15522	+	48–72 h	L	ORF10-F	CAAGAAGGCCCGTATTCGTTTC	15201–15222
						ORF10-R	GGATAGACTGCGAATACCCTGG	15299–15320
ORF7 + 8+9 + 10+11	ORF11	15791–16979	+	8 h	IE	ORF11-F	CGGAATGGCGCCCAA	16738–16752
						ORF11-R	GACGGGATGATCACTCGTGTT	16779–16799
ORF2+K2	K2	17227–17841	−	ND		K2-F	GGATGCTATGGGTGATCGATG	17700–17720
						K2-R	ACCCTTGCAGATGCCGG	17656–17672
ORF2	ORF2	18114–18519	−	NA		ORF2-F	TCGGTAGTCAGCTGTCGAAAAC	18398–18419
						ORF2-R	GTTGCGGTTGATACCAAACTCG	18474–18495
K3	K3	18574–19542	−	24 h	E	K3-F	AGCCCCATCGCCCG	18715–18728
						K3-R	TGAGCGGTATAGGGCCACTTAC	18754–18775
ORF70	ORF70	20023–21036	−	8 h	IE	ORF70-F	AGGCGCGGAAAGGGAC	20129–20144
						ORF70-R	AAACGCATATAGAGCCACTACGG	20167–20189
K4	K4	21480–21764	−	8 h	IE	K4-F	TTGTCCGGTCTATGCCAGG	21667–21685
						K4-R	CTGCCTTGCTTTGTTTGCAA	21708–21727
K5	K5	25865–26635	−	8 h	IE	K5-F	CTTTTTTGTGGGCGCGC	25901–25917
						K5-R	AACGACCGTGCGGGACT	25944–25960
K6	K6	27289–27576	−	8 h	IE	K6-F	GGCGTGTACGACACGAGTGA	27479–27498
						K6-R	GCGTACTGCTTGCCACGTT	27527–27545
K7	K7	27693–28818	+	ND		K7-F	CCTTTGTGGGCTTTGAGTTCTG	27891–27912
						K7-R	CAGATGACCCCTGAAACTACCC	27956–27977
K7+PAN	PAN	28868–29895	+	8 h	IE	PAN-F	GCCGCTTCTGGTTTTCATTG	28929–28945
						PAN-F	TTGCCAAAAGCGACGCA	28887–28906
ORF16	ORF16	30242–30769	+	8 h	IE	ORF16-F	ACCAGCTTGGGTTGAGCATG	30441–30460
						ORF16-R	GGCTCGCCCCCAGTTC	30487–30502
ORF17	ORF17	31838–32524	−	48 h	L	ORF17-F	GGACTGACGAAATTTGGTGTGG	31950–31971
						ORF17-R	AGTGGGTGGTTTCCAGATTCTC	32040–32061
ORF17 + 17.5	ORF17.5	30920–31786	−	24 h	E	ORF17.5-F	GAGCGACTGCTGGCTTCAAC	30978–30997
						ORF17.5-R	CGGTGGAGAAAGACGCTCC	31023–31041
ORF18	ORF18	32605–32967	+	24 h	E	ORF18-F	GGCGAATTATCTGTTTCACCGG	32885–32906
						ORF18-R	CAAAAAGTTTCCGTCCACCAGG	32945–32966
ORF19 + 20	ORF19	33535–34709	−	ND		ORF19-F	ATACCAGGTTCAAGCGGCG	33772–33790
						ORF19-R	TGGATTGCTGGAGTTTGGG	33815–33833
ORF20	ORF20	35049–35401	−	ND		ORF20-F	AATGGATCGAGTCGGAGAGTTG	35114–35135
						ORF20-R	TTGTCTGAGTCACCAGCTGATC	35181–35202
ORF21	ORF21	35672–37147	+	48–72 h	L	ORF21-F	CGTAGCCGACGCGGATAA	37059–37076
						ORF21-R	TGCCTGTAGATTTCGGTCCAC	37101–37121
ORF21 + 22	ORF22	37212–39351	+	48–72 h	L	ORF22-F	CACCTTGGCGGATTTGGGATC	37623–37643
						ORF22-R	ACGGCCATGACAATCATTGGG	37727–37747
ORF23 + 24	ORF23	39401–40615	−	48–72 h	L	ORF23-F	TGCCGTCACATATCAGTTCGA	39502–39522
						ORF23-R	CCCCAAAGACCGTCAAAGC	39548–39566
ORF24	ORF24	40665–42636	−	48–72 h	L	ORF24-F	AGAAGTCAAACAGGCCCCG	40924–40942
						ORF24-R	GTTCGTTTCTCAGGCTTGACG	40969–40989
ORF25	ORF25	43041–46982	+	48–72 h	L	ORF25-F	CTCGGCGACGTGCTATACAAT	46608–46628
						ORF25-R	TGCCGACAAGGACTGTACATG	46658–46678
ORF25 + 26	ORF26	47032–47873	+	48–72 h	L	ORF26-F	AGCCGAAAGGATTCCACCAT	47386–47405
						ORF26-R	TCCGTGTTGTCTACGTCCAG	47599–47618
ORF25 + 26+27	ORF27	47973–48845	+	48–72 h	L	ORF27-F	CACCACGTTTGGACGCATT	48413–48431
						ORF27-R	TAATCCGTAGGCCTGCCGT	48457–48475
ORF28	ORF28	49091–49399	+	48–72 h	L	ORF28-F	GGAGGAATGGTGGACGGC	49121–49138
						ORF28-R	AAGACCAATCACGGGAGGCT	49164–49183
ORF29	ORF29	54176–54665	−	72 h	L	ORF29-F	AACAATCCAAACAACTGGTGCG	54192–54213
						ORF29-R	TCCAACAGCAGCTAGTAATGCA	54255–54276
ORF30	ORF30	50723–50952	+	48–72 h	L	ORF30-F	GAGCAAGTGGTCGCGGG	50807–50823
						ORF30-R	TTTTGTGACATAGAGAGTCAGCGAG	50849–50873
ORF30 + 31	ORF31	51002–51452	+	48–72 h	L	ORF31-F	TGCGGTATTTGCAGACATGG	51227–51246
						ORF31-R	CCGTCCCCCAGGGCTAT	51274–51290
ORF30 + 31+32	ORF32	51537–52860	+	48–72 h	L	ORF32-F	GAGTCTTGTGGCATGCGTGA	52736–52755
						ORF32-R	CCCCCAGGTAACACAAGCC	52780–52798
ORF30 + 31+32 + 33	ORF33	52910–53865	+	48–72 h	L	ORF33-F	TTAACGCCTGCACCTCTATCTC	52952–52973
						ORF33-R	GCCACAGACACATTGTAAGCAG	53021–53042
ORF34	ORF34	54774–55688	+	24–48 h	E	ORF34-F	ACCCCCTTCCGTTGCTATG	55234–55252
						ORF34-R	ACAGTCGGCCCGACAAAA	55278–55295
ORF34 + 35	ORF35	55738–56075	+	24–48 h	E	ORF35-F	AGGCGGGCCAGAGGTTT	55937–55953
						ORF35-R	GCGGCTGGCGCAAA	55980–55993
ORF34 + 35+36	ORF36	56190–57269	+	24–48 h	E	ORF36-F	CACCGGCAAAGCCCAG	57193–57208
						ORF36-R	TGCTTCTGAAACGCCAGCT	57234–57252
ORF34 + 35+36 + 37	ORF37	57409–58396	+	24–48 h	E	ORF37-F	GAACGTCAAAACTGACACCGAG	58043–58064
						ORF37-R	CTCGGTGTCAGTTTTGACGTTC	58112–58133
ORF34 + 35+36 + 37+38	ORF38	58446–58972	+	24–48 h	E	ORF38-F	GGGAACCGCTCGACGTAGT	58842–58860
						ORF38-R	GCTCAAGCAACATGCCCTTT	58891–58910
ORF39	ORF39	59072–60274	−	24–48 h	E	ORF39-F	CGCCGACGGTCGATAGAA	59227–59244
						ORF39-R	TGGTCTTTGCTGGGAGGG	59182–59199
ORF40	ORF40	60407–61780	+	48–72 h	L	ORF40-F	GGTCTGGTGGCCGTGAATC	61514–61532
						ORF40-R	ACGAGACCCGCGATAATACG	61558–61577
ORF42 + 43	ORF42	62644–63235	−	48–72 h	L	ORF42-F	GACGAAGGCCGCGTCC	62877–62892
						ORF42-R	ATTATTTGTCGCGCCAGAAAG	62916–62936
ORF43	ORF43	63444–64991	−	48–72 h	L	ORF43-F	GGATATGGTGTCCTGAGAATAGGTG	63553–63577
						ORF43-R	GCTGGCTCCCGTTGTTGA	63602–63619
ORF44	ORF44	65052–67357	+	48–72 h	L	ORF44-F	GCCGGTGTCTCAAGAGCTG	66568–66586
						ORF44-R	TGTCCCCCTCCTGCCC	66617–66632
ORF45 + 46+47 + 48	ORF45	67452–68675	−	8 h	IE	ORF45-F	GCTTTGCGGCTTAAGTTTGG	67705–67724
						ORF45-R	CGCCTCCTCTGGTAGCGA	67750–67767
ORF46 + 47+48	ORF46	68963–69503	−	24 h	E	ORF46-F	TGAACCAATCCCAGCCAAG	69002–69020
						ORF46-R	GTTTTGACGGTGGAGAAGGG	69043–69062
ORF47 + 48	ORF47	69611–70014	−	24 h	E	ORF47-F	GCATGTTTCCACGGTAATGTCG	69793–69814
						ORF47-R	TTCTTTGCGGTCCACTCTATCC	69876–69897
ORF48	ORF48	70272–71480	−	NA		ORF48-F	CGGGCAAGCAAGCTGGT	70533–70549
						ORF48-R	CCCTGGCGATTTTGGGTAC	70576–70594
ORF49	ORF49	71729–72637	−	NA		ORF49-F	ACAAAATGGGAGAGGCACCA	72180–72199
						ORF49-R	GCGCCCCTGGAATCAGA	72225–72241
ORF50+K8/KbZIP	K8/KbZIP	74949–75422	+	8–24 h	E	K8/KbZIP-F	GCCAAATGCCCAGAATGAAGGAC	74944–74966
						K8/KbZIP-R	TGGTTGCCCGTTGAGGCTTAGATC	75030–75053
K8.1	K8.1	76014–76254	+	48 h	L	K8.1-F	ACAGATTCGCACAGAAATCCCT	76025–76046
						K8.1-R	CGAACGATACGTGGGACAATTG	76091–76112
ORF52	ORF52	76901–77296	−	48–72 h	L	ORF52-F	GGCACCAGGAGGCGGT	76955–76970
						ORF52-R	TCGCTTAGAATCGACGTCTGC	76994–77014
ORF53	ORF53	77432–77764	−	48–72 h	L	ORF53-F	GCAACGTCATAGAATCCTGGG	77580–77600
						ORF53-R	GCTCAGCGCCAGGCCT	77627–77642
ORF54	ORF54	77835–78722	+	48–72 h	L	ORF54-F	TTGCGCCATAGGAAGCTAGC	78543–78562
						ORF54-R	TCGCGAAAATGCACTCGAG	78586–78604
ORF55	ORF55	78864–79513	−	48–72 h	L	ORF55-F	ACGAATGCATCGCGGAA	78929–78945
						ORF55-R	CGGAGGCAACTTTACCCAAG	78969–78988
ORF56	ORF56	79831–82066	+	48–72 h	L	ORF56-F	TCACTCCCCGGGCCA	81574–81588
						ORF56-R	GCGATCCATGATGCTATAGATGAT	81611–81634
ORF56 + 57	ORF57	82326–83644	+	8 h	IE	ORF57-F	ACGAATCGAGGGACGACG	82350–82367
						ORF57-R	CGGGTTCGGACAATTGCT	82395–82412
K9/vIRF1	K9/vIRF1	83960–85309	−	48–72 h	L	K9/vIRF1-F	ATATCCTCGTCGGTTTTCCCTG	84187–84208
						K9/vIRF1-R	GCGTATGTTATCATAACGGGCC	84264–84285
K10/vIRF4	K10/vIRF4	86174–88442	−	48–72 h	L	K10/vIRF4-F	CCCAACAGGCCAGCTACATAA	86612–86632
						K10/vIRF4-R	CTTCGTGGAACTCTGAGACGC	86657–86677
K10.5/vIRF3	K10.5/vIRF3	89700–90945	−	48–72 h	L	K10.5/vIRF3-F	AATCTGAGCTTGATGACGAGGG	89725–89746
						K10.5/vIRF3-R	TATCGCATACAGGGACATGAGC	89813–89834
K11/vIRF2	K11/vIRF2	92066–93620	−	48–72 h	L	K11/vIRF2-F	ATCCGAGTCATATTCAGGCGA	92120–92140
						K11/vIRF2-R	AATCGAGAACCTGAAGGGTCC	92174–92194
ORF58 + 59	ORF58	94577–95650	−	24 h	E	ORF58-F	GCCAAAGGCAGGAGAACAAA	94818–94837
						ORF58-R	TGTCATGCGTGGGCGTAT	94861–94878
ORF59	ORF59	95700–96658	−	24 h	E	ORF59-F	GCCCACATCCACCGACTTC	96135–96153
						ORF59-R	AGCCAGAAACCAAACCCGTT	96180–96199
ORF60 + 61	ORF60	96976–97893	−	24–48 h	E	ORF60-F	ACTGCCGGTGTATGAGAGGG	97045–97064
						ORF60-R	TCGTTCCTCCCATATTTGGC	97087–97106
ORF61	ORF61	97922–98399	−	24–48 h	E	ORF61-F	CCCATCTTGTTTCATCCCAGA	97997–98017
						ORF61-R	CTGACGGCTCTTCAGTGCC	98044–98062
ORF62	ORF62	100458–101171	−	72 h	L	ORF62-F	CACACGGAGATAAACCAACGAG	100596–100617
						ORF62-R	TTGACCCTCCCTTGTGATATGC	100691–100712
ORF63	ORF63	101495–104100	+	NA		ORF63-F	GCGACTTCGTGCGCGT	103385–103400
						ORF63-R	ATGCGACAGATGTACGTGCG	103427–103446
ORF63 + 64	ORF64	104150–111927	+	NA		ORF64-F	TGAGGTAATAAGGCAGCTGTGG	109946–109967
						ORF64-R	GAAGCCGTGTCGGATTCATC	109992–110011
ORF65 + 66+67	ORF65	112037–112549	−	48–72 h	L	ORF65-F	GCCTGCGACATATTTCCCTG	112434–112453
						ORF65-R	GGAGCGACTGGATCATGACT	112495–112514
ORF66 + 67	ORF66	112651–113799	−	48–72 h	L	ORF66-F	TAGAACAGCCCTGATTACGACG	112750–112771
						ORF66-R	TACTGTGGCAGCGAACATATGA	112818–112839
ORF67	ORF67	113957–114614	−	48–72 h	L	ORF67-F	TTTTGGCAAGGTATACGTCCGA	114251–114272
						ORF67-R	CAACCCTGAAGTGCTTTCTGTG	114322–114343
ORF68	ORF68	115108–116295	+	48–72 h	L	ORF68-F	GTGGTCGCATCCCACGA	115876–115892
						ORF68-R	ATGGACCCTGTGAGGTGTCTG	115919–115939
ORF68 + 69	ORF69	116544–117452	+	48–72 h	L	ORF69-F	TGCAGTGCAGGTACACACCA	117347–117366
						ORF69-R	GCATCTCGTCGGTGCAGTCT	117389–117408
K12	K12	118025–118207	−	Latent	Latent	K12-F	GTTGCAACTCGTGTCCTGAA	118038–118057
						K12-R	AGTTCATGTCCCGGATGTGT	118159–118178
ORF71 + 72	ORF71	122393–122959	−	Latent	Latent	ORF71-F	TTTCCCCTGTTAGCGGAATGT	122783–122803
						ORF71-R	CTAAGTGAAGCAGGTCGCGC	122734–122753
ORF72	ORF72	123042–123815	−	Latent	Latent	ORF72-F	CATTGCCCGCCTCTATTATCA	123224–123244
						ORF72-R	ATGACGTTGGCAGGAACCA	123184–123202
ORF73	ORF73	126456–127446	−	Latent	Latent	ORF73-F	GGCCTTCCTGTAGGACTTGAAA	126983–127004
						ORF73-R	GGACCACGGATACTCATTCTCC	127058–127079
K14	K14	128264–129079	+	24–48 h	E	K14-F	TGGTGGGCCTATTTGGGATA	128979–128998
						K14-R	GATGCACCGCCCTGCTT	129023–129039
ORF74+K14	ORF74	129520–130548	+	24–48 h	E	ORF74-F	CTACTGGACACTCTGCTAAGGC	130333–130354
						ORF74-R	TACTGCCAGACCCACGTTTATC	130392–130413
ORF75+K15	ORF75	130699–134589	−	48–72 h	L	ORF75-F	GAGAACCCCGACAAGGACTG	131405–131424
						ORF75-R	ACACGGGCTTTGAGGTGG	131457–131474
K15	K15	134824–135287	−	NA		K15-F	ATCGGTGACATGGCTAGGTAAC	135079–135100
						K15-R	GGGACCAGAACTTACACCACAA	135170–135191
-	GAPDH		Host			GAPDH-F	TCGCTCTCTGCTCCTCCTGTTC
						GAPDH-R	CGCCCAATACGACCAAATCC

^aUnique CDS were from Bruce et al. (27). The reference sequence was that of KSHV strain GK18 (GenBank accession number NC_0089333).

^bThe timing of expression of each KSHV gene was described by Arias et al. (46). ND, not described; NA, not available.

^cThe KSHV genes were classified as an IE gene when expression was at <8 h, an E gene when expression was at 8 to 48 h, and an L gene when expression was at 48 h to 72 h, as described by Arias et al. (46).

^dThe reference sequence was that of KSHV strain GK18 (GenBank accession number NC_009333).

FIG 3.

ORF66 is required for late gene expression. iSLK-WT and iSLK-ΔORF66 cells were cultured for 72 h in medium with NaB and Dox to induce a lytic state. Total RNA was extracted from the cells and subjected to RT-qPCR. The mRNA expression level of each viral gene was normalized by the level of GAPDH mRNA expression, and columns indicate the fold change in the amount of iSLK-ΔORF66 cell transcripts compared with the amount of iSLK-WT cell transcripts. Classification of the KSHV genes was performed as described by Arias et al. (46). The KSHV genes were classified as an IE gene when expression was at <8 h, an E gene when expression was at 8 to 48 h, and an L gene when expression was at 48 h to 72 h. Expression levels were assessed using three independent samples, and error bars indicate standard deviations.

If ORF66 functions as a vPIC component, we hypothesized that it would be recruited to the TATT box, which is known to be located approximately 30 bp upstream of the transcription start site (TSS) of L genes (28). To demonstrate this, we evaluated whether ORF66 interacts with L gene promoters by a chromatin immunoprecipitation (ChIP) assay. iSLK-ΔORF66 and iVero-ΔORF66 cells stably expressing 3xFLAG-ORF66 were treated with or without NaB and Dox to induce lytic infection, and then cells were subjected to the ChIP assay using anti-FLAG antibody (Fig. 4). As a result, immunoprecipitated ORF66 was found to be bound to the promoters of K8.1 (L gene) but not to the promoters of a latent gene (ORF72), IE gene (ORF16), or E gene (ORF46/47). These results indicate that a protein complex bearing ORF66 may interact with L gene promoters via an interaction between ORF24 and ORF66.

FIG 4.

ORF66 associates with the transcription start site of late genes. iSLK-WT control and iSLK-ΔORF66-3xFLAG-ORF66 cells were treated with or without Dox and NaB for 72 h and subjected to ChIP-qPCR. The 3xFLAG-ORF66 protein was immunoprecipitated by anti-FLAG or control IgG antibody, and precipitates, including chromatin and viral DNA, were subjected to SYBR green real-time PCR for measuring the amount of promoter DNA of ORF72 (a latent gene), ORF16 (an IE gene), ORF46/47 (an E gene), or K8.1 (an L gene). The levels of immunoprecipitated viral promoter were normalized to the total level of input DNA.

Interaction of ORF66 and ORF34 and its function in virus production.

ORF66 was reported to interact with other vPIC components, such as ORF31, ORF18, and ORF34, by us and others (17, 19, 21). We also found that ORF34 operated as the vPIC hub, by bridging ORF24 and vPIC components, including ORF66. Furthermore, these interactions are thought to be important for the functions of vPIC in gene expression. To gain further insight into the interaction between ORF66 and ORF34, we identified the region within ORF66 responsible for binding with ORF34. We made ORF66 truncated mutants, each of which had approximately 60-amino-acid deletions spanning from the N terminus to the C terminus of ORF66 (Fig. 5a). The interaction between the ORF66 mutants and ORF34 was assayed by pulldown experiments (Fig. 5b). 293T cells were cotransfected with 2xS-ORF66 mutants and 6xMyc-ORF34 plasmids, and 2xS-ORF66 was precipitated from cell extracts by beads onto which S protein was immobilized. As a result, truncated mutants ORF66 Δ1 to ORF66 Δ4 interacted with ORF34. However, no interaction with truncated mutants ORF66 Δ5 to ORF66 Δ7 was observed, meaning that the C-terminal region (amino acids 241 to 429) of ORF66 is critical for ORF34 binding. Furthermore, we performed a trans-complementation assay using these ORF66 truncated mutants to assess how the loss of the ORF34-ORF66 interaction affected virus production. ORF66 mutant plasmids were transfected into iSLK-ΔORF66 and iVero-ΔORF66 cells, and the recovery of virus production was measured. Wild-type ORF66 expression significantly increased virus production in iSLK-ΔORF66 and iVero-ΔORF66 cells, while recovery of virus production was not detected by expressing any of the ORF66 truncated mutants (Fig. 5c and d). These results indicate that not only the ORF66 C-terminal domain (i.e., ORF34-binding region) but also the entire structure of ORF66 is indispensable for virus production.

FIG 5.

ORF66 physically interacts with ORF34 via its C-terminal regions, and the entire structure of ORF66 is indispensable for virus production. (a) Schematic representation of tagged ORF34 deletion mutants used in the mapping experiments. The amino acids truncated are indicated to the right of the corresponding mutant. (b) 293T cells were cotransfected with plasmids expressing the 2xS-ORF66 truncated mutant and 6xMyc-ORF34. Transfected cells were lysed, and cell lysates were subjected to pulldown assays using beads onto which S protein was immobilized, which capture 2xS-ORF66. The precipitates obtained, including those of the 2xS-ORF66 truncated mutants, were probed with the indicated antibodies to detect interactions. WCE, whole-cell extract. (c, d) iSLK-ΔORF66 (c) or iVero-ΔORF66 (d) cells were transfected with plasmids expressing the control, ORF66, or the truncated ORF66 mutant. Transfected cells were stimulated for 3 days. Progeny KSHV was purified from the harvested culture supernatant, and the KSHV genome was quantified by real-time PCR. Viral productivity was assessed using four independent samples, and error bars indicate standard deviations. Transfected virus-producing cells were lysed and subjected to Western blotting to confirm ORF66 mutant expression.

An amino acid sequence alignment of the ORF66 C-terminal domain of KSHV and other herpesvirus homologs is depicted in Fig. 6a. Conserved amino acids are indicated with a gray background. To identify the key residues in the ORF66 C-terminal region (CR) for interaction with ORF34, we constructed block alanine-scanning mutants (the CR1 mutant [CR1 mut] to CR9 mut), where several neighboring conserved amino acids were replaced by alanine (Fig. 6a), with the exception of proline, to avoid disruption of the whole protein structure. These mutants were subjected to pulldown assays to investigate the association of ORF34. The results showed that ORF66 CR2, CR3, CR4, and CR6 mut bound to ORF34, whereas ORF66 CR1, CR5, CR7, CR8, and CR9 mut did not (Fig. 6b). Therefore, we focused on the specific amino acid sequences of the mutants that lost the ability to bind to ORF34. In particular, CR1, CR5, and CR9 mut had mutations in the C-X-X-C consensus sequence. On the other hand, CR7 mut contained a single cysteine residue and CR8 mut contained three leucine residues.

FIG 6.

Several conserved residues of ORF66 are essential for the physical association with ORF34. (a) Amino acid sequence alignment of the C terminus of ORF66 (amino acids 241 to 429). Herpesvirus homolog amino acid sequences were translated from nucleotide sequences found in the NCBI database (KSHV ORF66 [JSC-1-BAC16; GenBank accession number GQ994935], MHV68 ORF66 [strain WUMS; GenBank accession number NC_001826], BHV4 ORF65 [strain V; GenBank accession number JN133502], EBV BFRF2 [strain B95-8; GenBank accession number V01555], CMV UL49 [strain Towne; GenBank accession number FJ616285], HHV6 U33 [strain japan-a1; GenBank accession number KY239023]). Raw data for the alignment were obtained by using the Clustal Omega program (EMBL-EBI; https://www.ebi.ac.uk/Tools/msa/clustalo/). Amino acids completely conserved between homologs are indicated by a gray background. Based on this information, several conserved amino acids were split into mutants with blocks of alanine-scanning ORF66 mutations (CR1 mut, ORF66 C295A/C298A/G299A; CR2 mut, ORF66 C301A/L302A/N303A/G305A; CR3 mut, ORF66 F314A/F320A; CR4 mut, ORF66 R323A/D324A/E327A/K328A; CR5 mut, ORF66 C341A/S342A/C344A/G345A; CR6 mut, ORF66 V371A/N375A; CR7 mut, ORF66 C393A; CR8 mut, ORF66 L412A/L413A/L415A; CR9 mut, ORF66 C424A/C427A). (b) The mutants with block alanine-scanning mutations were cotransfected with plasmids expressing 6xMyc-ORF34. Cell lysates were subjected to pulldown assays using beads onto which S protein was immobilized.

To obtain further insight into the effects of ORF34 binding via CR1 to CR9 mut of ORF66 on virus production, the ORF66 alanine substitution mutants (CR1 to CR9 mut) were subjected to a trans-complementation assay using iSLK-ΔORF66 and iVero-ΔORF66 cells (Fig. 7a and b). Compared with ORF66 wild-type expression, virus production in both iSLK-ΔORF66 and iVero-ΔORF66 cells could not be recovered by the expression of CR1, CR5, CR7, CR8, and CR9 mut, which failed to interact with ORF34 (Fig. 6b). On the other hand, viral expression for CR2, CR3, CR4, and CR6 mut, which maintained an interaction with ORF34, showed various levels of recovery, indicated by the red columns in Fig. 7a and b. CR3 and CR6 mut showed recovery of virus production to levels almost identical to those for the ORF66 wild-type in iVero-ΔORF66 cells and partial recovery in iSLK-ΔORF66 cells. The recovery of virus production by the CR4 mut was significant; however, it was less than that for CR3 and CR6 mut. CR2 mut had no effect. These data reveal that the binding between ORF66 and ORF34 via the conserved amino acids of the CR1, CR5, CR7, CR8, and CR9 regions in ORF66 is necessary but not sufficient for virus replication.

FIG 7.

An association between ORF66 and ORF34 is necessary but not sufficient for virus production. (a, b) iSLK-ΔORF66 (a) or iVero-ΔORF66 (b) cells were transfected with plasmids expressing the control, ORF66, or ORF66 block alanine-scanning mutant. The progeny KSHV were purified, and the KSHV genome was quantified. Viral productivities were assessed using three independent samples, and error bars indicate standard deviations. The color of each bar indicates the following: black, control; gray, ORF66 WT; blue, ORF66 alanine-scanning mutants not binding to ORF34; red, ORF66 alanine-scanning mutants binding to ORF34. Transfected virus-producing cells were lysed and subjected to Western blotting for the confirmation of ORF66 mutant expression. (c to e) Stable iSLK-Δ66-control, iSLK-Δ66-66WT, and iSLK-Δ66-CR mutant cell lines were also quantified for the levels of progeny KSHV (c), viral protein expression (d), and mRNA expression (e).

To elucidate the detailed effects of CR1 to CR9 mut of ORF66 on virus production and viral gene expression, we established 10 iSLK-Δ66 KSHV cell lines which stably expressed 3xFLAG-ORF66, the CR1 to CR9 mutants (mut) or wild type. These cells were treated with Dox and NaB, and virus production (Fig. 7c), viral protein expression (Fig. 7d), and mRNA expression (Fig. 7e) were analyzed. As expected, lysis-induced iSLK-Δ66 KSHV cells transfected with ORF66 CR3, CR4, and CR6 mut and wild-type ORF66 displayed stable high levels of expression of the K8.1 protein, K8.1 mRNA, and virus production (Fig. 7c to e). The ORF66 CR2, CR3, CR4, and CR6 mut could interact with ORF34 (Fig. 6). In addition, cells stably expressing ORF66 CR1, CR2, CR5, and CR9 mut lacked K8.1 transcriptional activity. The ORF66 CR2 mut interacted with ORF34 but did not rescue virus production. The transcription of the K5 (IE) and ORF59 (E) genes was not influenced in lytic-induced iSLK-Δ66 KSHV cells in the presence or absence of ORF66 mut and wild-type ORF66 expression (Fig. 7e). These results also indicate that the binding of ORF34 and ORF66 is necessary not only for virus production but also for late gene expression.

C-X-X-C sequences and zinc binding are important for the association with ORF34.

Based on our results obtained with the ORF66 alanine-scanning mutants, we found that the CR1, CR5, and CR9 regions of ORF66 contain a C-X-X-C sequence which is critical for ORF34 binding (Fig. 6a). Therefore, we generated expression plasmids with single amino acid mutants ORF66 C295A, C298A, C341A, C344A, C424A, and C427A, in which a cysteine in C-X-X-C was replaced by an alanine. Plasmids harboring the ORF66 G299A, S342A, G345A, and G345A mutants, in which the conserved glycine or serine around the C-X-X-C sequence was replaced by alanine, were also generated. Plasmids harboring the 6xMyc-ORF34 and ORF66 alanine mutants were cotransfected into the cells, and cell extracts were subjected to affinity purification using beads onto which S protein was immobilized, followed by Western blotting. As a result, the ORF66 C295A, C298A, C341A, C344A, C424A, and C427A mutants failed to interact with ORF34 (Fig. 8a, b, and d). Furthermore, two leucine residues at amino acids 412 and 413 in CR8 of ORF66 were also important for the association with ORF34 (Fig. 8c). These results suggest that a conserved leucine repeat (412L and 413L) in CR8 and three conserved C-X-X-C sequences in CR1, CR5, and CR9 are needed for binding to ORF34, leading to the appropriate formation of vPIC. Because the C-X-X-C motif in proteins is often related to binding to a bivalent cation, we speculate that the C-X-X-C sequence contributes to the formation of a higher-order structure, such as a zinc finger domain. Next, to elucidate whether zinc ions are necessary for the binding of ORF66 and ORF34, we performed pulldown assays using zinc chelator N,N,N′,N′-tetrakis (2-pyridylmethyl) ethylenediamine (TPEN) and 2xS-ORF66-conjugated beads onto which S protein was immobilized. The ORF66-conjugated beads were mixed with the cell extracts including overexpressed Myc-tagged ORF34 in the presence of TPEN and were pulled down. The precipitates were subjected to Western blotting with anti-Myc. As a result, TPEN decreased the interaction of ORF66 with ORF34, indicating that zinc ions are necessary for binding between ORF34 and ORF66 (Fig. 8e).

FIG 8.

Identification of individual amino acids of ORF66 responsible for binding to ORF34. The ORF66 wild type, block alanine-scanning mutants, and single alanine-scanning mutants were cotransfected with plasmids expressing 6xMyc-ORF34. Cell lysates were subjected to pulldown assays using beads onto which S protein was immobilized. (a to d) The blots show the association between ORF34 and CR1 mutants (CR1 mut; ORF66 C295A/C298A/G299A, ORF66 C295A, ORF66 C298A, ORF66 G299A) (a), CR5 mutants (CR5 mut; ORF66 C341A/S342A/C344A/G345A, ORF66 C341A, ORF66 S342A, ORF66 C344A, ORF66 G345A) (b), CR8 mutants (CR8 mut; ORF66 L412A/L413A/L415A, ORF66 L412A, ORF66 L413A, ORF66 L415A) (c), and CR9 mutants (CR9 mut; ORF66 C424A/C427A, ORF66 C424A, ORF66 C427A) (d). (e) Zinc ion chelation influences the ability of ORF66 to bind to ORF34. The pulldown assay used S-tagged ORF66 binding beads in the presence of the zinc chelator TPEN.

DISCUSSION

The C terminus of ORF66 is much more highly conserved among herpesvirus homologs than the N terminus (Fig. 6a). We prepared alanine-scanning mutants of ORF66 for amino acid residues that are conserved among MHV, bovine herpesvirus (BHV), EBV, human CMV (HCMV), and HHV6. The conserved amino acid regions of CR1, CR5, CR7, CR8, and CR9 in the ORF66 C-terminal domain are necessary for the binding between ORF66 and ORF34 and for virus production. In addition, three conserved C-X-X-C sequences in ORF66, CR1, CR5, and CR9, are needed for binding to ORF34 (Fig. 8). Our results indicate that ORF66 leads to appropriate vPIC formation by the interaction with ORF34 via the C-X-X-C sequences in the C-terminal domain of ORF66. We speculate that the conserved C-X-X-C sequences could form a higher-order structure, such as a zinc finger domain. Therefore, we approached the protein structure and functional prediction of ORF66 by the use of I-TASSER (iterative threading assembly refinement) server-based helical protein structure simulation (29–31) and generated a full-length homology model of ORF66 (Fig. 9). According to a meta-server approach to protein-ligand binding site prediction (COACH) (32, 33), four cysteine residues, C295, C298, C341, and C344, are predicted to associate with a zinc ion. Based on the location of each cysteine residue in the homology model, a pair of cysteines, C295 and C298 in CR1 and CR2 or C341 and C344 in CR5, might chelate a single molecule of zinc, which correlates with our experimental observation using a zinc chelator, TPEN (Fig. 8). The zinc chelator TPEN inhibited the association between ORF66 and ORF34. This result indicates that ORF66 binds zinc, which is important for its interaction with ORF34. We also performed a homology search of ORF66 by use of the SWISS model. Interestingly, amino acids 319 to 348 (including the CR5 region) of ORF66 have low homology with the TFIIB zinc ribbon domain of the hyperthermophilic archaeon Pyrococcus furiosus. As these zinc-associated cysteine residues were mostly located inside the protein structure, except for C341 (Fig. 9, middle), it is implied that the CR1, CR2, and CR5 residues are involved in maintaining the functional structure formation of ORF66 via zinc interaction.

FIG 9.

Protein structure and functional prediction of ORF66. Structural models of ORF66 simulated by use of the I-TASSER server, which completed protein structural and functional predictions, are shown. (Left) The cartoon model indicates the position of residues responsible for interaction with ORF34. (Middle) The surface model represents the residues exposed on the protein surface. (Right) The hydrophobicity of each residue is shown by a red color gradient, followed by the Eisenberg hydrophobicity scale. The predicted zinc ion-binding cysteines (C295 and C298, C341 and C344) are highlighted in brown. C344 in CR5 is highlighted in magenta. The LLQL in CR8 and C341 in CR5 are shown in cyan and green, respectively.

Two leucine residues (L412 and L413) in LLQL of CR8 are essential for binding between ORF66 and ORF34 (Fig. 6b and 8c), suggesting that the hydrophobicity of a leucine-rich sequence is also indispensable for ORF66 to be functional. In fact, the L412 and L413 residues in CR8 are predicted to be fully exposed on the protein surface, which represented a high degree of hydrophobicity in our structural model (Fig. 9). Two cysteine residues (C424 and C427) in CR9 were not predicted to be a zinc-binding domain by our protein-ligand binding prediction; however, these residues in ORF66 were also responsible for the interaction with ORF34 (Fig. 6). These residues are exposed on the protein surface (Fig. 9, middle) and are located at the hydrophobic surface, including the LLQL motif (Fig. 9, right). Thus, the hydrophobic region surrounding two leucine residues in CR8 and cysteine residues in CR9 is expected to make up the binding surface and to be directly involved in the hydrophobic interaction with ORF34.

We evaluated viral replication in KSHV ΔORF66-producing iVero and iSLK cells, into which an ORF66-deficient KSHV BAC clone was stably integrated. ORF66 plays a critical role in virus production and the transcription of L genes. KSHV ORF24 binds to ORF34, RNA Pol II, and the TATT box of the transcription start site (TSS) of the L gene (17, 19, 20). Because a direct or indirect interaction of ORF34 with ORF18, -30, -31, and -66 was indicated by coimmunoprecipitation and split luciferase experiments (17, 19, 20), ORF34 has been thought to function as a hub for interactions between ORF24 and other vPIC components. Therefore, these reports, in addition to our viral replication kinetics and viral gene expression data, imply that ORF66 engages in L gene expression as a vPIC component. These results are in line with those for other herpesvirus homologs of KSHV ORF66. For instance, EBV BFRF2 is essential for virus production and contributes to vPIC formation (34). HCMV UL49 is also essential for replication in human foreskin fibroblasts (35, 36).

To gain a better understanding of the role of ORF66 within the vPIC complex, we attempted to search for ORF34-binding regions within ORF66. Pulldown assays using truncated ORF66 mutants showed that the C-terminal region (from amino acids 241 to 429 at the C-terminal end) of ORF66 is responsible for binding to ORF34 (Fig. 5b). The truncated ORF66 mutants were subjected to a trans-complementation assay. However, none of the truncated ORF66 mutants rescued virus production in either the iVero or the iSLK KSHV-ΔORF66-transfected cell line (Fig. 5c and d). Therefore, we speculate that the entire structure of ORF66 is necessary for virus production through vPIC formation. Considering the complex structure of human PIC components (TBP and GTFs) and RNA Pol II (37), vPIC might be a crowded complex that consists of not only vPIC factors but also host proteins, such as RNA Pol II, RNA Pol II binding proteins, and other unknown host proteins. A lack of large regions of ORF66 may also influence the overall structure of the protein and affect the interaction with its binding partners. Another possibility is that the N-terminal and center regions of ORF66 are related to the interaction with other vPIC components or unknown host factors.

To evaluate the physiological function of conserved amino acids, block alanine-scanning mutants of ORF66 were subjected to a trans-complementation assay. The ORF66 mutants (CR1, CR5, CR7, CR8, and CR9 mut) failed to associate with ORF34 (Fig. 6) and could not rescue virus production in KSHV-ΔORF66-transfected cell lines (Fig. 7). In contrast, some ORF66 mutants (CR3 and CR6 mut), which associate with ORF34, showed full or partial rescue activities, while CR2 mut did not. Presumably, the conserved sequence (CLNXG) in the CR2 region is related to binding to other vPIC-associated factors. ORF66 rescued activity in the iSLK-Δ66 and iVero-Δ66 cell lines; however, the recovery rates were lower in iSLK-Δ66 cells. The differences in the virus recovery rates in both cell lines may be due to the KSHV production potential and/or the characteristics of each cell line. The efficiency of KSHV production in iSLK cells was 100-fold higher than that in iVero cells. Furthermore, iVero cells are derived from nonhuman primates (African green monkeys). Slight differences in mutated site structures and/or species differences in host factors may influence vPIC formation and the accumulation of other host factors, resulting in differences in recovery rates. Altogether, an association between ORF66 and ORF34 is necessary but not sufficient for virus production.

Our results show the importance and molecular machinery of ORF66 as a vPIC component in viral replication and L gene expression. Herpesvirus vPICs consist of viral factors as well as RNA Pol II and several host factors that are engaged in a complex. In murine CMV, RNA helicase and cellular factors related to splicing and translation interact with vPIC (38). The regulation of vPIC occurs through the posttranscriptional modification of vPIC components, such as phosphorylation, which is known to contribute to the physiological functions of vPIC in EBV (39). Controlling the dynamics of vPIC in viral replication, vPIC-targeted promoters on the KSHV genome have more complexity than was previously estimated. They are inextricably linked to genome DNA replication (28). Our efforts to unveil the vPIC machinery help to shed light on why beta- and gammaherpesviruses have incorporated the vPIC machinery into their genomes for survival.

MATERIALS AND METHODS

Plasmids.

The pCI-neo-3xFLAG-ORF66, pCI-neo-3xFLAG-ORF66, and pCI-neo-6xMyc-ORF34 expression plasmids were previously described (19). Truncation and alanine mutant ORF66-coding fragments were obtained by PCR or overlap extension PCR from the ORF66 expression plasmids using the primer sets noted in Table 2 and were cloned into the pCI-neo-2xS and pCI-neo-3xFLAG vectors, respectively. For pCI-blast plasmid construction, fragments coding a blasticidin S resistance (Bla^r) gene were obtained by PCR from pLKO.1-blast (Addgene plasmid 26655, a kind gift from Keith Mostov [40]) using the primer sets noted in Table 2 and were replaced by a neomycin resistance cassette (Neo^r) in the pCI-neo mammalian expression vector (Promega, WI, USA). The pCI-neo-3xFLAG, pCI-neo-3xFLAG-ORF66, and pCI-neo-3xFLAG-ORF66 CR mutants were digested at the NheI and NotI sites, and the protein-coding fragments were inserted into the multiple-cloning site of pCI-blast to construct the pCI-blast-3xFLAG, pCI-blast-3xFLAG-ORF66, and pCI-blast-3xFLAG-ORF66 CR mutants.

TABLE 2.

Primers for BAC mutagenesis, construction of expression plasmids, and ChIP-qPCR

Primer purpose and name	Primer sequence (5′ → 3′)
BAC mutagenesisa
S_dORF66_3stop_EP	ttttgtcatattcggggagcggggtttccagggaaacatgTAGTTAGATAGTccctcctgggccaggcacctTAGGGATAACAGGGTAATCGATTT
As_dORF66_3stop_EP	ccagcgacgggcggtccaacaggtgcctggcccaggagggACTATCTAACTAcatgtttccctggaaaccccGCCAGTGTTACAACCAATTAACC
Cloning pCI-blast plasmidb
S_StuI-BlaR-BstBI	gcctaggcctaggcttttgcaaaaagcttgattcttctgacacaacagtctcgaacttaaggctagagccaccatggccaagcctttgtctc
As_StuI-BlaR-BstBI	ggggttcgaaccccagagtcccgcttagccctcccacacataac
Cloning expression plasmidb
S_XbaI_ORF66	cattctagaATGGCCCTGGATCAGCGCTGGGATC
As_ORF66_NotI	gagcggccgcTCAGGAGGAACACTTCCC
ORF66 truncated mutant constructionc
S_XbaI_d1(2-60)_ORF66	catctagaATGctgttggaccgcccgtcgctg
S_d2(61-120)_ORF66	GGCCAGGCACtgggcaaaatacctgtcgc
As_d2(61-120)_ORF66	attttgcccaGTGCCTGGCCCAGGAGGG
S_d3(121-180)_ORF66	CACGTAcccacccgtcgcggtgtgg
As_d3(121-180)_ORF66	gacgggtgggTACGTGGCGCGGTATCG
S_d4(181-240)_ORF66	CCCTCGctgcccgcctgcgagcag
As_d4(181-240)_ORF66	ggcgggcagCGAGGGCACCTCCAG
S_d5(241-300)_ORF66	CCACGTAGTGtgtcttaactttggcaggggcaag
As_d5(241-300)_ORF66	agttaagacaCACTACGTGGCGGGACTTAATAAGGCTC
S_d6(301-360)_ORF66	GTGTGGACACgggaccctagcacgcgtc
As_d6(301-360)_ORF66	ctagggtcccGTGTCCACACTCCATGCAC
As_d7(361-429)_ORF66_NotI	aagcggccgctcaGCGTCCGGTAATATCGC
ORF66 block alanine-scanning mutant constructionc
S_CR1mut_ORF66	GCAatggagGCTGCAcactgtcttaactttggcag
As_CR1mut_ORF66	TGCAGCctccatTGCcacaaccgccc
S_CR2mut_ORF66	GCTGCTGCAtttGCCaggggcaagtttcatac
As_CR2mut_ORF66	GGCaaaTGCAGCAGCgtgtccacactccatg
S_CR3mut_ORF66	GCTcatactgtcaatGCTcctcccaccaacgtgtttttc
As_CR3mut_ORF66	AGCattgacagtatgAGCcttgcccctgccaaagttaag
S_CR4mut_ORF66	GCTGCAaggaaaGCAGCTcagttcaccatctgtgc
As_CR4mut_ORF66	AGCTGCtttcctTGCAGCgctgaaaaacacgttg
S_CR5mut_ORF66	GCTGCTtacGCTGCAagcgaacatatgagggtgtatc
As_CR5mut_ORF66	TGCAGCgtaAGCAGCgtagatcctccccgtg
S_CR6mut_ORF66	GCTctagctaacGCAgcggcccttgccattc
As_CR6mut_ORF66	TGCgttagctagAGCagccctgattacgacgcgtg
S_CR7mut_ORF66	cctGCCcttgggacgcccgactg
As_CR7mut_ORF66	gtcccaagGGCaggcactacaaaactgacagtttg
S_CR8mut_ORF66	GCAGCTcagGCAacctcacagctgctgg
As_CR8mut_ORF66	TGCctgAGCTGCtccgcgcacgtcac
As_CR9mut_ORF66_NotI	tagcggccgctcaggaggaAGCcttcccTGCacagaac
ORF66 single alanine-scanning mutant constructionc
S_CR1_C295A_ORF66	GCAatggagtgtggacactgtcttaactttggcag
As_CR1_C295A_ORF66	tccacactccatTGCcacaaccgccc
S_CR1_C298A_ORF66	tgcatggagGCTggacactgtcttaactttggcag
As_CR1_C298A_ORF66	tccAGCctccatgcacacaaccgccc
S_CR1_G299A_ORF66	tgcatggagtgtGCAcactgtcttaactttggcag
As_CR1_G299A_ORF66	TGCacactccatgcacacaaccgccc
S_CR5_C341A_ORF66	GCTtcttactgtggcagcgaacatatgagggtgtatc
As_CR5_C341A_ORF66	gccacagtaagaAGCgtagatcctccccgtg
S_CR5_S342A_ORF66	tgtGCTtactgtggcagcgaacatatgagggtgtatc
As_CR5_S342A_ORF66	gccacagtaAGCacagtagatcctccccgtg
S_CR5_C344A_ORF66	tgttcttacGCTggcagcgaacatatgagggtgtatc
As_CR5_C344A_ORF66	gccAGCgtaagaacagtagatcctccccgtg
S_CR5_G345A_ORF66	tgttcttactgtGCAagcgaacatatgagggtgtatc
As_CR5_G345A_ORF66	TGCacagtaagaacagtagatcctccccgtg
S_CR8_L412A_ORF66	GCActtcagctcacctcacagctgctgg
As_CR8_L412A_ORF66	gagctgaagTGCtccgcgcacgtcac
S_CR8_L413A_ORF66	ctgGCTcagctcacctcacagctgctgg
As_CR8_L413A_ORF66	gagctgAGCcagtccgcgcacgtcac
As2_CR8_L415A_ORF66_NotI	tagcggccgctcaggaggaacacttcccgcaacagaactccagcagctgtgaggtTGCctgaagcagtccgcgcacg
As_CR9_C424A_ORF66_NotI	tagcggccgctcaggaggaacacttcccTGCacagaac
As_CR9_C427A_ORF66_NotI	tagcggccgctcaggaggaAGCcttcccgcaacagaac
ChIP qPCR
ChIP-qPCR_ORF72-F	GGCGGGCCATTTGTACTTTC
ChIP-qPCR_ORF72-R	ATCTCAGGCCTTCCAGTTTG
ChIP-qPCR_ORF16-F	GACGGCAAGGTTTTTATCCC
ChIP-qPCR_ORF16-R	CGCAAGTCAAGACACAAGTC
ChIP-qPCR_ORF46/47-F	AGCCCCCTTCCGTAATATCTG
ChIP-qPCR_ORF46/47-R	TTTTCCGCGGAAGTATGTCG
ChIP-qPCR_K8.1-F	ACTCCCACCATGTTGAAGCTTG
ChIP-qPCR_K8.1-R	GGGATTTCTGTGCGAATCTGTG

^aLowercase indicates a sequence with homology to the KSHV BAC16 sequence, underlined uppercase indicates mutagenesis site, and uppercase indicates the pEP-KanS sequence.

^bUnderlined lowercase indicates a restriction enzyme site.

^cUnderlined lowercase indicates a restriction enzyme site, and underlined uppercase indicates a mutagenesis site.

Mutagenesis of KSHV BAC16.

KSHV BAC16 was a kind gift from Jae U. Jung, and mutagenesis of KSHV BAC16 was performed as described in previous publications (24, 25). The primer sequences used for mutagenesis are noted in Table 2. Insertion and deletion of kanamycin resistance cassettes (Kan^r) in each mutant were analyzed by digestion of EcoRV and agarose gel electrophoresis. The mutated sites of each BAC clone were confirmed by Sanger sequencing.

Establishment of doxycycline-inducible recombinant KSHV-expressing cells and stably ORF66-expressing cells.

For maintenance, iSLK cells were cultured in growth medium containing 1 μg/ml of puromycin (InvivoGen, CA, USA) and 0.25 mg/ml of G418 (Nacalai Tesque, Kyoto, Japan). iVero cells (19) were cultured in a growth medium containing 2.5 μg/ml of puromycin. BAC16 with wild-type KSHV (WT-BAC16) and BAC16 with mutant KSHV (ΔORF66-BAC16) were transfected into iSLK and iVero cells. The iSLK and iVero cells were transfected by a calcium phosphate method and a lipofection method, respectively. The transfected cells were selected under 1,000 μg/ml of hygromycin B (Wako, Osaka, Japan) and 2.5 μg/ml of puromycin to establish doxycycline-inducible recombinant KSHV-producing cell lines (iSLK-WT, iSLK- ORF66, iVero-WT, and iVero-ΔORF66 cells).

To establish stable ORF66-expressing cells, pCI-blast-3xFLAG-ORF66 and the empty vector (pCI-blast-3xFLAG) were transfected into iSLK-WT or iSLK-ΔORF66 cells, and the transfected cells were selected in 10 μg/ml of blasticidin S (InvivoGen) and maintained in 7.5 μg/ml of blasticidin S. Thus, the stable cell lines iSLK-WT-pCI-blast-3xFLAG, iSLK-ΔORF66-pCI-blast-3xFLAG, and iSLK-ΔORF66-pCI-blast-3xFLAG-ORF66 were established. To establish stable ORF66 CR mutant-expressing cells, pCI-blast-3xFLAG-ORF66 CR mutants were transfected into iSLK-ΔORF66 cells, and the transfected cells were selected in blasticidin S. iSLK-ΔORF66-pCI-blast-3xFLAG-ORF66 CR mutant stable cell lines were established as described above. To establish iVero-WT-pCI-neo-3xFLAG, iVero-ΔORF66-pCI-neo-3xFLAG, and iVero-ΔORF66-pCI-neo-3xFLAG-ORF66 stable cell lines, iSLK cells harboring each KSHV BAC clone were selected and maintained in 1.5 mg/ml of G418.

Measurement of virus production and viral DNA replication.

For quantification of virus production, KSHV virions in culture supernatant were quantified as previously described (19, 41, 42). Briefly, iSLK and iVero cells (iSLK-WT, iSLK-ΔORF66, iVero-WT, or iVero-ΔORF66 cells) were treated with sodium butyrate (NaB) and doxycycline (for iSLK cells, 0.75 mM NaB and 4 μg/ml Dox; for iVero cells, 1.5 mM NaB and 8 μg/ml Dox) for 72 h to induce lytic replication and the production of recombinant KSHV, and the culture supernatants were harvested. The culture supernatants (220 μl) were treated with DNase I (New England Biolabs, MA, USA) to obtain only enveloped and encapsidated viral genomes. Viral DNA was purified and extracted from 200 μl of the DNase I-treated culture supernatant using a QIAamp DNA blood minikit (Qiagen, CA, USA). To quantify the viral DNA copies, a SYBR green real-time PCR was performed using KSHV-encoded ORF11-specific primers.

For measurement of KSHV genome replication, each KSHV-producing cell line was treated with doxycycline and NaB for 48 h to induce lytic replication and harvested. Total cellular DNA containing the KSHV genome was purified and extracted from washed cells using the QIAamp DNA blood minikit (Qiagen). The number of cellular KSHV genome copies was determined by SYBR green real-time PCR and normalized to the total number of DNA copies.

Recovery of an exogenous gene in BAC-harboring cells.

iSLK cells were transfected with pCI-neo-3xFLAG as a control plasmid and plasmids harboring 3xFLAG-tagged full-length ORF66, truncated mutant ORF66, and alanine mutant ORF66 using the Screenfect A Plus reagent (Wako, Tokyo, Japan) according to the manufacturer’s instructions and simultaneously stimulated with medium containing 0.75 mM NaB and 4 μg/ml Dox. The iVero cells were transfected using polyethyleneimine (molecular weight, 40,000; PEI Max; Polysciences, Inc., Warrington, PA, USA) (43). After 1 day, transfected iVero cells were stimulated with medium containing 0.5 mM NaB and 8 μg/ml Dox. After 3 days of stimulation, the viral supernatant was harvested and the KSHV genome was evaluated by real-time PCR.

RT-qPCR array and RT-qPCR.

For the comparison of WT KSHV and ΔORF66-KSHV, mRNA was extracted from iSLK-WT and iSLK-ΔORF66 cells that had been treated with Dox and NaB using a FastGene RNA premium kit (Nippon Genetics Co. Ltd., Tokyo, Japan). cDNA was synthesized by use of a ReverTra Ace RT-qPCR kit (Toyobo, Osaka, Japan) and subjected to SYBR green real-time PCR. Transcript expression was analyzed by qPCR using specific primers based on those from the work of Fakhari and Dittmer (26) and modified to fit the unique CDS referred to by Bruce et al. (27), as described in Table 1. Relative KSHV mRNA expression levels were determined by GAPDH (glyceraldehyde-3-phosphate dehydrogenase) expression and ΔΔC_T threshold cycle (C_T) methods. For comparison of the ORF66 wild type and CR mutants, mRNA was extracted from iSLK-Δ66-control, iSLK-Δ66-66WT, and iSLK-Δ66-CR mutant stable cell lines using the RNAiso Plus reagent (TaKaRa, Osaka, Japan). cDNA was synthesized by use of ReverTra Ace qPCR RT master mix with genomic DNA remover (Toyobo) and subjected to SYBR green real-time PCR using the primers described in Table 1. Relative KSHV mRNA expression levels were determined by GAPDH expression and ΔΔC_T methods.

ChIP assay.

The chromatin immunoprecipitation (ChIP) assay was performed as described previously (44) with slight modifications. Briefly, iSLK-WT-control and iSLK-ΔORF66-3xFLAG-ORF66 cells were treated with or without 4 μg/ml of Dox and 0.75 mM NaB for 72 h. Formaldehyde-fixed cells were lysed by the use of Farnham lysis buffer {5 mM PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)], pH 8.0, 85 mM KCl, 0.5% NP-40}, and the nuclear pellet was collected. The pellet was lysed in SDS lysis buffer and sonicated. The supernatant containing DNA was diluted with ChIP dilution buffer and then subjected to immunoprecipitation with an anti-FLAG (DDDDK tag) monoclonal antibody (FLA-1; MBL, Nagoya, Japan) or mouse control IgG (Santa Cruz, CA, USA). Immunoprecipitates containing chromatin and viral DNA were subjected to SYBR green real-time PCR for measuring the amount of promoter DNA of each gene. The amount of immunoprecipitated viral DNA was normalized to 1% of the amount of input DNA. The sequences of the qPCR primer sets for each transcription start site of the open reading frames (ORFs) are noted in Table 2.

Pulldown assay, Western blotting, and antibodies.

Western blot analyses were performed as described previously (19). For pulldown assays, transfected 293T cells (RCB2202; Riken Bio Resource Center, Tsukuba, Japan) were lysed with HNTG buffer (20 mM HEPES [pH 7.9], 0.18 M NaCl, 0.1% NP-40, 0.1 mM EDTA, 10% glycerol) with protease inhibitors and sonicated. The cell extracts were subjected to affinity purification using beads onto which S protein was immobilized (Novagen, MA, USA), and the purified proteins (containing 2xS-tagged ORF66 or mutants) were subjected to Western blotting.

For zinc chelator TPEN [N,N,N′,N′-tetrakis (2-pyridylmethyl) ethylenediamine; TCI, Tokyo, Japan] treatment, 3xFLAG-taggged ORF66 overexpressed in 293T cells was purified with beads onto which S protein was immobilized in the presence of each dose of TPEN or the vehicle (ethanol) for 2 h. The beads were mixed with the lysate of 293T cells in which 6xMyc-ORF66 was overexpressed in the presence of each dose of TPEN or the vehicle (ethanol) for 2 h. The beads were washed 4 times and subjected to Western blotting.

Anti-Myc antibody (9E10; Santa Cruz, CA, USA), anti-S-tag polyclonal antibody (MBL, Nagoya, Japan), anti-FLAG (DDDDK-tag) antibody (FLA-1; MBL), anti-HHV8 K-bZIP antibody (F33P1; Santa Cruz), anti-HHV8 K8.1A/B antibody (4A4; Santa-Cruz), and anti-actin antibody (AC-15; Santa Cruz) were used as the primary antibodies. Horseradish peroxidase (HRP)-linked anti-mouse IgG antibody (GE Healthcare UK Ltd., Buckinghamshire, UK) or HRP-linked anti-rabbit IgG antibody (GE Healthcare UK Ltd.) was used as the secondary antibody. Anti-FLAG-HRP antibody (M2; Sigma-Aldrich, MO, USA) was also used.

Homology modeling.

The template structure for ORF66 was initially identified by collecting high-scoring structural templates from a local meta-threading server to generate a three-dimensional (3D) structural model. The protein-ligand predictions were then derived by threading the 3D models through a protein function database, BioLiP. All procedures were automatically processed on the I-TASSER server program provided by the Y. Zhang lab (https://zhanglab.ccmb.med.umich.edu/I-TASSER/) (29–33). The visualization of the homology model was performed by use of the molecular visualization open-source software PyMOL. The hydrophobicity in each amino acid was determined by use of the Eisenberg hydrophobicity scale (https://web.expasy.org/protscale/pscale/Hphob.Eisenberg.html) (45) and visualized by running a color_h.py script in PyMOL.

Statistics.

A two-tailed Student's t test was used to indicate the differences between the groups. P values are shown in each figure.