The SMC5/6 Complex Represses the Replicative Program of High-Risk Human Papillomavirus Type 31

The multi-subunit structural maintenance of chromosomes (SMC) 5/6 complex includes SMC6 and non-SMC element (NSE)3. SMC5/6 is essential for homologous recombination DNA repair and functions as an antiviral factor during hepatitis B (HBV) and herpes simplex-1 (HSV-1) viral infections. Intriguingly, SMC5/6 has been found to associate with high-risk human papillomavirus (HPV) E2 regulatory proteins, but the functions of this interaction and its role during HPV infection remain unclear. Here, we further characterize SMC5/6 interactions with HPV-31 E2 and its role in the HPV life cycle. Co-immunoprecipitation (co-IP) revealed that SMC6 interactions with HPV-31 E2 require the E2 transactivation domain, implying that SMC5/6 interacts with full-length E2. Using chromatin immunoprecipitation, we found that SMC6 is present on HPV-31 episomes at E2 binding sites. The depletion of SMC6 and NSE3 increased viral replication and transcription in keratinocytes maintaining episomal HPV-31, indicating that SMC5/6 restricts the viral replicative program. SMC6 interactions with E2 were reduced in the presence of HPV-31 E1, suggesting that SMC6 and E1 compete for E2 binding. Our findings demonstrate SMC5/6 functions as a repressor of the viral replicative program and this may involve inhibiting the initiation of viral replication.


Introduction
Human papillomaviruses (HPVs) are non-enveloped, circular double-stranded DNA viruses that infect basal keratinocytes of stratified squamous epithelia. Infection with certain "high-risk" HPV types (e.g., HPV-16 and -31) causes nearly all cervical cancers and an increasing number of oropharyngeal cancers [1]. While prophylactic vaccinations have reduced the incidence of HPV infection and attributable cancers, there is currently no cure for pre-existing HPV infection. As such, HPV remains a global health threat and a better understanding of HPV biology remains of significant medical importance for the identification of novel therapeutic targets.
The HPV genome is roughly eight kilobases in length, encoding for eight proteins essential for the viral life cycle that are classified as early (E1-E7) or late (L1 and L2), based upon their temporal expression during HPV infection. The HPV replicative program can be divided into three hypothetical phases: Establishment, maintenance, and amplification. During the initial establishment of infection, HPV E1 and E2 proteins, together with the host replicative machinery, replicate HPV genomes to low copy numbers in host cells. Following establishment, the HPV genome enters the maintenance phase, whereby viral genome copy numbers are maintained and replication occurs in synchrony with host cells. In the outer layers of the epithelium, the viral replicative program transitions from maintenance to amplification, whereupon viral genomes are replicated to thousands of copies per (C) FLAG-tagged SMC6 and HPV-31 V5-E2 were transfected into HEK293 cells. Forty-eight hours later, immunoprecipitation was performed using mouse M2 anti-FLAG agarose beads, and immunoblotting was then performed using mouse M2 anti-FLAG, mouse anti-V5, and rabbit anti-NSE3 (positive control) antibodies. (D) FLAG-tagged SMC6 and HPV-31 V5-E2 were transfected into HEK293TT cells. Forty-eight hours later, immunoprecipitation was performed using rabbit anti-V5 antibodies bound protein A/G sepharose beads, and immunoblotting was then performed using mouse M2 anti-FLAG, rabbit anti-SMC6, rabbit anti-V5, and rabbit anti-NSE3 antibodies. (E) Schematic of FLAG-tagged HR-HPV-31 full-length E2 wt (top) and truncated HPV-31 E2 DBD (bottom). (F) FLAG-SMC6, full-length FLAG-tagged HPV-31 E2, and FLAG-tagged HPV-31 truncated E2 were transfected into HEK293TT cells. Forty-eight hours later, immunoprecipitation was performed using either rabbit IgG or rabbit anti-SMC6, and immunoblotting was then performed using mouse M2 anti-FLAG, rabbit anti-SMC6, and rabbit anti-NSE3 antibodies. and HPV-31 V5-E2 were transfected into HEK293 cells. 48 hrs later immunoprecipitation was performed using rabbit anti-V5 antibodies bound protein A/G sepharose beads, then immunoblotting performed using mouse M2 anti-FLAG, rabbit anti-SMC6, rabbit anti-V5 and rabbit anti-NSE3 (positive control). (B) FLAG-tagged SMC6 and HPV-31 V5-E2 were transfected into HEK293 cells. 48 hrs later immunoprecipitation was performed using mouse M2 anti-FLAG agarose beads, then immunoblotting performed using mouse M2 anti-FLAG, mouse anti-SMC6 and rabbit anti-NSE3 (positive control). (C) FLAG-tagged SMC6 and HPV-31 V5-E2 were transfected into HEK293 cells. Forty-eight hours later, immunoprecipitation was performed using mouse M2 anti-FLAG agarose beads, and immunoblotting was then performed using mouse M2 anti-FLAG, mouse anti-V5, and rabbit anti-NSE3 (positive control) antibodies. (D) FLAG-tagged SMC6 and HPV-31 V5-E2 were transfected into HEK293TT cells. Forty-eight hours later, immunoprecipitation was performed using rabbit anti-V5 antibodies bound protein A/G sepharose beads, and immunoblotting was then performed using mouse M2 anti-FLAG, rabbit anti-SMC6, rabbit anti-V5, and rabbit anti-NSE3 antibodies. (E) Schematic of FLAG-tagged HR-HPV-31 full-length E2 wt (top) and truncated HPV-31 E2 DBD (bottom). (F) FLAG-SMC6, full-length FLAG-tagged HPV-31 E2, and FLAG-tagged HPV-31 truncated E2 were transfected into HEK293TT cells. Forty-eight hours later, immunoprecipitation was performed using either rabbit IgG or rabbit anti-SMC6, and immunoblotting was then performed using mouse M2 anti-FLAG, rabbit anti-SMC6, and rabbit anti-NSE3 antibodies.
Previously, SMC5/6 has been shown to inhibit the transcription of HBV [8][9][10][11] and HSV-1 [11]. We questioned whether these antiviral functions of SMC6 and its interactions with HPV E2 indicated that SMC5/6 functions as a repressor of HPV-31 replication or transcription. We began investigating this using the siRNA knockdown of SMC6 and NSE3 in CIN612-9E cells ( Figure 2A)-a cervical keratinocyte cell line derived from a patient CIN I lesion that maintains low copy numbers of episomal HPV-31 [15]. SMC6-and NSE3-depleted cells displayed an increased HPV-31 DNA content ( Figure 2B) and viral E8ˆE2 transcript abundance ( Figure 2C). SMC6 knockdown using an alternative siRNA also resulted in increased HPV-31 transcripts in CIN612-9E cells (Supplementary Figure S1). We performed an analogous experiment in episomal maintaining NIKS/HPV-31 keratinocytes. Similar results were found, as the depletion of either SMC6 or NSE3 resulted in an increased HPV-31 DNA content in NIKS/HPV-31 ( Figure 2D,E); however, only NSE3 knockdown increased E8ˆE2 viral transcripts ( Figure 2F). Taken together, these data show that the transient depletion of SMC6 and NSE3 in CIN612-9E and NIKS/HPV-31 results in increased HPV-31 replication and transcription.
Pathogens 2020, 9, x FOR PEER REVIEW 4 of 14 Previously, SMC5/6 has been shown to inhibit the transcription of HBV [8][9][10][11] and HSV-1 [11]. We questioned whether these antiviral functions of SMC6 and its interactions with HPV E2 indicated that SMC5/6 functions as a repressor of HPV-31 replication or transcription. We began investigating this using the siRNA knockdown of SMC6 and NSE3 in CIN612-9E cells ( Figure 2A)-a cervical keratinocyte cell line derived from a patient CIN I lesion that maintains low copy numbers of episomal HPV-31 [15]. SMC6-and NSE3-depleted cells displayed an increased HPV-31 DNA content ( Figure 2B) and viral E8^E2 transcript abundance ( Figure 2C). SMC6 knockdown using an alternative siRNA also resulted in increased HPV-31 transcripts in CIN612-9E cells (Supplementary Figure S1). We performed an analogous experiment in episomal maintaining NIKS/HPV-31 keratinocytes. Similar results were found, as the depletion of either SMC6 or NSE3 resulted in an increased HPV-31 DNA content in NIKS/HPV-31 ( Figure 2D,E); however, only NSE3 knockdown increased E8^E2 viral transcripts ( Figure 2F). Taken together, these data show that the transient depletion of SMC6 and NSE3 in CIN612-9E and NIKS/HPV-31 results in increased HPV-31 replication and transcription. SMC6 has been shown to associate with nuclear episomal HBV cccDNA [8] and HSV-1 DNA [11], but how SMC5/6 localizes to viral DNA remains unclear. An SMC6 association with HPV genomes has not been reported. Similar to HBV cccDNA and HSV-1, HPV genomes are maintained  SMC6 has been shown to associate with nuclear episomal HBV cccDNA [8] and HSV-1 DNA [11], but how SMC5/6 localizes to viral DNA remains unclear. An SMC6 association with HPV genomes has not been reported. Similar to HBV cccDNA and HSV-1, HPV genomes are maintained as episomes adjacent to promyelocytic leukemia (PML) nuclear bodies [16], which are known sites of SMC5/6 localization [10]. The association of SMC6 with viral DNA and its interaction with HPV E2 led us to question whether SMC6 is present on episomal HPV-31 genomes. To answer this, we first validated our ability to immunoprecipitate SMC6 in episomal HPV-31 maintaining CIN612-9E cells to determine the suitability of the cell line for subsequent chromatin immunoprecipitation (ChIP) experiments. Total protein was isolated from CIN612-9E cells and SMC6 immunoprecipitation (IP) was performed, followed by immunoblotting using the anti-SMC6 antibody. SMC6 was detectable in input samples and IP samples where the anti-SMC6 antibody was used, but not negative control rabbit IgG ( Figure 3A). SMC5/6 is present on DNA during the S phase of the cell cycle to facilitate DNA damage repair and restart stalled replication forks [17][18][19]. CIN612-9E cells were arrested in the S phase by a double-thymidine block to enrich for SMC6's association with DNA, and ChIP was then performed using HPV-31 E2, SMC6, or rabbit IgG antibodies. HPV-31 E2 served as our positive control, as it is well-established that E2 binds episomal HPV genomes and our lab has previously demonstrated that E2 binds to ChIP at various sites along the long control region (LCR) of HPV-31 episomes [2,3]. ChIP and qPCR determined that HPV-31 E2 was at E2 binding sites within the HPV-31 LCR, but not the viral L1 region ( Figure 3C). SMC6 was also not present in the HPV-31 L1 region, but was detected near E2 binding sites located towards the LCR 5 region and near the viral ori and early promoter ( Figure 3D). E2 is present on cellular DNA [20][21][22] and reportedly tethers HPV episomes to host common fragile sites [23]. We questioned whether SMC6 is associated with a cellular replication origin (GM-CSF) where E2 may be found. HPV-31 E2, but not SMC6, was present on the eukaryotic GM-CSF origin of replication ( Figure 3B). The absence of SMC6 at the GM-CSF origin led us to utilize the GM-CSF locus as a negative control for SMC6 binding and to evaluate the enrichment of SMC6 on viral DNA. These data show that SMC6 is present on HPV-31 episomes at E2 binding sites located within the LCR in the 5 region and near the viral ori and early promoter, but not the flanking L1 region.
HPV replication depends on E2-mediated recruitment of the viral E1 DNA helicase to the viral origin of replication, during which a transient E1/E2-ori ternary complex is formed prior to the release of E2 from the complex and viral genome [24][25][26][27]. The association of SMC6 with E2 and HPV-31 genomes, as well as its repression of viral replication, led us to speculate that SMC6 may interact with HPV-31 E1/E2 replicative complexes. To explore this, HEK293TT cells were transfected with FLAG-SMC6, HPV-31 V5-E2, and HPV-31 HA-E1 plasmids alone or in combination, and co-IP reactions were performed using the anti-SMC6 antibody. We observed HPV-31 FLAG-E2 and NSE3 co-IP with endogenous SMC6 ( Figure 4A), but not HPV-31 HA-E1, when E1 was expressed alone or co-expressed with E2, or when FLAG-SMC6 was co-transfected with E1 and E2. The co-expression of E1 with FLAG-SMC6 and E2 reduced the amount of E2 pulled-down by SMC6. To confirm that SMC6 does not co-IP HPV-31 E1/E2 complexes or E1 alone, we performed an analogous experiment using the anti-HPV-31 E1 antibody for IP. The IP of HPV-31 HA-E1 pulled-down V5-E2, but not SMC6 or NSE3 ( Figure 4B). The E1 co-IP of E2 was not influenced by the overexpression of FLAG-SMC6. Together, these data demonstrate that neither SMC6 nor NSE3 associate with HPV-31 E1 or E1/E2 complexes, but that E1 expression inhibits the SMC6 pulldown of E2.
1 Figure 3. SMC6 associates with episomal HPV-31 genomes in CIN612-9E cells. (A) Total protein was isolated from CIN612-9E cells and immunoprecipitation was performed overnight using either rabbit anti-SMC6 or control rabbit IgG antibodies, and immunoblotting was then conducted using the rabbit anti-SMC6 antibody. (B) CIN612-9E cells were synchronized in the S phase by a double-thymidine block and chromatin immunoprecipitation (ChIP) was performed using rabbit anti-HPV-31 E2, rabbit anti-SMC6, or control rabbit IgG antibodies. qPCR was conducted using primers to the cellular replication origin (GM-CSF) of replication with primers listed in Table 1. (C) As in (B), but examining E2's presence in the viral L1 region, at the 5 LCR E2 binding site, and at the E2 binding site proximal to the ori and early promoter (pE) of the HPV-31 genome. Approximate locations on the viral genome and corresponding primers are shown below their respective sites on the horizontal axis. (D) As in (C), with SMC6. C t values were normalized to the input with control IgG set to 1 and values are expressed as the mean fold change over IgG +/− SEM. ChIP was performed at least four independent times for all locations examined and statistical significance was determined by a one-tailed Student's t-test compared to control IgG, with * p ≤ 0.05 and ** p ≤ 0.005.
does not co-IP HPV-31 E1/E2 complexes or E1 alone, we performed an analogous experiment using the anti-HPV-31 E1 antibody for IP. The IP of HPV-31 HA-E1 pulled-down V5-E2, but not SMC6 or NSE3 ( Figure 4B). The E1 co-IP of E2 was not influenced by the overexpression of FLAG-SMC6. Together, these data demonstrate that neither SMC6 nor NSE3 associate with HPV-31 E1 or E1/E2 complexes, but that E1 expression inhibits the SMC6 pulldown of E2. FLAG-E2, and HPV-31 HA-E1 were transfected into HEK293TT cells. Forty-eight hours later, immunoprecipitation was performed using either rabbit IgG or rabbit anti-SMC6, and immunoblotting was then performed using anti-SMC6, rabbit anti-NSE3, mouse M2 anti-FLAG, or rat anti-HPV-31 E1 antibodies. (B) FLAG-SMC6, HPV-31 V5-E2, and HPV-31 HA-E1 were transfected into HEK293TT cells. Forty-eight hours later, immunoprecipitation was performed using rat anti-HPV-31 E1, and immunoblotting was then performed using anti-SMC6, rabbit anti-NSE3, mouse M2 anti-FLAG, rabbit anti-V5, or rat anti-HPV-31 E1 antibodies. All experiments were performed three independent times. HPV E2 proteins have been found to influence the host factor stability and transcriptional activity [28,29]. Having validated E2/SMC6 interactions led us to question whether E2 influenced the SMC6 abundance. To explore this, we generated HaCaT cells stably expressing HPV-31 FLAG-E2 or a control vector by G418 selection ( Figure 5A). We observed no change in SMC6 protein levels in response to HPV-31 E2 expression. To confirm this, we examined the SMC6 protein expression in  FLAG-E2, and HPV-31 HA-E1 were transfected into HEK293TT cells. Forty-eight hours later, immunoprecipitation was performed using either rabbit IgG or rabbit anti-SMC6, and immunoblotting was then performed using anti-SMC6, rabbit anti-NSE3, mouse M2 anti-FLAG, or rat anti-HPV-31 E1 antibodies. (B) FLAG-SMC6, HPV-31 V5-E2, and HPV-31 HA-E1 were transfected into HEK293TT cells. Forty-eight hours later, immunoprecipitation was performed using rat anti-HPV-31 E1, and immunoblotting was then performed using anti-SMC6, rabbit anti-NSE3, mouse M2 anti-FLAG, rabbit anti-V5, or rat anti-HPV-31 E1 antibodies. All experiments were performed three independent times. HPV E2 proteins have been found to influence the host factor stability and transcriptional activity [28,29]. Having validated E2/SMC6 interactions led us to question whether E2 influenced the SMC6 abundance. To explore this, we generated HaCaT cells stably expressing HPV-31 FLAG-E2 or a control vector by G418 selection ( Figure 5A). We observed no change in SMC6 protein levels in response to HPV-31 E2 expression. To confirm this, we examined the SMC6 protein expression in parental N/TERT cells and N/TERT/E2 cells stably expressing the HPV-16 E2 protein ( Figure 5B). Similarly, no effect on SMC6 levels was found in N/TERT cells with HPV-16 E2 compared to control cells.
Pathogens 2020, 9, x FOR PEER REVIEW 8 of 14 parental N/TERT cells and N/TERT/E2 cells stably expressing the HPV-16 E2 protein ( Figure 5B). Similarly, no effect on SMC6 levels was found in N/TERT cells with HPV-16 E2 compared to control cells.

Discussion
In this study, we investigated SMC6 interactions with HPV proteins and explored the role of SMC5/6 during the maintenance phase of the HPV-31 life cycle. We observed SMC6 co-IP HPV-31 E2 and further identified that this interaction requires the E2 transactivation domain. While we cannot rule out a contribution of the E2 DNA-binding domain, these data infer that SMC6 interacts with fulllength HPV-31 E2. Importantly, we found that the knockdown of SMC6 and NSE3 results in increased HPV-31 replication and transcription, implying that SMC5/6 represses these processes. We had speculated that the SMC5/6-mediated repression of HPV replication may occur through interactions of SMC6 with E1/E2 complexes as a mechanism utilized by E2 to prevent over-replication of viral genomes. This is unlikely as we did not co-immunoprecipitate SMC6 or NSE3 subunits with either HPV-31 E1/E2 complexes or E1. The association of SMC6 with E2 appeared to be reduced in the presence of E1 and together, these data imply that SMC6/E2 and E1/E2 complexes are exclusive. This suggests that SMC6 repression of the viral replicative program may involve inhibiting E2 recruitment of E1 to viral genomes for the initiation of replication. Our observations that viral replication increased in cells following the depletion of SMC6 and NSE3 support this explanation, as more E2 would be available to recruit E1 to viral genomes for initiating replication. Together, our study supports a role for the SMC5/6 complex as an endogenous antiviral factor and repressor of the HPV-31 replicative program. SMC5/6 associates with host chromatin to mediate homologous recombination DNA repair, rescue stalled replication forks [30], and topologically entrap extrachromosomal DNA [31]. Additionally, SMC5/6 has been shown to associate with hepatitis B [8,11] and herpes simplex-1 [11] viral episomes. The purported localization of SMC6 to viral DNA and its association with E2 led us to speculate that SMC6 may be present on episomal HPV-31 genomes. Indeed, we found SMC6 on episomal HPV-31 genomes at E2 binding sites located towards the LCR 5′ region and proximal to the viral ori and early promoter, but not the L1 region or GM-CSF origin of replication. These data demonstrate that SMC6 localization on the viral genome is specific to the LCR, inferring that SMC5/6 repressor functions may occur at the viral genome. Mechanistically, the repression of HPV-31 replication and transcription could involve SMC5/6 influencing the accessibility of viral DNA through steric hindrance, the recruitment of histone modifying proteins, or interactions with E2 at the viral genome. It is also possible that E2 recruits SMC5/6 to viral episomes to regulate transcription and replication, as this is an established function of E2 [32].

Discussion
In this study, we investigated SMC6 interactions with HPV proteins and explored the role of SMC5/6 during the maintenance phase of the HPV-31 life cycle. We observed SMC6 co-IP HPV-31 E2 and further identified that this interaction requires the E2 transactivation domain. While we cannot rule out a contribution of the E2 DNA-binding domain, these data infer that SMC6 interacts with full-length HPV-31 E2. Importantly, we found that the knockdown of SMC6 and NSE3 results in increased HPV-31 replication and transcription, implying that SMC5/6 represses these processes. We had speculated that the SMC5/6-mediated repression of HPV replication may occur through interactions of SMC6 with E1/E2 complexes as a mechanism utilized by E2 to prevent over-replication of viral genomes. This is unlikely as we did not co-immunoprecipitate SMC6 or NSE3 subunits with either HPV-31 E1/E2 complexes or E1. The association of SMC6 with E2 appeared to be reduced in the presence of E1 and together, these data imply that SMC6/E2 and E1/E2 complexes are exclusive. This suggests that SMC6 repression of the viral replicative program may involve inhibiting E2 recruitment of E1 to viral genomes for the initiation of replication. Our observations that viral replication increased in cells following the depletion of SMC6 and NSE3 support this explanation, as more E2 would be available to recruit E1 to viral genomes for initiating replication. Together, our study supports a role for the SMC5/6 complex as an endogenous antiviral factor and repressor of the HPV-31 replicative program. SMC5/6 associates with host chromatin to mediate homologous recombination DNA repair, rescue stalled replication forks [30], and topologically entrap extrachromosomal DNA [31]. Additionally, SMC5/6 has been shown to associate with hepatitis B [8,11] and herpes simplex-1 [11] viral episomes. The purported localization of SMC6 to viral DNA and its association with E2 led us to speculate that SMC6 may be present on episomal HPV-31 genomes. Indeed, we found SMC6 on episomal HPV-31 genomes at E2 binding sites located towards the LCR 5 region and proximal to the viral ori and early promoter, but not the L1 region or GM-CSF origin of replication. These data demonstrate that SMC6 localization on the viral genome is specific to the LCR, inferring that SMC5/6 repressor functions may occur at the viral genome. Mechanistically, the repression of HPV-31 replication and transcription could involve SMC5/6 influencing the accessibility of viral DNA through steric hindrance, the recruitment of histone modifying proteins, or interactions with E2 at the viral genome. It is also possible that E2 recruits SMC5/6 to viral episomes to regulate transcription and replication, as this is an established function of E2 [32].
The absence of NSE3 in the complex with E2 suggests that E2 may alter SMC5/6 complex formation by excluding NSE3. While this possibility accounts for the observed absence of NSE3, we believe it to be unlikely. SMC5/6 is comprised of a core SMC5/6 heterodimer and several non-SMC elements