Auxin-inducible Degron (AID) to Dissect Kaposi’s Sarcoma associated Herpesvirus (KSHV) LANA protein function

Protein knock-down with an inducible degradation system is a powerful tool to study proteins of interest in living cells. Here, we adopted the auxin-inducible degron (AID) approach to detail Kaposi’s Sarcoma-associated herpesvirus (KSHV) latency-associated nuclear antigen (LANA) function in latency maintenance and inducible viral lytic gene expression. We fused the mini-AID (mAID) tag at the LANA N-terminus with KSHV BAC16 recombination, and iSLK cells were stably infected with the recombinant KSHV encoding mAID-tagged LANA. Incubation with 5-phenyl-indole-3-acetic acid (5-Ph-IAA), a derivative of natural auxin, rapidly degraded LANA protein within 1.5 hours. In contrast to our hypothesis, depletion of LANA not only failed to trigger lytic reactivation but rather decreased inducible lytic gene expression when we triggered reactivation with a combination of ORF50 protein expression and sodium butyrate treatment. Decreased overall lytic gene induction seemed to associate with a rapid loss of KSHV genomes in the absence of LANA. Furthermore, we found that small cell fractions harbor non-depletable LANA dots in the presence of 5-Ph-IAA. In the cell population containing degradation-resistant LANA, induction of lytic reactivation was strongly attenuated. These results suggest that (i) there are at least two populations of LANA dots in cells, (ii) local nuclear environment and its epigenetic effects on the episomes are heritable to daughter cells; this biological had substantial effects in degree of KSHV reactivation, and finally (iii) LANA may have an additional function in protecting KSHV episomes from degradation. IMPORTANCE KSHV LANA protein plays a wide variety of roles in latency maintenance and lytic gene expression. We adapted the inducible protein knockdown approach to examine its role directly, and revealed that there are cell populations that possess viral episomes insensitive to reactivation stimuli. Viral reactivation is known to be highly heterogenic, and our observations suggest that LANA tethering sites on host chromatin may play a critical role in determining diverse responsiveness to the stimuli. We also demonstrated that depletion of LANA leads to rapid reduction of viral genome, which suggests that LANA might be actively protecting latent viral genome from degradation. These results add novel insights into the role of LANA in latency maintenance and regulation of lytic reactivation.

auxin or its derivatives, and polyubiquitinates AID-tagged (or mAID-tagged) target 150 protein which leads to proteasome-mediated degradation. We next verified the 151 virological function of mAID-tagged LANA. We first confirmed that iSLK-OsTIR1-mAID-152 LANA BAC16 cells produced amounts of virions in culture supernatant comparable to 153 those by iSLK.219 cells (Fig. 1B). We also examined the infectivity of progeny virions 154 produced from iSLK-OsTIR1-mAID-LANA BAC16 cells. For that, iSLK cells were 155 infected with purified mAID-LANA or BAC16 WT virions at a multiplicity of infection 156 (MOI) of 10, and GFP-positive iSLK cells were quantified by flow cytometry. The GFP-157 positive (infection) ratio was comparable with that of BAC16 WT virus (Fig. 1C). 158 Furthermore, we confirmed that KSHV episomes were maintained during cell passage. 159 These results suggested that the tagging with the 68-amino acid residue mAID at the N-160 terminus did not interfere with LANA function. concentrates LANA proteins on the viral genome that makes them visible as "LANA-172 dots" or "LANA-speckles" (42) as shown in Fig. 1E. Thus, a LANA-dot also indicates a 173 single viral episome in the nucleus. As expected, LANA-dots signal were drastically 174 reduced at 24 h after addition of 5-Ph-IAA (Figs. 1E and 1F). However, the GFP 175 fluorescence signal was also decreased at 24 h after addition of 5-Ph-IAA (Fig. 1E), 176 suggesting that depletion of LANA presumably decreased viral episome copies within 177 the cell. Accordingly, we next determined KSHV genomic copy number per cell, and 178 normalizing to the host genome. The results showed that the depletion of LANA protein 179 indeed induced rapid reduction of viral genomic DNA to approximately 20% of control within 24 h (Fig. 1G). The results were somewhat surprising since iSLK cells are 181 unlikely to divide twice within 24 hours in which case prevention of episome tethering 182 would dilute the viral genome to 25%, or ¼, of that present in the cells prior to LANA 183 depletion. We speculated that active viral episome degradation might be occurring in 184 the absence of LANA in the cells. approximately 30-40% of cells expressed early gene products such as ORF6 (40). To 194 our surprise, depletion of LANA itself did not induce lytic reactivation (the third row in Fig.  195 2A), but rather inhibited expression of K8α (the bottom row in Fig. 2A). Viral lytic gene 196 expression was also confirmed by Western blotting (Fig. 2B) and real time-qPCR (Fig. 197 2C). Expression of K8α was still seen at 1.5, 3, and 6 h after addition of 5-Ph-IAA 198 compared to the control when cells were reactivated. Consistent with the 199 immunofluorescence experiment shown in Fig. 2A, K8α expression was largely inhibited 200 at the 24 h time point (Fig. 2B). mRNA expression for lytic genes was reduced to less 201 than one-fourth to one-tenth of control levels at 24 h in reactivated, LANA-depleted cells 202 (Fig. 2C). On the other hand, two highly inducible non-coding RNAs (PAN RNA and 203 T1.5) showed increased, leaking expression in the absence of LANA (Fig. 2C, 5-Ph-IAA 204 (+) Dox/NaB (-)), suggesting that gene silencing effects were restricted to selective 205 genomic loci. In order to evaluate the effects of reduced DNA copy number on gene 206 transcription, we next calculated mRNA expression level per viral genome copy (Fig.  207   2D). The results demonstrated that KSHV genomes in the absence of LANA did not 208 alter inducible lytic gene expression in that after normalization, the transcription rate 209 was comparable to that in non-5-Ph-IAA treated cells. 210

LANA dots 213
During our study, we noticed small fractions of cells harboring non-depletable 214 LANA dots even in the presence of 5-Ph-IAA for more than 3 days. We decided to 215 isolate and characterize these cells harboring non-depletable LANA-dots (i.e., LANA cells was also very low and similar to that from parental cells treated with 5-Ph-249 IAA for 24 hours. These results suggested that there are nuclear spaces that are 250 presumably unreachable for the SCF-OsTIR1 E3 ubiquitin ligase. KSHV episomes 251 tethered within such a nuclear environment are still able to express LANA protein, but 252 the episomes become insensitive to reactivation signals and this phenotype is heritable 253 to the daughter cells. 254

DISCUSSION 256
In the present study, we applied the AID system for studying the role of KSHV 257 LANA protein in latency maintenance. By integrating an mAID-tag into KSHV BAC16 258 using a recombination-based method, we demonstrated the utility of this approach to 259 study the function of a KSHV viral protein in infected cells. RNA interference (RNAi)-260 mediated gene silencing has been utilized to study the function of specific 261 genes/proteins of interest in various organisms in the past two decades. Some genes, 262 however, cannot be knocked down because they are essential for cell viability. In 263 addition, the RNAi approach relies on the turnover (half-life) of the target protein for 264 efficient reduction in target protein levels. For silencing of most proteins, RNAi is usually 265 effective only after 48-72 h of transfection (43). On the other hand, the inducible protein 266 degradation approach allows us to quickly deplete the target protein (37, 38). Indeed, 267 we showed that mAID-tagged LANA protein was successfully depleted within 1.5 h after 268 addition of 5-Ph-IAA. We believe that rapid depletion of target protein enables us to 269 identify the biological effects more directly. After learning that mAID-tagged LANA can 270 be depleted rapidly and efficiently, we tagged several other viral proteins with mAID 271 using the same recombination approach. We found that, however, K-Rta (ORF50) could from ATP and GTP that binds to its adaptor protein named Stimulator of Interferon 289 Genes (STING). A recent study demonstrated that LANA binds to cGAS, and inhibits 290 the cGAS-STING pathway and thereby antagonizes the cGAS-STING-mediated 291 restriction of KSHV lytic replication (48). Our finding suggests that LANA may also 292 actively inhibit cGAS and protect the viral genome from the host cellular defense 293 system(s) during latency. Our previous study also suggests that LANA may be involved 294 in formation of 3D episome structure through tight binding with TRs and formation of 295 genomic loops with selected genomic loci in unique region. Rapid loss of LANA may 296 disrupt such 3D genomic structure to expose naked DNA to induce DNA damage and/or 297 recombination. Further study will be necessary to clarify how LANA plays a role in 298 protecting latent viral episomes from degradation. Nonetheless, this study suggested an 299 additional reason for which therapeutic targeting of LANA is an attractive approach for 300

KS-associated diseases. 301
Previous studies have revealed that LANA plays an important role in maintaining 302 latency by repressing the transcriptional activity of the K-Rta promoter (49, 50). K-Rta is 303 a key transcriptional regulator that controls the switch from latency to lytic replication, 304 and is sufficient to drive the completion of the viral lytic gene cascade (8). LANA 305 downregulates K-Rta's promoter activity in transient reporter assays, thus repressing K-306 Rta-mediated transactivation (51). It also was shown that LANA physically interacts with 307 K-Rta in vitro, and associates with K-Rta in KSHV-infected cells (51). In addition, a 308 study showed that KSHV can undergo spontaneous lytic reactivation, and K-Rta 309 transcription level was increased when LANA expression was knocked down with 310 specific siRNA (52). These results establish a model in which LANA is actively 311 suppressing viral lytic replication by antagonizing the functions of K-Rta. In the present 312 study, however, we showed that the depletion of LANA itself did not induce lytic gene L-Gln at 37°C with air containing 5% carbon dioxide. iSLK cells were maintained in 367 DMEM supplemented with 10% FBS, 1x Pen-Strep-L-Gln, and 2 μg/ml puromycin at 368 37°C with air containing 5% carbon dioxide. iSLK.219 cells were maintained in DMEM 369 supplemented with 10% FBS, 10 μg/ml puromycin, 400 μg/ml hygromycin B, and 370 250 μg/ml G418. iSLK cells harboring BAC16 WT were maintained in DMEM 371 supplemented with 10% FBS, 1x Pen-Strep-L-Gln, 1,000 μg/ml hygromycin B, and 2 372 μg/ml puromycin at 37°C with air containing 5% carbon dioxide. 373 374

Construction of mAID-LANA KSHV BAC16 375
Recombinant KSHV was prepared by following a protocol for en passant 376 mutagenesis with a two-step markerless red recombination technique (39). Briefly, 377 mAID-coding sequence was first synthesized (IDT DNA gBlock) and cloned into a 378 pBluescript SK vector. The pEPkan-S plasmid was also used as a source of the 379 kanamycin cassette, which includes an I-SceI restriction enzyme site at the 5' end of the 380 kanamycin resistance gene-coding region. The kanamycin cassette was amplified with 381 primer pairs listed in Table I. An amplified kanamycin cassette was then cloned into the 382 mAID-coding region. The resulting plasmid was used as a template for another round of 383 PCR to prepare a transfer DNA fragment for markerless recombination with BAC16. 384 Recombinant BAC16 clones with insertion and also deletion of the kanamycin cassette 385 in the BAC16 genome were confirmed by colony PCR with appropriate primer pairs. 386 The recombination junction and adjacent genomic regions were amplified by PCR, and 387 the resulting PCR fragments were directly sequenced with the same primers to confirm 388 in-frame insertion into the BAC DNA. Two independent BAC clones were generated as 389 biological replicates. BAC DNA was extracted from E. coli using the NucleoBond Xtra 390  The cells were then treated with 1 μg/ml of doxycycline and 1.5 mM sodium butyrate for 446 another 96 hours together with 5-Ph-IAA. Two hundred μl of cell culture supernatant 447 was treated with 12 μg/ml DNase I for 15 min at room temperature to degrade DNAs 448 that were not correctly encapsidated. The reaction was stopped by the addition of 5 mM 449 EDTA, followed by heating at 70°C for 15 minutes. Viral DNA was then purified using a 450 QIAamp DNA Mini kit according to the manufacturer's protocol. Five μl of eluate was 451 used for real-time qPCR to determine viral genomic DNA copy number. 452 453

Statistical analysis 454
Results are shown as means ± S.E.M. from at least three independent 455 experiments. Data were analyzed using unpaired Student's t-test. An FDR-corrected p 456 value less than 0.05 was considered statistically significant. 457