CRISPR-mediated activation of endogenous BST-2/tetherin expression inhibits wild-type HIV-1 production

The CRISPR technology not only can knock out target genes by using the RNA-guided Cas9 nuclease but also can activate their expression when a nuclease-deficient Cas9 (dCas9) is employed. Using the latter function, we here show the effect of the CRISPR-mediated pinpoint activation of endogenous expression of BST-2 (also known as tetherin), a virus restriction factor with a broad antiviral spectrum. Single-guide RNA (sgRNA) sequences targeting the BST-2 promoter were selected by promoter assays. Potential sgRNAs and dCas9 fused to the VP64 transactivation domain, along with an accessory transcriptional activator complex, were introduced into cells by lentiviral transduction. Increased expression of BST-2 mRNA in transduced cells was confirmed by real-time RT-PCR. Cells in which BST-2 expression was highly enhanced showed the effective inhibition of HIV-1 production and replication even in the presence of the viral antagonist Vpu against BST-2. These findings confirm that the physiological stoichiometry between host restriction factors and viral antagonists may determine the outcome of the battle with viruses.

methods using a catalytically inactive version of the clustered, regularly interspaced, short palindromic repeats (CRISPR)-associated protein 9 (Cas9)-based RNA-guided activation system 36 . This modified CRISPR-Cas9 system allows us to specifically target and activate an endogenous promoter of a gene of interest, with a nuclease-deficient mutant of Cas9 (dCas9) that binds to a single guide RNA (sgRNA) by recruiting transcription factor complex(es). By using this system, we herein attempt to enhance the endogenous expression of BST-2 to observe whether the CRISPR-based pinpoint activation of BST-2 expression is able to confer cells with inhibitory activity on HIV-1 virion production, even in the presence of the viral antagonist Vpu that counteracts BST-2 by binding and downregulating the restriction factor 37 .

Results
In a CRISPR-based RNA-guided activation system, the sgRNA targets, which are 20 nucleotides in addition to the PAM sequence (NGG), should be located within 200 base pairs of the transcription start sites of the promoter 36 , which are dispersed in the BST-2 promoter region 38 . We selected six different sgRNA targets against the BST-2 promoter (sgBST2#1-6) (Fig. 1A). To identify the best sgRNA target sequences in the BST-2 promoter, we cloned its promoter into a firefly luciferase reporter plasmid and cotransfected HeLa cells with this promoter-indicator plasmid together with plasmids expressing sgRNA appended to the phage MS2 RNA stem loop (sgRNA-MS2 loop), as well as dCas9 fused to the herpes simplex virus transcription factor VP16 minimal activation domain termed VP64 (dCas9-VP64), and the transcription factor fusion protein that is the NF-kB trans-activating subunit p65 with the activation domain from the human heat-shock factor 1 fused to the phage MS2 coat protein (MS2-p65-HSF1), and then we performed luciferase assays. The BST-2 promoter was activated by the CRISPR-dCas9 system using all six sgRNA targets, among which sgBST2#1 and #2 revealed a 30 to 40-fold activation (Fig. 1B). We therefore selected these two sgRNA target sequences for further experiments.
To examine whether these sgRNAs are indeed able to activate the endogenous expression of BST-2, we employed lentiviral vector systems for transduction. We first produced lentiviruses expressing dCas9-VP64 fusion proteins in HEK293T cells, next infected BST-2-negative HOS cells with the lentiviruses, and then selected the transduced cells with blasticidin whose resistance gene was expressed using the dCas9-VP64 vector. Next, we produced another lentivirus expressing the MS2-p65-HSF1 transcription factor fusion protein, transduced the blasticidin-selected cells with the viral vector, and selected the cells using hygromycin. Lastly, we generated a lentivirus expressing sgBST2#1 and/or sgBST2#2, and transduced the dually selected HOS cells with the viral vector and selected the cells with zeocin. By performing Western analysis, we successfully confirmed the activated expression of endogenous BST-2 protein in a polyclonal population of CRISPR-transduced HOS cells, although additive effect of two sgRNAs was not observed ( Fig. 2A). Then, we cloned the transduced cells using limiting dilution and subjected the resultant single cell clone to real-time RT-PCR, flow cytometry, and immunofluorescence. Real-time RT-PCR showed that endogenous expression of BST-2 mRNA was drastically activated in all single cell clones (Fig. 2B). Cell-surface BST-2 expression analyzed by flow cytometry was robustly increased in the cloned cells (Fig. 2C). Additionally, immunofluorescence revealed a high level of intracellular expression of BST-2 in the same cells (Fig. 2D). We conclude that this CRISPR-based system effectively activates BST-2 expression.
We next performed infection-based virus production assays. HeLa cells or CRISPR-modified HOS cells, as well as BST-2(−) control HOS cells, were infected with either Vpu-intact or deficient VSV-G-pseudotyped viruses prepared from HEK293T cells transfected with the corresponding plasmids, and viruses produced from the infected HeLa or HOS cells were subjected to HIV-1 p24 ELISA to determine the levels of virus production  www.nature.com/scientificreports www.nature.com/scientificreports/ ( Fig. 3A). Production of not only Vpu mutant viruses but also Vpu-intact viruses were effectively inhibited in all single clone cells (Fig. 3B). Importantly, electron microscopic analyses showed that Vpu-intact viruses were indeed accumulated at the surface of BST-2 positive HOS cells (Fig. 3C).
Finally, we transduced the CD4-positive T-cell line H9, which exhibit an intermediate level of expression of BST-2, using the lentiviral CRISPR system to enhance endogenous expression of this protein. Real-time RT PCR showed that CRISPR-transduced H9 clonal cells indeed expressed high levels of BST-2 mRNA (Fig. 4A) without affecting cell proliferation (Fig. 4B). We then performed viral replication assays using either wild-type HIV or Vpu-defective viruses. Consistent with the results obtained in the production assays, replication of not only the Vpu mutant but also the wild-type viruses were markedly decreased in BST-2 positive H9 cells (Fig. 4C). We therefore conclude that enhancement of endogenous BST-2 expression leads to inhibition of HIV-1 production as well as multiple-round replication, even in the presence of the BST-2 antagonist Vpu.

Discussion
In this study, we first identified the sgRNA target sequences at the promoter regions that can activate expression of BST-2, then transduced cells with lentiviruses expressing the sgRNAs plus dCas9 together with accessory transcriptional activators and induced enhanced expression of BST-2 at both the mRNA and protein levels, resulting in successful inhibition of viral production and replication. In many cases, host restriction factors that are able to inhibit HIV infection are normally counteracted by HIV accessory proteins that act as viral antagonists. It has however been shown that the outcome of the battle between the host restriction factors and viral antagonists may be determined not only by their intrinsic activities but also by their stoichiometric balance, since overexpression www.nature.com/scientificreports www.nature.com/scientificreports/ of restriction factors can relieve the blockade imposed by viral antagonists 12,15,16,[39][40][41][42] . This is also suggested by several studies showing that the levels of expression of restriction factors are linked to disease progression after viral infections [32][33][34][35] . In other words, even in the presence of virus-encoded antagonists, the more host restriction factors are expressed, the higher the chances to inactivate the viruses. Although type-I IFN can upregulate hundreds of IFN-stimulated genes that include many host restriction factors, the study of SIV-infected rhesus macaques showed that treatment with IFNα is effective in vivo only during the acute phase of infection and rather detrimental during the chronic stage due to the enhancement of systemic inflammation 43 . Even in human studies, type-I IFNs have been reported to upregulate CCR5 44 that provides target cells for HIV-1, potentially leading to loss of CD4+ T-cells [45][46][47] . In this sense, it is likely that pinpoint activation of restriction factors without using IFNs might be beneficial to hosts. Indeed, our present study showed that endogenous BST-2 expression enhanced by CRISPR methods led to remarkably decreased production and multiple-round replication of HIV-1 harboring the viral antagonist Vpu, although in the case of APOBEC3G gene activation, its inhibitory effect on wild-type HIV-1 infection was weak 48 . Higher expression of BST-2 has been implicated in the growth and progression of cancers due to its ability to promote cell-to cell interactions and to activate NF-kB-mediated signal transduction pathways 49 . In this case, it may be better to avoid traditional overexpression systems using exogenous transgenes that lead to uncontrollably robust expression of target proteins, while CRISPR activation results in endogenously maximized levels of expression, reflecting a more natural mechanisms of action with physiologically relevant phenotypes in vivo 50,51 . Also, instead of the lentiviral systems that we used in this study, adeno-associated viral vectors should be applied to the future experiments to readily prepare high-titer virus stocks and to transduce primary cells. Moreover, it may be necessary to achieve specific gene delivery of CRISPR-based vectors to the infected cells. Targeting restriction factors to enhance their endogenous expression may provide a new therapeutic strategy to combat HIV infections. By using this system, we will pursue the possibility that the CRISPR-based activation of antiviral proteins, especially in the case of BST-2, could be one of the novel therapeutic strategies for inhibiting infections and preventing of transmission with HIV-1 as well as a variety of enveloped viruses.
Virion production assays and transmission electron microscopy. HEK293T cells (5 × 10 5 ) were cotransfected with 1 μg of the proviral construct pNL-E(−), or pNL-E(−)U(−), together with 40 ng of pC-VSVg and 1 μg of the empty pCAGGS by using FuGENE 6. After 48 h the supernatants were harvested and subjected to p24-antigen capture ELISA. CRISPR-transduced HOS cells seeded at 5 × 10 4 cells were infected with 10 ng each of VSV-G-pseudotyped HIV-1. Sixteen hours later the cells were washed with PBS and 1 ml of fresh complete medium was added. For virion production assays, supernatants were harvested after 24 h and subjected to HIV-1 www.nature.com/scientificreports www.nature.com/scientificreports/ p24-antigen capture ELISA. For transmission electron microscopy, infected cells were harvested after 24 h with a cell scraper and washed twice with ice-cold PBS. Cells were then prefixed with 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, for 2 h at room temperature, postfixed in 1% osmium tetroxide, and embedded in Epon 812 (TAAB Laboratories). Ultrathin sections were stained with uranyl acetate and lead citrate and then observed under a transmission electron microscope (HT7700; Hitachi) at 80 kV.

Statistical analyses.
Values are presented as the mean ± s.d. for three or four independent experiments, determined on the basis of pilot experiments to estimate the effective numbers. Statistical comparisons were made using a paired two-tailed Student's t-test, and a P < 0.05 was considered statistically significant.