Analysis of the efficacy of HIV protease inhibitors against SARS-CoV-2′s main protease

Background The pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has resulted in millions of infections worldwide. While the search for an effective antiviral is still ongoing, experimental therapies based on repurposing of available antivirals is being attempted, of which HIV protease inhibitors (PIs) have gained considerable interest. Inhibition profiling of the PIs directly against the viral protease has never been attempted in vitro, and while few studies reported an efficacy of lopinavir and ritonavir in SARS-CoV-2 context, the mechanism of action of the drugs remains to be validated. Methods We carried out an in-depth analysis of the efficacy of HIV PIs against the main protease of SARS-CoV-2 (Mpro) in cell culture and in vitro enzymatic assays, using a methodology that enabled us to focus solely on any potential inhibitory effects of the inhibitors against the viral protease. For cell culture experiments a dark-to-bright GFP reporter substrate system was designed. Results Lopinavir, ritonavir, darunavir, saquinavir, and atazanavir were able to inhibit the viral protease in cell culture, albeit in concentrations much higher than their achievable plasma levels, given their current drug formulations. While inhibition by lopinavir was attributed to its cytotoxicity, ritonavir was the most effective of the panel, with IC50 of 13.7 µM. None of the inhibitors showed significant inhibition of SARS-CoV-2 Mpro in our in vitro enzymatic assays up to 100 µM concentration. Conclusion Targeting of SARS-CoV-2 Mpro by some of the HIV PIs might be of limited clinical potential, given the high concentration of the drugs required to achieve significant inhibition. Therefore, given their weak inhibition of the viral protease, any potential beneficial effect of the PIs in COVID-19 context might perhaps be attributed to acting on other molecular target(s), rather than SARS-CoV-2 Mpro.


Background
In December 2019, a novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was identi ed as the etiological agent of viral pneumonia cases that occurred in Wuhan, Hubei Province, China. As of October 29 thth 2020, the pandemic has resulted in more than 44 million infections, and 1 million deaths worldwide according to the World Health Organization..
There is currently no standardized treatment protocol, and there is no antiviral treatment of proven e cacy recommended for COVID- 19. Clinical management of patients is mainly supportive, including supplementary oxygen and mechanical ventilation if needed. However, given the overwhelming burden of the pandemic on national healthcare systems and the global economy, experimental therapies have been attempted, which are predominantly based on the repurposing of FDA approved antivirals, antimalarials, arthritis drugs, and blood plasma derivatives [1].
The HIV protease inhibitors (PIs) lopinavir and ritonavir have gained particular interest, having shown documented in vitro activity against SARS-CoV and the Middle East respiratory syndrome coronavirus (MERS), however, these studies did not identify a molecular target for the drugs, since their e cacy was solely determined based on the inhibition of cytopathic effects or viral replication, respectively [2 3]. Given the 94.4 % identity in amino acid sequence between SARS-CoV and SARS-CoV-2 [4], studying the e cacy of HIV protease inhibitors against SARS-CoV-2 would be of major relevance.
The genome of SARS-CoV-2 encodes for two viral cysteine proteases; nsp3 (papain-like protease) and nsp5 (main protease) [5]. The main protease (M pro ) of SARS-CoV-2; also named chymotrypsin-like protease (3CLpr), plays a crucial role in the viral life cycle, cleaving the initial polyproteins translated from the viral RNA at at least 11 of its 14 cleavage sites. M pro of SARS-CoV-2 shares 96 % sequence identity to that of SARS-CoV. The enzyme consists of three domains; two domains (I and II) which consist of antiparallel β-barrels, and an α-helical domain (I), which is responsible for dimerization and enzymatic activity [6 7]. Recent structure determination con rmed the similarities between the two enzymes [8].
One potential target for the HIV PIs is the M pro . In silico screening identi ed nel navir as its potential inhibitor [9], while lopinavir and ritonavir were found to be potential inhibitors of the viral enzyme by molecular dynamics simulation [10].
It is important to note that the HIV protease is a C2-symmetric homodimeric aspartyl protease, composed of two identical subunits that are 99 amino acids each. The active site is located at the interface between the two monomers, and contains the catalytic Asp-Thr-Gly residues [11]. M pro on the other hand, is a cysteine protease that can also potentially be targeted by peptide mimetics. Given the structural difference between the two proteases, the e cacy of HIV protease inhibitors against SARS-CoV and SARS-CoV-2 is questionable.
Previous studies reported that a combination of lopinavir/ritonavir and ribavirin was effective against SARS-associated coronavirus, with concentrations of 4 µg/ml and 50 µg/ml, respectively [12]. However, a recent clinical trial of 99 patients with laboratory-con rmed SARS-CoV-2 infection who were treated with lopinavir-ritonavir concluded that no signi cant bene t was observed in the treated group compared to those who received standard care [13]. Recently, a short communication reported that lopinavir inhibited SARS-CoV-2 replication in Vero E6 cells with IC 50 of 26.63 μM, ritonavir, however, showed no inhibition of viral replication [14].
Early in vitro reports from China showed that darunavir inhibited SARS-CoV-2 replication, although at a very high concentration (300 µM) [15]. Clinical trials are currently ongoing [16].
Our aim was to test the e cacy of a panel of HIV PIs against SARS-CoV-2 M pro , using a cell culture-based model. In this study, we determined the IC 50 of the PIs with the aid of a dark-to-bright GFP substrate system that had been developed and applied previously for the investigation of caspases [17]. Moreover, in vitro enzymatic inhibitory assays were also carried out using puri ed M pro and an oligopeptide substrate representing the AVLQ*SGFR cleavage site of SARS-CoV-2 polyprotein 1 ab (PP1ab).

Plasmids and inhibitors
Coding sequence of SARS-CoV-2 M pro (GenBank: MT291835.2) was cloned into pcDNA3.1(+) mammalian expression plasmid using BamHI/EcoRI restriction sites to create the SARS-CoV-2 M pro coding plasmid; thereafter referred to as CoV-2 M pro . The coding sequence of a dark-to-bright GFP reporter substrate; thereafter referred to as PR-Sub, was also cloned into pcDNA3.1(+) plasmid. The PR-Sub was designed to contain a sequence representing the N-terminal autoproteolytic cleavage site of SARS-CoV-2 M pro (TSAVLQ*SGFRKM); corresponding to the nsp4/nsp5 cleavage site, between the GFP and the In uenza A/M2 protein hydrophobic tail (CNDSSDPLVVAASIIGILHLILWILDRL). For in vitro expression of the protease, the coding sequence of His 6 -tagged M pro was cloned into pET11a bacterial expression plasmid using NdeI and BamHI enzymes. The above mentioned expression constructs were obtained using the gene synthesis service of GenScript.
The protease inhibitors darunavir, saquinavir, lopinavir, tipranavir, indinavir sulfate, and atazanavir sulfate were obtained through the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH. Ritonavir was obtained from Abbott laboratories, nel navir from Agouron, and atazanavir from Bristol-Myers Squibb.
A synthetic oligopeptide used in our in vitro enzymatic assay representing the N-terminal autoproteolytic cleavage site of SARS-CoV-2 M pro (AVLQ*SGFR) was obtained from a peptide synthesis service (BioBasic).

Inhibition pro ling in cell culture
On the day of transfection, HEK-293T cells were split and transferred into a 48-well plates (30,000 cells/well) containing serial dilutions of the inhibitor ranging from 200 µM to 5 nM in a total volume of 200 μL DMEM/well, supplemented with 10 % FBS, 1 % glutamine and 1 % penicillin-streptomycin. After 3 hours incubation at 37 °C, cells were transfected with 300 ng of CoV-2 M pro and PR-Sub plasmids using lipofectamine LTX reagent (Thermo Fisher Scienti c), then the cells were incubated for 24 hours. GFP uorescence was then measured by ow cytometry using FACS Calibur. The results were analyzed by FlowJo Software Version 10 (Becton, Dickinson and Company; 2019). Calculations of IC 50 were performed using GraphPad Prism 5.0 (GraphPad Software, Inc).

Cell viability assay
The day before the assay, HEK-293T cells were split into a 96-well plates (20,

In vitro protease assay
The AVLQ*SGFR oligopeptide was dissolved in distilled water and was used as substrate in activity measurements to test inhibitory potentials of the PIs. concentration of 0.12 µM, and the mixtures were incubated at 37°C for 10 minutes. The reactions were terminated by the addition of 180 µL 1% tri uoroacetic acid (TFA). The cleavage products were detected using high performance liquid chromatography (HPLC), utilizing a 0-100% water-acetonitrile gradient in the presence of TFA using Merck Hitachi instrument. Relative activity was determined at less than 20 % substrate hydrolysis. Activity measured in the presence of DMSO was considered to be 100 %. While no potent inhibitor of M pro was available to perform active-site titration, 100 % activity was assumed for the enzyme.

Modeling
Homology modeling of dark-to-bright GFP substrate was performed using Phyre2 web portal [19]. 97 % of residues were modelled at > 90% con dence. Structural gures were prepared PyMol Molecular Graphics System (Version 1.3 Schrödinger, LLC).

Inhibition pro ling in cell culture
To measure M pro activity in cell culture experiments, we applied a modi ed version of a dark-to-bright GFP reporter substrate [17] which was adapted in this study for SARS-CoV-2 M pro . The recombinant substrate consists of an N-terminal GFP, followed by a natural proteolytic cleavage site of SARS-CoV-2 polyprotein, and a C-terminal hydrophobic tail. Proteolysis at the inserted cleavage site releases the tail that serves as a hydrophobic quencher of uorescence and facilitates tetramerization of GFP, which prevents chromophore maturation; the uorescence is restored upon proteolysis (Fig. 1).
Firstly, we optimized the transfection of HEK-293T cells for the use of the SARS-CoV-2 M pro and the darkto-bright GFP substrate. Transfection of cells with only the PR-Sub plasmid resulted in a maximum of 1 % background uorescence after 24 hours incubation. When cells were transfected with both the PR-Sub and CoV-2 M pro plasmids, GFP uorescence ranged from 28-34 %, indicating processing of the substrate and the activity of the protease (Fig. 2).
We then analyzed the inhibition e cacy of a panel of HIV PIs against SARS-CoV-2 M pro . While none of the inhibitors was able inhibit the viral protease in nanomolar concentration; which is expected for effective transition state analogs, in micromolar range, ritonavir was the most effective (IC 50 = 13.7 ± 1.1 µM). Saquinavir, darunavir, and atazanavir were also able to inhibit SARS-CoV-2 M pro at higher concentrations (Table 1). Although a combination of lopinavir and ritonavir resulted in better inhibition of the viral enzyme as compared to ritonavir alone (10.9 µM vs. 13.7 µM, respectively), when we carried out cell viability assays after treatment of the cells with the inhibitors, interestingly, inhibition by lopinavir was found to be a result of the high cytotoxicity observed at concentrations above 50 µM (90 %) (Fig. 3).
In the case of ritonavir and saquinavir, cytotoxicity of >50 % was only observed at concentrations above 50 µM. High concentrations of darunavir and atazanavir on the other hand, were well tolerated by HEK-293T cells.
Indinavir, nel navir and tipranavir however, failed to inhibit SARS-CoV-2 M pro , even at 200 micromolar concentration of the inhibitors.

In vitro enzymatic assay
Following expression and puri cation of the untagged M pro , we determined its catalytic activity after incubation with AVLQ*SGFR oligopeptide substrate. Cleavage position within the substrate was determined using high-performance liquid chromatography coupled to electrospray ionization time-ofight mass spectrometry (HPLC-ESI-TOF MS) (Supplementary Fig. 1). Inhibition pro ling of the PIs was carried out after incubation of the inhibitors along with M pro and the substrate, and the relative e cacies of PIs were compared. None of the inhibitors showed a signi cant inhibition of the M pro in vitro (p values > 0.05) (Fig. 4).

Discussion
To our knowledge, direct determination of the inhibition e cacy of HIV PIs against SARS-CoV-2 M pro in cell culture has not yet been published, although many in silico studies analyzing interaction between SARS-CoV-2 M pro and potential inhibitors were published [20][21][22][23][24][25][26] (Supplementary Table 1). Antiviral assays using lopinavir and ritonavir in Calu-3 cells were previously carried out for MERS-CoV, and the IC 50 for lopinavir, ritonavir and their combination was 11.6, 24.9, and 8.5 µM respectively [3]. In our analysis, we found that a combination of lopinavir plus ritonavir achieved the lowest IC 50 , this however was due to the high cytotoxicity of lopinavir, and not as a result of direct inhibition of SARS-CoV-2 M pro . Ritonavir on the other hand was much more tolerable than lopinavir, and achieved the lowest IC 50 . This should be taken into consideration, given the current drug formulation of lopinavir which is in combination with ritonavir, where ritonavir is used as a pharmacokinetic enhancer due to its inhibition of the cytochrome P450 3A4 isoenzyme, thereby increasing the bioavailability of lopinavir [27]. As a result, administration of ritonavir in combination with other PIs was found to decrease its minimum blood plasma concentration level, as compared to the generic formulation of the drug [28]. Also, while our result regarding ritonavir was in direct contrast to what Choy et al. reported in their short communication [14], we believe that difference in methodologies is to blame for this discrepancy, as we examined the e cacy of ritonavir against the viral M pro protease per se.
It is important to note that the IC 50 of the inhibitors were in the micromolar range which is not considered optimal for the inhibition of the viral enzyme. Previous studies have reported that the minimum concentrations (C min ) for lopinavir, darunavir, saquinavir, and atazanavir in patient's serum under antiretroviral treatment was found to be 9.3, 3.3, 3.8, and < 1 µM, respectively [29][30][31][32] . It would indeed be challenging to achieve such high plasma levels of the inhibitors in order to block the viral replication, moreover, the cytotoxic effects of some of the inhibitors, in addition to the side effects commonly observed with PIs questions the use of anti-HIV PIs in the context of SARS-CoV-2.
Additionally, using an in vitro enzymatic assay, we were able to directly analyze any potential inhibition of M pro by the HIV PIs. Our results show that none of the inhibitors was able to signi cantly inhibit M pro in vitro.
A drawback of this study is that we were not able to assess the interaction between the PIs and the papain-like protease of SARS-CoV-2, as our methodology only enabled us to study the viral main protease. Whether or not these exert any inhibitory effect on the papain-like protease is a subject for future studies, although, a similar methodology may be adapted for SARS-CoV-2 papain-like protease, and other proteases as well. Also, while our cell culture assays were not performed in SARS-CoV-2 target cells, our methodology enabled us to directly examine any potential inhibition of the viral M pro by the PIs, therefore, it is unlikely that different results will be obtained in target cells.

Conclusion
In conclusion, to our knowledge, thorough analysis of the e cacies of PIs against SARS-CoV-2 remains scarce, and the targets of the drugs are yet to be veri ed. While few studies examined the e cacy of some PIs against the replication of SARS-CoV-2, we set out to study whether or not the inhibitors exert a direct effect on the viral protease. In our experiments, even though some of the PIs developed for the treatment of HIV were able to inhibit SARS-CoV-2 M pro , they were only able to do so at high concentrations. The combination of lopinavir plus ritonavir resulted in the lowest IC 50 in cell culture, albeit at the cost of cellular viability. Although, darunavir and atazanavir required a much higher concentration to achieve the inhibition, cytotoxicity was not observed even at a concentration of 200 µM. It should be noted that there might be other molecular targets for the HIV PIs, as nel navir was recently shown to inhibit spike protein-mediated fusion of SARS-CoV-2 [33]. Availability of data and materials All data generated during this study are included in this article and its supplementary information les

Competing interests
The authors declare no competing interests Funding