key: cord-0258971-lfj7t2ip
authors: Ma, Jianpeng; Acevedo, Adam Campos; Wang, Qinghua
title: High-Potency Polypeptide-based Inhibition of Enveloped-virus Glycoproteins
date: 2021-07-01
journal: bioRxiv
DOI: 10.1101/2021.04.05.438537
sha: bc17ee940274859049024e0ccb8391880c6a2f58
doc_id: 258971
cord_uid: lfj7t2ip

Specific manipulation of proteins post-translationally remains difficult. Here we report results of a general approach that uses a partial sequence of a protein to efficiently modulate the expression level of the native protein. When applied to coronavirus, human immunodeficiency virus, Ebolavirus, respiratory syncytial virus and influenza virus, polypeptides containing highly conserved regions of the viral glycoproteins potently diminished expression of the respective native proteins. In the cases of coronavirus and influenza virus where multiple strains were tested, the polypeptides were equally effective against glycoproteins of other coronavirus and influenza strains with sequence identity as low as 27%, underscoring their high insensitivity to mutations. Thus, this method provides a platform for developing high-efficacy broad-spectrum anti-viral inhibitors, as well as a new way to alter expression of essentially any systems post-translationally.

Proteins constitute the building blocks of the cells. Changes in native proteins including mutations, altered functions or levels of expression are fundamental to many forms of diseases including cancer and neurodegenerative diseases, just to name a few. Therefore, manipulation of proteins in a highly specific manner at the post-translational level in the cells is presumably the most straightforward approach for treating diseases. However, this remains a difficult task.

To address these unresolved challenges, here we report a new approach in which a partial sequence of a native protein can efficiently modulate its expression level. Application of this method to the highly conserved regions of glycoproteins from coronavirus, human immunodeficiency virus, Ebolavirus, respiratory syncytial virus and influenza virus resulted in potent polypeptide-based inhibition. Strikingly, the polypeptides developed for SARS-CoV-2 coronavirus and influenza A/H3N2 virus were equally effective against glycoproteins of other coronavirus and influenza strains with sequence identity as low as 27%, underscoring their high insensitivity to mutations. Considering that infectious diseases persist to be a major medical burden around the globe, as exemplified by the ongoing coronavirus disease 2019 (COVID-19) pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), this method provides a platform for developing high-efficacy broadspectrum anti-viral inhibitors. Furthermore, this approach represents a new way to alter expression of essentially any systems post-translationally.

The concept of polypeptide-based interference using coronavirus as an example Viral membrane fusion proteins such as coronavirus spike proteins are oligomeric Class-I transmembrane glycoproteins on the viral envelope 1 . Coronavirus spike (S) proteins are cleaved to give rise to N-terminal S1 regions and C-terminal S2 regions (Fig.1a) . The N-terminal S1 regions are the major target for neutralizing antibodies elicited by natural infection or vaccination, and therefore under constant positive selection for escape variants. For instance, since the onset of COVID-19, four major variants (a-d) of SARS-CoV-2 with extensive mutations have emerged. The mutations on SARS-CoV-2 spike (SARS2-S) proteins 2-4 were concentrated in the S1 regions (Fig.1a) , leading to tighter binding with host cell surface receptor, human angiotensin-converting enzyme 2 (hACE2), higher virulence, resistance to antibody neutralization, and partial escape from natural infection or vaccine induced sera 5-9 .

On the other hand, the C-terminal S2 regions responsible for oligomerization and membrane fusion [10] [11] [12] [13] [14] are more conserved among different coronavirus strains. Upon entry into a host cell, the viral genome will guide the synthesis of new coronavirus spike proteins by ribosomes, which are then folded, assembled and translocated into endoplasmic reticulum (ER) membranes and transit through the ER-to-Golgi intermediate compartment for interaction with newly replicated genomic RNA to produce new virions 15,16 .

We hypothesized that stably foldable fragments of the coronavirus spike proteins, for instance, polypeptides derived from the SARS2-S S2 regions, which maintain the same oligomeric interface as the native spike proteins, would form non-native oligomers with the wild-type spike proteins, thus significantly lowering the level of native spike oligomers on the envelope of new virions, and impairing their infectivity (Fig.1b) . More importantly, the fragment derived from the conserved S2 region would be more resistant to mutations and could potentially be a pan-coronavirus inhibitor. To test this hypothesis, we made two polypeptides derived from the SARS2-S S2 sequence, F1 and F2 (Fig.1a,   Fig.S1, S2) . F1 polypeptide encompassed amino-acid residues 911~1273, while F2 harbored residues 985~1273. Both polypeptides contained a N-terminal signal peptide (SP) as the wild-type SARS2-S for cell surface translocation.

We first assessed the impacts of F1 or F2 on the expression and cell surface translocation of SARS2-S protein. Transient transfection of SARS2-S-coding plasmid resulted in a good level of proteins detected in HEK293T whole cell lysate, with most of the expressed proteins were cleaved ( Fig.2a) . This high-efficiency cleavage of SARS2-S proteins agreed with the novel polybasic cleavage site at the S1/S2 boundary 10, 11, 17, 18 . Moreover, only the S2 fragments of the cleaved SARS2-S protein were labeled by biotin, affinity-purified by anti-biotin antibody and detected in the cell surface fraction ( Fig.2a) , suggesting that only properly cleaved SARS2-S protein were translocated to cell surface.

Most impressively, when F1-coding plasmid was co-transfected with SARS2-S-coding plasmid, even at a twofold molar ratio, the predominant cleaved S2 band of SARS2-S was almost completely diminished in whole cell lysate and in cell surface fraction (Fig.2a) . Thus, F1 strongly inhibited the expression and surface translocation of SARS2-S. In sharp contrast, F2 did not exhibit any significant interference even at a tenfold molar ratio (Fig.2a) . S) and COVID-19 SARS2-S uncovered a wide range of sequence identity levels among them (Fig.S2,   S3 ). For instance, compared with full-length COVID-19 SARS2-S, 2002 SARS-S was 77% amino-acid sequence identical, while 2012 MERS-S was only 35% (Fig.S2, S3) . If only considering the regions included in F1 polypeptide, the identical residues became 94% for 2002 SARS-S and 42% for 2012 MERS-S (Fig.S2, S3) . These spike proteins were thus ideal for testing the sensitivity of F1-induced inhibition to amino-acid mutations. Strikingly, even with F1-coding plasmid at a twofold molar ratio, the full-length SARS-S and MERS-S bands and the cleaved S2 bands were almost completely diminished in whole cell lysate and in cell surface fraction (Fig.2b,c) . Consistent with the robust inhibitory activity exhibited by F1 (Fig.2) , a high level of F1 polypeptide was constantly detected in cell surface fraction (Fig.S4a-c) . In marked contrast, a limited level of F2 polypeptide was detected in cell surface fraction (Fig.S4a ), in agreement with the non-interference of F2 (Fig.2a ).

In order to probe the mechanism of F1-mediated inhibition, we analyzed the mRNA levels of coronavirus spike proteins in the absence or presence of F1 or F2-coding plasmids (Fig.S5, S6) . Clearly co-transfection of F1-or F2-coding plasmids did not significantly change the mRNA levels of coronavirus spike proteins, which were kept at a relatively constant level (between 50%~150%) comparing to the samples that were transfected only with plasmids coding the respective coronavirus spike proteins (Fig.S5) . In sharp contrast, the mRNA levels of F1 were at the order of 2 6.5~212 when normalized against endogenous GAPDH (Fig.S6) , justifying the high potency of F1-mediated interference. The mRNA levels of F2 were comparable with those of F1 (Fig.S6a) , in marked contrast with the low level of F2 polypeptide detected in cell surface fraction (Fig.S4a) , suggesting that the noninhibition of F2 (Fig.2a) was likely due to the instability of the synthesized F2 polypeptide. These data suggested that F1-mediated inhibition of various coronavirus spike glycoproteins was not at the mRNA level.

We then investigated whether F1-mediated inhibition is at the protein level via direct interaction with SARS2-S to form non-native oligomers in the cell. SARS2-S and F1 were each tagged with a monomeric green fluorescent protein (GFP) variant at the extreme C-terminus, CFP for SARS2-S (termed as SARS2C) and YFP for F1 (named as F1Y) (Fig.S7 ). Since the expression of SARS2-S was completely diminished when F1-and SARS2-S-coding plasmids were co-transfected at a 2:1 molar ratio ( Fig.2a) , we used F1Y-and SARS2C-coding plasmids at a reduced 1:1 molar ratio to transiently transfect HEK293T cells, and monitored fluorescence resonance energy transfer (FRET) ratio between them by using the three-cube approach 21 . Although the non-native oligomers formed by SRARS2C and F1Y were expected to be highly unstable, FRET signals between them were robustly detected (Fig.S7) , supporting direct formation of non-native oligomers between SARS2-S and F1.

The high potency of F1 in inhibiting the expression and surface translocation of spike glycoproteins from human coronaviruses that caused severe outbreaks or pandemic between 2002 to 2021 suggests that F1 has a high promise to become an effective therapeutic agent against different coronavirus lineages over a long time period. Therefore, we sought to identify a convenient way to deliver F1 for therapeutic purpose. Since F1 directly interacts with its target spike protein to form nonnative oligomers concomitant with protein synthesis and folding, F1-coding gene needs to be delivered to the site of action. Minicircles are a type of newly developed DNA carriers for gene therapy 22 . The main features of minicircles include the cleaner gene background with minimal viral or bacterial gene elements, sustained high-level protein expression, and more importantly, the small size that may allow the use of aerosols for drug delivery 23 . The latter may be a distinct advantage against coronaviruscaused respiratory diseases.

We made a F1 minicircle by inserting the F1-coding sequence into the parental minicircle cloning vector pMC.CMV-MCS-SV40polyA (Fig.3a) , and tested its efficacy in inhibiting the expression and surface translocation of coronavirus spike glycoproteins. Compared to the controls where no minicircle was used, the presence of F1 minicircles, at a merely 4.5-fold molar ratio, almost completely abolished cell surface translocation of all three spike proteins ( Fig.3b-d) . It is important to note that this level of inhibition was achieved under the situation where pcDNA3.1-based plasmids harboring coronavirus spike-coding genes were efficiently replicated in HEK293T cells, while F1 minicircle cannot.

To investigate the consequences of the reduced cell surface translocation of coronavirus spike proteins by F1 minicircle, we compared the level of SARS2-S protein on pseudoviruses generated using luciferase-expressing, env-defective HIV-1 genome plasmid pRL4.3-Luc-R -Ein the presence of different molar ratios of control minicircle made from the empty parental vector (termed as MN501A) or F1 minicircle. In order to make sure that only spike proteins anchored on the pseudovirus envelope were accounted for, we employed QuickTiter Lentivirus Titer kit to precipitate intact pseudoviruses from cleared supernatant prior to analysis by western blot. Impressively, even with a twofold molar ratio of F1 minicircle, almost no SARS2-S was detected on the generated intact pseudoviruses (Fig.4a ).

Not surprisingly, these pseudoviruses completely failed to infect hACE2-expressing HEK293T cells ( Fig.4b) .

Using the same strategy, we designed polypeptide-based inhibitors for human immunodeficiency virus-1 (HIV-1) envelope glycoprotein (gp160), Zaire ebolavirus (EBOV) glycoprotein (GP), human respiratory syncytial virus (RSV) A fusion protein (F) and influenza virus type A and B hemagglutinin (HA) (termed as gp160i, GPi, Fi and HAi, respectively, where the letter "i" denotes their inhibitory activities, Fig.S1 ). When the inhibitor-coding plasmids were co-transfected with the corresponding glycoprotein-coding plasmid, at 5~15-fold molar ratio, the corresponding glycoprotein was almost completely diminished in whole cell lysate and in cell surface fraction (Fig.5a-d S3 ). If only considering the interfering polypeptide HAi region, the sequence identity was at 53%, 33% and 34%, respectively. These data further demonstrated that the use of a partial native sequence is an effective method for targeted reduction of protein expression of these important human pathogens, even when the level of sequence identity is as low as 27%.

In the current research, we present results of a new approach that uses a partial sequence of a protein to efficiently modulate its expression level. In the case of coronaviruses, the F1 polypeptide derived from COVID-19 SARS2-S potently reduced the expression and surface translocation of the spike proteins from all three human coronaviruses that caused major regional outbreaks or global pandemic in the last 20 years, despite as low as 35% amino-acid sequence identity among them.

Although extensive mutations were found on SARS2-S proteins in recent SARS-CoV-2 variants 2-4 ( Fig.1a) , the regions corresponding to F1 polypeptide were highly conserved, as exemplified by the nearly identical sequence of SARS2a-S, a variant of SARS2-S, with F1 ( Fig.S3b) . Therefore, F1 polypeptide should be effective on the spike proteins of almost any emerging COVID-19 SARS-CoV-2 variants in the future. Furthermore, since the spike proteins of other human coronaviruses (including HCoV-HKU1 and HCoV-OC43, HCoV-NL63 and HCoV-229E) and 2012 MERS-CoV have a similar level of sequence identity with F1 (Fig.S3b) , F1 polypeptide may be equally effective on all these spike proteins. Additionally, we demonstrated that the same strategy was also effective on glycoproteins from other enveloped viruses including HIV-1, EBOV, RSV and influenza viruses. Therefore, this polypeptide-based approach represents a new concept for targeted manipulation of viral and non-viral proteins with high potency and fine tunability. EBOV GP. c). RSV F. d). Influenza HAs. For panels a-d), the levels of envelope glycoproteins in whole-cell lysate (left) or in cell-surface fraction (right) were compared for HEK293T cells transfected with envelope-coding plasmid only, or together with 5~15-fold molar ratio inhibitor-coding plasmid.

Endogenous CD147 protein detected by anti-CD147 antibody was used as an internal control. 

Cell surface biotinylation and protein purification were performed using Pierce Cell surface Protein Biotinylation and Isolation Kit following the manufacturer's instruction. Briefly, cell surface proteins on HEK293T cells were first labeled with Sulfo-NHS-SS-Biotin at 4°C for 30 minutes, which were then stopped by adding Tris-buffered saline and further washed. After cells were lysed with Lysis Buffer, lysate was cleared by centrifugation. Cleared lysate was incubated with NeutrAvidin Agarose to allow binding of biotinylated proteins. After extensive wash, the bound proteins were eluted with Elution Buffer containing 10 mM DTT. The cleared lysate ("Whole cell" fraction) and eluted proteins ("Cell surface" fraction) were run on 10% SDS-PAGE and the spike proteins were detected by C9rhodopsin antibody 1D4 HRP. Endogenous membrane-anchored protein CNPase detected by anti-CNPase antibody or CD147 detected by anti-CD147 antibody were used as an internal control.

Total RNA was purified using Quick-RNA miniprep Kit. Reverse transcription was carried out using iScript Reverse Transcription Supermix. qPCR was performed using Bimake SYBR Green qPCR Master Mix with the following primers: 

Pseudovirus generation followed the protocol reported earlier 25 . Essentially, HEK293T cells were seeded on 6-well plates the night before. The next day, pcDNA3.1-SARS2-S (0.6 µg) and pRL4.3-Luc-R -E -(0.6 µg) were used to transfect one-well HEK293T cells using Lipofectine 3000. MN501A minicircle or F1 minicircle at indicated molar ratio was included in the transfection mixture. At 16 hours post-transfection, the HEK293T cells were fed with fresh medium. At 48 hours after medium change, the supernatant of each well of the 6-well plates was harvested, and centrifuged at 300 g for 5 minutes to remove cell debris. Intact pseudoviruses were purified using QuickTiter Lentivirus Titer kit following the manufacturer's instruction. The virus lysate was analyzed by western blot using rhodopsin antibody 1D4 HRP for spike proteins, and FITC-conjugated anti-p24 mAb and HRPconjugated anti-FITC mAb for p24, which served as an internal control. A portion of pseudoviruscontaining supernatant was concentrated by PEG8000 and used for luciferase assay of cell entry.

HEK293T cells were seeded on 100 mm dishes the night before. The next day, HEK293T cells were transfected with 10 µg pcDNA3.1-hACE2 using Lipofectine 3000. At 16 hours post-transfection, the cells were resuspended in FBS-free DMEM medium, and plated onto 96-well white plates to which 10 µL concentrated pseudoviruses was already added to each well. Two hours later, each well was added with an equal volume of DMEM containing 20% FBS. The cells were further incubated for 36 hours, then an equal volume of One-Glo EX Luciferase Assay Reagent was added, after incubation for 3 minutes, the luminescence signals were recorded.

High quality/high resolution automated imaging was performed on a GE Healthcare DVLive epifluorescence image restoration microscope using an Olympus PlanApoN 60X/1.42 NA objective and 

Anti-CNPase

Anti-C9-Rhodopsin

Whole cell Cell surface

➕ ➕ ➕ ➕ ➕ ➕ F1 plasmid ➖ 2x 10x ➖ 2x 10x

Anti 

Anti-C9-Rhodopsin 

RSV F ➕ ➕ ➕ ➕ ➕ ➕ Fi ➖ 5x 15x ➖ 5x

Anti-C9-Rhodopsin 

Ebola

Whole ALGKLQDVVNQNAQALNTLVKQLSSNFGAISSVLNDILSRLDKVEAEVQIDRLITGRLQSLQTYVTQQLIRAAEIRASANLAATKMSECVLGQSKRVDFCGKGYHLMSFPQSAPHGVVFL 1063 SARS2a ALGKLQDVVNQNAQALNTLVKQLSSNFGAISSVLNDILARLDKVEAEVQIDRLITGRLQSLQTYVTQQLIRAAEIRASANLAATKMSECVLGQSKRVDFCGKGYHLMSFPQSAPHGVVFL 1060 **************************************:************************************************************************:******** SARS HVTYVPSQERNFTTAPAICHEGKAYFPREGVFVFNGTSWFITQRNFFSPQIITTDNTFVSGNCDVVIGIINNTVYDPLQPELDSFKEELDKYFKNHTSPDVDLGDISGINASVVNIQKEI 1178 SARS2 HVTYVPAQEKNFTTAPAICHDGKAHFPREGVFVSNGTHWFVTQRNFYEPQIITTDNTFVSGNCDVVIGIVNNTVYDPLQPELDSFKEELDKYFKNHTSPDVDLGDISGINASVVNIQKEI 1183 SARS2a. HVTYVPAQEKNFTTAPAICHDGKAHFPREGVFVSNGTHWFVTQRNFYEPQIITTHNTFVSGNCDVVIGIVNNTVYDPLQPELDSFKEELDKYFKNHTSPDVDLGDISGINASVVNIQKEI 1180 ******:**:**********:***:******** *** **:*****:.******.**************:************************************************** SARS DRLNEVAKNLNESLIDLQELGKYEQYIKWPWYVWLGFIAGLIAIVMVTILLCCMTSCCSCLKGACSCGSCCKFDEDDSEPVLKGVKLHYT 1268 SARS2

DRLNEVAKNLNESLIDLQELGKYEQYIKWPWYIWLGFIAGLIAIVMVTIMLCCMTSCCSCLKGCCSCGSCCKFDEDDSEPVLKGVKLHYT 1273 SARS2a. DRLNEVAKNLNESLIDLQELGKYEQYIKWPWYIWLGFIAGLIAIVMVTIMLCCMTSCCSCLKGCCSCGSCCKFDEDDSEPVLKGVKLHYT 1270 ********************************:****************:*************.************************** a) S1 S2 

Coronavirus biology and replication: implications for SARS-CoV-2

Advances in Non-Viral DNA Vectors for Gene Therapy

Supercoiled Minivector DNA resists shear forces associated with gene therapy delivery

A robust system for production of minicircle DNA vectors

A safe and convenient pseudovirus-based inhibition assay to detect neutralizing antibodies and screen for viral entry inhibitors against the novel human coronavirus MERS-CoV