Endosomal acidication inhibitors broadly inhibit inuenza virus and coronavirus in vivo

Inuenza virus, coronavirus, and drug-resistant viruses are long-term threats to public health because of lacking effective antivirals. Thus, chemicals with broad-spectrum antiviral activities and low possibility to induce drug resistance are urgently needed. Here, we identify a peptidic inhibitor P16 signicantly inhibiting inuenza A/B virus by binding to HA to block viral fusion. Moreover, P16 antagonizes endosomal acidication to suppress inuenza virus and SARS-CoV-2 entry through the endocytic pathway. Importantly, endosomal acidication inhibitor P16 or chloroquine can broadly inhibit A(H1N1) virus, SARS-CoV and SARS-CoV-2 replication in mice and hamsters when administrated through intranasal inoculation or atomization inhalation, contrary to reported treatment failure by systemic route. Chloroquine can signicantly inhibit SARS-CoV-2 replication in ex vivo human lung tissues. In conclusion, endosomal acidication inhibitors (P16 and chloroquine) can broadly inhibit inuenza virus and coronavirus replication in vivo, which supports atomization inhalation of chloroquine for treating coronavirus and inuenza patients in clinical trials.


Introduction
Before COVID-19, limited attention was paid on discovering anti-coronavirus drugs, even after the SARS-CoV outbreak in 2003. In uenza viruses which caused pandemic and seasonal outbreaks have repeatedly overwhelmed healthcare institutions and affected socioeconomic activities. The suboptimal effectiveness of currently available anti-in uenza drugs against certain strains was evidenced by the high mortality rates (> 30%) of In uenza A(H5N1) and A(H7N9) virus infected patients (1,2). Drug resistant viruses can emerge quickly in patients while on treatment with speci c anti-in uenza drugs such as oseltamivir and baloxavir (3,4). Moreover, resistant viruses against anti-in uenza neutralizing monoclonal antibody could be identi ed after extensive virus passaging (5,6). Besides the endless in uenza outbreaks, the novel SARS-CoV-2 has emerged and disseminated globally since early 2020. Together with the circulating seasonal in uenza virus, SARS-CoV-2 may cause co-infection with increased severity during the in uenza season (7)(8)(9). These circulating in uenza virus, coronavirus and resistant virus mutants reveal our poor capability in responding to the threats of emerging/reemerging viruses with the currently available antivirals (1,2). Thus, broad-spectrum agents inhibiting both in uenza virus and coronavirus with low possibility to induce drug-resistance are urgently needed for combating the emergence of novel viruses. Antiviral peptides with broad-spectrum antiviral activities against in uenza virus and/or coronavirus have showed promising prospects with little metabolic toxicity and low possibility of inducing drug-resistant viruses (10)(11)(12)(13)(14)(15)(16). Due to the lack of proofreading activity of RNA polymerases in RNA viruses, SARS-CoV-2 mutants were not infrequently found in patients during the galloping pandemic (17). Similarly, drugresistant viral mutants emerged during treatment by speci c antivirals, especially the small molecular compounds including neuraminidase inhibitors, M2 and polymerase inhibitors (3,4). Defensins, naturally existing in almost all multicellular plants and animals, have been shown to have broad antiviral activities against in uenza virus, coronavirus and other viruses (18,19). Moreover, defensin-induced resistant viruses have not been reported which is consistent with our nding that defensin-derived peptide P9R did not induce drug-resistant virus even after extensive virus passaging in the presence of P9R (16).
Chloroquine with broad-spectrum antiviral activities targeting host factors was effective inhibiting pHdependent viruses in vitro and in vivo (20)(21)(22), but not effective in vivo in some other studies with unclear reason (23)(24)(25). The broad-spectrum antiviral activity and low chance of inducing drug-resistant viruses make the defensin-derived peptide and similar host-targeting antivirals being promising candidates for drug development.
In this study, we rst identi ed a peptidic P16, modi ed from frog defensin with more positive charges, which can inhibit in uenza A/B virus and SARS-CoV-2. In vitro mechanistic studies demonstrated that P16 could block the low-pH induced HA conformational changes of A(H1N1), A(H7N7) and FluB viruses.
In addition, P16 could inhibit endosomal acidi cation and block spike-ACE2-mediated viral entry of SARS-CoV-2 through the endocytic pathway. Importantly, we showed that endosomal acidi cation inhibitors (peptidic P16 and chemical chloroquine) could signi cantly inhibit A(H1N1) virus, SARS-CoV and SARS-CoV-2 replication in mice and hamsters when administrated through intranasal routes (intranasal inoculation or atomization inhalation), in contrary to previous reports of treatment failure by oral chloroquine in clinical trials. We illustrated that chloroquine could effectively inhibit viral replication when given at the same route as virus challenge. Moreover, chloroquine signi cantly inhibited SARS-CoV-2 replication in ex vivo human lung tissues. Overall, we developed a strategy to make antiviral peptide having dual antiviral functions by direct binding leading to inhibition (i.e. blocking in uenza HA conformational change) and inhibiting endosomal acidi cation to suppress pH-dependent viruses (i.e. blocking in uenza virus and SARS-CoV-2 fusion in endolysosomes). We further demonstrated that endosomal acidi cation inhibitors (P16 and chloroquine) could effectively inhibit in uenza virus and coronavirus replication in vivo when administrated through intranasal inoculation or atomization inhalation. The antiviral activity of atomization inhalation of chloroquine provided the evidence that chloroquine might be re-purposed for treating coronavirus and in uenza virus diseases when administrated through atomization inhalation to patients, in which atomization inhalation may achieve more effective drug delivery when compared with the conventional inhalation in animal models.

Basic peptide P16 inhibited in uenza A and B virus
Our previous studies of mouse beta defensin (30 amino acid peptide) indicated that putative virusbinding peptide without antiviral activity through virus binding could be modi ed to have more positive charges to enhance the antiviral activity by inhibiting endosomal acidi cation (16). Here, from another perspective, we aimed to show that a virus-binding antiviral peptide could be modi ed with more positive charges to acquire dual-functional activities against virus due to the inhibition by direct binding and the enhanced inhibition on endosomal acidi cation. We identi ed short peptide U4 and U5, derived from a frog defensin Urumin which could bind to HA stem of group 1 in uenza A virus (13), showing more potent antiviral activity than that of Urumin against A(H1N1) virus (Fig. 1A-1B). Based on U4 and U5, we identi ed a short 16 amino acid peptide (P16) with positive charge (+6.1), which could signi cantly inhibit A(H1N1) virus (IC 50 =3.9 μg ml -1 , Fig. 1C), which is lower than the IC 50 of U4 (6.6 μg ml -1 ) and U5 (12.9 μg ml -1 ). The potent anti-H1N1 activity of P16 was further con rmed by the inhibition on viral multicycle growth (Fig. 1D), which showed that P16 could inhibit viral replication by 41-fold. More interestingly, P16 could also inhibit A(H3N2) (IC 50 =1.6 μg ml -1 ) and FluB (IC 50 =7.1 μg ml -1 ) viruses (Fig.   1E). The cytotoxicity analysis indicated that no signi cant cytotoxicity of P16 on MDCK cells treated by 1 mg ml -1 of P16 (TC 50 > 1mg ml -1 , Fig. 1F) and no signi cant hemolysis was observed when Turkey red blood cells (RBC) were treated by P16 ( Supplementary Fig. 1). These results showed that the short basic peptide P16 could potently inhibit in uenza A (group 1 and group 2) and B viruses without obvious cytotoxic effect on host cells.

P16 inhibited HA fusion and endosomal acidi cation
To investigate the antiviral mechanism of P16 against in uenza virus, cells were treated with P16 before A(H1N1) virus infection, but no antiviral activity was detected ( Supplementary Fig. 2). When virus was pretreated by P16 before viral infection, P16 could signi cantly inhibit viral replication ( Fig. 2A). When infected cells were treated by P16 after viral infection, P16 did not inhibit viral replication in cells ( Supplementary Fig. 3A) and viral release in supernatants ( Supplementary Fig. 3B). These results indicated that the antiviral activity of P16 might rely mainly on targeting virus before viral entry. To further con rm this result, we treated A(H1N1) virus (1×10 6 PFU ml -1 ) with P16 (500 μg ml -1 ) and then diluted the P16-treated virus by 10, 000 folds for plaque assay. After dilution, P16 at the concentration of 0.05 μg ml -1 which is less than the IC 50 of 3.9 μg ml -1 , could still signi cantly reduce the plaque number (Fig. 2B), which further con rmed that the antiviral activity of P16 mainly relied on targeting virus and that the binding of P16 to virus was irreversible. Next, intact A(H1N1) virus treated by P16, similar to the untreated virus, could be detected by transmission electron microscopy (TEM). However, intact viral particle was not detected when virus was treated by Triton X-100 ( Supplementary Fig. 4). These results indicated that the antiviral activity of P16 mainly relied on direct targeting of virus which interfered with viral infection at early stage without disrupting the viral particles.
Next, we used A(H5N1) pseudovirus which only expressed HA and NA proteins to test whether P16 could affect the entry of pseudovirus. As shown in Fig. 2C, P16 could signi cantly inhibit the entry of A(H5N1) pseudovirus, which indicated that P16 most likely interfered with the early step of viral infection by targeting HA. Next, P16 did not reduce the viral attachment (Fig. 2D), which also suggested that P16 did not disrupt viral particles because the viral RNA copies of attached virus were signi cantly reduced on cell surface (Fig. 2D) when virus was treated and disrupted by Triton X-100. We further con rmed that P16 did not have hemagglutination inhibition (HAI) activity against A(H1N1) and FluB viruses when compared with neutralization antibody ( Supplementary Fig. 5). Furthermore, unlike P9R which bound to viral surface HA and captures viral particles(16), P16 could not capture viral particles as shown in capture EIA assay (Fig. 2E and Supplementary Fig. 6), which implicated that P16 was less likely to bind to the head region of HA. Considering the broad-spectrum antiviral activities of P16 against group 1 and 2 in uenza A virus and FluB virus, we hypothesized that P16 might bind to HA stem region so as to interfere HA conformational change. Consistently, we identi ed that P16 could signi cantly inhibit RBC hemolysis induced by group 1 A(H1N1) virus at pH 5.0 condition (Fig. 2F). Furthermore, it was con rmed that P16 could block group 2 A(H7N7)-HA and FluB virus mediated cell fusion triggered by the low pH in 293T cells (Fig. 2G) and MDCK cells ( Supplementary Fig. 7). The low pH-treated cells without P16 showed >2-fold larger size than the normal cells treated by pH 7.4. These results indicated that P16 could broadly bind to stem region to block the low pH-induced HA conformational change (26). Finally, it was demonstrated that basic P16 could inhibit endosomal acidi cation similar to the effect of ba lomycin A1 (16) in live cells (Fig. 2H). These results indicated that P16 could have dual functions which not only blocked HA fusion by binding but also inhibited endosomal acidi cation to interfere viral entry through the endocytic pathway.

P16 inhibited SARS-CoV-2 by interfering endosomal pH
The broad-spectrum antiviral activity of P16 was not restricted to in uenza A and B viruses. P16 could signi cantly inhibit SARS-CoV-2 infection with IC 50 (2.8 μg ml -1 ) as measured by plaque reduction assay  10), which was consistent to the activity of P16 against SARS-CoV-2 by inhibiting endosomal acidi cation. This is same as chloroquine which cannot inhibit SARS-CoV-2 replication in Calu-3 cells(27) because chloroquine inhibited endosomal acidi cation but not TMPRSS2 activity.

Endosomal acidi cation inhibitors against viruses in vivo
To evaluate the antiviral e cacy of P16 in vivo, we challenged mice with A(H1N1)pdm09 virus. P16 could signi cantly decrease lung viral load by 12-fold (Fig. 4A) and improve the survival of challenged mice by 80% (Fig. 4B). The protection conferred by P16 on infected mice was similar to that of zanamivir ( Fig. 4A-4B). Considering the drug resistant problems of anti-in uenza drugs, we tested whether the drug-resistant mutants of A(H1N1) virus could emerge when A(H1N1) virus was cultured in the presence of P16 (Fig.  4C). P16 could e ciently inhibit the replication of passaged virus (P15 and P20), with similar e ciency against the virus without any passage (P0).
Interestingly, intranasal inoculation of P16 could signi cantly inhibit 14-fold SARS-CoV-2 replication in hamster lungs (Fig. 4D), which indicated that blocking endosomal acidi cation could inhibit SARS-CoV-2 replication in vivo even though P16 did not inhibit TMPRSS2-mediated infection of SARS-CoV-2 in Calu-3 cells ( Supplementary Fig. 10). Thus, we also demonstrated that intranasal administration of endosomal acidi cation inhibitor chloroquine (2 mg kg -1 ) could inhibit 2-fold SARS-CoV-2 replication in hamster lungs (Fig. 4D). We further demonstrated that intranasal administration of chloroquine could also inhibit 3.5-fold A(H1N1) virus replication in mice (Fig. 4E) when compared with mock, which was also less potent than that of P16 (12-fold) against A(H1N1) virus. To understand the highly signi cant antiviral effect of chloroquine against in uenza virus and coronavirus in cell culture ( Supplementary Fig. 11), but poor treatment effect in vivo, we identi ed that chloroquine could not provide 1-day prophylactic activity inhibiting SARS-CoV and A(H1N1) virus replication in BALB/c mice ( Supplementary Fig. 12), even though chloroquine has a long half life time (28). This indicated that the accumulation of chloroquine in endosomes but not in serum and cytosolic environment of lungs (25) did not provide the antiviral activity against pH-dependent viruses in vivo. Chloroquine could effectively inhibit viral replication when virus was treated with chloroquine during viral infection (Fig. 4F). However, treating cells with chloroquine before or after SARS-COV-2 infection, which did not effectively let chloroquine enter cells with virus together, showed signi cantly less antiviral activity than treating cells during viral infection (Fig. 4F). Thus, we speculated that chloroquine can exhibit antiviral activity in vivo if lung cells are constantly exposed to chloroquine with effective concentration against pH-dependent viruses. We demonstrated that one dose of chloroquine could inhibit (3-5)-fold SARS-CoV and A(H1N1) virus replication in BALB/c mice when chloroquine was intranasally administrated into mouse lungs at the time before viral inoculation ( Supplementary Fig. 13).
Most importantly, we demonstrated that atomization inhalation of chloroquine could inhibit SARS-CoV and A(H1N1) virus replication in mice ( Fig. 4G-4H), which provided the evidence that atomization inhalation of chloroquine is an achievable method in clinical treatment for coronavirus patients and in uenza patients. These in vivo data provided the information that the endosomal acidi cation inhibitor chloroquine, which could not inhibit coronavirus or in uenza virus replication in vivo when administrated by oral inoculation (Supplementary Fig. 14) or intraperitoneal injection (24,25,29,30), might effectively inhibit coronavirus and in uenza virus replication in humans if chloroquine could be administrated to lungs by intranasal routes to provide effective chance of keeping lung cells in the bath of chloroquine environment. Thinking about the TMPRSS2-mediated entry pathway of SARS-CoV-2 in humans, we further demonstrated that chloroquine could signi cantly inhibit SARS-CoV-2 replication in ex vivo human lung tissues which were kept in the bath of chloroquine (Fig. 4I). Overall, we demonstrated that endosomal acidi cation inhibitors (P16 and chloroquine) could signi cantly inhibit SARS-CoV-2, SARS-CoV and in uenza A(H1N1) virus replication in vivo through intranasal administration when lung cells could be bathed in chloroquine with effective concentration.

Discussion
In this study, we identi ed that endosomal acidi cation inhibitors (peptidic P16 and chemical chloroquine) could suppress in uenza virus and coronavirus infection in vitro and in vivo. Dual-functional P16 could inhibit in uenza virus by blocking in uenza HA conformation change and inhibiting endosomal acidi cation to block pH-dependent in uenza and coronavirus fusion. Importantly, we demonstrated that both endosomal acidi cation inhibitors (P16 and chloroquine) could effectively inhibit A(H1N1) and SARS-CoV-2 replication in mice and hamsters. Chloroquine could signi cantly inhibit SARS-CoV-2 replication in ex vivo human lung tissues, which indicated that chloroquine might signi cantly inhibit SARS-CoV-2 replication in human lungs when chloroquine could be delivered into lungs with effective concentration.
Broad-spectrum antivirals, like peptides binding to HA stem region, have been found to inhibit group 1 in uenza A viruses (13,26). However, the stem region is not very conserved in in uenza virus (31), which may pose a challenge to nd an universal antiviral against in uenza virus by targeting the HA stem region. Here, we reported the peptidic P16 inhibiting HA-mediated cell fusion of group 1 A(H1N1), group 2 A(H7N7) and FluB viruses, which provided the evidence that an antiviral peptide could be a fusion inhibitor with broad-spectrum activities against both in uenza A and B viruses. Moreover, basic P16 could bind to viruses and inhibit in uenza virus and SARS-CoV-2 by preventing endosomal acidi cation. Broadspectrum antiviral, like chloroquine elevating the endosomal pH of host cells without binding to virus, had been identi ed to inhibit viruses in vitro and in vivo (20,21). However, a number of studies demonstrated the lack of antiviral activity of chloroquine in vivo (24,25,32). Here, we identi ed both endosomal acidi cation inhibitors (P16 and chloroquine) could inhibit A(H1N1) virus in mice and SARS-CoV-2 replication in hamsters when administrated through intranasal routes. We demonstrated that treating mice with chloroquine before viral inoculation could effectively inhibit in uenza and coronavirus replication in vivo. However, one-day prophylactic treatment with chloroquine could not show antiviral activity against in uenza and coronavirus in mice. We further illustrated that the effective way to use chloroquine against pH-dependent virus is to provide a high enough concentration of chloroquine when virus is entering cells (Fig. 4J). Thus, it is important to provide the effective chance of keeping lung cells in chloroquine environment, which could effectively inhibit viral RNA release by preventing virusendosome acidi cation.
Because of the high mutation rates resulting antigenic drift, in uenza vaccines could only provide 10 ~ 60% protection during the past decades. Since the discovery of neuraminidase inhibitors in 1990(33), M2 inhibitor (34), and polymerase inhibitor(4) have been found. However, viruses resistant to these inhibitors quickly emerged even during the clinical trials (3,4,35). In addition, with the circulation of pandemic SARS-CoV-2, different mutants have been reported. Thus, antivirals with new acting mechanism and broad-spectrum activities against viruses with less possibility to induce drug resistance are needed for treating in uenza virus and coronavirus. P16 derived from frog defensin could broadly inhibit in uenza A and B viruses and SARS-CoV-2 with low possibility to cause in uenza drug-resistant virus. No drugresistant virus was found after 20 viral passages in the presence of P16. Interestingly and importantly, chloroquine, despite not showing effective antiviral activity in vivo (23,24,29,30) when administrated through systemic routes in many studies, could target host endosomes to block pH-dependent viral infection and signi cantly inhibit in uenza virus and coronavirus replication in mice and hamsters when administrated by intranasal routes. The host-targeting antiviral activity of chloroquine indicates that chloroquine is less likely to induce drug-resistant problem. Currently, the rapid RT-PCR test can identify patients in the early period of infection which allows atomization of these endosomal acidi cation inhibitors to treat SARS-CoV-2 patients. For late presenters when the peak of viral load has passed, antiviral drugs given by any routes are unlikely to improve the outcome. Immunomodulatory agents such as steroid may play a more important therapeutic role. In addition, chloroquine could suppress the expression of in ammation cytokines in human lungs (36). Thus, the topical administration of chloroquine by atomization inhalation in the early period of infection may not only enhance the antiviral e cacy in lungs but also reduce the potential cardiac side effects when compared with systemic administration by oral or intravenous injection. These data give us evidences to support the use of endosomal acidi cation inhibitor chloroquine as a broad-spectrum antiviral candidate to treat coronavirus, in uenza or co-infection of coronavirus and in uenza virus in clinical trials through atomization inhalation.

Plaque reduction assay
Peptides were synthesized by ChinaPeptide. Antiviral activity of peptides was measured using a plaque reduction assay. Brie y, peptides or bovine serum albumin (BSA, 0.2-50.0 μg ml −1 ) were premixed with 50 PFU of virus in PBS at room temperature. After 45-60 min of incubation at room temperature, peptidevirus mixture was transferred to MDCK or Vero-E6 cells, correspondingly. At 1 h post infection, infectious media were removed and 1% low melting agar was added to cells. Cells were xed using 4% formalin at 2-3 day post infection. Crystal violet (0.1%) was added for staining, and the number of plaques was counted.
Viral RNA extraction and RT-qPCR Viral RNA was extracted by Viral RNA Mini Kit (QIAGEN, Cat # 52906, USA) according to the manufacturer's instructions. Extracted RNA was reverse transcribed to cDNA using PrimeScript II 1 st Strand cDNA synthesis Kit (Takara, Cat # 6210A) with GeneAmp® PCR system 9700 (Applied Biosystems, USA). The cDNA was then ampli ed using speci c primers (Supplementary Table 1) for detecting A(H1N1) virus using LightCycle® 480 SYBR Green I Master (Roach, USA). For quantitation, 10-fold serial dilutions of standard plasmid equivalent to 10 1 to 10 6 copies per reaction were prepared to generate the calibration curve. Real-time qPCR experiments were performed using LightCycler® 96 system (Roche, USA).

Antiviral multicycle growth assay
In uenza viruses were pretreated by peptide and then infected MDCK cells (0.005 MOI). After 1h infection, infectious media were removed and fresh media with supplemental peptides were added to infected cells for virus culture. At 18 h post infection, the supernatants of infected cells were collected for RT-qPCR assay to determine the viral titers in cell supernatants.

Cytotoxicity assay
Cytotoxicity of peptides was determined by the detection of 50% cytotoxic concentration (CC 50 ) using a tetrazolium-based colorimetric MTT assay. Brie y, MDCK cells were seeded in 96-well cell culture plate at an initial density of 2 × 10 4 cells per well in DMEM supplemented with 10% FBS and incubated for overnight. Cell culture media were removed and then DMEM supplemented with various concentrations of peptides and 1% FBS were added to each well. After 24 h incubation at 37 °C, MTT solution (5 mg ml −1 , 10 μl per well) was added to each well for incubation at 37 °C for 4 h. Then, 100 μl of 10% SDS in 0.01M HCl was added to each well. After further incubation at room temperature with shaking overnight, the plates were read at OD 570 using VictorTM X3 Multilabel Reader (PerkinElmer, USA). Cell culture wells without peptides were used as the experiment control and medium only served as a blank control.

Hemolysis and hemolysis inhibition assay
Serially diluted peptide P16 in PBS were incubated with turkey red blood cells for 1 h at 37°C. PBS was used as a 0% lysis control and 0.1% Triton X-100 as 100% lysis control. Plates were centrifuged at 350 g for 3 min to pellet non-lysed red blood cells. Supernatants used to measure hemoglobin release were detected by absorbance at 450 nm. For hemolysis inhibition assay, P16 (200 μg ml -1 ), P9R (200 μg ml -1 ) or arbidol (100 μg ml -1 ) were mixed with or without same volume of H1N1 virus (HA titer >128) for 1 h, and then 60 μl of 2% turkey red blood cells was added for 15 min. PBS and Triton X-100 (0.1%) were included as the negative and positive control of hemolysis. The precipitated erythrocytes were incubated with sodium citrate solution (pH of 4.9) for 25 min. The hemoglobin release in supernatants was detected at 450 nm.

Transmission electron microscopy assay
To determine the effect of P16 on viral particles, A(H1N1) virus was pretreated by 200 μg ml -1 of P16, PBS or Triton X-100 (0.15%) for 1h. The virus was xed by formalin for overnight and then applied to continuous carbon grids. The grids were transferred into 4% uranyl acetate and incubated for 1 min. After removing the solution, the grids were air-dried at room temperature. For each sample, two-three biological samples were done for taking TEM images by FEI Tecnal G2-20 TEM.
H5N1 pseudovirus assay H5N1 pseudotype virus (38) bearing H5N1 HA and NA was pretreated with PBS or P16 in PBS and then incubated at RT for 1h. MDCK cells were infected with the treated pseudotype virus for 1h. MDCK cells without pseudotype virus infection were served as the baseline control of luciferase protein. After 18h cell culture, cell lysates were collected and the luciferase protein was measured by Luciferase assay system (Promega) in a Victor X3 Multilabel reader (PerkinElmer). The luminescence reading was normalized to 1mg protein.
Virus induced cell fusion assay MDCK cells were transfected with pGFP. Eight hour later, cells were infected with 1MOI of FluB virus. At 18h post infection, cells were treated by PBS or P16 for 1h and then cells were treated by pH5.0 or pH 7.4 for 10 min. After removing the pH buffer, cells were cultured at 37 °C for 4 h with full media. Fusion pictures were taken at 4h after pH treatment.

HA mediated cell fusion assay
The 293T cells were co-transfected with pGFP and pH7N7-HA. At 24h post transfection, cells were treated by PBS or P16 for 1h and then treated by pH5.0 or pH 7.4 for 10 min. After removing the pH buffer, cells were cultured at 37 °C for 4 h with full media. Fusion pictures were taken at 4 h after pH treatment.

Spike-ACE2 mediated cell fusion assay
The pSpike of SARS-CoV-2, pACE2-human, or pGFP were transfected to 293T cells for protein expression. After 24 hours, to trigger the spike-ACE2 mediated cell fusion, 293T-Spike-GFP cell were co-cultured with 293T-ACE2 with or without the supplement of peptide or drug. The 293T-GFP cells were co-cultured with 293T-ACE2 cells as the no-fusion negative control. After 8 h of co-culture, ve elds were randomly selected in each well to take the cell fusion pictures by uorescence microscopes.

Peptide binding assay
Peptides (1.0-3.0 μg per well) dissolved in H 2 O were coated onto ELISA plates and incubated at 4 °C overnight. Then, 2% BSA was used to block plates at 4°C overnight. For virus or spike protein binding to peptides, viruses or spike protein were diluted in PBS and then were added to ELISA plate for binding to the coated peptides at room temperature for 1h. After washing the unbound viruses or spike protein, the bound viruses were lysed by RLT buffer of RNeasy Mini Kit (Qiagen, Cat# 74106) for viral RNA extraction. Viral RNA copies of binding viruses were measured by RT-qPCR. The bound spike protein detected by anti-His-HRP (Invitrogen, Cat# R93125, 1: 2,000) by reading OD 450 .
Endosomal acidi cation assay Endosomal acidi cation was detected with a pH-sensitive dye (pHrodo Red dextran, Invitrogen, Hemagglutination inhibition assay HA titers of H1N1 and FluB viruses were tested by TRBC. Viruses (8HA titer) were premixed with peptides or PBS for 1h and then equal volume of TRBC was added to virus for incubation at room temperature for 30 min. The precipitates of TRBC were record for calculating the HAI activity. TRBC with untreated virus and neutralization antibody from serum were served as negative and positive control of hemagglutination inhibition.

Human Ex Vivo Lung Tissues
Human lung tissues for ex vivo studies were obtained from patients undergoing surgical operations at Queen Mary Hospital, Hong Kong, as we previously described (39). All donors gave written consent as approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (UW13-364). The freshly obtained lung tissues were processed into small pieces and then were infected with SARS-CoV-2 (2 × 10 4 PFU) in 500 µl advanced Dulbecco's Modi ed Eagle's Medium (DMEM)/F12 medium (Gibco, Thermo Fisher Scienti c) supplemented with 100 U/mL penicillin, 100 μg/mL streptomycin and chloroquine (5 µg/ml). After 18 hours, the inoculum was removed and the specimens were washed one time with PBS. The infected human lung tissues were then cultured in 0.5 ml of advanced DMEM/F12 medium with chloroquine (5 µg/ml). Supernatants were collected at 60 hours post inoculation for plaque assays. Antiviral analysis in vivo BALB/c female mice (10-12 weeks for H1N1 virus and 10-12 months for SARS-CoV) and female hamsters (4-6 weeks for SARS-CoV-2) were kept in biosafety level 2 laboratory (housing temperature between 22~25 °C with dark/light cycle) and given access to standard pellet feed and water ad libitum. All experimental protocols followed the standard operating procedures of the approved biosafety level 2 animal facilities and were approved by the Committee on the Use of Live Animals in Teaching and Research of the University of Hong Kong (40). To evaluate the therapeutic effect, mice were intranasally inoculated with 3 LD 50 of A(H1N1) virus or hamsters were intranasally inoculated with 5000 PFU of SARS-CoV-2 and then intranasally inoculated with PBS, P16, zanamivir, or chloroquine at 6-8 h after the viral inoculation. Two more doses were given to the challenged mice on the following day.  could bind to S protein and U5 could block P16 binding to S protein (n = 4). S protein of SARS-CoV-2 and S treated by U5 (S + U5) were added to ELISA plate for binding to peptides coated on ELISA plate. * indicates P < 0.05 when compared with untreated S. (F) U5 showed weaker antiviral activity than that of P16 against SARS-CoV-2 (n = 4). The antiviral activity was measured by plaque reduction assay. ** indicates P < 0.01 when compared with P16. P values were calculated by the two-tailed Student's t test.
Data are presented as mean ± SD of independent biological samples. A