Identification of the SHREK Family of Proteins as Broad-Spectrum Host Antiviral Factors

Mucins and mucin-like molecules are highly glycosylated, high-molecular-weight cell surface proteins that possess a semi-rigid and highly extended extracellular domain. P-selectin glycoprotein ligand-1 (PSGL-1), a mucin-like glycoprotein, has recently been found to restrict HIV-1 infectivity through virion incorporation that sterically hinders virus particle attachment to target cells. Here, we report the identification of a family of antiviral cellular proteins, named the Surface-Hinged, Rigidly-Extended Killer (SHREK) family of virion inactivators (PSGL-1, CD43, TIM-1, CD34, PODXL1, PODXL2, CD164, MUC1, MUC4, and TMEM123) that share similar structural characteristics with PSGL-1. We demonstrate that SHREK proteins block HIV-1 infectivity by inhibiting virus particle attachment to target cells. In addition, we demonstrate that SHREK proteins are broad-spectrum host antiviral factors that block the infection of diverse viruses such as influenza A. Furthermore, we demonstrate that a subset of SHREKs also blocks the infectivity of a hybrid alphavirus-SARS-CoV-2 (Ha-CoV-2) pseudovirus. These results suggest that SHREK proteins may be a part of host innate immunity against enveloped viruses.

To test whether molecules with a similar structure would also have the ability to block virus infectivity, we selected a group of cellular proteins, including CD43, TIM-1, CD34, PODXL1, PODXL2, CD164, MUC1, MUC4, and TMEM123. These proteins have diverse coding sequences, tissue expression patterns, and functionalities, but share a similar structural feature, namely a highly extended and heavily glycosylated extracellular domain. CD43 is a sialomucin transmembrane protein expressed at high levels on T lymphocytes, monocytes, and some B lymphocytes [12]. TIM-1 is preferentially expressed on Th2 T cells, and also contains an extracellular mucin domain and an N-terminal immunoglobulin (Ig)-like domain [13]. Both proteins have previously been shown to block HIV-1 infection of T cells [6,7,14]. CD34, the common surface marker of stem and progenitor cells [15], is also a member of the sialomucin family of proteins, which contain an extensively O-glycosylated

HIV Env Incorporation Assay
HIV-1 particles were produced by cotransfection of HEK293T cells with pHIV-1(NL4-3) DNA (1 µg) plus 100 ng of vector expressing CD43, CD164, TMEM123, CD34, PODXL1, or PODXL2, or 200 ng of vector expressing MUC1 or MUC4. Empty vector was used to keep the amount of DNA equal in each cotransfection. Virus particles were harvested at 48 h and purified through 10% sucrose gradient by ultracentrifugation (10,000× g for 4 h at 4 • C). Particles were resuspended in LDS lysis buffer (Invitrogen) and subjected to Western blot analysis.

Viral Attachment Assay
HIV-1 virus particles produced in the presence of PSGL-1, CD43, CD164, TMEM123, CD34, PODXL1, PODXL2, TIM-1, MUC1, MUC4, or empty vector were incubated with HelaJC.53 cells (prechilled at 4 • C for 1 h) at 4 • C for 2 h. The cells were then washed extensively (5 times) with cold PBS buffer and then lysed with LDS lysis buffer (Invitrogen) for analysis by Western blot. The attachment of HIV-1 virus produced in the presence of TMEM123 was also separately tested by pre-incubating the virus with the mouse monoclonal anti-TMEM123 antibody (297617) (ThermoFisher, Waltham, MA, USA) (25 µg/mL) at 37 • C for 1 h, followed by incubation of the virus with HeLaJC.53 cells at 4 • C for 2 h and lysis of the cells with LDS buffer.

p24 ELISA
Detection of extracellular HIV-1 p24 was performed using an in-house p24 ELISA kit. Briefly, microtiter plates (Sigma-Aldrich) were coated with anti-HIV-1 p24 monoclonal antibody (183-H12-5C) (NIH AIDS Reagent Program). Samples were incubated for 2 h at 37 • C, followed by washing and incubating with biotinylated anti-HIV immune globulin (HIVIG) (NIH AIDS Reagent Program) for 1 h at 37 • C. Plates were then washed and incubated with avidin-peroxidase conjugate (R &D Systems) for 1 h at 37 • C, followed by washing and incubating with tetramethylbenzidine (TMB) substrate. Plates were kinetically read using an ELx808 automatic microplate reader (Bio-Tek Instruments, Winooski, VT, USA) at 630 nm.

SARS-CoV-2 S-Antigen ELISA
The SARS-CoV2-S antigen kit was purchased from Sinobiological (KIT40591). The ELISA procedure was performed according to the manufacturer's manual.

Inactivation of HIV-1 Virion Infectivity by Mucins and Mucin-Like Proteins
Previous studies have demonstrated that deletion of the highly glycosylated extracellular domain of PSGL-1 abolished its antiviral activity [6]. The extracellular domain of PSGL-1 consists of 14-16 decameric repeats (DR) with multiple O-glycosylated threonines (30%) and prolines (10%) [23,24]. DR plays a pivotal role in elongating and strengthening the extracellular N-terminal domain [24]. We found that when the DR domain was deleted from the N-terminus, the anti-HIV activity of PSGL-1 was also abolished ( Figure 1A,C) [6]. These results suggest that the heavily glycosylated, large extracellular domain of PSGL-1 plays a critical role in blocking virion infectivity [6,7].   To determine whether mucins and mucin-like proteins with a similar extracellular structure would also inhibit HIV-1 virion infectivity, we selected and tested a group of mucins and mucin-like proteins, including CD43, TIM-1, CD34, PODXL1, PODXL2, CD164, MUC1, MUC4, and TMEM123 ( Figure 1A). We cotransfected them individually with HIV-1(NL4-3) DNA into HEK293T cells to assemble viral particles, and then quantified the infectivity of the virions produced by infecting an HIV-1 Rev-dependent indicator cell, Rev-A3R5-GFP [6,25,26] (using an equal level of p24). The presence of each of these proteins in virus producer cells blocked the infectivity of the HIV-1 virions ( Figure 1D).
To determine whether transmembrane proteins with large extracellular domains would similarly inhibit HIV-1 infectivity, we selected several non-mucin proteins with various extracellular repeat domains. Both CD2 and ICAM-1 are members of the immunoglobulin (Ig) superfamily of proteins that contain heavily glycosylated extracellular Ig domains (2 and 5 Ig domain repeats for CD2 and ICAM-1, respectively) [27]. ICAM-1 has been shown to enhance virion infectivity in an ICAM-1/LFA-1-dependent fashion [28][29][30][31]. L-selectin and E-selectin are members of the selectin family of cell adhesion molecules that share a similar extracellular structure with a variable number of consensus repeats (2 and 6 for L-and E-, respectively) [32]. We also selected integrin beta chain-2 (ITGB2), which has a large extracellular domain with integrin epidermal growth factor like repeat domains (I-EGF repeats 1 to 4) [33]. Each of these proteins was similarly cotransfected with HIV-1(NL4-3) DNA into HEK293T cells to assemble viral particles. Virions released were harvested and quantified for infectivity in Rev-A3R5-GFP cells (using an equal level of p24). As shown in Figure 1E, none of these proteins, except for E-selectin (Supplementary Figure S1), was found to block the infectivity of HIV-1 virions. This is in contrast with the mucin and mucin-like proteins that we selected; all of these mucin and mucin-like proteins were found to block HIV-1 infectivity ( Figure 1D). These results demonstrate that the ability of mucin and mucin-like proteins to block HIV-1 infectivity is not a shared property of transmembrane proteins expressed in virus producer cells.
We further performed experiments to quantify and compare the anti-HIV activity of the mucin and mucin-like proteins (Supplementary Figures S2-S4). We observed dosagedependent inhibition of HIV-1 by all of them (Figure 2A,B, Supplementary Figures S5-S11), and the 50% inhibition dosages were determined ( Figure 2C). Among these proteins, CD34, the common stem cell maker, had the strongest inhibition, blocking HIV-1 virion infectivity at an IC 50 of 2.33 ng ( Figure 2C). The MUC1 vector was less effective, with an IC 50 of 216.06 ng ( Figure 2C). Based on the antiviral activities and shared structural feature of these proteins, we conveniently named them the Surface-Hinged, Rigidly-Extended Killer (SHREK) family of virion inactivators.  (Figure 2A,B, Supplementary Figures S5-S11), and the 50% inhibition dosages were determined ( Figure 2C). Among these proteins, CD34, the common stem cell maker, had the strongest inhibition, blocking HIV-1 virion infectivity at an IC50 of 2.33 ng ( Figure 2C). The MUC1 vector was less effective, with an IC50 of 216.06 ng ( Figure 2C). Based on the antiviral activities and shared structural feature of these proteins, we conveniently named them the Surface-Hinged, Rigidly-Extended Killer (SHREK) family of virion inactivators.

Effects of SHREK Protein Expression on HIV-1 Virion Release
To elucidate the anti-HIV mechanisms of SHREK proteins, we first examined the effects of SHREK expression on virion release in virus producer cells. Previous studies of PSGL-1 have demonstrated that expression of PSGL-1 in virus producer cells had minimal effects on HIV-1 virion release [6]. To examine effects of SHREK proteins on HIV-1 virion release, we cotransfected HEK293T cells with HIV(NL4-3) proviral DNA (1 µg) plus each of the SHREK protein expression vector at varying inputs from 0.5 to 400 ng ( Figure 3). The release of virions in the presence of each SHREK protein was quantified. All of the proteins except TIM-1 had minor effects on virion release at doses below 100 ng ( Figure  3). CD34, PODXL1, and CD164 [34] inhibited HIV-1 virion release only at high dosages (100 ng and above). TIM-1 potently inhibited virion release at doses as low as 5-10 ng (Figure 3 and Supplementary Figure S12), consistent with a previous report [14]. While not inhibiting virion release at low dosages, all of these proteins, except for MUC1 (IC50, 216.05), effectively blocked virion infectivity with an IC50 at approximately 10 ng and

Effects of SHREK Protein Expression on HIV-1 Virion Release
To elucidate the anti-HIV mechanisms of SHREK proteins, we first examined the effects of SHREK expression on virion release in virus producer cells. Previous studies of PSGL-1 have demonstrated that expression of PSGL-1 in virus producer cells had minimal effects on HIV-1 virion release [6]. To examine effects of SHREK proteins on HIV-1 virion release, we cotransfected HEK293T cells with HIV(NL4-3) proviral DNA (1 µg) plus each of the SHREK protein expression vector at varying inputs from 0.5 to 400 ng ( Figure 3). The release of virions in the presence of each SHREK protein was quantified. All of the proteins except TIM-1 had minor effects on virion release at doses below 100 ng ( Figure 3). CD34, PODXL1, and CD164 [34] inhibited HIV-1 virion release only at high dosages (100 ng and above). TIM-1 potently inhibited virion release at doses as low as 5-10 ng (Figure 3 and Supplementary Figure S12), consistent with a previous report [14]. While not inhibiting virion release at low dosages, all of these proteins, except for MUC1 (IC 50 , 216.05), effectively blocked virion infectivity with an IC 50 at approximately 10 ng and below ( Figure 2B,C). These results suggested that, for most of these SHREK proteins, inactivating virion infectivity rather than blocking virion release is a major mechanism. below ( Figure 2B,C). These results suggested that, for most of these SHREK proteins, inactivating virion infectivity rather than blocking virion release is a major mechanism.

Virion Incorporation of SHREK Proteins and Effects on HIV-1 Env Incorporation
The anti-HIV activity of PSGL-1 is mainly attributed to its incorporation into virion particles that sterically hinders virion attachment to target cells [6,7]. In addition, virion incorporation of PSGL-1 also inhibits the incorporation of the HIV-1 Env protein [6,35]. To study the anti-viral mechanisms of SHREK, we examined virion incorporation of SHERK proteins. For this purpose, we adopted a previously established particle pulldown assay [22], using magnetic beads-conjugated anti-SHREK antibodies to pull down virion particles that express SHREK on the surface ( Figure 4A). Magnetically purified virions were quantified for the presence of HIV-1 p24 in the virions. We were able to demonstrate that the anti-PSGL-1 antibody selectively pulled down p24+ virion particles produced from the PSGL-1 expression cells, but not from the cells co-transfected with the empty vector, recapitulating previous findings that PSGL-1 is incorporated into HIV-1 virion particles [5][6][7]. Using this pull-down assay, we were able to demonstrate that similar to PSGL-1, the anti-CD43, -CD164, -PODXL1, -PODXL2, and -CD34 antibodies selectively pulled down virion particles produced from these SHREK expression cells ( Figure  4A,B), demonstrating virion incorporation of SHREK proteins. The anti-TMEM123 antibody had a weak ability to pull down virion particles produced from the TMEM123 expression cells. We were not able to use the commercial anti-MUC1 and anti-MUC4 antibodies to pull down virion particles. To overcome these technique difficulties, we replaced TMEM123 and MUC1 with hemagglutinin-tagged TMEM123 or MUC1 to assemble viral

Virion Incorporation of SHREK Proteins and Effects on HIV-1 Env Incorporation
The anti-HIV activity of PSGL-1 is mainly attributed to its incorporation into virion particles that sterically hinders virion attachment to target cells [6,7]. In addition, virion incorporation of PSGL-1 also inhibits the incorporation of the HIV-1 Env protein [6,35]. To study the anti-viral mechanisms of SHREK, we examined virion incorporation of SHERK proteins. For this purpose, we adopted a previously established particle pull-down assay [22], using magnetic beads-conjugated anti-SHREK antibodies to pull down virion particles that express SHREK on the surface ( Figure 4A). Magnetically purified virions were quantified for the presence of HIV-1 p24 in the virions. We were able to demonstrate that the anti-PSGL-1 antibody selectively pulled down p24+ virion particles produced from the PSGL-1 expression cells, but not from the cells co-transfected with the empty vector, recapitulating previous findings that PSGL-1 is incorporated into HIV-1 virion particles [5][6][7]. Using this pull-down assay, we were able to demonstrate that similar to PSGL-1, the anti-CD43, -CD164, -PODXL1, -PODXL2, and -CD34 antibodies selectively pulled down virion particles produced from these SHREK expression cells ( Figure 4A,B), demonstrating virion incorporation of SHREK proteins. The anti-TMEM123 antibody had a weak ability to pull down virion particles produced from the TMEM123 expression cells. We were not able to use the commercial anti-MUC1 and anti-MUC4 antibodies to pull down virion particles. To overcome these technique difficulties, we replaced TMEM123 and MUC1 with hemagglutinin-tagged TMEM123 or MUC1 to assemble viral particles, and then used magnetic beads-conjugated anti-hemagglutinin antibody to pull down virion particles. Using this approach, we were able to demonstrate virion incorporation of TMEM123 and MUC1 ( Figure 4A,B). particles, and then used magnetic beads-conjugated anti-hemagglutinin antibody to pull down virion particles. Using this approach, we were able to demonstrate virion incorporation of TMEM123 and MUC1 ( Figure 4A,B). As a control, cells were also transfected with only SHREK-expressing vector (400 ng). Empty vector was used to maintain an equal DNA concentration. For MUC1 and TMEM123, vectors expressing hemagglutinin-tagged TMEM123 or MUC1 were used. Supernatants were harvested at 48 h, normalized for p24, and incubated with magnetic beads coated with antibodies for each individual SHREK or hemagglutinin. Captured particles were washed, eluted, and quantified for the p24 levels with HIV-1 p24 ELISA. Data are presented as the percentage of input particles captured by the beads. The experiment was independently repeated 3 times, and the means ± SD from experiment triplicate are shown. (C) Effect of SHREK proteins on HIV-1 Env incorporation. HEK293T cells were cotransfected with HIV(NL4-3) DNA plus each individual SHREK protein expression vector or an empty vector. Particles were harvested at 48 h, and purified through a sucrose gradient. Virions were lysed and analyzed with Western blot using antibodies against HIV-1 gp41 and p24. Representative blots from 3 independent experiment repeats are shown. The band intensities were quantified from the three blots and normalized for p24. p-values were calculated using the two-tailed T-test. Significance values are indicated using asterisks as follows; * = p < 0.05, *** = p < 0.001, **** = p < 0.0001.
We further examined effects of SHREK incorporation on HIV-1 Env incorporation. Virion particles were assembled in the presence or absence of SHREK proteins, and then analyzed by western blots for HIV-1 Env gp41 and p24. As shown in Figure 4C, all . As a control, cells were also transfected with only SHREK-expressing vector (400 ng). Empty vector was used to maintain an equal DNA concentration. For MUC1 and TMEM123, vectors expressing hemagglutinin-tagged TMEM123 or MUC1 were used. Supernatants were harvested at 48 h, normalized for p24, and incubated with magnetic beads coated with antibodies for each individual SHREK or hemagglutinin. Captured particles were washed, eluted, and quantified for the p24 levels with HIV-1 p24 ELISA. Data are presented as the percentage of input particles captured by the beads. The experiment was independently repeated 3 times, and the means ± SD from experiment triplicate are shown. (C) Effect of SHREK proteins on HIV-1 Env incorporation. HEK293T cells were cotransfected with HIV(NL4-3) DNA plus each individual SHREK protein expression vector or an empty vector. Particles were harvested at 48 h, and purified through a sucrose gradient. Virions were lysed and analyzed with Western blot using antibodies against HIV-1 gp41 and p24. Representative blots from 3 independent experiment repeats are shown. The band intensities were quantified from the three blots and normalized for p24. p-values were calculated using the two-tailed T-test. Significance values are indicated using asterisks as follows; * = p < 0.05, *** = p < 0.001, **** = p < 0.0001.
We further examined effects of SHREK incorporation on HIV-1 Env incorporation. Virion particles were assembled in the presence or absence of SHREK proteins, and then analyzed by western blots for HIV-1 Env gp41 and p24. As shown in Figure 4C, all SHREKs, except for TMEM123, had low-level inhibition of HIV-1 Env incorporation to various degrees; while TMEM123 did not affect Env incorporation, PODXL2 had the strongest inhibition.

SHREK Proteins Inhibit Virus Particle Attachment to Target Cells
We and others previously reported that PSGL-1-mediated inhibition of virion infectivity occurs mainly through virion incorporation, leading to PSGL-1-mediated steric hindrance of particle attachment to target cells [6,7]. We performed a similar virion attachment assay using particles produced from cells expressing each individual SHREK protein.
As shown in Figure 5A, as a control, particles produced from the PSGL-1-expressing cells were highly impaired in their ability to attach to target cells. Similar strong impairment was observed for virions produced from the CD43-, CD34-, and PODXL2-expressing cells. Of note, these four proteins (PSGL-1, CD43, CD34, and PODXL2) have the strongest inhibition of HIV-1 among SHREKs, with an IC 50 around 2 ng ( Figure 2C). The other proteins (CD164, PODXL1, MUC1, and MUC4) also inhibited virion attachment to varying degrees. Interestingly, TMEM123 did not significantly inhibit HIV-1 virion attachment to target cells, although it inhibited viral infection ( Figure 2B,C). One possible reason is that TMEM123 itself may interact with surface proteins, thereby promoting non-productive attachment of virus particles to target cells. Indeed, when the binding assay was performed in the presence of an anti-TMEM123 antibody to block possible TMEM123 interaction with cell surface proteins, virion attachment was decreased ( Figure 5B).
SHREKs, except for TMEM123, had low-level inhibition of HIV-1 Env incorporation to various degrees; while TMEM123 did not affect Env incorporation, PODXL2 had the strongest inhibition.

SHREK Proteins Inhibit Virus Particle Attachment to Target Cells
We and others previously reported that PSGL-1-mediated inhibition of virion infectivity occurs mainly through virion incorporation, leading to PSGL-1-mediated steric hindrance of particle attachment to target cells [6,7]. We performed a similar virion attachment assay using particles produced from cells expressing each individual SHREK protein. As shown in Figure 5A, as a control, particles produced from the PSGL-1-expressing cells were highly impaired in their ability to attach to target cells. Similar strong impairment was observed for virions produced from the CD43-, CD34-, and PODXL2-expressing cells. Of note, these four proteins (PSGL-1, CD43, CD34, and PODXL2) have the strongest inhibition of HIV-1 among SHREKs, with an IC50 around 2 ng ( Figure 2C). The other proteins (CD164, PODXL1, MUC1, and MUC4) also inhibited virion attachment to varying degrees. Interestingly, TMEM123 did not significantly inhibit HIV-1 virion attachment to target cells, although it inhibited viral infection ( Figure 2B,C). One possible reason is that TMEM123 itself may interact with surface proteins, thereby promoting non-productive attachment of virus particles to target cells. Indeed, when the binding assay was performed in the presence of an anti-TMEM123 antibody to block possible TMEM123 interaction with cell surface proteins, virion attachment was decreased ( Figure 5B). HIV-1 p24-normalized viral particles were then assayed for attachment to target HeLaJC.53 cells by Western blot for cell-bound p24. (B) Virions produced in the presence of the TMEM123 expression vector or the empty vector were assayed for attachment in the presence or absence of an anti-TMEM123 antibody. Representative blots from 3 experiment repeats are shown. The band intensities were quantified (for A and B) from the three blots and normalized with GAPDH. p-values were calculated using the two-tailed T-test. Significance values are indicated using asterisks as follows; * = p < 0.05, *** = p < 0.001.

SHREK Proteins Are Broad-Spectrum Host Antiviral Factors
PSGL-1 has been shown to be a broad-spectrum host antiviral factors [6]. To further determine whether these SHREK proteins also possess broad-spectrum antiviral activities, we tested their ability to block other viruses, including influenza A and a new hybrid alphavirus-SARS-CoV-2 pseudovirus [21]. To assemble influenza A virus, eight vectors p-values were calculated using the two-tailed T-test. Significance values are indicated using asterisks as follows; * = p < 0.05, *** = p < 0.001.

SHREK Proteins Are Broad-Spectrum Host Antiviral Factors
PSGL-1 has been shown to be a broad-spectrum host antiviral factors [6]. To further determine whether these SHREK proteins also possess broad-spectrum antiviral activities, we tested their ability to block other viruses, including influenza A and a new hybrid alphavirus-SARS-CoV-2 pseudovirus [21]. To assemble influenza A virus, eight vectors expressing each of the segments of the influenza A/WSN/33 (H1N1) genome plus a GFP-reporter vector were cotransfected with individual SHREK proteins into HEK293T ( Figure 6A). Viral particles were harvested and used to infect target MDCK cells. As shown in Figure 6B, the presence of each of the SHREK proteins in virus-producer cells inhibited the infection of target cells by the virions released. However, the degree of inhibition was different. MUC1, PODXL1, and MUC4 had the strongest inhibition of influenza A virus, although MUC1 and MUC4 were less effective against HIV-1. CD164 and TMEM123 had the weakest inhibition of influenza A. These results demonstrate that SHREK proteins are broad-spectrum and can block the infectivity of multiple viruses. Nevertheless, it is apparent that for each individual SHREK, its antiviral potency can vary among different viruses, with the differences likely related to possible viral antagonisms, the localization of SHREKs and sites of viral budding, and other unidentified factors.

Discussion
In this report, we identified a group of proteins, namely SHREK, that share similar structural characteristics and the ability to broadly inactivate virion infectivity (Figure 1). To test possible effects of SHREK on SARS-CoV-2 infection, we assembled a newly developed hybrid alphavirus-SARS-CoV-2 particles (Ha-CoV-2) [21] in the presence of individual SHREKs, using PSGL-1 as a control. Ha-CoV-2 is a non-replicating SARS-CoV-2 virus-like particle, composed of SARS-CoV-2 structural proteins (S, M, N, and E) and an RNA genome derived from an alphavirus vector. Our recent study has shown that PSGL-1 can block SARS-CoV-2 S protein virion incorporation, virus attachment, and Ha-CoV-2 infection of target cells [36]. We performed similar experiments on the inhibition of Ha-CoV-2 infection by SHREKs, and found that in addition to PSGL-1, CD164, TIM-1, MUC1, and MUC4 also inhibited the infection of Ha-CoV-2 virus ( Figure 6C); The MUC4 vector had the strongest inhibition among these SHREKs. However, we did not find inhibition of Ha-CoV-2 by CD43, CD34, PODXL1, PODXL2, and TMEM123 ( Figure S13), although they are potent inhibitors of HIV-1 ( Figure 2B,C).
We further examined possible effects of SHREK proteins on Ha-CoV-2 S protein incorporation and virion release by quantifying the amounts of S in the supernatant (Supplementary Figure S14). Expression of MUC4 greatly reduced the S protein concentration, likely through blocking virion release or S incorporation; expression of MUC1 or PSGL-1 also led to a roughly 50% reduction of S in the supernatant [36]. For the rest of SHREK proteins, their expression had minor effects on virion release or S protein incorporation.

Discussion
In this report, we identified a group of proteins, namely SHREK, that share similar structural characteristics and the ability to broadly inactivate virion infectivity (Figure 1). We also quantified their antiviral activity in a range of doses (from 0.5 ng to 400 ng of SHREK expression vector), using HIV-1 infection as a model, and observed a strictly dosagedependent inhibition of HIV-1 virion infectivity (Figure 2). The physiological relevance of SHREK proteins such as PSGL-1 in restricting HIV-1 virion infectivity has been previously confirmed in HIV-1 infection of primary blood CD4 T cells, and by shRNA knockdown of PSGL-1 in human CD4 T cells [6]. The slight reduction of endogenous levels of PSGL-1 in transformed and primary CD4 T cells led to an enhancement of virion infectivity, demonstrating that the presence of endogenous PSGL-1 can affect virion infectivity [6].
Although we mainly focused on proteins with mucin and mucin-like domains, we expect that SHREK proteins will likely include other proteins, such as E-selectin ( Figure S1B), with diverse molecular structures. The antiviral activities of SHREK can be achieved through at least three different mechanisms: (1) blocking virion release, e.g., TIM-1 inhibition of HIV-1 release [14]; (2) inhibition of virion incorporation of viral attachment proteins, e.g., PSGL-1 inhibition of HIV-1 gp160 incorporation [6]; (3) virion incorporation of SHREK that blocks progeny virus attachment to target cells through steric hindrance (e.g., PSGL-1 inhibition of HIV-1 virus attachment to target cells [6,7]). Our previous studies have also shown that viruses such as HIV-1 can evolve antagonisms to counteract the blockage imposed by SHREK proteins. For example, HIV-1 uses the accessory proteins Vpu and Nef to degrade and downregulate PSGL-1 on CD4 T cells [5,6]. However, it remains to be determined what other antagonisms may exist, by which different viruses antagonize individual SHREKs.
Previous studies have shown that several of the SHREK proteins, such as PSGL-1 [5], CD43 [37], CD164 [34,38], TMEM123 [39], MUC1 [40], and MUC4 [41], are induced by interferons. However, SHREK proteins in general are likely different from interferoninduced restriction factors, and can be constitutively expressed in specific cell populations. Interestingly, we noticed that multiple SHREKs (CD34, PODXL1, PODXL2, and CD164) identified in this study are heavily expressed on the surface of stem and progenitor cells. In particular, CD34, the common surface marker of stem and progenitor cells, is one of the most potent SHREKs against retroviruses such as HIV-1 in our study ( Figure 2C). It is likely that these SHREKs are a part of host innate antiviral mechanisms that limit retroviral replication in critical cell populations such as stem and progenitor cells. It has been known that HIV-1 can enter and express genes in CD34+ multipotent hematopoietic progenitor Viruses 2021, 13, 832 13 of 15 cells (HPCs), but viral replication is limited [42]. However, viral replication can occur in HPCs with GM-CSF and TNF-α treatment, which induces myeloid differentiation. Such cytokine-induced viral permissiveness coincides with cytokine-mediated CD34 removal from the cell surface [42].
We found that several SHREKs (PSGL-1, CD164, TIM-1, MUC1, and MUC4) inhibited Ha-CoV-2 infection. MUC1 and MUC4 are closely related mucins; MUC4 had the strongest inhibition among these SHREKs in our infection assay. Purified cell-free human breast milk mucins have been shown to possess anti-HIV activity [43]. A recent study has suggested that MUC4 expression plays a protective role in female mice in SARS-CoV infection [44]. In addition, in chikungunya virus (CHIKV) infection, the loss of MUC4 also results in augmented disease during early time points, indicating a possible broad role for MUC4 in viral infection and pathogenesis [44].
We found no inhibition of Ha-CoV-2 by CD43, CD34, PODXL1, PODXL2, and TMEM123 (Supplementary Figure S13), although these proteins are potent inhibitors of HIV-1 ( Figures 2B and 3). It is possible that the difference may result from the different sites of viral assembly and budding. SARS-CoV-2 budding occurs mainly at the membranes of ER-Golgi intermediate compartment [45], whereas HIV-1 buds from the plasma membrane [46]. Because of the differences, it is possible that different sets of cellular proteins may be incorporated into HIV-1 and SARS-CoV-2.
Our identification of the SHREK family of proteins offers potential new antiviral therapeutics that may be developed through the induction and modulation of SHREK activities and the inhibition of viral antagonisms.

Patents
A provisional patent application pertaining to the results presented in this paper has been filed by George Mason University.