SARS-CoV-2 spike ectodomain targets α7 nicotinic acetylcholine receptors

Virus entry into animal cells is initiated by attachment to target macromolecules located on host cells. The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) trimeric spike glycoprotein targets host angiotensin converting enzyme 2 to gain cellular access. The SARS-CoV-2 glycoprotein contains a neurotoxin-like region that has sequence similarities to the rabies virus and the HIV glycoproteins, as well as to snake neurotoxins, which interact with nicotinic acetylcholine receptor (nAChR) subtypes via this region. Using a peptide of the neurotoxin-like region of SARS-CoV-2 (SARS-CoV-2 glycoprotein peptide [SCoV2P]), we identified that this area moderately inhibits α3β2, α3β4, and α4β2 subtypes, while potentiating and inhibiting α7 nAChRs. These nAChR subtypes are found in target tissues including the nose, lung, central nervous system, and immune cells. Importantly, SCoV2P potentiates and inhibits ACh-induced α7 nAChR responses by an allosteric mechanism, with nicotine enhancing these effects. Live-cell confocal microscopy was used to confirm that SCoV2P interacts with α7 nAChRs in transfected neuronal-like N2a and human embryonic kidney 293 cells. The SARS-CoV-2 ectodomain functionally potentiates and inhibits the α7 subtype with nanomolar potency. Our functional findings identify that the α7 nAChR is a target for the SARS-CoV-2 glycoprotein, providing a new aspect to our understanding of SARS-CoV-2 and host cell interactions, in addition to disease pathogenesis.

The novel strain severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has caused a global pandemic of the potentially fatal coronavirus disease 2019 . SARS-CoV-2 has caused more than 650 million confirmed COVID-19 cases and more than 6.65 million deaths worldwide according to the Johns Hopkins Coronavirus Resource Center as of December 14, 2022. COVID-19 is a severe acute respiratory syndrome that initially infects respiratory epithelial cells by binding to the angiotensin converting enzyme 2 (ACE2) but other cellular targets have been proposed (1)(2)(3)(4)(5). SARS-CoV-2 infection results in severe inflammation and damage to peripheral organs (6)(7)(8). There is also increasing evidence that SARS-CoV-2 enters the central nervous system (CNS) and can have long-term effects including memory impairment and fatigue (9,10).
nAChRs are pentameric ligand gated ion channels that are composed of different combinations of α (α2-10) and β (β2-4) subunits (41). α4β2 and α7 nAChR subtypes have been identified to be targets for the glycoproteins of the rabies virus and HIV, respectively (42,43). The rabies virus, HIV, and SARS-CoV-2 glycoproteins all contain a conserved protein sequence similar to loop 2 of α-bungarotoxin (Table S1) (44). This neurotoxin-like region is critical for nAChR interactions with α-bungarotoxin, as well as the rabies virus and HIV glycoproteins (45)(46)(47)(48)(49). The SARS-CoV-2 glycoprotein is located on the surface of the virion, forming a homotrimer spike with two regions S1 and S2 which can interact with cellular targets to mediate host entry (50,51). A structural model of the SARS-CoV-2 glycoprotein shows this neurotoxin-like region, located at the junction between the S1 and S2 segments, is highly exposed to solvent, free of shielding glycans, and accessible to host cellular protein targets (4,52). Functional and binding studies have shown that the SARS-CoV-2 neurotoxin-like region functions either as a coagonist on α7 nAChRs and does not bind to the α7 nAChR orthosteric site, respectively (53,54).
In this work, we set out to determine if the SARS-CoV-2 glycoprotein neurotoxin-like region targets nAChRs and if nicotine modifies this potential interaction. To do so, we used a SARS-CoV-2 glycoprotein peptide (SCoV2P) comprised of residues Y660-S689 of the spike ectodomain (YECDIPIGA-GICASYQTQTNSPRRARSVAS) (Table S1) and the SARS-CoV-2 ectodomain (SCoV2ED) to investigate the functional effects on nAChRs using two-electrode voltage clamp (TEVC) electrophysiology. A neuronal-like cell culture model was used to verify the interaction of SCoV2P with α7 nAChRs, as well as confirm the lack of cytotoxicity for this peptide. The α7, α4β2, α3β4, and α3β2 nAChR subtypes were chosen as these are expressed in SARS-CoV-2 target tissues, including nose, lung, the CNS, and some immune cells (55)(56)(57)(58). We hypothesized that the ectodomain and SCoV2P would antagonize these nAChR subtypes, given the sequence similarities to α-neurotoxins and the rabies glycoprotein. Further, we anticipated that nicotine would enhance this potential inhibition effect.
Using SCoV2P, we defined that the neurotoxin-like region antagonizes nAChRs in a subtype selective manner, showing a high preference for the α7 subtype. Even more surprising, SCoV2P potentiates and inhibits acetylcholine (ACh)-induced α7 nAChR responses by a potential allosteric mechanism and nicotine enhances these effects. We further show that SCoV2P can reset desensitized α7 nAChRs by aiding receptor transition back to the resting closed state. Confocal imagining of α7 nAChR transfected cultured neuronal-like N2a and human embryonic kidney 293 (HEK293) cells confirmed that SCoV2P interacts with α7 nAChRs expressed on the cell surface. Importantly, we identified that the SCoV2ED potentiates and antagonizes the α7 nAChR subtype with nM potency. Our functional and in vitro cell culture findings confirm that the α7 nAChR is a target for the SARS-CoV-2 glycoprotein, providing a new aspect to our understanding of SARS-CoV-2 and host cell interactions.

Heteromeric nAChR SCoV2P inhibition
Modeling data has predicted that the SCoV2ED neurotoxinlike region interacts with α4β2 and α7 nAChRs (4). We performed SCoV2P concentration-response curves using heteromeric α4β2, α3β2, and α3β4 nAChRs and the α7 homomer (presented in the following sections) to test if SCoV2P functionally interacts with these nAChRs. Heteromeric nAChRs express in distinct stoichiometries. High agonist sensitivity isoforms form with two αβ pairs and one additional β subunit. In addition to the two αβ pairs, the low agonist sensitivity isoforms have an accessory α subunit. Preapplication of SCoV2P in absence of ACh did not activate any of the tested nAChRs (Fig. S1A). SCoV2P applied to non-circular RNA (cRNA) injected oocytes did not induce a response in the absence or presence of ACh (Fig. S1B). Application up to 100 μM of SCoV2P caused no change in ACh-induced responses, which was followed by a minimal antagonistic effect with 300 μM SCoV2P applied to heteromeric nAChRs in absence of nicotine (Fig. 1). For α4β2 and α3β2 nAChRs, the low agonist sensitivity isoforms experienced less inhibition than the high agonist sensitivity isoforms. The (α3β4) 2 β4 isoform ACh currents were completely unmodified by SCoV2P application (Fig. 1C).
It is also known that nicotine exposure can increase nAChR expression on the cell surface (63)(64)(65). Although unlikely, if our nicotine pretreatment chaperoned α7 nAChRs to the cell surface, then upon recovery from desensitization, the total ACh peak current would be augmented compared to the nicotine naïve group. To test this, α7 nAChRs were preincubated with 200 nM nicotine and tested as described above. ACh peak currents were recorded every 4 min until the α7 nAChRs recovered from desensitization (24 min) (Fig. S2). We observed no change in total ACh peak current response between the naïve and the recovered 200 nM nicotine treatment groups (Fig. 3B). These findings demonstrate that nicotine preincubation did not alter the total functional receptor pool, thus the SCoV2P potentiation we observed in Figure 2, B and C is likely due to allosteric actions. To determine if SCoV2P is an allosteric modulator of α7 nAChRs, ACh concentration-response curves were generated in the presence of either the maximum potentiating concentration of SCoV2P (10 μM) or a high inhibiting concentration (100 μM) and compared to SCoV2P naïve ACh responses (Fig. 4A). With preapplication of 10 μM SCoV2P, the ACh potency was significantly enhanced from 192 μM [CI 174, 212] to 119 μM [CI 104, 138]), as noted by the leftward shift of the response curve (**p = 0.0060) ( Fig. 4B and Table S3). The 200 nM nicotine pretreatment group also displayed enhanced SCoV2P potency (121 μM [CI 103, 150])) when compared with the ACh-only control (nicotine naïve, no SCoV2P) group (**p = 0.0014) ( Fig. 4B and Table S3). Preapplication of 10 μM SCoV2P increased ACh efficacy, as noted by the upward shift of the response curve, from the α7 nAChR control (99.9 ± 0.3%) to 144 ± 8% (****p < 0.0001). The 200 nM nicotine pretreatment further enhanced the SCoV2P potentiation to 163 ± 18% (****p < 0.0001) ( Fig. 4B and Table S3).

SCoV2P mechanism of action
Being a relatively small peptide, SCoV2P could interact with α7 nAChRs in many ways, including opportunistically. To demonstrate that SCoV2P interactions are mediated by specific residues, we tested the delta variant mutation (P681R). This single point mutation enhanced inhibition phase potency from 337 μM [CI 273, 490] to 100 μM [CI 91,111] without significantly changing potentiation in comparison to SCoV2P (Fig. 5A). These data demonstrate that residue 681 can mediate specific interactions with α7 nAChRs to alter, and in the case of P681R, enhance potency.
To further ensure that the effects observed with SCoV2P were α7 nAChR specific, we blocked SCoV2P potentiation with the competitive antagonist methyllycaconitine (MLA) (66). MLA coapplied with SCoV2P fully prevented α7 nAChRs response to ACh, verifying that the SCoV2P effects are mediated by α7 nAChRs (Fig. 5B). To evaluate if SCoV2P is a silent agonist as recently proposed by Chrestia et al. SARS-CoV-2 spike ectodomain targets α7 nAChRs explored if SCoV2P could induce receptor activation in the presence of the positive allosteric modulator PNU-120596 (53). To test this, we preapplied 3 μM SCoV2P followed by coapplication of 3 μM SCoV2P and 3 μM PNU-120596. We lastly stimulated the α7 nAChRs with ACh to demonstrate that the receptors were functional (Fig. 5C). Following this protocol, we observed no current responses when SCoV2P and PNU-120596 were both present. With application of ACh, the response was massively potentiated demonstrating that the α7 nAChRs were functional. Because SCoV2P did not alter or block the effects of PNU-120596, nor did the combination of these two molecules activate the receptors in the absence of ACh, it is likely that SCoV2P does not share a common binding pocket with ACh or PNU-120596 on the α7 nAChR. As SCoV2P facilitates α7 nAChR recovery from nicotine desensitization (Figs. 2 and 3), we wanted to determine if SCoV2P can directly reactivate desensitized receptors. We applied fully saturating and desensitizing concentrations of ACh (30 mM) for approximately 30 s until the steady state was reached ( Fig. 5D top). We then coapplied ACh with potentiating concentrations of SCoV2P and found that SCoV2P had no effect on the steady-state response (Fig. 5D left). The same experiment was replicated using a lower concentration of ACh (EC 20 60 μM) and again we observed no change in the steady SARS-CoV-2 spike ectodomain targets α7 nAChRs state response (Fig. 5D right). These experimental results demonstrate that for SCoV2P to modulate α7 nAChRs, it must first be able to interact with the receptors prior to application of ACh (Fig. 5D right). Further, when SCoV2P is applied alone to nicotine desensitized receptors, no activation occurs (Fig. S3). Based on the inability of SCoV2P to directly reactivate desensitized receptors (i.e., transitioning from the receptor desensitized state directly to the open state) (Fig. S3) it is likely that SCoV2P is acting allosterically via a mechanism that can transition receptors from the desensitized to the resting state.
To ensure that the effects of SCoV2P on α7 nAChRs is sequence specific, and not a result of opportunistic effects, we used a peptide of the Pasteur lab strain rabies viral glycoprotein (RVG-P) containing the neurotoxin-like motif. This peptide (RVG-P) is the same length as SCoV2P (30-mer), similar amino acid composition and has high sequence homology in the neurotoxin domain (Table S1). Functionally, the RVG-P, unlike SCoV2P, did not potentiate α7 nAChR ACh-mediated responses at low concentrations (Fig. S5). Instead, RVG-P antagonized ACh-induced currents and is approximately 30 times more potent that SCoV2P, with an IC 50 of 10 μM [9,11] than 337 μM [273, 460] for SCoV2P. These large differences in both potentiation and inhibition verify that the SCoV2P's effects on the α7 nAChR are unique.
To determine if Y674-S689 could function as an allosteric modulator, we performed ACh concentration response profiles in presence of 1 μM Y674-S689 (Fig. 6B). We observed significant potentiated shifts in both potency 93 μM [85,102] (****p < 0.0001) and efficacy (127 ± 11) (*p = 0.0378) as compared to peptide naïve α7 nAChRs ( Fig. 6B and Table S4). One difference between our work and Chrestia et al. (2022) is the use of the chaperone protein NACHO (53). When we coexpressed α7 nAChR with NACHO, we observed no difference in ACh potency or maximal currents compared to α7 nAChRs expressed without NACHO (Fig. 6C). This was not unexpected as Xenopus laevis oocytes contain endogenous chaperone proteins, allowing for α7 nAChRs to readily express. Using α7 nAChRs co-expressed with NACHO, the change in ACh potency with application of 1 μM Y674-S689 was the same as α7 nAChR expressed without NACHO ( Fig. 6C and Table S4).
Visualization of SCoV2P interacting with α7 nAChR-positive neuronal-like cells To further demonstrate SCoV2P interacts with α7 nAChRs in a cellular context, we applied FITC N-terminally labeled SCoV2P to N2a and HEK293 cells. N2a cells were chosen because they are neuronal-like and commonly used to study nAChRs, but do endogenously express low levels of nAChRs (67,68), while HEK293 cells are nAChR-negative. N2a and HEK293 cells transiently transfected with α7 nAChR and NACHO were abundantly labeled with FITC-SCoV2P (Fig. 7, A and C). Non-transfected N2a cells show slight fluorescence labeling but this is not unexpected as N2a cells endogenously express nAChRs (Fig. 7B). HEK293 cells, which do not endogenously express nAChRs, display no fluorescence (Fig. 7D). Analysis of corrected total cell fluorescence (CTCF) showed a significant increase in fluorescence for FITC-SCoV2P treated α7 nAChR transfected N2a and HEK293 cells compared to non-transfected cells (Fig. 7E). We evaluated SCoV2P cytotoxicity and found that SCoV2P did not induce cell death at any tested concentration (1-100 μM) (Fig. S6). As dimethyl sulfoxide (DMSO) is cytotoxic, a 15% DMSO control was used to demonstrate detection of dead cells in this assay.
To further validate that SCoV2P interacts with α7 nAChRs expressed on the cell surface, we used α7 nAChRs that have SARS-CoV-2 spike ectodomain targets α7 nAChRs the fluorescent probe pHuji attached to the α7 subunit Cterminal (α7-pHuji) and transfected N2a and HEK293 cells as above. Application of 50 μM FITC-SCoV2P shows colocalization with α7-pHuji nAChRs and further confirms our functional data demonstrating that SCoV2P interacts with α7 nAChRs (Fig. 8).

SCoV2ED potently modulates α7 nAChRs
Many of the commercially available SARS-CoV-2 glycoprotein ectodomains have mutations within the neurotoxinlike region, including the 682 RRAR 685 furin cleavage site, where it is hypothesized, and we have demonstrated (Fig. 5A) nAChR interactions occur (1,4). To ensure that the potential neurotoxin-like region interaction site was intact, we obtained an unmutated, furin-cleavable SCoV2ED and performed modified concentration-response experiments on the lung and CNS abundant α7 subtype (Fig. 9). To provide sufficient time for the ectodomain to interact with surface α7 nAChRs, each oocyte was incubated with a single SCoV2ED concentration (0.01-30 nM) for 5 min, followed by a 1 s ACh EC 90 stimulation (Fig. 9A). The resulting ACh-induced responses were both potentiated and inhibited at lower SCoV2ED concentrations (0.01-1 nM). With application of higher concentrations (3-30 nM), ACh responses were antagonized with a SCoV2ED potency of 12 nM [CI 8, 18] (Fig. 9B).

Discussion
Approximately one-third of COVID-19 patients experience neurological manifestations including delirium, psychiatric disorders, brain inflammation, hyposmia, cognitive deficits, stroke, and nerve damage (69)(70)(71). nAChRs are widely distributed across the CNS and are involved in neuroinflammation, psychiatric disorders, and other conditions (72,73). Identifying host cellular targets of the SARS-CoV-2 spike glycoprotein will aid the development of treatment strategies to help those infected with COVID-19. We have demonstrated SARS-CoV-2 spike ectodomain targets α7 nAChRs a functional interaction of the SARS-CoV-2 spike glycoprotein with nAChRs. Using a peptide of the SCoV2ED neurotoxinlike region, we identified that this region potentiates and inhibits α7 nAChRs in a concentration-dependent manner by an allosteric mechanism, while inhibiting α4β2, α3β4, and α3β2 subtypes minimally. We further demonstrate that SCoV2ED potentiates and inhibits α7 nAChR ACh-mediated currents with nanomolar potency. Nicotine treatments further enhanced the actions of both SCoV2P and SCoV2ED. Understanding how SCoV2P and SCoV2ED affect different nAChR subtypes and associated isoforms is likely relevant to understanding COVID-19 pathophysiology. Prediction that the neurotoxin-like region of the SARS-CoV-2 glycoprotein interacts with nAChRs was initially suggested by Changeux (44). Modeling data has since predicted that this region interacts with α7 and α4β2 nAChRs (4). Interestingly, this in silico study predicted that the neurotoxinlike glycoprotein region interacts with high positional and conformational variability with the α7 and α4β2 nAChR models. In the α4β2 model, the peptide was not able to bind deeply into the proposed orthosteric-binding site, keeping the C loops of the receptor in the open conformation. The ability of the C loop to close over the binding pocket is an important step in achieving gating (74)(75)(76). Agonists are proposed to stabilize a compact C loop conformation, whereas antagonists prevent C loop closure. In the α7 nAChR model, the peptide showed multiple modes of C loop movement, including an open conformation and a semiclosed structure as the peptide moved deeper into the binding pocket (4). These modeling predictions are consistent with our functional data. The heteromeric α4β2, α3β4, and α3β2 subtypes were antagonized slightly with high concentrations of SCoV2P (Fig. 1). However, our α7 nAChR data was more complex, showing that the SCoV2P potentiates ACh-induced currents at low peptide concentrations, followed by inhibition at higher concentrations (Fig. 2). A point of difference is that our functional data show that SCoV2P binds to nAChRs via an allosteric mechanism, rather than at the orthosteric-binding site (Fig. 4). However, it seems possible that since the modeled peptide showed high positional and conformational variability, SCoV2P may interact with nAChRs in multiple orientations and/or sites and thus cause different functional outcomes.
Our data show that SCoV2P is an allosteric modulator of α7 nAChRs. With potentiating concentrations of SCoV2P preapplied prior to ACh, we observed a significant enhancement in ACh potency and efficacy (Fig. 4). These data suggest that SCoV2P can bind to α7 nAChRs and prime them for agonist activation (Fig. 10B). Ligand-driven shifts in ACh potency have previously been identified for ligands interacting with nAChRs allosterically, as performed with PNU-120596 and desformylflustrabromine (66,77). Another clue that SCoV2P is an allosteric modulator can be observed in the SCoV2P concentration response data (Fig. 2). At the peak potentiating concentration (10 μM) of SCoV2P, the enhancement in potentiated response for the nicotine 200 nM and 300 nM treatment groups matches well with the percentage of receptors desensitized (Fig. 2D). As there is no increase in receptor pool due to nicotine incubation (Fig. 3B), these data show that with potentiating concentrations, SCoV2P can transition nicotine desensitized α7 nAChRs back to the resting state (Fig. 10C). We also observed a significant reduction in ACh potency with a high, inhibiting concentration of SCoV2P. These findings suggest that SCoV2P may bind with different orientations and likely occupies an allosteric site on α7 nAChRs to perform its dual functional actions.
Similar allosteric actions of SCoV2P were not observed on heteromeric nAChRs. At high concentrations, SCoV2P inhibited ACh-induced currents of α4β2, α3β2, and α3β4 nAChRs, with some isoform specific effects (Fig. 1). In the absence of nicotine, the α4β2, α3β2, and α3β4 subtypes were minimally inhibited. With pretreatment and cotreatment of 200 nM nicotine, SCoV2P inhibition of (α4β2) 2 α4 and (α3β2) 2 α3 ACh-mediated currents were enhanced. The inhibition of (α3β2) 2 α3 ACh-mediated currents reached similar  (38). The shown schematic is simplified and does not show the multiple ACh-binding sites and associated transition states. B, SCoV2P is an allosteric modulator capable of modifying α7 nAChR ACh potency and efficacy in a concentration-dependent manner. At low peptide concentrations, our data support the conclusion that SCoV2P can transition α7 nAChRs from the resting state to an intermediate state. The structural changes induced by SCoV2P to prime α7 nAChRs for rapid activation may allow conformation changes elicited by an agonist to be additionally stable or occur more rapidly, ultimately resulting in potentiated (O) ACh-mediated responses. C, SCoV2P is also able to transition α7 nAChRs from the nicotine-driven desensitized state back to the closed state. ACh, acetylcholine; SCoV2P, SARS-CoV-2 glycoprotein peptide; nAChR, nicotinic acetylcholine receptor.
SARS-CoV-2 spike ectodomain targets α7 nAChRs levels to the α7 nAChR (40% inhibition of ACh-induced currents). These results, in combination with the α7 nAChR data, suggest that α subunits, and perhaps α/α interfaces, may contain the SCoV2P-binding site. These findings demonstrate that nAChR subtypes interact with the SARS-CoV-2 neurotoxin-like region, which may have cellular consequences, such as mediating viral entry.
While preparing this article, Chrestia et al. (2022) published a nice body of work identifying that the neurotoxin-like region of SARS-CoV-2 (Y674-R685) acts as a silent-agonist to cause α7 nAChR currents when applied with PNU-120596 and inhibits α7 nAChRs by a proposed noncompetitive mechanism (53). To facilitate a more direct comparison to our work, we evaluated the functional outcomes of a shortened peptide containing the last 16 residues of SCoV2P (Y674-S689). We identified that Y674-S689 has similar positive allosteric modulation activities to that of SCoV2P, as it needs an agonist to have its effects (Fig. 6). One potential difference that may explain the macroscopic observational differences between the presented work and Chresita et al. (2022) is that we used a slightly longer peptide. We chose this as these residues were in SCoV2P and are found in the ectodomain. Another difference is the type of TEVC system. We use a relatively fast perfusion system where the oocyte is flushed with fresh drugs or buffer solutions at a rate of 4 ml/min, while Chresita et al. (2022) used a system where the oocyte is dunked into 230 μl of stationary solution and thus the drug to buffer exchange is slower (53). Importantly, both bodies of work contribute to our knowledge regarding how the SARS-CoV-2 glycoprotein interacts and modulates α7 nAChRs. Chrestia et al. (2022) single channel unitary studies identify that the low concentrations of the neurotoxin-like region enhance α7 nAChR transition to the active conformation in the presence of ACh or PNU 120596 (53). Godellas et al. (2022) also released a key paper while the presented work was under review which showed that the SARS-CoV-2 spike protein, specifically regions S375-L390 and Y674-R685, do not compete with ACh, choline or nicotine for binding to the orthosteric site on α7 nAChRs, which is consistent with our conclusions that the mechanism of action is via an allosteric site (54).
The work presented in this article demonstrates at low concentrations SCoV2P and SCoV2ED positively modulate ACh-mediated currents via facilitation of α7 nAChRs transition to the active conformation. The modulation activity switches to inhibition at higher concentrations of both SCoV2P and SCoV2ED. Further, we provide visual confirmation that SCoV2P interacts with α7 nAChRs transfected in neuronal-like N2a and HEK293 cultured cells (Figs. 7 and 8). Cell culture imaging results demonstrate that SCoV2P interacts preferentially to the surface of cells expressing α7 nAChRs, which is consistent with our functional data. Non-transfected HEK293 cells, which do not endogenously express nAChRs, showed little residual fluorescence. These findings demonstrate that α7 nAChRs are cellular targets of the SCoV2P.
COVID-19 disease severity has been linked to an overproduction of pro-inflammatory cytokines promoting what is referred to as a cytokine storm (78). α7 nAChRs are found prevalently in the immune system in a multitude of cell types including macrophages, T cells, dendritic cells, and B-cells (39,79,80). Activation of α7 nAChRs in the cholinergic antiinflammatory pathway suppresses the production of proinflammatory cytokines (81)(82)(83). Cells in the immune system express nAChR subunits including α3, α4, β2, and β3 (84)(85)(86) and applying the noncompetitive antagonist mecamylamine has been shown to reduce the pro-inflammatory response (87).
As α7 nAChR potentiation suppresses the immune system (83), it is tempting to speculate that α7 nAChR positive modulation by the SARS-CoV-2 neurotoxin-like region may allow for rapid and minimally restricted viral replication early in COVID-19 infection due to activation of the cholinergic anti-inflammatory response. In support of this theory, the delta SCoV2P enhanced ACh potentiation in two ways: an increase in total potentiation and a decrease in the concentration required to induce ACh potentiation. The potency for the inhibition phase was also enhanced. Together these factors may provide some explanation as to why patients infected with the SARS-CoV-2 delta variant experience reduced incubation time, worse symptoms, and increased fatality (88)(89)(90). These potential scenarios may be amplified in smokers, as these individuals have enhanced expression of nAChRs, many of which are desensitized. With the ability of SCoV2ED to potentiate α7 nAChRs (Fig. 9) and, in further detail, the SARS-CoV-2 neurotoxin-like region to resensitize nicotine desensitized α7 nAChRs (Fig. 2), tobacco users' cholinergic anti-inflammatory response may be further activated, thus allowing for even higher levels of viral replication to occur. Our data also suggests that with nicotine, the effects of SCoV2P become broader to include α4β2 and α3β2 subtypes in addition to α7 nAChRs. The increased disease severity and mortality rate for smokers may be due to these proposed modulatory effects of the SARS-CoV-2 on the immune system. Additional investigations probing the above speculation could shed light onto disease pathophysiology.
Nicotine pretreatment enhances SCoV2P and SCoV2ED modulation of α7 nAChR responses resulting in enhanced potentiation via a mechanism that resensitizes nicotine desensitized receptors. This conclusion has potentially large epidemiological consequences as there are 1.3 billion tobacco users worldwide according to the World Health Organization. This raises the concern that physiological processes modulated by α7 nAChRs, such as the immune system and the cholinergic anti-inflammatory response, may be further compromised in tobacco users who are infected with SARS-CoV-2. Understanding how the SARS-CoV-2 glycoprotein and the neurotoxin-like region affects different nAChR subtypes and associated isoforms provides an enriched understanding of COVID-19 pathophysiology and will likely facilitate the development of targeted therapeutics.

DNA constructs and cRNA synthesis
Human α3 (NM_000743.5), α4 (NM_000744.5), α7 (NM_000746.3), β2 (NM_000748.2), and β4 (NM_000750.5) nAChR subunit complementary DNAs were generously gifted from Drs Ron J. Lukas and Paul Whiteaker in mammalian expression vectors (pCI [Promega]) for all constructs, except the α7 subtype, which is in pSHE, a modified pGEMHE vector. The DNA plasmid for the chaperone protein NACHO in pREP9 was generously gifted by Dr R. Loring (Northeastern University). α7 pSHE DNA was transferred into pcDNA 3.1(+) by inserting the gene first into pCI using the restriction enzymes XbaI and NotI then into pcDNA3.1(+) using XbaI and XhoI. The α7-pHuji construct was purchased from Invitrogen GeneArt Gene Synthesis Services (Thermo Fisher Scientific) in the pMA-RQ vector and moved into the mammalian expression vector pcDNA3.1(+). DNA constructs were digested with their respective enzymes and separated on a 0.8% ethidium bromide agarose gel. Appropriate vector and insert bands were cut out and purified using the Wizard SV Gel and PCR Cleanup system. All cDNAs were amplified for use by transformations using NEB 5-α competent Escherichia. coli cells (New England Biolabs [NEB]). DNA isolation was accomplished using Qiagen QIAprep Spin Miniprep kits. Extracted DNA was verified by restriction enzyme digest (XbaI and NotI for constructs in the pCI vector and XbaI and EcoRV for constructs in pSHE vector and XbaI, NotI and PvuI for pcDNA3.1 (+)) (NEB) and visualized on a 1% ethidium bromide agarose gel. To achieve transfection-grade constructs for cell culture experiments, α7, α7-pHuji, and NACHO plasmid DNA was amplified using the EndoFree Plasmid Maxi Kit.
In preparation for cRNA synthesis, cDNA subunits in the pCI vector were linearized with the restriction enzyme SwaI, while the α7-pSHE construct used NheI. Linearized DNAs were visualized by gel electrophoresis, treated with proteinase K (NEB) to digest any contaminating proteins, and purified utilizing the Qiagen PCR clean-up kit. Linearized templates were then used to transcribe cRNA using the T7 mMESSAGE mMACHINE High Yield Capped RNA Transcription Kit (Ambion). Confirmation of cRNA purity was accomplished via quantification using the NanoDrop 2000 and electrophoresis gel imaging. Final products were stored at −80 C in subaliquots.

Oocyte preparation and cRNA injection
Stage IV and V X. laevis oocytes were purchased from IACUC-certified Ecocyte Bioscience for human nAChR expression. All efforts were made to minimize animal suffering and to reduce the number of animals used. The α7 homomeric nAChR was expressed in oocytes by injection of the individual subunit. For co-expression of α7 nAChRs with NACHO, a 40 ng α7 subunit to 0.4 ng NACHO cRNA injection ratio was used. To express the desired heteromeric nAChR subtype, cRNAs were injected into oocytes in biased ratios to enhance specific isoform expression. High agonist sensitivity nAChRs received a higher percentage of the β subunit cRNA than the α subunit. The low agonist sensitivity nAChRs were expressed by injecting higher percentage of the α subunit to the β subunit cRNA into oocytes, as outlined in Table S5. Each oocyte received a cRNA injection of 81 nl via impalement using a pulled micropipette with an outer diameter of 40 μm. Oocytes were incubated in buffer (82.5 mM NaCl, 2.5 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 1 mM Na 2 HPO 4 , 5 mM Hepes, 600 μM theophylline, 2.5 mM sodium pyruvate, 50 μg/ml each penicillin, streptomycin, neomycin, gentamycin sulfate, pH to 7.5 using NaOH) at 13 C for 72 to 120 h prior to recording, with daily buffer changes.

Electrophysiology
TEVC electrophysiology was used to measure nAChR function. No less than 72h post-cRNA injection, X. laevis oocytes expressing α7, α4β2, α3β2, or α3β4 nAChRs were voltage clamped at −70 mV using an Axoclamp 900A amplifier (Molecular Devices, LLC). Data acquisition and analysis were performed using pClamp 10.6 software (Molecular Devices; https://www.moleculardevices.com/). Direct current offset was accomplished using a 40 Hz high-pass filter and 10 kHz lowpass Bessel filter. Recording electrodes were pulled from thin wall capillary glass and filled with 3 M KCl. Electrode resistance ranged from 0.5 to 10 MΩ. Oocytes with leak currents below −100 nA were discarded.
Drug solutions were applied to clamped oocytes using a 16 channel, gravity-fed, perfusion system with automated valve control (AutoMate Scientific, Inc). All drug solutions were made in an oocyte ringer 2 (OR 2 ) recording buffer (92.5 mM NaCl, 2.5 mM KCl, 1 mM MgCl 2 ⋅6H 2 O, 1 mM CaCl 2 ⋅2H 2 O, 5 mM Hepes, pH to 7.5 using NaOH) containing atropine sulfate (1.5 μM) and 0.1% BSA. Atropine sulfate was used to block potential response from endogenous muscarinic receptors, and BSA was used to prevent SCoV2P and SCoV2ED from sticking to plastics in the experimental apparatus. Solutions were made fresh daily. The SCoV2ED came prepared in a PBS 7.2 pH buffer, which we diluted to 40% in our OR 2 recording buffer to minimize buffer effects.
nAChR ACh pharmacology is dependent on the individual subunits that assemble to form distinct subtypes and heteromeric receptor isoforms (91,92). ACh concentration-response profiles were generated to ensure the correct nAChR subtype and isoform expression. Potency of ACh was determined by applying increasing concentrations of ACh (0.010 μM -10 mM) for 1 s with an 84 s wash of OR 2 recording buffer in between each drug application (Fig. S4). Using the generated ACh peak current responses, we determined the α7 nAChR ACh potency to be 183 μM, which matches literature values ( Fig. S4A and Table S6) (33).
To determine the effects of the SCoV2P on the tested nAChR subtypes, increasing concentrations of peptide (0.01-300 μM) were preapplied for 30 s followed by 1 s of ACh at the subtype-specific EC 90 concentration, followed by 225 s of OR 2 recording buffer (Table S6). Responses were normalized to the ACh EC 90 prior to the SCoV2P application. Peptide concentration-response profiles were also performed on α7 nAChRs for delta SCoV2P and control peptide RVG-P, as described above (Fig. 5A). For experiments involving nicotine, nAChR-expressing oocytes were preincubated with 100, 200, or 300 nM nicotine for 80 to 110 min. Nicotine was present in all drug and buffer solutions.
To determine the SCoV2P modulation mechanism, α7 nAChR ACh concentration-response profiles were conducted in the presence of 10 or 100 μM SCoV2P. For each drug application, SCoV2P was preapplied for 30 s followed by 1 s of ACh at increasing concentrations (1 μM-10 mM), followed by 89 s of OR 2 recording buffer. Responses were normalized to the ACh EC 100 without peptide preapplication. For experiments involving nicotine, nAChR-expressing oocytes were preincubated for 80 to 110 min with 200 nM nicotine. During these experiments, nicotine was present in all drug and buffer solutions. To establish if nicotine chaperoned α7 nAChRs during our preincubation period, ACh-evoked currents were measured on either naïve or nicotine treated oocytes. To determine the length of time it took to wash off nicotine, 1 s ACh EC 90 currents were measured, followed by 4 min buffer washes. Recovery was determined when successive ACh applications produced equitable responses (difference < 2%).
ACh concentration-response profiles were also performed for 1 μM Y674-S689. Using our standard protocol, 1 μM Y674-S689 was preapplied for 30 s followed by 1 s of ACh at increasing concentrations (1 μM-10 mM), followed by 89 s of OR 2 recording buffer. Responses were normalized to the ACh EC 100 without peptide preapplication. These experiments were also performed for oocytes co-expressing α7 nAChRs and NACHO (Fig. 6).
In order to further elucidate SCoV2P's modulation mechanism, several other experiments were performed. Experiments were performed in gap-free mode with manual valve operation. To verify that SCoV2P effects were mediated by α7 nAChRs, 10 nM MLA was coapplied with 3 μM SCoV2P for 30 s followed by a 2 s ACh EC 90 application. To evaluate whether Y674-S689 exhibited the silent agonism seen by Chrestia et al (53), 1 μM Y674-S689 was coapplied for 30 s with either 3 μM or 10 μM PNU-120596. To demonstrate if SCoV2P shares a common binding site with PNU-120596 or is a silent agonist, 3 μM PNU-120596 was coapplied with 3 μM SCoV2P for 30 s. The resulting modulation of α7 nAChRs was evaluated using an ACh EC 90 -evoked response (2 s). To determine if SCoV2P could enhance steady-state responses, α7 nAChRs were exposed to either 60 μM (EC 20 ) or 30 mM (EC 100 ) for 30 s until steady-state was achieved. In attempts to evoke a response during steady-state, 3 μM SCoV2P was coapplied with ACh. Reactivation of nicotine desensitized receptors experiments were accomplished by preincubating α7-nAChR expressing oocytes as described, followed by a 30 s application of 3 μM SCoV2P and a 2 s ACh EC 90 stimulation.
For evaluation of SCoV2ED effects on α7 nAChRs, each oocyte was placed in incubation buffer with 40% PBS and a single concentration of the ectodomain for 5 min. Oocytes were then rapidly voltage clamped and the α7 nAChRs were acutely stimulated with a 1 s application of ACh (EC 90 ). At each SCoV2ED concentration (0.01-30 nM), 25 to 44 oocytes were used. Responses were normalized to a control group exposed to incubation buffer with 40% PBS without SCoV2ED. Another control group not exposed to PBS was included demonstrating differences in recording buffer composition.
To further explore the potentiation seen at 0.1 nM SCoV2ED, ACh concentration-response profiles were performed. For each drug application, SCoV2ED was preapplied for 60 s followed by 1 s of ACh at increasing concentrations (1 μM-10 mM), followed by 89 s of OR 2 recording buffer. Responses were normalized to the ACh EC 100 without SCoV2ED preapplication. Evaluation of the effects of the SCoV2ED on 200 nM nicotine desensitized α7 nAChRs, oocytes were exposed to 200 nM nicotine for 80 to 110 min, followed by a preincubation of 0.1 nM SCoV2ED with 200 nM nicotine present. The resulting effect on α7 nAChRs was measured at either the ACh EC 20 or EC 90 .

Cell culture and transient transfection
Mouse neuroblastoma N2a cells and HEK293 cells were purchased from American Tissue Culture Collection and cultured in Eagle's Minimum Essential Medium with 10% fetal bovine serum and 1× penicillin-streptomycin (VWR). Cultures were maintained at 37 C and 5% CO 2 in a humidified cell culture incubator. Cells were subcultured onto poly-Dlysine-coated glass coverslips (MatTek) and transiently transfected when approximately 70% confluent, using the Lipofectamine 2000 transfection reagent (Invitrogen) for N2a cells or Calcium Phosphate Transfection Kit (Millipore Sigma) for HEK293 cells. Mammalian α7 nAChR or α7-pHuji nAChR coding plasmid DNA and NACHO chaperone plasmid DNA were used at a 4:1 ratio following the manufacturers' protocols. Transfected cells were incubated for 24 h (N2a) or 48 h (HEK293) before addition of peptide.

Fluorescent peptide imaging
To assess interactions with N2a or HEK293 cells, 50 μM FITC-tagged SCoV2P was added to the growth medium 24 h or 48 h post transfection and incubated for an additional 24 h. Cells were extensively rinsed with pH 7.4 PBS three times before being imaged on the Olympus Fluoview FV10i Laser Scanning Confocal Microscope and processed using ImageJ (https://imagej.nih.gov/ij/download.html). For imaging of FITC tags, the excitation wavelength was 495 nm, while the emission wavelength was 519 nm. Cells transfected with α7-pHuji were also imaged at 566 nm excitation wavelength and 598-nm emission wavelength. FITC and pHuji bleedthrough were not observed at either emission wavelength. For colocalization imaging, images were taken at the cell surface level to show surface interactions. ImageJ was used for postprocessing of images (NIH) (94), including deconvolution using the DeconvolutionLab2 and PSF Generator plug-ins (both Biomedical Imaging Group, EPFL).
To determine CTCF with ImageJ, we calculated the mean integrated density of fluorescence of our cell groups, using 2D images, by free-hand tracing the outlines of cells of interest in the phase channel, to reduce operator bias. Area, mean, and integrated density were measured in the channel of the fluorescent tag. While viewing only with the phase channel on, visually healthy cells were selected. Three separate regions of interest were traced surrounding each selected cell to account for background fluorescence. Using these values, CTCF was calculated using the following formula: CTCF = integrated density -(area of cell x mean background fluorescence) (http://theolb.readthedocs.io/en/latest/imaging/measuringcell-fluorescence-using-imagej.html#measuring-cell-fluorescenceusing-imagej). This process was repeated for 30 cells for each group.

Cell viability assay
N2a cells were plated on 48-well plates (VWR) and α7 nAChR-positive groups were transfected as described above. Twenty-four hours post transfection, cells were exposed to varying concentrations of SCoV2P and incubated for another 24 h. As positive control for cell death, 15% DMSO was added to control wells instead of peptide. Media was changed post incubation and alamarBlue cell viability (Thermo Fisher Scientific) reagent was added to all wells according to the manufacturer's suggestions. Cells were incubated for 3 h before fluorescent measurements were collected on a Tecan Spark Multimode Microplate Reader (Tecan U.S., Inc).

Data analysis
All TEVC experiments were conducted on at least two oocyte isolations and two batches of cRNA synthesis. Throughout the article, large N indicates the number of replicates and small n indicates the number of individual oocytes. All data analysis was accomplished using GraphPad Prism 9.3 software (www.graphpad.com). TEVC EC 50 , IC 50 , Hill slopes (n H ), and ACh EC 20 and EC 90 current amplitude (I EC20 or I EC90 ) values were determined from individual oocytes. All generated curves were calculated using standard built in nonlinear curve fitting. A sum of squares F-test was used to determine whether an unconstrained monophasic sigmoidal or a constrained biphasic logistic equation best fit the ACh concentration-response data. The inhibitory phase of the SCoV2P concentration-response profiles were fit to monophasic nonlinear curves. SCoV2ED inhibition data was fit with a monophasic nonlinear regression curve with a constrained hill slop (n H = −1). Statistical analyses were performed using GraphPad Prism 9.3. Data was analyzed using Welch's two tailed t test to compare pairs of groups. One-way ANOVA with Tukey's or Dunnett's multiple comparison test were used to evaluate the means of three or more groups.

Data availability
All data are available in the main text or the supplementary materials.
Supporting information-This article contains supporting information.