key: cord-0963039-f6fhtk10
authors: Hancock, Trevor J.; Hickman, Peyton; Kazerooni, Niloo; Kennedy, Melissa; Kania, Stephen A.; Dennis, Michelle; Szafranski, Nicole; Gerhold, Richard; Su, Chunlei; Masi, Tom; Smith, Stephen; Sparer, Tim E.
title: Possible Cross Reactivity of Feline and White-tailed Deer Antibodies Against the SARS-CoV-2 Receptor Binding Domain
date: 2022-02-14
journal: bioRxiv
DOI: 10.1101/2021.12.17.473265
sha: 1e9d343c5ab77ae55cc1f6817dc5216669ded7f1
doc_id: 963039
cord_uid: f6fhtk10

In late 2019, a novel coronavirus began circulating within humans in central China. It was designated SARS-CoV-2 because of its genetic similarities to the 2003 SARS coronavirus (SARS-CoV). Now that SARS-CoV-2 has spread worldwide, there is a risk of it establishing new animal reservoirs and recombination with native circulating coronaviruses. To screen local animal populations in the United States for exposure to SARS-like coronaviruses, we developed a serological assay using the receptor binding domain (RBD) from SARS-CoV-2. SARS-CoV-2’s RBD is antigenically distinct from common human and animal coronaviruses allowing us to identify animals previously infected with SARS-CoV or SARS-CoV-2. Using an indirect ELISA for SARS-CoV-2’s RBD, we screened serum from wild and domestic animals for the presence of antibodies against SARS-CoV-2’s RBD. Surprisingly pre-pandemic feline serum samples submitted to the University of Tennessee Veterinary Hospital were ∼50% positive for anti-SARS RBD antibodies. Some of these samples were serologically negative for feline coronavirus (FCoV), raising the question of the etiological agent generating anti-SARS-CoV-2 RBD cross-reactivity. We also identified several white-tailed deer from South Carolina with anti-SARS-CoV-2 antibodies. These results are intriguing as cross-reactive antibodies towards SARS-CoV-2 RBD have not been reported to date. The etiological agent responsible for seropositivity was not readily apparent, but finding seropositive cats prior to the current SARS-CoV-2 pandemic highlights our lack of information about circulating coronaviruses in other species. Importance We report cross-reactive antibodies from pre-pandemic cats and post-pandemic South Carolina white-tailed deer that are specific for that SARS-CoV RBD. There are several potential explanations for this cross-reactivity, each with important implications to coronavirus disease surveillance. Perhaps the most intriguing possibility is the existence and transmission of an etiological agent (such as another coronavirus) with similarity to SARS-CoV-2’s RBD region. However, we lack conclusive evidence of pre-pandemic transmission of a SARS-like virus. Our findings provide impetus for the adoption of a One Health Initiative focusing on infectious disease surveillance of multiple animal species to predict the next zoonotic transmission to humans and future pandemics.

In late 2019, a novel coronavirus began circulating within humans in central China. It 24 was designated SARS-CoV-2 because of its genetic similarities to the 2003 SARS coronavirus 25 (SARS-CoV) . Now that SARS-CoV-2 has spread worldwide, there is a risk of it establishing 26 new animal reservoirs and recombination with native circulating coronaviruses. To screen local 27 animal populations in the United States for exposure to SARS-like coronaviruses, we developed 28 a serological assay using the receptor binding domain (RBD) from SARS-CoV-2. SARS-CoV-29 2's RBD is antigenically distinct from common human and animal coronaviruses allowing us to 30 identify animals previously infected with SARS-CoV or SARS-CoV-2. Using an indirect ELISA 31 for SARS-CoV-2's RBD, we screened serum from wild and domestic animals for the presence of 32 antibodies against SARS-CoV-2's RBD. Surprisingly pre-pandemic feline serum samples 33 submitted to the University of Tennessee Veterinary Hospital were ~50% positive for anti-SARS 34 RBD antibodies. Some of these samples were serologically negative for feline coronavirus 35 (FCoV) , raising the question of the etiological agent generating anti-SARS-CoV-2 RBD cross-36 reactivity. We also identified several white-tailed deer from South Carolina with anti-SARS-37

CoV-2 antibodies. These results are intriguing as cross-reactive antibodies towards SARS-CoV-2 38 RBD have not been reported to date. The etiological agent responsible for seropositivity was not 39 readily apparent, but finding seropositive cats prior to the current SARS-CoV-2 pandemic 40 highlights our lack of information about circulating coronaviruses in other species. 41 Importance Introduction Silver Stain Kit, Thermo Scientific, USA) 12% SDS-PAGE gel using a standard curve of BSA 176 (bovine serum albumin). Analysis was performed using Image Studio Lite ver. 5.3 Biosciences, Lincoln, NE, USA). 178 Table 1 ). Twenty cat samples were grouped into FCoV positive and 198 negative groups based on feline infectious peritonitis (FIP) serology using an 199 immunofluorescence assay (IFA) against FIP serotypes I and II, as well as TGEV (transmissible 200 gastroenteritis virus) (VMRD, Pullman, WA, USA). Normal cat serum was purchased from 201 Jackson ImmunoResearch (West Grove, PA, USA). Tennessee-resident cows (n=33) and tigers 202 (n=9) were collected pre-pandemic for routine diagnostic testing. Post-mortem, post-pandemic 203 samples were collected from East Tennessee elk (n=12) and South Carolina white-tailed deer 204 (n=22). 205

Anti-RBD ELISA was based on the published protocol by Amanat et al. and Stadlbauer 207 et al. [30, 59] . Purified RBD was diluted to 2ug/mL in PBS and 50uL was placed into each well 208 of a 96 well plate (Immulon 4HBX, Thermo Fisher, USA) and allowed to incubate overnight at 209 4 0 C. Unbound RBD was removed and wells were washed 3x with PBS-T (PBS with 0.1% 210 . Rinsed wells were blocked with 5% milk in PBS for 2 hours at room temp. Block 211 was removed and serum or plasma samples were added at 1:50 dilution for the initial screen in 212 PBS with 1% milk and incubated at room temp for 1 hour. After 1 hour, wells were washed 3x 213 with PBS-T and a secondary antibody for that species was added (i.e., HRP goat-anti-human 214 

For dot blots, 5-10uL of sample was applied directly onto nitrocellulose membranes and 235 allowed to dry. Western blots were loaded with 30uL (~3ug) of purified recombinant RBD, 236 resolved in a 12% SDS-PAGE gel and transferred to a nitrocellulose membrane. Blots were blocked overnight at 4 0 C with 5% milk in PBS. Mouse anti-6His-HRP conjugated antibody 238

(1:5,000) (Proteintech, Rosemont, IL, USA) or polyclonal serum samples (1:20) were incubated 239 with the blots at room temperature for 2 hours and subsequently washed 2x with TBS-T (tris 240 buffered saline with 0.1% Tween-20). For polyclonal serum, species specific HRP anti-IgG 241 antibodies (1:5,000 dilution) were incubated for an additional 2 hours and washed 2x as above. 242

Chemiluminescent substrate (Pierce SuperSignal West Pico PLUS, Thermo Fisher, USA) was 243 added and luminescence was detected using BioRad ChemiDoc (Bio-Rad, Hercules, CA, USA). 244

245 Serum samples were screened for neutralization using LEGENDplex SARS-CoV-2 246 neutralizing antibody assay (BioLegend, San Diego, CA, USA) following manufacturer 247 recommendations. Briefly, serum was diluted 1:100 and incubated with biotinylated SARS-CoV-248 2 S1 subunit containing the RBD and human ACE-2 (hACE-2) conjugated to fluorescent beads. 249

Streptavidin-PE (phycoerythrin) was added to detect SARS-CoV-2 S1 subunit bound to Table 2 ). 281

All graphs and statistical analysis were performed in GraphPad Prism ver. 9 (GraphPad 283 Software, San Diego, CA, USA). ELISA OD 450 results were normalized to an inter-plate 284 replicate run with all assays. Student's one-tailed t-tests with Welch's correction and one-way 285 ANOVA with multiple comparisons tests were performed on ELISA results and documented in 286 the respective figure legends. Descriptive statistics were provided for each ELISA group (mean, 287 median, and quartiles). Receiver operator curve (ROC) analysis was performed to determine 288 appropriate threshold values for human, cat, and deer serum samples. Area under the curve was 289 calculated for each titrated ELISA sample and graphed. Neutralization data was normalized with 290 negative control group (normal cat serum) representing 100% MFI. 291 292 Data will be made publicly available upon publication and upon request for peer review. 293 294

We developed an in-house ELISA to serologically screen companion animals based on a 296 protocol developed at Mt. Sinai [30, 59] . To examine cross-reactivity of our in-house anti-SARS-297

CoV-2 RBD indirect ELISA, we used polyclonal guinea pig serum raised against different 298 animal coronaviruses ( Fig 1A) . Consistent with previous reports, no cross-reactive antibodies for 299 any of the common coronaviruses were found [18, 32, 38, 56] . Only antibodies from SARS-CoV 300 or SARS-CoV-2 infected individuals reacted (Fig 1A, 1B) . Human serum collected from 301 individuals prior to the SARS-CoV-2 pandemic or plasma from recovered SARS-CoV-2 donors 302 were used to validate our ELISA screen ( Fig 1B) . ROC analysis determined the positive cutoff 303 threshold, using a value that gave highest specificity and sensitivity with pre-pandemic human 304 serum and serum from confirmed SARS-CoV-2 infected individuals. ROC analysis was in 305 agreement with the commonly used threshold determination method of three standard deviations 306 above the mean negative value. Our assay based on RBD screening showed high sensitivity 307 (96.96%) and specificity (95.45%) with 66 SARS-CoV-2 samples and 22 pre-SARS-CoV-2 308 samples ( Fig 1B) . While Stadlbauer et al. performed two diagnostic ELISAs, one with RBD and 309 the other with full-length spike, our results using only the RBD-based screen are in good 310 agreement with their published data. Others have also demonstrated the accuracy of an RBD-311 only based ELISA [33, 38] . A western blot using an anti-6His antibody (Fig 1C) shows the 312 expected size of purified RBD with a single band ~32kDa. This shows that our isolated RBD is 313 the correct size and runs as a monomer. Silver stain of the same affinity-purified SARS-CoV-2 314 RBD demonstrates relative purity ( Fig 1D) . However, there are co-purified proteins present at 315 lower levels. 316

To establish a baseline for future SARS-CoV-2 screening of companion animals, 128 317 pre-pandemic feline serum samples collected prior to December 2019 were retrospectively 318 screened using our in-house ELISA. Nineteen samples were of a known FCoV serological status, 319 with the remaining 109 of unknown FCoV status. Following the same protocol used for 320 screening human serum samples (Fig. 1 ), feline samples were tested for antibodies against 321 SARS-CoV-2's RBD (Fig 2A) . There were two batches tested. Serum samples from feral cats in 322

East Tennessee collected from 2007 to 2012 (2007-2012) (n=36) and convenience samples from 323 client-owned cats undergoing routine blood work (listed as Pre-pandemic) (n=92) (Fig 2A) . As 324 expected, SARS-CoV-2 experimentally infected cats [14] tested positive with high relative 325 OD 450 , and normal cat serum (i.e., negative control) with very low relative OD 450 (Fig 2A) . 326

Despite pre-dating the pandemic, 52% (67/128) of the cat samples tested positive for antibodies 327 against SARS-CoV-2 RBD. This is surprising as there was a lack of high cross-reactivity in 328 guinea pigs immunized with FCoV in Fig 1A. Several reports also showed a lack of similarity 329 and cross reactivity between alpha coronaviruses and SARS-CoVs [18, 32, 38, 56] . Indeed, two 330 other groups found that pre-existing immunity to FCoV had no impact on seropositivity of feline 331 samples [38, 58] . To ensure that the positive ELISA results were specific to the RBD and not to a 332 co-purified protein, a western blot was carried out using serum from a positive sample (Fig 2B) . 333

Positive cat serum bound a ~32 kDa protein, the size of the RBD protein ( Fig 1C) . Notably, 334 normal cat serum did not react with any other protein despite the presence of co-purified 335 proteins. To further show the specificity of the anti-RBD response, we titrated seropositive and 336 seronegative samples. Starting with serum from cats experimentally infected with SARS-CoV-2 337 (Fig 2A) and normal cat serum, saw a normalized OD 450 >3 standard deviations above the 338 negative control (i.e., normal cat serum) at all dilutions. This gives a titer >8100 (Fig 2C) . 17 339 seropositive and 10 seronegative pre-2020 cat samples were titrated and assayed in our ELISA 340 Table 1 . Both seropositive and seronegative samples had a 351 mean age of >3 years with no difference between the groups and contained similar ratios of 352 male: female animals (Table 1) . Seropositive samples were found in disparate geographic 353 locations from opposite coasts of the United States (i.e., New York to California (Table 1) ). This 354 observation indicates that seropositivity is not confined to a single geographic region (e.g., East 355 Tennessee). Based on our limited sampling, we were unable to identify any unique characteristic 356 or identifier for seropositive vs seronegative samples. 357

With our discovery of pre-existing antibodies against SARS-CoV-2's RBD, it was 358 pertinent to examine samples from dogs, another companion animal with high human contact. 359

Serum samples from dogs (n=36) were collected and retrospectively screened as part of a tick study during a 7-month period beginning in Jan 2020 and extending into July 2020. These 361 samples are considered post-pandemic because the timeframe straddles the arrival of SARS-362

CoV-2 in East Tennessee (~March 2020). The initial ELISA screen identified 97% seropositivity 363 in the dog samples (Fig 3A) with only 1 sample falling below the cutoff established on human 364 serum. Surprisingly, serum from purpose-bred research animals housed at the University of 365

Tennessee also showed high levels of reactivity (Fig 3A) . This raised suspicion about the 366 specificity of the response. To address this, western blot analysis with canine serum (Fig 3B)  367 identified a protein other than the RBD (see the ~32 kDa protein in Fig 1C and 2B) . The canine 368 serum recognized a ~60 kDa protein which is likely a co-purified protein present after RBD 369 purification and is faintly seen in the silver-stained gel in Fig 1D. This co-purified protein was 370 not detected in the blots performed for American cats, we began examining other regional animals. Serum from Tennessee resident, pre-376 pandemic cows (n=33) and tigers (n=9), post-pandemic East Tennessee elk (n=12), and post-377 pandemic South Carolina white-tailed deer (n=22) were tested for anti-SARS-CoV-2 RBD 378 antibodies (Fig 4A) . Of the four species tested, only the deer from South Carolina showed any 379 seropositive samples (2/22). Serum titrations show the two seropositive samples have a high titer 380 >8100 (Fig 4B) , and AUC of the titrations show a significant difference between seropositive 381 and negative deer samples ( Fig 4C) . Unfortunately, due to limited sample volume, we were 382 unable to run western blots to demonstrate the specificity for the RBD protein. The deer are post-383 pandemic and could represent recent transmission of SARS-CoV-2 into the deer population. 384

Although these animals probably have had limited contact with humans, white-tailed deer are 385 susceptible to and capable of transmitting SARS-CoV-2 [64]. Another possibility is that this 386 species was exposed to the same etiological agent as our pre-pandemic seropositive cats. 387

To address whether our ELISA-positive animal samples can neutralize SARS-CoV-2 388 infections, we measured the ability of cat serum to block the interaction of the spike protein with 389 the human ACE-2 (hACE-2) receptor using a commercially available flow cytometry-based bead 390 assay. In this assay, neutralization is characterized as the decrease in fluorescence when antibodies block the fluorescently labeled SARS-CoV-2 S1 subunit from binding to hACE-2 392 conjugated beads (Fig 5A) . Because this assay is not species specific or immunoglobulin type 393 dependent, it is applicable for assessing both human and feline serum. The internal antibody 394 control shows a decrease in fluorescence corresponding to levels of neutralizing monoclonal 395 antibody against SARS-CoV-2. Serum from experimentally infected cats showed potent 396 neutralization at 1:100 dilution. However, only one ELISA-positive, pre-pandemic cat sample 397 showed neutralization (Fig 5B) . One of the seropositive white-tailed deer samples, and a single 398 serum sample from mice immunized with PRCoV also showed slight neutralization, clearing the 399 determined ROC threshold/cutoff value (Fig 5B) . Notably, we were unable to detect high levels 400 of neutralization/neutralizing antibodies even in several of the human convalescent serum 401 samples ( Fig 5B) . 402

Because cross-reactivity of antibodies to SARS-CoV-2 RBD independent of SARS-CoV-403 2 infection has not been previously reported in felines, we suspected that the etiological agent 404 could be another coronavirus [38, 58] . Fecal samples were collected from healthy East 405 Tennessee cats and screened for coronaviruses using pan-coronavirus primers amplifying 406 conserved regions of the RNA-dependent RNA polymerase (RdRp), helicase (Hel), and spike (S) 407 genes [60] . Coronavirus viral RNA, whether common animal coronavirus or SARS-like 408 coronavirus, is potentially shed in feces [8, 9, 11, 13] . Collection of fecal samples represented a 409 non-invasive collection method, and SARS-CoV-2 has been reported to have prolonged shedding 410 within fecal samples of humans [8, 9, 11, 13] . Fifteen out of thirty samples (50%) tested positive 411 for at least one loci, with most yielding positive results for multiple loci (Table 2) . Not 412 surprisingly, sequences cluster within the alpha-coronavirus group and with high similarity to 413 previously identified FCoV strains. When all five loci were aligned and concatenated together, 414 the Maximum-Likelihood phylogenetic tree places the concatenated coronavirus sequences 415 within the alpha-coronavirus lineage, closely related to FCoV (Fig 6A) . We were unable to 416 amplify or identify any sequences which resemble SARS-like coronaviruses or beta-417 coronaviruses. Partial sequencing of the S1 region was able to amplify the RBD from several 418 coronavirus RNA positive samples. The sequenced RBDs were again highly similar to FCoV 419 based on Maximum-Likelihood phylogenetic tree (Fig 7A) . Along with the phylogenetic tree, a 420 similarity matrix demonstrates high RBD similarity between previous FCoV strains and those 421 sequenced here (~80%) (Fig 7B) . The RBD from these fecal samples displays low similarity to 422 betacoronaviruses such as MERS, SARS-CoV, and SARS-CoV-2 (~30%) as previously reported 423 ( Fig 7B) . while still maintaining sensitivity. A downside of this methodology is the lack of an up-to-date 433 picture of cross-species transmission, as serology trails initial infections by several days to weeks 434 [22, 33] . On the other hand, due to the lowered cost of serological testing there is a compensatory 435 increase in testing capability allowing a broader swath of animals and regions to be sampled with 436 more frequent re-sampling to track spillover into new species. Our adapted protocol yields 437 recombinant SARS-CoV-2 RBD protein allowing production of a low-cost indirect anti-RBD 438 ELISA. The recombinant RBD was sufficient for serological screening via ELISA and is 439 amenable to most labs with prior tissue culture capabilities and does not require large initial 440 investments in cell lines, culture media, or specialized incubators. We validated our method 441 demonstrating low cross-reactivity with other common animal coronaviruses ( Fig 1A) and >95% 442 sensitivity and specificity on human serum samples (Fig 1B) . 443

For the seropositive samples identified in our study, mean titers for positive cat samples 444 were relatively high at ~2700 (Fig 2E, G) , which based on reported rapid declines in anti-RBD 445 responses for SARS-CoV and FCoV points to exposure within the past few years [37, 38, 40] . 446 Further, based on FCoV studies, animals with high titers typically correlate with active viral 447 shedding and spread within a household, which highlights a potential overlap between 448 seropositivity and viral shedding [37] . This is in stark contrast to SARS-CoV-2 serosurveys on 449 pre-and post-pandemic feline samples from Central China. They found no evidence of exposure 450 before the outbreak, but also positivity levels post-pandemic were significantly lower than shown 451 here (~12% Central China vs >50% USA) (Fig 2A) [38, 58] . OD 450 and titers of pre-pandemic 452 seropositive cat samples, while high, were lower than SARS-CoV-2 experimentally inoculated cats (6 weeks post infection) (Fig 2A, C, E, G) . This likely represents a natural decline in titer 454 over time for the environmental samples but could also represent lower titers of cross-reactive 455 antibodies from another coronavirus. 456 Unfortunately, dog serum was shown to bind to a co-purified protein (Fig 3B) , leaving us 457 unable to utilize our assay for examining cross-species transmission of SARS-like coronaviruses 458 to canines. We can show that recombinant RBDs produced and purified by groups at both Mt. 459

Sinai and Emory contain co-purified proteins at approximately the same size as shown in Fig 1A  460 [30, 33]. As such, screens for SARS-CoV-2 exposure in canines would likely require producing 461 and purifying the RBD using a different strategy that eliminates non-RBD protein contamination. 462

Recently both SARS-CoV-2's RBD and soluble full-length spike have been produced and 463 purified in plants [65] . This alternative method may prove useful for animal SARS-CoV-2 464 screening by reducing or eliminating false positives due to co-purified proteins. 465

After the discovery of pre-pandemic seropositive cats, we examined other commercial 466 (cow) and convenience samples from local wild species (deer, elk, tiger) ( Fig 4A) . Two out of 467 twenty-two (9%) white-tailed deer from South Carolina were positive for antibodies against the 468 RBD ( Fig 4B) . Unlike our cat samples, the two seropositive deer could represent transmission of 469 SARS-CoV-2 into the local deer population because these samples were collected post-470 pandemic. Interestingly, a recent report showed that ~40% of white-tailed deer from 4 states 471 (Illinois, Michigan, New York, and Pennsylvania) were positive for SARS-CoV-2 antibodies 472 [29] . Seropositive animals were only observed from 2019 onward, with pre-pandemic deer 473 testing negative on their SARS-CoV-2 neutralization assay. This information supports our 474 finding in South Carolina deer (Fig 4) . SARS-CoV-2 sequences were also recently isolated from 475 the retropharyngeal lymph nodes of wild and captive deer [27, 28] . The dominant genotype of 476 deer-isolated SARS-CoV-2 genotypes closely corresponded to those circulating within humans 477 at the time, pointing to potential rapid transmission from humans to animals [28] . This highlights 478 the importance of the One Health Initiative to provide information on the potential exposure and 479 spillover into other species and whether there is recombination with native coronaviruses 480 occurring to generate new variants or establishment of new reservoirs in North America. Further 481 work is needed to determine the prevalence, spread, and identity of SARS and other 482 coronaviruses circulating within North American deer and associated species. 483

That samples from cats experimentally infected with SARS-CoV-2 displayed potent 484 neutralization (Fig 5B) is unsurprising because of their high ELISA titers (Fig 2C) . These 485 samples were collected at ~8 weeks post infection and likely represent peak titer and 486 neutralization capacity [14] . Neutralization of SARS-CoV-2 S1 subunit was variable for 487 environmental feline ELISA-positive samples (Fig 5B) . There was no significant difference in 488 MFI/neutralization between the anti-RBD seropositive and seronegative feline samples (Fig 5B) . 489

Even sera from convalescent, COVID-recovered individuals showed none to minimal 490 neutralization (Fig 5B) . Several groups have found that not all anti-RBD responses generate 491 neutralizing antibodies [50] [51] [52] [53] 66] . Indeed, even in convalescent serum, high levels of RBD 492 recognition does not guarantee high neutralizing titers, consistent with our own observations (Fig  493   5B ) [50] . Based on ELISA and neutralization results (Fig 2A, Fig 5B) , we suspect that these 494 animals contain antibodies recognizing SARS-CoV-2's RBD, but likely bind to non-neutralizing 495 epitopes of the RBD domain. 496

Because ~50% of cats surveyed were seropositive, we reasoned that isolation of the 497 suspected infectious agent or coronavirus might be possible. Based on fecal viral RNA shedding 498 following animal coronaviruses infections, PCR amplification with universal coronavirus 499

primers was used to screen for potential causative agents of anti-RBD seropositivity. This 500 allowed for non-invasive testing and isolation of coronavirus RNA from infected cats. In line 501 with previous studies on other wild animals, we did not identify any non-alphacoronaviruses 502 circulating in felines [67] [68] [69] [70] . The coronavirus sequences that were isolated and sequenced likely 503 represent normal circulating FCoVs (Fig 6) . Due to the opportunistic nature of our sampling, we 504 were unable to obtain any paired blood and fecal samples from the same animal. As such, we are 505 unable to conclude whether the cats with identified FCoVs would produce cross-reactive 506 antibodies against SARS-CoV-2's RBD. However, based on our ELISA results in Fig 1A and 2A 507 showing no correlation with SARS RBD antibodies and FCoV infection, we would suspect not. 508 Furthermore, not knowing when the seropositive cats were exposed (i.e., cats could have been 509 infected potentially years prior to any fecal sampling) fecal sampling and sequencing would not 510 detect a novel coronavirus if it had been cleared. Following the identification of coronavirus 511 positive fecal samples, we attempted to amplify and sequence the entire RBD region from 512 positive samples to look for similarity to SARS-CoV-2 or SARS-like viruses. Large portions of the S1 region spanning the RBD were sequenced and contained RBD regions similar to 514 previously isolated FCoV strains (Fig. 7) , with no similarity to SARS or betacoronaviruses. 515

The current study demonstrates cross-reactivity of pre-pandemic feline samples with the 516 RBD of SARS-CoV-2. Our indirect ELISA screen has provided evidence for seropositivity of 517 serum from North American cats and deer to a SARS-CoV-2 protein previously shown to be 518 highly specific to SARS coronaviruses [32, 38, 46, 49] . This is the first study to demonstrate 519 seropositivity of animal samples pre-pandemic. What induces this cross-reactive response was 520 not readily apparent. However, we propose several possibilities: exposure to another infectious 521 agent generating cross-reactive antibodies, infection with multiple common coronavirus strains 522 (Feline coronavirus or otherwise) generating cross-reactive antibodies, or exposure of animals to 523 a SARS-like coronavirus pre-pandemic. 524

There is evidence both for and against these explanations for the seropositivity observed. 525

While we cannot discount a non-coronavirus infection generating cross-reactive antibodies, the 526 RBD of SARS-like viruses is thought to be unique with no previous evidence of RBD cross-527 reactivity [32, 38, 46, 49] . A plausible explanation for seropositivity against SARS-CoV-2's 528 RBD is infections with coronaviruses generating an atypical response. To-date, cross-reactivity 529 against the RBD of SARS-CoV-2 has only been reported for SARS and SARS-like 530 coronaviruses [32, 46] . The common human and animal coronaviruses (both alpha and beta 531 coronavirus families) individually do not generate cross-reactive antibodies against this protein, 532 presumably making it a SARS-specific response [30, 32, 33, 38, 46, 49] . For example, prior 533

FCoV exposure did not impact SARS-CoV-2 RBD recognition [38] . Our own observations 534 further demonstrated that FCoV (both serotype I and II) and TGEV did not correlate with 535 serostatus for SARS-CoV-2's RBD (Fig 2A) (FCoV+, FCoV-) . Additionally, humans and 536 animals are exposed to multiple coronaviruses throughout their lifetime, generating measurable 537 immune responses to antigenic proteins [71, 72] . With humans, CoV infections occur on average 538 once a year, with protective immunity appearing to wane a few months after initial exposure [71, 539 72] . Despite reports of cross-reactivity towards portions of SARS-CoV-2's spike protein induced 540 by other human coronaviruses, there have been no reports of cross-reaction against the distinct 541 RBD region [32, 46, 49] . While it does not disprove multiple CoV infections generating cross 542 reactive antibodies, it does cast doubt on this possibility. This leads us to propose prior exposure 543 to a SARS-like coronavirus for North American cats.

There is limited evidence to support prior transmission of a SARS-like virus within 545 felines. However, cats are susceptible to infection and transmission of SARS-CoV-2 [14] . SARS 546 coronaviruses evolved from bat coronaviruses and maintain a high degree of sequence similarity, 547 even in the RBD region [32] . Conceivably, our definition of SARS-like coronavirus could also 548 encompass bat coronaviruses due to amino acid similarity and potential cross-reactivity in the 549 RBD region. At least within Tennessee there are numerous cave systems with several native 550 species of bats potentially leading to an inter-species transmission. Feline and bat interactions 551 could lead to direct transfer of a SARS-like agent, or an intermediary species could be involved. 552

North American deer mice were recently shown to be susceptible to human SARS-CoV-2 and 553 capable of mouse-to-mouse transmission representing another potential point of introduction into 554 the feline population [73] . The pre-pandemic exposure of cats to a SARS-like agent does have 555 detractions. Other groups examining pre-pandemic samples have failed to find evidence of RBD-556 reactive serum even within Central China [38, 56, 58] . The ELISA-positive samples from our 557 environmental samples (i.e., feline and deer) did not neutralize to the same degree as 558 experimentally inoculated felines (Fig 5B) [14]. However, significant neutralization was not 559 observed for most samples, even convalescent serum from COVID-positive individuals (Fig 5B) . 560

Despite ~50% seroprevalence in our samples, we were unable to identify any coronavirus 561 sequences capable of eliciting an anti-SARS RBD response (Fig 6 and 7) . Additionally, there is 562 currently no evidence of a circulating SARS-like or bat betacoronavirus in North America [67-563 70] . While we cannot conclusively demonstrate the origin of the anti-RBD responses shown 564 here, this study indicates that there could be a virus (or another infectious agent) that can 565 generate cross reactive antibodies to the SARS-CoV-2 RBD. 566

A major limitation of this study is the number of serum samples surveyed for all species. 568

Due to the small sample size, we have refrained from making extrapolations from our data to 569 larger animal populations or geographic regions. Similarly, a relatively small number of feline 570 fecal samples were screened for coronavirus viral RNA (n=30), limiting our ability to detect 571 coronaviruses. Our choice of sampling source may have also limited our ability to detect novel 572 coronaviruses or SARS-like coronaviruses. We chose to utilize fecal samples for coronavirus 573 screening because it is non-invasive, animal coronavirus shedding within feces is common, and 574 many are transmitted fecal to oral. Additionally, SARS-CoV and SARS-CoV-2 viral RNA has 575 been detected in stool samples of infected humans [8, 9, 11, 13] . This fecally shed RNA persists 576 longer than nasal or oral sources [13] . However, sampling nasal/oral sources rather than fecal 577 samples may allow for better detection of SARS-CoV-2. Indeed, isolation of SARS-CoV-2 RNA 578 and whole genome sequencing was possible from deer retropharyngeal lymph nodes [28] . Along 579 with sampling site limitations, infection and shedding of viral RNA is a transient process with a 580 relatively small window to detect and sequence the causative agent. While ~20 feline samples 581 were tested for anti-FCoV antibodies, the majority of our samples were of an unknown FCoV 582 status. Additionally, medical histories were not available for these animals and were not screened 583 for prior exposure to other infectious agents. As such, we cannot definitively rule out some other 584 infectious agent conferring the anti-RBD responses. 585

Our initial goal was to develop an ELISA assay for tracking the reverse zoonosis of 587 SARS-CoV-2. However, when establishing our baseline on pre-pandemic cat samples we 588 discovered seropositive serum for an antigen previously reported to be SARS-specific (i.e., 589 RBD). Seropositivity was ~50% in feline samples and could be found several years prior to the 590 genesis of the current SARS-CoV-2 pandemic. What generated the RBD-reactive antibodies is 591 unknown, but we proposed three possibilities: cross-reaction caused by another infectious agent, 592 multiple coronavirus infections creating a rare cross-reactive antibody, or existence and 593 circulation of a SARS-related virus containing the RBD sequence. Regardless, the high rate of 594 SARS-CoV-2 RBD seropositivity within a common companion animal further highlights our 595 need for a better understanding of the prevalence and crossover potential of wild coronaviruses. 596

Further investigations should address shedding of viral RNA from the seropositive species (i.e., 597 cats and deer) identified here to isolate, sequence, and identify the agent enabling cross-reactivity 598 against the RBD of SARS-CoV-2. 599 600 Acknowledgments 601 We would like to thank the following for their generous contributions making this work possible: 602

Mark Slifka -Oregon Health Sciences University 604 Jon Wall and Steve Foster -University of Tennessee Medical Center Knoxville

Emergence of a novel human coronavirus threatening 612 human health

Clinical Characteristics of Coronavirus Disease 2019 in China

Origin and cross-species transmission of bat coronaviruses in China

Emergence of SARS-CoV-2 through recombination and strong purifying 618 selection

The proximal origin of SARS-CoV-2

Prolonged Persistence of SARS-CoV-2 RNA in Body Fluids

Transmission of SARS-CoV-2: A Review of Viral, Host, and 624 Environmental Factors

Detection of SARS-CoV-2 in Different Types of Clinical Specimens. 626 JAMA

Detection and Isolation of SARS-CoV-2 in Serum, Urine, and Stool 628 Specimens of COVID-19 Patients from the Republic of Korea. Osong Public Health Res 629 Perspect

Probable Evidence of Fecal Aerosol Transmission of SARS-CoV-2 in a 631 High-Rise Building

Infectious SARS-CoV-2 in Feces of Patient with Severe COVID-19

Evidence of airborne transmission of the severe acute respiratory 635 syndrome virus

Defining the features and duration of antibody responses to SARS-637

CoV-2 infection associated with disease severity and outcome. Sci Immunol, 2020. 5(54)

Experimental infection of domestic dogs and cats with SARS-639

CoV-2: Pathogenesis, transmission, and response to reexposure in cats

Susceptibility of ferrets, cats, dogs, and other domesticated animals to 642 SARS-coronavirus 2

Broad host range of SARS-CoV-2 and the molecular basis for SARS-CoV-2 644 binding to cat ACE2

Detection of SARS-CoV-2 in a cat owned by a COVID-19-affected 646 patient in Spain

High prevalence of SARS-CoV-2 antibodies in pets from COVID-19+ 648 households. One Health

Evidence of exposure to SARS-CoV-2 in cats and dogs from 652 households in Italy

Households or Close Contacts

Transmission of SARS-CoV-2 in Domestic Cats

Transmission of SARS-CoV-2 on mink farms between 659 humans and mink and back to humans

Virology: SARS virus infection of cats and ferrets

Possible role of an animal vector in the SARS outbreak at Amoy Gardens. 663 Lancet

Animal-to-human SARS-associated coronavirus transmission? 665 Emerg Infect Dis

SARS-CoV-2 infection in free-ranging white-tailed deer

Multiple spillovers from humans and onward transmission of 668 SARS-CoV-2 in white-tailed deer

SARS-CoV-2 exposure in wild white-tailed deer (Odocoileus 670 virginianus)

A serological assay to detect SARS-CoV-2 seroconversion in humans

Antibody responses to SARS-CoV-2 in patients with COVID-19

The receptor binding domain of the viral spike protein is an 676 immunodominant and highly specific target of antibodies in SARS-CoV-2 patients

Rapid Generation of Neutralizing Antibody Responses in COVID-19

Experimental feline enteric coronavirus infection reveals an 681 aberrant infection pattern and shedding of mutants with impaired infectivity in enterocyte 682 cultures

SARS-CoV-2, SARS-CoV, and MERS-CoV viral load dynamics, 686 duration of viral shedding, and infectiousness: a systematic review and meta-analysis. 687 The Lancet Microbe

Use of a reverse-transcriptase polymerase chain reaction for 689 monitoring the shedding of feline coronavirus by healthy cats

A serological survey of SARS-CoV-2 in cat in Wuhan. Emerg Microbes 692 Infect

Two-year prospective study of the humoral immune response of patients 694 with severe acute respiratory syndrome

Disappearance of antibodies to SARS-associated coronavirus after 696 recovery

Duration of serum neutralizing antibodies for SARS-CoV-2: Lessons from 698 SARS-CoV infection

Lack of peripheral memory B cell responses in recovered patients with 700 severe acute respiratory syndrome: a six-year follow-up study

Biochemical characterization of SARS-CoV-2 nucleocapsid protein

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is 705 Blocked by a Clinically Proven Protease Inhibitor

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the 707 ACE2 receptor

Preexisting and de novo humoral immunity to SARS-CoV-2 in humans

A neutralizing human antibody binds to the N-terminal domain of the Spike 711 protein of SARS-CoV-2. Science

Cross-reactive serum and memory B-cell responses to spike protein in 713 SARS-CoV-2 and endemic coronavirus infection

Seasonal human coronavirus antibodies are boosted upon SARS-715

CoV-2 infection but not associated with protection

Neutralizing Antibody Responses in COVID-19 Convalescent Sera

Isolation of potent SARS-CoV-2 neutralizing antibodies and 719 protection from disease in a small animal model

A Structural Landscape of Neutralizing Antibodies Against SARS-CoV-2 721 Receptor Binding Domain. Front Immunol

Enhancement versus neutralization by SARS-CoV-2 antibodies from a 723 convalescent donor associates with distinct epitopes on the RBD

Cross-reactive neutralization of SARS-CoV-2 by serum antibodies from 726 recovered SARS patients and immunized animals

Cross-reactive Antibody Response between SARS-CoV-2 and SARS-CoV 728 Infections

Serological Screening for Coronavirus Infections in Cats. Viruses

Serological survey of SARS-CoV-2 for experimental, domestic, 732 companion and wild animals excludes intermediate hosts of 35 different species of 733 animals

SARS-CoV-2 Serological Survey of Cats in China before and after the 735 Pandemic

SARS-CoV-2 Seroconversion in Humans: A Detailed Protocol for a 737

Serological Assay, Antigen Production, and Test Setup

Identification of a novel coronavirus in bats

CLUSTAL W: improving the sensitivity 742 of progressive multiple sequence alignment through sequence weighting, position-743 specific gap penalties and weight matrix choice

Molecular Evolutionary Genetics Analysis across Computing 746 Platforms

Production of trimeric SARS-CoV-2 spike protein by CHO cells for 748 serological COVID-19 testing

Susceptibility of white-tailed deer (Odocoileus virginianus) to SARS-750 CoV-2

Rapid production of SARS-CoV-2 receptor binding domain (RBD) 752 and spike specific monoclonal antibody CR3022 in Nicotiana benthamiana

Neutralizing antibodies targeting SARS-CoV-2 spike protein. Stem Cell 755 Res

Possibility for reverse zoonotic transmission of SARS-CoV-2 to free-757 ranging wildlife: A case study of bats

Detection of group 1 coronaviruses in bats in North America

Emerg Infect Dis

Detection of polyoma and corona viruses in bats of Canada

Alphacoronaviruses in New World bats: prevalence, persistence, 763 phylogeny, and potential for interaction with humans

Occurrence and frequency of coronavirus infections in humans as 765 determined by enzyme-linked immunosorbent assay

Rises in titers of antibody to human coronaviruses OC43 and 229E 768 in Seattle families during 1975-1979

SARS-CoV-2 infection and transmission in the North American deer 770 mouse

Anti-SARS-CoV-2 ELISA Sensitivity and Specificity. (A) Cross reactivity of

CoV antibodies against SARS-CoV-2 RBD. Polyclonal sera from guinea pigs immunized with 774 common animal coronaviruses (turkey coronavirus, TCoV

BCoV) was 776 used in a SARS-CoV-2 RBD indirect ELISA. Positive samples consisted of polyclonal serum 777 from a SARS-CoV-2 infected patient and a monoclonal antibody to SARS-CoV (CR3022). The 778 negative control group was comprised of pre-pandemic human serum

Bars represent mean and standard deviation 781 (n>3 for all samples). (B) ELISA validation using 66 human Covid-positive plasma and 22 782 negative serum samples. Human antibodies against the SARS-CoV-2 RBD were detected with 783 an indirect RBD-specific ELISA

Positive plasma samples were donated COVID recovered patients and pre-786 pandemic serum samples were the negative controls. Based on the experimentally determined 787 cutoff value, 64 of the 66 positive samples were anti-RBD positive, giving a sensitivity value of 788 96.96%. All but one of the negative samples were below the cutoff value for a specificity of 789 95.45%. Adjacent tables list first and third quartiles along with mean and median OD 450 values 790 of COVID-positive and negative human samples. Bars represent mean and standard deviation 791 (n>3). (C) Anti

White light and 794 chemiluminescent images were overlaid and from left to right, ladder (lane 1) and purified RBD 795 (lane 2). (D) Silver stain of purified recombinant SARS-Cov-2 RBD

From left to right: Ladder, HEK-RBD under 5% serum conditions, HEK-RBD from 2% serum 797 conditions. Samples were denatured and run on a 12% SDS-PAGE gel and Silver Stained 798 (Thermo Fisher Scientific, USA). For B and C, representative data shown

Pre-Pandemic Feline Antibodies Cross-React with SARS-CoV-2 RBD. (A) ELISA 801 results of cat serum RBD reactivity. 93 pre-pandemic feline serum samples were tested for 802 reactivity in our anti-RBD ELISA with anti-felid IgG secondary

Cutoff values were determined by receiver operator curve (ROC) analysis. OD 450 for samples in 804 each group were plotted with the dotted line representing the positive threshold