Coronaviruses (CoVs) are enveloped viruses with a positive-stranded RNA genome and classified into four genera (alpha-, beta-, gamma- and deltacoronavirus
) within the subfamily Orthocoronavirinae
in the family Coronaviridae
of the order Nidovirales
. CoVs are found in a variety of mammals and birds, in which they can cause respiratory, enteric and systemic infections [1
]. Additionally, CoVs have proven ability for cross-species transmission, exemplified by the emergence of severe acute respiratory syndrome (SARS) coronavirus in 2002/2003, and of the Middle-East respiratory syndrome (MERS) coronavirus in 2012 [4
]. Both viruses belong to the Betacoronavirus
genus and have an animal origin. SARS coronavirus crossed over from bats via intermediate hosts to humans, became human-adapted and quickly spread worldwide before its containment. MERS coronavirus recurrently enters the human population via its dromedary camel reservoir host, with limited, non-sustained human-to-human transmission particularly in healthcare settings [5
]. Apart from SARS- and MERS-CoV, all four globally endemic human CoVs (HCoV-OC43, HCoV-NL63, HCoV-229E and HCoV-HKU1) originate from animals [8
]. In addition, cross-species transmission potential of CoVs is also illustrated by the occurrence of chimeric coronaviruses that resulted from recombination events between feline CoVs (FCoV) and canine CoVs (CCoV) [12
In order to get insight into the frequency of interspecies transmission of coronaviruses within and between animal and human populations and the risk of subsequent development of a pandemic, it is useful to screen for coronavirus infections in animal species; especially those that are in close contact with humans. Serological assays that can detect virus-specific antibody responses against infection play an important role in these epidemiological studies [14
Cats live in close contact with humans and often roam around freely in the environment. Hence cats are an interesting species to study for infections with coronaviruses. Infections with feline coronaviruses (FCoVs) are recognized and widespread [15
]. FCoVs are classified into two types, type 1 and type 2, based on the genetic and antigenic difference of their spike (S) protein [17
]. In the field, the majority of FCoV infections are caused by FCoV type 1, while FCoV type 2, derived from recombination events of type 1 FCoVs and CCoVs obtaining the S gene and some flanking regions of CCoVs, is less prevalent [18
]. Depending on the virulence of the FCoV strain and the immune response of the cat, the clinical presentation can range from apparently asymptomatic, through diarrhea, to full-blown feline infectious peritonitis [20
]. FCoVs are members of the genus alphacoronavirus
, to which also HCoV-229E, porcine transmissible gastroenteritis virus (TGEV), and CCoV belong. The latter three viruses and FCoV type 2 have been proven to use feline aminopeptidase N (fAPN) as a functional receptor in vitro [21
]. The receptor for type 1 FCoV has still not been identified [22
]. Notably, previous studies have shown that HCoV-229E and CCoV could infect cats after experimental inoculation, causing an asymptomatic infection [23
]. Thus, cats might potentially become naturally infected with CoVs of other species which may lead to virus-host adaptation e.g., mutation or recombination, resulting in emergence of novel coronaviruses and potentially new diseases [19
]. The extent to which infections with CoVs of other species occur in the field, has not been explored in previous epidemiological studies of CoV infections in cats [15
Being the main envelope protein of coronaviruses, the spike (S) protein mediates cell attachment and membrane fusion to allow viral entry. S functions as the main determinant of cell-, organ- and host-tropism. Additionally, it is also the major target of neutralizing antibodies. Spike comprises two functionally interdependent subunits, S1 and S2, with S1 responsible for receptor binding and S2 for membrane fusion [3
]. The S1 subunit is the least conserved and the most variable immunogenic antigen between coronavirus species [30
]. Therefore, the S1 subunit is well suited as an antigen to screen for coronavirus type specific antibodies [31
In this study, CoVs infection in cats were detected through profiling antibody presence in serum samples from cats. Recombinant CoV spike S1 subunits of different animal and human CoVs were expressed in a mammalian expression system and used for screening of cat sera for the presence of antibodies against the respective proteins. Positive samples were also tested by virus neutralization assays to support the specificity of the reaction [32
]. This investigation intends to extend our knowledge of CoV epidemiology, potential reservoirs, and cross-species transmission.
2. Materials and Methods
2.1. Serum Samples
Specific FCoV type 1 and FCoV type 2 sera were obtained from specific pathogen free (SPF) cats previously infected with strain UU2 or RM and FIPV-1146 respectively [35
]. In addition, for the serological survey, 137 feline sera were retrieved from the serum bank in our lab. These had all been collected from cats in the Netherlands. Most of the samples (>80%) were from a study on antibody titer testing for feline panleukopenia virus. The other samples were send to our lab for FIP or FeLV-FIV diagnostics. Sera of uninfected SPF cats were included as negative controls. All samples were stored at −20 °C until analysis.
2.2. Cells and Viruses
African green monkey kidney cells (Vero-CCL81), human hepatoma cells (Huh7), pig kidney epithelial cells (LLC-PK1), human embryonic kidney 293 cells stably expressing the SV40 large T antigen (HEK-293T) were maintained in Dulbecco modified Eagle medium (DMEM, Lonza, Basel, Switzerland) supplemented with 10% fetal bovine serum (FBS, Bodinco, Alkmaar, The Netherlands).
Virus strains used in this study have been described previously [37
]. Briefly, recombinant porcine epidemic diarrhea virus (PEDV) (rPEDV-SDR13
-GFP) was propagated and titrated in Vero cells, and HCoV-229E in Huh7 cells. PDCoV was propagated and titrated in LLC-PK1 cells, but supplemented with 1 μg/mL TPCK-treated trypsin (Sigma-Aldrich, Inc., St Louis, MO, USA) in DMEM.
2.3. Plasmids Constructs and Recombinant Protein Expression
Synthetic sequences of 12 coronavirus spike S1 subunits (HCoV-HKU1 (GB: YP_173238.1), MERS-CoV (GB:YP_009047204.1), SARS-CoV (GB: AAX16192.1), HCoV-OC43 (GB: AAR01015.1), HCoV-229E (GB: NP_073551.1), HCoV-NL63 (GB: YP_003767.1), TGEV (GB: ABG89325.1), PEDV (GB: AOG30832.1), BCoV (GB: P15777.1), PDCoV (GB: AML40825.1), FCoV type 1 (GB: FJ938060.1), FCoV type 2 (GB: AY994055.1)) and different domains of PEDV S1 subunit (S10
, as identified and described in [40
]) were cloned into pCAGGS expression plasmids as described previously [41
]. Similarly, the expression constructs encoding chimeric proteins in which S1s were fused to the Fc domain of mouse IgG2a. For protein production, HEK-293T cells were transfected with plasmid DNA conjugated to polyethyleneimine (Polysciences, Inc., Warrington, PA, USA). At 8–16 h post transfection, inoculum was removed and the transfection mixture was replaced by 293 SFM II expression medium (Gibco®, Life Technologies Inc., Grand Island, NY, USA). At 6–7 days post transfection, cell supernatants were harvested and proteins were collected by Protein A Sepharose beads (GE Healthcare Bio-Sciences AB, Uppsala, Sweden). Proteins were then eluted with 0.1 M citric acid, pH 3.0 and neutralized with 1 M Tris-HCl, pH 8.8. Concentrations of proteins were assessed by Nanodrop spectrophotometry (ThermoFisher Scientific Inc., Waltham, MA, USA) and confirmed by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) with bovine serum albumin (BSA, BioIVT, West Sussex, UK) as standard. Typical yields for proteins were 0.2–0.5 mg/mL. For long term storage, proteins were stored at −80 °C upon usage.
2.4. FCoV S Structure Modelling and S1 Domain Expression
To study the potential cross-reaction between FCoV type 1 and 2 in more detail, models of FCoV type 1 (strain: UU2; GenBank accession no.: FJ938060.1) and FCoV type 2 (strain: 79-1146; GenBank accession no.: AY994055.1) S proteins were generated via the automated protein structure SWISS-MODEL Homology Modelling server (https://swissmodel.expasy.org/
] using the elucidated HCoV-NL63 Cryo-EM structure (PDB code: 5SZS) as the input model. Figures were made with PymoL (The PyMOL Molecular Graphics System, Version 1.0 Schrödinger, LLC.). FCoV S1 domains of both type 1 and 2, namely S10-CD
, were expressed as murine Fc fusion proteins in HEK-293T cells as described above.
2.5. Enzyme-Linked Immunosorbent Assay (ELISA)
High binding microtiter plates (Greiner Bio-one BV, Alphen aan den Rijn, The Netherlands) were coated overnight at 4 °C with equal molar amount of protein (0.25 pmol per well, diluted in phosphate buffered saline (PBS, pH 7.4)). After three washes with washing buffer (PBS containing 0.05% Tween-20), the plates were blocked for 2 h at 37 °C with blocking buffer (PBS containing 5% milk powder (Protifar, Nutricia, Zoetermeer, The Netherlands), 0.05% Tween-20). Protein coating efficiency was assessed by binding of anti-mouse IgG antibodies in a direct ELISA, and confirmed the equimolar coatings of all proteins. To detect antigenic reaction with serum samples, sera were tested in duplicate at a 1:200 dilution in blocking buffer, and then incubated in the plates at 37 °C for 1 h. After washing, plates were incubated with a 1:4000 diluted horseradish peroxidase (HRP)-conjugated goat anti cat IgG (Rockland Immunochemicals, Inc., Pottstown, PA, USA) at 37 °C for 1 h. The peroxidase reaction was then visualized via adding TMB Super Slow One Component HRP Microwell Substrate (BioFX®, Surmodics IVD, Inc., Eden Prairie, MN, USA) for 10 min. Reaction was stopped with 12.5% sulfuric acid and optical densities (OD) were measured at 450 nm. Negative sera (from uninfected SPF cats) were included to determine the ELISA cut-off values; sera with OD values higher than 5-fold the OD of negative sera were considered positive. All 12 S1 proteins were coated on the same ELISA plates making it easy to screen and compare the OD values of individual sera in one assay. Hereby we excluded the sera that give high background OD values against all proteins being considered false positive.
2.6. Virus Neutralization Assay
Neutralization assays were performed with some of the CoVs to support the specificity of ELISA results. Cat sera were serially diluted 2-fold in DMEM and mixed 1:1 with rPEDV-SDR13-GFP, HCoV-229E or PDCoV (2000 50% tissue culture infective doses [TCID50]/mL). These mixtures were then incubated at 37 °C for 1h, and 100 µL of each mixture was used for inoculation with Vero, Huh7 and LLC-PK1 cell monolayers in 96-well plates, respectively. For PDCoV infection, TPCK-treated trypsin (Sigma-Aldrich, Inc., St Louis, MO, USA) was supplied to LLC-PK1 tissue culture medium at a final concentration of 1 μg/mL. At 2–5 days post infection, cytopathic effect (CPE) could be observed via microscopy. Virus neutralization titers (VNT) were expressed as the highest serum dilution resulting in 90% reduction of cytopathic effect (HCoV-229E and PDCoV) or virus-induced fluorescent cells (PEDV). Before virus neutralization, sera were inactivated through incubation at 56 °C for 30 min. Experiments were performed in triplicate.
Coronavirus infections are endemic and ubiquitous in feline populations. Two viral types, type 1 and 2, are distinguished and both of them could well sustain themselves in the cat reservoir [15
]. Both have been shown to have worldwide distribution, with the seropositivity rate up to 90% among animal shelter populations and in multi-cat households [20
]. The majority of natural infections are caused by type 1 FCoVs, while in the field type 2 FCoVs are less common and mainly occur in Asia [15
]. CoVs are generally considered to be host-specific; however, cross-species transmission does occur which may lead to incidental infections like the spillover of MERS-CoV from dromedary camel to humans, where humans function as an incidental and ultimately dead-end host [4
]. But CoVs might also adapt to the new host exemplified by the animal origin of all four endemic human CoVs (HCoV-OC43, HCoV-NL63, HCoV-229E and HCoV-HKU1) [8
]. Whereas in cats infections with FCoV are well recognized, studies regarding possible natural infections with other animal and human coronaviruses are lacking to the best of our knowledge. Knowing the genetic variability of coronaviruses and the use of orthologous receptors by non-feline CoVs, studies on cross-species transmission are desirable. This may provide insight regarding whether cross-species transmission does occur. In the present study we used the highly immunogenic S1 antigens to screen cat sera for the presence of antibodies against feline and non-feline coronaviruses, as a first indication of possible infections with these viruses.
In our study, 78 of the 137 cat sera were shown to be seropositive for coronaviruses. The seropositive rate (54.7%) against S1 of FCoV type 1 is consistent with previous studies [15
]. All of the FCoV type 2 S1 positive sera of naturally infected cats were also positive for FCoV type 1 S1, which might be the result of cross reaction between the two proteins, despite their low amino acid identity. ELISA with specific antisera from experimentally FCoV type 1 and type 2 infected cats showed that sera of several FCoV type 1 infected cats could cross-react with FCoV type 2 S1. Domain mapping ELISA results showed that FCoV type 1 specific sera react to different levels with the S1 domains of FCoV type 1 S1 protein, and also reacts with FCoV type 2 S1CD
. Vice versa, FCoV type 2 specific sera also reacted with S1CD
of FCoV type 1. These observations pose a potential two-way cross-reactivity between S1CD
domains. Interestingly, in parallel with our findings on feline coronaviruses, we identified a number of samples that were seropositive against the S1 of PEDV, a viral pathogen that mainly replicates in the porcine intestinal epithelium. To study the possibility of cross-reaction, samples derived from pre- and post- FCoV infected cats were screened against PEDV S1 in ELISA. The reactivity found against PEDV S1 with FCoV specific sera shows that cross-reaction can occur at the level of domain S1CD
; the other PEDV S1 domains showed no reaction with the FCoV positive sera. Judging from these observations, it seems that S1CD
plays an important role in cross-reaction between FCoV type 1 and 2, and also FCoV and PEDV. As S1CD
is the most conserved domain among FCoV and also between FCoV and other alphacoronaviruses (for a systematic assessment of sequence identities, see Table 3
), it is reasonable to hypothesize that antibodies can develop against conserved epitopes within this region and subsequently cause cross-reaction. This should be taken into account when developing and interpreting serological assays.
Noticeably, ELISA reactivity among cat sera towards the N-terminal FCoV S1 domains 0 and A was less consistent and generally lower compared to whole S1, which seems to correlate with the higher antigenic variation in those domains found among FCoV type 1 strains [48
] (Figure 3
). Especially the sera from FCoV-RM infected cats (cat 91, 93, 95 and 115) showed lower OD values against S1A
. This phenomenon could be explained as the samples displaying higher reactivity were from cats inoculated with FCoV-UU2 (cat 089, 131 and 129), the particular strain from which the S1 region was used as an antigen in the ELISA studies. In the meantime, the possibility of the variable ELISA reactivity might be due to the difference in individual antibody levels. In principle, the distinct antigenic reactivity of S10
between the two FCoV types might facilitate the development of a specific ELISA method which allows the serological discrimination of FCoV type 1 and type 2 infections in cats.
In order to provide further insight regarding cross-reactivity between FCoV type 1 and PEDV, we performed virus neutralization assays. Cross-neutralization of PEDV infection could be observed for some of the feline FCoV type 1 post-infection sera, in contrast to the pre-infection serum counterparts. Since FCoV specific PEDV neutralizing sera did not react with PEDV S10
, it is likely that the cross-neutralizing antibodies are targeting conserved epitopes in the S1CD
domain or the S2 subunit of the PEDV spike protein [49
]. Given the unknown TGEV infection background of the PEDV positive pigs, the cross-reaction of PEDV specific sera against FCoV type 1 could not be explored in our study, as TGEV positive pig samples would certainly influence the outcome [12
]. Of note, our findings cannot exclude the possibility that field cats might incidentally get naturally infected with PEDV or PEDV-like viruses, as there had been one report showing the detection of PEDV in one stray cat via PCR assay [51
]. It would be interesting to include more sera of cats from pig farms in future studies.
Considering the fact that cats play an important role in human society and have constant interaction with humans, it is of interest to conduct serological surveys for possible reverse zoonosis of human pathogens. In our study S1 antigens of several human coronaviruses were included and this led us to identify HCoV-229E seropositive feline samples in our ELISA survey (Table 1
); one serum in particular reacted solely with HCoV-229E S1 but not with any other coronavirus. Of the HCoV-229E S1 reactive feline sera one showed low neutralizing activity against HCoV-229E infection. This might suggest that positive cats were indeed exposed to HCoV-229E or related viruses. MERS seropositivity is also seen in other species besides the dromedary host [52
]. Rare cases of seropositivity might be considered as spill-over infections from the dromedary camel reservoir. Similar (perhaps dead-end) spill-over infections of 229E from the human reservoir to cats might also occur. A similar principle could also apply for PDCoV, a porcine pathogen that emerged rather recently. Both HCoV-229E and PDCoV use APN as their receptor and have been reported to also be able to use feline APN for cellular entry [21
]. Although reports are lacking regarding the natural infection of these two viruses in cats, HCoV-229E was shown to cause a priming effect of FCoV antibody in experimentally FCoV infected cats suggesting that infection occurred [24
]. Therefore, the detection of antibodies against S1 of HCoV-229E in a portion of the cats might be specific and due to the exposure to HCoV-229E through daily interaction with humans. Eight cats were seropositive for PDCoV of which two cats were seropositive only for PDCoV and not for any other CoVs. This could be caused by infection with PDCoV or PDCoV-related viruses through avian sources, considering the fact that cats are natural avian predators and the presumed avian origin of PDCoV [2
]. Our findings emphasize the potential role of cats as incidental hosts for non-feline coronaviruses and the need of in-depth study of naturally infected pathogens in cats. Besides serological studies, efforts should also focus on isolation and identification of these viruses in cats.
In conclusion, we presented a thorough serological survey in cats using S1 proteins of different animal and human coronaviruses. We demonstrated, despite the low amino acid identity, cross-reactivity between S1 proteins of FCoV type 1 and 2, and between that of FCoV type 1 and PEDV. This should be considered when developing FCoV serological assays as well as interpreting the results. Our observation that some feline sera displayed antibody reactivity exclusively against non-feline CoV S1 proteins warrant further research into the epidemiology and cross-species transmission of coronaviruses in cats and other animals that are in close contact with humans. Further large scale serological studies regarding coronaviruses infection across animal species using arrays of CoV S1 antigens can shed light into the hitherto unresolved host promiscuity of coronaviruses and the risk of cross-species transmission.