The effect of climbing chalk powder on the infectivity of human coronavirus OC43

Abstract There does not appear to be any studies in the published literature on the stability of SARS‐CoV‐2 in climbing chalk powder (magnesium carbonate and/or calcium carbonate), which has been hypothesized to pose a potential risk of fomite transmission of coronavirus disease 2019 (COVID‐19) within climbing gyms. The aim of this study was to determine the infectivity of a model human coronavirus HCoV‐OC43 in the presence of climbing chalk powder on a dry plastic surface. The stability of HCoV‐OC43 on a plastic surface dusted with climbing chalk powders (magnesium carbonate, calcium carbonate or a blended chalk) was determined by titration on BHK‐21 fibroblast cells. No chalk and no virus controls were included. HCoV‐OC43 was stable on the plastic surface for 48 h. The stability of HCoV‐OC43 was significantly (P ≤ 0·05) reduced in the presence of magnesium carbonate, calcium carbonate and the chalk blend; the infectivity was reduced by ≥2·29 log10 50% tissue culture infective dose (TCID50) immediately upon on contact and by ≥2·46 log10 TCID50 within 1 h of contact. These findings suggest that the infectivity of coronaviruses is reduced by climbing chalk, limiting the risk of potential fomite transmission.


Introduction
The emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and subsequent development of the coronavirus disease 2019 (COVID-19) pandemic has led to the widespread implementation and promotion of public health interventions, such as social distancing, closure of nonessential businesses, enhanced disinfection and handwashing to limit transmission of COVID-19 (Hartley and Perencevich 2020;Prather et al. 2020). Although COVID-19 is predominantly transmitted by respiratory droplets, it has been hypothesized that indirect transmission through contact with fomites could also occur because SARS-CoV-2, similarly to other coronaviruses, can remain infectious on dry surfaces for several hours (Pastorino et al. 2020).
It has been speculated that recreational climbing gyms could be high-risk environments for the transmission of SARS-CoV-2. In particular, concerns have been raised regarding potential fomite transmission by the use of shared climbing holds which are not cleaned between users in common practice (Debenham and Reynolds 2020). SARS-CoV-2 has been shown to remain infectious on plastic for around three days (Van Doremalen et al. 2020) to seven days (Chin et al. 2020), indicating that SARS-CoV-2 could persist on plastic climbing holds.
Another speculated potential source for the transmission of COVID-19 within climbing gyms is climbing chalk, which is applied to the hands of climbers to improve grip on holds (Bacon et al. 2018). Climbing chalk has been observed to deposit onto climbing holds during use (Li et al. 2001) and as such has been hypothesized to serve as another potential environmental reservoir of SARS-CoV-2. Climbing chalk powder is primarily comprised of magnesium carbonate (Li et al. 2001), with calcium carbonate present as an impurity or filler. There does not appear to be any studies in the published literature on the stability of SARS-CoV-2 or other coronaviruses in climbing chalk powder and therefore the risk of fomite transmission in this context is not well understood.
In this study, human coronavirus (HCoV-OC43), a common causative agent of upper respiratory tract infections in humans  was used as a model for SARS-CoV-2 to determine stability of coronaviruses in climbing chalk powder. HCoV-OC43 is in the same genus as SARS-CoV-2 (Betacoronavirus) and has a similar overall structure; both possess a single-stranded positive-sense RNA genome and are enclosed in a lipid envelope with spike proteins protruding from the surface (Chen et al. 2020). Previous studies suggest that the stability of SARS-CoV-2 in the environment is likely to be within the same order of magnitude of other coronaviruses (Aboubakr et al. 2020), suggesting that HCoV-OC43 could useful as a model for SARS-CoV-2 in environmental stability experiments. For example, Van Doremalen et al. (2020) reported that the half-lives of SARS-CoV-2 and SARS-CoV were similar on inanimate objects including copper (0Á8 vs 1Á4 h), stainless steel (5Á6 vs 4Á2 h) and plastic (6Á8 vs 7Á6 h).
The aim of this study was to determine the stability of a model human coronavirus (HCoV-OC43) in the presence of climbing chalk powder on a dry plastic surface.

Results and discussion
HCoV-OC43 remained relatively stable on the plastic surface without chalk, with 3Á74 log 10 50% tissue culture infectious dose (TCID 50 ) sample À1 infectious virus being detected on the surface after 1 h and ≤2Á68 AE 0Á35 after 24 h (Table 1). No significant (P ≤ 0Á05) decrease in infectious viral titre was observed until 48 h incubation, where no infectious virus was detected (Table 1). These results are in line with previous studies conducted on the survival of SARS-CoV-2 on plastic surfaces. Van Doremalen et al. (2020) recovered 2Á5 log 10 ml À1 SARS-CoV-2 from plastic surfaces after 24 h incubation at room temperature (initial inoculum, 5 log 10 ml À1 ), and after 48-72 h, only 0Á5-1Á0 log 10 ml À1 was detected. No infectious virus was recovered from the no virus controls.
The infectivity of HCoV-OC43 was significantly (P ≤ 0Á05) reduced by magnesium carbonate, calcium carbonate and the blended chalk (Table 2) compared to the no chalk control at 0 min, which was only reduced by 0Á63 log 10 TCID 50 (Table 1). Magnesium carbonate and the chalk blend both reduced HCoV-OC43 by ≥2Á29 log 10 TCID 50 immediately on contact, with ≤2Á50 log 10 TCID 50 Table 1 Stability of HCoV-OC43 on a PVC PUR surface (n = 3, mean AE standard error of the mean) Time Log 10 TCID 50 per sample (Log 10 reduction from initial inoculum*) *The initial inoculum was 4Á79 log 10 TCID 50 per sample. † Below detection limit (2Á33 log 10 TCID 50 per sample). Where one or more samples reached the detection limit of the assay (2Á33 log 10 TCID 50 per sample) the number of infectious virus is expressed as ≤log 10 TCID 50 .
*The initial inoculum was 4Á79 log 10 TCID 50 per sample except for calcium carbonate, where the inoculum was 5Á01 log 10 TCID 50 per sample. † Below detection limit (2Á33 log 10 TCID 50 per sample). Where one or more samples reached the detection limit of the assay (2Á33 log 10 TCID 50 per sample) the number of infectious virus is expressed as ≤log 10 TCID 50 per sample. per sample infectious virus being detected (Table 2). Calcium carbonate chalk reduced HCoV-OC43 to a lesser extent, by 2Á17 log 10 TCID 50 . In this study, infectious virus was recovered from the chalk by resuspension in PBS and centrifuging to remove excess chalk. This method was employed to minimize the cytotoxicity of the chalk and allow syringe filtration of the viral supernatant to avoid bacterial contamination. In order to validate the recovery method, the distribution of viral RNA between the supernatant and the chalk pellet after 0 min contact were quantified by quantitative reverse transcription PCR (RT-qPCR). A significantly greater (P ≤ 0Á05) number of viral RNA was present in the magnesium carbonate supernatant compared to the pellet, where 80Á3 AE 2Á4% of the total RNA was recovered in the supernatant compared to the pellet. There was no significant difference (P > 0Á05) in the recovery of HCoV-OC43 between calcium carbonate and blended chalk supernatants and pellets, with 50Á5 AE 16Á4% and 39Á4 AE 9Á7% of the total viral RNA being recovered in the respective supernatants compared to the pellets. Although a proportion of the viral RNA was lost within the chalk pellet, removal of excess chalk could not be avoided due to cytotoxicity of the chalk against the mammalian cell lines. These results do not change the observations of an overall loss in infectivity of HCoV-OC43 on exposure to the chalk; theoretically if virus from the pellet was also recovered, the maximum increase in infectivity in the blended chalk condition at 0 min would be to 2Á9 log 10 TCID 50 per sample (from ≤2Á5 log 10 TCID 50 per sample) assuming 39Á4% recovery, which still results in approximately a 2 log 10 reduction. It is important to note that detection of nucleic acids does not distinguish between infective and inactive virus and therefore it was not possible to ascertain the TCID 50 sample À1 of HCoV-OC43 within the pellet. Moreover the behaviour of the virus within the pellet would also be in line with those in the supernatant and therefore the majority are likely be noninfectious.
The number of infectious virus remained close to or below the detection limit (2Á33 log 10 TCID 50 per sample) up to 30 min contact with magnesium carbonate and the chalk blend. No infectious virus was detected after 1 h in the presence of chalk, indicating that ≤2Á33 log 10 TCID 50 infectious virus was present. The infectious viral titre did not change significantly (P > 0Á05) over the course of 1 h upon contact with all three chalks. One limitation of this study is the relatively low sensitivity of the infectivity assay, where the limit of detection was 2Á33 log 10 TCID 50 per sample. This can be attributed to cytotoxicity of the chalk supernatants upon BHK-21 cells, thus requiring a greater level of dilution to quantify infectious virus.
The results of this study indicate that human coronavirus HCoV-OC43 is unlikely to persist for extended periods of time within climbing chalk powders, and therefore the risk of fomite transmission is limited. There do not appear to be any studies in the published literature on the antiviral activity of climbing chalk powders. Magnesium carbonate (1 mg ml À1 suspension) was previously reported to inhibit the growth of Staphylococcus epidermidis, which was in part attributed to its alkalinity (Welch et al. 2016). Further research is required to determine the mechanism by which HCoV-OC43 is inhibited by climbing chalk powders. The inactivation of HCoV-OC43 could potentially be attributed to desiccation of the virus, due to the hygroscopic nature of magnesium carbonate and calcium carbonate, and/or the alkalinity. Desiccation has been hypothesized to lead to phase changes, peroxide formation and other reactions within the lipid envelope of coronaviruses, leading to denaturation (Casanova et al. 2010) and thereby preventing interaction with host cells. In accordance, previous research has indicated that HCoV-OC43 is more stable in suspension compared to dry conditions, remaining infectious for 2 to 3 days suspended in culture media compared to less than three hours dried onto aluminium, sponge or latex gloves; a similar pattern was also observed for HCoV-229E (Sizun et al. 2000). Another potential hypothesis is that binding of HCoV-OC43 to fine chalk particles may prevent attachment to and/or entry into the cells, thus rendering the virus noninfectious.
The initial viral titre used in this study (~5 log 10 TCID 50 per ml) is in line with that of similar studies measuring the environmental stability of coronaviruses (Sizun et al. 2000;Lai et al. 2005;Van Doremalen et al. 2020). There is limited published research on the quantity of infectious virus that is shed by COVID-19 patients and therefore a realistic infectious viral load for contaminated surfaces has not been fully ascertained. Previous work on influenza demonstrated that the infectious viral load of one cough was 5 plaque-forming units ml À1 (Lindsley et al. 2010), suggesting that 5 log 10 TCID 50 per ml may be greater than that encountered under realistic conditions. However, the use of a 5 log 10 TCID 50 per ml inoculum allows viral titre reductions to be monitored above the detection limit of this assay, and can therefore be used to infer the viral stability in the presence of chalk under real life conditions.
Overall, the infectivity of HCoV-OC43 is reduced by ≥2Á29 log 10 TCID 50 immediately on contact with magnesium carbonate, calcium carbonate and a blended chalk and by ≥2Á46 log 10 TCID 50 within 1 h contact. HCoV-OC43 reduces naturally on the plastic surface over time, decreasing by 0Á63 log 10 TCID 50 immediately on contact and 1Á05 log 10 TCID 50 within one hour, which is significantly less than on contact with the chalk. This suggests that coronaviruses such as HCoV-OC43 or SARS-CoV-2 are likely inactivated by climbing chalk, limiting the risk of potential fomite transmission. Further research into the stability of SARS-CoV-2 in climbing chalk would enhance our understanding of COVID-19 transmission risks.
Human coronavirus HCoV-OC43 (ATCC VR-1558) was cultured in HCT-8 epithelial cells in RPMI-1640 with 5% FBS, 100 IU ml À1 penicillin and 100 lg ml À1 streptomycin for 7 days at 33°C. HCoV-OC43 was harvested by aspirating the RPMI-1640 culture media and centrifuging (3000g, 4 min) to remove cell debris. The virus stocks were stored at À80°C before use. All experiments were performed within a Biosafety Level 2 laboratory.
Aliquots of 200 ll of the model human coronavirus HCoV-OC43 (4Á79 log 10 TCID 50 per sample) was transferred via 10 ll droplets to the surface of the chalk and incubated at room temperature for 0, 1, 5, 10, 30 and 60 min. No virus (chalk with 200 ll cell culture media) and no chalk (200 ll HCoV-OC43) controls were included.
Infectious virus was recovered in 3 ml phosphate buffered saline (PBS; Lonza, Basel, Switzerland) using cotton swabs, vortexed for 30 s and centrifuged at 2500g for 10 min to remove excess chalk. The supernatant was syringe filtered (polyether sulfone, 0Á45 lm; Fisher Scientific, Loughborough, UK), serially diluted in DMEM supplemented with 5% FBS, 100 IU ml À1 penicillin and 100 lg ml À1 streptomycin and plated onto BHK-21 fibroblast cells seeded in a 96-well format. The plates were incubated at 33°C with 5% CO 2 for four d before the cells were scored for cytopathic effect. The number of infectious viral particles present on each surface sample (log 10 TCID 50 per sample) was then determined according to the Karber method (Ramakrishnan 2016). The detection limit of the assay was 2Á33 log 10 TCID 50 per sample. The log 10 reduction of infectious virus for each sample was calculated from the number of infectious virus particles deposited onto the surface (4Á79 log 10 TCID 50 per sample).

RT-qPCR
Chalk powders were inoculated with 200 ll HCoV-OC43, immediately recovered in three ml PBS as described above and the supernatants harvested. The chalk pellet was resuspended in a further three ml PBS and three cycles of vortexing for 30 s and centrifuging at 2500g for 10 min were performed prior to recovering the supernatant. Resulting supernatants were syringe filtered prior to RNA extraction using the Meridian Bioscience (London, UK) ISOLATE II RNA Mini Kit according to the manufacturers' instructions. A no chalk control was included.
Total RNA samples were DNase-treated using the Promega (Chilworth, UK) RQ1 RNase-free DNase kit prior to reverse transcription using random primers and the GoScript TM reverse transcription system (Promega, Chiworth, UK) according to the manufacturers' instructions. No reverse transcriptase controls were included, where nuclease free water was included in the reaction mix in place of reverse transcriptase.
RT-qPCR was performed manually using the Meridian Bioscience SensiFAST TM Probe Lo-ROX Kit with HCoV-OC43-specific primers (forward primer, 5-AGC AAC CAG GCT GAT GTC AAT ACC-3; reverse primer, 5-AGC AGA CCT TCC TGA GCC TTC AAT-3) and probe ([6FAM]TGACATTGTCGATCGGGACCCAAGTA[TAM]) as described by Owczarek et al. (2018). No template and no reverse transcriptase controls were included. PCR was performed at 95°C for 2 min, followed by 40 cycles of 95°C for 10 s and 60°C for 20 s using an Applied Biosystems (Waltham, MA) QuantStudio TM 5 real-time PCR system. The amount of viral RNA was normalized by comparing to a standard curve of a known quantity of infectious HCoV-OC43 (8Á35 log 10 TCID 50 per ml).

Statistical analysis
Viral infectivity experiments were conducted in triplicate on separate occasions (n = 3). RT-qPCR was conducted with technical triplicates of at least two independent experiments (n = 2-3). Distribution of the data was ascertained using the Shapiro-Wilk test (Shapiro and Wilk 1965). Significant differences (P ≤ 0Á05) in recovered infectious virus (log 10 TCID 50 per sample) for each condition were determined by the independent samples Kruskal-Wallis test (Kruskal and Wallis 1952) with multiple comparisons using SPSS ver. 26 (IBM, Armonk, New York, USA). Significant differences (P ≤ 0Á05) in viral nucleic acid recovered from chalk supernatants and pellets was determined using the related-samples Wilcoxon signed rank test (Wilcoxon 1945).