Material Collection

We collected throat swab (BIOCOMA, http://www.biocomma.com) specimens from 91 mink culled for pelting at a mink farm in Pomorskie Voivodeship in northern Poland on November 17, 2020. The farm owner reported no respiratory symptoms in the animals. We collected blood samples directly from the heart by using cardiac puncture and a sterile 5-mL syringe immediately after death of the mink. After separation of the blood clot, we centrifuged samples at 5,000 rpm for 10 min by using an MPW High-Speed Brushless Centrifuge (MPW Med Instruments, https://mpw.pl). We collected serum and stored it at −80°C until the samples could be analyzed.

RNA Isolation

We added 150 μL of each swab specimen sample in inactivation buffer to 300 μL of RLT lysis buffer (RNeasy Mini Kit; QIAGEN, https://www.qiagen.com). We then mixed samples vortexing and incubated them for 10 min at room temperature. After incubation, we added 400 μL of 70% ethanol to each sample and mixed them by pipetting. We transferred lysates to RNeasy Mini spin columns with collection tubes (QIAGEN) and centrifuged them at 13,000 rpm for 1 min. We then washed the columns once with 700 μL of RW1 buffer and twice with 500 μL of RPE buffer. Between every wash, we centrifuged the columns and discarded the flow-through. We performed elution by adding 50 μL of PCR-grade water to the columns and incubating them for 2 min. We placed the columns into new tubes and centrifuged them at 13,000 rpm for 1 min. After isolation, we stored the samples for <2 hours at 4°C. No human-origin samples were processed at the same time.

SARS-CoV-2 Case Definition

We defined SARS-CoV-2‒positive animals as suggested by the World Organisation for Animal Health (10). We considered mink to be SARS-CoV-2 positive if SARS-CoV-2 was isolated from a sample taken directly from an animal (nasal or oropharyngeal swab sample) or if viral nucleic acid was identified in a sample taken directly from an animal, giving cause for suspicion of previous association or contact with SARS-CoV-2, by targeting >2 specific genomic regions at a level indicating presence of infectious virus or by targeting 1 genomic region, followed by sequencing of a secondary target.

Real-Time Reverse Transcription PCR

For each sample, we prepared the reaction mixture by using a TaqPath 1-Step quantitative RT-PCR (reverse transcription PCR) Master Mix (ThermoFisher Scientific, https://www.thermofisher.com), polymerase, diethyl pyrocarbonate‒treated water (EURx, https://eurx.com.pl) and primers and probes for the RNA-dependent RNA polymerase (RdRp) and envelope (E) genes (16) in white, 8-well, quantitative PCR strips with optical clear caps (Applied Biosystems, https://www.thermofisher.com). We also prepared positive control plasmids made in-house containing the RdRp and E genes and a no-template control containing diethyl pyrocarbonate‒treated water instead of template reactions. We mixed reactions and loaded them into a Light Cycler 480 (Applied Biosystems, https://www.thermofisher.com). Cycling conditions were incubation with uracil N-glycosylase for 2 min at 25°C, reverse transcription for 15 min at 50°C, and enzyme activation for 2 min at 95°C, followed by 40 amplification cycles consisting of 3 s at 95°C and 30 s at 60°C. After each amplification cycle, we measured the signal from each sample in both the FAM (RdRp gene) and HEX (E gene) channels. Samples with a crossing point (Cp) <35 for any gene were considered positive for SARS-CoV-2.

SARS-CoV-2 Antigen Detection in Mink

We performed 2 different antigen tests to confirm the presence of viral antigen in either the swab or serum samples. Antigen tests were conducted on samples positive by quantitative PCR. We used 3 negative samples from the same batch as the positive samples as controls. A total of 150 μL of transport medium from each swab specimen was transferred to tubes from a COVID-19 Antigen Detection Kit (Zhuhai Lituo Biotechnology Co., Ltd., https://www.lituo.com.cn) containing extraction buffer. This antigen test was in the form of a cassette with a lateral flow assay that targets the nucleocapsid protein of SARS-CoV-2. Samples were mixed and incubated for 1 min at room temperature. We added 2 drops of each sample to the sample window on the test cassettes. Results consisted of 2 bands: 1 for the control and 1 for the target. If both bands showed a burgundy line, the test result was considered positive. We read results after 12 min.

We tested 91 mink serum samples by using a SARS-CoV-2 antigen ELISA (COV-04-S; Salofa Oy, https://www.salofa.com) according to the kit instructions. This test is a double-antibody sandwich ELISA. The results were obtained according to the formula based on the concentration standards provided in the kit. The cutoff value for this test was 2.97 pg/mL. The tests were repeated twice, and additional dilutions were performed to determine the final concentration as suggested in the kit instructions.

Full SARS-CoV-2 Genome Sequencing and Classification

We performed SARS-CoV-2 genome sequencing at University of Gdansk in Poland and the University of Helsinki in Finland by using only samples containing RNA isolated from virus-positive swab specimens (amplification of target gene in an RT-PCR; this gene has a Cp <35) or that were inconclusive (only 1 target gene amplification with Cp <35). Samples with positive results in the SARS-CoV-2 antigen detection assays were also sequenced.

At Gdansk, 2 independent protocols were used for SARS-CoV-2 genome sequencing. The first protocol was an Illumina (https://www.illumina.com) RNA preparation with enrichment for respiratory virus oligos panel V2, followed by an Illumina MiniSeq medium output run that produced 150-nt paired-end reads. The second protocol was an ARTIC version 3 amplicon generation (https://bmcgenomics.biomedcentral.com), followed by an Oxford Nanopore Technology MinIONB run (https://nanoporetech.com). No human origin samples were processed at the same time. No DNA/rRNA depletion methods were used. Reads were base called, de-barcoded, and trimmed to remove adaptor, barcode, and PCR primer sequences. Oxford Nanopore Technology reads were used for SARS-CoV-2 genome assembly in ARTIC-nCoV-bioinformaticsSOP-v1.1.0 (https://artic.network/ncov-2019).

Illumina paired reads were used to prepare contigs by de novo assembly by using the Geneious Prime 2020.2 (https://www.geneious.com) software suite with integrated tools for variant calling (minimum coverage = 100, minimum variant frequency = 0.25) and consensus sequence generation. The fasta files generated by the Illumina procedure were further analyzed in Kraken2 2.1.1 software (https://www.ccb.jhu.edu/software/kraken2) to classify every read to the reference database containing viral and American mink genomes (17).

In Helsinki, the sequencing libraries were prepared by using the Illumina DNA Prep Kit (New England BioLabs, https://www.neb.com). We measured library fragment sizes by using agarose gel electrophoresis and concentrations by using the Qubit dsDNA HS Assay Kit (Life Technologies, https://www.thermofisher.com) and the NEBNext Library Quant Kit for Illumina (New England BioLabs). Sequencing was conducted by using the MiSeq V3 Reagent Kit (Illumina) with 250-bp reads. We trimmed raw sequence reads and removed low-quality (quality score <30) and short (<50 nt) sequences by using Trimmomatic (18). Trimmed sequence reads were mapped against the SARS-CoV-2 reference sequence (GenBank accession no. NC_045512.2) by using BWA-MEM version 0.7.17 (19), followed by sorting and removal of duplicate reads by using SAMTools version 1.10 (20).

Phylogenetic Analysis of SARS-CoV-2 Isolates

The dataset consisted of all genetic sequences of SARS-CoV-2 from this study (Poland, Germany, Lithuania, Latvia, Estonia, Russia, and Ukraine), which was completed as the representative pool Europe by Nexstrain (https://nextstrain.org/ncov/europe) and resulted in 5,778 entries. We performed phylogenetic analysis by using the procedure recommended by Nextstrain with modifications in the subsampling region filtering procedure, in which the number of sequences per country was 40 (21). We used Augur toolkit version 10.1.1 (Nextstrain) for phylogenetic analysis and Auspice version 2.10.1 (Nextstrain) for visualization. Possible time of divergence for samples was inferred by using the TreeTime pipeline (https://www.treetime.com) implemented in the Nextstrain analysis and presented in the phylogenetic tree (22).

Statistical Analysis and Ethics

We calculated 95% CIs by using published procedures (23,24). This study was conducted with due regard for European Union Principles and the Polish Law on Animal Protection. No permit from the Local Bioethical Committee for Animal Experimentation was obtained because animals were culled by the owner for production of pelts. Samples were collected postmortem.

Prevalence of SARS-CoV-2

We confirmed that 15 mink (16.5%, 95% CI 8.4%–28.6%) were positive for SARS-CoV-2. We summarized and provide the diagnostic results (Table 1).

SARS-CoV-2 Antigen Detection in Mink

Samples mink_4, mink_5, mink_48, mink_50, mink_77, and mink_83 had highly visible signals in both the control and test lines. Samples mink_20, mink_36, mink_42, mink_46, mink_49, mink_67, mink_76, and mink_88 had a highly visible control line and a much less pronounced test line. In all other samples, only the test line was visible. All 8 real-time RT-PCR‒positive samples were also positive by the antigen test. In addition, 5 E gene‒positive samples were also positive in the antigen test, but 4 were negative. The sample from mink_20 was positive in the antigen test, but SARS-CoV-2 RNA was not detected by RT-PCR.

Read Classification and SARS-CoV-2 Genome Sequences

The final validation of detection of SARS-CoV-2 in the mink was classification of the next-generation sequencing reads on the basis of the database containing reference viral, human, and American mink genomes. We used 3 independent approaches to obtain full viral genomes (Table 2). Only samples that had a complete SARS-CoV-2 genome sequence are shown. The number of Illumina reads generated for samples mink_4, mink_42, mink_49, mink_76, and mink_88 was not enough to produce full SARS-CoV-2 genomes. For these samples, the genomes were obtained by using the ARTIC procedure.

Phylogenetic Analysis of SARS-CoV-2 in Farmed Mink

Figure 2

Figure 2. Phylogenetic tree estimating the divergence time for severe acute respiratory syndrome coronavirus 2, Europe. Sequences from mink sampled on November 24, 2020, in Pomorskie Voivodeship, northern Poland, are shown on...

We checked the 12 mink-originated SARS-CoV-2 sequences for mink-specific mutations detected earlier in mink from the Netherlands and Denmark, but found none. This finding suggested recent and separate introduction of SARS-CoV2 into mink from Poland (Figure 2). Alignment of full-genome sequences from 12 samples showed multiple polymorphisms at different nucleotide sites. Many of these polymorphisms gave rise to changes in amino acids when compared with the reference Wuhan-Hu-1 sequence (GenBank accession no. MN908947). Two specific mutations present in all samples were found in the spike protein: G75V and C1247F. The G75V mutation is present in 199 isolates published in GISAID (https://www.gisaid.org), and the C1257F mutation in 83 isolates. Other rare amino acid variants present in every SARS-CoV-2 isolate from mink in Poland were found in 5 additional proteins: nonstructural protein (nsp) 2, nsp3, nsp14, nsp15, and nucleocapsid protein.

On the basis of the dataset, we inferred a phylogenetic relationship by estimating the divergence times between each isolate (Figure 2). The analysis estimated that the most recent common ancestor for SARS-CoV-2 from mink in Poland and the 2 most similar sequences (German/NW-HHU-340/2020 and Norway/4235/2020) diverged on approximately September 31, 2020. We recognized mutations in amino acid sequences. If the molecular evolution started after virus introduction into the farm, this incident is estimated to have occurred on approximately October 4, 2020. Complete genome sequences of SARS-CoV-2 isolated from farmed mink in Poland have been deposited in GISAID (accession nos. EPI_ISL_732948‒59).