Laboratory Worker Self-Contamination with Noninfectious SARS-CoV-2 DNA Can Result in False-Positive Reverse Transcriptase PCR-Based Surveillance Testing

LETTER

The pandemic caused by severe acute respiratory distress syndrome coronavirus 2 (SARS-CoV-2) prompted rapid development of research tools, including plasmid systems carrying cDNA copies of various portions of the viral genome, providing tractable tools for the study of the virus (1, 2). Here, we report evidence that noninfectious SARS-CoV-2 cDNA poses a contamination risk to laboratory personnel, resulting in false-positive diagnostic reverse transcriptase PCR (RT-PCR) results.

In the fall of 2020, weekly screening of students, staff, and faculty by self-administered nasal swabs using a high-throughput version of the CDC SARS-CoV-2 RT-PCR diagnostic test (performed off-site at the Broad Institute) was implemented at many universities, including the University of Vermont (UVM). At the same time, our research group initiated work using a SARS-CoV-2 reverse genetics (RG) system (1). Briefly, in this system, the viral genome fragments are encoded on 7 DNA plasmids, excised and ligated into a SARS-CoV-2 full-length cDNA, and transcribed to yield the full-length SARS-CoV-2 RNA genome. Additionally, this system requires separate PCR amplification and transcription of the N gene. Notably, N is the target gene used in many RT-PCR-based nasal swab screening tests.

At UVM, SARS-CoV-2 RG work commenced on 11 September 2020 and was completed on 11 December 2020. During this period, a cluster of five researchers working in a single departmental suite where the SARS-CoV-2 DNA was amplified/manipulated were identified as SARS-CoV-2 positive by RT-PCR-based weekly screening. None of these individuals had known positive contacts, developed COVID-19 symptoms, or transmitted virus to close contacts, and follow-up testing revealed that none had seroconverted (Table 1). In most cases, follow-up RT-PCR was also negative (Table 1). These results strongly suggest that the initial RT-PCR results were false positives.

TABLE 1.

Clinical results^a

Subject	Sex	Age (yrs)	Initial RT-PCR resultb	Symptomatic	Follow up RT-PCR	Serology date (mo/day/yr or mo/yr)	Serology result
R1f	F	36	10/27/2020 POS; N2 CT  = 33.1	No	10/28/2020 POSb	12/9/2020	Negatived
					N2 CT  = 36.5 10/29/2021 NEGb
R2	F	32	11/2/2020 POS; N2 CT  = 36.4	No	ND	11/2020	Negatived
R3	F	21	11/10/2020 POS; N2 CT  = 35.8	No	ND	12/2020	Negatived
R4	F	29	12/1/21 POS; N2 CT  = 34.1	No	12/2/20 NEGb; 12/2/21 NEGc	3/12/2021	Negatived
R5	M	24	12/8/2020 POS; N2 CT  = 35.4	No	12/10/20 NEGc	3/17/2021	Negativee

^aClinical record data were obtained from the electronic medical record system following a determination of exempt status by the UVM Institutional Review Board (CHRMS STUDY00001409). Abbreviations: F, female; M, male; POS, positive; NEG, negative; ND, not done.

^bUVM surveillance testing; Broad Institute high-throughput version of the CDC 2019-nCoV real-time RT-PCR (https://covid-19-test-info.broadinstitute.org).

^cHologic Panther Fusion SARS-CoV-2 assay (target: ORF1ab region 1 and region 2).

^dSARS-CoV-2 IgG Ab, S; VITROS Immunodiagnostic product anti-SARS-CoV-2 IgG reagent pack assay (Ortho-Clinical Diagnostics, Inc.).

^eSARS-CoV-2 Spike Ab, Interp, S; Roche Elecsys anti-SARS-CoV-2 S reagent assay from Roche Diagnostics.

^fWe note that R1 was the only researcher directly handling the SARS-CoV-2 cDNA samples, and hence the subsequent repeat positive test was likely due to a high level of exposure.

Based on these circumstances, and knowing that viral cDNA was being actively manipulated in the laboratory, we undertook quantitative PCR (qPCR)-based screening of the lab environment using the CDC N2 primer set (3), while intentionally omitting the reverse transcription step to detect only DNA and not RNA. To determine whether the noninfectious viral cDNA components of the RG system could function as a template for the N2 primer set, plasmid DNA encoding the N gene was used as input for qPCR analysis. We found that as little as 4 fg of viral N cDNA was sufficient for detection by the N2 primer set (Fig. 1A). We next conducted swab-based screening of laboratory equipment and surfaces used for SARS-CoV-2 RG work. Importantly, each of the selected swab sites had already been previously decontaminated several times with bleach and DNAZap. Nonetheless, multiple sites in the laboratory were still found to be positive for viral cDNA by qPCR (cycle threshold [C_T], ∼25 to 40) (Fig. 1B). These data suggest that that viral cDNA can persist on laboratory surfaces after decontamination.

FIG 1.

Noninfectious SARS-CoV-2 cDNA is detected by CDC diagnostic PCR primers and persists in the laboratory environment. (A) Direct SYBR green-based qPCR analysis using the CDC SARS-CoV-2 N2 primer set (without reverse transcription) was run on serially diluted RG system N plasmid DNA, with N gene equivalents ranging from 10¹ to 10⁷ copies (3.9 × 10⁻⁹ to 0.04 ng) input per reaction. Raw C_T values are reported. (B) Screening of laboratory equipment was conducted using sterile prewetted swabs, which were incubated in 50 μl of distilled H₂O (dH₂O) to elute the DNA. Swabs were incubated at 4°C for 4 h, vortexed, and centrifuged, and 2 μl of eluant was used as input for N2 qPCR, as described above.

Taken together, these data suggest that laboratory manipulation of SARS-CoV-2 RG cDNA poses a noninfectious contamination risk that results in false-positive RT-PCR results, causing unnecessary quarantine and contact tracing. This is likely true for other SARS-CoV-2 cDNAs, as reported by a recent preprint (4, 5). We note that when this work was initiated in our research laboratory, we were unaware of this possibility, and thus, no special precautions for handling SARS-CoV-2 cDNA fragments were taken (gloves and masks were worn, but no additional personal protective equipment [PPE]; work was performed on regular benchtop with no dedicated area, no aerosol containment, etc.). Strategies to reduce the likelihood of such contamination should include stringent standard operating procedures and the use of diagnostic tests that discriminate between cDNA and viral RNA, and/or target different regions of the viral genome, as exemplified by negative follow-up testing for subjects R4 and R5 using an assay that targets ORF1ab instead of N (Table 1).