Reinfection or reactivation: Genome‐based two distinct SNP profile of SARS‐CoV2 repositivity in an Indian case

To Editor

To date, a large number of case reports have emerged claiming COVID‐19 reinfection, some confirmed by the distinct viral genotypes in each episode. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ This communication highlights a case report from Delhi, India, with two instances of SARS‐CoV‐2 where the second episode witness moderate illness after 73 days of the first SARS‐CoV‐2 positive episode (asymptomatic). A 52‐year‐old male individual was tested positive at a COVID health care unit [RT‐PCR/Ct‐values (ORF1ab‐36.04/E gene‐36.74)—using Roche Cobas 6800] for SARS‐COV‐2 on June 12, 2020, upon contact tracing. At that time, the subject underwent home isolation and remained asymptomatic for the next 14 days, later detected negative for SARS‐CoV‐2 by RT‐PCR on June 27. The rest of the period remained uneventful until, 73 days later, on August 23, the patient developed low‐grade fever and body ache, without breathlessness/cough. The subject got positive for SARS‐CoV2 (ORF1ab/E gene; Ct values 17.9/17.8 respectively on Roche Cobas 6800) on August 24. During this episode, the oxygen saturation remained between 92% and 95%, and all the other blood biochemical investigations (complete blood cell count, liver function tests, blood urea, creatinine) were also within the normal values/indices range. On August 28, bilateral ground‐glass opacities on high‐resolution chest computed tomography images were noted. Of note, the serology investigation conducted for IgG (antibody ELISA) was detected negative on August 27th, possibly indicating either the lack of a strong detectable immune response or the waning of antibodies over time. ¹⁰ Later, on October 3rd, IgG was detected positive with a ratio of 6.2, reference value <0.8 (cut‐off OD/patient OD; 0.414/2.599) (Figure 1).

Figure 1.

The timelines of SARS‐CoV‐2 infection in the case under study

We estimated viral load on QuantStudio 6 and 7 Flex Real‐Time PCR systems using Labgun's Exofast SARS‐CoV‐2 RT‐PCR kit in both the samples. It indicated that episode‐1 sample, Ri‐Ep‐1 (92 × 10³ viral copies/ml; Ct value: 27.2) and episode‐2 sample, Ri‐Ep‐2 (187 × 10⁶ viral copies/ml; Ct value: 13.9) had a significant difference in viral load (Figure 2B). The difference in viral load of two episodes could also implicate that either the first episode was contaminated or a case of prolonged infection. Further, we confirmed, there was no sample contamination by short tandem repeat‐based analysis (GenePrint10 system; Promega) (Figure 2A). Consequently, we did viral genome sequencing to confirm the validity of the episodes being reinfection.

Figure 2.

Molecular detection, genome sequence analysis, and phylogenetic visualization of SARS‐CoV‐2 infection in two episodes. (A) The STR profile of both the samples displayed as multiple peaks of fluorescent signals of different sizes, showing a match between both the samples (Top: Episode 1; and Bottom: Episode 2 sample). (B) Amplification plot of N gene for Ri‐Ep‐1 and Ri‐Ep2 along with amplification curves of serially diluted synthetic controls. (C) The schematic representation shows the nucleotide variations profile of two samples of two different episodes (Top: Ri‐Ep‐1; and Lower: Ri‐Ep‐2 sample) and coverage read distribution across viral genomes has also been shown. Protein level changes are marked in red font. (D) The phylogenetic analysis by FigTree v.1.4.4 shows the clade distinction between two samples and similarity with another circulating clade (India/IIP‐309/ISL_528685) and in secondary contact at the time of the second episode (India/ST/ISL_581507). STR, short tandem repeat

QIASeq FX DNA library kit was used for viral genome sequencing of both the samples initially. Additionally, a modified ARTIC primer protocol was used for Ri‐Ep‐1 (due to low viral genome coverage by QIASeq FX). Sequencing was done on the MiSeq platform and data was analyzed using CLC genomics workbench 20.0.4 (Qiagen) and Freebayes for variant calling. The mean coverage and other statistics for each genome are as following: Ri‐Ep1‐(EPI_ISL_614157) [3070535 read, 1924x, N content −2871 (9.6%)] and for Ri‐Ep2‐(EPI_ISL_581506) [3958088 reads, 5815x, N content‐109 (0.003%)] was achieved (Figure 2C) (for detailed methods https://doi.org/10.6084/m9.figshare.13194515.v5).

In comparison to the Wuhan genome, the Ri‐Ep‐1 genome (Figure 2C) displayed six homoplasic variations (one upstream variation at C241T, three missense variations at 9419, 21456, 23403 (D614G), one deletion at 13993, and one insertion at 26381) whereas Ri‐Ep‐2 had 25 homoplasic variations (Figure 2C). At the variation level, both episodes' genome sequences shared three homoplasic variation sites, that is, C241T, A9419G, and A23403G, and another site was not covered at 3037 position in the Ri‐Ep‐1 genome. We evaluated all the variations in both the genomes called by Freebayes and confirmed them manually. A total of 118 variations were called and 46 variations had variant frequency above 30%. Interestingly, 15 out of 25 homoplasic variations of Ri‐Ep‐2 displayed a variable degree of heteroplasmy in the Ri‐Ep‐1 genome (Figure 2C). Besides, the first episode genome carries a significant number (n = 22) of other heteroplasmic sites (variant frequency between 0.31 and 0.7) (Figure 2C). We present a case of reinfection based on the clinical data and whole genome sequencing of two episodes of SARS‐CoV‐2 as the genomes of the two samples were distinct from each other. Also, different clades were predicted by phylogenetic analysis, Ri‐Ep‐1: 19A; Ri‐Ep‐2: 20A, and PANGOLIN lineage, B.1.0 and B.1.36.1, respectively (Figure 2D). This clearly suggests that these two viruses are of different origin with two distinct SNP profile of SARS‐CoV‐2 genomes, thus highlighting a possible scenario of reinfection against relapse. Yet, we cannot ignore the fact that the significant amount of heteroplasmy in the first episode could be host‐driven viral genome evolution and/or raises the possibility of reactivation. ¹¹

CONFLICT OF INTERESTS

The authors declare that there is no conflict of interests.

AUTHOR CONTRIBUTIONS

Mahesh S. Dhar, Partha Rakshit, and Mohammed Faruq conceived the study. Mahesh S. Dhar, Simmi Tiwari, RadhaKrishnan V. S., Robin Marwal, Ajit Shewale, and Tushar Nale planned and optimized diagnosis. Manoj Jais and Sanjib Gogoi for clinical analysis. Nishu Tyagi and Pooja Sharma performed the viral genome sequencing. Nishu Tyagi performed RT‐PCR for viral load estimation. Akshay Kanakan, Azka M. Khan, and Rajesh Pandey for sequencing validation. Vivekanand A. and Bharathram Uppili performed computational analysis for viral genome assembly. Variant calling and phylogenetic analysis performed by Bharathram Uppili, Mohammed Faruq, Vivekanand A., Nishu Tyagi, and Mahesh S. Dhar for data interpretation. Mohammed Faruq, Vivekanand A., Nishu Tyagi, Mahesh S. Dhar, and Partha Rakshit wrote the manuscript. Anurag Agrawal, Sandhya Kabra, and Sujeet Singh edited and reviewed the manuscript. All the authors read and approved the final manuscript.

DATA AVAILABILITY STATEMENT

The data analysis that supports the findings of this study is openly available in figshare at https://doi.org/10.6084/m9.figshare.13194515.v5. The consensus fasta generated from the Sequencing has been submitted in GISAID under the accessions EPI_ISL_614157 (Ri‐Ep1) and EPI_ISL_581506 (Ri‐Ep2) and the raw fastq data in Sequence Read Archive (SRR12991871 andSRR12991870) respectively.