Highly Multiplexed Reverse-Transcription Loop-Mediated Isothermal Amplification and Nanopore Sequencing (LAMPore) for Wastewater-Based Surveillance

Wastewater-based surveillance (WBS) has gained attention as a strategy to monitor and provide an early warning for disease outbreaks. Here, we applied an isothermal gene amplification technique, reverse-transcription loop-mediated isothermal amplification (RT-LAMP), coupled with nanopore sequencing (LAMPore) as a means to detect SARS-CoV-2. Specifically, we combined barcoding using both an RT-LAMP primer and the nanopore rapid barcoding kit to achieve highly multiplexed detection of SARS-CoV-2 in wastewater. RT-LAMP targeting the SARS-CoV-2 N region was conducted on 96 reactions including wastewater RNA extracts and positive and no-target controls. The resulting amplicons were pooled and subjected to nanopore sequencing, followed by demultiplexing based on barcodes that differentiate the source of each SARS-CoV-2 N amplicon derived from the 96 RT-LAMP products. The criteria developed and applied to establish whether SARS-CoV-2 was detected by the LAMPore assay indicated high consistency with polymerase chain reaction-based detection of the SARS-CoV-2 N gene, with a sensitivity of 89% and a specificity of 83%. We further profiled sequence variations on the SARS-CoV-2 N amplicons, revealing a number of mutations on a sample collected after viral variants had emerged. The results demonstrate the potential of the LAMPore assay to facilitate WBS for SARS-CoV-2 and the emergence of viral variants in wastewater.


Experimental description of sample processing
Each sample was processed via electronegative filtration as previously described. 1,2In brief, MgCl2 was added to 150−200 mL of sample to a final concentration of 25 mM and the pH was adjusted to 3−4 to facilitate attachment of virus particles to biomass via salt-bridging. 3An aliquot of 1 µL/mL of calf-guard cattle vaccine (Zoetis, Parsippany-Troy Hills, NJ) was spiked into each sample as a bovine coronavirus (BCoV) source to validate filtration and extraction recovery efficiency.Samples were filtered through 0.45-µm mixed cellulose ester membrane filters (Millipore, Billerica, MA) and the filters were folded and torn into ~1 cm 2 pieces using sterile forceps.RNA extraction was conducted following a TRIzol/ethanol-based nucleic acid extraction protocol in a 96-well format system (Zymo Research) as described previously. 2 As a positive control, RNA was extracted from heat-inactivated SARS-CoV-2 suspension (VR1986HK, ATCC) using a QIAamp mini viral RNA kit (Cat.No. 52904, Qiagen, Hilden, Germany).Negative controls were prepared using nuclease-free water (Invitrogen, Waltham, MA).

Reference RT-ddPCR analysis
For ddPCR analysis, subsets of RNA extracts were shipped on dry ice to Hampton Roads Sanitation District (HRSD), Virginia Beach, VA.Details on ddPCR analysis using the Bio-Rad QX200 (Bio-Rad, Hercules, CA, USA) are provided elsewhere.The samples were considered positive when the concentration was >2× the limit of detection (LOD).LODs were calculated by running serial dilutions of Twist Synthetic SARS-CoV-2 RNA Control 4 (Twist Bioscience, San Francisco, CA) in seven replicates over six orders of magnitude.The LOD was the concentration at which > 60% of the replicates were positive.
The concentration of positive RNA control that was extracted from heat-inactivated SARS-CoV-2 suspension was estimated to be 511.35gc/µL.The concentrations of SARS-CoV-2 in WW #1-15 and their status (i.e., positive/negative) as determined by the reference ddPCR analysis are summarized in Table S2.

Figure S3 Gel electrophoresis of the individual (left) and pooled (right) RT-LAMP products
Gel electrophoresis was run for 10 of individual (left) and 12 of pooled RT-LAMP products by columns (right) to validate sample amplification by the RT-LAMP procedure.Multi-ladder-like bands with strong intensity were observed, thus demonstrating successful amplification by RT-LAMP. 6Unlike PCR, gel electrophoresis of RT-LAMP products is challenging since the various loop concatemers make it difficult to discriminate between specific and non-specific amplification.threshold of 20, 20 out of 30 samples returned true positives/negatives against short, medium, and long amplicons.(Identity cutoff = 0.8 for amplicon and identity cutoff = 1.0 for barcoded-FIP sequences).There was no significant difference across different length of amplicons.Thus, short amplicon was chosen for the following parameter optimization.1197 With the read-count threshold of 20, 26 out of 30 samples returned true positives/negatives with the identity cutoff of 0.8 and 0.7 for amplicon.And 23 and 21 out of 30 samples returned true positives/negatives with the identity cutoff of 0.7 and 0.6.The identity cutoff of 0.8 for amplicon was chosen since it showed greater number of samples with true positives/negatives.

Figure S1 .
Figure S1.Alignments for RT-LAMP and RT-PCR primers (blue and red colored) against SARS-CoV-2 N gene.

Figure S2 .
Figure S2.Layout of 96-well plate that contains 96 RT-LAMP assays.Barcodes (BCs) from FIP for RT-LAMP and Rapid Barcoding Kit for nanopore sequencing were added to the RT-LAMP product before sequencing.
read-count threshold of 20, 5 and 6 out of 6 samples returned true positives at 10-and 20-fold dilution.But, only 4 and 3 out 6 samples returned true positives at 5-and 100-fold dilutions.

Figure S4
Figure S4 Quality score distribution over all sequences (left) and the boxplots of quality scores across all bases (right) for the sequencing reads from the samples in 2020 (top) and 2022 (bottom).

Table S2 .
RT-ddPCR analysis results targeting the SARS-CoV-2 N region for the wastewater samples

Table S3 .
RT-ddPCR analysis results targeting bovine coronavirus (BCoV) N region for the wastewater samples

Table S5 .
SARS-CoV-2 positive reads with alignment against barcoded-FIP sequences with different identities