Ultrasensitive RNA biosensors for SARS-CoV-2 detection in a simple color and luminescence assay

The gold standard diagnostic method for Covid-19 infections is the RT-qPCR test, which is very sensitive, detecting viral RNA at concentrations as low as 500-1000 copies/mL (1). Nasopharyngeal samples that test positive have viral RNA at about 100,000 copies/mL on average (2). RT-qPCR relies on equipment that can be cost limiting. For example, the way the RNA is amplified for detection requires continuous cycles of temperature change for enzymes in the reaction to function and a sophisticated instrument to detect amplification in real-time. It also entails some expertise to set up and analyse results. The read-out of the technique is quantitative and would require an understanding of the reaction kinetics to be interpreted correctly.

In this preprint, the authors optimise a specialised RNA biosensor and combine it with an isothermal RNA amplification technique. Their sensor produces a colour or luminescence output in the presence of SARS-CoV-2 RNA. This approach is promising as it appears sensitive to low RNA levels and the results could be interpreted with a phone camera.

 

 

Sensing the presence of viral RNA:

Toehold RNA biosensors are long RNA molecules comprising three parts: a sequence complementary to the RNA you want to detect, a stem loop containing a ribosome binding site and a translation start site, and finally the sequence that will code for your chosen reporter protein. Such sensors have been used previously, for example in the detection of Zika virus (3). A core challenge in designing them is to choose an RNA sequence that will fold on itself to obstruct the start site in the absence of relevant RNA (sensor OFF = no reporter protein). Only when the trigger RNA is present, the sensor will bind it to allow synthesis of the reporter (sensor ON). To achieve this, the authors screened target RNA sequences in silico and in vitro to create a reliable sensor that would detect SARS-CoV-2 RNA. Their chosen biosensors recognise Orf1ab, a part of the virus genome that is invariant in >99% sequenced Covid-19 strains thus far, suggesting that this could be a widely applicable method. However, by itself this biosensor would require a high copy number (10¹²-10¹³) of RNA to work.

Amplifying RNA to make it detectable:

To amplify the RNA found in clinical samples, the authors used an isothermal amplification method that does not involve dramatic temperature cycling, called Nucleic Acid Sequence Based Amplification or NASBA (4). It involves the activity of three enzymes – a reverse transcriptase, an RNase and an RNA polymerase. Its efficiency relies on selection of optimum primers (short pieces of DNA/RNA that help initiate amplification). The authors screened several primers for NASBA to arrive at a combination that allowed their sensor to detect as low as 100 copies of starting synthetic RNA.

Testing real samples:

Next, 39 clinical samples (nasopharyngeal swabs) that had been previously tested by RT‑qPCR were analysed by this method. Rather encouragingly, the samples designated negative by RT-qPCR did not activate the sensor in this new assay, whereas the positive ones turned the sensor on i.e. the reporter gave a predicted colour shift.

A note on the read-outs:

By testing more than one reporter the authors have also demonstrated the modularity of their system. Their colorimetric reporter relies on the translation of the lacZ gene when the sensor is on. This produces an enzyme β-Galactosidase that gives a coloured product when provided an appropriate substrate. Here two substrates ONPG and CPRG were tested (5) with the latter being preferred. It gives a yellow to red colour change when the enzyme is active, and this was detectable by a cell phone camera. As an alternate, a luminescence-based reporter called nano-lantern (6) was also shown to work. This can be detected using a standard microplate reader.

What I like about this preprint:

In addition to the obvious fact that this could be an invaluable tool in the real world, I found the preprint very clear both in the experimental approach and in the way it is written. Additionally, the key computational methods used to optimise the biosensor and primers are detailed out with available code. Hopefully, this will help other researchers adopt this technique or even expand it toward other pathogens more readily.

Questions for the authors:

1. What challenges do you foresee in this method being usable widely for diagnostics?
2. Generally in RT-qPCR reactions, 10-1000 copies of RNA are recommended for use as template. Would you say that your technique is comparable to diagnostic RT-qPCR in terms of absolute sensitivity or are there other factors to be considered?
3. This may not be practically feasible, but I was curious whether the 3 clinical samples with Ct 36-38 that tested negative in your assay could be reanalysed with one of your other sensors (1, 17, 10) and/or a different NASBA primer pair? Broadly I’m wondering if some clinical swabs may contain reaction inhibitors that impair some sensors or primers more than others, and if there is a way to address that.

References:

1. Wang et al., Clinical Chemistry, Volume 66, Issue 7, 977–979 (2020)
2. Wyllie et al., Engl. J. Med., 383, 1283-1286 (2020)
3. Pardee et al., Cell, 165, 1255-1266 (2016)
4. https://international.neb.com/applications/dna-amplification-pcr-and-qpcr/isothermal-amplification
5. Serebriiskii & Golemis, Analytical Biochemistry, Volume 285, Issue 1, 1-15 (2000)
6. https://blog.addgene.org/plasmids-101-imaging-with-nano-lanterns

Tags: diagnostics, sars-cov-2

doi: https://doi.org/10.1242/prelights.27284

Read preprint (1 votes)