Homebrew: Protocol for glassmilk-based nucleic-acid extraction for SARS-CoV-2 diagnostics

Summary The gold standard protocol for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection detection remains reverse transcription quantitative polymerase chain reaction (qRT-PCR), which detects viral RNA more sensitively than any other approach. Here, we present Homebrew, a low-cost protocol to extract RNA using widely available reagents. Homebrew is as sensitive as commercially available RNA extraction kits. Homebrew allows for sample pooling and can be adapted for automation in high-throughput settings. For complete details on the use and execution of this protocol, please refer to Page et al. (2022).

1. Suspend 21 g silicon dioxide 325 mesh (Sigma 342890) in 40 mL 10% HCl in a 50 mL tube. Polypropylene tubes are appropriate. CAUTION: 10% HCl is highly corrosive. Only use under a fume hood when handling HCl. If diluting from a concentrated stock, add acid to water. This acid wash of silica matrix ensures getting rid of contaminants that may interfere with downstream qRT-PCR (e.g., RNAseP detection due to contamination).
a. Agitate suspension for 4 h on a tube roller. Secure the cap with parafilm to minimize the risk of spillages. If no roller is available move every 10 min manually by gently inverting the tube 10 times. b. Centrifuge at 2,000 g for 5 min. c. Carefully remove HCl with a pipette and store in appropriate double container for reuse. HCl can be reused for this washing procedure multiple times. Do not dispose of HCl by pouring down the drain with copious amounts of water if it has not been previously neutralized. Spills may be neutralized with sodium bicarbonate or baking soda. 2. Resuspend the silica pellet with 40 mL single deionized water.
a. Resuspend silica in a 50 mL polypropylene tube by agitation, vortexing at $ 134 3 g or repeated tapping against a surface or hand so no HCl is 'trapped' in the pellet. b. Centrifuge at 2,000 g for 5 min. c. Remove wash water and dispose of through the drain. Flush with plenty of water. 3. Repeat step 2 for a total of 6 washes. 4. Before removing the final wash, measure pH of the supernatant with a strip, it should be between 7 and 8. If acidic pH is detected perform more water washes and measure pH after each wash until pH 7-8 is achieved. 5. Remove the supernatant, the pellet should now be white. 6. Resuspend silica in a final volume of 30 mL of H 2 O -if possible, use MilliQ or RNAse free H 2 O. 7. Aliquot in clean 1.5 mL tubes and keep at 16 C-40 C.
Note: We have kept the glass milk for 18 months without loss in performance at temperature ranging from 4 C-40 C.
Preparation of SDS, NaCl, and 70% ethanol Alternatives: SDS can also be prepared from powder. However, this is highly dangerous and needs to be performed under the fume-hood.

Timing: 5 min
Prepare the GM_MB fresh every time prior to use. Follow this table for calculating volumes (in microliter) to prepare enough glassmilk master buffer GM_MB) for all samples. Consider always an excess of 10% (i.e., prepare enough for 11 samples if you are analyzing 10). The table below shows calculations for one and twenty-four samples.
CRITICAL: Make sure the master buffer is properly mixed and shake vigorously every 3 samples to avoid glassmilk pelleting.

Timing: 5 min
Prepare the GM_MB fresh every time prior to use.
Pooling samples can accelerate processing and considerably reduce the costs of sample processing. We have tested Homebrew and determined that the method allows for pooling of up to 19 negative samples with one positive sample .
For use of the pooled protocol, we found that 20 mL glassmilk is sufficient to capture the nucleic acid material of up to 20 swabs and obtain a clearly visible pellet of glassmilk without losing sensitivity. If using less than 20 swabs, only adjust NaCl and isopropanol amounts. The glassmilk-master buffer recipe will change to:

STEP-BY-STEP METHOD DETAILS
Note: When performing this protocol for the first time, we recommend either using spiked swab material with a known positive control (e.g., commercially available or laboratory-grown virus) or using swabs that have previously been tested by conventional extraction methods. Every extraction should include a positive control (in our case we employed laboratory-grown virus) and a negative extraction control (H 2 O). All steps in the extraction are performed at ambient temperature (16 C-40 C) unless otherwise stated.
We have performed this protocol on swabs from various different sources that have been stored at various temperatures (16 C to 40 C or À80 C) before analysis. As far as we can attest, all different viral transport media that are currently in circulation allow for successful isolation of viral material. Viral transport medium can be prepared as per CDC recommendation (https://www.cdc.gov/ coronavirus/2019-ncov/downloads/Viral-Transport-Medium.pdf), i.e., 2% FBS 100 mg /mL Gentamicin 0.5 mg /mL Amphotericin B in Hanks Balanced Salt Solution (HBSS). Please refer to Page et al., 2022 for a more complete characterization of the robustness of this method.

Inactivation and lysis of swab material (individual samples)
Timing: 1 min This step inactivates potential SARS-CoV-2 viral particles and aids disruption of enveloped viral particles and subsequent solubilization of genetic material.
We recommend heat inactivation using a dry bead bath (either 70 C or 90 C for 10-30 min) prior to addition of SDS as it allows safe handling of otherwise infectious samples while preserving sensitivity of detection (Lista et al., 2021). Inactivated swabs must ALWAYS be opened in a microbiological safety cabinet class I when possible (class II when I not available) for the safety of the handler. Once swab material is non inactivated samples can be handled on a regular bench.

Timing: 6 min
In this step, RNA from the lysed sample binds to glassmilk.
GM_MB contains NaCl as chaotropic agent, isopropanol to increase binding to glassmilk and glassmilk as RNA-binding matrix.
3. Prepare GM_MB buffer as per instructions above. Prepare enough for all samples and consider an extra 10% for pipetting errors. 4. Add 610 mL of GM_MB to 200 mL of lysed swab (100 mL 4% SDS + 100 mL swab sample). 5. Incubate 5 min at 16 C-40 C to allow RNA binding.

Mastermix addition and RNA binding (pooled samples)
Timing: 6 min 6. Prepare GM_MB buffer as per instructions above. Prepare enough for all samples and consider an extra 10% for pipetting errors. This will contain 20 mL of glassmilk and a multiple per sample of 400 mL isopropanol and 200 mL 1.25 M NaCL. Due to the increased volume, pooled samples are mixed in a Falcon tube. 7. Mix well by vortexing. 8. Incubate 5 min at 16 C-40 C to allow RNA binding.

RNA cleaning
Timing: 3 min In this step, RNA from the lysed samples binds to glassmilk.
The bound RNA is pelleted and washed with 70% ethanol twice to remove excess proteins in the sample and increase the sensitivity of the PCR.
9. Spin the sample at 4,500 g for 15 s.

OPEN ACCESS
10. Discard the supernatant by decanting into a waste basin. The pellet will not detach (Methods video S1: Decanting). 11. Disaggregate the pellet of GM by tapping the plastic tube against a rack or by vortexing (Figure 1 and Methods video S2, Disaggregation by tapping). a. If using the pooled protocol, resuspend the pellet from the Falcon tube and transfer with a pipette to a microcentrifuge tube. Continue processing as described for individual samples. 12. Add 500 mL of 70% ethanol and mix by flicking the tube.

Note:
Steps 11 and 12 can be inverted, i.e., add 500 mL of 70% ethanol to the GM pellet and then tap the plastic tube against a rack or vortexing to resuspend (Methods video S3, GM resuspension with ethanol).
13. Spin the sample at 4,500 g for 15 s. 14. Discard the supernatant by decanting (Methods video S1, Decanting). 15. Repeat steps 11-14. 16. Spin the tube at 7,000 3 g for 15 s to allow complete ethanol removal. 17. Remove all remaining 70% ethanol with a pipette. 18. Air dry at 65 C for 5 min in a heat block or equivalent to remove all ethanol from the pelleted GM with bound RNA (Figure 2).

Timing: 1 min
In this step RNA is released from the GM matrix in nuclease-free water.
To elute RNA from the dry GM matrix: 19. Resuspend pellet in 50 mL of nuclease-free water (Methods video S4, GM resuspension with water). 20. Spin the sample at 4,500 g for 15 s. 21. Take off 45 mL from the supernatant and transfer to a clean tube (Methods video S5, Elution). 22. Proceed to downstream detection of viral RNA. We routinely use qRT-PCR but other methods such as RT-LAMP (reverse transcription and loop-mediated isothermal amplification) or CRISPR-based methods can also be employed. Note: We employed 50 mL of nuclease-free water to resuspend the GM matrix. Higher volumes such as 70 mL may be employed but lower volumes will likely not resuspend the GM matrix and/or carry over matrix to the downstream reaction.
Pause point: Extracted RNA can be frozen at this step. Long term storage is recommended at À80 C but short-term storage ($1 week) can be done at À20 C.

Reverse transcription and PCR
This step allows the detection of specific genes from SARS-CoV-2 in the eluted RNA, utilizing an internal human gene (RNAse P) as control of extraction.
Prepare the RT-PCR reaction master mix with the following format. Importantly, always include a non-template control, in which the RNA is replaced with H 2 O. When possible, prepare the PCR reaction and plate at a different location to that where swabs are handled to reduce risk of contamination. The positive control from the RNA extraction can be employed as positive control in the RT-PCR.  Note: For this step pooled and individual samples are treated equally. i.e., both are eluted in 50 ml water and 5 ml are used per technical replicate. For the purpose of establishing this protocol, we performed RT-PCR reactions in technical duplicates, for analytical purposes, it is custom to just perform one technical replicate. The limit of detection for our method, employing this master mix is Ct 36 . Other qRT-PCR conditions may be employed (Lista et al., 2021;Reijns et al., 2020).

EXPECTED OUTCOMES
Please refer to Page et al., 2022 for a complete description and troubleshooting of the method, including RNA binding capacity, limitation to isolate human cellular RNA or lack of effect of inhibitors such as blood in the quantification of SARS-CoV-2 in swab samples. Homebrew allows for effective isolation of RNA and not DNA. In the study by Page et al., we systematically compare Homebrew to QIAamp Viral RNA Mini Kit (QIAGEN) as per CDC-recommendation. We typically are within 1-3 Ct of the values obtained with RNA isolated using the QIAamp kit. Recovery of pure cellular total RNA is $30% from a 5 mg preparation .
Although we recommend prior heat inactivation of swabs prior to the addition of SDS, SDS inactivates SARS-CoV-2 at concentrations as low as 0.5% (Patterson et al., 2020). We recommend prior heat inactivation to increase the speed of processing since all steps (including the addition of 4% SDS) can then be performed safely on a bench.
Quantification of SARS-CoV-2 material in RNA isolated from swabs using Homebrew has been validated employing the primer-probe and qPCR reagents listed above. According to our parameters, we consider negative/positive/inconclusive/void samples: Inconclusive samples are those in which viral presence is difficult to determine, and void samples are those that have little presence of RNAseP, insufficient to determine a true negative result. In both cases, our recommendation is to ask for a new sample from the same individual, for testing as soon as possible. Void samples may be due to technical errors such as adding variable amounts of GM to each tube, which are reduced when preparing GM_MB rather than adding components separately. In our experience these are caused by insufficient biological material or wrong storage (e.g., repeated freeze-thaw cycles). We iterate the need for including a positive and a negative extraction control, as specified above. We recommend a positive control with a Ct of around 30 to minimize cross-contamination and demonstrate effective RNA extraction. Negative controls (extraction and qRT-PCR) should not amplify viral targets or RNAseP at all. Small contaminations with RNAseP may occur given its presence in plastics but should be kept at least at 5Ct difference from the maximum allowed.
As shown in the study by Page et al. (Page et al., 2022) we do not see a reduction in sensitivity when pooling up to 20 samples. Pooling is only recommended in low prevalence settings, since positive pools of samples must be reanalyzed to determine which sample(s) are positive among the pool, and it depends on the pool size (Cherif et al., 2020). In our case, positive rate should be less than 1 in 20 (5%), otherwise all pools of 20 samples will need to be re-tested given that each pool will (statistically) be positive

LIMITATIONS
We have validated our method for SARS-CoV-2 RNA detection from combined nose and throat swab samples, including those that contain inhibitors such as blood . Our method allows for the sub-optimal isolation of cellular RNA from cells and further modifications to the lysis step would be required to improve its efficiency and the quality of isolated RNA. We did not measure the variability between different batches of GM preparations. RNA extraction reproducibility was demonstrated employing serial dilutions of SARS-CoV-2 samples as well as comparable results to gold-standard clinically diagnosed swabs in $ 100 samples . We recommend anyone starting to use Homebrew to carefully validate this protocol with swabs that have already been tested with alternative isolation methods until the user feels confident in our method. Although greatly decreasing the cost per sample, we have used our method with conventional qRT-PCR analysis. Sourcing the reagents for qPCR analysis is in our experience less of a bottleneck, but sufficient provision of reagents should be carefully planned in. Availability of real time thermocyclers is a limitation on the use of homebrew for qRT-PCR, but colorimetric and isothermal methods may still benefit from it (Alcá ntara et al., 2021; Baba et al., 2021;Joung et al., 2020).
Homebrew employing GM cannot be automated. However, our data employing carboxylated magnetic beads  show comparable sensitivity, with those being amenable to automatization.

TROUBLESHOOTING
Problem 1 GM not resuspending after centrifugation.

Potential solution
Pull tube over a microtube rack (Methods videos S1 and S2). This will create a vigorous small shaking that in our experience is most effective to resuspend the GM. Shake vigorously, vortex or flick the tube until resuspended.

Problem 2
Positive extraction control failure.
GM contains ethanol in the last step. Carry over ethanol can negatively affect downstream processing i.e., transcription and/or PCR. A void result in the positive extraction control suggests that the extraction has not been efficient, and ethanol carry-over must be the first step to check. It is important to check if more than 50 mL is present in the tube after elution; typically GM absorbs 5 mL of the 50 mL and thus carry-over ethanol can be easily spotted at this stage. Figure 3 depicts a wet (left) vs. a dry (right) GM pellet.

Potential solution
Air dry the pellet for 2 extra minutes.

Problem 3
Multiple void results.
We typically consider a failed extraction batch when more than 10% of all samples extracted are 'void', i.e., there is no amplification of the internal control (RNAseP in our case). This is likely due to carry-over ethanol or deficient mixing of the GM-MB.

Potential solution
Repeat the extraction of those samples (with positive and negative extraction controls) ensuring adequate mixing of GM-MB, incubation with the samples and air-drying of the GM pellet.

OPEN ACCESS
Ensure that all reagents have been filtered through a 0.2 mm PES membrane to remove contaminating RNases. Repeat extraction with a confirmed positive swab in parallel.

Problem 4
Contamination of negative extraction control.
Given that PCR is an exponential reaction, it can be relatively easy to contaminate reagents with PCR products.

Potential solution
To minimize this risk we recommend performing RNA extractions and qRT-PCR set up in different benches/rooms when possible. Aliquoting of negative and positive controls will also minimize risks of cross-contamination. Use of filter tips is highly recommended. If there is suspected contamination of one of the reagents discard and aliquot from the stock. If there is suspected contamination of the glassmilk, perform an acid wash as in the preparation steps.
Problem 5 GM pellet is over dried.
We recommend air drying the GM pellet at 65 C for 5 min (Figure 2).

Potential solution
If sample appears over dried i.e., the GM pellet does not resuspend (Methods video S4, GM resuspension with water), consider warming up water at 65 C prior to resuspending again. If this does not resuspend the GM pellet, we recommend re-extraction and careful monitoring of the drying process.

RESOURCE AVAILABILITY
Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Rocio T Martinez-Nunez (rocio.martinez_nunez@kcl.ac.uk).

Materials availability
This study did not generate new unique reagents.