Enzymatic and Chemical-Based Methods to Inactivate Endogenous Blood Ribonucleases for Nucleic Acid Diagnostics

Abstract

There are ongoing research efforts into simple and low-cost point-of-care nucleic acid amplification tests (NATs) addressing widespread diagnostic needs in resource-limited clinical settings. Nucleic acid testing for RNA targets in blood specimens typically requires sample preparation that inactivates robust blood ribonucleases (RNases) that can rapidly degrade exogenous RNA. Most NATs rely on decades-old methods that lyse pathogens and inactivate RNases with high concentrations of guanidinium salts. Herein, we investigate alternatives to standard guanidinium-based methods for RNase inactivation using an activity assay with an RNA substrate that fluoresces when cleaved. The effects of proteinase K, nonionic surfactants, SDS, dithiothreitol, and other additives on RNase activity in human serum are reported. Although proteinase K has been widely used in protocols for nuclease inactivation, it was found that high concentrations of proteinase K are unable to eliminate RNase activity in serum, unless used in concert with denaturing concentrations of SDS. It was observed that SDS must be combined with proteinase K, dithiothreitol, or both for irreversible and complete RNase inactivation in serum. This work provides an alternative chemistry for inactivating endogenous RNases for use in simple, low-cost point-of-care NATs for blood-borne pathogens.

Nucleic acid amplification tests (NATs) are the gold standard of diagnosis for many infectious diseases and biomarkers, with more reaching the market every year.¹ There is a push to bring NATs to the point of care (POC) to rapidly provide diagnostic information in clinical settings without the need for off-site laboratory testing.² Specifically, recent efforts to thwart disease incidence in low-resource settings aim to decentralize testing for prominent RNA viruses, including HIV, hepatitis C, Ebola, Zika, and dengue.3, 4, 5 Development of effective POC NATs is challenging because of the technical complexities of detecting RNA. The typical steps required for NATs include lysis to release target RNA, nucleic acid extraction, and amplification using PCR or isothermal methods.⁶

Blood is a favorable sample for POC NATs because it is a well-mixed fluid continuously circulated through the cardiovascular system that carries many infectious pathogens and diverse human host cells.⁷ One of the challenges in using blood for molecular testing is the high concentration of ribonucleases (RNases) in blood that degrade exogeneous RNA as part of a natural defense system from infectious nucleic acids.⁸ Tsui et al⁹ found that free RNA incubated with blood plasma for only 15 seconds was degraded such that 99% of the RNA could not be amplified via RT-PCR. RNases are among the most stable enzymes known to microbiologists, making them difficult to inactivate.¹⁰ RNase A has been widely studied for decades, and some have posited that its compact structure, controlled by persistent disulfide bonds, lends increased protection from denaturation.^11,12 RNases are generally stable in pH extremes and weak denaturing solutions.¹³ Heat inactivation of RNase A is reversible for short incubations up to 90°C.^14,15 Furthermore, human blood serum RNases are diverse in size, structure, and function.^16,17 Completely inactivating a wide range of endogenous RNases with differing stability poses a challenge for POC NATs.

Commercial NATs targeting RNA biomarkers in blood samples commonly use lysis buffers that inactivate RNases to protect RNA from degradation.¹⁸ The typical chemistry fully denatures all proteins present in a sample with high concentrations of chaotropic salts, specifically 4 to 6 mol/L guanidinium chloride or guanidinium thiocyanate.¹⁹ Guanidinium-based lysis buffers can be paired with a silica column that adsorbs nucleic acids before subsequent elution, otherwise known as solid-phase extraction.²⁰ Qiagen (Hilden, Germany) DNA and RNA purification kits are the gold standard for conducting solid-phase extraction in laboratory settings.²¹ Solid-phase extraction adequately protects RNA from degradation and returns a high yield, making it the predominant sample preparation technique in blood-based NATs.¹⁸ Another widely used RNA purification protocol, developed by Chomczynski and Sacchi,²² uses guanidinium-based lysis and phenol-chloroform extraction. Several washing steps and/or buffer exchanges are required to remove guanidinium, phenol, chloroform, and alcohols before downstream assays, and automation of these processes necessitates centrifuges, pumps, or robotics, which complicates their use in POC devices.²³

There are ongoing research efforts to circumvent solid-phase extraction or other cumbersome nucleic acid purification techniques (eg, phenol extraction) to simplify POC NATs. Electrokinetic techniques, which leverage the negatively charged phosphate backbone of nucleic acids, have proved effective in extracting nucleic acids from a wide range of complex samples in microchips and paper-based devices that are well suited for POC applications.24, 25, 26, 27 Several studies have extracted nucleic acids from complex samples with electrostatic attraction to chitosan-coated surfaces that switch charge depending on pH, allowing for capture, washing, and elution without the use of chaotropes or phenol.^28,29 An alternate approach is to eliminate nucleic acid purification and amplify targets directly from crude lysates (eg, blood, urine, or saliva). Amplifying crude lysates has been enabled by a combination of novel additives and mutant forms of PCR polymerases with increased tolerance to traditional inhibitors, including whole blood, sputum, and stool.^30,31 For example, the Phusion Blood Direct PCR Kit (Thermo Fisher Scientific, Waltham, MA) claims polymerase activity is retained even in the presence of up to 40% whole blood, according to the product bulletin. This direct PCR method has been applied to pathogenic nucleic acid targets in whole blood and serum.32, 33, 34 Recent isothermal amplification assays have exhibited excellent tolerance of crude samples and have many features congruous with POC use.^35,36 The sample preparation strategies reviewed herein have not been employed to detect RNA viruses in blood and may not be effective because of endogenous RNases and the potential for RNA target degradation. For tests that detect RNA in blood, methods must be used to inactivate RNases in lysates so that target RNA is protected from degradation.

In the absence of guanidinium, many sample preparation protocols employ the use of enzymes and surfactants, such as proteinase K, pronase, Triton X, Tween, and SDS.³⁷ The lytic mechanisms and properties of these reagents are reported elsewhere,38, 39, 40 yet their utility for nuclease control in complex samples is not well characterized. Proteases irreversibly break down proteins by cleaving specific peptide bonds.¹⁰ Nonionic surfactants (eg, Triton X and Tween) disrupt viral envelopes or cellular membranes and solubilize proteins. SDS is a powerful anionic surfactant that at high concentrations denatures proteins by disturbing the noncovalent bonds that provide secondary protein structure.³⁹ Proteases and surfactants act in different ways to alter proteins and their native configurations. As such, they have been shown to remove RNase activity in samples such as HeLa and neuroblastoma cell cultures.^41,42 Most studies on RNase inactivation techniques were found to use samples with low concentrations of pancreatic RNase.43, 44, 45, 46 This is useful for identifying potential inhibiting reagents, yet human serum and other complex biological samples often contain high concentrations of salts, proteins, cells, and diverse RNases of varied structure. RNase inhibiting conditions are not necessarily universal across samples, so additional study is needed to identify chemistries that effectively inactivate diverse RNases in human serum, without the use of highly concentrated guanidinium.

This article investigates the effects of several prominent proteases and chemicals on RNase activity in human blood serum. The most common method for RNase detection is extended incubation of an RNA substrate with a test solution, followed by gel electrophoresis to identify possible degradation.⁴⁷ This technique is laborious and not amenable to high-throughput experimentation, so a commercially available RNA substrate that emits fluorescence when cleaved by an RNase enzyme was selected. This RNase activity assay allows for numerous parallel experiments on a microwell plate and generates real-time data of the activity over a 30-minute incubation time. This article reports on the performance of proteinase K, nonionic detergents, SDS, dithiothreitol (DTT), and other additives on inactivating RNases in human serum. This research explores the interplay between surfactants, nonspecific proteins, RNA, and RNases, which is crucial in selecting lysis and RNase inactivation chemistries in blood. Denaturing SDS concentrations must be combined with either proteinase K or DTT for irreversible and complete RNase inactivation in serum. Explanations for how these RNase inactivation schemes may be implemented in sample-to-answer diagnostic tests are provided. These findings provide valuable information for researchers looking for alternative strategies to guanidinium-based lysis and nuclease removal. This work presents a step toward simple sample preparation of RNA in blood samples for use in POC NATs.

Materials and Methods

Samples and Reagents

Samples are defined as volumes of pooled serum, individual serum, plasma, or RNase A that are tested for RNase activity. Pooled serum is sterile filtered from clotted whole blood of male donors with AB blood type (H6914; Sigma-Aldrich, St. Louis, MO). Individual serum was unfiltered and recovered from the whole blood of five different donors (BioIVT, Hicksville, NY). Donors included both males and females and different races, and ranged in age from 18 to 70 years, as detailed in Supplemental Table S1. Plasma was obtained from whole blood of a 23-year–old female donor pretreated with K₂EDTA anticoagulant (BioIVT). One milliliter of whole blood was centrifuged at 3000 × g for 10 minutes, and the plasma supernatant was removed with a pipette. The stock solution of RNase A (2250G; Thermo Fisher Scientific) at 30 U/L was diluted in diethyl pyrocarbonate–treated water (AM9906; Thermo Fisher Scientific) before testing. The manufacturer approximated 50 units of RNase A equivalent to 1 Kunitz unit.⁴⁸

Select experiments used serum pretreated by proteinase K (AM2546; Thermo Fisher Scientific) with or without additives. Samples were incubated for 1 hour in a water bath at 50°C. After incubation, the samples were immediately tested for RNase activity. All reagents used in this study are certified RNase free by the manufacturers. Reagents include dithiothreitol (D9779; Sigma-Aldrich), guanidinium chloride (G3272; Sigma-Aldrich), pronase (537088; Sigma-Aldrich), trypsin (T5266; Sigma-Aldrich), SDS (71725; Sigma-Aldrich), Triton X-100 (T8787; Sigma-Aldrich), poly(A) (27-4110; GE Healthcare, Chicago, IL), vanadyl ribonucleoside complexes (94742; Sigma-Aldrich), and Qiagen AVL viral lysis buffer (19073; Qiagen). Working solutions were diluted with diethyl pyrocarbonate–treated water.

RNase Detection Assay

The RNaseAlert Substrate Detection System (Integrated DNA Technology, Coralville, IA) was employed for testing RNase activity in samples. The detection method relies on short RNA oligonucleotides with fluorophores (fluorescein) and quenchers on either end. In their native form, the RNA substrates are quenched and do not fluoresce. RNase cleaves the RNA substrates, which separates the fluorophores from the quenchers and emits fluorescence (Figure 1). The intensity of the fluorescence is measured in real time using a fluorometer, providing temporal data that indicate the rate of RNase activity. Human serum was chosen as the primary sample in this study because it carries the five known human blood RNases.¹⁷ Whole blood contains red blood cells rich with hemoglobin that has high absorbance at the emission wavelengths of fluorescein. Unless significantly diluted, hemoglobin absorbance was observed to obscure the fluorescence signal from the RNaseAlert assay. Human serum is similar to plasma except that it has fibrinogen and clotting factors removed, making it easier to store and manipulate in laboratory settings.

Figure 1.

Illustration of the mechanism of the RNaseAlert assay used to test for RNase activity in serum and RNase samples. The short single-stranded RNA substrate is bookended with a fluorescein (F) molecule and quencher (Q). When the sugar-phosphate backbone is cleaved by an RNase, the F is spatially separated from the Q, allowing for fluorescence.

The manufacturer's recommended protocol for the RNaseAlert assay allows for some flexibility, depending on application. The current research supplements the recommended protocol and provides the necessary specifications to replicate this work. The RNaseAlert experiments were prepared in a lidded 96-well plate with black walls and clear bottom (3603; Corning Inc., Corning, NY). The total assay volume in each well was 100 μL. Quantities of 10 μL of RNaseAlert substrate and 10 μL of 10× RNaseAlert buffer were pipetted into each well; 60 μL of water or reagents were then added at the specified concentrations. When running experiments in parallel, it is crucial to simultaneously add samples to the wells so the incubation time with RNA oligonucleotides is equal. Toward this end, a 12-channel pipette was used to concurrently add 20 μL of samples to each well. The plate was immediately loaded into a plate reader (SpectraMax iD3; Molecular Devices, San Jose, CA). The excitation and emission wavelengths were 485 and 535 nm, respectively. The photomultiplier tube gain was set to low with an exposure of 140 milliseconds. The heating block in the plate reader was set to 37°C. The instrument agitated the plate and measured the fluorescence in the wells every 2 minutes over a 30-minute incubation time. For each set of data presented in this article, a positive control of RNase A (1.5 U/L) is presented. The maximum fluorescence of this positive control is used to normalize the fluorescence values of each data set.

Results

Differences in RNase Activity of Serum and Plasma Samples

Our goal was to compare the performance of common lytic chemicals and enzymes on inactivating serum RNases and shed light on the complex mechanisms of RNA degradation in serum. RNase activity was first compared in human plasma and a variety of serum samples, both pooled and individual donors. Figure 2A shows normalized fluorescence intensities from RNase activity assays of serum and plasma samples over a 30-minute incubation period. All data are normalized by the maximum average fluorescence of the RNase A (1.5 U/L) controls run in triplicate. In the RNase controls, fluorescence rapidly increases after adding RNase A samples to the wells containing RNA substrate. After 10 to 15 minutes of incubation, the fluorescence signal plateaus, indicating that all available RNA substrate was cleaved. For samples containing RNase activity, the initial fluorescence value at time = 0 is substantially higher than the respective value of the negative control. This is because of the short delay required to load the microplate into the reader and begin data acquisition, over which the RNA substrate in the assay begins to degrade. The serum and plasma experiments increased in fluorescence at differing rates, yet they all appear to plateau near a normalized fluorescence value of 0.75, notably lower than the RNase control. This lower plateau is attributed to two different sources. First, serum has some absorbance over the emission spectrum of fluorescein, which results in a decrease in measured fluorescence intensity. This hypothesis is supported in Supplemental Figure S1. Second, serum is highly proteinaceous, and it has been documented that nonspecific RNA-protein complexes are common in biological systems.⁴⁹ These complexes between serum proteins and the assay's RNA substrate may sequester a portion of the substrate, resulting in less total fluorescence.

Figure 2.

A: RNase activity in pooled serum (S), serum from five individual donors, and plasma from a donor. Samples were simultaneously added to a 96-well plate with RNaseAlert substrate, and fluorescence intensity was measured in 2-minute increments over a 30-minute incubation. B: Effect of guanidinium chloride (0.1 to 4 mol/L) on serum RNase activity. Serum with no guanidinium chloride is presented for comparison. The positive control (cntrl) is RNase A (1.5 U/L), and the negative control (Neg cntrl) is RNase-free water. Serum with 4 mol/L of guanidinium chloride (S + GC4) and S + QiagenAVL data are indistinguishable from the negative controls. RNase A (1.5 U/L) is the positive control, and its mean maximum fluorescence intensity is used to normalize all data points. The negative control is RNase-free water added to the RNaseAlert assay. All data plotted are averaged triplicates with error bars of 1 SD around the mean. Data points where no error bars are visible have errors too small to plot. N = 3.

Differences in the fluorescence curves between the individual serum samples are observed. This indicates that RNase activity differs between blood donors. There are documented differences in blood RNase activity, but this has primarily been observed in patients with certain conditions, such as pancreatic cancer.⁵⁰ The pooled serum exhibited the lowest RNase levels, which may be due to the particular collection technique. Pooled serum was chosen for use as the primary sample in this study because it is similar in RNase activity to plasma and other serum samples. This pooled serum is also commercially available so others may replicate these results.

Guanidinium-Mediated RNase Immobilization

Guanidinium is the most common chemical used in RNA purifications and considered the gold standard for inactivating diverse RNases.⁵¹ It is commonly used at concentrations between 4 and 6 mol/L to denature proteins.^19,20 The current article investigated the effects of guanidinium on inactivating serum RNases as a comparison benchmark and if lower concentrations may be used. Figure 2B plots the fluorescence from the RNase detection assays of serum treated with guanidinium chloride at 100 mmol/L to 4 mol/L. Lower concentrations of guanidinium reduce the activity of endogenous serum RNases, but complete inactivation is achieved at a minimum of 4 mol/L concentration. A widely used commercial lysis buffer for immobilizing RNases in blood samples is the Qiagen viral lysis buffer (AVL buffer), composed of guanidinium thiocyanate and a detergent. This buffer is also highly effective at eliminating activity of serum RNases (Figure 2B).

Effects of Nonionic Surfactants

The nonionic surfactant, Triton X-100, was tested for its RNase inactivation properties. Nonionic surfactants are well suited for simple sample preparation strategies because they do not inhibit PCR or other amplification assays, even at high concentrations.⁵² Therefore, there is no need for onerous purification methods to remove the surfactant. Nonionic surfactants solubilize proteins rather than fully denaturing them, so enzyme activity may be retained in the presence of even high concentrations of surfactant.³⁹ The effect of Triton X-100 on RNase activity were explored (Figure 3A). All concentrations of Triton X-100, ranging from 0.1% to 2% w/v, increased the rate of RNA degradation in serum when compared with an untreated sample. This same effect was observed with Tween-20, another common nonionic surfactant (Supplemental Figure S2). A possible explanation is that nonionic surfactants directly interact with RNases and increase their activity, which has been observed for other enzymes in molecular biology.⁵³ Figure 3B plots the activity of RNase A in the presence of Triton X-100. There are no distinguishable differences in the slopes of the fluorescence curves of the detection assay, suggesting Triton X-100 does not significantly affect RNase activity directly. Instead, nonionic surfactants are hypothesized to increase RNase activity in serum by disrupting protective complexes formed between serum proteins and RNA. This theory is expounded on in Discussion.

Figure 3.

A: RNase activity of untreated serum (S) versus serum treated with Triton X-100 (TX; 0.1% to 2% w/v). B: Effect of Triton X-100 on the activity of RNase A. The HighRNase sample is the typical RNase A control (cntrl; 1.5 U/L). The LowRNase samples with and without Triton X-100 all contain 0.3 U/L RNase A. Either 0.5% or 1% w/v Triton X-100 was added to RNase A. Neg, negative.

Protease-Mediated RNase Removal

Proteinase K is a broad-spectrum serine protease that has been used for decades in DNA and RNA preparations to remove nucleases from biological samples.⁴¹ The ability of proteinase K to inactivate serum RNases was examined here. Samples of serum were pretreated with proteinase K for 1 hour at 50°C, as advised in the product manual, and then tested for RNase activity. The concentrations of proteinase K during the pretreatments ranged from 5 to 1000 μg/mL. Figure 4A reports the normalized real-time fluorescence values of each digested serum sample over the 30-minute RNaseAlert assay, and Figure 4B plots the end point fluorescence to highlight the relationship between proteinase K in the digestions and resulting RNase activity. Increases in fluorescence are observed in all serum samples compared with negative controls, indicating proteinase K is unable to fully deactivate serum RNases. The proteinase K stock was confirmed to be free of RNase activity with a proteinase K–only control. There appears to be an optimum concentration of proteinase K at 50 μg/mL to protect RNA (Figure 4B). Incubations at high concentrations of proteinase K lead to extensive degradation of the RNA substrate, even when compared with untreated serum.

Figure 4.

A: Inactivation of serum RNases via proteinase K (PK; 5 to 1000 μg/mL). Serum (S) samples are pretreated with proteinase K for 1 hour at 50°C, and resulting digested serum (DS) is tested with the RNaseAlert assay. The real-time fluorescence from the assay during a 30-minute incubation with the substrate is plotted. B: Respective end point fluorescence values are provided to highlight the optimal proteinase K concentration for RNA protection in serum digestions. A positive control (cntrl) with RNase A (1.5 U/L) and a negative control (Neg cntrl) with nuclease-free water are presented. C and D: RNase activity in serum pretreated with proteinase K as well as various additives. Additives tested include 0.5 μg/μL poly(A; pA) carrier RNA, 2 mmol/L vanadyl ribonucleoside complex (VRC), 5 mmol/L dithiothreitol (DTT), and 0.5% w/v Triton X-100 (TX).

There are several additives that are known to inhibit RNase activity or improve the proteolytic performance of proteinase K. Several prominent additives were evaluated by digesting serum with 50 μg/mL proteinase K in the presence of 0.5 μg/μL poly(A), 2 mmol/L vanadyl ribonucleoside complex, 10 mmol/L DTT, or 0.5% w/v Triton X-100. Poly(A) serves as sacrificial carrier RNA that slows the degradation of RNaseAlert RNA substrate, and vanadyl ribonucleoside complex is a well-known RNase inhibitor. DTT is a disulfide bond reducer that helps unravel protein structures, making peptide bonds more susceptible to proteases. Triton X-100 similarly relaxes protein structures by weakening hydrophobic effects. Respective additive concentrations are typical values for inhibiting RNases or aiding proteinase K. Figure 4C plots real-time fluorescence over a 30-minute assay measuring RNase activity in the samples, with respective end point fluorescence displayed in Figure 4D. Data indicate proteolyic digestion to remove serum RNases is not significantly impacted by the selected additives, with the exception of Triton X-100.

The efficacy of protease-mediated RNase inactivation with RNase A samples was further explored. In one configuration, RNase A was pretreated with proteinase K for a 1-hour incubation before running the detection assay. In the other, both RNase A and proteinase K were combined at the initiation of the RNaseAlert assay. Pretreated RNase A samples were shown to have no detectable ribonucleic degradation (Figure 5A). When proteinase K and RNase A (1.5 U/L) is added at the initiation of the detection assay, RNase activity was eliminated within 10 minutes. It was also observed that a concentration of 500 μg/mL proteinase K degrades RNase A more rapidly than 50 μg/mL proteinase K. Proteinase K easily digests RNase A but is not able to remove endogenous RNases in serum. It has been noted in other studies that protease digestion efficiency varies widely across proteins. This may be due to differences in disulfide bonds, specific folds, glycosylation, or a combination thereof.⁵⁴ There may be certain serum RNases with structures resilient to proteases.

Figure 5.

A: Effect of proteinase K on RNase A activity. The RNase A activity is monitored in real time over a 30-minute period. For the RNase + PK500 (1.5 U/L RNase A and 500 μg/mL proteinase K) and RNase + PK50 (1.5 U/L RNase A and 50 μg/mL proteinase K) experiments, proteinase K and RNase A were not incubated together before the 0-minute mark of incubation. The other data set shown is RNase A (1.5 U/L) pretreated with proteinase K (50 μg/mL) for 1 hour at 50°C before adding to the RNaseAlert assay. These data are indistinguishable from the negative control (Neg cntrl). B: Shielding effect of bovine serum albumin (BSA) on RNase A. RNase A (1.5 U/L) activity was tested in the presence of BSA at various concentrations (0.5% to 0.01% w/v). TX, Triton X-100.

Figure 5B explores the effects of nonspecific proteins on shielding RNA from degradation. Increasing concentrations of bovine serum albumin (BSA) drastically reduce RNA degradation from RNase A. The minimum concentration of BSA needed for noticeable RNA protection was 0.05% w/v. It is notable that the highest concentration tested is 0.5% w/v or 5 mg/mL. Human serum albumin in blood is typically between 35 and 50 mg/mL.⁵⁵ Human serum albumin (66 kDa) and BSA (66 kDa) are similar in chemical makeup and function. Therefore, it is reasonable to assume that human serum albumin plays a role in protecting exogenous RNA from degradation by serum RNases. The addition of 1% w/v Triton X-100 was observed to result in faster ribonucleic degradation for the identical BSA concentration (Figure 5B). This highlights the detrimental effect of nonionic surfactants on RNA degradation in complex samples.

Immediate and Irreversible Inactivation of Serum RNases

The most effective reagent identified for RNase inactivation in serum samples is SDS. SDS is a powerful anionic surfactant and protein denaturant commonly used in lysis protocols.³⁹ Figure 6A reports RNase activity in serum samples in the presence of SDS concentrations ranging from 0.1% to 2% w/v. At concentrations of ≥0.5%, RNase activity could not be detected, which suggests the surfactant generated fully denaturing conditions. Serum with 0.3% SDS partially inhibited RNases compared with an untreated serum sample. Serum treated with 0.1% SDS resulted in rapid degradation of RNA substrate in the detection assay. This indicates RNA degradation in serum is more pronounced at low concentrations of SDS, where RNases are not fully denatured. It is hypothesized that this is due to the same phenomenon observed with nonionic surfactants, such as Triton-X. In terms of concentration, SDS was found to be more potent than the gold standard reagent, guanidinium chloride, that must be used at concentrations >4 mol/L to completely prevent RNase activity.

Figure 6.

A: Effect of SDS (0.1% to 2% w/v) on serum (S) RNase activity. The positive control (cntrl) is RNase A (1.5 U/L), and the negative control (Neg cntrl) is RNase-free water. Data for 0.5%, 1%, and 2% SDS are indistinguishable from those of the negative control. B: Irreversible inactivation of serum RNases using a combination of SDS and proteinase K, dithiothreitol (DTT), or both. The experimental data labeled S + SDS0.5/0.3 monitors serum in the presence of 0.5% SDS for 15 minutes, when a 1.66× dilution of RNase-free water is added to reduce the SDS concentration to 0.3%. Serum samples pretreated with a combination of SDS and DTT or proteinase K (PK). S + SDS + DTT is serum pretreated with 0.5% SDS and 75 mmol/L DTT for 30 minutes at either 50°C or room temperature. DS | SDS is serum digested with 0.5% SDS and 1000 μg/mL proteinase K at 50°C for 1 hour. DS | SDS + DTT includes 10 mmol/L DTT. Serum samples pretreated or digested with 0.5% SDS were diluted to a final concentration of 0.1% SDS in the RNaseAlert detection assay.

The crucial shortcoming of SDS is that it does not irreversibly inactivate serum RNases, which renature once SDS concentration is reduced. This effect is illustrated in Figure 6B, where an RNaseAlert assay containing serum and 0.5% SDS is diluted 1.66-fold after 15 minutes to a concentration of 0.3%. Serum RNases are observed to be initially fully inhibited, but the dilution of SDS allows RNases to renature and regain activity. Hilz et al⁵⁶ reported that SDS reversibly inactivated RNase A and a combination of proteinase K and SDS is needed for irreversible inactivation. Similarly, serum that is digested with 0.5% SDS and 1000 μg/mL proteinase K is irreversibly removed of RNase activity (Figure 6B). The 0.5% SDS in the digested serum is diluted to a final concentration of 0.1% SDS when added to the RNaseAlert assay. For irreversible RNase inactivation by proteolytic digestion with proteinase K, it was found that SDS must be present in the serum at denaturing concentrations of ≥0.5% (data not shown).

RNases are reliant on their disulfide bonds for sustained activity, so the reducing agent, DTT, has been shown to irreversibly inactivate RNases, most effectively at elevated temperatures.^45,57 A pretreatment of serum with 0.5% SDS and 75 mmol/L DTT for 30 minutes at 50°C was found to irreversibly inactivate serum RNases. This same incubation with SDS and DTT at room temperature does not irreversibly inactivate serum RNases (Figure 6B), indicating elevated incubation temperature is crucial. The most effective serum pretreatment protocol leveraged a 1-hour incubation at 50°C with 0.5% SDS, 1000 μg/mL proteinase K, and 10 mmol/L DTT, which resulted in digested serum with RNase activity indistinguishable from the negative control (Figure 6B). When removal of serum RNases solely with DTT was attempted, the serum firmly coagulated on heating and could not be manipulated. It is noted that there is a published technique for RNase inactivation in blood samples leveraging Triton X and highly alkaline conditions to accelerate DTT-mediated reduction of disulfide bonds at room temperature.⁵⁸ The current research replicated this recommended protocol and found that it did not inactivate serum RNases (Supplemental Figure S3).

Discussion

In this work, a variety of methods that may be used to inactivate or control RNase activity in human serum were investigated. First, serum RNases were confirmed to be immobilized by guanidinium, the gold standard reagent for such purposes. Guanidinium chloride was found to completely eliminate serum RNase activity at 4 mol/L and only partially reduce activity at lower concentrations. This finding is in agreement with a recent study observing partial unfolding or an expanded form of RNase HI at low concentrations of guanidinium and extensive protein unfolding of the hydrophobic core at concentrations of 3 to 4 mol/L.⁵⁹ Guanidinium must be purified away from RNA or significantly diluted before adding to PCR or other amplification assays. PCR can tolerate guanidinium up to approximately 50 mmol/L⁶⁰; therefore, a blood lysate at 4 mol/L guanidinium chloride would need to be diluted 80-fold to enable PCR amplification. A lysate dilution of this magnitude is not preferred for POC NATs because it drastically dilutes the analyte concentration, negatively affecting the limit of detection of the test. Dilution also may allow blood RNases to renature and regain activity.

The use of nonionic surfactants was discovered to lead to accelerated RNA degradation, in both untreated serum and serum digested with proteinase K. It is hypothesized that this surfactant-induced increase in RNase activity in serum may be a result of the disruption of protective protein-RNA complexes. IgG is present in high concentrations in serum, and Sidstedt et al⁶¹ observed that IgG readily binds to single-stranded DNA, which is similar in chemical structure to single-stranded RNA. Nonspecific ribonucleoprotein complexes have not specifically been studied in human serum, but free RNA is likely bound by multiple proteins in biological matrices, such as serum.⁴⁹ Ribonucleoprotein complexes generate a protection from RNase degradation.⁵⁶ Surfactants disrupt the hydrophobic interactions of ribonucleoprotein complexes and free the RNA, making it more susceptible for degradation. It is hypothesized that Triton X-100 prevents the formation of complexes between RNA substrate and serum proteins, allowing for rapid degradation of free RNA via serum RNases. Similarly, nondenaturing concentrations (specifically <0.3%) of the anionic surfactant, SDS, were found to accelerate RNA degradation in serum. This may be because of the same mechanism observed with Triton X-100.

Samples of RNase A spiked with BSA were observed to experience significantly less RNA degradation. This supports the theory that nonspecific complexes between proteins and RNA slow degradation. Although this finding does not address RNase inactivation, it frames a strategy that may be leveraged to protect RNA in samples with limited RNases, such as cell lysates or heavily diluted serum or blood. BSA is advantageous for RNA preservation because it is compatible with downstream nucleic acid amplification techniques, unlike other RNA preservation reagents, such as vanadyl ribonucleoside complexes. Only commercial BSA preparations that are confirmed to be RNase free should be used with RNA samples. Multiple commercial BSA products purified using the traditional heat shock fractionation method were found to have significant RNase activity (Supplemental Figure S4). There may be other proteins besides BSA that provide RNA protection in complex samples, including single-stranded binding proteins, which may have a higher affinity for exogenous RNA. Further study of the interplay between serum proteins, RNA, and surfactants could employ an electrophoretic mobility shift assay with carefully selected running buffer to mimic binding conditions in serum.⁶²

In this investigation of protease-mediated RNase inactivation, the data revealed two curious findings: proteinase K does not inactivate all serum RNases, and higher concentrations of protease result in more severe RNA degradation in serum. In addition to proteinase K, serum RNases were not removed by other broad-spectrum proteases, including pronase and trypsin (Supplemental Figure S2). In addition, serum RNases with an extended 24-hour proteinase K digestion could not be inactivated (Supplemental Figure S5). High concentrations of proteases result in more thorough digestion of proteins in a sample. This was confirmed to be true with SDS-PAGE experiments of digested serum samples (Supplemental Figure S6), suggesting that serum RNases are resistant to proteolysis. It is hypothesized that extensive degradation of serum proteins may remove protective ribonucleoprotein complexes, leading to a more rapid RNA degradation in the sample. Another possible explanation for the adverse effects of high proteinase K concentrations is that proteinase K may actually activate some endogenous RNases. A study from Simpson et al⁶³ found that endogenous RNase activity was activated by proteinase K in mitochondrial extracts from Leishmania tarentolae, a eukaryotic protozoan parasite. This finding has not been observed in serum RNases.

The most effective RNase inactivation strategy for serum samples identified is a combination of SDS, proteinase K, and DTT. Each reagent disrupts different aspects of protein structure, specifically denaturation, peptide bond cleavage, and disulfide bond reduction, respectively. Individually, the reagents are limited, but in concert, they are able to completely and irreversibly immobilize the diverse and abundant RNases present in human serum. This method of RNase inactivation was found to be equally as effective as gold standard guanidinium-based chemistries.

All digestions of serum with proteinase K were performed for 1 hour in the interest of experimental consistency. This incubation time is relatively long for use in POC tests, where rapid diagnostic results are desired. Although it was not the focus of this work, much shorter sample pretreatment times were observed to be possible by increasing the incubation temperature. All serum pretreatment protocols presented in this article dilute serum with reagents by no more than a factor of 1.1×. Higher serum dilution factors were found to reduce the incubation time necessary for complete RNase immobilization (data not shown).

As detailed in the beginning of this article, lysis and RNase inactivation chemistries that avoid high concentrations of guanidinium may not require solid- or liquid-phase extraction for RNA purification. For example, DTT is a common PCR additive and, at low concentrations, does not need to be purified away from target nucleic acids. DTT has been shown to be compatible with multiple reverse transcriptases and polymerases up to 20 mmol/L concentrations.⁴⁵ Proteinase K can inhibit nucleic acid amplification assays by digesting polymerases and other necessary enzymes, but it is permanently inactivated by a 5-minute incubation at 95°C. Proteinase K can also be removed from RNA samples by electrophoretic methods, such as isotachophoresis.²⁴ SDS is typically incompatible with amplification assays at high concentrations, but the inhibitory effect of SDS can be mitigated using nonionic surfactants, including Triton X and Tween 20.⁵² Potassium phosphate may also be used to precipitate SDS out of solution so that PCR can tolerate up to 0.3% w/v SDS.⁶⁴ A notable study lysed bacterial cultures with 0.5% SDS and 200 μg/mL proteinase K. The resulting lysate was heated to 95°C to inactivate proteinase K and added directly to a PCR master mix containing Tween 20 for direct amplification and detection of target bacterial DNA.⁵² This provides a helpful blueprint for how cumbersome RNA purification may be avoided when using SDS and proteinase K for lysis and RNase inactivation.

Conclusions

This article investigates alternative methods to the standard guanidinium-based buffers for inactivating endogenous RNases in human serum. The use of nonionic surfactants and proteases in blood is explored, as well as two approaches to completely and permanently inactivate serum RNases. Degradation of RNA substrate was more pronounced in serum samples treated with common nonionic surfactant, Triton X-100, than in samples without surfactant. It is hypothesized that this is due to the surfactants disrupting nonspecific ribonucleoprotein complexes in serum that protect RNA from degradation via RNases. Use of nonionic detergents in RNase inactivation chemistries should be avoided, as they appear to have a deleterious effect, even when used in combination with proteinase K. It was observed that proteinase K digests RNase A with ease yet is unable to remove all endogenous serum RNases. An optimum proteinase K concentration of 50 μg/mL for RNase inactivation in serum digestions was found, with higher and lower concentrations offering significantly worse results. These data suggest there may be certain serum RNase variants that are resistant to proteolytic digestion. This is an unexpected result that contradicts widespread a priori assumptions in microbiology that proteinase K eliminates nucleases in biological samples.

It was observed that proteinase K is only able to irreversibly inactivate serum RNases in the presence of denaturing concentrations of SDS (ie, ≥0.5%). SDS at these higher concentrations immediately inhibits serum RNases, but their activity is regained once the SDS concentration is lowered. Irreversible RNase inactivation by way of SDS requires an incubation combined with proteinase K, DTT, or both. These RNase inactivation techniques use reagents that do not necessitate subsequent solid- or liquid-phase extraction before real-time quantitative PCR, simplifying the diagnostic process. SDS, proteinase K, and DTT are relatively innocuous reagents, which allows for safe handling and disposal in resource-limited clinical settings. This is in contrast with solid- or liquid-phase extractions that use hazardous chemicals and flammable organic solvents. Ultimately, this work aims to assist researchers developing simple POC NATs for blood-borne RNA viruses or other RNA targets.

Supplemental Data

Supplemental Figure S1

Reduction in fluorescence from the RNaseAlert assay after the addition of human serum. The end point fluorescence of a 30-minute RNaseAlert assay is plotted after a sample of RNase A (1.5 U/L) has completely degraded the target substrate, labeled RNase control (cntrl). This is contrasted with the end point fluorescence after human serum is spiked into control. The serum reduces the intensity of the fluorescence signal emanating from the sample. The negative control (neg cntrl) is RNase-free water added to the RNaseAlert assay. All data plotted are averaged triplicates with error bars of 1 SD around the mean. N = 3.

Supplemental Figure S2

RNase activity in serum (s) samples in the presence of Tween 20 (Tw20) or pretreated with pronase. Samples were simultaneously added to a 96-well plate with RNaseAlert substrate, and fluorescence intensity was measured in 2-minute increments over a 30-minute incubation. There is a short delay between adding samples to the microplate and loading it into the plate reader, which results in some experiments with initial fluorescence values significantly greater than the negative control (neg cntrl). RNase activity in serum in the presence of nonionic surfactant, Tween 20, is higher than that of serum with no surfactant. Activity is greater in samples at 0.5% w/v Tween 20 than at 0.1%, as observed with Triton X-100. RNase activity of serum pretreated with the protease cocktail, pronase, for 1 hour at 50°C, was also measured. As with proteinase K, digestions at higher protease concentrations resulted in serum samples with increased RNase activity. RNase A (1.5 U/L) is the positive control, and its maximum fluorescence intensity is used to normalize all data points. The negative control is RNase-free water added to the RNaseAlert assay. N = 1.

Supplemental Figure S3

Dithiothreitol (DTT)–mediated inactivation of serum (s) RNases accelerated by highly basic conditions. A previously reported method for inactivating blood RNases claimed RNase activity was eliminated using a 5-minute incubation of human serum treated with 40 mmol/L DTT, 1% Triton X-100, and 0.4 mg/mL poly(A) carrier RNA at pH 12. This protocol was repeated, and the treated serum was tested with the RNaseAlert assay, which reported significant remaining RNase activity. As a control (cntrl), RNase-free water was treated at these same conditions, and its fluorescence does not increase over the RNaseAlert incubation period. An RNase A control, untreated serum control, and negative control (neg cntrl) are provided for comparison.

Supplemental Figure S4

RNase activity of the stock solutions of reagents used in this work. Herein, whether any reagents employed may have activity measurable via the RNaseAlert assay is investigated. Reagents tested include 200 μg/mL proteinase K (PK), 50 mg/mL UltraPure bovine serum albumin (BSA; Thermo Fisher Scientific), 50 mg/mL BSA 1 (B4287; Sigma-Aldrich), 50 mg/mL BSA 2 (A7030; Sigma-Aldrich), 0.5% w/v Triton X-100, 0.5% w/v SDS, 500 mmol/L guanidine hydrochloride (GH), 25 mmol/L EDTA, and 100 mmol/L ammonium chloride (AC). Two different BSA products from Sigma-Aldrich possess noticeable RNase activity. The UltraPure BSA has an elevated baseline fluorescence, but no noticeable increase in fluorescence over the incubation, indicating minimal or nonexistent RNase activity. Triton X-100 has no RNase activity but does emit some fluorescence at the same wavelength as the fluorophore in the RNaseAlert assay. Cntrl, control; neg, negative.

Supplemental Figure S5

RNase activity in serum samples digested with proteinase K (PK) for 1- and 24-hour durations. Digested serum (ds) contained either 100 or 800 μg/mL proteinase K with all incubations at 50°C. Serum RNase activity remains even after a 24-hour proteolytic digestion. RNase A (1.5 U/L) is the positive control (cntrl), and the negative control (neg cntrl) is RNase-free water. A serum-only control (s) is presented for comparison.

Supplemental Figure S6

SDS-PAGE of products from serum digested with various concentrations of proteinase K (PK). The protein ladder on the left side of the gel provides a reference for proteins ranging from 10 to 250 kDa. Undigested serum is presented as a no-digestion control. Digested serum (DS) with increasing amounts of PK shows more substantial proteolytic size reduction of serum proteins. The far-right lane of the gel shows products from a digestion with 1000 μg/mL PK and 0.5% w/v SDS. There is nearly complete digestion of serum proteins when SDS is incorporated with high PK concentration. All serum digestions were performed for an hour at 50°C.