Compatibility of Maximum-Containment Virus–Inactivation Protocols With Identification of Bacterial Coinfections by Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry

Abstract

Diagnostics and research analyses involving samples containing maximum-containment viruses present unique challenges, and inactivation protocols compatible with downstream testing are needed. Our aim was to identify a validated viral inactivation protocol compatible with bacterial identification by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). We assessed a panel of bacteria with 6 validated maximum-containment virus–inactivation protocols and report that inactivation with TRIzol or γ-irradiation is compatible with MALDI-TOF MS. The availability, simplicity, and rapidity of TRIzol inactivation make this method the more suitable choice.

The importance of developing maximum-containment virus–inactivation protocols that are compatible with diagnostic assays was demonstrated during the 2013–2016 West Africa Ebola epidemic, as 27 patients infected with Ebola virus (EBOV) were treated in advanced healthcare facilities in Europe and the United States [1]. These facilities and their associated laboratories had limited capacity to isolate and identify bacterial copathogens from clinical specimens recovered from patients with EBOV disease (EVD). A patient with EVD treated in Hamburg, Germany, developed severe gram-negative bacterial sepsis, but the etiological agent was not determined, because “[m]ore advanced tools for full identification of the organism and assessment of speciation were not accessible” [2p2397]. A recent study using an unbiased sequencing approach applied to blood specimens of patients with EVD cared for in Guinea suggests bacterial sepsis might frequently complicate EVD [3], further emphasizing the need for improved diagnostic assays to detect bacterial coinfection among patients with EVD. Current World Health Organization (WHO) EVD treatment protocols recommend empirical use of broad-spectrum antibiotics for patients treated in African Ebola treatment units [4], and 81% of patients with EVD treated in Europe or the United States during the 2013–2016 epidemic also received intravenous broad-spectrum antibiotics [1]. Colonization with multidrug-resistant bacteria, including Enterobacteriaceae species producing extended-spectrum β-lactamase, has been increasingly documented in the WHO African Region and has raised concern regarding the efficacy of empirical antibiotic use [5]. Improved diagnostic techniques and continued research may allow for more targeted and more effective use of antibiotics for the treatment of bacterial coinfections in patients with EVD or other viral hemorrhagic fevers.

Laboratory handling of specimens from patients with EVD requires effective virus inactivation and maintenance of specimen integrity for downstream analyses [6]. Here we report testing of 6 validated virus-inactivation protocols [7–9] for their downstream compatibility with bacterial identification by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). Each inactivation protocol has been validated with EBOV [7, 9] or vesicular stomatitis virus expressing EBOV glycoprotein (rVSV-EBOVgp-GFP) [8] as surrogates for a range of maximum-containment, enveloped, single-stranded, negative-sense RNA viruses, including other filoviruses, arenaviruses, bunyaviruses, orthomyxoviruses, and paramyxoviruses.

MATERIALS AND METHODS

Bacteria Species

Six representative bacteria commonly associated with bloodstream infections in humans were selected for evaluation in this study, including gram-negative and gram-positive species [2]: Escherichia coli (DH10B), Pseudomonas aeruginosa (PRD-10), Klebsiella pneumoniae (ATCC 13883), Acinetobacter baumannii (ATCC BAA-1605), Staphylococcus aureus (ATCC 29213), and Streptococcus pyogenes (ATCC BAA-947). To simulate protocols for bacterial isolation in clinical microbiology laboratories, aliquots of each bacterial species were first inoculated into Bactec Peds Plus aerobic medium culture bottles (Becton Dickinson, Franklin Lakes, NJ) and incubated in a Bactec FX instrument (Becton Dickinson) until positive for growth. Bacteria from culture positive for growth were subcultured by plating cells on nonselective sheep blood agar and incubating them overnight (35°C in 5% CO₂).

Inactivation Protocols

Six virus-inactivation protocols were assessed. The first is based on treatment with 8 megarads of γ-radiation. The remaining protocols are based on chemical lysis. Three rely on guanidinium thiocyanate–based lysis reagents: TRIzol (Life Technologies, Grand Island, NY), Buffer AVL (Qiagen, Hilden, Germany), and Buffer RLT (Qiagen). The final 2 protocols involve incubation in 10% neutral buffered formalin (NBF) and lysis with a sodium dodecyl sulfate (SDS) buffer, consisting of 4% SDS, 35% glycerol, 0.05% bromophenol blue, and 20% β-mercaptoethanol, buffered to pH 6.8 with 200 mM Tris, followed by heating for 10 minutes at 100°C.

In each inactivation protocol, approximately one-half loopful of a 10-µL loop of each bacterial species was collected in triplicate from the sheep blood agar as starting material. For γ-irradiation, the bacteria were suspended in tryptic soy broth containing 10% glycerol, vortexed, and, while frozen on dry ice, irradiated with a ⁶⁰Co source in a Model 484 irradiator (J. L. Shepherd and Associates, San Fernando, CA) until an 80,000 Gy dose was achieved. For TRIzol, the bacteria were placed directly in 1 mL of TRIzol, vortexed, and incubated at room temperature for 10 minutes. Proteins were then purified and pelleted according to the manufacturer’s instructions; however, the pellet was not resuspended in 1% SDS buffer and was instead frozen in pellet form. For AVL, the bacteria were first suspended in 250 µL of PBS. A total of 140 µL of this suspension was then transferred to 560 µL of AVL, vortexed, and incubated at room temperature for 10 minutes. The entire sample was transferred to a 2-mL tube containing 560 µL of 100% ethanol. For RLT, the bacteria were placed directly in 600 µL of RLT, vortexed, and incubated at room temperature for 10 minutes. The entire sample was then transferred to a 2-mL tube containing 600 µL of 70% ethanol. For 10% NBF and SDS buffer, the bacteria were placed directly in 1 mL of the respective reagent, vortexed, and incubated at room temperature for 10 minutes. The SDS samples were further heated at 100°C for 10 minutes.

MALDI-TOF MS

The MALDI-TOF MS testing was performed on a MALDI Biotyper Microflex LT system (Bruker Daltonik, Bremen, Germany). For TRIzol, the protein pellet was resuspended in 50 µL of 70% formic acid, and the suspension was mixed by pipetting; 50 µL of 100% acetonitrile was then added, and the suspension was vortexed to mix. For γ-irradiation, AVL, RLT, 10% NBF, and SDS buffer, the samples were centrifuged at 15000 × g for 2 minutes, and the supernatant was removed. A total of 50 µL of 70% formic acid was added to each pellet, and the suspension was mixed by pipetting; 50 µL of 100% acetonitrile was then added, and the suspension was vortexed to mix. Standards were prepared according to a routine MALDI-TOF protocol by suspension of half loopfuls of bacteria in 1 mL of 70% ethanol, centrifugation at 15000 × g for 2 minutes to pellet, discarding of supernatant, and resuspension of pellets in formic acid and acetonitrile as described above. Controls consisting of commercially prepared bacterial peptides were included on each run. One-microliter aliquots of each of these samples were spotted onto a microScout plate (Bruker), allowed to dry, and overlaid with Bruker Matrix HCCA solution (α-cyano-4-hydroxycinnamic acid in a standard solvent containing 50% acetonitrile, 47.5% water, and 2.5% trifluoroacetic acid). A log_(score) of ≥2.0 is recommended for reliable identification at the species level, and a log_(score) of 1.7 ≤ x < 2.0 is recommended by the manufacturer for confident identification at the genus level. A log_(score) of <1.7 does not provide any confident identification.

Statistical Analysis

It was required that all identifications within 10% of the top log_(score) value concorded with the reported genus (for genus-level identification) or species (for species-level identification) [10–12]. A 2-tailed t-test was used to assess significant differences in log_(score) values across inactivation methods. A P value of < .05 was considered statistically significant.

RESULTS

Sample Inactivation by TRIzol or γ-Irradiation Is Compatible With MALDI-TOF MS

Inactivation of bacterial samples from culture by γ-irradiation or TRIzol prior to analysis by MALDI-TOF MS allowed for confident and accurate bacterial speciation in all 18 samples tested. Although significant differences existed in some log_(score) values between the standard MALDI-TOF MS protocol and the samples inactivated by γ-radiation or TRIzol (Figure 1), all inactivated samples provided species-level confident log_(score) values of ≥2.0 that were also accurate. For E. coli and K. pneumoniae, the log_(score) values obtained after TRIzol inactivation were not significantly different from that of the standard protocol. Additionally, TRIzol inactivation provided significantly higher log_(score) values when compared to γ-irradiation for all gram-negative species tested. The other 4 inactivation protocols failed to consistently provide confident species-level or genus-level results and often produced no identifiable spectra (Table 1). Results from all inactivation protocols with a log_(score) value of ≥1.7 were accurate to the respective genus or species level and met the 10% rule.

Figure 1.

Bacterial identification log_(score) values for the 6 viral inactivation protocols and a standard protocol tested with 6 bacterial species. All samples were analyzed in triplicate, and data are presented as mean values with standard deviations. Manufacturer log_(score) recommendations are as follows: ≥2.0 indicates confident species-level identification, 1.7 ≤ x < 2.0 indicates confident genus-level identification, and <1.7 does not indicate confident identification. Statistical significance was calculated with a 2-tailed t-test. A. baumannii, Acinetobacter baumannii; E. coli, Escherichia coli; K. pneumoniae, Klebsiella pneumoniae; NBF, neutral buffered formalin; NS, not significant; P. aeruginosa; Pseudomonas aeruginosa; S. aureus, Staphylococcus aureus; SDS, sodium dodecyl sulfate; S. pyogenes, Streptococcus pyogenes.

Table 1.

Bacterial Species and Genus Identification Results, by Inactivation Protocol

	Inactivation method
	TRIzol	Buffer AVL	Buffer RLT	10% NBF	SDS Buffer	γ-Irradiation
Confident species identificationa (log(score) ≥2.0)	18/18	1/18	0/18	0/18	2/18	18/18
Confident genus identificationb (log(score) 1.7 ≤ x < 2.0)	NA	3/18	0/18	3/18	9/18	NA
No confident identification (log(score) < 1.7)	NA	14/18	18/18	15/18	7/18	NA

Data are no. of specimens with the characteristic/no. evaluated. See Materials and Methods for descriptions of the inactivation protocols.

Abbreviations: NA, not applicable; NBF, neutral buffered formalin; SDS, sodium dodecyl sulfate.

^aAll log_(score) values ≥2.0 produced accurate species level identification and met the 10% rule.

^bAll log_(score) values 1.7 ≤ x < 2.0 produced accurate genus level identification and met the 10% rule.

DISCUSSION

Each of the inactivation protocols described here is effective for the inactivation of maximum-containment, enveloped, single-stranded, negative-sense RNA viruses such as EBOV. However, the suitability of each protocol for maintaining bacterial proteins that are identifiable by MALDI-TOF MS was unknown. We expected that the γ-irradiation protocol would allow for the greatest preservation of protein integrity and thus produce the most robust MALDI-TOF MS results. The effect of γ-radiation on proteins in solution has been shown to be mediated primarily through an indirect mechanism involving radiolytic H⁺ and OH^- formed from water and can be mitigated 1000–10 000-fold if the proteins are irradiated while in a frozen solution [13], as was the case in our study. While the γ-irradiation protocol ultimately produced acceptable results, some damage to the proteins likely still occurred, as evidenced by the significantly decreased log_(score) values as compared to the standard and TRIzol protocols. Furthermore, the γ-irradiation protocol is likely of limited usefulness, owing to its lack of accessibility.

The TRIzol-based viral inactivation protocol is simple, economical, easily accessible, and rapid, and it is widely applicable in both clinical and research settings for speciation of bacteria in samples containing maximum-containment viruses, such as EBOV. The TRIzol inactivation protocol is optimized for the purification of protein and required only minor modifications for compatibility with MALDI-TOF MS analysis. The wash steps during purification of the protein pellet after TRIzol inactivation were adequate for removal of excess salts and other contaminants known to interfere with MALDI-TOF MS analysis by formation of adducts and signal suppression if present in sufficient concentrations [14, 15]. As the other inactivation protocols are not optimized for precipitation and purification of a protein pellet from the lysis solution, it is reasonable to posit that excess salts and other impurities remaining in solution were responsible for incompatibility with MALDI-TOF MS. Consistent with this hypothesis is that noticeable crystalline residue formed on the microScout plate from many of the sample aliquots. Given the success of the TRIzol inactivation protocol, further attempts at optimizing the other viral inactivation protocols for compatibility with MALDI-TOF MS analysis were not made. However, the negative results from the AVL, RLT, 10% NBF, and SDS buffer inactivation protocols are preliminary, and further optimization might render these inactivation methods suitable for MALDI-TOF MS.

A prerequisite for MALDI-TOF MS is the isolation of pure bacterial colonies, as polymicrobial samples generally produce spectral fingerprints that are not recognizable. The method described here thus requires sample culturing in liquid medium, followed by subculturing on solid medium, as is typical for other bacterial diagnostics. Although nonselective culture media are used, each of these steps potentially biases for bacteria possessing the most favorable growth kinetics at the given culture conditions. Further refinements in MALDI-TOF MS software algorithms and expanded spectral databases may allow for robust identification of polymicrobial samples, thus eliminating the need for subculturing on solid medium [16]. Samples must be handled appropriately within containment during culturing because these steps must take place prior to viral inactivation.

MALDI-TOF MS is a robust, high-throughput alternative to traditional biochemical analysis or 16S sequencing and is an emerging microbiological diagnostic platform of choice. Further studies with an expanded panel of bacteria and pathogenic yeast could further validate this promising method and ultimately improve diagnostic assays and research analyses involving copathogens in maximum-containment viral infections. The burgeoning threat of emerging viruses necessitates development of viral inactivation protocols that are compatible with a full range of downstream diagnostic analyses, such as that reported here, to maximize preparedness for inevitable future outbreaks.

Notes

Acknowledgments.  We thank Dr Ryan Relich at Indiana University School of Medicine for providing the S. aureus and A. baumannii isolates, Dr Frank DeLeo and Mr Brett Freedman in the Laboratory of Bacteriology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, for providing the E. coli, K. pneumoniae, and S. pyogenes isolates; and Cara Bushmaker at Marcus Daly Memorial Hospital, for providing the P. aeruginosa isolate.

Disclaimer. The funders had no role in study design, data collection, interpretation, or the decision to submit the work for publication. The content presented in this work is solely the responsibility of the authors and does not necessarily represent the official positions of the National Institutes of Health or the US government.

Financial support. This work was supported by the Intramural Research Programs of the National Institute of Allergy and Infectious Diseases, the National Institutes of Health Clinical Center, and the Marshall University Joan C. Edwards School of Medicine.

Potential conflicts of interest.  All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.