A Simple and Safe Protocol for Preparing Brucella Samples for Matrix-Assisted Laser Desorption Ionization–Time of Flight Mass Spectrometry Analysis

Abstract

We describe a simple protocol to inactivate the biosafety level 3 (BSL3) pathogens Brucella prior to their analysis by matrix-assisted laser desorption ionization–time of flight mass spectrometry. This method is also effective for several other bacterial pathogens and allows storage, and eventually shipping, of inactivated samples; therefore, it might be routinely applied to unidentified bacteria, for the safety of laboratory workers.

TEXT

Brucellosis is a zoonosis caused by bacteria of the genus Brucella, which are transmitted to humans through direct contact, ingestion of contaminated animal products, or aerosolization (1, 2). Brucellosis is endemic in several regions of the world, with more than 500,000 new cases each year (2–5). To date, 10 recognized species of Brucella have been described, with the most pathogenic for humans being Brucella melitensis (2). Diagnoses of clinical brucellosis are made initially using serological tests and are confirmed by isolation of the agent (6, 7). The recent introduction of matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry (MS) has revolutionized the identification of bacteria and yeasts (8, 9). However, the procedures recommended by the manufacturers for non-biosafety level 3 (BSL3) organisms are not adequate for the manipulation of Brucella strains, which are classified as BSL3 (potential bioterrorism pathogens) and represent potential health hazards for laboratory workers (1, 10). Any inactivation procedure must allow safe sample handling outside the BSL3 environment and must avoid destruction of the biomarkers used for identification. Therefore, we developed a simple, safe, and efficient sample preparation method for Brucella isolates that is compatible with their analysis by the current MALDI-TOF MS platforms.

Solvent inactivation of Brucella.

We first applied the direct transfer procedure recommended by the manufacturer to a panel of Brucella strains, working in class II microbiological safety cabinets in a BSL3 laboratory (Table 1). The target plate (Vitek MS DS slide; bioMérieux) was inoculated by picking a portion of a colony from a plate with a 1-μl disposable loop (Sarsted), and the deposit was overlaid with 1 μl of α-cyano-4-hydroxycinnamic acid (CHCA) matrix solution (saturated solution of CHCA in a solvent mixture composed of 33% acetonitrile, 33% absolute ethanol, 3% trifluoroacetic acid, and 31% water) and then air dried for 2 min at room temperature. The dried material was recovered with a sterile swab and spread on culture plates, which were then incubated for up to 3 weeks under the optimal growth conditions for the corresponding bacteria indicated in Table 1. Viable bacteria were recovered for most strains, showing that this protocol does not kill all Brucella strains, possibly because the bacterial deposit was not completely covered with the matrix solution. This risk is not acceptable for BSL3 pathogens. To overcome this problem, different liquid-phase inactivation protocols were tested in 1.5-ml Eppendorf tubes (see Fig. S1 in the supplemental material). Cunningham and Patel recently suggested that a short treatment with 70% ethanol effectively kills Brucella; however, they used Oligella ureolytica as a surrogate (11). In our hands, treating two full loops of Brucella with 100 μl to 1 ml of 70% ethanol did not consistently inactivate all bacteria after 10 min of incubation, which suggests that ethanol should not be used for short inactivation of Brucella. Vortex-mixing of bacteria in 5 μl of the CHCA matrix solution was systematically very efficient. However, this small volume could not ensure that the liquid would cover all bacteria consistently. Because the CHCA matrix is an expensive reagent, we tested the efficiency of the solvent alone, with an aim of reducing costs. Two full loops of bacteria of Brucella strains that form small (Brucella papionis F8/08-60), medium (B. melitensis 16M), or large (Brucella inopinata BO1) colonies (corresponding to 1.8 × 10¹¹, 9.9 × 10¹¹, or 2.2 × 10¹² CFU, respectively) were vortex-mixed in 200 μl of freshly prepared solvent mixture. No colonies grew for any of the solvent-treated bacteria, showing that this protocol can inactivate up to 2 × 10¹² CFU of Brucella. After inactivation, samples were manipulated under BSL2 containment.

TABLE 1.

Summary of results obtained in this study

Strain name	Strain description	Cultivation conditionsa		CFU/plate after inactivationb				No. identified/no. analyzed by MALDI-TOF MS (%)c
		Temperature and time	Medium	Direct transfer	70% ethanol	CHCA matrix	Solvent	Direct transfer	Solvent	Solvent with no centrifugation
16M	Brucella melitensis biovar 1, reference strain (ATCC 23456)	37°C, 48 h	COS-B	4	2	0	0	NT	10/11 (91)	4/4 (100)
544	Brucella abortus biovar 1, reference strain (ATCC 23448)	37°C, 48 h	COS-B	8	>200	0	0	NT	6/6 (100)	NT
1330	Brucella suis biovar 1, reference strain (ATCC 23444)	37°C, 48 h	COS-B	>200	0	0	0	NT	6/6 (100)	NT
513	Brucella suis biovar 5, reference strain (NCTC 11996)	37°C, 48 h	COS-B	NT	0	0	0	NT	6/6 (100)	NT
RM6/66	Brucella canis, reference strain (ATCC 23365)	37°C, 48 h	COS-B	2	18	0	0	NT	6/6 (100)	NT
5K33	Brucella neotomae, type strain (ATCC 23459)	37°C, 48 h	COS-B	0	0	0	0	NT	6/6 (100)	NT
BO1	Brucella inopinata (BCCN 07-01)	37°C, 24 h	COS-B	0	2	0	0	NT	12/12 (100)	NT
BO2	Brucella inopinata-like	37°C, 24 h	COS-B	NT	0	0	0	NT	10/10 (100)	NT
CCM4915	Brucella microti (BCCN 07-01)	37°C, 24 h	COS-B	0	0	0	0	NT	12/12 (100)	NT
F8/08-60	Brucella papionis (NVSL)	37°C, 96 h	COS-B	>200	0	0	0	NT	4/4 (100)	NT
B1/94	Brucella ceti, reference strain (NCTC 12891)	37°C, 5% CO2, 96 h	COS-B	NT	0	0	0	NT	7/8 (88)	2/2 (100)
B2/94	Brucella pinnipedialis, reference strain (NCTC 12890)	37°C, 5% CO2, 96 h	COS-B	NT	10	0	0	NT	4/4 (100)	NT
63/290	Brucella ovis, reference strain (ATCC 25840)	37°C, 5% CO2, 96 h	COS-B	NT	NT	NT	NT	NT	2/2 (100)	NT
NF2653	Brucella spp. (wild rodent, Australia)	37°C, 24 h	COS-B	NT	NT	NT	0	NT	13/13 (100)	NT
bIN1983	16M transformed with GFP-encoding plasmid	37°C, 48 h	COS-B	NT	NT	NT	NT	NT	7/7 (100)	NT
LMG 3301	Ochrobactrum intermedium	37°C, 24 h	COS-B	NT	NT	NT	NT	NT	2/2 (100)	NT
ATCC 49188	Ochrobactrum anthropi	37°C, 24 h	COS-B	NT	NT	NT	NT	NT	1/2 (50)	NT
BAA-747	Acinetobacter baumannii (ATCC BAA-747)	37°C, 24 h	COS-B	NT	NT	NT	NT	5/5 (100)	8/9 (89)	3/4 (75)
K56-2	Burkholderia cenocepacia	37°C, 24 h	COS-B	NT	NT	NT	NT	4/5 (80)	6/6 (100)	4/4 (100)
ATCC 29212	Enterococcus faecalis	37°C, 24 h	TSA-S	NT	NT	NT	NT	4/5 (80)	4/7 (57)	4/4 (100)
ATCC 25922	Escherichia coli	37°C, 24 h	COS-B	NT	NT	NT	NT	6/7 (86)	7/10 (70)	4/4 (100)
ATCC 10211	Haemophilus influenzae	37°C, 5% CO2, 24 h	CHOC	NT	NT	NT	NT	4/6 (67)	6/8 (75)	4/4 (100)
ATCC 19424	Neisseria gonorrhoeae	37°C, 5% CO2, 24 h	CHOC	NT	NT	NT	NT	3/6 (50)	5/6 (83)	4/4 (100)
ATCC 27853	Pseudomonas aeruginosa	37°C, 24 h	COS-B	NT	NT	NT	NT	7/8 (88)	9/10 (90)	4/4 (100)
ATCC 14028	Salmonella enterica serovar Typhimurium, reference strain	37°C, 24 h	COS-B	NT	NT	NT	NT	6/6 (100)	10/10 (100)	4/4 (100)
ATCC 25923	Staphylococcus aureus	37°C, 24 h	COS-B	NT	NT	NT	NT	6/6 (100)	8/8 (100)	4/4 (100)
LyfoCults	Escherichia coli strain used to calibrate Vitek MS system (ATCC 8739)	37°C, 24 h	COS-B	NT	NT	NT	NT	449/464 (97)	2/2 (100)	NT

^aStandard cultivation conditions for the bacterial strains used. COS-B, Columbia agar with 5% sheep blood (bioMérieux); TSA-S, trypticase soy agar with 5% sheep blood (Becton, Dickinson); CHOC, chocolate agar with vitox (Oxoid).

^bEfficiency of the indicated inactivation methods with several Brucella strains. Shown are the maximal colony numbers obtained from several independent repeats.

^cResults of identification by MALDI-TOF MS using the direct transfer procedure or the solvent-based method, with or without centrifugation. Shown are the numbers of successful identifications/numbers of spectra analyzed, as well as the rate of success. NT, not tested.

MALDI-TOF MS analysis of solvent-inactivated Brucella.

Solvent-inactivated bacteria were centrifuged (10,000 × g for 2 min) and resuspended in ∼10 μl of solvent left in the tube. One microliter was spotted on target slides, air dried, and overlaid with 1 μl of CHCA matrix solution. All data presented were acquired using a Vitek MS Plus system (bioMérieux) and were analyzed with Launchpad V2.8 software and the Vitek MS research use only (RUO) 4.10 database. Only spectra containing 100 to 300 peaks were retained for analysis (except for the stability assays, in which all spectra were analyzed). Correct identification with a confidence value above 75% was considered successful identification.

After using this inactivation protocol on strain 16M, we obtained high-quality MALDI-TOF MS spectra (with 13 peaks with resolutions of ≥500) (Fig. 1), which allowed its identification as Brucella sp. (Tables 1 and 2). All tested cultivation conditions for B. melitensis allowed its correct identification, which is of importance since these conditions may vary among laboratories. Strain bIN1983 (strain 16M overexpressing green fluorescent protein [GFP], which is widely used in research laboratories) was also identified correctly (Table 2). All other Brucella strains analyzed (also inactivated with solvent) were correctly identified (Table 1). When testing B. melitensis strain 16M, we found that concentration of the samples by centrifugation was not required (Table 1); the 200-μl suspensions contained enough biomass to allow direct spotting of 1 μl on the target slide, thus reducing the sample processing time. In the event of nonidentification, the remaining sample could be concentrated by centrifugation and retested. The solvent inactivation procedure was also applied to a panel of other bacterial pathogens that are regularly encountered in hospital laboratories. This method allowed their correct identification by MALDI-TOF MS (Table 1).

FIG 1.

Representative MALDI-TOF MS spectrum obtained for Brucella melitensis strain 16M that had been inactivated using the solvent-based protocol. Resolution (Resol.) values and signal/noise (S/N) ratios for the main masses are given in the table. Int., intensity.

TABLE 2.

Rate of identification of B. melitensis 16M, cultivated under different conditions and inactivated with the solvent mixture before MALDI-TOF MS

Strain	Mediuma	Incubation conditions	No. identified/ no. analyzed (%)b
16M	COS-B	24 h	2/2 (100)
16M	COS-B	48 h	10/11 (91)
16M	COS-B	48 h, +5% CO2	2/2 (100)
16M	COS-B	72 h	2/2 (100)
16M	COS-B	96 h	2/2 (100)
16M	COS-B	96 h, +5% CO2	2/2 (100)
16M	COS-O	48 h	2/2 (100)
16M	MHB	48 h	3/4 (75)
16M	MHF	48 h	3/4 (75)
16M	BAS	48 h	6/6 (100)
16M	BBA	48 h	4/4 (100)
16M	CHOC	48 h	4/4 (100)
16M	TS	48 h	2/2 (100)
16M	TSA-S	48 h	3/4 (75)
bIN1983c	COS-B	48 h	7/7 (100)

^aCOS-B, Columbia agar with 5% sheep blood (bioMérieux); COS-O, Columbia agar with 5% sheep blood (Oxoid); MHB, Mueller-Hinton agar with 5% sheep blood (Bio-Rad); MHF, Mueller-Hinton agar with 5% horse blood and β-NAD (Bio-Rad); BAS, Brucella agar with 5% sheep blood, hemin, and vitamin K₁ (Becton, Dickinson); BBA, Brucella blood agar (bioMérieux); CHOC, chocolate agar with vitox (Oxoid); TS, homemade trypticase soy agar; TSA-S, trypticase soy agar with 5% sheep blood (Becton, Dickinson).

^bShown are the numbers of successful identifications/number of spectra analyzed, as well as the rate of success.

^cbIN1983, strain 16M plus pMR10-GFP.

Sample stability.

Samples prepared from B. melitensis 16M in the solvent mixture were stored at room temperature (25°C) or at 4°C prior to analysis by MALDI-TOF MS. For samples stored at 4°C, the percentage of identified samples remained above 70% for 6 days; for samples stored at room temperature, the spectrum quality and hence the identification rate decreased after 3 days (Fig. 2). These findings suggest that these microbiologically safe samples could be shipped for analysis on MALDI-TOF MS platforms, provided that national and international regulations are respected. This is of particular interest for low-income countries, where brucellosis and other highly infectious bacterial infections are common but frequently missed because laboratory facilities cannot afford such modern tools.

FIG 2.

MALDI-TOF MS spectra for Brucella melitensis strain 16M samples that had been stored in solvent at 4°C or 25°C for up to 2 weeks before analysis. Shown are the number of successful identifications/number of spectra analyzed, as well as the rate of success. For each condition, a representative spectrum is also shown.

In conclusion, we describe a simple approach for inactivation of Brucella isolates prior to their analysis by MALDI-TOF MS. This protocol is considerably faster than previously described protocols, which require several reagents, centrifugation steps, and long incubations (12–16) or expensive gamma ray sources (17). Sample preparation with the solvent mixture allowed correct MALDI-TOF MS identification of all Brucella strains tested and several other bacterial pathogens. Further studies are needed to determine whether the solvent-based protocol can be used to inactivate other bacterial pathogens and allow their identification by MALDI-TOF MS, which would allow routine application of this method to all unidentified bacteria, as a safety measure for laboratory workers and/or for the shipment of erstwhile infectious samples to mass spectrometry platforms.

Funding Statement

Institut National de la Santé et de la Recherche Médicale (INSERM), Université de Montpellier, and Institut de Veille Sanitaire (InVS) provided funding to David O’Callaghan through recurrent (yearly) sources. This work was also partially sponsored by bioMérieux, through a private research contract.