Fluorescent Whole-Cell Hybridization with 16S rRNA-Targeted Oligonucleotide Probes To Identify Brucella spp. by Flow Cytometry

Abstract

A whole-cell hybridization assay with fluorescent oligonucleotide probes derived from the 16S rRNA sequence of Brucella abortus in combination with flow cytometry has been developed. With the three fluorescent probes selected, a positive signal was observed with all the representative strains of the species and biovars of Brucella and with a total of nine different Brucella clinical isolates. Using the B9 probe in the hybridization assay, it was possible to discriminate between Brucella suis biovars 2, 3, 4, and 5 and almost all the other Brucella spp. On the basis of differences in fluorescence intensities, no discrimination was established between Brucella spp. and other phylogenetically related microorganisms. No positive fluorescence signals were detected with any of the bacteria showing serological cross-reactions with Brucella spp. and with a total of 17 clinical isolates not belonging to the genus Brucella. These results suggest that the 16S rRNA whole-cell hybridization technique could be a valuable diagnostic tool for the detection and identification of Brucella spp.

Brucellae are gram-negative intracellular bacteria that are pathogenic for humans and many animal species. On the basis of differences in pathogenicity and host preferences, the genus Brucella has six recognized species: Brucella melitensis, B. abortus, B. suis, B. canis, and B. neotomae (8). At the DNA level, these Brucella spp. show a high degree of homology, thus forming a highly homogeneous and possibly monospecific group (36, 37). On the basis of the 16S rRNA sequence, Brucella has been found to be a member of the alpha-2 division of the class Proteobacteria and closely related to the Rhizobiaceae family and other related genera not included in this family (40).

A diagnosis of brucellosis in humans and animals is often difficult to establish. For the identification of Brucella spp., mainly bacteriological and immunological detection techniques have been used. However, since isolation of the etiological agent from clinical samples by conventional methods is not always possible (2), serological tests play a major role in the routine diagnosis of brucellosis (38). Although the sensitivity of these diagnostic tests for the detection of specific antibodies against Brucella spp. can range from 65 to 95%, their specificities can be low in areas where brucellosis is endemic, mainly due to the high prevalence of antibodies in the healthy population (6). Moreover, other gram-negative bacteria may cross-react with smooth Brucella spp., and vaccinated animals can also give false-positive results in such tests (9, 18). Thus, the development of new procedures for the detection and differentiation of brucellae is currently of great practical importance.

Nucleic acid-based detection methods are very promising tools for the diagnosis of brucellosis. The 16S rRNA gene sequence has been extensively used to elucidate the phylogenetic relationships among bacteria at intra- and intergeneric levels, and it is also an excellent target for diagnostic methods (3, 5). PCR assays using specific primer pairs derived from the 16S rRNA gene sequence of B. abortus have been developed for the detection and identification of Brucella spp. (10, 21, 30, 32, 33). Whole-cell hybridization with fluorescently labeled 16S rRNA-targeted oligonucleotide probes (3–5) has been applied for the detection and identification of different bacteria (19, 24, 26, 31, 34). In contrast to PCR, by rendering the probes complementary to more abundant 16S rRNA rather than to the genes encoding this RNA, the sensitivity obtained is more than sufficient for use in diagnostic tests (31). However, to date this method has never been used for both the detection and identification of Brucella spp. or the diagnosis of brucellosis.

Bacterial strains and growth conditions.

The bacterial strains used in this study and their sources are listed in Tables 1, 2, and 3. Reference Brucella strains and clinical isolates were grown on tryptic soy agar supplemented with 0.1% (wt/vol) yeast extract at 37°C for 48 h. Agrobacterium, Alcaligenes, Phyllobacterium, Mycoplana, and Ochrobactrum strains were grown on the same medium at 26°C. Rhizobium and Sinorhizobium strains were cultured in tryptone-yeast medium at 30°C. The remaining bacteria, including those showing serological cross-reactions with Brucella spp. (9), were grown on nutrient agar at 37°C for 24 h. Clinical samples were isolated from hospitalized patients and were identified by standard methods.

TABLE 1.

Fluorescence intensities obtained in whole-cell hybridization with reference microorganisms and clinical isolates unrelated to Brucella spp.

Bacterium	Strain	Source and origina	Mean fluorescence intensity with probeb:
			B6	B9	B13
Escherichia coli	25922	ATCC	24.1	5.1	31.4
Escherichia coli	35218	ATCC	20.5	3.5	31.1
Pseudomonas aeruginosa	27853	ATCC	17.1	5.5	18.6
Escherichia coli		HUS, human	13.3	3.4	15.3
Enterococcus faecalis		HUS, human	20.3	28.6	75.7
Klebsiella pneumoniae		HUS, human	41.1	38.9	16.7
Klebsiella pneumoniae		HUS, human	15.1	10.1	21.6
Pseudomonas aeruginosa		HUS, human	75.3	19.1	23.5
Pseudomonas aeruginosa		HUS, human	64.4	6.8	83.1
Serratia marcescens		HUS, human	59.7	33.6	52.7
Salmonella enterica serovar Typhi		HUS, human	35.7	10.6	24.1
Proteus vulgaris		HUS, human	31.2	13.7	19.4
Shigella flexneri		HUS, human	66.3	12.3	53.2
Shigella boydii		HUS, human	23.1	11.7	33.5
Staphylococcus aureus		HUS, human	40.3	15.8	35.1
Streptococcus faecalis		HUS, human	49.9	10.9	44.7
Yersinia enterocolitica		HUS, human	94.5	11.6	27.0

^aATCC, American Type Culture Collection, Manassas, Va.; HUS, Hospital Universitario de Salamanca, Salamanca, Spain. 

^bResults are expressed as mean fluorescence values of triplicate determinations, with the SD always being less than 12% of the mean values. 

TABLE 2.

Fluorescence intensities obtained in whole-cell hybridization with reference strains and clinical isolates of Brucella spp.

Bacterium	Strain	Source and origina	Mean fluorescence intensity with probeb:
			B6	B9	B13
Brucella abortus bv. 1	2308	G. Schurig	377.7	394.1	205.4
Brucella abortus bv. 1	544	INRA	215.8	482.9	268.1
Brucella abortus bv. 1	S99	NCTC	381.2	232.9	277.5
Brucella abortus bv. 1	S19	INRA	469.2	456.6	222.9
Brucella abortus bv. 1	RB51	G. Schurig	765.7	1,362.6	649.6
Brucella abortus bv. 2	86/8/59	INRA	254.6	356.6	212.1
Brucella abortus bv. 3	Tulya	INRA	380.3	328.6	273.1
Brucella abortus bv. 4	282	INRA	213.4	252.5	179.2
Brucella abortus bv. 5	B3196	INRA	313.8	328.6	262.2
Brucella abortus bv. 6	870	INRA	490.1	523.3	519.8
Brucella abortus bv. 9	C68	INRA	342.9	457.2	298.4
Brucella melitensis bv. 1	16M	NCTC	181.8	110.8	147.3
Brucella melitensis bv. 2	63/9	G. Schurig	270.3	286.6	196.1
Brucella melitensis bv. 3	Ether	INRA	166.7	172.8	122.7
Brucella suis bv. 1	1330	INRA	180.5	283.5	260.6
Brucella suis bv. 2	Thomsen	INRA	1,910.7	2,771.4	1,673.3
Brucella suis bv. 3	686	INRA	551.5	1,601.6	450.1
Brucella suis bv. 4	40	INRA	653.6	1,428.3	445.6
Brucella suis bv. 5	513	INRA	941	2,067.9	806.9
Brucella ovis	BOW	INRA	955.1	594.8	1,054.4
Brucella melitensis		HUS, human	255.6	188.3	248.4
Brucella melitensis		HUS, human	184.1	309.1	116.3
Brucella melitensis		HUS, human	167.5	272.6	122.7
Brucella melitensis		HUS, human	211.8	300.8	166.3
Brucella melitensis		HUS, human	206.1	256.8	156.9
Brucella melitensis		HUS, human	245.2	212.4	176.7
Brucella melitensis		HUS, human	187.3	234.7	145.6
Brucella melitensis		HUS, human	216.9	218.7	167.3
Brucella melitensis		HUS, human	209.7	198.4	201.7

^aINRA, Institut National de la Recherche Agronomique, Tours, France; NCTC, National Collection of Type Cultures, London, England; HUS, Hospital Universitario de Salamanca, Salamanca, Spain. 

^bResults are expressed as mean fluorescence values of triplicate determinations, with the SD always being less than 12% of the mean values. 

TABLE 3.

Fluorescence intensities obtained in whole-cell hybridization with bacteria phylogenetically or serologically related to Brucella spp.

Bacterium	Strain	Sourcea	Mean fluorescence intensity with probed:
			B6	B9	B13
Agrobacterium rhizogenesb	150T	LMG	2,537.4	186.9	2,079.3
Agrobacterium tumefaciensb	196	LMG	1,694.5	1,193.7	1,257.5
Alcaligenes denitrificansb	467	CECT	220.6	44.2	347.1
Alcaligenes faecalisb	928	CECT	92.1	38.3	87.1
Rhizobium leguminosarumb	1004	ATCC	1,184.1	1,572.8	1,753.8
Sinorhizobium melilotib	9930	ATCC	1,409.1	603.3	1,284.5
Ochrobactrum anthropi biotype Ab	3331	LMG	427.7	551.4	477.7
Ochrobactrum intermediumb	3301	LMG	416.2	446.5	447.7
Ochrobactrum anthropib	3309	LMG	1,357.2	1,295.7	1,265.5
Ochrobactrum anthropib	401	LMG	1,468.7	1,145.9	1,334.3
Ochrobactrum anthropib	33	LMG	1,844.7	1,983.9	1,769.4
Mycoplana dimorphab	4061(t1)	LMG	727.2	674.3	670.1
Phyllobacterium rubiacearumb	1(t1)	LMG	464.8	322.9	666.8
Phyllobacterium myrsinacearumb	4451(t2)	CECT	1,669.9	110.6	1,088.8
Yersinia enterocolitica O:9c	754	CECT	47.1	18.8	50.6
Salmonella enterica serovar Urbana O:30c	4151	CECT	50.2	11.7	40.5
Escherichia coli O:157c	4267	CECT	14.6	13.6	12.9
Xanthomonas maltophiliac	113	CECT	37.7	16.8	21.5
Vibrio cholerae O:1c	552	CECT	53.3	13.9	50.5

^aLMG, Culture Collection of the Laboratory for Microbiology, Ghent, Belgium; CECT, Colección Española de Cultivos Tipo, Valencia, Spain; ATCC, American Type Culture Collection, Manassas, Va. 

^bPhylogenetically related to Brucella spp. 

^cSerologically related to Brucella spp. 

^dResults are expressed as mean fluorescence values of triplicate determinations, with the SD always being less than 12% of the mean values. 

Probe design.

Multiple sequence alignments of the 16S rRNA sequences of B. abortus (EMBL accession number X13695), Rochalimaea quintana (M11927), Ochrobactrum anthropi (D12794), Phyllobacterium rubiacearum (D12790), Mycoplana dimorpha (D12786), Agrobacterium tumefaciens (D13294), Agrobacterium rubi (I67228), Rhodopseudomonas palustris (D25312), and Bradyrhizobium japonicum (D12781) were carried out, and on the basis of foreseeable specificity for Brucella spp., three probes were designed. The oligonucleotide probes selected were synthesized and 5′ labeled with fluorescein isothiocyanate by Isogen Bioscience BV, Maarsen, The Netherlands. All probes, together with their sequences and target positions, are listed in Table 4.

TABLE 4.

Sequences of all the probes used in this study

Probe	Nucleotide positionsa	Sequence
B6	73–90	5′-TCTGCCGCTCACCCTTGC-3′
B9	169–186	5′-GCCGATAAATCTTTCCCC-3′
B13	73–87	5′-GCCGCTCACCCTTGC-3′

^aAccording to the nucleotide sequence of the B. abortus 16S rRNA described by Dorsch et al. (13). Probes were chosen on the complementary strand. All probes are from the present study. 

Cell fixation and fluorescent whole-cell hybridization.

Bacterial cells grown according to their specific growth conditions were harvested in the logarithmic phase of growth (absorption at 650 nm [A₆₅₀] of 0.3; see “Results and discussion” below), washed twice with phosphate-buffered saline (PBS) (130 mM sodium chloride, 10 mM sodium phosphate buffer [pH 7.2]), and fixed with 4% paraformaldehyde as described previously (3, 4). Fixed cells were stored at −20°C for up to 8 weeks (3). For the fluorescent whole-cell hybridization, approximately 5 × 10⁵ fixed cells μl⁻¹ were hybridized in 100 μl of hybridization buffer containing 0.9 M sodium chloride, 20 mM Tris-HCl (pH 7.2), 0.1% sodium dodecyl sulfate, and 7 ng of fluorescent probe μl⁻¹ (see “Results and discussion” below) at 46°C for 2 h (39). Subsequently, cells were pelleted by centrifugation for 2 min at 8,000 × g and resuspended in 100 μl of hybridization buffer containing no probe. After a 20-min wash at 46°C, cells were mixed with 500 μl of PBS (pH 8.4), immediately placed on ice, and analyzed within 3 h.

Flow cytometry.

Analyses were performed with a FACScan flow cytometer (Becton Dickinson, San Jose, Calif.). The forward-angle light scatter, right-angle light scatter, and green fluorescence parameters were detected. Each measurement was made for 10,000 events. Data were collected in the list mode (individual measurements per cell were collected and stored). Subsequent analysis was done using Cell Quest software (Becton Dickinson). The fluorescence intensity conferred by each probe was determined as the mean of the green fluorescence values of single cells. In each measurement, the fluorescence of the cells was corrected by subtracting the background fluorescence of the cells hybridized with the same probe without fluorescence (negative control). The individual results obtained with each probe are expressed as mean fluorescence values of triplicate determinations, with the standard deviations (SD) always being less than 12% of the mean values.

Statistical methods.

Statistical analyses were performed with analysis of variance and Fisher's exact test. A P value of <0.05 was considered statistically significant.

Results and discussion.

Preliminary studies were carried out in order to optimize some parameters which might affect the results obtained in the whole-cell hybridization procedure (19). Thus, to select the optimum fluorescent probe concentration, hybridizations using 5 × 10⁵ fixed B. abortus 2308 and Escherichia coli ATCC 25922 (negative control) cells and concentrations of 1, 5, 10, 15, and 20 ng of each fluorescently labeled oligonucleotide probe μl of hybridization buffer⁻¹ were carried out. Similar fluorescence intensities of hybridized cells were observed with each probe at concentrations ranging between 5 and 10 ng μl⁻¹. At higher concentrations, an increase in fluorescence signals was observed in both B. abortus and E. coli cells (data not shown), probably due to nonspecific staining (39). Accordingly, all further hybridization experiments were performed using each probe at a concentration of 7 ng μl⁻¹ under the experimental conditions described above. Following this, attempts were made to determine the effect that the growth phase of cultured cells used in the hybridization assay had on the fluorescence signal obtained with each probe. The results are shown in Fig. 1. With the three probes used in the hybridization with B. abortus 2308 and E. coli 25922 (control) cells, the strongest fluorescence intensity signal was achieved with cells of both microorganisms fixed in the logarithmic phase of growth (A₆₅₀ ranging between 0.25 and 0.35). These results are in agreement with those described previously (12, 39), in which it was demonstrated that the signals conferred by fluorescently labeled rRNA-targeted probes on whole fixed cells are correlated with cellular rRNA contents and therefore reflect the cells' metabolic activities and their growth rate. Accordingly, the fixation of all the bacterial cells used in this study was performed on growing cells at an A₆₅₀ of 0.3.

FIG. 1.

Analysis by flow cytometry of the variation in the fluorescence intensity of hybridized B. abortus 2308 (——) and E. coli 25922 (------) cells, using probes B6 (□), B9 (○), and B13 (■), in relation to the growth phase of the cultured cells used for fixation. Each point represents the mean fluorescence of triplicate determinations, with the SD always being less than 10% of the mean values. a.u., arbitrary units.

To determine the specificity of the fluorescent probes selected, a total of 17 clinical isolates and reference strains of gram-positive and gram-negative microorganisms not belonging to the genus Brucella and unrelated either phylogenetically or serologically to this genus were assayed. The results from the flow cytometric quantification of the fluorescence signals obtained with each probe (Table 1) revealed that the means of the fluorescence intensity signals were as follows: B6 probe, 40.7 (SD, 23.3); B9 probe, 12.8 (SD, 8.2); and B13 probe, 37.5 (SD, 20.1). On the basis of these results, we established as positive, for each probe, mean fluorescence intensity obtained plus three times the SD obtained. Thus, throughout the subsequent part of the study, fluorescence intensity values equal to or greater than 110.6, 34.7, and 97.8 for the B6, B9, and B13 probes, respectively, were considered positive. Following this determination, a total of 22 Brucella strains representative of the different species and biovars and 9 clinical isolates previously identified as B. melitensis were analyzed. The results obtained in the flow cytometric quantification of the fluorescence intensities conferred by each probe (Table 2) revealed a positive signal, with the three probes, for all the strains of Brucella assayed. These results point to the high sensitivity of the whole-cell hybridization assay in the identification of microorganisms belonging to the genus Brucella and are in agreement with those of Verger et al. (36, 37), who demonstrated a high degree of homology between the microorganisms included in this genus at the DNA level. This has led to a proposal to classify the genus into a single species containing several serovars (36). Nevertheless, the statistical study carried out on the different fluorescence intensities detected by each molecular probe revealed that only with the B9 fluorescent oligonucleotide can statistically significant differences (P < 0.05) be established between B. suis biovars 2, 3, 4, and 5 and almost all the other Brucella spp. studied (Table 2). Therefore, the use of this probe in the hybridization assay can be useful for the identification of the microorganisms belonging to this species.

Finally, the specificity of the fluorescent probes used in the whole-cell hybridization assay was also evaluated to analyze a total of 14 different strains of bacteria phylogenetically related to Brucella spp. and 5 bacterial strains characterized by showing serological cross-reactions with Brucella spp. (Table 3). No positive fluorescence signals were observed with the five bacterial strains serologically related to Brucella spp. In contrast, all the microorganisms phylogenetically related to Brucella showed a positive fluorescence intensity signal. This is in agreement with the results of other authors (11, 29, 35, 40) who, on the basis of 16S rRNA sequence comparisons, reported that the genus Brucella would be a member of the alpha-2 subdivision of the class Proteobacteria, belonging to a tight phylogenetic group with members of the genera Mycoplana, Ochrobactrum, Rhodobacter, and Bartonella (Rochalimaea), which also show a close relationship with the genera Phyllobacterium, Rhizobium, and Agrobacterium from the Rhizobiaceae family. Despite this, although some of these microorganisms, such as Agrobacterium spp., Alcaligenes spp., O. anthropi, and O. intermedium, are currently recognized as emerging opportunistic human pathogens often isolated from immunocompromised patients (1, 7, 14–17, 20, 22, 23, 25, 27, 28), the incidence of such infections is for the time being still low (1), and hence it is unlikely that they could give a false-positive result in a test for Brucella spp. such as the whole-cell hybridization assay described here.