Development of a DNA Microarray Method for Detection and Identification of All 15 Distinct O-Antigen Forms of Legionella pneumophila

Abstract

Legionella is ubiquitous in many environments. At least 50 species and 70 serogroups of the Gram-negative bacterium have been identified. Of the 50 species, 20 are pathogenic, and Legionella pneumophila is responsible for the great majority (approximately 90%) of the Legionnaires' disease cases that occur. Furthermore, of the 15 L. pneumophila serogroups identified, O1 alone causes more than 84% of the Legionnaires' disease cases that occur worldwide. Rapid and reliable assays for the detection and identification of L. pneumophila in water, environmental, and clinical samples are in great demand. L. pneumophila bacteria are traditionally identified by their O antigens by immunological methods. We have recently developed an O serogroup-specific DNA microarray for the detection of all 15 distinct O-antigen forms of L. pneumophila, including serogroups O1 to O15. A total of 35 strains were used to verify the specificity of the microarray, including 15 L. pneumophila O-antigen standard reference strains and seven L. pneumophila clinical isolates as target strains, seven reference strains of other non-pneumophila Legionella species as closely related strains, and six non-Legionella bacterial species as nonrelated strains. The detection sensitivity was 1 ng of genomic DNA or 0.4 CFU/ml in water samples with filter enrichment and plate culturing. This study demonstrated that the microarray allows specific, sensitive, and reproducible detection of L. pneumophila serogroups. To the best of our knowledge, this is the first report of a microarray serotyping method for all 15 distinct O-antigen forms of L. pneumophila.

INTRODUCTION

In 1976, an outbreak of severe pneumonia among the participants of an American Legion convention in Philadelphia, PA, led to the recognition of Legionnaires' disease (1). The disease was later found to be caused by the Gram-negative bacterium Legionella pneumophila (Legionella for the legionnaires who were infected at the convention and pneumophila meaning “lung loving”), which belongs to the family Legionellaceae.

Legionella is widespread in freshwater habitats, and its ubiquity is probably due to the bacterium's capacity to survive under a wide range of environmental conditions (2). Parthuisot et al. studied the diversity and dynamics of Legionella species under different thermal and wastewater discharge conditions during an annual cycle, and the results suggest that Legionella spp. may be present at significant concentrations in many more freshwater environments than previously thought, highlighting the need for further ecological studies (3).

At least 50 Legionella species are known, and 70 serogroups have been identified, most notably, in L. pneumophila (4). The great majority (approximately 90%) of Legionnaires' disease cases that occur are caused by L. pneumophila, and among the at least 15 L. pneumophila serogroups identified, O1 alone is responsible for more than 84% of the Legionnaires' disease cases that occur worldwide (3, 5, 6). A study investigating 259 clinical and 3,128 environmental strains isolated in France in 2001 and 2002 showed that L. pneumophila serogroup O1 accounted for 28.2% of the environmental Legionella isolates, in contrast to 95.4% of the clinical isolates (7). Furthermore, the prospective surveillance of the extent of Legionella pollution conducted at three hot spring recreational areas in Beijing, China, in 2011 indicated that 51.9% of the spring water samples were Legionella positive, and their concentrations ranged from 1 to 2,218 CFU/liter. Again, L. pneumophila was the most frequently isolated species (98.9%), and serogroups 3 (25.3%), 6 (23.4%), 5 (19.2%), 1 (18.5%), 2 (10.2%), 8 (0.4%), 10 (0.8%), 9 (1.9%), and 12 (0.4%) were identified (8).

Of the methods currently used for the routine diagnosis of Legionella infection, in vitro culture is the most specific but its sensitivity is variable and unreliable because of the presence of a low number of viable organisms in most specimens (9). The detection of Legionella antigen in urine by enzyme immunoassays is very specific, and commercially available systems are able to detect L. pneumophila serogroup O1 but fail to detect other serogroups (10). Serological typing methods with monoclonal and multiclonal antibodies detect L. pneumophila only with the aid of intense preculturing (7, 8). Therefore, there is a need to establish a rapid and specific method that is at the same time more sensitive than the methods currently used. Amplification of Legionella-specific DNA sequences has been used to detect the bacterium in environmental water by using multilocus sequence analysis (11), the DNA gyrase subunit B gene (gyrB) (12, 13), the macrophage infectivity potentiator gene (mip) (14), and 5S rRNA (15). These technologies can typically provide much faster results but cannot distinguish the pathogen at the serogroup level.

The O antigen, which consists of many oligosaccharide unit (O unit) repeats, is part of the lipopolysaccharide (LPS) in the outer membrane of Gram-negative bacteria and contributes major antigenic variability to the cell surface (16). L. pneumophila uses the ATP-binding cassette transporter-dependent process for the synthesis and translocation of O antigens (17, 18, 19, 20, 21). In L. pneumophila, genes for O-antigen synthesis are normally located in a cluster that maps to the region between hisH and wecA on the chromosome (18).

Recently, PCR or PCR-based DNA microarray analysis of O-antigen-specific genes for several different serogroups of Escherichia coli and Shigella spp. have been developed (22, 23, 24, 25, 26). DNA microarrays detect thousands of specific DNA sequences simultaneously. In this study, a PCR-based DNA microarray based on the wzx, wzy, and wecA genes was established to detect all 15 distinct O-antigen forms (O1 to O15) of L. pneumophila. The microarray method described in this communication is specific, sensitive, and reliable and can be used as a better alternative to the traditional serotyping procedure, which is laborious and frequently cross-reactive.

MATERIALS AND METHODS

Bacterial strains.

The L. pneumophila strains used for whole-genome sequencing by the Solexa technology (27) are standard strains NCTC 11230, NCTC 11232, DSM 7514, ATCC 33216, NCTC 11406, ATCC 33823, NCTC 11985, NCTC 11986, NCTC 12000, NCTC 12179, NCTC 12180, NCTC 12181, NCTC 12174, and ATCC 35251 of serogroups O2 to O15, respectively. The bacterial strains used for microarray assays in this study are listed in Table 1. Fifteen L. pneumophila O standard reference strains, seven reference strains of other non-pneumophila Legionella species, six non-Legionella bacterial species, and seven L. pneumophila clinical isolates were used in this study (Table 1).

Table 1.

Bacterial strains used in this study

Application and bacterium	Serogroup	No. of strains from each source	Total no.
Targets for probe specificity testing
Legionella pneumophila	O1	1,a 1c	2
Legionella pneumophila	O2	1b	1
Legionella pneumophila	O3	1b	1
Legionella pneumophila	O4	1d	1
Legionella pneumophila	O5	1e	1
Legionella pneumophila	O6	1b	1
Legionella pneumophila	O7	1,e 1c	2
Legionella pneumophila	O8	1b	1
Legionella pneumophila	O9	1b	1
Legionella pneumophila	O10	1b	1
Legionella pneumophila	O11	1b	1
Legionella pneumophila	O12	1b	1
Legionella pneumophila	O13	1b	1
Legionella pneumophila	O14	1b	1
Legionella pneumophila	O15	1e	1
Other species used for probe specificity testing
Aeromonas hydrophila		1h	1
Legionella anisa		1d	1
Legionella bozemanii		1e	1
Legionella gormanii		1e	1
Legionella longbeachae		1d	1
Legionella micdadei		1b	1
Legionella steigerwaltii		1b	1
Legionella waltersii		1b	1
Pseudomonas aeruginosa		1g	1
Salmonella		1f	1
Shigella dysenteriae		1g	1
Staphylococcus aureus		1g	1
Yersinia enterocolitica		1d	1
Blind testing (n = 17)
Legionella pneumophila	O1	1a	1
Legionella pneumophila	O5	1e	1
Legionella pneumophila	O6	1b	1
Legionella pneumophila	O13	1b	1
Aeromonas hydrophila		1h	1
Legionella anisa		1d	1
Legionella bozemanii		1e	1
Legionella gormanii		1e	1
Legionella longbeachae		1d	1
Legionella micdadei		1b	1
Legionella steigerwaltii		1b	1
Legionella waltersii		1b	1
Pseudomonas aeruginosa		1g	1
Salmonella		1f	1
Shigella dysenteriae		1g	1
Staphylococcus aureus		1g	1
Yersinia enterocolitica		1d	1

^aCzech Collection of Microorganisms, Masaryk University, Brno, Czech Republic.

^bNational Collection of Type Cultures, Central Public Health Laboratory, London, United Kingdom.

^cCenter for Disease Control and Prevention of Nanshan District, Shenzhen, China.

^dGerman Collection of Microorganisms and Cell Cultures, Braunschweig, Germany.

^eAmerican Type Culture Collection, Manassas, VA.

^fSchool of Molecular and Microbial Biosciences, University of Sydney, Sydney, Australia.

^gNational Center for Medical Culture Collections, Beijing, China.

^hAcademy of Military Medical Sciences, Beijing, China.

Genomic DNA extraction.

All of the Legionella strains were cultured on buffered charcoal yeast extract (BCYE) agar (Hope Biotechnology Co., Ltd., Qingdao, China) in a 5% CO₂ incubator at 37°C for 48 h. The other strains were cultivated in Luria-Bertani medium (Oxoid Ltd., Basingstoke, Hampshire, England) at 37°C overnight with shaking. Genomic DNA was prepared with a DNA extraction kit (Tiangen, Beijing, China).

Target genes and oligonucleotide primer design.

Sequence alignment and comparison were performed with the ClustalW program. Phylogenetic trees were constructed by the neighbor-joining method and plotted by the Molecular Evolutionary Genetics Analysis (MEGA) 3.1 software package (http://www.megasoftware.net). Bootstrap analysis was carried out on the basis of 1,000 replicates. The target genes used to design primers and probes were wzm for L. pneumophila serogroups O4, O9, O10, O12, O13, and O15; wzt for L. pneumophila serogroups O1, O2, O3, O5, O6, O7, O8, and O14; and wecA for L. pneumophila serogroup O11. All of the primer pairs were designed with the Primer Premier 5.0 software (Premier Boost International). The primer sequences and concentrations used for the multiplex PCR are listed in Table 2.

Table 2.

Primers used in multiplex PCR assays and their concentrations

Primer	Target gene	Tm (°C)	Orientationa	Sequence (5′–3′)	Size (bp)	GenBank accession no.	Primer concn (μM) for:
							Multiplex PCR	Labeling
Group A
wl-52601	O1/O7 wzt	56.3	F	TGGGRTCAATGATAGATACTCC	817	Lab data	0.20
wl-52603		56.3	R	CCAATSCAACTGWGGTCTWATC			0.20	0.20
wl-41032	O2/O3 wzt	49.4	F	TTCAGAAATCCTCTGGAAG	908	Lab data	0.60
wl-41033		49.9	R	ATTTGCTTGGAGAACCTTA			0.60	0.60
wl-41037	O4 wzm	51.1	F	CAACTCCGGATTGGTAAA	243	Lab data	0.12
wl-41038		54.9	R	TTCAAATCGCGGTACCTG			0.12	0.12
wl-41388	O5 wzt	49.4	F	ATAATAAAGCAAGCCTTGAT	293	Lab data	0.12
wl-41389		47.7	R	TTCTGGATGAAAACCAGTC			0.12	0.12
wl-22688	O6 wzt	53.7	F	TAAAGATATTGTAGAGAGCCAGC	517	Lab data	0.12
wl-22689		52.7	R	CATAGAGAGATAACCCTCACATT			0.12	0.12
wl-52835	O9 wzm	47.1	F	ACAGATGGTTTGCCTTAC	420	Lab data	0.40
wl-41040		52.2	R	TTCATACCAAAACCGCAG			0.40	0.40
Group B
wl-47909	O8/O14 wzt	45.5	F	CAATACGAGATTAAAAGAAA	343	Lab data	0.20
wl-47910		40.3	R	CTTTGGTCTTAATAAGCCATC			0.20	0.20
wl-41046	O10 wzm	46.7	F	ACACTTTTAGGCTTTGGT	645	Lab data	0.40
wl-41047		49.1	R	CCCAAGCATAAAAACAATA			0.40	0.40
wl-50709	O11 wecA	55.0	F	TTGAATTCATTATTTCTTTTCG	228	Lab data	0.40
wl-50711		53.7	R	ATGAATAATAAACTAATTAACTGA			0.40	0.40
wl-41050	O12/O15 wzm	50.2	F	GGGGATATTCAACCGTTA	751	Lab data	0.40
wl-41051		45.1	R	TAAATCCTATTACAAATATAGC			0.40	0.40
wl-41044	O13 wzm	50.0	F	TTGTCATTTGTGCCACAG	297	Lab data	0.40
wl-41045		47.3	R	GCCAATTACCCTTTAAAC			0.40	0.40
Positive controls
Wl-3110	16S rRNA		F	TGTACACACCGCCCGTC	500–1,000	AB553285.1	0.08
Wl-3111			R	GGTACTTAGATGTTTCAGTTC		AY987650.1	0.08	0.13

^aF, forward primer; R, reverse primer.

Multiplex PCR and labeling of target genes.

The multiplex PCR was carried out with two groups; the first group consisted of L. pneumophila serogroups O1, O2, O3, O4, O5, O6, O7, and O9, and the second group consisted of L. pneumophila serogroups O8, O10, O11, O12, O13, O14, and O15. Amplification was performed with 50 μl of a reaction mixture consisting of 100 ng of DNA, 1× PCR buffer (50 mM KCl, 2.5 mM MgCl₂, 10 mM Tris-HCl [pH 8.3]); 100 μM deoxynucleoside triphosphates, 2.5 U of Taq DNA polymerase (TaKaRa Biotechnology [Dalian] Co. Ltd.), and each primer at the concentration shown in Table 2. The reaction parameters were 94°C for 5 min; 35 cycles of 94°C for 30 s, 57°C for 1 min, and 72°C for 1 min; and a final extension at 72°C for 5 min. An aliquot of the PCR product (2 μl) was subjected to agarose gel electrophoresis to examine the amplified DNA (Fig. 1). To label PCR products, 10 μM Cy3-dUTP (Amersham Biosciences UK Ltd., Little Chalfont, United Kingdom) and each reverse primer at the concentration shown in Table 2 were included in a PCR mixture. Twelve microliters of the amplification products generated from the above multiplex PCR was added as the template to 30 μl of the PCR mixture. The thermal cycling conditions were the same as for the multiplex PCR. All labeled DNA was stored at −20°C in the dark.

Fig 1.

Agarose gel electrophoresis of multiplex PCR products. (A) Lanes: 1, molecular size standards (DL2000 marker); 2, L. pneumophila serogroup O1; 3, L. pneumophila serogroup O2; 4, L. pneumophila serogroup O3; 5, L. pneumophila serogroup O4; 6, L. pneumophila serogroup O5; 7, L. pneumophila serogroup O6; 8, L. pneumophila serogroup O7; 9, L. pneumophila serogroup O9. (B) Lanes: 1, molecular size standards (DL2000 marker); 2, L. pneumophila serogroup O8; 3, L. pneumophila serogroup O10; 4, L. pneumophila serogroup O11; 5, L. pneumophila serogroup O12; 6, L. pneumophila serogroup O13; 7, L. pneumophila serogroup O14; 8, L. pneumophila serogroup O15.

Oligonucleotide probe design.

For each type of pathogen, one to four probes were designed by OligoArray 2.0 on the basis of the sequences in GenBank. One probe based on the 16S rRNA gene was designed as the positive control. A probe containing 40 poly(T) oligonucleotides was used as the negative control. A probe containing 40 poly(T) oligonucleotides and labeled with 3′-Cy3 was used as the positional reference and printing control. Each probe comprised a modified 5′ amino acid sequence followed by a spacer of 10 to 15 poly(T)s and a stretch of specific sequence (synthesized by AuGCT Biotechnology Corporation, Beijing, China). All of the oligonucleotide probes used are listed in Table 3.

Table 3.

Oligonucleotide probes used in this study

Probe name	Target gene or virulence factor	Tm (°C)a	Sequence (5′–3′)	GenBank accession no.
OA-3383	O1 wzt	67.8	CAGTCATGGATATCACTCGCGACTATCTC	CP000675.2
OA-3386		67.6	GCGAAATATAAATCGGAACAGGTTTGG
OA-3761	O2 wzt	63.9	GTTAGCAGTTGGAGATCAGGATTTTCA	Lab data
OA-3762		61.2	GCCATGATATGAGTGCTATTGAATCGATTTG
OA-3762	O3 wzt	61.2	GCCATGATATGAGTGCTATTGAATCGATTTG	Lab data
OA-3758		66.1	TATTAGCAGTAGGAGATCAAGATTTTCAAA
OA-3702	O4 wzm	65.8	CAACTCCGGATTGGTAAATAAAATTTATTTT	Lab data
OA-3703		71.4	CGGTTTAATTATAATTTGCGCCACCATTTATG
OA-3377	O5 wzt	69.9	CAAGCCTTGATGATGAGAATACTCAAAAATCC	Lab data
OA-3705		67.0	GAAACAGAAGAGTTCCAGAGTTTGTGTAAGG
OA-3706		74.8	CCAAATATCACCTGGCGAACGCATAGGTTTAA
OA-3707		66.7	ATCCAGAATAACTACTCCGACTTATGGTGAA
OA-3373	O6 wzt	67.0	TAGATTCAATTGCAAGATCTCATGCCC	Lab data
OA-3374		63.8	ACAAGGTTCTTATTCATTCTCTTAAGCTTTT
OA-3770	O7 wzt	66.0	CTGTGTAGAGCTTAATTGTGTGAGATTATTGG	Lab data
OA-3792		66.0	AAATGCAAATTCAGTCAGTAAAAGTTCAGTG
OA-3793		69.0	CCATAGTCCAATGCGTTTTTTAAACTGTGTAG
OA-3680	O8 wzt	64.2	TACAGATTTAGAAAATAATGGACCAGCAAT	Lab data
OA-3776		60.4	TTCACTTGTCATTACAGATTTAGAAAATAAT
OA-3359	O9 wzm	67.8	CCTACAGATCAGTAGAGGGTAACGCCG	Lab data
OA-3360		65.7	TCCCTACAGATCAGTAGAGGGTAACGC
OA-3799		68.2	GCTTGGTGGGACATATACCCTATCTGCT
OA-3800		61.6	CCTTTGTGCCATTTTATATCATAAACTT
OA-3801		64.4	TTTTCAGTTATTCGTGCTATTTCCTTTTT
OA-3802		76.0	GTTTCAGCTTCACCACATTGGCTGCGGTTT
OA-3366	O10 wzm	64.8	CCTTGTTGGATTTGGTTTAATAATAATTTGT	Lab data
OA-3367		64.3	TTTAGGATTAATGTTGGCTCCTTTTAGTACT
OA-3771	O11 wecA	64.1	TATTATTTATTTTCTTACACCGCATTATTGG	Lab data
OA-3772		59.0	TCATCTCATATTGATTCAGTTAATTAGTTT
OA-3773		59.4	CATATTGATTCAGTTAATTAGTTTATTATTCAT
OA-3774		67.3	GATTATAAATCACGAAAAATGGTTTAGCCCT
OA-3775		64.1	TTGAATTCATTATTTCTTTTCGATTCTACAG
OA-3369	O12 wzm	63.4	CAGTTATAGAGATGCACTAAATGGTCGAT	Lab data
OA-3370		65.5	TTCACCTCATTGGTTAAGATTCTGGTATG
OA-3361	O13 wzm	62.2	TGCCACAGTTTATGATAAACTTAATTATCA	Lab data
OA-3362		60.0	CCTACTATTTGTTATTCAAGCAATGTTT
OA-3363		63.1	TTTGTACCACTGGCATTACAATTTTATTAT
OA-3680	O14 wzt	64.2	TACAGATTTAGAAAATAATGGACCAGCAAT	Lab data
OA-3776		60.4	TTCACTTGTCATTACAGATTTAGAAAATAAT
OA-3369	O15 wzm	63.4	CAGTTATAGAGATGCACTAAATGGTCGAT	Lab data
OA-3370		65.5	TTCACCTCATTGGTTAAGATTCTGGTATG
OA-1993	16S rRNA gene	71.9	1380-TTGTACACACCGCCCGTCACACCAT-1404b	X80725
WL-4006			TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTc
Cy3			TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT-Cy3d

^aT_m was predicted using Primer Premier 5.0 software.

^bThe 16S rRNA gene-based probe was used as the positive control.

^cThe probe containing 40 poly(T) oligonucleotides was used as the negative control.

^dThe probe labeled with 3′-Cy3 was used as the positional reference and printing control.

DNA array preparation and hybridization.

The probes were dissolved in 50% dimethyl sulfoxide (DMSO) at a final concentration of 1 μg/μl and printed onto aldehyde group-modified glass slides (CapitalBio Corporation, Beijing, China) with SpotArray 72 (Perkin-Elmer Corporation). Each probe was spotted in triplicate to eliminate irregular data arising from physical defects in the glass slides. Printed slides were dried and stored at room temperature in the dark. Before use, the slides were scanned at 532 nm for spotting quality control. Each glass slide consisted of eight individual arrays framed by a 20-μl Geneframe (CapitalBio Corporation, Beijing, China), which consisted of individual reaction chambers. A schematic diagram of the probe positions on the microarray is shown in Fig. 2. Hybridization was performed according to the following procedure. All 30 μl of the labeled PCR product was heated for approximately 2 h at 65°C until dry and diluted in 20 μl of hybridization buffer (30% formamide, 0.5% SDS, 6× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 5× Denhardt's solution). The mixture was then applied to a hybridization chamber and incubated at 43°C for 12 h in a water bath. Subsequently, the slide was washed sequentially in solution A (1× SSC, 0.2% SDS) for 3 min, in solution B (0.2× SSC) for 3 min, and in solution C (95% alcohol) for 1.5 min. The slide was dried under a gentle air stream before it was scanned.

Fig 2.

Probe positions on the slide. OA-1993 is the positive-control probe based on the 16S rRNA gene. WL-4006 is the negative-control probe. Cy3 is the positional reference and printing control probe. The blank is 50% DMSO. The rest are the specific probes for the target strains.

Data acquisition and automated analysis.

The slide was scanned with 532-nm laser beams with a 4100A biochip scanner (Axon Corporation), a photomultiplier tube gain of 600, and a pixel size of 5 μm. The image files were saved as .tif files, and the signal intensity was saved as .gpr files. The signal-to-noise ratio was calculated for each spot with a Bactarray Analyzer 1.0 developed in house with the threshold set at 3.0. A detection result was recorded as positive only when all of the hybridization signals generated by the probes of the target genes were above the signal-to-noise ratio threshold.

Testing of mock samples.

BCYE medium was used for proliferation. Pure cultures of L. pneumophila serogroups O1, O10, and O12 were serially diluted to 10¹ to 10⁶ CFU/ml, and 1 ml of the diluent was mixed with 100 ml of fresh tap water from the laboratory and vacuum filtered with a 0.22-μm membrane. The membrane was treated with 500 μl of diluted HCl (pH 3.0) for 1 min, placed face down on BCYE agar plates, and incubated in a 5% CO₂ incubator at 37°C for 3 to 5 days. The genomic DNA was then extracted from the cells for microarray hybridization.

Testing of air conditioner condensed-water samples.

Eight hundred milliliters of filter-enriched air conditioner condensed-water sample was plated onto a GVPC agar plate (Hope Bio-Technology Co., Ltd., Qingdao, China), and after incubation in a 5% CO₂ incubator at 37°C for 48 h, the culture was collected and genomic DNA was extracted for use in the multiplex PCR.

Nucleotide sequence accession numbers.

The DNA sequences of the wzm and wzt genes of L. pneumophila serogroups O2, O3, O4, O5, O6, O7, O8, O9, O10, O12, O13, O14, and O15 have been deposited in GenBank under accession numbers KF536969 to KF536994.

RESULTS

Phylogenetic analysis of wzm and wzt genes.

The wzm and wzt gene sequences of 14 of the 15 L. pneumophila serogroups, with the exception of serogroup O11, were determined by Solexa sequencing technology. Sequencing failed to detect the wzm and wzt genes in serogroup O11. Two phylogenetic trees of 14 L. pneumophila serogroup strains based on the wzm and wzt gene sequences were constructed with E. coli IAI1 as the outgroup reference (Fig. 3A and B). In both cases, L. pneumophila serogroups O1 and O7 formed a small group and the other 12 serogroups clustered together as a large group. For the wzm tree, the large group was divided into two equal-size subgroups, with L. pneumophila serogroups O2, O3, and O9 grouping together with L. pneumophila serogroups O6, O12, and O15 in one subgroup and L. pneumophila serogroups O5 and O13 grouping together with L. pneumophila serogroups O10, O4, O8, and O14 in the other subgroup. In the wzt tree, L. pneumophila serogroups O6, O12, and O15 separated from the smaller subgroup of L. pneumophila serogroups O2, O3, and O9 and instead grouped together with L. pneumophila serogroups O8, O14, O4, O10, O5, and O13 to form a bigger subgroup.

Fig 3.

Unrooted phylogenetic trees constructed by the neighbor-joining method on the basis of the wzm and wzt genes. Bootstrap values were based on 1,000 replications, and only values greater than 50% are shown. (A) Unrooted wzm gene phylogenetic tree of 14 L. pneumophila serogroups constructed by the neighbor-joining method. (B) Unrooted wzt gene phylogenetic tree of 14 L. pneumophila serogroups constructed by the neighbor-joining method.

Optimization of PCRs.

A multiplex PCR was used to amplify and label the target genes in two groups. Group A targeted serogroups O1, O2, O3, O4, O5, O6, O7, and O9, and group B targeted serogroups O8, O10, O11, O12, O13, O14, and O15. The wzt genes of L. pneumophila serogroups O2 and O3 are very similar (97.5% identical); therefore, the same primer pair, wl-41032/wl-41033, was used to amplify the target wzt genes of both of these serogroups. Since the wzm and wzt genes of serogroups O8 and O14 are the same, primer pair wl-47909/wl-47910 was used to amplify wzm of L. pneumophila serogroups O8 and O14. Also, since the wzm and wzt genes of serogroups O12 and O15 are 100% identical, primer pair wl-41050/wl-41051 was used to amplify wzt of serogroups O12 and O15. The wecA gene was used to detect serogroup O11. The multiplex PCR was used to streamline the test and maintain the specificity of individual amplicons. Initially, all of the primers were used at 0.2 μM. However, several pathogens failed to generate the expected hybridization signals under the conditions used. Consequently, primer concentrations of 0.12 to 0.60 μM were tested to establish the optimal concentration (Table 2). The amplicons of the 11 groups of target pathogens were amplified under optimized conditions, and the PCR product lengths varied from 481 to 1,000 bp (Fig. 1A and B).

Probe specificity.

The DNA microarray was tested by using 35 strains, including 15 L. pneumophila O standard reference strains, seven reference strains of other non-pneumophila Legionella species, six non-Legionella bacterial species, and seven L. pneumophila clinical isolates (Table 1). A total of 38 probes were used in the microarray, including 35 probes for specific genes, 1 positive control, 1 negative control, and 1 positional reference and printing control (Table 3). The probe positions on the slide are shown in Fig. 2. All of the representative strains belonging to the 15 groups consistently hybridized to the corresponding probes with 100% specificity. In addition, in order to distinguish L. pneumophila serogroup O2 from serogroup O3, specific probe OA-3761 was designed for L. pneumophila serogroup O2, specific probe OA-3758 was designed for L. pneumophila serogroup O3, and probe OA-3762 was designed for both serogroups. The hybridization results are shown in Fig. 4, panels 1 to 16.

Fig 4.

Microarray differentiation of pathogens. Panels: 1, L. pneumophila serogroup O1; 2, L. pneumophila serogroup O2; 3, L. pneumophila serogroup O3; 4, L. pneumophila serogroup O4; 5, L. pneumophila serogroup O5; 6, L. pneumophila serogroup O6; 7, L. pneumophila serogroup O7; 8, L. pneumophila serogroup O8; 9, L. pneumophila serogroup O9; 10, L. pneumophila serogroup O10; 11, L. pneumophila serogroup O11; 12, L. pneumophila serogroup O12; 13, L. pneumophila serogroup O13; 14, L. pneumophila serogroup O14; 15, L. pneumophila serogroup O15; 16, non-pneumophila Legionella species; 17, L. pneumophila serogroups O1 and O6; 18, L. pneumophila serogroups O6 and O7; 19, L. pneumophila serogroups O1, O4, O10, and O13.

Sensitivity of detection with genomic DNA.

Serial 10-fold dilutions of 10, 1.0, 0.1, and 0.01 ng of genomic DNA of L. pneumophila serogroup O1 from the first PCR group and serogroup O12 from the second PCR group were used as the templates in a multiplex PCR to test its sensitivity. Positive signals were generated by using 1 ng of DNA for both groups A and B. Therefore, the sensitivity of detection with genomic DNA was set at 1 ng of DNA.

Simultaneous detection of multiple pathogens.

Genomic DNAs from two pathogens of L. pneumophila serogroups O1 and O6, L. pneumophila serogroups O6 and O7, and four pathogens of L. pneumophila serogroups O1, O4, O10, and O13 were mixed and used as templates to test the specificity of the microarray assay. Data demonstrated that the probes were able to hybridize with and detect multiple pathogens in the sample (Fig. 4, panels 17 to 19).

Blind test.

A blind test was performed in order to verify the reliability and specificity of the microarray method. A total of 17 environmental and clinical isolates (Table 1) were used for hybridization to the microarray without disclosure of their identities during testing. The test samples included L. pneumophila serogroup O1, L. pneumophila serogroup O5, L. pneumophila serogroup O6, L. pneumophila serogroup O13, L. anisa, L. bozemanii, L. gormanii, L. longbeachae, L. micdadei, L. steigerwaltii, L. waltersii, Aeromonas hydrophila, Pseudomonas aeruginosa, Salmonella, Shigella dysenteriae, Staphylococcus aureus, and Yersinia enterocolitica. The detection results are consistent with those of conventional methods.

Testing of mock samples.

L. pneumophila serogroups O1, O10, and O12 were detected at levels of 20 CFU/100 ml, 90 CFU/100 ml, and 11 CFU/100 ml, respectively. On average, as few as 40 CFU/100 ml or 0.4 CFU/ml of L. pneumophila can be detected by filter enrichment and culturing.

Testing of environmental isolates and confirmation by sequencing.

A total of five L. pneumophila isolates, G2759, G2761, G2762, G2763, and G2765, from air conditioner condensed-water samples were obtained from the Center for Disease Control and Prevention of Nanshan District, Shenzhen, China. The microarray profile identified them as belonging to L. pneumophila serogroups O1, O2, O3, O7, and O5, respectively. The identities of these samples were further validated by amplification and sequencing of the wzt gene of each sample (data not shown).

Testing of real water samples and confirmation by sequencing.

Seven batches of condensed-water samples from air conditioners collected, enriched, and provided by the Center for Disease Control and Prevention, Shanghai, China, were analyzed by the microarray method. The microarray profile showed that three of the seven samples were L. pneumophila serogroup O1 (Fig. 4, panel 1). The remaining four samples tested yielded positive-control probe signals only. Further validation was done by amplification and sequencing of the wzt gene of each sample. The overall results obtained by the microarray method produced 100% accuracy, indicating that the microarray method has the practical ability to detect and differentiate L. pneumophila serogroups.

DISCUSSION

In this report, we describe a microarray method for the determination of L. pneumophila O serogroups on the basis of the O serogroup-specific genes. This is the first report of the comprehensive detection and identification of all of the 15 serogroups of L. pneumophila. In comparison with other molecular-analysis-based detection methods, such as PCR and DNA sequencing, the advantages of the present method include (i) high throughput, (ii) high specificity, and (iii) greater efficiency. PCR is rapid and easy to process, but it cannot differentiate amplicons of similar sizes from different serogroups. DNA sequencing is very accurate, but it requires more expensive equipment, costs more, and takes a lot of time and hands-on work for library preparation and sequencing (28, 29).

The O antigen contains many repeats of an O unit and constitutes part of the LPS present in the outer membrane of Gram-negative bacteria (30). The O antigen exhibits variation in terms of the types of sugars, the arrangement of the sugars within the O unit, and different types of linkages within and between O units (26). This makes LPS one of the most variable cell constituents and provides a basis for the serotyping of many Gram-negative bacteria (31, 32, 33). So far, only L. pneumophila serogroup O1 LPS gene clusters have been determined. The L. pneumophila serogroup O1 LPS gene locus includes genes that are involved in LPS core oligosaccharide biosynthesis (rmlA to rmlD, rhamnosyltransferases, acetyltransferase), and LPS O-chain biosynthesis and translocation (neuC, neuB, neuA, wecA, wzt, and wzm) (18). In this study, wzt and wzm were used as the target genes for 14 of the 15 L. pneumophila serogroups, with the exception of serogroup O11, for which wecA was used as the target gene instead.

The resolving power of the microarray assay is comparable to that of traditional immunological methods; for example, this method is able to differentiate closely related serogroups O1 and O7 and closely related serogroups O2 and O3 by the use of serogroup-specific probes. The current version of the microarray failed to differentiate serogroup O8 from O14 and serogroup O12 from O15 because the target genes used are the same. Future addition of other target genes in the O-antigen gene cluster or in other regions of the genome should be able to differentiate these serogroups.

A two-step multiplex PCR was used for amplification and labeling. In the first step, the target genes were amplified with the forward and reverse primers, and in the second step, the single-stranded DNA was labeled with the reverse primers. The two-step PCR not only enhanced amplification efficiency but also generated single-stranded PCR products for hybridization. As little as 1 ng of DNA or 0.4 CFU/ml before filtering and culturing enrichment can be detected by the microarray. In the case of water samples obtained from the environment, the bacterial concentration, especially in bottled water, can be as low as <0.1 CFU/ml, which falls below the threshold of detection by immune- or molecular-analysis-based assays. To overcome this problem, a two-step filtering and culturing process was used here to enrich the target bacteria. In the first step, vacuum filtering was used to collect all of the bacteria in the samples, and in the second step, the filter membrane was treated with diluted HCl and plated on the selective culture medium BCYE. The acid treatment used is based on the acid resistance feature of Legionella spp. By this method, L. pneumophila could be detected in as little as 2 or 3 days, 9 to 10 days earlier than by the traditional detection and identification method (ISO 11731:1998).

The DNA microarray method described here is specific and sensitive and offers a promising tool for wide applications in basic microbiological research and epidemiological surveillance. The method described here will be further verified with various clinical samples of bronchoalveolar lavage fluid and sputum and other samples, as well as environmental water samples, in the next stage of our study. The microarray has many advantages over traditional bacterial culture and serotyping methods; (i) it allows high sample throughput, (ii) it is able to detect O rough strains raised as a consequence of mutations in the O-antigen gene cluster, and (iii) cross-reactions are less common.

In conclusion, this study presents a new multiplex PCR-based microarray assay for the comprehensive detection and identification of all 15 L. pneumophila O serogroups. This new method provides an accurate and reliable approach by which to differentiate L. pneumophila isolates at the serogroup level, contributes significantly to large-scale epidemiology studies, and could be used to monitor local, regional, and national trends in human legionellosis.