In Vivo Himar1 Transposon Mutagenesis of Burkholderia pseudomallei

Abstract

Burkholderia psedudomallei is the etiologic agent of melioidosis, and the bacterium is listed as a potential agent of bioterrorism because of its low infectious dose, multiple infectious routes, and intrinsic antibiotic resistance. To further accelerate research with this understudied bacterium, we developed a Himar1-based random mutagenesis system for B. pseudomallei (HimarBP). The transposons contain a Flp recombinase-excisable, approved kanamycin resistance selection marker and an R6K origin of replication for transposon rescue. In vivo mutagenesis of virulent B. pseudomallei strain 1026b was highly efficient, with up to 44% of cells transformed with the delivery plasmid harboring chromosomal HimarBP insertions. Southern analyses revealed single insertions with no evidence of delivery plasmid maintenance. Sequence analysis of rescued HimarBP insertions revealed random insertions on both chromosomes within open reading frames and intergenic regions and that the orientation of insertions was largely unbiased. Auxotrophic mutants were obtained at a frequency of 0.72%, and nutritional supplementation experiments supported the functional assignment of genes within the respective biosynthetic pathways. HimarBP insertions were stable in the absence of selection and could be readily transferred between naturally transformable strains. Experiments with B. thailandensis suggest that the newly developed HimarBP transposons can also be used for random mutagenesis of other Burkholderia spp., especially the closely related species B. mallei. Our results demonstrate that comprehensive transposon libraries of B. pseudomallei can be generated, providing additional tools for the study of the biology, pathogenesis, and antibiotic resistance of this pathogen.

Burkholderia pseudomallei is the etiologic agent of melioidosis, a disease that is endemic to tropical and subtropical regions of the world (6, 30). Research with this bacterium has significantly increased with its listing as a priority pathogen by the U.S. National Institutes of Health and as a select-agent pathogen by the Centers for Disease Control and Prevention and the United States Department of Agriculture. Despite the availability of complete annotated and draft genome sequences for several strains (reference 12 and several GenBank entries), efforts aimed at understanding the biology and pathogenesis of B. pseudomallei are still hampered by a lack of genetic tools and the strict regulations that govern their use in the United States. Although many genetic tools have previously been used to study the biology and virulence of B. pseudomallei (9, 10, 18, 25), most of them are not compliant with United States select-agent regulations because they involve the use of nonapproved antibiotic selection markers. We recently published select-agent-compliant tools for allele replacement and single-copy gene integration in B. pseudomallei which facilitate targeted gene mutations and complementation (7). What is still needed, however, is a select-agent-compliant method for the efficient creation of random, transposon-induced mutants. The availability of such a system would greatly facilitate low-throughput strategies such as the identification of virulence or antibiotic resistance factors, as well as high-throughput strategies such as the construction of ordered, genome-wide transposon mutant libraries. Although Tn5-based transposon mutagenesis systems were previously described for and successfully used with B. pseudomallei (9, 21), most of them use a tetracycline selection marker that cannot be used in the United States because it conflicts with the potential use of doxycycline to treat B. pseudomallei infections in human and veterinary medicine. A previously described Tn5-based plasposon system with a kanamycin resistance marker (8) has, to our knowledge, not yet been tested with B. pseudomallei. Furthermore, the resistance marker residing on previously constructed transposons cannot be excised once inserted into the chromosome. In this study, we evaluated the use of Himar1 mariner transposons for random mutagenesis of B. pseudomallei. Himar1 transposons have been used for random in vitro (1, 2, 20) and in vivo (3, 14, 22, 27, 31, 33) mutagenesis of numerous bacteria, including the select agent Francisella tularensis (16). mariner-based transposons do not require host-specific factors and, other than preference for a TA dinucleotide target, do not display target site specificity. We describe the development of an efficient in vivo Himar1 transposon mutagenesis system for B. pseudomallei and demonstrate its use for the isolation of auxotrophic and other mutants.

MATERIALS AND METHODS

Bacterial strains, media, and growth conditions.

Several Burkholderia strains were used in this study. Before working with B. pseudomallei, genetic constructs were routinely tested in B. thailandensis wild-type strain E264 (5). B. thailandensis is traditionally regarded as a naturally attenuated relative of B. pseudomallei (32) which can be handled at biosafety level 2 and is exempt from select-agent guidelines. B. pseudomallei 1026b was used as the wild-type strain (Table 1). The Escherichia coli strains used for routine cloning experiments were DH5α (15), DH5α(λpir) (laboratory strain), HPS1 (24), and S17-1 (26). All bacteria were routinely grown at 37°C. Strains containing temperature-sensitive plasmid derivatives were grown at 30°C (permissive temperature) or 37°C (nonpermissive temperature). Low-salt (5 g liter⁻¹ NaCl) Lennox LB broth and agar (MO BIO Laboratories, Carlsbad, CA) were used. M9 medium (17) with 10 mM glucose was used as the minimal medium. Nutritional supplements for auxotrophic mutants were added at the following concentrations: 20 μg ml⁻¹ l-phenylalanine, l-tyrosine, or l-tryptophan; 100 μg ml⁻¹ l-aspartic acid or l-glutamine; 40 μg ml⁻¹ uracil. Unless otherwise noted, antibiotics were added at the following concentrations: 100 μg ml⁻¹ ampicillin, 35 μg ml⁻¹ kanamycin, and 25 μg ml⁻¹ zeocin for E. coli; 1,000 μg ml⁻¹ kanamycin and 2,000 μg ml⁻¹ zeocin for wild-type B. pseudomallei; 200 μg ml⁻¹ zeocin and 500 μg ml⁻¹ kanamycin for B. thailandensis. Antibiotics were purchased from either Sigma, St. Louis, MO (ampicillin and kanamycin), or Invitrogen, Carlsbad, CA (zeocin).

TABLE 1.

Strains, plasmids, and primers used in this study

Strain, plasmid, or primer	Relevant property(ies)a	Reference or source
Strains
B. thailandensis E264	Wild-type strain	5
B. pseudomallei 1026b	Wild-type strain, clinical isolate	9
Plasmids
pFNLTP16 H1	Apr Kmr; source of tnp and nptII genes and Himar1 transposon with oriR6K	16
pFLPe2b	Zeor; source of Flp recombinase	7
pFKM2b	Apr Kmr; source of FRT-nptII-FRT cassette	7
pPS2163	Apr Kmr; source of ColE1 ori, oriT, and ori1600-rep(TsBt)	7
pPS2413	Apr Kmr; pCR2.1 (TA cloning kit; Invitrogen) with 1,456-bp PCR fragment amplified from pFKM2 with primers 596 and 1758	This study
pHBurk1	Kmr; ligation of fragment from pFNLTP16 H1 containing Himar1 (with oriR6K, npt gene, inverted repeats) and tnp gene and a pPS2163 fragment containing ColE1 ori, oriT, and ori1600-rep(TsBt)	This study
pHBurk-Link-2	Kmr; pHBurk1 with BglII-SmaI linker inserted at PvuI site upstream of tnp	This study
pHBurk2	Kmr; pHBurk-Link-2 with nptII gene replaced with FRT-nptII-FRT containing PCR fragment from pFKM2; nptII gene oriented away from tnp	This study
pHBurk3	Kmr; like pHBurk2 but with nptII facing toward tnp	This study
pHBurk4	Kmr; pHBurk2 with Placb downstream of nptII, which is oriented away from tnp	This study
pHBurk5	Kmr; like pHBurk4 but with nptII and Plac facing toward tnp	This study
pHBurk6	Kmr; pHBurk-Link-2 with B. thailandensis PS12 promoter obtained by annealing oligonucleotides 1690 and 1691 and inserting the double-stranded oligonucleotide fragment into PvuI-SmaI-digested pHBurk-Link-2	This study
Primers and other oligonucleotides
1668c	5′-CGCTGACATCGAGATCTCTAACCCGGGAT	This study
1669c	5′-CCCGGGTTAGAGATCTCGATGTCAGCGAT	This study
596	5′-CGAATTAGCTTCAAAAGCGCTCTGA	This study
1758d	5′-CACAACATACGAGCCGGAAGCATAAAGTGTAAAGCCGAATTGGGGATCTTGAAGTACCT	This study
511	5′-ATTAACCGCTTGTCAGCCGTTAAGTGTTCCT	This study
512	5′-ATTACCGCGGCAGTTCAACCTGTTGATAGTAC	This study
1398	5′-GTCAGCACGTTGATCGAGAA	This study
1399	5′-CGCTGTGATGTTCCTCTTCA	This study
1768	5′-AGGCTTTACCAGTAAGAAGGAG	This study
1769	5′-GATTTCGACCTTCAAACGCTCC	This study
1670e	5′-TCGGGTATCGCTCTTGAAGGG	This study
1722	5′-GACTTGGTTGAGTACTCACCAG	This study
1690f	5′-ATGTTGACTCGCTTGGGATTTTCGGAATATCATGCCGGT	This study
1691	5′-ACCCGGCATGATATTCCGAAAATCCCAAGCGAGTCAACATAT	This study
1829e	5′-GCATTTAATACTAGCGACGCC	This study
1832	5′-GTTCCCTTCAAGAGCGATACC	This study
1833	5′-AACGCACTGAGAAGCCCTTAG	This study

^aAbbreviations: Ap^r, ampicillin resistance; Km^r, kanamycin resistance; Zeo^r, zeocin resistance.

^bP_lac, E. coli lac operon promoter.

^cBglII and SmaI recognition sequences are underlined.

^dE. coli lac promoter −10 (AACATA) and −35 (TGTAAA) reverse complement sequences are in boldface.

^e1670 and 1829 are DNA sequencing primers P1 and P2, respectively, used for determination of transposon-chromosome junction sequences.

^fB. thailandensis P_S12 promoter predicted −10 (GAATATCAT) and −35 (TTGACT) sequences are in boldface.

DNA methods and transformation.

Routine procedures were employed for manipulation of DNA (23). Plasmid DNAs were isolated from E. coli and Burkholderia spp. with the Fermentas GenJet Plasmid Miniprep Kit (Fermentas, Glen Burnie, MD). Bacterial chromosomal DNA fragments (20 to 30 kb) were isolated with the QIAamp DNA Mini Kit, and the DNA was suspended in 200 μl of buffer AE (10 mM Tris-HCl, 0.5 M EDTA, pH 9). Plasmid DNA fragments were purified from agarose gels with the Fermentas DNA Extraction Kit (Fermentas, Glen Burnie, MD). E. coli strains were transformed by using chemically competent cells (23). Replicative plasmids were transformed into B. thailandensis and B. pseudomallei by a rapid electroporation procedure (7). Colony PCR with Burkholderia spp. was performed as previously described (7). Custom oligonucleotides were synthesized by Integrated DNA Technologies (Coralville, IA). DNA maps were constructed with Gene Construction Kit 2.5 (Textco, West Lebanon, NH) and exported to Microsoft Powerpoint for final annotation.

For determination of insertion sites, genomic Burkholderia DNA was initially extracted from selected clones with the QiAmpDNA Mini Kit (Qiagen, Valencia, CA) but later the Gentra Puregene DNA purification kit (Qiagen) was used because of superior yield, and 1 μg was digested overnight with NotI. DNA was purified with the Fermentas DNA Extraction Kit, treated with T4 DNA ligase (Invitrogen) overnight at 14°C, and transformed into DH5α(λpir), and Km^r transformants were selected. Plasmid DNA was prepared, and transposon-chromosomal junction sequences were determined by nucleotide sequencing with primers 1670 and 1829 (the primers and other oligonucleotides used are listed in Table 1).

HimarBP3 was transferred between chromosomes of transposon-containing strains and strain 1026b by DNA fragment transfer with naturally competent cells by previously described methods (7, 28).

Southern blot analysis.

For genomic Southern analysis, genomic DNA was isolated with the Centra Puregene DNA purification kit (Qiagen). DNA (4 μg) was digested with NotI overnight, electrophoresed on a 1% agarose gel, and transferred to positively charged nylon membranes (Roche Diagnostics Corp., Indianapolis, IN) by passive transfer as previously described (23). Following transfer and UV fixation, blots were probed with a PCR fragment biotinylated by random hexamer priming following the NEBlot Phototype labeling and detection kit protocols (New England BioLabs, Beverly, MA). The probe detected the HimarBP transposon with a 376-bp fragment recognizing the ori_R6K region.

Construction and transposition of Himar1 derivatives.

All of the Himar1 derivatives used, as well as other plasmids used for their construction, are listed in Table 1. pHBurk1 (Fig. 1), containing a temperature-sensitive Burkholderia sp. replicon, was derived by combining a blunt-ended 2,974-bp BpmI-NsiI fragment from pPS2163 with a blunt-ended 3,971-bp NotI fragment from pFNLTP16 H1 containing the Himar1 transposon and its transposase-encoding tnp gene. Next, pHBurk-Link-2 was constructed by ligating a linker composed of oligonucleotides 1668 and 1669 containing a BglII and a SmaI site into the single PvuI site located immediately upstream of the tnp gene of pHBurk1 such that a single PvuI site was recreated at the linker insertion site. The SmaI and PvuI sites were subsequently used to insert promoter-containing linkers. Plasmids pHBurk2 and pHBurk3 (Fig. 1) were derived from pHBurk-Link-2 by replacing a 1,206-bp blunt-ended MluI fragment containing the resident nptII gene with a FRT-nptII-FRT-containing 1,444-bp SmaI fragment from pFKM2. Plasmids pHBurk2 and pHBurk3 differ in the orientation of the nptII gene. Next, pHBurk4 and pHBurk5 (Fig. 1) were constructed by replacing the blunt-ended 1,206-bp MluI fragment of pHBurk-Link-2 containing the npt gene with a 1,476-bp blunt-ended EcoRI fragment from pPS2413 containing FRT-nptII-FRT-P_lac. Plasmids pHBurk4 and pHBurk5 differ in the orientation of the nptII gene and P_lac. Lastly, pHBurk6 was constructed by inserting a double-stranded oligonucleotide containing the B. thailandensis ribosomal S12 gene promoter (P_S12) between the PvuI and SmaI sites of pHBurk-Link-2 such that tnp transcription was promoted by P_S12.

FIG. 1.

Maps of two representative HimarBP-containing delivery plasmids. The plasmids contain the following shared features: IR, Himar1 inverted repeat; nptII, neomycin phosphotransferase-encoding gene; ori, E. coli ColE1 origin of replication; ori₁₆₀₀, pRO1600 origin of replication requiring the rep(Ts)-encoded replication protein which confers a temperature sensitivity phenotype on Burkholderia spp. at temperatures of 37°C and above; oriT, RK2-derived origin for conjugal plasmid transfer; ori_R6K, π protein-dependent R6K replication origin; tnp, transposase-encoding gene. Plasmid pHBurk3 additionally contains Flp recombinase targets (FRT) and two unique restriction sites (PvuI and SmaI) derived by insertion into the unique PvuI site of pHBurk1. pHBurk5 has the same features as pHBurk3 but contains the E. coli lac operon promoter (P_lac) for the transcription of genes adjacent to the promoter insertion site. Similarly, pHBurk6 is the same as pHBurk1 but tnp transcription is directed by the promoter for the B. thailandensis ribosomal S12 protein-encoding gene (P_S12). The transposons harbored by the individual plasmids are named after plasmid numbers, e.g., pHBurk1 harbors HimarBP1, pHBurk3 harbors HimarBP3, etc. The plasmids are not drawn to scale.

For Himar1 transposition, the previously described one-step protocol (16) was followed, with appropriate modifications. Briefly, 100 ng of each pHBurk plasmid was electroporated into freshly prepared electrocompetent B. thailandensis or B. pseudomallei cells. After incubation at 30°C in a shaker for 1 h, dilutions were plated on LB medium containing 1,000 μg ml⁻¹ kanamycin and incubated at 37°C to select for plasmid loss and chromosomal transposon integration. Dilutions were also plated on LB medium with and without kanamycin, and plates were incubated at 30°C to determine the total number of CFU transformed or the total number of viable cells, respectively. Colonies grown at 37°C for 24 to 96 h were picked and patched onto LB plates with kanamycin to recover individual clones containing Himar1 in the chromosome. For auxotrophy screening, colonies were also patched onto M9-glucose-kanamycin plates. Transposon presence in genomic DNA was assessed by PCR with primers 511 and 512 (specific for ori_R6K), and primers 1398 and 1399 (specific for the amrB efflux protein-encoding gene) were used to amplify a positive control fragment. Delivery plasmid loss was verified by PCR with primers 1768 and 1769, which are specific for the tnp gene located on the plasmid backbone.

Flp recombinase-mediated Km^r marker excision was performed with pFLPe2 by using a previously described protocol (7). The plasmid was cured by growing kanamycin-susceptible colonies at 37°C, a nonpermissive temperature for pFLPe2, resulting in markerless mutants.

Nucleotide sequence accession numbers.

The sequences of pHBurk3 and pHBurk5 were deposited in GenBank and assigned accession numbers EU919403 and EU919404, respectively.

RESULTS AND DISCUSSION

Construction of HimarBP and transposition in B. thailandensis and B. pseudomallei.

Initial experiments with a nonreplicative Himar1 delivery plasmid, i.e., pFNLTP16 H1 (16), yielded Km^r transposants at frequencies that were between 11,000 (B. thailandensis)- and 250 (B. pseudomallei)-fold lower than those obtained with the same plasmid backbone in F. tularensis (not shown). For construction of a Himar1 delivery system allowing efficient one-step transposon delivery and transposition, we therefore made use of the previously isolated conditional broad-host-range ori₁₆₀₀-rep(Ts) replicon (7). Plasmids containing this replicon can be efficiently electroporated into B. thailandensis and B. pseudomallei and maintained in single copy at the permissive temperature (30°C) but not at the nonpermissive temperature (37°C or greater) (7). The prototype plasmid pHBurk1 (Fig. 1) contains a Himar1 transposon named HimarBP1 in which transcription of the tnp gene is initiated by a mycobacterial promoter and an nptII gene, approved for use in B. pseudomallei, transcribed from its endogenous promoter (16). The nptII gene specifying Km^r is the sole selection marker present in pHBurk1 and all of its derivatives. By a previously described one-step transposition protocol (16), HimarBP1 was transposed into B. thailandensis and B. pseudomallei with efficiencies of 8 × 10⁻⁶ and 1.2 × 10⁻⁴ to 4.7 × 10⁻⁵ events per viable cell, respectively. Similar efficiencies were observed with HimarFT in F. tularensis, where tnp and nptII selection marker transcription was driven by Francisella-specific promoters (16). Thus, there is no significant difference in transposition into the AT-rich F. tularensis genome versus the GC-rich Burkholderia sp. genomes.

The next-generation pHBurk plasmids (pHBurk2, pHBurk3, and pHBurk6; Fig. 1) were designed with two goals in mind, (i) utilization of excisable Km^r selection markers because of the paucity of approved selection markers for use in B. pseudomallei and (ii) increased tnp transcription, which may result in increased transposition efficiencies. First, pHBurk1 was modified with a polylinker that would facilitate directed cloning of promoter-containing fragments (pHBurk-Link-2). Second, the resident nptII gene on pHBurk1 was replaced with a FRT-nptII-FRT cassette so that the nptII gene could be excised from transposon integrants with the help of Flp recombinase (pHBurk2 and pHBurk3, containing HimarBP2 and HimarBP3, respectively). Third, pHBurk6 (HimarBP6) was constructed such that tnp transcription would be promoted by the B. thailandensis P_S12 promoter, which was previously used for driving gene expression in B. thailandensis (4) and B. pseudomallei (7). By the one-step transposition protocol, pHBurk2, pHBurk3, and pHBurk6 (containing HimarBP2, HimarBP3, and HimarBP6, respectively; HimarBP6 was only studied in B. thailandensis) were transposed into B. thailandensis and B. pseudomallei with similar efficiencies (1 × 10⁻⁵ to 1 × 10⁻⁷), indicating that driving tnp transcription from the B. thailandensis P_S12 promoter did not result in significantly increased transposition efficiencies. As a matter of fact, promoting transcription from the strong P_S12 promoter may be counterproductive as sequencing of the P_S12-containing region of several pHBurk6 isolates with primer 1722 revealed a single base deletion in the −10 region. The data presented in Table 2 clearly indicate that transposition efficiencies were significantly higher with delivery vectors where the nptII and tnp genes are in the same orientation, e.g., pHBurk1, pHBurk3, and pHBurk5, versus those plasmids containing these genes in the opposite orientation, e.g., pHBurk2 and pHBurk4. Transposition efficiencies were not significantly increased with pHBurk5 containing the E. coli lac operon promoter reading toward tnp in addition to the nptII promoter.

TABLE 2.

Frequency of transposition of HimarBP derivatives into B. pseudomallei 1026b^a

Himar1 derivative	No. of transformantsb	No. of insertionsc	Avg no. of insertions/avg no. of transformants
pHBurk1	1.86 × 10−4 ± 1.35 × 10−4	8.15 × 10−5 ± 3.69 × 10−5	0.44
pHBurk2	2.45 × 10−4 ± 1.52 × 10−4	1.01 × 10−5 ± 1.18 × 10−6	0.04
pHBurk3	3.48 × 10−4 ± 2.43 × 10−4	1.2 × 10−4 ± 5.46 × 10−5	0.35
pHBurk4	1.46 × 10−4 ± 8.49 × 10−5	3.83 × 10−6 ± 2.61 × 10−6	0.03
pHBurk5	2.78 × 10−4 ± 1.53 × 10−4	1.00 × 10−4 ± 4.69 × 10−5	0.36

^aResults shown are averages from two (pHBurk2), three (pHBurk1 and pHBurk4), or five (pHBurk3 and pHBurk5) separate experiments.

^bShown is the average number of viable CFU per milliliter recovered after growth on selective medium at the permissive temperature (30°C).

^cShown is the average number of viable CFU per milliliter recovered after growth on selective medium at the nonpermissive temperature (37°C).

Because the highest transposition efficiencies were consistently obtained with pHBurk3, it was used for further studies.

HimarBP3 transposition in B. pseudomallei.

With the one-step delivery and transposition procedure, the efficiency of plating at the nonpermissive temperature was approximately 35% of that observed at the permissive temperature (Table 2). Frequencies of transposition ranged from 1.81 × 10⁻⁴ to 3.22 × 10⁻⁵. Chromosomal insertion versus plasmid maintenance was investigated by colony PCR with primers 511 and 512 and primers 1768 and 1769, respectively. These analyses revealed that the plasmid was lost in all of the investigated cases and that the Km^r phenotype was due to chromosomal HimarBP3 insertion. Km^r colonies obtained after 24 to 72 h of incubation at 37°C all contained HimarBP3 insertions, as assessed by colony PCR with primers 511 and 512, which was performed on 66 random colonies picked after 24, 48, and 72 h of incubation time. The longer incubation times needed to obtain a significant number of Km^r colonies are therefore not of concern. Similar observations were made with 62 colonies obtained with pHBurk5.

Verification of HimarBP3 transposition and stability in B. pseudomallei.

The one-step transposition protocol with pHBurk3 was used to obtain Km^r colonies of strain 1026b which were picked and purified after a 48-h incubation at 37°C. Genomic DNA was isolated from 14 randomly selected Km^r colonies and 4 auxotrophic colonies (see below), digested with NotI, and hybridized with a probe that recognized the ori_R6K sequences present on HimarBP3. Single bands of different sizes were obtained in all cases, as shown in Fig. 2 (lanes a) for five isolates, suggesting a single and random insertion of HimarBP3 into the B. pseudomallei genome. The same result was obtained when 15 Km^r B. thailandensis isolates mutagenized with HimarBP1 were analyzed (data not shown), suggesting that HimarBP transposons are functional in and can be used for random mutagenesis strategies of other Burkholderia spp., especially the closely related category B agent B. mallei, the etiologic agent of glanders (19, 29).

FIG. 2.

Transposition of HimarBP3 and stability in B. pseudomallei 1026b. Genomic DNA was prepared from mutants after initial isolation (a) or after ∼100 generations in the absence of kanamycin selection (b), digested overnight with NotI, and transferred to a nylon membrane. The membrane was hybridized with a probe that detected ori_R6K on HimarBP3. Isolates 1 to 5 are randomly selected Km^r colonies. Wild-type B. pseudomallei 1026b was included as a negative control (lane −). The positive control (lane +) was MluI-digested pHBurk3. The 10-, 8-, 6-, 5-, 4-, and 3-kb (top to bottom) fragments contained in the biotinylated 2-log DNA ladder (New England BioLabs) are in lane M.

To assess the stability of HimarBP3 insertions, five randomly selected Km^r mutants were grown for ∼100 generations in the absence of kanamycin selection, after which all five of the isolates recovered were still Km^r. Genomic DNA was extracted from the five original Km^r isolates and the five mutants that were grown in the absence of selection, and Southern analysis was performed as described above. Identical bands were present in the original Km^r isolates and the bacteria grown in the absence of selection (Fig. 2, lanes b). These results indicated that HimarBP3 insertions in the B. pseudomallei genome are stable in the absence of continued antibiotic selection.

Since currently only kanamycin and zeocin markers are approved for the genetic manipulation of wild-type B. pseudomallei bacteria (gentamicin is also approved, but its use is confined to efflux pump-deficient mutant derivatives), Km^r tagging of mutants severely impacts downstream genetic manipulations such as complementation, double-mutant isolation, reporter gene tagging, etc. This issue was overcome by equipping the HimarBP transposons with a Flp recombinase-excisable Km^r marker. To assess Flp-mediated Km^r marker excision, selected Km^r mutants were transformed with pFLPe2 containing a Zeo^r marker and Flp excision was performed as previously described (7). As expected, kanamycin-susceptible colonies were readily obtained with marker excision efficiencies ranging from 20 to 70%. Marker-free mutants were then obtained by growing kanamycin-susceptible colonies at 37°C, a nonpermissive temperature for pFLPe2. All of the marker-free mutants analyzed by sequence analysis of a 398-bp PCR fragment amplified with primers 1832 and 1833 had the expected physical structures, i.e., a single FRT site in place of the excised FRT-nptII-FRT cassette.

Because HimarBP3 mutants containing the Km^r selection marker were stable for ∼100 generations in the absence of antibiotic selection, the isogenic marker-free mutants should also be stable.

Determination of HimarBP3 insertion sites.

HimarBP3 insertion sites in B. pseudomallei strain 1026b were mapped by rescue of HimarBP3 and sequence analysis of insertions. This was achieved by ligation of NotI-digested DNA fragments and recovery of plasmid DNA from Km^r E. coli DH5α(λpir) transformants. Both transposon-chromosomal DNA junction sequences were obtained by priming sequencing reaction mixtures with transposon-specific oligonucleotides 1670 and 1829.

Because the annotated sequence of strain 1026b is not yet available, insertions were mapped relative to the strain 1710b genome. This mapping revealed that insertions were randomly distributed on both chromosomes with no apparent regional bias (Fig. 3A). The small number of insertions relative to the large genome allowed no conclusive predictions about the variety of insertions with respect to open reading frame (ORF) or transposon orientation, although there was a slight tendency toward having transposons inserted in genes whose orientation was the same as the chromosome, irrespective of transposon orientation (Fig. 3B). Of the 24 mapped insertions, 4 were in intergenic regions and 20 were within predicted ORFs (Table 3). Transposon insertions were observed in genes involved in biosynthetic pathways, metabolic pathways, DNA repair, gene regulation, and secretion. These observations are similar to those made during Himar1 mutagenesis of F. tularensis (16).

FIG. 3.

Mapping of HimarBP3 insertions in the B. pseudomallei genome. (A) Transposon HimarBP3 insertions were mapped to the chromosomes of 1710b, with the exception of two insertions (labeled K) that could only be mapped to K96243 chromosome 1 (GenBank accession number NC006350). Filled and open circles denote insertions where HimarBP3 is either inserted in the same direction as or opposite to the chromosome. (B) Graphical representation of the orientation of HimarBP3 insertions. Major features of HimarBP3 are shown, including the locations of the two sequencing primer (P1 and P2) binding sites, and the orientations of sequence extensions from these primers are indicated by arrows. The transposon was found in both orientations in the B. pseudomallei chromosome with little bias to the orientation of the ORF (arrows) at the insertion site based on the strain 1710b genome annotation. Numbers adjacent to the arrows denote isolates containing insertions in ORFs in the indicated orientations. IR, inverted repeat; nt, nucleotides.

TABLE 3.

HimarBP insertions within ORFs^a

Location of insertion	Chromosome	Gene name, putative function
BURPS1710b_3728	1	aroB, 3-dehydroquinate synthase
BURPS1710b_3718	1	gltB, glutamate synthase, large subunit
BURPS1710b_1228	1	ppc, phosphoenolpyruvate carboxylase
BURPS1710b_3425	1	pyrC, dihydroorotase, homodimeric type
BURPS1710b_A0679	2	sctV, type III secretion inner membrane protein SctV
BURPS1710b_A2481	2	ribA, GTP cyclohydrolase II
BURPS1710b_A1568	2	uvrA, excinuclease ABC, subunit A, form 2
BURPS1710b_A2192	2	Sensor histidine kinase
BURPS1710b_1949	1	Hypothetical protein (lipoprotein in strain K96243)
BURPS1710b_A2590	2	Cytochrome c family protein
BURPS1710b_0018	1	Indolepyruvate ferredoxin oxidoreductase
BURPS1710b_A2174	2	Short-chain dehydrogenase
BURPS1710b_A1577	2	plcN, phospholipase C
BURPS1710b_A1202	2	Serine/threonine protein kinase
BURPS1710b_0528	1	rbsR, transcription regulator, LacI family
BURPS1710b_A1366	2	Glutathione-dependent formaldehyde dehydrogenase
BURPS1710b_0918	1	dgoA, DgoA protein
BURPS1710b_2750	1	prlC, oligopeptidase A
BURPS1710b_1771	1	Rhs element Vgr protein subfamily, putative
BURPS1710b_1692	1	gp18

^aShown are locations of HimarBP insertions in genes and putative functions according the B. pseudomallei strain 1710b genome sequence (GenBank accession numbers CP000124 and CP000125 for chromosomes 1 and 2, respectively).

The typical TA insertion site for Himar1 transposons was observed in all 24 rescued and sequenced insertions. Deletions or duplications of flanking sequences were not detected in any of the mapped insertions.

Auxotrophic mutants obtained by HimarBP3 transposition.

To test the utility of HimarBP3 for mutant isolation, Km^r colonies were obtained after one-step transposition and 24 to 72 h of incubation at 37°C. Km^r colonies were transferred to M9 minimal-glucose kanamycin (MMGK) plates. From 2,781 Km^r colonies, 19 isolates were obtained that failed to grow on MMGK, which corresponds to 0.68% recovery of auxotrophs. Analysis of 6,124 Km^r colonies generated with HimarBP1, HimarBP2, HimarBP3, HimarBP4, and HimarBP5 yielded 44 colonies which failed to grow on MMGK plates, corresponding to 0.72% recovery of auxotrophs. This is comparable to the 1 to 2% recovery rate during isolation of auxotrophs in other bacteria (3, 13).

Four Km^r auxotrophs were selected for further characterization by genomic Southern analysis and mapping of genomic insertion sites. All mutants had single transposon insertions in the genome (not shown). This was verified by insertion site mapping, which showed that the four insertions were located in the aroB, gltB, ppc, and pyrC genes, respectively, all of which are located on chromosome 1 (Table 3). These genes encode dehydroquinate synthase, glutamate synthase (large subunit), phosphoenolpyruvate carboxylase, and dihyroorotase, respectively, which are involved in the phenylalanine, tyrosine, and tryptophan; glutamine; oxaloacetate; and pyrimidine biosynthetic pathways. The respective mutants are therefore phenylalanine, tyrosine, and tryptophan (aroB); glutamine (gltB); aspartic acid (ppc); and pyrimidine (pyrC) auxotrophs. These auxotrophies were experimentally confirmed since growth of the aroB, gltB, ppc, and pyrC mutants in MMGK medium was restored by the addition of phenylalanine, tyrosine, and tryptophan; glutamine; aspartic acid; and uracil, respectively.

These results demonstrated the utility of HimarBP3 for rapid mutant construction and characterization.

Transfer of HimarBP3 insertions between B. pseudomallei chromosomes.

We previously showed that 20- to 30-kb linear chromosomal DNA fragments tagged with an antibiotic resistance marker could be readily transferred from strain 1026b derivatives back to strain 1026b and, to a lesser extent, strain 1710b (7). To test the transfer of HimarBP3 insertions, fragmented chromosomal DNA from the four Km^r aroB, gltB, ppc, and pyrC mutants (Table 3) and a randomly selected Km^r prototroph was used to transform strain 1026b. Km^r 1026b transformants were obtained at a frequency of about 240 colonies per μg of DNA. All of the Km^r colonies obtained with DNA from the aroB, gltB, ppc, and pyrC mutants were auxotrophs, whereas all of the Km^r colonies obtained with DNA from the Km^r prototroph remained prototrophs. These data showed that HimarBP3-induced mutations can readily be transferred between strain 1026b derivatives. In this context, it should be noted, however, that not all B. pseudomallei strains are naturally transformable (28).

Conclusions.

We have developed an efficient HimarBP mutagenesis system for B. pseudomallei which continues to expand the arsenal of still fledgling select-agent-compliant tools that can be used with this bacterium. Its development takes advantage of previously constructed tools such as approved excisable selection markers and in vivo marker excision systems (7). The HimarBP elements are small (2,205 to 2,479 bp) and can thus be readily transferred between B. pseudomallei strains that are naturally transformable (7, 28), which facilitates double-mutant construction and mutant sharing by virtue of sharing sterile exempt genomic DNA rather than nonexempt live strains. The basic HimarBP transposons were engineered with ease of use (e.g., rapid and simple transposon rescue and insertion site mapping) and versatility (e.g., they can be readily equipped with other genetic elements such as other approved selection markers, outward reading promoters, reporter genes, affinity tags, etc.) in mind. Random mutagenesis strategies will greatly facilitate studies of the biology and pathogenesis of this and related understudied pathogens and perhaps facilitate the establishment of a comprehensive B. pseudomallei transposon mutant library. Such libraries have accelerated research with diverse other bacteria, including F. tularensis (11) and two Pseudomonas aeruginosa prototype strains (13, 14).