Burkholderia pseudomallei asd Mutant Exhibits Attenuated ... · plies) against biological agents and toxins in case of a bioter-rorism event. To combat potential foul play associated

INFECTION AND IMMUNITY, Oct. 2011, p. 4010–4018 Vol. 79, No. 100019-9567/11/$12.00 doi:10.1128/IAI.05044-11Copyright © 2011, American Society for Microbiology. All Rights Reserved.

The Burkholderia pseudomallei �asd Mutant Exhibits AttenuatedIntracellular Infectivity and Imparts Protection against Acute

Inhalation Melioidosis in Mice�

Michael H. Norris,2 Katie L. Propst,3 Yun Kang,1 Steven W. Dow,3

Herbert P. Schweizer,3 and Tung T. Hoang1,2*Department of Microbiology1 and Department of Molecular Biosciences and Bioengineering,2 University of

Hawaii at Manoa, Manoa, Hawaii 96822, and Department of Microbiology,Immunology, and Pathology, Colorado State University,

Fort Collins, Colorado 805233

Received 30 March 2011/Returned for modification 13 May 2011/Accepted 26 July 2011

Burkholderia pseudomallei, the cause of serious and life-threatening diseases in humans, is of nationalbiodefense concern because of its potential use as a bioterrorism agent. This microbe is listed as a select agentby the CDC; therefore, development of vaccines is of significant importance. Here, we further investigated thegrowth characteristics of a recently created B. pseudomallei 1026b �asd mutant in vitro, in a cell model, and inan animal model of infection. The mutant was typified by an inability to grow in the absence of exogenousdiaminopimelate (DAP); upon single-copy complementation with a wild-type copy of the asd gene, growth wasrestored to wild-type levels. Further characterization of the B. pseudomallei �asd mutant revealed a markeddecrease in RAW264.7 murine macrophage cytotoxicity compared to the wild type and the complemented �asdmutant. RAW264.7 cells infected by the �asd mutant did not exhibit signs of cytopathology or multinucleated giantcell (MNGC) formation, which were observed in wild-type B. pseudomallei cell infections. The �asd mutant wasfound to be avirulent in BALB/c mice, and mice vaccinated with the mutant were protected against acute inhalationmelioidosis. Thus, the B. pseudomallei �asd mutant may be a promising live attenuated vaccine strain and a biosafestrain for consideration of exclusion from the select agent list.

Burkholderia pseudomallei, a Gram-negative saprophyte andfacultative intracellular pathogen, is a common cause of envi-ronmentally acquired septicemia in Southeast Asia and north-ern Australia (7, 10, 44). It is the etiological agent of thedisease melioidosis and is listed as a category B select agent bythe U.S. Centers for Disease Control and Prevention. Bacterialselect agent research is currently focused on basic research intovirulence and pathogenesis to fulfill the five main points of thePublic Health Security and Bioterrorism Preparedness andResponse Act of 2002 (43). One of the goals is to develop andmaintain medical countermeasures (such as drugs, vaccines,and other biological products, medical devices, and other sup-plies) against biological agents and toxins in case of a bioter-rorism event. To combat potential foul play associated withintentional release of select agents, a focus on vaccine devel-opment for first responders, such as military and health serviceprofessionals, is of the utmost importance (26). Currently,there are no vaccines against B. pseudomallei, and treatmententails prolonged regimens of intravenous and orally adminis-tered antibiotic therapy (30).

In a recent publication, we described an engineered B. pseu-domallei strain with a deletional mutation in the aspartate-�-semialdehyde dehydrogenase (asd) gene, which is auxotrophicfor diaminopimelate (DAP) in rich medium and auxotrophicfor DAP, lysine, methionine, and threonine in minimal me-

dium (28); this is consistent with similar mutations in manyother bacterial species (13, 18–20). DAP is a diamino acid thatcross-links to D-alanine in neighboring peptidoglycan strands,and the �asd mutant exhibits the “pop-and-die” phenotypeassociated with an inability to synthesize DAP for cell wallbiosynthesis. Previous works have created asd mutants in Sal-monella enterica serovar Typhimurium (13) and Legionellapneumophila (17) and demonstrated a growth requirement forDAP. In addition, the S. Typhimurium �asd strain has beenextensively used in clinical studies with human subjects as avaccine delivery strain (25, 40). The pathway for synthesizingDAP from aspartate is absent in mammals; therefore, no DAPis present in mammalian hosts, including humans (13, 36). Theother amino acids (lysine, methionine, and threonine) madefrom aspartate via Asd are essential amino acids in humans,affording another possible level of nutrient limitation in vivo.Without a considerable exogenous concentration of DAP, the�asd mutant is unable to cross-link its cell wall and cannotreplicate. Even when supplied with high levels of DAP, intra-cellularly replicating L. pneumophila �asd did not recover towild-type levels of pathogenicity in either macrophage or pro-tozoan infection models (17).

Live attenuated vaccines are particularly effective vaccines,because live bacteria may replicate modestly in the host, sim-ilar to situations encountered during an actual infection. Inaddition, live attenuated vaccines contain complex epitopesnot found in subunit or heat-inactivated vaccines, and thus theystimulate parts of the immune system that could otherwise beneglected (e.g., a strong Th1 response) (12). Previous studiestesting the efficacy of auxotrophic B. pseudomallei and Burk-

* Corresponding author. Mailing address: Department of Microbi-ology, University of Hawaii at Manoa, Manoa, HI, 96822. Phone: (808)956-3522. Fax: (808) 956-5339. E-mail: [email protected].

� Published ahead of print on 1 August 2011.

4010

holderia mallei strains as live attenuated vaccines have resultedin various degrees of success (4, 42). In order to determine ifthe �asd strain may be appropriate as a future vaccine candi-date, we evaluated the growth and attenuation of the B. pseu-domallei 1026b �asd mutant in vitro and in cell culture. Animalstudies were carried out to determine virulence levels andattenuation of the �asd strain. Efficacy of the B. pseudomallei1026b �asd strain as a live attenuated vaccine against inhala-tion melioidosis was then ascertained in a BALB/c mousemodel.

MATERIALS AND METHODS

Bacterial strains, media, and culture conditions. All manipulations of B.pseudomallei were conducted in CDC/USDA-approved and -registered biosafetylevel 3 (BSL3) facilities at the University of Hawaii at Manoa and Colorado StateUniversity, and experiments with select agents were performed in accordancewith the recommended BSL3 practices (32). Derivatives of Escherichia colistrains EPMax10B (Bio-Rad), E1345, E1354, E1869, and E1889 (Table 1) wereroutinely used for cloning or plasmid mobilization into B. pseudomallei as de-scribed previously (24, 28). Luria-Bertani (LB) medium (Difco) was used toculture E. coli strains. B. pseudomallei strains were cultured in LB or 1� M9minimal medium supplemented with 20 mM glucose (MG). Antibiotics andnonantibiotic antibacterials in solid media were utilized as follows: for E. coli,glyphosate (GS) at 0.3% (wt/vol) and phosphinothricin (PPT) at 0.3% (wt/vol);for B. pseudomallei, GS at 0.3% (wt/vol) and PPT at 2.5% (wt/vol). Growth of E.coli �asd strains and preparation of DAP were carried out as previously de-scribed (2). Selections for bar and gat genes in E. coli and B. pseudomallei strainswere performed as previously described (28). B. pseudomallei �asd::gat strainswere grown on LB containing 200 �g/ml DAP or on MG containing 1 mM Lys,1 mM Met, 1 mM Thr, and 200 �g/ml meso-DAP, as described previously (28).

Molecular methods and reagents. Molecular methods, PCR conditions, andconjugation into select agents were conducted as described previously (2, 28, 35).

Engineering of B. pseudomallei �asdBp::gat-FRT. B. pseudomallei �asdBp::gat-FRT was engineered as described previously (28); briefly, the allelic replacementvector pBAKA-�asdBp::FRT-gat was conjugally introduced into B. pseudomalleistrain 1026b, and selection of the mutation was carried out on MG medium plus200 �g/ml DAP, 0.3% GS, and 1 mM (each) of Lys, Met, and Thr (these 3 aminoacids [3AA] are required for the specific �asd mutation). Colonies were streakedon the same medium supplemented with 0.1% p-chlorophenylalanine (cPhe) tocounterselect against pheS. GS-resistant mutants were purified once on LB plusDAP and patched again on MG plus 0.3% GS, 0.1% cPhe, and 1 mM 3AA withor without 200 �g/ml DAP to confirm the phenotype.

Construction of single-copy rfp-containing vectors. The red fluorescent pro-tein gene (rfp) was optimized for the codon preference of B. pseudomallei, andthe constitutive B. pseudomallei rpsL promoter (PS12) was incorporated upstreamof the gene (29). Constructed as previously described (29), rfp was cloned frompUC57-PS12-rfp into mini-Tn7-PCS12-bar to yield mini-Tn7-bar-rfp (Fig. 1). Themini-Tn7-PCS12-bar (28) construct was digested with PstI and SpeI and ligated tothe rfp fragment obtained from pUC57-PS12-rfp digested with PstI and XbaI,producing mini-Tn7-bar-rfp. Next, the complementation and fluorescent taggingtransposon was constructed by digesting mini-Tn7-bar-asdBp (28) with PstI andSpeI and ligating it to the rfp fragment from PstI- and XbaI-digested pUC57-PS12-rfp, yielding the single-copy complementation/fluorescence tagging vectormini-Tn7-bar-asdBp-rfp. In addition to the bar-based vector, the fluorescencetagging vector mini-Tn7-gat-rfp, based on gat and constructed as previously de-scribed (29), was also utilized.

Engineering of rfp-tagged B. pseudomallei strains and complemented mutants.E1354 was utilized as the conjugal donor to introduce the single-copy vector mini-Tn7-bar-rfp into the B. pseudomallei �asd mutant for fluorescent tagging, producingB. pseudomallei �asd::gat-FRT/attTn7-bar-rfp (�asd/rfp). The mini-Tn7-bar-asdBp-rfpconstruct was introduced into the �asd mutant for complementation and fluorescenttagging, yielding B. pseudomallei �asd::gat-FRT/attTn7-bar-asdBp-rfp (�asd/comple-ment/rfp). The mini-Tn7-gat-rfp construct was introduced into wild-type B. pseu-domallei strain 1026b for fluorescent tagging of the wild type (wt), resulting in B.pseudomallei attTn7-gat-rfp (wt/rfp). These strains were obtained from a triparentalmating experiment using the pTNS3-asdEc helper plasmid, and bacteria containingthe integrated transposon were selected and screened via PCR as described previ-

TABLE 1. Strains used in the study

Strain Lab ID Relevant properties Source

E. coli strainsEPMax10B-pir116/�asd::Gmr E1345 Gmr; F� �� mcrA �(mrr-hsdRMS-mcrBC)

�80lacZ�M15 �lacX74 deoR recA1 endA1araD139 �(ara leu)7697 galU galK rpsLnupG Tn-pir116-FRT2 �asd::Gmr-wFRT

Available lab straina

EPMax10B-pir116/�asd/�trp::Gmr/ mob-Kmr E1354 Gmr Kmr; F� �� mcrA �(mrr-hsdRMS-mcrBC) �80lacZ�M15 �lacX74 deoR recA1endA1 araD139 �(ara leu)7697 galU galKrpsL nupG Tn-pir116-FRT2 �asd::wFRT�trp::Gmr-FRT5 mob�recA::RP4-2Tc::Mu-Kmr

Available lab strain

EPMax10B-lacIq/pir E1869 F� �� mcrA �(mrr-hsdRMS-mcrBC)�80lacZ�M15 �lacX74 deoR recA1 endA1araD139 �(ara leu)7697 galU galK rpsLnupG lacIq-FRT8 pir-FRT4

Available lab strain

EPMax10B-lacIq/pir/leu E1889 F� �� mcrA �(mrr-hsdRMS-mcrBC)�80dlacZ�M15 �lacX74 deoR recA1 endA1galU galK rpsL nupG lacIq-FRT8 pir-FRT4

Available lab strain

B. pseudomallei strains1026b (wt) B0004 Wild-type strain, clinical melioidosis isolate 111026b attTn7-gat-rfp (wt/rfp) B0015 GSr; 1026b with mini-Tn7-gat-rfp inserted at

glmS2 site29

1026b-�asdBp::gat-FRT (�asd) B0011 GSr; 1026b with gat-FRT cassette inserted inasdBp gene

28

1026b-�asdBp::gat-FRT/attTn7-bar-rfp (�asd/rfp) B0024 GSr PPTr; 1026b �asdBp::gat-FRT mutant withmini-Tn7-bar-rfp inserted at glmS2 site

This study

1026b-�asdBp::gat-FRT/attTn7-bar-asdBp-rfp(�asd/complement/rfp)

B0022 GSr PPTr; 1026b �asdBp::gat-FRT mutant withmini-Tn7-bar-asdBp-rfp inserted at glmS2site

This study

a E. coli strains are available lab strains (when requesting, please use the Lab ID number).

VOL. 79, 2011 ATTENUATED B. PSEUDOMALLEI �asd MUTANT VACCINATION OF MICE 4011

ously (8, 24, 28). The mini-Tn7 system allows site-specific insertion of the transposonat a neutral site in the chromosome, downstream of any glmS gene, of which B.pseudomallei has three (8). In all cases the transposon had inserted at the highlyfavored glmS2 site. Fluorescence was verified by fixing the bacteria with fresh 1%paraformaldehyde in phosphate-buffered saline (PBS) for 30 min followed by im-aging with a Zeiss Axio Observer D.1 fluorescence microscope and the accompa-nying AxioVision release 4.7 software.

Construction of fluorescent strains, including selection for glyphosphate orphosphinothricin-resistant colonies, and transposon integration and screeningwere performed as previously described (8, 24, 28, 29). Sample preparation andfluorescent imaging were also carried out as previously described (29).

Growth analysis of the rfp-tagged B. pseudomallei �asd mutant, complemented�asd mutant, and wild-type strain. Growth curve experiments were performedon the three RFP-labeled B. pseudomallei strains engineered above (wt/rfp,�asd/rfp, and �asd/complement/rfp). These strains were grown overnight at 37°Cin LB medium, where the �asd mutant was supplemented with 200 �g/ml ofDAP. Overnight cultures were then washed twice with 1� M9 medium to removetrace amounts of DAP and resuspended in an equal volume of 1� M9 medium.Resuspended cultures were diluted 100-fold into fresh LB medium, withoutDAP, and shaken at 225 rpm at 37°C. At each time point, 300-�l aliquots wereremoved and diluted 2-fold in LB medium, and their optical densities weremeasured at 600 nm using an Eppendorf Biophotometer.

DAP dependency of B. pseudomallei �asd/rfp. A growth curve experiment wasperformed on the B. pseudomallei �asd/rfp and B. pseudomallei wt/rfp strains.The strains were grown overnight and washed with 1� M9 and inoculated intoLB media described above, supplemented with different concentrations of DAP(0, 50, 100, 200, and 500 �g/ml). At each indicated time point, 300-�l aliquotswere removed and the optical densities at 600 nm were determined.

Intracellular replication assays. Both murine macrophage RAW264.7 andhuman cervical carcinoma HeLa cell lines were grown in a 5% CO2 environmentat 37°C in Dulbecco’s modified Eagle’s medium (DMEM) with 10% (vol/vol)fetal bovine serum (FBS). Gibco 100� antibiotic/antimycotic was added at a 1�working concentration (containing 100 U/ml of penicillin, 100 �g/ml of strepto-mycin, and 250 ng/ml of amphotericin B) to the cell culture medium during cellculture growth but was omitted during the infection assay. Intracellular replica-tion assays were performed using a modified kanamycin protection assay aspreviously described (22). Briefly, cells (HeLa and RAW264.7 lines) were cul-

tured in DMEM to confluence, scraped from cell culture flasks, and seeded at1 � 105 cells per well into 24-well Corning CellBIND culture plates. To preparecells for infection study, cells were allowed to attach overnight and were washedtwice with 1� PBS in the morning.

The three bacterial strains used in this experiment were the same as usedabove for the complementation growth study. To investigate whether exogenousDAP allowed intracellular infection, two series of infection were carried out withthe �asd/rfp strain. One series was allowed to infect during the entire course ofthe study in the presence of DAP, and the other had DAP omitted from themedium after 1 h of infection (T � 1). During the first hour of infection, bothseries of the B. pseudomallei �asd/rfp strain were supplemented with 200 �g/mlof DAP in the cell culture medium, so as not to bias the invasion ability duringattachment and internalization. Assays with the wild-type and complemented�asd/rfp mutant strains were carried out essentially as those with the �asd/rfpmutant, except that no DAP was added. Briefly, B. pseudomallei strains weregrown to a high cell density, washed twice with 1� PBS, and then diluted to �1 �106 CFU/ml. At time zero, 1 ml of DMEM containing diluted bacteria was addedto the macrophage monolayers (multiplicity of infection [MOI], 10:1). Afterallowing the infection to progress for 1 h, the medium was removed and themonolayers were washed twice with 1� PBS to remove any unattached bacteria.Next, fresh DMEM with 700 �g/ml each of amikacin and kanamycin was addedto the monolayers to kill any noninternalized bacteria and inhibit extracellularbacterial replication. During the assay, medium was removed from the wells atthree time points (2, 6, and 24 h postinfection), and the infected cell monolayerswere washed twice with 1� PBS and then lysed with 0.1% Triton X-100. Serialdilutions of the lysates were plated on Brucella agar (Difco) plus 4% (vol/vol)glycerol (BAG) medium at 37°C, as described previously (6, 14). BAG mediumwas supplemented with 200 �g/ml of DAP when enumerating B. pseudomallei�asd/rfp colonies. Colonies were counted within 48 h. Experiments with bothHeLa and RAW264.7 cell lines, in combination with all bacterial strains, wereperformed in triplicate, and the standard errors of the means (SEM) werecalculated for each.

RAW264.7 macrophage cytotoxicity assay. Macrophages were cultured as de-scribed above and seeded into a 96-well CellBIND plate at �5 � 104 cells perwell. A kanamycin protection and infection assay was carried out, as describedabove for the intracellular replication assay, with bacteria infected at an MOI of10:1. At 2, 6, 12, and 24 h postinfection, the cellular supernatant was removed,and lactate dehydrogenase (LDH) levels were determined using the CytoTox 96nonradioactive cytotoxicity assay (Promega). LDH levels of infected monolayerswere compared and normalized to maximal LDH levels (after complete mono-layer lysis using 0.1% Triton X-100) to determine the percent cytotoxicity. Thecytotoxicity assay was carried out in triplicate, and the SEMs were calculated.

Light microscopy and time course of B. pseudomallei wt/rfp, �asd/rfp, and�asd/complement/rfp infection of RAW264.7 murine macrophages. Light micros-copy of infected cell monolayers was carried out as described previously (29),except for a few modifications. Glass coverslips were sterilized in 70% (vol/vol)ethanol and then treated for 4 h with 150 �g/ml poly-L-lysine in sterile double-distilled water (ddH2O). The glass coverslips were washed twice with ddH2O andallowed to air dry within a sterile petri dish overnight. In our experience, glasscoverslips treated with poly-L-lysine provided the best surface for cell attachmentand microscopic imaging. The 22- by 22-mm coverslips were placed at thebottoms of the wells in a 6-well Corning CellBIND plate prior to seeding.RAW264.7 macrophages were seeded at �8 � 105 cells per well and allowed toattach overnight, and the infection was initiated by adding different bacterialstrains at an MOI of 10:1. At 1 h postinfection, the coverslips were washed twicewith 1� PBS, and then fresh DMEM containing 700 �g/ml of kanamycin wasadded to inhibit extracellular bacterial replication. At 2, 6, 12, and 24 h postin-fection, the medium was removed and the cell monolayers were washed twicewith 1� PBS and fixed with 1% paraformaldehyde for 30 min. After 30 min, theparaformaldehyde was removed and the coverslips were washed twice with 1�PBS. For safe removal of fixed samples from the BSL3 cabinet for imaging, thismethod must be initially tested by incubating fixed coverslips for 5 days in LB toconfirm the absence of growth and viable bacteria. Coverslips were mounted witha slide-mounting buffer containing 50% glycerol in 1� PBS. Images were ob-tained as previously described (29).

Animal studies. BALB/c mice between 4 and 6 weeks of age were purchasedfrom Jackson Laboratories (Bar Harbor, ME). Animals were housed in microiso-lator cages under pathogen-free conditions. The Institutional Animal Care andUse Committee at Colorado State University approved the animal experimentsconducted for these studies. B. pseudomallei infections were done using intrana-sal (i.n.) inoculation (31). Animals were anesthetized with 100 mg of ketamine/kgof body weight plus 10 mg/kg xylazine. The desired challenge dose of B. pseu-domallei was suspended in PBS, and 20 �l was delivered i.n., into alternating

FIG. 1. Mini-Tn7-bar-rfp, single-copy tagging vector based on phos-phinothricin resistance, harboring rfp driven by the PS12 promoter.After insertion aided by pTNS3-asdEc (29), the non-antibiotic resis-tance marker, which is flanked by identical FRTs, can be removed byFlp-mediated excision. Abbreviations: bar, gene encoding bialaphos/phosphinothricin resistance; FRT, Flp recombination target sequences;oriT, RP4 conjugal origin of transfer; PCS12, promoter of the Burk-holderia cenocepacia rpsL gene; PS12, promoter of the B. pseudomalleirpsL gene; R6K ori, � protein-dependent R6K origin of replication;Tn7L/Tn7R, left and right transposase recognition sequences; ToT1,transcriptional terminator.

4012 NORRIS ET AL. INFECT. IMMUN.

nostrils. For the challenge studies, groups of 5 mice were challenged with thewild-type or mutant strain. For the vaccination studies, mice (n � 10) wereadministered 1 � 107 CFU �asd mutant B. pseudomallei intranasally and thenboosted in the same manner 3 weeks later. Two weeks following the boost, micewere challenged intranasally with 4 � 103 CFU wild-type B. pseudomallei 1026b.For all B. pseudomallei challenge and survival studies, animals were monitoredfor disease symptoms twice daily and were euthanized according to predeter-mined humane end points. Lungs, liver, and spleen were removed and homog-enized using a tissue stomacher (Teledyne Tekmar, Mason, OH), and homoge-nates were plated in serial dilutions to determine bacterial counts in the B.pseudomallei-challenged mice 75 days postinfection. Statistical differences insurvival times were determined by Kaplan-Meier curves followed by the log-ranktest (Prism5 software; GraphPad, La Jolla, CA).

RESULTS

Construction and growth analysis of the rfp-tagged B. pseu-domallei �asd, �asd/complement, and wild-type strains. In pre-vious work (28), we developed two nonantibiotic markers, barand gat, which are effective for the genetic manipulation of B.pseudomallei. In this study it became apparent that anothermarker besides gat was needed for fluorescent tagging of B.pseudomallei during infection studies. Therefore, we con-structed a new non-antibiotic-based single-copy transposonvector (mini-Tn7-bar-rfp) (Fig. 1) for stable site-specific inser-tion of rfp genes without the need for plasmid maintenance.We used the gat select agent-compliant nonantibiotic markerto fluorescently tag wild-type bacteria as previously described(29). However, mini-Tn7-bar-rfp-based constructs were used tofluorescently tag the B. pseudomallei �asd::gat-FRT strain, aswell as to tag and complement this mutant strain.

The amino acid requirements of the B. pseudomallei �asd/rfpmutant were previously demonstrated by showing that the mu-tant could not grow in the absence of methionine, threonine,

and DAP on minimal medium plates (28). In this study, wewanted to show that this mutant is unable to grow in rich liquidmedium in the absence of DAP. A growth curve experimentwas initiated to allow the comparison of growth between B.pseudomallei wt/rfp, B. pseudomallei �asd/rfp, and the B. pseu-domallei �asd/complement/rfp strains in a rich nutrient source(LB medium). The B. pseudomallei �asd mutant displayed aninability to grow compared to the wild-type strain (Fig. 2A).This was expected, as in previously published work the B.pseudomallei 1026b �asd mutant began to lyse after 6 h withoutDAP (28). When the B. pseudomallei �asd/rfp strain was com-plemented using a transposon containing a single copy of theB. pseudomallei asd gene, the growth defect of the mutant wasabolished and normal growth was restored. This indicated thatthe growth defect exhibited by the mutant was solely caused bydeletion of the asd gene.

We next investigated the effects of different concentrationsof DAP on growth of the �asd/rfp mutant in rich medium (Fig.2B). In the presence of DAP, however, all curves showed asignificant growth lag. Compared to the wild type, the opticaldensity at 600 nm eventually reached wild-type levels. Thisdemonstrated that the �asd mutant can grow well when DAPis added to the medium and not at all in the absence of DAP.

The B. pseudomallei �asd mutant is highly attenuated inintracellular replication. Assessment of the �asd mutant at-tenuation in HeLa and RAW264.7 cell infection models wasnecessary before animal vaccination. In agreement with previ-ous work (22), our experience suggested that an MOI of 10:1would initiate an infection that would maximally affect the cellmonolayer within 24 h. Internalization was very inefficient inHeLa cells, with only �500 CFU out of �1 � 106 CFU inter-

FIG. 2. Growth curve experiments performed with B. pseudomallei strains. (A) B. pseudomallei strains were grown in the absence of DAP. Thewild-type strain and the �asd strain complemented with a single copy of the asd gene on a site-specific transposon exhibited the same growth ratesand final optical densities, while the �asd mutant exhibited a typical DAP-dependent phenotype. (B) The B. pseudomallei �asd mutant was testedin different concentrations of DAP, ranging from 0 �g/ml to 500 �g/ml. Compared to wild type, the �asd mutant exhibited absence of growthwithout DAP. All other concentrations of DAP afforded a partial growth rate recovery and final optical density, albeit only after a lag in growth.

VOL. 79, 2011 ATTENUATED B. PSEUDOMALLEI �asd MUTANT VACCINATION OF MICE 4013

nalized by the monolayers (Fig. 3A). As shown in Fig. 3A, theB. pseudomallei �asd/rfp strain was unable to replicate in HeLacells. It was able to attach and become internalized as well asthe wild type, as indicated by similar intracellular CFU ob-tained at 2 h postinfection. However, by 6 h postinfection (T �6), the �asd/rfp mutant alone (without DAP) was only able toreplicate modestly. By 24 h postinfection, intracellular mutantbacteria were undetectable. However, when complementedwith a single wild-type copy of the asd gene, the �asd mutantstrain behaved exactly as the wild type, reaching a maximum of�1 � 106 CFU. Interestingly, the �asd/rfp strain could infectHeLa cell monolayers when the growth medium was supple-mented with 200 �g/ml of DAP and only slowed down the rateof decline in RAW 264.7 macrophages. This indicated that asufficient amount of DAP was transported into HeLa cells.

B. pseudomallei was internalized efficiently and replicatedmuch more significantly within RAW264.7 cells (Fig. 3B). TheB. pseudomallei �asd/rfp strain was not internalized efficientlyor was killed more efficiently by macrophages than the wildtype, as indicated by the 1-log difference in CFU at 2 h postin-fection. The �asd mutant could not sustain a wild-type level ofreplication even in the presence of 200 �g/ml of DAP, unlikeHeLa cell infection. By the end of the assay (24 h), the �asd/complement/rfp and the wt/rfp strains replicated to a similarlevel (�1 � 106 CFU) within the RAW264.7 macrophagemonolayer. This indicated that even a single copy of the asdBp

gene can restore the mutant’s abilities to grow within cells.Cytotoxicity and light/fluorescence microscopy time course

of the B. pseudomallei 1026b �asd mutant infection ofRAW264.7 murine macrophages. Although the �asd mutantdoes not replicate to high cell numbers like the wild-type, itwas of interest whether or not the �asd mutant damages thecell monolayers comparably to the wild-type. LDH assays of allstrains infecting RAW264.7 monolayers at 2 and 6 h postin-fection revealed little differences in cytotoxicity compared to

the noninfected control (Fig. 4A), while LDH levels of the wildtype and complement began to rise at 12 h postinfection. Thewild-type-infected and complemented �asd mutant-infectedmonolayers had surpassed the maximum cytotoxicity at 24 h(determined by lysing the initially seeded macrophages),reaching �100% (Fig. 4A). Independently of DAP, the �asdmutant did not damage the monolayers to the same level as thewild type and was still comparable to the noninfected control.LDH levels of the noninfected macrophage monolayer rose at24 h, indicating the spontaneous lysis of macrophages at highconfluence, a usual occurrence where rapid division leads tolow nutrient availability, macrophage death, and LDH release.

Intracellular replication and host cell cytotoxicity were thenplaced in a visual context by tracking the rfp-tagged bacteria viafluorescence microscopy. Visible in representative images from24 h postinfection was a pervasive red fluorescence indicativeof high numbers of intracellular bacteria (Fig. 4B). The ma-jority of macrophages were joined together in MNGCs. Uponcloser inspection (Fig. 5), the MNGCs were observed teamingwith bacteria in both the wt/rfp- and �asd/complement/rfp-infected monolayers. The bacteria-containing protrusions wereclearly visible, extending from the surfaces of the macrophages(Fig. 5). The �asd mutant-infected monolayers neither con-tained high numbers of replicating bacteria nor did they showany sign of MNGC formation in the monolayer at 24 h, and infact they appeared as healthy as the noninfected controls (Fig.4B). Although not unexpected, the data reaffirm that the mu-tant was internalized by the macrophages but was unable toproduce cytopathology (or MNGCs), with or without DAP,consistent with the wild-type infection.

Attenuation, vaccination, and acute protection of the �asdmutant in mice. We first tested the �asd mutant for attenua-tion in vivo. The 50% lethal dose (LD50) of B. pseudomallei1026b in BALB/c mice has been determined to be approxi-mately 900 CFU via the inhalation route (14). An i.n. dose of

FIG. 3. Infection of HeLa and RAW264.7 cells by B. pseudomallei and the �asd strain. HeLa (A) and RAW264.7 (B) cell monolayers wereinfected at an MOI of 10:1. The complemented �asd strain showed no decrease in its ability to invade and replicate within either cell line. However,the �asd mutant in the absence of DAP could not sustain an infection in either cell line, denoted by an overall drop in bacterial numbers.

4014 NORRIS ET AL. INFECT. IMMUN.

4,500 CFU has been experimentally determined to produce100% mortality in BALB/c mice after 3 days (21, 31). Intrana-sal inoculation mimics inhalation melioidosis and produces acharacteristic acute pneumonic infection to which BALB/cmice succumb within a few days. Five BALB/c mice werechallenged i.n. with 4,500 CFU of B. pseudomallei 1026b, andanother five BALB/c mice were challenged i.n. with 1 � 107

CFU (5 logs � LD50). Survival of the mice was then moni-tored. After 3 days, mice challenged with wild-type B. pseu-domallei had all been euthanized due to progressive infection(Fig. 6A). In contrast, mice infected with asd mutant B. pseu-domallei showed no outward signs of infection and were ob-served for 75 days postchallenge, and all remained healthyduring this period (Fig. 6A). Thus, the �asd mutant was highlyattenuated compared to the wild-type strain of B. pseudomallei.To assess possible bacterial persistence in vivo, mice chal-lenged with mutant B. pseudomallei were euthanized on day 75,and the lungs, livers, and spleens were homogenized, diluted,and plated on LB agar. Bacteria were not detected in anyorgan, based on assays with limits of detection of approxi-mately 50 CFU/organ, indicating the mutant bacteria did notpersist in organs typically infected during the chronic phase ofinfection with virulent B. pseudomallei. The numbers of mice

used in these studies were judged to be adequate (31) to ensurethat the mutant bacterium was avirulent in immunocompetentmice.

We then considered if the �asd mutant could be used as avaccine against inhalation melioidosis in BALB/c mice. Nu-merous publications support the fact that single vaccinationswith attenuated live B. pseudomallei vaccines are generallyunable to protect mice from developing chronic melioidosis (4,15, 41). Therefore, we investigated whether an i.n. prime-boostvaccination strategy could extend protection against develop-ment of chronic melioidosis. The i.n. route of infection andvaccination emulates aerosol exposure/vaccination, and typi-cally vaccination at the route of pathogen entry generally leadsto more effective disease prevention (3). Ten BALB/c micewere primed with an i.n. vaccination of 1 � 107 CFU of the B.pseudomallei 1026b �asd mutant. Three weeks later, the samemice were boosted with another i.n. vaccination of 1 � 107

CFU of the �asd mutant. The time period between the initialexposure and the boost would presumably allow for an adap-tive cellular and humoral immune response to occur. Twoweeks postboost, the mice were challenged with 4 � 103 CFUof wild-type strain 1026b, and survival was compared to unvac-cinated mice challenged with the same amount of the wild type.

FIG. 4. (A) Cytotoxicity of B. pseudomallei strains to the RAW264.7 murine macrophage cell line. RAW264.7 cells were infected with the �asdmutant (in the presence or absence of DAP), the complemented mutant strain, and wild-type strain. Between 2 h and 6 h postinfection there wasa slight increase in cytotoxicity associated with infection by the complemented and wild-type strains compared to the �asd mutant-infectedmonolayers. By 12 h postinfection, cytotoxicities of the complement- and wild-type-infected monolayers were more obvious, while cytotoxicitycaused by the �asd mutant remained similar to the noninfected control. At 24 h postinfection, the complement- and wild-type-infected monolayersexhibited maximal cytotoxicity. (B) Microscopy and time course of the cytopathic effects of B. pseudomallei �asd infection. Monolayers wereinfected at an MOI of 10:1 and then analyzed for red fluorescence at 2 h and 24 h postinfection. Differential interference contrast (DIC) imageswere overlaid with the red fluorescent channel. Red fluorescence indicates the presence of B. pseudomallei. Note the high bacterial levels in thecomplement- and wild-type-infected monolayers at 24 h and the confluent MNGC formation. This coincides with high levels of cytotoxicity at 24 hpostinfection. Abbreviations: CT, noninfected control; �asd, B. pseudomallei �asd/rfp; �asd DAP, B. pseudomallei �asd/rfp in the presence of200 �g/ml of DAP; �asd complement, B. pseudomallei �asd/complement/rfp; WT, B. pseudomallei wt/rfp. Error bars represent the SEM of threeexperiments. Statistical significance was determined by the two-tailed unpaired t test (***, P � 0.0005).

VOL. 79, 2011 ATTENUATED B. PSEUDOMALLEI �asd MUTANT VACCINATION OF MICE 4015

The boost was administered 2 weeks before the infection tofurther enhance the immune response, presumably allowingtime for enhanced adaptive immunity (33, 38). The datashowed that vaccinated mice survived significantly longer thanthe unvaccinated control mice (Fig. 6B). While the prime-boost strategy used in this study protected mice from acuteinfection, it failed to protect mice from development of chronicB. pseudomallei infection, as nearly all of the vaccinated andchallenged mice developed infection of organs at secondarysites, particularly the spleen (data not shown).

DISCUSSION

The essentiality of the asd gene in E. coli has been known forsome time, but its requirement for growth and infectivity ofselect agent species has not been thoroughly investigated. Thisstudy evaluated the growth and pathogenicity of the B. pseu-domallei �asd mutant produced in the previous work (28) andits potential for use as a live attenuated vaccine. By performinggrowth experiments, it was found that without DAP the mutantis unable to replicate and complementation, with a single copyof wild-type asdBp, is sufficient for in vitro and intracellularreplication compared to the wild-type bacterium in both HeLa

and RAW264.7 cells. These studies demonstrated that by add-ing DAP, the �asd strain can be easily propagated within alaboratory setting and, by complementing the asd gene, amarkerless balanced lethal system could be used for various B.pseudomallei studies (37). It is important to note that whileadding DAP during asd mutant infection can reestablish wild-type growth levels in some cell lines (i.e., HeLa cells), it is nothomologous to replication of the mutant after single-copycomplementation, where high levels of replication are seenwithin both cell lines. This may have important implications infuture subcutaneous vaccine experiments due to the epithelialnature of the HeLa cell line.

In the absence of DAP, the mutant was unable to replicatein either the HeLa or RAW 264.7 cell infection models. Cy-totoxicity data showed that the mutant did not cause increaseddeath or distress to the macrophages, indicating that the mu-tant is unable to replicate within or cause significant damage tohost cell macrophages via endotoxin or exoenzyme release.The link between cytotoxicity and inflammation has been

FIG. 5. Intracellular replication of B. pseudomallei. RAW264.7 mu-rine macrophage monolayers were visualized using a combination ofdifferential interference contrast and red fluorescence microscopy 24 hpostinfection with the B. pseudomallei �asd/complement/rfp strain (Aand B) and with B. pseudomallei wt/rfp strain (C and D). All macro-phages in the field of view are interacting with MNGCs and are filledwith intracellular bacteria about to burst into the extracellular milieu.Images in panels A and C were captured at 600� magnification, whileimages in panels B and D are zoomed-in images of the regions denotedin panels A and C, respectively. Note the large number of bacteriaprojecting out of the remaining macrophages in panels C and D. Therewere no bacteria and there was an absence of protrusions as well asMNGCs at all time points in monolayers infected with the �asd mutant(data not shown).

FIG. 6. (A) The B. pseudomallei 1026b �asd mutant is avirulent inmice. Mice (n � 5 animals per group) were challenged i.n. with either4,500 CFU B. pseudomallei 1026b (wild type) or 1 � 107 CFU B.pseudomallei 1026b �asd mutant, and survival was monitored. Statis-tical differences in survival times were determined by Kaplan-Meiercurves followed by log-rank test (**, P � 0.01 for B. pseudomallei1026b wt versus B. pseudomallei 1026b �asd mutant). (B) Intranasalvaccination with the B. pseudomallei 1026b �asd mutant protects micefrom lethal B. pseudomallei challenge. Mice (n � 10 animals pergroup) were primed i.n. with 1 � 107 CFU B. pseudomallei 1026b �asdand boosted in the same manner 3 weeks later. Two weeks postboost,mice were challenged i.n. with 4 � 103 CFU wild-type B. pseudomallei1026b. Survival was monitored, and statistical differences in survivaltimes were determined by Kaplan-Meier curves followed by a log-ranktest (***, P � 0.0001 for vaccinated versus nonvaccinated mice). Datarepresen two individual pooled experiments.

4016 NORRIS ET AL. INFECT. IMMUN.

known for some time and can be partially attributed to freeradical release during cellular damage both in vitro and in vivo(27). Although inflammatory modulators were not measured inthe cytotoxicity assay, a tentative hypothesis would place thecorresponding inflammatory modulator levels in the sametrend as LDH. By tagging the B. pseudomallei strains with RFPand tracking them in vitro during intracellular replication, wewere able to confirm the intracellular location of the mutantand further demonstrate the utility of non antibiotic selectablemarkers in pathogenesis research.

B. pseudomallei �asd mutants should be considered biosafestrains suitable for laboratory use and exclusion from theUSDA/CDC select agent lists. First, the �asd mutant was con-structed by deleting several hundred bases in the middle of theasd gene (28), producing a stable mutant unable to revert.Additionally, as DAP is not present within mammals, there isno source of exogenous DAP, affording another level of safetyfor this strain. On the other hand, it can be seen that the B.pseudomallei �asd mutant invades host cells although, like theB. pseudomallei purM mutant (31), the �asd mutant is unableto replicate in the host, and bacterial persistence cannot occur.These data, together with the use of �asd mutants of other spe-cies as vaccine delivery strains in humans (13, 41), provide strongevidence supporting the removal of this strain from select agentlists, as was previously done for the 1026b �purM strain Bp82(31).

Previous vaccination studies utilizing live attenuated strainsand a single vaccination were unable to prevent death fromchronic infection (1, 4, 9, 34, 39). However, we had reason tobelieve that the �asd mutant would be more effective thanprevious live attenuated strains. It has been shown that a moreprotective immune response can be achieved by increasingshort-term vaccine persistence, which we attempted with thebooster vaccine (5, 23, 45). Unfortunately, while vaccinationwith the �asd mutant did indeed protect against acute melioid-osis, the vaccine failed to protect against chronic melioidosis.This failure might have been because the �asd mutant vaccinewas unable to persist long enough or disseminate and prolif-erate enough, even after the boost, to induce systemic protec-tion. The route of vaccination can be important because ofincreased protection at the site of challenge (e.g., mucosalsurfaces); however, this may not generate systemic protection(3). This is a possible reason for why mice were protected fromthe initial lung infection but eventually succumbed to systemicinfection at secondary sites. Even so, protection from acutepneumonic melioidosis may provide a vital increase in survivaltime that could allow for the administration of antibiotic ther-apeutics.

Future investigations should be carried out to addresswhether short-term persistence, proliferation, and dissemina-tion of the �asd mutant, achieved by adding DAP to thevaccine, would provide systemic protection against chronic me-lioidosis. Within-host persistence of the �asd mutant wouldthen be contingent on the amount of DAP administered withthe vaccine. Addition of DAP to mutant-infected HeLa cells,an epithelial cell line, did allow some intracellular replication;therefore, it may be highly beneficial to incorporate a subcu-taneous vaccine containing DAP. Longer exposure to the cu-taneous and subcutaneous dendritic cells could prolong T-cellactivation at draining lymph nodes (33) and create the power-

ful cell-mediated immune response hypothetically necessaryfor sterile immunity (16). Other means of producing systemicdissemination and, perhaps, protection would be to incorpo-rate an intramuscular or subcutaneous vaccination along withthe inhaled vaccination. This two-pronged approach may giverise to longer protection from chronic or latent infection,which is proving more difficult to combat than acute melioid-osis. Seemingly, the greatest prospect for an effective vaccineagainst melioidosis is a live attenuated strain. In conclusion,this initial work suggests the utility of the B. pseudomallei �asdmutant as a live attenuated vaccine against acute melioidosisand further justifies its potential removal from the select agentlist.

ACKNOWLEDGMENTS

This project was supported by award number AI065359 from theNational Institute of Allergy and Infectious Diseases and by Center ofBiomedical Research Excellence grant P20RR018727 from the Na-tional Center for Research Resources of the National Institutes ofHealth. S.W.D. and H.P.S. were supported by NIH NIAID grantAI065357.

The content of this report is solely the responsibility of the authorsand does not necessarily represent the official view of the fundingagencies.

REFERENCES

1. Atkins, T., et al. 2002. Characterisation of an acapsular mutant of Burkhold-eria pseudomallei identified by signature tagged mutagenesis. J. Med. Micro-biol. 51:539–553.

2. Barrett, A. R., et al. 2008. Genetics tools for allelic-replacement in Burk-holderia species. Appl. Environ. Microbiol. 74:4498–4508.

3. Belyakov, I. M., and J. D. Ahlers. 2009. What role does the route of immu-nization play in the generation of protective immunity against mucosalpathogens? J. Immunol. 183:6883–6892.

4. Breitbach, K., J. Kohler, and I. Steinmetz. 2008. Induction of protectiveimmunity against Burkholderia pseudomallei using attenuated mutants withdefects in the intracellular life cycle. Trans. R. Soc. Trop. Med. Hyg. 2008:S89–S94.

5. Buckley, A. M., et al. 2010. Evaluation of live-attenuated Salmonella vaccinesexpressing Campylobacter antigens for control of C. jejuni in poultry. Vaccine28:1094–1105.

6. Burtnick, M. N., et al. 2008. Burkholderia pseudomallei type III secretionsystem mutants exhibit delayed vacuolar escape phenotypes in RAW 264.7murine macrophages. Infect. Immun. 76:2991–3000.

7. Cheng, A. C., and B. J. Currie. 2005. Melioidosis: epidemiology, pathophys-iology, and management. Clin. Microbiol. Rev. 18:383–416.

8. Choi, K.-H., et al. 2008. Genetic tools for select-agent-compliant manipula-tion of Burkholderia pseudomallei. Appl. Environ. Microbiol. 74:1064–1075.

9. Cuccui, J., et al. 2007. Development of signature-tagged mutagenesis inBurkholderia pseudomallei to identify genes important for survival and patho-genesis. Infect. Immun. 75:1186–1195.

10. Dance, D. A. B. 1991. Melioidosis: the tip of the iceberg? Clin. Microbiol.Rev. 4:52–60.

11. DeShazer, D., P. J. Brett, R. Carlyon, and D. E. Woods. 1997. Mutagenesisof Burkholderia pseudomallei with Tn5-OT182: isolation of motility mutantand molecular characterization of the flagellin structural gene. J. Bacteriol.179:2116–2125.

12. Drabner, B., and C. A. Guzman. 2001. Elicitation of predictable immuneresponses by using live bacterial vectors. Biomol. Eng. 17:75–82.

13. Galan, J. E., K. Nakayama, and R. Curtiss. 1990. Cloning and characteriza-tion of the asd gene of Salmonella typhimurium: use in stable maintenance ofrecombinant plasmids in Salmonella vaccine strains. Gene 94:29–35.

14. Goodyear, A., et al. 2009. Protection from pneumonic infection with Burk-holderia species by inhalational immunotherapy. Infect. Immun. 77:1579–1588.

15. Haque, A., et al. 2006. A live experimental vaccine against Burkholderiapseudomallei elicits CD4 T cell-mediated immunity, priming T cells specificfor 2 type III secretion system proteins. J. Infect. Dis. 194:1241–1248.

16. Haque, A., et al. 2006. Role of T cells in innate and adaptive immunityagainst murine Burkholderia pseudomallei infection. J. Infect. Dis. 193:370–379.

17. Harb, O. S., and Y. A. Kwaik. 1998. Identification of the aspartate-beta-semialdehyde dehydrogenase gene of Legionella pneumophila and character-ization of a null mutant. Infect. Immun. 66:1898–1903.

VOL. 79, 2011 ATTENUATED B. PSEUDOMALLEI �asd MUTANT VACCINATION OF MICE 4017

18. Hatten, L.-A., H. P. Schweizer, N. Averill, L. Wang, and A. B. Schryvers.1993. Cloning and characterization of the Neisseria meningitidis asd gene.Gene 129:123–128.

19. Haziza, C., P. Stragier, and J. C. Patte. 1982. Nucleotide sequence of the asdgene of Escherichia coli: absence of a typical attenuation signal. EMBO J.1:379–384.

20. Hoang, T., S. Williams, H. P. Schweizer, and J. S. Lam. 1997. Moleculargenetic analysis of the region containing the essential Pseudomonas aerugi-nosa asd gene encoding aspartate-�-semialdehyde dehydrogenase. Microbi-ology 143:899–907.

21. Jeddeloh, J. A., D. L. Fritz, D. M. Waag, J. M. Hartings, and G. P. Andrews.2003. Biodefense-driven murine model of pneumonic melioidosis. Infect.Immun. 71:584–587.

22. Jones, A. L., T. J. Beveridge, and D. E. Woods. 1996. Intracellular survival ofBurkholderia pseudomallei. J. Bacteriol. 64:782–790.

23. Kahl-McDonagh, M. M., and T. A. Ficht. 2006. Evaluation of protectionafforded by Brucella abortus and Brucella melitensis unmarked deletion mu-tants exhibiting different rates of clearance in BALB/c mice. Infect. Immun.74:4048–4057.

24. Kang, Y., M. H. Norris, A. R. Barrett, B. A. Wilcox, and T. T. Hoang. 2009.Engineering of tellurite-resistant genetic tools for single-copy chromosomalanalysis of Burkholderia spp. and characterization of the Burkholderia thai-landensis betBA operon. Appl. Environ. Microbiol. 75:4015–4027.

25. Kong, W., et al. 2008. Regulated programmed lysis of recombinant Salmo-nella in host tissues to release protective antigens and confer biologicalcontainment. Proc. Natl. Acad. Sci. U. S. A. 105:9361–9366.

26. Lane, H. C., J. L. Montagne, and A. S. Fauci. 2001. Bioterrorism: a clear andpresent danger. Nat. Med. 7:1271–1273.

27. McCord, J. M., and K. Wong. 1979. Phagocyte-produced free radicals: rolesin cytotoxicity and inflammation. John Wiley & Sons, Ltd., Hoboken, NJ.

28. Norris, M. H., Y. Kang, D. Lu, B. A. Wilcox, and T. T. Hoang. 2009.Glyphosate resistance as a novel select-agent-compliant, non-antibiotic-se-lectable marker in chromosomal mutagenesis of the essential genes asd anddapB of Burkholderia pseudomallei. Appl. Environ. Microbiol. 75:6062–6075.

29. Norris, M. H., Y. Kang, B. Wilcox, and T. T. Hoang. 2010. Stable, site-specificfluorescent tagging constructs optimized for Burkholderia species. Appl. En-viron. Microbiol. 76:7635–7640.

30. Peacock, S. J., et al. 2008. Management of accidental laboratory exposure toBurkholderia pseudomallei and B. mallei. Emerg. Infect. Dis. 14:e2.

31. Propst, K. L., T. Mima, K.-H. Choi, S. W. Dow, and H. P. Schweizer. 2010.A Burkholderia pseudomallei �purM mutant is avirulent in immunocompe-tent and immunodeficient animals: candidate strain for exclusion from se-lect-agent lists. Infect. Immun. 78:3136–3143.

32. Richmond, J. Y., and R. W. McKinney. 2007. Biosafety in microbiologicaland biomedical laboratories, 5th ed. Centers for Disease Control and Pre-vention, Atlanta, GA.

33. Roake, J. A., et al. 1995. Dendritic cell loss from nonlymphoid tissues aftersystemic administration of lipopolysaccharide, tumor necrosis factor, andinterleukin 1. J. Exp. Med. 181:2237–2247.

34. Rodrigues, F., et al. 2006. Global map of growth-regulated gene expressionin Burkholderia pseudomallei, the causative agent of melioidosis. J. Bacteriol.188:8178–8188.

35. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratorymanual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor,NY.

36. Schleifer, K. H., and O. Kandler. 1972. Peptidoglycan types of bacterial cellwalls and their taxonomic implications. Bacteriol. Rev. 36:407–477.

37. Spreng, S., and J.-F. Viret. 2005. Plasmid maintenance systems suitable forGMO-based bacterial vaccines. Vaccine 23:2060–2065.

38. Stavnezer, J. 1996. Immunoglobulin class switching. Curr. Opin. Immunol.8:199–205.

39. Stevens, M. P., et al. 2003. A Burkholderia pseudomallei type III secretedprotein, BopE, facilitates bacterial invasion of epithelial cells and exhibitsguanine nucleotide exchange factor activity. J. Bacteriol. 185:4992–4996.

40. Tacket, C. O., et al. 1997. Safety and immunogenicity in humans of anattenuated Salmonella typhi vaccine vector strain expressing plasmid-en-coded hepatitis B antigens stabilized by the asd-balanced lethal vector sys-tem. Infect. Immun. 65:3381–3385.

41. Titball, R. W., et al. 2008. Burkholderia pseudomallei: animal models ofinfection. Trans. R. Soc. Trop. Med. Hyg. 102:S111–S116.

42. Ulrich, R. L., K. Amemiya, D. M. Waag, C. J. Roy, and D. DeShazer. 2005.Aerogenic vaccination with a Burkholderia mallei auxotroph protects againstaerosol-initiated glanders in mice. Vaccine 23:1986–1992.

43. U.S. Code of Federal Regulations. 2002. Public Health Security and Bioter-rorism Preparedness and Response Act, 107th Congr. Public Law 107-18. 42CFR 73.21.

44. Wiersinga, W. J., T. van der Poll, N. J. White, N. P. Day, and S. J. Peacock.2006. Melioidosis: insight into the pathogenicity of Burkholderia pseudomal-lei. Nat. Rev. Microbiol. 4:272–282.

45. Zhao, Z., et al. 2008. Subcutaneous vaccination with attenuated Salmonellaenterica serovar Choleraesuis C500 expressing recombinant filamentoushemagglutinin and pertactin antigens protects mice against fatal infectionswith both S. enterica serovar Choleraesuis and Bordetella bronchiseptica.Infect. Immun. 76:2157–2163.

Editor: F. C. Fang

4018 NORRIS ET AL. INFECT. IMMUN.