Genome-wide identification of Francisella tularensis virulence determinants

Abstract

Francisella tularensis is a gram-negative pathogen that causes life-threatening infections in humans and has potential for use as a biological weapon. The genetic basis of the F. tularensis virulence is poorly understood. This study screened a total of 3,936 transposon mutants of the live vaccine strain for infection in a mouse model of respiratory tularemia by signature-tagged mutagenesis. We identified 341 mutants attenuated for infection in the lungs. The transposon disruptions were mapped to 95 different genes, virtually all of which are also present in the genomes of other F. tularensis strains, including human pathogenic F. tularensis strain Schu S4. A small subset of these attenuated mutants carried insertions in the genes encoding previously known virulence factors, but the majority of the identified genes have not been previously linked to F. tularensis virulence. Among these are genes encoding putative membrane proteins, proteins associated with stress responses, metabolic proteins, transporter proteins, and proteins with unknown functions. Several attenuated mutants contained disruptions in a putative capsule locus which partially resembles the poly-gamma-glutamate capsule biosynthesis locus of Bacillus anthracis, the anthrax agent. Deletional mutation analysis confirmed that this locus is essential for F. tularensis virulence.

FIG. 1.

Construction of STM strains in LVS by transposon mutagenesis. (A) The 52-bp oligonucleotide tags were separately cloned in the KpnI-digested EZ::TN transposon as marked by inverted repeats (IR). The tagged transposon was liberated by digestion with PvuII and mixed with EZ::TN transposase to form a transposome for transformation into LVS. The kanamycin-resistant transformants were sorted into various sets based on the sequence identity of the tag in each of the transposons. (B) Detection of transposons in STM strains. A DNA blot representing 10 STM strains was probed with a 270-bp probe representing the kanamycin resistance cassette (Kan^r) of the tagged transposon. The molecular sizes of DNA markers are marked in bases.

FIG. 2.

Negative screening of the STM pools in mice. Each mutant in a single set was grown separately to an OD₆₀₀ of 0.35 to prepare the input mutant pool. The pool was used to infect four BALB/c mice by intranasal inoculation. The mice were sacrificed 7 days postinfection to remove the lungs and recover the bacteria (output). The input and output pools of the same mutant sets were compared by amplifying the transposon with tag-specific primers and detection by agarose gel electrophoresis. The mutants that were missing in the output pools from multiple mice infected with the same input pools are indicated with asterisks.

FIG. 3.

In trans complementation of the capB and capC disruptive mutants. (A) Genetic organization of the capBCA locus in LVS. The genes and their coding directions are represented with open arrows and arrowheads, respectively. The transposon insertion sites of the capB (JS2512) and capC (JS2531) mutants are indicated with vertical arrowheads. The numbers of intergenic nucleotides are indicated at the bottom. (B) Generation of complementation constructs in plasmid pMP633. The wild-type capB and capC genes of LVS were inserted into the NdeI site of pMP633. (C) Bacterial load in lungs of mice infected with LVS and isogenic transposon mutants. BALB/c mice were intranasally infected with strains LVS (12,400 CFU/mouse), JS2512 (capB mutant; 30,840 CFU/mouse), JS2531 (capC mutant; 28,000 CFU/mouse), capB-complemented JS2512 (34,000 CFU/mouse), and capC-complemented JS2531 (30,000 CFU/mouse). The bars represent the mean CFU values + standard errors of bacterial levels in the lungs of three mice 7 days postinfection. P values were determined based on comparisons of the values between the transposon mutants and corresponding in trans-complemented strains using Student's t test in Microsoft Excel.

FIG. 4.

Construction and characterization of the capBCA deletion mutant. (A) Construction of the deletional mutation in the capBCA locus. The sequences flanking the capBCA genes were separately amplified and cloned in the SacI/XhoI-digested pDMK plasmid. The resulting plasmid was introduced into LVS by conjugation. Following sequential selection steps with kanamycin and sucrose, the resultant strains (as represented by strain ST938) were examined for loss of the capBCA genes by PCR. The sizes of the representative molecular markers (M) are indicated in kb. Approximate locations of the relevant primers are marked with small lateral arrows. (B) Growth of the capBCA deletion mutant in MHB. LVS and ST938 were separately cultured in 5 ml MHB. Optical absorbance levels (OD₆₀₀) were determined for each culture at the indicated time points. The value represents the means + standard errors of five independent cultures. (C) Bacterial load in mice infected with the capBCA deletion mutant ST938 (6,500 CFU/mouse) and LVS (4,000 CFU/mouse). BALB/c mice were intranasally infected with ST938 or LVS and sacrificed 7 days later to determine bacterial levels in the lungs, livers, and spleens. The bars represent the mean CFU values + standard errors of six mice. Asterisks indicate P < 0.05 as determined using Student's t test in Microsoft Excel.