- Email: [email protected]
Allelic exchange in Francisella tularensis using PCR products
FEMS Microbiology Letters 229 (2003) 195^202
www.fems-microbiology.org
Allelic exchange in Francisella tularensis using PCR products Crystal M. Lauriano a , Je¡rey R. Barker a , Francis E. Nano b , Bernard P. Arulanandam c , Karl E. Klose a; a
Department of Microbiology and Immunology, University of Texas Health Science Center, San Antonio, TX 78229-3900, USA b Department of Biochemistry and Microbiology, University of Victoria, Victoria, BC, Canada c Department of Biology, University of Texas San Antonio, San Antonio, TX, USA Received 5 August 2003 ; received in revised form 1 October 2003; accepted 28 October 2003 First published online 20 November 2003
Abstract We describe here a technique for allelic exchange in Francisella tularensis subsp. novicida utilizing polymerase chain reaction (PCR) products. Linear PCR fragments containing gene deletions with an erythromycin resistance cassette insertion were transformed into F. tularensis. The subsequent ErmR progeny were found to have undergone allelic exchange at the correct location in the genome; the minimum flanking homology necessary was 500 bp. This technique was used to create mglA, iglC, bla, and tul4 mutants in F. tularensis subsp. novicida strains. The mglA and iglC mutants were defective for intramacrophage growth, and the tul4 mutant lacked detectable Tul4 by Western immunoblot, as expected. Interestingly, the bla mutant maintained resistance to ampicillin, indicating the presence of multiple ampicillin resistance genes in F. tularensis. 7 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. Keywords : Tularemia ; Francisella tularensis ; Mutagenesis
1. Introduction Francisella tularensis, a Gram-negative coccobacillus, can cause the life-threatening disease tularemia in humans. It can be acquired through skin abrasions, ingestion of contaminated food or water, by a vector-borne route (e.g. ticks), or most seriously, via aerosols. Aerosol administration can lead to the pneumonic form of the disease, which has a high mortality rate. Because of this, F. tularensis was developed as a bioweapon in several national weapons programs. There is currently no approved vaccine available for protection against this disease. F. tularensis is classi¢ed as one of the most dangerous bioweapons (Category A) by the CDC (for review, see [1]). Several di¡erent subspecies of F. tularensis have been identi¢ed, including subspecies tularensis (Type A) and subspecies holarctica (Type B). Subsp. tularensis is the most virulent form of F. tularensis because it is associated with the highest morbidity and mortality in humans, and
* Corresponding author. Tel. : +1 (210) 567 3990; Fax : +1 (210) 567 9231. E-mail address : [email protected] (K.E. Klose).
also exhibits the highest virulence by all routes of administration in animal models, while subsp. holarctica causes infections that are rarely fatal in humans. An additional subspecies, novicida, is essentially avirulent for humans, but causes a disease in some animal models that is virtually indistinguishable from that caused by subsp. tularensis. All the subspecies of F. tularensis are closely related at the genomic level, and the reasons for the di¡erences in tropism and virulence are yet to be determined [2]. Relatively little is known about molecular mechanisms of Francisella pathogenesis. F. tularensis can survive and replicate within macrophages both in vitro and in vivo, and this appears to be its primary mechanism of pathogenesis [3,4]. The Francisella-containing vacuole evades phagosome^lysosome fusion [5], and Francisella infection can cause macrophage apoptosis [6] and abrogate Toll-like receptor signaling [7]. Several genes necessary for intramacrophage survival have been identi¢ed, including mglAB [8] and iglC [9,10], but their function(s) have not yet been elucidated. No Type III or Type IV secretion systems that may be responsible for altering the intramacrophage environment have been identi¢ed in the partially completed genome sequence of a subsp. tularensis strain [11]. The current knowledge of Francisella pathogenesis is
0378-1097 / 03 / $22.00 7 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/S0378-1097(03)00820-6
FEMSLE 11299 8-12-03
Cyaan Magenta Geel Zwart
196
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
based upon studies utilizing either an attenuated (for humans) subsp. holarctica ‘live vaccine strain’ (LVS), or a wild-type subsp. novicida strain, due to the relative ease of working with these organisms under less stringent biosafety conditions. The attenuating mutation(s) in the LVS strain are unde¢ned, and this strain demonstrates attenuated virulence by the intradermal route but not the intraperitoneal route in mice [12]. The wild-type subsp. novicida strain U112 is more virulent by the intradermal and intranasal routes than LVS in mice ([13] and KEK and BPA, unpublished data), and thus is more similar to subsp. tularensis in this animal model. The subsp. novicida strain has also proven to be more easily manipulated genetically than LVS, making this an attractive model for studying Francisella pathogenesis. Very few genetic tools exist with which to dissect virulence mechanisms of Francisella. The mutant strains that have been generated in subsp. novicida have been constructed by random transposon mutagenesis [8,9]. While this technique has been invaluable in identifying Francisella virulence genes, we have observed that transposon insertions are highly unstable in subsp. novicida, despite the lack of a transposase gene within the inserted transposon. Thus we have developed a technique to mutagenize speci¢c genes within subsp. novicida using polymerase chain reaction (PCR) products. We have used this technique to successfully create mutations in four di¡erent Francisella genes.
2. Materials and methods 2.1. Strains and media F. tularensis subsp. novicida strain U112 (ATCC) was
grown on tryptic soy broth/agar (Difco) supplemented with 0.1% cysteine (TSAcys) or on cysteine heart agar (Difco) supplemented with 5% horse blood (CHA). Escherichia coli strain DH5K [14] was used for cloning, and grown on Luria broth (Difco). The concentrations of antibiotics used, when appropriate, were: erythromycin 250 Wg ml31 , kanamycin 50 Wg ml31 , and tetracycline 10 Wg ml31 . 2.2. Plasmid construction The same general strategy was used to create all deletion/insertion constructs. A V500^600-bp fragment corresponding to the region 5P of the deletion was PCR ampli¢ed from U112 chromosomal DNA with primers 1 and 2 (Table 1), as was a V500^600-bp fragment corresponding to the region 3P of the deletion with primers 3 and 4. Primers 2 and 3 share a complementary sequence that overlaps the deletion to be created and also contains EcoRI and PstI restriction sites. The two PCR-ampli¢ed fragments were then annealed and used as a template in a second PCR reaction with primers 1 and 4, which resulted in ampli¢cation of the deletion of interest (which contains EcoRI and PstI sites) with 500^600 bp of £anking homology. Primers 1 and 4 typically contained BamHI and HindIII or SalI sites, respectively, so the ampli¢ed PCR product would then be cleaved with these two enzymes and ligated into pWSK30 [15] digested similarly. The ermC gene was PCR ampli¢ed from pIM13 [16] with primers ermC1 and ermC2 that contain EcoRI and PstI sites, and the resulting fragment was digested with these enzymes and ligated into the plasmid containing the deletion at the corresponding sites. This resulted in the construction of (vgene of interest : :ermC). The vmglA: :TetR construct was obtained by PCR ampli¢cation of the tet gene from pACYC184 [17] with
Table 1 Oligonucleotides used in this study MglA1 MglA2 MglA3 MglA4 IglC1 IglC2 IglC3 IglC4 Tul4-1 Tul4-2 Tul4-3 Tul4-4 Bla1 Bla2 Bla3 Bla4 ErmC1 ErmC2 KD46-1 KD46-2
5P-GCGGATCCGAGCTCTAATCGATGATGATG-3P 5P-CTTTTGCTTTCTGCAGGCGAATTCATTAACATATTCTAACAATATTC-3P 5P-ATATGTTAATGAATTCGCCTGCAGAAAGCAAAAGGAGCTTAATATG-3P 5P-GCGCAAGCTTATACCATTGAGTTTTATAGTAG-3P 5P-GCGGATCCTACCATTTGTAAGTGAAAAAGG-3P 5P-CAATATATCCCTGCAGCGCGAATTCCTGTTGTCTTGTTATCATCTC-3P 5P-AAGACAACAGGAATTCGCGCTGCAGGGATATATTGCAGCTGCATAG-3P 5P-GCGCAAGCTTTTAAGAGTTAATAGCGGGAAG-3P 5P-GCGGATCCGTTAACTATGTTATTATAATCAAG-3P 5P-TTACTTCTTTCTGCAGCGCGAATTCTAAAGATAAAAGACTAAGCTC-3P 5P-TTTATCTTTAGAATTCGCGCTGCAGAAAGAAGTAATCGCTTCTGATTC-3P 5P-GCGCGTCGACTTATTTTCTCATCAATCTTAACC-3P 5P-GCGGATCCAAACATCAAAAAGTATGCTAC-3P 5P-TCGTAGTATTCTGCAGCGCGAATTCCAGCGTAATTATTTTTTTCATATTC-3P 5P-AATTACGCTGGAATTCGCGCTGCAGAATACTACGAAAAATGCCTAAAAT-3P 5P-GCGCGTCGACATCAATAGCACTATCTATATC-3P 5P-GCGAATTCACGCGTTGGCTAACACACACGCCATTC-3P 5P-GCGGATCCCTGCAGTTACTTATTAAATAATTTATAGC-3P 5P-GCTCTAGATTATTATGACAACTTGACGGC-3P 5P-GCTCTAGAATTCTTCGTCTGTTTCTACTG-3P
FEMSLE 11299 8-12-03
Cyaan Magenta Geel Zwart
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
primers that contained EcoRI and PstI sites, with subsequent ligation into pKEK597 (contains vmglA). The vmglA: :KanR construct was obtained by ligation of the PstI fragment of pUC4K (Pharmacia) containing KanR from Tn903 into the PstI site of pKEK597. Sequence of the subsp. novicida mglAB locus was already available [8]. The sequences of the tul4, iglC, and bla loci were obtained from the F. tularensis strain Schu4 genome sequencing consortium (http://artedi.ebc.uu.se/ Projects/Francisella/). The locations of the genes were: tul4, contig 31, nucleotides 21 972^22 421; iglC, contig 36, nucleotides 45 440^46 069, and bla, contig 33, nucleotides 122 464^123 348. The deletion constructs removed coding sequence corresponding to amino acids 101^200 of MglA, amino acids 12^205 of IglC, amino acids 13^ 139 of Tul4, and amino acids 8^308 of L-lactamase. All constructs were veri¢ed by DNA sequencing. The VRed recombinase plasmid that replicates in F. tularensis was constructed by ¢rst PCR amplifying the araCPBAD -QLexo fragment from plasmid pKD46 [18] with primers KD46-1 and KD46-2 that incorporated XbaI restriction sites. This fragment was subsequently digested with XbaI and ligated into the Francisella plasmid pKK214 [19] that had been digested similarly to form pKEK612. 2.3. Transformation Deletion/insertions (constructed as described above) were PCR ampli¢ed utilizing KOD HiFi DNA polymerase (Novagen), with the exception of the long-range vmglA: :ermC PCR product described in Section 3, which was ampli¢ed with KOD XL DNA polymerase (Novagen). The PCR products were transformed into strain U112 via a cryotransformation technique [20], which was modi¢ed in the following manner. Brie£y, 2 Wg of PCR product was added to U112 cells at an OD600 of 65^70 in transformation bu¡er [20], incubated at room temperature for 10 min, then £ash frozen in liquid nitrogen for 5 min. Cells were then thawed at room temperature for 5 min, and placed on 0.2 W nitrocellulose ¢lter paper on CHA for 4^6 h at 37‡C. The bacteria on ¢lter paper were then resuspended in 0.2 M KCl and plated on TSAcys containing 250 Wg ml31 erythromycin and incubated at 37‡C. This resulted in the creation of strains KKF34 (vmglA: :ermC), KKF24 (viglC: :ermC), KKF29 (vtul4: :ermC), and KKF31 (vbla: :ermC). 2.4. Southern hybridization and Western immunoblot Southern blot was performed utilizing ECL detection reagent (Amersham Pharmacia), according to manufacturer’s directions. Whole cell lysates were separated by 15% sodium dodecyl sulfate (SDS)^polyacrylamide gel electrophoresis and probed by Western immunoblot with rabbit
FEMSLE 11299 8-12-03
197
KTul4 (kindly provided by M. Forsman), and detected utilizing ECL detection reagent (Amersham Pharmacia). 2.5. Intramacrophage survival assay Wild-type or mutant F. tularensis subsp. novicida strains were used to infect the J774.1 macrophage cell line (ATCC) at an MOI of V1:1. Wells were seeded with V105 J774 cells, and bacterial inocula ranged from 1.56^5.8U105 . After 1 h incubation at 37‡C, 5% CO2 , gentamicin (50 Wg ml31 ) was added to the medium to eliminate extracellular organisms. The macrophage cells were lysed with 0.2% deoxycholate 24 h post infection, and the lysate plated on TSAcys, incubated at 37‡C, and CFU enumerated.
3. Results 3.1. Transposon insertions are unstable in F. tularensis subsp. novicida F. tularensis subsp. novicida strain GB5 [8] was constructed by random mutagenesis with mTn10Km [21] into a plasmid containing the mglAB locus. A clone with a transposon insertion in mglA was subsequently selected and recombined into the F. tularensis subsp. novicida U112 genome and con¢rmed by Southern blot immediately after isolation. When utilizing strain GB5 as a control in Southern blot and PCR ampli¢cation, we were surprised that this strain appeared to lack an insertion element within the mglA gene, even though the strain maintains resistance to kanamycin (data not shown). We ampli¢ed and sequenced the mglA coding region from GB5, and found a 52-bp Tn10 excision scar at the location of the original transposon insertion, indicating that the transposon had hopped from its original location. Southern blot con¢rmed that the mTn10Km transposon is located elsewhere within the genome of GB5 (data not shown). The GB5 strain has undergone minimal passage (approximately three times) since its initial construction. Similar instability of other mTn10Km insertions in other F. tularensis subsp. novicida genes has since been observed (FEN, unpublished data). Given that this mini transposon lacks a transposase gene, this suggests a complementing transposase activity resident in the F. tularensis subsp. novicida genome. Insertion mutations constructed with the TnMax2 [22] transposon (a mini Tn1721 derivative) in subsp. novicida also appear to be unstable (R. Titball, personal communication). Multiple and various insertion elements with transposase genes are found in the partially completed F. tularensis subsp. tularensis genome [11]. Given the close genetic relationship between the subspecies [2], transposon instability may be a general phenomenon of Francisella spp. This led us to develop an alternative meth-
Cyaan Magenta Geel Zwart
198
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
od for targeted allelic replacement in F. tularensis subsp. novicida. 3.2. A method for allelic replacement in F. tularensis subsp. novicida using PCR products We chose the mglA gene as the target for allelic replacement, because this is one of the few known Francisella virulence genes. MglA is necessary for intramacrophage survival [8], and is a homolog of E. coli SspA, suggesting it functions as an activator of transcription [23]. A deletion of the mglA gene was constructed by a two-step PCR technique; the resulting vmglA construct was cloned into pWSK30 [15]. The deletion spans the residues encoding MglA amino acids 101^200 with 500-bp £anking homology. The ermC gene encoding erythromycin resistance [16] was ligated into the vmglA construct to create vmglA: :ermC. The vmglA: :ermC construct was PCR ampli¢ed, and the PCR product was used to transform F. tularensis subsp. novicida strain U112 via a cryotransformation technique, with selection for ErmR progeny. Two transformants were recovered. The two ErmR colonies from this initial transformation were further analyzed by Southern blot and PCR (Fig. 1). Southern analysis of the ErmR strains, utilizing a mglAP probe (upstream of the deletion/insertion) was consistent with the predicted vmglA: :ermC allelic replacement in the chromosome of both strains, resulting in a V4-kb HindIII fragment, compared to the V3-kb fragment of the parent strain. Likewise, long-range PCR, utilizing primers V5 kb upstream and downstream of mglA, and subsequent restriction analysis with EcoRI (which cuts within the ermC insertion, but not within the wild-type mglA locus), was also consistent with correct allelic replacement. The mglA gene of the long-range PCR fragment was sequenced, con¢rming the presence of the vmglA: :ermC allele (data not shown). 3.3. Length requirement of £anking homology/e¡ect of VRed recombinase on recombination
Fig. 1. Southern blot/PCR analysis of vmglA : :ermC mutant. A: The anticipated mglAB chromosomal loci in the wild-type and vmglA : :ermC mutant strains are depicted, including location of mglAP probe (‘probe’) used in Southern analysis, and position of primers used in long-range PCR. The vmglA: :ermC mutation is predicted to increase the size of the HindIII fragment containing mglAP by 1 kb, and to introduce an EcoRI site into the V10-kb long-range PCR product. B: Southern blot of HindIII digest of DNA from U112 (‘WT’; lane 1, V3-kb product) and vmglA : :ermC mutants (lanes 2 and 3, V4-kb products) utilizing mglAP probe. Migration of V/HindIII digestion is depicted on left, with 4.3- and 2.3-kb fragments noted. C: Long-range PCR products derived from U112 (‘WT’; lane 2, product V10 kb) and vmglA: :ermC mutant (lane 3, products V4 and V6 kb), which were digested with EcoRI and separated by agarose electrophoresis. One-kb ladder (Gibco BRL) is in lane 1, with 4- and 6-kb fragments noted.
Wanner and colleagues have developed a method for allelic replacement in E. coli, utilizing PCR fragments containing an antibiotic resistance cassette with only 50-bp £anking homology, that requires expression of VRed recombinase in the recipient strain [18]. We adapted this method to Francisella by creating a plasmid, pKEK612, with the Francisella-speci¢c origin of replication and the VRed recombinase gene under control of the arabinoseinducible promoter PBAD . We PCR ampli¢ed the ermC gene with primers containing 50-bp mglA £anking homology on either side, and transformed into F. tularensis subsp. novicida U112 either carrying pKEK612, or the vector pKK214, both were grown in the presence of arabinose. In the same experi-
ment, we transformed the same strains with PCR fragments containing the ermC cassette with either 500-bp mglA £anking homology (described above), or V5000bp mglA £anking homology (the long-range PCR fragment containing vmglA: :ermC, also described above). The results are shown in Table 2. No ErmR colonies were obtained with only 50-bp £anking homology, and only two ErmR colonies were obtained with 500-bp £anking homology, but the presence of V5000-bp £anking homology dramatically increased the number of ErmR colonies (V1000-fold increase). The presence of the VRed recombinase plasmid in the recipient strain had no e¡ect on number of ErmR colonies obtained, suggesting VRed
FEMSLE 11299 8-12-03
Cyaan Magenta Geel Zwart
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
199
recombinase either fails to promote recombination in F. tularensis, or that it was poorly expressed from the PBAD promoter in Francisella. We have not distinguished between these two possibilities. We constructed vmglA: :Kan and vmglA: :Tet alleles; these alleles contain an identical 500-bp £anking homology to the vmglA: :ermC allele described above. PCR products with these alleles were transformed into strain U112 in an identical manner to that used with the vmglA: :ermC fragment, but no KanR or TetR colonies were obtained, despite multiple attempts (more than four times each). Although the TetR gene used was the same as that carried on plasmid pKK214 [19], which results in TetR when carried by strain U112, we suspect the promoter is poorly expressed in single copy in Francisella. 3.4. Generation of iglC, tul4, and bla F. tularensis subsp. novicida mutants IglC has been shown to be important for intramacrophage survival both in subsp. novicida and subsp. holarctica (LVS) [9,10], while Tul4 was identi¢ed as a T-cellreactive protein with unknown function [24]. Francisella spp. are naturally ampicillin resistant, so the presence of a L-lactamase gene (bla) within the genome was predicted. We identi¢ed these open reading frames (ORFs) and the surrounding genes within the F. tularensis subsp. tularensis partial genome sequence (Schu4 genome sequencing consortium; http://artedi.ebc.uu.se/Projects/Francisella/). The coding sequences for iglC and tul4 have been previously cloned from LVS [25,26]. The bla gene was found in contig 33 (nucleotides 122 464^123 348), and the deduced ORF is 880 bp long, with 59% similarity with a class A L-lactamase of Yersinia enterocolitica [27]. Primers were designed to create in-frame deletions in iglC, tul4, and bla, as described for mglA. Due to the extremely high level of genetic relatedness between subsp. tularensis and subsp. novicida, it was possible to design primers based on the subsp. tularensis genome sequence and faithfully amplify the corresponding subsp. novicida genes in every case. viglC: :ermC, vtul4: :ermC, and vbla: :ermC alleles were constructed and PCR ampli¢ed, as outlined above. The £anking homology was V500^600 bp in each case, and the number of ErmR transformants
Table 2 E¡ect of £anking homology length and presence of VRed recombinase plasmid on recombination frequency between PCR products and F. tularensis subsp. novicida chromosome Flanking homology (bp)a
50 500 5000 a
# Transformants/recipient strain U112
U112/pKEK612
0 3 3556
0 2 2464
Transformations were performed as described in Section 2.
FEMSLE 11299 8-12-03
Fig. 2. PCR analysis of viglC: :ermC, vtul4: :ermC, and vbla: :ermC mutants. PCR analysis was performed with U112 (‘WT’) and mutant strains, each utilizing the same outward-facing primer that anneals to ermC (‘ermC primer+’) paired with a speci¢c primer that anneals downstream of the insertion site, but outside of £anking homology used to create mutation (‘iglC’, ‘tul4’, ‘bla’ primers). PCR products were separated by 1% agarose electrophoresis and stained with ethidium bromide. One-kb DNA ladder (Gibco BRL) is shown in lane 1, with the 1- and 2-kb fragments noted. With each primer pair, the wild-type strain failed to produce a PCR product (‘WT’, lanes 2, 4, 6), but the mutant strains produced the anticipated PCR product consistent with allelic replacement : 1632 bp for viglC: :ermC (‘iglC’, lane 3), 940 bp for vtul4: :ermC (‘tul4’, lane 5) and 769 bp for vbla: :ermC (‘bla’, lane 7).
for each PCR product into strain U112 was between 2 and 10 colonies. We con¢rmed the correct chromosomal allelic replacement in the resulting viglC : :ermC, vtul4: :ermC, and vbla: :ermC F. tularensis subsp. novicida strains by PCR and Southern blot. Utilizing an outward-facing primer that anneals to the ermC gene paired with a downstream primer that anneals outside of the £anking homology used to create the allelic replacement, it was clear that the correct chromosomal allelic replacement had occurred in all three strains (Fig. 2). Southern blot con¢rmed that a single copy of the ermC cassette had been incorporated into the genome of the three mutant strains (data not shown). 3.5. Characterization of the mglA, iglC, tul4, and bla F. tularensis subsp. novicida strains An intramacrophage survival assay in the murine macrophage cell line J774 was performed with the wild-type and all four mutant strains (Fig. 3). After 24 h, no iglC or mglA F. tularensis subsp. novicida were recovered from the macrophage cells, in contrast to the high numbers of the wild-type U112 strain recovered. This assay demonstrated that the vmglA: :ermC and viglC: :ermC strains had the anticipated defect in intramacrophage growth. In contrast, the vtul4: :ermC and vbla: :ermC strains were recovered at relatively high numbers from the macrophage cell line 24 h
Cyaan Magenta Geel Zwart
200
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
Fig. 3. Intramacrophage survival of F. tularensis subsp. novicida mutants. J774 cells were infected with wild-type and mutant strains, and intracellular bacteria were enumerated 24 h post-infection (Section 2). Infections were performed in triplicate for each strain, with standard deviations depicted by error bars. Bacterial inocula were in the range 1.56^5.8U105 . No bacteria were recovered for vmglA: :ermC and viglC : :ermC infections.
p.i. The amount recovered of the vtul4: :ermC mutant was approximately 10-fold lower than for the wild-type U112 strain. Clearly Tul4 and L-lactamase are not essential for survival and growth within J774 cells, but Tul4 may be required for full resistance to macrophage-killing mechanisms. Whole cell lysates of all four mutant strains were tested for the presence of Tul4 by Western immunoblot with K-Tul4 antisera (Fig. 4). The vtul4: :ermC mutant strain showed a complete lack of Tul4, while the other mutant strains resembled the wild-type strain in levels of Tul4 present. The vbla: :ermC strain was still able to grow in the presence of ampicillin at concentrations as high as 1 mg ml31 , similar to the wild-type strain (data not shown), indicating that additional ampicillin resistance gene(s) are present within the genome.
subspecies appears to lie primarily in host tropism, and not in fundamental di¡erences in pathogenic mechanisms, and therefore subsp. novicida infection of mice can be used as a model for subsp. tularensis infections. This has the advantages of (1) obviating the need for high-level biocontainment necessary for subsp. tularensis, (2) utilizing a fully virulent wild-type strain (vs. the unknown attenuating mutation(s) of the subsp. holarctica LVS), and (3) working with the subspecies that has proven to be the most amenable to genetic manipulation. The subsp. novicida mutants that have been constructed thus far (aside from spontaneous mutants) have been random transposon insertion mutants, utilizing mTnKm and TnMax2 [8,9]. As described here, mTnKm insertions are unstable in subsp. novicida, suggesting the presence of transposase(s) within the subsp. novicida genome that can act on mTnKm. This prompted us to develop an alternative method for targeted gene inactivation in subsp. novicida described here that would result in stable mutations. A recent report has described a method for targeted gene inactivation in LVS via conjugation [10] ; hopefully this method will also be successful in subsp. novicida. The allelic exchange method described here utilizes linear DNA for transformation, which allows one to select for a double recombination in the gene of interest, and therefore there is no need for subsequent selection or screening. We determined the minimum £anking homology necessary for this method is 500 bp, which prevents the use of the £anking homology being incorporated directly into the PCR primers (as can be done with the PCRbased method of Wanner, [18]). Because the selection for an ErmR transformant results in the mutant strain of interest, this is a relatively rapid means of generating targeted mutations in subsp. novicida. We do not yet know whether this technique will work in other Francisella subspecies; the LVS (subsp. holarctica) strain is naturally ErmR so ermC will not work as a selective marker in this strain, and we have not yet tested this method in subsp. tularensis (which is ErmS ). We attempted
4. Discussion F. tularensis subsp. novicida is closely related to the potential bioterrorist agent, F. tularensis subsp. tularensis [2]. This relationship is supported by our ability to design PCR primers based on the subsp. tularensis strain Schu4 genome sequence that accurately amplify the corresponding sequences in subsp. novicida. Although the novicida subspecies has extremely low virulence for humans, it can survive and replicate within macrophages in vitro, and has high levels of virulence and similar pathology to subsp. tularensis and subsp. holarctica in mice [3,13]. One of the few known Francisella virulence factors, IglC, was identi¢ed independently as being important for intramacrophage survival and growth in both subsp. novicida and subsp. holarctica (LVS) [9,10]. Thus the di¡erences in the
FEMSLE 11299 8-12-03
Fig. 4. Tul4 Western immunoblot of F. tularensis subsp. novicida mutants. Equivalent amounts of whole cell lysates of wild-type, vmglA: :ermC, viglC: :ermC, vtul4: :ermC, and vbla: :ermC strains were separated by 15% SDS^polyacrylamide gel electrophoresis, transferred to nitrocellulose membrane, and probed with KTul4.
Cyaan Magenta Geel Zwart
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
to use either TetR or KanR as a selective marker (either of which could be used in LVS) in subsp. novicida, but were unsuccessful, probably due to poor expression of the antibiotic resistance genes. Because Francisella spp. are naturally AmpR , we wished to create an AmpS subsp. novicida strain, since this would then allow L-lactamase to be used as a selective marker in this strain. However, the vbla: :ermC strain we created remained AmpR , suggesting additional L-lactamase gene(s) within the genome. In fact, we have recently identi¢ed an additional putative L-lactamase gene in Contig 32 of the partial subsp. tularensis genome sequence. Thus ErmR is currently the only reliable selectable antibiotic resistance marker for targeted chromosomal mutations in subsp. novicida. Since erythromycin is not recommended for therapeutic use in human tularemia, this marker should be useful for the generation of subsp. tularensis mutants. The vmglA: :ermC and viglC: :ermC subsp. novicida mutants had the anticipated phenotypes of failure to survive and grow within cultured macrophages [8^10]. These mutants will be useful in dissecting the roles these virulence factors play in Francisella virulence. The function of Tul4 had not previously been described, but this protein is a T-cell-reactive antigen and thus might have played a role in virulence [26]. However, our results with the vtul4: :ermC mutant indicate that this protein is not necessary for intramacrophage survival, although it appears to enhance growth within the macrophage. With the method of gene inactivation using PCR products described here, we can continue to perform targeted mutagenesis of speci¢c genes in F. tularensis subsp. novicida, and identify additional factors important for virulence.
Acknowledgements We thank Ake Forsberg and Mats Forsman for kindly providing plasmid pKK214 and KTul4 antisera, Rob Edwards for assistance with the F. tularensis genome sequence, and Erin Raulie for assistance with macrophage assays. Preliminary sequence data was obtained from the Francisella tularensis strain Schu4 genome sequencing consortium. This study was supported by National Institutes of Health grant AI50564 to K.E.K.
References [1] Ellis, J., Oyston, P.C.F., Green, M. and Titball, R. (2002) Tularemia. Clin. Microbiol. Rev. 15, 631^646. [2] Broekhuijsen, M., Larsson, P., Johansson, A., Bystrom, M., Eriksson, U., Larsson, E., Prior, R.G., Sjostedt, A., Titball, R. and Forsman, M. (2003) Genome-wide DNA Microarray Analysis of Francisella tularensis Strains Demonstrates Extensive Genetic Conservation within the Species but Identi¢es Regions that are Unique to the Highly Virulent F. tularensis subsp. tularensis. J. Clin. Microbiol. 41, 2924^2931.
FEMSLE 11299 8-12-03
201
[3] Anthony, L.S.D., Burke, R.D. and Nano, F.E. (1991) Growth of Francisella spp. in rodent macrophages. Infect. Immun. 59, 3291^ 3296. [4] Tarnvik, A. (1989) Nature of protective immunity to Francisella tularensis. Rev. Infect. Dis. 11, 440^451. [5] Fortier, A.H., Green, S.J., Polsinelli, T., Jones, T.R., Crawford, R.M., Leiby, D.A., Elkins, K.L., Meltzer, M.S. and Nacy, C.A. (1994) Life and death of an intracellular pathogen: Francisella tularensis and the macrophage. Immunol. Ser. 60, 349^361. [6] Lai, X.H., Golovliov, I. and Sjostedt, A. (2001) Francisella tularensis induces cytopathogenicity and apoptosis in murine macrophages via a mechanism that requires intracellular bacterial multiplication. Infect. Immun. 69, 4691^4694. [7] Telepnev, M., Golovliov, I., Grundstrom, T., Tarnvik, A. and Sjostedt, A. (2003) Francisella tularensis inhibits Toll-like receptor-mediated activation of intracellular signalling and secretion of TNF-alpha and IL-1 from murine macrophages. Cell. Microbiol. 5, 41^51. [8] Baron, G.S. and Nano, F.E. (1998) MglA and MglB are required for the intramacrophage growth of Francisella novicida. Mol. Microbiol. 29, 247^259. [9] Gray, C.G., Cowley, S.C., Cheung, K.K. and Nano, F.E. (2002) The identi¢cation of ¢ve genetic loci of Francisella novicida associated with intracellular growth. FEMS Microbiol. Lett. 215, 53^56. [10] Golovliov, I., Sjostedt, A., Mokrievich, A.N. and Pavlov, V.M. (2003) A method for allelic replacement in Francisella tularensis. FEMS Microbiol. Lett. 222, 273^280. [11] Prior, R.G., Klasson, L., Larsson, P., Williams, K., Lindler, L., Sjostedt, A., Svensson, T., Tamas, I., Wren, B.W., Oyston, P.C.F., Andersson, S.G.E. and Titball, R. (2001) Preliminary analysis and annotation of the partial genome sequence of Francisella tularensis strain Schu4. J. Appl. Microbiol. 91, 614^620. [12] Elkins, K.L., Rhinehart-Jones, T., Culkin, S.J., Yee, D. and Winegar, R.K. (1996) Minimal requirements for murine resistance to infection with Francisella tularensis LVS. Infect. Immun. 64, 3288^3293. [13] Kie¡er, T.L., Cowley, S.C., Nano, F.E. and Elkins, K.L. (2003) Francisella novicida LPS has greater immunobiological activity in mice than F. tularensis LPS, and contributes to F. novicida murine pathogenesis. Microbes Infect. 5, 397^403. [14] Hanahan, D. (1983) Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166, 577^580. [15] Wang, R.F. and Kushner, S. (1991) Construction of versatile lowcopy-number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 100, 195^199. [16] Monod, M., Denoya, C. and Dubnau, D. (1986) Sequence and properties of pIM13, a macrolide-lincosamide-streptogramin B resistance plasmid from Bacillus subtilis. J. Bacteriol. 167, 138^147. [17] Rose, R.E. (1988) The nucleotide sequence of pACYC184. Nucleic Acid Res. 16, 355. [18] Datsenko, K.A. and Wanner, B.L. (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97, 6640^6645. [19] Kuoppa, K., Forsberg, A. and Norqvist, A. (2001) Construction of a reporter plasmid for screening in vivo promoter activity in Francisella tularensis. FEMS Microbiol. Lett. 205, 77^81. [20] Norqvist, A., Kuoppa, K. and Sandstrom, G. (1996) Construction of a shuttle vector for use in Francisella tularensis. FEMS Immunol. Med. Microbiol. 13, 257^260. [21] Way, J.C., Davis, M.A., Morisato, D., Roberts, D.E. and Kleckner, N. (1984) New Tn10 derivatives for transposon mutagenesis and for construction of lacZ operon fusions by transposition. Gene 32, 369^ 379. [22] Haas, R., Kahrs, A.F., Facius, D., Allmeier, H., Schmitt, R. and Meyer, T.F. (1993) TnMax - a versatile mini-transposon for the analysis of cloned genes and shuttle mutagenesis. Gene 130, 23^31. [23] Hansen, A.M., Lehnherr, H., Wang, X., Mobley, V. and Jin, D.J. (2003) Escherichia coli SspA is a transcription activator for bacteriophage P1 late genes. Mol. Microbiol. 48, 1621^1631.
Cyaan Magenta Geel Zwart
202
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
[24] Sjostedt, A., Sandstrom, G. and Tarnvik, A. (1992) Humoral and cell-mediated immunity in mice to a 17-kilodalton lipoprotein of Francisella tularensis expressed by Salmonella typhimurium. Infect. Immun. 60, 2855^2862. [25] Golovliov, I., Ericsson, M., Sandstrom, G., Tarnvik, A. and Sjostedt, A. (1997) Identi¢cation of proteins of Francisella tularensis induced during growth in macrophages and cloning of the gene encoding a prominently induced 23-kilodalton protein. Infect. Immun. 65, 2183^ 2189.
FEMSLE 11299 8-12-03
[26] Sjostedt, A., Sandstrom, G., Tarnvik, A. and Jaurin, B. (1990) Nucleotide sequence and T cell epitopes of a membrane protein of Francisella tularensis. J. Immunol. 145, 311^317. [27] Seoane, A. and Garcia Lobo, J.M. (1991) Nucleotide sequence of a new class A beta-lactamase gene from the chromosome of Yersinia enterocolitica : implications for the evolution of class A beta-lactamases. Mol. Gen. Genet. 228, 215^220.
Cyaan Magenta Geel Zwart
www.fems-microbiology.org
Allelic exchange in Francisella tularensis using PCR products Crystal M. Lauriano a , Je¡rey R. Barker a , Francis E. Nano b , Bernard P. Arulanandam c , Karl E. Klose a; a
Department of Microbiology and Immunology, University of Texas Health Science Center, San Antonio, TX 78229-3900, USA b Department of Biochemistry and Microbiology, University of Victoria, Victoria, BC, Canada c Department of Biology, University of Texas San Antonio, San Antonio, TX, USA Received 5 August 2003 ; received in revised form 1 October 2003; accepted 28 October 2003 First published online 20 November 2003
Abstract We describe here a technique for allelic exchange in Francisella tularensis subsp. novicida utilizing polymerase chain reaction (PCR) products. Linear PCR fragments containing gene deletions with an erythromycin resistance cassette insertion were transformed into F. tularensis. The subsequent ErmR progeny were found to have undergone allelic exchange at the correct location in the genome; the minimum flanking homology necessary was 500 bp. This technique was used to create mglA, iglC, bla, and tul4 mutants in F. tularensis subsp. novicida strains. The mglA and iglC mutants were defective for intramacrophage growth, and the tul4 mutant lacked detectable Tul4 by Western immunoblot, as expected. Interestingly, the bla mutant maintained resistance to ampicillin, indicating the presence of multiple ampicillin resistance genes in F. tularensis. 7 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. Keywords : Tularemia ; Francisella tularensis ; Mutagenesis
1. Introduction Francisella tularensis, a Gram-negative coccobacillus, can cause the life-threatening disease tularemia in humans. It can be acquired through skin abrasions, ingestion of contaminated food or water, by a vector-borne route (e.g. ticks), or most seriously, via aerosols. Aerosol administration can lead to the pneumonic form of the disease, which has a high mortality rate. Because of this, F. tularensis was developed as a bioweapon in several national weapons programs. There is currently no approved vaccine available for protection against this disease. F. tularensis is classi¢ed as one of the most dangerous bioweapons (Category A) by the CDC (for review, see [1]). Several di¡erent subspecies of F. tularensis have been identi¢ed, including subspecies tularensis (Type A) and subspecies holarctica (Type B). Subsp. tularensis is the most virulent form of F. tularensis because it is associated with the highest morbidity and mortality in humans, and
* Corresponding author. Tel. : +1 (210) 567 3990; Fax : +1 (210) 567 9231. E-mail address : [email protected] (K.E. Klose).
also exhibits the highest virulence by all routes of administration in animal models, while subsp. holarctica causes infections that are rarely fatal in humans. An additional subspecies, novicida, is essentially avirulent for humans, but causes a disease in some animal models that is virtually indistinguishable from that caused by subsp. tularensis. All the subspecies of F. tularensis are closely related at the genomic level, and the reasons for the di¡erences in tropism and virulence are yet to be determined [2]. Relatively little is known about molecular mechanisms of Francisella pathogenesis. F. tularensis can survive and replicate within macrophages both in vitro and in vivo, and this appears to be its primary mechanism of pathogenesis [3,4]. The Francisella-containing vacuole evades phagosome^lysosome fusion [5], and Francisella infection can cause macrophage apoptosis [6] and abrogate Toll-like receptor signaling [7]. Several genes necessary for intramacrophage survival have been identi¢ed, including mglAB [8] and iglC [9,10], but their function(s) have not yet been elucidated. No Type III or Type IV secretion systems that may be responsible for altering the intramacrophage environment have been identi¢ed in the partially completed genome sequence of a subsp. tularensis strain [11]. The current knowledge of Francisella pathogenesis is
0378-1097 / 03 / $22.00 7 2003 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/S0378-1097(03)00820-6
FEMSLE 11299 8-12-03
Cyaan Magenta Geel Zwart
196
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
based upon studies utilizing either an attenuated (for humans) subsp. holarctica ‘live vaccine strain’ (LVS), or a wild-type subsp. novicida strain, due to the relative ease of working with these organisms under less stringent biosafety conditions. The attenuating mutation(s) in the LVS strain are unde¢ned, and this strain demonstrates attenuated virulence by the intradermal route but not the intraperitoneal route in mice [12]. The wild-type subsp. novicida strain U112 is more virulent by the intradermal and intranasal routes than LVS in mice ([13] and KEK and BPA, unpublished data), and thus is more similar to subsp. tularensis in this animal model. The subsp. novicida strain has also proven to be more easily manipulated genetically than LVS, making this an attractive model for studying Francisella pathogenesis. Very few genetic tools exist with which to dissect virulence mechanisms of Francisella. The mutant strains that have been generated in subsp. novicida have been constructed by random transposon mutagenesis [8,9]. While this technique has been invaluable in identifying Francisella virulence genes, we have observed that transposon insertions are highly unstable in subsp. novicida, despite the lack of a transposase gene within the inserted transposon. Thus we have developed a technique to mutagenize speci¢c genes within subsp. novicida using polymerase chain reaction (PCR) products. We have used this technique to successfully create mutations in four di¡erent Francisella genes.
2. Materials and methods 2.1. Strains and media F. tularensis subsp. novicida strain U112 (ATCC) was
grown on tryptic soy broth/agar (Difco) supplemented with 0.1% cysteine (TSAcys) or on cysteine heart agar (Difco) supplemented with 5% horse blood (CHA). Escherichia coli strain DH5K [14] was used for cloning, and grown on Luria broth (Difco). The concentrations of antibiotics used, when appropriate, were: erythromycin 250 Wg ml31 , kanamycin 50 Wg ml31 , and tetracycline 10 Wg ml31 . 2.2. Plasmid construction The same general strategy was used to create all deletion/insertion constructs. A V500^600-bp fragment corresponding to the region 5P of the deletion was PCR ampli¢ed from U112 chromosomal DNA with primers 1 and 2 (Table 1), as was a V500^600-bp fragment corresponding to the region 3P of the deletion with primers 3 and 4. Primers 2 and 3 share a complementary sequence that overlaps the deletion to be created and also contains EcoRI and PstI restriction sites. The two PCR-ampli¢ed fragments were then annealed and used as a template in a second PCR reaction with primers 1 and 4, which resulted in ampli¢cation of the deletion of interest (which contains EcoRI and PstI sites) with 500^600 bp of £anking homology. Primers 1 and 4 typically contained BamHI and HindIII or SalI sites, respectively, so the ampli¢ed PCR product would then be cleaved with these two enzymes and ligated into pWSK30 [15] digested similarly. The ermC gene was PCR ampli¢ed from pIM13 [16] with primers ermC1 and ermC2 that contain EcoRI and PstI sites, and the resulting fragment was digested with these enzymes and ligated into the plasmid containing the deletion at the corresponding sites. This resulted in the construction of (vgene of interest : :ermC). The vmglA: :TetR construct was obtained by PCR ampli¢cation of the tet gene from pACYC184 [17] with
Table 1 Oligonucleotides used in this study MglA1 MglA2 MglA3 MglA4 IglC1 IglC2 IglC3 IglC4 Tul4-1 Tul4-2 Tul4-3 Tul4-4 Bla1 Bla2 Bla3 Bla4 ErmC1 ErmC2 KD46-1 KD46-2
5P-GCGGATCCGAGCTCTAATCGATGATGATG-3P 5P-CTTTTGCTTTCTGCAGGCGAATTCATTAACATATTCTAACAATATTC-3P 5P-ATATGTTAATGAATTCGCCTGCAGAAAGCAAAAGGAGCTTAATATG-3P 5P-GCGCAAGCTTATACCATTGAGTTTTATAGTAG-3P 5P-GCGGATCCTACCATTTGTAAGTGAAAAAGG-3P 5P-CAATATATCCCTGCAGCGCGAATTCCTGTTGTCTTGTTATCATCTC-3P 5P-AAGACAACAGGAATTCGCGCTGCAGGGATATATTGCAGCTGCATAG-3P 5P-GCGCAAGCTTTTAAGAGTTAATAGCGGGAAG-3P 5P-GCGGATCCGTTAACTATGTTATTATAATCAAG-3P 5P-TTACTTCTTTCTGCAGCGCGAATTCTAAAGATAAAAGACTAAGCTC-3P 5P-TTTATCTTTAGAATTCGCGCTGCAGAAAGAAGTAATCGCTTCTGATTC-3P 5P-GCGCGTCGACTTATTTTCTCATCAATCTTAACC-3P 5P-GCGGATCCAAACATCAAAAAGTATGCTAC-3P 5P-TCGTAGTATTCTGCAGCGCGAATTCCAGCGTAATTATTTTTTTCATATTC-3P 5P-AATTACGCTGGAATTCGCGCTGCAGAATACTACGAAAAATGCCTAAAAT-3P 5P-GCGCGTCGACATCAATAGCACTATCTATATC-3P 5P-GCGAATTCACGCGTTGGCTAACACACACGCCATTC-3P 5P-GCGGATCCCTGCAGTTACTTATTAAATAATTTATAGC-3P 5P-GCTCTAGATTATTATGACAACTTGACGGC-3P 5P-GCTCTAGAATTCTTCGTCTGTTTCTACTG-3P
FEMSLE 11299 8-12-03
Cyaan Magenta Geel Zwart
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
primers that contained EcoRI and PstI sites, with subsequent ligation into pKEK597 (contains vmglA). The vmglA: :KanR construct was obtained by ligation of the PstI fragment of pUC4K (Pharmacia) containing KanR from Tn903 into the PstI site of pKEK597. Sequence of the subsp. novicida mglAB locus was already available [8]. The sequences of the tul4, iglC, and bla loci were obtained from the F. tularensis strain Schu4 genome sequencing consortium (http://artedi.ebc.uu.se/ Projects/Francisella/). The locations of the genes were: tul4, contig 31, nucleotides 21 972^22 421; iglC, contig 36, nucleotides 45 440^46 069, and bla, contig 33, nucleotides 122 464^123 348. The deletion constructs removed coding sequence corresponding to amino acids 101^200 of MglA, amino acids 12^205 of IglC, amino acids 13^ 139 of Tul4, and amino acids 8^308 of L-lactamase. All constructs were veri¢ed by DNA sequencing. The VRed recombinase plasmid that replicates in F. tularensis was constructed by ¢rst PCR amplifying the araCPBAD -QLexo fragment from plasmid pKD46 [18] with primers KD46-1 and KD46-2 that incorporated XbaI restriction sites. This fragment was subsequently digested with XbaI and ligated into the Francisella plasmid pKK214 [19] that had been digested similarly to form pKEK612. 2.3. Transformation Deletion/insertions (constructed as described above) were PCR ampli¢ed utilizing KOD HiFi DNA polymerase (Novagen), with the exception of the long-range vmglA: :ermC PCR product described in Section 3, which was ampli¢ed with KOD XL DNA polymerase (Novagen). The PCR products were transformed into strain U112 via a cryotransformation technique [20], which was modi¢ed in the following manner. Brie£y, 2 Wg of PCR product was added to U112 cells at an OD600 of 65^70 in transformation bu¡er [20], incubated at room temperature for 10 min, then £ash frozen in liquid nitrogen for 5 min. Cells were then thawed at room temperature for 5 min, and placed on 0.2 W nitrocellulose ¢lter paper on CHA for 4^6 h at 37‡C. The bacteria on ¢lter paper were then resuspended in 0.2 M KCl and plated on TSAcys containing 250 Wg ml31 erythromycin and incubated at 37‡C. This resulted in the creation of strains KKF34 (vmglA: :ermC), KKF24 (viglC: :ermC), KKF29 (vtul4: :ermC), and KKF31 (vbla: :ermC). 2.4. Southern hybridization and Western immunoblot Southern blot was performed utilizing ECL detection reagent (Amersham Pharmacia), according to manufacturer’s directions. Whole cell lysates were separated by 15% sodium dodecyl sulfate (SDS)^polyacrylamide gel electrophoresis and probed by Western immunoblot with rabbit
FEMSLE 11299 8-12-03
197
KTul4 (kindly provided by M. Forsman), and detected utilizing ECL detection reagent (Amersham Pharmacia). 2.5. Intramacrophage survival assay Wild-type or mutant F. tularensis subsp. novicida strains were used to infect the J774.1 macrophage cell line (ATCC) at an MOI of V1:1. Wells were seeded with V105 J774 cells, and bacterial inocula ranged from 1.56^5.8U105 . After 1 h incubation at 37‡C, 5% CO2 , gentamicin (50 Wg ml31 ) was added to the medium to eliminate extracellular organisms. The macrophage cells were lysed with 0.2% deoxycholate 24 h post infection, and the lysate plated on TSAcys, incubated at 37‡C, and CFU enumerated.
3. Results 3.1. Transposon insertions are unstable in F. tularensis subsp. novicida F. tularensis subsp. novicida strain GB5 [8] was constructed by random mutagenesis with mTn10Km [21] into a plasmid containing the mglAB locus. A clone with a transposon insertion in mglA was subsequently selected and recombined into the F. tularensis subsp. novicida U112 genome and con¢rmed by Southern blot immediately after isolation. When utilizing strain GB5 as a control in Southern blot and PCR ampli¢cation, we were surprised that this strain appeared to lack an insertion element within the mglA gene, even though the strain maintains resistance to kanamycin (data not shown). We ampli¢ed and sequenced the mglA coding region from GB5, and found a 52-bp Tn10 excision scar at the location of the original transposon insertion, indicating that the transposon had hopped from its original location. Southern blot con¢rmed that the mTn10Km transposon is located elsewhere within the genome of GB5 (data not shown). The GB5 strain has undergone minimal passage (approximately three times) since its initial construction. Similar instability of other mTn10Km insertions in other F. tularensis subsp. novicida genes has since been observed (FEN, unpublished data). Given that this mini transposon lacks a transposase gene, this suggests a complementing transposase activity resident in the F. tularensis subsp. novicida genome. Insertion mutations constructed with the TnMax2 [22] transposon (a mini Tn1721 derivative) in subsp. novicida also appear to be unstable (R. Titball, personal communication). Multiple and various insertion elements with transposase genes are found in the partially completed F. tularensis subsp. tularensis genome [11]. Given the close genetic relationship between the subspecies [2], transposon instability may be a general phenomenon of Francisella spp. This led us to develop an alternative meth-
Cyaan Magenta Geel Zwart
198
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
od for targeted allelic replacement in F. tularensis subsp. novicida. 3.2. A method for allelic replacement in F. tularensis subsp. novicida using PCR products We chose the mglA gene as the target for allelic replacement, because this is one of the few known Francisella virulence genes. MglA is necessary for intramacrophage survival [8], and is a homolog of E. coli SspA, suggesting it functions as an activator of transcription [23]. A deletion of the mglA gene was constructed by a two-step PCR technique; the resulting vmglA construct was cloned into pWSK30 [15]. The deletion spans the residues encoding MglA amino acids 101^200 with 500-bp £anking homology. The ermC gene encoding erythromycin resistance [16] was ligated into the vmglA construct to create vmglA: :ermC. The vmglA: :ermC construct was PCR ampli¢ed, and the PCR product was used to transform F. tularensis subsp. novicida strain U112 via a cryotransformation technique, with selection for ErmR progeny. Two transformants were recovered. The two ErmR colonies from this initial transformation were further analyzed by Southern blot and PCR (Fig. 1). Southern analysis of the ErmR strains, utilizing a mglAP probe (upstream of the deletion/insertion) was consistent with the predicted vmglA: :ermC allelic replacement in the chromosome of both strains, resulting in a V4-kb HindIII fragment, compared to the V3-kb fragment of the parent strain. Likewise, long-range PCR, utilizing primers V5 kb upstream and downstream of mglA, and subsequent restriction analysis with EcoRI (which cuts within the ermC insertion, but not within the wild-type mglA locus), was also consistent with correct allelic replacement. The mglA gene of the long-range PCR fragment was sequenced, con¢rming the presence of the vmglA: :ermC allele (data not shown). 3.3. Length requirement of £anking homology/e¡ect of VRed recombinase on recombination
Fig. 1. Southern blot/PCR analysis of vmglA : :ermC mutant. A: The anticipated mglAB chromosomal loci in the wild-type and vmglA : :ermC mutant strains are depicted, including location of mglAP probe (‘probe’) used in Southern analysis, and position of primers used in long-range PCR. The vmglA: :ermC mutation is predicted to increase the size of the HindIII fragment containing mglAP by 1 kb, and to introduce an EcoRI site into the V10-kb long-range PCR product. B: Southern blot of HindIII digest of DNA from U112 (‘WT’; lane 1, V3-kb product) and vmglA : :ermC mutants (lanes 2 and 3, V4-kb products) utilizing mglAP probe. Migration of V/HindIII digestion is depicted on left, with 4.3- and 2.3-kb fragments noted. C: Long-range PCR products derived from U112 (‘WT’; lane 2, product V10 kb) and vmglA: :ermC mutant (lane 3, products V4 and V6 kb), which were digested with EcoRI and separated by agarose electrophoresis. One-kb ladder (Gibco BRL) is in lane 1, with 4- and 6-kb fragments noted.
Wanner and colleagues have developed a method for allelic replacement in E. coli, utilizing PCR fragments containing an antibiotic resistance cassette with only 50-bp £anking homology, that requires expression of VRed recombinase in the recipient strain [18]. We adapted this method to Francisella by creating a plasmid, pKEK612, with the Francisella-speci¢c origin of replication and the VRed recombinase gene under control of the arabinoseinducible promoter PBAD . We PCR ampli¢ed the ermC gene with primers containing 50-bp mglA £anking homology on either side, and transformed into F. tularensis subsp. novicida U112 either carrying pKEK612, or the vector pKK214, both were grown in the presence of arabinose. In the same experi-
ment, we transformed the same strains with PCR fragments containing the ermC cassette with either 500-bp mglA £anking homology (described above), or V5000bp mglA £anking homology (the long-range PCR fragment containing vmglA: :ermC, also described above). The results are shown in Table 2. No ErmR colonies were obtained with only 50-bp £anking homology, and only two ErmR colonies were obtained with 500-bp £anking homology, but the presence of V5000-bp £anking homology dramatically increased the number of ErmR colonies (V1000-fold increase). The presence of the VRed recombinase plasmid in the recipient strain had no e¡ect on number of ErmR colonies obtained, suggesting VRed
FEMSLE 11299 8-12-03
Cyaan Magenta Geel Zwart
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
199
recombinase either fails to promote recombination in F. tularensis, or that it was poorly expressed from the PBAD promoter in Francisella. We have not distinguished between these two possibilities. We constructed vmglA: :Kan and vmglA: :Tet alleles; these alleles contain an identical 500-bp £anking homology to the vmglA: :ermC allele described above. PCR products with these alleles were transformed into strain U112 in an identical manner to that used with the vmglA: :ermC fragment, but no KanR or TetR colonies were obtained, despite multiple attempts (more than four times each). Although the TetR gene used was the same as that carried on plasmid pKK214 [19], which results in TetR when carried by strain U112, we suspect the promoter is poorly expressed in single copy in Francisella. 3.4. Generation of iglC, tul4, and bla F. tularensis subsp. novicida mutants IglC has been shown to be important for intramacrophage survival both in subsp. novicida and subsp. holarctica (LVS) [9,10], while Tul4 was identi¢ed as a T-cellreactive protein with unknown function [24]. Francisella spp. are naturally ampicillin resistant, so the presence of a L-lactamase gene (bla) within the genome was predicted. We identi¢ed these open reading frames (ORFs) and the surrounding genes within the F. tularensis subsp. tularensis partial genome sequence (Schu4 genome sequencing consortium; http://artedi.ebc.uu.se/Projects/Francisella/). The coding sequences for iglC and tul4 have been previously cloned from LVS [25,26]. The bla gene was found in contig 33 (nucleotides 122 464^123 348), and the deduced ORF is 880 bp long, with 59% similarity with a class A L-lactamase of Yersinia enterocolitica [27]. Primers were designed to create in-frame deletions in iglC, tul4, and bla, as described for mglA. Due to the extremely high level of genetic relatedness between subsp. tularensis and subsp. novicida, it was possible to design primers based on the subsp. tularensis genome sequence and faithfully amplify the corresponding subsp. novicida genes in every case. viglC: :ermC, vtul4: :ermC, and vbla: :ermC alleles were constructed and PCR ampli¢ed, as outlined above. The £anking homology was V500^600 bp in each case, and the number of ErmR transformants
Table 2 E¡ect of £anking homology length and presence of VRed recombinase plasmid on recombination frequency between PCR products and F. tularensis subsp. novicida chromosome Flanking homology (bp)a
50 500 5000 a
# Transformants/recipient strain U112
U112/pKEK612
0 3 3556
0 2 2464
Transformations were performed as described in Section 2.
FEMSLE 11299 8-12-03
Fig. 2. PCR analysis of viglC: :ermC, vtul4: :ermC, and vbla: :ermC mutants. PCR analysis was performed with U112 (‘WT’) and mutant strains, each utilizing the same outward-facing primer that anneals to ermC (‘ermC primer+’) paired with a speci¢c primer that anneals downstream of the insertion site, but outside of £anking homology used to create mutation (‘iglC’, ‘tul4’, ‘bla’ primers). PCR products were separated by 1% agarose electrophoresis and stained with ethidium bromide. One-kb DNA ladder (Gibco BRL) is shown in lane 1, with the 1- and 2-kb fragments noted. With each primer pair, the wild-type strain failed to produce a PCR product (‘WT’, lanes 2, 4, 6), but the mutant strains produced the anticipated PCR product consistent with allelic replacement : 1632 bp for viglC: :ermC (‘iglC’, lane 3), 940 bp for vtul4: :ermC (‘tul4’, lane 5) and 769 bp for vbla: :ermC (‘bla’, lane 7).
for each PCR product into strain U112 was between 2 and 10 colonies. We con¢rmed the correct chromosomal allelic replacement in the resulting viglC : :ermC, vtul4: :ermC, and vbla: :ermC F. tularensis subsp. novicida strains by PCR and Southern blot. Utilizing an outward-facing primer that anneals to the ermC gene paired with a downstream primer that anneals outside of the £anking homology used to create the allelic replacement, it was clear that the correct chromosomal allelic replacement had occurred in all three strains (Fig. 2). Southern blot con¢rmed that a single copy of the ermC cassette had been incorporated into the genome of the three mutant strains (data not shown). 3.5. Characterization of the mglA, iglC, tul4, and bla F. tularensis subsp. novicida strains An intramacrophage survival assay in the murine macrophage cell line J774 was performed with the wild-type and all four mutant strains (Fig. 3). After 24 h, no iglC or mglA F. tularensis subsp. novicida were recovered from the macrophage cells, in contrast to the high numbers of the wild-type U112 strain recovered. This assay demonstrated that the vmglA: :ermC and viglC: :ermC strains had the anticipated defect in intramacrophage growth. In contrast, the vtul4: :ermC and vbla: :ermC strains were recovered at relatively high numbers from the macrophage cell line 24 h
Cyaan Magenta Geel Zwart
200
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
Fig. 3. Intramacrophage survival of F. tularensis subsp. novicida mutants. J774 cells were infected with wild-type and mutant strains, and intracellular bacteria were enumerated 24 h post-infection (Section 2). Infections were performed in triplicate for each strain, with standard deviations depicted by error bars. Bacterial inocula were in the range 1.56^5.8U105 . No bacteria were recovered for vmglA: :ermC and viglC : :ermC infections.
p.i. The amount recovered of the vtul4: :ermC mutant was approximately 10-fold lower than for the wild-type U112 strain. Clearly Tul4 and L-lactamase are not essential for survival and growth within J774 cells, but Tul4 may be required for full resistance to macrophage-killing mechanisms. Whole cell lysates of all four mutant strains were tested for the presence of Tul4 by Western immunoblot with K-Tul4 antisera (Fig. 4). The vtul4: :ermC mutant strain showed a complete lack of Tul4, while the other mutant strains resembled the wild-type strain in levels of Tul4 present. The vbla: :ermC strain was still able to grow in the presence of ampicillin at concentrations as high as 1 mg ml31 , similar to the wild-type strain (data not shown), indicating that additional ampicillin resistance gene(s) are present within the genome.
subspecies appears to lie primarily in host tropism, and not in fundamental di¡erences in pathogenic mechanisms, and therefore subsp. novicida infection of mice can be used as a model for subsp. tularensis infections. This has the advantages of (1) obviating the need for high-level biocontainment necessary for subsp. tularensis, (2) utilizing a fully virulent wild-type strain (vs. the unknown attenuating mutation(s) of the subsp. holarctica LVS), and (3) working with the subspecies that has proven to be the most amenable to genetic manipulation. The subsp. novicida mutants that have been constructed thus far (aside from spontaneous mutants) have been random transposon insertion mutants, utilizing mTnKm and TnMax2 [8,9]. As described here, mTnKm insertions are unstable in subsp. novicida, suggesting the presence of transposase(s) within the subsp. novicida genome that can act on mTnKm. This prompted us to develop an alternative method for targeted gene inactivation in subsp. novicida described here that would result in stable mutations. A recent report has described a method for targeted gene inactivation in LVS via conjugation [10] ; hopefully this method will also be successful in subsp. novicida. The allelic exchange method described here utilizes linear DNA for transformation, which allows one to select for a double recombination in the gene of interest, and therefore there is no need for subsequent selection or screening. We determined the minimum £anking homology necessary for this method is 500 bp, which prevents the use of the £anking homology being incorporated directly into the PCR primers (as can be done with the PCRbased method of Wanner, [18]). Because the selection for an ErmR transformant results in the mutant strain of interest, this is a relatively rapid means of generating targeted mutations in subsp. novicida. We do not yet know whether this technique will work in other Francisella subspecies; the LVS (subsp. holarctica) strain is naturally ErmR so ermC will not work as a selective marker in this strain, and we have not yet tested this method in subsp. tularensis (which is ErmS ). We attempted
4. Discussion F. tularensis subsp. novicida is closely related to the potential bioterrorist agent, F. tularensis subsp. tularensis [2]. This relationship is supported by our ability to design PCR primers based on the subsp. tularensis strain Schu4 genome sequence that accurately amplify the corresponding sequences in subsp. novicida. Although the novicida subspecies has extremely low virulence for humans, it can survive and replicate within macrophages in vitro, and has high levels of virulence and similar pathology to subsp. tularensis and subsp. holarctica in mice [3,13]. One of the few known Francisella virulence factors, IglC, was identi¢ed independently as being important for intramacrophage survival and growth in both subsp. novicida and subsp. holarctica (LVS) [9,10]. Thus the di¡erences in the
FEMSLE 11299 8-12-03
Fig. 4. Tul4 Western immunoblot of F. tularensis subsp. novicida mutants. Equivalent amounts of whole cell lysates of wild-type, vmglA: :ermC, viglC: :ermC, vtul4: :ermC, and vbla: :ermC strains were separated by 15% SDS^polyacrylamide gel electrophoresis, transferred to nitrocellulose membrane, and probed with KTul4.
Cyaan Magenta Geel Zwart
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
to use either TetR or KanR as a selective marker (either of which could be used in LVS) in subsp. novicida, but were unsuccessful, probably due to poor expression of the antibiotic resistance genes. Because Francisella spp. are naturally AmpR , we wished to create an AmpS subsp. novicida strain, since this would then allow L-lactamase to be used as a selective marker in this strain. However, the vbla: :ermC strain we created remained AmpR , suggesting additional L-lactamase gene(s) within the genome. In fact, we have recently identi¢ed an additional putative L-lactamase gene in Contig 32 of the partial subsp. tularensis genome sequence. Thus ErmR is currently the only reliable selectable antibiotic resistance marker for targeted chromosomal mutations in subsp. novicida. Since erythromycin is not recommended for therapeutic use in human tularemia, this marker should be useful for the generation of subsp. tularensis mutants. The vmglA: :ermC and viglC: :ermC subsp. novicida mutants had the anticipated phenotypes of failure to survive and grow within cultured macrophages [8^10]. These mutants will be useful in dissecting the roles these virulence factors play in Francisella virulence. The function of Tul4 had not previously been described, but this protein is a T-cell-reactive antigen and thus might have played a role in virulence [26]. However, our results with the vtul4: :ermC mutant indicate that this protein is not necessary for intramacrophage survival, although it appears to enhance growth within the macrophage. With the method of gene inactivation using PCR products described here, we can continue to perform targeted mutagenesis of speci¢c genes in F. tularensis subsp. novicida, and identify additional factors important for virulence.
Acknowledgements We thank Ake Forsberg and Mats Forsman for kindly providing plasmid pKK214 and KTul4 antisera, Rob Edwards for assistance with the F. tularensis genome sequence, and Erin Raulie for assistance with macrophage assays. Preliminary sequence data was obtained from the Francisella tularensis strain Schu4 genome sequencing consortium. This study was supported by National Institutes of Health grant AI50564 to K.E.K.
References [1] Ellis, J., Oyston, P.C.F., Green, M. and Titball, R. (2002) Tularemia. Clin. Microbiol. Rev. 15, 631^646. [2] Broekhuijsen, M., Larsson, P., Johansson, A., Bystrom, M., Eriksson, U., Larsson, E., Prior, R.G., Sjostedt, A., Titball, R. and Forsman, M. (2003) Genome-wide DNA Microarray Analysis of Francisella tularensis Strains Demonstrates Extensive Genetic Conservation within the Species but Identi¢es Regions that are Unique to the Highly Virulent F. tularensis subsp. tularensis. J. Clin. Microbiol. 41, 2924^2931.
FEMSLE 11299 8-12-03
201
[3] Anthony, L.S.D., Burke, R.D. and Nano, F.E. (1991) Growth of Francisella spp. in rodent macrophages. Infect. Immun. 59, 3291^ 3296. [4] Tarnvik, A. (1989) Nature of protective immunity to Francisella tularensis. Rev. Infect. Dis. 11, 440^451. [5] Fortier, A.H., Green, S.J., Polsinelli, T., Jones, T.R., Crawford, R.M., Leiby, D.A., Elkins, K.L., Meltzer, M.S. and Nacy, C.A. (1994) Life and death of an intracellular pathogen: Francisella tularensis and the macrophage. Immunol. Ser. 60, 349^361. [6] Lai, X.H., Golovliov, I. and Sjostedt, A. (2001) Francisella tularensis induces cytopathogenicity and apoptosis in murine macrophages via a mechanism that requires intracellular bacterial multiplication. Infect. Immun. 69, 4691^4694. [7] Telepnev, M., Golovliov, I., Grundstrom, T., Tarnvik, A. and Sjostedt, A. (2003) Francisella tularensis inhibits Toll-like receptor-mediated activation of intracellular signalling and secretion of TNF-alpha and IL-1 from murine macrophages. Cell. Microbiol. 5, 41^51. [8] Baron, G.S. and Nano, F.E. (1998) MglA and MglB are required for the intramacrophage growth of Francisella novicida. Mol. Microbiol. 29, 247^259. [9] Gray, C.G., Cowley, S.C., Cheung, K.K. and Nano, F.E. (2002) The identi¢cation of ¢ve genetic loci of Francisella novicida associated with intracellular growth. FEMS Microbiol. Lett. 215, 53^56. [10] Golovliov, I., Sjostedt, A., Mokrievich, A.N. and Pavlov, V.M. (2003) A method for allelic replacement in Francisella tularensis. FEMS Microbiol. Lett. 222, 273^280. [11] Prior, R.G., Klasson, L., Larsson, P., Williams, K., Lindler, L., Sjostedt, A., Svensson, T., Tamas, I., Wren, B.W., Oyston, P.C.F., Andersson, S.G.E. and Titball, R. (2001) Preliminary analysis and annotation of the partial genome sequence of Francisella tularensis strain Schu4. J. Appl. Microbiol. 91, 614^620. [12] Elkins, K.L., Rhinehart-Jones, T., Culkin, S.J., Yee, D. and Winegar, R.K. (1996) Minimal requirements for murine resistance to infection with Francisella tularensis LVS. Infect. Immun. 64, 3288^3293. [13] Kie¡er, T.L., Cowley, S.C., Nano, F.E. and Elkins, K.L. (2003) Francisella novicida LPS has greater immunobiological activity in mice than F. tularensis LPS, and contributes to F. novicida murine pathogenesis. Microbes Infect. 5, 397^403. [14] Hanahan, D. (1983) Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166, 577^580. [15] Wang, R.F. and Kushner, S. (1991) Construction of versatile lowcopy-number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 100, 195^199. [16] Monod, M., Denoya, C. and Dubnau, D. (1986) Sequence and properties of pIM13, a macrolide-lincosamide-streptogramin B resistance plasmid from Bacillus subtilis. J. Bacteriol. 167, 138^147. [17] Rose, R.E. (1988) The nucleotide sequence of pACYC184. Nucleic Acid Res. 16, 355. [18] Datsenko, K.A. and Wanner, B.L. (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97, 6640^6645. [19] Kuoppa, K., Forsberg, A. and Norqvist, A. (2001) Construction of a reporter plasmid for screening in vivo promoter activity in Francisella tularensis. FEMS Microbiol. Lett. 205, 77^81. [20] Norqvist, A., Kuoppa, K. and Sandstrom, G. (1996) Construction of a shuttle vector for use in Francisella tularensis. FEMS Immunol. Med. Microbiol. 13, 257^260. [21] Way, J.C., Davis, M.A., Morisato, D., Roberts, D.E. and Kleckner, N. (1984) New Tn10 derivatives for transposon mutagenesis and for construction of lacZ operon fusions by transposition. Gene 32, 369^ 379. [22] Haas, R., Kahrs, A.F., Facius, D., Allmeier, H., Schmitt, R. and Meyer, T.F. (1993) TnMax - a versatile mini-transposon for the analysis of cloned genes and shuttle mutagenesis. Gene 130, 23^31. [23] Hansen, A.M., Lehnherr, H., Wang, X., Mobley, V. and Jin, D.J. (2003) Escherichia coli SspA is a transcription activator for bacteriophage P1 late genes. Mol. Microbiol. 48, 1621^1631.
Cyaan Magenta Geel Zwart
202
C.M. Lauriano et al. / FEMS Microbiology Letters 229 (2003) 195^202
[24] Sjostedt, A., Sandstrom, G. and Tarnvik, A. (1992) Humoral and cell-mediated immunity in mice to a 17-kilodalton lipoprotein of Francisella tularensis expressed by Salmonella typhimurium. Infect. Immun. 60, 2855^2862. [25] Golovliov, I., Ericsson, M., Sandstrom, G., Tarnvik, A. and Sjostedt, A. (1997) Identi¢cation of proteins of Francisella tularensis induced during growth in macrophages and cloning of the gene encoding a prominently induced 23-kilodalton protein. Infect. Immun. 65, 2183^ 2189.
FEMSLE 11299 8-12-03
[26] Sjostedt, A., Sandstrom, G., Tarnvik, A. and Jaurin, B. (1990) Nucleotide sequence and T cell epitopes of a membrane protein of Francisella tularensis. J. Immunol. 145, 311^317. [27] Seoane, A. and Garcia Lobo, J.M. (1991) Nucleotide sequence of a new class A beta-lactamase gene from the chromosome of Yersinia enterocolitica : implications for the evolution of class A beta-lactamases. Mol. Gen. Genet. 228, 215^220.
Cyaan Magenta Geel Zwart