|Home | About | Journals | Submit | Contact Us | Français|
The group A streptococcus (GAS) is a strict human pathogen responsible for a wide spectrum of diseases. Although GAS genome sequences are available, functional genomic analyses have been limited. We developed a mariner-based transposon, osKaR, designed to perform Transposon-Site Hybridization (TraSH) in GAS and successfully tested its use in several invasive serotypes. A complex osKaR mutant library in M1T1 GAS strain 5448 was subjected to negative selection in human blood to identify genes important for GAS fitness in this clinically relevant environment. Mutants underrepresented after growth in blood (output pool) compared to growth in rich media (input pool) were identified using DNA microarray hybridization of transposon-specific tags en masse. Using blood from three different donors, we identified 81 genes that met our criteria for reduced fitness in blood from at least two individuals. Genes known to play a role in survival of GAS in blood were found, including those encoding the virulence regulator Mga (mga), the peroxide response regulator PerR (perR), and the RofA-like regulator Ralp-3 (ralp3). We also identified genes previously reported for their contribution to sepsis in other pathogens, such as de novo nucleotide synthesis (purD, purA, pyrB, carA, carB, guaB), sugar metabolism (scrB, fruA), zinc uptake (adcC), and transcriptional regulation (cpsY). To validate our findings, independent mutants with mutations in 10 different genes identified in our screen were confirmed to be defective for survival in blood bactericidal assays. Overall, this work represents the first use of TraSH in GAS to identify potential virulence genes.
Streptococcus pyogenes (group A streptococcus; GAS) is an important human pathogen responsible for infecting over 750 million people around the world each year. Most of these infections are mild and target the upper respiratory tract (pharyngitis) or skin (impetigo); however, invasive GAS strains can gain access to deeper tissues, leading to life-threatening diseases (necrotizing fasciitis and toxic shock syndrome) that result in over 500,000 deaths worldwide (1). Since GAS strain diversity appears to be a key contributor to its broad pathogenesis, significant progress has been made over the last decade in analyzing GAS genome content. This has resulted in 17 completed GAS genome sequences representing 12 different M serotypes being publically available, with more studies in progress (2). Comparative genomics have revealed important information regarding GAS genome diversity and have also shed light upon the relationship between GAS tissue tropism and strain evolution and dissemination (3, 4).
The current challenge is to integrate the wealth of information provided by GAS genome sequences with its pathogenesis to establish a functional framework. Mutagenesis of individual genes and phenotypic screening remain the most common methods to identify the function of any particular gene. Typically, a gene is considered essential for cell adaptation or survival (fitness) in a particular environment if its loss through mutation is detrimental or lethal under the given conditions. Genetic manipulations and the use of different animal models of infection have been key to understanding GAS pathogenesis (5), and the field has advanced through the development and use of genome-wide transcriptomic and proteomic approaches (2). Unfortunately, phenotypic analysis of GAS mutant strains remains a time-consuming process and most studies are limited to a small number of genes that are often selected based on prior knowledge (reverse genetics). Furthermore, although expression analyses are able to reveal global regulons and stimulons, they fail to directly establish which regulated genes are important for bacterial survival in that particular environment.
Forward genetics approaches screen random mutants to identify those with an altered phenotype, and transposons have been the tool of choice for the generation of random mutant libraries. In GAS, there have been a number of forward genetic screens using transposon libraries that were screened for phenotypes important in virulence, including hemolytic activity (6, 7), protease activity (8, 9), capsule production (10), and global regulation (11, 12). Although effective, these initial studies required manual screening to identify mutants altered in the appropriate phenotype.
Transposon-based high-throughput technologies were developed to monitor larger pools of mutants for those unable to survive in a given environment such as the host. Transposon Site Hybridization (TraSH) (13) and related technologies were developed to perform en masse analysis of the output pool using DNA microarrays to rapidly identify mutants with reduced fitness in vivo (14, 15). TraSH has been quite effective, although it requires access to a relevant microarray. More recently, Tn-seq (16) and related techniques were developed that use massively parallel deep sequencing to precisely identify transposon insertion points within mutant pools. There have been only two genome-wide functional genomic screens of virulence in GAS. A modified signature-tagged mutagenesis (STM) screen using the insertion sequence IS256 from Tn4001 combined with a murine model of invasive skin disease was used to identify the streptococcal invasive locus sil (17). More recently, Kizy and Neely performed a more traditional STM screen to identify GAS genetic determinants important for virulence in a zebrafish model using Tn4001 mutants (18).
Several transposition systems have been successfully used in GAS (5), including those based on Tn4001 (9, 11), Tn916 (6, 10), Tn917 (7, 8), and ISS1 (12) as well as Tn5-based (19) and Mu-based (20) transposome systems. Since GAS is not naturally competent for DNA uptake, all of these approaches rely on in vivo transposition. Although each system has been effective for generating mutations in GAS, they all share a basic limitation: significant insertion site bias of the transposon leads to genetic “hot spots” for mutation and an overall reduction in randomness. This is a critical aspect for the production of mutant libraries used in TraSH and Tn-seq, where possessing multiple mutations in each gene increases confidence in the results. The mariner Himar1 transposon has become the state-of-the-art genetic tool for complex and random mutagenesis in Gram-positive genomes due to its simple transposition mechanism (no host factors required) and a ubiquitous target sequence: the dinucleotide “TA” (21–23). In addition, mariner has been predominately used for high-throughput mutant screens such as TraSH (13) and Tn-seq (16).
In this study, we adapted a mariner-based transposon (osKaR) for in vivo transposition in GAS and with attributes that allow its use in TraSH. Mutant libraries were produced in the invasive M1T1 GAS strain 5448 and used to perform TraSH analyses to identify genes important for growth and/or survival in whole human blood. Thus, we have developed a new transposon system for GAS that will allow screening and rapid detection of genes functionally important for fitness of GAS in vivo on a genome-wide scale.
Bacterial strains used in this study are shown in Table 1. GA19681 (serotype M6) and GAS 5448 (serotype M1T1) are clinical GAS isolates obtained from patients with invasive infection. GAS strains were routinely cultured in Todd-Hewitt medium (Alpha Biosciences) supplemented with 0.2% yeast extract (THY). Escherichia coli strains DH5α and C43[DE3] were used as strain hosts for plasmid construction and were cultured in Luria-Bertani (LB) medium (EMD Chemicals). Antibiotics (Fisher Scientific; Gold Biotechnology) were used at the following concentrations: ampicillin (Ap) at 100 μg/ml for E. coli, spectinomycin (Sp) at 100 μg/ml for both E. coli and GAS, and kanamycin (Km) at 50 μg/ml for E. coli and 300 μg/ml for GAS.
Oligonucleotides used in this study were synthesized by Integrated DNA Technologies, Inc., and are shown in Table S1 in the supplemental material. Plasmids were isolated using a Wizard Plus SV Minipreps kit (Promega) or a Qiagen Plasmid Purification Midi kit (Qiagen). Restriction enzymes, Antarctic phosphatase, T4 DNA ligase, RNase inhibitor, Moloney murine leukemia virus (M-MLV) reverse transcriptase, and DNA polymerase I, large (Klenow) fragment (New England BioLabs), were used according to the manufacturer's instructions. PCR was performed using either Taq DNA polymerase (New England BioLabs) or High-Fidelity AccuPrime Pfx DNA polymerase (Life Technologies) with 1 μg of DNA template and 10 pmol of the appropriate primers (see Table S1 in the supplemental material). When necessary, PCR products were purified using a Wizard SV Gel and PCR Clean-Up System kit (Promega). DNA sequencing was carried out by Genewiz, Inc. Transformations were performed with a Gene Pulser Xcell System apparatus (Bio-Rad) as recommended by the manufacturer, using electrocompetent cells of E. coli or GAS prepared as described by Ausubel et al. (24) or Simon and Ferretti (25), respectively. Genomic DNA (gDNA) from GAS was purified using a MasterPure Complete DNA Purification kit (Epicentre Biotechnologies).
Plasmids used in this study are shown in Table 1. The E. coli/GAS temperature-sensitive shuttle vector pCRS was generated as follows: a 1.28-kb DNA fragment (AMR) containing the ColE1 origin of replication and the multiple-cloning site was amplified from pUC19 using primers AMR1 and AMR2 (see Table S1 in the supplemental material). A 1.66-kb DNA fragment (OTS) containing the thermosensitive pWV01 origin of replication was amplified from pJRS233 using primers OTS1 and OTS2 (see Table S1 in the supplemental material). A 1.32-kb DNA fragment (SPR) containing a spectinomycin-resistance (Spr) cassette was amplified from pJRS525 using primers SPR1 and SPR2 (see Table S1 in the supplemental material). PCR products OTS and SPR were digested by SpeI and ligated together. The resulting ligation product was PCR amplified using primers SPR1 and OTS2, digested by NotI and XhoI, and finally ligated to the NotI-XhoI-digested AMR PCR fragment, resulting in the conditional integration plasmid pCRS (see Fig. S1A in the supplemental material).
A second E. coli/GAS Kmr temperature-sensitive shuttle vector was constructed as follows: a 1.49-kb DNA fragment containing aphA3 (Kmr) along with its promoter was amplified from pUC4ΩKm2 (Table 1) using primers KmR1bis and KmR2bis (see Table S1 in the supplemental material). PCR products OTS and Kmr were digested by SpeI and ligated together. The resulting ligation product was PCR amplified using primers KmR1bis and OTS2 (see Table S1 in the supplemental material), digested by NotI and XhoI, and ligated to the NotI-XhoI-digested AMR fragment, resulting in the conditional integration plasmid pCRK (see Fig. S1B in the supplemental material).
A composite mariner transposon, osKaR (outward sequencing Kanamycin-Resistance transposon), with attributes suitable for TraSH experiments was generated using successive PCRs: first, the aphA3 gene and its promoter region were PCR amplified from pUC4ΩKm2 (Table 1) using primers T7KmR1 and T7KmR2 (see Table S1 in the supplemental material) to introduce outward-facing consensus T7 promoter sequences at both ends of the aphA3 gene. The resulting 1.53-kb PCR product was then used as a template using primer ITR-T7 (see Table S1 in the supplemental material) to include the inverted terminal repeats (ITR) targeted by the Himar1 mariner transposase, resulting in the 1.6-kb osKaR transposon (Fig. 1A). PstI sites were included around osKaR using PCR with the primer IRP (see Table S1 in the supplemental material). Finally, the osKaR transposon was cloned as a PstI-digested fragment into PstI-cut pCRS to generate pCTL2 (Table 1).
A hyperactive allele of the Himar1 transposase gene (tnpC9) was amplified by PCR along with an optimal ribosome-binding site from pMarA (22) (Table 1) using primers RBStnpF and PAtnpRev (see Table S1 in the supplemental material). The resulting PCR product was digested by BamHI and ligated to BamHI-cut pIB166 (Table 1) containing the Lactococcus lactis promoter P23. The ligation was used for PCR amplification with the primers P23F and PAtnpRev (see Table S1 in the supplemental material) to generate the P23-tnpC9 allele. We were not able to clone directly the EcoRI-digested PCR product P23tnpC9 into pCTL2. To circumvent this issue, the PstI-digested osKaR PCR fragment and the P23tnpC9 PCR fragment were cloned into pUC19 to produce pP23tnpO (Table 1), digested by KasI and SphI, and the resulting fragment containing osKaR and the mariner transposase gene was moved into pCRS to produce pOSKAR (Fig. 1B). pOSKAR was found to be unstable in E. coli DH5α, and its propagation was facilitated using E. coli strain C43[DE3].
pOSKAR was electroporated into E. coli C43[DE3], and transformants were grown for 24 to 48 h on LB agar containing Km and Sp at 30°C. Multiple 200-ml cultures from individual transformants were grown overnight in LB containing Km and Sp at 30°C under gentle agitation. Plasmid was isolated from 5 ml of each culture, and restriction analyses were conducted using EcoRI and PstI to identify those containing the expected pOSKAR digest profile (2 bands with EcoRI [5.86 kb and 1.32 kb] and 2 bands with PstI [5.57 kb and 1.61 kb]) to eliminate cultures with rearranged plasmid. Validated cultures were then subjected to large-scale isolation.
GAS GA19681 was transformed with 20 μg pOSKAR and allowed to grow in THY broth at 30°C for 4 h, and 100-μl aliquots were plated onto THY agar containing either Km and Sp or Km alone to select for pOSKAR transformants or osKaR transposants, respectively. GAS 5448 was transformed with 300 μg pOSKAR and allowed to grow in THY broth at 30°C for 4 h. Cells were then plated on THY agar containing Km and Sp and incubated at 30°C for 48 h. Individual colonies were tested for the presence of intact plasmids by assaying for Kmr and Spr at the permissive temperature for plasmid replication (30°C), and Kmr and Sps at the restrictive temperature (37°C). For osKaR transposant selection, an aliquot of the frozen cultures was grown on THY agar containing Km and Sp at 30°C for 48 h. Colonies were resuspended in 2 ml of THY broth and inoculated into 150 ml THY containing Km for overnight growth at 37°C. Transposant libraries were further tested by patching Kmr clones on THY agar containing Km or Sp for overnight growth at 37°C. Libraries with 10% or less of Kmr Spr clones were chosen, individual clones (Kmr Sps) were further tested by PCR to verify the presence of the osKaR transposon, and arbitrarily primed PCR (AP-PCR) was used to identify osKaR insertion sites and determine overall library randomness. Libraries with fewer than 10% siblings were frozen in 10-ml aliquots.
The protocol described by Li et al. (26) was modified as follows. First-round PCR was performed with 100 ng of gDNA using primers oPCR1 and Deg3 (see Table S1 in the supplemental material) under the following conditions: 95°C for 5 min; six cycles of 95°C for 30 s, 30°C for 30 s, and 72°C for 1 min; 72°C for 5 min; 30 cycles of 95°C for 30 s, 45°C for 30 s, and 72°C for 1 min; and 72°C for 5 min. For the second round of PCR, 2 μl of purified first-round PCR product was amplified using the primers Anchor1 and Deg4 (see Table S1 in the supplemental material) under the following conditions: 95°C for 5 min; 30 cycles of 95°C for 30 s, 45°C for 30 s, and 72°C for 1 min; and 72°C for 5 min. The resulting PCR product was subjected to nucleotide sequence analysis (Genewiz) using primer Anchor2 (see Table S1 in the supplemental material). osKaR insertion sites were identified within relevant genome sequences using the Comprehensive Microbial Resource (CMR) BLAST program from the J. Craig Venter Institute (http://blast.jcvi.org/cmr-blast/).
A 10-ml aliquot of a random osKaR mutant library in GAS 5448 was grown overnight in 250 ml of THY containing Km at 37°C. A 10-ml volume of the resulting culture was inoculated in 250 ml THY containing Km and incubated overnight at 37°C again. Finally, 10-ml aliquots of the resulting osKaR mutant pool after two overnight incubations in THY broth at 37°C were collected and cells harvested by centrifugation for gDNA isolation.
Using a protocol approved by the University of Maryland Institutional Review Board (IRB), two 10-ml samples of whole blood were collected from healthy human donors. A 10-ml aliquot of a random osKaR mutant library in GAS 5448 was grown overnight in 40 ml THY broth containing Km at 37°C, 800 μl was then inoculated in 40 ml of THY broth containing Km, and cells were grown at 37°C to an optical density at 600 nm (OD600) of ~0.8 (late exponential phase). A 5-ml aliquot was removed, and cells were washed and resuspended in saline solution to obtain a suspension of ~5 × 108 CFU/ml. The remaining culture (45 ml) was further grown overnight at 37°C, and 10-ml aliquots of the resulting osKaR mutant pool were harvested (input culture). Approximately 2 × 108 CFU of the mutant library grown to mid-exponential phase were grown in whole human blood as follows: 400 μl of GAS mutant suspension in saline solution was inoculated into 4 ml of fresh human blood and grown for 4 h at 37°C in a rotary shaker (4 h blood). A 400-μl volume of the 4-h blood culture was inoculated into 4 ml of fresh human blood for an additional 4 h (8 h blood), and this step was repeated one additional time (12 h blood). Finally, 500 μl of the 12-h blood culture was inoculated into 500 ml of THY broth containing Km and grown overnight at 37°C, and 10-ml aliquots of osKaR mutant pool were harvested by centrifugation (output culture).
Sequences flanking both ends of the osKaR transposon (osKaR insertion tags) were generated using the procedure described by Murry et al. (27) with the following modifications: gDNA from osKaR mutant pools was purified, and 30 μg was partially digested using Tsp509I (Tsp). DNA fragments with a size under 3 kb were purified from a 1% agarose gel, treated with the Antarctic phosphatase (NEB), and subsequently ligated to a 1,000-fold molar excess of a Tsp-Adapter generated by annealing primers Adapter-1 and Adapter-Tsp (see Table S1 in the supplemental material). The ligation mixture was used for PCR amplification using primers Adapter-1 and oID3 (see Table S1 in the supplemental material) under the following conditions: 95°C for 5 min; 30 cycles of 95°C for 30 s, 45°C for 30 s, and 72°C for 1 min; and 72°C for 5 min. Approximately 4 μg of the resulting PCR products were used for in vitro transcription with T7 RNA polymerase using a MEGAshortscript kit (Ambion). The synthesized RNA was treated with DNase I (Ambion) for 1 h followed by a phenol-chloroform extraction and ethanol precipitation. Total synthesized RNA was used for reverse transcription in a 20-μl reaction mixture using 50 nmol of primer Adapter-1 (see Table S1 in the supplemental material), 10 nmol (each) of dATP, dCTP, and dGTP, 0.5 nmol of dTTP, 1 nmol of Cy(Cy3 or Cy5)-dUTP, 10 U of RNase inhibitor (NEB), and 400 units of M-MLV reverse transcriptase (NEB) at 37°C for 12 h. After degradation of the RNA, labeled cDNA was purified using a Vivacon 500 concentrator (Sartorius Biotech) as recommended by the manufacturer.
A 20-μg aliquot of GAS 5448 gDNA was digested with HindIII for 90 min, and the resulting DNA fragments were purified by phenol-chloroform extraction. Cy5-labeled gDNA probes were produced by random primer extension using 5 μg of HindIII-generated DNA fragments, 1 nmol (each) of dATP, dCTP, and dGTP, 1 nmol of Cy5-dUTP, and 5 U of DNA polymerase I, large (Klenow) fragment (NEB). Labeled gDNA probes were finally purified using a Vivacon 500 concentrator as recommended by the manufacturer.
The GAS pan-genome microarray used in this study (3) possesses 3,388 amplicons (size ranging from 51 to 1,170 bp) targeting GAS open reading frames found in 12 different genomes representing 11 distinct serotypes. Each amplicon is spotted in triplicate. Microarray slides were incubated in prehybridization solution (5× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 0.1% SDS, 8% bovine serum albumin [BSA]) at 42°C for 2 h, washed with water and then with isopropanol, and dried immediately before hybridization. Cy3- and Cy5-labeled DNA probes (100 pmol each) were mixed together, dried, and resuspended in 35 μl of hybridization solution (50% formamide, 5× SSC, 0.1% SDS, 0.6 μg/μl salmon sperm DNA). Samples were heated twice at 95°C for 10 min, snap-cooled on ice, and applied to the array slide under raised coverslips (Fisher Scientific). Microarray slides were hybridized overnight at 42°C in slide chambers (ArrayIt) under conditions of gentle agitation. After hybridization, arrays were washed 2 times for 10 min in Wash Solution 1 (2× SSC, 0.1% SDS), Wash Solution 2 (0.1× SSC, 0.1% SDS), and then Wash Solution 3 (0.1× SSC) and finally rinsed in distilled water (dH2O) for 5 min and dried.
Arrays were scanned at a 5-μm resolution with a GenePix Personal 4100A microarray scanner (Axon). Data were subsequently acquired with GenePixPro software (version 6.0) by generating GenePix result files (gpr) that were subsequently transferred along with their associated .jpg images into Acuity 4.0 software (Axon). For data analyses, a linear ratio-based normalization was performed using the ratio of means, and data sets were sorted by the mean of the ratios (Cy5/Cy3). Spots of interest were visually screened to identify those of low quality, which were then removed from the data set. Statistical analysis was performed on the mean of the ratio for each hybridization, and the results are presented as standard deviations (Table 2; see also Tables S2, S3, and S4 in the supplemental material).
An internal fragment from 10 different genes identified as important for fitness in donor A (Spy0078, adcC; Spy0220, tatD; Spy0221; Spy0414; Spy0662, fruA; Spy0701, cpsY; Spy0752, acoB; Spy0907, citF; Spy1794; Spy1857, guaB [see Table S2 in the supplemental material]) was PCR amplified from GAS 5448 gDNA using relevant primers (see Table S1 in the supplemental material), digested with BamHI, and cloned into pCRS to generate an insertional inactivation construct, pCRS.xxxx (Table 1). GAS 5448 was transformed with 5 μg of each mutagenic plasmid (pCRS.xxxx) and grown on THY agar containing Sp at 30°C. Potential integration mutants with mutations in each gene (GAS 5448.xxxx) were identified following growth on THY agar containing Sp at 37°C (Table 1). The presence of inserted plasmid was verified by PCR analysis of each GAS mutant gDNA using primers SPR1 and SPR2 (see Table S1 in the supplemental material). Mutant GAS strains were subjected to passage overnight twice in THY broth at 30°C to allow excision of the integrated plasmid from the chromosome and then grown on THY agar at 37°C to generate the rescued strain GAS 5448.xxxxR (Table 1). Absence of plasmid was verified by PCR analyses using primers SPR1 and SPR2 (see Table S1 in the supplemental material).
The ability of GAS strains to survive in whole human blood was tested as previously described (28): GAS were grown to early mid-exponential phase (OD600 ~ 0.15) and serially diluted in saline. A 50-μl volume of a 10−4 dilution (ca. 50 to 200 CFU) was added to 500 μl of fresh whole human blood and rotated for 3 h at 37°C. The multiplication factor (MF) was calculated by dividing the CFU obtained after blood challenge by the initial CFU inoculated. Data are presented (see Fig. 5) as percent growth in blood corresponding to the MF of the mutant divided by the MF of the wild type (WT) × 100.
To generate a tool for creating random and complex mutant libraries for high-throughput genetic screens in GAS, we developed a Himar1-based mariner transposable element called osKaR (outward-sequencing Kanamycin-Resistance transposon). osKaR consists of an aphA3 kanamycin-resistant (Kmr) cassette flanked upstream and downstream by outward-facing T7 promoter regions contained within ITR sequences recognized by the Himar1 transposase (Fig. 1A). For recruitment of osKaR to occur in GAS, a new spectinomycin-resistant temperature-sensitive shuttle vector, pCRS (stable at 30°C and unstable at 37°C), was created (see Fig. S1A in the supplemental material). The conditional replication of pCRS in GAS was verified in a serotype M6 strain, GA19681 (data not shown).
As a source of transposase, a hyperactive allele of the Himar1 gene (tnpC9) modified to contain a ribosomal binding sequence for optimal translation in Gram-positive bacteria was amplified from pMarA (22) (Table 1). A P23-tnpC9 transposase allele cloned with osKaR into pCRS was found to be stable when transformed into E. coli C43[DE3] (Table 1). The resulting pOSKAR plasmid (Fig. 1B) was used for all osKaR transpositions performed in this study. A pCTL2 control plasmid containing osKaR, but not the tnpC9 transposase gene, was also created (Table 1).
Transposition using pOSKAR was initially tested in the M6 GA19681 GAS strain due to its high transformation efficiency (29). pOSKAR and pCTL2 (Table 1) were separately introduced into GAS GA19681, and Kmr Sps colonies at 37°C, representing osKaR transposants, were obtained at an average transposition frequency of 8.7 × 10−3. No antibiotic-resistant GAS clones were observed at 37°C when transformation was performed with pCTL2, showing that osKaR is unable to transpose in the absence of the Himar1 transposase. Transposition of osKaR into M6 GAS GA19681 occurs rapidly after transformation, and growth for 4 h at 30°C is sufficient for transposition to occur. These results show that osKaR can transpose from pOSKAR in GAS through transient Himar1 expression.
To demonstrate that osKaR integrated into the GAS chromosome randomly, 60 Kmr Sps transposants were further investigated. Southern blot analysis using an osKaR-specific probe showed profiles consistent with the integration of one copy of osKaR into each GAS GA19681 chromosome (data not shown). Arbitrarily primed PCR (AP-PCR) followed by DNA sequencing was used to identify the osKaR insertion points in the genome of the 60 mutants. BLAST analyses of resulting sequence against the available M6 MGAS10394 genome revealed that osKaR inserted at 58 (96.7% randomness) unique positions distributed across the GAS chromosome (Fig. 1C) and that the insertions were exclusively at TA dinucleotides (data not shown). Thus, osKaR is able to transpose with a high degree of randomness in GAS comparable to mariner functionality in other bacterial species.
To investigate whether osKaR can produce complex mutant collections in GAS, the GA19681 transposon library was screened for mutants affected in the production of streptolysin S (SLS), the cytolytic toxin important for virulence and responsible for ß-hemolysis surrounding GAS colonies cultured on blood agar plates. SLS is the product of sagA, the first of nine sag genes required for synthesis and export of the toxin (6, 7). Approximately 12,000 osKaR mutants in GA19681 were patched onto blood agar plates, and 21 distinct ß-hemolysin-deficient mutants (0.18% randomness) were isolated based on significantly reduced or absent zones of clearing (Fig. 2A). AP-PCR analyses revealed that 15 of the hemolysin deficiency mutations mapped to osKaR insertions within the sag operon (Fig. 2C), including multiple insertions within several genes (4 in sagB, 3 in sagC, 3 in sagD, 2 in sagF), strongly suggesting that osKaR can potentially be used to perform saturating forward genetic screens in GAS. Although we did not find insertions within sagE, sagG, or sagI in our screen, this may reflect the qualitative nature of assessing halo size on plates. Eight novel hemolysin-deficient mutants of GA19681 were also identified with osKaR insertions outside the sag operon, including insertions in genes encoding the histidine kinase SalK and an ABC transporter protein, M6_Spy1879 (data not shown). These results confirm that osKaR mutagenesis can produce multiple transposon insertions in a single gene, a property required when performing TraSH analyses.
Although useful for initial assessment of osKaR transposition in GAS, the value of GA19681 for studying pathogenesis is limited due to its lack of virulence in most in vivo models of infection. The M1T1 GAS strain 5448 (Table 1) is a well-studied clinical isolate from a patient with an invasive infection, is virulent in mouse models, and represents the most common serotype associated with severe GAS diseases worldwide. Therefore, GAS 5448 was chosen to generate complex osKaR mutant libraries that would then be used for high-throughput TraSH analyses. Initial attempts to perform the in vivo osKaR transposition in GAS 5448 using the protocol developed for GAS GA19681 were unsuccessful, primarily due to low transformation efficiency. Separating the initial step of pOSKAR transformation into GAS 5448 (step 1) and the selection for osKaR transposition (step 2) proved to be successful (see Materials and Methods). GAS 5448 transformants were first screened for pOSKAR transformation (Spr Kmr at 30°C; Kmr Sps at 37°C). To produce osKaR mutant libraries, pOSKAR-containing GAS 5448 clones were grown on THY plates with Km and Sp for 48 h at 30°C. Cells were then plated on THY plates with Km at 37°C to select for osKaR transposants. Four different mutant libraries were produced, generating on average 3.8 × 104 transposants per library with an average frequency of transposition of approximately 3.4 × 10−3. The libraries were then pooled, and the resulting master library was analyzed for complexity.
AP-PCR analysis of 58 randomly picked transposants from the pooled master library of GAS 5448 identified 35 unique osKaR insertion sites (60.3% randomness), indicating a reduction of the randomness of osKaR mutagenesis in GAS 5448 compared to that observed with GAS GA19681 (96.7%). The extended 2-step protocol for in vivo transposition in the GAS M1T1 background compared to the more direct 1-step method used for the GAS M6 background likely resulted in an increase in sibling isolation and the observed reduction in overall randomness.
Approximately 25,000 osKaR mutants from the master library generated in GAS 5448 were screened for a capsule-overproducing phenotype (mucoidy) on plates (Fig. 2B). A total of 17 unique osKaR-mediated capsule-overproducing mutants mapping to 4 different locations in the GAS genome were identified. These included single insertions in Spy1478 (1 mutant) and Spy0349 (1 mutant), and multiple insertions mapping to the 2-component histidine kinase genes covS (8 mutants) and rocA (7 mutants) (Fig. 2D). Mutations in both covS and rocA are known to be linked to increased capsule production (10, 12). The presence of multiple hypermucoid mutations mapping within single genes suggests complete coverage for at least some genes (>5× insertions) in the osKaR GAS 5448 mutant library.
TraSH analyses rely on DNA microarray hybridization to identify mutants with fitness deficiency during en masse phenotype screens. The GAS pan-genome DNA microarray previously designed by Bessen et al. (3) was tested for its ability to detect DNA probes corresponding to (i) the GAS 5448 genome (gDNA) and (ii) osKaR insertion tags generated from the master osKaR mutant library in GAS 5448 grown in THY broth. Cy5-labeled GAS 5448 gDNA probes were produced using random primer extension, and Cy3-labeled osKaR insertion tags were generated using gDNA from the master osKaR mutant library subjected to passage in THY (see Materials and Methods). Equal amounts of Cy5- and Cy3-labeled DNA probes were mixed and hybridized onto the GAS pan-genome DNA microarray (3) and signal intensities determined from two biological replicates (six independent hybridizations). An amplicon was defined as “present” when its hybridization signal value was at least 2.5 times greater than the value of the background signal in 4 of the 6 hybridizations. Using these criteria, GAS 5448 gDNA probes identified 1,338 amplicons corresponding to 1,129 genes (Fig. 3, red lines). The result emphasizes a technical bottleneck encountered in using the GAS pan-genome DNA microarray, as it appears to detect only ca. 65% of the 1,865 genes present on the reference annotated genome of the M1T1 GAS strain MGAS5005 (30) (Fig. 3, blue lines). The osKaR insertion tags hybridized to 1,093 amplicons corresponding to 950 different genes (Fig. 3, green lines). This finding suggests that osKaR mutagenesis of the GAS 5448 genome was not saturating, as 179 genes detectable using the GAS pan-genome DNA microarray were not found. However, we were able to identify osKaR-generated insertion tags from our TraSH THY library representing over half the GAS 5448 gene content and 84% of genes detectable on the array with even coverage (Fig. 3, green lines), supporting the use of osKaR for TraSH analyses.
The ability of GAS to grow in human blood is a major virulence phenotype, and we used the TraSH technique to identify genes required for GAS survival and/or growth in this clinically relevant environment. The Lancefield blood bactericidal assay has been a standard model to study GAS virulence (28); however, this assay relies on very small GAS inoculums (50 to 200 CFU) that would not provide the mutant library complexity required to perform TraSH. Therefore, we developed a modified assay to support increased inoculums and look for GAS growth defects in human blood. Larger inoculums of GAS (see Materials and Methods) were used to inoculate whole human blood from a given donor prior to incubation at 37°C for 4 h. The master library of GAS 5448 osKaR mutants was grown to late exponential phase, and a 10−1 dilution of this culture was inoculated into fresh blood for an additional incubation at 37°C for 4 h and this was repeated again for a total passage time of 12 h (see Materials and Methods). To validate this assay, we used an osKaR mutation in mga (GAS 5448.mga, Table 1) encoding the multiple virulence gene regulator Mga known to play an important role for GAS survival in human blood (31). In a Lancefield bactericidal assay, wild-type GAS 5448 grew ca. 10-fold over 3 h of incubation in whole human blood whereas GAS 5448.mga showed an almost 30-fold decrease in CFU (Fig. 4A). To test our modified assay, approximately 2 × 105 CFU of wild-type GAS 5448 were mixed with 5 × 103 CFU of an isogenic GAS 5448.mga (40/1 ratio) to mimic the likely ratios found in a TraSH competition experiment. The resulting mix was used to inoculate 500 μl of whole human blood from a single donor (donor A) in the assay described above. Wild-type GAS 5448 showed an increase (1.0 to 1.5 logs) in CFU during each 4-h passage, whereas the mga mutant exhibited a significant decrease in numbers during the second and third passages, resulting in over 2 logs reduction from the input (Fig. 4B) and a competitive index (mutant/WT) of 4 × 10−5. Thus, our 12-h competitive assay was sufficient to significantly reduce survival of a known mutant defective for growth in human blood.
For the full-scale TraSH experiment, the master GAS 5448 osKaR mutant library was grown to mid-exponential phase in THY broth (input pool), and approximately 2 × 108 CFU was used to inoculate 4 ml of human blood (donor A) and subjected to three successive 4-h passages in human blood as described above. Viable cell counts were performed to verify the ability of the library to expand during each passage in blood (data not shown). After the final passage in human blood (12 h total), the osKaR mutant library was expanded in THY broth for 8 h in order to dilute the blood and recover GAS cells suitable for gDNA extraction (output pool). TraSH insertion tags were produced from gDNA extracted from both the output and input mutant pools using in vitro transcription and reverse transcription in the presence of Cy3- and Cy5-labeled dUTP, respectively. Equal amounts of labeled tags were hybridized on the pan-genome GAS microarray (3) in technical duplicates, generating a total of six probe hybridizations. The Cy5/Cy3 (input/output) signal ratio was averaged from the six hybridization values, and genes with an average Cy5/Cy3 signal ratio of over 1.5 were regarded as underrepresented in the output osKaR mutant pool.
A total of 73 genes were identified that fit the above criteria for donor A (see Table S2 in the supplemental material). Clusters of orthologous groups (COG) category classification from the JCVI CMR website shows a prevalence of genes with putative functions in energy production (8%), carbohydrate transport and metabolism (12%), amino acid transport and metabolism (13%), nucleotide transport and metabolism (8%), and DNA replication, recombination, and repair (10%). As expected from our preliminary study (Fig. 4), mga (Spy1720) was identified as important for GAS fitness in blood from donor A (see Table S2 in the supplemental material). We also identified the genes perR (Spy0161) and hasA (Spy1851), which have been previously reported for their role in GAS resistance to killing by human blood leukocytes (32) and opsonization (33), respectively. The TraSH approach also identified adcC (Spy0078), which was recently shown to play a role in Streptococcus pneumoniae survival within human serum (34), and cpsY (Spy0701), which was found to be essential for S. iniae survival in human blood (35).
To validate the TraSH findings, we performed confirmatory analyses on 10 genes from our TraSH list in blood donor A (Spy0221, Spy1794, adcC, tatD, Spy0414, fruA, cpsY, acoB, citF, and guaB [see Table S2 in the supplemental material]) that have not yet been established as playing a role in GAS fitness in human blood. To test whether disruption of these genes could have an effect on GAS fitness in blood, GAS 5448 mutants were constructed by insertional inactivation using conditionally replicative plasmids derived from the pCRS vector (Table 1; see also Fig. S1A in the supplemental material). Insertional mutants with mutations in each gene were then cured for their integrated plasmid to produce a corresponding rescue strain (Table 1). All strains grew at levels comparable to that of wild-type GAS 5448 in THY broth (data not shown), indicating that the genes were dispensable for GAS growth in rich medium. Lancefield bactericidal assays performed with donor A blood revealed that all insertional mutants showed a significant reduction in growth, presented as percent growth in blood (multiplication factor of mutant/multiplication factor of wild type × 100), compared to that of the parental GAS 5448 (Fig. 5A). Strains that had been rescued for the integration mutations to reestablish a wild-type genetic background led to the restoration of a percentage of growth in blood comparable to that observed with GAS 5448 (Fig. 5A). These results validate our TraSH findings and confirmed the role of 10 previously uncharacterized genes in GAS fitness in human blood.
Additional TraSH analyses using the master osKaR mutant library were performed as described above using blood obtained from two additional donors (blood B and C). Our data identified 73 genes important for GAS 5448 fitness in donor B blood and 104 genes in the blood from donor C (see Tables S3 and S4 in the supplemental material). Comparing the results from all 3 donors, a total of 139 total genes were identified (Table 2; see also Tables S2, S3, and S4 in the supplemental material and Fig. 6). The results obtained from the TraSH experiments using the three different bloods reveal a discrepancy in the numbers of genes identified. The TraSH technique is known to introduce bias leading to variability among replicates (36), and this is likely the case in the results presented here. It is also possible that the 3 different bloods used in these experiments may have different compositions and therefore represent different environments for which GAS 5448 fitness could be different.
A total of 30 genes were found common to all 3 experiments (Table 2). Of the genes found critical in all three experiments, 4 genes (purD, purA, fhs2, and guaB) are involved in nucleotide transport and biosynthesis, 4 genes are linked to DNA replication, recombination, and repair (tatD, Spy0221, Spy0414, and Spy1716), 5 genes relate to energy production and conversion (acoB, acoC, citF, Spy0932, and Spy0933), and 5 genes are associated with amino acid transport and metabolism (Spy0596, carB, bcaT, Spy0829, and artP). Additional Lancefield bactericidal assays performed with insertional mutants with mutations in Spy0221 and Spy1794 confirmed that these 2 genes were important for GAS 5448 fitness in blood from all 3 donors (Fig. 5).
Our data also identified 51 genes in 2 of the 3 experiments (Table 2 and Fig. 6), with mga and the Mga-regulated scpA being found in this category (Table 2). Interestingly, however, Lancefield bactericidal assays showed that mga was important for GAS fitness in blood from all three donors (Fig. 5). In this category are also other genes (adcC, carA, mefE, pyrB, pyrG, ralp3, trxS) known for their role in the fitness of GAS, or closely related bacteria, in blood (Table 2). The 81 total genes found in at least 2 of the 3 experiments (Table 2; Fig. 3, purple lines; Fig. 6) are likely to represent a conserved set of genes important for GAS 5448 fitness in human blood.
Invasive GAS diseases such as septicemia, necrotizing fasciitis, and streptococcal toxic shock syndrome, although rare, are life-threatening conditions that present very limited treatment options. Forward genetic screens (e.g., TraSH) using both in vivo and ex vivo conditions relevant to such severe infections can be invaluable for gaining a deeper molecular understanding of GAS pathogenesis. However, these approaches typically depend on mariner-based transposons for complex random mutagenesis of the pathogen genome prior to screening. We report here the development of a composite mariner transposon (osKaR) designed for efficient in vivo transposition within GAS. Complex osKaR mutant libraries were successfully generated in the invasive M1T1 GAS strain 5448, and a TraSH screen was performed using a pan-genome GAS microarray to identify genes important for growth in whole human blood. Thus, we have produced a new genetic tool for GAS research and used it to reveal genes that are important for GAS dissemination through the bloodstream.
The primary requirement for genome-wide genetic screens such as TraSH is the ability to produce a comprehensive or saturating random mutant library in clinically relevant strains of the pathogen. Although several transposon systems have been successfully used in GAS over the years, they all exhibit various amounts of insertion site bias leading to limited mutant library complexity. The mariner-based system (osKaR) developed here (Fig. 1), as with previously tested mariner systems in other bacteria, was found to insert exclusively at the dinucleotide TA regardless of the serotype of GAS tested (M6, M1, M49, M3; data not shown). Given the low G+C content of GAS, the potential for complex mutagenesis of the genome by osKaR insertions is quite high. This expectation was supported by our screen of osKaR mutant libraries for hemolysin-deficient (M6, GA19681) and hypermucoidy (M1T1, GAS 5448) phenotypes (Fig. 2), whereby multiple independent insertions across a large operon (sagA to sagI) and 6 to 7 independent insertions in a single gene (covS, rocA) were found, respectively. Therefore, osKaR seems to represent an effective mariner-based transposon for GAS mutagenesis that should also be functional for closely related genera (e.g., Streptococcus, Enterococcus, and Lactococcus).
Interestingly, we observed that the randomness of osKaR insertions was linked to the transformation efficiency of the GAS strain being tested. Easily transformable strains (M6 GA19681) resulted in almost complete randomness (96.7% randomness). This was likely due to transposition concomitant with efficient transformation allowing for direct selection of transposants in one step, leaving very little time for siblings (duplicates) to accumulate. In contrast, strains with lower transformation efficiencies (M1T1 5448) required an initial selection for transformation of pOSKAR (step 1) followed by a separate selection for transposants (step 2). This resulted in levels of random insertions (48% to 60% randomness) in the libraries that were lower than those seen with GA19681. Similar observations were seen with other clinical GAS strains with low efficiencies of transformation, including M1T1 MGAS5005 and M3 MGAS315 (data not shown). High numbers of siblings in a mutant library are more of a problem for direct screening of the library for phenotypes (Fig. 2) and less of an issue for genome-wide negative selections such as TraSH (Table 2; see also Tables S2, S3, and S4 in the supplemental material). Nevertheless, we have taken steps to optimize osKaR transposition while limiting sibling growth and combine only those independent mutant libraries that show optimal randomness into “master libraries” to enhance coverage of the GAS genome.
We also observed early on that pOSKAR was very unstable in E. coli, leading to plasmid rearrangements during the process of purification. This is not surprising, considering that osKaR is fully capable of transposing in E. coli. To overcome this issue, we utilized the E. coli strain C43[DE3] normally reserved for overexpression of potentially toxic membrane proteins (37). This provided a significant improvement in pOSKAR stability over the use of normal cloning strains such as DH5α (data not shown). Still, precautionary measures are necessary to monitor the stability of pOSKAR by restriction mapping prior to plasmid isolation for use in generating GAS mutant libraries.
osKaR was designed with outward-facing T7 promoters to generate transposon-specific insertion tags based on in vitro transcription from mutant library gDNA en masse for screening by DNA microarray (TraSH) as initially described by Sassetti et al. (13). Our “proof-of-principle” TraSH experiment, although constrained by the number of GAS 5448 genes identified using the GAS pan-genome microarray (3), confirmed that osKaR was suitable for TraSH analyses. The TraSH method used here allowed the monitoring of mutants representing over half the GAS 5448 gene content, but a full 84% of genes actually detectable on the array (Fig. 3). Thus, using sequencing-based detection of insertions (Tn-seq) to remove the array bottleneck would be expected to reveal comprehensive coverage of the full GAS genome in our osKaR libraries.
Since Rebecca Lancefield first developed the blood bactericidal assay (28), it has been a gold standard for determining the virulence capacity of GAS ex vivo. The low inoculum (50 to 200 CFU) and 3-h growth period test for the ability of GAS to survive phagocytic killing, and inocula multiply based on the immune status of the donor. In order to modify the Lancefield assay for TraSH, we increased the GAS inoculum (see Materials and Methods) to allow a more complex input pool, increased the volume of human blood used (4 ml), and grew the bacteria for 3 serial passages of 4 h each (12 h total) to enhance mutant growth defects. A GAS 5448 mga mutant control killed in the Lancefield assay was also selected against in the modified competition assay with wild-type GAS 5448 (10:1 WT/mutant) comparable to what occurs during a TraSH negative selection (Fig. 4). Thus, the modified assay appears to function in selecting against GAS mutants unable to grow in blood. Since the increased inoculum size might overwhelm phagocytic killing, we hypothesized that two main classes of mutations could be selected against in the modified assay: both metabolic genes required to use human blood as a growth medium (more likely) and virulence genes important for resistance to phagocytic killing (less likely). Although mutants from both classes of mutations appear to be evident in our TraSH analysis in blood (Table 2; see also Tables S2, S3, and S4 in the supplemental material), there is a preponderance of genes involved in metabolism, nutrient uptake, and energy production. Thus, our modified blood growth assay clearly favors identifying mutants unable to use blood for growth. Importantly, a Lancefield assay using independent mutants with mutations in 10 genes identified in the TraSH screen showed that they were all diminished for blood survival (Fig. 5), validating the idea that our screen does reveal genes important for GAS fitness in this ex vivo environment.
TraSH analyses were performed using whole blood obtained from 3 different individuals (2 male, 1 female), and our criteria identified totals of 73, 73, and 104 genes important for GAS 5448 fitness in bloods A, B, and C, respectively (see Tables S2, S3, and S4 in the supplemental material) (Fig. 6). These represent a total of 139 individual genes, 30 genes being common in all 3 TraSH experiments plus 51 genes identified in 2 of the 3 experiments (Table 2). Although technical variability has been reported in performing TraSH analyses (36), one can also speculate that the variability described here could be the result of inherent interindividual and gender-related differences in terms of blood composition (blood cell types, plasma) that might have an impact in terms of GAS fitness and growth requirements. To support this idea, transcriptome analyses following growth of GAS in human blood have reported inherent interindividual and gender-related variation in their donor blood (38). Although genes conserved in all three donor experiments (Table 2) are likely to play a key role in GAS fitness in human blood, even those genes found only in a single donor (see Tables S2, S3, and S4 in the supplemental material) are still important to pursue further.
Genes known to be important for GAS virulence were found in the TraSH analysis, providing further support for the validity of the screen. Mga is a global virulence regulator that activates expression of emm1 and was used to establish the modified blood growth assay due to its inability to grow in blood (donor A). mga was identified in only 2 of the 3 TraSH experiments in blood (Table 2). Bactericidal assays with a defined mga mutant later confirmed its importance for GAS survival in blood from all three donors when not in competition with the wild type (Fig. 5). Interestingly, the antiphagocytic hyaluronic acid capsule (hasA operon) was also found in our screen, but only in donor A (see Table S2 in the supplemental material). We identified the stand-alone regulator RofA-like protein 3 (ralp3, Spy0559), which was recently shown to have a role in GAS survival in human blood and plasma (39). The gene perR (Spy0161) encoding the peroxide resistance regulator was found conserved in all donor bloods and has been shown to be important for GAS fitness within the host and necessary for resistance to phagocytic killing in human blood (32). PerR was also recently reported to enhance GAS oxidative stress resistance and virulence in the host (40). The TraSH experiments also identified mefE (Spy0450), encoding a macrolide-efflux pump, previously characterized as important for GAS virulence in a STM screen using a zebra fish model (18).
Genes important for virulence in other GAS-related Gram-positive pathogens were also found. The cpsY gene (Spy0701) encoding a LysR-family transcriptional regulator was recently identified in S. iniae as a virulence determinant important for systemic infections and intracellular survival in neutrophils (35). CpsY controls steps involved in the modification of S. iniae peptidoglycan composition (41). In S. mutans and S. agalactiae, cpsY is involved in the regulation of amino-acid biosynthesis (42). Genes with virulence functions in S. pneumoniae were also identified. The adcC (Spy0078) gene encodes a component of a Zn ABC transporter recently shown for its role in S. pneumoniae resistance to human serum (34). The clpE (Spy1240) gene encodes an ATP-dependent protease previously shown for its role in the virulence of S. pneumoniae and Listeria monocytogenes (43). Individual GAS 5448 mutants with mutations in cpsY and adcC confirmed the role of these 2 genes in GAS fitness in blood (Fig. 5A).
Many genes found in our TraSH screen in blood were related to metabolism and nutrient uptake (Table 2; see also Tables S2, S3, and S4 in the supplemental material), providing insight into the growth needs of GAS in blood. Human blood is considered to be rich in sugars and peptides, but lacking in free amino acids (44). Key genes involved in nucleotide biosynthesis were identified, including guaB, purA, and purD in purine synthesis and carA, carB, pyrB, pyrD, and pyrG in pyrimidine synthesis. The de novo biosynthesis of nucleotides has been shown to be critical for growth in blood and sepsis for both Gram-positive (Bacillus anthracis, S. aureus, and S. pneumoniae) and Gram-negative (E. coli, Francisella tularensis, Salmonella sp.) pathogens (45, 46). Our results support the idea that blood is also a poor source of nucleotides and that their biosynthesis is essential for GAS persistence in the bloodstream. Several phosphotransferase system (PTS) transporters and carbohydrate metabolism genes (fruA, acoB, citF, scrB, lacZ, mipB, Spy0780, Spy0212, Spy0179, and Spy1794) were also identified in the TraSH screen, suggesting the importance of sugar metabolism in GAS fitness in human blood. Among these is scrB, which encodes a sucrose-6-phosphate hydrolase that also contributes to S. pneumoniae in vivo fitness (47). Members of a third group of metabolic genes likely to indicate a key process in blood survival were those involved in amino-acid transport and biosynthesis (bcaT, artP, pepF, glyA, Spy0596, Spy0829, Spy1513). Individual GAS 5448 mutants with mutations in guaB, acoB, citF, and fruA confirmed the role of these genes in GAS fitness in blood (Fig. 5A).
Probably the most interesting of the genes identified by TraSH were those with no known association with growth in blood; among these, 4 genes, Spy0221, Spy1794, Spy0414, and tatD, were chosen for individual validation. Spy0221 encodes a homolog of the B. subtilis RNase M5 that functions as a 5S rRNA maturase and appears to possess a limited number of RNA targets (48). Interestingly, Spy0414 also encodes putative RNase, suggesting that RNA degradation might serve a function for GAS fitness in blood. RNases have been shown to contribute to virulence regulation (49), including the recent demonstration that the GAS RNase CvfA (RNase Y) regulates several virulence factors in response to nutritional stimuli (50). Individual GAS 5448 mutants with mutations in Spy0221 and Spy0414 confirmed the role of these 2 genes in GAS fitness in blood (Fig. 5A). The Spy1794 and tatD genes, predicted to encode a putative membrane-spanning protein with no known function and a putative Sec-independent translocase protein, respectively, were also further analyzed and the corresponding insertional mutants exhibited a significant defect for survival in blood in the Lancefield assay (Fig. 5A). The in vivo role of these genes and other genes identified in our screen with no obvious association with bloodstream infections are currently being explored in the laboratory.
In conclusion, this work describes the development and use of a new mariner transposon, osKaR, for the first genome-wide TraSH analyses to identify genes important for GAS fitness in human blood. Although many of the genes found have an established role in the dissemination of GAS and other pathogens in blood, there were also a number of novel genes identified that should provide new insights into GAS growth in the bloodstream. Since osKaR has the ability to generate complex random mutant libraries in GAS, the creation of an ordered mutant library in the M1T1 GAS 5448 is currently in progress as a useful resource to the GAS pathogenesis community. Finally, osKaR has been modified (MmeI restriction site in its ITR) for the use of deep sequencing of insertion tags (Tn-seq) to increase the dynamic range of our fitness studies and allow assessment of genetic interactions in relevant host environments.
We thank members of the McIver laboratory for critical review of the manuscript and Patrick Curry, Rena Bernstein, and Jason Glass for technical assistance.
This work was supported in part by a UMB/UMCP Seed Grant with Mark Shirtliff at UMB Dental School (K.S.M. and Y.L.B.), an NIH F31 predoctoral fellowship (K.M.V., AI00576), and a grant from the NIH National Institute of Allergy and Infectious Diseases (K.S.M., AI47928).
Published ahead of print 7 January 2012
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.00837-12.