Open Access

Defining the ABC of gene essentiality in streptococci

  • Amelia R. L. Charbonneau1, 2Email author,
  • Oliver P. Forman1,
  • Amy K. Cain3, 4,
  • Graham Newland1,
  • Carl Robinson1,
  • Mike Boursnell1,
  • Julian Parkhill3,
  • James A. Leigh5,
  • Duncan J. Maskell2 and
  • Andrew S. Waller1
BMC Genomics201718:426

Received: 24 March 2017

Accepted: 14 May 2017

Published: 31 May 2017



Utilising next generation sequencing to interrogate saturated bacterial mutant libraries provides unprecedented information for the assignment of genome-wide gene essentiality. Exposure of saturated mutant libraries to specific conditions and subsequent sequencing can be exploited to uncover gene essentiality relevant to the condition. Here we present a barcoded transposon directed insertion-site sequencing (TraDIS) system to define an essential gene list for Streptococcus equi subsp. equi, the causative agent of strangles in horses, for the first time. The gene essentiality data for this group C Streptococcus was compared to that of group A and B streptococci.


Six barcoded variants of pGh9:ISS1 were designed and used to generate mutant libraries containing between 33,000-66,000 unique mutants. TraDIS was performed on DNA extracted from each library and data were analysed separately and as a combined master pool. Gene essentiality determined that 19.5% of the S. equi genome was essential. Gene essentialities were compared to those of group A and group B streptococci, identifying concordances of 90.2% and 89.4%, respectively and an overall concordance of 83.7% between the three species.


The use of barcoded pGh9:ISS1 to generate mutant libraries provides a highly useful tool for the assignment of gene function in S. equi and other streptococci. The shared essential gene set of group A, B and C streptococci provides further evidence of the close genetic relationships between these important pathogenic bacteria. Therefore, the ABC of gene essentiality reported here provides a solid foundation towards reporting the functional genome of streptococci.


Transposon Sequencing Essentiality Barcode


Strangles, caused by Streptococcus equi subspecies equi (S. equi), is one of the most frequently diagnosed infectious equine diseases worldwide. S. equi is a Gram positive bacterium belonging to the Lancefield group C family of streptococci [1]. S. equi is closely related to the group A Streptococcus, Streptococcus pyogenes (S. pyogenes) [2] and the group B Streptococcus, Streptococcus agalactiae (S. agalactiae) [3], both of which are important human pathogens. S. pyogenes causes impetigo, pharyngitis, scarlet fever and necrotising fasciitis [46] and S. agalactiae causes meningitis, pneumonia and sepsis in neonates [7], in addition to mastitis in cattle [8] and streptococcosis in fish [9].

The increased accessibility of next-generation sequencing (NGS) technologies has facilitated the development of a variety of transposon-genome junction sequencing techniques, which combine dense mutant libraries and sequencing to identify essential bacterial genomes and assign gene function. The precise details of these methods: TraDIS, Tn-seq, HITS, INSeq and PIMMS vary from one another [1014], yet all produce similar end-point data [15]. Each technique employs a transposon delivery vector to produce a library of random transposition mutants within the bacterial genome. Viable mutants contain transposons that have inserted into non-essential genes, with insertions into essential genes proving lethal. NGS of transposon-genome junctions in saturated transposon mutant libraries permits the simultaneous identification of potentially hundreds of thousands of unique insertion sites, providing data pertaining to gene essentiality at the most basic level. Exposure of mutant libraries to varying experimental conditions, however, enables the relative fitness and conditional essentiality of each gene to be determined. In recent years, a range of essential bacterial genomes have been published using transposon directed sequencing methods [10, 11, 14, 1621]. Interrogating genomes in this way provides an unprecedented insight into genome-wide fitness, especially when libraries are subjected to disease relevant conditions either in vitro [10, 16, 22] or in vivo [2327].

Here, we present a transposon directed insertion-site sequencing (TraDIS) system which is conducted using standard Illumina sequencer protocols. Dense mutant libraries utilising the plasmid pGh9 carrying the insertion element ISS1 (pGh9:ISS1) [28], have previously been utilised with success in Streptococcus uberis (S. uberis) [14]. We modified pGh9:ISS1 within the 5’ terminal of ISS1 to create six barcoded transposons. The six independent libraries were generated and sequenced after growth in rich media. Data for each library was compared and combined, providing a blue-print data set for the subsequent analysis of conditional fitness and gene essentiality assignment in S. equi. The agreement of gene essentiality between our S. equi TraDIS data and Tn-Seq data from the close relatives S. pyogenes and S. agalactiae was determined. KEGG (Kyoto encyclopaedia of genes and genomes) pathways were attributed to the essential gene sets of S. equi, S. pyogenes and S. agalactiae to unveil the key biochemical pathways in which they are involved.


Barcoding ISS1

Five barcoded variants of the plasmid pGhost9:ISS1 [28] were generated using the primers listed in (Additional file 1: Table S1) to mutate the two bases (CA) located three and four bases downstream of the ISS1 inverted repeat (Additional file 1: Figure S1 for plasmid map and Additional file 1: Figure S2 for PCR design). The new plasmids: pGh9:ISS1:TC, pGh9:ISS1:AG, pGh9:ISS1:AC, pGh9:ISS1:CT and pGh9:ISS1:GA contained the alternative bases TC, AG, AC, CT or GA, respectively at these positions. For clarity, the original pGhost9:ISS1 will be referred to as pGh9:ISS1:CA in this manuscript with final libraries referred to as CA, TC, AG, AC, CT and GA. The methodology utilised for this process is provided in Additional file 1.

Generation of transposon libraries

S. equi strain 4047 (Se4047) cells were transformed with the desired pGh9:ISS1 plasmid by electroporation as previously described [29]. Transformants were grown for 3 h in Todd-Hewitt broth (THB) at 28 °C, permitting extrachromosomal plasmid replication. Transformants were grown on Todd-Hewitt agar (THA) supplemented with 0.5 μg/ml erythromycin (THAE) for 3 days at 28 °C. A colony of transformants was then grown overnight at 28 °C in THB supplemented with 0.5 μg/ml erythromycin (THBE). Overnight cultures were heat shocked at 40 °C for 3 h resulting in random transposition of ISS1 and the plasmid into the bacterial chromosome. Transposition frequency was determined by counting the colony forming units per millilitre of transposants on THAE versus THA. Transposants were selected by overnight growth at 37 °C in a humidified atmosphere containing 5% CO2 on 30 large (150 mm diameter) THAE plates supplemented with 0.03 μg/ml of hyaluronidase. Pools of random transposon mutants (transposon libraries) were harvested from the plates by washing with THB containing 25% glycerol and the bacterial suspension stored at -20 °C. The transposon libraries were then grown at 37 °C in a humidified atmosphere containing 5% CO2 to an OD600nm of 0.3 in THBE, 2.5 ml of the culture was centrifuged at 10,000 xg for 5 min and the bacterial pellet stored at -20 °C.

Effect of barcoded ISS1 on library growth

Each of the six barcoded libraries were grown overnight in THBE at 37 °C in a humidified atmosphere containing 5% CO2 alongside wild-type Se4047, which was grown in THB. Cultures were diluted to an initial OD600nm of approximately 0.08 and incubated under the same conditions. The OD600nm was measured every 30 min until stationary phase. The growth curves were completed in triplicate, with each replicate conducted on different days and from different stored aliquots. Doubling times were calculated from the mean exponential phase data for each library and Se4047. The mean doubling times of the libraries were tested for statistical significance using the Student’s t-test.

Stability of integrated pGh9:ISS1

Ninety-five colonies recovered from library CA were grown overnight in THBE at 37 °C in a humidified atmosphere containing 5% CO2, before they were combined to generate P0. The 95 mutant pool was then passaged twice overnight under the same conditions to produce P1 and P2. 2.5 ml of each culture was centrifuged at 10,000 xg for 5 min and the bacterial pellet stored at -20 °C.

DNA preparation and sequencing by TraDIS

DNA was extracted from the six barcoded mutant libraries and the three stability libraries cell pellets using a GenElute column kit according to the manufacturer’s instructions for Gram positive bacteria (Sigma-Aldrich). DNA was quantified using the Qubit dsDNA BR assay kit. 1.5 μg DNA was fragmented by sonication using a Misonix XL 2020 Ultrasonic Liquid Processor (cup horn arrangement) to produce fragments in the range of 200-800 bp, with 800 bp fragments being most prevalent. A Y-adaptor generated in-house using Illumina multiplexing adaptor sequences (Oligonucleotide sequences © 2007- 2012 Illumina, Inc. All rights reserved) was ligated to 1 μg of fragmented DNA using the NEBNext Ultra II DNA library prep kit for Illumina (New England Biolabs) according to the manufacturer's instructions for End Repair and Adaptor Ligation (see Additional file 1 for adaptor generation protocol). Fragments were purified using AMPure XP beads (Agencourt, Beckman Coulter) with a bead to DNA ratio of 1:1, according to manufacturer’s instructions.

Incubation of adaptor ligated DNA with the restriction enzyme SmaI for two hours at 25 °C was used to cleave the pGh9:ISS1 plasmid 33 bp upstream of the sequence encoding ISS1 in order to minimise the amount of TraDIS reads mapping to plasmid. AMPure XP beads with a bead to DNA ratio of 1.8:1 were used to clean up the digest reaction, according to manufacturer’s instructions. The amount of DNA recovered was quantified using the Qubit dsDNA HS assay kit. As recommended by Langridge et al. 100 ng of library DNA was PCR amplified for 20 cycles according to 1.4C of the NEBNext Ultra II DNA library prep kit protocol. Amplification utilised the specific ISS1 primer and indexing PCR primer, which facilitated the attachment of the resultant product to the sequencing flow cell (Additional file 1: Table S1). The regions that were amplified span the 5' end of ISS1 and the site of transposition in the S. equi genome. The use of a Y-adaptor enabled amplification of ISS1 containing fragments only, as reverse amplification could not occur until the specific ISS1 primer had amplified, generating a complementary Y-adaptor sequence for the indexing PCR primer to bind (see Additional file 2 for a figure illustrating PCR strategy).

AMPure XP beads with a bead to DNA ratio of 0.8:1 were used to remove small PCR products, non-ligated adaptors and primer dimers. The concentrations of the libraries were calculated using the Kapa Biosystems library quantification kit, with average fragment sizes estimated from gel electrophoresis. The amplified libraries were then single-end sequenced using the Illumina MiSeq, with the six barcoded libraries as two triplex runs and the stability libraries as one run in triplex. All libraries were loaded at 10 pM. The barcoded libraries were combined with 40% PhiX (Illumina) and the stability libraries combined with 90% PhiX to increase cluster diversity. For each run, 3.4 μl of the custom Read 1 primer (Additional file 1: Table S1) was added to the Read 1 primer mix of the MiSeq cartridge (Illumina) to enable sequencing of PhiX and to generate reads beginning with the barcoded ISS1. A custom Index Read primer (Additional file 1: Table S1) was also loaded into the MiSeq cartridge according to the manufacturer’s instructions. Fastq only files were generated according to the following settings; TruSeq LT, 1 index read, 76 cycles, adaptor trimming unchecked and custom indexing primer selected.

Analysis of sequencing data

Raw demultiplexed fastq files were analysed using the Bio-TraDIS scripts made available by the Sanger Wellcome Trust Institute [30] ( Initially, the single command pipeline script, bacteria_tradis, was utilised. The pipeline filtered and removed reads according to the transposon tag specified (e.g. CAGAAAACTTTGCAACAGAACC for library CA). After tag removal, the remaining 46 bp of S. equi DNA were mapped to the Se4047 reference genome using SMALT short read mapper, producing a plot file of insertion sites for viewing in the Artemis genome browser [31], and for downstream analysis. The default transposon tag mismatch of 0 was maintained, however a mapping threshold of 100% was set (SMALT parameter y = 1) to improve accuracy and confidence in the assignment of insertion sites. Next, the plot files generated from bacteria_tradis were analysed by tradis_gene_insert_sites, generating a readable document of unique insertion sites, total read counts and insertion indices, per gene. Unique insertion sites represented by two or fewer reads were not included in the analysis. The output file from tradis_gene_insert_sites was then used in tradis_essentiality to determine the essential genome of S. equi. Tradis_essentiality uses the empirically observed bimodal distribution of the insertion indices (essential and non-essential peaks) to fit gamma distributions. Log2 likelihood ratios (LLR) are calculated between the gamma distributions, with genes assigned a LLR of less than -2 identified as essential, more than 2 as non-essential and between the two values as ambiguous [30]. Essential and ambiguous changepoints were calculated from these LLRs to categorise genes into essential, ambiguous and non-essential groups. Essentialities of genes with multiple genomic copies were called as ‘not defined’ due to reduced confidence in read mapping. The fastq files from each library were combined, clipped of their first 2 bp to standardise the ISS1 tag at the beginning of each read and re-analysed to generate a master library, from which final gene essentiality is reported in this study. To identify any insertion site bias, master library mapped reads, with duplicates removed, were parsed through WebLogo, to determine the probability of each nucleotide occurring at positions 1-20 (the insertion site to 20 bp downstream) [32].

Comparative analysis of S. equi TraDIS to S. pyogenes and S. agalactiae Tn-Seq data

Gene essentiality calls of S. pyogenes strain M1T1 5448 and S. agalactiae strain A909 were retrieved from the supplementary information provided by Le Breton et al. and Hooven et al [19, 20]. In these studies, each gene of S. pyogenes and S. agalactiae was reported as essential, critical, non-essential or not defined/non-conclusive. KEGG pathway enrichment was completed on the essential and critical genes of S. pyogenes and S. agalactiae in addition to the essential and ambiguous genes of S. equi, using the gene set enrichment analysis available as an online tool on Genome 2D ( [33]. The KEGG pathways attributed to the essential, critical and ambiguous genes were compared between the three bacteria. Gene orthologues were also identified between Se4047 and S. pyogenes strain MGAS5005 (reference strain used by Le Breton et al. for M1T1 5448), Se4047 and S. agalactiae strain A909 and between S. pyogenes strain MGAS5005 and S. agalactiae strain A909 using the online tool OrtholugeDB ( [34]. The essentiality calls of each orthologous gene pair were compared to determine concordance. All results generated from OrtholugeDB were included in the analysis, except for duplicated calls where multiple copies of a gene exist in either bacterium or when gene essentiality is not defined or non-conclusive.

Results and discussion

Insertion of barcoded pGh9:ISS1 is random, stable and dense in S. equi

To generate our six S. equi mutant libraries, we utilised six barcoded pGh9:ISS1 plasmids. There were no significant differences in the mean doubling time of Se4047 relative to those of the six barcoded libraries (p = 0.48) (see Additional file 1 for average growth curve plot). Transposition frequencies of between 3.5-3 and 7.8-3 were observed across the six barcoded libraries, which is comparable to the frequency of 4.9-3 reported by Magiun et al. where pGh9:ISS1 was transposed into Lactococcus lactis (L. lactis) strain IL1403. The transposition frequency of pGh9:ISS1 in S. equi was also comparable to that of the transposon, Krmit, in S. pyogenes (4-3) [14], but was higher than Himar1, a mini-transposon, in S. agalactiae (10-4–10-6) [13]. In common with previous studies that identifed ISS1 transposition sites [14, 28], no specific sequence motif was observed at the transposition sites of ISS1 in S. equi (Fig. 1). The probability of either an A or a T occuring at any position between the insertion site and 20 bp downstream, was between 54% to 70% per bp highlighting a modest preference of ISS1 for AT rich regions, which is in agreement with the overall AT content of the S. equi genome (58.7%) [2].
Fig. 1

WebLogo of ISS1 insertion sites in S. equi. Data from six barcoded ISS1 mutant libraries in S. equi were combined to generate a master library. Unique sequence reads were isolated from the master library data set and parsed through WebLogo [32] to identify any insertion site bias between the insertion site and 20 bp downstream. No insertion site bias was found

To determine the stability of pGh9:ISS1 transposition, 95 colonies from library CA were pooled (P0) and passaged twice. Sequencing of P0 identified 95 insertion sites, representing 84 genes. Ninety-five insertion sites were also identified in P1, in the same 84 genes, except that an additional mutant was identified in SEQ_1253 and a SEQ_0705 mutant was lost. For P2, 92 mutants were identified, representing 83 of the same genes. The SEQ_1253 mutant gained in P1 was lost, in addition to two other mutants in SEQ_1270 and SEQ_1697. The gain then loss of a mutant in SEQ_1253 is likely due to sample preparation/sequencing differences with the remaining losses due to fitness effects following transposition of ISS1. Our data support the stability of pGh9:ISS1 in the S. equi genome and provide evidence that any onward translocation of pGh9:ISS1 post-transposition occurs at an undetectable level.

Our technique for the generation of transposon libraries, in common with the PIMMS method utilised for the identification of ISS1 insertion sites in S. uberis [14], does not attempt to eliminate the plasmid after transposition. ISS1 duplicates on transposition generating a copy of pGh9, flanked on both sides by ISS1, resulting in the presence of undesirable ISS1-plasmid fragments in library DNA [28]. PIMMS employs an inverse PCR of re-circularised DNA fragments to identify genomic sequences flanking ISS1 insertion sites [14]. Our TraDIS approach utilised Y-adapters to specifically amplify from ISS1 generating both ISS1-plasmid and ISS1-genome fragments (see Additional file 2 for a figure illustrating PCR strategy). Incubation of Y-adaptor ligated DNA with SmaI before PCR cleaved ISS1-plasmid fragments, such that these undesirable sequence reads accounted for only 5 to 10% of the final dataset. Thirteen SmaI restriction sites are present in the Se4047 genome and it is predicted that sequence reads mapping to the immediate regions surrounding these sites will similarly be lost from the final TraDIS data set. An alternative restriction enzyme may be preferred for use in other bacteria. A list of restriction enzymes that cut within 200 bp of the 5’ end of ISS1 is provided in (Additional file 1: Table S2).

The fastq files from each barcoded library were combined and reanalysed to generate a master library (Table 1). The master library represents sequencing data from two MiSeq runs, from which 37.6 million reads were obtained. Reads that contained the desired ISS1 tag totalled 32.6 million of which 17.2 million (53%) mapped with 100% identity to the Se4047 genome. ISS1-plasmid reads account for some of the unmapped reads, however the majority are likely attributable to insufficient mapping quality using the high stringency criteria described above or through mapping to repetitive sequences within the S. equi genome [2].
Table 1

Summary of TraDIS data obtained from sequencing six barcoded S. equi mutant libraries, generated with pGh9:ISS1


Unique insertion sites in genes

Total reads in genes

Genes containing insertions (% of total genes)

Library saturation (insertion every n bp in genes)




1,787 (87.6)





1,804 (88.5)





1,792 (87.9)





1,797 (88.1)





1,804 (88.5)





1,815 (89)





1,935 (94.9)


Data from the six libraries were combined to generate the master library

On average, the master library contained an insertion every 9 bp in genes, representing a 79% increase in saturation when compared to insertions in the individual barcoded libraries. This considerable increase in library saturation did not greatly increase the number of genes represented in the master library, which was an average of 6.6% more than was found in the individual barcoded libraries. Our data demonstrate that ISS1 transposition occurred reproducibly across the S. equi genome regardless of the barcoded ISS1 that was used.

The widespread distribution of ISS1 transposition is evident from Fig. 2a, which shows common regions of increased and decreased transposition (insertion index (number of unique insertions/size of the gene)) across the six libraries. A low insertion index was observed in genes encoding ribosomal proteins, with increased insertion indices evident in regions of low GC content for example in the integrative conjugative element (ICE) Se1 (ICESe1) and ICESe2 (Fig 2a). The pooling of data to generate the master library was supported by the increased interquartile range observed in Fig. 2b. Pooling the data elevated the lower quartile range increasing the robustness of the data set from which gene essentiality was determined.
Fig. 2

Insertion indices of S. equi genes disrupted by barcoded pGh9:ISS1. a. Insertion indices (log10) per gene is replicable between the six barcoded libraries. Each library is identified by its barcode on the right of the figure. The data was combined to generate a master library. Common peaks and troughs are evident; a decreased insertion index is clear in all libraries in a region of ribosomal proteins, with peaks in the integrative conjugative elements ICESe1 and ICESe2 visible. b. Box and whisker plot of the insertion indices of each barcoded library and the master library. The pooling of data to generate the master library was supported by the increased interquartile range and the elevated lower quartile range, increasing the robustness of the data set from which gene essentiality was determined. Average insertion indices from master library data in a region of ribosomal proteins, ICESe1 and ICESe2 are shown

The essential genome of S. equi is comparable to that of group A and B streptococci

Analysis of the master library with the tradis_essentiality TraDIS toolkit script [30] identified essential, ambiguous and non-essential genes based on the insertion index attributed to each gene. The tradis_essentiality script calculates the essential and ambiguous changepoints, from which gene essentiality is categorised. For the master data set, the essential and ambiguous changepoints were 0.0314 and 0.0408, respectively. Diagnostic plots and the results files generated by tradis_essentiality are available in Additional files 3 and 4. Using these thresholds, 19.5% of the Se4047 genome was found to be essential, 1.2% ambiguous, 73.4% non-essential and 5.8% not defined. The proportion of essential genes in Se4047 is similar to the 12% and 13.5% essential genes in S. pyogenes [14] and S. agalactiae [13], respectively. The essential gene sets for Se4047 were compared to those reported for S. pyogenes M1T1 5448 [20] and S. agalactiae A909 [19] (Additional file 5). There was 90.2% concordance of gene essentiality between S. equi and S. pyogenes; 89.4% between S. equi and S. agalactiae; 90.9% between S. pyogenes and S. agalactiae and 83.7% between the three species (Fig. 3). Our data highlight the similarities of the functional genomes of these different pathogens in support of previous studies that identified shared core and accessory genomes [2, 3]. In each species, libraries were generated using different transposons, prepared and analysed in different ways and yet identified common essential gene sets. Our data illustrate the compatibility of these methodologies and the reproducibility of essentiality assignments across these streptococci.
Fig. 3

Gene essentiality concordance between a Group A, B and C streptococci. Essentiality between orthologous gene pairs in S. equi, S. pyogenes and S. agalactiae were compared. Orthologues were classified as either essential/critical/ambiguous concordant (E/C/A) or non-essential (NE) concordant. Non-concordances are also shown for 2-species comparisons only

The biosynthetic pathways attributed to each species’ essential/critical/ambiguous gene set were identified by KEGG pathway analysis. Our analysis revealed that the essential/critical/ambiguous genes of S. equi, S. pyogenes and S. agalactiae were attributed to 45, 41 and 41 KEGG categories, respectively, 39 of which were shared between the three species (Fig. 4a) (Additional file 6). The 10 most prevalent essential/critical/ambiguous KEGG pathways in each species were compared (Fig. 4b). The highest-ranked categories were involved in key cellular processes such as aminoacyl-tRNA biosynthesis, purine and pyrimidine metabolism, glycolysis and gluconeogenesis, the pentose phosphate pathway and peptidoglycan biosynthesis. The top KEGG categories in each species were consistent with one another. However, the S. equi essential genome contained noticeably more genes implicated in purine and pyrimidine biosynthesis. This may reflect the larger essential gene set of Se4047 or may be attributed to the in vitro conditions in which our libraries were grown. A potential lack of purines and pyrimidines within our Todd Hewitt media could provide an alternative explanation for these findings. Interestingly, a broad transcriptional regulator codY [35], reported as non-essential in S. agalactiae [19] was found to be ambiguous and critical in S. equi and S. pyogenes [20], respectively.
Fig. 4

KEGG analysis of the essential/critical/ambiguous genes of Group A, B and C streptococci. a. Venn diagram showing the comparison of the KEGG categories assigned to the essential/critical/ambiguous genes of S. equi, S. pyogenes and S. agalactiae. The overlap of genes concludes that the essential pathways employed by the three different species are conserved. b. Barchart of the calls within most highly ranked KEGG pathways. The top KEGG categories in each species were consistent with one another

Novel features of the S. equi essential gene set

Although the majority of essential genes in S. equi were similarly important in S. pyogenes and S. agalactiae, our analysis also identified some essential genes that were restricted to Se4047. S. equi produces a secreted molecule provisionally named equibactin, which aids the acquisition of iron in vitro [36] and is required for the full virulence of S. equi in a susceptible natural host [37]. Equibactin is synthesised by a non-ribosomal peptide synthesis system encoded in an operon (eqbB to eqbN) on the integrative conjugative element ICESe2 (Fig. 5a), which is regulated by the iron-dependent transcriptional repressor, EqbA [2, 36]. None of the genes eqbB to N were identified as essential in S. equi, in agreement with the free availability of iron in Todd-Hewitt media [29]. However, eqbA was essential for growth in vitro (Fig. 5b). Our results concur with those of Heather et al. who found that deletion of eqbA led to a slow-growth phenotype that was caused by excessive import of iron following de-regulation of the equibactin operon [36].
Fig. 5

Sequence data from S. equi mutant libraries generated with ISS1. a. Overview of the integrative conjugative element, ICESe2. Green and red peaks indicate reads mapping on the forward and reverse strand, respectively. ISS1 insertion is dense in the region, except in two distinct genes, eqbA and SEQ_1258. The labels indicate the areas zoomed into in B and C of the figure. b. eqbE to SEQ_1247. ISS1 insertion is dense, except for in eqbA, the regulator of the equibactin locus. Equibactin aids the acquisition of iron, which if unregulated leads to excessive iron import and a slow growth phenotype. c. SEQ_1254 to SEQ_1262. ISS1 insertion is dense, except for in SEQ_1258, a putative antitoxin. ICESe2 encodes a putative toxin-antitoxin system which maintains the ICE in the bacterial genome. Both eqbA and SEQ_1258 were identifed as essential genes. Data is viewed in window size 9 for (a) and 3 for (b) and (c), with a maximum display value of 100 reads for ease of viewing. Data is viewed in Artemis [31]

ICESe2 also contained a second essential gene, SEQ_1258 (Fig 3b). SEQ_1258 and SEQ_1257 are predicted to encode a novel toxin-antitoxin system in S. equi [2]. Toxin-antitoxin systems comprise a stable toxin and a labile antitoxin, which promote the maintenance of the element on which they are encoded within the bacterial genome [38]. Our data suggest that SEQ_1258 encodes the antitoxin in this system (Fig. 5c). The gene encoding the MosA antitoxin of the integrative conjugative element, SXT, of Vibrio cholerae was found to be essential, while mosT, encoding the toxin component could be deleted [38]. Recircularised extra-chromasomal copies of ICESe2 could not be recovered from Se4047 [36]. One possible explanation for this finding is that recircularisation of ICESe2 halts the production of the labile antitoxin, which cannot then neutralise the stable toxin still present in the cell. S. equi and Streptococcus zooepidemicus (S. zooepidemicus) share over 97% genetic identity [2], yet ICESe2 is not present in any strains of S. zooepidemicus studied to date [36, 39]. The maintenance of ICESe2 by its toxin-antitoxin system may restrict it to S. equi. Interestingly, decay of the equibactin locus in some S. equi isolates that were recovered from persistently infected horses did not include decay in SEQ_1258 [30], in agreement with the importance of the antitoxin as measured by TraDIS.


We have successfully customised a barcoded TraDIS method based on the original method developed by Langridge et al. [10]. The barcoded TraDIS technique described herein is easily transferrable between laboratories and is conducted using accessible Illumina sequencer protocols, without the need for software alteration. Our barcoded technique will be of value to other researchers as it could be easily applied to other transposon systems for the study of a wide range of pathogenic bacteria. TraDIS and other transposon directed methods, represent a major advance in the study of gene function in bacteria. Utilising dense mutant libraries yields significant time and cost savings over the generation of traditional knockout strains, not only due to the speed at which saturated libraries can be generated, but also due to the ability to simultaneously identify conditionally essential genes. The use of barcoded pGh9:ISS1 plasmids to generate mutant libraries of S. equi has provided a highly useful tool for the progression of TraDIS studies in this important bacterium. In particular, the ability to combine barcoded mutant libraries, challenge animals and then deconvolute the data generated has the potential to minimise the effects of animal to animal variation, enhance data quality and reduce the total number of animals required in future studies in accordance with the principles of the 3Rs; replacement, reduction and refinement [40]. Data from such in vivo studies, which are possible in the natural host, will provide an unprecedented insight into the pathways that underpin the virulence of S. equi, which will help to direct future vaccine research.

The shared essential gene set of group A, B and C streptococci provides further evidence of the close relationships of these important pathogenic bacteria. Our data suggest that the determination of gene essentiality for S. equi in the natural equine host is likely to also shed light on pathways of importance to the virulence of other streptococci. Therefore, this ABC of essential genes provides a solid foundation upon which to begin the process of reading the functional genomes of streptococci.



replacement, reduction and refinement


Integrative conjugative element


Kyoto encyclopaedia of genes and genomes


Next-generation sequencing


Todd-hewitt agar


Todd-hewitt agar with erythromycin


Todd hewitt broth


Todd-hewitt broth with erythromycin


Sequencing of transposon-genome junctions


Transposon directed-insertion site sequencing



Valentina Stefanetti, University of Perugia, isolated the 95 individual colonies for library stability testing, for which we are grateful. We would also like to acknowledge Karen Steward, Animal Health Trust, for initial laboratory training of ARLC, general supervision and assistance.


This work was funded through a pump prime grant from the PetPlan Charitable Trust (ref: S14-51) and a follow-on grant from the Horse Trust (ref: G4104). ARLC receives a stipend from the University of Cambridge Doctoral Training Partnership scheme, which is funded by the BBSRC (ref: 1503883).

Availability of data and materials

Raw Illumina fastq files have been made available in Genbank at the sequence read archive (SRA accession number: SRP102055). Outputs files of tradis_essentiality and from orthologue and KEGG analysis are available in the Additional files.

Authors’ contributions

ARLC completed all experiments, analysed and interpreted data, and drafted the manuscript. OPF contributed to the initial method design, including sequencing and oversaw library preparation. AKC advised and assisted in the design of the barcoded pGh9:ISS1 plasmids and plasmid depletion step. CR oversaw the generation of the barcoded pGh9:ISS1 plasmids and library generation. GN and MB implemented the bioinformatics tools and assisted with their use. JP, JAL and DJM advised on the initial study design, the concept of utilising ISS1 for TraDIS and the optimisation of sequencing strategies. ASW conceived TraDIS in S. equi, oversaw the project and assisted in interpreting the data. AKC, CR, JAL and ASW contributed to the final manuscript. All authors read and approved the final manuscript.

Competing interests

The authors declare they have no competing interests.

Consent for publication

Not applicable.

Ethics approval and consent to participate

Not applicable.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

Authors’ Affiliations

Animal Health Trust, Lanwades Park
Department of Veterinary Medicine, University of Cambridge
The Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus
Department of Chemistry and Biomolecular Sciences, Macquarie University
The School of Veterinary Medicine and Science, University of Nottingham


  1. Farrow JAE, Collins MD. Taxonomic studies on Streptococci of serological Groups C, G and L and possibly related taxa. Syst Appl Microbiol. 1984;4:483–93.View ArticleGoogle Scholar
  2. Holden MT, Heather Z, Paillot R, Steward KF, Webb K, Ainslie F, Jourdan T, Bason NC, Holroyd NE, Mungall K, Quail MA, Sanders M, Simmonds M, Willey D, Brooks K, Aanensen DM, Spratt BG, Jolley KA, Maiden MC, Kehoe M, Chanter N, Bentley SD, Robinson C, Maskell DJ, Parkhill J, Waller AS. Genomic evidence for the evolution of Streptococcus equi: host restriction, increased virulence, and genetic exchange with human pathogens. PLoS Pathog. 2009;53, e1000346.View ArticleGoogle Scholar
  3. Lefébure T, Richards VP, Lang P, Pavinski-Bitar P, Stanhope MJ. Gene repertoire evolution of Streptococcus pyogenes inferred from phylogenomic analysis with Streptococcus canis and Streptococcus dysgalactiae. PLoS ONE. 2012;75:e37607.View ArticleGoogle Scholar
  4. Olsen RJ, Musser JM. Molecular pathogenesis of necrotizing fasciitis. Annu Rev Pathol. 2010;5:1–31.View ArticlePubMedGoogle Scholar
  5. Cunningham MW. Pathogenesis of group A streptococcal infections and their sequelae. Adv Exp Med Biol. 2008;609:29–42.View ArticlePubMedGoogle Scholar
  6. Katz AR, Morens DM. Severe streptococcal infections in historical perspective. Clin Infect Dis. 1992;141:298–307.View ArticleGoogle Scholar
  7. Gibbs RS, Schrag S, Schuchat A. Perinatal infections due to group B streptococci. Obstet Gynecol. 2004;1045(Pt 1):1062–76.View ArticleGoogle Scholar
  8. Bohnsack JF, Whiting AA, Martinez G, Jones N, Adderson EE, Detrick S, Blaschke-Bonkowsky AJ, Bisharat N, Serotype GM, III. Streptococcus agalactiae from bovine milk and human neonatal infections. Emerg Infect Dis. 2004;108:1412–9.View ArticleGoogle Scholar
  9. Amal MN, Zamri-Saad M, Siti-Zahrah A, Zulkafli AR, Nur-Nazifah M. Molecular characterization of Streptococcus agalactiae strains isolated from fishes in Malaysia. J Appl Microbiol. 2013;1151:20–9.View ArticleGoogle Scholar
  10. Langridge GC, Phan MD, Turner DJ, Perkins TT, Parts L, Haase J, Charles I, Maskell DJ, Peters SE, Dougan G, Wain J, Parkhill J, Turner AK. Simultaneous assay of every Salmonella Typhi gene using one million transposon mutants. Genome Res. 1912;2009:2308–16.Google Scholar
  11. van Opijnen T, Bodi KL, Camilli A. Tn-seq: high-throughput parallel sequencing for fitness and genetic interaction studies in microorganisms. Nat Methods. 2009;610:767–72.View ArticleGoogle Scholar
  12. Goodman AL, McNulty NP, Zhao Y, Leip D, Mitra RD, Lozupone CA, Knight R, Gordon JI. Identifying genetic determinants needed to establish a human gut symbiont in its habitat. Cell Host Microbe. 2009;63:279–89.View ArticleGoogle Scholar
  13. Gawronski JD, Wong SM, Giannoukos G, Ward DV, Akerley BJ. Tracking insertion mutants within libraries by deep sequencing and a genome-wide screen for Haemophilus genes required in the lung. Proc Natl Acad Sci U S A. 2009;10638:16422–7.View ArticleGoogle Scholar
  14. Blanchard AM, Egan SA, Emes RD, Warry A, Leigh JA. PIMMS (Pragmatic Insertional Mutation Mapping System) Laboratory Methodology a Readily Accessible Tool for Identification of Essential Genes in Streptococcus. Front Microbiol. 2016;7:1645.View ArticlePubMedPubMed CentralGoogle Scholar
  15. Barquist L, Boinett CJ, Cain AK. Approaches to querying bacterial genomes with transposon-insertion sequencing. RNA Biol. 2013;107:1161–9.View ArticleGoogle Scholar
  16. Dembek M, Barquist L, Boinett CJ, Cain AK, Mayho M, Lawley TD, Fairweather NF, Fagan RP. High-throughput analysis of gene essentiality and sporulation in Clostridium difficile. MBio. 2015;62:e02383–14.Google Scholar
  17. Wong YC, Abd El Ghany M, Naeem R, Lee KW, Tan YC, Pain A, Nathan S. Candidate Essential Genes in Burkholderia cenocepacia J2315 Identified by Genome-Wide TraDIS. Front Microbiol. 2016;7:1288.PubMedPubMed CentralGoogle Scholar
  18. Pechter KB, Gallagher L, Pyles H, Manoil CS, Harwood CS. Essential Genome of the Metabolically Versatile Alphaproteobacterium Rhodopseudomonas palustris. J Bacteriol. 1985;2015:867–76.Google Scholar
  19. Hooven TA, Catomeris AJ, Akabas LH, Randis TM, Maskell DJ, Peters SE, Ott S, Santana-Cruz I, Tallon LJ, Tettelin H, Ratner AJ. The essential genome of Streptococcus agalactiae. BMC Genomics. 2016;17:406.View ArticlePubMedPubMed CentralGoogle Scholar
  20. Le Breton Y, Belew AT, Valdes KM, Islam E, Curry P, Tettelin H, Shirtliff ME, El-Sayed NM, McIver KS. Essential Genes in the Core Genome of the Human Pathogen Streptococcus pyogenes. Sci Rep. 2015;5:9838.View ArticlePubMedPubMed CentralGoogle Scholar
  21. Moule MG, Hemsley CM, Seet Q, Guerra-Assunção JA, Lim J, Sarkar-Tyson M, Clark TG, Tan PB, Titball RW, Cuccui J, Wren BW. Genome-wide saturation mutagenesis of Burkholderia pseudomallei K96243 predicts essential genes and novel targets for antimicrobial development. MBio. 2014;51:e00926–13.Google Scholar
  22. van Opijnen T, Camilli A. A fine scale phenotype-genotype virulence map of a bacterial pathogen. Genome Res. 2012;2212:2541–51.View ArticleGoogle Scholar
  23. Chaudhuri RR, Morgan E, Peters SE, Pleasance SJ, Hudson DL, Davies HM, Wang J, van Diemen PM, Buckley AM, Bowen AJ, Pullinger GD, Turner DJ, Langridge GC, Turner AK, Parkhill J, Charles IG, Maskell DJ, Stevens MP. Comprehensive assignment of roles for Salmonella typhimurium genes in intestinal colonization of food-producing animals. PLoS Genet. 2013;94:e1003456.View ArticleGoogle Scholar
  24. Subashchandrabose S, Smith S, DeOrnellas V, Crepin S, Kole M, Zahdeh C, Mobley HL. Acinetobacter baumannii Genes Required for Bacterial Survival during Bloodstream Infection. mSphere 2016;11:e00013–15.Google Scholar
  25. Grant AJ, Oshota O, Chaudhuri RR, Mayho M, Peters SE, Clare S, Maskell DJ, Mastroeni P. Genes Required for the Fitness of Salmonella enterica Serovar Typhimurium during Infection of Immunodeficient gp91-/- phox Mice. Infect Immun. 2016;844:989–97.View ArticleGoogle Scholar
  26. Moule MG, Spink N, Willcocks S, Lim J, Guerra-Assunção JA, Cia F, Champion OL, Senior NJ, Atkins HS, Clark T, Bancroft GJ, Cuccui J, Wren BW. Characterization of New Virulence Factors Involved in the Intracellular Growth and Survival of Burkholderia pseudomallei. Infect Immun. 2015;843:701–10.Google Scholar
  27. Gutierrez MG, Yoder-Himes DR, Warawa JM. Comprehensive identification of virulence factors required for respiratory melioidosis using Tn-seq mutagenesis. Front Cell Infect Microbiol. 2015;5:78.View ArticlePubMedPubMed CentralGoogle Scholar
  28. Maguin E, Prévost H, Ehrlich SD, Gruss A. Efficient insertional mutagenesis in lactococci and other gram-positive bacteria. J Bacteriol. 1996;1783:931–5.View ArticleGoogle Scholar
  29. Hamilton A, Robinson C, Sutcliffe IC, Slater J, Maskell DJ, Davis-Poynter N, Smith K, Waller A, Harrington DJ. Mutation of the maturase lipoprotein attenuates the virulence of Streptococcus equi to a greater extent than does loss of general lipoprotein lipidation. Infect Immun. 2006;7412:6907–19.View ArticleGoogle Scholar
  30. Barquist L, Mayho M, Cummins C, Cain AK, Boinett CJ, Page AJ, Langridge GC, Quail MA, Keane JA, Parkhill J. The TraDIS toolkit: sequencing and analysis for dense transposon mutant libraries. Bioinformatics. 2016;327:1109–11.View ArticleGoogle Scholar
  31. Carver T, Harris SR, Berriman M, Parkhill J, McQuillan JA. Artemis: an integrated platform for visualization and analysis of high-throughput sequence-based experimental data. Bioinformatics. 2012;284:464–9.View ArticleGoogle Scholar
  32. Crooks GE, Hon G, Chandonia JM, Brenner SE. WebLogo: a sequence logo generator. Genome Res. 2004;146:1188–90.View ArticleGoogle Scholar
  33. Baerends RJ, Smits WK, de Jong A, Hamoen LW, Kok J, Kuipers OP. Genome2D: a visualization tool for the rapid analysis of bacterial transcriptome data. Genome Biol. 2004;55:R37.View ArticleGoogle Scholar
  34. Whiteside MD, Winsor GL, Laird MR, Brinkman FS. OrtholugeDB: a bacterial and archaeal orthology resource for improved comparative genomic analysis. Nucleic Acids Res. 2013;41(Database issue):D366–76.View ArticlePubMedGoogle Scholar
  35. Sonenshein AL. CodY, a global regulator of stationary phase and virulence in Gram-positive bacteria. Curr Opin Microbiol. 2005;82:203–7.View ArticleGoogle Scholar
  36. Heather Z, Holden MT, Steward KF, Parkhill J, Song L, Challis GL, Robinson C, Davis-Poynter N, Waller AS. A novel streptococcal integrative conjugative element involved in iron acquisition. Mol Microbiol. 2008;705:1274–92.View ArticleGoogle Scholar
  37. Harris SR, Robinson C, Steward KF, Webb KS, Paillot R, Parkhill J, Holden MT, Waller AS. Genome specialization and decay of the strangles pathogen, Streptococcus equi, is driven by persistent infection. Genome Res. 2015;259:1360–71.View ArticleGoogle Scholar
  38. Wozniak RA, Waldor MK. A toxin-antitoxin system promotes the maintenance of an integrative conjugative element. PLoS Genet. 2009;53:e1000439.View ArticleGoogle Scholar
  39. Webb K, Barker C, Harrison T, Heather Z, Steward KF, Robinson C, Newton JR, Waller AS. Detection of Streptococcus equi subspecies equi using a triplex qPCR assay. Vet J. 1953;2013:300–4.Google Scholar
  40. Russell WMS, Burch RL. The principles of human experimental technique. London: Methuen; 1959.Google Scholar


© The Author(s). 2017