Knockout of an outer membrane protein operon of Anaplasma marginale by transposon mutagenesis

Transformation of Anaplasma marginaleby transposon mutagenesis

Attempts to transform A. marginale by transposon mutagenesis using the Himar1 transposon/transposase system delivered in two separate plasmids were not successful. The probability that two plasmids are introduced at once into A. marginale organisms could be very low, especially when viability in the extracellular environment might be highly compromised, resulting in a low fraction of cells competent to take up DNA.

Therefore in order to promote transposon mutagenesis in these bacteria, the transposase was provided in cis with the Himar1 transposon sequences (R. F. Felsheim unpublished data). The pHimarcisA7mCherry-SS contains the hyperreactive allele A7 transposase and the Himar1 TIR flanking the mCherry reporter gene and the aadA gene, which confers resistance against spectinomycin and streptomycin. Expression of the transposase and the reporter and antibiotic selection genes is driven by the A. marginale tr promoter [41, 42] (Figure 1A). Antibiotic selection pressure of electroporated bacteria with this construct resulted in the isolation of red fluorescent and antibiotic resistant bacteria (Figure 1B).

Mapping of transposon insertion within the A. marginalechromosome

We used Roche/454 and Illumina high-throughput genome sequencing to determine: 1) the location of plasmid sequences within the A. marginale chromosome, 2) the recombination mechanism that allowed the segregation of mutant bacteria and 3) if these recombinant organisms correspond to a population containing insertions in different genomic locations or in a single genome site.

Mutations produced by the integration of the Himar1 transposon into the A. marginale chromosome will generate new junction sequences that are absent in the wild-type. These new sequences should include the Himar1 terminal inverted repeats (TIR) followed by the sequence of the regions in which the transposon is integrated. Based on this, the strategy that we used to map the Himar1 insertion site involved alignment of the sequencing reads obtained by Roche/454 and Illumina methods to two reference sequences, the A. marginale str. St Maries genome sequence (CP000030) and the Himar1 TIR sequence. The Himar1 TIR-A. marginale genome junctions were identified by extracting reads that aligned to the A. marginale genome at one end and to the Himar1 TIR at the other end.

Analysis using Illumina reads mapped the Himar1 TIR-A. marginale genome junctions into a region of omp6 and omp10 genes. Interestingly these reads contained the same mutated sequence. The omp6 and omp10 genes share a large stretch of identity 456 nt/459 nt (99%) [43]. The short, 100 nt length of the Illumina reads, made it difficult to differentiate which gene contained the Himar1 transposon. Additional analysis using longer reads obtained on the Roche/454 platform revealed that the Himar1 transposon was integrated within the omp10 gene. These reads contained a region of omp10 that is not shared with omp6. Based on this sequencing analysis the genomic location of the Himar1 transposon in the chromosome of the transformed A. marginale, is at position 245 considering 1 as the first base of the omp10 start codon (Figure 2A).

These results were verified by PCR amplification of gDNA from ISE6 cells infected with wild-type and transformed A. marginale using omp6 and omp10 specific primers (Figure 2A-B). The size of omp6 amplicons (492 bp) in wild-type and transformed A. marginale was the same. However the size of the omp10 amplicon in transformed A. marginale was increased by 1836 bp when compared to the wild-type (969 bp), indicating that the transposon was integrated within the omp10 gene.

The genome sequence of A. marginale str. Virginia is available only as unannotated contigs with gaps. Therefore for our analysis we used the A. marginale str. St Maries genome as reference. For this reason we wanted to confirm that the transposon location in the mutated Virginia strain was the same as the one mapped in the reference genome. For this, combined Roche/454 and Illumina reads were assembled and a contig of 21,324 nucleotides identified. Alignment of this contig with the A. marginale str. St Maries genome showed that this sequence contained part of omp10 and upstream genes (99% identity) (Figure 3) and that the transposon insertion site in the A. marginale str. Virginia matches the same region mapped using the reference genome.

Further analysis of sequencing reads determined that there is only one transposon insertion in the chromosome of recombinant A. marginale. The reads containing the Himar1 TIR-A. marginale junctions aligned to a single genome site. Although these transformed organisms were not cloned, data suggest that they are isogenic for the transposon insertion site within the omp10 gene.

The mobilization of the Himar1 transposon from one locus to another is mediated by a transposase using a cut and paste mechanism [27, 30]. It has been shown previously in other organisms that the Himar1 transposon integrates preferentially into a TA site and leads to duplication of this dinucleotide upon integration into the target site [30]. This was found to be true also for A. marginale. Sequencing analysis revealed that the Himar1 transposon targeted a TA dinucleotide in omp10 (Figure 4A) and upon integration it is flanked by a TA dinucleotide sequence (Figure 4B). Thus, the mobilization of the Himar1 transposon into the omp10 gene of A. marginale was mediated by means of the A7 transposase in a cut and paste mechanism. This transformant of A. marginale will be referred to as omp10::himar1 mutant.

Evidence for expression of omp10as part of an operon

We hypothesize that the transposon insertion could alter the expression of omp10 and downstream genes. This hypothesis is based on recent work in which the transcriptome profile of A. marginale using RNAseq indicated that omp10 is expressed as part of a six-gene operon in erythrocytes of infected cattle [25]. This operon includes AM1225, omp10, omp9, omp8, omp7 and omp6 (Figure 5A).

Because of this, we wanted to determine if omp10 is expressed within a polycistronic message in A. marginale replicating in ISE6 tick cells. The intergenic region between AM1225 and omp10 is 440 bp long, while intergenic regions between omp10-9, omp9-8, omp8-7 and omp7-6 are 62 bp, 63 bp, 64 bp and 36 bp respectively (Figure 5A). To test whether AM1225 through omp7 are expressed as a single transcriptional unit, total RNA isolated from ISE6 cells infected with wild-type A. marginale was reverse transcribed and template cDNA was used for amplification of intergenic regions with primers that connect neighboring genes (Figure 5A). The omp6 gene was not included in these experiments, because previous work [43] and work in our lab showed that transcripts from this gene are not detected in A. marginale during infection of tick cells. Appropriate size amplicons of the intergenic regions between omp7-8, omp8-9, omp9-10 and omp10-AM1225 gene were detected (Figure 5B), providing evidence that these genes are transcribed as a single mRNA in A. marginale infected tick cells.

RNA transcript analysis

Next, we determined if insertion of the Himar1 sequences resulted in alteration of omp10 expression and the expression of genes downstream. For this, total RNA from ISE6 tick cells infected with A. marginale wild-type and omp10::himar1 mutant was reverse transcribed and cDNA used as template for PCR amplification with specific primers that were designed to anneal to omp6, omp7, omp8, omp9, and omp10 in wild-type and omp10::himar1 mutant respectively (Figure 6A). The omp10, 9, 8, and 7 genes, but not omp6, are transcriptionally active in wild-type A. marginale, although at low levels (Figure 6B). The Himar1 transposon insertion into the coding sequence of omp10, disrupted its expression and that of omp9, omp8, and omp7 since transcripts from these genes were not detected in omp10::himar1 mutants of A. marginale by this method (Figure 6B). To ensure integrity, cDNA samples from A. marginale wild-type and omp10::himar1 mutant were used for amplification with specific primers of a region of 131 bp of the 16S rRNA. Amplicons from this region were detected in both wild-type and omp10::himar1 mutant. No bands were visualized in negative controls (Figure 6B).

Since omp10 through omp7 are expressed at low levels in ISE6 tick cells, RT-qPCR was used to quantitatively determine differences of expression between A. marginale wild-type and omp10::himar1 mutant. For this, cDNA generated from ISE6 tick cells infected with A. marginale wild-type and omp10::himar mutant was used for real time PCR amplification using primers and probes targeting omp8, omp9, and the 3’ and 5’ ends of omp10 (Figure 7A).

In order to compare these gene expression results between wild-type and omp10:himar1 A. marginale, Ct values were normalized to the rpoH, msp5 and 16S rRNA genes. Changes in expression of these genes were calculated by the 2^-ΔΔCt method, and results were expressed as percentage of expression, with a 100% expression level being assigned to the calibrator or control group, which in this case is wild-type A. marginale.

Although three different reference genes were used, RT-qPCR data normalization led to similar results in which there was a significantly reduced expression for omp8 (97–99%), omp9 (90–99%) and omp10 3’ end (85-98%) relative to their counterparts in wild-type A. marginale (Figure 7B). These results show that Himar1 transposon insertion into omp10 affected its expression and the expression of genes downstream, confirming the results obtained by RT-PCR and agarose gel electrophoresis. A second experiment investigated the possibility of the same effect occurring in regions of omp10 before the Himar1 transposon insertion site. For this, a primer and probe set was designed to anneal with a region at the 5’ end of omp10 (Figure 7A). Even though there was a significant reduction in the detection of transcripts from this region (27-57%) relative to the 5’ end of omp10 in wild-type, this reduction was not as great as with the sequences located in omp10 downstream of the Himar1 transposon insertion site.

Western immunoblot analysis

To determine if the decreased expression of mRNA in genes downstream of omp10 correlated with protein expression a Western immunoblot analysis using anti Omp9 antibody was performed.

To compare the protein expression of omp9 between A. marginale omp10::himar1 and wild-type, the number of organisms per sample was quantified by qPCR using the opag2 single copy gene to determine the copy number of A. marginale. Equal amounts (10⁸) of organisms of A. marginale wild-type and omp10::himar1 mutant were loaded per lane. A. marginale str. Virginia initial bodies and uninfected ISE6 cells were used as positive and negative controls respectively.

Western immunoblot showed a reduced expression of Omp9 in omp10::himar1 A. marginale mutant compared with wild-type (Figure 8A). The Omp9 band of 40 kDa was present in wild-type and initial bodies but was not detected in the mutant or using negative control antibody Tryp1E1 (Figure 8B). Antibody F16C1 that reacts with major surface protein 5 (Msp5) was used as a loading control. Anti-Msp5 detected this protein (19 kDa) in wild-type and omp10::himar1 A. marginale (Figure 8C).

These results correlated with results obtained from the RNA transcript analysis, showing that the transposon insertion severely affected the expression of both mRNA and protein from downstream genes such as omp9.

A. marginalecultivation

Cultures of A. marginale str. Virginia wild-type and omp10::himar1 mutant were maintained in tick ISE6 cells derived from embryonated eggs of the blacklegged tick, Ixodes scapularis at 34°C in non-vented 25-cm² cell culture flasks (NUNC). A. marginale-infected cell cultures were maintained in L15B300 medium supplemented with 5% fetal bovine serum (FBS, BenchMark, Gemini Bio-Products), 5% tryptose phosphate broth (TPB, Difco, Becton Dickinson), 0.1% bovine lipoprotein concentrate (LPC, MP-Biomedical), 0.25% NaHCO₃, and 25 mM HEPES buffer, adjusted to pH 7.8, as previously described [49]. The cell culture medium for ISE6 cells infected with the A. marginale omp10::himar1 mutant was supplemented with spectinomycin (Sigma Aldrich) and streptomycin (Sigma Aldrich) to a final concentration of 50 μg/ml each.

Isolation of the A. marginalemutant by transposon mutagenesis

To maximize chances of obtaining a transformant using transposon mutagenesis, we used a single plasmid construct that encoded both the transposon and the transposase in cis configuration as described [50], except that the fluorescent marker was replaced by sequences encoding a monomeric red fluorescent protein, mCherry[51] (Figure 1A). A. marginale bacteria passaged 53 times in ISE6 cells were harvested from one 25-cm² culture in 5 ml of medium when ~80% of cells were infected, and many cells were undergoing lysis. The cells were recovered in 2 ml of culture medium, and added to a 2-ml microcentrifuge tube containing 0.3 ml of sterile silicon carbide abrasive (60/90 grit; Lortone, Inc), vortexed at maximum speed for 30 sec, and the lysate transferred to a fresh 2-ml tube on ice. Bacteria were collected by centrifugation at 11,000 g for 10 min at 4°C, and washed twice in ice-cold 300 mM sucrose. They were then resuspended in 50 μl of 300 mM sucrose containing 3 μg of plasmid DNA, and incubated on ice for 15 min before being electroporated (Biorad Gene Pulser II) at 2 kV, 400 Ohm and 25 μF in a 0.2 cm gap cuvette. The electroporation mixture was recovered in 1.5 ml of an ISE6 cell suspension (~2×10⁶ cells), and centrifuged in a microcentrifuge tube at 1,000 g for 10 min at room temperature. The tube was left undisturbed for 30 min at room temperature, and the pellet then resuspended in the supernatant medium and added to a 25-cm² flask containing ~5×10⁶ ISE6 cells in 3 ml of L15B300 medium supplemented as described for Anaplasma-infected cultures. The culture was incubated at 34°C in a tightly capped flask. Three days after electroporation, the culture medium was replaced with 5 ml of medium additionally containing 50 μg/ml of spectinomycin and streptomycin (selection medium). Subsequently, the culture was fed twice weekly with selection medium and examined weekly on an inverted microscope (Diaphot, Nikon) fitted for epifluorescence using a Texas Red filter. The first fluorescent colonies of bacteria were noted 6 wk following electroporation, and the culture was maintained in selection medium with twice-weekly medium changes until ~90% of cells were infected. At that time, the mutant was passaged (ten-fold dilution) to fresh cells, and the remainder was stored in liquid nitrogen.

Preparation of host cell-free A. marginale wild-type and omp10::himar1 mutant from ISE6 tick cells

Isolation of A. marginale wild-type and omp10::himar1 mutant was performed by disruption of ISE6 tick cells with 1 mm diameter glass beads (BioSpec Technologies) in a Minibead beater (BioSpec technologies) as described elsewhere [52], with the exception that cells were shaken only once for 10s and immediately placed on ice. Cell lysates were transferred to 1.5 ml centrifuge tubes and centrifuged at 100 g for 5 min at 4°C to pellet cell debris. The supernatant was then carefully removed and transferred to clean 1.5 ml centrifuge tubes. A. marginale organisms (wild-type and omp10::himar1 mutant) were pelleted at 11,000 g for 10 min at 4°C, and stored at −20°C.

DNA isolation and Phi29 amplification of the A. marginale omp10::himar1mutant

Before DNA isolation, pelleted A. marginale omp10::himar1 mutants were treated with RNaseA (QIAGEN) and DNase I (Sigma Aldrich) to remove ISE6 host cell contaminant nucleic acids. DNA isolation was performed using the QIAamp DNA Mini kit (QIAGEN) as per manufacturer’s instructions, but in this case the DNA was eluted in 50 μl of 1 mM Tris pH 9.0. DNA concentration was determined using the Qubit dsDNA HS assay kit (Life technologies) on a Qubit fluorometer (Life technologies). 5 reactions of 10 ng of DNA were used for whole genome amplification using the Genomi Phi V2 DNA amplification kit (GE Healthcare) according to manufacturer’s instructions. Following amplification, aliquots were pooled together and the DNA purified with GelElute Extraction Kit (5 PRIME) by adsorption to silica particles and eluted with 10 mM Tris pH8.2.

Genome sequencing and bioinformatics

Samples from 2.0 to 3.6 μg of amplified DNA derived from the omp10::himar1 mutant, were provided for library construction and sequencing by the Roche/454 (GS-FLX) method to the Interdisciplinary Center for Biotechnology Research (ICBR) at the University of Florida. Also, samples of equivalent amounts were provided to the Scripps Research Institute, La Jolla, California for sequencing by the Illumina (HiSeq) method.

A total of 374,151 and 207,288,916 reads of Roche/454 and Illumina sequencing data, respectively, were obtained. The FASTQ files provided by the sequencing facilities were uploaded to the UF GALAXY web site http://galaxy.hpc.ufl.edu, and analyzed separately.

Uploaded Illumina FASTQ files were groomed, filtered and formatted into FASTA files using the FASTQ Groomer, Filter FASTQ and FASTQ to FASTA converter tools located in the NGS: QC and manipulation toolbox of GALAXY. FASTA files were then aligned to the A. marginale str. St Maries reference genome sequence (CP000030) using the Megablast alignment tool (NCBI BLAST + blastn (version 0.0.12) in GALAXY) to obtain sequencing reads that contained A. marginale sequences.

These A. marginale sequencing reads were then used for a second Megablast alignment using as a reference sequence 28 nucleotides from the Himar1 terminal inverted repeats (TIR). The transposon insertion locus within the A. marginale chromosome was then determined, since the reads obtained contained the A. marginale-Himar1 TIR junctions.

A similar strategy was used for the analysis of the Roche/454 sequencing reads. CLC genomics workbench, version 6.5 was used for assemblies of Roche/454 and Illumina reads.

RNA isolation

For RNA isolation, three samples of ISE6 cells infected with A. marginale wild-type and three omp10::himar1 samples were used. Each sample derived from separate cultures grown in T-25 cell culture flasks. Samples containing approximately equal numbers of infected cells were collected in RNA stabilization reagent RNAlater (AMBION-Life technologies) and stored at −80°C. Total RNA was isolated using the RNeasy kit (QIAGEN) with an added “on-column” DNase I treatment (QIAGEN) according to manufacturer’s instructions. Aliquots of extracted RNA were used to measure contaminant DNA concentration using the Qubit dsDNA HS assay kit (Life technologies). Additionally, RNA was treated three times with RNase-free Dnase I (AMBION-Life technologies) to remove any trace of contaminant DNA in the sample. RNA concentration was measured with the Qubit RNA assay kit (Life technologies), and samples were stored at −80°C.

RT-PCR and RT-qPCR experiments

RT-qPCR experiments

Transcript differences between omp8, omp9, omp10-5’ end, and omp10-3’ genes in A. marginale wild-type and omp10::himar1 mutant were determined using the comparative 2^-∆∆Ct method [53, 54] and the results were based on the mean of three biological samples (individual RNA extracts). For Taqman quantitative PCR, cDNA obtained from ISE6 cells infected with A. marginale wild-type and the omp10::himar1 was used with primers and probes (Table 1) designed to amplify omp8, omp9, omp10-5’ end, omp10-3’ end, msp5, rpoH and the 16S gene sequences. Reaction conditions are described in Additional file 1: Table S3, specificity of primers and probes is shown in Additional file 1: Figure S1 and the amplification efficiencies for each target are reported in Additional file 1: Table S4. For a valid 2^-∆∆Ct calculation, relative efficiencies of target vs. reference genes were calculated and are reported in Additional file 1: Table S5.

Significant differences between the A. marginale wild-type and omp10::himar1 mutant were calculated by Student’s t test (P < 0.05), comparing ∆Ct values (target gene- reference gene) of the omp10::himar1 mutant and the wild-type. The fold difference was based on ∆∆Ct (omp10::himar1 mean ∆Ct – wild-type mean ∆Ct) and calculated as 2^-∆∆Ct which yields the expression ratio. The expression ratio was then expressed as percentage of expression by multiplying the 2^-∆∆Ct values by 100. For normalization of relative gene expression data msp5[55], rpoH, and 16S were used as reference genes.

Western immunoblots

Expression of the Omp9 protein in A. marginale wild-type and omp10::himar1 mutant was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblotting using equal amounts (10⁸) of host-free bacteria. Membranes were incubated with three different antibodies; the anti-Omp9 monoclonal antibody (121/1055) [43], the monoclonal antibody F16C1 (reacts with the Msp5 protein and served as a loading control) [56] and the monoclonal antibody Tryp1E1 (exhibits specificity for a variable surface glycoprotein of Trypanosoma brucei) [56]. This last antibody served as a negative control. Final concentrations of each antibody used were 4 μg/ml, 2 μg/ml and 4 μg/ml. Antibody binding was detected with the secondary antibody goat anti-mouse IgG, horseradish peroxidase labeled and diluted to 1:10,000 using the Pierce ECL Western blotting substrate (Thermo scientific) as described in manufacturer’s instructions.

Quantification of the number of A. marginale wild-type and omp10::himar1 organisms was performed as described elsewhere [41].

GenBank accession numbers

for assembled contigs containing the Himar1 transposon sequences integrated within omp10 and upstream genes (KJ567138) and omp10 (partial 3’ end) and omp9 genes (KJ567139).