To study the nature of heterology blocks, we obtained the genomic DNA sequence from the Plasmodium genome database, PlasmoDB. We retrieved the var gene sequences and grouped them according to their upstream promoter sequence (UPS) region. Sequence alignment was carried out to get the heterologous blocks among var genes using the bioinformatics tool CLUSTALW. We analysed two types of heterologous blocks: nucleotide substitution and insertion/deletion. The length and the occurrence frequencies of the heterologous blocks were plotted using GraphPad Prism.

Synchronous early ring-stage parasites with 6%–8% parasitaemia resuspended in pre-warmed Cytomix buffer (10 mM K₂HPO₄ pH 7.6, 120 mM KCl, 0.15 mM CaCl₂, 25 mM HEPES pH 7.6, 2 mM EGTA pH 7.6, and 5 mM MgCl₂) were electroporated with 100 µg plasmid DNA (PfCENV3–PfRad51^K143R) using a Bio-Rad Gene Pulsar (0.31 kV, 950 µF, and ∝ resistance). Post-transfection, once the parasitaemia reached 4%–8%, the cells were kept under blasticidine (Sigma-Aldrich) pressure (2.4 μg/mL) until the transfectants appeared. The presence of the plasmid PfCENV3–PfRad51^K143R was confirmed by amplifying the PCR fragment size of 1,325 bp using primer pairs OMKB 198 (GGT GAG CTA GCT AAT AGG C) and OMKB 462 (TTA CAC TTT ATG CTT CCG GC). Similarly, the transfectants possessing the empty PfCENV3 vector were confirmed with vector-specific primers OMKB 501 (ACA ACC ATT ACC TGT CCA C) and OMKB 462 that yielded an amplicon of 1,171 bp.

The parasite culture with 6%–8% parasitaemia was harvested and treated with saponin (0.15%) (Sigma-Aldrich) at 37°C for 30 min. The parasite pellet was resuspended in the lysis buffer (10 mM Tris-HCl pH 8, 20 mM EDTA pH 8, 0.5% SDS, and 0.1 mg proteinase K) and incubated for 3 h with intermittent mixing. After incubation, equal volumes of PCIA [phenol, chloroform, and isoamyl alcohol at a ratio of 25:24:1 (v/v/v)] were added, and the aqueous layer was collected upon centrifugation at 12,000 rpm for 15 min at room temperature. The aqueous layer was treated with RNase for 30 min at 37°C, followed by PCIA treatment. The sample was subjected to precipitation by addition of sodium acetate (1/10th vol) and 100% ethanol (2.2 vol) and incubated at −80°C overnight, followed by centrifugation at 12,000 rpm at 4°C, and the DNA were resuspended in 1x TE buffer.

The previously published protocol was followed for cloning and re-cloning parasite lines using the limiting dilution method (Butterworth et al., 2011). Briefly, parasite strains of Pf3D7, PfCENV3, and PfDNRad51 were cultured routinely in vitro. All isogenic lines were subjected to re-cloning at their designated time points, including the 0th, 50th, and 100th generation. It was made sure that the culture had >90% of the singly infected erythrocytes. Parasite cultures were seeded at 0.2 to 0.5 parasites per well in a 96-well plate in 100 µL with 2% HCT Complete medium. The culture media were changed on days 4, 7, 11, 14, and 18. Sufficient number of parasites was visible after 21 days, and the culture was utilised for further experiments. It is worth mentioning that each re-cloning process, from the time of seeding to the time of harvesting parasites for DNA isolation, took approximately 10 generations time. The Nth generation is defined as the day of seeding and not the day of harvesting parasites.

A measure of 2 μg of genomic DNA samples from each time point was used for whole-genome shotgun (WGS) library preparation with a fragment length ranging from 300 to 500 bp. The generated DNA library was subjected to paired-end sequencing of 150 bp using the Illumina NextSeq (150 × 2) sequencing platform (Genotypic Technology, Bangalore, India). The WGS analysis entailed the sequencing of the 23-MB Plasmodium falciparum genome using 150-bp reads, which resulted in a genome coverage of ×20. This sequencing approach yielded an adequate quantity of high-quality reads, exceeding 2 million reads per clone for both the wild-type and mutant parasite lines. The extensive dataset provided a comprehensive perspective of the genomic landscape, facilitating a thorough examination of the genetic composition and variations. These high-quality reads were aligned with the reference genome sequence of Pf3D7 (PlasmoDB.), with the coverage of each sample at a depth of ×10 and ×20. We identified both nucleotide variations that include single-nucleotide polymorphisms (SNPs) and insertions/deletions (INDELs) between 1 and 20 bp size and structural variations (insertions, deletions, translocations, and inversions with the size of >300 bp) using SAM, BCF, and DELLY tools.

To visualise the WGS data for the easy identification of similarities and differences between genomic sequences, we prepared Circos plots for each sample using an online tool Circa (http://circos.ca). The circular ideogram layout facilitates the display of all chromosomes and the position of SNPs, INDELs, and structural variations on each chromosome. The data were represented with identified SNPs, INDELs, and translocations of six different samples from both wild-type and PfRad51 mutant parasites in six different plots. A typical Circos plot contained all 14 chromosomes of P. falciparum arranged in an ideogram track. Both SNPs and INDELs in the core genome and var genes are represented as a histogram and Scatchard plot, respectively. The translocations are represented in blue and orange lines for the core genome and var genes, respectively.

We estimated the rate of change of both nucleotide variations (SNPs and INDELs) and structural variations per generation per nucleotide in wild-type and PfRad51 mutant parasites. The rate of change was calculated by taking the total number of unique changes in nucleotide or structural variations present in that particular strain per generation per nucleotide. The genome sequences of the 0th generation of both wild-type and PfRad51 mutant parasites were taken as the reference, and any new or missing nucleotide variations or structural variations were taken as changes. The rate of change for the var gene region was calculated by considering the total size of the var genes, while the rate of change for the rest of the genome was calculated considering the genome size.

As var sequences are homeologous in nature with respect to each other, it is important to estimate the length and the frequency of such heterologous blocks at var genes. It is intitutive that shorter heterologous blocks could be bypassed during recombination events between var sequences, while the longer blocks may not be tolerated. To analyse var genes, we retrieved var gene sequences (P. falciparum 3D7 genome) from the PlasmoDB database. Sequence alignment was carried out for each UPS group of var genes using CLUSTALW (DNASTAR software), and the length of the heterologous blocks was analysed. After the alignment of the sequence, deletions/insertions or substitutions of the base pair that caused sequence divergence were identified as heterologous blocks. Every single-base substitution from each UPS group was analysed and represented in graphs by taking the length of heterologous blocks and the total number within the group (Figures 1A–E). The consensus sequences from each UPS group were also analysed to determine the lengths and frequencies of heterologous blocks across UPS groups (Figure 1F). Similarly, insertions/deletions from each UPS group and across UPS groups were also analysed (Figures 2A–F). From the data, it is evident that var genes mostly have short heterologous blocks, ranging from 1 to 6 bp. Additionally, the length of such heterologous blocks was inversely proportional to the frequency of their occurrence. There are very few base substitution blocks with lengths greater than 6 bps, and a similar pattern was also observed with insertions/deletions.

Since the lengths of heterologous blocks were found to be below the permissible limit of heterology bypass by eukaryotic Rad51, we hypothesised that PfRad51 should be able to carry out recombination between homeologous var sequences. To investigate the role of PfRad51 in var gene recombination, we have expressed a dominant negative allele of PfRad51^K143R in the P. falciparum in vitro culture. This particular mutant was reported to be defective in DSB repair (Jha et al., 2021). We have chosen a Pf3D7 strain for our investigation of gene diversification studies because of its complete annotated genome (Gardner et al., 2002). We sought to identify nucleotide and structural variations at the var loci and in the core genome. To this end, we successively cultured isogenic lines of both wild-type and mutant parasites for 100 generations. At designated time intervals (0th, 50th, and 100th generation), we re-cloned both parasite lines and isolate genomic DNA from individual clones, and representative clones were subjected to Illumina platform-based WGS to obtain the mutation score as described in methodology. The total number of nucleotide variants (SNPs and micro-INDELs with the size of <20 bp) and structural variations (translocations, inversions, and duplications with >300 bp) distributed throughout the genome were identified at all the time points of both parasite lines using SAM, DELLY, and BCF tools (Morel et al., 1994; Holmes et al., 2001). A detailed inspection of NVs generated in both wild-type and DN parasites were represented as a histogram on all 14 chromosomes in Circos plots (Figure 3). Similarly, structural variations are represented as blue and orange lines in Circos maps of the respective generation for the core genome and var gene region, respectively, of both parasite lines (Figure 3).

To compare the rate of rearrangement between wild-type and mutant parasite lines, we scored the unique change and calculated the rate of change in nucleotide and structural variations. We carried out detailed inspections of NVs and SVs generated in both wild-type and DN parasites. By taking the 0th generation as a reference, the appearance or disappearance of any variation was considered a single change. We scored all such novel and missing NVs or SVs from WGS of both parasite lines in subsequent generations using BCF tools. Taking the 0th generation as the reference point, we estimated the genome-wide rate of change of NVs and SVs per generation per nucleotide in both parasite lines and observed nearly a 30% reduction in the rate of change in HR-deficient parasites (Figures 4A, B).

To gain insights into recombination events at the var loci, we estimated the rate of change of SVs at var genes. After normalisation with the size of the core genome, the total number of unique SVs was found to be more at the var region than the core genome, and it was found that in HR-deficient parasites, nearly a 70% reduction in the rate of change at var genes was observed (Figure 4C). Together these results suggest that PfRad51-deficient parasites undergo fewer recombination events and thereby decrease the rate of change. In conclusion, the lack of a functional PfRad51 led to a significant decrease in the rate of var gene diversity in Plasmodium.