Geographical and temporal dynamics of genetic diversity of Plasmodium falciparum merozoite surface proteins 1/2 in India

Abstract

The high genetic diversity of Plasmodium falciparum (Pf) is a big obstacle to successful vaccine development programs. Here, the geographical and temporal dynamics of the genetic diversity of Indian Pf isolates from patients living in Ranchi, Raipur, Mewat, and Rourkela were analyzed. Typing and frequency of merozoite surface protein 1 and 2 genes (pfmsp1/2), their genotypes, clonality, heterozygosity, multiplicity of infection, and neutral evolution metrics were computed. A phylogenetic analysis was also performed for these two genes. The dominant allelic types were K1 (55%) and MAD20 (55%) for msp1, and FC27 (64.7%) for msp2. Infections were mainly monoclonal in Ranchi and Mewat while polyclonal in Raipur and Rourkela. Polyclonal infections dropped from 57.1 to 71.3% in 2013 to 33.3–33.4% in 2016 in Raipur. K1 and MAD20 sequences were highly diverse due to the organization of the amino acid units SGG, SVA, SVT, and SGN. The IC/3D7-related G,S,A-rich region showed a large variation of four to eight amino acid repeats, including mostly GAVASA (81.8%), GSGA (54.5%), and GASGSA (45.5%). The 32-amino acid sequence of the FC27 type was present in all isolates with several mutations. The msp1/2 sequences were not under neutral evolution, except the K1 family, which is under balancing selection. The msp1/2 sequences are phylogenetically closer to previous Indian sequences than those from Africa, Asia, the Americas, and Oceania. This study outlines a high genetic diversity of Pf infections with complex structure, and evolutionary signature changed with time.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12639-024-01698-8.

Introduction

Malaria remains one of the main poverty-amplifying infectious diseases in the world. This arthropod-borne disease is caused by parasites of the Plasmodium genus transmitted to humans through the infecting bite of female Anopheles mosquitoes (Cowman et al. 2016). Five species can cause human infections, with Plasmodium falciparum (Pf) being the most dangerous species (Cowman et al. 2016). The other species such as P. vivax, P. malariae, P. knowlesi, and P. ovale spp., may cause severe malaria and deaths, but at a lower extent than P. falciparum (Kojom Foko et al. 2023). In their latest report, the World Health Organization (WHO) estimated that malaria was responsible for 249 million cases and 608,000 deaths worldwide, respectively (World Health Organization 2023). A large number of efforts regarding malaria control have been made in these last two decades, with a huge decrease in malaria burden mostly in endemic areas of the globe (World Health Organization 2023). These achievements are translation of the implementation and scaling up of control strategies for prevention (long-lasting insecticide nets, indoor residual spraying) and treatment (artemisinin-based drugs). However, there are still a large number of countries dramatically affected by malaria in Africa and Asia (World Health Organization 2023).

The high genetic diversity of Pf is a major impediment to efficient malaria control, as it is at the basement of antimalarial drug resistance and host immunity, and explains the difficulties for successful vaccine design (Takala and Plowe 2010; Arya et al. 2021). In this context, it is of utmost importance to determine the genetic structure of malaria parasites. The two leading candidate vaccine genes, merozoite surface protein genes (msp1 and msp2), are highly polymorphic in allele sizes and sequence polymorphism, and this makes them interesting markers for studying the genetic structure of Pf populations. The high polymorphism of genes is mostly contained in block 2 for msp1, and block 3 for msp2, respectively. The msp1 block 2 includes four allelic types (i.e., K1, MAD20, RO33, and MR) while the msp2 block 3 consists of two types (i.e., IC/3D7 and FC27) (Smythe et al. 1990; Takala et al. 2002).

Malaria is also a cause of concern in India, which accounts for ~ 50% of all vivax cases in the Southeast Asian region. A recent systematic review and meta-analysis showed that P. vivax mono-infection was involved in severe malaria cases reported in the country, with a prevalence of 22.9% (Kojom Foko et al. 2021). P. falciparum species is also highly prevalent in the country and balances with P. vivax with a ratio of about 1:1 at the country level, but with region-specific differences, especially in urban areas (Anvikar et al. 2016). Other “minor/neglected” Plasmodium species (P. ovale spp., P. knowlesi, P. malariae) are also reported, often at submicroscopic levels, in the country (Kojom Foko et al. 2024). The Government of India has developed a fight plan to curve malaria burden by 2030 to achieve the WHO malaria elimination milestones (WHO 2020). Thus, studying the genetic diversity of malaria parasites is a prerequisite to adequately defining and implementing context-adapted control strategies, including the design of effective vaccines against Pf parasites. There is a dearth of data on the genetic structure of Pf populations in India. This study was therefore designed to determine the genetic polymorphism and evolutionary patterns of the msp1/2 genes of Pf isolates from four Indian regions.

Materials and methods

Study areas and P. falciparum samples

Samples were collected from symptomatic individuals during the years 2008–2016 in four areas viz. Ranchi in Jharkhand, Raipur in Chhattisgarh, Mewat in Haryana, and Rourkela in Odisha state. The details of samples by area and year of collection are shown in Fig. 1. According to the data from National Vector Borne Control Disease Programme (NVBCDP), Pf is the main malarial species in Raipur and Rourkela. In Mewat, Pv is predominant with high prevalence during the years 2008–2016 (80 to 100%). In Ranchi, Pf and Pv are involved in malaria burden with a prevalence ratio ~ 1 (NVBDCP 2013, 2018). Recent data from the NVBCDP indicate that malaria transmission is still high in Jharkhand, Odisha, and Chhattisgarh states with annual parasite indexes above 2, despite the huge reduction in malaria burden these last two decades across all Indian regions (Kojom Foko and Singh 2023). Thus, areas (Raipur, Ranchi, and Rourkela) located in these states were selected as study sites. Also, a low-endemic area (i.e., Mewat) was included in the analysis for a comparative analysis of the genetic diversity of msp1/2 genes between low- and high-endemic areas (NVBDCP 2018) (Fig. 1). The present study was approved by Institutional Ethics Committee (IEC) of ICMR-National Institute of Malaria Research (NIMR) (N°PHB/NIMR/EC/2020/55).

Fig. 1.

Map of India showing study sites for P. falciparum sample collection. Note: The map depicted here is taken from the official website of Ministry of External Affairs, India (https://mea.gov.in/india-at-glance.htm accessed 15/11/2021)

DNA extraction

The Pf genomic DNA was extracted from the dried blood spot (DBS) using the Qiagen QIAamp DNA Blood Mini Kit (Hilden, Germany) as per the manufacturer’s instructions. An elution volume of 100 µL was used, and the DNA was kept at –20°C until further molecular analysis.

PCR assay—18S rRNA and pfmsp1/2 genes

The extracted DNA was used to confirm infection with Pf through the amplification of the 18S rRNA gene as described previously (Snounou et al. 1993). Briefly, a first PCR using genus-specific primers (rPLU5 and rPLU6) was run to target the pan-Plasmodium 18S gene. The PCR products of the first PCR round were used as templates for the nested PCR using Pf-specific primers (rFAL-F and rFAL-R) (Table 1). The PCR mixture and cycling conditions were the same in the two PCR rounds. PCR products were run for 50 min on a 2% agarose gel, then stained with ethidium bromide (EtBr) solution (0.5 µg/mL) and visualized under an ultraviolet (UV) trans-illuminator. The expected size for Pf positive PCR products is 205-bp.

Table 1.

Primers and PCR conditions of 18sRNA and msp1/2 genes

PCR	Genes	Primer sequences (5'–3')	Denaturation		Annealing		Elongation		No. of cycles
			Temp (°C)	Time (min)	Temp (°C)	Time (min)	Temp (°C)	Time (min)
Primary	18sRNA	rPLU5: CCTGTTGTTGCCTTAAACTTC	94	1	58	2	72	5	30
		rPLU6: TTAAAATTGTTGCAGTTAAAACG
Nested		rFAL-F: TTAAACTGGTTTGGGAAAACCAAATATATT	94	1	58	2	72	5	30
		rFAL-R: ACACAATGAACTCAATCATGACTACCCGTC
Primary	msp1	M1-OF: CTAGAAGCTTTAGAAGATGCAGTATTG	95	1	58	2	72	2	25
		M1-OR: CTTAAATAGTATTCTAATTCAAGTGGATCA
Nested	K1	M1-2KF: AAATGAAGAAGAAATTACTACAAAAGGTGC	95	1	61	2	72	2	25
		M1-2KR: GCTTGCATCAGCTGGAGGGCTTGCACCAGA
	MAD20	M1-2MF: AAATGAAGGAACAAGTGGAACAGCTGTTAC	95	1	61	2	72	2	25
		M1-2MR: ATCTGAAGGATTTGTACGTCTTGAATTACC
	RO33	M1-2RF: TAAAGGATGGAGCAAATACTCAAGTTGTTG	95	1	61	2	72	2	25
		M1-2RR: CATCTGAAGGATTTGCAGCACCTGGAGATC
Primary	msp2	M2-OF: ATGAAGGTAATTAAAACATTGTCTATTATA	95	1	58	2	72	2	30
		M2-OR: CTTTGTTACCATCGGTACATTCTT
Nested	IC/3D7	M2-ICF: AGAAGTATGGCAGAAAGTAAKCCTYCTACT	95	1	61	2	72	2	30
		M2-ICR: GATTGTAATTCGGGGGATTCAGTTTGTTCG
	FC27	M2-FCF: AATACTAAGAGTGTAGGTGCARATGCTCCA	95	1	61	2	72	2	30
		M2-FCR: TTTTATTTGGTGCATTGCCAGAACTTGAAC

Samples positive for the Pf-specific 18S gene were eligible for pfmsp1/2 amplification (Fig. 2). Nested PCR protocols were used to amplify these genes as described previously (Table 1). The first PCR round consisted of amplification using gene-specific primers, while the second PCR round used allelic type-specific primers (K1, MAD20, and RO33 for msp1, and IC/3D7 and FC27 for msp2) (Table 1). All PCR amplifications were performed on a thermal cycler in a final volume of 25 µL. The PCR mixture and cycling conditions of the pfmsp1/2 genes are shown in Table 1. PCR products were run for 50 min on a 3% EtBr-stained agarose gel and then visualized with an UV trans-illuminator. The expected size of the PCR products is variable due to the existence of a size polymorphism as mentioned above.

Fig. 2.

Workflow of the study. Note: DBS Dried blood spot, msp Merozoite surface protein, MOI Multiplicity of infection, PCR Polymerase chain reaction, Pf Plasmodium falciparum

Sequencing

Samples with good-quality agarose gel bands were purified using the DNA GeneJET PCR purification kit (ThermoFisher Scientific, Lithuania) and sequenced on an ABI3037XL analyzer (Applied Biosystems). The sequences have been submitted to the NCBI GenBank database.

Pfmsp1/2 allele typing

After successful amplification for msp1/2 genes, the total number of Pf genotypes for each gene and their respective proportions were deduced by binning at a 20 bp interval. PCR products with a size difference of 20 bp were considered as different genotypes (Gupta et al. 2014).

Clonality of infections

An infection was considered monoclonal (MI) if only one allelic type (K1, MAD20, or RO33 for msp1, IC/3D7, or FC27 for msp2) was PCR-identified, while polyclonal infection (PI) was defined as the presence of at least two allelic types for msp1 (K1 + MAD20, MAD20 + RO33, K1 + RO33, and K1 + MAD20 + RO33) and the two allelic types for msp2 (IC/3D7 + FC27) in the same isolate. Accordingly, infections were categorized into MI and PI.

Multiplicity of infection and heterozygosity

The multiplicity of infection (MOI) was defined as the number of genotypes per infection and estimated by dividing the total number of msp1 or msp2 fragments detected by the number of positive samples (Apinjoh et al. 2015). Heterozygosity (H_E) was determined as follows: H_E = [n/(n – 1)]*[(1 – Ʃpi²)],where n is the number of samples, and pi, the allele frequency at a given locus (Apinjoh et al. 2015).

Genetic diversity and structural organization

The genetic diversity of msp1/2 sequences was assessed through several parameters, including number of segregating sites, nucleotide diversity (π), minimum number of recombination events (Rm), heterozygosity per site (θ), and haplotype diversity (Hd). The nucleotide msp1/2 allelic family sequences were translated into amino acid (aa) using BioEdit software v7.0.5.3 and their structural patterns were then analyzed and compared with reference Pf strains (Smythe et al. 1990, 1991; Zhang et al. 2019).

Departure from neutral evolution

Tajima’s D, Fu and Li’s D, and Fu and Li’s F tests were used to test the departure of msp1/2 sequences from neutral evolution using DnaSP v6 software (Rozas et al. 2017; Kumar et al. 2018). In practice, positive values of these statistics outline population size shrinking or a balancing selection, while negative values outline the evolution of sequences driven by purifying selection forces (Tajima 1989; Fu and Li 1993).

Statistical analysis

All data collected were entered into an Excel spreadsheet and exported to GraphPad v5.03 (GraphPad Prism, Inc., San Diego, CA, USA) and the statistical package for social sciences v16 (SPSS, IBM, Chicago, IL, USA) for statistical analysis. Descriptive statistics (percentage, mean) were used to depict qualitative and quantitative variables in figures and tables where appropriate. Pearson’s chi square (χ²) test was used to compare proportions, while correlation test was used to study the association between the different variables. Non-parametric Kruskal–Wallis and Mann–Whitney tests were used to compare quantitative variables between groups. Statistical significance was set at a p-value less than 0.05.

Results

Frequency of different pfmsp1/2 allelic types

Of the 455 samples analyzed, a total of 75 Pf-positive samples were included in the study, among which 69 and 68 were positive for msp1 and msp2, respectively (Supplementary file 1 & 2). Of the 75 Pf-positive samples, 13, 14, 33, and 15 were from Raipur, Rourkela, Ranchi, and Mewat, respectively (Supplementary file 2). Overall, the K1 and MAD20 types were mainly found in the study (55% each), followed by RO33 (15.9%). Regarding the msp2 gene, the FC27 family was predominant (64.7%), while the IC/3D7 family was found in 60.3% of the isolates (Supplementary file 1). The msp1 MAD20 allele was found in all samples from Mewat, while 85.7% of them belonged to the msp2 FC27 type (Fig. 3). Regarding Raipur, PCR was positive for K1 in 71.4% and 50.0% of the samples from this area in 2013 and 2016, respectively. PCR positive results for the MAD20 and RO33 families were as follows: MAD20 [42.9% in 2013 and 66.7% in 2016] and RO33 [57.1% in 2013 and 16.7% in 2016] (Fig. 3).

Fig. 3.

Proportion of msp1/2 allelic types by area and year of data collection. Note: Pie charts represent the proportion of K1 (blue), MAD20 (red), RO33 (green) for msp1, and IC/3D7 (brown) and FC27 (orange) for msp2. A fully colored circle indicates a proportion of 100%. For a given gene (i.e., msp1 or msp2), the sum of the proportions of its allelic families may exceed 100% given the fact that a same sample may be positive for at least two families for msp1 gene or both families for the msp2 gene

Genotype diversity

On analysis of msp1 gene, a total of 20 genotypes were found (7 genotypes for K1, 7 genotypes for MAD20, and 6 genotypes for RO33). The size range for the three allelic types were 120–280 bp for K1, 100–240 bp for MAD20, and 120–260 bp for RO33 (Fig. 4 a-d). The 180–200 bp and 140–160 bp alleles were the most frequently found among the K1 and MAD 20 alleles, with a proportion of 11.94% and 14.93%, respectively. All RO33 alleles were found in proportion less than 5% (Fig. 4a). On analysis by area and year of sample collection, no clear trend was observed for genotype diversity of K1 and MAD20 allelic families across the study sites (Supplementary file 3). Several observations were noted, with either a decrease, increase, appearance, or disappearance of some K1 and MAD20-related genotypes. In contrast, a temporal reduction in genotype diversity and frequency was found for the RO33 family in Raipur and Ranchi (Supplementary file 3).

Fig. 4.

Proportion of genotypes of allelic types of a msp1, b PCR band patterns of K1 type, c PCR band patterns of MAD20 type, d PCR band patterns of RO33 type. Note: NC: Negative control, M: 100 bp molecular weight marker. Field samples are depicted on lanes 1–7 in Figs. 4b for K1 family, 4c for MAD20 family, and 4d for RO33 family

A total of 24 genotypes were found based on msp2 analysis, 13 genotypes for IC/3D7 and 11 genotypes for FC27. The size of PCR products ranged from 360 to 660 bp for IC/3D7 and 260–520 bp for FC27 (Fig. 5 a-c). The 500–520 bp allele (16.42%) was mainly found among IC/3D7 alleles, while the 280–300 bp (11.94%) and 300–320 bp (13.43%) were the dominant FC27 alleles (Fig. 5a). A higher number of genotypes were found in Ranchi for the IC/3D7 family (12 genotypes), while the highest number of genotypes for FC27 family was found in Rourkela (9 genotypes) (Supplementary file 4). Overall, three area- and time-dependent trends in genotype diversity were observed for both msp2 allelic families: i) reduction in diversity with the appearance of some genotypes (e.g., in Raipur for the IC/3D7 family, and Ranchi for the FC27 family), ii) maintenance of diversity with appearance of new genotypes (e.g., in Raipur and Rourkela for the IC/3D7 family, and Raipur for the FC27 family), and iii) increase in genotype diversity (e.g., in Rourkela for the FC27 family). For instance, four FC27-like genotypes were reported in 2008 in Rourkela (i.e., 300–320 bp, 340–360 bp, 360–380 bp, and 380–400 bp), but these disappeared, with the exception of genotypes 300–320 bp and 360–380 bp. In 2010, five genotypes emerged in the same area (i.e., 260–280 bp, 320–340 bp, 400–420 bp, 420–440 bp, and 500–520 bp) (Supplementary file 4).

Fig. 5.

Proportion of genotypes of allelic types of a msp2, b PCR band patterns of IC/3D7 type, c PCR band patterns of FC27 type. Note: M: 100 bp molecular weight marker. Field samples are depicted on lanes 1–12 in Figs. 5c for the IC/3D7family, and 5c for the FC27 family

Clonality of P. falciparum samples

Most of Ranchi samples were found to be monoclonal in regardless of the genetic markers used (msp1: 78.12%, msp2: 68.75%). The same pattern was also found in Mewat samples (Supplementary file 5). Contrasting results were found for Raipur and Rourkela. Based on msp1 and msp2, most infections were monoclonal across the study areas, with the exception of Raipur, where the infections were mainly polyclonal (Supplementary file 4). The temporal analysis of the clonality in each study site revealed that infections were mostly monoclonal in Ranchi irrespective of the year of sample collection and genetic marker (Fig. 6). In Rourkela, the proportion of polyclonal infections was reduced from 50% in 2008 to 20% in 2010 using msp1, and from 75 to 20% using msp2. Likewise, the proportion of polyclonal infections in Raipur dropped from 71.3% in 2013 to 33.4% in 2016 using msp1, and from 57.1 to 33.3% using msp2 (Fig. 6).

Fig. 6.

Frequency distribution of allelic types of msp1 and msp2 by study sites and year of data collection

Multiplicity of infection and heterozygosity

The MOI was highest in Rourkela irrespective of the genetic marker used, with values of 1.60 (msp1) and 1.61 (msp2) (Supplementary file 6). There was a statistically significant difference between the mean values of msp1-based MOI (Kruskal–Wallis test, H = 17.61, p-value = 0.0002). A total of 11 and 10 genotypes were identified in Ranchi and Rourkela using msp1, respectively. Using msp2, the total number of genotypes in these two areas was 18 and 16. The genetic diversity was high in all study areas, as H_E values were close to 1 (Supplementary file 6). Overall, the msp1 gene revealed exhibited higher MOI values than those for msp2, with the exception of Mewat in 2016, and Rourkela in 2010 (Table 2). In Ranchi, the MOI estimates in samples collected in 2010 were statistically higher than those for the year 2013, irrespective of the polymorphic marker (e.g., 1.54 in 2010 vs 2.17 in 2013, p = 0.01 for msp1, 1.38 in 2010 vs 1.54 in 2013, p = 0.03 for msp2). In contrast, a statistically significant reduction in MOI was found in Rourkela between 2008 and 2010, but no difference was found for samples from Raipur (Table 2).

Table 2.

MOI of Pf infections by areas, year of data collection and polymorphic markers

Areas	Years	msp1				msp2
		K1	MAD20	RO33	Overall	IC/3D7	FC27	Overall
Mewat	2016	–	1	–	1	–	1.17	1.17
Ranchi	2010/11	1.17	1.17	1	1.54	1	1.11	1.38
	2013	1.33	1	1	2.17	1.50	1	1.54
Raipur	2013	1.40	1.30	1	2.30	1	1	1.57
	2016	1.30	1	1	2.20	1	1.2	1.60
Rourkela	2008	1.67	1	1	2.00	1.33	1.50	2.00
	2010	1	1	1	1.33	1	1.17	1.38

MOI Multiplicity of infection, msp Merozoite surface protein, H_E Heterozygosity, Pf: Plasmodium falciparum

Genetic structure of msp1 allelic families

The MSP1 protein sequences for K1 (n = 10) and MAD20 (n = 18) obtained from Raipur, Rourkela, Ranchi, and Mewat were structurally different and highly diverse from each other with eight alleles for K1 and 11 alleles for MAD20. The sequence and length diversity were due to repetition in the varying number and organization of trimeric amino acid units, including mainly SGG, SVA, SVT and SGN (Fig. 7a & Supplementary file 7). Only 36.3% of the MAD20 started with the SVT trimer as seen in the reference sequence. Few variants of the common trimers were found in the field sequences (e.g., PRG in K1 sequences, PVA and AVT in MAD20 sequences) (Fig. 7b). In total, seven good-quality sequences were obtained from the 11 samples positive for RO33. Seven alleles were found and presented several mutations with four conserved regions (NPPGAT, PSGTA, TKG, and SPGAANP) (Fig. 7c).

Fig. 7.

Structural organization of msp1 families for K1 (a), MAD20 (b), and RO33 (c). Note: The sequences were aligned with reference (Accession numbers: K1—M59766, MAD20—JX283490.1, RO33—AB276005). In Fig. 7c, dashes represent gaps introduced to maximize the alignment. Substitutions are highlighted in pink. The conserved regions in RO33 sequences are framed

Genetic structure of msp2 allelic families

As seen for msp1, msp2 sequences obtained for IC/3D7 (n = 12) and FC27 (n = 11) also exhibited a high sequence diversity with allelic diversity of 100%. Analysis of the G,S,A-rich region of IC/3D7 sequences showed a large variation in repeat units consisting of four to eight amino acids. The main units included GAVASA, GSGA, GASGSA, and VAGSGA, which were found in 81.8%, 54.5%, 45.5%, and 27.3% of the IC/3D7 sequences analyzed (Fig. 8 & Supplementary file 7). The GAVASA unit was found once in the sequences while VAGSGA and GASGSA were found at least twice in each sequence. Variants of reference units were also found, and these included ESGA, GSGD, GAVARA, GAVAESGA, GADD, and SGRA____. The FC27 sequences showed several sequence, structure, and length differences in comparison to the reference 3D7 sequence. The 32-amino acid sequence characterizing the FC27 type was present with several mutations in all isolates. The majority of the sequences carried two repeats of the 32-aa sequence, while one repeat was deleted in three isolates. Two isolates presented three repetitions of this sequence. The 7-amino acid sequence (ADPTAT) was well conserved except in one isolate where a variant of this sequence was noted at the first two amino acids (DKPTAT). All isolates presented variants of the 12-amino acid sequence (ESISPSPPITTT), and these included KSNSPSPPITTT and ESNSPSPPITTT. A complete deletion of the sequence was found for two isolates, while the sequence was repeated 1–4 times for the rest of the isolates (Fig. 8 & Supplementary file 8).

Fig. 8.

Structural organization of msp2 allelic families IC/3D7 (a) and FC27 (b). Note: The (G, S, A-rich)_n R1 region of IC/3D7 family was compared for field sequences with reference IC/3D7 (X53832) (Chaorattanakawee et al. 2018). The block 3 of FC27 sequences from field isolates is compared to that of reference sequence FC27 (J03828). This block consists of two repeat regions, R1 with two 32-aa sequences and R2 with 12-aa sequence, spaced by non-repeat region (NR) (Chaorattanakawee et al. 2018). In Fig. 8a, the letters in red represent the substitution points and red dash lines represent deletions

Phylogenetic and evolutionary patterns

The analysis of the msp1/2 sequences revealed a diverse phylogenetic relatedness, with some sequences close to previous Indian sequences, while others were phylogenetically closer from sequences isolated in different geographical areas including sub-Saharan Africa (Cameroon, Ghana, The Gambia, Senegal, Kenya, the Democratic Republic of the Congo, Sao Tome and Principe), Asia (Iran, Myanmar, Thailand, Vietnam), the Americas (Brazil, Honduras-Nicaragua border), and Oceania (Papua New Guinea, Solomon Islands, Vanuatu) (Supplementary files 9 & 10).

The nucleotide diversity of msp1/2 sequences was high, with values of 0.67 to 0.72. The sequences were not under neutral evolution as no statistically significant differences were found for Tajima, and Fu & Li’s D tests (Table 3). In contrast, the K1 sequences seem to be under balancing selection (Fu and Li’s D = 0.0466, p < 0.05). Few recombination events were found for MAD20 (Rm = 3), IC/3D7 (Rm = 4), and FC27 (Rm = 3) (Table 3).

Table 3.

Genetic diversity and neutrality tests of pfmsp1/2 allelic families in Pf isolates

Families	n	N	S	θ	π	Hd ± Sd	Fu and Li’s D	Fu and Li’s F	Tajima's D	Rm
K1	10	318	129	0.93	0.72	1 ± 0.002	0.4440a	0.0466b	−1.1620a	0
MAD20	18	182	102	0.85	0.67	1 ± 0.001	0.3499a	-0.0209a	1.0794a	3
RO33	7	127	101	0.93	0.71	1 ± 0.001	−0.0214a	-0.5218a	−1.3974a	3
IC/3D7	12	520	387	0.89	0.71	1 ± 0.001	0.5010a	0.1286a	−0.9674a	4
FC27	11	394	261	0.83	0.67	1 ± 0.001	0.3715a	0.0271a	−0.9764a	3

Pf Plasmodium falciparum, msp Merozoite surface protein, n Number of isolates, N Total number of sites, S Segregating sites, θ Heterozygosity per site, π Nucleotide diversity, Hd Haplotype diversity, Sd Standard deviation, Rm Minimum number of recombination events

^ap-value > 0.05

^bp-value < 0.05

Discussion

The development of an effective vaccine against Pf malaria disease is greatly hindered by the high genetic diversity of this species, which also plays a crucial role in shaping the host immune response. In this context, it is important to investigate the genetic diversity of Pf populations in malaria-endemic areas, especially in India, where malaria is highly prevalent and there is a paucity of such studies (Supplementary file 11). We therefore determined the genetic diversity, sequence variation, and evolutionary and phylogenetic patterns of pfmsp1/2 vaccine candidates in four areas in India.

The proportion of different msp1 and msp2 allelic types varied with the study area. In Ranchi (Jharkhand), Rourkela (Odisha), and Raipur (Chhattisgarh), K1 type was predominant, while MAD20 type was predominant in Mewat (Haryana). Most of the studies in India reported a predominance of K1 type (Supplementary file 11). Similarly, a global genetic analysis of the msp1 gene outlined the predominance of K1 allelic type in sub-Saharan African and American regions (Lê et al. 2019). In contrast, one study reported a predominance of MAD20 type in 13 villages in Odisha state. Likewise, Hussain and colleagues found the MAD20 type in Jharkhand state (Hussain et al. 2011). The RO33 type was mainly reported in the Odisha and Arunachal Pradesh states (Sarmah et al. 2017). The same pattern was observed for msp2 allelic types, for which we found a predominance of the IC/3D7 types in Ranchi (Jharkhand), Rourkela (Odisha), and Raipur (Chhattisgarh), but in contrast, a predominance of FC27 type in Mewat (Haryana). Previous studies have reported the predominance of the FC27 type in the Odisha and West Bengal states (Supplementary file 11). These discrepancies observed could be due to an area-specific variability of malaria transmission and differences in study population, geographical area, and period. In the present study, we worked on feverish individuals, while other studies worked on individuals with either asymptomatic, uncomplicated, or severe malaria. Sahu and colleagues found higher prevalences and significant associations of RO33- and 3D7-allelic types with two forms of severe malaria namely cerebral malaria and severe malarial anemia in patients living in Odisha (Sahu et al. 2008). Also, working on patients with mild or severe malaria, Rout and colleagues reported a higher proportion of msp2 FC27 family among mild malaria cases, while an equal distribution of FC27 and IC/3D7 families was observed among severe malaria cases (Rout et al. 2009).

The clonality of infections also varied by study area, with a predominance of polyclonal infections in some areas. Some studies found that a large number of infections were polyclonal, while other studies reported a predominance of monoclonal infections (Supplementary file 11). Polyclonal infections imply the infection of an individual with several Pf strains which may collectively infect mosquito vectors during their blood meal. Thus, this increases the risk of genetic recombination events in the mosquito, thereby increasing the chances of generating new genotypes in the study areas, including drug resistant strains (Camponovo et al. 2023). MOI is generally used as a proxy for the level of malaria transmission in endemic areas. In this study, MOI was below 2, and this is consistent with previous studies (Gupta et al. 2014). These low MOI values can be a reflection of the implementation and scaling up of malaria control interventions in these areas.

This study also pinpointed dynamics in the genetic polymorphisms of msp1 and msp2 allelic families over time and area, with the appearance of new alleles and disappearance of pre-existing alleles. Similar observations were reported in Myanmar (Lê et al. 2019). Furthermore, a spatial and temporal variation in the clonality of Pf infections was also found in this study, especially in Rourkela, Ranchi, and Raipur where msp1- or msp2-based clonality dropped over time. This finding is not consistent with our findings on MOI estimates in these three areas, for which an increase in MOI was observed in Ranchi, no variation in Raipur, and a reduction in Rourkela.

The important diversity in K1 and MAD20 isolates was due to the high variation in number and arrangements of tripeptide repeats, which is supporting studies in sSA, the Americas, and Asia (Ferreira et al. 2003; Lê et al. 2019). It was theorized that interallelic diversity of K1 and MAD20 families could be due to unequal exchange during mitosis and meiosis, and DNA strand slippage (Miller et al. 1993). Besides, several repeat motifs were seen in the G,S,A-rich region of IC/3D7-like isolates (e.g., GAVAGSGA, GAVASA). Some of these motifs have been reported from India (Patel et al. 2017), and other settings such as Cameroon (Basco et al. 2004), and Panama (Santamaría et al. 2020). The NR and R2 regions of FC27-like isolates were well conserved, although some of them presented complete deletions. In Cameroon, it was seen that FC27 sequences from the central region were conserved (Basco et al. 2004).

The spatial and temporal dynamics of genetic diversity and structure observed in the study has implications in the design, implementation, or scale-up of control strategies (e.g., vaccines) in the study areas. Indeed, the genetic structure of Pf populations modulates greatly the natural acquisition of host immune responses in a given area (Langhorne et al. 2008; Beeson et al. 2016). Given the high genetic polymorphism, which is partially maintained by natural selection processes such as balancing selection, our findings suggest the need to develop a strain-transcending vaccine to cover the maximum of strains observed in the study sites. Analyzing the global diversity and evolutionary patterns of 23 leading Pf candidate vaccine antigens, Naung and coworkers have reported a high genetic diversity with evidence of balancing selection (Naung et al. 2022), which may impact the efficacy of the vaccines against the natural Pf populations. Besides, MOI were below 2 irrespective of the study areas, msp1/2 allelic families, and sample collection period. As above-mentioned, these low MOI values could reflect a reduction in malaria transmission in the study areas, which is likely due to the control strategies implemented by the Government of India in these last two decades. Thus, these findings support the effectiveness of current control strategies in the country (NVBDCP 2018).

Limitations of the study

This study has several limitations that should be taken into account when interpreting the findings. The main reason was the small sample size of Pf-infected samples with lack of information on patients, which made it impossible to do a fine analysis of msp1/2 evolutionary patterns in study sites for instance. The sample size limited the analysis of temporal dynamics of Pf genetic diversity using sophisticated statistical analyses such as generalized estimating equations or Bayesian analysis. Again, the sample size of Pf-positive samples was not similar between the chosen sites, which may have introduced bias in the study findings. The geographical distance and difference in demography between the low-endemic area (i.e., Mewat) and the high-endemic areas (i.e., Raipur, Ranchi, and Rourkela) also may limit the comparability of findings between the study areas. PCR and agarose electrophoresis were used to differentiate the msp1/2 fragments in the present study. Although this methodological approach is most commonly used in the literature, it has limitations by its ability to distinguish minor fragments, unlike more sensitive techniques such as capillary electrophoresis (Liljander et al. 2009; Gupta et al. 2010). Finally, the samples analyzed in this study covered the frame time of 2008 – 2016, which further limits our evaluation of the current genetic diversity of Pf parasites in the study areas. However, given the dearth of data on the genetic diversity of Pf populations in India, especially in these areas where malaria transmission is still high, the findings from the present study could serve as baseline data for comparative analysis in further genetic diversity studies.

Conclusion

This study provides information on the genetic diversity and evolutionary patterns of Pf populations in some areas of India, a malaria endemic country, after the deployment of a large number of control interventions over these last two decades. The Pf infections are highly genetically diverse, with an area-dependent predominance of the different msp1/2 allelic types. Also, the isolates exhibited a high level of sequence variation. The K1 sequences were under balancing selection. It is needed to continuously follow the evolution of genetic diversity of Pf isolates in order to evaluate the impact of malaria control and elimination strategies implemented in country.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

DBS

Dried blood spot

DNA

Deoxyribonucleic acid

MI

Monoclonal infection

MOI

Multiplicity of infection

msp

Merozoite surface protein

NA

Not available

NC

Negative control

NCBI

National Center for Biotechnology Information

NVBCDP

National vector borne control disease programme

PCR

Polymerase chain reaction

Pf

Plasmodium falciparum

PI

Polyclonal infection

sSA

Sub-Saharan Africa

SP

Sulfadoxine-pyrimethamine

WHO

World Health Organization

Author’s contribution

LPKF and VS conceived and designed the study. LPKF conducted laboratory experiments, performed statistical analysis, and drafted the first version of the manuscript. JJ and GN helped in laboratory experiments, literature review, and data analysis. LPKF interpreted data with the help of JJ, GN and VS. LPKF conceived maps and retrieved the msp1/2 sequences from NCBI for performing phylogenetic analyses. JJ and GN participated in msp1/2 sequencing. LPKF and VS validated the reliability of sequencing data. VS provided material and reagents, critically revised the manuscript for intellectual content, and supervised the work at all stages. All authors read and approved the final version of the manuscript before submission.

Funding

This study has received no external financial support.

Declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Ethical approval

An ethical clearance was issued by the Ethics Committee of the ICMR-National Institute of Malaria Research (NIMR), India (N°PHB/NIMR/EC/2020/55).

Consent to participate

Authors, declare that they have participated in this work.

Consent to publication

Authors declare that they know the content of this manuscript and agreed to submit it for publishing.