The samples for this study were collected in a community-based cross-sectional study conducted from December 2022 to July 2023. Two regions, Geita and Kigoma, were purposefully chosen due to their high malaria transmission rates in Tanzania. In 2022, malaria prevalence in both regions was 13%. From each region, two districts were randomly selected (Geita; Nyang’hwale and Chato, Kigoma; Kibondo and Kasulu), and from each district, two villages were randomly chosen ( Figure 1 ). In total, the study covered eight villages across Tanzania’s high-endemic areas for P. falciparum.

A research team consisting of a medical doctor and laboratory technician visited house-to house and collected whole blood samples from asymptomatic P. falciparum participants. After obtaining consent, a structured questionnaire was used to gather demographic information from participants. Asymptomatic cases were defined as individuals with a positive microscopic examination for any level of parasitemia, who had been afebrile, showed no malaria symptoms in the past 4-5 days, and had not used antimalarial medication in the past 7 days. Immediately after collection, blood was tested for malaria infections using the Bioline™ Malaria Ag P.f/Pan rapid diagnostic test (RDT) (Abbott, Chicago, IL, USA). Additionally, thin blood smears were prepared and stained with Giemsa for examination under a light microscope by two experienced microscopists. Parasitemia was counted according to the WHO Malaria Microscopy Standard Operating Procedure, based on the observation of 200 white blood cells. Patients identified as asymptomatic cases with positive P. falciparum tests were treated with artemether-lumefantrine in accordance with Tanzania’s malaria treatment policy. The specimens were then processed, with whole blood preserved as dried blood spots (DBS) on Whatman 3MM filter paper and serum separated and stored in a laboratory near the collection site. Serum from asymptomatic patients was kept at -20°C until use. The DBS and serum were subsequently transported to Kangwon National University for molecular diagnosis, PfRh5 sequencing, and antigenicity screening. A total of 242 whole blood samples were collected, and 242 serum (100%) were separated for antigenicity screening. However, due to the low quality of DBS and low parasitemia, only 164 isolates (67.8%) were successfully analyzed for the full-length of pfrh5 sequence. Healthy sera (n = 56) for the protein array negative control were obtained from individuals under the age of 10 living in malaria non-endemic areas of the Republic of Korea during general health check-ups at Kangwon National University Hospital.

All clinical samples were collected following ethical guidelines and approved protocols from the National Institute for Medical Research (NIMR), a division of the Ministry of Health (MoH) in Tanzania (NIMR/HQ/R.81/Vol.IX/4114). The protocol for clinical samples handling was also approved by the Ethical Review Board at Kangwon National University (KWNUIRB-2022-06-008). Informed written consent was obtained from all adult participants, while for child participants, consent was secured from a parent or guardian.

Genomic DNA was extracted from 50 µL of dried blood spots (one spot) using the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany), following the manufacturer’s protocol, with a final elution volume of 50 µL. A total of 164 isolates from Tanzania were processed to amplify and sequence the pfrh5 gene. PCR was performed to amplify the target gene for sequencing and gene variation analysis. To cover the entire pfrh5 gene sequence, the sequencing primers were designed for three regions ( Supplementary Table S1 ). The PCR mixture, prepared in a total volume of 20 μL, contained KOD one PCR Master mix (Toyobo, Osaka, Japan), 1 μL of template DNA (gDNA from asymptomatic patients), 16 μL of distilled deionized water (DDW), and 1.0 μL of each 10 μM primer. The PCR conditions included an initial denaturation at 98°C for 7 minutes, followed by 45 cycles of 98°C for 10 seconds, 58°C for 5 seconds, and 68°C for 10 seconds, with a final extension at 68°C for 10 minutes. The amplicons were purified and visualized using electrophoresis with 1.2% agarose gel. The purified fragments were then subjected to Sanger sequencing (Nbit, Chuncheon, Republic of Korea) using sequencing primers ( Supplementary Table S1 ). The sequences from each isolate were confirmed for quality by chromatogram analysis and assembled using SnapGene software version 2.3.2 and the Lasergene package to obtain the complete pfrh5 gene sequence. The low-frequency genetic mutations were confirmed by performing three independent experiments. The detailed whole gene sequences of pfrh5 are available at GenBank (accession numbers: PQ337615 - PQ337778).

Pfrh5 nucleotide diversity (π) measures the average number of nucleotide differences per site between two sequences. The number of polymorphic sites, haplotypes, and haplotype diversity (Hd) were calculated using DnaSP software version 5.10 to evaluate sequence diversity within the group (24, 25). To investigate evidence of natural selection (non-random evolution) within the species, we applied Tajima’s D, Fu and Li’s D* and F*, and Fu’s Fs tests. Under neutral evolution, Tajima’s D is expected to be zero. Significant positive values of Tajima’s D may suggest deviations from neutral mutation, potentially reflecting population contraction or balancing selection, while significant negative values could indicate recent population expansion following a bottleneck or a recent selective sweep (26). Fu and Li’s D* and F* tests, with significant positive values, generally point to population contraction due to natural selection pressure, whereas negative values often suggest population expansion and an excess of singletons (27). Fu’s Fs test evaluates allele or haplotype distribution, with negative values indicating an excess of alleles, which may suggest recent population expansion or genetic hitchhiking (28). Furthermore, Nei and Gojobori’s method was used to calculate the proportion of synonymous substitutions per synonymous site (d_S) and non-synonymous substitutions per non-synonymous site (d_N), with 1,000 bootstrap replications in MEGA 11 software. A d_N-d_S greater than zero typically indicates positive selection, while a ratio less than one suggests purifying selection. The McDonald-Kreitman (MK) test was used to compare within-species variation to divergence between species to assess neutrality. In this test, P. reichenowi rh5 (PRCDC_0421300), which is a species in Laverania clade that related to P. falciparum, was used as an out-group species and the analysis was performed with DnaSP software. Haplotype diversity of Pfrh5 was analyzed using DnaSP, and clustering patterns were visualized using the median-joining method in Network 10.0 software.

The tertiary structure of PfRh5 was obtained from the Protein Data Bank (PDB, https://www.rcsb.org/) using IDs 8CDD and 4u0q. The 8CDD structure represents the PfRh5-PfCyRPA-PfRipr complex, while the 4u0q structure contains the PfRh5-Basigin binding complex. Both tertiary structures were aligned using ClustalW based on their sequences by PyMol, and a predicted model of the Basigin-PfRh5-PfCyRPA complex was generated. The structure was then visualized in PyMol to highlight the positions of non-synonymous mutations, particularly on the interaction surface of PfRh5 residues.

The recombinant PfRh5-ecto construct was designed based on the 3D7 strain sequence (PF3D7_0424100) and included amino acids 25-526 with an expected molecular weight of 64.4 kDa including the tag. This construct comprises all non-synonymous mutation residues necessary for antigenicity screening. The protein was produced using the HEK293 EBNA1-6E cell-based expression system (29). To generate the protein, the codon-optimized pfrh5 gene was amplified with a specific primer set ( Supplementary Table S1 ) and subcloned into the expression vector pTT5-8x His using the In-Fusion^® HD Cloning Kit (Clontech, Mountain View, CA, USA). The final construct was then transfected into HEK293 EBNA1-6E cells using transfection grade linear polyethylenimine hydrochloride. After five days for transfection, the secreted proteins were collected from the culture supernatant and purified using Ni-NTA agarose (QIAGEN) with elution buffer (350 mM imidazole, 50 mM HEPES, 5% glycerol, 150 mM NaCl). The purified recombinant protein (1 µg) was prepared with 2x reducing buffer and separated by 4-12% gradient Bis-Tris Mini Protein Gel (Invitrogen, Waltham, MA, USA). After electrophoresis, the Bis-Tris gels were stained with Coomassie Brilliant Blue (Sigma-Aldrich).

Protein microarray analysis was performed to assess the antigenic variation of PfRh5. Three aminopropyl-coated slides were prepared as described previously (30). Each slide was printed with PfRh5-ecto at an optimized concentration of 50 ng/µL per spot and incubated for 2 hours at 37°C. After incubation, the slides were blocked with a solution of 5% BSA in PBS-T (0.1% Tween 20) for 1 hour at 37°C. Sera from healthy individuals and asymptomatic P. falciparum patients were diluted 1:50 in PBS-T and applied to each spot in duplicate on the slides for 1 hour at 37°C. The arrays were then visualized with 10 ng/µL of Alexa Fluor 546 goat anti-human IgG (Invitrogen) in PBS-T for 1 hour at 37°C, followed by scanning with the InnoScan 300 (INNOPSYS, Carbonne, France). The positive cut-off values were determined by adding two standard deviations to the average mean fluorescence intensity (MFI) of healthy sera, which were used as a negative control.

The antigenicity screening results were analyzed and visualized using GraphPad Prism v8 (GraphPad Software, San Diego, CA, USA) and SigmaPlot v12.0 (Systat Software Inc., San Jose, CA, USA) software. For the protein array data, an unpaired Student’s t-test was used to compare the measured values between groups. Pearson correlation (ρ) analysis was applied to evaluate the relationship between antigenicity and clinical factors such as age and parasitemia (parasites/μL). Significant differences were considered as p < 0.05 at 95% CI.

This study collected P. falciparum isolates from asymptomatic participants in Geita and Kigoma regions in Tanzania, among of the regions showed high P. falciparum prevalence in 2022. From each region, two districts were randomly selected followed by a random selection of two villages from each district. In Geita region, Chato and Nyang’hwale districts were chosen. From Chato district, selected villages were Rwantaba (GCR) and Ihanga (GCI) whereas Kayenze (GNK) and Nyangalamila (GNN) were respectively chosen from Nyang’hwale district. In Kigoma region, Kasulu and Kibondo districts were selected. From Kasulu district, the villages chosen were Mugombe (KKM) and Nyamnyusi (KKN), while from Kibondo district, the selected villages were Bunyambo (KKB) and Kumuhasha (KKK) ( Figure 1A ).

Baseline information and parasitemia from asymptomatic participants who were positive for P. falciparum were compared with demographic factors ( Supplementary Table S2 ). Significant differences in parasitemia were observed between regions, with Geita showing a mean parasitemia of 4,400 ± 20,343 parasites/µL and Kigoma showing 9,381 ± 38,880 parasites/µL ( Figure 1B ). However, there was no significant differences between parasitemia and gender ( Figure 1C ). When comparing by age groups, younger than 5 years old exhibited lower parasitemia levels, whereas those aged in 11 to 25 years showed significantly higher parasitemia compared to the younger age group (<5 years old) ( Figure 1D ).

The complete pfrh5 (PF3D7_0424100) genomic DNA sequence spans 1,788 base pairs and is located on chromosome 4. It comprises one intron and two exons, with the coding sequence in total 1,581 base pairs (527 amino acids) and a predicted molecular weight of 62.9 kDa ( Figures 2A, B ). Nucleotide diversity (π) among clinical isolates showed limited variation in codons ( Figure 2A ). Most mutations were non-synonymous, with Y147H (16.5%), H148D (19.5%), C203Y (91.5%), and K429N (36.0%) being relatively high in prevalence ( Figure 2B ). Non-synonymous mutations with low frequencies were observed at V371I (0.6%) and I410M (1.2%). Additionally, novel mutations were identified at D65H (1.2%), H170N (1.2%), and I227M (1.2%), both of which occurred at low frequencies ( Figure 2B ).

The protein tertiary structure, obtained from PDB (ID: 8CDD), represents the PfRh5-PfCyRPA-PfRipr (RCR) complex and highlights the non-synonymous mutation residues within the PfRh5 structure (159-516 aa.) ( Figure 2C ). PfRh5 directly interacts with PfCyRPA, while another region of PfRh5 binds with Basigin (CD147) as a host cell receptor. We also obtained the PfRh5-Basigin complex structure from the PDB (ID: 4u0q) and reconstructed the Basigin-PfRh5-PfCyRPA complex ( Figure 2D ). The amino acid residues involved in the interaction with PfCyRPA and Basigin were well conserved ( Figure 2D ).

Among the 164 isolates from eight villages, we identified 12 polymorphic sites in pfrh5, comprising 3 synonymous and 9 non-synonymous sites. The highest nucleotide diversity (π) was observed in Kibondo district (π ± SD., 0.00105 ± 0.00011), with Kumuhasha (KKK, 0.00109 ± 0.00009), and Bunyambo (KKB, 0.00098 ± 0.00021) ( Table 1 ; Supplementary Figure S1A ). Following Kibondo district, Kasulu district (0.00082 ± 0.00012) showed high nucleotide diversity in Mugombe (KKM, 0.00099 ± 0.00020), with the lowest diversity observed in Nyamnyusi (KKN, 0.00070 ± 0.00016) ( Table 1 ; Supplementary Figure S1A ). Comparatively, Geita region (0.00080 ± 0.00007) showed lower nucleotide diversity than Kigoma region (0.00093 ± 0.00009) ( Table 1 ; Supplementary Figure S1A ). Rwantaba village (GCR, 0.00086 ± 0.00013) in Chato district exhibited the highest nucleotide diversity within Geita region. This was followed by Kayenze (GNK, 0.00080 ± 0.00016), Nyangalamila (GNN, 0.00079 ± 0.00012), and Ihanga (GCI, 0.00073 ± 0.00016) ( Table 1 ; Supplementary Figure S1A ). Overall, the population-wide nucleotide diversity was 0.00086 ± 0.00006 (π ± S.D.) in the study areas ( Table 1 ; Supplementary Figure S1A ).

To investigate whether allele frequencies at the pfrh5 polymorphisms reflect evidence of selection rather than random mutation, several tests were conducted at both intra- and inter-species levels. Among the intra-species tests, Tajima’s D (-0.88994) indicates an excess of low-frequency alleles, suggesting recent population expansion. Similarly, the negative values for Fu and Li’s D* (-1.27193) and Fu and Li’s F* (-1.35296) reinforce this interpretation, pointing to an increase in rare mutations. However, these values did not reach statistical significance, indicating weak evidence for that interpretation ( Table 1 ). In contrast, the strong negative value of Fu’s Fs (-8.71) further supports the notion of population expansion. The significant d_N/d_S ratio of 1.85 (p = 0.03) suggests positive selection, where beneficial non-synonymous mutations are favored, indicating adaptive evolution in the population ( Table 2 ). The inter-species McDonald-Kreitman (MK) test ratio for PfRh5 was greater than 1, indicating purifying selection, but it did not reach statistical significance (p = 0.89) ( Table 2 ). Together, these metrics suggest that PfRh5 is experiencing natural selection characterized by recent population expansion and positive selection, where advantageous mutations are proliferating while deleterious mutations are being purged from the population.

Geographical haplotype networks of PfRh5 were constructed using the Median-Joining method, revealing a total of 17 haplotypes with mixed geographic patterns ( Figure 3 ). The network analysis indicated that PfRh5 underwent population expansion within the study area, without showing a distinct geographical pattern.

PfRh5-ecto (25-526 aa.) domains were expressed for antigenicity screening ( Figure 1B ). Recombinant proteins were evaluated for antigenicity across eight villages ( Figure 4A ). Total IgG reactivity was detected in both healthy sera (MFI ± S.D., 13,198 ± 5,556) and asymptomatic patient sera (23,189 ± 13,663) for PfRh5-ecto within the total population. A PfRh5-specific humoral immune response was detected in 34.7% of asymptomatic patients following exposure to P. falciparum ( Table 3 ). The humoral immune response was higher in patients from Geita (57.9%) region compared to those from Kigoma (11.6%) region ( Figure 4B ; Table 3 ). A direct comparison of the total IgG response by area revealed significant differences between Geita and Kigoma regions ( Supplementary Figure S1B ). Furthermore, when comparing the total population IgG response at the district level (Chato, Nyang’hwale, Kibondo, and Kasulu), variations in antigenicity were observed based on geographical location ( Supplementary Figure S1B ).

Correlation analysis of PfRh5 humoral immune response by gender showed no significant differences ( Figure 4C ). The parasitemia with antigenicity showed negative correlation, however no significant differences (ρ = 0.170, p = 0.103) ( Figure 4D ). In contrast, antigenicity showed a significant positive correlation with patient age (ρ = 0.170, p = 0.008) ( Figure 4E ). The most important factors considered were nucleotide diversity (π) and total IgG response. A comparison of the correlation between total IgG response and nucleotide diversity clearly showed a significant negative correlation (ρ = -0.463, p = 0.006) across different geographical locations ( Figure 4F ). To clearly verify the correlation between genetic polymorphism and antigenicity, we conducted an analysis within regions ( Figure 4G ). For data normalization, we divided each nucleotide diversity (π) and mean fluorescence intensity values by the overall mean, normalizing them to “1”. Then, we compared the genetic polymorphism and antigenicity by each region. The results showed that, Geita region had a fold change of mean nucleotide diversity versus antigenicity greater than 1, while the Kigoma region exhibited a fold change of less than 1 ( Figure 4G ). This result indicate that those regions with lower genetic polymorphism maintained higher antigenicity, whereas regions with higher genetic polymorphism displayed lower antigenicity.