In-silico classification and antigen library expression of Plasmodium falciparum STEVOR hypervariable infected erythrocyte surface-expressed multivariant protein family

doi:10.21203/rs.3.rs-3591844/v1

Download PDF

Research Article

In-silico classification and antigen library expression of Plasmodium falciparum STEVOR hypervariable infected erythrocyte surface-expressed multivariant protein family

https://doi.org/10.21203/rs.3.rs-3591844/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Malaria pathogenesis is dependent on complex interactions between host and parasite factors, where variant surface antigens such as the PfEMP1 protein family play a critical role in disease severity through various mechanisms, including immune evasion, cytoadherence and sequestration. The under characterised infected erythrocytes variant surface-expressed antigens of the STEVOR protein family are also implicated in cytoadherence and rosette formation exhibiting high antigenic variability, potentially contributing to parasite immune evasion. This study describes a novel approach for the construction of a comprehensive library of STEVOR recombinant antigens.

Methods

This study used all available STEVOR protein sequence data from the PlasmoDB database to classify the variability between STEVOR members within isolates and between isolates. We have used bioinformatic and mathematical approaches to design an in-silico model to study the protein family variability with 100% reproducibility when performed on the same data set. Using information from the model, we have designed constructs and have expressed them with the CyDisCo co-expression plasmid to create the first STEVOR recombinant antigen library in a competent E. coli expression system. Finally, we have proven the recombinants antigenicity using the multiplex magnetic bead-based assay: Luminex.

Results

The large hypervariable domain of STEVOR protein family exhibited the highest variability with a mean diversity of 52.1%, as compared to the semi-conserved and the conserved STEVOR domains. The variability was captured in a library of 13 representative sequences, mostly derived from West African isolates. Those variants were expressed as recombinant proteins in BL21(DE3) E. coli competent cells together with the CyDisCo co-expression plasmid. The recombinants varied in expression levels, but not in antigenicity. Three semi-conserved recombinant antigens were also expressed as controls and those although with smaller size, demonstrated higher reactivity as compared to the variable domain recombinants.

Conclusions

This study presents an in-silico model that effectively elucidates the spatial relationship between amino acid sequences, applicable to sequence data from any organism. Moreover, it presents the first library of STEVOR hypervariable domain recombinant antigens. Expressed antigens have potential applications in serological studies as indicators of exposure to infection and to further dissect STEVOR variants associated with severe malarial disease outcome.

Plasmodium falciparum

STEVOR

recombinant proteins

protein library

markers of infection

clustering

Malaria is a vector borne disease caused by an infection with Plasmodium spp., transmitted via bites of female pregnant Anopheles mosquitos (1). Of the 5 species causing disease in humans, Plasmodium falciparum is responsible for more than 90% of the world malaria mortality, with the greatest burden felt in Sub-Saharan Africa (1).

Malaria pathogenesis depends on multiple socio-economic, host and parasite factors, resulting in varying severity of the disease from mild and asymptomatic to severe and death (2). Some parasite virulence factors are associated with evasion of the host immune response through antigenic variation and prolonged and efficient infection via cytoadherence and formation of rosettes (2)^,(3).

Variant surface antigens (VSA) represent several protein families synthesized at distinct phases of the parasite blood stage life cycle and subsequently translocated and expressed on the surface of the infected erythrocytes (IEs) (4). Members of these hypervariable surface antigen families are characterized by a high degree of variation between members, are highly antigenic and are found to be associated with a number of important biological processes in malaria pathogenesis including sequestration and cytoadherence. (3)(5). The most well-characterized protein family is the P. falciparum Erythrocyte Membrane Protein1 family (PfEMP1). PfEMP1 protein variants have been shown to be important for parasite survival, transmission, and virulence due to their role in IE cytoadherence in the microvasculature and sequestration in rosettes. Ongoing research focuses on deciphering the associations between PfEMP1 variants, severe malaria, and their common pathogenic mechanisms with a goal of targeting these for therapeutic interventions (9). PfEMP1 sequence classification and diversity characterization could enable the predictions of host receptor interactions and provide foundation for the development of anti-adhesion strategies (9).

The Repetitive Interspersed Protein family (RIFIN; n = 160), have also been studied but with less research conducted than the var gene products. By contract, the Sub-Telomeric Variable Open Reading-frame protein family (STEVOR; n = 40) remains relatively uncharacterised (5). The var genes protein products: PfEMP1 protein family is encoded of approximately 60 var genes per parasite with limited genotypic overlap, enabling the parasite to escape the host immune system and to exhibit different tissue tropisms, particularly in cerebral and placental malaria (6). This remarkable diversity of variants is concentrated in few protein domains such as the Duffy Binding-Like (DBL) and Cysteine-Rich Interdomain region (CIDR) domains, also associated with binding specific endothelial receptors (7). A subgroup of PfEMP1 is found to be often expressed in severe malaria cases and particular var gene subset are linked to adverse clinical outcomes in malaria across all ages and geographical population, with specific var phenotypes associated with cerebral malaria in children (8).

STEVOR is a multi-copy gene family coded on the sub-telomeric region of all P. falciparum chromosomes, except for chromosome 5, with multiple variants per chromosome. A single parasite genome encodes approximately 40 stevor genes each coding for a different variant, with each parasite expressing a single variant within a polyclonal infection (5). The primary linear protein structure of a STEVOR protein is composed of a small signal peptide (SP), followed by a short variable domain (V1), a 5 amino acid long conserved motif encoding a protein trafficking peptide (PEXEL), a semi conserved domain (SC), a large variable domain (V2) flanked by two transmembrane domains (TM), and a short, terminal conserved domain (C), graphically represented in Fig. 1 (10).

Functionally STEVORs are found to be trafficked to the membrane of IE, localizing in proximity to knobs, shown to be involved in the mediation of the formation of rosettes and the cytoadherence of IE in the microvasculature (11),(12). STEVOR protein members differ mostly by their large variable domain (V2) which is the part of the protein that protrudes to the extracellular space and has been shown to be antigenic (10)(13). This high antigenic variability is an adaptive parasitic mechanism to help evade the host immune response during the infection (5)(14). Studies using peptide arrays have shown that there is an age and disease severity dependent seroreactivity and serorecognition to the STEVOR V2 and SC domains (10). Additional studies using recombinant STEVOR V2 and SC proteins have suggested the generation of an age and exposure dependent antibody responses against them, indicating their potential as markers of infection (15). However, to be able to further explore the antibody responses to these antigens and their potential as vaccine candidates or for use in serosurveilance further research is required to comprehensively characterize the diversity of this family in order to express recombinant antigens which will capture the highest level of diversity.

The aim of this study is to generate a library of STEVOR recombinant antigens which represents the P. falciparum STEVOR protein family for future serological studies in exploring potential of STEVORs as markers of P. falciparum infection exposure and potential for disease protection.

Methods

Protein sequence alignment and inspection

A total of 546 STEVOR protein sequences from 14 sequenced isolates were analyzed: 1 reference strain, 6 laboratory strains and 7 clinical isolates, were obtained from the PlasmoDB database, corresponding to all available protein sequence ‘hits’ under the search terms of “STEVOR and Plasmodium falciparum” (16). Sequences from ‘ML01’ and ‘TG01’ laboratory strains were not downloaded due to the large number of corresponding ‘hits’ of 236 and 272, respectively, annotated as STEVOR/RIFIN representing both families. Moreover, ‘ML01’ and ‘TG01’ strains are known to come from complex infections with more than one parasite strain, possibly explaining the large number of sequences under STEVOR/RIFIN annotation. Another 19 sequences were excluded from this database due to wrong annotation, or considerably too long or too short sequences in comparison to the rest, leaving a database of 527 protein sequences. The new protein database was subjected to multiple alignment with Mafft on Windows 10 terminal, using G-INS-I strategy with a global pair pairwise alignment, computed with the Smith-Waterman algorithm (17)^,(18). This algorithm was chosen as it is specifically tailored for sequences that have global homology (the entire length of each sequence is related to the entire length of every other sequence), as such is the case of STEVORs, and the global pair is aligning the entirety of each two sequences against each other in a pairwise matter, rather than aligning only the portions of the sequences where they best match (19). The alignment included a gap extension penalty for group-to-group with a default value of 0.123, to deter the introduction of gaps unless they are truly needed, and to refine the alignment when dealing with a mixture of closely and distantly related sequences (20). Subsequently, 34 shorter protein sequences with large gaps and/or annotated as ‘pseudogenes’ were excluded, resulting in the final database of 493 STEVOR aligned protein sequences, summarized in Table 1.

Domain sequences isolation

Isolation of the aligned large hypervariable domain (V2), semi-conserved domain (SC) and conserved domain (C), of all 493 STEVOR protein sequences were manually performed briefly as follows. The two transmembrane domains (2xTM) flanking the V2 domain were identified using TMHMM2.0 trans-membrane domain prediction software (21) isolating the approximately 70 amino acids long V2 protein sequences being between 158th and 215th amino acid position, also supported by literature (22). The C domain situated right after the second TM domain was isolated, identified to be 17 amino acids long. The SC domain was isolated by taking the amino acid sequences between the detected 5 amino acid long PEXEL motif and the first TM domain, identified to be approximately 123 amino acids long (23). Amino acid sequences per isolated domain were realigned using the same method outlined above and no further sequences were removed from either domain group. All isolated domains sequences for V2, SC and C domains can be found as Additional files 1, 2 and 3, respectively.

Clustering Model and Variant Library Selection

Amino acid sequences per domain were digitized into binary Boolean vectors. Each amino acid position for each sequence was challenged to be one of the 21 amino acids found in Plasmodium falciparum with the results recorded as present = 1 and not present = 0, resulting in 21 binary vectors, one for each amino acid. (24). All Boolean vectors were then stored into a matrix, which was further subjected to Singular Value Decomposition (SVD), resulting in a distance matrix of Euclidean distances (sPC) between sequences. sPC were used in a Principal Component Analysis (PCA) and visualized as 2D plots. All analysis and data visualization were performed on R 4.3 computational platform. Alignment diversity for each of the isolated domains was calculated once as Shannon Entropy index, measuring the diversity at each position and second as overall mean diversity of the pairwise distances, analyzed using “Biostrings” package, Bioconductor on R 4.3 computational platform (25). The V2 PCA was further subdivided into 9 equal quadrants and 13 sequences were selected to represent the STEVOR protein family, briefly as follows. The selection of sequences was based on (i) their two-dimensional position on the PCA plot and (ii) the geographical location of the strains. In the case of variants from multiple isolates clustering together, the variant coming from a Western sub-Saharan African isolate was selected. Three SC sequences from the P. falciparum reference strain 3D7 found to have potential as markers of infection were also selected to serve as controls for antigenicity (15).

Recombinant Antigens Library Expression

E. coli BL21(DE3) competent cells (Trans, China) were initially transformed with pMJS226 CyDisCo plasmid (University of Oulu, Finland), containing sulfhydryl oxidase (Evr1p), disulfide bond isomerase (PDI) and chloramphenicol (CMP) resistance cassette (26). Transformed cells were grown in LB media with 100 µg/ml CMP and were further transformed with pGEX-5X-1(RBS) plasmids (GenScript, UK) each containing one of the 13 STEVOR V2 variants, or one of the three Pf3D7 STEVOR SC variants, selected to be expressed as N-terminal GST-tag recombinant proteins. Transformed bacteria was grown in ZY-autoinduction media with 100 µg/mL Ampicillin (Amp) and 100 µg/mL CMP, further lysed using LM20 microfluidizer (Analytik LTD, UK). Recombinant proteins were affinity purified in a batch mode using Glutathione Sepharose 4B beads and quantified using Bradford protein quantification assay (Bio-Rad, UK).

To confirm recombinant antigenicity, the proteins were chemically coupled to MagPlex microsphere beads (Luminex, UK), following established protocols (27). Briefly, coupled beads were titrated at 6-point 8-fold dilutions against 5-point 2-fold dilutions of positive control serum pool of malaria hyperimmune Ugandan adults (PRIMS), starting from 1/100 down to 1/1600 serum concentration (15). Data in form of medium fluorescence intensity (MFI) was obtained for each titration point and a 4-parameter logistic regression was applied to each titration curve to obtain the EC50 point on sigmoidal curve. The median EC50 point across all dilutions was selected as the optimum protein coupling concentration for each recombinant.

STEVOR protein panel selection

The total number of 493 STEVOR sequences used in this study represent 90.30% of all protein sequences ‘hits’ under the defined search terms from PlasmoDB database, summarized in Table 1. The remaining 9.7% (n = 53) have been removed from the database due to detected mis annotations.

Table 1

Summary of *Plasmodium falciparum* STEVOR protein sequences used in the study with the number of variants per isolate directly taken from PlasmoDB and number of variants per isolate selected for the analysis.
Isolate	Location	Type	PlasmoDB hits (n)	Analysis sequences (n)
3D7	Netherlands	Reference strain	43	41
Dd2	Cambodia	Laboratory strain	40	37
GA01	Gabon	Laboratory strain	43	38
GB4	Cambodia	Laboratory strain	40	36
HB3	Cambodia	Laboratory strain	32	28
IT	Cambodia	Laboratory strain	41	37
SD01	Sudan	Laboratory strain	25	22
7G8	Brazil	Clinical isolate	25	21
CD01	Congo	Clinical isolate	44	40
GN01	Guinea	Clinical isolate	46	41
KE01	Kenya	Clinical isolate	37	34
KH01	Cambodia	Clinical isolate	43	39
KH02	Cambodia	Clinical isolate	41	38
SN01	Senegal	Clinical isolate	46	41

STEVOR V2 region presented higher variability with overall mean diversity of pairwise distances in the alignment of 52.1% as compared to 47.8% and 2.9% for STEVOR SC and STEVOR C, respectively (Fig. 3). Overall mean diversity of SC and V2 did not appear significantly different, however the amino-acid alignment length was significantly different of around 150 amino-acid compared to around 90 amino-acids, respectively, as demonstrated in the Shannon Entropy plot in Fig. 3. These differences in variability are also observed in the PCA analysis where V2 domain sequences are more dispersed compared to the SC sequence which most cluster together. The majority of the C domain sequences completely overlap in the PCA, which is expected since there is an overall mean diversity of 2.9% of the amino acids in the C domain alignment.

To select sequences which encompass the diversity across the STEVOR V2 sequences we selected 13 sequences (Fig. 4) which represent 8 variants from West Sub-Saharan Africa (5 from Guinea and 3 from Senegal), 2 variants from Central Africa (1 from Gabon and 1 from Congo), 1 variant from Kenia, East Africa, and 2 Cambodian variants. The STEVOR library panel was selected to be able to develop recombinant serological tools to help interrogate immunity to this protein family and expand our understanding of their function, in addition proving a template for the generation of protein products for other hypervariable protein families of pathological importance. Since P. falciparum malaria is mostly found in Sub-Saharan Africa, most of the library representatives were selected from isolates from the same region. The other African variants are selected from countries with high P. falciparum prevalence. The remaining 2 variants from Cambodia have been included to explore potential geographical differences in exposure to STEVOR variants.

Expression of the selected recombinant proteins in E. coli

The 13 V2 variants were then expressed as GST-tag recombinant proteins in BL21(DE3) E. coli competent cells, co-transformed with the CyDisCo plasmid (pMJS226). The sequences ranged between 272 and 283 amino-acids in length and expressed as recombinant protein products between 32 and 33 kDa, including the 211 amino acids long, 26 kDa GST-tag, summarised in Table 2. Although of similar sizes, the V2 recombinants varied substantially in terms of expression concentrations between 1.45 mg/ml (SN01_115) and 10.51 mg/ml (IT01_13776 and KE01_158). However, the differences in expression concentration were not reflected in the optimum protein-bead coupling concentrations for Luminex assay (median EC50 point), demonstrating the recombinant specific antigenicity properties, and excluding coupling bias in Luminex serological assays.

Table 2

Summary of protein name, amino acids length, molecular size, and protein concentrations after protein expression of V2 STEVOR library recombinants. The length and weight of the recombinants represent the V2 domain including the 211 amino acid long 26 kDa GST molecular tag.
Protein Name (PlasmoDB)	Abbreviated protein name	Quadrate Number (PCA)	Length (amino acids)	Molecular size (kDa)	Recombinant concentration (mg/ml)	EC50 point (mg/ml)
PfGN01_004003110-t41_1	GN01_4311	Q1	283	33	1.87	0.46
PfIT01_130077600-t41_1	IT01_13776	Q1	273	32	10.51	0.48
PfSN01_140006300-t41_1	SN01_1463	Q1	277	32	6.19	0.33
PfGN01_100006100-t41_1	GN01_161	Q4	283	33	4.12	0.15
PfKE01_100005800-t41_1	KE01_158	Q4	272	32	10.51	1.17
PfSN01_030005600-t41_1	SN01_356	Q4	277	32	3.28	1.56
PfCD01_008003800-t41_1	CD01_838	Q5	276	32	7.69	2.12
PfGN01_000800376-t41_1	GN01_8376	Q6	277	32	9.19	0.34
PfHB3_040028900	HB3_4289	Q6	279	33	8.63	0.20
PfGA01_070034600-t41_1	GA01_7346	Q7	281	32	5.62	1.97
PfGN01_020006800-t41_1	GN01_268	Q7	278	32	3.77	0.21
PfSN01_000011500-t41_1	SN01_115	Q7	273	32	1.45	0.73
PfGN01_130006600-t41_1	GN01_1366	Q8	279	32	4.33	0.58

The SC recombinant proteins ranged between 334 and 340 amino acids in length with size ranging between 40 and 41 kDa (Table 3). Although with much lower EC50 concentration points, the SC recombinant proteins exhibited much higher antigenicity than the V2 recombinants, illustrated in the titration plots of Fig. 5.

Table 3

Summary of protein name, amino acids length, molecular size, and protein concentrations after protein expression of SC STEVOR library recombinants. The length and weight of the recombinants represent the SC domain including the 211 amino acid long 26 kDa GST molecular tag.
Protein Name (PlasmoDB)	Abbreviated protein name	Length (amino acids)	Molecular size (kDa)	Recombinant concentration (mg/ml)	EC50 point (mg/ml)
Pf3D7_1300900	3D7_139	336	40	1.91	0.01
Pf3D7_0832000	3D7_832	340	41	5.34	0.24
Pf3D7_0832600	3D7_8326	334	41	3.28	0.04

We have developed an in-silico model, simplifying the complex amino-acid polymorphism inherent in hypervariable gene families. This novel approach, which is 100% reproducible when applied on the same sequence alignment data, has allowed us to develop the first recombinant protein STEVOR variant library, providing novel serological tools with which to explore the human immune response to the understudied but associated with pathology STEVOR protein family.

Other studies investigating the variable domain of STEVORs and dissecting the protein structure have been mostly focused on laboratory strains such as 3D7, Dd2 and HB3 (10)(28). However, the variability between STEVORs is firstly between variants within isolates, termed antigenic variability, and secondly between the variants between isolates, due to high recombination rate in the P. falciparum genome, which we have demonstrated in the more conserved SC and C domains using the PCA analysis in Fig. 4B and 4C (28)(29). Therefore, it is important to include both laboratory strains and clinical isolates to develop biologically relevant library.

Since most of the protein sequences between STEVOR variants are highly conserved, exploring the variability relationship between them using the whole protein sequences would reduce the resolution level in examining the differences. Hence, after the initial alignment the V2 domain was isolated for the variability analysis. The SC and C domains were also isolated to test the reliability of the model, expecting closer relationship due to higher conservation resulting in lower variability. Furthermore, it was worth considering that the expression of proteins in E. coli that encompass domains such as signal peptides, PEXEL motifs, and trans-membrane domains would most likely result in packaging of the proteins in inclusion bodies, thereby rendering them insoluble (30).

Since protein sequences are qualitative data, the numerical conversion of the sequences into binary Boolean vectors with a quantifiable score allowed for the estimating the relationships between variants in a quantitative way (24). The Single Value Decomposition method was used to construct the Euclidean distance matrix from the Boolean vectors data. This method was chosen as it effectively converts interrelated variables into a set of orthogonal variables, while identifying and arranging the dimensions along which data points display the greatest variability, therefore finding the optimum approximation of the original data points using reduced number of dimensions. (31). Furthermore, protein sequence data is a multivariable data due to the multiple types of amino acids present, therefore using a multivariate analysis of the distance matrix such as PCA was the most appropriate to observe the matrix data without harming its structure (32). The direct PCA analysis rotates the data and projects it to set of diagonal axes finding direction of differences considering both the amino acid composition and the genetic distance between the samples (32). Moreover, this analysis demonstrated 100% reproducibility when performed on the same dataset, unlike other sequence data clustering methods such as phylogenetic trees and the use of k-means, exhibiting variance in their outcomes with each permutation or with the change of numbers of k-means used, respectively (33).

The importance of conducting variant down selection process for the STEVOR library from the clustering model was derived from practical constrains, specifically the inability to express every variant from each strain due to resource and time limitations, as well as the concern of high likelihood of cross reactivity between closely related variants when further tested on samples. Alike RIFIN and PfEMP1 research, there is no accepted level of cross reactivity. A method of creating specific monoclonal antibodies to variants appearing close to each other in the PCA analysis could be performed to further calculate the difference in reactivity to measure the distance needed between variants for the lowest cross-reactivity. From the PCA analysis in Fig. 4A most of the variation in the V2 domain is explained by the second principal component (PC2) with sequences clustering mostly in quadrant 1 (Q1), Q4 and Q7. From each of these three quadrants we have selected 2 to 4 variants for the library, one of which, ‘GN01_1366’ is practically part of Q8, majority coming from West Sub-Saharan African stains, summarised in Table 2. The Congolese variant ‘CD01_838’ was the only variant selected from Q5 which lies exactly in the middle of all variants in the quadrant. The last 2 variants: ‘GN01_8376’ and ‘HB3_4289’ were selected from Q6. The former originated from a clinical isolate from Guinea, representing West Sub-Saharan Africa, an argument discussed previously. The latter comes from a laboratory strain from Cambodia, selected to introduce geographical diversity into the recombinant library, as well as due to the high prevalence of Cambodian laboratory strains variants in the sequence database. Q2, Q3 and Q9 had no STEVOR variants, while in Q8 the remaining 3 variants closely aligned with ‘CD01_838’ position from Q5 according to the first principal component (PC1), hence were not selected for the library. The further three variants selected to be expressed, represented the SC domain of STEVOR variants from the P. falciparum reference strain 3D7. The reason behind expressing SC domains alongside the V2 library is to serve as library controls, expected to be recognised by a diverse panel of malaria immune sera due to the conservation of the domain, also demonstrated in the titration plots of Fig. 5(10). The specific selection of the three variants: ‘Pf3D7_1300900’, ‘Pf3D7_0832000’ and ‘Pf3D7_0832600’ was based on previously published study showing a potential of those SC variants recombinants as markers of P. falciparum exposure (10), (15).

The decision was made to create a library of recombinant STEVOR proteins instead of isolates sequencing or employing peptide arrays. This choice was based on several factors. Firstly, the stevor genes are located on the sub telomeric region of P. falciparum chromosomes, adjacent to the telomer, where abundant repetitions of heterogeneous sequences, pose challenges for mapping and sequencing (34). Moreover, using recombinant proteins for the detection of antibody responses to the variants would offer a more precise evaluation of serorecognition compared to the peptide arrays, since recombinant proteins possess a fully folder structure while peptide arrays are made of only primary liner structure. Using a CyDisCo plasmid technology to transform E. coli competent cells for the expression of the recombinants was selected to improve the chances of proper protein folding and post-translational modifications (26). Proteins expressed in the cytoplasm of E. coli often require the formation of disulfide bonds between an even number of cysteine molecules (35). Manual examination of the V2 variant amino acid sequences confirmed that they contain 2 cysteine molecules (Additional file 1). The CyDisCo plasmid is therefore likely to benefit the correct folding of the proteins in the cytoplasm of E. coli due to the co-expression of sulfhydryl oxidase (Evr1p), allowing the production of disulfide-bonded proteins. Additionally, the expression of disulfide bond isomerase (PDI) ensures the editing of the disulfide bonds during protein folding via cleavage and formation cycles, a type of post-translational modification (26).

The study utilizes the Luminex serological assay to verify the antigenic properties of the expressed recombinant proteins. Figure 5 shows that all proteins possess differing reactivity characteristics against the PRISM serum pool positive control from Uganda (15). Specifically, ‘SN01_115’ and ‘SN01_356’ exhibit the lowest reactivity while ‘GA01_7346’ displays the highest for the V2 recombinants. The GA01 isolate comes from Gabon, geographically closer to Uganda, whereas SN01 strain comes from Senegal. This geographical difference may account for the observed disparity in seroreactivity between variants from the two strains. As expected, the SC recombinants exhibit considerably higher serorecognition when compared to all the V2 recombinants, due to the substantial conservation of the SC domains across all STEVORs demonstrated in Fig. 2B, and an expected cross reactivity between SC domains from different STEVOR variants.

Limitations

Selecting variants for the library out of the database of 493 STEVOR V2 protein sequences was based on their position according to the PCA analysis and biological reasoning. The selected PRISM positive control to assess the reactivity of these variants is uncertain as we don’t know whether the serum pool from the Ugandan hyperimmune individuals would specifically recognise these variants. This is observed in Fig. 5, where the reactivity against the V2 variants was found to be relatively low. On the contrary, we also anticipated some degree of cross-reactivity among the V2 variants. Unfortunately, similar to the case for tother variable antigen families, such as RIFIN and PfEMP1, these is no definitive mathematical or biological method to gauge the necessary level of divergence between the variants to minimize the potential cross-reactivity. Generation of monoclonal antibodies against each of the variants identified would facilitate further characterisation and epidemiological surveillance.

The novel approach described in this study provides a robust in-silico model to help interrogate complex sequence dynamics by highlighting the spatial relationship between sequences. This model is applicable to any amino acid sequence alignment data with 100% reproducibility when performing permutations, irrespective of species. Additionally, this is the first study to generate a recombinant antigen library representing the variable domain of the neglected surface expressed protein family STEVOR. The protein library tool of V2 variants would be a valuable resource for subsequent investigations, such as serosurveillance. It can be employed to explore the degree of seroreactivity and recognition using serum or plasma samples from various populations, potentially identifying variants associated with enhanced parasite fitness, or severe clinical disease, warranting further exploration through functional assays in vitro.

STEVOR

Sub telomeric Open Reading Frame protein family

PfEMP1

Plasmodium falciparum Erythrocyte Membrane Protein 1

WHO

World Health Organisation

LLIN

Long-Lasting Insecticidal Net

MSP1

Merozoite Surface Protein 1

AMA1

Apical Membrane Antigen 1

VSA

Variable Surface Antigens

Infected Erythrocytes

RIFIN

Repetitive Interspersed protein family

DBL

Duffy-Binding Like domain

CIDR

Cysteine-Rich Interdomain Region

Signal Peptide

Small Variable domain

PEXEL

Translocating Element

Semi-Conserved domain

Large Variable domain

Trans-Membrane domain

Conserved domain

SVD

Singular Value Decomposition

sPC

Principal Component

PCA

Principal Component Analysis

CMP

Chloramphenicol

Lysogeny Broth

GST

Glutathione S-transferase

MFI

Median Fluorescent Intensity

Quadrant

Ethical approval and consent to participate.

The serum samples used in this study come from the Program for Resistance, Immunology, Surveillance, and Modelling of Malaria in Uganda longitudinal cohort (PRISM), approved by the Ethics Committee of LSHTM (reference number: 15823) and the Research and Ethics Committee of the Makerere University School of Medicine in Kampala, Uganda (reference number: 2011-167).

Availability of data

The data sets supporting the conclusions of this article are included within the article and its additional files.

Competing interests

All authors declare that the research was conducted in the absence of any financial, personal, or professional relationships that could be construed to have influenced the work.

Funding

This project is nested within the clinical trial: “Adjunctive Ivermectin Mass Drug Administration for Malaria Control (MATAMAL)” ClinicalTraials.gov Identifier: NCT04844905, funded by the Joint Global Health Trial scheme (JGHTs) with PI Dr Anna Last, (Grant Code MR/S005013/1). This scheme is jointly funded by the Foreign, Commonwealth and Development Office (FCDO), the Medical Research Council (MRC), the National Institute for Health Research (NIHR) and the Wellcome Trust foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Authors contributions

HV was responsible for the design, acquisition and analysis, interpretation of results and drafting of the work. KT was involved in the design of the work and interpretation of results. EB was involved in the design of the work. AL and KT were responsible for securing funds for the study. All authors have approved the submitted version and have agreed to be personally accountable for their own contributions of this work.

Acknowledgement

Special acknowledgements to Dr James Ashall for contributing to the graphical illustrations of this study.

World malaria report 2022 - World Health Organization. - Google Books [Internet]. [cited 2023 Mar 13]. Available from: https://books.google.gm/books/about/World_Malaria_Report_2022.html?id=ST-hEAAAQBAJ&printsec=frontcover&source=kp_read_button&hl=en&redir_esc=y#v=onepage&q&f=false.
Miller LH, Baruch DI, Marsh K, Doumbo OK. The pathogenic basis of malaria. Nature [Internet]. 2002 Feb 7 [cited 2021 Nov 17];415(6872):673–9. Available from: https://pubmed.ncbi.nlm.nih.gov/11832955/.
Niang M, Bei AK, Madnani KG, Pelly S, Dankwa S, Kanjee U et al. The variant STEVOR protein of Plasmodium falciparum is a red cell binding protein important for merozoite invasion and rosetting. Cell Host Microbe [Internet]. 2014 Jul 7 [cited 2023 Aug 29];16(1):81. Available from: /pmc/articles/PMC4382205/.
Chan JA, Fowkes FJI, Beeson JG. Surface antigens of Plasmodium falciparum-infected erythrocytes as immune targets and malaria vaccine candidates. [Internet]. Vol. 71, Cellular and molecular life sciences: CMLS. Springer; 2014 [cited 2020 Oct 16]. p. 3633–57. Available from: https://link.springer.com/article/10.1007/s00018-014-1614-3.
Scherf A, Lopez-Rubio JJ, Riviere L. Antigenic Variation in Plasmodium falciparum. Annu Rev Microbiol [Internet]. 2008 Oct 11 [cited 2019 May 20];62(1):445–70. Available from: http://www.annualreviews.org/doi/10.1146/annurev.micro.61.080706.093134.
Kraemer SM, Kyes SA, Aggarwal G, Springer AL, Nelson SO, Christodoulou Z et al. Patterns of gene recombination shape var gene repertoires in Plasmodium falciparum: Comparisons of geographically diverse isolates. BMC Genomics [Internet]. 2007 Feb 7 [cited 2023 Nov 7];8(1):1–18. Available from: https://bmcgenomics.biomedcentral.com/articles/10.1186/1471-2164-8-45.
Smith JD. The role of PfEMP1 adhesion domain classification in Plasmodium falciparum pathogenesis research. Mol Biochem Parasitol [Internet]. 2014 [cited 2023 Nov 7];195(2):82. Available from: /pmc/articles/PMC4159067/.
Duffy F, Bernabeu M, Babar PH, Kessler A, Wang CW, Vaz M et al. Meta-analysis of Plasmodium falciparumvar Signatures Contributing to Severe Malaria in African Children and Indian Adults. MBio [Internet]. 2019 Mar 1 [cited 2023 Nov 7];10(2). Available from: /pmc/articles/PMC6495371/.
Chan JA, Boyle MJ, Moore KA, Reiling L, Lin Z, Hasang W et al. Antibody Targets on the Surface of Plasmodium falciparum–Infected Erythrocytes That Are Associated With Immunity to Severe Malaria in Young Children. J Infect Dis [Internet]. 2019 Mar 3 [cited 2023 Nov 7];219(5):819. Available from: /pmc/articles/PMC6376912/.
Zhou AE, Berry AA, Bailey JA, Pike A, Dara A, Agrawal S et al. Antibodies to Peptides in Semiconserved Domains of RIFINs and STEVORs Correlate with Malaria Exposure. mSphere. 2019;4(2).
Singh H, Madnani K, Lim YB, Cao J, Preiser PR, Lim CT. Expression dynamics and physiologically relevant functional study of STEVOR in asexual stages of Plasmodium falciparum infection. Cell Microbiol [Internet]. 2017 Jun 1 [cited 2023 Nov 8];19(6). Available from: https://pubmed.ncbi.nlm.nih.gov/28030753/.
Sherman IW, Eda S, Winograd E. Cytoadherence and sequestration in Plasmodium falciparum: defining the ties that bind. Microbes Infect. 2003;5(10):897–909.
Cheng Q, Cloonan N, Fischer K, Thompson J, Waine G, Lanzer M et al. stevor and rif are Plasmodium falciparum multicopy gene families which potentially encode variant antigens. Mol Biochem Parasitol [Internet]. 1998 Nov 30 [cited 2020 Oct 7];97(1–2):161–76. Available from: https://pubmed.ncbi.nlm.nih.gov/9879895/.
Lavazec C, Sanyal S, Templeton TJ. Hypervariability within the Rifin, Stevor and Pfmc-2TM superfamilies in plasmodium falciparum. Nucleic Acids Res [Internet]. 2006 Dec [cited 2020 Oct 5];34(22):6696–707. Available from: /pmc/articles/PMC1751529/?report = abstract.
Vasileva H, Chopo-Pizarro A, Ooko M, Dumont E, Wu L, Ssewanyana I et al. Novel hypervariable erythrocyte surface expressed recombinant proteins show promise as serological markers of exposure to Plasmodium falciparum infection. bioRxiv [Internet]. 2023 Nov 2 [cited 2023 Nov 7];2023.10.31.564916. Available from: https://www.biorxiv.org/content/10.1101/2023.10.31.564916v1.
PlasmoDB. : The Plasmodium Genomics Resource [Internet]. [cited 2020 Mar 30]. Available from: https://plasmodb.org/plasmo/.
Rozewicki J, Li S, Amada KM, Standley DM, Katoh K. MAFFT-DASH: integrated protein sequence and structural alignment. Nucleic Acids Res [Internet]. 2019 Jul 7 [cited 2023 Jun 20];47(W1):W5. Available from: /pmc/articles/PMC6602451/.
Katoh K, Kuma KI, Toh H, Miyata T. MAFFT version 5: Improvement in accuracy of multiple sequence alignment. Nucleic Acids Res [Internet]. 2005 [cited 2020 Oct 16];33(2):511–8. Available from: /pmc/articles/PMC548345/?report = abstract.
Katoh K, Standley DM. MAFFT Multiple Sequence Alignment Software Version 7: Improvements in Performance and Usability. Mol Biol Evol [Internet]. 2013 Apr [cited 2023 Oct 31];30(4):772. Available from: /pmc/articles/PMC3603318/.
Durbin R, Eddy SR, Krogh A, Mitchison G. Multiple sequence alignment methods. Biol Seq Anal [Internet]. 1998 Jul 6 [cited 2023 Oct 31];135–60. Available from: https://www.cambridge.org/core/books/biological-sequence-analysis/multiple-sequence-alignment-methods/217DF949061D16E0D2A4B034233DA7F4.
Krogh A, Larsson B, Von Heijne G, Sonnhammer ELL. Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol [Internet]. 2001 Jan 19 [cited 2023 Jan 5];305(3):567–80. Available from: https://pubmed.ncbi.nlm.nih.gov/11152613/.
Hypervariability within the Rifin., Stevor and Pfmc-2TM superfamilies in Plasmodium falciparum [Internet]. [cited 2020 Mar 9]. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1751529/.
Boddey JA, Moritz RL, Simpson RJ, Cowman AF. Role of the Plasmodium export element in trafficking parasite proteins to the infected erythrocyte. Traffic [Internet]. 2009 [cited 2020 Oct 7];10(3):285–99. Available from: /pmc/articles/PMC2682620/?report = abstract.
Konishi T, Matsukuma S, Fuji H, Nakamura D, Satou N, Okano K. Principal Component Analysis applied directly to Sequence Matrix. Sci Rep [Internet]. 2019 Dec 1 [cited 2021 Apr 8];9(1):1–13. https://doi.org/10.1038/s41598-019-55253-0.
Pagès H, Aboyoun PGR, Biostrings. Efficient manipulation of biological strings version 2.58.0 from Bioconductor [Internet]. 2003 [cited 2023 Nov 1]. Available from: https://rdrr.io/bioc/Biostrings/.
Gaciarz A, Khatri NK, Velez-Suberbie ML, Saaranen MJ, Uchida Y, Keshavarz-Moore E et al. Efficient soluble expression of disulfide bonded proteins in the cytoplasm of Escherichia coli in fed-batch fermentations on chemically defined minimal media. Microb Cell Fact [Internet]. 2017 Jun 15 [cited 2023 Aug 28];16(1):1–12. Available from: https://microbialcellfactories.biomedcentral.com/articles/10.1186/s12934-017-0721-x.
Tetteh KKA, Wu L, Hall T, Ssewanyana I, Oulton T, Patterson C et al. Optimisation and standardisation of a multiplex immunoassay of diverse Plasmodium falciparum antigens to assess changes in malaria transmission using sero-epidemiology. Wellcome Open Res [Internet]. 2020 Apr 23 [cited 2020 Oct 15];4:26. https://doi.org/10.12688/wellcomeopenres.14950.1.
Lavazec C, Sanyal S. and TJT. Hypervariability within the Rifin, Stevor and Pfmc-2TM superfamilies in Plasmodium falciparum. Nucleic Acids Res [Internet]. 2006;34(22):6696–707. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1751529/.
Jiang H, Li N, Gopalan V, Zilversmit MM, Varma S, Nagarajan V et al. High recombination rates and hotspots in a Plasmodium falciparum genetic cross. Genome Biol [Internet]. 2011 Apr 4 [cited 2023 Sep 13];12(4):R33. Available from: /pmc/articles/PMC3218859/.
Bhatwa A, Wang W, Hassan YI, Abraham N, Li XZ, Zhou T. Challenges Associated With the Formation of Recombinant Protein Inclusion Bodies in Escherichia coli and Strategies to Address Them for Industrial Applications. Front Bioeng Biotechnol [Internet]. 2021 Feb 10 [cited 2023 Jul 18];9. Available from: /pmc/articles/PMC7902521/.
Baker K. Singular Value Decomposition Tutorial. 2005.
Jolliffe I, Jolliffe. Ian. Principal Component Analysis. In: Wiley StatsRef: Statistics Reference Online [Internet]. Chichester, UK: John Wiley & Sons, Ltd; 2014 [cited 2017 Oct 18]. http://doi.wiley.com/10.1002/9781118445112.stat06472.
k-Means Advantages and Disadvantages. | Machine Learning | Google for Developers [Internet]. [cited 2023 Oct 10]. Available from: https://developers.google.com/machine-learning/clustering/algorithm/advantages-disadvantages.
Kwapisz M, Morillon A. Subtelomeric Transcription and its Regulation. J Mol Biol [Internet]. 2020 Jul 7 [cited 2023 Aug 29];432(15):4199. Available from: /pmc/articles/PMC7374410/.
Berkmen M. Production of disulfide-bonded proteins in Escherichia coli. Protein Expr Purif. 2012;82(1):240–51.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

In-silico classification and antigen library expression of Plasmodium falciparum STEVOR hypervariable infected erythrocyte surface-expressed multivariant protein family

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Background

Results

Discussion

Conclusion

Abbreviations

Declarations

Ethical approval and consent to participate.

Availability of data

Competing interests

Funding

Authors contributions

Acknowledgement

References

Additional Declarations

Supplementary Files

Status:

Version 1