Sub-grouping of Plasmodium falciparum 3D7 var genes based on sequence analysis of coding and non-coding regions

Abstract

Background: The variant surface antigen family Plasmodium falciparum erythrocyte membrane protein-1 (PfEMP1) is an important target for protective immunity and is implicated in the pathology of malaria through its ability to adhere to host endothelial receptors. The sequence diversity and organization of the 3D7 PfEMP1 repertoire was investigated on the basis of the complete genome sequence.

Methods: Using two tree-building methods we analysed the coding and non-coding sequences of 3D7 var and rif genes as well as var genes of other parasite strains.

Results: var genes can be sub-grouped into three major groups (group A, B and C) and two intermediate groups B/A and B/C representing transitions between the three major groups. The best defined var group, group A, comprises telomeric genes transcribed towards the telomere encoding PfEMP1s with complex domain structures different from the 4-domain type dominant of groups B and C. Two sequences belonging to the var1 and var2 subfamilies formed independent groups. A rif subgroup transcribed towards the centromere was found neighbouring var genes of group A such that the rif and var 5' regions merged. This organization appeared to be unique for the group A var genes

Conclusion: The grouping of var genes implies that var gene recombination preferentially occurs within var gene groups and it is speculated that the groups reflect a functional diversification evolved to cope with the varying conditions of transmission and host immune response met by the parasite.

Figure 1

Schematic representation of head-to-head genomic organisation of rif and upsA flanked var genes. Nine genes are flanked by a rif gene, which has its initiation codon approximately 3 or 4 kb upstream from the var initiation codon, and one var gene by another var gene at -2 kb. Punctured lines represent upsA, dotted lines upsA-rif and full line upsBsh. The diamond marks the putative termination site of upsA characterised by a stretch of TA repeats. Sizes of genes are not in scale.

Figure 2

Distance tree of 3D7 var gene 500 bp 3' region generated using the p-distance/NJ method. The four dense clusters A through D were supported by both bootstrapping and maximum parsimony (MP) tree (not shown). The relationship of the remaining sequences could not be verified by the MP tree making method. Numbers at the nodes represent bootstrap proportions (BP) on 1000 replicates. The scale bar represents the proportion of different nucleotide compared. PlasmoDB accession numbers are shown. Genes with assigned cluster are collected in figure 7.

Figure 3

Distance tree of DBLαCIDR1 domains of 3D7 PfEMP1 and pseudogene PFE1640w generated using the p-distance/NJ method. The clusters A through E were supported by both bootstrapping and maximum parsimony tree (not shown). Numbers at the nodes represent bootstrap proportions (BP) on 1000 replicates. The scale bar represents the proportion of different amino acids compared. PlasmoDB accession numbers are shown. Genes with assigned cluster are collected in figure 7.

Figure 4

Distance tree of ATS domains of 3D7 PfEMP1 generated using the p-distance/NJ method. Though A, B and C clusters were vaguely supported by the bootstrapping, they were supported by maximum parsimony tree making. Numbers at the nodes represent bootstrap proportions (BP) on 1000 replicates. The scale bar represents the proportion of different amino acids compared. PlasmoDB accession numbers are shown. Genes with assigned cluster are collected in figure 7.

Figure 5

Distance tree of all 3D7 DBLx and randomly chosen 50 DBL domains representing all DBL subtypes generated using the p-distance/NJ method. All clusters were supported by their bootstrap proportions and by maximum parsimony tree making method (data not shown) Numbers at the nodes represent bootstrap proportions (BP) on 1000 replicates. The scale bar represents the proportion of different amino acids compared. PlasmoDB accession numbers are shown.

Figure 6

Distance tree of 3D7 RIFINs generated using the p-distance/NJ method. MAL6P1.251 was left out as it aligned closer to STEVORs in preliminary alignments (data not shown). Red dots mark RIFINs flanked by upsA-rif. Though, the tree topology could not be confirmed by bootstrapping, the clusters containing the marked RIFINs were verified by the maximum parsimony tree (not shown). The scale bar represents the proportion of different amino acids compared. No accession numbers are given to simplify the graphics.

Figure 7

A) Schematic presentation of all 3D7 var gene sequence analyses. Gene names, chromosomal location, transcriptional direction and domain structure are shown along with the cluster to which each gene was assigned by the sequences analyses. Sequences that could not be assigned to any cluster were named X. Three major var gene groups (group A-C), two intermediate groups group B/A and group B/C and two unique genes representing var1 and var2 var gene families were defined (framed). B) Sequence analyses of var genes from other P. falciparum strains than 3D7. Protein accession numbers, originating strain, domain structure and the closest related 3D7 var 5' sequence are shown along with sequence group allocations as defined in 3D7. *) The genes were assigned to group A, as their DBL1α sequences clustered together with other group A sequences in analysis of DBLα sequences. **) Pseudogene, belongs to the var1 family ^) Upstream sequences with atypically low similarity to upsB or upsC sequences.