Ectopic recombination of a malaria var gene during mitosis associated with an altered var switch rate

Abstract

The Plasmodium falciparum var multigene family encodes P. falciparum erythrocyte membrane protein 1, which is responsible for the pathogenic traits of antigenic variation and adhesion of infected erythrocytes to host receptors during malaria infection. Clonal antigenic variation of P. falciparum erythrocyte membrane protein 1 is controlled by the switching between exclusively transcribed var genes. The tremendous diversity of the var gene repertoire both within and between parasite strains is critical for the parasite's strategy of immune evasion. We show that ectopic recombination between var genes occurs during mitosis, providing P. falciparum with opportunities to diversify its var repertoire, even during the course of a single infection. We show that the regulation of the recombined var gene has been disrupted, resulting in its persistent activation although the regulation of most other var genes is unaffected. The var promoter and intron of the recombined var gene are not responsible for its atypically persistent activity, and we conclude that altered subtelomeric cis sequence is the most likely cause of the persistent activity of the recombined var gene.

Supplementary Fig. 1

(a) PCR of the four rif genes present on chromosome 12 of E8B (E) and lost following recombination in CS2 (C); a PCR of It4_var8 exon 1 was included as a positive control. (b) PCR across the var2csa intron using primers specific for the E8B wild-type sequence or the CS2 mutant sequence: CS2 (C), E8B (E), CS2 ICAM-1-selected subclones CS2isB10 (B10), CS2isC9 (C9), CS2isD8 (D8), negative control (-). Subclone C8 could not be recovered from cryopreserves for this analysis.

Fig. 1

(a) Diagrams of the left-hand telomeric fragment of chromosome 7 and the var2csa locus on the left-hand telomeric fragment of chromosome 12 in wild-type E8B clonal parasites and in their clonal progeny CS2 following ectopic recombination between chromosome 7 and chromosome 12. The filled bar indicates wild-type chromosome 12, and the unfilled bar indicates the sequence from chromosome 7 introduced following recombination. Positions are indicated for TARE 1–5, TARE 6 (rep20), the second telomere-associated repeat sequence from subtelomeric block 2 (SB2-2), several members of the rif and var multigene families, and the restriction sites AccI (A), EcoRI (E), HindIII (H), PvuII (P), XbaI (X), SalI (S), and PstI (Ps). The positions of probes used for Southern and Northern blots and regions that were directly sequenced are indicated beneath the diagrams. (b) Southern blots of CS2 and E8B gDNA digested with the enzymes indicated above the figure and separated by conventional gels (AccI, HindIII, and XbaI) and PFGE (XbaI, PvuII, and EcoRI) were probed with sequence from the 3′ end of var2csa exon 1. (c) Southern blots of PFGE-separated chromosomes probed with sequence from the 3′ ends of var2csa and It4_var44. (d) Southern blots of CS2 and E8B gDNA digested with the enzymes indicated above the figure were probed with sequence from the 3′ end of It4_var8 exon 1.

Fig. 2

(a) The numbers of CS2 and E8B parasites that bound per square millimeter to immobilised CSA and CD36 at different time points following selection on CSA. Means of three replicates for each of two assays conducted on separate days; error bars are standard error of the mean. (b) Northern blots of RNA from E8B and CS2 parasites collected at different time points following selection on CSA and probed with sequence from the centre of the first exon of var2csa (Fig. 1a), from the 3′ end of IT4_var44, and with a conserved var exon 2 probe.

Fig. 3

Comparative genome hybridisation log2ratios of (a) CS2 compared to E8B, (b) CS2 compared to P1B5, and (c) E8B compared to P1B5. The genes (vertical black bars) are plotted as physical order on the x-axis whilst the log ratios are on the y-axis. The zero line shows the distribution of the genes if no variation was to be found between the two strains compared. The horizontal lines at + 1 and − 1 are limits indicating amplification (log2ratio > 1) and deletion (log2ratio < − 1), respectively. The grey vertical bars delimit the chromosomes. (d) The skeleton binding protein gene sbp and genes identified as possibly deleted or duplicated by microarray were quantitated in both CS2 and E8B gDNA using standard curves of serially diluted gDNA. The levels of sbp were used to normalise the levels of the other genes so that the resulting relative values compare the level of each gene in equivalent amounts of CS2 and E8B gDNA.

Fig. 4

(a) CS2 parasites were selected on ICAM-1 and then cloned, and the number of infected erythrocytes of each clone that bound per square millimeter to immobilised CSA, CD36, and ICAM-1 was determined. (b) The numbers of erythrocytes infected with CS2 clones and E8B parasites that bound per square millimeter to immobilised CSA, CD36, and ICAM-1 at different time points following selection on ICAM-1. Means of three replicates for each of two assays conducted on separate days; error bars are standard error of the mean.

Fig. 5

(a) Northern blot of RNA collected from E8B and CS2 parasites at different times after selection for adhesion to ICAM-1 and sequentially probed with sequences from var2csa, the 3′ end of exon 1 of It4_var44, and the conserved var exon 2. (b) Western blot probed with rabbit antiserum to the conserved exon 2 of PfEMP1; samples were CS2, CS2 selected on ICAM-1 and cultured until both var2csa and It4_var44 were transcribed (CI), E8B soon after selection on ICAM-1 when It4_var44 was transcribed (EI44+), and E8B selected on ICAM-1 and grown until transcription of It4_var44 had ceased (EI44-).

Fig. 6

Q-RT-PCR was used to determine the levels of 17 var genes in the cDNAs of E8B and CS2 parasites at different time points during continuous growth following selection for adhesion to CSA and ICAM-1. Quantitation was performed using efficiency correction for each PCR and expressed relative to the levels of each gene in E8B gDNA following normalisation with the skeleton binding protein gene sbp. Linear regression curves are indicated on the plots of E8B selected on CSA and CS2 selected on CSA and ICAM-1.

Fig. 7

(a) Plasmids transfected into 3D7-infected erythrocytes carried the P. falciparum heat shock protein 86 promoter (5′ hsp86) driving transcription of the blasticidin deaminase gene (bsd) followed by the Plasmodium berghei dihydrofolate reductase transcriptional terminator (Pb dhfr 3′) followed by the ItG upsE driving transcription of human dihydrofolate reductase (hdhfr) followed by the P. falciparum histidine-rich protein 2 transcriptional terminator (hrp2 3′) followed by either the E8B or CS2 var2csa intron in plasmids pHBEiE and pHBEiC, respectively. The plasmid pHBEiER is the same as pHBEiE except for the inclusion of the TARE 6 (rep20) upstream of the upsE. (b) Plasmids were maintained in 3D7 parasites by growth in the presence of blasticidin-S, and the levels of hdhfr transcription were determined at different time points following transfection by Q-RT-PCR. Absolute quantitation was performed using standard curves of serially diluted, cloned hdhfr DNA of known concentration. The results were normalised using skeleton binding protein 1 cDNA levels to allow comparison of equivalent quantities of parasite material and then expressed as relative units of hdhfr cDNA per plasmid by dividing by the number of plasmids per genome. Plasmid number was determined by using Q-RT-PCR of transfected parasite gDNA to determine hdhfr normalised with sbp. (c) Northern blots of RNA from transfected parasites and nontransfected 3D7 parasites serially probed with sequence from bsd, hdhfr, and a conserved var exon 2.

Fig. 8

Fold differences in the cDNA levels of the transcribed var repertoire of transfected parasites relative to 3D7 nontransfected parasites calculated from 2^− ΔΔCt values normalised using arginyl tRNA synthetase.