Successful Profiling of Plasmodium falciparum var Gene Expression in Clinical Samples via a Custom Capture Array

Abstract

var genes encode Plasmodium falciparum erythrocyte membrane protein-1 (PfEMP1) antigens. These highly diverse antigens are displayed on the surface of infected erythrocytes and play a critical role in immune evasion and sequestration of infected erythrocytes. Studies of var expression using non-leukocyte-depleted blood are challenging because of the predominance of host genetic material and lack of conserved var segments. Our goal was to enrich for parasite RNA, allowing de novo assembly of var genes and detection of expressed novel variants. We used two overall approaches: (i) enriching for total mRNA in the sequencing library preparations and (ii) enriching for parasite RNA with a custom capture array based on Roche's SeqCap EZ enrichment system. The capture array was designed with probes based on the whole 3D7 reference genome and an additional >4,000 full-length var gene sequences from other P. falciparum strains. We tested each method on RNA samples from Malian children with severe or uncomplicated malaria infections. All reads mapping to the human genome were removed, the remaining reads were assembled de novo into transcripts, and from these, var-like transcripts were identified and annotated. The capture array produced the longest maximum length and largest numbers of var gene transcripts in each sample, particularly in samples with low parasitemia. Identifying the most-expressed var gene sequences in whole-blood clinical samples without the need for extensive processing or generating sample-specific reference genome data is critical for understanding the role of PfEMP1s in malaria pathogenesis. IMPORTANCE Malaria parasites display antigens on the surface of infected red blood cells in the human host that facilitate attachment to blood vessels, contributing to the severity of infection. These antigens are highly variable, allowing the parasite to evade the immune system. Identifying these expressed antigens is critical to understanding the development of severe malarial disease. However, clinical samples contain limited amounts of parasite genetic material, a challenge for sequencing efforts further compounded by the extreme diversity of the parasite surface antigens. We present a method that enriches for these antigen sequences in clinical samples using a custom capture array, requiring minimal processing in the field. While our results are focused on the malaria parasite Plasmodium falciparum, this approach has broad applicability to other highly diverse antigens from other parasites and pathogens such as those that cause giardiasis and leishmaniasis.

Keywords: P. falciparum; PfEMP1; Plasmodium falciparum; RNA-Seq; malaria; var gene.

FIG 1

Overview of pipeline for enrichment of parasite RNA, sequencing, and assembly of transcripts. P. falciparum mRNA was enriched in all 12 samples using both globin and rRNA depletion as well as capture. Ten samples were subjected to all three methods, including globin and rRNA depletion followed by poly(A) selection. For the Depletion and Depletion + poly(A) method, enrichment took place during library preparation, and for Capture, libraries were prepared, and then Capture was applied to enrich for parasite cDNA. Reads were mapped to a concatenated reference of human and P. falciparum. We used reads that mapped to P. falciparum and unmapped reads to de novo assemble transcripts, open reading frames for the transcripts were subjected to a protein blast search to identify var transcripts, and then the protein sequences were annotated with the VarDom online database.

FIG 2

Reads mapping to P. falciparum, known var genes, and number and length of assembled var gene transcripts. Results were quantified for Capture (orange; first row), Depletion (blue; second row), and Depletion + poly(A) (green; third row). Reads mapping to P. falciparum were quantified for the three methods (A, D, and G); percentages of the random selection of 20 million reads mapping to known var genes were quantified for the three methods (B, E, and H); and the number and length of unique transcripts were quantified for the three methods (C, F, and I). Capture retained the greatest percentages of P. falciparum and var reads, and more unique var transcripts were assembled from the Capture-prepared samples. Samples are arranged from least (left) to greatest parasitemia. CM, cerebral malaria; UM, uncomplicated malaria; UMC, uncomplicated malaria control; SMA, severe malarial anemia.

FIG 3

Comparison of expression and assembly results from Capture and Depletion + poly(A) mapping to genomic sequences. Each scatterplot compares transcripts per million (TPMs) in Depletion + poly(A) on the x axis to Capture on the y axis by mapping the reads from each library to the genomic reference. The bar graph in the inset of each panel shows the length of the most-expressed transcripts de novo assembled from the RNA-Seq data [Capture, darker shade; Depletion + poly(A), lighter shade] and the length of the corresponding genomic sequence (dotted black line). For each uncomplicated malaria sample, the same var genes were identified in the two libraries, but Capture yielded greater transcripts per million (scatterplots). Predominantly expressed var genes de novo assembled from RNA-Seq data were similar in length (bar graphs) to the genome reference var genes (dotted black line), with de novo-assembled sequences from capture typically longer than those from Depletion + poly(A). Panel A (UM 8, UM 3, and UM 6) includes monoclonal (one infecting parasite clone) samples, and panel B includes polyclonal (more than one infecting parasite clone) samples (UM 11, UM 9, and UM 1). Samples are arranged from least to greatest parasitemia within each row, from left to right. The color of the dots in the scatterplot (red, blue, or green) corresponds to the same transcript (T1, T2, or T3) shown assembled in the inset bar graph. UM, uncomplicated malaria; T1, transcript 1; T2, transcript 2; T3, transcript 3.