Surface co-expression of two different PfEMP1 antigens on single plasmodium falciparum-infected erythrocytes facilitates binding to ICAM1 and PECAM1

Abstract

The Plasmodium falciparum erythrocyte membrane protein 1 (PfEMP1) antigens play a major role in cytoadhesion of infected erythrocytes (IE), antigenic variation, and immunity to malaria. The current consensus on control of variant surface antigen expression is that only one PfEMP1 encoded by one var gene is expressed per cell at a time. We measured var mRNA transcript levels by real-time Q-PCR, analysed var gene transcripts by single-cell FISH and directly compared these with PfEMP1 antigen surface expression and cytoadhesion in three different antibody-selected P. falciparum 3D7 sub-lines using live confocal microscopy, flow cytometry and in vitro adhesion assays. We found that one selected parasite sub-line simultaneously expressed two different var genes as surface antigens, on single IE. Importantly, and of physiological relevance to adhesion and malaria pathogenesis, this parasite sub-line was found to bind both CD31/PECAM1 and CD54/ICAM1 and to adhere twice as efficiently to human endothelial cells, compared to infected cells having only one PfEMP1 variant on the surface. These new results on PfEMP1 antigen expression indicate that a re-evaluation of the molecular mechanisms involved in P. falciparum adhesion and of the accepted paradigm of absolutely mutually exclusive var gene transcription is required.

Figure 1. Surface expression of PfEMP1 on single 3D7 infected erythrocytes.

(A) erythrocytes infected with a 3D7_PFD1235w, (B) a 3D7_{PF11_0008}, (C) a 3D7_{PFD1235w/PF11_0008}, and (D) a NF54 _VAR2CSA sub-line. Localisation of PfEMP1 by confocal microscopy was done using (A1–D1) rat PFD1235w-DBL4γ antisera and (A2–D2) rabbit PF11_0008-CIDR2β antisera. Double surface staining was done using the following combinations (A3–D3) rat PFD1235w-DBL4γ and rabbit PF11_0008-CIDR2β antisera; (A4–D4) rat PFD1235w-DBL4γ and mouse VAR2CSA-DBL5ε antisera; (A5–D5) rabbit PF11_0008-CIDR2β and mouse VAR2CSA-DBL5ε antisera; (A6–D6) rat PFD1235w-DBL4γ and rabbit PFD1235w-CIDR1α; (A7–D7) rabbit PF11_0008-CIDR2β and rat PF11_0008-DBL4β; (A8–D8) rabbit VAR2CSA-DBL5ε-DBL6ε and mouse VAR2CSA-DBL5ε antisera. Antisera staining of PFD1235w expressed by the 3D7_PFD1235w (A1, A3, A4, A6) and 3D7_{PFD1235w/PF11_0008} (C1, C3, C4, C6) sub-lines was detected using a secondary antibody labelled with Alexa 488 (green). Antisera staining of PF11_0008 expressed by the 3D7_{PF11_0008} (B2, B3, B5, B7) and 3D7_{PFD1235w/PF11_0008} (C2, C3, C5, C7) was detected using a secondary antibody labelled with Alexa 568 (red). Staining of VAR2CSA expressed by NF54_VAR2CSA is red (D4, D8) and green (D5, D8). DAPI staining of DNA in the nuclei is blue. Scale bar 5 µm.

Figure 2. Simultaneous surface expression of PfEMP1 on single 3D7 infected erythrocytes.

(A) erythrocytes infected with a 3D7_PFD1235w, (B) a 3D7_{PF11_0008}, and (C) a 3D7_{PFD1235w/PF11_0008} sub-line. Double surface staining and localisation of PfEMP1 by flow cytometry (dotplots) and confocal microscopy (photo inserts) was done using rabbit sera against CIDR2β of PF11_0008 in combinations with rat antisera against the following PFD1235w domains (A1–C1) DBL1α-CIDR1α, (A2–C2) CIDR1α*, (A3–C3) DBL3β, (A4–C4) DBL4γ, (A5–C5) DBL5δ, and (A6–C6) DBL5δ-CIDR2β. Similarly rat sera against DBL4β of PF11_0008 was used in combinations with rabbit sera against (A7–C7) CIDR1α and (A8–C8) DBL4γ of PFD1235w. For confocal microscopy rat antisera staining of PFD1235w expressed by the 3D7_PFD1235w (A1–A6) and 3D7_{PFD1235w/PF11_0008} (C1–C6) and PF11_0008 expressed by the 3D7_{PF11_0008} (B7–B8) and 3D7_{PFD1235w/PF11_0008} (C7–C8) sub-lines was detected using a secondary antibody labelled with Alexa 488 (green). Rabbit antisera staining of PF11_0008 expressed by the 3D7_{PF11_0008} (B1–B6) and 3D7_{PFD1235w/PF11_0008} (C1–C6) and PFD1235w expressed by the 3D7_PFD1235w (A7–A8) and 3D7_{PFD1235w/PF11_0008} (C7-C8) sub-lines was detected using a secondary antibody labelled with Alexa 568 (red). Flow cytometry plots include both infected and uninfected erythrocytes (see Materials and Methods). Flow cytometry settings and capture parameters for confocal microscopy were identical for all images. DAPI staining of DNA in the nuclei is blue. Scale bar 5 µm.

Figure 3. Simultaneous surface expression of PfEMP1 on 3D7_{PFD1235w/PF11_0008} clone 3 infected erythrocytes.

Double surface staining and localisation of PfEMP1 by flow cytometry (dotplots) and confocal microscopy (photo inserts) was done using (A) buffer, (B) rat and rabbit pre-bleeds, (C) rat sera against PFD1235w-DBL4γ and rabbit sera against PF11_0008-CIDR2β, (D) rabbit sera against PFD1235w-DBL4γ and rat sera against PF11_0008-DBL4β, (E) rat sera against PFD1235w-DBL4γ and rabbit pre-bleed, (F) rat pre-bleed and rabbit sera against PF11_0008-CIDR2β, and (G) rat sera against PFD1235w-DBL5δ-CIDR2β and rabbit sera against PF11_0008-CIDR2β. For confocal microscopy rat antisera staining of PFD1235w (C, E, G) and of PF11_0008 expressed by the 3D7_{PFD1235w/PF11_0008} clone 3 (D) was detected using a secondary antibody labelled with Alexa 488 (green). Similarly rabbit antisera staining of PFD1235w (D) and PF11_0008 (C, F, G) was detected using a secondary antibody labelled with Alexa 568 (red). Flow cytometry plots include both infected and uninfected erythrocytes (see Materials and Methods).Flow cytometry settings and capture parameters for confocal microscopy were identical for all images. DAPI staining of DNA in the nuclei is blue. Scale bar 5 µm.

Figure 4. Individual var gene transcripts relative to the total var transcript copy number in ring-stage IE.

The analysis was done on cultures of the (A1) 3D7_PFD1235w, (B1) 3D7_{PF11_0008}, (C1) 3D7_{PFD1235w/PF11_0008}, and (D1) 3D7_VAR2CSA sub-lines. A similar analysis was done on Dynabeads antibody enriched 3D7_{PFD1235w/PF11_0008} (A2) IE prior to enrichment, (B2) IE enriched using rabbit PFD1235w-DBL4γ antisera, (C2) IE enriched using rabbit PFD1235w-CIDR1α antisera. (D2) IE enriched using rabbit PF11_0008-CIDR2β antisera. In another experiment we used unsorted (A3) 3D7_{PFD1235w/PF11_0008} IE and (B3) FACS sorted 3D7_{PFD1235w/PF11_0008} IE stained using rabbit PF11_0008-CIDR2β and rat PFD1235w-DBL4γ antisera followed by FITC- and Alexa Fluor 610-R-PE conjugated secondary antibodies as described in Materials and Methods. The 3D7_{PFD1235w/PF11_0008} sub-line was cloned by limiting dilution and the transcript profile of nine different clones was analysed. The transcript profiles of three representative clones are shown (A4-C4).

Figure 5. Nuclear transcripts of PFD1235w and PF11_0008 in single erythrocytes.

FISH analysis was done on single ring stage parasites in 100 single-cells of (A–C) 3D7_PFD1235w IE, (D–F) 3D7_{PF11_0008} IE and (G–I) 3D7_{PFD1235w/PF11_0008} clone 3 IE using a (B, E, H) PFD1235w (green) and (C, F, I) PF11_0008 specific probe (red). The overlaying of the two colours identified transcripts of PFD1235w in (A) 3D7_PFD1235w IE, PF11_0008 in (D) 3D7_{PF11_0008} IE, and simultaneous transcription of PFD1235w and PF11_0008 in (G) 3D7_{PFD1235w/PF11_0008} clone 3 IE. Ethidium bromide staining (J) and Northern blotting (K and L) verification of the specificity of the probes used for FISH was done on RNA purified from 3D7 sub-lines transcribing PFD1235w (J1, K1, L1), PF11_0008 (J2, K2, L2), co-transcribing PFD1235w and PF11_0008 (J3, K3, L3), and FCR3 transcribing var2CSA (J4, K4, L4). The blots were probed with (K) a PFD1235w and (L) a PF11_0008 specific probe. M: RNA molecular weight marker; < PFD1235w transcript (10.66 kb); ≪ PF11_0008 8.98 kb transcript. R: ribosomal band. Scale bar 5 µm.

Figure 6. Infected erythrocytes bind CD54/ICAM1, CD31/PECAM1 or both receptors.

Binding of (1) 3D7_PFD1235w, (2) 3D7_{PF11_0008}, (3) 3D7_{PFD1235w/PF11_0008}, (4) 3D7_PFD1235w mixed 1∶1 with the 3D7_{PF11_0008} sub-line and (5) non selected knob positive 3D7 (3D7k+) to (A) CHO expressing CD54/ICAM1, (B) CHO expressing CD36, or (C) wildtype CHO cells was done in the absence or presence of anti-CD54/ICAM1 (clone 15.2 and My13) and anti-CD36 antibodies. Similarly binding of the different 3D7 sub-lines to (D) human umbilical vein endothelial cells (HUVEC) was done in the absence or presence of anti-CD54/ICAM1 (clone 15.2 and My13), anti-CD36 or anti-CD31/PECAM1 antibodies. The data represents the mean of at least three independent experiments except for the αICAM1 (My13) experiment in B, C, and D which was only done once. Error bars are ± SD. Buffer: background control with no IE added; no Ab: binding assay done using IE, but without addition of antibodies.