GDV1 induces sexual commitment of malaria parasites by antagonizing HP1-dependent gene silencing

See allHide authors and affiliations

Science  16 Mar 2018:
Vol. 359, Issue 6381, pp. 1259-1263
DOI: 10.1126/science.aan6042

Sexual development in Plasmodium

Malaria-causing parasites (Plasmodium) have complex life histories in the tissues of humans. For the most part, the parasites focus their efforts on replication within the human host cells. However, occasionally, some replicating cells release gametes into the bloodstream, which are picked up by biting mosquitoes. Filarsky et al. discovered that the Plasmodium parasite keeps the production of gametes under tight epigenetic control using heterochromatin protein 1 (HP1). Plasmodium gametocytogenesis is initiated when HP1 is evicted from upstream of gamete-specific genes by gametocyte development 1 (GDV1) protein. GDV1 is in turn regulated by its antisense RNA. What triggers GDV1 expression remains unclear. Elucidating this pathway could provide a target for interrupting malaria transmission.

Science, this issue p. 1259


Malaria is caused by Plasmodium parasites that proliferate in the bloodstream. During each replication cycle, some parasites differentiate into gametocytes, the only forms able to infect the mosquito vector and transmit malaria. Sexual commitment is triggered by activation of AP2-G, the master transcriptional regulator of gametocytogenesis. Heterochromatin protein 1 (HP1)–dependent silencing of ap2-g prevents sexual conversion in proliferating parasites. In this study, we identified Plasmodium falciparum gametocyte development 1 (GDV1) as an upstream activator of sexual commitment. We found that GDV1 targeted heterochromatin and triggered HP1 eviction, thus derepressing ap2-g. Expression of GDV1 was responsive to environmental triggers of sexual conversion and controlled via a gdv1 antisense RNA. Hence, GDV1 appears to act as an effector protein that induces sexual differentiation by antagonizing HP1-dependent gene silencing.

Heterochromatin protein 1 (HP1) is a conserved regulator of heterochromatin formation, heritable gene silencing, and variegated gene expression (1). In Plasmodium falciparum, HP1-dependent clonally variant expression allows parasites to adapt rapidly to environmental challenges encountered during infection (24). For example, immune evasion via antigenic variation of PfEMP1 is the hallmark of Plasmodium survival. Other processes, such as expression of red blood cell (RBC) invasion ligands or nutrient transporters, are similarly regulated in this parasite (4). Most clonally variant genes cluster in subtelomeric domains, but some also occur in internal heterochromatic regions of chromosomes. In addition, HP1 forms microdomains at some euchromatic genes (2). One of these encodes the transcription factor AP2-G, which is required for sexual conversion and differentiation (2, 57). HP1-dependent regulation of ap2-g controls the rate at which parasites commit to sexual differentiation (7).

To explore the mechanisms regulating HP1 occupancy in P. falciparum, we identified HP1-interacting proteins by liquid chromatography–tandem mass spectrometry (LC-MS/MS) analysis of native HP1 complexes that were purified by coimmunoprecipitation (co-IP) from parasites expressing green fluorescent protein (GFP)–tagged HP1 (7) (Fig. 1A and table S3). We consistently observed GDV1 among the potential HP1 interaction partners (table S1). GDV1 is a nuclear protein implicated in sexual commitment and early gametocytogenesis, but its exact function remains unknown (8). We therefore created a parasite line for the conditional expression of fluorescently labeled ectopic GDV1 (designated GDV1-GFP-DD) (Fig. 1B). Proteins tagged with the destabilization domain (DD) of the immunophilin protein-folding chaperone FK506 binding protein (FKBP) are proteolytically degraded unless cells are cultured in the presence of Shield-1, a small-molecule ligand stabilizer (9, 10). Thus, GDV1-GFP-DD is barely detectable in parasites cultured in the absence of Shield-1 (designated 3D7/GDV1-GFP-DDOFF parasites), but its expression is markedly induced in parasites grown in the presence of Shield-1 (3D7/GDV1-GFP-DDON parasites) (Fig. 1, B and C). In agreement with the co-IP results, GDV1-GFP-DD colocalized with HP1 at the nuclear periphery (Fig. 1C and fig. S1). Furthermore, we found that recombinant HP1 and GDV1 formed a complex (fig. S1) and that HP1 copurified with GDV1-GFP-DD in reverse co-IPs (Fig. 1D and tables S2 and S4). The chromodomain-helicase-DNA-binding protein 1 (CHD1) and a protein of unknown function (PF3D7_1451200) also consistently copurified with both HP1 and GDV1-GFP-DD (tables S1 and S2). Given that CHD1 plays important roles in cell fate decision and heterochromatin remodeling in other organisms (11, 12) and that GDV1 is implicated in gametocytogenesis (8), this putative regulatory complex may function in activating sexual commitment.

Fig. 1 GDV1 interacts with HP1.

(A) Endogenous hp1 locus in 3D7/HP1-GFP parasites and Western blots with antibodies to HP1 (anti-HP1 [α-HP1]) of the anti-HP1-GFP co-IP and negative (neg.) control samples. Results are representative of three biological replicates. α-GFP, antibodies to GFP (anti-GFP). (B) GDV1-GFP-DD expression plasmid and anti-GFP Western blots of 3D7/GDV1-GFP-DDOFF and 3D7/GDV1-GFP-DDON parasites. Anti-HP1 served as a loading control. (C) GDV1-GFP-DD/HP1 colocalization IFAs with 3D7/GDV1-GFP-DDOFF and 3D7/GDV1-GFP-DDON trophozoites (24 to 32 hpi). DAPI, 4′,6-diamidino-2-phenylindole; DIC, differential interference contrast; Shld, Shield-1. Scale bars, 2.5 μm (0.5 μm for the magnified views in the rightmost images). Results are representative of three biological replicates. (D) Anti-GFP and anti-HP1 Western blots of the anti-GDV1-GFP-DD co-IP and negative control samples. Results are representative of three biological replicates.

Malaria parasites proliferate by iterative rounds of intraerythrocytic replication through schizogony, merozoite release, and RBC reinvasion. The decision to enter gametocytogenesis is made in the cell cycle before sexual differentiation; sexually committed schizonts release merozoites that invade RBCs and differentiate all into either female or male gametocytes (13, 14) (Fig. 2A). To test whether GDV1 triggers sexual commitment, we divided samples of 3D7/GDV1-GFP-DDOFF parasites and cultured them in the absence or presence of Shield-1. After reinvasion, stage I gametocytes were quantified by immunofluorescence assays (IFAs) with antibodies to the gametocyte marker Pfs16 (15) (anti-Pfs16). The 3D7/GDV1-GFP-DDON population displayed a sexual conversion rate of 57.2% (±10.0% SD), compared with 11.0% (±2.4% SD) for 3D7/GDV1-GFP-DDOFF parasites, and these gametocytes differentiated normally into both male and female gametocytes and showed a typical female-biased sex ratio (Fig. 2B and fig. S2). Moreover, Shield-1 titration revealed a positive correlation between ectopic GDV1-GFP-DD expression levels and sexual conversion rates (Fig. 2C). To test whether endogenous GDV1 expression levels similarly correlate with gametocyte conversion, we used CRISPR-Cas9–based gene editing to append a triple hemagglutinin (3×HA) tag to the N terminus of GDV1 (yielding 3D7/3×HA-GDV1 parasites) (fig. S3). Endogenous 3×HA-GDV1 colocalized with HP1 as expected (Fig. 2D and fig. S3) but was expressed only in some parasites. We next quantified 3×HA-GDV1 expression under conditions that either suppress or favor sexual conversion. To this end, we made use of the recent discovery of choline as an inhibitor of sexual commitment (16). 3D7/3×HA-GDV1 parasites cultured in the presence or absence of 2 mM choline displayed sexual commitment rates of 1.8% (±0.3% SD) or 30.9% (±3.8% SD), respectively (Fig. 2E). Parasites cultured in the absence of choline showed markedly increased 3×HA-GDV1 expression levels (Fig. 2F). This increase was accounted for by a higher proportion of 3×HA-GDV1–positive cells [48.6% (±3.4% SD) in the absence of choline compared with 16.4% (±1.8% SD) in the presence of choline] (Fig. 2F) and higher 3×HA-GDV1 expression levels in individual 3×HA-GDV1–positive parasites (fig. S3). Together, these results show that GDV1 activates sexual conversion in a dose-dependent manner and that endogenous GDV1 expression can be induced by environmental signals triggering sexual commitment.

Fig. 2 GDV1 induces sexual commitment and differentiation.

(A) Schematic illustrating the iterative cycles of schizogony and RBC reinvasion (top) or sexual commitment, RBC reinvasion, and gametocyte differentiation (bottom). ER and LR, early and late ring stages; T, trophozoites; ES and LS, early and late schizonts; gen, generation; *, time point of anti-Pfs16 (α-Pfs16) IFAs. (B) Top panels show results from anti-Pfs16 IFAs identifying stage I gametocytes. Quantification of Pfs16-positive parasites is shown at the right [results are the means for three biological replicates (200 infected RBCs counted per sample); error bars indicate SD]. Bottom panels show Giemsa-stained blood smears revealing stage V gametocytes. Scale bars, 5 μm. (C) Western blot showing GDV1-GFP-DD expression in the presence of increasing Shield-1 concentrations. Antibodies to glyceraldehyde-3-phosphate dehydrogenase (α-GAPDH) and HP1 served as loading controls. Percentages of Pfs16-positive parasites are shown at the bottom (400 infected RBCs were counted per sample). (D) Endogenous gdv1 locus in 3D7/3×HA-GDV1 parasites and 3×HA-GDV1/HP1 colocalization IFAs in trophozoites (24 to 32 hpi). Scale bar, 2.5 μm. (E) Sexual conversion rates in 3D7/3×HA-GDV1 parasites cultured in the presence or absence of choline [results are the means for three biological replicates (>190 infected RBCs counted per sample); error bars indicate SD]. (F) Western blot showing 3×HA-GDV1 expression levels in 3D7/3×HA-GDV1 parasites cultured in the presence or absence of choline. The arrowhead indicates the intact full-length protein. Anti-HP1 served as a loading control. Percentages of 3×HA-GDV1-positive (3×HA-GDV1-pos.) parasites in the presence or absence of choline are shown on the right [results are the means for three biological replicates (>100 infected RBCs counted per sample); error bars indicate SD]. (G and H) Comparison of mean expression levels for all genes in 3D7/GDV1-GFP-DDON parasites versus 3D7/GDV1-GFP-DDOFF parasites (G) and F12/GDV1-GFP-DDON parasites versus F12/GDV1-GFP-DDOFF parasites (H). Significantly deregulated genes are indicated by circles [mean fold change (FC) cutoff, >1.5; q-value (false discovery rate) cutoff, <0.15]. Known early gametocyte markers (7, 8, 17) are labeled orange. Line graphs show fold changes in expression across seven consecutive time points. pfs16, pfg27, and gexp02, early gametocyte markers (15, 22, 32); pk4 (PF3D7_0628200) and glu-tRNA-l. (PF3D7_1331700), control genes (7).

We next performed comparative transcriptome analyses using two-color microarrays. 3D7/GDV1-GFP-DDOFF ring-stage parasites were split and cultured separately in the absence or presence of Shield-1, and total RNA was harvested at seven paired time points spanning the remaining 24 hours of generation 1 [24 to 32 hours postinvasion (hpi), 32 to 40 hpi, and 40 to 48 hpi] and the first 40 hours after reinvasion in generation 2 (8 to 16 hpi, 16 to 24 hpi, 24 to 32 hpi, and 32 to 40 hpi) (Fig. 2A). As expected, GDV1-GFP-DD expression triggered a transcriptional response characteristic of sexual commitment and early differentiation. This effect was evident from the induction of ap2-g in generation 1, followed by the activation of early gametocyte markers (5, 7, 8, 17) after reinvasion (Fig. 2G, fig. S4, and table S5). In the parasite line F12, a 3D7-derived gametocyte-deficient clone carrying a loss-of-function mutation in ap2-g (5, 18), GDV1-GFP-DD expression still activated ap2-g but failed to launch a sexual differentiation response (Fig. 2H and table S6). Next to ap2-g, only eight other genes, all of which are marked by HP1, were significantly induced in F12/GDV1-GFP-DDON parasites. This set included dblmsp2, which was also induced in 3D7/GDV1-GFP-DDON parasites (Fig. 2, G and H). Given that DBLMSP2 is a merozoite surface antigen expressed only in a small subpopulation of schizonts (19, 20), the GDV1-dependent activation of the dblmsp2 locus suggests that DBLMSP2 may be expressed specifically in sexually committed schizonts. In summary, these findings show that GDV1 is an upstream activator of sexual commitment and likely triggers this process by antagonizing HP1-dependent silencing of ap2-g.

To test whether GDV1 associates with heterochromatin in vivo, we conducted comparative chromatin immunoprecipitation–sequencing (ChIP-seq) experiments. 3D7/GDV1-GFP-DDOFF parasites were split at 28 to 34 hpi and cultured in parallel in the absence or presence of Shield-1, and paired chromatin samples were harvested 2 (30 to 36 hpi), 6 (34 to 40 hpi), and 10 (38 to 44 hpi) hours after Shield-1 addition. We found that (i) GDV1-GFP-DD associates specifically with heterochromatin throughout the genome (Fig. 3A, fig. S5, and table S7), (ii) GDV1-GFP-DD occupancy was markedly higher in 3D7/GDV1-GFP-DDON parasites than in 3D7/GDV1-GFP-DDOFF parasites (fig. S5 and table S7), and (iii) GDV1-GFP-DD occupancy was highly correlated with that of HP1 (Fig. 3B). Moreover, GDV1-GFP-DD occupancy peaked 6 hours postinduction and decreased substantially thereafter (Fig. 3A, fig. S5, and table S7). This drop in GDV1-GFP-DD signal coincided with reduced HP1 occupancy over heterochromatic genes in 3D7/GDV1-GFP-DDON parasites compared with that in 3D7/GDV1-GFP-DDOFF parasites (Fig. 3, A and C, and table S7). Although the vast majority of heterochromatic loci, in particular those displaying high HP1 occupancy, such as the var genes encoding PfEMP1, displayed only slightly decreased HP1 levels, some genes exhibited as much as 40% reduction in HP1 occupancy (Fig. 3C and table S7). This group of genes included ap2-g and most known HP1-associated early gametocyte markers, including geco (21), pfgexp17 (22), and pfg14_748 (8, 17) (Fig. 3, C and D, and table S7). These data are consistent with the microarray results, where GDV1-GFP-DD expression activated ap2-g and early gametocyte genes but had no effect on the expression of the bulk of heterochromatic loci, including var genes (Fig. 2, G and H). Given the 50 to 60% sexual conversion rate observed for 3D7/GDV1-GFP-DDON parasites (see above), a 30 to 40% reduction in HP1 occupancy indicates that HP1 may be depleted at these loci specifically in sexually committed parasites, but single-cell approaches are required to confirm this hypothesis. Overall, we suggest that GDV1 destabilizes heterochromatin and thus allows specific transcription factors to activate expression of ap2-g and other gametocyte-specific heterochromatic genes, and this process may play an important role in the positive autoregulatory feedback loop proposed to reinforce AP2-G expression in committed parasites (5, 6, 23). How GDV1 achieves specificity in unlocking specific HP1-associated genes despite binding heterochromatin genome-wide is a challenging question to be addressed in the future.

Fig. 3 GDV1 associates with heterochromatin throughout the genome and triggers HP1 removal at ap2-g.

(A) Plots of the ratio of the HP1 level to the input for 3D7/GDV1-GFP-DDOFF schizonts [38 to 44 hpi; time point 3 (TP3)] (gray), ChIP-seq subtraction tracks displaying relative enrichment of GDV1-GFP-DD in 3D7/GDV1-GFP-DDON schizonts at 2 (30 to 36 hpi; TP1), 6 (34 to 40 hpi; TP2), and 10 (38 to 44 hpi; TP3) hours after Shield-1 addition (blue), and the relative depletion of HP1 in 3D7/GDV1-GFP-DDON parasites at TP3 (green). chr12, chromosome 12; bp, base pairs. (B) Correlation between GDV1-GFP-DD enrichment in 3D7/GDV1-GFP-DDON schizonts (34 to 40 hpi; TP2) and HP1 enrichment in 3D7/GDV1-GFP-DDOFF schizonts at each coding region. r, Pearson correlation coefficient. (C) Fold change in HP1 enrichment upon GDV1-GFP-DD overexpression in relation to HP1 enrichment in 3D7/GDV1-GFP-DDOFF schizonts for each heterochromatic gene. (D) Zoom-in view of the enrichment and subtraction tracks at the ap2-g locus. Numbers above the graph indicate the position in base pairs.

Because GDV1 activates sexual commitment, the question arises of how parasites limit GDV1 expression to prevent sexual conversion in asexual schizonts. A recent study identified a multiexon-long noncoding gdv1 antisense RNA (asRNA) that initiates downstream of the gdv1 locus and overlaps with the ATG start codon of gdv1 (24), which is a hallmark feature of regulatory asRNAs (25). To investigate whether the gdv1 asRNA participates in regulating sexual commitment, we created a gdv1 asRNA loss-of-function [asRNA knockout (asKO)] mutation in F12 parasites (yielding F12/gdv1-asKO parasites) (Fig. 4A and fig. S6). Strand-specific RNA sequencing (RNA-seq) analysis identified a small set of genes that were consistently differentially expressed between F12/gdv1-asKO and F12 wild-type parasites (17 genes were up-regulated and 23 genes were down-regulated in the mutants) (Fig. 4B and table S8). Similar to F12 parasites expressing ectopic GDV1-GFP-DD (Fig. 2H), F12/gdv1-asKO parasites showed marked induction of ap2-g, dblmsp2, and two early gametocyte genes (pfg14_748 and PF3D7_1477400) (8, 17), and all except one up-regulated gene are HP1-associated genes (Fig. 4B, fig. S6, and table S8). gdv1 sense transcripts were slightly increased in the F12/gdv1-asKO population, and gdv1 antisense transcripts were undetectable, as expected (Fig. 4, B and C; fig. S6; and table S8). These results indicated that the gdv1 asRNA acts as a negative regulator of GDV1 expression. To confirm this hypothesis, we tagged endogenous GDV1 with HA in these parasites (yielding an F12/3×HA-GDV1/gdv1-asKO population) and observed that indeed almost all parasites (96.7% ± 2.5% SD) expressed 3×HA-GDV1 (Fig. 4D and fig. S7). Lastly, we showed that deletion of the gdv1 asRNA locus in a conditional AP2-G mutant resulted in markedly increased production of gametocytes (fig. S8 and supplementary text). Together, these findings demonstrate a central role for the gdv1 asRNA in regulating GDV1-dependent activation of sexual commitment. We anticipate that this mechanism likely involves inhibiting GDV1 expression by interference with gdv1 mRNA transcription, stability, or translation, similar to asRNA-mediated gene regulation in other organisms (26).

Fig. 4 A gdv1 asRNA antagonizes GDV1-dependent sexual commitment.

(A) gdv1 locus in F12 wild-type (wt) and F12/gdv1-asKO parasites. The gdv1 sense transcript (blue), five-exon gdv1 asRNA (24) (red), and human dhfr (hdhfr) resistance marker (brown) are highlighted. chr 9, chromosome 9. (B) Comparison of gene expression levels in F12 wild-type and F12/gdv1-asKO early and late schizonts. Genes deregulated >5-fold at both time points are indicated by circles. RPKM, reads per kilobase million. (C) University of California–Santa Cruz Genome Browser screenshots of RNA-seq coverage plots over the gdv1 locus in F12 wild-type and F12/gdv1-asKO early and late schizonts. The human dhfr resistance cassette that is present downstream of the gdv1 locus in F12/gdv1-asKO parasites and absent in the 3D7 reference genome is indicated by a brown rectangle. Numbers along the top indicate the position in base pairs. RPM, reads per million. (D) Endogenous gdv1 locus in F12/3×HA-GDV1/gdv1-asKO parasites and results from overview IFAs with antibodies to HA (α-HA) in early schizonts (32 to 40 hpi). Scale bar, 5 μm. The percentage of 3×HA-GDV1-positive parasites is shown at the right [results are the mean for three biological replicates (100 infected RBCs counted per sample); error bars indicate SD].

We identified GDV1-mediated heterochromatin destabilization as an epigenetic control strategy regulating sexual cell fate decision in P. falciparum. Our discovery of the gdv1 asRNA as a negative regulator of sexual commitment is reminiscent of long noncoding RNA–mediated control of gametogenesis in yeasts (27, 28). In Saccharomyces cerevisiae, nutritional stress triggers gametogenesis by activating the transcriptional regulator inducer of meiosis 1 (IME1) (28). A long noncoding RNA in the ime1 promoter and antisense transcription of ime4 are key factors in preventing IME1 expression under noninducing conditions (29, 30). These parallels raise the exciting possibility that evolutionarily divergent unicellular eukaryotes employ conceptually similar regulatory logic to control entry into the sexual phases of their life cycles. All Plasmodium species infecting humans possess a GDV1 ortholog, suggesting that GDV1-based regulation of sexual commitment is conserved in all human-infective malaria parasites. In conclusion, our study contributes to understanding of the molecular pathway underlying the formation of malaria transmission stages and provides opportunities for the development of intervention strategies targeting transmission of human malaria.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S8

Tables S1 to S9

References (3358)

References and Notes

Acknowledgments: We are grateful to M. van de Vegte-Bolmer and R. Sauerwein for determining gametocyte sex ratios and providing anti-Pfs16, to D. Richard for providing the pL6-3HA_glmS-246 plasmid, and to T. Haefliger for technical assistance. Funding: This work was supported by the Swiss National Science Foundation (grant numbers 31003A_143916, 31003A_163258, and BSCGI0_157729), the Fondation Pasteur Suisse, and the Netherlands Organization for Scientific Research (NWO-Vidi 864.11.007). S.A.F. is supported by a Ph.D. fellowship from the European Community’s Seventh Framework Program (grant no. FP7/2007–2013) under grant agreement no. 242095 and no. ParaMet 290080. Author contributions: M.F. designed and performed experiments, analyzed data, prepared illustrations, and wrote the paper. S.A.F. performed and analyzed ChIP-seq and RNA-seq experiments. I.N. designed and cloned CRISPR-Cas9 mother plasmids and performed experiments involving recombinant proteins. N.M.B.B. performed experiments related to the 3D7/3xHA-GDV1 and F12/3xHA-GDV1/gdv1-asKO parasites. E. Carrington performed and analyzed quantitative reverse transcription polymerase chain reaction experiments. E. Carrió performed experiments involving 3D7/AP2-G-GFP-DDglmS parasites. S.M. performed LC-MS/MS experiments. P.J. provided conceptual advice. P.J., R.B., and T.S.V. provided resources. R.B. designed, supervised, and analyzed experiments. T.S.V. conceived the study; designed, supervised, and analyzed experiments; and wrote the paper. All authors contributed to editing of the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data and code to understand and assess the conclusions of this research are available in the main text and supplementary materials and via the Gene Expression Omnibus ( (31) under accession numbers GSE95549 (microarray data) and GSE94901 (ChIP-seq and RNA-seq data).

Stay Connected to Science

Navigate This Article