Histone H3K9 acetylation level modulates gene expression and may affect parasite growth in human malaria parasite Plasmodium falciparum

Abstract

Three-dimensional positioning of the nuclear genome plays an important role in the epigenetic regulation of genes. Although nucleographic domain compartmentalization in the regulation of epigenetic state and gene expression is well established in higher organisms, it remains poorly understood in the pathogenic parasite Plasmodium falciparum. In the present study, we report that two histone tail modifications, H3K9Ac and H3K14Ac, are differentially distributed in the parasite nucleus. We find colocalization of active gene promoters such as Tu1 (tubulin-1 expressed in the asexual stages) with H3K9Ac marks at the nuclear periphery. By contrast, asexual stage inactive gene promoters such as Pfg27 (gametocyte marker) and Pfs28 (ookinete marker) occupy H3K9Ac devoid zones at the nuclear periphery. The histone H3K9 is predominantly acetylated by the PCAF/GCN5 class of lysine acetyltransferases, which is well characterized in the parasite. Interestingly, embelin, a specific inhibitor of PCAF/GCN5 family histone acetyltransferase, selectively decreases total H3K9Ac acetylation levels (but not H3K14Ac levels) around the var gene promoters, leading to the downregulation of var gene expression, suggesting interplay among histone acetylation status, as well as subnuclear compartmentalization of different genes and their activation in the parasites. Finally, we found that embelin inhibited parasitic growth at the low micromolar range, raising the possibility of using histone acetyltransferases as a target for antimalarial therapy.

Abbreviations

CSA

chondroitin sulfate A

DAPI

4′,6-diamidino-2-phenylindole

FISH

fluorescence in situ hybridization

HS

hybridization solution

IFA

immunofluorescence assay

KAT

lysine acetyltransferase

SSPE

saline-sodium phosphate-EDTA

Introduction

Plasmodium falciparum, an apicomplexan parasite, causes cerebral malaria and kills approximately 660 000 individuals per year. The nuclear architecture and associated epigenetic state of the given domain maintain any gene either in an active or repressed state depending on the particular cell types or the stage of development [1-4]. Movement of developmentally regulated genes from one to another chromatin domain occurs during the course of differentiation, such as hematopoietic differentiation and astrocyte differentiation [5-7]. Although, in the malarial parasite, the genome is organized into highly evolved nucleosomal structure, the organization in the context of gene expression remains to be determined [8]. Recent studies have highlighted the importance of the nuclear periphery and associated protein complexes for the regulation of telomere associated var gene families in the parasite [9-11]. Antigenic variation of var genes on the surface of infected erythrocyte is a major determinant of the immune evasion leading to the parasite pathogenicity [12-14]. The molecular events leading to the random selection of var genes are not properly understood in P. falciparum. The P. falciparum genome encodes approximately 60 var genes showing mutually exclusive expression during the life cycle. Only one var gene is expressed, whereas 59 other var genes remain silent at one time [15-20]. Epigenetic regulation is one of the major determinants for the mutually exclusive expression of var genes. The major determinants for the active and repressed state of these var genes include modifications of chromatin structure and post-translational modification of the N-terminal histone tails [9, 21-23]. The perinuclear positioning of var gene families represents their repressed state [24]. The var genes still occupy the nuclear periphery upon activation but change their positions to unique compartments for transcription [24]. In yeast, the region beneath the nuclear pore complexes plays key roles in determining the active status of a gene and has to move away from silenced zone to a particular region below the nuclear pore complexes for expression [3]. In P. falciparum, however, positioning of the var gene beneath the nuclear pore complex is not required for gene expression, as it is in yeast [25].

The specific histone modification signatures at different sites in diverse locations (enhancer, promoter, exons, introns and also UTRs) of a gene have been causally related to the active or repressed state of the genes [26]. A highly complicated but poorly understood epigenetic signaling network regulates the activation or repression of domain specific gene(s) [27]. Notably, the histone lysine acetylation is a very dynamic cellular phenomenon that plays crucial role in the regulation of gene expression [28]. On the other hand, histone deacetylases that remove acetyl groups from histone also play a major role in the transcriptional regulation in parasite [29]. The lysine acetyltransferases (KATs) catalyze the acetylation of lysine residues present in the N-terminal tail of histones, resulting in a relaxed chromatin conformation leading to transcriptional activation. KATs can be categorized into five major families, such as GNATs (GCN5 N-acetyltransferases), MYSTs (Tip60, MOZ, Sas2 and Ybf1/Sas3), p300/CBP (CREB binding protein), nuclear hormone related KATs and general transcription factor KATs [30, 31]. In P. falciparum, PfGCN5 (PF08_0034) and PfMYST (PF11_0192) KATs have been characterized in detail [32, 33]. PfGCN5 acetylates histone H3 at the ninth and fourteenth positions in vitro [32]. Unlike PfGCN5, PfMYST acetylates histone H4 at the fifth, eight, twelfth and sixteenth lysine of the N-terminal tail [33].

Different histone modifications occupy distinct nuclear environments, such as H3K9Ac (active) and H3K9me3 (repressed), show perinuclear localization in the parasite [34]. Other chromatin modulators involved in var gene silencing, such as PfSir2, ORC1 [9, 10, 35] or HP1 [36, 37], also show distinct perinuclear positioning. Histone modifications are also recognized by different downstream effector proteins. For example, H3K4me3 modification recognized by the plant homeodomain containing effector protein leads to dramatic changes in methylation or acetylation status near H3K4me3 [38].

In recent years, studies of the different modifications of histone H3 have shown that the dynamic distribution of these modified histones throughout the genome regulates transcription during different developmental stages of P. falciparum [39, 40]. In the present study, we report the distinct perinuclear positioning of H3K9Ac marks that may be involved in the active transcription of associated gene promoters. We further reveal the localization of a developmentally regulated asexual stage-specific promoter, such as tubulin-1 (Tu1), at the nuclear periphery enriched with H3K9Ac marks. However, the perinuclear localization of promoters of sexual stage-specific genes such as Pfg27 (early gametocyte marker) and Pfs28 (ookinete marker) is devoid of H3K9Ac marks during the asexual stage. Furthermore, to correlate the H3K9Ac status with transcriptional activation, we used embelin, a specific GNAT family KAT inhibitor [41] that specifically inhibits PCAF/GCN5 dependent H3K9Ac acetylation in mice. We found that embelin treatment affected H3K9Ac levels but not H3K14Ac levels in the parasites. In addition, embelin decreased H3K9Ac levels around the var gene promoter (upsE) with a concomitant decrease in var2csa gene expression levels in the parasites. Finally, we found that embelin inhibited parasitic growth at the low micromolar range, raising the possibility of using histone acetyltransferases as a target for anti-malarial therapy. These results suggest that an interplay between H3K9Ac levels around the promoter and its subnuclear organization in controlling gene expression in the parasites, which may have far-reaching effects with respect to the design of new drugs for antimalarial therapy.

Results

P. falciparum histone mark H3K9Ac localizes to the nuclear periphery

In P. falciparum, histone acetylation is considered to be involved in the transcription of telomere associated var gene families, as well as other genes responsible for parasite cell cycle progression similar to the regulation in other organisms [23, 42]. Previous studies suggest the differential localization of these modifications in the parasite nuclei [33, 43]. To investigate further the presence of H3K9Ac mark in the parasite, we performed western blot analysis using core histones isolated from parasites (Fig. 1A). Antibodies against H3K9Ac specifically recognized the acetylated protein band (Fig. 1B, second panel, lane b) at the similar position of histone H3 (Fig. 1B, second panel, lane a). The expression of another predominant histone acetylation mark, H3K14Ac in parasite was also detected (Fig. 1C, second panel, lane d) at the similar position of histone H3 (Fig. 1C, second panel, lane c). After confirming the substantial level of H3K9Ac and H3K14Ac in parasites, we performed immunofluorescence assays to investigate the subcellular localization patterns of H3K9Ac and H3K14Ac marks in the parasites. Immunofluorescence assays were performed in ring stage of the parasite development because the ring stage is crucial for mutually exclusive selection and the expression of var genes related to parasite pathogenesis. We detected a horseshoe-shaped pattern as well as a sharp perinuclear foci like pattern closely attached to nuclear periphery in the ring stage parasites for H3K9Ac (Fig. 1D,E). Furthermore, H3K9Ac showed perinuclear foci in all the erythrocytic developmental stages (data not shown). Because H3K9Ac marks are often correlated with the transcriptional activation, the above results suggest the presence of specific sites/zones for active transcription of genes in the parasites. Analysis for H3K9Ac foci in approximately 100 ring stage parasites show an average of three or four perinuclear foci in the majority of the rings (approximately 70%) (data not shown). Interestingly, H3K14Ac was detected all over the nuclei compared to the distinct foci found for H3K9Ac (Fig. 1F,G). Taken together, these data reveal a distinct pattern of epigenetic marks, especially for the H3K9Ac in the parasites.

Figure 1

Expression and subcellular localization patterns of H3K9Ac and H3K14Ac histone marks. (A) Purification of core histones from P. falciparum. Native core histones were acid extracted followed by SDS/PAGE analysis on 15% polyacrylamide gel. Molecular mass markers (kDa) are shown on the left. (B) Acetylation status of histone H3 in parasite. Parasite lysates were separated on 15% SDS/PAGE followed by transfer to the polyvinylidene difluoride membrane (left panel). Lane (a) was used for western blot analysis with anti-H3 polyclonal sera. Lane (b) was used for western blot analysis with anti-H3K9Ac sera. (Right two panels). (C) Lanes (c) and (d) were used for Western blot analysis with antibodies against histone H3 and H3K14Ac, respectively. The left panel shows coomassie stained gel after transfer. (D) Perinuclear localization pattern of acetylated H3K9Ac marks. Immunofluorescence analysis using antibodies specific for H3K9Ac confirms the horseshoe-shaped pattern around the nuclear periphery. H3K9Ac also localizes as sharp three to seven perinuclear foci in majority of the parasites. (E) Schematic representation of H3K9Ac mark in parasite nuclei. (F) Localization pattern of H3K14Ac mark in the parasite nuclei. Unlike H3K9Ac, the H3K14Ac signal shows an even distribution on the DAPI. (G) Schematic diagram of H3K14Ac marks in the parasite nuclei. Scale bar = 1 μm. DAPI stains ring nuclei.

H3K9Ac marks localize to specific active sites at the nuclear periphery, which is different from the repressive sites marked with H3K9me3

To identify the relative position of H3K9Ac, with respect to PfSir2, which deacetylates histone H3 at the ninth and fourteenth lysine positions [44], immunofluorescence experiments were performed. The perinuclear localization of PfSir2 and H3K9Ac was specifically detected but no-colocalization was observed, suggesting the presence of different territories of activation (H3K9Ac) and repression (PfSir2) (Fig. 2A). It will be interesting to investigate the localization of PfGCN5 with H3K9Ac marks in the parasites. Unfortunately, we could not fulfill this objective because of the non-availability of antibodies against PfGCN5.

Figure 2

Differential localization patterns of H3K9Ac, PfSir2 and PfHP1. (A) Differential distribution of PfSir2 and H3K9Ac in P. falciparum. Dual immunofluorescence analysis of PfSir2 (green) and H3K9Ac (red) reveals PfSir2 as sharp foci around nuclei and H3K9Ac as sharp foci closely attached to the nuclear periphery. There is no colocalization of PfSir2 and H3K9Ac in ring stage. Scale bar = 1 μm. DAPI stains nuclei. (B) Differential distribution of H3K9Ac (green) and PfHP1, specific for H3K9me3 repressive histone marks (red). Dual immunofluorescence assay shows no colocalization between these two proteins. Scale bar = 1 μm. DAPI stains nuclei.

Previous work has revealed perinuclear localization of repressive H3K9me3 mark as two or three sharp polar foci. These repressive modifications are recognized by HP1 as specific effector protein [34, 36, 37]. Co-immunofluorescence experiments using antibodies specific for these marks revealed perinuclear positioning of both H3K9Ac and HP1 marks without any colocalization, suggesting differential and distinct distribution of transcriptionally active and repressed sites in the parasite nuclei (Fig. 2B).

Active promoters occupy specific zones at the nuclear periphery enriched with H3K9Ac marks

After visualizing the differential pattern of active (H3K9Ac) and repressive (HP1) marks around the parasite nuclei, we were interested in studying the localization pattern of developmentally regulated stage-specific gene promoters and the possible relationship of these promoters with respect to active H3K9Ac marks. Accordingly, double synchronized ring stage parasites were used for fluorescence in situ hybridization (FISH) analysis. DNA fragments corresponding to specific promoters like tubulin-1 (Tu1) (active in asexual blood stage), Pfg27 promoter (active in gametocytes) and Pfs28 promoter (active in ookinetes) were amplified by PCR and used as probes for the FISH analysis. Interestingly, we found perinuclear punctate foci for both the asexually active Tu1 promoter and the inactive Pfg27 and Pfs28 promoters (Fig. 3A). FISH experiments were repeated several times and FISH signals were mostly found at the nuclear periphery under our experimental conditions (data not shown). These results raise the possibility that both the transcriptionally active and inactive gene promoters may present at the nuclear periphery. However, a comprehensive FISH analysis using several inactive and active gene promoters will be required to evaluate the relative position of the active and inactive promoters at the nuclear periphery in P. falciparum.

Figure 3

Active promoter localizes to H3K9Ac enriched zones located at the nuclear periphery. (A) FISH analysis shows perinuclear localization of Tu1 promoter (specifically expressed in asexual blood stages), Pfg27 promoter (expressed in gametocyte stage) and Pfs28 (a marker for ookinete stage). (B) Tu1 promoter colocalizes with active H3K9Ac marks at the nuclear periphery. FISH-IFAs were performed with Tu1 (red) labeled DNA probe and antibodies specific for H3K9Ac (green) marks. We observed colocalization of Tu1 promoter with one of the perinuclear H3K9Ac active marks. White arrow-heads indicate colocalization between Tu1 and H3K9Ac. (C) FISH-IFA analysis using Pfg27 (red) DNA probe and antibodies against H3K9Ac (green) shows that Pfg27 promoter (inactive in asexual stage) does not colocalize with active H3K9Ac marks at the nuclear periphery. (D) No colocalization takes place between TARE3 subtelomeric regions and H3K9Ac marks. FISH-IFA experiments were performed using labeled TARE3 (red) DNA probe and antibodies specific for H3K9Ac (green) marks. (E) Diagram showing interaction of Tu1 promoter with H3K9Ac. (F) Schematic representation of differential localizations of TARE3 and H3K9Ac marks. DAPI stains ring nuclei. Scale bar = 1 μm.

Active histone modifications such as H3K9Ac and H3K4me3 are commonly distributed throughout the genome as opposed to the clustered distribution of repressive histone modification, such as for H3K9me3 at silenced telomeres, TAREs1-6 and a few other regions [11]. It has been proposed that H3K9Ac is absent at telomeres and noncoding TAREs [11]. To examine the epigenetic state around the active or repressed promoters, ring stage parasites were processed for FISH-immunofluorescence assay (FISH-IFA) experiments with probe specific for Tu1 and the Pfg27 promoter region and antibodies against H3K9Ac marks. Remarkably, we found colocalization of the perinuclear H3K9Ac active marks and single copy Tu1 promoter active in the asexual stage (Fig. 3B,E). However, no colocalization was observed between the Pfg27 promoter (inactive in asexual stage) and active H3K9Ac marks (Fig. 3C,F). As a control, FISH-IFA experiments were performed with probe specific for the noncoding TARE3 region and H3K9Ac antibodies. No colocalization between multiple TARE3 foci and perinuclear H3K9Ac marks (Fig. 3D,F) could be observed. Collectively, these data suggest that active promoters in the specific regions of nuclear periphery are enriched with H3K9Ac marks, whereas developmentally regulated Pfg27 promoter and noncoding TARE sequences are devoid of active H3K9Ac marks.

Specific inhibition of H3K9 acetylation in parasites

The malaria parasites possess highly evolved epigenetic system. The higher eukaryotic homolog of GCN5 lysine acetyltransferase (KAT2B) predominantly acetylates histone H3K9 in the parasite [32]. To clarify further the relationship between H3K9Ac active mark and gene expression in vivo, we employed a recently discovered small molecule inhibitor specific for the PCAF/GCN5 family of acetyltransferases, embelin, and one of its modified inactive derivatives, MJTK-1 (Fig. 4A,B) [41]. Synchronized early ring stage (approximately 10–12 h post invasion) parasites were treated with embelin (25 μm), followed by incubation for an additional 18–20 h. Parasites were harvested at the trophozoite stage (approximately 30 h) for western blotting analysis with antibodies specific for histone H3 acetylation at the K9 and K14 positions (Fig. 4C). Antibodies against histone H3 were also used as control. We find a considerable decrease in H3K9Ac level upon embelin treatment compared to inactive derivative MJTK-1 treatment (Fig. 4C, first panel). Interestingly, no visible change was observed in H3K14Ac level following embelin treatment under similar experimental conditions confirming specificity of embelin for H3K9Ac marks in the parasites (Fig. 4C, second panel). Western blot analysis using antibodies against Histone H3 did not show any significant change in the embelin treated parasites compared to the MJTK-1 treated parasites (Fig. 4C, third panel). The bottom panel shows the coomassie stained gel as control following transfer of proteins on polyvinylidene difluoride membrane (Fig. 4C). These experiments were repeated several times and a similar result was obtained each time.

Figure 4

Embelin specifically inhibits H3K9Ac levels in parasite. (A, B) Structures of embelin and inactive derivative of embelin, MJTK-1. (C) Western blot analysis of parasites treated with inactive (MJTK-1) or active (embelin) histone acetyl transferase inhibitors in early ring stage for 18–20 h. Parasite lysates were resolved in 15% SDS/PAGE followed by transfer of proteins on polyvinylidene difluoride membrane. Polyvinylidene difluoride membranes were probed with anti-H3K9Ac (first panel) or anti-H3K14Ac polyclonal sera (second panel), respectively. Immunoblotting with anti-H3 polyclonal sera was used as loading control (third panel). The coomassie stained gel after transfer is also shown (fourth panel).

Effect of embelin on var gene expression

The FISH-IFA data clearly suggest that the active promoters are associated with H3K9Ac modification. To evaluate this in the in vivo context of gene expression, we selected 3D7 wild-type parasites for the expression of var2csa (PFL0030c), var gene specific for adherence to chondroitin sulfate A (CSA) [45, 46]. The quantitative PCR analysis revealed drastic upregulation of var2csa transcripts in CSA panned parasites compared to wild-type 3D7 control parasites (P = 0.003) (Fig. 5A, CSA unpanned versus CSA panned). CSA selected parasites were treated with 25 μm MJTK-1 and embelin in early ring stage (approximately 6 h post invasion) and harvested at the trophozoite stage (24 h post invasion) for quantitative RT-PCR analysis. It was found that the expression of var2csa mRNA transcripts was significantly decreased upon embelin treatment but not with the inactive derivative, MJTK-1 (P = 0.048) (Fig. 5B).

Figure 5

Embelin inhibits var2csa transcript levels. (A) Quantitative PCR analysis of unpanned or CSA panned parasites indicates increase in var2csa mRNA levels in comparison to constant GAPDH mRNA levels. (B) Quantitative PCR analysis of var2csa mRNA levels relative to GAPDH transcript level (internal control) in CSA selected parasites in response to MJTK-1 and embelin. (C) qChIP PCR confirms a decrease of H3K9Ac levels at the upsE promoter. The histone H3 level was used as an internal control for qChIP PCR experiments. Error bars represent the SD of three or more independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001) (Student's t-test).

Reduced expression of var2csa in the presence of embelin prompted us to check the level of H3K9Ac at the upsE promoter. Accordingly, we performed ChIP analysis using antibodies against histones H3K9Ac followed by quantitative PCR analysis using primers specific for upsE promoter. We found that the levels of H3K9Ac at the upsE promoter were decreased significantly in the presence of embelin compared to MJTK-1 (P = 0.0008) (Fig. 5C). Interestingly, after embelin treatment, the level of H3K14Ac at the upsE promoter did not change drastically compared to the H3K9Ac level (data not shown). This is consistent with the western blot data where embelin treatment does not show an overall change in the H3K14Ac level compared to the H3K9Ac level (Fig. 4C). Taken together, these data suggest the possible role of the parasite GCN5 KAT activity in maintaining the H3K9 acetylation at the promoter, which is one of the key factors for the transcriptional regulation of var2csa.

Embelin inhibits parasite growth

The specific effect of embelin on H3K9Ac level is intriguing. Because H3K9Ac marks at the various promoters modulate the transcription level of the corresponding genes, it would be interesting to explore the effect of embelin on parasite growth and survival. Accordingly, synchronized ring stage parasites were grown in the absence of any drug or in the presence of different concentration of embelin or MJTK-1 drug. The parasitemia in each case was calculated after one life cycle. The growth of embelin treated parasites was affected significantly with an increasing concentration, whereas the growth of MJTK-1 treated parasites was affected only marginally compared to the untreated parasites under the same experimental conditions (Fig. 6). The half maximal inhibitory concentration (IC₅₀) value of embelin was found to be approximately 10–12 μm. The control solvent DMSO treated parasites did not show any effect on growth. It is important to note that no visible growth arrest was found in mammalian cells for up to a 40 μm concentration of embelin under our experimental conditions (data not shown). These results clearly suggest that embelin that affects H3K9Ac levels, confering a significant effect on parasitic growth at the low micromolar range.

Figure 6

Embelin inhibits parasite growth in vitro. Synchronized ring stage parasites were treated with different concentrations of embelin or inactive drug MJKT-1 as indicated. The parasites were incubated in the presence of drug for complete one life cycle and parasitemia was calculated after approximately 52 h of drug treatment and plotted accordingly. The parasitemia for untreated (control) and solvent (DMSO) treated parasites were also plotted. The results indicate significant decrease in parasitemia for the embelin treated parasites in a concentration dependent manner. Error bars represent the SD of three independent experiments.

Discussion

Perinuclear positioning of different chromosomal domains is a characteristic feature of eukaryotic nuclei [47]. These domains occupy specific locations below the inner nuclear membrane depending on an active or repressed nature [48]. The parasite nucleus is compartmentalized into different zones below the nuclear periphery depending on the presence of specific proteins or the location and state (active or repressed) of genes. The present study focuses on the differential localization patterns of distinct histone modifications and other effector proteins in subnuclear environments of the parasite. Our data also reveal how the status of histone modifications at specific promoters/regions residing in the nuclear periphery modulates the transcriptional activity of the associated genes in the parasite.

We find punctate staining of H3K9Ac during all the stages (data not shown), including the ring stage. Previously, Petter et al. [49] reported a diffuse staining pattern for H3K9Ac (undefined stage). Subsequently, Volz et al. [43] reported distinct foci for H3K9Ac during the late troph/schizont stage. Our results are consistent with the studies reported by Volz et al. [43]. Earlier studies used commercial antibodies available against H3K9Ac. We have used antibodies against H3K9Ac that have been characterized thoroughly for their specificity [50]. The difference between the H3K9Ac pattern reported in our studies and those of earlier studies [49] may be attributed to the different source of antibodies used.

Our findings suggest differences in the localization pattern between two different active histone modifications such as H3K9Ac and H3K14Ac in the parasite nuclei. During the ring stage, H3K9Ac occupies the nuclear periphery as distinct sharp foci, whereas H3K14Ac shows diffused staining over the parasite nuclei under the same experimental conditions. A recent study showed that, despite co-occurrence of H3K9Ac and H3K14Ac at active promoters, H3K14Ac is also found in inactive inducible promoters [51]. In Plasmodium, differential distribution of H3K9Ac and H3K14Ac indicates that, apart from gene activation, these two modifications may have different roles during the ring stage of the erythrocytic parasite development. The P. falciparum genome shows global distribution of active H3K9Ac marks, leaving only telomeres and noncoding TARE regions [11]. By contrast, another histone tail modification, H3K9me3, epigenetically silences distinct var gene families with the help of specific repressive centers at the nuclear periphery. The P. falciparum telomeres and TAREs1-6 enriched with H3K9me3 are responsible for heterochromatin formation at these regions [11]. We find that Sir2, which is involved in telomeric heterochromatinization, and HP1 are separated from active H3K9Ac marks, suggesting that H3K9Ac marks are not associated with transcriptionally repressed sites.

Perinuclear positioning of genes is independent of their transcriptional status in Plasmodium [11, 52]. Consistent with earlier studies, we find perinuclear positioning of asexual stage-specific active gene promoters such as Tu1 and inactive gene promoters such as Pfg27 (active in gametocytes) and Pfs28 (active in ookinetes) (Fig. 3A). We also find active Tu1 promoter colocalized with H3K9Ac marks, whereas repressed promoters Pfg27 (active in gametocytes) and Pfs28 (active in ookinetes) behave differently in asexual blood stages, occupying a distinct region at the nuclear periphery devoid of H3K9Ac marks. It will be interesting to determine whether repressed promoters during asexual stages overlap with H3K9Ac marks at the nuclear periphery upon activation of these genes during the sexual development of the parasite.

Genes move to transcriptionally permissive perinuclear domains prior to the formation of accessible euchromatin that is essential for transcriptional activation [11, 52]. Therefore, H3K9Ac deposition is probably dependent on the locus moving to a perinuclear site for activation.

The use of GNAT family inhibitor embelin and its effect on var2csa expression at the transcript level is also intriguing because enrichment of H3K4me3 and H3K9Ac at the var gene promoter is essential for its activation at the ring stage of the parasite [11, 23]. The decrease in the H3K9Ac levels at the upsE promoter following embelin treatment suggests a role for the H3K9Ac active marks with respect to the upsE promoter associated with var2csa gene expression. However, compared to a highly significant change in H3K9Ac levels at the upsE promoter, transcription of var2csa is less significant. It is possible that other regulators may also present at the upsE promoter together with H3K9Ac for regulation of the transcriptional activation of var2csa. However, this needs to be investigated further. Embelin, specifically downregulates H3K9Ac levels in mice and recombinant PCAF-mediated acetylation in vitro [41]. PCAF belongs to the GNAT family of KATs and specifically acetylates the histone H3K9 residue. P. falciparum contains a clear homolog of yeast GCN5 histone acetyltransferase [32]. It will be interesting to investigate whether the effect of embelin in downregulating H3K9Ac levels in the parasite is mediated through inhibition of PfGCN5 activity or any other histone acetyltransferase.

Finally, specific inhibition of parasitic growth at a low micromolar concentration (IC₅₀ 10–12 μm) by embelin raises the possibility of using lysine acetyl transferases as a potential target for antimalarial therapy. It may have a far-reaching effect amidst the growing incidences of drug-resistant malaria strains, including artimisinin. It is important to note that no visible growth defect was observed in mammalian cells with up to a 40 μm concentration of embelin under our experimental conditions (data not shown). However, a comprehensive analysis of toxicity will be required both in mammalian cells and an animal model for the establishment of embelin as a potent antimalarial.

Collectively, the present study reveals differential distribution patterns of active histone modifications H3K9Ac and H3K14Ac in parasite nuclei (Fig. 7). We find that the parasite has the tendency to segregate different histone modifications relative to promoters of var genes and developmentally regulated genes. We postulate the localization of active promoters enriched with active H3K9Ac marks at the nuclear periphery, whereas developmentally regulated promoters (i.e. inactive state) are devoid of an H3K9Ac enriched nuclear periphery (Fig. 7). It would be interesting to investigate the movement of gene promoters in active or repressed states in the different developmental stages of the parasite cycle.

Figure 7

Model for differentially distributed histone modifications (H3K9Ac and H3K14Ac) and associated promoters (Pfg27, Tu1 and Pfs28). H3K9Ac shows perinuclear localization as three to seven foci (yellow circles). Association between promoter (Tu1) and H3K9Ac decides the expression of associated genes at the nuclear periphery decorated with H3K9Ac marks. Repressed promoters (Pfg27, Pfs28) occupy the region beyond the H3K9Ac enriched periphery devoid of acetylated histone marks beneath the nuclear membrane.

Materials and methods

P. falciparum culture

P. falciparum (3D7 strain) asexual stage parasites were cultured as described previously [53]. Parasites were synchronized using sorbitol (5% solution in water).

Reagents

The GNAT inhibitors embelin and MJTK-1 were purified and produced in the laboratory of T. K. Kundu (Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore, India). Rabbit anti-H3, rabbit anti-H3K9Ac and rabbit anti-H3K14Ac antibodies were raised and characterized in the same laboratory. The FISH Tag DNA Multicolor Kit (catalog number F32951), Alexa Flour 488 anti-rabbit IgG, Alexa Flour 594 anti-mice IgG and real-time PCR kit were purchased from Invitrogen (Carlsbad, CA, USA). CSA sodium salt from bovine trachea (catalog number C9819) was purchased from Sigma (St Louis, MO, USA). All other reagents were purchased from Sigma, if not stated otherwise.

Antibody production

Polyclonal antibodies against PfSir2 (full-length) were raised in mice using purified recombinant PfSir2 proteins as described previously [54]. Antibodies against PfHP1 (full-length) were also raised in mice using purified recombinant PfHP1 proteins as described previously [35].

The polyclonal antibodies against bacterially expressed recombinant histone H3 were raised in rabbit and characterized using recombinant histones and HeLa cell core histones. The acetylated site specific histone H3 antibodies were also raised in rabbit using KLH-conjugated peptides. For anti-H3K9Ac and H3K14Ac, the peptides used were: ARTKQTARK[Ac]-STGG-C-KLH and KSTGGK(Ac)APRKQ-C-KLH, respectively [50]. The specificity of these antibodies was confirmed by peptide challenge assays.

Western blot analyses using Plasmodium cell extract

Infected erythrocytes containing parasites were treated with 0.1% Saponin (Sigma) for 5 min at 4 °C followed by centrifugation at 1000 g. Parasite pellet was washed with ice cold 1 × PBS four times to remove residual hemoglobin and finally resuspended in 2 × SDS loading buffer and boiled at 95 °C for 5 min. The lysate was resolved on SDS/PAGE and transferred to polyvinylidene difluoride membrane (Millipore, Billerica, MA, USA). Western blot analysis was performed with anti-H3 (dilution 1 : 5000), anti-H3K9Ac (dilution 1 : 5000) and anti-H3K14Ac (dilution 1 : 5000) antibodies using a standard protocol.

Real-time PCR

A real-time PCR was performed as described previously [55]. Briefly, unpanned or CSA panned 3D7 parasites were isolated using saponin lysis. Parasites were resuspended in TRIzol (Invitrogen) and RNA was extracted subsequently using a standard protocol (Invitrogen). RNA samples were used for cDNA synthesis using SuperScript III Reverse transcriptase (Invitrogen). cDNA samples were used for real-time RT-PCR reactions using an absolute SYBR Green Mix (Applied Biosystems, Foster City, CA, USA), PCR primers (as shown in Table S1; quantitative RT-PCR primers) and an Applied Biosystems instrument.

Core histones isolation from P. falciparum

Core histones were isolated from parasites as described previously [56]. Briefly, asynchronous parasites were washed with ice-cold PBS followed by washing twice with hemoglobin removal buffer (25 mm Tris-HCl, pH 8.0, 1 mm EDTA, 0.2% NP40) to remove residual hemoglobin. Parasites were washed again twice with cold 1 × PBS. Parasites were washed with 800 mm NaCl three times to remove loosely attached chromatin associated proteins. Finally, parasites were extracted twice with ice-cold 0.25 m HCl. Core histones were dialyzed with water to remove acid from the preparation and stored at −80 °C.

Parasite panning for selection of var2csa expressing parasites

3D7 parasites were selected for var2csa expressing parasites by panning method on plastic dishes coated with CSA as described previously [45]. Briefly, plastic petri plates were coated with CSA (50 μg·mL⁻¹ in 1 × PBS) from bovine trachea. Petri-plates were incubated with 0.5% BSA in PBS for 1 h to block the remaining sites. Parasitized red blood cells containing mature pigmented trophozoites (> 5% parasitimea) were then incubated with immobilized CSA at 37 °C for 1 h with gentle shaking at 15-min intervals. Mature pigmented trophozoites were used for the selection process because maximum expression of var2csa specific protein on the red blood cell surface takes place in mature trophozoites. After binding, petri plates were washed gently with RPMI media six or seven times to remove unbound cells. Bound cells were again grown for 7 days. Panning was repeated six or seven times at 7-day intervals to increase var2csa expressing parasites.

IFA

Glass slides containing thin smears of P. falciparum infected erythrocytes from the synchronized stages were prepared. Cells were fixed in methanol for 10 s. Fixed slides were incubated for 1 h at room temperature in the blocking buffer containing 3% BSA and 0.01% Saponin in 1 × PBS. The slides were incubated either in the presence of immune sera raised against different proteins or with pre-immune sera at 4 °C overnight. The slides were washed further three times followed by incubation with secondary antibody (Alexa Flour 488 anti-rabbit IgG or Alexa Flour 594 anti-mice IgG) for 45 min. DNA was counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen) at a concentration of 1 μg·mL⁻¹ followed by three washes with PBS and mounted in Antifade reagent (Invitrogen) and then analyzed further using an Axioscope fluorescent microscope (Carl Zeiss, Oberkochen, Germany) for the detection of a fluorescent signal.

The antibody dilutions were: mice anti-Sir2 (dilution 1 : 500), mice anti-HP1 (dilution 1 : 500), rabbit anti-H3K9Ac (dilution 1 : 5000), rabbit anti-H3K9Ac (dilution 1 : 2500), Alexa Flour 488 anti-rabbit IgG (dilution 1 : 1000) and Alexa Flour 594 anti-mice IgG (dilution 1 : 1000).

FISH-IFA

Parasites were dried on glass slides for from 30 min to overnight at room temperature. Air dried parasites were fixed in 4% para formaldehyde solution for 10 min followed by four washes in 1 × PBS for 5 min each. Next, the parasites were permeabilized with 0.1% Triton X-100/1 × PBS for 10 min followed by washing with 1 × PBS. The parasites were subjected to FISH-IFA treatment.

FISH probes were prepared using a multi DNA FISH kit in accordance with the manufacturer's instructions (Millipore), with a few modifications. We PCR amplified the DNA sequence of interest with modified dNTPs supplied with Millipore kit for FISH probe preparation and omitted DNAse treatment for probe synthesis. All of the primers used for PCR amplification of FISH DNA probes are shown in Table S1. Fixed parasites were first incubated with hybridization solution (HS) containing 50% formamide, 10% dextran sulfate and 2 × saline-sodium phosphate-EDTA (SSPE) solution at 42 °C for 42 min. Next, 30 ng of probe was resuspended in HS and incubated with fixed parasites at 94 °C for 2 min followed by overnight incubation at 37 °C for hybridization in humidified chamber. The slide was covered with coverslip to avoid evaporation of HS. Following hybridization, the coverslip was removed and slide was washed with 2 × SSPE solution containing 50% formamide for 30 min at 37 °C followed by washing with 2 × SSPE at 50 °C for 10 min, 60 °C for 10 min and 4 × SSPE for 10 min at room temperature. Slide was washed again with 1 × PBS and fixed again in 4% paraformaldehyde solution followed by 0.1% Triton-X 100/1 × PBS treatment as described above to better preserve the complex of FISH probe with protein epitopes.

Next, the parasites were incubated with primary antibody at 4 °C overnight followed by incubation for 1 h with secondary antibody conjugated with fluorochrome. The slide was washed finally in 1 × PBS and mounted with antifade reagent (Invitrogen) and further analyzed using an Axioscope fluorescent microscope for the detection of a fluorescent signal.

ChIP assay

The ChIP assay was performed as described previously [35]. Briefly, parasites were treated with 1% formaldehyde for 10 min at 37 °C followed by treatment with 125 mm glycine to stop the reaction. Cross-linked parasites were thoroughly washed with ice cold 1 × PBS and subjected to saponin lysis. Parasites were resuspended in SDS lysis buffer (1% SDS, 10 mm EDTA, 50 mm Tris-HCL, pH 8.1) supplemented with protease inhibitor cocktail and incubated for 10 min on ice. Lysed parasites were then sonicated 10 times for 10 s to obtain sheared chromatin fragments ranging from 200 bp to 1000 bp. Parasite lysate was centrifuged at 1000 g for 15 min followed by 10-fold dilution in ChIP dilution buffer (0.01% SDS, 0.1% Triton-X 100, 1.2 mm EDTA, 16.7 mm Tris, pH 8.1, 150 mm NaCl). The diluted samples were first pre-cleared with Protein-A sepharose beads (Sigma) followed by immunoprecipitation with specific antibodies (5 μL of rabbit anti-H3, 4 μL of rabbit anti-H3K9Ac or 5 μL of rabbit anti-H3K14Ac). Immunoprecipitated chromatin samples were subjected to quantitative PCR analysis using the primers shown in Table S1 (ChIPqRT-PCR primers).

Acknowledgements

This work is supported by a Centre of Excellence (CoE) Grant (Department of Biotechnology), a Swarnajayanti Fellowship (Department of Science and Technology) and a National Biosciences Award for Career Development (Department of Biotechnology) awarded to SKD. SKD also acknowledges DST-PURSE, UGC-SAP and ICMR core funding to SCMM as financial support. Research in the laboratory of TKK is supported by grants from the Department of Biotechnology, Govt of India, through Chromatin and Disease: Program Support (Grant No. BT/01/CEIB/10/111/01 dated 30.09.2011). TKK is a recipient of Sir J. C. Bose National Fellowship, Department of Science & Technology, Govt of India. SS thanks CSIR, India, and DST, India, for fellowship. JB was supported by JNCASR, Bangalore, India. We thank Rahul Modak for technical help. Abhijit Subhashrao Deshmukh is acknowledged for his help with designing the primers for the upsE promoter used in the present study, as well as for the standardization of the quantitative PCR technique. The authors acknowledge DBT-BUILDER programme in Chemical Biology for funding.

Author contributions

SKD, TKK, SS and KB designed the experiments and analyzed the data. SS and KB performed all of the experiments. JB isolated embelin and synthesized the derivatives and SC evaluated their effect in mammalian cells. SS, KB, SKD and TKK wrote the manuscript.