Transcription factor AP2-Sp and its target genes in malarial sporozoites

Authors


*E-mail m-yuda@doc.medic.mie-u.ac.jp; Tel. (+81) 59 231 5430; Fax (+81) 59 231 5430.

Summary

The malarial sporozoite is the stage that infects the liver, and genes expressed in this stage are potential targets for vaccine development. Here, we demonstrate that specific gene expression in this stage is regulated by an AP2-related transcription factor, designated AP2-Sp (APETALA2 in sporozoites), that is expressed from the late oocyst to the salivary gland sporozoite. Disruption of the AP2-Sp gene did not affect parasite replication in the erythrocyte but resulted in loss of sporozoite formation. The electrophoretic mobility-shift assay showed that the DNA-binding domain of AP2-Sp recognizes specific eight-base sequences, beginning with TGCATG, which are present in the proximal promoter region of all known sporozoite-specific genes. Promoter assays demonstrated that these sequences act as cis-acting elements and are critical for the expression of sporozoite-specific genes with different expression profiles. In transgenic parasites that express endogenous AP2-O (APETALA2 in ookinetes), but whose AP2 domain had been swapped with that of AP2-Sp, several target genes of AP2-Sp were induced in the ookinete stage. These results indicate that AP2-Sp is a major transcription factor that regulates gene expression in the sporozoite stage.

Introduction

During a complex life cycle, Plasmodium parasites invade different types of host cell and significantly change gene expression in each stage. Until recently, however, the mechanisms of stage-specific gene regulation remained unclear and transcription factors (TFs) participating in this regulation were unknown.

Recently, Apetala2 (AP2) family genes were identified in Plasmodium species and were suggested to encode Plasmodium sequence-specific TFs (Balaji et al., 2005). This family is characterized by the AP2 domain, a DNA-binding domain composed of approximately 60 amino acids that was first identified in Arabidopsis APETALA2 protein (Jofuku et al., 1994). In human and rodent Plasmodium parasites, this family is composed of 26 members, and the amino acid sequences of the AP2 domains are highly conserved among orthologues.

AP2-related genes have been reported to be expressed in asexual intra-erythrocytic stages of Plasmodium parasites. In Plasmodium falciparum their expression during asexual replication is co-ordinated with cell cycle progression (Balaji et al., 2005; De Silva et al., 2008). Among these AP2-related genes two (PlasmoDB identifier: PF14_0633 and PFF0200c) have been suggested to participate in the induction of a group of genes that are required in a cell cycle-specific manner (De Silva et al., 2008). In a previous paper we reported that an AP2 family protein, designated AP2-O (Apetala2 in ookinetes), regulates stage-specific gene expression in the ookinete, which is a motile stage that invades the mosquito midgut (Yuda et al., 2009). AP2-O has a single AP2 domain and binds to the proximal promoter region of target genes with this domain, thereby inducing several genes involved in mosquito midgut-invasion. These findings indicate that AP2-related TFs are widely used for gene regulation in the Plasmodium life cycle.

The Plasmodium sporozoite is the stage that infects the mammalian liver. Genes expressed in this stage are potential targets for vaccine development. Identification of TFs and elucidation of the mechanisms of gene regulation in this stage should yield novel means to discover vaccine antigens. Here, we show that an AP2-related protein, designated AP2-Sp (AP2 in sporozoites), plays a central role in gene expression in the sporozoite stage using the rodent malaria parasite Plasmodium berghei. We also show that AP2-Sp is not necessary for parasite proliferation in the blood stage but is essential for formation of sporozoites.

Results

AP2-Sp is expressed from the late oocyst to the salivary gland sporozoite

To search for AP2-related genes expressed in sporozoites, we screened our P. berghei EST (expressed sequence tag) database [available at the PlasmoDB website (http://www.plasmodb.org)] and found that ESTs of AP2-Sp (PlasmoDB identifier, PB000752.01.0) are present in oocyst and salivary gland sporozoites (1 and 3 ESTs, respectively). We therefore prepared transgenic parasites that expressed GFP-tagged AP2-Sp in P. berghei (AP2-Sp::GFP parasites, Fig. 1A and B) and investigate the expression profile of AP2-Sp. In AP2-Sp::GFP parasites, GFP signals were not observed in the blood stages, including gametocytes and ookinetes cultured in vitro (Fig. S1). In the mosquito vector, however, fluorescent signals appeared in some oocysts 6 days after an infective blood meal (or fertilization), and the proportion of oocysts expressing AP2-Sp::GFP increased with time (Table S1). In these oocysts signals were weak and localized in Hoechst-stained spotted nuclei of sporoblasts (Fig. 1C, upper panels). Around 10 days after an infective blood meal, new, strong GFP signals appeared in oocysts (Fig. 1C, lower panels). These signals were localized to the nuclei of sporozoites budding from sporoblasts and therefore displayed ring-like appearances around the core of the sporoblasts. Strong GFP signals were further observed in the nucleus of sporozoites released from oocysts (oocyst sporozoites) and in sporozoites harvested from mosquito salivary glands (salivary gland sporozoites) (Fig. 1D). These results suggested that this protein would function as a TF in the sporozoite stage.

Figure 1.

AP2-Sp is necessary for sporozoite formation.
A. Preparation of parasites expressing AP2-Sp tagged with GFP. The targeting construct (upper) was inserted into the 3′-end of the endogenous AP2-Sp gene locus (middle) by homologous recombination. Transformed parasites (lower) were selected by pyrimethamine-resistance.
B. Southern blot analysis of AP2-Sp::GFP parasites. WT and AP2-Sp::GFP genomic DNA was digested with EcoRV and SacI and hybridized with the probe indicated by a solid bar in (A). WT, original wild-type parasites.
C. Upper panels: Oocyst of AP2-Sp::GFP parasites 8.5 days after an infective blood meal. Fluorescent signals of GFP were colocalized with spotted nuclei stained with Hoechst. Phase, phase-contrast image. Lower panels: Oocyst 10 days after infective blood meal. Scale bars, 5 µm.
D. Oocyst and salivary gland sporozoites of AP2-Sp::GFP parasites. GFP signals are localized in the nuclei of both stages. Scale bars, 2 µm.
E. Targeted disruption of the AP2-Sp gene. Targeting construct (upper) was inserted into the AP2-Sp gene locus (middle) by homologous recombination.
F. Southern blot analysis was carried out on AP2-Sp (−) parasite populations (1 and 2) that were independently prepared. WT and AP2-Sp (−) genomic DNA was digested with EcoRV and BglII and hybridized with the probe indicated by a solid bar in (E).
G. Oocysts of wild-type (WT) and AP2-Sp (−) 1 parasites 10 days after an infective blood meal. Midguts infected with wild-type or knockout parasites were stained with Hoechst. Budding sporozoites were not observed around sporoblasts in AP2-Sp (−) oocysts (upper panels), while numerous sporozoites budding from sporoblasts were stained with Hoechst in WT oocysts (lower panels). Scale bars, 5 µm.

AP2-Sp is necessary for the formation of sporozoites in oocysts

Next, we performed targeted disruption of AP2-Sp in P. berghei (Fig. 1E and F). Two knockout populations were obtained from independent transfection experiments. The knockout populations replicated normally in the blood, developed into ookinetes and infected the mosquito midgut; the number and sizes of oocysts formed in the midgut were essentially the same as in the wild type (Table 1). In the oocysts, however, sporozoites were not observed, even 14 days after the infective blood meal and were barely detectable after harvest from the midgut (Table 1). Nuclear staining with Hoechst showed that the nucleus of each oocyst was divided into several small nuclei 10 days after the infective blood meal in knockout as in wild-type parasites (Fig. 1G). Semithin sections stained with Giemsa, however, revealed that the subsequent invagination of the plasma membrane, which is necessary for formation of sporoblasts, did not occur in knockout parasites even 14 days after infective blood meal (Fig. S2). These results suggested that AP2-SP is involved in the expression of genes necessary for sporozoite formation.

Table 1.  Formation of oocysts and sporozoites in AP2-Sp (−) parasites.
Parasite populationsNumber of oocysts per mosquito (mean ± SE, n = 20)Diameter (µm) of oocysts (mean ± SE, n = 50)Number of oocyst sporozoites per mosquitoNumber of salivary gland sporozoites per mosquito
  • a. 

    Twenty mosquitoes were dissected, and sporozoites were counted using a haemocytometer. Each value is a mean of three independent experiments ± SE.

  • b. 

    Fifty mosquitoes were dissected, and sporozoites were counted using a haemocytometer. Two independent experiments were performed in each population.

Wild type112.95 ± 36.17133.75 ± 3.7626235 ± 2095a10750 ± 472a
AP2-Sp (−) 1115.05 ± 43.00130.91 ± 3.281.8, 0.9b0, 0b
AP2-Sp (−) 2107.65 ± 30.21138.06 ± 2.593.6, 0.9b0, 0b

AP2-Sp binds to eight bp sequences containing TGCATG

In our previous paper (Yuda et al., 2009), we reported that AP2-O binding sites in a promoter have the following two features; they are located immediately upstream of the transcription start site (TSS), and usually two or more binding sites are present. According to these observations, we explored putative cis-acting elements in sporozoite-specific genes. We found that the six-base sequence, TGCATG, and its reverse complimentary sequence, CATGCA, were frequently found in the upstream regions of these sporozoite-specific genes. In particular, a conjugated palindromic sequence, TGCATGCA, was found at high frequencies in them (see also Fig. 3A). The sequence is identical to the sequence that was recently reported to be a common upstream motif of sporozoite genes in P. falciparum (Young et al., 2008). The conjugated eight-base sequence is identical to the sequence that has been reported as a binding sequence for the AP2 domain of P. falciparum AP2-Sp (PF14_0633) (De Silva et al., 2008).

Figure 3.

Promoter activities of sporozoite-specific genes are dependent on AP2-Sp binding to sequences in the upstream region.
A. Binding sequences for AP2-Sp (rectangles) and start sites of most upstream ESTs (arrows), which indicate TSS, in the 1 kb region upstream of sporozoite-specific genes. Binding sequences are coloured according to the groups to which they belong (see also Fig. 2D). In some genes with a long 5′-untranslated region (lower right), 1.2 kb upstream regions are shown. In most genes transcription starts 100–300 bp downstream of a binding site for AP2-Sp, and two or more binding sites are usually within the 1 kb region upstream of the first ATG codon. EST data are available at the PlasmoDB website. Arrowheads indicate binding sequences that were mutated in reporter assays in (B) and (C).
B. Reporter assays of MAEBL and SPECT2 were performed in vivo with pCen-GFP. Promoter activities in oocyst (OC) and salivary gland (SG) sporozoites were compared between original (WT) and mutated (M1) promoters. In both genes, a point mutation was added to the binding sequence for AP2-Sp (TGCATGCA to TGGATGCA) in the upstream region.
C. The reporter assays of SPECT and UIS3 were performed using pCen-Luc. A mutation was added to one [numbered 1 in (A), M1] or two binding sites (1 and 2, M2) in SPECT and to one binding site (M1) in UIS3. Results were expressed as a percentage of the original activity in salivary gland sporozoites.

We next examined whether the AP2 domain of P. berghei AP2-Sp binds to the six-base sequence using EMSA (electrophoretic mobility-shift assay). Because the AP2 domain of AP2-Sp has an adjacent AT-hook motif (Balaji et al., 2005) and this region might participate in binding, recombinant AP2 domains with and without the AT-hook region were tested. As a probe, a 192 bp region upstream of the MAEBL gene, a sporozoite-specific gene essential for salivary gland infection (Kariu et al., 2002), was used. This region contained a single TGCATGCA site. As shown in Fig. 2A, EMSA generated clearly shifted bands for both recombinant proteins. However, addition of a single mutation to the putative binding sequence (TGCATGCA to TGGATGCA) virtually abolished these bands. This result indicated that the AP2 domain specifically recognizes this sequence.

Figure 2.

AP2 domain of AP2-Sp recognizes eight-base sequences beginning with TGCATG.
A. EMSA was performed with the GST-fused AP2 domain of AP2-Sp, with and without the AT-hook region (GST-AT-AP2 and GST-AP2 respectively). The upstream region of MAEBL (192 bp), which contains one TGCATGCA sequence, was used as a probe (WT). A point mutation (TGCATGCA to TGGATGCA) in this probe (M1) nearly abolished the shifted band (arrow head). GST was used as a control.
B. EMSA was performed with the GST-fused AP2 domain of AP2-Sp without the AT-hook region. Oligonucleotide probes (29 bp) were designed based on the upstream region of MAEBL around the TGCATGCA sequence. Mutations added to this probe are indicated under each lane with small characters. A mutation to CATG at the center of TGCATGCA abolished the shifted band (left panel). A mutation to either TG or CA at the end of TGCATGCA significantly reduced the shifted band, and mutations to both ends together abolished the shifted band (centre panel). Mutation of the bases beside TGCATGCA had no affect on binding (right panel).
C and D. Three independent EMSAs were performed with oligonucleotide probes that are different from the original probe in (B) at either or both of the two bases following TGCATG. One of the electrophoresis images obtained in these assays is shown in (C). Intensities of shifted bands were measured with a densitometer, and the averages of three assays are shown in a bar graph with standard errors in (D). The sequences of oligonucleotide probes were roughly divided into three groups according to the band intensities, as indicated under the graph.

Next, we designed short (29 bp) double-stranded oligonucleotide probes from the region around the TGCATG site upstream of MAEBL. These probes contained various point mutations, to determine the sequence necessary for binding to the AP2 domain (Fig. 2B). Only GST-fused proteins, without the AT-hook region, were used in these assays, because there seemed to be no difference in binding properties between the two recombinant proteins. The results indicated that CATG is a core sequence essential for binding (Fig. 2B, left and centre panels) and that the two bases either side of this sequence also participate in binding (Fig. 2B, left panel), i.e. the whole palindromic sequence of eight bases, TGCATGCA, is involved in the binding of the AP2 domain. The minimum sequence essential for binding the AP2 domain was TGCATG, because, binding was abolished only when both TGCATG and its reverse complementary sequence CATGCA were mutated at the same time (Fig. 2B, centre panel). A mutation to either side of the sequence had no affect on binding (Fig. 2B, right panel).

These results suggest that AP2-Sp could bind to several eight-base sequences beginning with TGCATG but that binding properties would vary among them. We therefore compared binding affinities among these sequences using 29 bp short oligonucleotide probes (Fig. 2C and D). The results showed that binding sequences for the AP2 domain could be roughly divided into three groups according to binding properties. The first group is composed of TGCATGCA and TGCATGCG, to which the AP2 domain binds most tightly. The second group contains four sequences, TGCATGCC/T and TGCATGTA/G, which are identical to the sequences of the first group at seven bases and have lower affinities. The third group includes other sequences containing TGCATG that also resulted in a shifted EMSA band, although with lower binding affinities than the former two groups.

Figure 3A illustrates these binding sequences in 1 kb upstream regions (and 1.2 kb regions in some genes) of sporozoite-specific genes and the most upstream ESTs, which indicate the putative TSS of each gene. All of these genes have one of the eight-base sequences in the upstream region (over half of them have the eight-base sequences belonging to the first group), and transcription of these genes starts 100–300 bp downstream of the binding sequences. Computational analyses in 1 kb upstream regions of 13 genes, which have been reported sporozoite-specific, demonstrated that the six-base sequence occurs at high frequencies in these genes (Table S2). These results strongly suggest that binding sequences for AP2-Sp act as common cis-acting elements in the upstream regions of sporozoite-specific genes and that they are usually located in the proximal promoter region of these genes. These features are similar to those of AP2-O, which we reported previously (Yuda et al., 2009). We also found that some eight-base sequences beginning with CGCATG can be binding sequences for AP2-Sp (data not shown). However, we omitted these sequences from the following analyses because the frequencies of these GC-rich sequences in the genome are very low (an example that such sequences act as cis-acting motifs in vivo is shown in the promoter analysis of PB000863.01.0 in Fig. 4B).

Figure 4.

Induction of AP2-Sp target genes in ookinetes.
A. Preparation of AP2-O::Sp parasites. DNA construct containing the coding region of the AP2 domain of AP2-Sp (upper) was inserted in frame into the 3′-end of the endogenous AP2-O gene (middle) by homologous recombination, resulting in the swap of the region encoding the AP2 domain of AP2-O with the region encoding the AP2 domain of AP2-Sp (lower). Preparation of AP2-O::Sp::GFP parasites was carried out with the same procedure but using a construct containing the coding region of the AP2 domain of AP2-Sp followed by the GFP gene. Results of Southern hybridization with the probe indicated by a solid bar are shown in the right panel.
B. Reporter assays were performed with the artificial chromosome pCen-Luc. Promoter activities in oocyst (OC) and salivary gland (SG) sporozoites were compared between the original upstream region (WT) and regions containing point mutations at binding sites for AP2-Sp. The binding sites mutated were TGCATGCG and TGCATGCT (M2, see also Fig. 3S) in PF11_0545 orthologue, and two TGCATGTGs (1 and 2 in Fig. 3S, M2) and CGCATGTT and CGCATGTG (3 and 4 in Fig. 3S, M4) in PB000863.01.0. Results were expressed as a percentage of the original activity in salivary gland sporozoites.
C. ChIP was performed in AP2-O::Sp::GFP parasites with anti-GFP antibodies. Precipitated DNA fragments were analysed by qPCR using primers for upstream regions of 11 genes including those induced in AP2-O::Sp parasites (SPECT, SPECT2, PF11_0545 orthologue, PB000522.03.0 and PB000863.01.0), other known sporozoite-specific genes (TRAP, MAEBL, UIS3 and SERA8) and ookinete-specific genes (chitinase and SOAP) as controls. Percentage of input of each gene is shown in Table S8.

TGCATGCA is a cis-acting element specific for the sporozoite stage

We next examined whether the eight-base sequences act as cis-acting elements in the sporozoite stage using in vivo promoter assays (Fig. 3B and C). First, we tested TGCATGCA in the upstream region of MAEBL (Fig. 3B). For this assay we used the P. berghei centromere plasmid pCen-GFP, which contains the GFP reporter gene (Fig. S3A) (Yuda et al., 2009). Reporter constructs made in this plasmid are retained through the sporozoite stage by approximately 90% of parasites without drug selection, having been introduced into asexual blood stages by electroporation (S. Iwanaga et al., unpubl. results). In parasites transfected with pCen-GFP containing the MAEBL promoter, strong GFP signals were observed in oocyst sporozoites. The signal intensity reached a peak in this stage, which was in accordance with the expression profile of the MAEBL gene in vivo. Addition of a point mutation to the eight-base sequence (TGCATGCA to TGGATGCA) decreased the signals intensity to the background level, indicating that this sequence is critical for promoter activity. Next, we tested the upstream region of SPECT2, which also contains a single TGCATGCA site (Fig. 3B, right). SPECT2 is necessary for skin and liver invasion and is predominantly expressed in the salivary gland sporozoite (Ishino et al., 2005; Amino et al., 2008). In parasites transfected with pCen-GFP containing the SPECT2 promoter, GFP signals were weak in oocyst sporozoites but increased approximately 20 times after salivary gland infection. This steep increase in promoter activity coincides well with the expression profile of SPECT2. Addition of a point mutation to the motif decreased GFP signals to background levels in both oocyst sporozoites and salivary gland sporozoites.

The promoter region of sporozoite-specific genes usually has multiple binding sites for AP2-Sp. To examine how these individual sequences contribute to the overall promoter activity, we next performed promoter assay with pCen-Luc in the SPECT promoter (Fig. 3C, left) (Ishino et al., 2004) where two putative cis-acting elements different from TGCATGCA tested above (TGCATGCG and TGCATGTT) are present (Fig. 3A). In these assays luciferase was used as a reporter (pCen-Luc, Fig. S3B) instead of GFP to assess promoter activity more quantitatively. The activity of the SPECT promoter decreased according to the number of binding sites that were mutated. This indicated that both TGCATGCG and TGCATGTT act as cis-acting elements and that these elements additively affect promoter activity. The additive effect on promoter activity was similar to that reported for AP2-O binding sequences in the ookinete stage (Yuda et al., 2009).

We also tested the promoter of UIS3 (upregulated in infective sporozoites gene 3) (Mueller et al., 2005) (Fig. 3C, right). UIS3 is upregulated in salivary gland sporozoites and is also expressed in the liver stage. Therefore, the expression profile and function is different from the three genes tested above. Addition of a mutation to the single TGCATGCA site significantly decreased promoter activity of the UIS3 upstream region in both oocyst and salivary gland sporozoites. These results together with those described above demonstrated that the eight-base sequence acts as a critical cis-acting element, irrespective of whether the peak activities of the promoter are present in oocyst or salivary gland sporozoites.

Induction of AP2-Sp target genes in the ookinete stage

The results presented above indicated that AP2-Sp is a trans-acting factor that induces sporozoite-specific genes. However, it was unknown whether this protein directly binds to the promoter of these genes in vivo. ChIP-qPCR (chromatin immunoprecipitation with quantitative PCR) analysis is an appropriate technique to demonstrate this. However, we were unable to obtain sufficient parasites for this analysis. Indeed, we attempted a ChIP assay with sporozoites that were harvested from 300 infected mosquitoes (approximately 6 × 106 sporozoites), but insufficient chromatin DNA for subsequent qPCR analysis was recovered (data not shown).

To conquer this problem, we prepared parasites that expressed endogenous AP2-O, but whose AP2 domain had been swapped with that of AP2-Sp (AP2-O::Sp parasites and AP2-O::Sp::GFP parasites, Fig. 4A), and examined (i) whether the chimeric TF might induce target genes of AP2-Sp in the ookinete stage and (ii) whether the TF would bind to the promoter regions of sporozoite-specific genes. To assess the induction, we analysed gene expression by microarray analysis of ookinetes using AP2-O::Sp parasites and selected genes whose expression increased over threefold compared with AP2-O (−) parasites. The microarray analysis showed that expression increased in 33 genes (Table S3). They were composed of six sporozoite-specific genes [SPECT, SPECT2, CS, TLP (TRAP-like protein), S10 and S23], whose expression in sporozoites has been demonstrated or suggested (Matuschewski et al., 2002; Ishino et al., 2004; 2005; Moreira et al., 2008), and 27 novel genes whose expression has not been reported in the sporozoite stage. Computational analysis of upstream regions of the 30 genes (upstream regions of three genes were not obtained from the P. berghei genome database) showed that sequences containing TGCATGCA occurred at significantly high frequencies, indicating that direct binding of the TF to these regions could induce these genes (Table S4). Figure S3 illustrates the binding sites for AP2-Sp in the upstream regions of the 26 genes that are not in Fig. 3A. Of the 30 genes, 19 had at least one binding site belonging to the first group (TGCATGCA, TGCATGCG and their reverse complementary sequences, see also Table S3), and 17 genes had at least one binding site belonging to the second group (TGCATGTG, TGCATGTA, TGCATGCT, TGCATGCC and their reverse complementary sequences). The most upstream ESTs of each gene (Fig. S3) indicated that transcription of these genes starts approximately 100–300 bp downstream of the binding sequences, as was observed for the sporozoite-specific genes in Fig. 3A.

We performed further promoter assays to confirm that these genes really are under the control of AP2-Sp in the sporozoite stage (Fig. 4B). We tested two genes. One of the genes we tested currently has no gene ID in PlasmoDB (in P. falciparum PlasmoDB identifier, PF11_0545, has been attributed to the orthologue). The upstream region of this gene had a single binding sequence, TGCATGCG, and ESTs of this gene were mainly present in salivary gland sporozoites (Fig. S4 and Table S3). The promoter activity of this gene increased after salivary gland infection, and addition of a mutation (TGCATGCG to TGGATGCG) significantly decreased the activity in both oocyst and salivary gland sporozoites. Another gene we tested (PlasmoDB identifier: PB000863.01.0) was a member of the AP2-related gene family and has an orthologue in P. falciparum (PlasmoDB identifier: PFD0985w). This gene has four putative cis-acting elements in the upstream region, two TGCATGTG and two CGCATGTG (Fig. S4), and ESTs of this gene were present mainly in salivary gland sporozoites (Table S3). Promoter activity of this gene increased steeply after salivary gland infection, and point mutations to the motifs (in two and four motifs) reduced the activity stepwise. These results indicated that expression of these two genes is controlled by AP2-Sp in the sporozoite stage.

We then examined direct binding of the chimeric TF to the promoter region of these genes by ChIP-qPCR assays using parasites expressing the chimeric TF fused with GFP at the C-terminus (AP2-O::SP::GFP parasites). In these parasites, GFP signals were detected in the nucleus of ookinetes, as was seen in AP2-O::GFP parasites (data not shown). As shown in Fig. 4C, IP with antibodies against GFP significantly enriched the upstream regions of sporozoite-specific genes, including those induced in AP2-O::SP parasites (SPECT, SPECT2, CS, P. berghei orthologue of PF11_0545, PB000863.01.0 and PB000522.03.0) and also MAEBL, UIS3, TRAP and SERA (serine repeat antigen) 8/egress cysteine protease 1 (Aly and Matuschewski, 2005), which were not significantly induced in AP2-O::SP parasites. In contrast, binding to the upstream regions of SOAP (secreted ookinete adhesive protein) and chitinase, which are ookinete-specific genes (Dessens et al., 2003), were not observed. These results demonstrated that the AP2 domain of AP2-Sp specifically binds to the upstream region of sporozoite-specific genes in vivo.

Discussion

Proteins expressed in sporozoites are potential targets for vaccine development, but the mechanisms of gene regulation in this stage remain poorly understood. Here we have provided evidence that the AP2 family TF, AP2-Sp, widely regulates gene expression in the sporozoite stage. This TF is expressed in the period from the late oocyst to the salivary gland sporozoite and induces the expression of genes with various functions and expression profiles. Target genes of this TF include all known genes specifically expressed in sporozoites, suggesting that this TF plays a central role in gene regulation of the sporozoite stage. We also demonstrated that AP2-Sp induces these target genes by binding to specific cis-acting elements in proximal promoter regions and that these elements are critical for promoter activity of each gene.

In a previous paper we reported that another AP2 family TF, AP2-O, regulates gene expression in the ookinete stage. Similar to AP2-Sp, Ap2-O directly activates several genes, including all known genes specifically expressed in the ookinete stage. AP2-O also shares some unique features of gene activation with AP2-Sp, such as binding to the proximal promoter region and additive effects of binding sites with respect to promoter activity. This indicates that mechanisms of gene regulation are closely related in these two stages.

However, a difference in gene expression between these two stages is that the sporozoite stage is composed of two forms with distinct infection ability, salivary and oocyst sporozoites, and thus target genes of AP2-Sp include those exhibiting distinct expression patterns such as MAEBL and SPECT2. Because such distinct expression patterns are difficult to explain solely by interactions between AP2-Sp and its binding sites on the promoter, there may be other TFs or co-regulators that modulate the function of AP2-Sp. Whereas at present we have not yet succeeded in identifying them, further investigation of AP2 family TFs in this stage and promoter analysis of genes exhibiting distinct expression patterns using the centromere plasmid might answer the question of how Plasmodium sporozoites change infection capability in the mosquito.

AP2-Sp (−) parasites form oocysts of normal size and number but they fail to generate sporozoites. This phenotype resembles to that of CS (−) parasites (Menard et al., 1997), suggesting that loss of CS protein in AP2-Sp (−) parasites could be the cause of the phenotype. However, the possibility cannot be excluded that other genes involved in sporozoite formation were also regulated by AP2-Sp and that the reduced expression of these genes contributed to the phenotype. Indeed, expression of AP2-Sp starts 5 or 6 days before sporozoite budding, therefore a set of genes encoding cytoskeletal proteins, including those necessary for apical complex formation, could be under the control of AP2-Sp. In a previous paper (Yuda et al., 2009), we reported that AP2-O (−) parasites could not form ookinetes of normal elongated shape, which suggested that AP2-O controls genes that participate in the morphogenesis of the invasive stage. Considering the analogous role of AP2-Sp and AP2-O in the respective motile stages it is possible that AP2-Sp regulates a group of genes necessary for formation of the sporozoite.

The chimeric TF induced several sporozoite-specific genes in ookinetes. ChIP-qPCR assays demonstrated that the chimeric TF activated these genes by directly binding to their promoter regions. This indicated that AP2-Sp is the trans-factor that interacts with the cis-elements. In this assay, however, the expression levels of the induced genes were not as high as those of the original ookinete-specific genes (data not shown) and significant induction was not observed in many genes, including MAEBL, UIS3 and TRAP. It remains unclear why this expression system did not work well on these promoters, but it is possible that co-activators of AP2-Sp, which would not be expressed in the ookinete, might be required for full activation of these promoters. Although further improvement may be required, this system could be a useful tool to determine target genes of AP2-family TFs. On the other hand, the results of the ChIP-qPCR assays suggested that the proximal promoter regions of sporozoite-specific genes are readily accessible in the ookinete stage where these genes are not primarily expressed.

Our study has demonstrated that AP2-Sp plays a central role in stage-specific gene expression in sporozoites and that cis-acting motifs in the proximal promoter region determine gene expression in this stage. Because the eight-base binding sequences for AP2-Sp are rarely found in the genome, these elements could mark sporozoite-specific genes on the genome. The present findings should pave the way for predicting genes expressed in this stage, based solely on genome sequence.

Experimental procedures

Parasite preparations

Female BALB/c mice (6–10 weeks old, Japan SLC, Hamamatsu, Japan) were infected with P. berghei ANKA strain as described previously (Yuda et al., 2009). Oocyst sporozoites and salivary gland sporozoites were harvested 14 and 24 days after an infective blood meal respectively. For nuclear staining of oocysts, mosquitoes were dissected 7, 10 and 14 days after an infective blood meal, and their midguts were stained with Hoechst 34580 (Molecular Probes, Eugene, OR, USA) (0.2 µg ml−1 PBS final concentration) for 10 min at RT. Ookinete culture was performed as described previously (Yuda et al., 2009).

Preparation of GFP-tagged AP2-Sp expressing parasites

For the targeted insertion construct (Fig. 1A), a DNA fragment containing the 3′ part of the AP2-Sp coding region was amplified by PCR, using genomic DNA as a template, and inserted into the XhoI/NheI site of the construct in frame with the GFP gene. The downstream region of the AP2-Sp gene was also amplified by PCR and inserted into the pBluescript BamHI/NotI site. Plasmids containing the construct were digested with XhoI and NotI and then used for transfection. The PCR primer pairs used for preparing the construct and Southern hybridization probe are listed in Table S5.

Targeted disruption of the AP2-Sp gene

The targeting construct was prepared by PCR using essentially the same procedure as described previously (Yuda et al., 2009). The PCR primer pairs used for preparing the targeting construct and Southern hybridization probe are listed in Table S5.

Electrophoretic mobility-shift assays

The AP2 domain of AP2-SP was produced as a glutathione S–transferase fusion protein using the GST Gene Fusion System (Amersham Bioscience, Tokyo, Japan) essentially as described previously (Yuda et al., 2009). Briefly, the coding regions were amplified from P. berghei genomic DNA with the primer set: 5′-CGGGATCCGCCTTTAGGGTATTTGATGTAGAC-3′ and 5′-CCGCTCGAGTTAATACTTTAGTTTCATCATTTCGC-3′ for the AP2 domain with AT-hook region (amino acid residues 144–248) and with the primer set: 5′-CGGGATCCGCCTTTAGGGTATTTGATGTAGAC-3′ and 5′-CCGCTCGAGTTAATACTTTAGTTTCATCATTTCGC-3′ for the AP2 domain without AT-hook region (amino acid residues 157–248). Probes were prepared by PCR with 5′-biotinylated primers using the cloned promoter region of the MAEBL gene as template. The primer pairs used were 5′-TAAAAATGTAAAGCATTTGAATTAAGAACC-3′ and 5′-GATGTATTTTTTGTGTAGAAAAACTGAAGG-3′. Short double-strand oligonucleotide probes were prepared from synthetic 5′-biotinylated oligonucleotides. The original oligonucleotide sequence was 5′-TAATATTATTATGCATGCATTCTTATAAG-3′. EMSA was performed as described previously (Yuda et al., 2009).

Reporter assays with P. berghei centromere plasmids

Reporter assays were performed with the P. berghei centromere plasmid, pCen-GFP and pCen-Luc (Fig. S3). The upstream region of the MAEBL or SPECT2 gene was inserted into the EcoRV/BamHI site, upstream of the GFP gene of pCen-GFP. For luciferase assay, the upstream promoter region was inserted into the KpnI/XhoI site upstream of the luciferase gene in pCen-Luc. The PCR primer pairs used for amplification of the upstream regions are listed in Table S6. Transfection and sporozoite preparations were performed as described previously (Yuda et al., 2009). The rate of GFP-positive parasites was calculated in each sporozoite preparation. Luciferase activity was measured using 10 000 sporozoites with the Luciferase Assay System (Promega Corp., Madison, WI, USA) and normalized to rates of GFP-positive parasites.

Preparation of mutant parasites expressing AP2-O with the AP2 domain of AP2-Sp (AP2-O::Sp parasites and AP2-O::Sp::GFP parasites)

The targeted insertion construct for AP2-O::Sp parasites was prepared as follows. Part of the AP2-O coding region, just upstream of the AP2 domain, was amplified by PCR and ligated in frame to a DNA fragment encoding the AP2 domain of AP2-SP. This fragment was inserted into the XhoI/BglII site of the targeted insertion construct for AP2-O::GFP (Fig. 4A) in place of the 3′ part of the AP2-Sp coding region and the GFP gene. To prepare the targeted insertion construct for AP2-O::Sp::GFP parasites, the same fragment with XhoI/NheI sites at the termini was inserted into the XhoI/NheI site of the targeted insertion construct for AP2-O::GFP, in frame with the GFP gene. Plasmids containing these constructs were separated from plasmid backbone by digestion with XhoI and NotI and then used for transfection. The PCR primer pairs used for preparing the targeted insertion constructs and Southern hybridization probe are listed in Table S5.

DNA microarray analysis

Ookinetes were cultured for 12 h. Five biologically independent samples were prepared from AP2-O::Sp and AP2-O (−) parasites. AP2-O (−) parasites were used as controls. Extraction of total RNA and the following microarray analyses were performed using the same procedure as described in Yuda et al., (2009). Genes that increased more than threefold compared with AP2-O (−) parasites, for at least two probes, were selected as being induced in AP2-O::Sp parasites. From the selected genes, those having no orthologues in P. falciparum were excluded, because such genes are usually located in the subtelomeric regions and belong to multigene families such as the bir gene family. Microarray data have been submitted to Gene Expression Omnibus (GEO) under Accession No. GSE18910.

ChIP-quantitative PCR assays

ChIP was performed using essentially the same procedure as described previously (Yuda et al., 2009). DNA fragments obtained by IP were analysed by real-time PCR using an iCycler iQ Real-Time Detection System (Bio-Rad, Hercules, CA, USA) with the primers listed in Table S7.

Acknowledgements

This work was supported by the Ministry of Education, Science, Culture, and Sports of Japan (Grant 20249023, 21022019 and 21659105 to M.Y., Grant 21022032 and 21790406 to S.I. and Grant 21790403 to I.K.) and by the Ministry of Health, Labor, and Welfare (Research Grant for Research on Emerging and Re-emerging Infectious Diseases to H.K.).

Ancillary