The Plasmodium serine-type SERA proteases display distinct expression patterns and non-essential in vivo roles during life cycle progression of the malaria parasite

Parasite proteases play key roles in several fundamental steps of the Plasmodium life cycle, including haemoglobin degradation, host cell invasion and parasite egress. Plasmodium exit from infected host cells appears to be mediated by a class of papain-like cysteine proteases called ‘serine repeat antigens’ (SERAs). A SERA subfamily, represented by Plasmodium falciparum SERA5, contains an atypical active site serine residue instead of a catalytic cysteine. Members of this SERAser subfamily are abundantly expressed in asexual blood stages, rendering them attractive drug and vaccine targets. In this study, we show by antibody localization and in vivo fluorescent tagging with the red fluorescent protein mCherry that the two P. berghei serine-type family members, PbSERA1 and PbSERA2, display differential expression towards the final stages of merozoite formation. Via targeted gene replacement, we generated single and double gene knockouts of the P. berghei SERAser genes. These loss-of-function lines progressed normally through the parasite life cycle, suggesting a specialized, non-vital role for serine-type SERAs in vivo. Parasites lacking PbSERAser showed increased expression of the cysteine-type PbSERA3. Compensatory mechanisms between distinct SERA subfamilies may thus explain the absence of phenotypical defect in SERAser disruptants, and challenge the suitability to develop potent antimalarial drugs based on specific inhibitors of Plasmodium serine-type SERAs.


Introduction
Intracellular pathogens have evolved numerous strategies to exit their host cells after completion of replication and growth and depletion of host cell nutrients (Hybiske and Stephens, 2008). Cellular exit is often an active biological process triggered by the pathogen and accompanied by consecutive breaching of the membrane of the parasitophorous vacuole (PV) that harbours the pathogen and the host cell plasma membrane.
Plasmodium and other apicomplexan parasites are obligate intracellular pathogens that need to efficiently enter and exit their respective host cells in order to propagate and progress along the life cycle. Studies with broadspectrum cysteine inhibitors have indicated central roles for proteolytic events during egress of merozoites, the invasive stage of the malarial parasite in the pathogenic red blood cell cycle, out of the PV and the erythrocyte plasma membrane (Salmon et al., 2001;Wickham et al., 2003). Plasmodium appears to compartmentalize proteins that function specifically in parasite egress in specialized electron-dense secretory organelles termed 'exonemes' (Yeoh et al., 2007). Exonemes contain the subtilisin-like serine protease subtilase 1 (SUB1) that is essential for parasite growth and can proteolytically activate a family of papain-like proteases termed 'serinerepeat antigens' (SERAs), which in turn may mediate parasite egress through subsequent processing of cellular substrates. Therefore, exoneme discharge may trigger a proteolytic cascade that ultimately leads to cytolysis and parasite exit (Yeoh et al., 2007). Understanding the cellular roles of SERAs, which constitute major substrates of SUB1, may ultimately lead to the identification of parasite and/or host cell substrates and the underlying molecular mechanisms of proteolysis. Direct support for the proposed roles for SERAs in parasite egress comes from experimental genetics. Loss of PbSERA5/ECP1 function results in viable and motile sporozoites that are defective in exiting the midgut oocyst in the insect vector (Aly and Matuschewski, 2005). Remarkably, members of the SERA multigene family appear to have arisen from multiple gene duplication events. In P. falciparum eight out of nine SERAs are located in tandem on chromosome 2 (Aoki et al., 2002;Miller et al., 2002). Similarly, in the rodent malaria model parasite P. berghei the five SERAs are tandemly arranged in a head-to-tail fashion on chromosome 3 (Kooij et al., 2005).
This gene organization is evolutionary conserved and is a hallmark of the SERA multigene family (Arisue et al., 2007;McCoubrie et al., 2007).
All Plasmodium SERAs contain a central, papain-like protease domain and numerous cysteine residues. Intriguingly, this class of proteins appears to be absent in a number of related apicomplexan parasites, such as Toxoplasma gondii or Cryptosporidium parvum, suggesting that their respective roles are restricted to malaria parasites. Despite their overall sequence similarity in their central protease domain, Plasmodium SERA proteins can be classified into four major groups that form two distinct phylogenetic clusters (Hodder et al., 2003;Kooij et al., 2005;Arisue et al., 2007;McCoubrie et al., 2007). The active site cysteine SERAs (SERAcys) form three separate groups within one cluster, whereas those with an active site serine (SERAser) form a fourth monophyletic group.
The three orthologous SERAcys groups appear to be well conserved across the genus Plasmodium. Two groups, represented by P. falciparum PfSERA6 and PfSERA7, respectively, are expressed in asexual parasites, whereas the third and most ancestral group, represented by PfSERA8, is not (Aoki et al., 2002;Miller et al., 2002). Targeted gene deletion of the P. berghei orthologue of PfSERA8, termed 'egress cysteine protease 1' (ECP1), confirmed a dispensable role in the mammalian host and instead revealed an essential function for sporozoite egress from oocysts in the mosquito vector (Aly and Matuschewski, 2005). By analogy, members of the PfSERA6 and PfSERA7 groups may function in parasite egress out of mammalian host cells.
In contrast, the cellular roles of SERAser proteins, which together form the most diverse group, remain largely unsolved. The founding member PfSERA5 localizes to the PV of mature schizonts (Delplace et al., 1987;Miller et al., 2002). Purified recombinant PfSERA5 protein exhibits only limited chymotrypsin-like autoproteolytic activity and cleavage of polypeptide substrates is negli-gible (Hodder et al., 2003). However, this group stands apart, because (i) expression analysis revealed that SERAser genes, particularly SERA5, are very abundantly expressed in P. falciparum late trophozoites and schizonts, the parasite stages preceeding parasite egress from their host erythrocytes (Aoki et al., 2002;Lasonder et al., 2002;Miller et al., 2002), (ii) antibodies against PfSERA5 inhibit parasite erythrocytic growth in vitro through agglutination of merozoites and ruptured schizonts (Pang et al., 1999), (iii) infected individuals in malaria-endemic areas exhibit high antibody titres against SERAser proteins, and most prominently SERA5 (Okech et al., 2001;Aoki et al., 2002;Okech et al., 2006), and (iv) high anti-PfSERA5 antibody titres correlate with protection against severe disease (Okech et al., 2006). In a monkey model immunization with a purified recombinant PfSERA5 fragment induces protection against challenge infection (Inselburg et al., 1991). Therefore, SERAser proteins represent the most promising group of all SERAs for potential therapeutic and vaccine targets.
In this study, we investigated the cellular roles of SERAser by experimental genetics in the model rodent malaria parasite P. berghei. Unexpectedly, we could exclude an essential role for all SERAser during the P. berghei life cycle. Our data suggest that this monophyletic SERA group evolved in the absence of vital roles for the parasite.
Direct sequencing of cDNA from asynchronous blood stages permitted identification of the complete coding sequences (GenBank accession numbers: EU917224 and EU917225 for PbSERA1 and PbSERA2 respectively). Comparison of the P. berghei orthologues with the founding member PfSERA5 (PFB0340c) illustrates the overall amino acid sequence similarity (Fig. 1A) of~35% to the human malaria protein. A hallmark of this subfamily is the replacement of the catalytically active cysteine residue by a corresponding serine residue (Fig. 1B). The other amino acids of the catalytic centre, i.e. an aminoterminal glutamine and a carboxyterminal asparagine, are well conserved, apart from a histidine residue, which is changed to a methionine residue in the homologues of rodent malaria parasites (Hodder et al., 2003;Arisue et al., 2007;McCoubrie et al., 2007).
In order to initiate a genetic characterization of the SERAser subfamily, we profiled their expression by RT-PCR analysis ( Fig. 2A). In good agreement with P. falciparum expression data (Aoki et al., 2002;Miller et al., 2002), both transcripts are readily detectable in blood stage merozoites. Both transcripts are absent during sporozoite maturation and expression commences again during liver stage development. Both PbSERA1 and PbSERA2 appear to be abundantly expressed late in liver stage development ( Fig. 2A), as observed previously (Schmidt-Christensen et al., 2008). Together, these data suggest that the two SERAser genes are expressed during formation of liver stage and blood stage merozoites.
We further quantified the relative transcript abundance for the five P. berghei SERA genes using real-time RT-PCR (Fig. 2B). The two SERAser genes, PbSERA1 and PbSERA2, were the most abundantly expressed in blood stages. In late liver stages, PbSERA1, PbSERA2 and PbSERA3 were expressed at a similar level, whereas PbSERA4 transcripts were less abundant. As expected (Aly and Matuschewski, 2005), PbSERA5 was not expressed in liver or blood stages (Fig. 2B). These data confirm that the SERAser genes (and particularly PbSERA2) are the SERAs most prominently expressed in blood stages, consistent with P. falciparum expression data (McCoubrie et al., 2007).

Cellular localization of PbSERA1 and PbSERA2
We next investigated the localization and the expression timing of PbSERA1 and PbSERA2. For this purpose we generated parasite lines expressing the endogenous SERA1 and SERA2 proteins fused to the mCherry red fluorescent protein (Shaner et al., 2004). This was achieved by transfection of targeting vectors that contain an amino-terminally truncated PbSERA1 or PbSERA2 copy and in-frame fusion of the mCherry coding region, followed by the DHFR/TS 3′ untranslated region (Fig. 3A). Upon a single cross-over event, integration of these constructs is predicted to result in an allelic duplication, resulting in a mCherry-tagged full-length copy and a non-transcribed 5′ truncated version of the PbSERA1 or PbSERA2 gene.
Transfection was performed in P. berghei ANKA parasites expressing GFP (Janse et al., 2006), leading to green fluorescent parasites that express a red fluorescent SERA1 or SERA2 protein. Genotyping by PCR using specific primer combinations confirmed the desired integration events (Fig. S1).
We first performed live cell imaging of blood stages of the transgenic PbSERA1/mCherry and PbSERA2/ mCherry parasites. Similarly to SERA5 in P. falciparum (Delplace et al., 1987), both PbSERA1/mCherry and PbSERA2/mCherry were detected in late schizonts, but not in early blood stages (Fig. 3B). PbSERA1/mCherry and PbSERA2/mCherry parasites developed normal asexual and sexual blood stages, and could be transmitted to Anopheles mosquitoes, resulting in the formation of sporozoites (data not shown). In good agreement with our transcription analysis, PbSERA1/mCherry and PbSERA2/mCherry were barely detectable during the mosquito stages (data not shown). A. Primary structure of Plasmodium SERA proteins with active site serines (SERAser). The central papain-like cysteine protease domains are boxed in grey. Overall amino acid sequence identities of the P. berghei SERA1 (EU917224) and SERA2 (EU917225) sequences are indicated as percentages of identical residues compared with the P. falciparum SERA5 (PFB0340c). Fragments used for protein expression, purification and antibody production are indicated by bars. B. Partial conservation of the catalytic residues of the papain family within the central serine protease domain. P. berghei SERA1 and SERA2 and P. falciparum SERA5 are shown together with papain and a P. berghei cysteine family SERA protein, PbECP1/SERA5. The strictly conserved amino acid residues are boxed in grey, and the putative active-site serine or cysteine is shown in bold and boxed in black. Additional residues of the active site are marked with an 'o'. Note that the catalytic histidine is replaced by a methionine. In contrast, the carboxy-terminal asparagine and the amino-terminal glutamine, which form the oxyanione hole, are conserved.
We next investigated expression of mCherry-tagged SERAs during liver stage development in vitro. As observed with blood stages, the fusion proteins were not detected in early liver stages, but were abundantly expressed in late liver stages ( Fig. 3C and D). Remarkably, PbSERA1/mCherry and PbSERA2/mCherry showed clearly distinct expression patterns in liver stages. PbSERA1/mCherry was detected in mid and late liver stages, and localized predominantly to the PV, which constitutes the parasite/host interface (Fig. 3C). In contrast, PbSERA2/mCherry became detectable only at the end of liver stage development, with an intracellular distribution in the parasite (Fig. 3D). Interestingly, PbSERA1/mCherry was not detected in merosomes, in contrast to PbSERA2/ mCherry, which gave a strong signal associated with individual merozoites inside merosomes (Fig. 3E). Collectively, these data indicate that both PbSERA1 and PbSERA2 are expressed in late blood and liver stages. Importantly, the two proteins apparently distribute to distinct compartments, and only PbSERA2 remains associated with merozoites after PV membrane (PVM) rupture.
To get further insights into the distribution of PbSERAser proteins, we performed immunofluorescence analysis of late liver stages, using antibodies generated against the C-terminus of PbSERA1 (anti-SERA1C) or the central domain of PbSERA2 (anti-SERA2M) respectively (Fig. 1A). In late liver stages, A. RT-PCR analysis of SERA1 and SERA2 expression in mosquito midgut (mg) and salivary gland (sg) sporozoites, mid (24 h) and late (48 h, 72 h) liver stages, liver stage merosomes (mer) and blood stage merozoites (BS mer). The merozoite and sporozoite-specific transcripts, MSP1 and TRAP, and the constitutive GAPDH transcript were added as controls. cDNAs were synthesized from mRNA in the presence (+) or absence (-) of reverse transcriptase. Genomic DNA (gDNA) was added as an amplification control. B. Quantitative RT-PCR analysis of P. berghei SERA gene expression in WT mixed blood stages (BS), purified schizonts (late BS) or infected HuH7 cell cultures 65 h post-infection (late LS). Relative gene expression was normalized to MSP1 expression level, and is shown as the mean of two independent experiments (ϮSD). staining with anti-SERA1C antibodies was mostly restricted to the PVM, as shown by colocalization with the PVM marker exported protein 1 (EXP1) (Fig. 4A, upper panels). This pattern was also observed at more advanced stages of development, in cytomeres and fully differentiated merozoite-containing parasites (Fig. 4A, middle and lower panels). This distribution is reminiscent of the fluorescence pattern observed in parasites harbouring a mCherry tag at the C-terminus of PbSERA1 (Fig. 3C). In contrast, anti-SERA2M antibodies showed a more complex distribution in late liver schizonts and cytomeres, staining both the PVM and more internal structures (Fig. 4B, upper and middle panels).
Interestingly, in terminal liver stages, just prior to the release of merozoites, SERA2M was detected at the periphery of the parasites, in the PV and/or PVM compartment, as well as the host cell cytoplasm (Fig. 4B, lower panels). This distribution differs from that of the C-terminal fragment of SERA2, based on the fluorescence pattern in PbSERA2/mCherry parasites (Fig. 3D). Together, these results strongly suggest that SERA2 is processed towards the end of liver stage development, with the central puta-tive papain-like domain being released in the vacuolar space, while the C-terminal domain (visualized by the mCherry tag) remains associated with the merozoites. In good agreement with this hypothesis, Western blot analysis of purified blood stage schizonts demonstrated that SERA2 processed forms were enriched in saponin extracts, as compared with Triton X-100 ( Fig. S2), consistent with localization to the PV compartment of cleaved products that contain the central domain recognized by the anti-PbSERA2M antibodies.

Generation of PbSERA1 and PbSERA2 knockout parasites
We next wanted to study the cellular functions of PbSERA1 and PbSERA2 in the Plasmodium life cycle. Based on the refractoriness to gene knockout of PfSERA5 (McCoubrie et al., 2007), we expected that at least one member of the P. berghei SERAser family plays a vital role during asexual growth and therefore cannot be targeted by classical reverse genetics. We generated replacement vectors that were designed to disrupt the corresponding A. Insertion strategy to generate the SERA1/mCherry and SERA2/mCherry parasites. The PbSERA1 and PbSERA2 genomic loci were targeted with integration plasmids containing the 3′ SERA1 and SERA2 terminal fragments (dark grey box) that is fused in frame to the mCherry coding sequence (red box), the 3′ UTR of PbDHFR/TS (light grey box) and the TgDHFR/TS selectable marker (white box). Upon a single cross-over event, the region of homology is duplicated, resulting in a functional, endogenous PbSERA1 or PbSERA2 copy tagged with mCherry, followed by a truncated and non-expressed copy. B. Expression of the mCherry fusion proteins (red) was analysed by confocal fluorescence microscopy of SERA1/mCherry and SERA2/mCherry P. berghei blood stage parasites constitutively expressing GFP (green). Parasite stages are indicated by arrow heads. T, ring/trophozoite; S, schizont. Nuclei were stained with Hoechst 33342. Bars, 5 mm. C and D. Liver stage expression of the mCherry fusion proteins (red) was analysed by confocal fluorescence microscopy of SERA1/mCherry (C) and SERA2/mCherry (D) P. berghei parasites constitutively expressing GFP (green), at 24, 48 and 65 h after infection of HuH7 cells with sporozoites. Bars, 10 mm. open reading frames by double homologous recombination after transfection into P. berghei parasites (Fig. 5A). Unexpectedly, we could select recombinant parasite populations with the antifolate pyrimethamine. Subse-quent cloning of single parasites resulted in multiple clonal sera1(-) and sera2(-) parasite lines. Genotyping by PCR using specific primer combinations confirmed the expected recombination events in sera1(-) and sera2(-) parasite lines (Fig. 5C), and RT-PCR demonstrated the complete absence of SERA1 transcripts and SERA2 transcripts in sera1(-) and sera2(-) blood stage parasites respectively (Fig. 5D). Furthermore, Western blot analysis of purified blood schizonts demonstrated the absence of PbSERA1 and PbSERA2 proteins in sera1(-) and sera2(-) blood stage parasites respectively (Fig. 6). Collectively, these results confirm the successful disruption of the corresponding gene.
PbSERA1 and PbSERA2 are dispensable during the parasite life cycle We then analysed the phenotypes of sera1(-) and sera2(-) parasites. The successful disruption of SERA1 and SERA2 genes in P. berghei blood stages indicated that both are dispensable during blood stage multiplication of the parasite. Both sera1(-) and sera2(-) parasites produced gametocytes and exflagellation of male gametocytes was similar to WT parasites (data not shown). Transmission to Anopheles stephensi mosquitoes and oocyst development were also normal when compared with WT parasites (Table 1). Both sera1(-) and sera2(-) oocysts produced sporozoites, which invaded mosquito salivary glands as efficiently as WT parasites (Table 1). These findings demonstrate that SERA1 and SERA2 are dispensable for the P. berghei life cycle in the mosquito vector, in good agreement with the absence of gene expression in the mosquito stages ( Fig. 2A). Sporozoites from sera1(-) and sera2(-) parasites displayed normal gliding motility (Table 1), and were infective to rats (Table 2). Importantly, after intravenous injection of sporozoites or administration through mosquito bites, the natural transmission route, we observed no delay in patency as compared with WT parasites (Table 2). This clearly shows that hepatic merozoites are formed and released normally in sera1(-) and sera2(-) P. berghei parasites. This was confirmed by in vitro experiments, were determined using specific rat antibodies (red). As a marker for the PVM, we used chicken anti-EXP1 antibodies (green), and DNA was labelled with DAPI (blue). Scale bar: 5 mm. Note the distribution of SERA1C to the PV compartment in all stages, whereas SERA2M localizes both to the PV and internal parasite structures in schizonts and cytomeres, but only to the PV compartment after merozoite differentiation. which demonstrated that sera1(-) and sera2(-) parasites form exoerythrocytic forms (EEFs) in cultures, in numbers comparable with WT parasites (Table 1). Together, these findings demonstrate that individual SERAser are dispensable during the P. berghei life cycle.

P. berghei lacking both SERAser progress normally through the parasite life cycle
We hypothesized that compensatory mechanisms within the SERAser subfamily may explain the absence of phe-Plasmodium berghei serine-type SERAs 731 notypical defect in single gene disruptants. Therefore, we generated parasite lines with a double PbSERA1 and PbSERA2 gene deletion. To this end, we used a replacement vector containing the 5′ region of PbSERA1, the selection cassette and the 3′ region of PbSERA2 (Fig. 5B). Remarkably, after transfection of P. berghei parasites with this targeting construct, we could select and isolate sera1(-)/2(-) parasite populations. Genotyping by PCR using specific primer combinations confirmed the expected recombination events (Fig. 5C). Further-more, RT-PCR and Western blot analysis demonstrated the complete absence of SERA1 and SERA2 transcripts (Fig. 5D) and proteins (Fig. 6), respectively, in sera1(-)/ 2(-) blood stage parasites, confirming the simultaneous disruption of both SERA1 and SERA2 genes. As expected, immunofluorescence analysis confirmed the absence of staining of sera1(-)/2(-) late liver stages with anti-SERA1 and anti-SERA2 antibodies, which also demonstrates the specificity of the staining pattern observed in WT parasites with these antibodies (Fig. S3 and Fig. 4). Fig. 6. Western blot analysis of PbSERA1 and PbSERA2 protein expression in P. berghei blood stages. Triton X-100 extracts of enriched schizont preparations from WT, sera1(-), sera2(-) and sera1(-)2(-) parasites were analysed by Western blot using antibodies specific for PbSERA1 and PbSERA2. Anti-HSP70 antibodies were used as a loading control. Note the absence of specific bands (arrows) in the knockout parasite lines, confirming complete gene disruption. Non-specific bands are indicated with an asterisk. a. Gliding motility was visualized by CSP labelling of salivary gland sporozoites on glass slides, and motile sporozoites were counted using a fluorescence microscope. ++, continuous, multiple trails in more than 30% of sporozoites.
As observed with the single gene mutants, sera1(-)/ 2(-) displayed no obvious defect during asexual blood stage growth (Fig. 7) and sexual stage differentiation (data not shown). The double disruptants could be transmitted to mosquitoes, where parasite development was not affected (Table 1). sera1(-)/2(-) parasites produced normal numbers of sporozoites (Table 1), which were motile and as infective to rodents as WT parasites ( Table 2). As observed with single gene disruptants, sera1(-)/2(-) parasites formed EEFs in vitro, in similar numbers as WT (Table 1). Furthermore, the number of detached infected cells released in culture supernatants was similar between WT (473 Ϯ 200 merosomes per well) and sera1(-)/2(-) parasites (433 Ϯ 250 merosomes per well), and mice injected with 500 WT merosomes (n=3) or sera1(-)/2(-) merosomes (n=3) all developed a patent blood stage infection at day 2 post injection. Together, these data establish that in the absence of SERAser, P. berghei parasites progress normally through their life cycle. In particular, parasites lacking SERA1 and SERA2 display no detectable defect in egress from infected host cells.

P. berghei lacking SERAser display increased expression of the cysteine-type SERA3
Finally, we analysed the impact of SERAser gene disruption on the expression of the cysteine-type SERAs PbSERA3 and PbSERA4. Quantitative PCR on cDNA prepared from purified blood schizonts showed a modest reduction of SERA1 transcript levels in sera2(-) parasites, whereas SERA2 transcript levels were not significantly modified in sera1(-) parasites. In sharp contrast, in both sera1(-) and sera2(-) parasites there was an upregulation of PbSERA3, but not SERA4, expression (Fig. 8A). The increase in PbSERA3 transcript levels was even more pronounced in parasites lacking both SERAser (Fig. 8A). To confirm a compensatory upregulation of PbSERA3 in the SERAser loss-of-function parasite lines we analysed the protein levels in merozoites (Fig. 8B). PbSERA3 was detected mainly in saponin extracts, confirming its localization to the PV compartment in blood stages, as reported previously (Schmidt-Christensen et al., 2008). In good agreement with the transcript profiling we detected an increase in PbSERA3 protein by 2and 2.5-fold in sera1(-) and sera1(-)/2(-) parasites respectively. These results indicate a potential functional link between distinct SERA subfamilies, and provide a possible explanation for the absence of phenotypical defect in PbSERAser-deficient parasites.

Discussion
The most important finding from our study is the nonessential role of the SERAser subfamily in vivo in the model rodent malaria parasite P. berghei. We successfully deleted the single genes of SERA1, SERA2 and generated a SERA1/2 double knockout. Both single mutants and the double mutant showed no apparent defect at any phase of the Plasmodium life cycle under standard conditions. Most importantly, the successful generation of the double mutant demonstrates that redundancy is not an essential feature of the SERAser family in vivo. This finding was unexpected, because SERAser members are abundantly expressed in asexual blood stages (Aoki et al., 2002;Miller et al., 2002) and at least one member  (-) 10 000 3/3 3.3 sera2 (-) 10 000 2/2 3.5 25 000 2/2 3.5 50 000 1/1 3.0 Mosquito bite 1/1 3.0 sera1(-)/2(-) 10 000 5/5 3.4 a. Sporozoites were either injected intravenously at the doses indicated or delivered by natural mosquito bite via exposure of anesthetized SD rats to 5 infected Anopheles mosquitoes. b. Prepatency is the time until the first detection of an erythrocytic stage parasite in Giemsa-stained thin blood smears after sporozoite infection. Fig. 7. Blood stage growth of sera1(-)/2(-) double knockout P. berghei parasites. Naïve NMRI mice (n = 5) were injected intravenously with 1000 wild-type or 1000 sera1(-)/2(-) P. berghei infected erythrocytes. Infection was then monitored by daily examination of Giemsa-stained blood smears to determine the parasitemia.
of the P. falciparum SERAser family, SERA5, appears to be refractory to targeted gene deletion under in vitro culture conditions McCoubrie et al., 2007). Our findings suggest that the monophyletic group of SERAser enzymes serves a specialized and nonessential role across the genus Plasmodium. Our experimental genetics evidence is supported by the SERA phylogeny (Arisue et al., 2007;McCoubrie et al., 2007). Based on available genome sequence data, the avian parasite P. gallinaceum encodes only for SERAcys members and appears to have diverged from a common Plasmodium ancestor prior to a gene duplication event that formed the two separate phylogenetic clusters. In good agreement, we now show that the rodent malaria parasite does not rely on SERAser genes for parasite life cycle progression. We propose that members of this monophyletic group evolved gradually and independent of strong selection pressure for parasite growth. However, although SERAser do not play any vital function in rodent malaria parasites, we cannot exclude a role of immunological pressure in evolution of this family, including its expansion in the human parasites. Notably, one important feature distinguishes the P. falciparum SERAser proteins from their other relatives (Hodder et al., 2003). In their central domain they retain a histidine residue that is thought to be part of the catalytic triad. This residue is generally substituted by leucine in P. vivax and P. knowlesi SERAser proteins, except for P. vivax SERA2/P. knowlesi SERA5, and by methionine in the rodent Plasmodium species. Hence, we currently cannot formally exclude distinct roles of the histidinecontaining P. falciparum proteins, although their in vitro proteolytic activity is weak (Hodder et al., 2003). This possibility can now be addressed experimentally by generating a trans-species complementation of the PfSERA5 deletion that consistently appears to be required for P. falciparum growth under in vitro culture conditions McCoubrie et al., 2007). If complementation with PbSERA1 or PbSERA2 rescues the growth defect, an essential role for the histidine residue, and perhaps the PfSERA5 gene as a whole, can be excluded in vivo, and vice versa. Because of the absence of an apparent phenotype in the PbSERA1(-)/2(-) double mutant, the reverse experiment, i.e. complementation with PfSERA5, would not be informative. Although indirect, such an approach will eventually substantiate an important in vivo role for PfSERA5.
Our expression and localization studies with specific antibodies and fluorescently tagged parasite lines show that PbSERA1 and PbSERA2 have distinct expression patterns, both in terms of expression timing and subcellular localization. After proteolytic processing, N-terminal and C-terminal parts of SERA proteins remain covalently linked, while the central domain is released . The functions of the SERA N-and C-terminal domains remain unknown. It should be noted that the Nand C-terminal regions of PbSERA1 and PbSERA2 show only 34% and 29% amino acid sequence identity, respectively, whereas the central protease-like domain is more conserved (67% identity). Because in our mCherry constructs the fluorescent moiety is fused to the C-terminal end of SERA1 and SERA2, the fusion proteins only allow tracking of the C-terminal fragments of the corresponding SERA, and not the papain-like central domain after proteolytic processing.
Interestingly, by combining C-terminal tagging of PbSERA2 and antibodies specific for the central domain of PbSERA2, we found that the two domains localize to A. SERA gene expression was analysed by quantitative RT-PCR in purified schizonts from WT, sera1(-), sera2(-) and sera1(-)/2(-) parasites. Relative gene expression was normalized to MSP1 expression level, and is shown as the Log2 of the ratio knockout/WT (mean of two experiments Ϯ SD). n.a., no amplification. B. Saponin (S) or Triton X-100 (T) extracts of purified schizonts from WT, sera1(-), sera2(-) and sera1(-)/2(-) parasites were analysed by Western blot using antibodies specific for PbSERA3. As a loading control an anti-HSP70 antibody was used. distinct compartments after formation of liver stage merozoites. While the C-terminal region of SERA2 remains associated with merozoites, the central putative papainlike domain localizes predominantly to the vacuolar compartment. This was also evidenced by differential Western blot analysis of vacuolar versus parasite protein fractions, which showed preferential accumulation of SERA2 processed forms (containing the central domain) in the vacuolar compartment of asexual blood stage parasites. This suggests that PbSERA2 is proteolytically cleaved at the end of Plasmodium schizogony, the central domain being released in the vacuolar space while the C-terminus remains associated with merozoites.
Differently from PbSERA1/mCherry, which was no longer detected after PVM rupture, PbSERA2/mCherry was clearly found associated with liver and blood stage merozoites. This is reminiscent of PfSERA5 localization on the surface of free merozoites after processing (Pang et al., 1999;Okitsu et al., 2007). Although the different expression patterns of PbSERA1 and PbSERA2 might be explained by differential stability of their N-and C-terminal domains after proteolytic processing, our observations raise the possibility that different SERAs fulfill different functions. PbSERA1 and PbSERA2 are particularly abundant in the final stages of liver and erythrocytic schizont maturation, which is compatible with a potential role in merozoite egress. Based on their differential localization, the two P. berghei SERAser may act at different steps of egress. In such a scenario, PbSERA1, which localizes predominantly in the PV, could be involved in the rupture of the PVM. The PbSERA2 C-terminal domain, which remains associated with merozoites after the PVM rupture, may rather play a role during subsequent steps, such as rupture of the host cell membrane or preparation of merozoites for invasion. Nevertheless, such roles, if any, can only be auxiliary, because parasites that lack SERA1 and SERA2 display no defects in egress of liver and blood stage merozoites, or natural malaria transmission in general.
So far, only PbSERA5/ECP1 (the orthologue of PfSERA8) was demonstrated to play a role during parasite egress. Interestingly, PbSERA5/PfSERA8 stands apart in the SERA family as it is expressed only in the mosquito stages and is the only SERA that apparently lacks a SUB1 cleavage site (Yeoh et al., 2007). In this regard, it is noteworthy that inhibitors of P. falciparum SUB1, which cleaves (and presumably activates) SERAs, are more potent inhibitors of merozoite invasion than of merozoite egress (Yeoh et al., 2007). While this observation could be due to different pharmacodynamic properties of the inhibitors, it is also compatible with a predominant role of SERAs during merozoite invasion of erythrocytes rather than parasite egress. The potential use of several invasion pathways may result in compen-satory mechanisms explaining why single (and double) knockout parasites display no obvious defect during their progression through the life cycle. Because of the absence of detectable phenotype in parasites lacking SERAser genes, the loss-of-function approach cannot discriminate between a role of SERAser proteins during egress versus merozoite invasion. Gain-of-function approaches may constitute an alternative strategy, although it is probably difficult to overexpress SERAser, which are already highly expressed in normal parasites. Other strategies such as using blocking antibodies may help unraveling the function of the SERAser proteins.
Interestingly, we found that P. berghei parasites lacking both SERAser had increased levels of the cysteine-type SERA3, both at transcript and protein levels. Although we cannot formally exclude an effect of the modification in cis of the SERA gene locus in the knockout parasite lines, this observation suggests a potential functional link between SERAs belonging to distinct subfamilies. It has been shown before that P. falciparum parasites with a deletion of PfSERA4 gene display increased RNA levels of PfSERA5, which is refractory to gene deletion (McCoubrie et al., 2007). Similarly, attempts to knockout SERA3 gene in P. berghei remained unsuccessful so far (E.D. Putrianti, F.F. Masduki, and K. Matuschewski, unpubl. data), suggesting that PbSERA3 may play an essential role during P. berghei blood stage infection.
The crystal structure of the central protease-like domain of PfSERA5 was recently solved, revealing several anomalies in the active site, in addition to the serine substitution (Hodder et al., 2009). These structural features question the role of SERA5 as an actual protease, and may explain the limited proteolytic activity of recombinant PfSERA5 in vitro (Hodder et al., 2003). In the absence of a clear proteolytic activity of SERAser proteins under physiological conditions, multiple non-catalytic cellular roles, including regulatory functions, have to be considered. An attractive hypothesis is a potential function in substrate recognition. In analogy to ubiquitin E2-like variants of the ubiquitin/proteasome pathways that lack active site cysteine residues and form heterodimers with E2 enzymes (VanDemark et al., 2001), SERAser, while catalytically inactive on their own, could act in concert with canonical SERAcys proteases and provide crucial substrate binding sites. In such a scenario, the diverse group of abundant and degenerate SERAser proteins would bind and recruit substrates to the proteases, thereby enhancing their overall proteolytic activity in the cell. While not essential, the presence of multiple substrate recognition proteins may greatly enhance the adjustment of the parasite to a changing environment prior to parasite exit of the respective host cell.
Importantly, a non-vital role in vivo does not exclude the potential of SERAser members for antimalaria subunit Plasmodium berghei serine-type SERAs 735 vaccine development. There is precedence for a nonessential target antigen, namely MSP3 (Mills et al., 2002), which shows similar promising characteristics in a monkey challenge trial and functional assays with sera from immunized individuals (Hisaeda et al., 2002;Druilhe et al., 2005). The choice of candidate antigens for incorporation into a subunit vaccine does not necessitate an essential role in the parasite but should rather be based on functional assays (Matuschewski, 2006). Antibodies against PfSERA5 inhibit parasite erythrocytic growth in vitro through agglutination of merozoites and ruptured schizonts (Pang et al., 1999). Similarly to PfSERA5, the C-terminal region of PbSERA2 associates with merozoites, and therefore constitutes a potential target for inhibitory antibodies. In this regard, P. berghei mouse infection may thus represent a valuable in vivo model to evaluate antimalarial vaccines targeting SERA antigens. In contrast, target validation by reverse genetics to ultimately prove or disprove that loss of gene function results in non-viable malaria parasites in vivo is a prerequisite for preclinical development of tailor-made inhibitors against a target protein. Our results show that parasites grow normally in the absence of any SERAser members and cast profound doubt on the suitability to translate potential specific PfSERA5 or other SERAser inhibitors into potent antimalarial drugs.

Experimental animals
Animals were from Charles River Laboratories. All animal work was conducted in accordance with European regulations and approved by the state authorities (Regierungspräsidium Karlsruhe).

Generation of the mCherry-tagged SERA1 and SERA2 parasite lines
For targeted fluorescent tagging of SERA1 and SERA2, an integration vector was generated by amplification of a PCR fragment using P. berghei genomic DNA as template and primers mCherry-SERA1for (5′-ATAAGAATGCGGCCGCTACCACATG AGAATGAATTTGCAGGG-3′; NotI site is underlined) and mCherry-SERA1rev (5′-GGACTAGTCACATAACAAAAGTAGCA ATCGTCTG-3′; SpeI site is underlined), or mCherry-SERA2for (5′-ATAAGAATGCGGCCGCATGTTGGTGATTCATGCCCCG-3′; NotI site is underlined) and mCherry-SERA2rev (5′-TGC TCTAGATACAGCGCAAAAGTTACATTCATTATCACC-3′; XbaI site is underlined). Cloning into the P. berghei transfection vector that contained the mCherry sequence and PbDHFR/TS 3′ UTR resulted in plasmids pEDP05 and pEDP06, for tagging of PbSERA1 and PbSERA2 respectively. The targeting plasmids were linearized with HpaI and AarI respectively, and parasite transfection, positive selection, and parasite cloning was performed as described previously (Janse et al., 2006). Integrationspecific PCR amplification of the mCherry-tagged SERA1 or SERA2 was generated using specific primer combinations. We obtained one parasite population each that was used for a systematic expression and localization analysis. Expression of the mCherry fusion proteins was analysed through direct detection of the red fluorescence of mCherry by confocal microscopy. Hoechst 33342 (Molecular Probes) was used to stain nuclei. Images were acquired on a Zeiss LSM510 confocal system (Zeiss, Germany) equipped with visible and UV laser lines, and processed with Adobe Photoshop software (Adobe Systems).

Generation of anti-PbSERA1 and anti-SERA2 antibodies
DNA fragments corresponding to the coding sequence of PbSERA1 C-terminal region (Leu 877 -Ser 1122 ) and PbSERA2 central (M) domain (Lys 546 -Pro 714 ) were amplified from P. berghei cDNA by RT-PCR and cloned into pGEX6P-1 vector (Amersham, Buckinghampshire, England). Recombinant proteins were expressed in Escherichia coli BL21 cells (Stratagene) as glutathione S-transferase (GST) fusion proteins, and purified using glutathione-agarose as described by the manufacturer (Amersham Biosciences). Purified proteins were used to immunize Lewis rats along with complete Freund adjuvant, followed by multiple boosting immunizations. A rat monoclonal antibody was generated against PbSERA1-C. To this purpose, B cells were isolated from lymph nodes of one rat immunized against PbSERA1-C protein and fused to the mouse myeloma cell line P3X63.Ag8.653. The positive pools of hybridoma cells reacting with SERA1-C were screened by indirect ELISA. Single-cell clones were isolated by limited dilution, leading to the isolation of clone C65 (SERA1C), which was characterized further and found to bind an epitope (D892-EPASISTQ-E901) at the C-terminus of PbSERA1.

Western blot analysis
Parasite protein extracts were obtained from P. berghei Nycodenz-enriched blood stage schizont preparations, after lysis in saponin followed by Triton X-100, to differentiate the vacuolar compartment from the parasite fraction, as described (Schmidt-Christensen et al., 2008). Proteins were separated on 10% SDS-PAGE reducing gels and transferred to PVDF membranes (Amersham). Membranes were probed with anti-SERA1 and anti-SERA2 rat antibodies, anti-SERA3 mouse antibodies (Schmidt-Christensen et al., 2008) or anti-HSP70 mouse antibodies (Tsuji et al., 1994). Horseradish peroxidase-conjugated goat anti-rat or anti-mouse antibodies (Sigma) were used for detection, and bands were visualized by enhanced chemiluminescence (Amersham).

Immunofluorescence assay
For analysis of PbSERAser localization in late liver stages, infected HepG2 cells were fixed with 4% paraformaldehyde, permeabilized with ice-cold methanol and incubated with primary antibodies against PbSERA1C or PbSERA2M (rat). A chicken anti-EXP1 antibody was used to stain the PVM. Bound antibodies were detected using anti-rat Alexa Fluor 594-or anti-chicken Cy5-conjugated secondary antibodies (Molecular Probes, Leiden, the Netherlands). Nuclei were visualized with DAPI (Sigma-Aldrich, Germany). Immunofluorescence labelled cells were examined by confocal microscopy using the Olympus FV1000 (SIM scanner and spectral detection).

Phenotypical analysis during the Plasmodium life cycle in vivo
Blood stage development was analysed in vivo in asynchronous infections using NMRI mice. Gametocyte differentiation and exflagellation of microgametes were detected in mice before mosquito feedings. Anopheles stephensi mosquito rearing and maintenance was carried out under a 14 h light/10 h dark cycle, 75% humidity and at 28°C or 20°C respectively. Sporozoite populations were separated and analysed as described previously (Vanderberg, 1975). For determination of sporozoite infectivity, and numbers of midgut-and salivary gland-associated sporozoites, infected mosquitoes were dissected at days 10, 14 and 17 after feeding respectively. For determination of the infectivity of sporozoites, infected mosquitoes were dissected at day 17 after feeding. Sporozoites were liberated from salivary glands and injected intravenously at the numbers indicated into young Sprague/Dawley (SD) rats. Patency was checked daily by Giemsa-stained blood smears.

In vitro experiments
For analysis of gliding motility, sporozoites isolated from infected mosquito salivary glands were deposited on glass slides coated with bovine serum albumin, and incubated at 37°C for 30 min. Trails left behind gliding parasites were then visualized using anti-CSP antibodies (Potocnjak et al., 1980). For analysis of EEF development, we used HuH7 cells cultured in DMEM supplemented with 10% FCS and antibiotics. P. berghei sporozoites were added in triplicate wells, incubated for 2 h at 37°C, and washed off. After 24 or 48 h, EEFs were revealed using primary antibodies against Plasmodium heat shock protein 70 (HSP70) (Tsuji et al., 1994).  S1. Genotyping of PbSERA1/mCherry (A) and PbSERA2/ mCherry (B) clonal parasites. Confirmation of the predicted integration events by diagnostic PCR fragments, tests 1-4 (see Fig. 3A). Absence of the WT signal confirmed purity of the clonal parasite lines. Fig. S2. Processed forms of PbSERA2 localize preferentially to the PV compartment. Purified schizonts from WT P. berghei parasites were lysed in saponin (PV fraction), followed by Triton X-100 (parasite fraction). Protein extracts were analysed by Western blot using rat antibodies specific for the central domain of PbSERA2. Processed forms of SERA2 enriched in the PV fraction are indicated with an arrowhead, whereas other specific bands are shown with an arrow. Fig. S3. Immunofluorescence analysis confirms the absence of SERA1 and SERA2 expression in P. berghei sera1(-)/2(-) parasites. HepG2 cells infected with P. berghei WT or sera1(-)/2(-) parasites at the cytomere stage were stained with antibodies against PbSERA1 (A) or PbSERA2 (B). PVM was stained with anti-EXP1 antibodies (cyan), and DNA was labelled with DAPI (blue). Scale bar: 5 mm.
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.