TgSUB2 is a Toxoplasma gondii rhoptry organelle processing proteinase


  • Steven A. Miller,

    1. Departments of Medicine and of Microbiology and Immunology, Albert Einstein College of Medicine, Bronx NY 10461, USA.
    Search for more papers by this author
    • S.A.M. and V.T. contributed equally to this work.

  • Vandana Thathy,

    1. Departments of Medicine and of Microbiology and Immunology, Albert Einstein College of Medicine, Bronx NY 10461, USA.
    Search for more papers by this author
    • S.A.M. and V.T. contributed equally to this work.

  • James W. Ajioka,

    1. Department of Pathology, Cambridge University, Cambridge, UK.
    Search for more papers by this author
  • Michael J. Blackman,

    1. Division of Parasitology, National Institute for Medical Research, Mill Hill, London NW7 1AA, UK.
    Search for more papers by this author
  • Kami Kim

    Corresponding author
    1. Departments of Medicine and of Microbiology and Immunology, Albert Einstein College of Medicine, Bronx NY 10461, USA.
      E-mail; Tel. (+1) 718 430 2611; Fax (+1) 718 430 8968.
    Search for more papers by this author

E-mail; Tel. (+1) 718 430 2611; Fax (+1) 718 430 8968.


All parasites in the phylum Apicomplexa, including Toxoplasma gondii and Plasmodium falciparum, contain rhoptries, specialized secretory organelles whose contents are thought to be essential for successful invasion of host cells. Serine proteinase inhibitors have been reported to block host cell invasion by both T. gondii and P. falciparum. We describe the cloning and characterization of TgSUB2, a subtilisin-like serine proteinase, from T. gondii. Like its closest homologue P. falciparum PfSUB-2, TgSUB2 is predicted to be a type I transmembrane protein. Disruption of TgSUB2 was unsuccessful implying that TgSUB2 is an essential gene. TgSUB2 undergoes autocatalytic processing as it traffics through the secretory pathway. TgSUB2 localizes to rhoptries and associates with rhoptry protein ROP1, a potential substrate. A sequence within TgSUB2 with homology to the ROP1 cleavage site (after Glu) was identified and mutated by site-directed mutagenesis. This mutation abolished TgSUB2 autoprocessing suggesting that TgSUB2 is a rhoptry protein maturase with similar specificity to the ROP1 maturase. Processing of secretory organelle contents appears to be ubiquitous among the Apicomplexa. As subtilases are present in genomes of all the Apicomplexa sequenced to date, subtilases may represent a novel chemotherapeutic target.


The protozoan parasite Toxoplasma gondii commonly infects humans and other mammals, causing clinical disease in immunocompromised individuals and children infected in utero. After ingestion of environmentally resistant oocysts or tissue cysts, the tachyzoite form of the parasite undergoes multiple rounds of rapid replication before the host is able to mount an immune response to control the infection. T. gondii is an obligate intracellular pathogen, replicating only within a specialized parasitophorous vacuole formed in the cytoplasm of nearly all cell types. One means of developing new strategies to treat this infection involves targeting critical events in the tachyzoite life cycle, including host cell invasion, scavenging of essential metabolites, and host cell lysis and release of invasive parasites (Wastling et al., 2000).

The process of host cell invasion used by Toxoplasma appears to be well conserved across the phylum of Apicomplexan parasites, including Cryptosporidium and Plasmodium species, the agents of malaria. Invasive stages contain specialized apical organelles, called micronemes and rhoptries whose secretion is coupled to successful host cell invasion (Carruthers et al., 1999). Micronemes are small apical organelles containing several adhesin proteins and secrete during attachment to the host plasma membrane. Rhoptries are unique elongated organelles whose secretion is coincident with formation of the nascent parasitophorous vacuole. Once safely inside the host cell, the parasite secretes proteins that modify the vacuole membrane from organelles known as dense granules (Sibley et al., 1995; Carruthers and Sibley, 1997). The proteins and events associated with this sequence of regulated secretion offer insights into the biology of these parasites and potential means for control of infection (Carruthers, 1999).

Due to the importance of these specialized secretory organelles for T. gondii development, much attention has been directed towards the identification and function of their protein contents. One common theme emerging from these studies is that proteolytic processing plays a central role in the maturation of many of these proteins, both during trafficking through the secretory pathway and upon secretion from the parasite. Cleavage of proteins destined for rhoptry compartmentalization during biogenesis of the organelle is seen throughout the phylum Apicomplexa (Sam-Yellowe, 1996).

Initial characterization of rhoptry proteins in T. gondii focused upon ROP1 (Saffer et al., 1992). ROP1, or penetration enhancement factor, was originally identified as a protein implicated in invasion. Subsequent studies revealed that ROP1 is not essential (Kim et al., 1993), but most studies to characterize rhoptry protein trafficking have been performed with ROP1. In T. gondii an amino terminal fragment from ROP1 is removed within or in close proximity to nascent rhoptries (Soldati et al., 1998). Other rhoptry family members ROP2, ROP3 and ROP4 also contain a propeptide of similar size that is removed intracellularly (Sadak et al., 1988). Several microneme proteins are also processed within the secretory pathway. For most rhoptry and microneme proteins, the functional significance of their cleavage events is unknown, but their ubiquitous presence suggests a biological role for processing.

Subtilisin-like serine proteinases have potential importance in Apicomplexan biology, as inhibitor studies have shown that serine proteinase activity is necessary for successful invasion of host cells (Conseil et al., 1999; Blackman, 2000). The first subtilases identified in Apicomplexan parasites were Plasmodium falciparum enzymes PfSUB-1 and PfSUB-2 (Blackman et al., 1998; Barale et al., 1999; Hackett et al., 1999). Both are localized to merozoite dense granules. PfSUB-2 has been proposed to be a maturase for merozoite surface protein 1 (MSP-1) whose processing is critical for entry into host erythrocytes but this hypothesis has not yet been proven (Barale et al., 1999; Blackman, 2000). We have previously described the T. gondii subtilisin-like serine proteinase TgSUB1 (Miller et al., 2001) that is a homologue of PfSUB1 and Neospora caninum NcSUB1 (Louie and Conrad, 1999). TgSUB1 is a subtilase secreted from the parasite micronemes in a calcium-dependent manner (Miller et al., 2001). Although subtilases pose another potential novel chemotherapeutic target for the Apicomplexa, to date no biological substrates of these proteinases have been identified. In this report, we describe TgSUB2, a rhoptry processing subtilase of T. gondii that is probably essential and is a strong candidate to be the ROP1 processing proteinase.


Cloning of TgSUB2, a subtilisin proteinase

TgSUB2 was cloned by homology PCR with primers used previously to clone PfSUB-1 from P. falciparum (Blackman et al., 1998). Significant features of TgSUB2 are illustrated in Fig. 1. TgSUB2 has significant similarity to other subtilases within the 300 residues encompassing the protease catalytic domain (Fig. 2). Within the catalytic domain, TgSUB2 is more similar to PfSUB-2 and homologues from other malaria species (35% identical) than it is to TgSUB1 or NcSUB1 (30% identical). This region contains the highly conserved catalytic triad residues at D783, H836, S999 and oxanion hole N931 with locally conserved residues and expected spacing (Rawlings and Barrett, 1993). Five highly conserved glycine residues at space-limited sites (785, 837, 855, 899 and 997) are present in expected positions. There is an insertion of 20 amino acids relative to subtilisin BPN′ at the second extended β-sheet region (beginning at residue 43 of BPN), a feature commonly seen in members of the thermitase family of subtilases (Siezen and Leunissen, 1997). TgSUB2 contains the calcium-binding loop N849-V853, indicating that it is not a member of the proteinase K family (Siezen and Leunissen, 1997).

Figure 1.

TgSUB2 sequence. Nucleotide and deduced amino acid sequence of the composite TgSUB2 cDNA clone. Non-coding sequence is in lowercase and coding sequence is in uppercase. The putative signal peptide is underlined and the predicted transmembrane domain is double underlined. The catalytic triad residues D783, H836 and S999 are boxed. The TAG stop codon is indicated by an asterisk. The three potential autocatalytic cleavage sites that conform to the SΦXE motif are in grey boxes.

Figure 2.

Sequence comparison between Apicomplexan subtilisins. Multiple sequence alignment of the catalytic domains of Apicomplexa subtilisins TgSUB2 (GenBank™ Accession Number AF420596), PfSUB2 (GenBank™ Accession Number CAB43592), TgSUB1 (GenBank™ Accession Number AY043483), PfSUB1 (GenBank™ Accession Number CAA05261), NcSUB1 (GenBank™ Accession Number AAF04257) and B. amyloliquefaciens subtilisin BPN′ (SubBPN, GenBank™ Accession Number P00782) using the CLUSTAL method. The alignment begins at amino acid 28 of mature subtilisin BPN′. Residue numbering for each sequence is shown at the right. Sites with greater than 50% identity for all sequences are shaded. Catalytic triad residues aspartic acid, histidine and serine and oxanion hole residue asparagine are indicated by an asterisk.

Primary sequence alignment is significantly different from members of the kexin family, as is the DSG rather than DDG motif seen at the active site aspartic acid (Siezen and Leunissen, 1997). A predicted signal peptide cleavage site is present after G28, indicating likely involvement in the secretory pathway (Nielsen et al., 1997). The amino terminal extension of approximately 750 amino acids is relatively long for subtilases although there are other eukaryotic examples of propeptides of similar size. The extension of about 250 amino acids at the carboxy terminus is not unusual, and this region can confer new functionality onto the proteinase domain (Kim et al., 1997). The protein is largely hydrophilic, with the exception of internal regions of the core proteinase domain and residues 1178–1197, which are predicted to form a single transmembrane spanning segment near the carboxy terminus.

Southern blot analysis of RH strain genomic DNA and TgSUB2 BAC clones was performed with different cDNA probes designed to encompass the catalytic domain and 3′ regions of the gene. Hybridization to genomic DNA and BAC clones digested with a variety of restriction enzymes yielded identical results, with most identifying a single DNA fragment (data not shown). The TgSUB2 gene locus is 9.5 kb and contains 14 introns from 0.15 to 0.6 kb in length.

TgSUB1 and TgSUB2 are distinct tachyzoite proteins with minimal cross-reactivity

Mouse antiserum specific for the catalytic domain of TgSUB2 (MαTgSUB2) and rabbit antiserum to TgSUB2 C-terminal peptide (RαTgSUB2) were used to characterize expression of TgSUB2 in tachyzoites (Fig. 3A). Neither antiserum showed any reaction to host HFF cell lysate, but specifically bound to RH strain tachyzoite proteins (Fig. 3B). Three major species of Mr 140, 90 and 85 kDa were observed on Western blot with both antisera. Unlike TgSUB1 (Miller et al., 2001), TgSUB2 could not be detected in excreted-secreted antigens (ESA), proteins secreted from micronemes or dense granules.

Figure 3.

TgSUB2 is antigenically distinct from TgSUB1.
A. Schematic of TgSUB2 with the positions of the antigens used for MαTgSUB2 or RαTgSUB2. The signal peptide and transmembrane region are shown in black. The D, H, and S are the catalytic triad residues.
B. Western blot of tachyzoite cell lysate (RH) or HFF lysate probed with pre-immune mouse (RH lysate), MαTgSUB2, pre-immune rabbit (RH lysate), or RαTgSUB2. Pre-immune sera also did not recognize HFF proteins (data not shown).
C. Immunoprecipitation-Western of TgSUB1 and TgSUB2. Extracellular tachyzoites were lysed and immunoprecipitated with RαPfSUB1 (lanes 1, 4, 7), RαTgSUB2 (lanes 2, 5, 8) or MαTgSUB2 (lanes 3, 6, 9). Immunoprecipitated proteins were analysed by Western blot with RαPfSUB1 [(lanes 1, 2, 3; reactive with TgSUB1 (Miller et al., 2001)], RαTgSUB2 (lanes 4, 5, 6) or MαTgSUB2 (lanes 7, 8, 9). Positions of molecular weight standards (kDa) are indicated. Bands corresponding immunoglobulin heavy (HC) and light (LC) chains are indicated.

We analysed the specificity of the antisera by Western blot analysis of immunoprecipitated proteins (Fig. 3C). Proteins immunoprecipitated with RαPfSUB1, which recognizes TgSUB1 (Miller et al., 2001), are not recognized by RαTgSUB2 and MαTgSUB2 (Fig. 3C, lanes 4 and 7) although RαPfSUB1 reacts slightly with the 90 kDa band seen in RαTgSUB2 immunoprecipitates (Fig. 3C, lane 2).

TgSUB2 cannot be disrupted

We attempted to disrupt TgSUB2 to determine if invasion would be impaired. Screening of over 200 individual clones in three separate attempts using two different selectable markers, chloramphenical acetyl transferase (Kim et al., 1993) or hypoxanthine xanthine guanine phosphoribosyl transferase (Donald et al., 1996), failed to identify any disruptants of TgSUB2 (data not shown). Clones were screened by PCR, Western blot, and a subset by Southern. Because interruption of an essential gene will be lethal, our inability to identify any knockouts of TgSUB2 implies that TgSUB2 is essential.

TgSUB2 localizes to rhoptries

The subcellular localization of TgSUB2 was determined by immunoelectron microscopy (Fig. 4A). Labelling was primarily in the body of the rhoptries, elongated club-shaped organelles in T. gondii (15 nm gold). Occasional particles were also seen in rhoptry necks, closer to the apical complex or in vesicular structures that could represent the parasite secretory apparatus. The labelling of rhoptries was confirmed by co-localization of TgSUB2 with rhoptry protein ROP1 [10 nm gold; (Saffer et al., 1992)].

Figure 4.

TgSUB2 localizes to rhoptry organelles. Immunoelectron microscopy was performed on LR white embedded extracellular tachyzoites with affinity purified RαTgSUB2 (followed by GαRabbit IgG coupled to 15 nm gold beads) and mouse monoclonal Tg49 (specific for ROP1; followed by GαMouse IgG coupled to 10 nm gold beads). Both antibodies specifically label over the elongated rhoptry organelles of the parasite. The scale bar indicates 500 nm. The inset shows a higher magnification of the boxed region demonstrating large (15 nm; arrow) and small (10 nm; arrowhead) gold particles. The scale bar indicates 50 nm.

TgSUB2 is autocatalytically processed at the N terminus

TgSUB2 was not efficiently radiolabelled in short duration pulses. Therefore pulse chase experiments could not definitively establish whether the multiple proteins recognized by TgSUB2 antisera represented proteolytically processed versions of the same initial translation product (data not shown). Because TgSUB2 is highly interrupted, alternative splicing to generate separate proteins is possible. We therefore expressed a composite TgSUB2 cDNA clone tagged with the 9-amino-acid haemagglutinin epitope tag (HA) at the C terminus (Fig. 5A). Immunofluorescence of the SUB2-HA cassette driven by the ROP1 promoter revealed localization similar to ROP1 (Fig. 5B). Western blot of tachyzoites transfected with SUB2-HA was similar to that seen with native TgSUB2 (Fig. 5C) suggesting that the three major forms seen are not due to alternative splicing, but some other post-transcriptional event.

Figure 5.

TgSUB2 is autocatalytically processed and is a candidate rhoptry processing enzyme.
A. Schematic of TgSUB2 and constructs used for transfections. Domains of TgSUB2 are as illustrated in Fig. 3A. SUB2-HA was generated by adding an HA epitope tag to the C terminus of TgSUB2. Positions of the ROP1 cleavage site and hypothesized TgSUB2 cleavage sites are shown. The catalytically inactive mutant SUB2-HA S999A and P1 residue mutant SUB2-HA E686R mutation are also shown.
B. Double immunofluorescence of SUB2-HA. SUB2-HA driven by the ROP1 promoter was transiently transfected into RH strain tachyzoites. Twenty-hour hours after transfection, cells were labelled with ROP1 monoclonal antibody (Tg49; visualized with Texas Red goat anti-mouse IgG) and Rat α HA-FITC.
C. Tachyzoites were transiently transfected with no DNA (mock), pSUB2-HA, pSUB2-HA S999A, or pSUB2-HA E686R expression constructs driven by the GRA1 promoter. Western blot was performed with RαTgSUB2 or with RatαHA 48 h after transfection. Processed TgSUB2 migrating at 90 kDa (arrowhead) and 85 kDa (arrow) are indicated. Similar results were seen with constructs driven by the ROP1 promoter (data not shown).

A catalytically inactive mutant S999A-HA was well expressed in T. gondii tachyzoites with the most prominent product the size of the 140 kDa hypothesized precursor (Fig. 5C). The 90 kDa form was not seen and only small amounts of the 85 kDa product were evident. These results argue that the three forms of TgSUB2 are derived from the same initial translation product with the 90 kDa form derived from autocatalytic cleavage of the 140 kDa proprotein. The small amount of detectable 85 kDa TgSUB2 may be due to the presence of endogenous TgSUB2 that can process S999A-HA in trans. In some experiments, an extra band migrating at 100 kDa was seen only in transfected parasites, particularly those transfected with mutant TgSUB2. This is probably incorrectly processed TgSUB2 that is seen when TgSUB2 is overexpressed and cannot be correctly processed.

TgSUB2 shares features with the ROP1 processing proteinase

Rhoptry proteins in Plasmodium have been shown to assemble in macromolecular complexes that are important for proper trafficking to rhoptries (Baldi et al., 2000). Of the T. gondii rhoptry proteins, only ROP1 processing has been carefully investigated [(Bradley and Boothroyd, 1999, 2001; Bradley et al., 2002); Fig. 5A] although most other rhoptry proteins appear to be processed in a similar manner (Sadak et al., 1988).

The cleavage site of ROP1 occurs in the sequence SFVE↓APVR [P4-P4′ residues (Bradley and Boothroyd, 1999; Bradley et al., 2002)]. Mutations of the P1′–P3′ residues of ROP1 processing did not affect processing, but mutation of the P1 E83 residue dramatically inhibited processing (Bradley et al., 2002). Examination of the TgSUB2 sequence revealed three sites with homology to the ROP1 processing site (Fig. 5A).

Because the initial autocatalytic event for subtilases usually occurs rapidly in the endoplasmic reticulum and is less likely to be affected by trafficking through the secretory pathway, we tested whether mutation of E686 affected processing of SUB2-HA. The position of this residue is in approximately the predicted position for the initial autocatalytic cleavage. An E686R mutation of SUB2-HA was transfected into T. gondii tachyzoites. Processing of E686R-HA was dramatically reduced with only trace amounts of the 90 kDa form evident (Fig. 5C). As for S999A-HA, some 85 kDa protein was seen, and there was precursor accumulation. Thus it is likely that the autocatalytic cleavage site of TgSUB2 shares sequence specificity with ROP1, and TgSUB2 is a strong candidate for the ROP1 processing proteinase.

TgSUB2 associates with ROP1

Immunoprecipitation experiments were performed to test whether ROP1 associates with TgSUB2 (Fig. 6). TgSUB2 is not an abundant protein and may be present in limiting amounts sufficient for necessary biological activity. At any given time, only a fraction of the total candidate substrate would be expected to be associated with TgSUB2. Western blot of TgSUB2 immunoprecipitated from tachyzoite lysates revealed co-precipitation of ROP1 with TgSUB2 (Fig. 6A). As expected, the amounts of ROP1 co-precipitated were not quantitative and represented only a fraction of the total ROP1 present in lysate. Surprisingly, the form of ROP1 that was found was the mature form rather than the precursor. It is possible that the protease associates with precursor but retained activity during the steps of immunoprecipitation leading to detection of the mature form. Alternatively, there may be a site within the mature ROP1 that binds TgSUB2. The microneme protein MIC2 was not detected in the TgSUB2 immunoprecipitates (data not shown).

Figure 6.

TgSUB2 associates with ROP1, a candidate substrate.
A. ROP1 co-precipitates with TgSUB2. Immunoprecipitation of TgSUB2 was performed with tachyzoite lysate using RαTgSUB2. Immunoprecipitates were immunoblotted with Tg49, the ROP1 mouse monoclonal antibody. Lane 1: total cell lysate prior to immunoprecipitation; lane 2: TgSUB2 immunoprecipitation; lane 3 immunoprecipitation with pre-immune serum; lane 4: immunoprecipitation with lysate alone without antibody; lane 5: immunoprecipitation without tachyzoite lysate. The arrow indicates the position of ROP1.
B. ROP1 antibody Tg49 co-immunoprecipitates TgSUB2 and the S999A mutant. Immunoprecipitation of ROP1 was performed with tachyzoite lysate using mouse monoclonal antibody Tg49. Tachyzoites were transfected with SUB2-HA (wild type or WT), the catalytically dead S999A-HA construct, or mock transfected. The left panel (input) shows a short exposure of the total lysates probed with rat HA antibody. The right panel shows a longer exposure of the same gel probed with rat HA antibody. The first three lanes show the input lysates (exactly as in left panel). Subsequent lanes show ROP1 immunoprecipitations performed with the indicated lysates. Arrows show the position of the TgSUB2 precursor and two processed TgSUB2 bands.

To confirm the specificity of the interaction, immunoprecipitation of ROP1 followed by Western blot for TgSUB2 was performed. Both SUB2-HA (wild type or WT) and catalytically inactive S999A-HA were immunoprecipitated with ROP1 (Fig. 6B). Although TgSUB1 is a much more abundant protein than TgSUB2, TgSUB1 was not detectable in Western blots of ROP1 immunoprecipitates (data not shown).


TgSUB2, a subtilase from T. gondii, localizes to the rhoptry, a unique secretory organelle conserved throughout the phylum Apicomplexa. TgSUB2 contains all of the structural units known to be essential for generation of active enzyme (Siezen and Leunissen, 1997) and is homologous to other known Apicomplexan subtilases, with most similarity in primary sequence and domain structure to P. falciparum PfSUB-2. PfSUB-2 localizes to merozoite dense granules (Barale et al., 1999) while TgSUB2 is associated with rhoptries.

TgSUB2 appears to be an essential gene in tachyzoites. Attempts to generate parasites containing disruption of the TgSUB2 gene were unsuccessful. T. gondii tachyzoites are haploid so deletion of any essential gene would be lethal. Attempts to disrupt PfSUB-2 have also been unsuccessful to date suggesting that these proteinases have essential functions throughout the phylum (R. A. O’Donnell, M. J. Blackman, and B. S. Crabb, unpubl. obs.).

These experiments also suggest that TgSUB1, a microneme subtilase (Miller et al., 2001), cannot compensate for lack of TgSUB2 and that TgSUB2 substrate specificity or targeting results in it having a unique and essential function in T. gondii. PfSUB-2 has been proposed as a processing proteinase for MSP-1, the major merozoite surface antigen (Barale et al., 1999) based upon sequence homology to the MSP1 processing site within potential autocatalytic cleavage sites in PfSUB-2. Due to technical difficulties with PfSUB-2 expression, this hypothesis has not yet been proven. MSP-1 processing is required for successful invasion by merozoites (Blackman, 2000), but analogous processing of surface proteins of T. gondii during invasion has not been described.

Rhoptry protein ROP2 has sorting motifs in its cytoplasmic tail (Hoppe et al., 2000), and the malaria rhoptry proteins RAP1 and RAP2 are targeted to rhoptries as part of a complex (Baldi et al., 2000). Similarly, trafficking to T. gondii micronemes requires the presence of cytoplasmic tail targeting sequences, and soluble microneme proteins require association with transmembrane escorters to target correctly (Di Cristina et al., 2000; Reiss et al., 2001; Meissner et al., 2002).

TgSUB2 does not contain rhoptry targeting sequences YXXΦ or NYXP as defined for ROP2 and ROP4 (Hoppe et al., 2000). TgSUB2 targeting to the rhoptries may be due to association with other rhoptry proteins or due to the presence of alternative rhoptry targeting sequences. ROP1 does not have a transmembrane domain. Sequences within mature ROP1 or the pro-domain alone are sufficient to target ROP1 to the rhoptries (Soldati et al., 1998; Bradley and Boothroyd, 2001; Striepen et al., 2001; Bradley et al., 2002). In addition, ROP1 processing is not essential for ROP1 targeting to rhoptries (Bradley et al., 2002). It seems likely that, as for T. gondii microneme proteins, rhoptry proteins assemble as macromolecular complexes that are trafficked as a complex to the rhoptries. Proteinases have also been hypothesized as sorting receptors/chaperones within the secretory pathway (Cool et al., 1997) and it is possible that such a function for TgSUB2 is as important as its proteolytic activity.

Because subtilases undergo autocatalytic cleavage, mapping their endogenous cleavage sites enables identification of other biological substrates. Like PfSUB1 and PfSUB-2 (Hackett et al., 1999; Sajid et al., 2000), TgSUB2 undergoes two steps of processing during activation in the secretory pathway. Although further studies with active TgSUB2 are needed to completely define the substrate specificity of TgSUB2, our results are consistent with TgSUB2 being the ROP1 processing enzyme. TgSUB2 co-precipitates with ROP1, and mutagenesis studies suggest that it is autocatalytically cleaved after an E (glutamate), like ROP1. Attempts to prove this by N-terminal sequencing (or mass spectrometric analysis) are ongoing. TgSUB2 is not an abundant protein and is very insoluble, complicating our efforts to purify enough for successful protein microsequencing.

The residues surrounding the potential autocatalytic cleavage site are similar to ROP1, enabling us to hypothesize SΦXE as a cleavage motif, where Φ represents bulky hydrophobic residues and X is any amino acid. E686 and E734 are in the expected position for autocatalytic cleavage sites, 50–100 residues before the catalytic D at residue 783. Cleavage initially after E686 and then E734 would be expected to result in proteins that differ in mass by about 5 kDa, the observed difference in the mature forms of TgSUB2 seen on Western blot.

TgSUB2 processing activity on other substrates is likely to be more biologically important than its role as a ROP1 processing enzyme as TgSUB2 appears to be essential but ROP1 is not (Kim et al., 1993). Processing proteinases typically have more than one biological substrate. One other potential substrate for TgSUB2 is ROP2 and related family members. ROP2 is probably an essential gene and is hypothesized to be necessary for acquisition of essential nutrients such as lipids or cholesterol (Sinai and Joiner, 2001; Coppens et al., 2000). The exact site of proteolytic processing has not been determined for ROP2 family members. For ROP2 the site has been mapped to the N-terminal pro-domain close to residue 98 (Sinai and Joiner, 2001). Removal of the N-terminal prodomain of ROP2 exposes sequences important for host mitochondria association with the parasitophorous vacuole (Sinai et al., 1997; Sinai and Joiner, 2001). The sequences of ROP2, ROP4 and ROP8 have a similar sequence (SWLE/QE) to the ROP1 and TgSUB2 cleavage sites, as would be predicted if TgSUB2 is a general rhoptry maturase. ROP9 does not contain the motif, but may not be proteolytically processed (Reichmann et al., 2002).

Shaw et al. (2002) have reported that treatment of tachyzoites with cathepsin inhibitor III, TPCK, or subtilisin inhibitor III disrupt the parasite's secretory pathway and disrupt rhoptry formation. The cathepsin inhibitor target is likely to be toxopain I, a cathepsin B-like cysteine proteinase that localizes to the rhoptries (Que et al., 2002). TgSUB2 is an excellent candidate for the target of the subtilisin serine proteinase inhibitor. Thus TgSUB2 may have a role in rhoptry protein processing and rhoptry biogenesis.

Cleavage after acidic residues is highly unusual for subtilases with the only other known example being PfSUB1, which cleaves preferentially after VXXD (Sajid et al., 2000; Withers-Martinez et al., 2002). Most mammalian subtilases cleave after basic residues with the only exception being S1P/SKI1, which cleaves itself or its biological substrate SREBP after leucine or lysine (Cheng et al., 1999). Studies with recombinant enzyme and other substrates suggest broader specificity with a recognition motif that has been determined to be [R/K]-x-hydrophobic-Z where x is any amino acid and Z is any amino acid except P, D, E or C (Toure et al., 2000; Elagoz et al., 2002). Protease inhibitors usually mimic active substrates. Assuming that future studies with active enzyme verify that TgSUB2 cleaves rhoptry proteins after the acidic residue E, the unusual substrate specificity of TgSUB2 may facilitate synthesis of parasite-specific subtilase inhibitors. Similar subtilase genes are present in all apicomplexan genomes sequenced to date, and processing of secretory organelle contents is commonly seen in the Apicomplexa. Proteinase inhibitors are an active area of investigation for treatment of many human diseases, and inhibitors of the malaria falcipain and plasmepsin proteinases are being evaluated as anti-malarials. Subtilases may represent another unique chemotherapeutic target in T. gondii, malaria and other medically significant protozoan infections.

Experimental procedures

Cloning of the TgSUB2 gene

Degenerate oligonucleotide primers previously used to clone P. falciparum PfSUB1 were used to clone TgSUB2 (Blackman et al., 1998). RT-PCR using T. gondii RNA yielded a single 511 bp fragment used as a probe to screen a λ-ZAPII T. gondii cDNA library (Wan et al., 1997). Sequence of the largest cDNA (3.4 kb) revealed a conserved subtilisin proteinase domain and in frame stop codon but no start codon. Genomic clones were isolated by screening a T. gondii genomic BAC library (Wan et al., 1997). HindIII BAC subclones of TgSUB2 were sequenced. Comparison of genomic and RT-PCR products confirmed the absence of additional introns upstream of the original cDNA clone. The sequence revealed an in frame translation start codon meeting Kozak's criteria (Pedersen and Nielsen, 1997) and a predicted signal peptide cleavage site (Nielsen et al., 1997). A full-length recombinant TgSUB2 cDNA construct without introns was cloned into pBluescript (Stratagene) by ligating the 1.3 kB SacII–EcoRI genomic fragment to the 3.4 kB EcoRI fragment from the cDNA clone. The nucleotide sequence for TgSUB2 has been deposited in the GenBank database under GenBank Accession Number AF420596.

Southern blot analysis

Genomic DNA or genomic TgSUB2 BAC DNA was analysed by high stringency Southern blot (68°C) using digoxigenin-labelled probes (Boehringer Mannheim) made with PCR primer pairs KK178 (AACAACAGGATGGTCGAG)-KK182 (GCCATAAGAACAGCATCTC) for Probe A or KK162 (ACTACACGGCGGACTTAAC)-KK192 (CTACCAAGCTTAC GACTCA) for Probe B. Bound probe was detected with disodium 3-(4-methoxyspiro{1,2-dioxetane-3,2′-(5′-chloro)tricyclo[,7]decan}-4-yl) phenyl phosphate (CSPD®, Boehringer Mannheim).

Antibody production

Polyclonal mouse antisera (MαTgSUB2) were raised against a TgSUB2 glutathione-S-transferase (GST) fusion protein. The catalytic domain region was amplified with primers KK172 (ATGCTTGGATCCCATGGAGA) and KK173 (GACA CAACACTCGAGAATGC) and cloned into the BamHI and XhoI sites of pGEX-KG (Guan and Dixon, 1991). After protein purification, BALB/c mice were immunized at the tail base with 50 µg GST-TgSUB2-CAT mixed 1:1 with Titermax gold adjuvant (Sigma). Boosts (3) were performed at 2 week intervals.

Polyclonal rabbit antipeptide antiserum (RαTgSUB2) was raised against a peptide (CLEQFALSTPNENES) containing the last 14 amino acids of TgSUB2 (Laboratory for Macromolecular Analysis at Albert Einstein College of Medicine). The peptide was covalently linked to Imject® malemide activated keyhole limpet haemocyanin (KLH) (Pierce) and purified by gel filtration according to the manufacturer's protocol. Immunization of one elite rabbit was performed by Covance Research Products, Inc. Crude RαTgSUB2 antiserum was used for Western blotting and immunoprecipitations, and affinity purified antibody was used for immunoelectron microscopy.

Affinity purification was performed by covalently coupling the peptide used for immunization to thiopropyl sepharose (Sigma). Crude antiserum was applied to the column, washed, and peptide-specific antibody was eluted with low pH buffer, which was immediately neutralized with 1 M Tris-HCl pH 9.5 and stored at −20°C. Mouse monoclonal antibody Tg49 specific for rhoptry protein ROP1 (MαROP1; gift of Joe Schwartzman, Dartmouth University, UK) was used as indicated.

Electron microscopy

Immunoelectron microscopy of RH strain tachyzoites was performed as described (Miller et al., 2001) using affinity purified RαTgSUB2 (diluted 1:2) and/or mouse monoclonal antibody Tg49 (MαROP1; 1:500) as primary antibodies. Mouse monoclonal antibody Tg49 was a gift of Joe Schwartzman, Dartmouth University, UK (Saffer et al., 1992). Secondary antibodies were goat anti-rabbit (GαR) IgG coupled to 15 nm gold and/or goat anti-mouse (GαM) IgG coupled to 10 nm gold (diluted 1:20; Polysciences, Inc.). Stained sections were viewed on a JEOL 1200EX transmission electron microscope at the Albert Einstein College of Medicine Analytical Imaging Facility.

TgSUB2 gene disruption constructs

The plasmid TUB1 CAT (Soldati and Boothroyd, 1993) containing the chloroamphenical acetyl transferase (CAT) selectable marker gene under control of the Toxoplasmaα-tubulin promoter with 3′SAG1 transcription termination sequence was used to create knockout constructs. A 3.6 kb SacI–BamHI 5′ genomic TgSUB2 fragment was cloned into the same sites in TUB1 CAT. 3′ genomic fragments HindIII–KpnI 1.2 kb and KpnI 1.1 kb were then cloned sequentially into the HindIII and KpnI sites and correct orientation was confirmed by digestion and sequencing. The final construct has 3.6 kb of 5′ and 2.3 kb of 3′TgSUB2 genomic sequence separated by 1.4 kb TUB1 CAT in the opposite orientation. Next, 3.4 kb of genomic sequence encoding for the core proteinase domain was deleted and replaced by the selectable marker in this plasmid.

A second disruption construct with the hypoxanthine-xanthine-guanine phosphoribosyl transferase (HXGPRT) selectable marker with flanking BamHI and HindIII sites was created by PCR using primers KK324 (CAATAAAAGCT TGATCAGCAC) and KK325 (CTTCAAGGATCCCCCTC CACC) with pHLEM as a template (Black and Boothroyd, 1998). The amplified DNA has the HXGPRT open reading frame (ORF) flanked by 5′ and 3′ dihydrofolate reductase (DHFR) sequence from T. gondii. The 1.4 kb BamHI–HindIII TUB1 CAT marker was replaced with the 1.9 kb BamHI–HindIII fragment containing HXGPRT. Knockout constructs were linearized with SacI and 100 µg transformed into RH strain tachyzoites and selected with 20 µM chloramphenical (CAT construct) (Kim et al., 1993) or 25 µg ml−1 mycophenolic acid supplemented with 5 µg ml−1 xanthine (HXGPRT construct) (Donald et al., 1996). After 2–3 passages, stable transformants were subcloned by limiting dilution (two rounds) and grown in the presence of drug. These parasites were examined for the presence of endogenous TgSUB2 gene by PCR and Southern blot or TgSUB2 protein by Western blot.

Generation of HA-tagged TgSUB2

TgSUB2 was amplified by PCR with sense primer SUB2H3NsiF, CCCAAGCTTATGCATATGGCTAGACGCA GACCATCT and antisense primer SUB2NheR, CGCGCT AGCCGACTCATTCTCGTTGGGCGTCGATAA, replacing the stop codon with a NheI site. This cassette was cloned into HindIII and NheI-digested pHA9F [gift of Jay Bangs, University of Wisconsin (Bangs et al., 1996)] to yield plasmid pSUB2HA9F with TgSUB2 fused in frame to the HA9 epitope.

HA-tagged TgSUB2 was PCR-amplified from pSUB2HA9F using the same sense primer and antisense primer HAPacR, CGCTTAATTAACTACGCGTAATCTGGGACGTCGTAT to introduce a PacI site at the stop codon downstream of the HA9 epitope. This PCR product was cloned into the NsiI and PacI sites of GRA1-GFP replacing GFP [original vector gift of David Sibley, Washington University; (Kim et al., 2001)] to yield the expression plasmid pSUB2-HA. Constructs driven by the ROP1 promoter were constructed by cloning into the NsiI and PacI sites (Soldati and Boothroyd, 1993).

The catalytically inactive and processing site mutants of SUB2-HA were generated using the QuikChange® site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions with pSUB2-HA as the DNA template. The catalytically inactive mutant SUB2-HA S999A was produced by replacing the S codon at residue 999 with one for A (underlined) using the oligonucleotide primers: sense, CATATGTCTGGAACGGCCATGGCGACACCGCAC and antisense, GTGCGGTGTCGCCATGGCCGTTCCAGA CATATG. SUB2-HA E686R was produced by replacing the E codon at residue 686 with an R codon (underlined) with: sense CCAACAATGGTCATTCTCTAGAATAGTTCAACTC CGG and antisense CCGGAGTTGAACTATTCTAGAGAAT GACCATTGTTGG. All clones were verified by sequencing.

Immunoprecipitation and Western analysis of tachyzoite lysates

Immunoprecipitations were performed as previously described (Miller et al., 2001) with the following modifications. For the ROP1 co-immunoprecipitations, parasites (∼108) were lysed in cold lysis buffer (75 mM NaCl, 1% Triton X-100, 50 mM Tris-HCl, pH 7.4) in the presence of protease inhibitors. Immune complexes were collected with ImmunoPure Plus Immobilized Protein A slurry (Pierce). Controls included lysate incubated in the absence of antibody and RαTgSUB2 incubated in the absence of lysate.

Western blots were performed with Tg49 (1:1000), 50 mU ml−1 rat anti-HA-peroxidase conjugate monoclonal antibody 3F10 (Roche), MαTgSUB2 1:500 or RαTgSUB2 (1:1000). Bound antibodies were visualized using Super Signal chemiluminescent substrate (Pierce). For detection of total TgSUB2 present in transfected cell lysates, blots were first probed with HA antibody, incubated for 10 min in 0.1% sodium azide to inhibit activity of the HA-peroxidase conjugate and re-probed with RαTgSUB2.


We thank Peter Bradley and John Boothroyd for communicating results prior to publication, Joe Schwartzman for Tg49 monoclonal antibody, Vern Carruthers and Lou Weiss for helpful discussions throughout the course of this work, and the Albert Einstein College of Medicine Analytical Imaging facility for technical assistance. Supported by the Howard Hughes Medical Institute Biomedical Research Support Program for Medical Schools to Albert Einstein College of Medicine no. 513504, NIH grant RO1-A146985 and a Burroughs Wellcome New Investigator in Molecular Parasitology Award (K.K.); NIH MSTP training grant T32-GM07288 (S.A.M.); and NIH training grant T32-AI07501 (V.T.). The AECOM Analytical Imaging Facility is supported by NIH Cancer Center grant P30CA13330. Some of the data in this paper are from a thesis submitted by S.A.M. in partial fulfilment of the requirements for the Degree of Doctor of Philosophy in the Sue Golding Graduate Division of Medical Sciences, Albert Einstein College of Medicine, Yeshiva University.