A Glycosylphosphatidylinositol-Anchored Carbonic Anhydrase-Related Protein of Toxoplasma gondii Is Important for Rhoptry Biogenesis and Virulence

Toxoplasma gondii is an intracellular pathogen that infects humans and animals. The pathogenesis of T. gondii is linked to its lytic cycle, which starts when tachyzoites invade host cells and secrete proteins from specialized organelles. Once inside the host cell, the parasite creates a parasitophorous vacuole (PV) where it divides. Rhoptries are specialized secretory organelles that contain proteins, many of which are secreted during invasion. These proteins have important roles not only during the initial interaction between parasite and host but also in the formation of the PV and in the modification of the host cell. We report here the identification of a new T. gondii carbonic anhydrase-related protein (TgCA_RP), which localizes to rhoptries of mature tachyzoites. TgCA_RP is important for the morphology of rhoptries and for invasion and growth of parasites. TgCA_RP is also critical for parasite virulence. We propose that TgCA_RP plays a role in the biogenesis of rhoptries.

patients (2), and serious malformations in congenitally infected children (3). T. gondii utilizes a complex secretory system, which is essential for host cell invasion and the establishment of a parasitophorous vacuole (PV). This system is composed of multiple organelles: rhoptries, micronemes, and dense granules. These organelles release their contents when the parasite invades a host cell. Rhoptries are club-shaped organelles which share some characteristics with secretory lysosomes (4) and are comprised of two distinct regions termed the rhoptry neck and the rhoptry bulb, each compartment containing distinct protein constituents (5,6).
The ␣-carbonic anhydrases (␣-CAs) are zinc metalloenzymes found in a variety of organisms that catalyze the reversible hydration of CO 2 , which is important for a number of biological functions, including respiration, photosynthesis, renal tubular acidification, and bone resorption (7)(8)(9). In addition to these enzymatically active CAs, there are CA isoforms known as carbonic anhydrase-related proteins (CARPs), which share sequence and structural similarity to active CAs but lack enzymatic activity. CARPs can exist either as independent proteins or as domains of other proteins. Their lack of activity is due to the absence of one or more histidine amino acids that are required in the active site for zinc ion (Zn 2ϩ ) coordination (9). CARPs have been described in vertebrates, insects, nematodes, and viruses (9), but there are no reports of their presence in early branches of the tree of life.
The Plasmodium falciparum carbonic anhydrase (PfCA; PF3D7_1140000, Genbank accession no. CZT99063.1) was recently characterized and classified into a new class of CAs, -CAs. An interesting peculiarity of the PfCA sequence is the presence of the amino acid glutamine in the Zn 2ϩ coordination site, which replaces one of the three canonical histidine residues. PfCA is catalytically active in spite of this substitution (10,11).
In the present work, we investigated one of the annotated carbonic anhydrase proteins of T. gondii (TGME49_297070). Analysis of the sequence of the Toxoplasma protein showed that it is more similar to PfCA than to the ␣-CAs. We characterize and localize TgCA_RP in T. gondii tachyzoites and analyze the phenotypic profile of mutant parasites (⌬carp) isolated after the clean deletion of the TgCA_RP gene.

RESULTS
Identification of a carbonic anhydrase-related protein, TgCA_RP. A gene originally annotated as a carbonic anhydrase and subsequently as a hypothetical protein (TGME49_297070, NCBI accession no. XP_002371137.1) was cloned and sequenced. The open reading frame (ORF) of the annotated gene encodes a protein of 519 amino acids (aa) with a predicted molecular mass of 58 kDa and an isoelectric point of 6.29. We named the gene product T. gondii carbonic anhydrase-related protein (TgCA_RP), because close analysis of the amino acid sequence showed that two histidine (His) residues typically required for activity (see below) were replaced by a phenylalanine and a glutamine. In addition, recombinant TgCA_RP was inactive (see below). Analysis of the TgCA_RP amino acid sequence revealed that it contained an ␣-carbonic anhydrase domain (amino acids 121 to 445) and a predicted signal peptide (amino acids 1 to 39) at its N terminus.
Alignment of the amino acid sequence of TgCA_RP with those of other ␣-CA sequences ( Fig. 1A; see also Table S1 in the supplemental material) and with orthologous sequences from other apicomplexan parasites (Fig. 1B) illustrates that TgCA_RP is more similar to the Plasmodium -CA, PfCA (PF3D7_1140000), than the ␣-CAs (Fig. 1). The position of TgCA_RP in a neighbor-joining tree of alpha, beta, eta, and zeta family CAs supports this conclusion (Fig. S2). TgCA_RP also shares several important features with PfCA (12). The canonical ␣-CA zinc coordination domain consists of three histidines identified by their position in the human carbonic anhydrase 2 (HuCA2) sequence (His 94 , His 96 , and His 119 ) (Fig. 1A, *, **, and ***). The third histidine (His 119 ) of the zinc coordination domain is replaced with a glutamine (Gln) in the -class enzyme PfCA (Gln 320 , Fig. 1B, ***) (11). In the case of TgCA_RP, two of the three histidines of the Zn 2ϩ coordination domain are substituted. Like PfCA, the position of the third histidine is occupied by a Gln (Gln 254 , Fig. 1A, ***), but in addition, the first His is changed to a phenylalanine (Phe 233 , Fig, 1A, *). These two substitutions in an essential domain for Zn 2ϩ binding indicated that TgCA_RP would be inactive (13) (Fig. 1A and B). We compared the Zn 2ϩ coordination domain of TgCA_RP with the same domain of other apicomplexan CA orthologs (Fig. S1). Neospora caninum (NcCAh), Eimeria tenella (EtCAh), Theileria parva (TpCAh), and Theileria equi (TeCAh) show a replacement of the first His in the Zn 2ϩ binding domain with Phe, Leu, or Asp (Phe 236 , Leu 209 , Asp 200 , and Asp 196 , respectively). The Plasmodium berghei (PbCAh) and Plasmodium yoelii (PyCAh) sequences replaced the His in the Zn 2ϩ coordination domain with asparagine (Asn 328 and Asn 480 , Fig. 1A, *). PfCA is, the only protein of this group that possesses a histidine in this position. These substitutions, specifically with asparagines, do not necessarily preclude activity. Previously, Lesburg et al. (14) substituted an asparagine for His 94 in HuCA2, and the enzyme retained activity, although activity was markedly reduced (14). It is also possible that PfCA is the only enzymatically active carbonic anhydrase of this group and this question will be answered only once the other enzymes are characterized. TgCA_RP also shows other features that define the class. The gateway residues Glu 106 and Thr 199 in HuCA2 are responsible for the orientation of the substrate during catalysis. Both TgCA_RP (Glu 239 and Thr 414 ) and PfCA (Glu 305 and Thr 500 ) are predicted to have these gateway residues in the appropriate position for substrate orientation (14,15) (Fig. 1C, blue amino acids). TgCA_RP and PfCA lack the His 64 proton shuttle present in most ␣-CA active sites, which is responsible for increased catalysis via the transport of protons from catalytic intermediates to the environment ( Fig. 1A and B, ===) (16)(17)(18). This was also the case for the other apicomplexan orthologs that were aligned (Fig. 1B). An alignment of the I-TASSER predicted structure of the putative TgCA_RP CA domain (amino acids 121 to 446) (19)(20)(21) with the crystal structure of HuCA2 (22) illustrates the above features and their location in the predicted active site (Fig. 1C).
Two prediction systems for glycosylphosphatidylinositol (GPI)-anchored proteins, PredGPI (23) and GPI-SOM (24), predicted that TgCA_RP has a C-terminal GPI anchor cleavage/attachment consensus sequence (Table S2). We also analyzed the sequences of TgCA_RP orthologs from related apicomplexan parasites. All of the orthologs, except for PfCA of P. falciparum, were predicted to contain GPI anchor attachment sites, including orthologs of P. yoelii and P. berghei (Table S2) (23,24). These predictions suggest that the GPI anchor modification of TgCA_RP and its orthologs may be conserved from a common ancestor of the Apicomplexa.
To investigate the CA activity of TgCA_RP, we expressed two truncated forms, rTgCA_RPa (amino acids [aa] 121 to 445) and rTgCA_RPb (aa 94 to 488), in Escherichia coli. The truncated proteins were soluble, unlike the full-length protein, which was expressed and purified for antibody generation (see below). rTgCA_RPa contains the predicted CA domain, and rTgCA_RPb includes, in addition, the N-terminal region downstream of a cleavage motif, SXL2Q ( 91 SLLQ 94 ), previously characterized in toxolysin, a rhoptry metalloprotease (25). Both fragments were cloned in the Pet32Lic/EK vector, and the resulting constructs were transformed into E. coli for expression. Nickel affinity purification of soluble protein products followed with the aim of measuring rTgCA_RP catalytic activity. Two protocols for activity were used: hydrolysis of p-nitrophenyl acetate (pNPA) (26) and in-gel CO 2 hydration assays (27). Commercially available bovine carbonic anhydrase (Sigma catalog no. C3934) was used as a positive control for activity. We were unable to detect any measurable activity from rTgCA_RPa or rTgCA_RPb when measured at various pHs (pH 6.0 to 8.0) and buffer compositions (Tris-HCl, Tris-SO 4 , Tricine-KOH, HEPES-KOH, morpholineethanesulfonic acid [MES], and morpholinepropanesulfonic acid [MOPS]) (data not shown). It was reported that sitedirected mutagenesis of the Zn 2ϩ coordination site restored enzymatic activity of some ␣-type CARPs (28,29). We performed site-directed mutagenesis of rTgCA_RPa and rTgCA_RPb to alter the predicted Zn 2ϩ coordination residues to the His 94 -His 96 -His 119 of the ␣-CAs (as in HuCA2) or to His 299 -His 301 -Glu 320 as in PfCA, but these alterations did not restore activity to rTgCA_RP as measured with either activity assay (data not shown). In summary, sequence analysis and experimental results indicated that TgCA_RP lacks activity. However, we were unable to restore its activity by mutating specific amino acids of rTgCA_RP, making it difficult to definitively state that TgCA_RP is inactive.
TgCA_RP localized to the rhoptry bulbs. To investigate the localization of TgCA_RP, an in situ epitope-tagging technique was used to avoid the problem of overexpression and potential abnormal distribution of the tagged protein ( Fig. 2A). Western blot analysis of a clonal parasite line expressing TgCA_RP-HA showed three bands around the predicted molecular mass (58 kDa) of TgCA_RP (Fig. 2B). Immunofluorescence analysis experiments with the parasite clone expressing TgCA_RP-HA showed labeling of rhoptry bulbs in extracellular tachyzoites (Fig. 2C, green), determined by colocalization with the rhoptry bulb marker anti-ROP1 (Fig. 2C, red). Localization to rhoptries of TgCA_RP was previously done by expressing an extra copy of the gene with a c-Myc tag (30). To provide fine details of TgCA_RP localization, we performed cryo-immunoelectron microscopy of TgCA_RP-HA extracellular tachyzoites. Gold particles were observed in rhoptries (Fig. 2D, a), mainly detected at the basal bulbous portion of these organelles (Fig. 2D, b) and at the limiting membrane (Fig. 2D, c). In intracellular tachyzoites, TgCA_RP-HA showed more diffuse and less specific localization, and antihemagglutinin (anti-HA) labeled other uncharacterized structures, in addition to the rhoptries (Fig. 2E). Interestingly, the morphology of the rhoptries appeared to be affected by the expression of TgCA_RP-HA (Fig. 2E). To investigate the localization of TgCA_RP in wild-type tachyzoites, we generated specific antibodies to recombinant TgCA_RP in mice (Fig. 3A, lanes M) and guinea pigs (Fig. 3A, lane GP). Western blot analysis of lysates from parental cell lines (⌬ku80) showed one prominent band flanked by faint upper and lower bands; the lower band was partially covered by the intense middle band (Fig. 3A, Parental). In contrast, Western blot analysis of lysates from the tagged cell line (carp-HA) showed three strong bands around the predicted size of 50 kDa, in agreement with the results obtained with anti-HA antibodies (Fig. 2B). The increased presence of these higher-and lowermolecular-mass forms suggested that the tag could be altering the maturation/processing of TgCA_RP (Fig. 3A, lanes M, Parental and carp-HA). Immunofluorescence assays (IFAs) using antibodies to TgCA_RP showed localization of the protein to the rhoptry bulbs of both extracellular (Fig. 3B) and intracellular (Fig. 3C) tachyzoites, as confirmed by its colocalization with ROP7 ( Fig. 3B and C). This was further confirmed with cryo-immunoelectron microscopy (Fig. 3D). Using superresolution structured illumination microscopy (SR-SIM), we observed that TgCA_RP labels the bulbs of mature rhoptries. Interestingly, we observed regions of the rhoptry bulbs in which there was no colocalization between TgCA_RP and ROP7 (Fig. 3E, arrowheads). TgCA_RP localized specifically to the peripheral membrane of nascent rhoptries in dividing tachyzoites, surrounding the luminal proteins ROP7 and ROP4 (Fig. 3F). In contrast, TgCA_RP-HA did not specifically localize to the membrane of nascent rhoptries but instead showed labeling similar to ROP7 and ROP4 (Fig. 3G).
Tachyzoites pretreated with cytochalasin D successfully attach to host cells, but they are not able to invade. Under these conditions, rhoptry bulb contents are deposited in a trail of vesicles within the host cell cytosol. These vesicles were termed evacuoles because they do not contain parasites (31). We investigated if TgCA_RP was part of the evacuole contents according to a published protocol (31). Anti-TgCA_RP did not label evacuoles in attached RH strain tachyzoites, suggesting that TgCA_RP may not be secreted into the host cell (data not shown).

TgCA_RP-knockout mutants exhibited reduced invasion in vitro.
To establish the role of TgCA_RP in the life cycle of T. gondii, we generated tachyzoite knockout mutants (⌬carp) by targeting the native TgCA_RP locus in the ⌬ku80 strain, which favors homologous recombination (32, 33) (Fig. 4A). Successful gene deletion was confirmed by PCR amplification of genomic DNA, extracted from subclones of the transfectants, using upstream and downstream primers ( Fig. 4A; Table S3). A 2.1-kb product was amplified from the genomic DNA of a ⌬carp mutant subclone, confirming the replacement of TgCA_RP with a chloramphenicol acetyltransferase (CAT) cassette (Fig. 4B). Successful gene deletion was confirmed by Western blot analysis of parasite lysates using mouse polyclonal antibody to TgCA_RP (Fig. 4C). Restoration of gene expression was achieved by transfecting ⌬carp tachyzoites with a plasmid containing the entire TgCA_RP gene locus, including the predicted untranslated regions (UTRs) and the potential promoter region (500 bp upstream of the annotated 5= UTR), which was amplified from genomic DNA via PCR. The expression of TgCA_RP in complemented mutants (⌬carp-cm) appeared to be at appropriate levels (Fig. 4C). ⌬carp-cm mutants retained chloramphenicol resistance, and PCR of the genomic DNA showed bands of appropriate size for both the CAT cassette and the complemented TgCA_RP locus, confirming random integration (Fig. 4B). We compared plaque sizes of the parental (⌬ku80), knockout (⌬carp), and complemented (⌬carp-cm) clones in parallel plaque assays and found a significant difference in plaque size ( Fig. 4D and E). ⌬carp parasites formed smaller plaques ( Fig. 4D and E) and had an invasion defect, as shown by plaquing efficiency (Fig. 4F) as well as red-green invasion assays (Fig. 4G). The invasion defect observed in plaquing efficiency assays was exacerbated by extracellular stress consisting of a 1-h preincubation in Dulbecco's modified Eagle's medium (DMEM) (Fig. 4F). Interestingly, the number of ⌬carp tachyzoites per vacuole was not affected compared to the number of tachyzoites formed by ⌬ku80 and ⌬carp-cm strains, after 25 h of intracellular growth (data not shown). These results indicated that invasion and  not intracellular replication is the reason for reduced plaque size. This result is consistent with the fitness score of TgCA_RP, 0.31, which suggests that this gene is not relevant for intracellular replication (34).
TgCA_RP is required for effective in vivo infection. After establishing that the ⌬carp mutant exhibited an in vitro growth phenotype, we tested the virulence of mutant tachyzoites in a mouse model. We first infected mice intraperitoneally (i.p.), but because of the highly virulent nature of the parental RH strain, it was difficult to see a clear difference in virulence when comparing survival between mutant and parental lines, although we did observe a small increase in survival time of mice inoculated with the ⌬carp mutants (not shown). Dissemination through the circulation is important during natural infection, and this transit must expose tachyzoites to stressful conditions. Considering that the ⌬carp mutant showed sensitivity to stress (Fig. 4F), we tested virulence by infecting mice intravenously (i.v.) (35). We inoculated mice i.v. with 50 tachyzoites of the ⌬carp, parental, or ⌬carp-cm strain. We observed that 70% of mice infected with ⌬carp tachyzoites survived more than 30 days postinfection, while none of the mice infected with the parental or complemented parasites survived more than 12 days (Fig. 4H). Surviving mice and phosphate-buffered saline (PBS) control mice were challenged with 1,000 RH-rfp tachyzoites. All of the surviving mice previously inoculated with ⌬carp tachyzoites survived the RH challenge without showing illness, in contrast to a 100% mortality of the PBS control mice (Fig. 4H). When mice were inoculated with 150 or 500 ⌬carp tachyzoites, there was only a 40% and 10% survival rate, respectively, but there was also a delayed morbidity in mice that did not survive (Fig. S4). In conclusion, our results indicate that the infectivity of ⌬carp tachyzoites is attenuated compared to the parental and ⌬carp-cm strains.
TgCA_RP is posttranslationally modified with a GPI anchor. To determine whether TgCA_RP is modified with a GPI anchor, we first investigated whether TgCA_RP was found in detergent-resistant membranes (DRMs). TgCA_RP was found in the pellet after incubation with 1% Triton X-100 at 4°C and in the soluble fraction after incubation at 37°C (Fig. 5A). This is consistent with the behavior of SAG1, a known GPI-anchored protein (36,37) (Fig. 5A). TgCA_RP-HA showed no clear difference in solubility after 4°C or 37°C incubations, suggesting that it is not found in DRMs. We next investigated whether TgCA_RP is modified with a GPI anchor, by utilizing Bacillus cereus phosphatidylinositol-specific phospholipase C (PLC), an enzyme that specifically cleaves GPI anchors (38), followed by detergent extraction with Triton X-114. After PLC treatment and Triton X-114 phase separation, the aqueous and detergent phases were analyzed by Western blotting, probing for TgCA_RP or TgCA_RP-HA and for SAG1, as a GPI-anchored positive control (39). TgCA_RP and SAG1 were both detected in the aqueous phase after treatment with PLC but not after mock treatment, providing strong evidence for the GPI anchor modification of TgCA_RP. TgCA_RP-HA was detected only in the detergent phase, behaving as an integral membrane protein, but this behavior did not change after PLC treatment, suggesting that TgCA_RP-HA is not GPI modified, which could be a consequence of the disruption of the GPI consensus domain by the C-terminal tag (Fig. 5B). To radiolabel the GPI anchor of TgCA_RP, we grew mutants (carp-HA and ⌬carp) and parental tachyzoites in the presence of [ 3 H]inositol. Western blot analysis of lysates was performed with anti-TgCA_RP, and the membranes were subsequently exposed to X-ray film for autoradiography. [ 3 H]inositol labeled several bands, which probably correspond to other GPI-modified proteins, such as the surface antigen SAG1 (39). A band was observed in lysates of the parental strain that corresponded to the band labeled by TgCA_RP on the Western blot. This band was absent in lysates from the ⌬carp tachyzoites (Fig. 5C, right panel). Furthermore, this band was not observed in the autoradiography of lysates from [ 3 H]inositol-labeled carp-HA mutants (Fig. 5D, right panel), supporting our hypothesis that the hemagglutinin (HA) tag interferes with the GPI modification. We also complemented ⌬carp tachyzoites with a TgCA_RP that was modified by replacing the predicted GPI anchor cleavage site with a randomized amino acid sequence (⌬carpGPI) (Fig. S4). [ 3 H]inositol did not label TgCA_RP in these mutants ( Fig. 5C and D). In summary, our results support the conclusion that TgCA_RP is GPI anchored. The results with carp-HA indicate that the C-terminal 3ϫHA tag may disrupt the GPI anchor modification, despite its distance from the predicted GPI anchor attachment site. This phenomenon has been previously described for the Plasmodium protein PfASP (40).
Rhoptry morphology is altered in TgCA_RP-knockout mutants. Immunofluorescence analysis of ⌬carp intracellular tachyzoites (Fig. 6A, b) using superresolution structured illumination microscopy revealed dramatic changes in rhoptry labeling compared to that of parental tachyzoites (Fig. 6A, a) infecting the same host cell Tachyzoites were treated with 1% Triton X-100 at 4°C or 37°C. After centrifugation, immunoblot analysis of the pellet (P) and aqueous (S) fractions with anti-TgCA_RP revealed that TgCA_RP is only partially soluble in 1% Triton X-100 at 4°C and is completely soluble at 37°C. This behavior is mirrored by SAG1, a known GPI-anchored protein. The solubility of TgCA_RP-HA is largely unaffected by incubation at 37°C. (B) Immunoblot of tachyzoite lysates with anti-TgCA_RP after Triton X-114 detergent extraction. TgCA_RP is found in the detergent fraction (Tx) of mock-treated lysates (Ϫ) but is partially released into the aqueous fraction (Aq) when the lysate is treated with GPI-PLC (ϩ) prior to extraction. SAG1, a known GPI-anchored protein, showed a similar pattern. C-terminally tagged TgCA_RP-HA is not released into the aqueous fraction by GPI-PLC. (C) Autoradiography of lysates from parental and ⌬carp tachyzoites labeled with [ 3 H]myoinositol. The arrow points to a band of identical size as the one observed in the Western blot (anti-TgCA_RP) of the same membrane, and this band is absent in the lysate of ⌬carp parasites. This absence supports the conclusion that TgCA_RP is GPI anchored. (D) The TgCA_RP band is also absent in autoradiography of carp-HA cells and mutants with a mutated GPI anchor attachment region, ⌬carpGPI (also in panel C).
monolayer. Three-dimensional (3D) reconstruction shows the fragmented nature of the rhoptries in the ⌬carp tachyzoites (Fig. 6B). The percentage of parasitophorous vacuoles (PVs) containing at least one "normal" linear rhoptry (rhoptries in Fig. 3B) was significantly reduced in ⌬carp tachyzoites compared to parental and ⌬carp-cm parasites (Fig. 6C). PVs containing nascent rhoptries were labeled with the marker anti-proROP4 and excluded from counting. Transmission electron microscopy of intracellular tachyzoites revealed typical rhoptry morphology with a bulb and neck in parental tachyzoites, in contrast to the fragmented globular rhoptries, which do not connect to a rhoptry neck structure, in the majority of ⌬carp tachyzoites (Fig. 6D). A minority of ⌬carp tachyzoites did show a connected rhoptry bulb and neck, but the rhoptry bulb in these cases was swollen (Fig. 6D). Freeze fracture electron microscopy of ⌬carp extracellular tachyzoites provided further evidence of the stunted rhoptry morphology compared to the elongated rhoptries of the ⌬carp-cm mutants (Fig. 6E). Similar rhoptry morphology defects were observed in carp-HA tachyzoites (Fig. 6F). Student's t test, *, P Ͻ 0.05; ***, P Ͻ 0.0001 (means Ϯ standard errors of the means, n ϭ 3). (D) Transmission electron microscopy of intracellular tachyzoites showed typical thin and elongated rhoptry bulb morphology in a parental tachyzoite (left, green arrowhead) and a representative example of the atypical fragmented rhoptry bulbs that do not connect to a rhoptry neck structure found in the majority of ⌬carp tachyzoites (middle, red arrows). A minority of ⌬carp tachyzoites contained fully formed rhoptry bulb and neck structures, but in these cases, the rhoptry bulb is atypically swollen (right, red arrowhead). (E) Freeze fracture electron microscopy of ⌬carp-cm extracellular tachyzoites (top) shows typical elongated rhoptries in contrast to the short and thick rhoptries seen in the ⌬carp mutants (bottom). (F) IFAs of parental (left) and carp-HA (right) tachyzoites infecting the same host cell labeled with anti-ROP7 (green) and anti-HA (red) that were analyzed using SR-SIM followed by deconvolution and 3D reconstruction to reveal typical rhoptry bulb morphology in parental tachyzoites (c) and atypical rhoptry morphology in carp-HA tachyzoites (d). (G) IFA of an evacuole secretion assay, which shows a ⌬carp tachyzoite successfully secreting evacuoles labeled by anti-ROP7. This suggests that ⌬carp rhoptries are capable of secreting rhoptry evacuoles after treatment with cytochalasin D.
To establish whether the rhoptries of ⌬carp tachyzoites were still capable of secreting ROP proteins into the host cell, we performed an evacuole secretion assay with the ⌬carp mutants. ⌬carp tachyzoites were capable of secreting evacuoles (labeled with anti-ROP7) into the host cell, suggesting that the rhoptries of ⌬carp tachyzoites are functional in secretion despite their altered morphology (Fig. 6G).

DISCUSSION
We report here that the gene TgCA_RP, present in the T. gondii genome, encodes a catalytically inactive carbonic anhydrase-related protein (CARP) that is closely related to the recently identified -class carbonic anhydrase present in P. falciparum (PfCA) (10,11). Given that the closest active relative of TgCA_RP, PfCA, shares only 26% sequence identity, it was not surprising that our attempts to reconstitute activity by changing 1 to 2 amino acids of the Zn 2ϩ binding domain with His were insufficient to restore activity to rTgCA_RP. PfCA is the only active member of the -class that has been characterized, and no mutagenesis studies have been performed to identify essential residues. There are likely important residues for catalytic activity in -class CAs other than the Zn 2ϩ coordination site.
TgCA_RP localizes to both mature and nascent rhoptries. An antibody to the proregion of ROP4 was used for identification and colocalization to the nascent rhoptry vesicles (41). TgCA_RP appears to localize to the outer membrane of nascent rhoptries, surrounding the luminal localization of ROP7 and ROP4, suggesting that TgCA_RP may possess a localization signal that is distinct from other rhoptry bulb proteins. TgCA_RP is posttranslationally modified at its C-terminal region with a glycosylphosphatidylinositol (GPI) anchor, and blocking this C-terminal region with an epitope tag altered localization in intracellular but not extracellular tachyzoites. Tachyzoites lacking TgCA_RP display a growth and invasion phenotype in vitro, have atypical rhoptry morphology, and exhibit reduced virulence in a mouse model. Ionic stress is a major challenge that T. gondii confronts when exiting the host cell into the extracellular medium. Dealing with these stressful changes is important for the survival of the parasite, which needs to actively invade other host cells to continue its lytic cycle. Tachyzoites lacking TgCA_RP showed increased sensitivity to extracellular stress, perhaps because poorly developed rhoptries are less stable and more difficult to maintain in a functional state for invasion.
While several GPI-anchored proteins that localize to secretory organelles in apicomplexan parasites have been described, TgCA_RP is the first GPI-anchored protein to localize to the rhoptry bulb of T. gondii. Rhoptry-associated membrane antigen (RAMA) and Pf34, both GPI-anchored proteins, localize to the bulb and neck of the P. falciparum rhoptries, respectively, while T. gondii subtilisin (TgSUB1) localizes to the micronemes (42). The atypical sushi protein (ASP) of P. falciparum (orthologous to T. gondii RON1) is also GPI anchored and localizes to the rhoptry neck (43). Other T. gondii RON proteins are predicted to have a GPI anchor as well, although this has not been experimentally verified (44). The propensity to associate with DRMs has been shown to be important for the appropriate trafficking of some GPI anchor proteins through the secretory pathway (45).
Parasites expressing the tagged version of TgCA_RP in carp-HA mutants express Tg-CA_RP without the GPI anchor. This conclusion is based on the results of two experiments. The first result was the failure of TgCA_RP-HA to be released into the aqueous phase by pretreatment of the parasite lysate with GPI-PLC, while both surface antigen SAG1 and TgCA_RP were released by this method. Second, [ 3 H]inositol did not label TgCA_RP-HA in carp-HA tachyzoites. Disruption of the TgCA_RP GPI anchor attachment domain by the HA tag resulted in mislocalization of a large portion of TgCA_RP-HA protein in intracellular tachyzoites and a rhoptry biogenesis phenotype identical to the one obtained after deletion of the TgCA_RP gene. This defect was not rescued by the partial localization of TgCA_RP-HA to nascent rhoptries. The localization of TgCA_RP-HA was diffuse, and we did not see the clearly defined localization to the periphery of the nascent rhoptries as observed in the parental cell line. Our hypothesis is that the GPI anchor is acting as a localization signal. It is possible that the outer membrane of the nascent rhoptries has a different lipid composition, which favors the specific localization of GPI-anchored proteins (46). The presence of TgCA_RP specifically at the nascent rhoptry membrane or the GPI anchor itself may be necessary for its role in rhoptry biogenesis. We attempted to isolate T. gondii lipid rafts using a density gradient, but the results were not interpretable due to the lack of appropriate lipid raft markers for these parasites. However, we found that TgCA_RP separates with detergent-resistant membranes, typically enriched in lipid raft proteins (47)(48)(49).
There is evidence that CARPs coordinate the function of other proteins by proteinprotein interaction (9). A BLASTp search of the TgCA_RP sequence against the human proteome showed that the CA-like domain of the gamma-type protein tyrosine phosphatase matches TgCA_RP with the highest similarity score. This protein acts as a hydrophobic binding pocket for contactin, a GPI-anchored neuronal surface protein that is important for cell adhesion and outgrowth (50,51). A similar attachment role has been attributed to the CA-like N-terminal domain of the vaccinia virus (9). In addition, some CARPs interact with calcium channels (52). It is possible that TgCA_RP interacts in a similar way with an unknown binding partner in the nascent rhoptry membranes, promoting adhesion and fusion of the nascent rhoptry vesicles to form the mature organelle. The region containing the annotated CA domain of TgCA_RP was previously shown to be exposed to the cytosol, which would be consistent with this hypothesis (30). In this hypothetical scenario, the absence of TgCA_RP would cause a delay in the assembly of nascent rhoptry vesicles, which would become disorganized. This would explain the delayed maturation and unusual shape of the rhoptries that we observed in our extracellular ⌬carp mutants.
In summary, this is the first report of a GPI-anchored CARP in T. gondii. TgCA_RP localizes to the outer membrane of nascent rhoptries, where it plays a role in the biogenesis of the organelle. TgCA_RP appears to be important for rhoptry shape and maturation. When TgCA_RP is not expressed, parasites become less fit for growth in vitro and in vivo. Rhoptries in these mutants are probably suboptimally functional. An interesting and unique aspect of TgCA_RP is its similarity to the -class carbonic anhydrase described in Plasmodium, and its inactivity, even when the proposed Zn 2ϩ coordinating histidine residues were restored. Our interpretation is that other features, apart from the proposed Zn 2ϩ coordination domain, differentiate TgCA_RP from PfCA. The T. gondii CARP is the first CARP described in protists that does not belong to the ␣-class, suggesting that the evolutionary origin of these carbonic anhydrase-related proteins may be earlier than previously anticipated.

MATERIALS AND METHODS
Analysis of sequence and alignment. Alignment was performed using the T-Coffee web server (53). One unified alignment was performed, and the sequences used are shown in Table S1 in the supplemental material. The resulting alignment was edited in Jalview2 (54). Poorly aligned N-terminal regions were trimmed.
Parasite cultures and generation of mutants. Tachyzoites of the T. gondii RH ⌬ku80 ⌬hxgprt (33) and RH ⌬ku80 TATi (55) strains were cultured in human foreskin fibroblasts (HFF) or hTert (human telomerase reverse transcriptase) fibroblasts in Dulbecco's modified Eagle's medium (DMEM) supplemented with 1% fetal bovine serum (FBS), 1 mM sodium pyruvate, and 2 mM glutamine. Tachyzoites were obtained from infected hTert cells by passing them through a 25-gauge needle or otherwise collected from the supernatant of infected cells after natural egress. The RH ⌬ku80 TATi strain was obtained from Boris Striepen (University of Georgia), and the RH ⌬ku80 ⌬hxgprt strain was obtained from David Bzik (Dartmouth Medical School).
For C-terminal tagging of the TgCA_RP gene, the 3= 1,459 bp (minus the stop codon) of the gene annotated as a hypothetical protein, TgME49_297070, was amplified using primers P1 and P2 (see Table S3 in the supplemental material), which added the sequence required for ligation-independent cloning. The PCR product was purified using the Qiaex II gel extraction kit (Qiagen) and cloned into the pLIC-3ϫHA-CAT plasmid. The purified PCR product and plasmid were treated and combined as described in the work of Huynh and Carruthers (33). Eighty micrograms of the sterilized plasmid pTgCA_RP-3ϫHA-CAT was transfected into 1 ϫ 10 7 RH ⌬ku80 TATi (56). Transfected tachyzoites were selected with 20 M chloramphenicol, and clones were isolated by limiting dilution. Genomic DNA of clones was isolated and screened by PCR using a primer upstream of the original amplification from TgCA_RP using forward primer P3 and a downstream pLIC-3ϫHA-CAT reverse primer, P4. Clones were further confirmed by Western blot analysis.
Deletion of the TgCA_RP gene was achieved by transfecting ⌬ku80 ⌬hxgprt tachyzoites with a plasmid (pCTY) containing a chloramphenicol cassette flanked by a 1-kb region of the TgCA_RP 5= UTR and 1 kb downstream of the 3= UTR. The parasites were selected with chloramphenicol followed by subcloning. Complementation of TgCA_RP was accomplished by cloning the TgCA_RP gene, including the UTRs and potential promoter region, into the pDTM3 plasmid. The construct was transfected into ⌬carp tachyzoites and selected using pyrimethamine, followed by subcloning.
TgCA_RP expression. TgCA_RP was expressed in E. coli without the 36-amino-acid signal peptide. A PCR product was generated by amplifying a DNA fragment from T. gondii cDNA with primers P5 and P6 (Table S3). The fragment was cloned into the pTRCHisB expression vector (Invitrogen) using BamHI and XhoI restriction sites. The TgCA_RP-pTRCHis construct was sequenced and transformed into E. coli BL21(DE3) RIPL. Cells were induced for 2 h at 37°C with 1 mM isopropyl-␤-D-thiogalactopyranoside (IPTG). The inclusion bodies containing rTgCA_RP were isolated using the protocol described in the GE Healthcare recombinant protein purification manual. Inclusion bodies were solubilized with 0.02% SDS in PBS, pH 11. pH was adjusted to 7.4 prior to the inoculation.
For enzyme activity, soluble recombinant protein (TgCA_RP) was expressed using the Pet32Lic/EK vector (Novagen) system. Two fragments from TgCA_RP of the previously amplified TgCA_RP cDNA were used. rTgCA_RPa (aa 121 to 445) was amplified using P7 and P8, and rTgCA_RPb (aa 94 to 488) was amplified using primers P9 and P10 (Table S3; Fig. S2). Mutagenesis of F233H (primers P11 and P12) and Q254H (primers P13 and P14) ( Table S3) was performed using the QuikChange mutagenesis kit (NEB) and confirmed via sequencing. LB was supplemented with 125 M ZnSO 4 prior to induction with IPTG to ensure adequate Zn 2ϩ for proper folding of the recombinant CA.
Carbonic anhydrase activity assays. The p-nitrophenyl acetate hydrolysis assays were performed as described by Armstrong et al. (26) with some modifications. A fresh 3 mM stock pNPA solution was prepared for each experiment by dissolving the compound in acetone followed by water dilution to the final concentration of 3 mM. Readings were taken in a plate or cuvette in a SpectraMax e 2 plate reader at the isosbestic wavelength (348 nm) of p-nitrophenol and the p-nitrophenylate ion. Numerous buffers (25 mM Tris-HCl, 25 mM Tris-SO 4 , 25 mM Tricine-KOH, 25 mM HEPES-KOH, 50 mM MES, and 50 mM MOPS) and a range of pH conditions (6.0 to 8.0) were tested. In-gel CO 2 hydration assays were performed as described by Ramanan et al. (27) with the addition of 1 M ZnSO 4 in the Triton X-100 refolding buffer. Commercially available bovine carbonic anhydrase (Sigma catalog no. C3934) was used as a positive control for activity.
Antibody production and affinity purification. Antibodies to recombinant TgCA_RP were generated in mice and guinea pigs. Mouse antibodies were produced via intraperitoneal inoculation of six CD-1 mice (Charles River, Inc.) with 100 g of rTgCA_RP. Mice were boosted twice with 50 g of rTgCA_RP. Final serum was collected by cardiac puncture after CO 2 euthanasia. Guinea pig antibodies were produced via subdermal inoculation of Hartley strain guinea pigs (Charles River, Inc.) with 200 g of solubilized rTgCA_RP and mixed with complete Freund's adjuvant. Guinea pigs were boosted twice with 100 g of solubilized rTgCA_RP. Final serum was collected using cardiac puncture. Inoculations and final serum collections were performed under anesthesia with isoflurane.
The antibodies used in IFAs were either affinity purified (mouse anti-TgCA_RP) or affinity purified and adsorbed (guinea pig anti-TgCA_RP). Adsorption was performed by cross-linking lysate from ⌬carp tachyzoites to cyanogen bromide-activated resin (Sigma) according to the commercial Sigma protocol. One milliliter of anti-TgCA_RP guinea pig serum was incubated with beads for 2 h at room temperature, and the supernatant was collected for affinity purification. Affinity purification was performed by running 500 g of purified inclusion bodies on a single-well SDS-PAGE gel followed by transfer to a nitrocellulose membrane. The TgCA_RP band region was excised and incubated overnight with 200 l (mouse) or 1 ml (guinea pig) of serum diluted with 5 ml of incubation buffer (50 mM Tris-HCl, pH 10.2, 0.45 M NaCl, 0.01% Tween 20). The membrane was washed with PBS, pH 7.4. Antibody was eluted from the membrane twice using 500 l of stripping buffer (0.2 M glycine-HCl, pH 2.4, 0.5 M NaCl), which was pipetted over the membrane for 2 min and then equilibrated to pH 7.5 with 1 M Tris-HCl, pH 9.5.
Western blot analysis and immunofluorescence assays. Purified tachyzoites were treated with cell lysis buffer M (Sigma) and 25 units of Benzonase (Novagen) for 5 min at room temperature followed by addition of an equal volume of 2% SDS-1 mM EDTA solution. Total protein was quantified with a NanoDrop spectrophotometer (Thermo Scientific). Samples were resolved using a 10% bis-acrylamide gel in a Tris-HCl SDS buffer system (Bio-Rad). Gels were either transferred for Western blot analysis or stained with amido black. Primary antibody dilutions were as follows: anti-HA, 1:50 (monoclonal rat; Roche); mouse anti-TgCA_RP, 1:1,000; guinea pig anti-TgCA_RP, 1:500. Secondary horseradish peroxidase (HRP) antibodies were used at 1:10,000 dilutions.
Immunoelectron microscopy. Extracellular T. gondii endogenously expressing the C-terminal 3ϫHA tag (TgCA_RP-HA) was washed twice with PBS before fixation in 4% paraformaldehyde (Electron Microscopy Sciences, PA) in 0.25 M HEPES (pH 7.4) for 1 h at room temperature and then in 8% paraformaldehyde in the same buffer overnight at 4°C. Parasites were pelleted in 10% fish skin gelatin, and the gelatin-embedded pellets were infiltrated overnight with 2.3 M sucrose at 4°C and frozen in liquid nitrogen. Ultrathin cryosections were incubated in PBS and 1% fish skin gelatin containing mouse anti-HA antibody at a 1/5 dilution and then exposed to the secondary antibody that was revealed with 10-nm protein ␣-gold conjugates. Sections were observed and images were recorded with a Philips CM120 electron microscope (Eindhoven, the Netherlands) under 80 kV. For confirmation of rhoptry localization of TgCA_RP, ultrathin cryosections of T. gondii-infected cells were processed as described for extracellular parasites and immunolabeled with guinea pig anti-TgCA_RP and mouse anti-ROP7 antibodies at 1:25 and 1:750 dilutions, respectively, and detected with protein A-gold conjugates of 10 nm for TgCA_RP and 5 nm for ROP7.
Invasion and growth assays. Red-green invasion assays were performed as described in reference 58 with modifications. The number of tachyzoites used was 2 ϫ 10 7 , and invasion was for 5 min. Plaque growth and plaquing efficiency assays were performed as previously described (35) with modifications. For plaque growth assays, 125 tachyzoites were used for infection of hTert fibroblasts followed by an incubation time of 10 days prior to fixing and staining. Plaquing efficiency assay plates were infected with 2,000 tachyzoites for 1 h prior to washing and then allowed to incubate for 6 days prior to fixation and staining with crystal violet. For intracellular growth assays, coverslips with hTert fibroblasts were cooled to 4°C and infected with 5 ϫ 10 5 tachyzoites. Coverslips were first incubated for 15 min on ice followed for 20 min at 37°C. Coverslips were washed three times with sterile PBS. Fresh DMEM, high glucose (DMEM HG), plus 1% fetal bovine serum (FBS) was added and incubated at 37°C for 25 h prior to fixation with 3% paraformaldehyde. Giemsa staining was done as described previously (61). Vacuoles containing 2, 4, 8, 16, and 32 tachyzoites were counted.
Virulence tests in mice. For each experiment, cohorts of 5 female adult Swiss Webster mice (Charles River, Inc.) were inoculated intravenously via the tail vein with PBS or 50, 150, or 500 tachyzoites of the parental (⌬ku80 Δhxgprt), ⌬carp, or ⌬carp-cm strain. The tachyzoites were mechanically released via syringe from the host cell monolayer and resuspended in 100 l of sterile PBS at the proper concentration. Mice were observed several times daily. Mice were humanely euthanized when the illness score exceeded the level deemed acceptable by the established animal use protocol. Mice that survived the initial challenge with 50 tachyzoites were rechallenged by intraperitoneal inoculation with 1,000 RH-rfp tachyzoites, examined daily for signs of illness, and euthanized if serious illness was observed. Mouse experiments in this work followed a reviewed and approved protocol by the Institutional Animal Care and Use Committee (IACUC). Animal protocols followed the U.S. Government principles for the Utilization and Care of Vertebrate Animals. The protocol was reviewed and approved by the University of Georgia IACUC (protocol number A2015-02-025-R2).
PI-PLC treatments and detergent extractions. Naturally egressed tachyzoites were lysed via freeze-thawing and treated (or mock treated) with 0.4 units of GPI-specific phospholipase C (PI-PLC) from Bacillus cereus (Sigma) for 2 h at 37°C. Detergent extractions and acetone precipitations were performed as previously described (42). Isolation of detergent-resistant domains from naturally egressed tachyzoites was performed as previously described (59). myo-[ 3 H]inositol labeling. Parental (⌬ku80 ⌬hxgprt), ⌬carp, and carp-HA tachyzoites were used to infect hTert monolayers in T12.5 flasks containing DMEM-high-glucose medium with 1% FBS. The monolayer was washed 24 h after the initial infection with Hanks balanced salt solution, and the medium was replaced with DMEM HG without inositol (MP Biologicals) containing 100 Ci myo-(2,3)-[ 3 H]inositol (American Radiolabeled Chemicals). Parasites were allowed to egress naturally (~18 h) and were purified and washed as described previously (57). SDS-PAGE and Western blot analyses followed, and the nitrocellulose membranes were exposed to a radiography film (Denville Scientific) for 40 days before development.
Rhoptry morphology. For freeze fracture analysis, tachyzoites were fixed in 2.5% glutaraldehyde and processed as previously described (60). The material was washed in 0.1 M sodium cacodylate buffer and infiltrated in ascending concentrations of glycerol. The cells were maintained in 15% (vol/vol) glycerol for 12 h at 4°C. Cells were then infiltrated with 30% glycerol for 3 h, pelleted, mounted on support disks, and frozen in the liquid phase of Freon 22 cooled by liquid nitrogen. Frozen specimens were transferred and fractured in a BAF 060 freeze fracture apparatus (Baltec). Immediately after fracturing, the exposed surfaces were shadowed with platinum at an angle of 45°and carbon at an angle of 90°. Metal replicas were cleaned in sodium hypochlorite, collected on Formvar-covered nickel grids, and examined in a Tecnai Spirit Biotwin microscope (FEI, the Netherlands) operating at 120 kV.
For routine transmission electron microscopy, cells were fixed in 2.5% glutaraldehyde and postfixed in 1% osmium tetroxide, 1.25% potassium ferrocyanide, and 5 mM CaCl 2 in 0.1 M cacodylate buffer (pH 7.2). Samples were washed in the same buffer, dehydrated in an ascending acetone series, and embedded in epoxy resin. Ultrathin sections were collected and observed in a JEOL JEM-1210 transmission electron microscope (JEOL, Tokyo, Japan).
For rhoptry morphology quantification assays, tachyzoites were allowed to invade host cell monolayers on glass coverslips. After 24 h, infected cells were fixed and IFAs was performed with anti-ROP7 as described above. For each coverslip, 100 parasitophorous vacuoles were counted. A vacuole containing at least one rhoptry of "normal," linear morphology (Fig. 3B) was counted as normal.
Statistical analyses. All statistical analyses were performed using GraphPad Prism software (version 7).