Sequential processing of the Toxoplasma apicoplast membrane protein FtsH1 in topologically distinct domains during intracellular trafficking

Molecular & Biochemical Parasitology 166 (2009) 126–133

Contents lists available at ScienceDirect

Molecular & Biochemical Parasitology

Sequential processing of the Toxoplasma apicoplast membrane protein FtsH1 in topologically distinct domains during intracellular trafficking Anuradha Karnataki a,b , Amy E. DeRocher a , Jean E. Feagin a,b,1 , Marilyn Parsons a,b,∗ a b

Seattle Biomedical Research Institute, 307 Westlake Ave. N., Seattle, WA 98109, USA Interdisciplinary Program in Pathobiology, Department of Global Health, University of Washington, Seattle, WA 98195, United States

a r t i c l e

i n f o

Article history: Received 17 December 2008 Received in revised form 5 March 2009 Accepted 6 March 2009 Available online 21 March 2009 Keywords: Apicoplast Toxoplasma Protease Protein processing Membrane protein Chloroplast AAA protein FtsH

a b s t r a c t FtsH proteins are hexameric transmembrane proteases found in chloroplasts, mitochondria and bacteria. In the protozoan Toxoplasma gondii, FtsH1 is localized to membranes of the apicoplast, a relict chloroplast present in many apicomplexan parasites. We have shown that although T. gondii FtsH1 lacks the typical bipartite targeting presequence seen on apicoplast luminal proteins, it is targeted to the apicoplast via the endoplasmic reticulum. In this report, we show that FtsH1 undergoes processing events to remove both the N- and C-termini, which are topologically separated by the membrane in which FtsH1 is embedded. Pulse-chase analysis showed that N-terminal cleavage precedes C-terminal cleavage. Unlike the processing of the N-terminal transit peptide of luminal proteins, which occurs in the apicoplast, analysis of ER-retained mutants showed that N-terminal processing of FtsH1 occurs in the endoplasmic reticulum. Two of four FtsH1 mutants bearing internal epitope tags accumulated in structures peripheral to the apicoplast, implying that FtsH1 trafficking is highly sensitive to changes in protein structure. These mutant proteins did not undergo C-terminal processing, suggesting that this processing step occurs after localization to the plastid. Mutation of the peptidase active site demonstrated that neither processing event occurs in cis. These data support a model in which multiple proteases act at different points of the trafficking pathway to form mature FtsH1, making its processing more complex than other FtsHs and unique among apicoplast proteins described thus far. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Apicomplexan parasites are responsible for considerable suffering worldwide, both directly in the case of the human pathogens and indirectly in the case of pathogens affecting farm animals. This phylum of obligate intracellular parasites includes Toxoplasma gondii, an important opportunistic pathogen of AIDS patients [1] and a causative agent of severe birth defects [2]. It also includes Plasmodium spp., which cause approximately two millions deaths each year from malaria [3]. The drugs currently available to treat these diseases are either not very effective, poorly tolerated, or facing resistance [4]. Thus there is an urgent need to identify new drugs and drug targets. The apicoplast is an organelle present in most medically important apicomplexans, with the exception of

Abbreviations: CP, C-terminally processed; DAPI, 4,6-diamidino-2-phenylindole; ER, endoplasmic reticulum; mAb, monoclonal antibody; NP, N-terminally processed; TMD, transmembrane domain. ∗ Corresponding author at: Seattle Biomedical Research Institute, 307 Westlake Ave. N., Seattle, WA 98109, USA. Tel.: +1 206 256 7315; fax: +1 206 256 7229. E-mail address: [email protected] (M. Parsons). 1 Current address: Dept. of Pharmacy, University of Washington, Seattle, WA 98195, USA. 0166-6851/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.molbiopara.2009.03.004

Cryptosporidium parvum [5], but it is absent in the animal host. The apicoplast is home to several metabolic pathways, including the type II fatty acid synthesis pathway [6] as well as part of the type II heme biosynthesis pathway [7]. The type II isoprenoid synthesis pathway is also apicoplast-localized in P. falciparum [8], although its presence in the apicoplast of T. gondii is not yet confirmed [9]. The apicoplast and the pathways it compartmentalizes have been shown to be essential in both T. gondii and P. falciparum [8,10,11], and hence are of considerable interest as potential drug targets. The apicoplast is a remnant chloroplast that is thought to have been acquired by secondary endosymbiosis wherein an ancestral apicomplexan engulfed an alga and appropriated its chloroplast. The result is a plastid surrounded by four membranes [12]. The two inner membranes are thought to be homologues of chloroplast membranes, the third is proposed to be homologous to the plasma membrane of the algal cell, and the outermost membrane is thought to be derived from the endomembrane system of the apicomplexan [for review see [13]]. The apicoplast has its own genome, whose organization resembles that of chloroplasts, albeit highly reduced (∼35 kb). It mainly encodes housekeeping genes such as rRNAs and tRNAs and thus most proteins required for the organelle’s function are encoded in the nucleus. Those destined for

A. Karnataki et al. / Molecular & Biochemical Parasitology 166 (2009) 126–133

127

Fig. 1. Scale schematic of the FtsH1 protein and expression constructs. Unique regions at N-terminus (aa 1-347) and C-terminus (aa 903-1250) as well as AAA (aa 386-691) and peptidase (aa 698-902) domains are indicated. The transmembrane domain (TMD), Walker A and B motifs that function in ATP binding (W), and zinc binding (Zn) sites are marked. Beneath the schematic, the major isoforms discussed in the manuscript are shown as bars (the N-terminal cleavage site is approximate). The table below presents diagrams of the expression constructs described herein. V5 tags are indicated by triangles and HA tags are indicated by balls. Mutation at the peptidase active site residue is indicated by “X”. Shown at right is the localization of the expressed proteins as revealed by anti-V5 mAb (except for the mutants that lacked a V5 tag): Api, apicoplast; ER, endoplasmic reticulum. Cleavage (+) or lack thereof (−) is indicated for each protein. Proteins which were not assessed for cleavage due to the position of the tag and truncated proteins that lack the cleavage site are marked “NA”.

the apicoplast lumen are targeted there by virtue of an N-terminal targeting sequence. This consists of a signal sequence followed by a transit peptide, both of which are essential for proper localization [14–16]. The signal sequence is required for entry into the endoplasmic reticulum (ER), and is cleaved upon import. The exposed transit peptide then targets the protein to the apicoplast, where it too is cleaved, resulting in the formation of the mature protein [17]. Recently, we identified the first apicoplast membrane proteins in Toxoplasma: APT1 (apicoplast phosphate translocator 1) [18] and the zinc-dependent protease FtsH1 (so named for the filamentous phenotype observed in the original E. coli mutants) [19]. Both proteins lack the targeting sequences characteristic of luminal proteins, but appear to traffick to the organelle via the ER. Immunoelectron microscopy showed apparent residence of both proteins in multiple membranes of the apicoplast, as well as in vesicles that may serve to transport the molecules to the apicoplast. Furthermore, both proteins behaved as integral membrane proteins in biochemical studies. Recently, Tic20, a protein of the innermost membrane of apicoplast, was identified in T. gondii and shown to bear a signal and transit peptide [20]. The authors showed that Tic20 is essential for the survival of the parasite and that it likely plays an indirect role in import of proteins across the innermost membrane of the apicoplast. FtsHs are ubiquitous proteins found in prokaryotes as well as in the mitochondria and chloroplasts of eukaryotes. All FtsHs described to date exist as hexamers that are either homo-oligomers or hetero-oligomers [21,22]. FtsHs are transmembrane metalloproteases that possess AAA domains (ATPase associated with several cellular activities), and require ATPase function for protease activity. The main function of FtsHs identified thus far is to maintain quality control by degrading misassembled and damaged membrane proteins. For example, chloroplast FtsHs degrade proteins damaged by photooxidation, and mutants are typically defective in photosynthesis [23]. Similarly, Saccharomyces cerevisiae lacking the mitochondrial FtsH Yta10p is defective in respiration [24]. Escherichia coli FtsH (HflB) as well as FtsH homologues in the mitochondrial inner membrane of Saccharomyces cerevisiae, Yme1 and the Yta10-12 complex, have additionally been shown to function as chaperones [25–27]. Recently, Yme1 was also shown to be required

for import of polynucleotide phosphorylase to the intermembrane space of mitochondria [28]. FtsH is essential for division of E. coli, but the function of T. gondii FtsH1 and whether it is essential remain unknown. T. gondii FtsH1 encodes a 1250 aa protein with a single transmembrane domain (TMD) (Fig. 1). It has a long N-terminal extension that does not show sequence homology to any predicted proteins. This region, which contains the TMD, is followed by the hallmark AAA and peptidase domains. The protein ends in another unique region. We have previously shown that both the TMD and the peptidase domain are required for proper targeting of FtsH1 to the apicoplast [19], although the protein’s topology within the apicoplast membranes remains unknown. FtsH1 tagged at the N-terminus localizes to the apicoplast, and shows two bands on immunoblot analysis, which suggested that the protein undergoes either post-translational modification or processing at the C-terminus. Here we demonstrate that FtsH1 is proteolytically processed at the N-terminus as well as the C-terminus. In addition to being the first FtsH to undergo multiple processing events, T. gondii FtsH1 is also the first example of an apicoplast protein to exhibit processing at both termini. 2. Materials and methods 2.1. Cell culture and transfections T. gondii strain RHhxgprt and its derivatives [29], including a line expressing apicoplast-targeted Heteractis crispa red fluorescent protein (S+TACP -HcRed) [18], were grown and maintained in human foreskin fibroblasts as described previously [30]. Plasmids digested with NotI [31] were transfected into these lines and stable transfectants were obtained using xanthine and mycophenolic acid as selection reagents. Limiting dilution was used to obtain clonal lines. In previous studies, a single V5-FtsH1 transfectant showed a weak band at 151 kDa on immunoblot analysis using anti-V5 mAb, in addition to the bands described in this report [19]. Since this band appeared in only one of many transfectants, it is likely an integration artifact, and so other transfectants were used for the current studies.

128

A. Karnataki et al. / Molecular & Biochemical Parasitology 166 (2009) 126–133

2.2. Plasmids The derivations of T. gondii FtsH1 expression constructs pFtsH1HA, as well as pV5-FtsH1, have been described previously [19]. All of the FtsH1 constructs, which are diagrammed in Fig. 1, employ the FTSH1 promoter to drive expression and for each the correct open reading frame was verified by DNA sequencing. pV5-FtsH1HA was obtained by adding two copies of V5 tag at the N-terminus of pFtsH1-HA using a 2V5 linker (annealed and phosphorylated oligomers 2V5/s and /as) after insertion of Mun1 and Spe1 restriction sites (F1-Mun1Spe1-N-ter/s and/as) at the N-terminus (see Table 1 for all primers). Site-directed mutagenesis was performed using the indicated primers in PCR reactions employing Pfu polymerase, followed by digestion with DpnI. A deletion construct encoding the first 354 with four copies of HA tag at the Cterminus of FtsH1 was generated from pFtsH1-HA by mutagenesis using primers F1-354-4HA/s and /as. The transmembrane domain of FtsH1(1-354)-HA was deleted using mutagenesis primers F1TMD/s and /as. The proteolytically inactive mutant E732Q was obtained by site-directed mutagenesis of pV5-FtsH1-HA using F1E732Q/s and /as primers. Several constructs were generated with internal tags. A 3HA linker (annealed and phosphorylated oligomers 3HA/s and 3HA/as) was introduced at aa 325 after insertion of Nhe1 and Pst1 restriction sites (primers F1-Nhe1Pst1/s and /as) to obtain V5-FtsH1-HA325 . Two HA tags were added by site-directed mutagenesis, replacing aa 932 to 950 (F1-2HA 932-950/s and /as primers) to yield V5-FtsH1-HA932 . Two copies of the V5 tag were inserted at aa 233 FtsH1-HA using linker 2V5 (annealed and phosphorylated oligomers 2V5/s and /as) after insertion of Mun1 and Spe1 restriction sites (F1-Mun1Spe1-233/s and /as)) at aa 233 to yield FtsH1-V5233 -HA. 2.3. Pulse-chase, immunoprecipitation and immunoblot analysis Intracellular parasites (∼108 ) were labeled for the indicated times with 100 ␮Ci/ml [35 S] trans label (methionine and cysteine, MP Biomedicals, Irvine, CA) using methionine- and cysteine-free medium under standard growth conditions as described [32]. Subsequently the labeling medium was replaced with complete medium and the incubation continued for various times. After rinsing, the fibroblast layer was scraped from the flask and cells were pelleted by centrifugation at 2300 × g for 2 min. Pellets were lysed in 0.5 ml SK lysis buffer (150 mM NaCl, 50 mM Tris HCl pH 7.5, 2 mM EDTA, 1% NP-40, 0.25% deoxycholate, 1.7 ␮g/ml aprotinin, 5 ␮g/ml leupeptin, 1 ␮M pepstatin, 0.1 mM PMSF). Epitope-tagged FtsH1

was immunoprecipitated using monoclonal anti-V5 mouse IgG2a at 10 ␮g/ml (Invitrogen) followed by Protein G coupled to magnetic beads (Invitrogen). The washed immune complexes were resuspended in Laemmli sample buffer and placed in a boiling water bath for 5 min. Samples were separated by SDS-PAGE on 7.5% acrylamide gels, then transferred to nitrocellulose membrane. Radiolabeled proteins were detected by phosphorimaging using a Storm 860 (Molecular Dynamics). Approximately 107 parasites were used for immunoblot analyses, unless otherwise noted. Cell lysates were separated by SDS-PAGE on 7.5% gels except where indicated, followed by transfer to nitrocellulose membranes. After blocking in Odyssey block (LI-COR Biosciences), blots were probed with mouse anti-V5 monoclonal antibody (mAb) at 0.1 ␮g/ml (Invitrogen), mouse anti-HA mAb Mono HA.11 MMS-101P (0.4 ␮g/ml, Covance), polyclonal rabbit anti-HA (1 ␮g/ml, Thermo Scientific) or rat anti-HA 3F10 mAb (Roche, 0.4 ␮g/ml). This was followed by goat anti-mouse Ig coupled to IRDye 800CW (1:10,000, LI-COR) or goat anti-rabbit Ig coupled to IRDye 680 (1:10,000, LI-COR). Analysis of blots was done on the LI-COR Odyssey infrared imaging system. For the immunoblot in the pulse-chase experiment (Fig. 4, left), the anti-V5 mAb was covalently conjugated to DyLight 800 (Thermo Scientific). 2.4. Microscopy Immunofluorescence analysis (IFA) of intracellular parasites was performed as described previously [18]. The V5 tag was detected using anti-V5 mouse mAb IgG2a at 1 ␮g/ml (Invitrogen), followed by goat anti-mouse IgG2a FITC 2 ␮g/ml or goat anti-mouse IgG2a Texas Red 2 ␮g/ml (Southern Biotechnology). The HA tag was detected using FITC-coupled rat anti-HA mAb 3F10 3 ␮g/ml (Roche). S+TACP -HcRed (S+T-Red) was used as an apicoplast luminal marker [18] and 4,6-diamidino-2-phenylindole (DAPI) was used to stain the DNA. A Deltavision RT deconvolution microscope with an Olympus UPlan/Apo 100× 1.35NA objective was used to view the slides. Images were deconvolved using standard parameters and a conservative ratio algorithm. Single deconvolved planes are shown. 3. Results 3.1. FtsH1 is processed at both the N- and C-termini TgFtsH1 encodes a 1250 amino acid protein that has unique extensions at both the N- and C-termini as compared to other FtsHs, as diagramed in Fig. 1. This cartoon shows the transmembrane,

Table 1 Oligonucleotides. Primer ID

Primer sequence (5 to 3 )a

F1-Mun1Spe1-N-ter/s F1-Mun1Spe1-N-ter/as F1-354-4HA/s F1-354-4HA/as F1-TMD/s F1-TMD/as F1-E732Q/s F1-E732Q/as F1-Nhe1Pst1/s F1-Nhe1Pst1/as 3HA/s 3HA/as F1-2HA 932-950/s F1-2HA 932-950/s F1-Mun1Spe1-233/s F1-Mun1Spe1-233/as 2V5/s 2V5/as

GTTTTCTCTTCGACCACCAATTGTATACTAGTATGAGTTCTCGTAACG CGTTACGAGAACTCATACTAGTATACAATTGGTGGTCGAAGAGAAAAC CTCACCTTCAACGATTTGGCCCCTAGGATGTACCCTTACGATGTCC GGACATCGTAAGGGTACATCCTAGGGGCCAAATCGTTGAAGGTGAG GAGGACGCATGTGGGCAGCCCGTCTTCTCCACTCTCTCCACGATGG CCATCGTGGAGAGAGTGGAGAAGACGGGCTGCCCACATGCGTCCTC GGCAGCTGACTGCGTACCACCAGGCTGGGCATGCAATC GATTGCATGCCCAGCCTGGTGGTACGCAGTCAGCTGCC CATATGATGCATAATACGACAGATTCAGCTAGCAAACTGCAGTCGTCGGGTTCAAACGTC GACGTTTGAACCCGACGACTGCAGTTTGCTAGCTGAATCTGTCGTATTATGCATCATATG CTAGCTACCCTTACGATGTCCCTGATTACGCTTACCCTTACGATGTCCCTGATTATGCATACCCTTACGATGTCCCTGATTACGCACTGCA GTGCGTAATCAGGGACATCGTAAGGGTATGCATAATCAGGGACATCGTAAGGGTAAGCGTAATCAGGGACATCGTAAGGGTAG CGAGCTTCCTGCAACGCCTTCCTAGGTACCCTTACGATGTCCCTGATTACGCTTACCCTTACGATGTCCCTGATTATGCAGAGTTTGCGGCAGTTGAGGC GCCTCAACTGCCGCAAACTCTGCATAATCAGGGACATCGTAAGGGTAAGCGTAATCAGGGACATCGTAAGGGTACCTAGGAAGGCGTTGCAGGAAGCTCG GGAGTAACCCCGTTCTTCCAGAACAATTGTATACTAGTAAGACTCAAGGCGACGCTCCT AGGAGCGTCGCCTTGAGTCTTACTAGTATACAATTGTTCTGGAAGAACGGGGTTACTCC AATTGAAAATGGGCAAGCCTATCCCTAACCCTCTCCTCGGCCTCGACTCTACGGGCAAGCCTATCCCTAACCCTCTCCTCGGCCTGGACTCGACGA CTAGTCGTCGAGTCCAGGCCGAGGAGAGGGTTAGGGATAGGCTTGCCCGTAGAGTCGAGGCCGAGGAGAGGGTTAGGGATAGGCTTGCCCATTTTC

a

Restriction sites, both complete and partial, are underlined.

A. Karnataki et al. / Molecular & Biochemical Parasitology 166 (2009) 126–133

AAA, and peptidase domains of FtsH1, as well as the constructs used in this study. Fig. 1 also summarizes the results presented below on the localization and processing of the encoded proteins. We had previously observed that cell lines expressing FtsH1 with an N-terminal V5 tag, one of which is depicted in Fig. S1A, showed bands at 170 kDa and 140 kDa on immunoblots probed with anti-V5 mAb. The 170 kDa species is considerably larger than the predicted mass (140 kDa). A cytosolic mutant FtsH1 lacking the transmembrane domain (TMD) of V5-FtsH1 [19] also showed aberrantly slow migration (Fig. S1B). Since this protein did not enter the secretory system, it is unlikely that glycosylation is responsible for the slow migration of FtsH1. Analysis of deletion mutants mapped the region responsible for this aberrant migration to the first 354 aa (Fig. S1A). These data suggested that the 170 kDa molecule is the full-length protein and that the 140 kDa molecule represents a C-terminally processed species. If so, the 170 kDa molecule should possess both termini and the 140 kDa species would lack the C-terminus. To test this, we created a construct bearing two copies of the V5 tag at the N-terminus and four copies of the HA tag at the C-terminus (V5-

129

FtsH1-HA). We previously demonstrated that in cells expressing V5-FtsH1, anti-V5 staining shows a predominantly circumplastid distribution. As the plastid enlarges during the cell cycle, circumplastid staining is accompanied by some dispersed staining [19]. In stable transfectants expressing the double-tagged mutant protein, a similar pattern was seen with anti-V5 mAb, with most cells showing a circumplastid distribution (Fig. 2A, top). In this and other IFA analyses, a red fluorescent protein that localizes to the apicoplast (S+T-Red) [18] was used as the luminal marker, and cells were also stained with the DNA dye DAPI. When lysates from cells expressing V5-FtsH1-HA were simultaneously probed for the N-terminal V5 and C-terminal HA tags on immunoblot analysis only the 170 kDa band was detected by both antibodies. Thus it corresponds to the full-length protein (Fig. 2B). The 140 kDa species did not bear the C-terminal HA tag, indicating that the C-terminus had been removed. Surprisingly, anti-HA detected an additional band at 154 kDa (Fig. 2B). This protein was not seen with anti-V5 mAb, indicating that it lacked the N-terminus. Thus, the 154 and 140 kDa proteins correspond to N-terminally processed (NP) and C-terminally processed (CP) FtsH1, respectively.

Fig. 2. FtsH1 is processed at both termini. Stable transfectants were generated which expressed FtsH1 tagged at the C-terminus with HA plus an N-terminal (V5-FtsH1-HA) or internal V5 tag (FtsH1-V5233- HA). (A) Localization of encoded proteins. For this and other IFA experiments, schematics of the constructs are shown above the pseudocolored IFA images. The cells were stained with the indicated antibodies and DNA was stained with DAPI (blue). In the top and middle panels, plasmids encoding V5-FtsH1-HA or FtsH1-V5233 -HA were transfected into RH strain parasites expressing S+T ACP -HcRed (S+T-Red), an apicoplast luminal marker. Stable transfectants were stained with anti-V5 mAb (green), and, S+T-Red was detected by intrinisic fluorescence (red). In the bottom panels, RH parasites expressing FtsH1-V5233 -HA were stained with anti-V5 mAb (red) and anti-HA (green). Note that HA staining is very weak at the apicoplast, but that areas of anti-HA staining also stain with anti-V5. The apicoplast DNA as revealed by DAPI staining is used as an apicoplast luminal marker. Bar = 2 ␮M. (B) Immunoblot analysis of V5-FtsH1-HA simultaneously probed with mouse anti-V5 mAb and rabbit anti-HA antibodies. The antibodies were visualized with secondary antibodies coupled to anti-mouse Ig IRDye 800 (pseudocolored green) and anti-rabbit Ig IRDye 680 (pseudocolored red), respectively, which allowed concurrent imaging on the Odyssey infrared system. Predicted molecular masses for each band are indicated and were calculated from the migration of molecular markers (not shown). The region of the gel above 100 kDa is shown, and no smaller bands were reproducibly detected, although occasionally bands around 50 kDa were seen as in C. F, Full-length; NP, N-terminally processed; CP, C-terminally processed. (C) Immunoblot analysis of FtsH1-V5233 -HA and V5-FtsH1-HA using anti-V5 mAb showing relative abundance of the detected isoforms. FtsH1 is abbreviated as “F”. The 115 kDa (NPCP) band is processed at both termini. The N-terminally processed (NP) forms of the internally tagged protein (shown here) and the N-terminally tagged protein (B) differ slightly in size since processing removes the tag in the latter. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

130

A. Karnataki et al. / Molecular & Biochemical Parasitology 166 (2009) 126–133

Although we are unable to predict with certainty the location of the N-terminal cleavage event due to the aberrant migration conveyed by sequences near the N-terminus, we estimate that cleavage occurs prior to aa 125. The C-terminal cleavage site lies around aa 1015, approximately 90 aa after the peptidase domain (see bars in Fig. 1). The intensities of the different bands varied somewhat in different experiments, but for this construct (V5-FtsH1-HA) the CP species was always more abundant than the full-length molecule, which in turn was more abundant than the NP species. To determine whether many FtsH1 subunits were dually processed, we inserted two copies of the V5 tag at aa 233 of FtsH1-HA to yield FtsH1-V5233 -HA. A tag at this location, which is prior to the predicted TMD, should be present on both NP and CP FtsH1. IFA analysis was performed on stable transfectants to determine whether the tagged protein trafficked correctly to the apicoplast. Antibodies to the internal V5 tag revealed a predominantly circumplastid pattern, with staining surrounding the apicoplast lumen (Fig. 2A, middle panel). Interestingly, when these cells were probed with antibodies to the C-terminal HA tag, only dispersed spots were detected (Fig. 2A, bottom panel; due to the fluorochromes used, the apicoplast luminal marker was the DAPI-stained apicoplast genome). Anti-V5 also weakly detected most of these same regions, arguing that they represent proteins bearing both tags. A predominantly dispersed localization was also seen with FtsH1 bearing only the C-terminal HA tag (as detected by anti-HA, not shown). These findings suggest that the FtsH1 isoforms localized to the apicoplast lack the C-terminus bearing the HA tag. We therefore used N-terminal or internal tags to assess apicoplast localization. Immunoblots of cells expressing FtsH1-V5233 -HA were probed with anti-V5 mAb. The results showed the presence of the expected bands corresponding to the full-length, NP, and CP species previously seen, plus an additional, prominent band at about 115 kDa (Fig. 2C), which had not been seen with antibodies to tags at the Nterminus or C-terminus. Thus, the 115 kDa band represents protein processed at both termini. Similar results were obtained when two HA tags were inserted at the same position (data not shown). Thus, much of the tagged FtsH1 was processed at both termini. Interestingly, for these proteins, which had an unmodified N-terminus, the NP species was much more prominent than the CP species. This is the reverse of what was seen when the N-terminus was modified by the addition of the V5 tag (compare Fig. 2B and C). Thus, the N-terminal V5 tag appeared to inhibit N-terminal processing of FtsH1 This was confirmed by immunoblot analysis of untransfected RH strain parasites using a polyvalent antiserum raised against the C-terminal unique portion of FtsH1. This antiserum revealed a strong band corresponding to the N-terminally processed form, and a faint band corresponding to the full-length protein (unpublished data). However, the antiserum detected additional non-specific bands and was of low titer, precluding further studies. 3.2. Processing at the N-terminus precedes processing at the C-terminus To assess the precursor-product relationship of the FtsH1 cleavage products, we performed pulse-chase analysis. Intracellular parasites expressing FtsH1-V5233 -HA were pulsed with 35 S methionine and cysteine for one hour and either harvested or further incubated for up to four hours in complete medium (chase). A parallel set of cells was pulse-labeled for 15 min only. Untransfected RH parasites were used as the negative control. The samples were immunoprecipitated with anti-V5 mAb, followed by SDS-PAGE and transfer to nitrocellulose. Blots were also probed with anti-V5 mAb to identify the migration of the various tagged species (Fig. 3, left). Phosphorimaging (Fig. 3) detected the full-length and the 154 kDa

Fig. 3. Pulse-chase analysis of FtsH1 processing. Fibroblast monolayers infected with T. gondii expressing FtsH1-V5233 -HA (∼108 ) were labeled with 35 Smethionine/cysteine for the indicated times in minutes (pulse). RH is the untransfected parental line. Complete medium was subsequently added for various times (chase). Samples were immunoprecipitated with anti-V5 mAb, separated by SDS-PAGE, and transferred to nitrocellulose. The left panel shows the same lane detected by Western blot using anti-V5 mAb (W) or phosphorimaging (35 S). The right panel shows phosphorimaging only; electrophoresis was extended by half an hour to better separate bands on this gel.

NP species at the shortest pulse time, whereas the 140 kDa CP band and a band at about 117 kDa were seen during a longer 1 h pulse (a faint band at this position was detected at steady state, Fig. 3 left and Fig. 2C). The 115 kDa species accumulated during the four hour chase period, but did not reach the steady state level. During the course of the chase period, the full-length species was depleted in favor of the proteolytically processed forms demonstrating a precursor-product relationship. Thus these data indicate that Nterminal cleavage occurs first and C-terminal cleavage occurs later, yielding a form of FtsH1 that is processed at both N- and C-termini. One last cleavage event yields the 115 kDa species. 3.3. Subcellular localization of processing events We previously demonstrated that C-terminal deletion mutants having only the first 354 or 694 aa of FtsH1 were arrested in the ER [19]. These mutants only had N-terminal tags, so we could not assess their N-terminal processing. To assess whether N-terminal processing occurs in the ER, we generated FtsH1(1-354)-HA, which has four HA tags at the C-terminus. Like the N-terminally tagged version of the same protein [19], this protein was retained in the ER (note the perinuclear rim marked by yellow arrows in Fig. 4A, top). As shown in Fig. 4B, this mutant was processed efficiently at the N-terminus, although a small amount of uncleaved protein remained (similar to FtsH1-V5233 -HA). Comparable results were obtained for FtsH1(1694)-HA (not shown). Thus exit from the ER is not required for the N-terminal cleavage. However, a subsequent experiment showed that entry into the secretory system is essential for the N-terminal cleavage to take place. When the predicted transmembrane domain was deleted from the FtsH1(1-354)-HA, the protein was dispersed in the cytosol (Fig. 4A, notice the lack of a perinuclear rim). This protein did not show any N-terminal processing on immunoblot analysis (Fig. 4B). If N-terminal processing occurs while FtsH1 is in the ER, where does C-terminal processing occur? The immunofluorescence studies using antibodies to the C-terminal HA tag (Fig. 2A, bottom panels) suggested that FtsH1 at the apicoplast lacked the Cterminus. We created two insertion mutants whose proteolytic processing and intracellular location supported this conjecture. These proteins had HA tags inserted either after the AAA domain at aa 325 or after the peptidase domain at aa 932. Although they were processed at the N-terminus, they did not undergo C-terminal cleavage (note lack of band, arrowhead) (Fig. 5A). IFA using antibod-

A. Karnataki et al. / Molecular & Biochemical Parasitology 166 (2009) 126–133

131

Fig. 4. N-terminal processing occurs in the ER. Stable transfectants expressing C-terminally truncated mutants FtsH1(1-354)-HA and V5-FtsH1(1-354)TMD were generated. (A) Schematics and IFAs. Anti-HA antibody staining shows localization to the ER for FtsH1(1-354)-HA (note characteristic perinuclear staining) or to the cytoplasm for V5FtsH1(1-354)TMD. Yellow arrows mark the perinuclear rim staining. Bar = 2 ␮M. (B) Immunoblot analysis of FtsH1(1-354)-HA and FtsH1(1-354)TMD-HA using anti-HA antibody on a 10% gel. F, Full-length mutant protein; NP, N-terminally processed protein. Note that the NP form is lacking for the cytosolic protein. A lower laser setting was used to detect FtsH1(1-354)-HA. Both samples were present on the same gel. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

ies to the N-terminal V5 tag showed that these proteins were not spread throughout the ER like the C-terminal deletion mutant in Fig. 4A, nor did they show the plastid localization characteristic of the wild-type protein. Instead the proteins accumulated in what appeared to be vesicles near the apicoplast (Fig. 5B and C). We conducted a quantitative analysis of the location of wild-type and the two mutant proteins to solidify this finding, examining parasites in at least 100 vacuoles for each construct (Fig. 5C). The wild-type protein V5-FtsH1-HA showed a circumplastid distribution in 50% of vacuoles, a pattern rarely seen for the 325 and 932 tag insertion proteins (<5% of vacuoles). About 40% of vacuoles with cells express-

ing the mutant proteins exhibited a vesicle-like staining that was clearly not at the apicoplast, a pattern seen much less frequently for the wild-type protein (10% of vacuoles). About 40% of the vacuoles containing cells expressing either wild-type or mutant constructs showed a localization pattern we describe here as “mixed”, with some dispersed staining and some staining around or at the plastid. Blind analysis of the “mixed” population of wild-type and of the 325 insertion mutant indicated that the mutant protein showed less staining adjacent to the apicoplast. Thus little of this mutant protein was located at the apicoplast, correlating with the lack of C-terminal processing.

Fig. 5. C-terminal processing correlates with apicoplast localization. Localization and processing of expression constructs with internal HA tags, V5-FtsH1-HA325 and V5FtsH1-HA932 . (A) Immunoblot analysis. The lysates were separated by SDS-PAGE and probed with anti-V5 mAb and anti-HA to examine C-terminal and N-terminal processing, respectively. For V5-FtsH1-HA325 , the same blot was probed with both anti-V5 mAb and rabbit anti-HA antibodies. For V5-FtsH1-HA932 , separate blots were probed with anti-V5 mAb and the high affinity rat anti-HA mAb (the two lanes are from the same gel). F, Full-length; NP, N-terminally processed; CP, C-terminally processed. (B) Schematics and IFA of expression constructs with internal HA tags, as in Fig. 2A. Bar = 2 ␮M. (C) Tabulation of staining patterns. At least one hundred vacuoles were analyzed for each clonal cell line expressing the indicated construct. The staining is categorized as plastid (circumplastid and/or colocalizing), mixed (plastid-like plus dispersed staining), dispersed (not circumplastid or colocalizing), ER-like (punctate, highly dispersed, perinuclear) and weak (little or no staining seen).

132

A. Karnataki et al. / Molecular & Biochemical Parasitology 166 (2009) 126–133

Fig. 6. Processing and localization of FtsH1 peptidase mutant. (A) Schematics and localization of proteolytically inactive mutants. IFA was performed as in Fig. 2. Bar = 2 ␮M. (B) Extracts of the indicated cell lines were separated by SDS-PAGE and probed with anti-V5 mAb or rabbit anti-HA antibodies. F, Full-length; NP, Nterminally processed; CP, C-terminally processed.

3.4. FtsH1 does not undergo self proteolysis E. coli FtsH undergoes self-cleavage, removing the last seven amino acids from its C-terminus [33]. The cleavage is in cis, since the proteolytically inactive form of FtsH was not processed to a noticeable level even when it was expressed in the wild-type background. It has been shown previously that the glutamic acid residue in the zinc binding motif (HEXXH) of the peptidase domain is essential for the FtsH proteolytic activity [34]. To determine whether T. gondii FtsH1 has the capacity to undergo self-cleavage, the glutamic acid of the zinc binding site was changed to glutamine (V5-FtsH1E732Q HA), which is expected to abolish the proteolytic activity of the protein as it does in other FtsHs [34,35]. Clonal lines expressing the mutant protein were selected. We examined the localization of the tagged mutant protein by IFA, which demonstrated that it correctly localized to the apicoplast (Fig. 6A). Since the peptidase and N-terminal region are topologically separated by the TMD, we thought it unlikely that FtsH1 cleaves its own N-terminus. Indeed, when immunoblots were probed with anti-HA antibodies, both the full-length and NP isoforms were detected, and the ratios of the two bands appeared similar in mutants and wild-type. Additionally, as revealed by anti-V5 mAb, the abundance of the CP form in the peptidase mutant was similar to wild-type (Fig. 6B). These data suggest that C-terminal cleavage is not mediated in cis, but rather by an adjacent wild-type subunit in the hexamer or by some other protease. 4. Discussion Eukaryotic FtsHs are integral membrane proteins that degrade damaged membrane proteins in chloroplasts and mitochondria, as well as fulfilling other functions. T. gondii FtsH1 is one of three FtsHs

in the parasite, none of which have been functionally analyzed. None of the mutant proteins we tested conferred an obvious cellular phenotype when expressed in the wild type background, and thus these studies do not illuminate the function of FtsH1. However, we demonstrate here that FtsH1 undergoes an elaborate set of processing events. It is tempting to speculate that these cleavages modulate the function of FtsH1. For example, they may control activation of the protein by promoting assembly of the active complex or by removing an inhibitory domain as happens with many other proteases. FtsH1 is one of three apicoplast membrane proteins to be described thus far in T. gondii, the others being the sugar phosphate translocator APT1 [18] and the Tic20 protein involved in protein import [20]. On immunoblot analysis, C-terminally tagged APT1 showed a single band of expected size, indicating that it is unlikely to undergo proteolytic processing. Tic20 undergoes Nterminal processing [20], as expected since it bears signal and transit peptides. In contrast, here we show that FtsH1 undergoes processing at both termini. The proteolytic cleavages pare much of the “unique” sequence from the parasite protein, with the final product more closely resembling the FtsHs of other organisms. Since the sites of proteolytic cleavage are topologically separated by a membrane, it is clear that at least two distinct proteases participate in processing of T. gondii FtsH1. No FtsH homologues in other organisms have been shown to undergo multiple processing events. Our data indicate that FtsH1 must enter the secretory system for processing to initialize. Unlike the N-terminal processing of transit peptides from chloroplast FtsHs and from apicoplast luminal proteins, which is catalyzed by the stromal processing peptidase in the lumen [36,37], N-terminal processing of FtsH1 occurs in the ER. Unlike the rapid cleavage of N-terminal signal sequences of proteins as they enter the ER, the N-terminal processing of FtsH1 is slow enough that uncleaved molecules are readily detected. Furthermore, N-terminal processing of FtsH1 is not required to exit the ER. Little of FtsH1 bearing an N-terminal V5 tag was N-terminally processed, yet the tagged protein escaped the ER and localized to the apicoplast. Until now, the only FtsH known to undergo C-terminal processing is that of E. coli. Akiyama et al. [33] demonstrated that E. coli FtsH has the ability to undergo self-cleavage, which is dependent on ATP hydrolysis. In contrast, the C-terminal cleavage of T. gondii FtsH1 is not autocatalytic in cis, as shown by our analysis of protein mutated at the well-characterized peptidase active site motif. The protease-inactive mutant protein was able to localize to the apicoplast, indicating peptidase activity of an individual subunit is not required for proper localization. The exact subcellular location of C-terminal processing of FtsH1 is not yet clear, although we propose that it occurs around the time of localization to the apicoplast. The pulse-chase analysis demonstrates that it occurs after N-terminal cleavage, which occurs in the ER. C-terminal processing was absent in the two internal HA taginsertion mutants (at aa 325 or 932) that accumulated close to, but not at, the apicoplast. Finally, when antibodies to the C-terminal tag were used (shown in Fig. 2A for FtsH1-V5233 -HA), little staining was observed at the plastid, although protein residing at the plastid was clearly detectable by virtue of the internal tag. Thus protein at the apicoplast lacked its C-terminus. The last detected processing event, which generates the 115 kDa band, likely occurs within the apicoplast, as this band is not observed during a one hour labeling and is much more abundant after a 4 h chase period. However, even at that time it has not reached the steady state level. Since FtsH1 resides in multiple membranes within the apicoplast, the cleavage could be associated with localization to a specific membrane. Whether the last cleavage is at the N- or C-terminus is not yet known. Future analyses of immuno-

A. Karnataki et al. / Molecular & Biochemical Parasitology 166 (2009) 126–133

precipitated FtsH1-V5233 -HA could reveal the precise cleavage sites. Recently the apicoplast peripheral membrane protein ATrx1 was described [38]. Like FtsH1, this protein lacks a canonical signal sequence and transit peptide. It undergoes several processing events at the N-terminus. As with FtsH1, some of the processed forms occur relatively early after synthesis, but others appear much later. Indeed, some of the processed forms of ATrx1 do not appear until eight hours after synthesis. Taken together, these data indicate that protein processing associated with apicoplast proteins is more complex than previously suspected. We observed an unusual localization of two of the mutant proteins bearing internal HA tags, V5-FtsH1-HA325 and V5-FtsH1HA932 . The proteins accumulated in a compartment that may be either ER exit sites or vesicles that would normally traffic between the ER and the apicoplast. Further analysis of this compartment may shed light on the mechanisms of trafficking of membrane proteins to the apicoplast. Interestingly, although both of these tag insertions are C-terminal to the TMD, the crystal structure of bacterial FtsHs [39–41] predicts they are far apart in the folded protein. Hence changes in disparate regions of the FtsH1 can affect the steady state localization of the protein after it enters the ER. Acknowledgements The authors thank Ms. Pashmi Vaney for excellent technical assistance. We also thank Malcolm Gardner for critical reading of the manuscript. This work was supported in part by NIH R01 AI50506. The content is solely the responsibility of the authors. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.molbiopara.2009.03.004. References [1] Ajioka JW, Fitzpatrick JM, Reitter CP. Toxoplasma gondii genomics: shedding light on pathogenesis and chemotherapy. Expert Rev Mol Med 2001;2001:1–19. [2] Sukthana Y. Toxoplasmosis: beyond animals to humans. Trends Parasitol 2006;22:137–42. [3] Guinovart C, Navia MM, Tanner M, et al. Malaria: burden of disease. Curr Mol Med 2006;6:137–40. [4] Soldati D. The apicoplast as a potential therapeutic target in and other apicomplexan parasites. Parasitol Today 1999;15:5–7. [5] Zhu G, Marchewka MJ, Keithly JS. Cryptosporidium parvum appears to lack a plastid genome. Microbiology 2000;146:315–21. [6] Waller RF, Keeling PJ, Donald RGK, et al. Nuclear-encoded proteins target to the plastid in Toxoplasma gondii and Plasmodium falciparum. Proc Natl Acad Sci USA 1998;95:12352–7. [7] Sato S, Clough B, Coates L, et al. Enzymes for heme biosynthesis are found in both the mitochondrion and plastid of the malaria parasite Plasmodium falciparum. Protist 2004;155:117–25. [8] Jomaa H, Wiesner J, Sanderbrand S, et al. Inhibitors of the nonmevalonate pathway of isoprenoid biosynthesis as antimalerial drugs. Science 1999;285: 1573–6. [9] Wilson RJ. Progress with parasite plastids. J Mol Biol 2002;319:257–74. [10] Fichera ME, Roos DS. A plastid organelle as a drug target in apicomplexan parasites. Nature 1997;390:407–9. [11] Seeber F, Aliverti A, Zanetti G. The plant-type ferredoxin-NADP+ reductase/ferredoxin redox system as a possible drug target against apicomplexan human parasites. Curr Pharm Des 2005;11:3159–72. [12] Kohler S, Delwiche CF, Denny PW, et al. A plastid of probable green algal origin in Apicomplexan parasites. Science 1997;275:1485–9.

133

[13] Waller RF, McFadden GI. The apicoplast: a review of the derived plastid of apicomplexan parasites. Curr Issues Mol Biol 2005;7:57–79. [14] DeRocher A, Hagen CB, Froehlich JE, et al. Analysis of targeting sequences demonstrates that trafficking to the Toxoplasma gondii plastid branches off the secretory system. J Cell Sci 2000;113:3969–77. [15] Waller RF, Reed MB, Cowman AF, et al. Protein trafficking to the plastid of Plasmodium falciparum is via the secretory pathway. EMBO J 2000;19:1794–802. [16] Yung S, Unnasch TR, Lang-Unnasch N. Analysis of apicoplast targeting and transit peptide processing in Toxoplasma gondii by deletional and insertional mutagenesis. Mol Biochem Parasitol 2001;118:11–21. [17] van Dooren GG, Su V, D’Ombrain MC, et al. Processing of an apicoplast leader sequence in Plasmodium falciparum and the identification of a putative leader cleavage enzyme. J Biol Chem 2002;277:23612–9. [18] Karnataki A, DeRocher A, Coppens I, et al. Cell cycle-regulated vesicular trafficking of Toxoplasma APT1, a protein localized to multiple apicoplast membranes. Mol Microbiol 2007;63:1653–68. [19] Karnataki A, Derocher AE, Coppens I, et al. A membrane protease is targeted to the relict plastid of Toxoplasma via an internal signal sequence. Traffic 2007;8:1543–53. [20] van Dooren GG, Tomova C, Agrawal S, et al. Toxoplasma gondii Tic20 is essential for apicoplast protein import. Proc Natl Acad Sci USA 2008;105:13574–9. [21] Krzywda S, Brzozowski AM, Verma C, et al. The crystal structure of the AAA domain of the ATP-dependent protease FtsH of Escherichia coli at 1.5 A resolution. Structure 2002;10:1073–83. [22] Niwa H, Tsuchiya D, Makyio H, et al. Hexameric ring structure of the ATPase domain of the membrane-integrated metalloprotease FtsH from Thermus thermophilus HB8. Structure 2002;10:1415–23. [23] Sakamoto W, Zaltsman A, Adam Z, et al. Coordinated regulation and complex formation of yellow variegated1 and yellow variegated2, chloroplastic FtsH metalloproteases involved in the repair cycle of photosystem II in Arabidopsis thylakoid membranes. Plant Cell 2003;15:2843–55. [24] Tauer R, Mannhaupt G, Schnall R, et al. Yta10p, a member of a novel ATPase family in yeast, is essential for mitochondrial function. FEBS Lett 1994;353:197–200. [25] Leonhard K, Stiegler A, Neupert W, et al. Chaperone-like activity of the AAA domain of the yeast Yme1 AAA protease. Nature 1999;398:348–51. [26] Arlt H, Tauer R, Feldmann H, et al. The YTA10-12 complex, an AAA protease with chaperone-like activity in the inner membrane of mitochondria. Cell 1996;85:875–85. [27] Shirai Y, Akiyama Y, Ito K. Suppression of ftsH mutant phenotypes by overproduction of molecular chaperones. J Bacteriol 1996;178:1141–5. [28] Rainey RN, Glavin JD, Chen HW, et al. A new function in translocation for the mitochondrial i-AAA protease Yme1: import of PNPase into the intermembrane space. Mol Cell Biol 2006;26:8488–97. [29] Donald RG, Roos DS. Gene knock-outs and allelic replacements in Toxoplasma gondii: HXGPRT as a selectable marker for hit-and-run mutagenesis. Mol Biochem Parasitol 1998;91:295–305. [30] Roos DS, Donald RG, Morrisette NS, et al. Molecular tools for genetic dissection of the protozoan parasite Toxoplasma gondii. Meth Cell Biol 1994;45:27–63. [31] Black M, Seeber F, Soldati D, et al. Restriction enzyme-mediated integration elevates transformation frequency and enables co-transfection of Toxoplasma gondii. Mol Biochem Parasitol 1995;74:55–63. [32] DeRocher A, Gilbert B, Feagin JE, et al. Dissection of brefeldin A-sensitive and -insensitive steps in apicoplast protein targeting. J Cell Sci 2005;118:565–74. [33] Akiyama Y. Self-processing of FtsH and its implication for the cleavage specificity of this protease. Biochem 1999;38:11693–9. [34] Leonhard K, Herrmann JM, Stuart RA, et al. AAA proteases with catalytic sites on opposite membrane surfaces comprise a proteolytic system for the ATPdependent degradation of inner membrane proteins in mitochondria. EMBO J 1996;15:4218–29. [35] Kotschwar M, Harfst E, Ohanjan T, et al. Construction and analyses of mutant ftsH alleles of Bacillus subtilis involving the ATPase- and Zn-binding domains. Curr Microbiol 2004;49:180–5. [36] Richter S, Lamppa GK. A chloroplast processing enzyme functions as the general stromal processing peptidase. Proc Natl Acad Sci USA 1998;95:7463–8. [37] van Dooren GG, Waller RF, McFadden GI, et al. Traffic jams: protein transport in Plasmodium falciparum. Parasitol Today 2000;16:421–7. [38] Derocher AE, Coppens I, Karnataki A, et al. A thioredoxin family protein of the apicoplast periphery identifies abundant candidate transport vesicles in Toxoplasma gondii. Eukaryot Cell 2008;7:1518–29. [39] Suno R, Niwa H, Tsuchiya D, et al. Structure of the whole cytosolic region of ATP-dependent protease FtsH. Mol Cell 2006;22:575–85. [40] Bieniossek C, Schalch T, Bumann M, et al. The molecular architecture of the metalloprotease FtsH. Proc Natl Acad Sci USA 2006. [41] Kim SH, Kang GB, Song HE, et al. Structural studies on Helicobacter pylori ATPdependent protease, FtsH. J Synchrotron Radiat 2008;15:208–10.