Planta DOI 10.1007/s00425-008-0697-1 O R I G I N A L A R T I CL E Processing and traYcking of a single isoform of the aspartic proteinase cardosin A on the vacuolar pathway Patrícia Duarte · José Pissarra · Ian Moore Received: 13 November 2007 / Accepted: 22 January 2008  Springer-Verlag 2008 Abstract Cardosin A is the major vacuolar aspartic proteinase (APs) (E.C.3.4.23) in pistils of Cynara cardunculus L. (cardoon). Plant APs carry a unique domain, the plant-speciWc-insert (PSI), and a pro-segment which are separated from the catalytic domains during maturation but the sequence and location of processing steps for cardosins have not been established. Here transient expression in tobacco and inducible expression in Arabidopsis indicate that processing of cardosin A is conserved in heterologous species. Pulse chase analysis in tobacco protoplasts indicated that cleavage at the carboxy-terminus of the PSI could generate a short-lived 50 kDa intermediate which was converted to a more stable 35 kDa intermediate by removal of the PSI. Processing intermediates detected immunologically in tobacco leaves and Arabidopsis seedlings conWrmed that cleavage at the amino-terminus of the PSI either preceded or followed quickly after cleavage at its carboxy-terminus. Thus removal of PSI preceded the loss of the prosegment in contrast to the well-characterised Electronic supplementary material The online version of this article (doi:10.1007/s00425-008-0697-1) contains supplementary material, which is available to authorized users. P. Duarte · J. Pissarra IBMC - Instituto de Biologia Molecular e Celular, Universidade do Porto, Rua do Campo Alegre, 823, 4150-180 Porto, Portugal P. Duarte (&) · I. Moore Department of Plant Sciences, University of Oxford, South Parks Rd, Oxford OX1 3RB, UK e-mail: pduarte@ibmc.up.pt J. Pissarra Department of Botany, Faculty of Sciences, University of Porto, Rua do Campo Alegre, 1191, 4150-181 Porto, Portugal barley AP, phytepsin. PreprocardosinA acquired a complex glycan and its processing was inhibited by brefeldin A and dominant-inhibitory AtSAR1 or AtRAB-D2a mutants indicating that it was transported via the Golgi and that processing followed ER export. The 35 kDa intermediate was present in the cell wall and protoplast culture medium as well as the vacuole but the 31 kDa mature subunit, lacking the amino-terminal prosegment, was detected only in the vacuole. Thus maturation appears to occur only after sorting from the trans-Golgi to the vacuole. Processing or transport of cardosin A was apparently slower in tobacco protoplasts than in whole cells. Keywords Arabidopsis · Dexamethasone · Nicotiana · Phytepsin · pOp6 LhGR · Rab1 Abbreviations AP Aspartic proteinase ER Endoplasmic reticulum PSI Plant speciWc insert SAPLIP Saposin-like protein Introduction Aspartic proteinases (APs) (E.C.3.4.23) constitute one of the four main classes of endoproteases, the other three being serine, cysteine and metallo proteinases. APs are widely present in nature, having been found in retroviruses, bacteria, fungi, and plants as well as animals (Dunn 2002). In spite of their widespread distribution, aspartic proteinases share common features, such as inhibition by pepstatin A, an acidic pH optimum, and preferential cleavage of peptide bonds between residues with hydrophobic side chains. These common features reXect the high degree of similarity 123 Planta between AP amino acid sequences. The tertiary structure of these enzymes is usually bilobal, with two domains separated by a deep and large cleft where the active site is located. Despite the overall similarity, substantial diVerences exist in catalytic properties, cellular localisation and, consequently, in the biological functions of this family of enzymes (Koelsch et al. 1994). Eukaryotic aspartic proteinases are synthesised as zymogens which undergo proteolytic processing to the mature and active form, whereas retroviral APs are synthesised as part of a structural poly-protein (Koelsch et al. 1994). AP precursors of eukaryotic origin are commonly made up of a short signal peptide, which directs the protein to the endoplasmic reticulum (ER), a pro-segment (or activation segment) and a subsequent polypeptide that will compose the mature protein (Khan and James 1998). In the plant kingdom, aspartic proteinases have been puriWed and characterised from numerous species of gymnosperms, monocotyledons and dicotyledons (Mutlu and Gal 1999). In the Arabidopsis genome, 50 genes coding for APs of the pepsin type were identiWed, and further assigned to three diVerent categories: typical, nucellin-like, and atypical proteinases (Faro and Gal 2005). Most of the available knowledge on plant APs relates to those in the typical category and has been obtained mostly from the study of phytepsin from Hordeum vulgare (barley) seeds and cardosins from the Xowers of Cynara cardunculus (cardoon) (Simões and Faro 2004). Plant APs have been primarily isolated from seeds but were also identiWed in other organs such as leaves, roots and Xowers. Subcellularly, plant APs are mainly localised either in the vacuole/protein body or the extracellular space (Mutlu and Gal 1999). Their biological functions, however, are not as well assigned or characterized as those of mammalian, microbial or viral APs (Dunn 2002). Nevertheless, plant APs have been implicated in protein processing and/or degradation in diVerent plant organs, as well as in plant senescence, stress responses, programmed cell death and reproduction (Simões and Faro 2004). Cardosins are APs originally isolated from Cynara cardunculus L. (cardoon) pistils. Cardosins are expressed at unusually high levels in cardoon Xoral organs where Cardosins A and B account for the majority of the total soluble protein content (Ramalho-Santos et al. 1997). This is unusual since most plant APs described to date were mainly isolated from seeds or leaves, where they exist at much lower concentrations (Mutlu and Gal 1999). The abundance of cardosins in cardoon Xowers and their milk-clotting activity gives the proteins an economic signiWcance in Portugal where the Xowers are traditionally used in cheese manufacture (Ramalho-Santos et al. 1996; Veríssimo et al. 1996). Among plant vacuolar APs, barley phytepsin, is the best characterised (Glathe et al. 1998; Kervinen et al. 1999; Tormakangas et al. 2001). In keeping with other eukaryotic 123 APs, phytepsin is synthesised as a precursor which needs to lose its signal peptide, pro-region and plant speciWc segment, to become mature (Glathe et al. 1998). Indeed, phytepsin has been used as the model for maturation of plant AP enzymes and for identiWcation of vacuolar sorting determinants and routes (Glathe et al. 1998; Kervinen et al. 1999; Tormakangas et al. 2001; Pimpl et al. 2003). The APs in plants are distinguished from those in microbes by the presence of an additional segment of about 100 amino acids, the plant speciWc insert (PSI). The PSI sequence shows no homology with mammalian or microbial AP domains, but is highly similar to that of saposins and saposin-like proteins (SAPLIPs) which are lysosomal sphingolipid-activator proteins (Simões and Faro 2004). The maturation process of APs involves the removal of the PSI and the Pro region to generate the mature molecules (Mutlu and Gal 1999). Depending on the plant AP, this PSI is either partially or completely removed during maturation. For example, while mature phytepsin retains an N-terminal portion of the PSI, cardosin A maturation involves the complete removal of both PSI and the Pro segment (Glathe et al. 1998; Ramalho-Santos et al. 1998). The biological function(s) of saposins have not yet been completely established, however, it was already demonstrated that the correct sorting of cathepsin D, a mammalian AP, to the lysosome, requires the association of this protein with a saposin, namely saposin C (Zhu and Corner 1994; Simões and Faro 2004). Additionally, the PSI of phytepsin was shown to be essential for the correct vacuolar targeting of this enzyme, since deletion of this region causes phytepsin’s secretion and also alters the way in which this AP exits the ER (Tormakangas et al. 2001). Although it is already documented that cardosin A is extensively processed after synthesis, so far, very little is known about the maturation process or the intracellular location of processing events. Ramalho-Santos et al. (1998), using antibodies raised against diVerent regions of the protein, proposed a sequence of processing events in maturation of cardosin A based on the analysis of two time points in cardoon Xoral development. This sequence of events diVers from that of phytepsin (Glathe et al. 1998) but has not been conWrmed by direct methods. Furthermore, nothing is known about the subcellular localisation of the diVerent processing events to which cardosin A is subjected. The processed, mature form of cardosin A was, however, shown to be located in the vacuoles (RamalhoSantos et al. 1997). It is known that protein sorting to plant vacuoles is dependent on a considerable variety of protein motifs and this sorting can involve either traYc from the ER through the Golgi apparatus or direct ER-to-vacuole transport since there is evidence for the existence of a Golgi-independent route to storage vacuoles directly from the endoplasmic Planta reticulum (Levanony et al. 1992; Hara-Nishimura et al. 1998; Bassham and Raikhel 2000; Chrispeels and Herman 2000; Hadlington and Denecke 2000). The sorting pathway followed by cardosin A from the ER towards the vacuole is still unclear. Transmission electron microscopy (TEM) observations of developing cardoon pistils revealed swollen cisternae of RER with similar appearance to cardosin A-containing vacuoles (Duarte et al. 2006). This observation led to the hypothesis that cardosin A could reach the vacuole through a Golgi independent route, although the presence of modiWed N-glycans indicates involvement of the Golgi complex (Costa et al. 1997). The work presented here further characterises the sorting and processing of cardosin A using pulse-chase approaches involving several expression systems in whole tissues or protoplasts of heterologous species where other endogenous cardosins do not interfere with the analysis. Materials and methods Cardosin A cloning into the binary expression vectors Standard methods were employed for molecular cloning and bacterial culture (Sambrook et al. 1989). The original cardosin A cDNA clone (Faro et al. 1999) was modiWed by PCR with Platinum Pfx DNA Polymerase (Invitrogen), the forward primer CdAFwd1: 5⬘AAAACTCGAGCCACCAT GGGCACCTCAATCAAAGCAAACG3⬘ and the reverse primer 5⬘CGGGTTGTATCTTAGATCGG3⬘. This introduced an optimal eukaryote translation initiation site (underlined) (Kozak 1984), two new restriction sites (Xho I and Nco I), and a mutated nucleotide (bold) to remove a Kpn I restriction site. The generated fragments were digested with Ava I and Xho I, and the resulting 200-base pair (bp) fragment was inserted into the original construct pCR2.1-CdA at Xho I and Ava I sites and conWrmed by sequencing. To express cardosin A in tobacco leaves using Agrobacteriummediated transformation, the modiWed cardosin A cDNA was digested with BamH I and Xho I, and further cloned into the binary expression vector pVKH18-EN6 (Batoko et al. 2000), between BamH I and Sal I sites. The modiWed cardosin A cDNA was also digested with Ecl 136II and Xho I and cloned between Sma I and Xho I restriction sites of the binary expression vector pH-TOP for dexamethasoneinducible expression (Craft et al. 2005). Agrobacterium-mediated transient expression in Nicotiana tabacum Agrobacterium-mediated transient expression of cardosin A in Nicotiana tabacum L. cv. Petit Havana SR1 (provided by Dr. P. Czernilofsky, MPI Cologne, Germany) was carried out as described in Batoko et al. (2000) with the following modiWcation: instead of YEB-medium, LB medium supplemented with 25
g/mL of kanamycin and 0.4 mM of isopropyl-
-D-thiogalactopyranoside (IPTG) was used at all times. For experiments requiring co-infection of more than one construct, bacterial strains containing the constructs were mixed before performing inWltration, with the titre of each construct adjusted to the required Wnal OD600. Dexamethasone treatments and GUS staining of the Arabidopsis pOp/LhGR stable transformants Arabidopsis thaliana L. ecotype Columbia (provided by Dr. C. Dehio, MPI Cologne, Germany) plants of the 4c-S5 and 4c-S7 CaMV35S::LhGR activator lines were transformed and analysed according to Craft et al. (2005). Dexamethasone (Sigma-Aldrich) was dissolved at 100 mM in dimethylsulfoxide (DMSO) and kept at ¡20°C. Typically, 20
M dexamethasone was used for induction. All experiments were performed with T2 Arabidopsis seedlings, induced either in liquid or solid media, as described elsewhere (Craft et al. 2005). Putative transformants were tested for GUS expression through histochemical GUS assays as described in JeVerson (1987), except that repeated vacuum inWltrations were performed (up to 5£) which resulted in more staining of internal mature leaf tissues. For GUS staining purposes, dexamethasone induction was carried out for 2 days. For immunodetection of cardosin A, induction periods varied from 3 to 7 days. Protein extraction, endoglycosidase treatment, and protein gel blot analysis Protein extracts for endoglycosidase H (endoH) digestion were prepared with 50–300 mg fresh weight of tobacco infected leaf samples or Arabidopsis seedlings frozen in liquid N2 according to the method described in Batoko et al. (2000). EndoH digestions were carried out as described in Batoko et al. (2000). For SDS-PAGE we used 12% acrylamide gels in the Mini-Protean 3 apparatus (Bio-Rad), and electrophoresis was performed according to the Laemmli method (Laemmli 1970). After electrophoresis, proteins were transferred into polyvinylidene diXuoride (PVDF) membranes for 1 h 30 min in transfer buVer, according to the method described by Burnette (1981). The membrane was blocked for 1 h at room temperature (RT) in TrisbuVered saline (TBS) containing 0.1% (v/v) Tween-20 (Sigma, Poole, UK) (TBS-T) with 5% (w/v) skimmed milk powder, 1% (w/v) BSA (Fraction V; Roche Biochemicals) and 0.1% (v/v) goat serum (Sigma). Rabbit anti-sera raised against recombinant cardosin A (Faro et al. 1999), calreticulin (gift from J. Denecke) and -TIP (gift from 123 Planta A. SchaeVner) were used at 1:600, 1:10,000 and 1:200 dilutions respectively, in blocking solution, to probe the membranes for 1 h at RT or over-night (O/N) at 4°C. Alkaline-phosphatase conjugated secondary antibody (Sigma) was used at 1:5,000 dilution in TBS-T for 30 min at RT. Activity was revealed using the Sigma Fast 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium (BCIP/NBT) chromogenic reagent, according to manufacturer’s instructions (Sigma). Protoplast isolations Isolation of tobacco mesophyll protoplasts was performed using the method described by Denecke and Vitale (1995), with some modiWcations. Tobacco leaves were inWltrated with A. tumefaciens containing the pVKHEN6-cardosin A construct, and 24 h after inWltration, leaf digestion was started. Leaves were treated for 12–15 h, at RT and in the dark, with a mixture of 0.4% (w/v) cellulase (Onozuka R10) and 0.2% (w/v) macerozyme (Onozuka R10) dissolved in TEX medium, supplemented with carbenicillin (200
g/mL) and timentin (2
g/mL). The living protoplast fraction (Xoating) was obtained after 2 centrifugations (100g 10 min at RT) in TEX medium. The isolated fraction was then transferred into a 15 mL falcon tube containing two volumes of mannitol/W5 [0.4 M D(¡)-mannitol, 1 mM D(+)-glucose, 30.8 mM NaCl, 25 mM CaCl2, 1 mM KCl, 0.3 mM Mes (BDH Chemicals); pH 5.6–5.8] and gently mixed. Protoplasts were pelleted at 100g for 5 min at RT and the supernatant was carefully removed. TEX medium was added up to 4–5 mL, and the protoplasts were gently resuspended in this volume. A measure of 10
L of the cell suspension were taken for cell counting with a haemocytometer under a light microscope. The protoplasts resuspended in TEX medium were allowed to Xoat freely, and the medium underneath the cell layer was removed. The protoplast fraction was then resuspended in the appropriate TEX volume in order to achieve the required cell concentration for the subsequent experiments. Arabidopsis protoplasts were isolated from 2-week-old seedlings grown on inductive medium as described in Zheng et al. (1997) with some modiWcations. Total seedlings were digested with 1% cellulase (Onozuka R10) and 0.25% of macerozyme (Onozuka R10) in MM buVer (0.4 M mannitol and 20 mM Mes; pH 5.8) for 3 h in the dark. The protoplast suspension was Wltered through a 100
m nylon mesh and collected in 15 mL falcon tubes. Protoplasts were pelleted at 65g for 5 min at 20°C, resuspended in 5 mL of MM buVer and centrifuged as before. The puriWed protoplast pellet was then resuspended in the appropriate MM volume in order to achieve the required cell concentration for the subsequent experiments. 123 Metabolic labelling and pulse-chase experiments Metabolic labelling of tobacco protoplasts was performed as described by Denecke and Vitale (1995), with some modiWcations. Protoplasts were labelled for 3 h with Trans35S-Label No-Thaw Metabolic Labelling Reagent (ICN Biochemicals). For the labelling of 300
L of cell suspension (ca. 1.2 £ 106 protoplasts), 20
L of the labelling mixture (37,000 Bq/
L) diluted in 280
L of TEX medium were used. The labelling reaction was performed in microfuge tubes kept vertically, without agitation, at RT, and in the dark. After the labelling step, protoplasts were allowed to Xoat freely, and the medium below the cell layer was collected and replaced with approximately the same volume of TEX chase medium supplemented with unlabelled 10 mM methionine and 5 mM cysteine (Sigma). For BFA treatments the chase medium was supplemented with the BFA solution (10
g/mL) (Sigma). At the appropriate chase time points, protoplasts were resuspended and 100
L of cell suspension taken. The protoplasts present in the suspension fraction were allowed to Xoat freely, and the medium underneath the protoplast layer was removed. A measure of 1 mL of 250 mM NaCl was added before pelleting by centrifugation at 1,466g for 2 min at 4°C. The resulting sediment was frozen in liquid N2 and stored at ¡80°C until homogenization. Tobacco protoplast pellets were resuspended and lysed in four volumes of freshly prepared protoplast homogenization buVer [200 mM Tris–HCl pH 8.0, 300 mM NaCl, 1% (v/v) Triton X-100, 1 mM EDTA] supplemented with the appropriate volume of protease inhibitor cocktail (Sigma), and submitted to two freeze/thaw steps. The tubes were centrifuged at 15,493 g for 5 min at 4°C, and the colleted supernatant was frozen in liquid N2 and stored at ¡80°C until immunoprecipitation. Immunoprecipitation The protocol followed for immunoprecipitation was described by Denecke and Vitale (1995). The primary antibody was a rabbit antiserum raised against recombinant cardosin A (Faro et al. 1999) at a 1:200 dilution, for 2 h on ice. SDS-PAGE (see above) was followed by Wxation with a mixture of isopropanol:H2O:acetic acid (25:65:1, by vol.) for 30 min, with agitation and in the dark. This solution was discarded and the gel was agitated in the dark for 30 min with Amplify™ Fluorographic Reagent (Amersham Biosciences), rinsed with deionised water for 10 min with agitation, dried on 3MM paper, and exposed to HyperWlm™ MP (Amersham Biosciences) in a Kodak BioMax TranScreen LE (Amersham Biosciences). Planta Vacuole isolation and medium sampling Tobacco vacuole isolation and quantiWcation of vacuole recovery using -mannosidase activity assays were performed according to the methods described by Hadlington and Denecke (2001). Before loading the gel, cellular and vacuolar protein extracts were equalized for -mannosidase activity. Arabidopsis vacuole isolation was performed as described in Di Sansebastiano et al. (1998). To assay for cardosin A secretion, the TEX medium in which the tobacco protoplasts were incubated was collected from the region between the Xoating protoplasts and the lysed protoplasts at the bottom of the tube and subjected to trichloroacetic acid (TCA) precipitation of proteins. Results Cardosin A is stably accumulated in tobacco leaves To follow the processing of a single cardosin isoform, a cDNA encoding cardosin A was expressed in two diVerent heterologous systems. In the Wrst, cardosin A was expressed in tobacco leaf cells from an enhanced CaMV 35S promoter using Agrobacterium-mediated transient expression (Batoko et al. 2000). This results in a rapid and synchronous increase in gene expression that commences approximately 30 h after inWltration of leaf tissue with the bacteria (Zheng et al. 2005). In an alternative approach, the pOp/LhGR dexamethasone-inducible promoter system (Craft et al. 2005) was used to express cardosin A in Arabidopsis seedlings. Protein extracts made from inWltrated tobacco leaves between 48 and 106 h post-inWltration (hpi) were analysed by immunoblot using an antiserum raised against recombinant procardosin A. This antiserum detects epitopes in the mature 31 kDa subunit (Fig. 1a) (Faro et al. 1999). As shown in Fig. 1b, at 48 hpi the antibody detected a predominant 35 kDa form but a 66 kDa form was also detected. The 66 kDa species disappeared over the subsequent 12 h, followed by the 35 kDa form, while the 31 kDa mature form accumulated and appeared to be stable for at least 106 hpi. These observations suggest that the 31 kDa subunit is generated by an initial cleavage at the N-terminus of the PSI followed by removal of the amino-terminal pro segment. Cleavage at the N-terminus of the PSI must either precede or follow closely upon cleavage at the C-terminus as we rarely saw the 50 kDa intermediate predicted by initial cleavage of procardosin A at the C-terminus of the PSI. To investigate cardosin A processing in Arabidopsis, we introduced the cardosin A cDNA under control of a dexamethasone inducible promoter in plasmid pH-TOP (Fig. 1c) (Craft et al. 2005). T2 seedlings were initially screened for dexamethasone-inducible expression of a GUS reporter construct borne on the same T-DNA (Fig. 1d). Selected lines were then grown for 7 days in the presence or absence of dexamethasone and were analysed for cardosin A expression by immunoblotting (Fig. 1e). In two GUSpositive lines, the 35 and 31 kDa forms of cardosin A were detectable in seedlings grown on dexamethasone but were absent from non-induced controls and from seedlings of a GUS negative line. Thus, cardosin A can be processed to the mature 31 kDa form in Arabidopsis seedlings as well as in tobacco leaves. It was noticeable that after 48 h of incubation in dexamethasone, the precursor and intermediate forms of cardosin A were already detectable, and that the longer the induction period, the stronger was the signal for cardosin A presence (Fig. 1f). In contrast to the tobacco transient expression system, in Arabidopsis the mature form of cardosin A did not predominate over the intermediate form even after 7 days of induction (Fig. 3b, bottom panel). This may reXect either more rapid processing of the intermediate form to the mature form in tobacco or the transient nature of the tobacco expression system. Cardosin A expressed in both heterologous systems is correctly sorted to the vacuole To test whether cardosin A is ultimately targeted to the vacuole in these heterologous expression systems as it is in cardoon, protoplasts and vacuoles were prepared from transfected tobacco leaves and Arabidopsis seedlings, and were analysed by immunoblotting. In preparations from tobacco leaves, the activity of the vacuolar protein -mannosidase was used to normalise loading for the quantity of vacuolar protein (Fig. 2a). The vacuolar preparations were apparently free from contaminating endomembranes as they lacked calreticulin, an abundant component of the endoplasmic reticulum (Fig. 2b). The 31 kDa mature fragment of cardosin A was equally abundant in the protoplasts and vacuoles indicating that the majority of the mature protein present intracellularly is indeed localised in the vacuole. The 35 kDa intermediate form was also detected in the vacuole suggesting that cleavage occurs in this location to produce the 31 kDa form that ultimately accumulates. Consistent with this, the ratio of the 35kDa to 31 kDa form is slightly higher in the protoplast extract than in the vacuoles. Interestingly, the ratio is signiWcantly higher still in the whole leaf samples. As discussed later, this suggests that some of the 35 kDa form may exist in the cell wall of these samples. As shown in Fig. 2c, in Arabidopsis seedlings, cardosin A is also transported to the vacuole where the 35 and 31 kDa forms were both detected. Protoplast and vacuolar preparations were normalised by probing for -TIP (Fig. 2d). Thus, conversion to the 31 kDa form appears to 123 Planta C a 2kDa 4 kDa Pre Pro 31 kDa DTG C CC 15 kDa amino-terminal region LB b CC DSG 15 kDa carboxy-terminal region PSI nos RB EN6 Hygr CardosinA p(A) 48hpi 54hpi 60hpi 72hpi 84hpi 96hpi 106hpi Uninf. 66kDa RuBisCO 35kDa 31kDa pOp6 Promoter c LB RB (6xOp lac) nos Hygr Cardosin A GUS p(A) p(A) -Dex d A6 A8 A4 A6 A8 +Dex A4 - Dex A6 e A4 A8 A4 A4 + Dex A6 A6 A8 35kDa 31kDa f 35kDa 31kDa 123 72 h 48 h Dex - + - 96 h + - + Planta 䉳 Fig. 1 Expression of cardosin A in Nicotiana tabacum and in Arabidopsis thaliana. a Schematic representation of preprocardosin A and of the processing events endured by the protein during maturation, according to observations obtained while transiently expressing cardosin A in tobacco leaves. The rabbit antiserum used in this work is represented in purple and detects epitopes in the mature 31 kDa chain. Pre signal peptide, Pro pro-segment, PSI plant speciWc insert. Glycans are represented by (). The six conserved cysteine residues present in the PSI region of plant APs and in saposins are represented by (C). The arrows mark preprocardosin A’s cleavage sites. b Schematic representation of the construct used for cardosin A expression in Nicotiana tabacum leaf epidermis via Agrobacterium and immunodetection of cardosin A in tobacco samples from 48 h post-inWltration (hpi) onwards. 48 hpi, the immature (66 kDa), intermediate (35 kDa) and mature (31 kDa) forms of cardosin A could be observed (arrows). An uninfected leaf was used as negative control (Uninf.). Cardosin A expressed in tobacco leaves is stably accumulated up to 5 days after inWltration. The band marked with an arrowhead is present in all samples and corresponds to the large subunit of Rubisco (55 kDa), that cross occur in the vacuole in Arabidopsis as well as in tobacco cells. The relative abundance of the 31 and 35 kDa forms was similar in vacuole and protoplast fractions, consistent with the view that each species is resident in the vacuole. The higher relative abundance of the 35 kDa form in Arabidopsis vacuoles compared to tobacco leaf vacuoles may reXect slower processing in at least some Arabidopsis vacuoles or the sustained delivery of new 35 kDa precursor molecules to the vacuole together with turn over of the 31 kDa form in the induced Arabidopsis seedlings. Cardosin A N-glycans are complex and acquire partial Endoglycosidase H (EndoH) resistance during transport Cardosin A has two potential N-linked glycosylation sites (at Asn 70 and Asn 363, Fig. 1a), and in cardoon extracts both of these are occupied by oligosaccharides that are predominantly resistant to removal by endoglycosidase H indicative of Golgi processing (Costa et al. 1997). To determine whether cardosin A travels through the Golgi when expressed in tobacco cells, transfected tissue was analysed for endoH resistance at various times after inoculation (Fig. 3a). At 48 hpi the 66 kDa procardosin A form was visible and was converted by endoglycosidase H treatment to two species of approximately 64 and 62 kDa that are consistent with removal of one or two glycans, respectively (Fig. 3a, 48 hpi). Thus, one of the two glycans was endoglycosidase H resistant on approximately 50% of cardosin A molecules indicating that at least 50% of molecules had passed through the Golgi and that procardosin A can reach the Golgi prior to the Wrst cleavage. At 48 hpi, the single glycan on the 35 kDa intermediate was also at least 50% endoglycosidase H resistant. The endoglycosidase H resistant proportion of the 66 and 35 kDa forms was similar at this time point so it is possible that the glycan on the 15 kDa form remained largely unmodiWed in this system. reacts with the anti-recombinant cardosin A antiserum. The unspeciWc labelling of Rubisco occurred in all Western blots and will be marked in all Wgures. c Schematic representation of the construct used for the generation of the pop/LhGR Arabidopsis stable transformants with dexamethasone (Dex) inducible cardosin A expression. d GUS staining of T2 seedlings showing dexamethasone-induced GUS activity. The GUS-positive lines (A4 and A6) were tested for cardosin A expression upon addition of dexamethasone (e). Line A8 was negative for GUS staining and was used as a negative control in further experiments. e Immunodetection of cardosin A in pOp/LhGR Arabidopsis seedlings treated with dexamethasone. Samples were collected after 7 days of induction. The intermediate (35 kDa) and mature (31 kDa) forms of cardosin A were detected in the two GUS-positive lines (A4 and A6) but were absent from the non-induced controls and from seedlings of the GUS-negative line A8. f After ca. 48 h of Dex induction, cardosin A reaches levels detectable by western blotting. Immunodetection of cardosin A in Arabidopsis seedlings becomes more evident with increased induction periods At later time-points, the proportion of resistant molecules increased and was approximately the same for the 35 and 31 kDa forms (Fig. 3a, 84 and 106 hpi). Similar observations were made with Arabidopsis seedlings following induction of cardosin A expression with dexamethasone (Fig. 3b). Three days after induction, cardosin A was found principally in the 35 kDa form, most of which was resistant to endoglycosidase H activity (Fig. 3b, d induction). After 5–7 days of induction, when signiWcantly more cardosin A had accumulated and the 31 kDa form, most but not all of each form was resistant to endoglycosidase H (Figs. 3b, 5d, 7d induction). Inhibition of ER-Golgi traYc leads to an accumulation of procardosin A To test the conclusion that cardosin A is transported to the vacuole via the Golgi, we asked whether its transport was dependent on two GTPases that are required for normal ER-to-Golgi transport in tobacco epidermal cells (Andreeva et al. 2000; Batoko et al. 2000; Saint-Jore et al. 2002). Dominant inhibitory forms of NtSAR1 and AtRAB-D2a (ARA5, AtRab1b) inhibit the function of the respective GTPases and inhibit ER to Golgi traYc of secreted, Golgilocalised, and vacuolar markers (Andreeva et al. 2000; Batoko et al. 2000; Zheng et al. 2005). It is hypothesised that NtSAR1 is required for COPII vesicle formation at the ER and AtRAB-D2a for accurate vesicle targeting to the Golgi. Cardosin A was co-expressed with dominant inhibitory AtRAB-D2a[NI] and NtSar1[HL] mutants. AtRAB-D2a[NI] clearly slowed the processing of the 66 kDa form of procardosin A which was still present at 72 hpi although almost absent by 48 hpi in controls (Fig. 4a). The appearance of the 31 kDa mature form was also delayed though interestingly the accumulation of the 35 kDa form was 123 Planta a Leaf Unif Leaf CdA Ppts 䉳 Fig. 2 Intracellular location of cardosin A in Nicotiana tabacum and Vac * 66kDa 35kDa 35kDa b Leaf Cardosin A - Ppts + - Vac - + + * 35kDa 31kDa Calreticulin c Ppts Dex Vac - + - + * 66kDa 31kDa Ppts d + Vac - + - a-TIP barely aVected. Similar results were obtained with a dominant inhibitory mutant of a second member of the plant Rab-D group (Duarte et al. unpublished observations). The dominant inhibitory NtSar1[HL] mutant also delayed the processing of procardosin A up to 72 hpi (Fig. 4b) though its eVect was less pronounced than that of AtRAB-D2a[NI]. These results show that normal processing of procardosin A is dependent on RAB-D2 and Sar1 activity, consistent with transport to the Golgi in COPII-coated vesicles. It is also consistent with the view that the Wrst processing event occurs after export from the ER. As discussed later, the failure of either mutant to completely inhibit processing of pro- 123 cardosin A may simply reXect the probability of coexpression in the transfected cell population. Brefeldin A inhibits the processing of cardosin A in tobacco protoplasts 35kDa Dex in Arabidopsis thaliana. a Cardosin A expressed in tobacco leaves is correctly sorted to the vacuole. Samples were collected 7 days after inWltration and an uninfected leaf (Leaf Uninf) was used as negative control. The intermediate (35 kDa) and mature (31 kDa) forms of cardosin A were observable in leaf (Leaf CdA), protoplast (Ppts) and vacuolar (Vac) extracts. The asterisk labels the BSA (67 kDa) present in the vacuole buVer which cross-reacts with the anti-recombinant cardosin A antiserum. This was previously tested running a lane with vacuole buVer only (results not shown). The unspeciWc labelling of BSA occurred in all lanes with vacuolar extracts and will be marked when applicable, although it will not be referred again throughout this paper. b Samples from tobacco leaves transfected with cardosinA (+) or from mock transfections (¡) were collected 7 days after inWltration. The intermediate (35 kDa) and mature (31 kDa) forms of cardosin A were observable in leaf, protoplast (Ppts) and vacuolar (Vac) extracts. The bottom panel shows an immunoblot of the ER marker calreticulin, with approximately 60 kDa (arrow), which demonstrates the absence of non-speciWc endomembranes in the vacuolar prep. c Immunodetection of cardosin A in protoplasts (Ppts) and in vacuoles (Vac) isolated from Arabidopsis pOp/LhGR seedlings showed that most of the intracellular protein is located in the vacuolar compartment. Samples were collected after 7 days of induction and the intermediate (35 kDa) and mature (31 kDa) forms of cardosin A (arrows) were speciWcally detected in the protoplasts and vacuoles isolated from dexamethasone treated seedlings. d Immunodetection of the vacuolar marker -TIP ( isoform of the tonoplast intrinsic protein) with approximately 29 kDa (arrow) was performed for all Arabidopsis extracts To further test the involvement of the Golgi in procardosin A transport and processing we investigated the eVect of the fungal metabolite brefeldin A (BFA). In tobacco epidermal and cultured cells, this drug causes the Golgi to fuse with the ER inhibiting further anterograde transport of secreted and vacuolar markers (Matsuoka et al. 1995; Boevink et al. 1998; Saint-Jore et al. 2002). Since BFA can be applied homogeneously to populations of protoplasts, the eVect of this inhibitor on processing of cardosin A was followed by pulse-chase analysis of protoplasts from inWltrated leaves (Fig. 5). As shown in Fig. 5 at the end of the 3 h labelling period, corresponding to 54 hpi, almost all cardosin A was in the 66 kDa precursor form. In control protoplasts, processing to the 35 kDa form was detected within 3 h of chase and was almost complete by 18 h. In contrast, in samples treated with brefeldin A there was no detectable conversion of the 66 kDa cardosin A precursor to the 35 kDa intermediate during the 18 h chase (Fig. 5). During the chase period, we also detected the accumulation of a 50 kDa band that was much reduced from the BFA treated lanes. This species was not detected in immunoprecipitation from control transfections (Suppl. Fig. 1) so it appears to be a processing intermediate of cardosin A Planta a Nicotiana tabacum Arabidopsis thaliana b 48hpi EndoH - 3d induction + - 66kDa 35kDa 33kDa + - 72hpi + - secGFP + + 35kDa 31kDa A6 +Dex - 60hpi b 5d induction EndoH 35kDa 33kDa 84hpi + 60hpi + 66kDa 31kDa - 48hpi - AtRab-D2a[NI] 62-64kDa 31kDa EndoH a A6 +Dex + NtSAR1[HL] - Sar1 wt Sar1 HL 72hpi - Uninf. Sar1 wt Sar1 HL 66kDa 35kDa 35kDa 33kDa 31kDa 33kDa 31kDa 29kDa 29kDa 35kDa 31kDa A6 +Dex 106hpi EndoH 35kDa 31kDa - 7d induction + EndoH 33kDa 35kDa 29kDa 31kDa - + 33kDa 29kDa Fig. 3 Cardosin A expressed in Nicotiana tabacum and in Arabidopsis thaliana exhibited partial EndoH resistance. a Extracts of tobacco leaf tissues expressing cardosin A were incubated either with (+) or without (¡) endoH and were analysed by SDS-PAGE and immunoblotting with the anti-recombinant cardosin A antiserum. Samples were obtained 48, 84 and 106 h post-inWltration. The extracts tested for endoH resistance are the same as in Fig. 1b. 48 hpi the precursor and intermediate forms of cardosin A (arrows; 66 and 35 kDa) appear as partially resistant to endoH digestion. At later stages the intermediate and mature forms of the protein (arrow; 35 and 31 kDa) show almost complete resistance to endoH action. b Protein extracts were obtained from Arabidopsis pOp/LhGR seedlings subjected to diVerent induction periods with Dex (3, 5 and 7 days), incubated either with (+) or without (¡) endoH and further analysed by SDS-PAGE and immunoblotting with the anti-recombinant cardosin A antiserum. After 5 days of Dex induction, the intermediate and mature forms of cardosin A (arrows; 35 and 31 kDa) have signiWcantly accumulated and both appeared as partially resistant to endoH digestion that was not detected by immunoblotting of leaf extracts. Unexpectedly, however, we did not detect any of the mature 31 kDa form (Fig. 5) even if the chase was extended to 24 h, corresponding to 78 hpi (data not shown). The 35 kDa form of cardosin A is secreted from protoplasts Previous fractionation data (Fig. 2) had suggested that a portion of the 35 kDa intermediate may be secreted during transient expression in the tobacco leaf. To investigate the sorting of cardosin A sorting in tobacco leaf cells, protoplasts expressing cardosin A were isolated from previously Fig. 4 EVects of dominant-inhibitory mutants of diVerent GTPases involved in ER-Golgi traYcking over cardosin A transport and processing. a Co-expression of cardosin A in tobacco leaves with the AtRabD dominant-negative mutant AtRab-D2a[NI] delayed the transport and processing of the AP. Cardosin A and secGFP were expressed alone (¡), or in the presence of the mutant (+). Expression of secGFP alone was the negative control in these experiments. Samples were collected 48 hpi onwards; 60 and 72 hpi, accumulation of the precursor form of cardosin A (arrow; 66 kDa) was still evident in the cases where the mutant was present (+). In comparison to the controls, the overall processing of cardosin A was clearly delayed in the samples where co-expression of this AP and the dominant-inhibitory mutant was performed. b Co-expression of cardosin A alone (¡), with the wt (SAR1wt), or with the dominant inhibitory mutant form of NtSAR1 (Sar1 [H74L]). Samples were collected 60 and 72 hpi and an uninfected leaf (Uninf.) was used as negative control; 72 hpi, accumulation of the precursor form of cardosin A (arrow; 66 kDa) was still clearly visible but only in samples where the mutant form of NtSAR1[HL] was present inWltrated tobacco leaves. Protoplast preparation was completed at 48 hpi at which time a sample of the medium and the protoplast fraction was taken. The same samples were taken 6 h later and all four were analysed by immunoblot. Figure 6a and b show that cardosin A is detected in the medium fraction within 6 h (54 hpi) indicating that a portion of the total protein synthesised is indeed secreted. The absence of the 66 kDa precursor in the medium fraction at 54 hpi when it is abundant in the protoplasts argues against release of cardosin A to the medium through rupturing of protoplasts. The presence of the 35 kDa form in the medium indicates that secretion from the protoplasts can occur before processing to the 31 kDa form. Cardosin A processing is slower in protoplasts The observation that secretion occurs before processing to the 31 kDa form is consistent with the view that processing 123 Planta 0hr - BFA 3hr - + 18hr - + + 66kDa 35kDa Fig. 5 EVects of the fungal metabolite BFA over cardosin A transport and processing. Protoplasts (1.2 £ 106) were labelled with 35S for 3 h (pulse) and cardosin A was immunoprecipitated with anti-recombinant cardosin A antiserum from samples collected at the indicated timepoints of chase, in the presence and absence of BFA. 18 h into the chase period, and in the presence of BFA, all of cardosin A appeared in the precursor form (arrow; 66 kDa). Conversely, at the same timepoint and in the absence of BFA, cardosin A processing was not inhibited and most of cardosin A molecules were at an intermediate processing stage (arrow; 35 kDa). In this set of experiments an additional band with 50 kDa (white block arrow) appeared, consistent with the cleavage of cardosin A at the C-terminus of the PSI (see Fig. 1a) a b Protoplasts 48hpi 54hpi 67hpi Medium 48hpi 73hpi 54hpi 66kDa 35kDa 35kDa Fig. 6 Cardosin A expressed in tobacco endures partial secretion. a Immunodetection of cardosin A in tobacco protoplasts transiently expressing cardosin A. Samples were collected at the indicated timepoints after inWltration. The precursor (65 kDa) and intermediate (35 kDa) forms of cardosin A were present intracellularly at all the tested time-points. b Immunodetection of cardosin A in the protoplast medium. Extracellular accumulation of the 35 kDa intermediate was evident. Altogether, the absence of the 66 kDa precursor and presence of the 35 kDa intermediate indicate that most probably cardosin A is secreted as the later form occurs after the Golgi, in the vacuole or a prevacuolar compartment. However, this interpretation was complicated by the unexpected observation that whenever cardosin A processing was monitored in protoplasts either by pulse-chase (Fig. 5) or Western blotting (Fig. 6), the mature 31 kDa vacuolar form was never observed even at time points when its presence would be expected in intact cells in inWltrated leaves (Fig. 1b). For example in inWltrated leaves, there was signiWcant conversion of the 35 kDa form to the 31 kDa form over 12 h (see Fig. 1b) whereas in the protoplast system no conversion was observed between 3 h, when the 35 kDa form was already visible, and 18 h (Fig. 5). Thus, transport of cardosin A to the vacuole, its subsequent processing, or stability within the vacuole, appear to be aVected by the protoplasting process. To investigate this 123 further approximately half the tissue from inWltrated leaves were subjected to 16 h protoplasting from 53 to 69 hpi while the other half was subjected to rapid protoplasting between 63 and 68 hpi. A sample of intact leaf was taken at 67 hpi. Each sample was analysed by immunoblot for cardosin A processing (Fig. 7). In the sample subjected to protoplasting at 53 hpi, the 66 kDa form was still visible at 69 hpi and none of the mature form was visible (Fig. 7a). In the intact leaf, the 66 kDa form was already gone at 67 hpi and the major species was the mature 31 kDa form (Fig. 7b). Similarly, in samples subjected to processing at 63 hpi, the predominant form of cardosin A in the protoplasts was the mature 31 kDa form (Fig. 7c). Thus, protoplasting appears to reduce the rate at which cardosin A is processed to the mature form but does not prevent the detection of pre-existing mature form in the vacuole. Again the protoplasts prepared at 63–68 hpi showed a higher ratio of 31 kDa to 35 kDa forms than the intact leaf tissue, consistent with the hypothesis that much of the 35 kDa form in the intact leaf is apoplastic. Discussion Here we show that cardosin A, a soluble vacuolar aspartic proteinase originally isolated from Cynara cardunculus (Veríssimo et al. 1996), is correctly targeted to the vacuoles and highly stable when expressed in two distinct heterologous systems: tobacco leaf epidermis and Arabidopsis seedlings. Furthermore, our observations indicate that the processing of cardosin A in these two expression systems proceeds similarly, showing that the targeting and processing of cardosin A can occur through conserved mechanisms. In cardoon, cardosin A is synthesised as a preproenzyme of 66 kDa which is translocated into the RER where the signal peptide (Pre) is excised and the protein becomes glycosylated at its two predicted N-glycosylation sites (Asn70 and Asn 363) (Costa et al. 1997). Procardosin A then undergoes a sequence of proteolytic events to generate the mature form of the enzyme (Ramalho-Santos et al. 1998). In cardoon, the analysis is complicated by the presence of at least Wve isoforms of vacuolar and secreted cardosins (Pimentel et al. 2006). In this work, the processing of a single cardosin A isoform was followed in two heterologous expression systems and the resulting data suggested that cleavage at the N-terminus of the PSI preceded removal of the prosegment, producing the 35 kDa intermediate and the mature 31 kDa forms, respectively. As the 50 kDa intermediate predicted by initial cleavage of procardosin A at the C-terminus of the PSI was rarely seen, cleavage at the N-terminus of the PSI must either precede or closely follow upon cleavage at Planta a Ppts early b Leaf intact c 66 kDa 66 kDa 35 kDa 35 kDa 35 kDa 31 kDa 31 kDa 31 kDa 66 kDa Ppts late Fig. 7 Cardosin A processing is slower in protoplasts than in whole tissues. a–c Immunodetection of cardosin A in tobacco leaves and protoplasts, transiently expressing cardosin A. a Samples were collected 53 hpi and isolation of protoplasts was carried out with a long digestion of leaves (ca. 16 h). Only the precursor (66 kDa) and intermediate (35 kDa) forms of cardosin A (arrows) are present. b Leaf sample collected 67 hpi. Both intermediate (35 kDa) and mature (31 kDa) forms of cardosin A (arrows) are present in the leaf extract. c Samples were collected 63 hpi and protoplasting was performed with a short digestion period (5 h). In the protoplast fraction, the mature form of cardosin A (31 kDa) is the predominant (arrow) the C-terminus. Interestingly, by pulse chase analysis in tobacco protoplasts, where the transport or processing of cardosin A appeared to be slower than in intact cells, we did observe a 50 kDa intermediate that is consistent with cleavage at the C-terminus of the PSI. This form accumulated during the Wrst 3 h of chase but was diminished at 18 h relative to the 35 kDa form, so it is likely to be an early intermediate. Thus, it may be that cleavage at the C-terminus of the PSI can occur before cleavage at its N-terminus but if so these two events must follow closely relative to the subsequent removal of the prosegment. The results obtained in this study support the sequence of processing events for cardosin A which had been previously inferred by Ramalho-Santos et al. (1998) from the analysis of cardoon extracts. In contrast to procardosin A, the Wrst processing steps of prophytepsin are cleavage downstream of the prosegment and cleavage within the PSI (Glathe et al. 1998). Removal of the PSI to generate the mature forms is delayed by 24 h. For phytepsin it was described that the presence of the prosegment is essential for the control of unwanted proteolysis by this protein (Kervinen et al. 1999). The propetide and the N-terminus of the protein are anchored in the active site cleft by ionic interaction between Lys11 and the catalytic aspartates 36 and 223 determining inactivation (Kervinen et al. 1999). This interaction is destabilised at lower pH, found in the vacuolar compartment, leading to enzyme activation. The role of an internal anchor to active site aspartates, assumed in animal APs by Lys36 p (pepsin numbering) of the propeptide, is played by Lys11 in the mature part of prophytepsin (Kervinen et al. 1999). Homology studies between plant APs suggest that this is the case for plant APs in general, however, cardosin A does not Wt in this picture since it does not possess a Lys11 homologue (Kervinen et al. 1999). Nevertheless, the late removal of the prosegment in cardosin A, presumably upon arrival to vacuole, suggests that this region may also take part in an inactivation strategy in cardosin A. In fact, it was already described that for recombinant cardosin A to be active, all it takes is the partial removal of the prosegment (Castanheira et al. 2005). Our results, obtained in vivo, diVer however to the ones described by Castanheira et al. (2005) on the processing of recombinant procardosin A in vitro, that show the partial excision of the pro segment prior to the incomplete removal of the PSI. The authors suggest that complete maturation of cardosin A in vivo requires the action of additional proteases or exopeptidases (Castanheira et al. 2005). Prophytepsin processing was also shown to be incomplete in vitro and markedly distinct to the maturation pattern observed in the in vivo situation (Glathe et al. 1998). Our results indicate that not all plant APs are processed similarly but the maturation process can be conserved in heterologous species. It is worth noting too that in contrast to the extensively studied phytepsin which is degraded in the vacuoles of tobacco protoplasts (Tormakangas et al. 2001), cardosin A is stable for several days in vacuoles of tobacco leaves, tobacco protoplasts, and Arabidopsis seedlings so it may be a better marker for such studies than phytepsin. In cardoon, both glycans become mostly resistant to removal by endoglycosidase H (Costa et al. 1997) but in tobacco and Arabidopsis extracts acquisition of endoglycosidase H resistance is less complete. In tobacco, the single glycan on the 31 kDa subunit is eYciently converted to the resistant form, particularly at later stages of transient expression, though sensitive forms always remain. The observation that not all of the N-glycans on the 31 kDa subunit acquire endoH resistance in either heterologous system is in agreement to what had been previously described for the homologous system (Costa et al. 1997). The 66 kDa procardosin A is converted by endoglycosidase H to a mixture of 64 and 62 kDa forms in tobacco leaves indicating that this form does reach the Golgi but that only one of the two glycans can be rendered endoglycosidase H resistant. As the proportion of 64 and 62 kDa endoglycosidase H digestion products of procardosin A is similar to the proportion of resistant to sensitive forms of the 35 kDa intermediate at the corresponding time point, the simplest interpretation is that the glycan on the 15 kDa subunit of cardosin A remains unprocessed in the tobacco leaf system. In contrast, in cardoon pistils, the N-glycan on the 15 kDa subunit is almost fully converted to the endoglycosidase H resistant form. 123 Planta Several observations indicate that processing of procardosin A occurs only after the protein has been exported from the ER. Firstly, endoglycosidase H resistant forms of procardosin A can be detected. Secondly, the 66 kDa procardosin A does not undergo any processing in protoplasts treated with brefeldin A which causes the ER and Golgi to fuse and inhibits further anterograde transport (Saint-Jore et al. 2002). Third, the processing of cardosin A is inhibited by co-expression with dominant inhibitory forms of Sar1, and Rab-D2 GTPases that inhibit ER to Golgi traYc via diVerent mechanisms (Andreeva et al. 2000; Batoko et al. 2000; Takeuchi et al. 2000; Phillipson et al. 2001). Although it cannot be excluded, it seems unlikely that these disparate inhibitors of membrane traYc would all act additionally on an ER processing activity that converts procardosin A to its intermediate forms. Finally, we observed the 35 kDa intermediate in the vacuole fraction suggesting that the Wnal step in processing occurs in this compartment. The presence of endoglycosidase-H-sensitive forms of the 35 and 31 kDa forms of cardosin A together with the failure of the dominant-inhibitory Rab GTPases to completely inhibit processing of procardosin A could be interpreted as evidence for processing within the ER. However, we propose that a simpler explanation, consistent with the other data, is that conversion to the endoglycosidase H resistant form is simply ineYcient, particularly when rates of traYc are high and that the mutant proteins result in only partial inhibition of transport simply because not all cells are co-infected by bacteria harbouring the cardosin A and inhibitory GTPase constructs. In support of this interpretation, it should be noted that cardosin A was expressed using a particularly high bacterial titre (OD600 0.3) to aid immunodetection while the inhibitory GTPase mutants were expressed using a 10-fold lower bacterial titre. The relatively high expression levels employed in this study may also explain why some cardosin A was missorted to the apoplast in the tobacco transient expression system, as sorting to the vacuole is presumably a saturable receptor-mediated process (daSilva et al. 2005). Similar observations were made with protoplasts expressing phytepsin (Tormakangas et al. 2001; daSilva et al. 2005). The higher ratio of 35:31 kDa form in whole tissue vs. protoplasts suggests that it is the 35 kDa form which is secreted, consistent with the view that the 31 kDa form arises after the vacuolar and secretory pathways diverge at the Golgi. Attempts to isolate pure apoplastic Xuid from intact tobacco leaf tissue were unsuccessful but the 31 and 35 kDa forms could be detected in such extracts at similar abundance to the intact leaf (data not shown). Although we found that the protoplast incubation medium contained only the 35 kDa form, we also found that processing to the 31 kDa form was inhibited in protoplasts. Therefore, the protoplast system does not allow us to 123 establish whether the 31 kDa species is also secreted. It is not clear whether the slower processing of cardosin A in protoplasts reXects a speciWc property of this molecule or a wider eVect of protoplasting on the traYcking or proteolytic activities of tobacco cells. An alternative possibility that is consistent with much of the data is that cardosin A is Wrst secreted as the 35 kDa intermediate and targeted to the vacuole only after subsequent endocytosis from the apoplast. In this view, protoplasting would inhibit the subsequent endocytosis by dilution of the secreted 35 kDa species or by interfering with endocytosis directly. While we cannot exclude such a mechanism; we note that the 66 kDa ER-localised precursor persists longer in transfected protoplasts than intact leaves, arguing that the rate of transport is indeed slower in protoplasts. The PSI domain is present only in plant APs and its role is yet to be determined. Its presence does not appear to be essential for the maintenance of AP enzymatic activity (Asakura et al. 2000; Tormakangas et al. 2001), however, some data point to a possible role for the PSI in the correct vacuolar sorting of some plant APs (Mutlu and Gal 1999; Tormakangas et al. 2001; Simões and Faro 2004). This is the case for barley phytepsin and for the soybean aspartic proteinase soyAP2, but not for soyAP1 in which deletion of the PSI has no eVect on the vacuolar targeting (Terauchi et al. 2006). The PSI sequence shows no homology with mammalian or microbial APs, but is highly similar to that of saposins and saposin-like proteins (SAPLIPs) (Simões and Faro 2004), lysosomal sphingolipid-activator proteins. Tormakangas et al. (2001) have shown that deletion of the PSI domain of phytepsin, a model vacuolar AP, causes eYcient secretion of the truncated version of the protein. Phytepsin’s PSI is highly homologous to saposin C. In mammalian cells, it has been demonstrated that a portion of the synthesized prosaposin is associated with newly synthesized procathepsin D in the ER (Zhu and Corner 1994) and both the proteins have been reported to be associated with the membrane during their biosynthesis (Rijnboutt et al. 1991). The formation of the prosaposin–procathepsin D complex could explain the proposed M6P-independent lysosomal targeting of cathepsin D mediated by saposin C (Glickman and Kornfeld 1993). Based on these facts Tormakangas et al. (2001) suggested that the targeting of phytepsin to the plant vacuole may resemble events in the mammalian system, with the exception that in plants the saposin-like PSI and phytepsin are encoded as a single precursor. Although it is yet to be conWrmed that deletion of cardosin A’s PSI domain would cause secretion of the protein, Egas et al. (2000) showed that the PSI of procardosin A induces the release of content of vesicles with membranes containing acidic phospholipids in a similar way to saposin C. Therefore, it has been suggested that Planta procardosin A binding to membranes is accomplished via the PSI while being a part of the precursor protein (Egas et al. 2000). The fact that the 35 kDa form was detected in the vacuole of both species used in this study and also in the cell wall of tobacco indicates that the PSI segment is covalently attached at the late Golgi and in most if not all cardosin A molecules delivered to the vacuole. In conclusion, we have shown that cardosin A can be targeted to the vacuoles in diverse heterologous systems indicating that conserved mechanisms are likely to be involved. Processing of cardosin A is also conserved in the native and heterologous species but it appears to diVer from that of phytepsin. As the cardosin family contains vacuolar and secreted forms, heterologous expression of these proteins and mutant derivatives may elucidate the targeting signals and vesicular transport pathways that direct plant aspartic proteases to their destinations. Acknowledgments We are grateful to Jürgen Denecke (University of Leeds, UK) for the gift of anti-calreticulin antibody, to Tony SchaeVner (GSF Research Centre, München, Germany) for the antiAt--TIP antibody and to Sandro Vitale for advice on pulse chase and immunoprecipitation. We thank an anonymous reviewer for constructive comments that improved the manuscript. This research was supported by the Portuguese Science and Technology Foundation— Fundação para a Ciência e a Tecnologia (FCT), project POCTI/BME/39765/ 2001. The corresponding author, Patrícia Duarte, was beneWciary of a PhD grant from FCT. References Andreeva A, Zheng H, Saint-Jore C, Kutuzov M, Evans D, Hawes C (2000) Organization of transport from endoplasmic reticulum to Golgi in higher plants. Biochem Soc Trans 28:505–512 Asakura T, Matsumoto I, Funaki J, Arai S, Abe K (2000) The plant aspartic proteinase-speciWc polypeptide insert is not directly related to the activity of oryzasin 1. Eur J Biochem 267:5115–5122 Bassham D, Raikhel N (2000) Unique features of the plant vacuolar sorting machinery. Curr Opin Cell Biol 12:491–495 Batoko H, Zheng H, Hawes C, Moore I (2000) A Rab1 GTPase is required for transport between the endoplasmic reticulum and Golgi apparatus and for normal Golgi movement in plants. Plant Cell 12:2201–2218 Boevink P, Oparka K, Cruz SS, Martin B, Beterridge M, Hawes C (1998) Stacks on tracks: the plant Golgi apparatus traYcs on an actin/ER network. Plant J 15:441–447 Burnette W (1981) Western blotting: electrophoretic transfer of proteins from sodium dodecyl sulphate polyacrylamide gels to unmodiWed nitrocellulose and radiographic detection with antibody and radioiodinated protein A. Anal Biochem 112:195–203 Castanheira P, Samyn B, Sergeant K, Clemente JC, Dunn BM, Pires E, Van Beeumen J, Faro C (2005) Activation, proteolytic processing, and peptide speciWcity of recombinant cardosin A. J Biol Chem 280:13047–13054 Chrispeels M, Herman E (2000) Endoplasmic reticulum-derived compartments function in storage and as mediators of vacuolar remodeling via a new type of organelle, precursor protease vesicles. Plant Physiol 123:1227–1233 Costa J, Ashford D, Nimtz M, Bento I, Frazão C, Esteves C, Faro C, Kervinen J, Pires E, Veríssimo P, Wlodawer A, Carrondo M (1997) The glycosilation of the aspartic proteinases from barley (Hordeum vulgare L.) and cardoon (Cynara cardunculus L.). Eur J Biochem 243:695–700 Craft J, Samalova M, Baroux C, Townley H, Martinez A, Jepson I, Tsiantis M, Moore I (2005) New pOp/LhG4 vectors for stringent glucocorticoid-dependent transgene expression in Arabidopsis. Plant J 41:899–918 daSilva LLP, Taylor JP, Hadlington JL, Hanton SL, Snowden CJ, Fox SJ, Foresti O, Brandizzi F, Denecke J (2005) Receptor salvage from the prevacuolar compartment is essential for eYcient vacuolar protein targeting. Plant Cell 17:132–148 Denecke J, Vitale A (1995) The use of protoplasts to study protein synthesis and transport by the plant endomembrane system. Method Cell Biol 50: 335–348 Di Sansebastiano G, Paris N, Marc-martin S, Neuhaus J (1998) SpeciWc accumulation of GFP in a non-acidic vacuolar compartment via a C-terminal propetide-mediated sorting pathway. Plant J 15:449–457 Duarte P, Figueiredo R, Pereira S, Pissarra J (2006) Structural characterization of the stigma-style complex of Cynara cardunculus (Asteraceae) and immunolocalization of cardosins A and B during Xoral development. Can J Bot 84:737–749 Dunn B (2002) Structure and mechanism of the pepsin-like family of aspartic peptidases. Chem Rev 102:4431–4458 Egas C, Lavoura N, Resende R, Brito R, Pires E, Lima M, Faro C (2000) The saposin-like domain of the plant aspartic proteinase precursor is a potent inducer of vesicle leakage. J Biol Chem 275:38190–38196 Faro C, Gal S (2005) Aspartic proteinase content of the Arabidopsis genome. Curr Protein Pept Sci 6:493–500 Faro C, Ramalho-Santos M, Vieira M, Mendes A, Simões I, Andrade R, Veríssimo P, Lin X, Tang J, Pires E (1999) Cloning and characterization of cDNA encoding cardosin A, an RGD-containing plant aspartic proteinase. J Biol Chem 274:28724–28729 Glathe S, Kervinen J, Nimtz M, Li G, Tobin G, Copeland T, Ashford D, Wlodawer A, Costa J (1998) Transport and activation of the vacuolar aspartic proteinase phytepsin in barley (Hordeum vulgare L.). J Biol Chem 273:31230–31236 Glickman JN, Kornfeld S (1993) Mannose-6-phosphate independent targeting of lysosomal-enzymes in I-cell disease B-lymphoblasts. J Cell Biol 123:99–108 Hadlington J, Denecke J (2000) Sorting of soluble proteins in the secretory pathway of plants. Curr Opin Plant Biol 3:461–468 Hadlington J, Denecke J (2001) Transient expression, a tool to address questions in plant cell biology. In: Hawes C, Satiat-Jeunemaitre B (eds) Plant cell biology: a practical approach. Oxford University Press, Oxford, pp 107–126 Hara-Nishimura I, Shimada T, Hatano K, Takeuchi Y, Nishimura M (1998) Transport of storage proteins to protein storage vacuoles is mediated by large precursor-accumulating vesicles. Plant Cell 10:825–836 JeVerson R (1987) Assaying chimeric genes in plants: the GUS gene fusion system. Plant Mol Biol Rep 5:387–405 Kervinen J, Tobin G, Costa J, Waugh D, Wlodawer A, Zdanov A (1999) Crystal structure of plant aspartic proteinase prophytepsin: inactivation and vacuolar targeting. EMBO J 18:3947–3955 Khan A, James M (1998) Molecular mechanisms for the conversion of zymogens to active proteolytic enzymes. Protein Sci 7:815– 836 Koelsch G, Mares M, Metcalf P, Fusek M (1994) Multiple functions of pro-parts of aspartic proteinase zymogens. FEBS Lett 343:6–10 Kozak M (1984) Compilation and analysis of sequences upstream from the translational start site in eukaryotic mRNAs. Nucleic Acids Res 12:857–872 Laemmli U (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 277:680–685 123 Planta Levanony H, Rubin R, Altschuler Y, Galili G (1992) Evidence for a novel route of wheat storage proteins to vacuoles. J Cell Biol 119:1117–1128 Matsuoka K, Bassham D, Raikhel N, Nakamura K (1995) DiVerent sensitivity to wortmannin of 2 vacuolar sorting signals indicates the presence of distinct sorting machineries in tobacco cells. J Cell Biol 130:1307–1318 Mutlu A, Gal S (1999) Plant aspartic proteinases: enzymes on the way to a function. Physiol Plant 105:569–576 Phillipson B, Pimpl P, daSilva L, Crofts A, Taylor J, Movafeghi A, Robinson D, Denecke J (2001) Secretory bulk Xow of soluble proteins is eYcient and COPII dependent. Plant Cell 13:2005–2020 Pimentel C, Pires E, Faro C, Rodrigues-Pousada C (2006) Genomic organization and functional analysis of Cynara cardunculus L. aspartic protease gene family. FEBS J 273:PP1071 Pimpl P, Hanton S, Taylor J, Pinto-da Silva L, Denecke J (2003) The GTPase ARF1p controls the sequence-speciWc vacuolar sorting route to the lytic vacuole. Plant Cell 15:1242–1256 Ramalho-Santos M, Veríssimo P, Faro C, Pires E (1996) Action on bovine (s1-casein of cardosins A and B, aspartic proteinase from the Xowers of the cardoon Cynara cardunculus L. Biochim Biophys Acta 1297:83–89 Ramalho-Santos M, Pissarra J, Veríssimo P, Pereira S, Salema R, Pires E, Faro C (1997) Cardosin A, an abundant aspartic proteinase, accumulates in protein storage vacuoles in the stigmatic papillae of Cynara cardunculus L. Planta 203:204–212 Ramalho-Santos M, Veríssimo P, Cortes L, Samyn B, Van Beeumen J, Pires E (1998) IdentiWcation and proteolytic processing of procardosin A. Eur J Biochem 255:133–138 Rijnboutt S, Aerts H, Geuze HJ, Tager JM, Strous GJ (1991) Mannose 6-phosphate-independent membrane association of cathepsin-D, glucocerebrosidase, and sphingolipid-activating protein in Hepg2 cells. J Biol Chem 266:4862–4868 Saint-Jore C, Evins J, Batoko H, Brandizzi F, Moore I, Hawes C (2002) Redistribution of membrane proteins between the Golgi apparatus 123 and endoplasmic reticulum in plants is reversible and not dependent on cytoskeletal networks. Plant J 29:661–678 Sambrook J, Fritsch E, Maniatis T (1989) Molecular cloning: a laboratory manual. CSHL Press, Cold Spring Harbor Simões I, Faro C (2004) Structure and function of plant aspartic proteinases. Eur J Biochem 271:2067–2075 Takeuchi M, Ueda T, Sato K, Abe H, Nagata H, Nakano A (2000) A dominant negative mutant of Sar1 GTPase inhibits protein transport from the endoplasmic reticulum to the Golgi apparatus in tobacco and Arabidopsis cultured cells. Plant J 23:517–525 Terauchi K, Asakura T, Ueda H, Tamura T, Tamura K, Matsumoto I, Misaka T, Hara-Nishimura I, Abe K (2006) Plant-speciWc insertions in the soybean aspartic proteinases, soyAP1 and soyAP2, perform diVerent functions of vacuolar targeting. J Plant Physiol 163:856–862 Tormakangas K, Hadlington J, Pimpl P, Hillmer S, Brandizzi F, Teeri T, Denecke J (2001) A vacuolar sorting domain may also inXuence the way in which proteins leave the endoplasmic reticulum. Plant Cell 13:2021–2032 Veríssimo P, Faro C, Moir A, Lin Y, Tang J, Pires E (1996) PuriWcation, characterization and partial aminoacid sequencing of two new aspartic proteinases from fresh Xowers of Cynara cardunculus L. Eur J Biochem 235:762–768 Zheng HQ, Wang GL, Zhang L (1997) Alfalfa mosaic virus movement protein induces tubules in plant protoplasts. Mol Plant Microbe In 10:1010–1014 Zheng HQ, Camacho L, Wee E, Henri BA, Legen J, Leaver CJ, Malho R, Hussey PJ, Moore I (2005) A Rab-E GTPase mutant acts downstream of the Rab-D subclass in biosynthetic membrane traYc to the plasma membrane in tobacco leaf epidermis. Plant Cell 17:2020–2036 Zhu Y, Corner G (1994) Intermolecular association of lysosomal protein precursors during biosynthesis. J Biol Chem 269:3846–3851