Planta (2001) 212: 728±738

Di€erential tissue-speci®c expression of cysteine proteinases forms the basis for the ®ne-tuned mobilization of storage globulin during and after germination in legume seeds Jens Tiedemann, Armin Schlereth, Klaus MuÈntz Institut fuÈr P¯anzengenetik und Kulturp¯anzenforschung, Corrensstrasse 3, 06466 Gatersleben, Germany Received: 14 February 2000 / Accepted: 16 August 2000

Abstract. The temporal and spatial distribution of cysteine proteinases (CPRs) was analyzed immunologically and by in situ hybridization to identify the CPRs involved in the initiation of storage-globulin degradation in embryonic axes and cotyledons of germinating vetch (Vicia sativa L.). At the start of germination several CPRs were found in protein bodies in which they might have been stored in the mature seeds. Cysteine proteinase 1 was predominantly found in organs like the radicle, which ®rst start to grow during germination. Cysteine proteinase 2 was also present at the start of germination but displayed a less-speci®c histological pattern. Proteinase B was involved in the globulin degradation of vetch cotyledons as well. The histological pattern of CPRs followed the distribution of their corresponding mRNAs. The latter were usually detected earlier than the CPRs but the in situ hybridization signals were histologically not as restricted as the immunosignals. Proteolytic activity started in the radicle of the embryonic axis early during germination. Within 24 h after imbibition it had also spread throughout the whole shoot. At the end of germination, newly synthesized CPRs might have supplemented the early detectable CPRs in the axis. In the cotyledons, only the abaxial epidermis and the procambial strands showed proteinase localization during germination. Both CPR1 and CPR2, as well as the less common proteinase B, might have been present as stored proteinases. Three days after imbibition, proteolytic activity had proceeded from the cotyledonary epidermis towards the vascular strands deeper inside the cotyledons. The histochemical detection of the CPRs was in accordance with the previously described

Abbreviations: CPR ˆ cysteine proteinase; dai ˆ days after imbibition; DIG ˆ digoxigenin; hai ˆ hours after imbibition; LLP ˆ legumain-like proteinase; PCR ˆ polymerase chain reaction; VPE ˆ vacuolar processing enzyme Correspondence to: K. MuÈntz; E-mail: [email protected]; Fax: +49-39482-5523

histological pattern of globulin mobilization in germinating vetch [Tiedemann J, et al. (2000)]. A similar link between the distribution of CPRs and globulin degradation was found in germinating seeds of Phaseolus vulgaris L. The coincidence of the histological patterns of globulin breakdown with that of the CPRs indicates that at least CPR1, CPR2 and proteinase B are responsible for bulk globulin mobilization in the seeds of the two legumes. Key words: Cysteine proteinase ± Germination ± Globulin breakdown ± Phaseolus (cysteine proteinase) ± Seedling ± Vicia (cysteine proteinase)

Introduction Mobilization of stored compounds in the seeds of dicotyledonous plants does not occur in all cells and tissues simultaneously but follows speci®c spatial and temporal patterns as summarized for legumes by Smith (1981). In seeds of legumes and rape, storage globulins are ®rst mobilized in the embryonic axis during germination (Tiedemann et al. 2000). According to Bewley and Black (1994), germination is regarded as ®nished when the radicle has broken through the seed coat in 50% of the seeds. At this time, seedling growth starts after germination. Only after protein reserves are depleted in the axis do the bulk of globulins become mobilized in the cotyledons after germination. During the entire germination period the axis seems self sucient in amino acid supply (Schlereth et al. 2001a; Tiedemann et al. 2000). Protein mobilization in seeds of vetch (Vicia sativa L.) starts where growth and di€erentiation begin, e.g. in the radicle of the embryonic axis, and in the procambial tissue and epidermis of the embryonic axis and the cotyledons. Speci®c histological patterns of globulin mobilization have also been found in garden beans (Phaseolus vulgaris L.) and rape (Brassica napus L.) although their mobilization patterns

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination

di€er species-speci®cally from those found in vetch (HoÈglund et al. 1992; Tiedemann et al. 2000). Cysteine proteinases (CPRs) are supposed to be responsible for endoproteolytic cleavages during storage-protein mobilization in cotyledons of most of the dicotyledonous seeds so far investigated (for reviews, see Wilson 1986; MuÈntz 1996) including the vetch, V. sativa (Shutov and Vaintraub 1987). Four members of the papain-like and two members of the legumain-like CPR families have been described for vetch seed (Becker et al. 1994, 1995, 1997; Fischer et al. 2000). The papain-like CPRs are CPR1, CPR2, proteinase A and CPR4 (EMBL Data Library accession numbers X75749, Z30338, Z34859 and Z99172, respectively). VsPB2 and proteinase B (accession numbers AJ007743 and Z34899) belong to the recently described new class of legumainlike CPRs. These are vacuolar processing enzymes (VPEs) with strict asparagine and aspartic acid speci®city for the amino acid residue in the P1 position of the peptide bond to be cleaved (Hara-Nishimura et al. 1998). Except for VsPB2, similar members of the CPR families have been found in the garden bean, P. vulgaris (Tanaka et al. 1991, 1993; Rotari et al. 1997; Senyuk et al. 1998; Fischer et al. 2000) [EMBL Data Library accession numbers: CP1, Z99952; CP2, Z99953; the proteinase A-like EP-C1, S16251; CP4, Z99955; and the proteinase B-like legumain-like proteinase (LLP), Z99956]. Immunoblot analysis of extracts from developing and germinating seeds indicated that a VPE corresponding to VsPB2 from vetch should also be present in P. vulgaris (Okamoto et al. 1996). Immunoblots and mRNA revealed di€erential developmental patterns for the di€erent CPRs during seed maturation, germination and seedling growth of vetch (Fischer et al. 2000; Schlereth et al. 2001a). Cysteine proteinase 2, CPR4 and VsPB2 were shown to be present in protein bodies isolated from embryonic axes as well as from cotyledons of mature dry seeds. The vacuolar processing enzyme VsPB2 is the only legumain-like CPR found in dry seeds. Cysteine proteinase 2 (Fischer et al. 2000) and the legumain-like CPR contained in protein bodies of dry seeds degraded globulins and synthetic substrates, respectively, in in vitro assays. In lysates of such protein bodies, globulins were degraded in a similar way as during in vivo mobilization of storage globulins. Globulin degradation only took place at the acidic pH optimum of both enzymes and was prohibited by CPRspeci®c inhibitors (Schlereth et al. 2001b). These results showed that within the cysteine proteinase families of vetch seed the papain-like CPR1, CPR2, CPR4 and the legumain-like vacuolar processing enzyme VsPB2 are the most likely endopeptidases catalyzing storage-globulin mobilization in embryonic axes and cotyledons during germination. Provided that this supposition is correct, then in tissues of germinating vetch seeds the pattern of CPRs detectable by immunohistochemistry should coincide with the above-described histological pattern of globulin mobilization. Using antibodies and antisense-mRNA probes we studied CPR formation during and after germination of vetch seed. Garden beans were analyzed only immunohisto-

729

chemically. The results con®rm that the expression pattern of at least the CPR1, CPR2 and proteinase B genes must be responsible for the ®nely tuned histological pattern of storage-globulin mobilization during and after seed germination in legumes. Materials and methods Plant material Seeds of vetch [Vicia sativa L. cv. Consentini (Guss.) Arcang.], obtained from the Gatersleben Gene Bank (IPK Gatersleben, Germany, accession number: VIC133) and garden bean [Phaseolus vulgaris L. var. nanus (Jusl. Aschers.) ``Imuna''], obtained from the Gatersleben Gene Bank (accession number: PHA6017) were imbibed and grown as described previously (Tiedemann et al. 2000). Immunohistochemistry Fixation, embedding, sectioning and immunohistochemistry were performed as described previously (Tiedemann et al. 2000). Polyclonal antisera against CPR1, CPR2 (Becker et al. 1994; Fischer et al. 2000) and proteinase B (Schlereth et al. 2001a) were used to detect sites where proteinases were present in the tissues.

In situ hybridization Strand-speci®c riboprobes were synthesized by in vitro transcription of polymerase chain reaction (PCR)-derived templates (Urrutia et al. 1993). Primer pairs were designed for CPR1, CPR2, CPR4 and proteinase B, all carrying the promoter sequences for T7- and T3-RNA polymerase (Table 1). The base numbers for start and end points of the primers were determined according to the demands of appropriate cloning techniques and the CPR-speci®c cDNA sequences (Becker et al. 1994, 1995; for sequence information in the EMBL Data Library, see Introduction). Isolated and puri®ed cDNA templates were ampli®ed by PCR under conditions shown in Table 2. Using a mixture of digoxigenin (DIG)-UTP labeled nucleotides, antisense probes (+probe) were generated by transcription with T3-RNA polymerase, and non-hybridizing sense probes (±probe) were generated with T7-RNA polymerase. At the end of transcription the remaining template cDNA was digested by DNase 1, and the synthesized RNA was precipitated by 3 M Naacetate and 70% ethanol overnight at )20 °C. After centrifugation at 12,000 g for 20 min at 4 °C the pellet was resuspended in 20 ll diethylpyrocarbonate (DEPC)-treated water containing 1 U/ ll RNAsin. The concentration of the transcription product was determined by dot-blot techniques (Leitch et al. 1994) and photometry. The size of the translation products was controlled by agarose gel electrophoresis. In situ hybridization was performed on semi-thick sections of paran-embedded tissue samples collected on glass slides (Tiedemann et al. 2000). After deparanation and rehydration the sections were subjected to a partial proteinase-K digestion for 30 min (5 lg/ml in 100 mM Tris, pH 7.4) to increase permeability. Subsequently, the slides were washed twice with autoclaved DEPC/ H2O (0.025% v/v) and then acetylated (100 mM triethanolamine pH 8.0 plus 0.25% (v/v) acetic anhydride) to avoid background hybridization. After dehydration the slides were allowed to air-dry for at least 2 h before 150 ll of hybridization mixture (50% formamide, 10% dextran sulfate, 400 lg/ml tRNA, 1 ´ Denhardt's, 2 ´ standard saline citrate (SSC), 10 U/ml RNAsin), containing 1 ng RNA probe/ll, was spread over each slide. The slides were covered with autoclaved plastic foil and incubated at 60 °C for 12 h. Subsequently the slides were washed in 2 ´ SSC,

730

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination

Table 1. Primer sequences used for synthesis of transcription templates by PCR. In the case of the promoter sequence (underlined) of T7 a guanine (printed in bold italics), which was missing in Sample

Promoter sequence

CPR1-T7 CPR2-T7 CPR4-T7 Proteinase B-T7 CPR1-T3 CPR2-T3 CPR4-T3 Proteinase B-T3

Sample-speci®c primer sequence

5¢-AAA CGA CGG CCA GTG AAT TGT AAT ACG ACT CAC TAT AGG GCG-3¢ 5¢-AAG CGC GCA ATT AAC CCT CAC TAA AGG GAA CAA AAG CTG GGT-3¢

Table 2. Polymerase chain reaction conditions used for template synthesis Duration

Temperature (°C)

Procedure

Treatment

10 30 30 70 10

95 95 50 72 72

Denaturation Denaturation Annealing Elongation Final and complete elongation Maintenance

1 40 40 40 1

min s s s s

¥

4

the published sequence data of Urrutia et al. (1993), has been inserted. Numbers in superscript and subscript give the primer positions in accordance with the published cDNA sequences

Cycle Cycles Cycles Cycles Cycle

50% formamide at 60 °C to detach the plastic foil, followed by two additional 90-min washes with the same solution. Non-hybridized single-strand RNA was digested with 5 lg/ml RNase-A (in 5 mM NaCl, 1 mM EDTA, 10 mM Tris, pH 7.5) for 30 min at 37 °C. After two washes with 2 ´ SSC the sections were placed for an additional 90 min at 60 °C in 1 ´ SSC containing 50% (v/v) formamide for the high-stringency wash. Hybridization stringency, as calculated according to the method described by Leitch and co workers (1994), was about 72%; stringency of the ®nal washing step was about 82% for all probes used. Detection of the DIG-labeled probes was performed using alkaline phosphatase-coupled antibodies and 4-nitro blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl-phosphate 4-toluidine salt (NBT/BCIP) as a substrate (these and all other chemicals used for in situ hybridization were obtained from Roche, Mannheim, Germany). Photomicroscopy Light microscopy was performed using a Zeiss (Oberkochen, Germany) Axiovert microscope. Photographs were taken on Kodak EPY 64T ®lms.

Results Proteinases in the axis of vetch Already at 3 hours after imbibition (hai) a signal for CPR1 (Fig. 1) was found in protein bodies in the tip of the radicle (Fig. 1a, arrowhead). Within 18 hai the signal had spread throughout the radicle and was even weakly present in the procambial strands of the central cylinder (Fig. 1b, arrowheads). At 24 hai CPR 1 was

5¢-(1181)CTG CAG AGT TTT TGA TGG TG(1162)-3¢ 5¢-(1256)TGA TCG TGG GTG GCA GTC TCT TGA(1234)-3¢ 5¢-(1007)CCG AAG CCT AAG AAT GGA GAA ACC(984)-3¢ 5¢-(1253)AGC AGT TCA GTA CCC TTT TCA ATG(1230)-3¢ 5¢-(19)ACC CAT CCT TTC TCT TCT TAG(39)-3¢ 5¢-(58)GAC CGC CGT TTC ATC TTC GCT CTC(81)-3¢ 5¢-(30)ACA GAG ATG GTG GCG AAA CAA AAC(53)-3¢ 5¢-(32)TCT CCA CTC TTC TCT TTT TCA CCA(55)-3¢

also found in the embryonic shoot (Fig. 1c). Labeling was strongest in the calyptra and in the cortex. Only weak signals, which were mainly restricted to single cells of leaf primordia, were found in the rhizodermal cells of the radicle and in the epidermal cells of the shoot. Cysteine proteinase 2 (Fig. 2) was ®rst detected as a weak but homogeneous signal in the shoot of vetch at about 18 hai (Fig. 2a). At 24 hai (Fig. 2b) the rhizodermis was intensely stained (Fig. 2b, arrowhead). In the shoot there was a clear staining of cortex and epidermis (Fig. 2c) together with a weaker staining of the procambial strands (Fig. 2c, arrowheads). In contrast to CPR1 and CPR2, proteinase B was absent from the embryonic axis during the ®rst 48 hai (not shown). Proteinases in the cotyledons of vetch Within the ®rst 2 dai there were strong individual di€erences in immunostaining patterns between plants of the same germination stage. To eliminate e€ects of individual divergence, all of the pictures presented are of the same specimens, which were treated with the antibodies in the same experiment under identical conditions. Both CPR1 (Fig. 3) and CPR2 (Fig. 4) were detected in the epidermal and near-subepidermal cell layers as well as in the procambial strands of the cotyledons at 1 dai (Figs. 3a and 4a, 3b and 4b, respectively). At this stage, only a sparse and weak proteinase B signal was found in some individual cells of the epidermis (Fig. 5a) and of the procambial strands (Fig. 5b). From 3 dai onwards, the signals for CPR1 and CPR2 increased and a front of proteinase activity began to move deeper into the cotyledons towards the vascular strands near the center (Figs. 3d, 4d). The signal for proteinase B, however, did not change (Fig. 5d). Figures 3c, 4c and 5c show the preimmune controls for the CPRs and reveal no background staining within the tissues of cotyledons at 3 dai. At 5 dai, CPR1 and CPR2 showed identical distribution patterns extending deep into the tissues of the cotyledons (Figs. 3e, 4e). By now the proteinase B signal had also moved into the cotyledon tissues although not as far as CPR 1 and CPR 2 (Fig. 5e). Di€erences in signal intensity between the proteinases should not

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination

731

Fig. 1a±c. Immunohistochemical detection of proteinases in the embryonic axis of vetch. Detection of CPR1. a At 3 hai (hours after imbibition), a distinct immunostaining is visible in the root tip (arrowhead). b At 12 hai, the signal for CPR1 has spread through the whole radicle tissue and is even detectable within the prevascular strands of the central cylinder (arrowheads). c At 24 hai, CPR1 is detectable within the shoot. The prevascular strands (arrowhead), and the epidermal cells show only weak staining

Fig. 2a±c. Immunohistochemical detection of proteinases in the embryonic axis of vetch. Detection of CPR2. a At 18 hai, CPR2 is present within the whole axis tissue. b The rhizodermis (arrowhead) shows a high signal strength at 24 hai. c The signal strength of CPR2 within the axis increases up to 24 hai. CPR2 is located within the prevascular strands (arrowheads)

primarily re¯ect quantitative di€erences but should mainly be caused by di€erences in antibody quality and titer. At 7 dai (Figs. 3f, 4f, 5f) the three proteinases were found in protein bodies showing di€erent degrees of storage-protein degradation. The signal for CPR2 (Fig. 4f), however, tended to be restricted to areas of active globulin degradation.

Detection of proteinase mRNAs by in situ hybridization of axis tissue Hybridization with sense probes gave no background staining, as shown for the CPR1 sense probe in Fig. 6a. With the exception of proteinase B, all other proteinase mRNAs were detected in axis tissues. The signal

732

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination

distribution was nearly homogeneous with the exception of the prevascular strands, and epidermal and rhizodermal cells which showed stronger signals (Fig. 6b: CPR1, 12 hai; Fig. 7a: CPR2, 12 hai; Fig. 7b: CPR2, 24 hai; Fig. 8: CPR4 12 hai; arrowheads). Temporal changes in signal strength were not observed, as exempli®ed for CPR2 (Fig. 7a,b). The signal for CPR2 within the rhizodermis in Fig. 7a corresponded well with the immunological detection of the enzyme in Fig. 2b.

Cysteine proteinase 2 remained present in the radicle even 24 hai, at which time globulin degradation had reached an advanced stage. Besides the homogeneous staining of radicle cells, cross-hybridization with nuclear DNA occurred in cells of the calyptra and near the central cylinder (Fig. 7b, arrowheads). This latter staining pattern might represent degenerating nuclei of dying calyptra cells and di€erentiating cells of the vascular system.

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination b

Fig. 3a±f. Immunological detection of proteinases within the cotyledons of vetch. Detection of CPR1. a At 1 dai (day after imbibition), CPR1 is localized within the epidermis and the ®rst 1±2 subepidermal cell layers. b At 1 dai, CPR1 is also present within prevascular strands. c Preimmune control of CPR1 at 3 dai. d At 3 dai, the signal for CPR1 slightly intensi®es in the area of the abaxial epidermis. e At 5 dai, the CPR1 signal has already reached the sixth to seventh subepidermal cell layer. f At 7 dai, CPR1 is still detectable within protein bodies of degraded stages Fig. 4a±f. Immunological detection of proteinases within the cotyledons of vetch. Detection of CPR2. a At 1 dai, CPR2 is localized within the epidermis and the ®rst subepidermal cell layer. b At 1 dai, CPR2 is also present within prevascular strands. c Preimmune-control of CPR2 at 3 dai. d The signal strength for CPR2 does not increase but proceeds to the third subepidermal cell row. e At 5 dai, the CPR2 signal has already reached the sixth to seventh subepidermal cell layer. f At 7 dai, CPR2 is closely correlated with areas where globulin degradation still proceeds Fig. 5a±f. Immunological detection of proteinases within the cotyledons of vetch. Detection of proteinase B. a At 1 dai, proteinase B is only sparsely localized in single cells of the epidermis. b At 1 dai, staining in prevascular strands is weak. c Preimmune-control of proteinase B at 3 dai. d At 3 dai, the signal for proteinase B slightly intensi®es in the area of the abaxial epidermis. e At 5 dai, proteinase B is still detectable in emptied protein bodies. f At 7 dai, like CPR1 and CPR2, proteinase B is still detectable and has proceeded farther into the cotyledon mesophyll

Detection of proteinase mRNA within cotyledons Sense controls of the four CPRs tested gave no signals in cotyledon tissues (CPR1, Fig. 9a; CPR2, Fig. 10a; CPR4, Fig. 11a; proteinase B, Fig. 12a). In general the in situ-hybridization signals showed a less di€erentiated histological pattern than that of the immunological detection of the proteinase polypeptides. However, it must be stressed that the proteinase mRNAs were always found in areas where corresponding CPRs were immunodetected (arrowheads) and globulins were degraded (Tiedemann et al. 2000). At 1 dai, mRNAs for all proteinases investigated were found throughout the entire cotyledon (Figs. 9±12, parts b). The signals were strongest in prevascular strands (half-arrowheads), epidermis, and subepidermal tissues (arrowheads). While the signals for CPR1 (Fig. 9), CPR2 (Fig. 10), and proteinase B (Fig. 12) had increased at 3 dai (Figs. 9c, 10c, 12c) the signal for CPR4 had slightly decreased (Fig. 11c). At 5 dai (Figs. 9±12, parts d) labeling had become absent from areas where protein degradation had been completed. Proteinases in germinating garden bean Cotyledons of garden bean seedlings contain CPRs that are homologous to vetch CPR1, CPR2 and proteinase B (for EMBL Data Library accession numbers, see Introduction). These CPRs are recognized by antibodies raised against the corresponding vetch enzymes, e.g. LLP can be detected by antibodies raised against proteinase B from vetch (Senyuk et al. 1998). An immunohistochemcial analysis of garden bean seeds,

733

during and after germination, revealed the presence of CPR1- and CPR2-homologous enzymes in the axis of germinating P. vulgaris at 12 hai (Figs. 13a and 14a, respectively). The immunosignals remained stable during the whole period of globulin degradation and showed no further histological di€erentiation. Proteinase B-like CPR was absent from the axis at all stages of germination (data not shown). In the cotyledons, CPR1-homologous proteinase (Fig. 13b) was present in procambial strands and epidermis at 1 dai. CPR2- (Fig. 14b) and proteinase Bhomologous proteinases (Fig. 15a) were found in the same tissues although the signals were relatively weak, especially in the case of CPR2. At 5 dai, a nearly identical distribution and strength of staining was found for all three proteinases in circular areas equidistant from the bundles in the center of the area (Figs. 13c, 14c, 15b). In the following days staining for these proteinases progressed in a concentric manner towards the vascular strands (Figs. 13d, 14d, 15c). Discussion According to Shutov and Vaintraub (1987) proteinases contributing to storage-protein mobilization must meet at least three criteria: (A) they have to be present together with their substrates inside protein bodies under conditions permitting catalytic activity of the enzyme (s); (B) they have to be able to degrade storage protein in vitro; (C) the time pattern of in vitro determined proteinase activity has to correspond to the in vivo pattern of storageprotein degradation. In vetch seeds, CPR1, CPR2 and proteinase B obey all three criteria (Fischer et al. 2000; Schlereth et al. 2001a). We therefore focussed our study on the temporal and spatial distribution of these CPRs in relation to the histological pattern of globulin mobilization in vetch and garden bean (Tiedemann et al. 2000). We here provide evidence that the temporal and spatial patterns of gene expression of these CPRs, which have been thought to be responsible for storage-protein mobilization, coincide with the histological pattern of globulin mobilization. These histological patterns in the embryonic axes and cotyledons during and after seed germination of vetch had previously been described by Tiedemann et al. (2000). The histochemical results show that at least two papain- and one legumain-like CPR are involved in the mobilization of stored protein reserves. Cysteine proteinases 1 and 2 are the papain-like CPRs which, in coincidence with the starting site of globulin degradation, ®rst appear in the embryonic axis. They were ®rst detected in the root tip (CPR1), and epidermal and procambial tissues (CPR2) by immunohistochemistry. Also in cotyledons, CPR signals were ®rst observed in procambial strands. Towards the end of germination, CPR1 and CPR2 were also detectable in other cotyledon tissues, their signal intensity strongly increasing during postgerminative growth of the seedlings. The legumainlike proteinase B was not detectable in the axis during the period of storage-protein mobilization but was found in cotyledons during seedling growth.

734

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination

Fig. 6a,b. In situ hybridization of embryonic axes. Detection of CPR1. a No hybridization is detectable with the sense probe used as control. b At 12 hai, CPR1 mRNA is present in the complete axis tissue. Central cylinder and rhizodermis show slightly more intense signals (arrowheads) Fig. 7a,b. In situ hybridization of embryonic axes. Detection of CPR2. a At 12 hai, CPR2 mRNA is localized in the rhizodermis and the central cylinder (arrowheads). A weaker signal is present in the

During vetch seed germination some di€erences were found between the temporal patterns of CPR1 and CPR2 as determined by immunochistochemical (this paper) or biochemical methods (Schlereth et al. 2001a). While at 3 hai the concentration of CPR1 in the radicle was locally high enough to be detected immunologically, for a biochemical analysis the total amount of CPR1 in

cortex. b At 24 hai, CPR2 mRNA is present in the whole radicle. Nuclei of the calyptra and near the central cylinder show crosshybridization (arrowheads) Fig. 8. In situ hybridization of embryonic axes. At 12 hai, CPR4 mRNA is present in the whole embryonic axis. Epidermal cells and prevascular strands show slightly stronger signals than parenchymatic cortical tissues

an axis extract was just high enough to result in a faint signal in immunoblot analysis. By means of immunohistochemistry, CPR2 was detected later in time than CPR1 and seemed to be present at low concentrations throughout the axis. The total amount of CPR2 in an axis extract, however, was high enough to be detected by immunoblotting. Also, VsPB2 could be detected in

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination

735

736

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination

b

Fig. 9a±d. In situ hybridization of vetch cotyledons. Detection of CPR1 mRNA. a Sense controls. b Antisense labeling at 1 dai. c Antisense labeling at 3 dai. d Antisense labeling at 5 dai. The signals were strongest in prevascular strands (half-arrowheads), epidermis, and subepidermal tissues (arrowheads) Fig. 10a±d. In situ hybridization of vetch cotyledons. Detection of CPR2 mRNA. See legend to Fig. 9 for further explanations Fig. 11a±d. In situ hybridization of vetch cotyledons. Detection of CPR4 mRNA. See legend to Fig. 9 for further explanations Fig. 12. In situ hybridization of vetch cotyledons. Detection of proteinase B mRNA. See legend to Fig. 9 for further explanations

extracts from organs of dry and germinating seeds but could not be immunologically identi®ed in tissue sections (see below). In vetch cotyledons, globulin mobilization proceeds from the abaxial and, with some delay, also from the adaxial epidermis towards the central vascular bundles. In garden bean cotyledons the bundle distribution pattern is di€erent from that in vetch. Phaseolin mobilization starts in circular zones of nearly equal distance from the bundles, from where it proceeds towards the central bundles. This di€erence in histological pattern of globulin mobilization was shown using speci®c immunoprobes for vetch and garden bean globulins (Tiedemann et al. 2000). However, despite these di€erences in histological patterns of globulin mobilization in vetch and garden bean, the temporal and spatial distribution of CPRs coincides exactly with the speci®c pattern of globulin mobilization in these species. This strongly underlines the probability that these CPRs catalyze globulin degradation, though other proteinases might be involved as well. In our study, co-distribution of papain- and legumain-like CPRs was observed in the cotyledons of both legumes. This ®nding supports the hypothesis of Shutov and Vaintraub (1987), according to which papain-like CPRs must initiate globulin mobilization by limited proteolysis before the legumain-like proteinase B can attack the modi®ed storage protein. Afterwards, both papain- and legumain-like CPRs, together with carboxypeptidase present in protein bodies, catalyze the complete breakdown of their substrate (Dunaevsky and Belozersky 1989). We discovered that vetch proteinase B, which degrades legumin only after its previous modi®cation by papain-like CPR (Shutov et al. 1978, 1981), was only found in cotyledon tissues where the papain-like CPR1 and CPR2 were already present. In embryonic axes of vetch, proteinase B was only detectable after germination had been accomplished and protein reserves had already been degraded quantitatively (Schlereth et al. 2001a). In garden beans the major storage protein phaseolin is degraded by an LLP present in the cotyledons. In our histochemical analysis of garden bean seedlings, LLP, which is homologous to proteinase B from vetch (Senyuk et al. 1998), was simultaneously present with papain-like CPRs in cotyledons. In embryonic axes of both species no immunosignals for a legumain-like CPR were observed during the period

of storage-globulin mobilization. But it is known from biochemical analysis that VsPB2 is present in the axis at this time until 2±3 dai, and that it can be detected by proteinase B-speci®c antibodies (Schlereth et al. 2001a). Most probably the local concentration of VsPB2 remained below the detection limits of immunohistochemistry whereas in axis extracts the total amount of VsPB2 polypeptide permitted immunodetection. Consequently, we have co-distribution of papain- and legumain-like CPRs in the axis, too, which, again, is in agreement with the hypothesis of Shutov and Vaintraub (1987). The patterns of immunodetection of CPRs closely followed the pattern of mRNA detection by in situ hybridization with CPR-speci®c antisense-mRNA probes in tissue sections of vetch seed. Biochemical analysis had shown that the time pattern and banding strength of CPR polypeptides in extracts of vetch embryonic axes and cotyledons closely follow the qualitative and quantitative pattern of their mRNAs (Becker et al. 1994, 1995, 1997; Fischer et al. 2000; Schlereth et al. 2001a). The availability of mRNA thus seems to control the presence and amount of translated CPRs. Furthermore, Dietrich and co-workers (1989) found that in epidermal cells and procambial tissue of the embryonic axis the presence of a CPR mRNA preceded the appearance of the corresponding proteinase during rape seed germination. Unfortunately, the type of papain-like CPR and its relation to rape globulin breakdown were not identi®ed. Here we show that the mRNAs of three CPRs, known to be involved in globulin degradation, are present in tissue areas where the corresponding CPRs have also been detected in coincidence with zones of globulin mobilization (Tiedemann et al. 2000). The mRNA of CPR4 was also found in these tissue areas. It is unknown whether CPR4 is active against storage globulins but we have found it in the globulin fraction of isolated protein bodies (Schlereth et al. 2000b). Unfortunately, our CPR4 antibody (Fischer et al. 2000) and even the Cys-EP antibodies from Schmid et al. (1998) that were used by Schlereth et al. (2001a,b) to detect CPR4 in immunoblots turned out to be useless for immunohistochemistry (data not shown). Cysteine proteinase mRNA signals temporally and spatially preceded CPR immunostaining but were absent again from cell layers that had been depleted of globulins (Tiedemann et al. 2000). Discrepancies between histological patterns of mRNAs and corresponding CPR immunostaining might have several explanations. First of all, the sensitivity of detection by in situ hybridization might be higher than that by immunohistostaining. From biochemical analysis we know that CPR mRNAs can always be detected approximately 1 d before the corresponding polypeptides are found on immunoblots (Fischer et al. 2000; Schlereth et al. 2001a). The balance of translation and degradation control might further contribute to the more limited histolocalization of CPRs compared with that of CPR mRNAs. The level of mRNA availability results from the combined activities of mRNA formation and degradation. The change in the amount and distribution of

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination

737

738

J. Tiedemann et al.: Histological pattern of cysteine proteinases during germination

b

Fig. 13a±d. Immunodetection of proteinases in garden bean. Detection of CPR1. a Embryonic axes at 12 hai. b Cotyledons at 1 dai. c Cotyledons at 5 dai. d Cotyledons at 7 dai. Half-arrowheads Prevascular strands; arrows front of proteinase detection Fig. 14a±d. Immunodetection of proteinases in garden bean. Detection of CPR2. a Embryonic axes at 12 hai. b Cotyledons at 1 dai. c Cotyledons at 5 dai. d Cotyledons at 7 dai. Half-arrowheads Prevascular strands; arrows front of proteinase detection Fig 15a±c. Detection of proteinase B. a Cotyledons at 1 dai. b Cotyledons at 5 dai. c Cotyledons at 7 dai. Half-arrowheads Prevascular strands; arrows front of proteinase detection

mRNA does not necessarily re¯ect only transcriptional control of CPR gene expression, although this presumably is a major factor determining the spatial and temporal pattern of CPR formation. Additional translational and posttranslational control of globulin mobilization by CPRs cannot be excluded. The search for such control mechanisms should become the subject of future studies. This research was supported by grants from the Deutsche Forschungsgemeinschaft as part of the special research project SFB 363/TP12. The authors thank Dr. Twan Rutten (Structural Cell Biology Group, Institut fuÈr P¯anzengenetik und Kulturp¯anzenforschung, Gatersleben, Germany) for proof reading the manuscript.

References Becker C, Fischer J, Nong VH, MuÈntz K (1994) PCR cloning and expression analysis of cDNAs encoding cysteine proteinases from germinating seeds of Vicia sativa L. Plant Mol Biol 26: 1207±1212 Becker C, Shutov AD, Nong VH, Senyuk VI, Jung R, Horstmann C, Fischer J, Nielsen NC, MuÈntz K (1995) Puri®cation, cDNA cloning and characterization of proteinase B, an asparaginespeci®c endopeptidase from germinating vetch (Vicia sativa L.) seeds. Eur J Biochem 228: 456±462 Becker C, Senyuk VI, Shutov AD, Nong VH, Fischer J, Horstmann C, MuÈntz K (1997) Proteinase A, a storageglobulin-degrading endopeptidase of vetch (Vicia sativa L.) seeds, is not involved in early steps of storage-protein mobilization. Eur J Biochem 248: 304±312 Bewley JD, Black M (1994) Seeds. Physiology of development and germination, 2nd edn. Plenum, New York Dietrich RA, Maslyar DJ, Heupel RC, Harada JJ (1989) Spatial patterns of gene expression in Brassica napus seedlings: identi®cation of a cortex-speci®c gene and localization of mRNAs encoding isocitrate lyase and a polypeptide homologous to proteinases. Plant Cell 1: 73±90 Dunaevsky YE, Belozersky MA (1989) The role of cysteine proteinases and carboxypeptidases in the breakdown of storage protein in buckwheat seed. Planta 179: 316±322 Fischer J, Becker C, Hillmer S, Horstmann C, Neubohn B, Schlereth A, Senyuk VI, Shutov AD, MuÈntz K (2000) The family of cysteine proteinases from Vicia seeds. Plant Mol Biol 43: 83±101 Hara-Nishimura I, Kinoshita T, Hiraiwa N, Nishimura M (1998) Vacuolar processing enzymes in protein storage vacuoles and lytic vacuoles. J Plant Physiol 152: 668±674

HoÈglund A-S, RoÈdin J, Larsson E, Rask L (1992) Distribution of napin and cruciferin in developing rape seed embryos. Plant Physiol 98: 509±515 Leitch AR, Schwarzacher T, Jackson D, Leitch IJ (1994) In situHybridisierung. Series: Labor im Fokus. Spektrum Akademischer Verlag, Heidelberg MuÈntz K (1996) Proteases and proteolytic cleavage of storage proteins in developing and germinating dicotyledonous seeds. J Exp Bot 47: 605±622 Okamoto T, Miura-Izu Y, Ishii S-I, Minamikawa T (1996) Asparaginyl endopeptidase in developing and germinating legume seeds: immunological detection and quantitation. Plant Sci 115: 49±57 Rotari V, Senyuk V, Horstmann C, Jivotovskaya A, Vaintraub I (1997) Proteinase A-like enzyme from germinated kidney bean seeds. Its action on phaseolin and vicilin. Physiol Plant 100: 171±177 Schlereth A, Becker C, Horstmann C, Tiedemann J, MuÈntz K (2001a) Comparison of globulin mobilization and cysteine proteinases in embryonic axes and cotyledons during germination and seedling growth of vetch (Vicia sativa L.) J Exp Bot 51: 1423±1433 Schlereth A, Standthardt D, Mock H-P, MuÈntz K (2000b) Stored cysteine proteinases start globulin mobilization in protein bodies of embryonic axes and cotyledons during vetch (Vicia sativa L.) seed germination. Planta 212: 718±727 Schmid M, Simpson D, Kalousek F, Gietl C (1998) A cysteine endopeptidase with a C-terminal KDEL motif isolated from castor bean endosperm is a marker enzyme for the ricinosome, a putative lytic compartment. Planta 206: 466±475 Senyuk V, Rotari V, Becker C, Zakharov A, Horstmann C, MuÈntz K, Vaintraub I (1998) Does an asparaginyl-speci®c cysteine endopeptidase trigger phaseolin degradation in cotyledons of kidney bean seedlings? Eur J Biochem 258: 546±558 Shutov AD, Koroleva TN, Vaintraub IA (1978) Participation of the protease of dormant vetch seeds in the degradation of reserve proteins during germination (in Russian). Fiziol Rast 25: 735±742 Shutov AD, Bulmaga VP, Boldt EK, Vaintraub IA (1981) A study of modi®cation of vetch seed reserve proteins during germination and limited proteolysis (in Russian). Biochimija 46: 841±850 Shutov AD, Vaintraub IA (1987) Degradation of storage proteins in germinating seeds. Phytochemistry 26: 1557±1566 Smith DL (1981) Cotyledons of leguminosae. In: Polhill RM, Raven PH (eds) Advances in legume systematics. Royal Botanical Gardens, London, pp 927±940 Tanaka T, Yamauchi D, Minamikawa T (1991) Nucleotide sequence of cDNA for an endopeptidase (EP-C1) from pods of maturing Phaseolus vulgaris fruits. Plant Mol Biol 16: 1083±1084 Tanaka T, Minamikawa T, Yamauchi D, Ogishi Y (1993) Expression of an endopeptidase (EP-C1) in Phaseolus vulgaris plants. Plant Physiol 101: 412±428 Tiedemann J, Neubohn B, MuÈntz K (2000) Di€erent functions of vicilin and legumin are re¯ected in the histopattern of globulin mobilization during germination of vetch (Vicia sativa L.) Planta 211: 1±12 Urrutia R, McNiven MA, Kachar B (1993) Synthesis of RNA probes by the direct in vitro transcription of PCR-generated DNA templates. J Biochem Biophys Methods 26: 113±120 Wilson KA (1986) Role of proteolytic enzymes in the mobilization of protein reserves in the germinating dicot seed. In: Dalling MJ (ed) Plant proteolytic enzymes, vol II. CRC Press, Boca Raton, pp 19±47