First published online December 19, 2003; 10.1105/tpc.016378
The Plant Cell, Vol. 16, 270-290, January 2004,
www.plantcell.org ©2004, American Society of Plant Biologists
Storage Protein Accumulation in the Absence of the Vacuolar Processing Enzyme Family of Cysteine Proteases
Darren Gruis,
Jan Schulze and
Rudolf Jung1
Pioneer Hi-Bred International, A DuPont Company, Johnston, Iowa 50131-1004
1 To whom correspondence should be addressed. E-mail rudolf.jung{at}pioneer.com; fax 515-254-2619
 |
ABSTRACT
|
|---|
The role(s) of specific proteases in seed protein processing is only vaguely understood; indeed, the overall role of processing in stable protein deposition has been the subject of more speculation than direct investigation. Seed-type members of the vacuolar processing enzyme (VPE) family were hypothesized to perform a unique function in seed protein processing, but we demonstrated previously that Asn-specific protein processing in developing Arabidopsis seeds occurs independently of this VPE activity. Here, we describe the unexpected expression of vegetative-type VPEs in developing seeds and test the role(s) of all VPEs in seed storage protein accumulation by systematically stacking knockout mutant alleles of all four members ( VPE,  VPE,  VPE, and  VPE) of the VPE gene family in Arabidopsis. The complete removal of VPE function in the vpe vpe vpe vpe quadruple mutant resulted in a total shift of storage protein accumulation from wild-type processed polypeptides to a finite number of prominent alternatively processed polypeptides cleaved at sites other than the conserved Asn residues targeted by VPE. Although alternatively proteolyzed legumin-type globulin polypeptides largely accumulated as intrasubunit disulfide-linked polypeptides with apparent molecular masses similar to those of VPE-processed legumin polypeptides, they showed markedly altered solubility and protein assembly characteristics. Instead of forming 11S hexamers, alternatively processed legumin polypeptides were deposited primarily as 9S complexes. However, despite the impact on seed protein processing, plants devoid of all known functional VPE genes appeared unchanged with regard to protein content in mature seeds, relative mobilization rates of protein reserves during germination, and vegetative growth. These findings indicate that VPE-mediated Asn-specific proteolytic processing, and the physiochemical property changes attributed to this specific processing step, are not required for the successful deposition and mobilization of seed storage protein in the protein storage vacuoles of Arabidopsis seeds.
|
INTRODUCTION
|
|---|
Storage protein reserves in seeds provide a source of reduced nitrogen for seedling growth during germination. The cellular mechanisms involved in providing this resource appear to be highly specialized, including dedicated compartments termed protein storage vacuoles (PSVs), conserved storage proteins, and specific post-translational processing of these proteins within the PSV compartment (reviewed by Muntz, 1998 ). Although the post-translational polypeptide cleavage of storage proteins upon arrival at the PSV is common, the specific function(s) of that processing has not been elucidated in vivo. An attractive hypothesis for proteolytic processing suggests it triggers protein conformational changes that enable dense packaging and long-term stable storage of reserves within the PSV. This may be of particular importance in a compartment that also presumably contains proteolytic enzymes for the mobilization of this nitrogen reserve during germination. The lytic environment of the PSV in developing seeds appears to impose restrictions on the accumulation of both ectopically expressed nonstorage and modified storage proteins, typically resulting in low levels of accumulation and fragmentation of the desired protein (Hoffman et al., 1988; Jung et al., 1993, 1998; Kermode et al., 1995; Pueyo et al., 1995; Philip et al., 2001). Therefore, to fully use seeds as a protein resource, it is necessary to further delineate the proteolytic pathway(s) involved in storage protein processing and to understand seed physiology and the environmental interactions driving the conservation of these proteolytic processing mechanism(s) for protein storage.
Arabidopsis seeds contain legumin-type globulin and napin-type albumin proteins as the predominant seed storage protein reserves. Four genes encode legumin-type globulins (also referred to as 12S globulins or cruciferins) (Pang et al., 1988) and five genes encode napin-type albumins (also referred to as 2S albumins or arabidins) (Krebbers et al., 1988; van der Klei et al., 1993). Arabidopsis legumin-type globulins, like homologs in other plant species, are believed to be inserted cotranslationally first into the endoplasmic reticulum lumen, where the formation of specific disulfide bonds and assembly into trimers (sedimenting as a 9S complex on sucrose density gradients) occur. After transport to the PSV, proteolytic processing at a conserved Asn-Gly peptide bond by an asparaginyl endopeptidase converts the pro-form legumin-type globulin subunits to two disulfide-linked mature polypeptides referred to as - and -chains (Muntz, 1998). This processing step is accompanied by further assembly of the trimer protein complexes into hexamers that sediment as a 12S complex (Barton et al., 1982). The requirement of proteolytic processing for legumin-type globulins to assemble into hexamers has been demonstrated both in vitro using assembly and processing assays of in vitro–synthesized globulins (Dickinson et al., 1989) and in vivo by demonstrating that mutation of Asn at the P1 position of the precursor cleavage site of Vicia faba legumin B prevented its processing and precluded the assembly of subunits into hexamers (Jung et al., 1998). However, the specificity of the processing required for this step remains unclear, because studies, in which proglycinin was subjected to mild proteolysis with proteases not specifically cleaving at Asn in the P1 position, showed - and -chain–like derivatives to participate in hexameric assembly in vitro (Dickinson et al., 1989).
Proteolytic processing of napin-type albumins is more extensive and results in the removal of the N-terminal processed fragment, the internal processed fragment, and the C-terminal processed fragment to convert the pro-form to two disulfide-linked mature polypeptides referred to as small and large chains that sediment as a 2S complex in sucrose density gradients (Krebbers et al., 1988). Several of the napin-type albumin proteolytic processing steps also involve a conserved Asn residue in the P1 position of the cleavage site (Krebbers et al., 1988). Additionally, an aspartic endopeptidase purified from Brassica napus has been shown to cleave synthetic peptides corresponding to propeptide regions of the Arabidopsis napin-type albumins, suggesting the involvement of members of this family in processing (D'Hondt et al., 1993). Proteolytic processing removes the vacuolar sorting signal from the napin-type albumin of Brazil nut (Kirsch et al., 1996), which, like sorting signals in barley lectin and phaseolin (Bednarek and Raikhel, 1991; Frigerio et al., 1998), resides in a C-terminal processed fragment. Processing of the C-terminal propeptides appears not to engage Asn-specific proteases, but in at least one case it has been demonstrated to involve an aspartic protease (Runeberg-Roos et al., 1994).
An Asn-specific subclass (C13; EC 3.4.22.34) of the Cys endopeptidase family, referred to as the vacuolar processing enzymes (VPEs) or legumains, has been identified in seeds of several plant species and shown to catalyze Asn-specific storage protein processing in vitro (Hara-Nishimura et al., 1991, 1993b; Ishii, 1994; Hara-Nishimura, 1998). Interestingly, several members of the VPE gene family have been identified in other tissues, including leaf, root, nucellus, and cotyledons, of mature seeds and have been associated with various functions, including responses to various stresses, tissue senescence, and the mobilization of storage protein reserves (Becker et al., 1995; Kinoshita et al., 1995a; Linnestad et al., 1998; Fischer et al., 2000; Hayashi et al., 2001). In Arabidopsis, the VPE gene family is composed of four genes referred to as VPE, VPE, VPE, and VPE (Kinoshita et al., 1995a, 1995b; Gruis et al., 2002). VPE promoter–-glucuronidase (GUS) reporter fusion construct analysis identified the VPE promoter as driving GUS expression primarily in seeds, whereas the VPE and VPE promoters drive GUS expression in vegetative tissues (Kinoshita et al., 1999). VPE transcript was detected primarily in developing seeds by semiquantitative reverse transcriptase–mediated (RT) PCR analysis, but with peak transcript levels occurring before seed storage protein accumulation (Gruis et al., 2002). Phylogenetic analysis of the VPE gene family initially demonstrated that the gene family could be clustered into two subfamilies, with members of each group sharing a common expression location in either seed or vegetative tissues (Kinoshita et al., 1999). This evidence led to the widely held belief that seed-type VPEs constituted the subfamily of proteases involved in seed protein maturation and vegetative-type VPEs constituted the subfamily of proteases involved in senescence and the mobilization of protein reserves in germinating seedlings.
However, it appears now that this view of two subfamilies of VPEs, each with its own seed and vegetative functions, may need to be reevaluated. Recently, a third VPE subfamily emerged, referred to as the early embryogenesis type, and other members of the VPE family have been described that do not cluster in any of the three groups (Muntz et al., 2002). Furthermore, our functional in planta analysis of the seed-type VPEs in Arabidopsis (VPE and VPE) using gene knockouts did not support the concept that seed-type VPEs represent the only asparaginyl-endopeptidase processing activity in seeds (Gruis et al., 2002).
During a reinvestigation of vegetative-type VPE expression, contrary to expectations, we detected significant amounts of VPE and VPE transcripts in developing Arabidopsis seeds. Because this finding suggested the potential involvement of these proteases in processing seed storage proteins, we attempted to clarify their roles by genetic stacking of null alleles of all four identified members of the Arabidopsis VPE gene family. Here, we show that the removal of VPE activity has a profound impact on seed protein processing and legumin-type globulin physical properties in vivo. However, our results argue against the essentiality of other proposed VPE functions, including the mobilization of nitrogen reserves during germination.
|
RESULTS
|
|---|
Detection of Vegetative-Type VPE Gene Expression in Developing Seeds
Contrary to expectations, we previously observed Asn-specific processing of seed storage proteins in vivo in the absence of seed-type VPE activity (Gruis et al., 2002). Given that vegetative-type VPE gene expression (VPE and VPE) has been reported to be induced in vegetative tissues under stress conditions (Kinoshita et al., 1999), we hypothesized that the induction of vegetative-type VPE expression in developing seeds of seed-type VPE mutants may be responsible for the observed Asn-specific processing. To test this hypothesis, semiquantitative multiplexed RT-PCR was performed using VPE-specific primers in combination with primers specific for a constitutively expressed transcript (cytosolic ribosomal protein S11). As shown in Figure 1A, this analysis detected VPE transcript in a vegetative control sample (leaf) known to express VPE (Kinoshita et al., 1995a, 1999). However, contrary to expectations, prominent VPE-specific amplification products also were detected in developing seeds of wild-type plants. The ratio of the intensity of the VPE-specific band compared with the S11-specific band indicated that similar amounts of VPE transcript were present in leaf and developing seed samples of wild-type and vpe vpe double mutants. To confirm and quantify VPE transcript in developing seeds, quantitative real-time PCR was performed using independently isolated RNA from developing seeds of both wild-type and vpe vpe double-mutant plants. This analysis (Figure 1B) also detected VPE transcript in developing wild-type seeds and showed no statistically significant change (t test; P < 0.01; n = 9) of VPE transcript level in the mutant sample.
View larger version (38K):
[in this window]
[in a new window]
|
Figure 1. Expression of Arabidopsis VPE Genes in Developing Seeds.
(A) Semiquantitative multiplexed RT-PCR showing the expression of VPE () and the internal standard, cytosolic ribosomal protein S11 (R), in leaves (Leaf) and seeds at 12 DAA (12) and seeds at 15 DAA (15) of wild-type (Wt) and vpe vpe plants.
(B) Quantitative real-time RT-PCR showing relative amounts (arbitrary units) of VPE transcript in seeds of wild-type and vpe vpe plants at 15 DAA.
(C) MPSS expression profiles of VPE, VPE, VPE, and VPE genes plotted in parts per million of transcript levels in wild-type tissues. Tissues sampled were shoot inflorescences (SI), seeds at the cell division phase (CDS), seeds at the beginning of storage protein accumulation (ESS), seeds at the storage protein accumulation phase (MSS), germinating seeds at radicle protrusion (RP), stage-1 seedlings (S1), rosette leaves of 4-week-old plants (Lf), and whole roots of 2-week-old plantlets (Rt).
|
|
To further determine the significance of these observations and to relate the quantity of vegetative-type VPE gene expression (VPE and VPE) to the quantity of seed-type VPE gene expression (VPE and VPE) in seeds, queries of several Arabidopsis Massively Parallel Signature Sequencing (MPSS) high-resolution gene expression data sets with conceptual MPSS ESTs of Arabidopsis VPE genes (see Methods) were performed. MPSS gene expression data sets are essentially EST sequencing sets consisting of 1 to 2 million independently derived MPSS ESTs from a single tissue source (Brenner et al., 2000). Therefore, these comprehensive EST sequence libraries provide quantitative gene expression data reported in parts per million for each transcript. Results of the MPSS analysis are illustrated in the graph shown in Figure 1C. Corroborating the RT-PCR results, VPE transcripts were present in developing seeds concurrently with VPE and VPE transcripts. Moreover, the second Arabidopsis vegetative-type VPE gene, VPE, also was expressed in developing seeds, albeit at much lower levels than VPE (between 25-fold less in the cell division phase [CDS] and 4-fold less in the storage protein accumulation phase [MSS]). As reported previously (Gruis et al., 2002), the VPE expression profile was similar to the expression profile of seed storage protein genes, showing peak expression in seeds at 14 days after anthesis (DAA). At this stage (Figure 1C, sample MSS), VPE was the most prominent VPE gene transcript detected, approximately threefold more prevalent than VPE transcript. VPE transcript was the second most abundant VPE gene transcript detected at this stage (MSS); however, twofold to threefold higher levels of this transcript were detected earlier during seed development (Figure 1C, samples CDS and ESS). VPE also was the only VPE gene for which significant levels of transcript were detected in vegetative tissues, including leaves and roots (Figure 1C, samples Lf and Rt). The VPE gene was the most abundant VPE gene transcript during the cell division stage of seed development (CDS) and in germinating seeds (Figure 1C, sample RP). VPE transcript also was present at significant levels in all other developing seed stages assayed. Together, these data indicate that all four Arabidopsis VPE genes, including vegetative-type VPE family members, are expressed significantly in developing seeds during storage protein accumulation.
Isolation of Vegetative-Type VPE Gene Knockout Mutants
To investigate potential functions of the two Arabidopsis vegetative-type VPE genes during seed development, plants containing DNA insertion alleles in the VPE and VPE genes were isolated. A putative dSpm transposon insertion allele of VPE (vpe::dSpm1) was identified in pool 5.38 of the Sainsbury Laboratory collection by reverse screening using SLAT (Sainsbury Laboratory Arabidopsis thaliana dSpm Transposants) blots probed with DNA of VPE essentially as described for the isolation of VPE and VPE mutant alleles (Tissier et al., 1999; Gruis et al., 2002). DNA flanking the insertion site was cloned and sequenced to determine the location of the dSpm element within the gene. As illustrated in Figure 2A, the dSpm insertion in vpe::dSpm1 is located 249 bp downstream of the translational start codon in the intron after the first exon of the gene. The dSpm element used to create the Sainsbury mutant collection was designed to contain transcriptional termination sites in either orientation such that intronic insertion events would interfere with gene transcription (Tissier et al., 1999). To test whether vpe::dSpm1 is a knockout allele, multiplexed RT-PCR using VPE-specific primers annealing downstream of the dSpm insertion site in combination with primers specific for a control transcript (cytosolic ribosomal protein S11) was performed with RNA isolated from 14-DAA seeds of two homozygous vpe::dSpm1 plants and from two wild-type plants. As shown in Figure 2A, a PCR product corresponding to VPE transcript was amplified only in wild-type seed samples and not in samples of seeds homozygous for the vpe::dSpm1 allele, classifying the vpe::dSpm1 allele as a null allele.
A putative T-DNA insertion allele of VPE (vpe::T-DNA1) was identified by querying the SIGnAL World Wide Web site (http://signal.salk.edu). Seeds from the corresponding mutant line (Salk_010372) were obtained from the ABRC, and plants homozygous for the vpe::T-DNA1 allele were identified subsequently using allele-specific PCR. Analysis of the T-DNA adjacent DNA sequence was used to identify the T-DNA integration site as located within exon 5 of the VPE gene (Figure 2B). To test whether vpe::T-DNA1 is a null allele, RT-PCR was performed essentially as described above for vpe::dSpm1. As shown in Figure 2B, VPE transcript was detected clearly in wild-type control plants but not in homozygous vpe::T-DNA1 plants, indicative of a knockout allele.
Mutants homozygous for either vpe::dSpm1 or vpe::T-DNA1 were examined for visible phenotypes under normal growth conditions. No effects were observed on germination rate, vegetative growth rate, plant architecture, seed set, or senescence compared with wild-type controls. Moreover, no differences between protein profiles of mutant and wild-type seeds were detected (data not shown).
Genetic Stacking of VPE Mutant Alleles
To investigate the potential functional redundancy of VPE genes, genetic stacking of null alleles of the four unlinked Arabidopsis VPE genes was performed. The previously characterized (Gruis et al., 2002) seed-type VPE double mutant (vpe vpe) was first crossed to the vpe mutant, and triple-mutant plants (vpe vpe vpe), homozygous for the respective null alleles at each locus, were identified by allele-specific PCR analysis of the segregating F2 progeny after F1 self-pollination (data not shown). The vpe vpe vpe triple mutant then was crossed to the vpe mutant, and after F1 self-pollination, a total of 1132 F2 progeny plants were screened for the absence or presence of wild-type and mutant alleles at each VPE locus (see supplemental data online). This screen identified two vpe vpe vpe vpe quadruple-mutant plants (referred to here as vpe-quad) homozygous for null alleles at all four VPE loci, as well as plants with all possible combinations of homozygous triple-mutant alleles and homozygous double-mutant alleles of VPE genes. Figure 3A shows the results of the PCR allele analysis of specific mutant plants selected for further analysis. A minimum of two plants from each genotype were isolated (data not shown). Progeny of these plants, including vpe-quad plants, were grown for two generations under normal growth conditions side by side with wild-type plants and inspected closely for any phenotypic variation compared with the wild-type controls. In all cases, no effects were observed on germination rate, vegetative growth, plant architecture, flowering time, seed set, senescence, and light-microscopic seed morphology and PSV structure (data not shown).
View larger version (42K):
[in this window]
[in a new window]
|
Figure 3. Genotypes and Seed Protein Profiles of VPE Mutant Plants.
(A) PCR identification of each VPE allele (VPE, vpe::dSpm1, VPE, vpe::dSpm1, VPE, vpe::T-DNA1, VPE, and vpe::dSpm1) contained within individual plants numbered 1 to 8. The presence (+) or absence (-) of wild-type alleles at each VPE locus is indicated at top.
(B) SDS-PAGE analysis of protein profiles of mature seeds from mutant plants and from a homozygous vpe::dSpm1/VPE knockdown plant (KD). The genotypes shown are in the same order as in (A) and are identified at the bottom of the gel. The approximate gel locations of legumin-type proglobulins (Pro), - and -chains (Alpha and Beta), as well as napin-type large and small chains are indicated with brackets. The asterisks in lane 8 indicate alternatively processed polypeptides. Wt, wild type.
|
|
Seed Protein Profiles of VPE Mutants
The impact of the removal of VPE expression on seed storage protein processing was examined with seed protein extracts (Figure 3B) from plants with the mutant allele combinations shown in Figure 3A. A minimum of two plants from each genotype were analyzed to ensure that the SDS-PAGE protein profiles shown in Figure 3B were representative of each investigated genotype. Several observations can be made from this gel analysis. The double null mutant of the vegetative-type VPE genes (vpe vpe) did not detectably alter seed protein processing (lane 2). Mutants of seed-type VPEs, either vpe (lane 3) or vpe vpe (lane 4), showed subtle changes in the mature seed protein profiles, as reported previously (Gruis et al., 2002). The combination of the vpe vpe double mutant with the vegetative-type vpe mutant (vpe vpe vpe) did not result in any discernible additional change in the protein profile (lane 5) beyond what was observed for vpe vpe alone. However, dramatic differences in protein profiles were observed in seeds of plants homozygous for null alleles at both the VPE and VPE loci (lanes 6 to 8). The accumulation of polypeptides of the apparent molecular mass predicted for proprotein forms of the legumin-type globulin proteins was increased, whereas polypeptides corresponding to mature - and -chains were decreased significantly. Additionally, the accumulation of the mature small chains of napin-type albumins was decreased, and polypeptides of apparent molecular mass greater than that observed for mature large chains accumulated significantly. Interestingly, comparison of the protein profile in lane 6 (vpe vpe vpe) with the protein profile in lane 8 (vpe-quad) revealed subtle additional changes of legumin-type globulin and napin-type albumin accumulation that can be attributed to the vpe null allele. Therefore, both vegetative-type VPEs are involved in seed protein processing. Lastly, the seed protein profile of vpe vpe vpe (lane 7) appears to be indistinguishable from the profile of vpe-quad (lane 8), suggesting that VPE does not perform a relevant function with regard to seed protein accumulation.
To independently corroborate the observed null-allele phenotype of vegetative-type VPEs in a VPE null background, a vpe mutant plant was transformed with an RNA-silencing construct to suppress VPE expression. The seed protein profile from a resulting VPE knockdown/vpe plant (Figure 3B, lane KD) was similar to that observed for vpe vpe vpe triple mutants (lane 6), supporting the conclusion that the observed seed protein profile phenotypes of the vegetative-type VPE mutants are a direct result of the insertional interruption of VPE genes.
Alternative Proteolytic Processing of Seed Proteins
In addition to detecting polypeptides of an apparent molecular mass consistent with pro-forms of legumin-type globulins, several novel polypeptides of lower molecular mass were observed in vpe-quad seeds under reducing SDS-PAGE conditions (Figures 3B, lane 8, asterisks, and 4A, lane 2). At least some of these polypeptides cross-reacted with -chain–specific legumin antibodies (Figure 4B, lane 2), identifying them as alternatively processed legumin-type globulin polypeptides containing -chain epitopes. To determine if any of the other novel polypeptides are disulfide linked to these legumin -chain–related polypeptides, seed proteins were extracted in the presence of iodoacetamide (IAA) and separated by SDS-PAGE under oxidizing conditions. Alkylation of free sulfhydryl groups with IAA was necessary to prevent disulfide interchange reactions in legumin-type globulin subunits (Inquello et al., 1993). Without IAA added, even under oxidizing conditions, these reactions caused extensive breakage of disulfide bonds between - and -chains of Arabidopsis legumin-type globulins (data not shown). As expected, under oxidizing SDS-PAGE conditions, wild-type seed protein bands shifted to apparent molecular masses consistent with legumin-type proglobulins ( 50 kD) and napin-type proalbumins (12 kD), indicative of disulfide-linked chains for each class of storage proteins (Figures 4A and 4B, lanes 3). When IAA-treated protein from vpe-quad seeds was analyzed, it was likewise evident that many of the novel polypeptides observed under reducing SDS-PAGE conditions (Figure 4A, lane 2) were size-shifted under oxidizing conditions. Most polypeptides appeared to migrate at sizes similar to those of proproteins (Figure 4A, lane 4), including the bands that corresponded to legumin-type globulin polypeptides with -chain epitopes (Figure 4B, lane 4). However, at least one of these legumin-specific bands (40 kD) appeared to be smaller than legumin-type proglobulins, indicating alternative cleavage results in the loss of a polypeptide chain (10 kD) that is not disulfide linked to the alternatively processed subunit. Additionally, napin-type albumins accumulated in vpe-quad seeds also were size-shifted under oxidizing conditions (Figure 4A, lane 4); however, these proteins had a slightly greater apparent molecular mass than the napin-type polypeptides accumulated in the wild type (Figure 4A, lane 3). This observation is consistent with efficient VPE-independent cleavage of napin-type propolypeptides into disulfide-linked large and small chains, which still have amino acids from the N-terminal processed and internally processed propeptide fragments that remain covalently bound.
View larger version (57K):
[in this window]
[in a new window]
|
Figure 4. Alternatively Processed Storage Proteins in vpe-quad Seeds Remain as Disulfide-Bonded Polypeptides.
SDS-PAGE analysis of mature seed protein extracted under reducing (+DTT) or oxidizing (-DTT) conditions from wild-type (Wt) or vpe-quad (Q) plants. The approximate gel locations of legumin-type proglobulins (P-L), -chains (), and -chains () as well as napin-type proalbumins (P-N) and large and small chains (L&S) are indicated with brackets.
(A) Coomassie blue–stained gel.
(B) Immunoblot analysis of the same samples probed with -chain–specific legumin antibodies.
|
|
N-Terminal Amino Acid Sequence Analysis
To further investigate the nature of alternative processing in developing vpe-quad seeds, Edman degradation (Matsudaira et al., 1993) was attempted with several prominent polypeptide bands that appeared to be novel compared with wild-type seeds. Separation of seed proteins using linear sucrose density gradients (see below) and SDS-PAGE was used to further enrich protein bands before sequencing. Figure 5 shows the results of this analysis (summarized in Table 1) that was performed on polypeptides obtained from the density gradient 9S and 2S fractions of vpe-quad seeds. The wild-type seed protein profiles of density gradient 12S and 2S fractions also are shown in Figure 5 for reference. All polypeptides successfully identified from the vpe-quad 9S and 2S fractions were derivatives of legumin-type globulins and napin-type albumins, respectively. The majority of identifications corresponded to the two most highly expressed seed storage protein genes, legumin-type globulin Cruciferin1 (At4g28520) and napin-type albumin 3 (At4g27160) (F. Gruis, unpublished data).
Six polypeptides were successfully sequenced and identified from the 9S fraction of vpe-quad (Figure 5A, Table 1). The N-terminal sequences of two polypeptides with an apparent molecular mass consistent with pro-forms of legumin-type globulins (Table 1, band identifiers [IDs] 1 and 3) corresponded to the sequence of a different legumin-type globulin immediately downstream of the predicted signal peptide. Therefore, sequence and molecular mass identify these two legumin-type globulin proteins as unprocessed precursors, consistent with our previous study of the vpe null mutant (Gruis et al., 2002).
Instead of mature -chains of legumin-type globulins, vpe-quad seeds accumulated prominent polypeptides that were 1 kD greater in molecular mass (Figure 5A, band IDs 9 and 10) than the -chains that accumulated in wild-type seeds (left lane). Like wild-type -chains, these proteins failed to bind -chain–specific legumin antiserum (see Figures 9A and 9B, lanes 4, asterisks). The N-terminal sequence obtained for one of these polypeptides (Table 1, band ID 10) corresponded to a variable sequence region (Delseny and Raynal, 1999), also referred to as the "hypervariable" region (Nielsen et al., 1989), of a legumin-type globulin encoded by At4g28520, 11 residues upstream of the Asn-Gly polypeptide bond that normally is cleaved in wild-type seeds by VPE. A second At4g28520-specific polypeptide (Table 1, band ID 6) matched the N-terminal sequence immediately downstream of the signal peptide. However, the apparent mass of this polypeptide was 32 kD, which is 1 to 2 kD less than the calculated mass for the mature -chain derived from this protein. Therefore, the masses and sequences of the two polypeptides with band IDs 6 and 10 are consistent with the same alternative cleavage event that occurs in the hypervariable region of the At4g28520 protein, upstream of the normally processed Asn-Gly bond.
In addition to proteolytic cleavage of legumin-type globulins yielding novel - and -chain–like fragments, other fragments of lower molecular mass than either - and -chains also were identified. One fragment with an apparent mass of 7 kD matched the At4g28520 protein at the N-terminal sequence immediately downstream of the predicted signal peptide (Table 1, band ID 13). A second small polypeptide (13 kD) had an N-terminal sequence corresponding to the central portion of the -chain within the second variable sequence region (Delseny and Raynal, 1999) of At5g44120 (Table 1, band ID 11). The identification of several polypeptides all derived from a single legumin-type globulin gene indicated that no single preferred alternative processing pathway appeared to exist to compensate for the lack of VPE activity.
N-terminal amino acid sequencing of napin-type albumin polypeptides isolated from vpe-quad seeds (Figure 5B) was successful in identifying most of these polypeptides, with the results reported in Table 1. One of these polypeptides (Table 1, band ID 15), with an apparent molecular mass approaching the mass predicted for napin-type albumin precursors, had an N-terminal sequence corresponding to the central portion of the N-terminal processed propeptide. This identification confirmed an Edman degradation result reported previously for a peptide isolated from vpe (Gruis et al., 2002). However, the vast majority of napin-type albumin did not accumulate as a precursor-like form; instead, it was processed to novel forms (see above). The small chain–related peptide sequences determined contained seven additional N-terminal amino acids of the proprotein region (Table 1, band ID 19). This region normally is removed in the wild type by VPE proteolytic processing. Similarly, for large chain–related sequences isolated from vpe-quad seeds (Figure 5B), additional amino acid residues matching the internal processed propeptide fragment were found still attached to the large chain (Table 1, band ID 17). Interestingly, the apparent molecular mass of one napin-related polypeptide fragment remained unchanged in vpe-quad compared with the wild type (Table 1, band ID 18). Sequence analysis identified it as the At4g27160 large chain, with cleavage occurring at Phe-80/Asp-81 in the junction region between the internal processed propeptide and the large chain. The At4g27160 napin-type albumin does not contain an Asn residue in the internal processed fragment between the small and large chains. Therefore, in the wild type, cleavage by a non-VPE protease appears to occur at Phe-80/Asp-81.
All cleavage sites of napin-type albumins identified to date by N-terminal sequencing in vpe-quad seeds involved a Phe residue at the P1 or P1' position. Additionally, the cleavage of at least one legumin-type polypeptide (Table 1, band ID 11) also occurred at a Phe in the P1' position. Proteolysis at these locations is consistent in sequence context with cleavage by a member(s) of the aspartic protease gene family (D'Hondt et al., 1993; Muren and Rask, 1996).
SDS-PAGE Analysis of Polypeptide Content in Seeds during Development and Germination
To determine whether the lack of VPE processing had an effect on the rates of storage protein accumulation and/or mobilization, protein profiles from several stages of developing and germinating seeds from wild-type and vpe-quad plants were analyzed (Figure 6). As illustrated above (Figure 3), mature wild-type and vpe-quad seeds contained different forms of storage protein polypeptides. With the exception of precursor forms of legumin-type globulins, which peaked in vpe-quad seeds at 12 DAA and decreased slightly toward maturity (Figures 6A and 6B, right gels), all other polypeptide forms (bands) of storage proteins appeared to accumulate gradually throughout seed development. In both genetic backgrounds, legumin-type globulin bands (Figures 6A and 6B) and napin-type albumin bands (Figures 6A and 6C) were detected starting at 9 DAA and, in relation to total seed protein, increased most rapidly from 9 to 15 DAA. No obvious differences in the relative rates of protein deposition were observed between wild-type and vpe-quad plants.
View larger version (63K):
[in this window]
[in a new window]
|
Figure 6. SDS-PAGE and Immunoblot Analysis of Seed Protein during Development and Germination.
(A) Total protein extracted from seeds of wild-type or vpe-quad plants during seed development was compared using Coomassie blue–stained SDS-PAGE gels.
(B) and (C) Seed protein identity was confirmed on immunoblots using antiserum that cross-reacts with the -chain of legumin-type globulin (B) or the large chain of napin-type albumins (C).
(D) Total protein extracted from seeds of wild-type or vpe-quad plants during seed germination.
(E) and (F) Immunoblots of the same samples using antiserum that cross-reacts with the -chain of legumin-type globulin (E) or the large chain of napin-type albumins (F).
The developing seed stages (in DAA) are indicated at top in (A) to (C). After vernalization at 4°C for 2 days, the seeds were germinated at 25°C for the time periods (in hours) indicated at top ([D] to [F]). The approximate gel locations of legumin-type proglobulins (P-L), -chains (), and -chains () as well as napin-type large chains (L) and small chains (S) are identified with brackets.
|
|
Examination of the protein profiles of germinating seeds is illustrated in Figures 6D to 6F. In wild-type seeds (left gels), both legumin-type globulin bands and napin-type albumin bands decreased rapidly and steadily during the first 36 h, and only trace amounts were immunodetectible after 48 h of germination. A similar decline rate of legumin-type globulin bands and napin-type albumin bands was evident in germinating vpe-quad seeds; only a negligible amount of storage protein–related polypeptides was detectible after 48 h of germination (Figures 6D to 6F, right gels). Therefore, the lack of VPE activity appears not to grossly impact the stability of the novel vpe-quad–specific seed storage protein polypeptide forms in germinating seeds, nor does it impact seed storage protein mobilization in general in germinating seeds.
Assembly of Seed Proteins in VPE Mutants
A relationship between legumin-type globulin processing and assembly has been shown (Chrispeels et al., 1982; Dickinson et al., 1989; Jung et al., 1998). Because a large portion of globulins in vpe-quad seeds showed VPE-independent alternative processing, the question of storage protein assembly in VPE mutants was addressed. Protein was extracted from mature seeds and fractionated by ultracentrifugation using linear sucrose density gradients (Figures 7 and 8). The protein content in each fraction (Figures 7A and 7E) was measured with the bicinchoninic acid assay (Smith et al., 1985). The storage proteins from wild-type seeds sedimented as two major peaks, with sedimentation coefficients of 2S and 12S (Figure 7A). As expected, the napin-type albumins (Figures 7B and 7D) were identified in the 2S fractions of the gradient, and the legumin-type globulin - and -chains (Figures 7B and 7C) represented the predominant polypeptides in the 12S fractions. Interestingly, in contrast to the seed protein sedimentation profiles of many other plant species but similar to those of Brassica spp (Youle and Huang, 1981), only very minor amounts of polypeptides were observed in the 7S fractions of the gradient. The 25- to 30-kD polypeptides detected in these fractions (Figure 7B, fractions 13 to 18; see also Figure 9A, lane 1) likely represent the previously identified processed N- and C-terminal chains of vicilin-like proteins (Gruis et al., 2002), which are minor storage proteins in Arabidopsis seeds (F. Gruis, unpublished data).
View larger version (51K):
[in this window]
[in a new window]
|
Figure 7. Sedimentation Analysis of Wild-Type and vpe-quad Seed Proteins.
Protein extracted from mature seeds of wild-type (left) and vpe-quad (right) plants was separated using 6 to 20% linear sucrose density gradients.
(A) and (E) Graphs of the quantity of protein recovered in each gradient fraction (fraction number) plotted as a percentage of the total protein recovered. Sedimentation constants in Svedberg units (S) of specific fractions derived from known marker proteins are provided at top. The protein contents of selected fractions and the pellet (P) from each genotype were analyzed by SDS-PAGE and immunoblotting.
(B) and (F) Coomassie blue–stained SDS-PAGE gels.
(C) and (G) Immunoblots probed with -chain–specific legumin-type globulin antibody.
(D) and (H) Immunoblots probed with large chain–specific napin-type albumin antibody.
|
|
The seed protein sedimentation profile of vpe-quad was very different from that of the wild type (Figure 7E). In addition to the prominent 2S peak, two new peaks of 5.5S and 9S were seen; however, the prominent peak with a sedimentation coefficient of 12S was no longer detected. Similar to those in the wild type, napin-type polypeptides in vpe-quad (Figures 7F and 7H) migrated in the 2S fractions of the gradient. The peak observed at 5.5S did not coincide with major Coomassie blue–stained polypeptide bands on the SDS-PAGE gel (Figure 7F), and neither napin- nor legumin-specific antibodies detected strong immunoreactive signals in the corresponding fractions (Figures 7G and 7H). However, the 9S peak of the vpe-quad gradient contained very prominent polypeptides (Figure 7F, fractions 15 to 19). As described above (Figure 5A, Table 1), polypeptides from the vpe-quad 9S fractions were analyzed by Edman degradation. In agreement with the immunoblot analysis of the same fractions (Figure 7G), all six successfully sequenced peptides were identified as derivatives of legumin-type globulins, pro-forms, and alternatively processed polypeptides. The observed sedimentation coefficients of mature legumin - and -chains found in the wild type (12S) and of legumin pro-form polypeptides and alternatively processed chains found in vpe-quad (9S) are indicative of the assembly of the corresponding legumin subunits into hexamers and trimers, respectively (Barton et al., 1982).
Previous in vitro assembly experiments using legumin (Jung et al., 1997) suggested that mixed hexameric complexes might be formed in vivo between legumin pro-forms or alternatively processed legumin polypeptides and VPE-processed legumin polypeptides when the latter is in excess. To test this hypothesis, seed proteins from vpe and vpe vpe vpe were analyzed by sucrose gradient centrifugation (Figure 8). In vpe seeds, the majority of legumin-type globulin was processed at the conserved Asn-Gly bond, and only a small fraction of legumin-type globulin was deposited as either pro-form or alternatively processed polypeptides (Gruis et al., 2002). In vpe vpe vpe, this relationship was reversed, with the majority of legumin-type globulin being deposited as either pro-form or alternatively processed polypeptides, and only trace amounts of the protein were VPE processed into mature - and -chains (Figure 3B, lane 6). As expected, the sedimentation profile of vpe (Figure 8A) was very similar to that of the wild type, showing prominent 2S and 12S peaks. Also similar to the wild type, the predominant polypeptide bands in the 12S peak (Figure 8B) were consistent in mass with mature - and -chains of legumin-type globulin. However, as revealed by the legumin immunoblot (Figure 8C), the vast majority of the 50-kD legumin pro-forms also appeared to sediment in the 12S fractions of the gradient. The sedimentation profile of vpe vpe vpe proteins (Figure 8E) showed four peaks. In addition to the three peaks observed for vpe-quad (2S, 5.5S, and 9S), a distinct peak at 12S was observed. The legumin immunoblot of gradient samples (Figure 8G) detected the majority of legumin pro-forms and alternatively processed forms in the 9S fractions and VPE-processed legumin polypeptides (- and -chains) in the 12S factions. However, like vpe, a minor portion of legumin pro-forms exhibited a sedimentation coefficient of 12S. Figure 9 shows an SDS-PAGE separation of proteins contained in 9S and 12S peak samples obtained from each of the four density gradients (Figures 7 and 8) and analyzed in lanes next to each other. This analysis illustrates the virtually exclusive presence of alternatively processed legumin polypeptides in the 9S fractions (Figures 9A and 9B, lanes 3 and 4) and of VPE-processed legumin polypeptides in the 12S fractions (Figures 9A and 9B, lanes 5 to 7). Despite the excess of legumin pro-forms and alternatively processed forms in the vpe vpe vpe mutant, no VPE-processed legumin peptides were detected in the 9S fractions of this mutant (Figure 9B, lanes 3 and 7), indicating the efficient assembly of VPE-processed peptides into 12S hexamer complexes. The alternatively processed legumin polypeptides in 9S trimers appeared not to be competent for assembly into hexamers.
The protein sedimentation profiles plotted in the graphs in Figures 7 and 8 depict relative protein amounts and are the results of a minimum of two experimental repetitions per genotype. A trend for the apparent decrease in the net amount of legumin-type globulin extracted from seeds of vpe-quad (size or integral of the 9S peak) compared with the wild type (size or integral of the 12S peak) can be shown by plotting the protein amount detected in each peak as a percentage of the total protein detected (see supplemental data online). The 9S peak of vpe-quad (fractions 14 to 22) contained 28.11 ± 0.75% of the detected protein versus a baseline amount of 2.99 ± 0.72% in the same fractions in the wild type. The 12S peak in the wild type (fractions 23 to 31) contained 35.80 ± 1.77% of the detected protein compared with 3.11 ± 0.59% of the same fractions of vpe-quad. Although the net difference in legumin-type globulin protein accumulation was fairly small, these numbers indicate a 15 to 30% decrease of legumin-type globulin in vpe-quad seeds. Because a shift between the relative amounts of protein fractions might be reflected in an altered amino acid content, amino acid analysis was conducted with wild-type and vpe-quad seed samples (see supplemental data online). Interestingly, despite the observed protein differences between wild-type and vpe-quad seeds, no significant variation of the amino acid composition was detected.
Impact of Processing on Legumin-Type Globulin Solubility
A profound decrease in legumin-type globulin protein solubility under acidic conditions (pH 4.5 to 5.5) was observed after VPE-specific processing of pro-forms into mature - and -chains (K. Roesler, personal communication). To determine if the legumin-type globulin accumulated in vpe-quad seeds (a significant amount of which undergoes alternative proteolytic processing) shares similar solubility properties with wild-type VPE-processed protein, the solubility profile of the wild-type 12S proteins was compared with that of the 9S proteins of vpe-quad (Figure 10). The solubility profile of VPE-processed legumin-type globulin (wild type) showed the protein to be largely soluble at pH 7 to 8.5 and pH 3.5 to 4. At intermediate pH ranges, the solubility of the wild-type protein fraction was reduced gradually, with the majority of protein being insoluble at pH 5.5 to 6.0. By contrast, the solubility profile of legumin-type globulin accumulated in vpe-quad seeds showed the protein to be mostly soluble at pH 7.5 to 8.5 and mostly insoluble at pH 3.5 to 5. The solubility of the protein at intermediate pH 5.5 to 6.0 was 60 to 70%. Therefore, the solubility profile of the legumin-type globulin accumulated in vpe-quad seeds was altered markedly compared with that in wild-type seeds, supporting a function for proteolytic processing in determining this physiochemical property.
View larger version (17K):
[in this window]
[in a new window]
|
Figure 10. Solubility Properties of Legumin-Type Globulin Protein Isolated from Mature Wild-Type and vpe-quad Seeds.
Legumin-type globulin was isolated from sucrose density gradient fractions 24 to 30 of wild-type (Wt) gradients (Figure 7A) and from fractions 15 to 21 of vpe-quad gradients (Figure 7E). The solubility of protein obtained from these fractions was determined under low-ionic-strength conditions at various pH levels. After incubation of the protein sample at a given pH, the amount of protein remaining in solution was quantified and graphed as a percentage of the total protein added to the reaction. Error bars indicate 1 SD (three replications) at each data point.
|
|
To determine, whether specific subunits of the legumin-type globulin polypeptides showed distinct solubility characteristics, the pattern of the insoluble versus soluble peptides at different pH values was assessed by SDS-PAGE. However, in the wild type as well as in vpe-quad, no pH-dependent partitioning of individual legumin subunit species into either pellet or supernatant fractions was observed (data not shown).
|
DISCUSSION
|
|---|
Vegetative-Type VPE Expression in Developing Seeds
Storage protein deposition in the PSV of plant seeds involves post-translational processing of pro-form polypeptides by proteolytic cleavage at Asn residues in the P1 position of cleavage sites (Muntz, 1998). We recently showed that seed-type VPEs are not the only proteases that participate in seed protein maturation in vivo and suggested a number of protease genes, but not vegetative-type VPE genes, as being potentially involved in the redundant Asn-specific storage protein processing observed in vpe vpe plants (Gruis et al., 2002). Vegetative-type VPEs were excluded from this list, mainly because earlier studies strongly implied that vegetative-type VPE genes encode isoforms of VPE that are not expressed in seeds but are specific to vegetative tissues (Kinoshita et al., 1995a, 1995b, 1999). Therefore, the RT-PCR detection of significant amounts of VPE message in developing seeds of wild-type plants was a surprising result. However, this result was firmly supported by the MPSS transcript profiles obtained for the VPE genes. The analysis clearly showed that expression of VPE genes is not mutually exclusive to specific tissues, as implied previously. MPSS identified the expression of all four VPE genes in developing seeds, with transcript levels of each VPE gene exceeding those measured in nonseed tissues (root, leaf, shoot, and inflorescence). This result may appear contrary to previous reports, but it is supported by the demonstration of the role of the so-called vegetative VPE in seed protein processing. The discrepancy with earlier reports, which used transgenic plants expressing chimeric VPE promoter–GUS reporter genes (Kinoshita et al., 1999), may be attributable to differences in the temporal expression of the VPE genes. Our results indicate that peak levels of VPE and VPE occur at time points earlier in development than in VPE, with vegetative-type VPE expression decreasing to levels significantly less than that of VPE in seeds during peak storage protein gene expression. It is unclear whether previous studies closely examined the expression of the chimeric VPE-GUS genes at the onset and during earlier stages of seed storage protein accumulation. At late embryonic stages, it is possible that the VPE and VPE promoters drive only insignificant GUS expression, such that histochemical staining would not detect GUS activity in mature seeds. Similar reasoning also may explain why seed-type VPEs have been more easily isolated and identified in seeds from other plant species. Most if not all of these efforts may have isolated Asn-specific processing activity from seeds later in development, therefore correctly identifying seed-type VPEs as the predominant processing activity during later stages of seed development (Hara-Nishimura et al., 1991; Abe et al., 1993; Muramatsu and Fukazawa, 1993).
Functions of VPE Genes
Interestingly, the expression patterns of the VPE genes appear to be significantly different from each other, yet at least three of the four genes in Arabidopsis seem to be involved in seed storage protein processing. This variance in gene expression patterns also tends to suggest that the VPE genes may perform additional functions. The results presented here show VPE as the only VPE that had no detectable impact on seed protein processing, despite being an abundantly expressed VPE, especially during early seed development. It is possible that VPE is precluded from participating in storage protein processing by a subcellular localization that does not include the PSV. A nonvacuolar VPE localization has been reported for a VPE homolog from barley (Linnestad et al., 1998).
It may be expected in many cases that VPE gene functions would be difficult to identify from single or even double mutants, because overlapping or induced expression will act in a compensatory manner similar to what we observed with single-gene VPE mutants with regard to seed protein processing. However, this should not occur in the vpe-quad mutant, for which all VPE genes identified in the Arabidopsis genome are knocked out. Evidence confirming this hypothesis is provided by this report. Surprisingly, despite VPE being implicated in several processes throughout plant growth and development (Muntz et al., 2002), we were unable to detect any deleterious or pleiotropic effects of not having a functional VPE protease in Arabidopsis. This is not to say that under specific environmental conditions (different from those used here) the function of VPE is not essential. Multiple VPE gene paralogs can be identified in each of the plant genomes (including the moss Physcomitrella patens) (R. Jung and F. Gruis, unpublished data), for which extensive gene or transcript databases are available (www.ncbi.nlm.nih.gov/PMGifs/Genomes/PlantList.html). Some selective pressure must exist to maintain several functional members per genome of this highly conserved gene family. Directed assays for each potential VPE function might reveal what that selective pressure is.
One such function, in which VPEs have long been implicated, is the mobilization of storage protein reserves in germinating seeds and seedlings either by direct proteolysis (Shutov and Vaintraub, 1987; Schlereth et al., 2001) or indirectly by the activation of other proteolytic enzymes, such as SH-EP (Okamoto and Minamikawa, 1999). Suggestive of a functional role in seed germination, we detected significant levels of VPE, VPE, and VPE transcripts in germinating seeds during radicle protrusion and of VPE transcript in young seedlings. However, contrary to expectations, examination of the relative rates of mobilization of seed storage protein reserves during germination identified no appreciable differences between vpe-quad and wild-type controls. Germination rates and seedling vigor were unchanged between vpe-quad and the wild type. Therefore, Arabidopsis has other proteolytic mechanisms in place that efficiently mobilize reduced nitrogen reserves in the absence of VPE.
Seed Proteins Are Processed by Vegetative-Type VPE
Examination of vpe and vpe single or double mutants failed to detect any significant effect on seed storage protein processing, indicating that VPE was capable of compensating for the loss of these genes. To measure the specific contributions of VPE and VPE to storage protein processing, it was necessary to obtain seeds from plants homozygous for additional combinations of VPE mutant alleles. Investigation of the seed protein profiles from either vpe vpe or vpe |
TOP
ABSTRACT
INTRODUCTION
