Sorting of SBP in the Golgi apparatus
The results reported here indicate that SBP is sorted in DV, that is, along the same route into the PSV as the major storage proteins, legumin and vicilin. Furthermore, SBP-labeled DV were always labeled with legumin, suggesting that there is only one population of DV, transporting all cargo proteins. Until now, no conserved vacuolar sorting signal has been identified in the primary sequence of storage proteins. Common to the pea as well as the faba bean and two of the three soybean sucrose binding protein homologs is the presence of three hydrophobic amino acid residues (AVV for the pea, AFV for the faba bean, AVA for the GmSBP2 and GmSBP3) at the C-terminus that are similar to the C-terminal vacuolar sorting sequences of the garden bean seed storage protein phaseolin (AFVY; 18) and of the soybean seed storage protein β-conglycinin α′ subunit (PLSSILRAVY; 19). One might speculate that a hydrophobic motif at the C-terminus could serve as vacuolar sorting signal. However, as neither legumin or vicilin, or the third soybean homolog, GmSBP1, possess this hydrophobic C-tail, the function of hydrophobic amino acids as a putative vacuolar sorting sequence remains to be elucidated.
Stratification of cargo within the lumen of the DV
A stratified distribution of cargo proteins within the lumen of DV cannot, in our opinion, be explained by a receptor-mediated sorting mechanism as described for CCV. The sorting of storage proteins into the vacuole has been shown to be saturable (18,28), leading to the assumption that this sorting might also depend on the interaction of storage proteins with certain receptor proteins. In accordance with this hypothesis, two membrane proteins have been identified in Arabidopsis thaliana which might play a role in this process: the so-called RMR protein (3,29), and a member of the BP-80/AtELP family, AtVSR1 (30). In the case of the CCV-mediated sorting of mammalian lysosomal hydrolases, receptors are type I membrane proteins that recruit cargo proteins with a stoichiometry of about 1 to the lumenal leaflet of the membrane of the budding CCV (2,31). Similarly, the available data on the sorting of plant vacuolar hydrolases suggest that the plant vacuolar sorting receptors may function in a similar way. Cargo proteins, like the aleurain, have two binding domains which interact simultaneously with two domains of the receptor protein to give a high affinity interaction (3,4,32).
However, the lumen of the DV is completely filled with large electron opaque protein aggregates (6) and sorting of storage proteins into these vesicles is accompanied by an increase in their aggregation state (7), which increases as the DV pass through the Golgi stack (7,9). Crystal structure analysis of the precursor of the soybean 11S globulin, glycinin, a closely related homolog of pea legumin, has revealed that the precursor trimer is arranged around a 3-fold symmetry axis with dimensions of 9.5 × 9.5 × 4.5 nm (33). Under the steric conditions present in the DV it seems very unlikely that a ligand might bind via the same mechanism to a receptor located in the membrane of the DV as to a receptor located in the membrane of CCV. It seems unlikely that every single cargo molecule is bound to a single receptor molecule, and that aggregation might instead be part of the sorting process (5,34).
The observation, made in this report, that the cargo of the DV is stratified within the lumen of the vesicle adds support to this assumption. A stratification of storage protein aggregates has been described earlier. Early in developing pea cotyledons the major storage globulins form stratified protein lumps attached to the tonoplast, correlated with the time-course of their expression (35). These results might support the hypothesis that the stratification of cargo in the lumen of the pea cotyledon DV might reflect the sequence of sorting events of the different storage proteins into the DV while they are transported from the cis- to the trans side of the Golgi stack. Legumin and nonglycosylated vicilin enter the DV at the cis-Golgi cisternae where they aggregate. After the DV reach the medium and trans cisternae of the stack, due to cisternal progression, SBP is sorted into the attached DV. Because the lumen of the DV is already filled SBP stays attached to the DV membrane. This mechanism would certainly exclude the possibility that every single cargo molecule remains attached to a single receptor in the membrane of the DV.
However, stratification might also be the result of interactions between the different storage proteins present, and these interactions in turn might be important for correct sorting. Two reports have revealed that interactions between storage proteins in the secretory pathway do indeed influence the sorting process itself. In Hordeum vulgare and in soybean deletion of one of the storage proteins abolished vacuolar sorting and led to a retention of other storage proteins in the endoplasmic reticulum (36,37). In both cases it is discussed that the deleted storage protein influences the solubility of the other storage protein partners with the deletion leading to a pristine aggregation thus inhibiting the exit out of the ER. Because it is noticeable that SBP is concentrated at the membranes of DV but not of PSV it is tempting to speculate that membrane-attached SBP might influence the aggregation behavior of other storage proteins.
It has also been discussed that sorting of storage proteins in pea morphologically resembles the sorting of regulated secretory proteins into secretory granules in mammalian glands (7,8,38). The condensation of regulated secretory proteins and the formation of immature secretory granules occur within the lumen of the trans-Golgi network. According to the so-called sorting by entry model a subpopulation of the regulated secretory proteins themselves bind to the membrane, constituting a kind of ‘nucleus’ for further aggregation of other secretory proteins (39,40). Several candidate proteins for such unconventional receptors have been described, although their function is still a matter of dispute (41,42). The results obtained by the Triton X-114 phase partitioning together with the observed distribution of SBP in the DV presented in this report might indicate that some of the SBP is recruited to the membrane of the pea cotyledon Golgi. The mechanism of this recruitment remains to be elucidated. Because its biosynthesis starts relatively early during seed development, SBP is already present at the onset of globulin expression. One might speculate that SBP might thus act as a scavenger for the storage proteins, recruiting them to the membrane of nascent DV.