Cell, Vol. 68, 613-616,
0 1992 by Cell Press
Short Peptide Domains Target Proteins to Plant Vacuoles Maarten .I. Chrispeels’ and Natasha *Department of Biology University of California, San Diego La Jolla, California 92093-0116 tDepartment of Energy Plant Research Laboratory Michigan State University East Lansing, Michigan 48824-1312
The secretory system of plant cells delivers proteins to the plasma membrane for secretion as well as to the tonoplast (vacuolar membrane) for sequestration in the vacuole. In many cell types, the secretory system operates at a low level, delivering a daily quota of vacuolar hydrolases, cell wall proteins, and acomplement of channels, pumps, and signal receptors to the plasma membrane and the tonoplast. In some plant ceils, protein traffic in the secretory system is very intense. For example, in the storage parenchyma cells of developing seeds, more than 50% of the newly synthesized proteins enter the secretory system and accumulate in protein storage vacuoles. These proteins include not only the seed storage proteins but also plant defense proteins such as lectins and enzyme inhibitors that are toxic to beetles and mammals. Certain hormonal signals and environmental stresses, such as wounding and pathogen invasion, result in a dramatic up-regulation of protein sorting in the secretory system. For example, when a plant tissue is invaded by fungal or bacterial pathogens, different isozymes of chitinase or 8-glucanase are synthesized simultaneously. lsozymes that contain vacuolar sorting determinants will be directed to the vacuole, while others will be secreted to the cell wall. The complex system of membrane-bounded compartments that makes up the plant secretory system has been studied at the biochemical level for more than 20 years. Its pivotal role in the biosynthesis and the secretion of polysaccharides and proteins was well established before its role in protein sorting to the cell wall and the vacuole became known (reviewed in Chrispeels, 1976). The secretory system delivers many different proteins to four principal locations: the cell wall, the vacuole, the plasma membrane, and the tonoplast. The system consists of endoplasmic reticulum cisternae to which polysomes are attached (rough ER), numerous small stacks of Golgi cisternae, as well as vesicles, tubules, and membrane networks that carry proteins from the ER to the Golgi and from the Golgi to the plasma membrane or tonoplast. Events in the Endoplasmic Reticulum In the ER, the signal peptide is removed, cotranslational attachment of high mannose glycans occurs, followed by the removal of three glucose residues, protein folding and the formation of disulfide bonds, the modification of certain amino acids such as proline, and finally the formation of oligomers (see Chrispeels, 1991). Globular proteins that have neither retention signals to cause them to be retained along the secretory pathway, nor vacuolar sorting informa-
tion, will be secreted. In plant cells as in yeast and mammalian ceils, secretion is the bulk-flow pathway (Hunt and Chrispeels, 1991; see Chrispeels, 1991, and references therein). The ER residents, such as binding protein (BiP) and protein disulfide isomerase (PDI), that are involved in these posttranslational modifications appear to be retained in the ER by virtue of a C-terminal HDELlKDEL retention signal. In plant cells, BiP with its C-terminal HDEL is encoded by a small gene family, with different members expressed in different tissues (Denecke et al., 1991). Proteins with typical ER retention signals (KDEL or HDEL) include protein disulfide isomerase, an ER-localized auxin-binding protein, and a protease whose subcellular location is not known (reviewed in Chrispeels, 1991). When the C-terminus of the vacuolar lectin phytohemagglutinin is changed from KL to KDEL, the phytohemagglutinin-KDEL protein expressed in tobacco is retained in the nuclear envelope and the ER (Herman et al., 1990). However, about half the phytohemagglutinin still progresses to the storage vacuoles in the tobacco seeds, possibly due to either a poor display of the KDEL motif or its removal by carboxypeptidase. A more unexpected result was obtained by Wandelt et al.(1992), whochangedthec-terminusofthepeavacuolar storage protein vicilin to include KDEL. Expression in mesophyll cells of alfalfa resulted in a 1 OO-fold increase in the accumulation of the vicilin protein due to the presence of KDEL-the retention signal kept the protein from advancing to a compartment where it could be degraded (presumably the vacuole) and resulted in the formation of “protein bodies.” These ER-derived, membrane-bound, electron-dense structures measure 0.5-l .O urn in diameter, contain vicilin, and resemble the ER-derived protein bodies found in the endosperm cells of certain cereals such as maize and sorghum. Storage proteins such as zeins, which accumulate in ER-derived protein bodies, have no C-terminal KDEL or HDEL, demonstrating that certain proteins accumulate in the ER even in the absence of typical ER retention signals. Perhaps, zeins are associated with proteins such as BiP that do have an ER retention signal. This possibility is suggested by the recent observation that the maize mutant floury2, which has unusual protein bodies, overexpresses BiP in the endosperm (Fontes et al., 1991). The co-and posttranslational modifications that proteins undergo in the ER culminate in the formation of oligomers (see Chrispeels, 1991). Why should storage proteins form oligomers? Oligomerization greatly enhances the resistance of proteins to digestion by proteases (Ceriotti et al., 1992) and potent proteases are known to lurk in the vacuoles of plant cells. Thus, it is advantageous for the cell if monomers are not allowed to proceed out of the ER. By injecting various amounts of mRNAs of the vacuolar storage protein phaseolin into Xenopus oocytes, Ceriotti et al. (1992) showed that when the level of injected mRNA was low, only monomers of phaseolin were made, and
these monomers remained in the secretory system. However, when the level of mRNA and protein was high, trimers of phaseolin were made, and these were in part secreted into the oocyte culture medium and in part sequestered in the cells. A mutant phaseolin, in which a C-terminal a-helical domain of 59 amino acids was deleted, failed to form trimers in the oocytes, and the monomers were not secreted. While these experiments should be repeated with plant cells, they affirm the important role of oligomer formation for transport of plant secretory proteins. Sorting to the Vacuole From the ER the proteins are transported to the Golgi, where their glycans may be modified by glycosidases and glycosyltransferases. Subsequent sorting to the vacuoles requires positive targeting information. Secretion in plant cells as in other eukaryotic cells constitutes the default or bulk-flow pathway (see Chrispeels, 1991). In mammalian cells, sorting to the lysosomes, the mammalian equivalent of vacuoles, depends on information in the glycans (see Kornfeld and Mellman, 1989). But in yeast and plant cells, the vacuolar targeting information is contained in polypeptide domains. Many vacuolar proteins are proteolytically processed. Barley lectin is an example of a vacuolar protein that is processed at its C-terminus. Initially synthesized as a glycosylated 23 kd polypeptide, barley lectin dimerizes within the lumen of the ER to form an active lectin proprotein (reviewed by Raikhel and Lerner, 1991). During transport or after deposition in the vacuoles, the glycosylated C-terminal propeptide is removed from the proprotein to yield the mature lectin. In transgenic tobacco cells, barley lectin is correctly assembled, processed, and targeted to the vacuoles, indicating that the sorting machinery is similar in monocots and dicots (Wilkins et al., 1990). Deletion of the C-terminal propeptide from the prolectin results in missorting and secretion, indicating that the C-terminal propeptide is necessary for vacuolar targeting (Bednarek et al., 1990). To test the hypothesis that the C-terminal propeptide is a sufficient vacuolar signal, the sequences encoding it were fused to the C-terminus of a gene encoding a secreted protein, cucumber chitinase; the resulting chimeric gene was introduced into tobacco. The fusion protein was redirected to the vacuole, confirming that the C-terminal propeptide is a vacuolar sorting determinant. The redirection of cucumber chitinase-C-terminal propeptide fusion protein to the vacuole was 70%-750/o. However, when cucumber chitinase was fused to the entire barley prolectin sequence, the efficiency of redirection to the vacuole was 95%. These data indicate that targeting element(s) within the C-terminal propeptide of the cucumber chitinase-C-terminal propeptide fusion protein may not be presented properly and result in partial (25%-30%) secretion of chimeric protein (Bednarek and Raikhel, 1991). Other gramineae lectins, such as wheat germ agglutinin and rice, contain similar proteolytically processed C-terminal domains (reviewed in Raikhel and Lerner, 1991) and these probably serve the same function. Different isozymes of certain acid hydrolases are found
. .. ... .
. . . . . . DCMAAILANNQSiSFECiiES+iEi+
N. tabncum j3-1,3-glucanase Figure
. . .
in the vacuole and the extracellular space of plant cells. A comparison of their deduced amino acid sequences indicates the presence of C-terminal extensions on the vacuolar forms. Such is the case for the vacuolar and extracellular chitinases of tobacco and a number of other proteins induced by signals from invading pathogens. The vacuolar form of chitinase has Glu-Leu-Leu-Val-Asp-Thr-Met at its C-terminus, and this domain is absent from the homologouscell wall chitinase (Shinshi et al., 1990). To determine whether this heptapeptide is a vacuolar sorting domain, tobacco plants were transformed with a tobacco chitinase gene, with or without this C-terminus. Analysis of the intercellular leaf fluid, which contains secreted proteins, showed that the absence of the heptapeptide causes chitinase to be secreted. Furthermore, a C-terminal extension of 12 amino acids, of which the last 7 are the heptapeptide, is sufficient to redirect cucumber chitinase, which is normally secreted, to the vacuole (Neuhaus et al., 1991). The existence of C-terminal extensions on the vacuolar forms of a number of other proteins that have both vacuolar and cell wall forms, such as f3,1-+3 glucanase (reviewed in Bol et al., 1990) makes it likely that these termini also contain the vacuolar sorting information, but this remains to be shown. A comparison of the C-terminal extensions of the lectins and the vacuolar hydrolases shows no amino acid identities. However, as shown in Figure 1, these C-terminal extensions are rich in hydrophobic amino acids, and this characteristic may be recognized by the sorting machinery. In addition to having C-terminal extensions, many plant and yeast vacuolar proteins are synthesized as precursors with an N-terminal propeptide that is proteolytically removed. Sporamin, a storage protein from sweet potato, is synthesized as a preproprotein with a signal sequence followed by a 16 residue propeptide. Deletion of the propeptide leads to secretion of sporamin by transformed tobacco cells, indicating that the N-terminal propeptide is necessary for vacuolar targeting (Matsuoka and Nakamura, 1991). Similar to sporamin, the vacuolar thiol protease of barley, aleurain, is processed into its mature form by the removal of an N-terminal propeptide (Holwerda et al., 1990) whose amino acid sequence shares identities with the sporamin prodomain. Other plant proteins, such as potato 22 kd protein (Yamagishi et al., 1990; Suh et al., 1990) and potato cathespin D inhibitor (Strukelj et al., 1990; Ritonja et al., 1990), have proteolytically processed N-terminal propeptides that contain sequence identity to the N-terminal domain of sporamin, but their intracellular location is not yet known.
Figure 2. Common Motif N-Termmal Propeptides
protein D inhibitor
Comparison of the sequence of the propeptide of sporamin to other propeptide sequences (Figure 2) reveals that several share a short region of hydrophobic amino acids with a hydrophilic residue, Arg, in the center (Asn-Pro-lleArg-Leu-Pro); Asn and Ile are conserved, and Pro, Arg, and Leu can be substituted. Experiments to evaluate whether this region is also sufficient for vacuolar sorting are now in progress. Similar to sporamin, the N-terminal propeptides of two yeast vacuolar proteins, carboxypeptidase Y and proteinase A, contain elements necessary and sufficient for sorting to the yeast vacuole (Johnson et al., 1987; Klionsky et al., 1988; Valls et al., 1990). However, the vacuolar targeting signals of these yeast hydrolases do not appear to be similar to the N-terminal propeptide of sporamin. In contrast to the proteins discussed above, certain plant vacuolar proteins are synthesized without a cleaved propeptide, and targeting information must be present within the mature protein. Studies on two such proteins, phyfohemagglutinin of bean (Tague et al., 1990) and legumin of broadbean (Saalbach et al., 1991), have revealed internal polypeptide regions with vacuolar targeting information. Phytohemagglutinin is a typical legume lectin, and the crystal structure of such lectins shows a conserved P-bar-
rel structure with 13 8 sheets connected by loops at the surface of the protein (e.g., Cunningham et al., 1975). A 20 amino acid loop near the N-terminus of phytohemagglutinin includes the sequence Leu-Gln-Arg-Asp, which is similar to the sorting signal for the yeast vacuolar hydrolase, carboxypeptidase Y (Leu-Gln-Arg-Pro). Similar motifs are found in a number of seed storage proteins and lectins (Tague et al., 1990). Although a 43 amino acid N-terminal domain of phytohemagglutinin is sufficient to direct yeast invertase to the yeast vacuole (Tague et al., 1990) a larger N-terminal domain of phytohemagglutinin is needed to direct the same enzyme to the vacuoles of Arabidopsis protoplasts (Chrispeels, 1991). The vacuolar sorting information in another loop of phytohemagglutinin between amino acids 95-i 15 is presently under investigation. These findings indicate that vacuolar targeting information for plants and yeast are not the same, and different sorting mechanisms may be utilized. Sorting signals located on a C-terminal or N-terminal propeptide may enhance exposure of these signals without disrupting the secondary structure of the protein. In addition, the targeting information should be presented in the proper context to be recognized by the sorting apparatus. In yeast (Valls et al., 1990) and in tobacco (Bednarekand Raikhel, 1991) the efficiency of vacuolar targeting is influenced by the context in which the targeting signal is presented. Therefore, it is obvious that the secondary structure of protein in the region surrounding the target signal is a critical component for proper sorting and targeting and affects the signal’s accessibility to the sorting machinery. In summary, several short polypeptide domains may act as vacuolarsorting signals in plants: an N-terminal propeptide (Matsuoka and Nakamura, 1991) a C-terminal extension (Bednarek and Raikhel, 1991; Neuhaus et al., 1991), or an exposed region of a mature protein. Although no consensus sequence has been identified, there are some common features, such as a hydrophobic core or the AsnPro-lie-Arg motif, whose significance can be addressed by site-directed mutagenesis. Protein targeting through the secretory system is a fundamental biological process and appears to be conserved in eukaryotes. However, the diversity of sorting signals in mammals, yeast, and plants is evident, and there may be multiple, independent mechanisms for vacuolar protein sorting. References Bednarek.
Figure 3. A Summary and Plants
S. Y., and Raikhel,
N. V. (1991).
Bednarek, S. Y.. Wilkins, T. A., Dombrowskl, (1990). Plant Cell 2, 1145-1155.
Plant Cell 3, 1195-1206. J E., and Raikhel,
601, J. F., Linthorst, H. J. M., and Cornelissen, Rev. Phytopathol. 28, 113-138. Ceriotti, A., Pedrazzini, E., Fabbrini, Vitale, A. (1992). Eur. J. Biochem., Chrispeels,
M. J. (1976).
M. J. (1991).
B. J. C. (1990).
M. S., Zoppe, in press.
M., Bollini, FL, and
Annu. Rev. Plant Physiol.
27, 19-38. Plant Mol. Biol. 42,
Cunningham, B. A., Wang, J. L., Waxdal, (1975). J. Biol. Chem. 250, 1503-1512.
M. J., and Edelman,
Denecke, J., Goldman, M. H. S., Demolder, man, J. (1991). Plant Cell 3, 1025-1035.
J., and Botter-
Fontes, E. 8. P., Shank, B. B., Wrobel, Ft. L., Moose, S. P., O’Brian, G. R., Wurtzel, E. T., and Boston, R. S. (1991). Plant Cell 3, 483-496. Herman, E. M., Tague, B. W., Hoffman, L. M., Kjemtrup, Chrispeels, M. J. (1990). Planta 782, 305-312. Holwerda, B. C., Galvin, Plant Cell 2, 1091-1106. Hunt,
N. J., Baranski,
D. C., and Chrispeels,
Klionsky, D. J., Banta, 2105-2116. Kornfeld,
T. J., and Rogers,
M. J. (1991).
L. M., Bankaitis,
L. M., and Emr, S. D. (1988).
S., and Mellman,
I. (1989). Annu. K. (1991).
J. C. (1990).
V. A., and Emr, S. D. (1987).
Matsuoka, M., and Nakamura, 88, 834-830.
S. E., and
98, 18-25. Cell 48, 675-
Mol. Cell. Biol. 8,
Rev. Cell Biol. 5483-525. Proc.
Neuhaus, J.-M., Sticher, L., Meins, F., Jr., and Boller, T. (1991). Natl. Acad. Sci. USA 88, 10362-10366. Raikhel,
N. V., and Lerner,
D. R. (1991).
Ritonja, A., Krizaj, I., Mesko, P., Kopitar, M., Lucovnik, P., Strukelj, B., Pungercar, J., Buttle, D. J., Barrett, A. J., and Turk, V. (1990). FEBS Lett. 267, 13-15. Saalbach, K. (1991).
G., Jung, R., Kunze, G., Saalbach, Plant Cell 3, 695-708.
Shinshi, H., Neuhaus, J.-M., Mol. Biol. 14, 357-368. Strukelj, B., Pungercar, I,, and Turk, V. (1990). Suh, S.-G., Peterson, (1990). Plant Physiol.
J., and Meins,
F., Jr. (1990).
J., Ritonja,A., Krizaj, I., Gubensek, Nucl. Acids Res. 18, 4605. J. E., Stiekema, 94, 40-45.
Tague, B. W., Dickinson, Cell 2, 533-536. Valls, L. A., Winther, 361-366.
I., Adler, K., and Miintz,
J. R., and Stevens,
M. J. (1990).
T. H. (1990).
Wandelt, C. I., Khan, M. R. I., Craig, S., Schroeder, D., and Higgins, T. J. V. (1992). Plant J., in press. Wilkins, T. A., Bednarek, 2, 301-313. Yamagishi. K., Mitsumori. 17, 287-288.
W. J., and Hannapel,
C. D., and Chrispeels,
S. Y., and Raikhel, C., and Kikuta,
D. J. Plant
J. Cell Biol. 17 7, H. E., Spencer,
N. V. (1990).
Plant Mol. Biol.