Structure-activity of deleted and substituted systemin, an 18-amino acid polypeptide inducer of plant defensive genes.

The primary structure-activity relationships of systemin, an 18-amino acid polypeptide from tomato leaves that regulates the expression of two wound-inducible proteinase inhibitor genes in tomato and potato plants, were investigated. Analogs of systemin, the only example of a polypeptide signal from plants, were synthesized with progressive deletions of amino acids from both the NH2 terminus and COOH terminus and assayed in young excised tomato plants. All of the analogs exhibited severely decreased proteinase inhibitor-inducing activities, indicating that the entire 18-amino acid sequence is necessary for maximal activity. Deletion of the COOH-terminal Asp abolished inducing activity. Progressive replacement of each amino acid of the entire polypeptide with Ala revealed two regions, near residues Pro13, where Ala substitution reduced activity to less than 0.2%, and Thr17, which totally inactivated the analog. Other replacements with Ala had little or only moderate effects on activity. The two inactive analogs, des-Asp18 systemin and Ala17 systemin, were potent inhibitors of the inducing activity of the native systemin. These analogs, therefore, contain structural conformations sufficient for competition with systemin, but they are not competent for proteinase inhibitor gene induction. A synthetic COOH-terminal tetrapeptide, Met-Gln-Thr-Asp, retained low proteinase inhibitor inducing activity, but virtually any replacements with other amino acids either eliminated activity or reduced the activity to very low or nearly undetectable levels. These results indicate that residues near the COOH terminus of systemin are necessary for activity, possibly involving a phosphorylation at Thr17, and that other regions of the systemin sequence are important for interacting with a receptor(s) but are not sufficient to activate proteinase inhibitor gene expression.

The primary structure-activity relationships of systemin, an 18-amino acid polypeptide from tomato leaves that regulates the expression of two woundinducible proteinase inhibitor genes in tomato and potato plants, were investigated. Analogs of systemin, the only example of a polypeptide signal from plants, were synthesized with progressive deletions of amino acids from both the NH2 terminus and COOH terminus and assayed in young excised tomato plants. All of the analogs exhibited severely decreased proteinase inhibitor-inducing activities, indicating that the entire 18amino acid sequence is necessary for maximal activity. Deletion of the COOH-terminal Asp abolished inducing activity. Progressive replacement of each amino acid of the entire polypeptide with Ala revealed two regions, near residues Prol3, where Ala substitution reduced activity to less than 0.2%, and Thr17, which totally inactivated the analog. Other replacements with Ala had little or only moderate effects on activity. The two inactive analogs, des-Aspl'systemin and Ala17 systemin, were potent inhibitors of the inducing activity of the native systemin. These analogs, therefore, contain structural conformations sufficient for competition with systemin, but they are not competent for proteinase inhibitor gene induction. A synthetic COOH-terminal tetrapeptide, Met-Gln-Thr-Asp, retained low proteinase inhibitor inducing activity, but virtually any replacements with other amino acids either eliminated activity or reduced the activity to very low or nearly undetectable levels. These results indicate that residues near the COOH terminus of systemin are necessary for activity, possibly involving a phosphorylation at Thr17, and that other regions of the systemin sequence are important for interacting with a receptor@) but are not sufficient to activate proteinase inhibitor gene expression.
An 18-amino acid polypeptide, called systemin, was recently isolated from tomato leaves that is the most powerful inducer of proteinase inhibitor genes reported from plants and is the only plant polypeptide hormone-like signaling molecule presently known (1,2 (3)(4)(5)(6). These genes are systemically induced in tomato and potato leaves in response to attacking insects and pathogens (7,8) and are considered to be part of the inducible defensive chemicals of the plants. %-Labeled systemin placed in wounds on tomato leaves was shown to be rapidly transported throughout plants (1) and is therefore a primary candidate for being a systemic signal, released in response to pest or pathogen attacks.
Systemin is proteolytically processed from an internal sequence of a ZOO-amino acid precursor called prosystemin (2). The cDNA and gene coding for prosystemin have been isolated and characterized (2). An antisense construct driven by the CaMV promoter was stably integrated into the genome of tomato and expressed, where it severely inhibited the systemic wound induction of the two proteinase inhibitors (2). This directly demonstrated that the synthesis of prosystemin is necessary for the systemic wound induction to occur in tomato plants.
We have now investigated the amino acid residues in the systemin structure that are necessary for induction of proteinase inhibitor genes. We have found that all deletions within the 18-amino acid structure of systemin severely decrease or eliminate its inducing activity. Substitutions of individual residues in systemin with other amino acid residues have revealed regions that can be modified without loss of activity and other regions that severely decrease or eliminate inducing activity. Two totally inactive systemin analogs, substituted or deleted in the penultimate and terminal amino acids at the COOH terminus, are powerful inhibitors of the proteinase inhibitor-inducing activity of native systemin. MATERIALS  212 piperidine in dimethylformamide for 2 h at room temperature, after which the resin was washed twice with 1-ml aliquots of dimethylformamide followed by three washes with 1 ml of methanol. The resin was dried by vacuum centrifugation. The peptide was cleaved by incubation in a solution of 1 ml of 5% HzO, 5% ethylmethyl sulfide, and 90% triflouroacetic acid for 2 h at room temperature with stirring. The volume was reduced in a rotary evaporator to 0.1 ml, and the liquid was removed with a syringe and transferred to a 13 X 100-mm glass tube and chilled. Cold ether (3 ml) was added, thoroughly mixed, and then centrifuged at 1300 X g for 5 min in a Sorvall Rt 6000B centrifuge. The resulting pellet was lyophilized and stored at 2 "C. All peptides were purified on reverse phase HPLC usjng a Vydac Cla semipreparative column (10 X 250 mm, 5 pm, 300 A) with a 0-30% acetonitrile gradient in 0.1% trifluoroacetic acid. The flow rate was 2.0 ml/min. The carboxyl-terminal tetrapeptide analogs were purified using isocratic elution at 5% acetonitrile in 0.1% trifluoroacetic acid. The equipment for HPLC included two Beckman model 112 pumps with a Beckman model 420 gradient controller. Eluted peaks were detected at 225 nm with a Beckman model 165 variable wavelength detector and displayed on a Hewlett-Packard model 3390A Integrator. Protein content was quantified with the BCA protein assay (Pierce Chemical Co.). Purity of all analogs was confirmed by analytical HPLC. In Fig. 1 is shown a typical CIS HPLC elution profile of systemin and of two polypeptides containing residues 5-18 and 11-18 of systemin, respectively. All of the synthetic polypeptides reported in this study were purified in this manner and analyzed by amino acid analyses to confirm their chemical compositions.
Tomato plants were grown to a two-leaf stage (12-14 days after planting) under 17-h days at 28 "C with >300 microeinsteins. m-*. s-' of light and 7-h nights at 18 "C. The plants were excised at the base of the stem with a double-edged razor blade, and the cut stem was placed in a small vial containing 90 p1 of a solution containing various concentrations of the peptides buffered with 15 mM sodium phosphate, pH 6.5. After the plants have taken up the solution (about 20-30 min), they were transferred to 20-ml vials filled with distilled water and placed in a sealed plexiglass box containing a COz trap (9) and incubated for 24 h under constant light of 300 microeinsteins.

rn-'.s"
at 28 "C. The low CO, level enhances the synthesis of inhibitor proteins in leaves of excised plants in response to signal molecules such as oligouronides and systemin, but the reasons are not known. Leaves of individual plants were mascerated with a small mortar and pestle, and the juice was expressed and assayed immunologically for Inhibitor I content (10,11).

RESULTS AND DISCUSSION
Systemin had been shown previously to be a powerful inducer of the expression of genes coding for proteinase Inhibitor I and Inhibitor I1 in leaves of young excised tomato  plants, when supplied to the plants through their cut stems (1). T h e maximum levels of accumulation of Inhibitor I protein in the leaves of plants supplied with systemin in the femtomole to picomole range was about two times that of Inhibitor I1 protein (1). In Fig. 2 is shown the accumulation of Inhibitor I protein in leaves of young tomato plants in response to increasing concentrations of systemin. The induction of Inhibitor I in leaves of excised tomato plants has therefore provided a convenient and quantitative assay to examine the structure-activity relationships of several synthetic derivatives of systemin in order to identify individual amino acids, or sequences of amino acids, that may by important t o t h e biological activity of the native molecule.
Effects of Deletions and Substitutions in the NH2and COOH-terminal Regions on the Biological Activity of Systemin-Polypeptides were synthesized that represented stepwise deletions of systemin from both the COOH terminus and the NHp terminus. The Inhibitor I-inducing activity of each polypeptide was determined in the detached tomato plant bioassay. The quantity of each polypeptide that induces 50% maximal Inhibitor I accumulation was determined. In Table  I is    activity from the remaining 17-amino acid polypeptide. The cumulative evidence indicates that the tetrapeptide Met-Gln-Thr-Asp is sufficient for biological activity but not for the full activity found in native systemin. A recent analysis of the solution structure of systemin by proton NMR (14) did not reveal any persistent common secondary or tertiary structural elements in the polypeptide, but two weak distinct structural features were detected at the carboxyl terminus that may be due to internal hydrogen bonding. Whether these conformations are related to the activity found associated with this region of systemin remains to be established.

S K P P S K ' R D P P K M Q T D
To further investigate the role of alanine as the NH2terminal amino acid in systemin, analogs of systemin were synthesized that contained several different single amino acid substitutions at the NH2 terminus and assayed. In Table I1 is shown the various substitutions and their effects on the inducing activities, compared with native systemin (NH2terminal Ala). Substitutions of Ala with D-Ala or Ser had little effect on the inducing activity of the modified systemins, whereas substitutions with @-Ala, acetyl-L-Ala, Tyr, Leu, Asp, and Gly severely reduced the activities below that of native systemin. The activities of these latter derivatives were similar to systemin in which Ala has been deleted (Tables I and  11). The effects of the different substitutions on activity could be due to conformational constraints necessary for the interaction of systemin with a receptor. The data can be interpreted to mean that the free NH2-terminal a-amino group, together with a small R group, i.e. Ala or Ser, are together necessary for full activity, whereas modifications that eliminate the a-amino positive charge, i.e. @-Ala or acetyl-L-Ala, or that introduce large bulky R groups, i.e. Tyr, Leu, or Asp, or no R group, i e . Gly, severely reduce the activity. On the other hand, the activities of the analogs may reflect the halflives of the different polypeptides when supplied to the plants during assay. The pattern of activities generally fits the pattern expected by the N-end rule of protein degradation (15), with NHp-terminal Ala and Ser being about equal in activity but more active than the other substitutions. In plants, however, the degradation of polypeptides has not been studied in detail.
The smallest fragment of systemin that retained some inducing activity was the Met-Gln-Thr-Asp at the COOH terminus (Table I). Several tetrapeptide analogs of this motif were synthesized and assayed for inducing activities. In Table  I11 is summarized the analogs and their activities when supplied to plants at the level at which the wild type polypeptide produces maximal activity. Deletion of the COOH-terminal Asp completely abolished activity, similar to the COOHterminal deletion in the native systemin. All substitutions severely reduced the activity of the tetrapeptide, but substitutions of Asp for Met completely abolished activity. Interestingly, several substitutions for Met other than Asp did not completely eliminate activity.
Effects of Scanning the Systemin Sequence with Alanine Substitutions on Biological Actiuity-The powerful proteinase inhibitor-inducing activity of systemin suggested that the polypeptide may initially interact with a receptor on or in tomato leaf cells. Such an interaction would likely involve conformational changes that would result in the positioning of systemin at the receptor site. In seeking such a receptor through the use of affinity chromatography or photoaffinity labels derived from systemin, identifying regions of systemin that can be modified without affecting biological activity were of greatest interest. In order to determine whether the modification of individual amino acids within systemin would affect inducing activity, analogs of systemin were synthesized that were singly substituted with Ala at each of the positions of the polypeptide, and the biological activities of each were determined. The biological activity of the native systemin with Ala as its NH:! terminus was used as the standard for comparing the activities of the seventeen analogs containing Ala at each position in the sequence. In Fig. 3 is shown these comparisons, reported as the ratio of the activity (femtomoles required to produce 50% maximal activity) of the wild type (NHp-terminal alanine) over the activities of the various analogs. Substitution of Ala for residues at positions 2-6, 8, 9, 10, 14, and 15 produced only modest changes in activity, indicating that these residues appear to be relatively unimportant to the activity of systemin. Ala substitutions a t positions 7, 11, 12, 16, and 18 produced moderate changes in activity; substitutions at residue 13 (Pro) caused a major reduction in activity; and Ala at position 17 (Thr) totally eliminated biological activity. Thus, the activities of the Alasubstituted analogs indicate that the amino acid sequences of at least two regions of systemin are vitally important for biological activity, i.e. residues 11-13 and 16-18.
Substitution of the penultimate Thr in systemin with Ala Alanine Substitution Sites  FIG. 3. Induction of Inhibitor 1 synthesis in leaves of young tomato plants by systemin and systemin analogs, as described in Fig. 1. One-half maximal activity for each analog was calculated from activity plots generated over 3 orders of magnitude using 6 plants/concentration. Assays were repeated three times each. totally abolished activity. The substitution of Thr with Ser in the tetrapeptide Met-Gln-Thr-Asp severely reduced activity, inviting speculation that Thr (or Ser) is phosphorylated as part of the signaling process and that any other substitution at this residue totally inactivates the signaling pathway, although no direct evidence for a phosphorylated form of systemin has yet been detected.
Assays of the proteinase inhibitor activity of systemin analogs Ala13 and Ala'7 are shown in Fig. 4, compared with activity of wild type systemin. Ala13 activity requires about 3 orders of magnitudes more polypeptide to achieve 50% maximal activity as the native systemin, whereas Ala17 is totally inactive, even at nanomole levels.
Since the possibility had been proposed that the Thr at position 17 might be the substrate for a kinase as part of the signaling process, it was considered that the two analogs of systemin that were totally inactive, i.e. the Ala17 and the deleted COOH-terminal aspartic acid derivative (cf . Table I), might still bind to a receptor and might compete with the activity of native systemin. Fig. 5 shows that both inactive analogs are potent inhibitors of the proteinase inhibitor inducing activity of wild type systemin. The Ala" analog (SYS-MQAD) inhibits wild type systemin much more potently than the analog with the deleted COOH-terminal Asp, which is itself a strong inhibitor. This indicates that although the modifications in the COOH-terminal sequence at positions 17 and 18 totally inhibit inducing activity, they retain the property to strongly inhibit the native systemin, presumably by interacting with the binding sites of the systemin receptor.

CONCLUSIONS
The evidence presented here indicates that the entire 18amino acid sequence of systemin is important for maximal proteinase inhibitor-inducing activity. Although many of the individual amino acids could be substituted with negligible or moderate changes in activity, substitutions of Ala at and ThrI7 caused major reductions in activity. The regions of the systemin primary structure that can be modified without seriously impairing activity are obvious candidates for modification to generate affinity probes to identify and isolate the putative systemin receptor. The total loss of biological activity of systemin analogs that are modified at the COOH-terminal Thr" or deleted at the adjacent COOH-terminal Asp position suggests these residues are crucial for activating the expression of the proteinase inhibitor genes, perhaps via phosphorylation of systemin. This possibility is somewhat supported by the evidence that the biologically inactive Thr17 and des-Asp" analogs are powerful inhibitors of the activity of native sys-temin, interacting with the signaling process initiated by systemin but not capable of activating the genes. The inactive analogs with potent inhibitory activities should be valuable tools for the continuing studies of the mode of action of systemin and for the search for the components of the signaling cascade with which systemin interacts in tomato and potato cells.