The Pto Kinase of Tomato, Which Regulates Plant Immunity, Is Repressed by Its Myristoylated N Terminus*

Specific recognition of the Pseudomonas syringae effector proteins AvrPto and AvrPtoB in tomato is mediated by Pto kinase resulting in induction of defense responses, including hypersensitive cell death via a signaling pathway requiring the nucleotidebinding leucine-rich repeats protein Prf. Pto is a myristoylated protein, and N-myristoylation is required for signaling. Here we demonstrated a role for N-myristoylation in controlling Pto kinase activity. A myristoylated peptide corresponding to Pto residues 2-10 significantly impaired the kinase activity of N-truncated Pto. We show that kinase inhibition was specific to the myristoylated form of the peptide and that free myristate supplied in trans was a potent suppressor of Pto kinase activity. Thus, myristate, but not Pto residues 2-10, contributes to suppression of kinase activity in vitro. Accordingly, elimination of the in vivo myristoylation potential of Pto de-repressed kinase activity. The increased potency of free myristate relative to the myristoylated N-peptide inhibitor suggested that the peptide moiety is antagonistic to repression by myristate. Suppression of related protein kinases by myristate declined with similarity to Pto, and the inhibitory activity could be attributed to hydrophobicity. We present evidence that inhibition of Pto by the myristoylated N-peptide is mediated through a previously identified surface regulatory patch. The data show a role for negative regulation of Pto by N-myristoylation, in addition to the previously demonstrated positive role, and are consistent with a model in which the acylated N terminus is sequestered in the catalytic cleft prior to release by Pto activation.

Specific recognition of the Pseudomonas syringae effector proteins AvrPto and AvrPtoB in tomato is mediated by Pto kinase resulting in induction of defense responses, including hypersensitive cell death via a signaling pathway requiring the nucleotidebinding leucine-rich repeats protein Prf. Pto is a myristoylated protein, and N-myristoylation is required for signaling. Here we demonstrated a role for N-myristoylation in controlling Pto kinase activity. A myristoylated peptide corresponding to Pto residues 2-10 significantly impaired the kinase activity of N-truncated Pto. We show that kinase inhibition was specific to the myristoylated form of the peptide and that free myristate supplied in trans was a potent suppressor of Pto kinase activity. Thus, myristate, but not Pto residues 2-10, contributes to suppression of kinase activity in vitro. Accordingly, elimination of the in vivo myristoylation potential of Pto de-repressed kinase activity. The increased potency of free myristate relative to the myristoylated N-peptide inhibitor suggested that the peptide moiety is antagonistic to repression by myristate. Suppression of related protein kinases by myristate declined with similarity to Pto, and the inhibitory activity could be attributed to hydrophobicity. We present evidence that inhibition of Pto by the myristoylated N-peptide is mediated through a previously identified surface regulatory patch. The data show a role for negative regulation of Pto by N-myristoylation, in addition to the previously demonstrated positive role, and are consistent with a model in which the acylated N terminus is sequestered in the catalytic cleft prior to release by Pto activation.
Race-specific disease resistance in plants involves precise recognition events between avirulence (Avr) 3 genes in patho-genic organisms and cognate resistance (R) genes in the host that induce a variety of host defenses, including cell death known as the hypersensitive response (1). The physical events underlying pathogen recognition are not well understood. In some cases, Avr products appear to interact directly with R proteins (2,3), whereas several examples of indirect association, in which the Avr product contacts a linker protein associated with the R component, have been described (4,5). The molecular events underlying activation of R proteins, which commonly possess a central nucleotide-binding motif known as the NB-ARC domain and C-terminal leucine-rich repeats (NBARC-LRR proteins (6)), are poorly understood, as are the immediate proximal events in signal transduction.
The tomato (Lycopersicon esculentum Mill.) R gene Pto confers race-specific resistance to Pseudomonas syringae pv. tomato bacteria carrying avrPto or avrPtoB (7). Pto, a Ser/Thr protein kinase, apparently interacts directly with both Avrs upon delivery into host cells through the type III secretion system, although physical association in planta has yet to be demonstrated. Pto-mediated resistance requires the NBARC-LRR gene Prf (8). We have found that Pto and Prf associate constitutively in planta to regulate Avr perception and basal signaling. 4 In addition, various potential downstream members of the Pto signaling pathway have been reported (7,11), such as the protein kinase Pti1, which is phosphorylated by Pto in vitro and provides an informative measure of Pto kinase activity (12).
Pto and its homologues Fen, Pth2, Pth3, and Pth5 encode consensus N-terminal myristoylation motifs (13). Protein N-myristoylation involves the typically co-translational, covalent, and irreversible attachment of the 14-carbon saturated fatty acid myristate at the N-terminal glycine residue (14). N-Myristoylation often contributes to membrane localization of target proteins, acting either constitutively or induced by myristate exposure through so-called myristoyl switches (15). We showed previously that Pto is myristoylated in vivo dependent on the Gly-2 residue, but we were unable to demonstrate that Pto myristoylation status determines subcellular localization (16). Nevertheless, Gly-2 is essential for signaling both by wild type Pto and constitutive gain-of-function (CGF) forms in Nicotiana benthamiana (16). In addition, many Avr and R proteins possess consensus N-myristoylation motifs (17), and several Avr proteins are myristoylated in vivo (18,19). N-Myristoylation may also enhance intra-or inter-molecular protein interactions (20) and influence protein structural and thermal stability (21). Of interest is the effect of acylation on enzyme activity. Abolishing myristoylation of the ␤1 subunit of AMPactivated protein kinase resulted in a significant increase in kinase activity (22). This is reminiscent of an autoinhibitory mechanism in the nonreceptor tyrosine kinase c-Abl, for which elimination of the myristoylation potential also de-represses kinase activity (23). In c-Abl, the myristate group is sequestered in a hydrophobic pocket within the kinase domain, inducing a bend within a C-terminal ␣-helix and promoting docking of domains distal to the c-Abl active site, thus locking the enzyme in a regulated, autoinhibited conformation (23,24). Upon activation, removal of the myristate group from the hydrophobic pocket might mimic the unmyristoylated form, thus enhancing enzymic activity.
We recently identified a surface-exposed, hydrophobic region on Pto that confers negative regulation of Pto signaling (the negative regulatory patch (NRP) (25)). This surface is required for Pto-Avr interaction and spans the Pto catalytic cleft and adjacent residues. Mutations within the NRP result in CGF forms of Pto that elicit Avr-independent cell death in N. benthamiana. We proposed a model in which the NRP is normally occupied by a negative regulatory molecule that acts to repress Pto kinase activity. Avr binding displaces the repressor, allowing regulatory phosphorylation and a conformational change in the Pto tertiary structure leading to activation (25). The identity of the putative negative regulator is unknown.
In this paper we describe a role of the myristoylated Pto N terminus in suppressing Pto kinase activity in vitro. We show that myristate-modified peptides are potent inhibitors of the auto-and trans-phosphorylation activities of Pto. Free myristate severely inhibited kinase capability in a dose-dependent manner. Removal of the myristoylation potential of Pto expressed in planta resulted in stimulation of kinase activity relative to the wild type protein. We show that the inhibitory effect of the myristoylated N-peptide is mediated by a key residue within the NRP. Our data suggest that the myristoylated N terminus may be associated with the catalytic cleft of Pto kinase, consistent with our previous prediction of a peptide or molecule suppressing kinase activity.

EXPERIMENTAL PROCEDURES
Cloning and Expression of Pto Mutants in N. benthamiana and Escherichia coli-All constructs were cloned into pCRII (Invitrogen) and sequenced to exclude PCR errors. For expression in planta, constructs were cloned in the pTFS40 binary vector (26). Recombinant binary vectors were transferred to Agrobacterium tumefaciens strain GV2260 for expression in planta, and the presence of the vector was confirmed by PCR.
Manipulation of A. tumefaciens strain GV2260 containing the required binary plasmid for transient in planta expression of Pto was as described (16). Six-week-old N. benthamiana plants grown under standard greenhouse conditions were inoculated by pressure infiltration using a disposable syringe.
For expression of Pto mutants, Fen and Pti1 in E. coli as glutathione S-transferase (GST) fusions, constructs were cloned in the pGEX-2TK vector (GE Healthcare) and transferred either to E. coli strains DH5␣ or BL21-CodonPlus (Stratagene, La Jolla, CA). Transformed bacterial cultures were grown in SOB media supplemented with appropriate antibiotics to A 600 ϳ0.4 at 37°C, and expression was induced with isopropyl ␤-D-1-thiogalactopyranoside at 16°C overnight. Alternatively, Pto was expressed with a C-terminal His 6 epitope tag with the pBAD expression system (Invitrogen). Expression of recombinant Pto-His 6 was induced with L-arabinose at 16°C overnight as for the GST fusion proteins.
Recombinant Protein Purification-GST fusion proteins were purified over glutathione-Sepharose columns (GE Healthcare) as described (16,25). GST-pti1 K96N to be used as the substrate in Pto kinase assays was expressed as described (26) and purified over GSTrap FF columns as for Pto. Where required, GST-pti1 K96N was treated with thrombin protease (GE Healthcare) to cleave the GST moiety (25).
Protein Immunoprecipitation Using Epitope Tags-FLAG epitope-tagged proteins transiently expressed in N. benthamiana leaves were immunoprecipitated from crude extracts with anti-FLAG M2 affinity gel (Sigma) as described (16). Immune complexes were collected by centrifugation, washed with 1ϫ Tris-buffered saline (TBS), resuspended in 1ϫ TBS, 50% (v/v) glycerol, and stored at Ϫ20°C for subsequent kinase assays.
Activated AtMAPK3 and AtMAPK6 were detected by Western blotting using ␣-DP-ERK (Sigma) as the primary antibody. This reacts specifically with the epitope containing the phosphorylated Thr and Tyr residues within the regulatory site of active MAPK. Immunodetection was performed with an ECL chemiluminescence reagent (GE Healthcare).
In Vitro Kinase Activity Assays-Pto in vitro kinase assays contained the following in a total volume of 50 l: 1 g of recombinant Pto or 10 l of immunoprecipitated Pto-FLAG (1:1 (v/v) suspension in 1ϫ TBS, 50% (v/v) glycerol), 25 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 0.5 mM MnCl 2 , 1 mM dithiothreitol, 20 M ATP, 183 kBq of [␥-32 P]ATP (PerkinElmer Life Sciences). For assays of Pto trans-phosphorylation activity, either full-length GST-pti1 K96N or cleaved pti1 K96N (1 g) was also included in kinase reactions as required. Reactions were initiated by addition of the substrate, incubated at 25°C for 7.5 min, and terminated by addition of SDS-PAGE loading buffer and boiling for 5 min. Under these assay conditions, incorporation of radiolabel was found to be linear with time and the enzyme concentration used. To assess the effect of inhibitors on Pto kinase activity, each inhibitor was included in the kinase assays at the concentrations indicated under "Results." At the end of each assay, samples were loaded on SDS-PAGE. Postelectrophoresis, proteins were transferred onto polyvinylidene difluoride membranes, which were either stained with Coomassie Brilliant Blue R-250 to confirm equivalent loading or used for Western blots with specific antisera against Pto (25). Subsequently, they were subjected to autoradiography with a FUJI Film FLA5000 PhosphorImager (Fuji, Tokyo, Japan). Specific activity was estimated as the ratio between radioactive incorporation (PhosphorImager signal) and amount of protein added, except for Western blots where protein density was estimated with the QuantityOne software (Bio-Rad) after specific chemiluminescent detection of each protein. Results were expressed as % of the uninhibited enzyme. In vitro kinase activity of GST-Fen and GST-Pti1 kinases was assayed as described above for Pto, but omitting pti1 K96N .

The in Vitro Kinase Activity of Pto Is Inhibited in Trans by the
Myristoylated N-peptide-We previously identified the Pto NRP, a hydrophobic surface on Pto overlapping the kinase catalytic cleft that negatively regulates signaling (25). We hypothesized that this patch is normally occupied by an unknown molecule acting to repress kinase activity. This could be part of Pto itself or derived from another molecule. Removal of the negative regulator was predicted to be sufficient to activate downstream signaling. We found that N-myristoylation was important for the activity of CGF forms of Pto, downstream of kinase activity (16). Given its important functions and location outside of the kinase domain, we investigated whether the myristoylated N terminus of Pto is a candidate for the proposed regulatory molecule. To test this, we explored the ability of a peptide comprising amino acids 2-10 to repress Pto kinase activity in vitro.
We expressed pto[N-10] in E. coli as a C-terminal fusion with the GST protein. This protein lacks 10 N-terminal amino acids and was shown previously to be an active kinase in vitro (16). To test the ability of residues 2-10 (not including Met-1, which is cleaved after translation) to repress Pto kinase activity, we synthesized the unmyristoylated (Pto 2-10 ) and myristoylated (myr-Pto 2-10 , modified on Gly-2) forms of this peptide. Both were included in in vitro kinase activity assays of GSTpto[N-10] with pti1 K96N as the substrate. This assay allows measurement of both the auto-and trans-phosphorylation activities of Pto. Increasing the concentration of the myr-Pto 2-10 peptide effected significant inhibition of the transphosphorylation activity of GST-pto[N-10] against pti1 K96N (Fig. 1a). Pto trans-phosphorylation activity was reduced by more than 90% at myr-Pto 2-10 concentrations higher than 100 M. The half-maximal inhibitory concentration (IC 50 ) was estimated to be ϳ50 M. Conversely, neither Pto 2-10 nor the unrelated hemagglutinin peptide (HA) had a significant effect on Pto kinase activity. Pto autophosphorylation activity was also suppressed by myr-Pto 2-10 , albeit to a lesser extent than trans-phosphorylation (Fig. 1b). We also tested whether Pto expressed in planta was inhibited by myr-Pto 2-10 . pto [N-10] fused to a sequence encoding a C-terminal FLAG epitope was transiently expressed in N. benthamiana leaves under control of the cauliflower mosaic virus 35S promoter. The protein was immunoprecipitated from crude extracts with anti-FLAG M2-agarose beads and used for in vitro kinase assays in the presence of myr-Pto 2-10 and Pto 2-10 . As for GST-pto[N-10], severe inhibition of kinase activity was effected at increasing concentration of myr-Pto 2-10 but not Pto 2-10 (Fig. 1c).
We next asked whether the inhibition caused by myr-Pto 2-10 could be observed in the presence of a myristoylated peptide of random sequence. To test this, we used two peptides unrelated to Pto 2-10 , myr-PKC and PKC , differing only for myristoylation status. These peptides function as pseudosubstrate inhibitors of protein kinase C, isoform (PKC ) activity (29). A progressive reduction of both the trans- (Fig. 1d) and auto-phosphorylation (Fig. 1e) activities of GST-pto[N-10] was observed in the presence of increasing concentrations of the myristoylated peptide (myr-PKC ).
The rate and the extent of inhibition caused by myr-PKC were comparable with that effected by myr-Pto 2-10 . The estimated IC 50 for the inhibition of GST-pto[N-10] trans-phosphorylation activity by myr-PKC was ϳ55 M, similar to myr-Pto 2-10 (ϳ50 M). Comparable with Pto 2-10 , the nonmyristoylated PKC peptide did not suppress Pto kinase activity. Therefore, we conclude that Pto residues 2-10 do not contribute to suppression of Pto kinase activity in vitro.
Free Myristate Inhibits Pto Kinase Activity in Vitro-The results above suggest that inhibition of Pto kinase activity by myr-Pto 2-10 was because of the myristate moiety on Gly-2. To test this, we added myristate directly to kinase assays in the absence of conjugated peptide. These experiments used myristate-coenzyme A (myr-CoA) as the inhibitor to overcome the insolubility of myristate in aqueous solution. Severe inhibition of both the trans-and auto-phosphorylation activities of recombinant Pto was observed in the presence of myr-CoA (Fig. 2). Conversely, kinase activity was not significantly inhibited by CoA alone; therefore, inhibition was because of the myristate moiety. Less than 10% of kinase activity was retained at myr-CoA levels higher than 20 M. The corresponding IC 50 for inhibition of trans-phosphorylation activity by myr-CoA was ϳ15 M, substantially lower than the IC 50 of myr-Pto 2-10 (ϳ50 M). Similar results were obtained with Pto-FLAG purified from N. benthamiana leaf extracts (data not shown).
We reasoned that the inhibition of kinase activity by myristate might be attributable to its hydrophobic nature. In this case, other fatty acids might cause similar repression of Pto kinase activity. Therefore, we substituted myr-CoA in the kinase assays with CoA derivatives of both saturated (lauric acid, C12:0; palmitic acid, C16:0) and unsaturated (oleic acid, C18:1; linoleic acids, C18:2) fatty acids. Severe inhibition of Pto kinase activity was effected by all fatty acids tested (Fig. 3). The extent of suppression in all treatments was similar to that caused by myristate. These results indicate that the hydrophobic nature of the inhibitor is important for suppression of kinase activity in vitro.
Inhibition of Kinase Activity by N-Myristoylation Declines with Evolutionary Distance from Pto-The protein kinase family is monophyletic, and all crystallized eukaryotic protein kinases possess a common fold (30). Therefore, we were interested to test the generality of kinase activity repression by myr-Pto 2-10 or myristate. If the observed inhibition is specific to Pto, then less closely related protein kinases should not be affected to the same extent by either inhibitor. We first investigated the in vitro activity of two mitogen-activated protein kinases (MAPK) from A. thaliana, AtMAPK3 and AtMAPK6. These enzymes are stimulated by a wide array of abiotic stresses and biotic elicitors, including the peptide flg22 derived from bacterial flagellin (27,31,32), and share little identity to Pto (Table 1). AtMAPK3 and AtMAPK6 were immunoprecipitated from flg22-treated A. thaliana cell suspension cultures and the immune complexes used directly in in vitro kinase assays with Phos32 protein as the substrate. 5 Inclusion of either myr-Pto 2-10 (not shown) or myr-CoA in the kinase assays did not have a pronounced effect on the activity of AtMAPK3. At myr-CoA levels that caused more than 90% inhibition of Pto kinase, AtMAPK3 enzymic activity was repressed only by 20 -25% (Fig. 4). Results similar to the AtMAPK3 data were also obtained with AtMAPK6 (data not shown).
We next investigated the effect of myr-CoA on protein kinases more closely related to Pto. Fen is a Pto family member with ϳ80% identity to Pto, is capable of Prf-dependent signaling (13), and its N terminus is an efficient substrate for A. thaliana myristoyl-CoA:protein N-myristoyltransferase (33). Pti1 interacts with Pto in yeast two-hybrid assays and may participate in defense-related signaling in planta (7). Pti1 shares ϳ39% amino acid identity with Pto within the kinase domain (12). Fen and Pti1 were expressed in E. coli as C-terminal GST fusions, and their autophosphorylation potentials were assessed in vitro in the presence of myr-CoA. Similar to Pto, increasing the concentration of myr-CoA effected significant suppression of both GST-Fen and GST-Pti1 autophosphorylation capabilities (Fig.  4a). GST-Fen and GST-Pti1 activities were suppressed by ϳ40% at myr-CoA levels corresponding to the IC 50 of Pto inhibition. The estimated IC 50 values were ϳ25 M (GST-Fen) and 22 M (GST-Pti1) ( Table 1). We also examined the effect of myr-CoA on the autophosphorylation activity of the kinase domain of SYMRK of Lotus japonicus, which shares 37% identity with Pto (28). Recombinant SYMRK kinase domain (fused to an N-terminal His 6 epitope tag; His 6 -SYMRK) activity against MBP was moderately suppressed by myr-CoA. Approximately 25% of activity was retained at myristate levels that resulted in almost complete inhibition of Pto activity (Fig. 4). Taken together, the data indicate that inhibition of kinase activity by myristate declines with evolutionary distance from Pto.
In Vivo Myristoylation on Gly-2 Reduces Pto Kinase Activity-Because exogenous myristate was found to strongly inhibit Pto kinase activity, it follows that removal of the fatty acid from the native protein should de-repress enzymic activity. Pto is myristoylated on the glycine residue at position 2, and mutation of Gly-2 to Ala abolishes myristoylation in vivo (16). Hence, we expressed Pto-FLAG and pto G2A -FLAG transiently in N. benthamiana leaves under control of the cauliflower mosaic virus 35S promoter. Recombinant proteins were purified from crude extracts by immunoprecipitation with anti-FLAG M2-agarose beads, and both auto-and trans-phosphorylation capabilities were assessed as before. Elimination of in vivo myristoylation in the pto G2A mutant resulted in de-repression of kinase activity in vitro. The nonmyristoylated protein exhibited ϳ30 and 40% higher auto-and trans-phosphorylation activity, respectively, relative to the myristoylated wild type Pto (Fig. 5a).
We then determined the kinetic properties of the Pto and pto G2A proteins purified from plant extracts. Substrate saturation curves were prepared for both Pto-FLAG and pto G2A -FLAG (Fig. 5b). Pto activity increased with increasing concentration of GST-pti1 K96N , and pto G2A -FLAG activity was consistently higher than that of Pto-FLAG. The initial data were transformed to double-reciprocal plots for estimation of   apparent kinetic parameters (Fig. 5c). The K m value estimated for pto G2A -FLAG was ϳ1.9 M and is of the same order of magnitude as that estimated by Sessa et al. (34) for the similarly nonmyristoylated GST-Pto (4.1 M). However, it is significantly lower than the K m estimate for the myristoylated wild type Pto-FLAG (ϳ13 M). Together, the data suggest that elimination of the myristoylation potential of Pto results in higher kinase activity and affinity for the peptide substrate in vitro. This is consistent with occupation of the catalytic cleft by the myristoylated N terminus of Pto.
The Myristoylated N Terminus of Pto Suppresses Kinase Activity through the Negative Regulatory Patch-In previous work we hypothesized that Pto signaling is suppressed by a molecule or peptide that occupies the NRP that overlaps the catalytic cleft of Pto (25). Mutations within the NRP confer CGF activity, and most of them abolish kinase activity. However, we identified one such CGF mutant, pto I214D , which retained kinase activity. To test the possibility that the myristoylated N terminus of Pto suppresses kinase activity through interaction with the NRP, we expressed GST-pto[N-10] I214D in E. coli as described for GST-pto[N-10]. Purified GSTpto[N-10] I214D exhibited both auto-and trans-phosphoryla-tion activities in vitro as reported previously for the full-length GST-Pto fusions (25) (Fig. 6a). Inclusion of myr-Pto 2-10 but not Pto 2-10 in the kinase assays resulted in only moderate inhibition of GST-pto[N-10] I214D activity (Fig. 6b). Thus, trans-phosphorylation activity of GST-pto[N-10] I214D was suppressed to a lesser extent than GST-pto [N-10]. At the maximum inhibitor concentration, almost 40% of kinase activity was retained by the GST-pto[N-10] I214D protein. This is significantly higher than the activity of GST-pto[N-10] which, at the same inhibitor concentration, exhibited less than 10% activity. These differences were also reflected in both the rate of inhibition and the estimated IC 50 values (GST-pto[N-10] ϳ50 M and GST-pto[N-10] I214D ϳ100 M). No significant differences between GST-pto[N-10] and GST-pto[N-10] I214D were evident in either the rate or the extent of inhibition of autophosphorylation activity by myr-Pto 2-10 (cf. Fig. 1b and data not  shown). Thus, Ile-214 is required for inhibition of Pto transphosphorylation activity by myr-Pto 2-10 .
Data above show that the repressive effect of myr-Pto (2-10 is because of the myristate moiety. To investigate the effect of myristate on pto I214D activity, we included myr-CoA in in vitro kinase assays of full-length Pto and pto I214D proteins. These were expressed in E. coli fused to C-terminal His 6 tags (Pto-His 6 and pto I214D -His 6 , respectively) and purified by IMAC. Recombinant pto I214D -His 6 was an active kinase comparable with Pto-His 6 ( Fig. 6c) (25). We assessed the in vitro kinase capabilities of both these proteins, together with GST-pto[N-10] and GST-pto[N-10] I214D . myr-CoA severely inhibited the kinase activities of pto I214D -His 6 and GSTpto[N-10] I214D . The extent of inhibition of the I214D forms was similar to that of Pto-His 6 and GSTpto[N-10], respectively (Fig. 6c). We estimated the IC 50 values for inhibition of trans-phosphorylation activ-   Fig. 2b). We interpret the differential potency of myr-Pto 2-10 and myr-CoA on pto I214D inhibition as evidence of an antagonistic interaction between the mutated residue and the N-peptide sequence in kinase repression by myristate. This is consistent with previous observations suggesting that N-terminal Pto residues have both positive and negative influence on Pto kinase activity (16). Overall, our data imply a direct link between the myristoylated N terminus and the catalytic cleft of Pto, and identify Ile-214 as a key NRP residue for myr-Pto 2-10 -mediated inhibition of Pto kinase activity.

DISCUSSION
Protein kinases are potent signaling molecules whose activity is under acute control to prevent aberrant signaling. Although the structures of protein kinases in the active conformation are highly similar, inhibitory mechanisms tend to be specific to individual kinases (30). The tomato Pto kinase expressed either in bacteria or in plants possesses intrinsic enzymic activity in vitro so is not misfolded; yet its signaling in vivo requires regulated kinase capability. Therefore, Pto enzymic activity must be repressed in vivo by a mechanism that has not previously been detectable in in vitro experiments. Moreover, signaling by Pto in the active conformation is independent of kinase activity (25). Thus, activated Pto must be distinguished from the inactive form by a change in conformation, rather than by enhanced kinase activity. We hypothesized that the Pto NRP is normallyoccupiedbyaregulatorymoleculethatrepresseskinasedependent activation and is released by ligand-dependent kinase activity (25). Here we show that the myristoylated Pto N terminus is a candidate for this molecule. N-Myristoylation is a co-translational modification specific to eukaryotes (15), so recombinant Pto protein produced in E. coli is not acylated. We show that myristate is a repressor of Pto kinase activity, either supplied in trans as free myristate or as an acylated peptide or in cis as part of the native molecule expressed in planta. Furthermore, the importance of N-myristoylation for signaling of CGF forms of Pto (16), downstream of the requirement for kinase activity, demonstrates that acylation also plays a positive regulatory role in Pto signaling. Thus, the myristoylated N terminus of Pto fits the previously proposed model, in which expulsion of a negative regulatory molecule from the NRP is important for signaling.
Several lines of evidence demonstrate that myristate is a suppressor of Pto kinase activity. We showed that the myr-Pto 2-10 peptide is a potent inhibitor of Pto trans-phosphorylation. Surprisingly, the N-peptide itself was incapable of kinase inhibition. Thus, Pto residues 2-10 do not contribute to suppression of kinase activity in vitro. This is different from pseudosubstrate inhibition observed in a number of protein kinases, such as the protein kinase C (PKC) (29) and the myosin light chain kinase (35). Our data further indicate that there is no sequence requirement for the inhibitory activity of the myristoylated peptide, because both myr-Pto 2-10 and myr-PKC suppressed Pto kinase activity to the same extent. In contrast, an acylated PKC pseudosubstrate peptide, but not a similarly acylated unrelated peptide, was an efficient and selective inhibitor of PKC activity (9).
The inhibitory activity of myr-Pto 2-10 can be ascribed to the fatty acid moiety, because free myristate delivered as myr-CoA severely suppressed Pto kinase capability. Importantly, acylated Pto expressed in planta was less active than the myristoylation-deficient G2A mutant. Deletion of Pto acylation capability was associated with enhanced kinase activity and a significant increase in substrate specificity. Similarly, myristoylation-deficient forms of the nonreceptor tyrosine kinase c-Abl and AMP-activated protein kinase were more active compared with the acylated wild type forms (22,23). We suggest that the repressive effect of myris- tate is a consequence of increased hydrophobicity, because a range of both saturated and unsaturated fatty acids resulted in comparable suppression. This further indicates that there is no specific fatty acid chain length requirement. Ramdas et al. (10) investigated the effect of fatty acid chain length of an acylated peptide inhibitor of the Src protein tyrosine kinase. Contrary to Pto, increasing the carbon chain length both of free fatty acids and of the acylated peptide resulted in increased inhibitor potency.
Interestingly, the Pto N-peptide corresponding to amino acids 2-10 appeared to be antagonistic to inhibition of kinase activity by myristate. The IC 50 of myr-CoA for Pto trans-phosphorylation activity was five times lower than that of myr-Pto 2-10 . This difference is potentially due to the hydrophilic nature of the amino acid sequence, because a similar IC 50 was seen for myr-PKC , which possesses an unrelated amino acid sequence to Pto 2-10 . Furthermore, free myristate was a potent inhibitor of both N-truncated and full-length Pto (Fig.  6). de Vries et al. (16) observed that deletion of N-terminal residues 1-10 decreased in vitro kinase activity relative to fulllength Pto. Thus, the hydrophilic peptide may decrease or interfere with the binding of myristate to the kinase domain. This is consistent with a model for activation in which the N-peptide contributes to release of the myristate moiety from a hydrophobic cleft. Myristate was not sufficient for full repression of Pto autophosphorylation activity, suggesting that another molecule may be involved. Recently we have found that Pto interacts constitutively with the NBARC-LRR protein Prf. 4 It is likely that Prf also makes an important contribution to regulation of Pto kinase activity in vivo.
Repression of kinase activity was not specific to Pto but declined with evolutionary distance. For example, subtle differences in the degree of myristate-mediated inhibition were observed between Pto and its close homologue Fen. Despite the high degree of identity and overall similarity between Pto and Fen, some Pto NRP residues are polymorphic in Fen (supplemental Fig. 1). Conversely, AtMAPK3 is distantly related to Pto and was not significantly inhibited by myr-CoA. This is consistent with a model in which myristate interacts with a conserved feature of the kinase domain, possibly within the catalytic cleft itself, but the degree of repression is modified by individual side chain-myristate interactions.
Several lines of evidence suggest that the myristoylated N terminus lies within or close to the catalytic cleft of Pto. We showed previously that pto[N-10] was a less active kinase than wild type Pto, implying a role for residues 2-10 in kinase activity (16). Similarly, N-acylation decreased substrate affinity and enzyme activity (Fig. 5). Residues 11-20 of Pto were required for optimal interaction with AvrPto (but not AvrPtoB) in yeast (16). This is striking because the main site of AvrPto interaction overlaps the catalytic cleft (25). The NRP residue Ile-214 provides an important point of intersection between regulatory mechanisms of Pto in vivo and in vitro. Acidic substitution of this residue results in the CGF phenotype in vivo, interpreted previously as loss of negative regulation due to interruption of an inhibitory intermolecular interaction. Conversely, the same mutation reduces the inhibitory effect of myr-Pto 2-10 . Synthesis of these observations allows a model in which Ile-214 is a point of attachment with the inhibitory N terminus. In the absence of the Avr proteins, we suggest that the myristoylated Pto N terminus occupies the NRP and represses kinase activity in concert with the Prf protein (Fig. 7). Myristate attachment at Gly-2 allows docking of the N terminus onto the hydrophobic surface of the patch. The Avr proteins are delivered into the host cell, where they interact with the Pto-Prf complex, 4 targeting the same region on Pto that is normally occupied by the myristoylated N terminus. The negative regulator is displaced from the NRP, allowing de-repression of kinase activity, regulatory phosphorylation, and subsequent Pto activation.