Characterization of Cucurbita maxima Phloem Serpin-1 (CmPS-1)

We report on the molecular, biochemical, and functional characterization of Cucurbita maxima phloem serpin-1 (CmPS-1), a novel 42-kDa serine proteinase inhibitor that is developmentally regulated and has anti-elastase properties. CmPS-1 was purified to near homogeneity from C. maxima (pumpkin) phloem exudate and, based on microsequence analysis, the cDNA encoding CmPS-1 was cloned. The association rate constant (k a ) of phloem-purified and recombinant His6-tagged CmPS-1 for elastase was 3.5 ± 1.6 × 105 and 2.7 ± 0.4 × 105 m − 1 s− 1, respectively. The fraction of complex-forming CmPS-1, Xinh, was estimated at 79%. CmPS-1 displayed no detectable inhibitory properties against chymotrypsin, trypsin, or thrombin. The elastase cleavage sites within the reactive center loop of CmPS-1 were determined to be Val347-Gly348 and Val350-Ser351 with a 3:2 molar ratio. In vivo feeding assays conducted with the piercing-sucking aphid,Myzus persicae, established a close correlation between the developmentally regulated increase in CmPS-1 within the phloem sap and the reduced ability of these insects to survive and reproduce onC. maxima. However, in vitro feeding experiments, using purified phloem CmPS-1, failed to demonstrate a direct effect on aphid survival. Likely roles of this novel phloem serpin in defense against insects/pathogens are discussed.

The phloem long-distance translocation system of plants appears to function both as a nutrient delivery system and as an information superhighway (1)(2)(3). A central role for the phloem in the translocation of nutrients has long been recognized. The presence of plant hormones in the phloem sap (4,5) implicated this long-distance transport pathway in the delivery of signaling molecules. Recent studies provided new insights into the nature of the information molecules being transported from mature leaves, via the phloem, to distant plant organs. Irrefutable evidence has been obtained for the translocation of certain proteins (6 -9). In addition, it has also been demonstrated that specific RNA molecules are present in the phloem sap (10,11) and some move to distant tissues, where they appear to influence post-transcriptional events (9,12).
Given the importance of this nutrient/information delivery system to the functioning of the plant, it was axiomatic that plants had to evolve mechanisms to protect the operational integrity of the phloem. Maintenance of structural integrity required the development of systems able to rapidly respond to physical damage, imposed either by environmental forces or herbivory; the sealing of disrupted sieve tubes involves deposition of material at the level of the sieve plate pore (1). In the enucleate sieve tube system of angiosperms, maintenance of membrane integrity has also been transferred to the neighboring companion cells and likely involves the delivery of essential constituents via plasmodesmata (6,13,14).
To ensure the integrity of the signaling components, the plant needs also to protect against protein and RNA degradation occurring within the phloem sap. This capacity appears to have been achieved through the development of a control system that regulates the plasmodesmal-mediated exchange of macromolecules between companion cells and the sieve tube system (9,12,13). The absence of proteinase activity (15) within the phloem sap is consistent with this model. A range of small molecular weight proteinase inhibitors (PIs 1 ; 3-10 kDa) has also been isolated from the phloem sap of several species (16 -18). The smallest of these (3 kDa) are members of the serine PI family, whereas the 7-10 kDa PIs belong to the potato PI 1 family. Although nothing is known concerning the mechanism(s) by which these PIs enter the sieve tube system and their role remains conjectural, they likely serve to protect the proteins within the phloem sap against the action of endogenous proteinases.
An important challenge to the integrity of the phloem system is also posed by piercing-sucking insects, whose primary nutrition is gained through uncontrolled access to the phloem sap. In this regard, it is of interest to note that numerous plant PIs have been shown to modify plant-arthropod interactions, via their role as digestibility reducers, toxins, or modifiers of feeding behavior (19,20). In addition, in vitro assays have estab-lished that phloem PIs (3-10 kDa) are able to inhibit a wide range of proteinases (18,21,22), including a number of such enzymes extracted from the midgut of lepidopteran (chewing) larvae (23,24).
In the present study, we report on the molecular, biochemical, and functional characterization of Cucurbita maxima phloem serpin-1 (CmPS-1), a 42-kDa serine PI present in the C. maxima (pumpkin) phloem translocation stream. Biochemical studies revealed that CmPS-1 represents a novel plant serpin having anti-elastase properties. A close correlation was established between the developmentally regulated increase in CmPS-1 within the phloem sap and the reduced ability of the piercing-sucking aphids, Myzus persicae to survive and reproduce on C. maxima. However, in vitro feeding experiments using highly purified phloem-derived CmPS-1 failed to demonstrate a direct effect on aphid survival.

EXPERIMENTAL PROCEDURES
Plant Material-C. maxima Duch. cv. Big Max (pumpkin) plants were grown in an insect-free greenhouse under natural daylight conditions (14-h photoperiod). Light intensities at mid-day ranged from 1200 to 1500 mol m Ϫ2 s Ϫ1 and day/night temperatures were 26 Ϯ 3/22 Ϯ 2°C, respectively.
Protein Microsequencing, PCR, and cDNA Cloning of CmPS-Phloem exudate (sap) from cut stems of 4-week-old pumpkin plants was collected, dialyzed, and stored as described previously (13). Phloem sap proteins were fractionated by SDS-PAGE, and those at the 40-kDa range were excised from gels, electroeluted, and concentrated by partial lyophilization. This protein concentrate was digested overnight with lysyl endopeptidase (EC 3.4.21.50; Wako, Richmond, CA) and the resultant products separated by high pressure liquid chromatography on a 25-cm-long C18 column. Peptides were eluted, and a major peak was sequenced (model 477A, Applied Biosystems, Foster City, CA).
For cloning of CmPS-1, stem poly(A) ϩ RNA isolated from 4-week-old pumpkin plants was used to synthesize first-strand cDNA by reverse transcription (FastTrack 2.0 kit; Invitrogen, Carlsbad, CA); this cDNA was then used as template for PCR. The following degenerate primers were employed: forward, 5Ј-TICCITAYWSICARGGICCNGA-3Ј (I ϭ deoxyinosine, n ϭ A ϩ C ϩ T ϩ G, Y ϭ C ϩ T, S ϭ C ϩ G, W ϭ A ϩ T); and reverse, 5Ј-TCIGTICCYTCYTCRTTNACYTC-3Ј. The resultant 342-bp DNA fragment was labeled with 32 P nucleotides by random priming (high prime DNA labeling kit, Roche Molecular Biochemicals) and then used to screen a pumpkin stem cDNA library. This library was constructed in Uni-ZAP XR (Stratagene, La Jolla, CA) using the above described poly(A) ϩ RNA to synthesize double-stranded cDNA (ZAP synthesis kit, Stratagene). Plaques (10 5 ) were screened, and five positive clones were excised into phagemids. Each putative CmPS-1 clone was then sequenced in both directions. Editing, analysis, and sequence alignments were performed using the SeqEd1.03, SeqVu1.0.1, DNASIS-Mac v2.0, DNA Strider1.2 programs and GAP, FASTA, and BLAST search engines from the Genetics Computer Group (Wisconsin sequence analysis package).
Purification of CmPS-1 from C. maxima Phloem Exudates-All procedures were carried out at 4°C, and chromatography was performed using fast protein liquid chromatography. Purification was followed by Western analysis with polyclonal antibodies (R306, raised against serpin Z-type proteins from barley seed (25) and kindly provided by Dr. Jøern Hejgaard, Dept. of Biochemistry and Nutrition, Technical University of Denmark, Lyngby). Phloem exudate collected as described above was dialyzed overnight against buffer A (50 mM Tris, pH 7.5, 1 mM EDTA, and 50 mM 2-mercaptoethanol). After clarification by centrifugation (17,000 ϫ g for 30 min), the exudate was applied to HiTrap Q-Sepharose (Amersham Pharmacia Biotech) equilibrated with buffer A. CmPS-1 did not bind to the column and thus was collected from the flow-through fractions. These fractions were dialyzed against buffer B (25 mM HEPES, pH 7.0, 1 mM EDTA, and 14 mM 2-mercaptoethanol), clarified (as described above), and loaded onto HiTrap SP-Sepharose equilibrated with buffer B. CmPS-1 was present in the flow-through fractions. A second cation exchange chromatography fractionation was performed in the presence of a lower pH buffer to fractionate CmPS-1. The flow-through fractions were dialyzed against buffer C (30 mM MES, pH 5.5, 1 mM EDTA) and subjected to HiTrap SP equilibrated with buffer C. CmPS-1 was eluted at 120 mM NaCl using a linear gradient. The fractions containing CmPS-1 were pooled, supplemented to 1.7 M ammonium sulfate, and subsequently loaded onto a HiTrap phenyl-Sepharose column. CmPS-1 was eluted using a reverse gradient. Fractions of the highest purity, as judged by SDS-PAGE, were pooled and dialyzed against 25 mM HEPES, pH 7.5, and 100 mM NaCl. Dialyzed proteins were concentrated by ultrafiltration using Centricon (Millipore, Bedford, MA) and stored at 4°C. Protein concentration was measured by the Bio-Rad protein assay using bovine serum albumin as standard.
Expression and Purification of Recombinant CmPS-1-The expression vector used to produce recombinant, His 6 -tagged CmPS-1 was derived from pET-15b, and the Escherichia coli host was BL21(DE3)pLysS (26). To construct the expression vector, pET-15b/ CmPS-1, the CmPS-1 open reading frame was amplified by PCR using 5Ј primer, 5Ј-GCGGATCCAATGGA-CATCAAAGAAGCAATCAG-3Ј and 3Ј primer, 5Ј-GCGGATCCTCAATCCACAAGAGGG-TTTAACACC-TG3Ј. BamHI sites (underlined) were included to facilitate the cloning procedures. A PCR-amplified fragment was digested with BamHI and ligated into pET-15b, previously digested with BamHI, and dephosphorylated. After transformation, the orientation of the insert was verified by restriction enzyme analysis and the integrity of the cloning by sequencing. To express His 6 -tagged CmPS-1, transformed BL21(DE3)-pLysS harboring pET-15b/CmPS-1 was grown overnight at 37°C in Luria broth from a single colony, which was subsequently used as the primary culture. The secondary culture was initiated with a 1/500 dilution of primary culture in 750 ml of M9TB media (26) at 37°C and continued until it reached 0.8 A 600 . The culture was then cooled in a 16°C water bath and expression induced, overnight, with 1 mM isopropyl-1-thio-␤-D-galactopyranoside. Cells were then harvested and resuspended in 50 ml of B-PER solution (Pierce) supplemented with 0.5 M NaCl and 20 g/ml DNase I. Cell lysis (15 min at room temperature) was followed by centrifugation (27,000 ϫ g for 30 min), and the supernatant was batch-incubated with Ni-agarose (His⅐Bind resin, Novagen, Madison, WI) charged previously with 50 mM NiSO 4 and equilibrated with 50 mM Tris, pH 7.8, 0.5 M NaCl, and 5 mM imidazole. The resin was washed extensively with equilibration buffer followed by a second wash with 30 mM imidazole in equilibration buffer. Bound proteins were eluted with 400 mM imidazole and dialyzed against 50 mM Tris, pH 7.5, and 1 mM EDTA. Proteins were further purified by loading onto HiTrap Q-Sepharose. His 6 -tagged CmPS-1 was present in the flow-through fractions, which were pooled, dialyzed against 50 mM Tris, pH 7.5, and 100 mM NaCl, and stored at 4°C.
Inhibition Assay-To measure proteinase inhibitory activity, 10 l (0.5 pmol) of stock solution of each serine proteinase substrate was incubated at 24°C for 0.25-15 min with 10 l of an equimolar amount of CmPS-1 in 50 mM Tris, pH 8.0, 100 mM NaCl, and 0.01% Tween 20. After incubation, 80 l of 0.5 mM chromogenic substrate solution was added, and residual proteinase activities were monitored by detecting the time-dependent change in absorbance at 405 nm. Proteinases and substrates (Sigma) used in these assays were as follows: porcine pancreatic elastase and N-succinyl-Ala-Ala-Pro-Leu-p-nitroanilide; porcine pancreatic trypsin and N-p-tosyl-Gly-Pro-Arg-p-nitroanilide; bovine pancreatic ␣-chymotrypsin and N-succinyl-Ala-Ala-Pro-Phe-p- nitroanilide; human plasma thrombin and N-p-tosyl-Gly-Pro-Arg-pnitroanilide.
Association Rate Constants-Association rate constants were determined according to the scheme described by Dahl et al. (25). Active site titration was carried out as described by Jiang and Kanost (27). Trypsin was titrated first by using p-nitrophenyl-pЈ-guanidinobenzoate (Sigma) (28), and this was then used to titrate ␣-1-antitrypsin (Calbiochem). This ␣-1-antitrypsin was subsequently used as a secondary standard to titrate elastase and chymotrypsin. Titrated elastase was then used to titrate CmPS-1 and His 6 -tagged CmPS-1.
Complex Formation and Cleavage Site Determination-Serpin-proteinase complexes were formed by incubating 50 pmol of CmPS-1 with 50 pmol of proteinases at 24°C for 5 min in 20 mM Tris, pH 8.4, and 100 mM NaCl. The reactions were stopped by the addition of hot SDS-PAGE sample buffer, and the mixture was then subjected to Tricine-SDS-PAGE using a 10 -20% gradient gel (Novex, San Diego, CA). After blotting onto polyvinylidene difluoride membrane, the amino-terminal residues of the cleaved 4-kDa peptide were determined by sequencing analysis. Alternatively, the 4-kDa peptide fragment was first extracted from the gel using a combination of 50% acetonitrile plus 5% formic acid followed by 70% isopropanol plus 5% formic acid. The molecular weight of the pooled extracts was then determined by matrix-assisted laser desorption ionization/time of flight (MALDI-TOF) mass spectrometry on a Biflex III (Bruker, Billerica, MA). Theoretical values for the molecular weight of the peptide were calculated using the PeptideMass program.
In Vivo and in Vitro Aphid Feeding Assays-The aphid species M. persicae was used in feeding studies. As a cucurbit-adapted control, Aphis gossypii was raised on Cucurbita pepo L. (squash) and employed in parallel feeding experiments. Aphid isolates were collected locally and maintained in cages within a greenhouse (16 h light:8 h dark at 27°C); M. persicae was reared on Rhaphanus sativus L. (radish). The in vivo influence of CmPS-1 and plant age on aphid survival were investigated by placing M. persicae and A. gossypii on newly expanded leaves of 10-, 14-, 21-, 27-, 35-, and 42-day-old pumpkin plants. Aphids were confined to the feeding surface (second-through fifth-order veins) using clip cages mounted on the abaxial leaf surface, with 50 individuals in each cage and two cages per plant. After 3 days, aphid survival and fecundity (nymphs produced per adult) were recorded, and then phloem sap was collected from the petiole of the treated leaf. Each experiment was performed on five plants, and experiments were repeated in triplicate. The level of CmPS-1 was analyzed by SDS-PAGE and Western analysis.
For in vitro aphid feeding experiments, 10 neonate M. persicae (aphids less than 12 h old) were placed in a sterile 1.5-ml Eppendorf tube, which was then covered with a thin layer of Parafilm®. A 20-l drop of feeding solution was then placed onto the Parafilm® surface and immediately covered by an additional layer of Parafilm®, forming a feeding sachet. The feeding solution was composed of 20% sucrose supplemented with each protein being tested. Aphid survival was re- corded at 24-h intervals for 3 days. The enzymatic activity of each treated protein, maintained under feeding conditions within control sachets, was tested to confirm protein stability under these experimental conditions. Each feeding experiment was repeated at least 10 times.

Cloning of Developmentally Regulated 40-kDa Phloem
Protein-SDS-PAGE analysis performed on phloem exudates collected from different aged C. maxima plants revealed the presence of a developmentally regulated 40-kDa protein; the level of other phloem proteins appeared more or less invariant with plant age (Fig. 1A). For further study, this protein was subjected to internal microsequencing after lysyl endopeptidase digestion. A protein data base search, using the resultant 20amino acid peptide sequence, VLALPYSQGPDPRRFSMYFF, identified homologous regions within the serpin superfamily of serine PIs (see Ref. 25 and references therein). Degenerate PCR primers, based on the internal peptide sequence LPYSQGPD and a sequence within plant serpins, EVNEEGTE (29 -31), were used to amplify a 342-bp fragment. This PCR product was cloned and sequenced and was confirmed to contain the expected internal peptide sequence, PRRFSMYFF.
A pumpkin stem cDNA library was then used to obtain a full-length (1363 bp) clone, CmPS-1, encoding a 389-amino acid polypeptide with a predicted molecular mass of 42.8 kDa. In the deduced CmPS-1 open reading frame, amino acid residues 221-240 precisely matched the microsequenced peptide fragment (Fig. 2, dashed line). During the cDNA library screening, a second gene (1470 bp), CmPS-2, was cloned, and the encoded polypeptide (389 amino acid residues) was found to have 72% identity to CmPS-1. However, the internal microsequence obtained from the 40-kDa phloem sap protein failed to match any region within the predicted CmPS-2 open reading frame. In addition, immunological analyses revealed that CmPS-2 was undetectable in phloem exudate (data not shown). Collectively, these results suggested that CmPS-2 is not a phloem sap protein. Characterization of this protein will be presented elsewhere.

FIG. 5. Inhibition of elastase by phloem-purified and recombinant CmPS-1. CmPS-1 (A) and recombinant His 6 -tagged CmPS-1 (B)
was pre-incubated with elastase (ࡗ) or trypsin (Ⅺ) for the indicated time, and residual proteinase activity was then measured. Equivalent experiments were performed with chymotrypsin and thrombin, and the level of inhibition was found to be Ͻ 5% (data not shown). Inhibition is defined as the ratio of the decreased proteinase activity in the presence of the added serpin to that measured in the absence of serpin. Values are averages from three independent experiments. Fitted curves for elastase inhibition were derived based on Ref. 25.

2, horizontal line).
Confirmation of the relatedness between CmPS-1 and the serpins present in barley and wheat seeds was obtained through Western analysis. Polyclonal antibodies raised against a barley serpin (25) cross-reacted with CmPS-1 present in phloem sap collected from C. maxima plants (Fig. 1B) and confirmed the developmental up-regulation of CmPS-1. An additional, but much weaker, immunoreactive band was also detected at the 70-kDa region; the level of this protein remained constant over the developmental period examined (data not shown).
Purification of CmPS-1 from C. maxima Phloem Exudate-Biochemical studies of native proteins obtained from the phloem sap of plants are rare, largely because of the difficulty in collecting sufficient amounts of sap to permit biochemical purification. In this regard, cucurbits represent an excellent system because they allow the efficient collection of large amounts of exudate. To purify CmPS-1 from phloem sap, dialyzed exudate was subjected to a series of chromatographic steps (Fig. 3A). Anion followed by cation exchange chromatography (at pH 7.0) was used to separate CmPS-1 from phloem protein 1 (PP1; 96 kDa) and phloem protein 2 (PP2; 24 kDa), the two major constituents of the C. maxima phloem sap. As a result of Q-Sepharose chromatography, a number of minor proteins were retained in the anion exchange column (data not shown), whereas most proteins, including CmPS-1, PP1, and PP2, were present in the flow-through fraction (Fig. 3B, compare lanes 1 and 2). The SP-Sepharose cation exchange chromatographic step, carried out at pH 7.0, removed PP1, PP2, and a number of additional proteins from the CmPS-1 fraction (Fig. 3B, lane 3). A second cation exchange fractionation, carried out at pH 5.5, removed the remaining high molecular weight proteins; CmPS-1 was eluted using a salt gradient (Fig.  3B, lane 4). Finally, phenyl-Sepharose with reverse gradient was used to purify CmPS-1 to near homogeneity (Fig. 3B, lane  5). Western analysis confirmed the identity of the protein preparation (Fig. 3C), and the purity of CmPS-1 was confirmed by mass spectrometry (data not shown). Special attention was taken to separate CmPS-1 from low molecular weight proteins because the phloem sap contains a number of small (3-10 kDa) PIs (16 -18), which would have complicated any further biochemical analyses.
Purification of Recombinant CmPS-1-An amino-terminal His 6 -tagged recombinant form of CmPS-1 was overexpressed in and purified from E. coli (Fig. 4). To obtain soluble cytosolic CmPS-1, low temperature induction was employed during expression. As a first purification step, soluble E. coli extracts were loaded onto Ni-agarose columns (Fig. 4A, lanes 1-5). Based on Western analyses (data not shown), the majority of the His 6 -tagged CmPS-1 was retained on the column. Elution of CmPS-1 was achieved using 400 mM imidazole in equilibration buffer, and following dialysis, the eluate was subjected to anion exchange chromatography that allowed purified CmPS-1 to be collected from the flow-through fraction (Fig. 4A, lane 6); only two minor contaminants were still present. Western analysis showed that these contaminants were not the result of proteolytic degradation of CmPS-1 (Fig. 4B).
CmPS-1 Displayed Antielastase Activity-The P1 residue in the reaction center loop generally serves to determine the inhibitory specificity of serpins (27,33). As CmPS-1 has a valine in the putative P1 position, it was expected that it would exhibit inhibitory specificity against the elastase class of serine proteinases. Trypsin, chymotrypsin, and thrombin were employed as controls for our serine proteinase inhibition assays. As predicted, CmPS-1 displayed specific inhibition against elastase but was inactive against trypsin, chymotrypsin, or thrombin (Fig. 5A). The fraction of complex-forming CmPS-1, X inh , was estimated as 79% (Table I). Values for X inh could not be obtained for trypsin, chymotrypsin, and thrombin, as the percent inhibition was negligible (Ͻ5%). The calculated association rate constant, k a , for elastase was 3.5 Ϯ 1.    (Fig. 5B and Table I).
The specificity of inhibition between CmPS-1 and elastase was supported by the proteinase-serpin complex formation assay (25). CmPS-1 formed an inhibitory complex (C) with porcine pancreatic elastase (Fig. 6A). By mixing CmPS-1 and elastase at a molar ratio of 1:1, a fraction of the CmPS-1 was carboxyl-terminally cleaved (I*), and the released carboxylterminal peptide (P) was detected at 4 kDa. CmPS-1 fragments (18 -22 kDa) produced by substrate cleavage were also detected. In addition, CmPS-1 formed a complex with human neutrophil elastase (data not shown). However, CmPS-1 was unable to interact with trypsin to form an equivalent complex; rather, CmPS-1 was degraded by trypsin, resulting in the production of many lower molecular weight bands (Fig. 6B). Neither chymotrypsin nor thrombin formed an inhibitory complex with CmPS-1 (data not shown).
Cleavage Site Determination-The porcine pancreatic elastase cleavage site on CmPS-1 was first investigated using amino-terminal sequencing of the released carboxyl-terminal peptide (approx. 4 kDa). Two amino acid peaks were obtained in each Edman degradation cycle, and the molar ratio of the major to the minor peak was approx. 3:2. The amino-terminal sequences for these major and minor peaks were GIVSLP and SLPINR, respectively. These results indicated that CmPS-1 cleavage occurred at both the Val 347 -Gly 348 and Val 350 -Ser 351 peptide bonds; these cleavage site determinations were further confirmed by using mass spectrometry. The carboxyl-terminal 4-kDa peptides were extracted from the acrylamide gel and subjected to MALDI-TOF analysis. Two peaks at 4742.68 and 4473.41 Da were obtained; these were close to being identical to the predicted values of 4742.61 and 4473.26 Da for CmPS-1 carboxyl-terminal peptides starting at Gly 348 and Ser 351 , respectively. The collective results from these analyses are presented in Fig. 7. The Val 350 -Ser 351 site matches the predicted P1-P1Ј site based on the amino acid sequence alignment. An alternate cleavage site exists at Val 347 -Gly 348 and likely represents an inhibitory cleavage site.
In Vivo and in Vitro Effects of CmPS-1 on Aphid Survival-A potential role for serpins in plants is as a feeding deterrent/ inhibitor of piercing-sucking and/or chewing insects. The piercing-sucking aphid, M. persicae, was used in experiments to ascertain whether the developmentally regulated increase in CmPS-1 within the phloem of pumpkin plants (Fig. 1) had any deleterious effect on this insect pest. For these experiments, aphids were placed on newly expanded leaves of various-aged pumpkin plants (Fig. 8) and observed over a 3-day period. The survival rate of M. persicae declined from 48% on 14-day-old plants to a plateau of approx. 18% on 28 -42-day-old plants; a parallel and significant decline in fecundity was also observed (Fig. 8). These results established a strong negative correlation between the level of CmPS-1 in the phloem sap ( Fig. 1) and aphid survival. Analysis of phloem sap collected from control and aphid-infested leaves established that CmPS-1 expression was not altered in these aphid-feeding treatments (data not shown). Next, in vitro feeding experiments were conducted to ascertain directly the influence of CmPS-1 on the survival of M. persicae. For these experiments, CmPS-1 was employed at 200 g ml Ϫ1 (4.6 M), which reflected the level present within the phloem sap of 42-day-old pumpkin plants. As ␣-1-antitrypsin (A1AT) has anti-elastase, -trypsin, and -chymotrypsin activities (34), it was included along with bovine serum albumin as a control for these feeding studies. Interestingly, in these experiments, neither phloem-purified CmPS-1 nor A1AT had any significant effect on the survival of M. persicae (Table II). DISCUSSION In the present study, we establish that the phloem sap of C. maxima contains a novel 42-kDa PI, CmPS-1, belonging to the serpin superfamily. Sequence analysis and biochemical assays conducted both with phloem-purified and recombinant CmPS-1 confirmed that this protein is an active serpin capable of specifically inhibiting elastase. Confinement of CmPS-1 mRNA to companion cells, 2 in conjunction with the presence of CmPS-1 in the phloem sap, strongly implicates a role for CmPS-1 in the operation of the enucleate sieve tube system. Expression of CmPS-1 is developmentally regulated, as CmPS-1 within the phloem sap of 7-10-day-old C. maxima seedlings was either absent or on the threshold of detection by Western analysis. Over the ensuing 4 weeks of seedling/plant development, the level of CmPS-1 increased by three orders of magnitude (10 nM-5 M). As chymotrypsin/trypsin PI activity was detectable within the phloem sap of all plants tested, general proteolytic 2 B. Xoconostle-Cá zares and W. J. Lucas, unpublished results.
FIG. 8. Influence of plant age on the survival and fecundity of phloem-feeding insects. The piercing-sucking aphid species, M. persicae, was used to demonstrate that a negative correlation exists between aphid survival/fecundity and the amount of CmPS-1 present in C. maxima phloem sap. (Plants used for these experiments were common to those employed for the studies presented in Fig. 1.) All experiments were performed on true leaves, and hence, earlier time points were precluded. Values represent mean Ϯ S.E. Control infestation experiments, performed with A. gossypii (raised on C. pepo), resulted in 50% aphid survival and 1.5 offspring/adult over the 28 -42-day growth period.

TABLE II
Effect of CmPS-1 on Myzus persicae survival during in vitro feeding assays In vitro aphid feeding assays were performed as described under "Experimental Procedures." In control experiments, aphids were fed on distilled water, 20% sucrose, or 20% sucrose solution supplemented with either protein preparation buffer (2.5 mM Hepes, pH 7.5, and 10 mM NaCl) or bovine serum albumin (BSA) (200 g ml Ϫ1 ). The concentration of CmPS-1 and human ␣-1-antitrypsin (HsA1AT) was 200 g ml Ϫ1 delivered in a 20% sucrose solution. Ten neonate M. persicae were used per treatment, and survival data collected after 24-, 48-, and 72-h feeding periods represent the mean Ϯ S.E. for 10  protection of the proteins within the functional sieve tubes (both mobile and immobile) may well be afforded by the small molecular weight PIs (16 -18). Thus, CmPS-1 may well function to protect the phloem against proteolytic activities associated with an endogenous elastase-like proteinase. The present study implicated CmPS-1 in the protection of the pumpkin phloem sap against the intrusive feeding activities of aphids. Members of the serpin superfamily of serine PIs have been identified in animals, insects, and plants. In animals, serpins play pivotal roles in many physiological processes (35,36), with perhaps the most well characterized being antithrombin III, whose function is central to the control of blood coagulation. Little is known concerning the function of the plant serpins, although a role in the protection of seeds against insects and pathogens has been proposed (25, 29 -31, 37). Given that anti-elastase activity has been shown to impart resistance to bacteria (38), it is possible that CmPS-1 confers protection against bacteria or phytoplasma that invade the tissues of the phloem (39,40).
A central role for PIs in the defense response of plants to chewing insects is well supported by a wide range of experimental evidence (24,(41)(42)(43)(44). The applicability of this paradigm to piercing-sucking insects has remained unresolved (24,45). These insects appear to lack endoproteinase activity within their digestive tract (46) and, therefore, have not been considered likely targets for PIs. However, as our in vivo aphid infestation experiments established that increasing insect mortality and CmPS-1 levels were highly correlated ( Fig. 1 and Fig.  8), it would appear that CmPS-1 may well play a role in the defense of the phloem against piercing-sucking insects. As phloem-purified CmPS-1 had no detectable effect on aphid survival (Table II), it might well be that this serpin requires additional phloem proteins to form an active complex. Future experiments will be performed to test this hypothesis. Additionally, CmPS-1 may act upon homopteran insects other than aphids. As an alternative mode of action, the possibility cannot be discounted that CmPS-1 functions, in concert with other phloem and/or insect proteins, in the sealing (possibly by occlusion) of the stylet, once the insect has probed into a sieve element. Such a mode of action would be consistent with the general observation that insect feeding is impaired on resistant plant lines (47,48).