Rat mast cell carboxypeptidase: amino acid sequence and evidence of enzyme activity within mast cell granules

The amino acid sequence of rat mast cell carboxypeptidase has been determined. The major form has 308 residues; a minor form has an additional (glutamyl) residue at the amino terminus that may indicate an alternate cleavage site during zymogen activation. The enzyme is homologous to pancreatic carboxypeptidases A and B, with conservation of the functional amino acid residues of the active site. The putative substrate binding site resembles that of carboxypeptidase A, although other structural features bear more similarity to carboxypeptidase B. Mast cell carboxypeptidase retains enzymatic activity toward a peptide substrate (angiotensin I) while bound within the granular matrix of the rat connective tissue mast cells. Evidence is presented to suggest that a cluster of positively charged lysyl and arginyl residues binds the enzyme to the negatively charged heparin of the granular matrix but leaves the active site exposed to bind and cleave peptide substrates.     

Crayfish carboxypeptidase. Affinity chromatography, characterization and amino-terminal sequence.

In an effort to trace the evolutionary history of the pancreatic metalloexopeptidases, carboxypeptidase has been isolated from the cardia of the crayfish Astacus fluviatilis. The isolation procedure included affinity chromatography on a column of potato carboxypeptidase inhibitor covalently linked to Sepharose. Approximately 25 mg of pure enzyme can be obtained by the present procedure from 50 ml of cardia fluid. The pure enzyme resembles bovine carboxypeptidase B in specificity and is inhibited both by 3-phenyllactate and by 6-aminohexanoate. The pH optimum of activity is about pH 6.5, and the isoelectric point,pH 4.0. Inhibition by typical metal chelating agents (i.e. ethylenediamine tetraacetate and 1,10-phenanthroline) and neutron activation analysis indicate that, like the mammalian enzyme, crayfish carboxypepetidase is a zinc metalloenzyme. The purified enzyme migrates as a single band in cellulose acetate, disc gel and sodium dodecylsulfate gel electrophoresis. The amino acid composition is similar to that of pancreatic carboxypeptidases except for a higher content of acidic amino acid residues. The amino acid sequence of the first 19 amino-terminal residues reveals significant homology to that of pancreatic carboxypeptidases A and B.     

Molecular cloning and sequencing of the cDNA for human membrane-bound carboxypeptidase M. Comparison with carboxypeptidases A, B, H, and N

Carboxypeptidase M, a widely distributed membrane-bound carboxypeptidase that can regulate peptide hormone activity, was purified to homogeneity from human placenta (Skidgel, R. A., Davis, R. M., and Tan, F. (1989) J. Biol. Chem. 264, 2236-2241). The NH2-terminal 31 amino acids were sequenced, and two complementary oligonucleotide probes were synthesized and used to isolate a carboxypeptidase M clone from a human placental cDNA library. Sequencing of the cDNA insert (2009 base pairs) revealed an open reading frame of 1317 base pairs coding for a protein of 439 residues. The NH2-terminal protein sequence matched the deduced amino acid sequence starting with residue 14. Hydropathic analysis revealed hydrophobic regions at the NH2 and COOH termini. The NH2-terminal 13 amino acids probably represent part of the signal peptide, and the COOH-terminal hydrophobic region may act either as a transmembrane anchor or as a signal for attachment to a phosphatidylinositol glycan moiety. The carboxypeptidase M sequence contains six potential Asn-linked glycosylation sites, consistent with its glycoprotein nature. The sequence of carboxypeptidase M was 41% identical with that of the active subunit of human plasma carboxypeptidase N, 41% identical with bovine carboxypeptidase H (carboxypeptidase E, enkephalin convertase), and 15% with either bovine pancreatic carboxypeptidase A or B. Many of the active site residues identified in carboxypeptidases A and B, including all of the zinc-binding residues (2 histidines and a glutamic acid), are conserved in carboxypeptidase M. These data indicate that all of the metallocarboxypeptidases are related, but the nondigestive carboxypeptidases with more specialized functions, present in cell membranes, blood plasma, or secretory granules (i.e., carboxypeptidase M, carboxypeptidase N and carboxypeptidase H), are more closely related to each other (41-49% identity) than they are to carboxypeptidase A or B (15-20% identity).     

Carboxypeptidase T--intracellular carboxypeptidase of Thermoactinomycetes--a distant analog of animal carboxypeptidase

Carboxypeptidase T, an extracellular carboxypeptidase from Thermoactinomyces sp. was isolated and purified by affinity chromatography on bacitracin adsorbents. The enzyme homogeneity was established by SDS electrophoresis (Mr = 38 000) and isoelectrofocusing in PAAG (pI 5.3). Carboxypeptidase T reveals a mixed specificity in comparison with pancreatic carboxypeptidases A and B and cleaves with nearly the same efficiency the peptide bonds formed by the C-terminal residues of basic and neutral hydrophobic amino acids. The enzyme is insensitive to serine and thiol proteinase inhibitors but is completely inhibited by EDTA and o-phenanthroline. The maximal enzyme activity is observed at pH 7-8. With an increase of temperature from 20 to 70 degrees C the enzyme activity is enhanced approximately 10-fold. In the presence of 1 mM Ca2+ the enzyme thermostability is also increased. In terms of some properties, e.g. substrate specificity carboxypeptidase T is similar to metallocarboxypeptidase secreted by Streptomyces griseus. The N-terminal sequence of carboxypeptidase T: Asp-Phe-Pro-Ser-Tyr-Asp-Ser-Gly- Tyr-His-Asn-Tyr-Asn-Glu-Met-Val-Asn-Lys-Ile-Asn-Thr-Val-Ala-Ser-Asn-Tyr- Pro-Asn - Ile-Val-Lys-Thr-Phe-Ser-Ile-Gly-Lys-Val-Tyr-Glu-Gly-Xaa-Gly-Leu- coincides by 21% with that of pancreatic carboxypeptidases A and B. Thus, it may be concluded that these enzymes originate from a common precursor.     

Isolation and molecular cloning of mast cell carboxypeptidase A. A novel member of the carboxypeptidase gene family

Mast cell carboxypeptidase A has been isolated from the secretory granules of mouse peritoneal connective tissue mast cells (CTMC) and from a mouse Kirsten sarcoma virus-immortalized mast cell line (KiSV-MC), and a cDNA that encodes this exopeptidase has been cloned from a KiSV-MC-derived cDNA library. KiSV-MC-derived mast cell carboxypeptidase A was purified with a potato-derived carboxypeptidase-inhibitor affinity column and was found by analytical sodium dodecyl sulfate-polyacrylamide gel electrophoresis to be a Mr 36,000 protein. Secretory granule proteins from KiSV-MC and from mouse peritoneal CTMC were then resolved by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transblotted to polyvinylidine difluoride membranes. Identical aminoterminal amino acid sequences were obtained for the prominent Mr 36,000 protein present in the granules of both cell types. Based on the amino-terminal sequence, an oligonucleotide probe was synthesized and used to isolate a 1,470-base pair cDNA that encodes this mouse exopeptidase. The deduced amino acid sequence revealed that, after cleavage of a 15-amino acid hydrophobic signal peptide and a 94-amino acid activation peptide from a 417-amino acid preproenzyme, the mature mast cell carboxypeptidase A protein core has a predicted Mr of 35,780 and a high positive charge [Lys + Arg) - (Asp + Glu) = 17) at neutral pH. Although critical zinc-binding amino acids (His67, Glu70, His195), substrate-binding amino acids (Arg69, Asn142, Arg143, Tyr197, Asp255, Phe278), and cysteine residues that participate in intrachain disulfide bonds (Cys64-Cys77, Cys136-Cys159) of pancreatic carboxypeptidases were also present in mast cell carboxypeptidase A, the overall amino acid sequence identities for mouse mast cell carboxypeptidase A relative to rat pancreatic carboxypeptidases A1, A2, and B were only 43, 41, and 53%, respectively. RNA and DNA blot analyses revealed that mouse peritoneal CTMC, KiSV-MC, and bone marrow-derived mast cells all express a prominent 1.5-kilobase mast cell carboxypeptidase A mRNA which is transcribed from a single gene. We conclude that mouse mast cell carboxypeptidase A is a prominent secretory granule enzyme of mast cells of the CTMC subclass and represents a novel addition to the carboxypeptidase gene family.     

cDNA cloning and complete primary structure of the small, active subunit of human carboxypeptidase N (kininase 1)

The human plasma metallo-protease carboxypeptidase N of Mr 280,000 consists of two small, enzymatically active subunits of Mr 50,000 and two large subunits. Only the large subunits are glycosylated. They may have a function in stabilizing the complex in plasma. The N-terminal sequence of the small subunit was determined from the isolated protein and used to specify a unique 59-mer oligonucleotide probe. A cDNA clone of 1.7 kbp containing the entire coding sequence of the small subunit of carboxypeptidase N was isolated from a human-liver cDNA library. The cDNA clone encodes a signal sequence of 20 amino acids and the 438 amino acids of the mature subunit. There is a remarkable primary structure similarity of 49% to bovine carboxypeptidase E (enkephalin convertase). A more distant relationship to the bovine pancreatic, digestive carboxypeptidases A and B or even to the metallo-endopeptidases is based mainly on the occurrence of conserved, mechanistically important residues.     

Crystal structure of avian carboxypeptidase D domain II: a prototype for the regulatory metallocarboxypeptidase subfamily.

The crystal structure of domain II of duck carboxypeptidase D, a prohormone/propeptide processing enzyme integrated in a three repeat tandem in the natural system, has been solved, constituting a prototype for members of the regulatory metallocarboxypeptidase subfamily. It displays a 300 residue N-terminal alpha/beta-hydrolase subdomain with overall topological similarity to and general coincidence of the key catalytic residues with the archetypal pancreatic carboxypeptidase A. However, numerous significant insertions/deletions in segments forming the funnel-like access to the active site explain differences in specificity towards larger protein substrates or inhibitors. This alpha/beta-hydrolase subdomain is followed by a C-terminal 80 residue beta-sandwich subdomain, unique for these regulatory metalloenzymes and topologically related to transthyretin and sugar-binding proteins. The structure described here establishes the fundamentals for a better understanding of the mechanism ruling events such as prohormone processing and will enable modelling of regulatory carboxypeptidases as well as a more rational design of inhibitors of carboxypeptidase D.     

Purification of a human urinary carboxypeptidase (kininase) distinct from carboxypeptidases A, B, or N

A carboxypeptidase which cleaves basic C-terminal amino acids from peptides was purified from concentrated human urine by a three-step procedure: chromatography on Affi-Gel Blue, arginine-Sepharose affinity chromatography, and gel filtration by HPLC on a TSK-G3000SW column. Urinary carboxypeptidase was purified 406-fold with an 11% yield and a specific activity of 49 mumol/min/mg with benzoylglycylargininic acid as substrate. It migrated as a single band of Mr 75,700 in polyacrylamide gel electrophoresis with sodium dodecyl sulfate. It cleaved benzoylglycylarginine, benzoylglycyllysine, benzoylglycylargininic acid, benzoylalanyllysine, and benzoylphenylalanyllysine at different relative rates than human plasma carboxypeptidase N, the Mr 48,000 active subunit of carboxypeptidase N or human pancreatic carboxypeptidase B. Urinary carboxypeptidase did not hydrolyze benzoylglycylphenylalanine, a substrate of carboxypeptidase A, but readily cleaved bradykinin with a Km of 46 microM and a Kcat of 32 min-1. Its activity was enhanced by CoCl2 and inhibited by cadmium acetate, o-phenanthroline, or DL-2-mercaptomethyl-3-guanidinoethylthiopropanoic acid. The enzyme had a pH optimum of 7.0 and its activity dropped at pH 6.0 by 60%. It was stable for at least 2 h at 37 degrees C (pH 8.0) but was unstable at room temperature below pH 4.5. The molecular weight, electrophoretic mobility, and activity of urinary carboxypeptidase was not affected by trypsin. The effect of pH and stability further distinguished the urinary carboxypeptidase from other human carboxypeptidases. Urinary carboxypeptidase was immunologically distinct from carboxypeptidase N when analyzed by the "Western blot" technique. Thus, human urine contains a basic carboxypeptidase, different from known carboxypeptidases, which may be released into the urine by the kidney. Here it could inactivate kinins and other peptides containing a basic C-terminal amino acid.     

Carboxypeptidase inhibitor from ripened tomatoes; Purification and properties

A polypeptide which inhibits pancreatic carboxypeptidases has been purified from extracts of tomato fruit (Lycopersicum esculentum). The purification procedure involves chromatography on immobilized carboxypeptidase A and gel filtration, and yields homogeneous inhibitor as judged by disc gel electrophoresis and amino acid analysis. The inhibitor from tomatoes contains 38 amino acid residues and is related to one from potato tubers as determined by a comparison of their amino acid compositions and by cross-reactivity against anti-potato inhibitor antibodies. Inhibition studies indicate that bovine carboxypeptidase A, porcine carboxypeptidase B, and a carboxypeptidase from Streptomyces griseus bind to the inhibitors from tomato and potato with approximately the same free energy of association.     

A novel rat carboxypeptidase, CPA2: characterization, molecular cloning, and evolutionary implications on substrate specificity in the carboxypeptidase gene family

A new member of the carboxypeptidase gene family, carboxypeptidase A2 (CPA2), has been identified from the predicted amino acid sequence of a rat pancreatic cDNA clone. In vivo recombination and in situ hybridization techniques employing the CPA2 cDNA resulted in the isolation of two genomic clones spanning the 25-kilobase pair rat CPA2 gene. Evolutionary trees built from the amino acid sequences of the known pancreatic carboxypeptidases show that CPA2 and carboxypeptidase A1 (CPA1) are the products of genes which duplicated before the mammalian radiation, and that bovine CPA is of the A1 type. The substrate specificities of CPA1 and CPA2 isolated from rat pancreas are similar to bovine CPA in that carboxyl-terminal amino acids with aromatic or branched aliphatic side chains are preferred. However, the substrate preference of rat CPA1 is skewed toward smaller amino acids, while that of rat CPA2 is skewed toward bulkier amino acids as compared to bovine CPA. The differences in the substrate specificities of these three carboxypeptidases are compatible with the nature of the amino acid replacements in their binding pockets for the carboxylterminal amino acid of the substrate.     

The three-dimensional structures of tick carboxypeptidase inhibitor in complex with A/B carboxypeptidases reveal a novel double-headed binding mode

The tick carboxypeptidase inhibitor (TCI) is a proteinaceous inhibitor of metallo-carboxypeptidases present in the blood-sucking tick Rhipicephalus bursa. The three-dimensional crystal structures of recombinant TCI bound to bovine carboxypeptidase A and to human carboxypeptidase B have been determined and refined at 1.7 A and at 2.0 A resolution, respectively. TCI consists of two domains that are structurally similar despite the low degree of sequence homology. The domains, each consisting of a short alpha-helix followed by a small twisted antiparallel beta-sheet, show a high level of structural homology to proteins of the beta-defensin-fold family. TCI anchors to the surface of mammalian carboxypeptidases in a double-headed manner not previously seen for carboxypeptidase inhibitors: the last three carboxy-terminal amino acid residues interact with the active site of the enzyme in a way that mimics substrate binding, and the N-terminal domain binds to an exosite distinct from the active-site groove. The structures of these complexes should prove valuable in the applications of TCI as a thrombolytic drug and as a basis for the design of novel bivalent carboxypeptidase inhibitors.     

Crystal structure of a novel mid-gut procarboxypeptidase from the cotton pest Helicoverpa armigera.

The cotton bollworm Helicoverpa armigera (Hubner) (Lepidoptera: Noctuidae) is one of the most serious insect pests in Australia, India and China. The larva causes substantial economical losses to legume, fibre, cereal oilseed and vegetable crops. This pest has proven to be difficult to control by conventional means, mainly due to the development of pesticide resistance. We present here the 2.5 A crystal structure from the novel procarboxypeptidase (PCPAHa) found in the gut extracts from H. armigera larvae, the first one reported for an insect. This metalloprotease is synthesized as a zymogen of 46.6 kDa which, upon in vitro activation with Lys-C endoproteinase, yields a pro-segment of 91 residues and an active carboxypeptidase moiety of 318 residues. Both regions show a three-dimensional structure quite similar to the corresponding structures in mammalian digestive carboxypeptidases, the most relevant structural differences being located in the loops between conserved secondary structure elements, including the primary activation site. This activation site contains the motif (Ala)(5)Lys at the C terminus of the helix connecting the pro- and the carboxypeptidase domains. A remarkable feature of PCPAHa is the occurrence of the same (Ala)(6)Lys near the C terminus of the active enzyme. The presence of Ser255 in PCPAHa instead of Ile and Asp found in the pancreatic A and B forms, respectively, enlarges the S1' specificity pocket and influences the substrate preferences of the enzyme. The C-terminal tail of the leech carboxypeptidase inhibitor has been modelled into the PCPAHa active site to explore the substrate preferences and the enzymatic mechanism of this enzyme.     

The crystal structure of the inhibitor-complexed carboxypeptidase D domain II and the modeling of regulatory carboxypeptidases

The three-dimensional crystal structure of duck carboxypeptidase D domain II has been solved in a complex with the peptidomimetic inhibitor, guanidinoethylmercaptosuccinic acid, occupying the specificity pocket. This structure allows a clear definition of the substrate binding sites and the substrate funnel-like access. The structure of domain II is the only one available from the regulatory carboxypeptidase family and can be used as a general template for its members. Here, it has been used to model the structures of domains I and III from the former protein and of human carboxypeptidase E. The models obtained show that the overall topology is similar in all cases, the main differences being local and because of insertions in non-regular loops. In both carboxypeptidase D domain I and carboxypeptidase E slightly different shapes of the access to the active site are predicted, implying some kind of structural selection of protein or peptide substrates. Furthermore, emplacement of the inhibitor structure in the active site of the constructed models showed that the inhibitor fits very well in all of them and that the relevant interactions observed with domain II are conserved in domain I and carboxypeptidase E but not in the non-active domain III because of the absence of catalytically indispensable residues in the latter protein. However, in domain III some of the residues potentially involved in substrate binding are well preserved, together with others of unknown roles, which also are highly conserved among all carboxypeptidases. These observations, taken together with others, suggest that domain III might play a role in the binding and presentation of proteins or peptide substrates, such as the pre-S domain of the large envelope protein of duck hepatitis B virus.     

Molecular cloning, nucleotide sequence, and expression of a carboxypeptidase-encoding gene from the archaebacterium Sulfolobus solfataricus.

Mammalian metallocarboxypeptidases play key roles in major biological processes, such as digestive-protein degradation and specific proteolytic processing. A Sulfolobus solfataricus gene (cpsA) encoding a recently described zinc carboxypeptidase with an unusually broad substrate specificity was cloned, sequenced, and expressed in Escherichia coli. Despite the lack of overall sequence homology with known carboxypeptidases, seven homology blocks, including the Zn-coordinating and catalytic residues, were identified by multiple alignment with carboxypeptidases A, B, and T. S. solfataricus carboxypeptidase expressed in E. coli was found to be enzymatically active, and both its substrate specificity and thermostability were comparable to those of the purified S. solfataricus enzyme.     

Characterization and cDNA cloning of midgut carboxypeptidases from Trichoplusia ni

Carboxypeptidase A and carboxypeptidase B activities from the midgut of Trichoplusia ni larvae were characterized. In the T. ni larval midgut, the primary digestive carboxypeptidase activity was attributed to carboxypeptidase A, which was eight times more active than carboxypeptidase B. Both the midgut carboxypeptidase A and carboxypeptidase B exhibited maximal activities at pH 8.0-8.5 and were similarly susceptible to inhibition by potato carboxypeptidase inhibitor and phenanthroline. The midgut carboxypeptidase activities were analyzed in T. ni larvae fed on various diet sources and the results indicated that midgut carboxypeptidase activities per milligram of gut were similar regardless of the amount of dietary proteins or amino acids. However, midgut carboxypeptidase A activity was significantly higher in larvae exposed to soybean trypsin inhibitor and was significantly lower in larvae fed on broccoli foliage. From the T. ni larval midgut, five putative carboxypeptidase cDNAs were cloned, demonstrating that midgut carboxypeptidase activities are composed of multiple carboxypeptidase types. Sequence analysis indicated that the midgut carboxypeptidases were produced as secreted proenzymes which could be activated after removal of an N-terminal activation fragment by a trypsin. Two cloned cDNAs are predicted to code for carboxypeptidase A and one cDNA is predicted to code for a putative carboxypeptidase B. The other two cDNAs are highly similar to carboxypeptidase A and carboxypeptidase B in sequences, but their activity was not predictable.     

Design of potent and specific inhibitors of carboxypeptidases A and B.

The combination in one molecule of functional groups that can interact specifically with different substrate binding areas at the active site of carboxypeptidases A and B has led to the development of potent and specific inhibitors of these enzymes. 2-Benzyl-3-mercaptopropanoic acid (SQ 14,603) has a Ki of 1.1 x 10(-8) M vs. carboxypeptidase A and a Ki of 1.6 x 10(-4) M vs. the B enzyme. 2-Mercaptomethyl-5-guanidinopentanoic acid (SQ 24,798) has a Ki of 4 x 10(-10) M vs. carboxypeptidase B and a Ki of 1.2 x 10(-5) M vs. carboxypeptidase A. It is proposed that the sulfhydryl groups of these inhibitors bind to the catalytically important zinc ions of these enzymes, and that, in conjunction with the benzyl and guanidinopropyl side chains, they are responsible for their specificity.     

Purification and characterization of a thermostable carboxypeptidase from the extreme thermophilic archaebacterium Sulfolobus solfataricus.

A carboxypeptidase was purified to electrophoretic homogeneity from the thermoacidophilic archaebacterium Sulfolobus solfataricus. Molecular masses assessed by SDS/PAGE and gel filtration were 42 kDa and 170 kDa, respectively, which points to a tetrameric structure for the molecule. An isoelectric point of 5.9 was also determined. The enzyme was proven to be a metalloprotease, as shown by the inhibitory effects exerted by EDTA and o-phenanthroline; furthermore, dialysis against EDTA led to a complete loss of activity, which could be restored by addition of Zn2+ in the micromolar range, and, to a lesser extent, by Co2+. The enzyme was endowed with a broad substrate specificity, as shown by its ability to release basic, acidic and aromatic amino acids from the respective benzoylglycylated and benzyloxycarbonylated amino acids. An esterase activity of the carboxypeptidase was also demonstrated on different esterified amino acids and dipeptides blocked at the N-terminus. The enzyme displayed broad pH optima ranging over 5.5-7.0, or 5.5-9.0, when using an acidic or a basic benzyloxycarbonylated amino acid, respectively. With regard to thermostability, it was proven to be completely stable on incubation for 15 min at 85 degrees C. Furthermore, thanks to its relatively low activation energy, i.e. 31.0 kJ/mol, it was still significantly active at room temperature. At 40 degrees C, the enzyme could withstand 0.1% SDS and different organic solvents: particularly ethanol up to 99%. Amino acid and N-terminal sequence analyses did not evidence any similarity to carboxypeptidases A nor thermolysin. A weak similarity was only found with bovine carboxypeptidase B.     

The isolation and partial characterization of precursor forms of ostrich carboxypeptidase

Ostrich carboxypeptidases A and B were recently purified and characterized. The aim of this study was to isolate and purify, and partially characterize in terms of molecular weight, pI, amino acid composition and N-terminal sequencing, the precursor forms of carboxypeptidases from the ostrich pancreas. Inhibition studies with soybean trypsin inhibitor and activation studies with three proteases (bovine trypsin, bovine chymotrypsin and porcine elastase) were performed on crude ostrich acetone powder and the carboxypeptidase A and B activities were determined. SDS-PAGE was carried out after every incubation to monitor the rate and degree of conversion of a M(r) 66K component to procarboxypeptidase and carboxypeptidase A and B. The precursor forms were purified by Toyopearl Super Q and Pharmacia Mono Q chromatography. All three proteases converted the M(r) 66K component to procarboxypeptidases and carboxypeptidases over a set time interval, with carboxypeptidase A and B activities being detected in the acetone powder. Chymotrypsin was the preferred protease since it exhibited a more controlled activation of the procarboxypeptidases. The amino acid composition of procarboxypeptidase A revealed 525 residues. The N-terminal sequence of procarboxypeptidase A showed considerable homology when compared with several other mammalian sequences. M(r) and pI values of 52K and 5.23 were obtained for procarboxypeptidase A, respectively. This study indicated that ostrich procarboxypeptidase A is closely related to other mammalian procarboxypeptidase A molecules in terms of physicochemical properties.     

A potato carboxypeptidase inhibitor gene provides pathogen resistance in transgenic rice

A defensive role against insect attack has been traditionally attributed to plant protease inhibitors. Here, evidence is described of the potential of a plant protease inhibitor, the potato carboxypeptidase inhibitor (PCI), to provide resistance to fungal pathogens when expressed in rice as a heterologous protein. It is shown that rice plants constitutively expressing the pci gene exhibit resistance against the economically important pathogens Magnaporthe oryzae and Fusarium verticillioides . A M. oryzae carboxypeptidase was purified by affinity chromatography and further characterized by mass spectrometry. This fungal carboxypeptidase was found to be a novel carboxypeptidase B which was fully inhibited by PCI. Overall, the results indicate that PCI exerts its antifungal activity through the inhibition of this particular fungal carboxypeptidase B. Although pci confers protection against fungal pathogens in transgenic rice, a significant cost in insect resistance is observed. Thus, the weight gain of larvae of the specialist insect Chilo suppressalis (striped stem borer) and the polyphagous insect Spodoptera littoralis (Egyptian cotton worm) fed on pci rice is significantly larger than that of insects fed on wild-type plants. Homology-based modelling revealed structural similarities between the predicted structure of the M. oryzae carboxypeptidase B and the crystal structure of insect carboxypeptidases, indicating that PCI may function not only as an inhibitor of fungal carboxypeptidases, but also as an inhibitor of insect carboxypeptidases. The potential impact of the pci gene in terms of protection against fungal and insect diseases is discussed.