Human Skin Mast Cell Carboxypeptidase: Functional Characterization, cDNA Cloning, and Genealogy

Masaru Natsuaki, Caro-Beth Stewart, Peter Vanderslice, Lawrence B. Schwartz, Megumi Natsuaki, Bruce U. Wintroub, William J. Rutter, and Sanford M. Goldstein Departments of Dermatology, Hormone Research Institute (C-BS, WJR) and Cardiovascular Research Institute (PV) , University of California, San Francisco, and Department of Medicine (LBS), Medical College of Virginia, Richmond, Virginia, U .S.A.

We functionally characterized human skin mast cell carbox­ in substrate specificity, relative to bovine CPA, than might be ypeptidase A (MC-CPA), and explored its evolutionary rela­ expected from studies using rat CPA 1 and CP A2. W e noted tionship to other carboxypeptidases to understand further the two consensus N-linked glycosylation sites in human MC­ structural basis for the substrate preferences of this . CPA that are not found in rat and mouse MC-CPA, or in Purified human skin MC-CPA displayed more activity than bovine CPA; that at least one of these sites is glycosylated in did bovine pancreatic carboxypeptidase A (CPA) against car­ vivo was verified by N-glycosidase F treatment, lentil lectin boxyl-terminal leucine residues, about equal activity with binding, and Concanavalin A - Sepharose chromatography. phenylalanine and tyrosine residues, and no activity with ~volutionary trees constructed from the known carboxypep_ tryptophan or alanine. To correlate kinetic data with struc­ tIdase sequences suggested that MC-CPA most likely ture, we isolated and sequenced a cD N A encoding MC-CPA evolved from a carb?xypeptidase B~like enzyme, indepen­ from human skin, and directly sequenced 30% of the purified dent of the pancreatic CPA. Thus, III the carboxypeptidase protein. These sequences agreed with that of human lung gene family, MC-CPA displays a unique genealogy and sev­ MC-CPA, and further support the evidence for a single MC­ eral amino acid replacements in its S l' binding pocket that CPA gene in humans. Four amino acid replacements, result­ result in substrate specificity quite similar to bovine CPA. ing in a net positive change in non-hydrogen atoms in the S l' ] [,west Dermatol99:138-145, 1992 subsite of MC-CPA, were associated with less alteration

uman mast cell carboxypeptidase (MC-CPA) [1,2] by carboxypeptidase B. Yet, MC-CPA has not been as extensively is one of four neutral proteases identified in human characterized with regard to substrate preferences as have pancreatic mast cell s, along with [3] , [4-6], carboxype12tidases [2,11]. In addition, the sequences of human [9] , and a G - like protease [7]. These mast cell mouse [12J, and rat [13] MC-CPA have not been analyzed with are homologous with pancreatic proteases regard to Sl' subsite* replacements that appear to correlate with [H2,8- 10]. MC-CPA has substrate specificity similar to pancreatic substrate preferences of CPA enzymes [11], although the M C-CPA carboxypeptidases A in that it recognizes hydrophobic carboxyl­ sequences have revealed an amino acid replacement in the substrate terminal residues, and not the basic residues, Arg or Lys, recognized that appears critical for the hydrolysis of hydrophobic

Manuscript received October 23, 199 1; accepted for publication March Cbz-Gly-Gly-Trp: carbobenzoxy-glycyl-L glycyl-L-tryptophan 30, 1992. Cbz-Gly-Gly-Tyr: carbobenzoxy-glycyl-L glycy l-L-tyrosine This work was su pported by grants AM31901 and AI20487 from the C PA: carboxypeptidase A National Institutes of Health. CPB: carboxypeptidase B Caro-Beth Stewart is a recipient of a National Science Foundation post­ LEP: lysylendopeptidase doctoral fellowship (BSR 8700-192) and an Alfred P. Sloan Foundation MC-CPA: mast cell carboxypeptidase A

postdoctoral fe ll owship (89-1-8 ME). MCT : mast cell s staining positive for tryptase and negative for chy­ Caro-Beth Stewart's present address: Department of Biological Sciences, mase, carboxypeptidase, and a - li ke enzy me SUNY-Albany, Albany, NY 12222. MCTC : mas t ce ll s staining positive for tryptase, chymase, carboxy- Masaru Natsuaki and Megomi Natsuaki 's present address: 1-11-9 Mina­ peptidase, and a cathepsin G - like enzyme mitsukaguchi-cho, Amagasa ki , Hyogo 661, Japan. NBT: p-nitro-blue-tetrazolium chloride Presented in part at the 51st Annual Meeting of the Society for Investiga­ PCR: polymerase chain reaction tive Dermatology, W as hington, D.C., 1990. PEG: polyethylene glycol Reprint requests to: Dr. Sanford M. Goldstein , Department of Dennatol­ SBTI: soybean inhibitor ogy, University of California, Sa n Francisco, CA 94143-0536. SDS-PAGE: sod ium dodecyl sulfate-polyacryl amide gel electropho­ Abbreviations: reSiS AI: angiotensin I TFA: trifluoroacetic acid BCIP: 4-ch loro-1-nap htol, 5-bromo-4-chl oro-3-indolyl phosphate- toluidine sa lt Chz-G ly-G ly-Ala: ca rbobenzoxy-glycyl-L glycyl-L-ala nine • The S1' subsite of the enzyme refers to the binding pocket for the Cbz-Gly-Gly-Leu: carbobenzoxy-glycyl-L glycyl-L-leucine carboxyl-terminal amino acid of the substrate in accordance with the no­ Cbz-Gly-Gly-Phe: carbobenzoxy-glycy l-L glycyl-L-phenylalanine menclature of Schechter and Berger [1 4].

0022-202X/92/S05.00 Copyright © 1992 by The Society for Investigative Dermatology, Inc.

138 VOL. 99, NO. 2 AUGUST 1992 HUMAN MAST CELL CARBOXYPEPTIDASE FROM SKIN 139

residues [9,12,13,15]. Finally, human mast cell tryptase and chy­ protein concentration of MC-CP A and the specific activity of the mase are glycosylated in vivo [16,17], but MC-CPA has not been preparation we used (1 2.0 U/ f1 g protein) compared to our prepara­ studied in this regard. Glycosylation appears to be characteristic for tion with the highest specific activity (36.4 U/f1g protein) [2]. human and dog mast cell neutral proteases, but not for the corre­ sponding rat enzymes [18] . In this work we have compared the Enzyme Assays MC-CPA and bovine CPA activities eluting substrate preferences of MC-CPA purified from skin to that of from Concanavalin A-Sepharose were measured by the hydrolys is 9 bovine pancreatic CPA, and we have also demonstrated that MC­ of the carboxy terminal His -Leu 10 bond of angiotensin I as de­ CPA is glycosy lated in vivo. In order to inves tigate the relationship scribed previously [1]. MC-CPA and bovine CPA hydrolysis of of the kinetic and glycosylation data with the sequence of MC­ sy nthetic tripeptide substrates (Cbz-Gly-Gly-X, where X was the CPA, we iso lated and sequenced a cDNA encoding MC-CPA from substituted carboxy terminal amino ac id Leu, Phe, Tyr) were mea­ skin and directly sequenced several protease-generated fragments of sured twice at several enzyme concentrations and four to fi ve sub­ MC-CPA, constituting approximately 30% of the purified protein. strate concentrations in a 500-f1l volume of 0.01 M Tris HCl (pH The sequence of human MC-CPA cDNA from skin was identical to 8.0),0.15 M N aCl, for 7 min at room temperature. Cbz-Gly-Gly­ the enzyme from human lung [12] . This is consistent with evidence Phe was used at 10 to 80 f1M, Cbz-Gly-Gly-Leu at 50 to 800 f1M, suggesting that only one MC-CPA gene is expressed in humans and Cbz-Gly-Gly-Tyr at 25 to 200 f1M . Cbz-Gly-Gly-Trp and [12] . The amino acid sequences of mast cell carboxypeptidases are Cbz-Gly-Gly-Ala were used at 100 f1M and at 5 f1M and 50 f1M, more similar in overall sequence to the ca rboxypeptidase B (CPB) respectively. The reaction was stopped with an equal volume of enzymes rather than to the CPA [2,9,12], suggesting a closer evolu­ HPLC buffer (0 .017 M H3HP04 , 0.1 M NaC10 4 in 35% C H3CN), tionary relationship. To es tablish the genealogy of the carboxypep­ and assayed on a Beckman model 144 HPLC system by using a tidase fami ly, we built parsimony trees from the known sequences; 4.6 X 250 mm Altex ultrasphere 5 f1m ODS C18 column employ­ these trees suggest that the ge ne duplication that gave rise to the ing an isocratic perchlorate ion-pai r reverse-phase system using mast ce ll carboxypeptidases occurred along the CPB lineage. HPLC buffer at a flow rate of 1.0 mljmin [21]. The product, Cbz­ Gly-Gly, was monitored by absorbance at 200 nm, identified by MATERIALS AND METHODS retention time, and quantified by integration of peak area by using a Hewlett Packard HP 3390E integrating recorder. With the use of Materials The following oligonucleotides were synthesized by this system, Cbz-Gly-Gly eluted at approximately 4.15 min. The the Biomolecular Resource Center at UCSF: #7506 5'-GAG ATT kinetic constants K", and kcar were obtained by direct fit of 4 -10 data TCC AAG CTT GGG GAA CGA TTT TTT CTC CAC ATG points for each enzyme concentration using a non-linear least­ CGG TTC TTT GTC CAT GAC C -3'; #7505 5-' GAG CGG square regression analysis [22]. CCGCCTCGAGTCGACCCCAGGCAGGCACAGCTA CGC AAA ATA C -3'. Oligonucleotide #BS09 5'-TAC AAC Binding of MC Proteases to Concanavalin A-Sepha­ AAC TGG GAG AAG ATI GTI GCI TGG ACI GAG AAG ATG rose Bovine pancreatic CPA and purified MC-CPA were dia­ ATG GA-3' was synthesized by the Hormone Research Institute at lyzed against wash buffer (6 mM KH2P04 , 30 mM N 2 HP04 , pH U CSF. 7.6,0.1 mM CaCI2 ,0.5 M NaCl, 0.1 mMMnCI2 ' 4H20,and 0.1% Acetonitrile (Pierce), Guanidine-HCl (Aldrich) , Gigapack Gold PEG). A I-ml aliquot of each enzyme was independently mixed (Stratagene), Leucine aminopeptidase (Worthington) ; Tween 20, with 1 ml of Concanavalin A - Sepharose (Pharmacia) that was pre­ 4-chloro-l-naphtol, 5-bromo-4-chloro-3-indolyl phosphate-to Iui­ equilibrated in wash buffer. The slurry was mixed for 40 min at dine salt (BCIP); p-nitro-blue-tetrazolium chloride (NBT) 25 ° C using a multi-purpose rotator (Scientific Industries) and (Biorad); biotin-conjugated lentil lectin, avidin-conjugated horse poured into a Quik-sep column (Isolabs). The effl uent was collected rad ish peroxidase (EY Laboratories); Centricon 10 microconcentra­ and the column was washed with 15 vol of wash buffer, and then tors (Amicon) ; MgCl2 (Mallinckrodt) ; TFA (Pierce); Chromozyme was eluted with 15 ml of 1 M a-methyl -D-mannopyranoside PL (Tosyl-Gly-Pro-Lys-p-nitranilid-acetate) (Boehringer Mann­ (Sigma) in wash buffer. Aliquots weJ;,e assayed for carboxypeptidase heim); QUIK-SEP columns (ISOLAB); Angiotensin I (Star Bio­ activity using AI as described above. chemicals Inc. ); SBTI (Calbiochem); Nitroplus 2000 membranes (Fisher); Concanavalin A Sepharose (Pharmacia); Nonidet p-40, N-Glycosidase F Treatment of MC-CPA Purified MC-CPA .a -methyl-D-mannopyranoside, gelatin (Sigma); phenol, Blugene, (3 6 f1 g) was concentrated and desalted using a Centricon 10 micro­ concentrator (Amicon), dried under vaccum, and redissolved in 30 protease K, Kpn I, Hind Ill, T 4 polynucleotide kinase (BRL); Se­ quenase (US Biochemical); Erase-a-base (Promega); ca rbobenzoxy­ f1l of 0.5 % SDS, and trea ted with N-glycosidase F as described [1 8]. glycyl-L-glycyl-L-alanine (Cbz-Gly-Gly-Ala), Cbz-Gly-Gly-Phe, Aliquots of MC-CPA treated identically but with or without N­ Cbz-Gly-Gly-Leu, Cbz-Gly-Gly-Tyr, and Cbz-Gly-Gly-Trp (Ba­ glycosidase F were analyzed on a 12.5% polyacrylamide-SDS gel. A chem); N-glycosidase F (Genzyme) were obtained as noted. 2-f1l sample of each aliquot was also applied to nitrocellulose, incu­ bated with biotin-conjugated lentil lectin overnight, and visualized METHODS as described [18]. Purification ofMC-CPA and Bovine CPA MC-CPA was pu­ Protein Sequencing Peptide fragments of MC-CP A were made rified to a single band on SDS-PAGE using a single-step affiniry­ using lysyl- and leucine aminopeptidase digestion, purification procedure as previously desc ribed [2] . Bovine pancre­ lysylendopeptidase digestion and separation of peptides [23]' 70 Jig atic CPA (Sigma) was repurified by HPLC on a MonoQ HR5/5 of purified MC-CPA was incubated in 3.55 M guanidine-HC 1, 0.05 anion exchange column (Pharmacia). Protein was applied to this M Tris-HC1, pH 8.9, and 4.4 mU of lysylendopeptidase (LEP) for column in 20 mM Tris, pH 8.0, 20 mM NaCl, and eluted with a 16 h at 30°C, after which a second aliquot of lysylendopeptidase linear gradient up to 200 mM NaCI. Bovine pancreatic CPA was was added and the incubation continued for 18 h. A control was stored at 4 ° C. Protein concentrations were determined spectropho­ performed containing all components except MC-CPA. Three tometrically using the value of 6.42 X 104 M- l cm- 1 as the extinc­ hundred fifty microliters of the sample was then subjected to re­ tion coefficient at 278 nm for bovine CPA [19], and by Coomassie verse-phase HPLC using a Rainin Rabbit binary pumping system blue binding (Biorad) for MC-CPA usi ng bovine serum albumin as and Hewlett Packard 1040A diode array detector, equipped with a a standard. The determination of MC-CPA concentration by the 79994a ChemStation. The pep tides were separated by HPLC on a Biorad assay correlated well with that determined by titration of Brownlee C-8, RP300, 2.1 X 300 mm column, using a gradient MC-CPA against potato carboxypeptidase inhibitor [20], when program that consisted of injection and 5-min wash with solvent A compared using another preparation of MC-CPA. The concentra­ (0.1 % T FA in water), foll owed by a linear ramp to 100% solve nt B tion of catalytic centers of MC-CPA was determined using the (70% CH3CN/ 30% water/ O.l % TFA) at 200 f1l/min in 45 min. 140 NATSUAKI ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Fractions were dried to completeness in a speed vacuum concentra­ with Kpn I and Hind III, and incubated with Exonuclease III at tor, and those containing the most protein were subjected to se­ 25 ° C. Aliquots removed at I -min intervals, and subjected to agar.ose quencing as described below. gel electrophoresis, demonstrated deletions of increasing ~IZ~ . Leucine aminopeptidase digestion: an aliquot of one HPLC frac­ These deleted clones were ligated and used to transform Eschencl/la tion from the LEP digestion, whose amino acid sequence began at coli HBI01. For sequencing in the reverse direction, the clone was residue 224, was dissolved in a solution that consisted of 200,u1 of first digested using Xba I, then treated with alpha phosphor­

0.01 M Tris/HCI, pH 8.8, 4 mM MgCI2 , and 0.6 U of leucine othioate, and finally digested with Bam HI. Exonuclease III diges­ aminopeptidase, which had been previously pre-incubated for 2 h at tion at 30°C was followed by ligation, transformation, and se­ 37"C. A sample containing all components except peptide substrate quencing as described above. served as a control. The reactions were incubated for 14 h at 37" C, Polymerase Chain Reaction Total RNA was extracted from then subjected to HPLC, exactly as described above. several preparations of dispersed skin mast cells of 11-13% pUrIt)' Protein sequencing was performed at the Biomolecular Resource in a single step using acid guanidinium and phenol-chloroform [271j Center at the University of California San Francisco. Fractions were Polyadenylated RNA was prepared using a kit (Pharmacia). Tota sequenced by automated Edman degradation on a model 470 A RNA (5 ,ug) or mRNA (1 ,ug) was reverse transcribed using ~V gas-phase protein sequencer (Applied Biosystems, Inc., Foster City, reverse transcriptase (BRL) and 5 pmol of primer 75 06 encodl?g CAl equipped with an on-line 120 PTH Analyzer. nucleotide 341 to 388 ofMC-CPA plus EcoR I and Hind III reStriC­ Cloning and Screening tion sites. The cDNA was diluted 1/50 in water. One and t.en microliters were amplified using 30 cycles of PCR in a Perkln­ Lung Mast Cell Library: A cDNA library from dispersed lung cells Elmer Cetus DNA thermal cycler using primer 7506 and an up­ containing 40% mast cells constructed in A ZAP (Stratagene) was stream primer (7505). Primer 7505 included MC-CPA se qu e n c~ used [24] . A 47-bp degenerate oligonucleotide (oligonucleotide extending from BP 3 to BP 30 coupled to 5 -Not I, Xho I, and Sal BS09) was synthesized based on residues 12 to 27 of human MC­ ' sites. PCR was initiated with 2.5 U T AQ polymerase, and run at an CPA [2]. The oli gonucleotide was end labeled with 32p y-ATP and 6 annealing temperature of 60°C and extension at 72°C. Both reaC­ T4 polynucleotide kinase and used to screen approximately 10 tions showed single bands at - 450 BP. The products were ethanol clones of the lung mast cell library [25]. precipitated, di gested by Hind III and xho I, and cloned into pla Dermal Cell Library A eDNA library was constructed as de­ mid bluescript KS. The clones were directly sequenced in both sc ribed from dispersed human dermal cells, containing 4 X 106 mast directions from nucleotides 31 to 327. ce ll s of 1 % purity, from a single donor [10]. The library was ampli­ Evolutionary Analysis Amino acid parsimony trees were built fied prior to sc reening using a protocol provided by the manufac­ from the carboxypeptidase sequences shown in Fig 4 by the com­ turer (Stratagene). 8 X 105 independent recombinants were ob­ puter program PROTOP ARS in the phylogenetic-analysis package tained prior to amplification. Approximately 2 X 106 plaques were PAUP [28]. All possible unrooted trees (135,135) for nine sequencdS screened. Briefly, plaque lifts were made using Nitroplus 2000 (l eaving out either mouse MC-CPA or rat MC-CPA) were teste· membranes. After denaturing and neutralization, filters were Those residues at the N- and C-termini that are missing for some treated with 0.5 mg/ml protease K in 2 X SSC, 0.1 % SDS, at 37"C sequences due to differing lengths of the proteins were coded as for 1 h to decrease bac kground prior to washing and prehybridiza­ llJ " unknown" for the purposes of tree building. The number of"w - tion. The cDNA (hL 4-3) obtained from the human lung mast cell ning sites" were counted for the four shortest trees (from the "tre)e library was biotinylated using the supplier's protocol (Blugene, steps" column under "charac ter diagnostics in the PAUP output , BRL). Hybridization was carried out with 0.1 ,ug/ml of probe at and a statistical test applied as described previously [11]. The treeS 42°C overnight in 45% formamide, 5 X SSC, 1 X Denhardt's so­ were rooted by "outgroup" using the Streptomyces griseus sequences. lution, 20 mM NaP04 (pH 6.5), 0.2 ,ug/ml herring sperm DNA. The filters were washed in 0.16 X SSC, 0.1 % SDS, for 15 min at RESULTS 50°C, then incubated with a streptavidin-alkaline phosphatase con­ M C-CPA and Bovine Pancreatic CPA Have Similar but Dis' jugate. The clones were visualized using NBT and BCIP, using the tinct Substrate Specificities The activities of purified hU!11~n supplier's protocol. MC-CPA and bovine pancreatic CPA were determined on a serIes cDNA Sequencing Clone hL 4-3 isolated from the human lung of synthetic tripeptide substrates, Cbz-Gly-Gly-X, where X wad mast cell library was sequenced directly in bluescript SK using dou­ Al a, Leu, Phe, Tyr, or Trp. The calculated kinetic constants Km an • ble-stranded di-deoxynucleotide chain termination sequencing [26] kc t are shown in T able I. MC-CP A and bovine CPA have silUllar, [Sequenase (USB)]. but not identical, substrate preferences. Specifically, MC-CPA d e ~: Nested deletions were made of the clone hS-I in bluescript SK onstrated no measurable activity toward the substrate with the slUal from the human dermal cell library using a kit (Erase-a-Base) ac­ es t C-terminal amino acid, Ala, as well as none against the s ub s ~ratf cording to the supplier's protocol. Briefly, the clone was digested with Trp, the largest. Bovine CPA demonstrated no hydrolYS IS 0

Table I. .Kinetic Constants for Substrate Hydrolysis by MC-CPA and Bovine CPA" ------~------~------MC-CPA Bovine CPA 5 Lt/V (10- 5) ------(10- ) _ n ..", k"./K"'_INC I Substrate kc" (second- I) Km (Ilm) second- 1M-I kc" (second- I) K.", (Ilm) seco n---~ Cbz-Gly-Gly-Ala NOb NO NO NO NO ND Cbz-Gly-Gly-Leu 21 ± 0.7 160 ± 14.5 1.3 52 ± 7.2 930 ± 260 0.56 Cbz-Gly-G ly-Phe 23.5 ± 1.6 26.5 ± 2.3 8.9 72 ± 5.9 98 ± 19 7.3 Cbz-Gly-Gly-Tyr 19 ± 0.4 27 ± 0.1 7.0 56 ± 2.0 102 ± 2.3 5.5 Cbz-Gly-G ly-Trp NO NO NO ~

• Purified MC-CPA and bovine CPA were incubated with four to five concentrations of each substrate, and the product Cbz-Gly-Gly was measured by HPLC as described 10 experimental procedures. Values are given as mean ± SE. ! NO, no hydrolys is was detected. , Hydrolysis was measurable but too low for determination of kinetic constants. VOL. 99, NO.2 AUGUST 1992 HUMAN MAST CELL CARBOXYPEPTIDASE FROM SKIN 141

100 Purified MC-CPA and bovine CPA were independently applied to Con A-Sepharose. Whereas 76% of purified MC-CPA was re­ covered in the eluate from Con A - Sepharose, only 3% of CPA was 80 recovered in the eluate, the rest of the activity appearing in the effluent and washes (Fig 1). This suggests that human MC-CPA,

60 but not bovine CPA, is glycosylated. When purified MC-CPA was 13 EFFLUENT treated with N-glycosidase F to remove N-linked polysaccharides, D WASil its apparent Mr on SOS-PAGE decreased by 3500 from 39,800 to 40 I!2J ELUATE 36,300 (n = 2) (Fig 2a). MC-CPA was then immobilized on nitro­ cellulose and incubated with lentil lectin. The lectin binding of MC-CPA treated with N-glycosidase F was markedly reduced 20 compared to control enzyme (Fig 2b) . T aken together, these find­ ings suggest that human MC-CPA is glycosylated in vivo. The Sequence of Skin MC-CPA Reveals It Is Identical to Lung MC-CPA Filter hybridization screening of the lung mast call cONA library using oligonucleotide BS09 yielded one clone ENZYME with 500-bp insert (hL 4-3). Partial sequencing of this clone and ~1ure 1. Binding of purified human skin MC-CPA and bovine pancreatic comparison of its deduced amino acid sequence to those of CPA and Sto Concanavalin A-Sepharose. The percent of starting activity recov­ CPB in the Dayhoff protein sequence data base, using the computer ere In effluent, wash, and elute fractions is shown. algorithm dfastp, revealed that hL 4-3 encoded part of the signal peptide, the activation peptide, and the N-terminal portion of a carboxypeptidase-like enzyme to amino acid 48 (numbering in Fig 3). The deduced amino acid sequence of the clone was identical to C~z-GlY-GlY_Ala, but did hydrolyze Cbz-Gly-Gly-Trp. kat and K.n that of the N-terminal amino acid sequence of MC-CPA obtained ~b Ues Were measurable for three substrates, Cbz-Gly-Gly-Len, previously (2). This cDNA was used to screen the dermal cell li­ Cpz-GlY-Gly-Phe, and Cbz-Gly-Gly-Tyr. The kc>t values ofMC­ brary. Screening yielded one cDNA clone of approximately 1.6-Kb I A Were similar for all three substrates, and were two- to threefold (hS-I). This clone was sequenced fully in both directions. Compari­ ~;er than those of bovine CPA. However, the Km values of MC­ son of the nucleotide sequence of hS-I with that of other carboxy­ GI A Were four- to sixfold lower for Cbz-Gly-Gly-Phe, Cbz-Gly- peptidases showed the sequence extended from part of the signal Tyr t , and Cbz-Gly-Gly-Leu compared to CPA. The kcaJK.n peptide through the 3' untranslated region (not shown). Compari­ vhues of MC-CPA and CPA were similar for these substrates. AI­ son of the deduced protein sequence of hS-I with sequences for the \ ough Cbz-Gly-Gly-Leu was the most poorly hydrolyzed of these other carboxypeptidases revealed a likely deletion of nucleotides ~ree substrates by both enzymes, the higher catalytic efficiency of coding for amino acids 78 through 91 (i.e., bp 232-273 in Fig 3). dC-CPA compared to bovine CPA with this substrate was largely Therefore, a PCR fragment extending from bp 31 to 327 was ob­ Ue to lower Km values, perhaps reflecting tighter binding. tained using messenger and total RNA; the sequence revealed the ~~'CPA Is a Glycoprotein To determine whether MC-CPA additional nucleotides present in the mRNA population, suggesting sr. ycosylated, we examined the binding ofMC-CPA to Concana­ the deletion was a cloning artifact. Figure 3 shows the composite ~.~ In A-Sepharose, an immobilized lectin that binds polysaccha­ cDNA sequence encoding mature MC-CPA from human skin mast / es. We also hydrolyzed asparagine {N)-linked polysaccharides cells and its deduced protein sequence. The deduced amino acid torn MC-CPA using N-glycosidase F, and investigated the subse­ sequence was compared to amino acid sequence obtained from pro­ ~.~nt change in mobility of MC-CPA on SOS-PAGE and in its tein sequencing of approximately 30% of the enzyme. Where pro­ a t tty to bind lentil lectin. tein sequence data was available, the amino acids are underlined in

A B c o

94K A B 67K

43K

30K

20.1K

b Bigu 1v!C-~ep2. aJ Denaturing gel electrophoresis of glycosidase-treated MC-CPA. Lane B: MC-CPA treated overnight without glycosidase treatment. Lane C: 36,300 t treated with glycosidase; they are flanked by molecular weight markers (lan es A and D). The calculated Mr of the band in [atle B is 39,800 versus (tnagn'/i Or Ialle C. b) Nitrocellulose dot blot of MC-CPA incubated with lentil lectin. A) No glycanase treatment; B) with N-glycosidase F treatment 1 catlonX15). 142 NATSUAKI ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

1 : ATe CCA GGC AGG CAe AGe 'lAC GC.A AAA TAC AAT AAT TGG GAA AAG 10 20 30 40 50 60 I G R H S YAK Y NNW E K 15 Bovine CPAl ARSTNTfN't!TYHTLOEI't~fHDLLVAEH!QLVSKLQIG!S't£GRPl'tV~KfSTG ___ GSNR!AIWIO 46 : ATT GTG GeT TGG ACT GM AAG ATG ATG GAT AAG TAT ceT GAA ATG Rat CPAl ALS TOSfN't AT'tHT Lo£I 't EfMoLLVAEHpQLVSKIQIGNTf£GRp 1 HVLK fSTG- --GTNRPA I WID V A WorK M M D K Y P E M 30 Rat CPA2 111 GGNfNf£A YHT LEE IYOEMDNLV AEN PGLVSKVNLGSSfEN RPMNVLK fSTG-- __ GoKPAIWLD Rat MC-CPA 111 AGRHSYAKYNoWNK I VSWTEKHVEKH PEMVSRI KIGSTVEDNP LYVLK I GRK-- -OGERKAI fMD Mouso MC-CPA 11 t AGRHSYAK'tNoWoK IVSWT£KH LEKHpEMVSR I KIGSTVEDNP LYVLK I GKK-- - DGERKAI rMD 91 : GTe TCT CGT ATT AAA ATT eGA 'lCT ACT GTT GAA GAT AAT CCA eTA Human HC- CPA 11 I PGRHSYAK'tNNW£K IVAWTEKHMoK'tP£MVSRI K IGS TVEoNP L YVLK I GEK---N ERRKA I r~ v S R K I G S V E D N P L 45 Rat CPB 111 ASGHSYTK'tNKWET I EAW I QOVATDNPoLVTOSVIGTTF'EGRNHYVLK I GKT---RPNKPA 1 F' Bovino CPB 111TTGHS'tEK'tNNW £TIEAWT£QVAS£NpDLISRSAIGTTfLCNTIYLLKVGKP___ GSNKP A Vf~ 136: TAT GTT CTG AAG ATT GGG GAA AAG A.A.T GAA AGA AGA AAG GCT ATT crayfish CPB 11111 1HoWTS 'tHoyo£INAWLoS LATo'tpELASV£oVG LSYEGRT HK LLKLCKG- --GAoKp I I f I YVLK G KNERRICAI 60 S. grisElus CP 111 oFpPAoSR'tHNYAEMNAA I DAR IAANPS I MSKRVI GKTYCX;RoV I AVKVSoNVAAOEA£PEVLf1'

70 BO 90 100 110 120 130 181: TTT ATG GAT TGT GGC ATT CAC GCA CGA GAA TGG GTC TCC CCA GCA I I I I I I I FHDCG oAR WVS A 75 Bovine CPAl LG I HSREWI TQATGVWf AKKfT£N'tGONPSfT A I LoSHo I fLE I VTNPNG- -f AfTHSENRLWRKTRS Rat CPAl TGIHSR£WVTOASGVWfAKKITKoYGOoPTfTAVLD NM olfLEIVTNPOG --fAYTHKTNRMWRKT~ 226: TTe TGC CAG TGG TTT GTe TAT CAG GCA Ace AAA ACT TAT GGG AGA Rat CPA2 AG I HAREWVtOAT ALWT ANKIASo'tGTDP A I TS LLNT LD If LLPVTNPOG--'tVfSOTTNRMWRKT F C Q W F V Y 0 A K T Y G R 90 Rat MC-CPA CG I HAR£WVSPAF'COWfWOAAKS'tGKNN IHTK LLDRMNfYV LPVfNVOG--'t I WSWTKDRKWRKNflS MOUSEl MC-CPA CGI IlAREWI SPAF'COWfVYQAT KS'tGKNKIMTK LLoRMNf'tV LpVfNVDG--'t I WSWTODRMWRKN~ 271: Me AAA ATT ATG Ace AAA eTC TTG GAC eGA ATG A.A.T TTT 'lAC ATT Human HC-CPA CGI HAREWVSP AfCQWfWOATKTYGRN K I MTKLLoRMNf't t LPVfNVOG--'t IWSWTKNRMWRKN flS N It I H T K LOR M N FYI 105 Rat CPB CGf'HAREWI SPAfCQWfVREAVRT'tNO£ I HMKOLLD£LDfYVLPVVN I OG--YV'tTWTKoRMWRKT Bovine CPB CGf'HAREW ISPAfCOWf'VREAVRTYGRE I HHT£fLDKLof'YVLPVVN I OG--Y I YTWTTNRMWRKT: GGI HAREWI APSTVTY I VNEfVSN---SATYoD I LSNVNfYVMPT I NPOG--YA 'tTfToDRLWRKT 316 : CTT ceT GTG TTe A.A.T GTT GAT GGA TAT ATT TGG TCA TGG ACA AAG Crayfish CPB S. g r iseus CP AHOHAREHLTVEMALYLLRELGOG'tGSoSRITOAVNGRELWIVPOHNPOGGEYDIASGS'tRSWRKNRO L VFNVDGY wsw K 120

l~O .. 1~0 1~0 1 ~o 1~0 1~0 361: Me CGC ATG TGG AGA AAAI AAT CGT Teel AAG Me CAA Me Tee AAA N R M W R K. N R S. K N Q N S K 135 Bovine CPAl VTSSSL-CVGVoANRNWoAGfGK-AGASSSPCSETYHGKYANSEVEVKSIVDfVKNH __ ___ GNf~ Rat CPAl HTOGSL- CVGVOPNRNWDAGfGH-AGASSNPCS£TYRGKFPNSEVEVKSIVofVTSH _____GNlrv> KRSGSG- CVGVDPNRNWDANfGG-pG ASSS PCSoS'tHGPKPNSEVEVKSIVofIKSH _____ GKVItA 406 AiC A;T Ra t CPA2 T~ ~ G~C C~C~~T T;T[~T ~T T~[T~ ~C 150 Rat MC-CPA KNPSST - CIGTDLNR Nf DVS Wo S -SPN ToNPCLSV 'tRGPAP£ S£KETK AVTNflRSHL----NSI~ House HC-CPA 451 : Tee ATT ceT Me Ace A.A.T GAC CCA TGT GCA GAT M e TAT eGG GGC iluman MC-CPA =~~~~!=~~~~~~~::~~~~:~=~::~~~:~~~~~:~~::~~~~~~~~~~~i~~ ~=::=:~~KV Rat CPB TMAGSS-CLGVRpNRNfNAGWCE-VGASRSPCS£TYCGPAPESEKETKALADFIRNNL ____ STlltA 51 NTND CADNYRG 165 Bovine CPB TRAGSS-CTGToLNRNf'oAGWCS-IGASNNPCSET'tCGSAA£SEK£SKAV"'DFIRNHL____ SSI~ Crayfish CPB ETGSVLGCKGAoPNRNWSfHWD£- VGASoSPCSoI'tAGP£PFS£V£MRNVRDOlLEYA ____ ANII( 496 : TCT GCA CCA GAG Tee GAG AAA GAG ACG AAA GeT GTe ACT AAT TTe S. griseus CP PNAGSS-AVGTDLNRNWAYKWGCCGG SSSSPSS£TYRGAAAESAP£TKVV ... ofVRSRWGGKOOJTA SAP ESE K T K A V T N F 180 200 210 220 230 240 250 541: ATT AGA AGe CAe CTG A.A.T GAA ATe A.AG GTT TAC ACT ACC TTC CAT I 1 I I I·· •••• RS LNE KVYTTF 195 Bovine CPAl F' LS I HS't SQLLL't p'tG'tTTQS I PoKTE LNOVAKSAV AALKSL 'tCTS 'tKYGS I I TTl YQASC;C;S I OW Rat CPAl F' I 5 I HS't SOLLLYP'tGYTSEPAPOQA£LoOLAKSAVTA L TS LHGTKFK'tGS J lOTI YOASCSTI OW 586 : TCC TAC TCC CAG ATG CTA TTG TTT CCC TAT GGA TAT ACA TCA AAA Rat CP A2 F'I T LIlS't SOLLHFPYG't KCTKPDDF'NELOEVAOKAAOA LKRLHGTSY KVGP I CSV I YOASGGS I pW Rat MC-CPA 't I TfHS't SQMLLfPYGYTI K LP PNHOoLLKVARI ATDVLSSR't£TRY I 'tCP lAST I YKTSCS SLD" SYSQHLL YGYTSK 210 Houso MC-CPA 't I TfHSYSOMLLI P'tGYTfKLPpNHOoLLKVARI ATDA LSTRY£TRY I YGP J AST I YKTSGSSLD" Human MC-CPA Y I TFHS'tSQMLLfP'tGYTSK LPPNH EDLAKVAK IGToV LSTRY£TRY I 'tGP I EST I YP ISCSSLDN 631: CTG CCA CCT Me CAT GAG GAC TTG GeC AAA GTT GCA AAG ATT GGC Rat CP8 Y L T1 HSYSQMML'tP YSYO't KLP£NYE£LNALVKGAAKELA T LHGTKYTYGPGATTI YPMGGSDOIf LPN EDLAKVAKIG 225 Bovino CPB 'tL T1 HSY SQMHL't P'tS'lo'lKLp KNNV£LNTLAKGAVKK LAS LHGTTYS'tG PGATT 1 YPASGGSoo~ Crayfish CPB YLTfHS'tSOLWM'tpWGf'TSoLpooWQDLoTLATNAVoALTAVHGTRY£IGSSTNTIYMAGGSoD .. 676: ACT GAT GTT eTA TCA ACT CGA TAT GAA ACC CGA TAC ATe TAT GGC S . grlseus CP AlOfHTYS£LVLWPfGYTYNDTAPGHTAoDRNAFAAVGOKHAASNGYTA£-OSSoLYIToCSIODN DVL TRY RYIYG 240 260 270 280 290 300 721; I • I I 1 I 255 Bovine CPAl S'lNQC- I K't SfTf ELRDTGRY --GfL LPASQI J pT AQETW LGVL T I MEHTVNN 1 1 1" 11" 117 77 Rat CPAl T'l SQG - I KYSfTfELRDTGLR--GfLLPASQIl pTAE£TW LA. LLT 1 MDHTVKHPY? 1 7 7 1 1 1 7 11 Rat CPA2 A YDLG- I KYSf AfELRoTAf"t - -GfLLPAKOJ LPT A£ETW LG LKT I MEHVRoHP'l.,,,,, 7 1 11 1 77 766: TGG GeT TAT GAC C'PG GGC ATC A.AA CAC ACA TTT~ TTT GAG CTC Rat HC-CPA A 'tDLC- I KHTfAFELRDKGKS--GfLLP £SR I KPTCKETMLSVKF' lAKY 1 LKHTS., 1 11111111 1 WAY D L G I K H T F ~ F E L 270 House MC-CPA VYDLC- I KHTfAf£LRDKGKS--GfLLP£SR I KPTCK£TM LSVKF I AKY I LKNTS 1 1 1 7 1 1 7 77 Human MC-CPA A'lDLG- I KHTf AfE LRoKGKf--GF'LLPESR I KPTCRETHLAVKf lAKY I LKHTS 1 1 1 7 711 117 8 11 : CGA GAT AM GGC AM TTT GGT TTT CTC CTT CCA GAA TCC CGG AT A Rat CPB S 'tDQG- I K'tSF'TfE LRoTGff--Gf' LLP ESOI ROTCEETMLAVK'l t ANYVREH LY1 1 17 7 17 7 11 R 0 It G K G F L L R 285 Bovine CPB A 'tDQG- I KYSfTf£LRDKGR 't --GfV LP ESOI OPTCEETH LA I KYVTS 'lVLE H L 7 1 7 1" 1 17 7 77 Crayfish CPB AKGEGGVKYA 'tTl £LRDTGN't --GF LLP ENOl I PTGEETfF.GVKVVANfVKoTYS? 1? 1 1 7 7171 856: AAG CCA ACG TGC AGA GAG ACC ATG CTA GCT GTC AM TTT ATT GCC S. gr isous CP LWGSQ-K1 FG'tTfEM'lP RSASGGGf'tP I.' DEV I ERETSRNRDAVI.OL 1 EN A. DCH't RS I GKEAOYCS K TCRE HLAVKFIA 300

901 : AAG TAT ATC CTC AAG CAT ACT TCC TAA Figure 4. Sequences of carboxypeptidases used to construct evolutionary " y K H S stop trees, aligned with regard to the primary sequence for bovine CPA. Dashes indicate gaps in the sequences required for ali gnment, and thus represelJ1 Figure 3. Composite cDN A sequence and deduced protein sequence of the amino acid deletions. Asterisks highlight amino acid residues localized to [he catalytic portion of MC-CPA. The amino acid sequences determined by 51' subsite of bovine CPA [11]. Question marks indicate differences in sed direct protein sequencing are lIllderlilled. Consensus N-linked glycosylation quence due to length variation at the N- and C- termini of the enzymes. an. sites are boxed. Amino acids corresponding to residues in the 51' subsite of code these positions as "uncertain" for the purposes of parsimony analysl~ bovine pancreatic CPA are circled. [28]. Sequences for bovine CPA1 [44,55,56]; rat CPA1 [57]; rat CPA2 (1J . rat MC-CPA [13]; mouse MC-CPA [12]; human MC-CPA [9] (this stu ~ Fig 3); rat ePB [57]; bovine CPB [45]; crayfish ePB [58], and S. grisellS C [59]. Fig 3. At all positions the deduced and directly obtained amino acid sequences agreed. The sequence of the skin MC-CPA cDNA is identical to that reported for human lung MC-CPA cDNA [9]. surface of the molecule (not shown), and therefore is potentiallY Analysis of DNA and Protein Sequence Comparison of the available for glycosylation [33]. The other site is not located 011 the sequence of bovine pancreatic CPA with that ofMC-CPA indicates surface of bovine CPA, and may not be glycosylated. four replacements of amino acid residues that have been localized by X-ray crystallography to the Sl' binding site of pancreatic carboxy­ Evolutionary Tree Analysis The most parsimonious evo1u: peptidase A [11,15,29]. These replacements in MC-CPA occur at tionary tree (i.e., the one that requires the fewest amino acid rep~ace, positions lie 249, Ser 252, and Ala 267 (Fig 3), which correspond to ments and gene duplications to explain the modern sequence d1ve\ positions Ala 250, Gly 253, Thr 268, respectively, of bovine CPA sity and known distribution) is shown in Fig Sa. This tree re9Ulreo (Fig 4), in addition to the previously discussed reflacement of Leu 1054 replacements. It places the divergence of the lineage lead!l1g t., 254 in MC-CPA for lie 255 of bovine CPA [9,12. Thus MC-CPA crayfish CPB and lineage leading to the mammalian carboxypeptl, has one isomeric, one smaller, and two larger amino acids in the SI' dases prior to the gene duplications that gave rise to the four !11a~t site compared to bovine CPA. malian genes, CPA 1, CPA2, CPB, and MC-CPA. The next shortte The presence of two consensus sequences for N-linked glycosy­ trees (Fig 5b,c) each require 1062 amino acid replacements. 1" ,t !ation sites [30-32] at Asn 127-Arg 128-Ser 129, and Asn 146- shortest tree was not statistically significantly better than these ne~e Ala 147 -Ser 148 (Fig 3) is consistent with the observation that two shortest trees. These trees place crayfish CPB either on tl e MC-CPA is glycosylated in vivo. Using the computer program lineage leading to MC-CPA and CPB (Fig 5bjtree 2), or 011 ~;g INSIGHT II (Biosym Technologies, Inc.), the tertiary structure of lineage leading to the pancreatic CPAI and CPA2 enzymeS (I a[ bovine CPA was displayed, and we determined that Thr 148 (the 5cjtree 3), Importantly for this study, all three trees docum~l1t~~J3 amino acid corresponding to MC-CPA Asn 146) is located on the MC-CPA arose from a gene duplication along the pancreatiC VOL. 99, NO.2 AUGUST 1992 HUMAN MAST CELL CARBOXYPEPTIDASE FROM SKIN 143

a. Tree 1 Rat MCCPA quences and substrate specificity, using the substrate specificity of bovine pancreatic CPA as a reference [11]. It was proposed that the Human MCCPA substrate preferences of these enzymes may be correlated with the size of the Sl' , because the substitution of a smaller resi­ Rat CPB due, Ala, for Thr268 in rat CPA 2 (Fig 4) correlated with a change Bovine CPB in specificity toward higher-molecular-weight hydrophobics such as tryptophan, and the substitution of two bulkier residues, Ser for Rat CPA1 Gly 253, and Thr for Ser 254, in rat CPAI correlated with prefer­ ences skewed away from larger substrates. We were interested to see Bovine CPA if this hypothesis predicted the behavior of human MC-CPA, in which Ile 247, Ser 252, Leu 254, and Ala 267 of MC-CPA (Fig 3) Rat CPA2 replace Ala 250, Gly 253, Ile 255, and Thr 268, respectively, of Crayfish CPB bovine CPA (marked by asterisks in Fig 4) . The Sl' subsite replacements of two bulkier, one isomeric, as well S. griseus CP as one smaller residue in MC-CPA, represent a more complicated model of the S I' subsite than described for rat CPA 1 and CPA2, which have only bulkier or smaller amino acid substitutions, respec­ tively. Relative to bovine CPA, these S1' subsite replacements in b. Tree 2 c. Tree 3 Rat MCCPA Rat MCCPA MC-CPA result in a greater net positive change in non-hydrogen Human MCCPA Human MCCPA atoms and in volume in A3 [38] than the replacements in rat CPA1. Rat CPB Rat CPB Such changes might be expected to be associated with greater alter­ Bovine CPS Bovine CPS ations in specificity, similar to that exhibited by rat CPA1, than are ----- Crayllsh CPB Rat CPA1 Rat CPA1 Bovine CPA actually observed between MC-CPA and bovine CPA (Table I). Bovine CPA Rat CPA2 Like rat CPAl [11], MC-CPA did not hydrolyze carboxyl-terminal Rat CPA2 1_____ Crayfish CPB Trp, but unlike rat CPA1, MC-CPA hydrolyzed carboxy-terminal '-______S. griseus CP ------S. griseus CP Tyr with similar overall catalytic efficiency compared to Phe (kc.J Km in Table I), and did not hydrolyze carboxyl-terminal Ala. There­ Figure 5 G I tio . enea ogy of the carboxypeptidase gene family. a) The evolu- fore, it is apparent that the presumed overall size of the S l' pocket as rntaYbtree depicted here is the most parsimonious evolutionary tree deter­ predicted by amino acid replacements, may not be the sole determi­ rn ne y ammo aCid parsimony analysis (requiring 1054 amino acid replace­ s nant of the C-terminal amino acid substrate preferences of these (e;h ), b) and c) show the next two shortest trees for these nine sequences enzymes. It is reasonable to speculate that backbone differences the\ reqUiring 1062 amino acid replacements). These trees are rooted using between MC-CPA and bovine CPA will be larger than those be­ Cati actenal carboxypeptidase as an outgroup. Triallgles indicate gene dupli­ tween rat CPA 1 and bovine CPA, in order to compensate for these ber °f~ an~ the other divergences represent speciations; the minimum num­ of th upltcatlons that is required to explain the modern species distribution replacements, and still exhibit similar substrate preferences. This esc proteins is shown. would also be consistent with the genealogy of MC-CP A, as larger backbone differences are expected between products of ar,cient gene duplications (i.e., MC-CPA and bovine pancreatic CPA) than be­ linea h tween products of orthologous genes (i.e., rat CPAl and bovine MC-~ rat er than along the pancreatic CPA lineage. When mouse pancreatic CPA). It is also likely that the position of the amino acid sitnil PA [12] was used (rather than rat MC-CPA) to build trees, replacements in the S l' active site may also be critical, because this ar results were found. determines the shape of the pocket. For instance, in rat CPA1 the substitution ofThr for Ser 254 is in t~e base of the pocket, near the DISCUSSION hydroxyl group of glycyltyrosine, and the position of this group Mast c 1I d hav e b e proteases are useful markers for human mast cells, an may account for the fivefold lower hydrolytic activity of rat CPAI tase een used to phenotype them. MCTC mast cells contain tryp­ toward Cbz-Gly-Tyr, compared to Cbz-Gly-Phe [11]. Such a re­ rea' chymase [34], carboxypeptidase [35], and a proteinase cross- placement does not occur in MC-CPA. o cellst~~ With antibodies against cathepsin G [7], whereas MCT mast Our data demonstrate that MC-CPA is probably glycosylated in cel! ntam only tryptase [34]. The content of MC-CPA in mast vivo. Although the shift in apparent Mr on SDS-PAGE ofN-glyco­ lun s obtamed from skin appears to be ten- to fortyfold higher than sidase F treated MC-CPA (Fig 2a) was relatively small (-3500), it of ~~ast cells [36], presumably in part because of a higher percent was almost identical in magnitude to that reported for similarly o Tc mast cells in skin compared to lung (> 99% versus 7%). treated human and dog mast cell and [16,17,39]. focu ur analysis of the amino acid sequence of human MC-CPA It is interesting to note that rat [13] and mouse MC-CPA [12] lack in p.ssed on those residues in the SI' subsite (marked by an asterisk the consensus site most likely to be glycosylated [33], replacing Asp enc~g 4} that may be important in determining the substrate prefer­ for Asn 147 (Fig 4). Although some rodent mast cell neutral pro­ cryst:l~ the carboxypeptidases with A-like activity. Because the teases are not glycosylated [40,41]' purified rat and mouse MC-CPA ditn .tructure of MC-CPA has not been reported, only the three­ have not yet been so characterized. Pancreatic proteases are not rece~nslonal structure of bovine pancreatic CPA [15,29], and the glycosylated [30,42,43]. Further, bovine CPA and CPB lack both of B (p t ~port of the crystal structure of porcine procarboxypeptidase the N-linked glycosylation sites identified in MC-CPA: Thr 129 cel! ro PE) [37], may be used to model the SI' subsite in the mast destroys one site, and Asp 148 replaces Asn in the second [44,45], as sitnilcnzyme. The main chain conformation of this site is quite shown in Fig 4. The function of glycosylation of mast cell proteases COtn ar m bovine CPA and porcine proCPB, and the amino acids is unknown, although it has been postulated that glycosylation may pnsmgth' . II d ' I db' atic car b IS SIte are we conserve m t 1e rat an ovme pancre- protect dog mast cell chymase from autodegradation, and may playa graph Oxypeptidases (Fig 4). Therefore, pending future crystallo­ role in intercellular processing and packaging in the mast cell secre­ subs it Ie S?uctural data ofMC-CPA, it is reasonable to model the S1' tory granule [39]. In this light, it is interesting that purified rat detailedo MC-CPA on bovine CPA, for which we have the most MC-CPA appears to be less stable than its human homolog [13,46]. residue itruc~ure, and to examine the replacements among the nine es Pancreatic and mass cell carboxypeptidases are members of the data w. hocaltzed to this subsite [11], in order to correlate sequence same gene family and arose through gene-duplication events l'w It substrate preferences. [2,9,11-13]. Of these enzymes, MC-CPA exhibit a unique rela­ and r 0 pancreatic carboxypeptidases with A-like activity, rat CPA1 tionship between amino acid sequence and substrate specificity: at CPA2, have been studied in order to correlate their se- they are more similar in overall sequences to the pancreatic CPB 144 NATSUAKI ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

than to the CPA [2,9,12,13)' yet they have CPA-like substrate speci­ Rutter WJ: A novel rat carboxypeptidase, CPA2: characterization. fiCIty [1,2,46,47). Our study of MC-CPA has provided further in­ molecular cloning. and evolutionary implications on substrate specI - sight into the evolutionary divergence within this family. The ficiry in the carboxypeptidase gene family. J BioI Chern gene-duplication event that gave rise to MC-CPA appears to have 263: 17828 -17836. 1988 occurred along the vertebrate CPB lineage subsequent to its diver­ 12. Reynolds DS. Stevens RL. Gurley DS, Lane WS, Austen KF. Serafin gence from the lineage leading to CPAI and CPA2 (Fig 5). This WE: Isolation and molecular cloning of mast cell carboxypeptid~se may help explain why the residue considered critical for conferring A: a novel member of the carboxypeptidase gene family. J BioI Chern 264:20094-20099.1989 CPA-ltke substrate specificity to the pancreatic CPA (i.e., lie 255 in bovine CPA) [15,45,48) is Leu in human, mouse, and rat MC-CPA. 13. Cole KR. Kumur S, Le Trong HL, Woodbury RG, Walsh KA. Neur­ ath H: Rat mast cell carboxypeptidase: amino acid sequence and CPA-like specificity appears to have been invented independently evidence of enzyme activiry within mast cell granules. Biochem by MC-CPA, a gene duplicate along the CPB lineage. Because this 30:648 - 655. 1991 duplication event apparently occurred before later speciation diver­ 14. Schechter I. Berger A: On the size of the active site in proteases J gence of CPB, it is not surprising that MC-CPA is found in other Papain. Biochem Biophys Res Commun 27:157 -162.1967 mammals. Although mast cells are found in lower vertebrates, such 15. Lipscomb W, HartsuckJ. Reeke G, Quiocho F.Bethage P, LudwigM. as carp (Cyprinus carpio), leopard frogs (Rana pidiens), and turtles Steitz T. Muirhead H. Cuppola J: The structure of carboxypeptida.se (Chrysel'llyn picta), these species have not been investigated for the A. VII. Th 20-angstrom resolution studies of the enzyme and of ItS presence of MC-CPA [49); thus we do not know how early in complex with glycylryrosine. and mechanistic deductions. Brook­ vertebrate evolution MC-CPA arose. haven Symp BioI 21:24-90. 1968 In the mouse and rat there is substantial evidence for the existence 16. CromlishJA. Seidah NG. Marcinkiewicz M. HamelinJ,Johnson DA. of multiple chymase genes that are differentially expressed in differ­ Chretien M: Human pituitary tryptase: molecular forms. NHz-ter­ ent mast cell types [8,40,41,50-53). Our previous work demon­ minal sequence. immunocytochemical localization, and specifiClty strated multiple human mast cell tryptase cDNA [10,54), some of with prohormone and f1uorogenic substrates. J BioI chem which may be differentially expressed by mast cells in different 262:1363-1373. 1987 human tissues [15). In contrast, the current evidence is consistent 17. Schechter NM: Human chymase. Monogr Allergy 27:114-131.1990 with the view that there is a single gene for Me-CPA in human or 18. Caughey GH, Raymond WW, Vanderslice P: Dog mast cell chyrnase: mouse [9,12). molecular cloning and characterization. Biochemistry 29:5166- 5171,1990 19. Simpson RT. Riordan JF, Vallee BL: Functional ryrosyl residues in the Wegrattju l/y acktlOwledge the advice ofDrs. jeffEd mal" Danielle Melloul, and Meg active center of bovine pancreatic carboxypeptidase A. Biochemistry 2:616-622.1963 Phil/ips; Dr. Betsy Goldsmith for critica l commwts; Bill Raymond for studies using es glycosidase; jane Leong alld Carol Ka empfer Jor excel/flit technical assistance; jim 20. Hass GM. Ryan CA: Carboxypeptidase inhibitor from potato .­ Methods Enzymol 80:778-791.1981 Kealey for performing the protein sequwcing; Patricia Roloiffor preparing the mam.­ script; and Drs. Wil/iam WolJendon and Peter POllagotacos for contributing the skill 21. Klickstein LB. Wintroub BU: Separation of antiotensins and assay of angiotensin-generating enzymes by high-performance liquid chro' from which mRNA or purified MC-CPA was obtailled. matography. Anal Biochem 120:146-150. 1982 22. Cleland WW: Statistical analysis of enzyme kinetic data. Methods REFERENCES EnzymoI 63:103-138.1979 1. Goldstein SM, Kaempfer CE, Proud D, Schwartz LB, Irani AM, Win­ 23. Masaki T. Tanabe M. Nakamura K, Soejima M: Studies of a ne'v troub BU: Detection and partial characterization of a human mast proteolytic enzyme from achromobacter lyticus M497-1.1. purifica­ cell carboxypeptidase. J Immunol 139:2724 - 2729, 1987 tion and some enzymatic properties. Biochim Biophys Acta 660:44-50.1981 2. Goldstein SM, Kaempfer CE, Kealy JP, Wintroub BU: Human mast 24. Miller JS. Westin EH, Schwartz LB: Cloning and characterization of ce ll carboxypeptidase. Purification and characterization. J Clin In­ 88 vest 83: 1630 - 1636, 1989 complementary DNA for human tryptase. J Clin Invest 84: 11 - 1195.1989 3. Schwartz LB, Lewis RA, Austen KF: Tryptase from human lung mast Maniatis T. Fritsch FF. SambrookJ: In Molecular Cloning: A Labora­ ce ll s by heparin: stabilization of the active tetramer. J BioI Chem 25. tory Manual. 1st ed. Cold Spring Harbor Lab, Cold Spring Harbor. 256:11939- 11943, 1981 NY. 1982 4. Schechter NM, Fraki JE, Geesin JC, Lazarus GS: Human skin chymo­ ter tryptic proteinase. J BioI Chern 258:2973 - 2978, 1983 26. San~er F, N.icklen S, Coulson AR: DNA sequencing with chain- ; mll1atmg 1I1111bltors. Proc Nat! Acad Sci USA 74:5463-5467. 197 5. Schechter NM, Choi JK,Slavin DA, Dereskienski DT, Sayama S, Dong G, Lavker RM, Proud D, Lazarus GS: Identification of a 27. Chomczynski P~ Sacchi N: Single-step method of RN A extraction bJ chymostrypsin-like proteinase in human mast cell s. J Immunol aCld-gual1ldmlUm thIOcyanate-phenol-chloroform extraction. An 137:962-970, 1986 Biochem 162: 156 -159. 1987 Swofford DL: PAUP: Phylogenetic Analysis Using Parsimony. ver­ 6. Wintroub BU, Kaempfer CE, Schechter NM, Proud DA: Human lung 28 . sion 30. Computer program, distributed by the Illinois Natural mast cell chymostrypsin-like enzyme: identification and partial char­ HIS' tory Survey. Champaign. II. 1991 acterization. J Clin Invest 77:196-201,1986 29. Christianson DW. Lipscomb WN: X-ray crystallographic investiga: 7. Schechter NM. Irani AM. Sprows JL. Abernethy J. Wintroub B. tion of substrate binding to carboxypeptidase A at subzero tempera Schwartz LB : Identification of a cathepsin G-like proteinase in the ture. Proc Nat! Acad Sci USA 83:7568 - 7572. 1986 MCTe type of human mast cel l. J Immunol 145:2652-2661. 1990 89 30. Eylar EH: On the biologic role of glycoproteins. J Theor BioI 10: - 8. Benfey PN. Yin FH. Leder P: Cloning of the mast cell protease. 113,1966 RMCP II. Evidence for cell specific expression and a multi-gene -de fami ly. J BioI Chern 265:5377 - 5384. 1987 31. Marshall RD: The nature and metabolism of the carbohydrate-peptl linkages of glycoproteins. Biochem Soc Symp 40: 17 - 26. 1974 9. Reynolds DS, Gurley DS. Stevens RL, Sugerbaker DJ, Austen KF, Serafin WE: Cloning of human mast ce ll carboxypeptidase A and 32. Struck DK. Lennarz WJ. Brew K: Primary structural requirementSfor comparison with mouse mast cell carboxypeptidase A and compari­ the enzymatic formation of the N -glycosidic bond in glycoprotelns- J BioI Chem'253:5786-5794. 1978 d son with mouse mast cel l carboxypeptidase A and rat pancreatic -Ilke carboxypeptidases. Proc Nat! Acad Sci USA 86:9484-9490.1989 33. Hu bb ar d SC, Ivatt RJ: Synthesis and processing of asparagine-lI 10. Vanderslice p. Ballinger SM, Tam EK. Goldstein SM, Craik CS. oligosaccharides. Annu Rev Biochem 50:555 - 583 1981 Caughey GH: Human mast ce ll tryptase: multiple cDNAs and genes 34. Irani AM. Schechter NM. Craig SS. DeBlois G. Sch:Vartz LB: T~~ reveal a multigene family. Proc Nat! Acad Sci USA rypes of human mast cells that have distinct neutral protease comP 87:381 1- 3815.1990 sitions. Proc Nat! Acad Sci USA 83:4464-4468,1986 , 11. Gardell SJ. Craik CS. Clauser E. Goldsmith EJ. Stewart C-B. GrafM, 35. Irani AA. Goldstein SM. Wintroub BU. Bradford T. SchwartZ 1)3- VOL. 99, NO.2 AUGUST 1992 HUMAN MAST CELL CARBOXYPEPTIDASE FROM SKIN 145

Human mast cell carboxypeptidase: selective localization to MCte 48. Schmid MF, Herriott ]R: Structure of carboxypeptidase B at 28-ang­ cells. ] ImmunoI147:247-253, 1991 strom resolution.] Mol BioI 103:175-190, 1976 36. Goldstein SM, Wintroub BU: Human mast cell carboxypeptidase. 49. Chiu H , Lagunoff D: Histochemical comparison of vertebrate mast Monogr Allergy 27:132-145,1990 cel ls. Histochem] 4:135-144,1972 37. Coli M, Guasch A, Aviles FX, Huber R: Three dimensional structure 50. Woodbury R, Gruzenski G, LagunoffD: Immunofluorescent localiza­ of porcine procarboxypeptidase B: a structural basis of it inactivity. tion of a serine protease in rat small intestine. Proc Nad Acad Sci EMBO] 10:1-9, 1991 USA 75:2785-2789,1978 38. Zamyatnin AA: Protein volume in so lution. Prog Biophys Mol BioI 51. Woodbury R, Neurad1 H: Structure, specificity and localization of the 24:107 -123, 1972 serine proteases of connective tissue. FEBS Lett 114: 189 - 196, 1980 39. Caughey GH: Tryptase and chymase in dog mast ce lls. Monogr Al­ 52. LeTrong H, Neurad1 H , Woodbury RG: Substrate specificity of the lergy 27:67 -89,1990 -like protease in secretory granules iso lated from rat 40. Woodbury RG, Katunuma N, Kobayashi K, Titani K, N eurath H, mast cells. Proc Nad Acad Sci USA 84:364-367, 1987 Anderson W, Matthews B: Covalent structure of a group-specific 53. Reynolds DS, Stevens RL, Lane WS, Carr MH, Austen KF, Serafin protease from rat small intestine. Biochemistry 17:811-819, 1978 WE: Different mouse mast cell populations express various combina­ 41. Le Trong H, Parmelee DC, Walsh KA, Neurad) H, Woodbury RG: tions of at least six distinct mast cell serine proteases. Proc Nad Acad Amino acid sequence of rat mast cell protease I (Chymase). Biochem­ Sci USA 87:3230-3234,1990 Istry 26:6988 - 6994, 1987 54. Miller ]S, Moxley G, Schwartz LB: Cloning and characterization of a 42. Tartakoff A, Greene L], Palade GE: Studies on the Guinea pig pan­ second complementary DNA for human tryptase. ] Immunol creas. ] BioI Chem 249:7420- 7431, 1974 130:2352-2356,1990 43. Struck DK, Lennarz WJ: The function of saccharide-lipids in synthesis 55. Bradshaw RA, Ericsson LH, Walsh KA, Neurath H: The amino acid o~ glycoproteins. In: Lennarz W] (ed.). The Biochemistry ofGlycopro­ sequence of bovine carboxypeptidase A. Proc Nat! Acad Sci USA telMS and Proleoglycalls. Plenum Press, New York and London, 1980, 63:1389-1394, 1969 p 37 56. Petra PH, Hermodson MA, Walsh KA, Neurad1 H : C haracterization 44. Bradshaw RA, Walsh KA, Neurath H: Amino acid sequence of bovine of bovine carboxypeptidase A (Allen). Biochemistry 10:4023- carboxypeptidase A. Isolation and characterization of selected peptic 4025,1971 and nagarse peptides and the complete sequence of fragment F-l. 57. Clauser E, Gardell Sj, Craik CS, MacDonald Rj, Rutter WJ: Structural Biochemistry 10:961 - 972, 1971 characterization of the rat carboxypeptidase A 1 and B genes. ] BioI 45. T' Itani K, Ericsson LH, Walsh KA, Neurath H: Amino-acid sequence Chem 263:17837 -17845, 1988 of bovine carboxypeptidase B. Proc Nad Acad Sci USA 72:1666- 58. Titani K, Ericsson LH, Kumar S, Jakob F, Neurath H , Zwilling R: 1670, 1975 Amino acid sequence of crayfish (Astacus fluviatilis) carboxypepti­ 46. E veritt .MT, Neurad, H : Rat peritoneal mast cell carboxypeptidase: dase B. Biochemistry 23:1245 -1250, 1984 locahzation, purification, and enzymatic properties. FEBS Lett 59. Narahashi Y: The amin o acid sequence of zinc-carboxypeptidase from 110:292-296,1980 Streptomyces griseus. ] Biochem 107:879-886, 1990 47 . S erafin WE, Dayton ET, Granallese PM, Austen KF, Stevens RL: Carboxypeptidase A in mouse mast ce ll s: identification, characteriza­ tion, and use as a differentiation marker. ] Immunol 139:3771- 3776,1987

ERRATA

Please note the following corrections to "Acute or chronic topical retinoic acid treatment of human skin in vivo alters the expression of epidermal transglutaminase, loricrin, involucrin, filaggrin, and keratins 6 and 13 but not keratins 1, 10, and 14," Volume 98, March 1992, pp 343-350. Figure 1 legend: The bar should be 1.1 inches in length, representing 0.1 mm. Figure 6 legend: The sentence "Right: RA-treated epidermis (X50)" should be deleted. Figure 7 legend: "Top row" should read "Left" and "Bottom row" should read "Right". The words "left columlt," "middle column," and "right column" should all be deleted.