Báo cáo khoa học: A novel metallocarboxypeptidase-like enzyme from the marine annelid Sabellastarte magnifica – a step into the invertebrate world of proteases pdf

The enzyme, which is a minor component of the molecularly complex animal body, as shown by 2D gel electrophoresis, has been purified from crude extracts to homogeneity by affinity chromato

A novel metallocarboxypeptidase-like enzyme from the marine annelid Sabellastarte magnifica – a step into the invertebrate world of proteases

Maday Alonso-del-Rivero1, Sebastian A Trejo3, Mo´nica Rodrı´guez de la Vega3, Yamile Gonza´lez1, Silvia Bronsoms3, Francesc Canals2, Julieta Delfı´n1, Joaquin Diaz1, Francesc X Aviles3 and Marı´a

A Cha´vez1

1 Centro de Estudio de Proteı´nas, Facultad de Biologı´a, Universidad de la Habana, Cuba

2 Institut de Recerca Hospital Vall d’Hebron, Barcelona, Spain

3 Institut de Biotecnologı´a i Biomedicina and Departament de Bioquı´mica i Biologı´a Molecular, Universitat Autonoma de Barcelona, Spain

Introduction

Natural evolution has frequently generated a large

adaptative variety of forms among protein functional

families, and metallocarboxypeptidases (MCPs) have also followed this trend Such enzymes are

exopeptid-Keywords

enzyme specificity; marine annelid;

metallocarboxypeptidases; metalloproteins;

Sabellastarte magnifica

Correspondence

F X Aviles, Institut de Biotecnologı´a i

Biomedicina (IBB) and Departament de

Bioquı´mica i Biologia Molecular, Universitat

Autonoma de Barcelona, 08193 Bellaterra

(Barcelona), Spain

Fax: +34 93 581 2011

Tel: +34 93 581 1231

E-mail: francescxavier.aviles@uab.es

(Received 16 March 2009, revised 16 June

2009, accepted 30 June 2009)

doi:10.1111/j.1742-4658.2009.07187.x

After screening 25 marine invertebrates, a novel metallocarboxypeptidase (SmCP) has been identified by activity and MS analytical approaches, and isolated from the marine annelid Sabellastarte magnifica The enzyme, which is a minor component of the molecularly complex animal body, as shown by 2D gel electrophoresis, has been purified from crude extracts to homogeneity by affinity chromatography on potato carboxypeptidase inhib-itor and by ion exchange chromatography SmCP is a protease of

33792 Da, displaying N-terminal and internal sequence homologies with M14 metallocarboxypeptidase-like enzymes, as determined by MS and auto-mated Edman degradation The enzyme contains one atom of Zn per mole-cule, is activated by Ca2+and is drastically inhibited by the metal chelator 1,10-phenanthroline, as well as by excess Zn2+or Cu2+, but moderately so

by EDTA SmCP is also strongly inhibited by specific inhibitors of metallo-carboxypeptidases, such as benzylsuccinic acid and the protein inhibitors found in potato and leech (i.e recombinant forms, both at nanomolar levels) The enzyme displays high peptidase efficiency towards pancreatic carboxypeptidase-A synthetic substrates, such as those with hydrophobic residues at the C-terminus but, remarkably, also towards the acidic ones This property, previously described as for carboxypeptidase O-like activity, has been shown on long peptide substrates by MS The results obtained in the present study indicate that SmCP is a novel member of the M14 metal-locarboxypeptidases family (assignable to the M14A or pancreatic-like subfamily) with a wider specificity that has not been described previously

Abbreviations

AAFP, N-(4-methoxyphenylazoformyl)- L -phenyl-alanine; AAFR, N-(4-methoxyphenylazoformyl)- L -Arg; ACTH fragment (18–39),

adrenocorticotropic hormone (RPVKVYPNGAEDESAEAFPLEF); BAEE, benzoyl arginyl ethyl ester; BTEE, benzoyl tyrosine ethyl ester; CP, carboxypeptidase; CPA, carboxypeptidase A; CPB, carboxypeptidase B; CPO, carboxypeptidase O; DIGE, difference gel electrophoresis; E-64, L -carboxy-trans-2,3-epoxypropyl-leycylamido (4-guanidino) butane; FAAK, [3-(2-furyl)acryloyl] - L-alanyl- L -lysine; FAPP,

N-(3-[2-furyl]acryloyl)-Phe-Phe; Hippuryl-Phe, N-benzoyl-Gly-Phe; MCP, metallocarboxypeptidase; rLCI, recombinant leech carboxypeptidase inhibitor; rPCI, recombinant potato carboxypeptidase inhibitor; V15E, synthetic substrate [VKKKARKAAGC(Amc)AWE].

ases that catalyze the hydrolysis of peptide bonds at

the C-terminus of peptides and proteins They belong

to the catalytic classes of either metalloproteases (clan

MC, family M14) or serine proteases (clan SC, family

S10) [1] and their action causes strong effects in the

biological activity of their peptide and protein

sub-strates [2] M14 MCPs, including those from animals,

plants and bacteria, have been divided into three main

subfamilies based on structural similarity and sequence

homology The first one, which includes the digestive

enzymes carboxypeptidase (CP) A (CPA) 1, CPA2,

carboxypeptidase B (CPB) 1 and mast cell CPA3, as

well as CPA4, CPA5 CPA6 and carboxypeptidase O

(CPO) (known at the gene level), has been termed

sub-family M14A or A⁄ B; the second one, including the

bioactive peptide-processing or regulatory enzymes

(e.g carboxypeptidases N, E, M and D, amongst

oth-ers) has been termed subfamily M14B or N⁄ E [3] Very

recently, a novel subfamily composed of enzymes of

larger size and apparently with a predominant

cyto-solic location, termed M14D, Nna-like or CCPs, has

been proposed [4] Furthermore, three main classes

may be distinguished according to their substrate

spec-ificity: (a) for aromatic⁄ hydrophobic residues (A-like),

(b) for basic residues (B-like) and (c) for acidic

resi-dues (O-like) [3,5]

MCP enzymes have been isolated from different

sources [3,5,6], mainly from vertebrates, but a few of

them have come from marine invertebrate organisms:

the digestive crayfish carboxypeptidase (CPB) [7], the

carboxypeptidase E-like enzyme from the sea hare

Aplysia californica, with important regulatory

func-tions in this organism [8], two CPs (A and B types)

from the hepatopancreas of the crab Paralithodes

cam-tschatica[9], the CPA-like protease from squid

hepato-pancreas of Illex illecebrosus [10], and CPs (two A and

one B type) isolated from the pyloric ceca of the

starf-ishes Asterias amurensis [11,12] and Asterina pectinifera

[13]

More than 95% of the Earth’s animal species are

invertebrates [14] The ecological services provided by

invertebrates are immeasurable; life as we know it

would be quite different or decline without them (see

Center for Applied Biodiversity Science;

http://sci-ence.conservation.org) Overall, our knowledge about

MCPs in invertebrates is very limited given the

tremen-dous variety of such organisms and compared to the

much larger number of characterized CP from

verte-brates [6] In the present study, we screened for the

presence of CP activity in marine invertebrates

belong-ing to the Phyla Cnidaria, Annelida, Mollusca,

Echi-nodermata, Arthropoda and Chordata, amongst

others, collected on the coasts of Havana, Cuba The

study has been based on the use of N-(4-meth-oxyphenylazoformyl)-l-phenylalanine (AAFP), a sensi-tive, specific and known colorimetric substrate for CPA enzymes One of the highest activity levels was detected in extracts from the marine annelid S magni-fica.This marine invertebrate, also termed ‘magnificent feather duster’, was obtained from coral reefs It belongs to the Phylum Annelida, Class Polychaeta, which shows a clear delimitation between its tentacle crown and its body (Fig 1) [15] Some studies per-formed on another annelid, belonging to the Sabellidae family, have only detected proteolytic activity assign-able to serine proteases, which appeared to be involved

in reproduction [16] despite their digestive origin The presence of a carboxypeptidase-like enzyme in Annelida marine invertebrates has not been described

so far

The present study describes the enzymatic activity and MS detection of a novel MCP (termed SmCP) from S magnifica, and its occurrence as a minor com-ponent within the animal body extracts by 2D- PAGE The enzyme has been isolated and purified, and then characterized by size, metal content, location, basic interactions, sequence analysis of different regions of the enzyme, and by a description of the main parame-ters related to enzyme kinetics, specificity and inhibi-tion ranges, as well as other basic molecular features From this, it is apparent that SmCP is a novel M14 MCP (belonging to the pancreatic-like subfamily), showing simultaneous CPA- and CPO-like activities, which is an unusual feature The present study com-prises an attempt to expand the growing field of the M14 family of proteolytic enzymes, which is now quite diverse and contains more than 25 different variants

Fig 1 S magnifica Phylum Annelida, Class Polychaeta, Subclass Palpata, Order Canalipalpata, Suborder Sabellida, Family Sabellidae, Genus Sabellastarte [14] The ‘tentacle crown’ and the ‘body’ parts

of the animal are clearly visible.

[4–6], but for which only very few members from

invertebrates have been characterized until now

Results

Detection of MCP activities in marine organisms

Twenty-five marine species belonging to different

invertebrate Phyla were screened for CPA activity

using AAFP as a substrate: four species of Mollusca

(Aplysia dactylomela, Aplysia juliana, Isognomun

radia-tus and Lima scabra); four species of Chordata

(Pallu-sia nigra, Microcosmus gamus, Molgula occidentalis

and Pyura vittata); 11 species of Cnidaria

(Bartholo-mea annulata, Budonosoma granulifera, Cassiopea

xamachana, Condylactys gigantea, Gorgonia ventalina,

Lebrunia danae, Palythoa caribaeorum, Physalia

phy-salis, Plexaura homomalla, Stichodactyla helianthus and

Zoanthus pulchellus); two species of Annelida

(Sabellas-tarte magnificaand Hermodice carunculata); two species

of Echinodermata (Holothuria mexicana and

Isostisch-opus badionotus); and two species of Arthropoda

(Lito-peaeus schmittiand Litopenaeus vannamei)

Among them, only the three species S magnifica

(Phyllum Annelida), B granulifera (Phyllum Cnidaria)

and P vittata (Phyllum Chordata) gave rise to positive

results, with specific activity values of 56.0, 1.6 and

1.8 UÆ100 mg)1 extract, respectively In these three

cases, we found a linear relationship between CP-like

activity and the quantity of extract used in the assay

Given that the material of the annelid S magnifica

showed by far the highest specific activity, it was

selected for further characterization studies In this

case, it was also found that extracts from the ‘body’ showed CP activity, whereas the feather-like ‘crown’ was devoid of it

‘Intensity fading’ MALDI-TOF MS Once we focused our attention on S magnifica body extracts, we found there direct evidence of at least one MCP enzyme, of approximately 35 kDa by ‘intensity fading’ MALDI-TOF MS [17] In the present study, the added ‘binder’ was the recombinant form of potato carboxypeptidase inhibitor (rPCI) (4.5 kDa), immobi-lized on agarose beads, with the aim of both perturb-ing the MS spectrum and capturperturb-ing the MCP in the body extract The control spectra, as well as the ‘per-turbed’ one (by rPCI addition, followed by removal of the captured targets by sedimentation of the beads), are shown in Fig 2A,B It is apparent that some of the ion signals of the spectra were faded when the extract was treated with immobilized PCI Subse-quently, MS analysis of the protein eluted from the beads (Fig 2C) detected a molecular ion of 34 kDa This molecular species, which is able to strongly inter-act with PCI, presumably represents the CP-like enzyme activity found in S magnifica body extract The experiment indicates not only the occurrence in the extract of the strong ligand (the enzyme SmCP) for the added protease inhibitor, but also that this ligand

is probably functional in the very complex extract (i.e not in the zymogen state) It is worth noting that the apparent simplicity of the MALDI-TOF spectrum of the extract shown in Fig 2C is most likely caused not only by the low expansion scale used, but also by

1000

A

0 500

2000

+PCI

0 500 1000 1500

100

0 50

m/z

B

C

Fig 2 MALDI-TOF MS of the ‘intensity

fading’ experiment (A) Mass spectra of the

S magnifica body extract (control sample)

before rPCI-agarose addition (B) Unbound

proteins mass spectra obtained after

rPCI-agarose addition (C) MS spectra of

recov-ered m ⁄ z signal after elution of the sample,

corresponding to CP-like enzyme The arrow

indicates the ‘perturbed’ signal by

rPCI-aga-rose addition.

‘signal suppression effects’; such phenomena usually

affect visualization of signals in media crowded in

mol-ecules [17–19], as will be reported and discussed

subse-quently

Molecular complexity of the S magnifica body

extract by 2D-PAGE

The molecular complexity of the S magnifica extracts

(both from the body and from the crown, or mixed)

was demonstrated by 2D-PAGE analysis (Fig 3) A

great number of visible protein bands [as revealed

either by staining with silver or using difference gel

electrophoresis (DIGE)] appeared in the analysis of

both parts of the animal, with a major presence of

bands in the body (upper part) versus the crown (lower

part) In Fig 3, we show, in the uncombined

(Fig 3A,B) or in the combined way (Fig 3C), the

pro-tein components of both parts of the animal labeled

with fluorescent dyes using the DIGE approach That

is, the different materials (i.e crown and body extracts,

purified enzyme) were pre-labeled independently with

DIGE reagents before they were mixed and run

simul-taneously in a single 2D-PAGE separation The

inde-pendent labeling of the crown and body extracts was

performed not only to allow the differential tracking

of their components, but also to deal with the very high content of dyes and interfering materials from the crown, which required a harsh cleaning (and denatur-ing) procedure Such interfering materials strongly per-turbed the electrophoretic separation, and also gave rise to severe band strikes and decreased resolution Only after testing several pre-cleaning and staining procedures (not shown), and selecting an adequate one, were we able to unveil the real band complexity

of the extracts (see Experimental procedures) We hope that this experience might be useful for the analysis of other invertebrates with a high content in dyes and other similar problems

Overall, more than 200 protein species are detected

by this procedure, among which those in the 17–37 kDa range are the most prominent To facilitate identification, we repeated the 2D-PAGE with three different initial samples from the body, after passing them through microcolumns with immobilized protein-aceous inhibitors of serine (soya bean protease inhibi-tor, SBTI), cysteine (chicken cystatin) and aspartic (pepstatin) proteases The intact, flow-through (depleted) and captured (released) materials were deriv-atized with DIGE and run in the same 2D-PAGE gel for each case (see Experimental procedures) The anal-ysis of the ‘captured’ spots allowed us to potentially

C

Fig 3 2D gel electrophoresis of pre-labeled protein extracts from S magnifica The gel contained 30 lg of total protein, separated by IEF using a pH 3–10 IPG strip in the first dimension and 15% SDS ⁄ PAGE in the second dimension The gel was first stained with the DIGE approach (see Experimental procedures), and subsequently checked by silver staining (A) Labeling with Cy5 fluorofor for the tentacle crown (B) Labeling with Cy2 fluorofor for the body (C) Body and tentacle crown alltogether (overlapped images) In the light box, the corresponding position of SmCP enzyme is shown when it was run in an individual 2D-PAGE (and visualized by immunostaining) The spots labeled with numbers correspond to molecular species affected by affinity capture on the immobilized inhibitors cystatin C (3, 4, 5, 6, 7 and 14) and soybean trypsin inhibitor (8, 9, 10, 11, 12 and 13), or on both (1 and 2).

identify at least 14 proteins captured differentially for

the first two microcolumns, which are labeled with

numbers in Fig 3B (1 and 2 by both; 3, 4, 5, 6, 7 and

14 by the cystatin one; and 8, 9, 10, 11, 12 and 13 by

the SBTI one) An initial validation of these

assign-ments as proteolytic enzymes (awating MS⁄ MS

analy-sis) was made by ‘intensity fading’ MALDI-TOF MS

using the mentioned set of immobilized inhibitors,

employing a strategy similar to the one for PCI

described above

It is important to note that the band corresponding

to the SmCP enzyme, the target of the present study,

did not appear at around 34 kDa, which is the mass

assigned to it as a potential MCP (see MALDI-TOF

MS analysis and below), when the extracts (either from

the body or body + crown) were analyzed However,

such a band is clearly visible when the enzyme is

puri-fied, concentrated and subsequently applied to the

2D-PAGE (Fig 3, encircled region) We assume that such

a difference is a result of the very low abundance of

SmCP in the animal Also, it is relevant that the use of

an antibody raised against the sequence around

Asn144-Arg145, preserved in CPs [4], gave rise to a

spot in the same location by immunostaining (not

shown), confirming its assignment

Purification and partial molecular characterization

of SmCP

After detection of carboxypeptidase activity in the

annelid worm (‘bodies’) of S magnifica, SmCP was

fractionated to homogeneity using affinity

chromatog-raphy on a PCI-Sepharose column as the first step of

purification The enzymatic activity was detected in the

eluted fraction with a 79% yield and a 286-fold

purifi-cation with respect to the crude extract (Table 1) The

second step of purification comprised anion exchange

chromatography on a TSK-DEAE 5PW column

(FPLC) (Tosoh Bioscience LLC, Montgomeryville,

PA, USA) (Fig 4A) SmCP eluted in a single fraction

with a specific activity of 322 UÆmg)1 and 1150-fold purification (Table 1) The purified enzyme was submit-ted to metal analysis by inductive coupled plasma-MS, which indicated that it contains 0.96 atoms of Zn per molecule

A single band with a molecular mass of 34 kDa was detected by SDS⁄ PAGE (Fig 4B) This result agrees with the molecular mass of 33 792 Da that was obtained when it was analyzed by MALDI-TOF MS (Fig 4C)

In addition, Edman degradation analysis revealed a unique N-terminal sequence, confirming the homogene-ity of SmCP at this end of the molecule Despite the rather limited size of the N-terminal region sequenced (19 residues: AFDLNDFNTLEDTYDQMNV), a blast search for this sequence revealed a consistent

Table 1 Summary of a typical purification procedure for SmCP

The assays were carried out as described in the Experimental

procedures Substrate AAFP at 0.1 m M , pH 7.5, 25 C.

Step

Protein

(mg)

Enzymatic activity (U)

Specific sctivity (UÆmg)1)

Yield (%) Purification (n-fold)

Affinity

chromatography

Ion exchange

chromatography

14.2

1 2

28 34.1 51 90 120

20.0

UV1/

15.0

10.0

5.0

–5.0

0 100 200

300 Unit·m

400 500

mL 0.0

0 200 400 600 800

15 000 20 000 25 000 30 000 35 000

16 928.956

33 792.855.

A

B

C

Fig 4 Purification of SmCP from the body extract of S magnifica and its molecular weight (A) Ion exchange chromatography on a TSK-DEAE gel (7.5 · 7.5 cm) column Buffer A: 20 m M Tris–HCl (pH 8.0); buffer B: 1 M Tris–HCl (pH 8.0) (I) Equilibration: 0% B for

45 min; (II) 60% B for 20 min; and (III) gradient 60% to 80% B for

170 min; flow rate: 68 cmÆh)1–––, A280; - - - -, Enz Act; –––, Conc NaCl (B) SDS ⁄ PAGE gel (125%) of the purified enzyme Lane 1, Standard molecular weights [myosin (203 kDa), galactosidase (120 kDa), bovine serum albumin (90 kDa), ovoalbumin (51 kDa), carbonic anhydrase (34.1 kDa), soybean trypsin inhibitor (28 kDa) and lysosyme (14.2 kDa)] Lane 2: Fraction of S magnifica purified

by PCI-Sepharose and anionic exchange chromatography (C) MS spectrum (MALDI-TOF) of SmCP.

homology with other MCPs, such as porcine and bovine

carboxypeptidase A1 precursor, mosquito Aedes

ae-gipty CPA, and the carboxypeptidase homolog from

Bothrops jaraca, amongst others (Fig 5) Subsequently,

and as a result of SmCP trypsin digestion followed by

LC-MS⁄ MS analyses, we identified nine internal

pep-tides (termed T1–T9), which showed identity to internal

sequences of different CPs (Fig 5) Some of them

include important ‘canonical’ residues of the catalytic

site of these enzymes [3] Thus, in peptides T2 and T6,

respectively, His69 and His196 (using canonical

num-bering) were found, which are tetrahedrally coordinated

to the catalytic zinc ion in all MCPs (i.e the numbering

system corresponds to bovine pancreatic CPA and is

used throughout) The other three most important resi-dues found in the sequenced peptides are Glu270 (T9), Asn144 and Arg145 (T2) Glu270, in the S1 subsite, acts

as a general base for catalysis, whereas Asn144 and Arg145, in the S1¢ subsite, bind the C-terminal carboxyl-ate group of the substrcarboxyl-ate The peptide T6 appears to contain Tyr198, which usually belongs to the S2 CP sub-site In addition, peptides T4 and T5 appear to contain two cysteine residues conserved in all members of MCP

A⁄ B subfamily, forming the disulfide bridge Cys138-Cys161 [1] Any peptide assignable to the putative speci-ficity site [3] was found Overall, these results indicate that SmCP represents a CP-like enzyme of the M14A subfamily [1,4]

Fig 5 Alignment of the amino terminal and internal sequences of SmCP with the sequences of carboxypeptidases from other organisms SmCP sequences were derived after trypsin treatment of the purified enzyme followed by LC-MS ⁄ MS (de novo sequencing) and bioinfor-matics analyses (see Experimental procedures) Similar and identical residues are shown in light and dark grey, respectively ‘Canonical’ resi-dues of CP (based on bovine CPA1) that are present in the trypsin peptides of SmCP are labeled with an asterisk The sequences are CPA from Aedes aegypti (yellow fever mosquito) (Q9U9K2 AEDAE); Carboxypeptidase A1 precursor from Mus musculus (CBPA1 MOUSE); car-boxypeptidase A2 from Paralichthys olivaceus (Japanese flounder) (Q8QAXN5 PAROL); carcar-boxypeptidase A1 precursor from Sus scrofa (CBPA1 PIG); carboxypeptidase A1 precursor from Bos taurus (CPBPA1 BOVIN); carboxypeptidase homolog from B jaraca (Q9PUF2 BOT-JA); CPO from Homo sapiens (CBPO HUMAN); CPB from Astacus fluviatilis (broad-fingered crayfish) (CBPB ASTFL); CPA precursor from

H armigera (cotton bollworm) (097434_HELAM); carboxypeptidase precursor from H armigera (cotton bollworm) (Q6H962_HELAM); MCP from Culicoides sonorensis (Q5QBL3_9DIPT); and carboxypeptidase A2 precursor from H sapiens (CBPA2_HUMAN).

Kinetic characterization of SmCP

Kinetic analyses for isolated SmCP was performed using

different types of standard synthetic substrates for

carb-oxypeptidases that were clearly cleaved by the enzyme

The Km, kcat and kcat⁄ Km determined for the enzyme

against AAFP, N-benzoyl-Gly-Phe (Hippuryl-Phe) and

N-(3-[2-furyl]acryloyl)-Phe-Phe (FAPP) as substrates

are shown in Table 2 Such kinetic parameters indicate

that SmCP is highly efficient against the three CPA type

substrates used On the other hand, we found that

SmCP is unable to cleave CPB type substrates such as

[3-(2-furyl)acryloyl]-L-alanyl-l-lysine (FAAK) or

N-(4-methoxyphenylazoformyl)-l-Arg (AAFR) Therefore,

SmCP appears to be more related to the A-type than to

the B-type MCPs [1–4]

The influence of pH on SmCP activity was also

ana-lyzed using the AAFP substrate, and indicated an

opti-mum pH value in the range 7.0–7.5 The effect of

various protease inhibitors on the SmCP enzymatic

activity is shown in Table 3 Inhibitors of cysteine

proteases

(l-carboxy-trans-2,3-epoxypropyl-leycylami-do (4-guanidino) butane, E-64; cystatin), aspartic pro-teases (pepstatin) and serine proteases (Pefabloc, soybean trypsin–chymotrypsin inhibitor, soybean tryp-sin inhibitor, aprotinin) did not have noticeable effects

on SmCP activity The enzyme was drastically inhib-ited by the chelating agent 1,10-phenanthroline at

1 mm However, EDTA at 10 mm, which might act by metal chelation, did not produce any inhibition at sim-ilar concentrations and inhibitor⁄ enzyme (Io⁄ Eo) rela-tionships (3 · 105m) Nevertheless, EDTA partial inhibitory effects were observed when preincubation times were increased By contrast, benzylsuccinic acid,

a well-known organic inhibitor of A-type carboxypep-tidases, fully cancelled the enzyme activity, at 1 mm Furthermore, the addition of the protein inhibitor of carboxypeptidases PCI (in fact rPCI, a recombinant form, reactive towards CPA and CPB type enzyme) at 0.4 lm produced a 70% inhibition of SmCP activity The apparent Kivalue for this inhibitor towards SmCP was 7.37· 10)8m; however, the adjusted value considering the substrate-induced dissociation was 2.45· 10)8m Another protein inhibitor from leech (rLCI, also recombinant) at 13.5 lm produced a 70% inhibition of SmCP activity The estimated Ki value for rLCI was 2.95· 10)8m, and its adjusted value considering the substrate induced dissociation was 1.45· 10)8m (Table 4) Preincubation of the inhibi-tors with the enzymes for various periods of time did not affect its inhibitory activity, suggesting that rLCI and rPCI are fast tight binding inhibitors

Table 2 Kinetic parameters for substrate hydrolysis catalyzed by SmCP in comparison with data reported for bovine pancreatic CPA (bCPA) The assays were carried out under the same conditions as those described for AAFP Substrate concentrations in the range 0.11–1.2 m M (3.29 n M of the enzyme in assay), 0.1–2 m M (24 l M of the enzyme in assay) and 0.02–0.25 n M (3.29 n M of the enzyme in assay) were used for AAFP, Hippuryl-Phe and FAPP, respectively.

Enzyme

Km(m M ) kcats)1 kcat⁄ K m M )1Æs)1 K

a

Mock et al [23].bCho et al [24]

Table 3 Effect of protease inhibitors on the relative activity of

SmCP SmCP: 3.29 n M ; AAFP: 0.1 m M ; pH 7.5, 25 C The enzyme

was preincubated with the inhibitors for 10 min at 25 C.

% Enzymatic activity I o ⁄ E o

M

M

M Trypsin-chymotrysin

inhibitor (soybean)

M

M Benzylsuccinic acid 1 m M < 1 3.03 · 10 5

M

M

M

M

M Trypsin inhibitor

(soybean)

M

Table 4 Kivalues of rPCI and rLCI against SmCP compared to pre-vious data obtained for bovine pancreatic CPA (bCPA) SmCP: 3.29 n M ; AAFP: 0.1 m M ; pH 7.5, 25 C The enzyme was preincu-bated with the inhibitors for 10 min at 25 C.

Carboxypeptidase

Ki(n M )

a Ryan et al [25] b Reverter et al [27].

On the other hand, we evaluated the effect on SmCP

of metal ions after overnight dialysis against EDTA at

10 mm (followed by the removal of excess EDTA by

dialysis against metal-free buffers; see Experimental

procedures) After this, SmCP only retains 40% of its

initial activity This apoform subsequently was used as

a control for the studies with metals We observed that

1 mm Ca2+, Mn2+ or Mg2+ enhanced the enzyme

activity of apoSmCP above 100% of the control

activ-ity, whereas the addition of Cd2+at 1 mm or Co2+at

1 mm or 10 mm did not affect the enzymatic activity

of the control (Fig 6) However, Cu at 1 mm and

10 mm reduced the apoenzyme activity to 11% and 15% of its residual activity Noteworthy, under our conditions, the addition of Zn2+ at 1 mm or 10 mm brought the activity to 100% (full rescue) and to 70%, respectively, with the latter assignable to inhibition by this metal

Specificity of cleavage Two different long peptides were used as substrate models to analyze the ability of SmCP to cleave differ-ent kinds of residues at the C-terminus, in comparison

Fig 6 Effect of divalent metals on SmCP activity The concentrations used in the assays were 329 n M for the enzyme SmCP and 0.1 m M for the substrate AAFP, at pH 7.5 and 25 C The enzyme, after EDTA treatment and dialysis against metal-free buffer (see Experimental procedures), was preincubated with the different ion metal salts at 1 m M , for 10 min at 25 C The assays were also performed, under the same conditions, at 10 m M for Zn 2+ , Co 2+ and Cu 2+

SmCP vs ACTH

A

B

SmCP vs V15E

E F

E F

F

F

F

2188 2317 2466 1427 1529 1541 1563 1587 1619 1693 1716 1748

ACTH control 60 min

bCPA vs ACTH

ACTH control 60 min

bCPA vs V15E

1793 1748

V15E control 60 min

V15E control 60 min

SmCP + PCI 60 min

15 min

30 min

60 min

bCPA + PCI 60 min

15 min

30 min

60 min

Fig 7 Determination of SmP specificity for C-terminal substrate residues Comparative analysis by MALDI-TOF MS of the degrada-tion of two synthetic substrates by SmCP and bovine pancreatic CPA (bCPA) The assays were performed in 10 m M Tris–HCl buffer (pH 8.0) with 1 l M of peptides and 2.19 n M of SmCP or 1 n M of bCPA in 10 lL

of final volume for 60 min (A) represents the enzymatic activity of SmCP against the ACTH fragment and V15E peptide, whereas (B) represents the enzymatic activity of bCPA against the same substrate.

Sequence of the ACTH fragment (residues 18–39): RPVKVYPNGAEDESAEAFPLEF, MW: 2466 Da; ACTHdes-F, MW: 2317 Da; ACTHdes-EF, MW: 2188 Da; V15E peptide sequence, VKKKARKAAGC(Amc)AWE: MW

1716 Da; V15Edes-E, MW: 1587 Da; V15Edes-WE, MW: 1400 Da;

V15Edes-AWE, MW: 1329 Da.

with bovine pancreatic CPA (a reference enzyme in the

field) After 15 min of incubation of SmCP with the

adrenocorticotropic hormone (ACTH) fragment used

as substrate (residues 18–39, 2466 Da), the enzyme was

able to release phenylalanine (ACTHdes-F, 2317 Da)

and glutamic acid (ACTHdes-EF, 2188 Da) residues

from the substrate C-terminus (Fig 7A) No further

amino acids were released after a 30-min incubation

period Under the same conditions, bovine pancreatic

CPA was only able to hydrolyze the C-terminal

phen-ylalanine residue from ACTH to obtain the

ACTHdes-F (2317 Da) The addition of the protein inhibitor

rPCI prevented cleavage in all cases

To confirm the capability of SmCP to hydrolyze

acidic residues from the C-terminus of peptides, the

specificity of SmCP against synthetic substrate

[VKKKARKAAGC(Amc)AWE] (V15E peptide)

(resi-due 15, 1716 Da) was evaluated (Fig 7B) After

15 min of incubation, the release of glutamic acid from

the peptide was observed and, after 60 min, the new

C-terminus residues formed and tryptophan and

ala-nine were further released, as shown by the trimming

scale: 1716, 1587 and 1329 Da However, bovine

pan-creatic CPA was unable to hydrolyze the first of such

C-terminal residues, glutamic acid, even after 60 min

of incubation Again, the addition of rPCI prevented

any kind of hydrolysis by the enzyme The release of a

glutamic acid residue from the C-terminus of peptides

is a very unusual capability of a CPA-like enzyme and

is reminiscent of the so-called CPO forms [3,5]

Discussion

The growing application of genomics and related

tech-nologies is facilitating an expanding view of the

pres-ent enzymatic families, including proteases [20] and

CPs in particular [4] However, such an advance is

lim-ited in the invertebrate world because of the great

diversity of organisms within it, which complicates the

study, but has the potential to generate enzyme

vari-ants of great biological and biotechnological values

To gain insight into the field of MCPs, one of the

most unknown among proteases in invertebrates, we

have used a mix of both modern and more classical

approaches to identify and characterize them,

estab-lishing comparisons with the vertebrate species (i.e the

reference ones) The present study started with a

sys-tematic screening in extracts from 25 invertebrates

from marine Caribbean species, using a specific and

sensitive enzymatic assay; this allowed us to detect the

presence of CPA-like activity in the body extract of

the marine annelid S magnifica Given that we

previ-ously reported the successful use of MALDI-TOF MS

for the initial detection of CPs and carboxypeptidase inhibitors in other crude biological extracts [17–19], we have applied such approaches to the S magnifica case The use of affinity capture on microbeads or microcol-umns derivatized with a recombinant carboxypeptidase inhibitor from potatoes, specific for such class of enzymes, and the use of MALDI-TOF MS signal anal-ysis approaches, allowed us to quickly identify in this annelid a 35-kDa species as a potential MCP, which

we named SmCP

Different fractionation methods have been per-formed to purify SmCP from the body extract of

S magnifica In initial attempts, using anion exchange and gel filtration chromatographies, we found a frac-tion with clear carboxypeptidase activity, which, inter-estingly, conveyed two additional activities against typical substrates for trypsin-like (benzoyl arginyl ethyl ester; BAEE) and chymotrypsin-like (benzoyl tyrosine ethyl ester; BTEE) serine proteases (data not shown) This suggests that, in the fractionation, SmCP could co-elute with serine proteases, perhaps establishing bin-ary or ternbin-ary complexes with such enzymes, as shown

in other organisms [21,22] Nevertheless, the substitu-tive use of affinity chromatography on rPCI-agarose,

in subsequent experiments, allowed the selective cap-ture of SmCP and contributed to its separation from the other enzymes Potentially, rPCI could promote the dissociation of SmCP from ‘complexes with serine proteases’ that it might establish in the crude extracts This is an issue that merits further research

The 2D-PAGE analysis of the crude extracts indi-cates that they are very complex in protein species, and that a stainable band at around 35 kDa, attribut-able to SmCP, is not directly visible with such approach unless high sensitivity approaches (i.e immu-nostaining) are employed This is probably a result of the low representation of this enzyme in the animal extracts, in agreement with its subsequent analysis and visualization in the purified form

Additionally, we obtained evidence by affinity cap-ture on three different kinds of immobilized proteina-ceous inhibitors (soybean trypsin inhibitor, cystatin, pepstatin), indicating that different main 2D-PAGE protein bands around 20–55 kDa correspond to cyste-ine and sercyste-ine protease enzymes present in the S mag-nifica body extract At least 14 species that gave stainable and clearly visible bands were detected by this approach They were provisionally validated by

‘intensity fading’ MALDI-TOF MS perturbation stud-ies carried out by the addition of such protein inhibi-tors on the extracts Full validation would require either direct isolation or MS⁄ MS analyses The later type of study is under way in our laboratory, but is

proving more difficult than expected because of the

very low homologies shown by S magnifica proteases

with respect to equivalent ones found in databases

Given the poor representation of invertebrate proteases

in databases, this is not an unexpected problem when

carrying out identification proteomics

It is worth noting that the preliminary detection of

serine and cysteine proteases species in the body

extracts correlates with the measure of their activities

by enzymatic analysis of the crude samples

Interest-ingly, neither approach revealed evidence of the

occu-rence of aspartic proteases Overall, although the

presence of pigments and other interfering products

initially constituted a very serious problem, once this

was technically solved, the feasibility and data

genera-tion capability of both the 2D-PAGE and ‘intensity

fading’ MALDI-TOF MS of this annelid indicated

that such proteomic-like approaches (and probably

related ones) are very promising for the analysis of

proteolytic enzymes in marine invertebrates

A central question in the analysis of novel MCPs

from biological sources is whether they occur in their

precursor or mature forms [2–5] In the present study,

using direct extracts from S magnifica, we found only

a monomeric and activated form of SmCP, as shown

by its enzymatic activity, molecular mass, derived

N-terminal sequence and homology analysis

Procarb-oxypeptidases are usually activated by proteolytic

removal of their activation segment by serine

prote-ases, mostly trypsin Studies on procarboxypeptidases

from several species have indicated that its activation

is dependent of the environmental ionic conditions

and, sometimes, the influence of quaternary structure

[2,5] Under our experimental conditions, quick

activa-tion of SmCP by autologous serine-like proteases,

which appeared to be present in large quantities in the

extract, could be favored On the other hand, the

coincidence between the N-terminal sequences of

SmCP and those from several other MCPs included in

alignments (Fig 5) also suggests that SmCP has been

purified in the active mature form In addition, we

found that the sequences of a number of SmCP

inter-nal peptides included important residues that belong to

catalytic site and domain of this enzyme family,

confirming our interpretation

All the experimental data reported in the present

study indicate that SmCP belongs to the M14A

sub-family of metalloproteases [6], the so-called

pancreatic-like forms (or A⁄ B), favoring its potential digestive

function in the marine annelid Its molecular weight

(33.7 kDa), N-terminal sequence and behavior towards

a panel of substrates and inhibitors are similar to those

of mammalian pancreatic CP (i.e the best known)

These types of enzymes have molecular masses close to

35 kDa after the removal of the propeptide, whereas the regulatory CPs (or N⁄ E) display higher mass val-ues as a result of the presence of other domains in addition to the CP domain [2,3] On the other hand, SmCP shows sequence homology with some CPs iso-lated from different vertebrates and invertebrates, belonging to the A⁄ B subfamily with CPA substrate preferences Only a few CPs have been isolated from marine invertebrates, and in not one case have the whole or extended sequences been disclosed This would be the case for the two CPAs and CPBs isolated from the hepatopancreas of the crab P camtschatica [9] and the CPA-like enzyme from the squid hepato-pancreas of I illecebrosus [10]

SmCP is able to cleave different types of CPA sub-strates such as AAFP, Hippuryl-Phe and FAPP, with

an overall efficiency similar to bovine pancreatic CPA, but with some significant differences in kcat, Km and

kcat⁄ Km for certain substrates [23,24] In addition, SmCP has a maximum activity at pH 7.5, in agreement with the optimum pH activity of almost all M14A CP-like forms, including marine enzymes [7–13], which lie

in the neutral range (pH 6.5–8.5), and is consistent with the pH at their sites of biological action [1,2]

As previously shown for mammalian CPs [25–27], potato and leech proteinaceous inhibitors efficiently inhibit SmCP, displaying similar Kivalues In addition, two smaller organic molecules (benzylsuccinic acid and 1,10-phenantroline) known to act on MCPs are also able to inhibit the enzyme By contrast, EDTA, which chelates metal ions, at 10 mm, failed to inhibit SmCP activity significantly after 10 min of preincubation, which is in agreement with the reported properties of other invertebrate MCPs isolated from the gut of Tion-ela bisselliella[28] and from Helicoverpa armigera larvae [29] for which EDTA effects are also time dependent The capability of divalent metal ions to substitute the essential active site Zn2+of MCPs [30,31], or bind

a second atom nearby [32], interfering with the cata-lytic mechanism, is well known We also observed diverse effects by the addition of such metals to SmCP After its dialysis against EDTA at 10 mm, SmCP reduced its activity to 40% of initial activity Starting from this state, the capacity of different metal ions to regenerate SmCP activity demonstrated that, in certain cases [Mn, Mg and Ca], there is an enhancement of activity of the enzyme; in others [Cd and Co], no changes are observed; and, in a third case [Cu], a clear inhibition is produced Such results are quite congru-ent with the well-known properties of mammalian CPs [33] In the case of Zn, an enhancement of SmCP activity was observed when added at 1 mm, whereas,