Báo cáo khoa học: Regulation of calpain B from Drosophila melanogaster by phosphorylation pot

Because structural predictions hinted at the presence of several potential phosphorylation sites in this enzyme, we investigated the in vitro phosphorylation of the recombi-nant protein

Regulation of calpain B from Drosophila melanogaster by phosphorylation

La´szlo´ Kova´cs1,*, Anita Alexa2,*, Eva Klement3, Endre Ko´kai1, A´ gnes Tantos2

, Gergo¨ Go´gl2, Tama´s Sperka4, Katalin F Medzihradszky3,5, Jo´zsef To¨zse´r4, Viktor Dombra´di1,6and

Pe´ter Friedrich2

1 Department of Medical Chemistry, Faculty of Medicine, University of Debrecen, Hungary

2 Institute of Enzymology, Biological Research Center, Hungarian Academy of Sciences, Budapest, Hungary

3 Proteomics Research Group, Biological Research Center, Hungarian Academy of Sciences, Szeged, Hungary

4 Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Debrecen, Hungary

5 Department of Pharmaceutical Chemistry, University of California at San Francisco, CA, USA

6 HAS-DU Cell Biology and Signaling Research Group, Department of Medical Chemistry, Research Center for Molecular Medicine, University of Debrecen, Hungary

Keywords

calcium-dependent protease; Drosophila

melanogaster ; enzyme kinetics; epidermal

growth factor; protein kinase

Correspondence

V Dombra´di, Department of Medical

Chemistry, Faculty of Medicine, University

of Debrecen, 98 Nagyerdei krt, Debrecen,

H-4032, Hungary

Fax: +36 52 412 566

Tel: +36 52 412 345

E-mail: dombradi@med.unideb.hu

*These authors contributed equally to this

work

(Received 23 April 2008, revised 15 June

2009, accepted 6 July 2009)

doi:10.1111/j.1742-4658.2009.07198.x

Calpain B is one of the two catalytically competent calpain (calcium-acti-vated papain) isoenzymes in Drosophila melanogaster Because structural predictions hinted at the presence of several potential phosphorylation sites

in this enzyme, we investigated the in vitro phosphorylation of the recombi-nant protein by protein kinase A as well as by the extracellular signal-regu-lated protein kinases (ERK) 1 and 2 By MS, we identified Ser845 in the

Ca2+binding region of an EF-hand motif, and Ser240 close to the autocat-alytic activation site of calpain B, as being the residues phosphorylated by protein kinase A In the transducer region of the protease, Thr747 was shown to be the target of the ERK phosphorylation Based on the results

of three different assays, we concluded that the treatment of calpain B with protein kinase A and ERK1 and ERK2 kinases increases the rate of the autoproteolytic activation of the enzyme, together with the rate of the digestion of external peptide or protein substrates Phosphorylation also elevates the Ca2+ sensitivity of the protease The kinetic analysis of phos-phorylation mimicking Thr747Glu and Ser845Glu calpain B mutants con-firmed the above conclusions Out of the three phosphorylation events tested in vitro, we verified the in vivo phosphorylation of Thr747 in epi-dermal growth factor-stimulated Drosophila S2 cells The data obtained suggest that the activation of the ERK pathway by extracellular signals results in the phosphorylation and activation of calpain B in fruit flies

Structured digital abstract

l MINT-7214239 : ERK1 (uniprotkb: P40417 ) phosphorylates ( MI:0217 ) CalpainB (uniprotkb: Q9VT65 ) by protein kinase assay ( MI:0424 )

l MINT-7214216 , MINT-7214228 : PKA (uniprotkb: P12370 ) phosphorylates ( MI:0217 ) CalpainB (uniprotkb: Q9VT65 ) by protein kinase assay ( MI:0424 )

l MINT-7214325 : CalpainB (uniprotkb: Q9VT65 ) cleaves ( MI:0194 ) MAP2C (uniprotkb: P11137 ) by protease assay ( MI:0435 )

Abbreviations

CaMKII, calcium⁄ calmodulin-dependent protein kinase II; CID, collision-induced dissociation; EGF, epidermal growth factor; ERK, extracellular signal-regulated protein kinase; LY-AMC, N-succinyl-Leu-Tyr-7-amino-4-methyl-coumarin; MAP, microtubule-associated protein; Ni-NTA, nickel–nitrilotriacetic acid; PKA, protein kinase A.

The regulation of intracellular signaling pathways

involves an intricate interplay of various enzyme

sys-tems One intriguing example is the interaction

between calpains (i.e the Ca2+-dependent

SH-prote-ases) and protein kinases (i.e the enzymes of protein

phosphorylation) Regarding their modes of action,

calpains catalyze the irreversible, limited proteolysis of

their substrate proteins, whereas protein

phosphoryla-tion can be reversed by the protein phosphatases via

the hydrolytic elimination of the phosphate group

from Ser, Thr or Tyr residues Calpains play crucial

roles in controlling various cellular processes, such as

cytoskeletal remodeling, cell cycle, apoptosis and cell

motility [1,2] The ubiquitous mammalian l- and

m-calpains (calpain-1: EC 3.4.22.52; calpain-2: EC

3.4.22.53) are the best-characterized members of the

family The regulation of these essential proteases still

remains an open question The micromolar to

millimo-lar Ca2+concentrations required for the effective

acti-vation of calpains in vitro is not in the physiological

range It has been suggested that the presence of

addi-tional regulatory substances (e.g phospholipids) or

post-translational protein modifications (e.g

phosphor-ylation) can modulate the Ca2+-sensitivity of these

proteases [3,4] Phosphorylation of mammalian

calpains has been intensively studied Because

recombi-nant calpains, which are devoid of phosphate groups,

are fully active, it can be concluded that

phosphoryla-tion is not essential for their activity On the other

hand, l- and m-calpain extracted from different tissues

contain two to four phosphate group⁄ molecule, which

are distributed over eight or nine different Ser, Thr

and Tyr residues [1] The main phosphorylation sites

in m-calpain are Ser50 and Ser369⁄ Thr370 In vitro

and in vivo studies show that phosphorylation of Ser50

by extracellular signal-regulated protein kinases

(ERKs) enhances calpain activity, as well as

calpain-mediated physiological processes [5], whereas

phos-phorylation of Ser369⁄ Thr370 by protein kinase A

(PKA) inhibits calpain action [6] Nicotine-induced

phosphorylation by an isoform of protein kinase

C, PKCi, enhances both the activity and secretion of

l- and m-calpain in human lung cancer cells [7]

In the present study, we examined calpain B from

Drosophila melanogaster We chose a fly enzyme

because Drosophila is a handy model organism and, out of its four calpains, three (calpain A, B and C) have been characterized in our laboratory [2] As far

as we are aware, only calpain A and B exhibit protease activity in the fruit flies Both of them are activated at millimolar free Ca2+ concentration The active Dro-sophila calpains consist of a single polypeptide chain, for which the domain structure shows strong similarity

to the catalytic subunits of mammalian l- and m-cal-pain They are composed of an N-terminal regulatory domain (I), a catalytic domain (II), a C2-like domain (III) and a calmodulin-like calcium binding domain (IV) The main difference between mammalian and Drosophila calpains is represented by the length of the N-terminal domains; for example, calpain B has a 240 amino acid long N-terminal region with an extended, disordered structure [8] Upon Ca2+-dependent activa-tion, this N-terminal inhibitory region is first clipped off in an autoproteolytic process Because the bioinfor-matic analysis of the primary structure of calpain B suggested a number of potential phosphorylation sites that correspond to consensus recognition motifs of sev-eral protein kinases, we initiated the investigation of the phosphorylation of this protein under both in vitro and in vivo conditions In the present study, we report

on the identification of cAMP-dependent protein kinase (protein kinase A, PKA; EC 2.7.11.11) and mitogen-activated protein kinase (ERK1 and ERK2;

EC 2.7.11.24) phosphorylation sites in calpain B and describe the effects of phosphorylation on the proteo-lytic activation and activity of the enzyme

Results and discussion

Phosphorylation of calpain B in vitro The phosphorylation of any of the Drosophila calpains has not been reported yet According to motif scan (http://scansite.mit.edu/motifscan_id.phtml) prediction, there are five putative PKA consensus sites and several ERK target sites in calpain B, when screened at a low stringency level (data not shown) The feasibility of ERK action was further supported by the presence of three ERK1 binding sites and three ERK2 binding sites in the protein To confirm the structural

predic-l MINT-7214275 : ERK2 (uniprotkb: P40417-2 ) phosphorylates ( MI:0217 ) CalpainB (uni-protkb: Q9VT65 ) by protein kinase assay ( MI:0424 )

l MINT-7214319 : CalpainB (uniprotkb: Q9VT65 ) and CalpainB (uniprotkb: Q9VT65 ) cleave ( MI:0194 ) by protease assay ( MI:0435 )

tions, purified recombinant calpain B was

phosphory-lated with PKA, as well as with the two isoforms of

the extracellular signal-regulated protein kinase, ERK1

and ERK2, in the presence of [32P]ATP[cP] in vitro

(Fig 1) PKA incorporated 0.20 ± 0.09 (n = 5) mol

PÆmol protein)1 The phosphorylation with ERK1 was

more effective, and 0.62 ± 0.27 (n = 6) mol PÆmol

protein)1 was achieved within 2 h, whereas ERK2

built in 0.73 ± 0.17 (n = 7) mol PÆmol protein)1

under the same conditions (Fig 1B) In all of the

experiments, more than 95% of the total protein

bound radioactivity resided in the band corresponding

to the apparent molecular mass of the recombinant

calpain B (Fig 1A)

For the identification of the phosphorylated amino

acid residues, both the wild-type active protease and

an inactive calpain B mutant were phosphorylated as

described above; with the exception that

nonradioac-tive ATP was used instead of [32P]ATP[cP] The inac-tive mutant was generated (aiming to avoid unwanted autoproteolytic degradation during sample handling)

by replacing Cys314 with Ala in the active center of the enzyme The active calpain B and the inactive C314A mutant were purified by SDS⁄ PAGE after

in vitro phosphorylation The proteins were in-gel digested with trypsin and analyzed by MS The phos-phopeptides were identified from the digests using precursor ion scanning and affinity enrichment Phosphopeptides yield diagnostic m⁄ z 79 (PO3 )) ions

in negative ion mode The precursors of this fragment were identified in nanoLC⁄ MS ⁄ MS experiment on a QTRAP mass spectrometer (Applied Biosystems, Fos-ter City, CA, USA) Collision-induced dissociation (CID) data acquired in positive ion mode from these precursor ions provided sufficient information for sequence and modification site assignment These

A

B

Fig 1 Phosphorylation of calpain B in vitro (A) Recombinant calpain B was phosphorylated with PKA, ERK1 and ERK2 protein kinases in the presence of [32P]ATP[cP] Samples were analyzed by SDS ⁄ PAGE followed by autoradiography The left-hand lanes show the molecular mass standards (St) and 2 lg of unphosphorylated calpain B protein (P) stained with Coomassie brilliant blue The molecular mass of the standards is given in kDa The right-hand lanes present the autoradiorams of the samples taken at the indicated time points after the initia-tion of phosphorylainitia-tion (B) The phosphorylainitia-tion reacinitia-tion by PKA ( ), ERK1 (.) and ERK2 ( ) was also monitored by counting the radioactiv-ity incorporated into the TCA insoluble protein The mean ± SD of five to seven independent experiments is shown.

assignments were confirmed by phosphopeptide

enrich-ment using TiO2 affinity chromatography followed by

LC⁄ MS ⁄ MS analysis in information-dependent

acqui-sition mode on an ion trap mass spectrometer In the

wild-type calpain B enzyme, two PKA-phosphorylation

sites at Ser240 and Ser845 were revealed by the

MS⁄ MS spectra of the precursors at m ⁄ z 764.8 (2+)

and m⁄ z 591.7 (2+), respectively The MS ⁄ MS

spec-trum of the precursor at m⁄ z 764.8 (2+) represents

the phosphopeptide 238qNS(p)VSKGDFQSLR250 The

phosphorylation site assignment is based on fragments

observed at m⁄ z 393.3 (b3), 519.3 (y92+), 1037.0 (y9)

and 569.0 (y102+) (Fig 2A) The phosphorylation site

at Ser845 was unambiguously identified from the

fragment observed at m⁄ z 512.7 (y8) in the MS⁄ MS

spectrum of the precursor at m⁄ z 591.7 (2+)

corre-sponding to 842TGS(p)IDGFHLR852 (Fig 2B) For

the ERK2 kinase, phosphorylation at Thr747 was

determined from the MS⁄ MS spectrum of the

pre-cursor at m⁄ z 689.9 (3+) corresponding to 739

IA-PSLPPPT(p)PKEEDDPQR756; the fragments at m⁄ z 482.0 (b5) and 793.5 (y13) clearly indicate phosphoryla-tion at Thr747 (Fig 2C) The same phosphorylaphosphoryla-tion sites were identified in the inactive calpain B mutant, with the exception that Ser240 phosphorylation was not found In addition, we demonstrated that ERK1 phosphorylated the same Thr747 residue as the ERK2 isoenzyme (data not shown) The sites of in vitro phos-phorylation in calpain B are summarized in Fig 2D Although the 3D structure of calpain B has not been solved yet, from the available atomic coordinates of m-calpain [9], the sites of phosphorylation can easily

be assigned to the structural domains of the enzyme

by homologous modeling In agreement with the motif scan prediction, the PKA consensus site, Ser845, lies

in domain IV within the second EF-hand motif (resi-dues 831–859) It is preceded by two basic resi(resi-dues, Arg841 and Arg842, that create a favorable environ-ment for PKA recognition The second PKA site at Ser240 is at the end of domain I, close to the

x 5

Fig 2 Identification of the in vitro phosphorylation sites in calpain B by MS (A) CID spectrum of the precursor at m ⁄ z 764.8 (2+) represent-ing qNS(p)VSKGDFQSLR and (B) CID spectrum of the precursor at m ⁄ z 591.7 (2+) corresponding to TGS(p)IDGFHLR, both observed in the TiO2 enrichment of the PKA phosphorylated calpain B digest (C) CID spectrum of the precursor at m ⁄ z 689.9 (3+) representing IAPSLPPPT(p)PKEEDDPQR detected in the TiO2enrichment of the ERK2 phosphorylated calpain B digest Water loss is marked with a hash symbol (#); )98 indicates phosphoric acid loss Nomenclature is used in accordance with Biemann [22] The peaks used for the identification

of the phosphorylation sites are marked with an arrow (D) Summary of the in vitro phosphorylation sites (shown in bold) in calpain B.

ing scission site located between amino acids 224–225.

Because its environment (Fig 2D) is less preferred by

the kinase, we consider this residue as a secondary site

of phosphorylation Although Ser845 appears to be

the preferential PKA site within the molecule, it should

be noted that both the rate and the extent of PKA

phosphorylation are rather low (Fig 1B) Thr747 was

identified as the site of phosphorylation of either

ERK1 or ERK2 In agreement with the consensus

sequence of the two kinase isoforms, there are three

consecutive Pro residues immediately at the N-terminal

side of Thr747 Indeed, this site is recognized by all of

the so called Pro-directed protein kinases Thr747 is

situated at the surface of the molecule between

domains III and IV in an extended structural element

termed ‘transducer’ [10,11] This region of the

polypep-tide was suggested to transmit the Ca2+ signal from

the Ca2+binding EF-hands of domain IV to the active

site cleft that is situated between the IIa and IIb

sub-domains The phosphorylation of additional potential

ERK consensus sites was not supported by the

experi-ments

In comparison with the mammalian counterparts,

none of these phosphorylation sites of calpain B have

been conserved in calpain l or m and, vice versa, the

known phosphorylation sites of mammalian calpains

are missing from Drosophila calpain B From the first

part of our studies, we conclude that calpain B can be

phosphorylated by three kinases at three different

resi-dues, although the sites of phosphorylations are

dis-tinct from those reported for the mammalian enzymes

[5,6] Consequently, the regulation of the Drosophila

protease can be different from the well-known

mam-malian calpains

Effects of phosphorylation on the activation and

activity of calpain B

Because the location of the phosphorylation sites

suggested an effect of phosphate incorporation on the

Ca2+regulation of calpain B, we compared the kinetic

properties of the nonphosphorylated (control) and

phosphorylated enzymes using three independent

methods

Fluorimetric assay with a peptide substrate

At high Ca2+ concentration, a continuous assay was

applied using the fluorescent dipeptide substrate

N-suc-cinyl-Leu-Tyr-7-amino-4-methyl-coumarin (LY-AMC)

In an earlier study [10], we determined that a

8.6 ± 0.8 mm free [Ca2+] concentration was required

for the half maximal activation of calpain B

Accord-ingly, at 9 and 19 mm free Ca2+ concentrations, the enzyme works at 50% and 90% of its full activity, respectively Under these conditions, the reaction is fast enough to reach the maximal velocity (vmax), a parameter that can be used to compare the phosphory-lated forms with the nonphosphoryphosphory-lated one As an example, two calpain progress curves are presented in

Fig 3A, demonstrating the effect of PKA-treatment Both progress curves start with a lag-phase, corre-sponding to the autoproteolytic activation; later, the sigmoid-like curves reach maximal activity In the pres-ent study, the progress curves were fitted with a log-istic curve (see Experimental procedures), which provided the kactand vmax values The phosphorylated calpain B forms were found to be activated faster (Fig 3B), and had a greater activity (vmax) at a 9 mm free Ca2+concentration (Fig 3C) Similar results were obtained at a 19 mm free Ca2+ concentration (data not shown) In all of our experiments, ERK2 exerted a more pronounced effect than ERK1, in accordance with the fact that the stoichiometry of phosphorylation was somewhat higher with the former kinase The rela-tively small effect of PKA can be attributed either to the different sites of modification or, more likely, to the lower level of phosphorylation

Activity assay with a protein substrate Microtubule-associated protein (MAP) 2c is readily digested by mammalian m-calpain as well as calpain B, even at lower Ca2+ concentrations The proteolytic reaction was carried out at a 350 lm free Ca2+ con-centration and monitored by SDS⁄ PAGE followed by the densitometric scanning of the 62 kDa intact MAP2c band (Fig 4A) Simple visual inspection of the results demonstrates that calpain B phosphorylated with PKA digested MAP2c faster than the nonphosph-orylated form For quantitative evaluation of the data,

we plotted the logarithm of the optical density of the MAP2c band at a given time divided by the optical density measured at the beginning of the reaction [i.e ln(A⁄ A0)] against the reaction time and determined the slope of the linear curve (Fig 4B) Assuming that the quantity of the active enzyme is constant during the reaction, the slope gives the rate constant of the first-order reaction We used this constant for the charac-terization of the enzyme activity and expressed the effect of phosphorylation as described above (Fig 4C)

In this independent assay, again, ERK2 increased cal-pain B activity more vigorously than the other kinases The Ca2+ dependence of PKA action on calpain B activity was investigated in more detail (Fig 4D) The results obtained clearly demonstrate that the effect of

PKA is larger at lower Ca2+ concentrations Because, using this method, we were unable to collect data at the very beginning of the reaction, the plots were not suited to analyze the effect of phosphorylation on the activation of calpain B

Autolysis assay

To circumvent the above limitation of MAP2c diges-tion assay, we designed an alternative approach for the determination of the effects of phosphorylation on autolysis The approach was based on our observation

of a slight autocatalytic processing of calpain B in the MAP2c based assay (Fig 4A) Under modified condi-tions, and in the absence of MAP2c, we monitored the disappearance of the 104 kDa intact calpain B band as

a function of time (Fig 5A) and used the same kinetic approach as applied before for the determination of the apparent first-order rate constant of the reaction (Fig 5B) According to our assumption, the first auto-catalytic cleavage between amino acid residues 224–

225 is sufficient for the activation of calpain B; thus, the rate of the elimination of the 104 kDa inactive form is approximately proportional to the rate of acti-vation Figure 5C shows that phosphorylation by either ERK2 or PKA slightly elevated the rates of autolysis in the presence of a 19 mm free Ca2+ concen-tration The effect was negligible when the free Ca2+ concentration was set at 1 mm (data not shown) ERK1 was not studied in this test because it phospho-rylates the same site as ERK2, but with slightly lower efficiency When calpain B was phosphorylated with [32P]ATP[cP], the distribution of the radioactive label was monitored by autoradiography during autolysis (Fig 5D) Although the overall rate of autolysis was comparable (i.e the rate constants were kERK2= 0.039 s)1and kPKA= 0.034 s)1, respectively), the gen-eral picture was quite different for the two kinases tested The radioactive phosphate incorporated by PKA was eliminated very quickly from the 104 kDa band, and most of the total radioactivity (more than 85%) accumulated in the 75 kDa fragment after the first steps of autolysis According to the activation model of calpain B [12], this fragment represents the C-terminal portion of the protein Thus, the distribu-tion of radioactivity is in agreement with our previous result indicating that the PKA phosphorylation sites of calpain B reside inside the 75 kDa fragment Regard-ing the ERK2 phosphorylation, approximately 75% of the radioactivity incorporated by the kinase dis-appeared within less than 30 s from the gel Only 21–26% of 32P remained in the 75 kDa C-terminal proteolytic fragment The most likely explanation is

A

B

C

Fig 3 Effect of phosphorylation on calpain B activity as measured

with a peptide substrate The progress curve of calpain activity

measurement with the fluorimetric assay (see Experimental

proce-dures) in the presence of a 9 m M free Ca 2+ concentration is shown

as an example (A) Calpain B was phosphorylated by PKA (broken

line) for 60 min to the stoichiometry of 0.21 mol PÆmol protein)1, or

treated under the same conditions without the kinase (solid line) in

a control experiment The fluorescent signal was recorded after the

addition of the enzyme to the reaction mixture From the progress

curves, the rate of activation (kact) as well as the maximal activity

(v max ) of calpain B was calculated The average k act for the

non-phosphorylated form was 8.5 · 10 3

M )1Æs)1 at a 9 mM free Ca2+

concentration, whereas the average vmaxwas 10)8M )1Æs)1, and the

k cat , which can be determined from the v max , was 1.5 · 10)2s)1.

The increase in the rate of activation (B) and in the activity (C) upon

phosphorylation with PKA, ERK1 and ERK2 is given as the

percent-age of the unphosphorylated controls The extent of

phosphoryla-tion was 0.19 ± 0.03, 0.63 ± 0.18 and 0.75 ± 0.18 mol PÆmol

protein)1 for the three kinases, respectively The mean ± SD of

three or four independent experiments is shown.

that, beside Thr747, ERK2 effectively phosphorylated

another residue(s) in the unstructured N-terminal

regu-latory domain that was very quickly degraded during

the autoproteolytic activation process

Site-directed mutagenesis mimicking

phosphorylation of calpain B

Although the direct phosphorylation of calpain B with

the selected protein kinases provides realistic results,

the explanation of the data is complicated by the low

stoichiometry of phosphorylation in the case of PKA,

and by the existence of multiple phosphorylation sites

in all cases To assess the contribution of a well

estab-lished site to the regulation of calpain B, we adopted

the approach of Smith et al [13] and replaced the

target sites of the kinases with Glu by site-directed

mutagenesis The negative Glu side-chain mimics the

effect of phosphorylation In addition to being specific,

site-directed mutagenesis has two additional

advanta-ges: the modifications are stoichiometric and

perma-nent The mutated forms of calpain B carrying the

point-mutations T747E and S845E were expressed and

purified in exactly the same way as the wild-type

recombinant protein The yield and the purity of the

three calpain B variants was the same (Fig 6A) The effect of the phosphorylation mimicking mutations on calpain B activity was first analyzed by the fluorimetric assay (Fig 6B) Both mutations activated the protease

in a wide range of free Ca2+ concentrations The mathematical analysis of the Ca2+ response curves is given in Table 1 The data can be fitted well to sig-moid curves and the parameters of the curves provide

an excellent tool for the characterization of the Ca2+ dependence of the three enzyme forms On the basis of the [Ca2+]1⁄ 2and dx parameters, we conclude that the mutants are activated at lower Ca2+ concentrations, but are less sensitive to changes in the Ca2+ concentra-tion than the wild-type enzyme (Table 1) Although phosphorylation-mimicking mutations can enhance the

Ca2+ sensitivity of calpain B, the values of [Ca2+]1⁄ 2 are still far above the physiological range Both muta-tions raise the maximal activity of calpain B, and this activity enhancement is greater at low Ca2+(Table 1); thus, the effect of phosphorylation could be more pro-nounced at Ca2+ concentrations that are closer to the physiological conditions

The effects of the mutations on the autocatalysis of calpain B were tested using two independent methods The elimination of the intact 104 kDa protein band

A

B

C

D

Fig 4 Effect of phosphorylation on calpain B activity as measured with MAP2c substrate The digestion of MAP2c by calpain B (see Experi-mental procedures) at a 50 l M free Ca 2+ concentration is shown as an example (A) The time-course of proteolysis with calpain B that had been either phosphorylated by PKA (0.21 mol PÆmol protein)1), or not phosphorylated (control) was monitored by SDS ⁄ PAGE The arrow points towards the calpain B bands The densities of the intact MAP2c bands (indicated by an arrowhead) were estimated by densitometry and the apparent first-order rate constants of MAP2c digestion were determined for both the phosphorylated ( ) and nonphosphorylated (h) protease (B) The percentage increase in the reaction rate upon phosphorylation by PKA, ERK1 and ERK2 was calculated as in Fig 3 The proteolytic reactions were carried out in the presence of a 350 l M free Ca2+concentration, and the average first-order rate constant was 2.2 · 10)3Æs)1 for the nonphosphorylated protease The average stoichiometry of phosphorylation was the same as that in Fig 3 The mean ± SD of seven to eleven independent experiments is shown (C) The effect of PKA-mediated phosphorylation on calpain B activity was also measured as a function of Ca 2+ concentration (D) The free Ca 2+ concentration is given in molÆdm)3, and the mean ± SD for three

to five experiments is shown The stoichiometry of calpain B phosphorylation was 0.19 ± 0.03 mol PÆmol protein)1in this experiment.

was determined by SDS⁄ PAGE and densitometry (Fig 6C) as described above (Figs 5A,B) The data obtained by this approach are in agreement with the

kact values calculated from the fluorimetric progress curves Both approaches resulted in comparable data indicating that the T747E mutation causes a larger (and the S845E mutation a small but reproducible) increase in the activation of calpain B (Fig 6D) It is readily apparent that the Glu mutations and the incorporation of a phosphate into the Thr or Ser side-chains are not fully equivalent modification; neverthe-less, the mutations tested here confirm our previous results obtained with the phosphorylated proteins, at least in qualitative terms

In vivo phosphorylation of calpain B

To determine the possible physiological significance of our in vitro findings, we investigated whether calpain B was phosphorylated in the S2 Drosophila cell-line First, we isolated the protein by immunoprecipitation and SDS⁄ PAGE from the untreated cells (Fig 7A) and analyzed the putative phosphorylation sites by

MS None of the phosphopeptides presented in Fig 2

D were detected; thus, we concluded that calpain B was not phosphorylated in resting cells The same result was obtained when the cells were treated with the phosphatase inhibitor calyculin A Next, we inves-tigated whether calpain B become phosphorylated upon the stimulation of the cells When epidermal growth factor (EGF) was used to activate the MAP kinase⁄ ERK pathway and the dephosphorylation of proteins was prevented by calyculin A, we noted that,

on SDS⁄ PAGE, the calpain B band was split into a doublet of two closely migrating bands, suggesting a postsynthetic modification of the protein (Fig 7A)

MS revealed that two residues (Thr747 and Ser240) were phosphorylated in the EGF-treated cells The ion chromatograms corresponding to the m⁄ z 690 and 765 values are shown in Fig 7B The identity of the two phosphopeptides was confirmed by MS⁄ MS experi-ments, which gave mass spectra very similar to those shown in Figs 2A,C The phosphopeptide peaks were missing from the ion chromatograms of the untreated sample, despite the fact that, in the tryptic digest, many unphosphorylated calpain B peptides were detected with higher relative abundance than in the +EGF sample Thus, we demonstrated that the main ERK1⁄ ERK2 site, Thr747, was phosphorylated in vivo upon EGF stimulation No additional ERK sites expected from the motif scan prediction and from the

in vitro phosphorylation experiment (Fig 5D) were verified in vivo Surprisingly, Ser240, a putative PKA

A

B

C

D

Fig 5 Effect of phosphorylation on the autolysis of calpain B The

autolysis of recombinant calpain B without phosphorylation (h) or

after treatment with ERK2 (m) was tested at a 19 m M free Ca2+

concentration The process was monitored by SDS ⁄ PAGE (A) and

the rate constants for the disappearance of the 104 kDa band were

determined (B) The increase of autolysis rate caused by

phosphor-ylation with ERK2 and PKA is shown in (C) as the mean ± SD of

three to four independent experiments The average first-order rate

constant of autolysis was 0.028 s)1for the nonphosphorylated

cal-pain B The stoichiometry of phosphorylation was 0.77 and

0.13 mol PÆmol protein)1for ERK2 and PKA, respectively (D) After

phosphorylation of the recombinant protein with [ 32 P]ATP[cP], the

autolysis of radioactive calpain B was analyzed by SDS ⁄ PAGE and

Coomassie staining (upper panel) followed by autoradiography

(lower panel) The arrows point towards the intact calpain B bands.

site, was also phosphorylated in the same experiment,

despite the fact that the treatment of S2 cell with

fors-kolin and calyculin A was not sufficient to induce in

vivo calpain B phosphorylation (data not shown)

Obviously, the activation of the MAP kinase signaling

was more effective than the stimulation of the PKA

pathway alone A positive interaction between the two

signaling pathways cannot be excluded, but it is not

clear why the other, more potent in vitro PKA site (i.e

Ser845) remained unaffected One likely explanation is

that Ser240 was phosphorylated not by PKA, but by

another kinase in the living cells In this respect, it is

important to note that the stoichiometry of in vitro

PKA phosphorylation has always been rather low and, according to the results of the motif scan, the peptide sequence 233ATSARQNSVSKGDFQ247, containing Ser240, is a preferred target of the calcium⁄ calmodu-lin-dependent protein kinase II (CaMKII kinase; EC 2.7.11.17) as well Thus, it is possible that, after EGF treatment, the EGF induced Ca2+ influx activated the latter kinase [14], which in turn incorporated the phos-phate into Ser240 in a cAMP-independent manner To make things even more complicated, a motif scan test indicates that CaMKII can also phosphorylate the PKA site Ser845, but with lower efficiency However, modification at the latter site was not observed in any

Table 1 Parameters characterizing the Ca 2+ concentration dependence of enzyme activity [Ca 2+ ]1⁄ 2denotes the Ca 2+ concentration which corresponds to half-maximal activity of the enzyme, whereas parameter dx is the width of the logistic function characterizing the depen-dence of the activity on the logarithm of the Ca 2+ concentration, lg[Ca 2+ ] It is a dimensionless quantity and gives the sensitivity of enzyme activity to small changes in the ion concentration around the value [Ca2+] 1⁄ 2 Greater values of dx correspond to a lower sensitivity to changes in Ca 2+ concentration.

Calpain B

[Ca 2+ ]1⁄ 2

Activity ratio

at high [Ca 2+ ]

Activity ratio

at low [Ca 2+ ]

A

B

C

D

Fig 6 The effects of phosphorylation mimicking mutations on calpain B Thr747 (ERK site) and Ser845 (PKA site) were substituted with Glu

in calpain B by site-directed mutagenesis The purified mutated proteins T747E and S845E behaved as the wild-type (Wt) recombinant protein

on SDS⁄ PAGE (A) St, standards; molecular mass values are given in kDa The Ca 2+

-dependent activity of the wild-type (h) as well as the T747E ( ) and S845E ( ) mutant calpain B was determined with the fluorescent LY-AMC substrate (B) The maximal activity of the wild-type enzyme was taken as 100% The mean ± SD of eight independent experiments is shown The autolysis of the native (h) and mutated ( ), ( ) calpain B was monitored by SDS ⁄ PAGE (C) The effect of the mutations on the activation of the protease was estimated from the progress curves, as in Fig 3 (white columns), or from the SDS ⁄ PAGE patterns, as in Fig 5 (black columns), in the presence of a 19 m M free Ca 2+

concentration (D) The mean ± SD of three to seven experiments is shown.

of our in vivo experiments Taking these arguments together, the phosphorylation of Ser240 can take place

in living cells and may contribute to the regulation of the enzyme, although, most probably, this reaction is not catalyzed by PKA Additional experiments are required for the clarification of the role for this site

On the other hand, it is clear that Thr747 can be phos-phorylated in vitro and in vivo by the ERK enzymes The extracellular signal regulated modification of cal-pain B, as identified in the present study, represents a physiologically relevant regulatory tool

Conclusions

The identification of the in vitro phosphorylation sites comprised the first step towards a better understanding

of the regulation of calpain B According to homolo-gous modeling, the amino acid residues phosphory-lated by PKA and the ERK isoforms are situated in sensitive regions of the molecule and can be involved, either directly or indirectly, in the regulation of enzyme activity We demonstrated that the postsynthetic modi-fication of these sites increases the rate of autocatalytic activation, as well as the activity and Ca2+ sensitivity

of recombinant calpain B Although the changes were moderate, they could contribute to the modulation of regulatory networks in a more significant way Struc-tural predictions and in vitro experiments with recom-binant proteins reveal biochemically feasible mechanisms that are not necessarily operating in a living organism We found that the phosphorylation of Thr747 in the transducer region of the protease occurs

in EGF-stimulated S2 cells The results obtained in the present study suggest that the activation of the MAP kinase⁄ ERK pathway by extracellular signals, among many diverse changes, results in the phosphorylation and activation of calpain B in D melanogaster We suggest that the regulation of calpain B in fruit flies must be different from that in mammalian counter-parts Although calpains catalyze the same proteolytic reaction, the position and function of the critical phos-phorylation sites have not been conserved Indeed, instead of the Thr747 residue that is phosphorylated in Drosophila calpain B, Asp524 or Glu524 are found in the corresponding position in different mammalian m-calpain enzymes [1], suggesting that a natural drift

in the protein sequence mimics the effect of phosphor-ylation in mammals This example confirms that evolu-tion operates at the level of regulaevolu-tion Without understanding the small structural alterations affecting the regulatory potential of a protein, it would be pre-mature to suggest functional equivalence based on overall structural similarities

A

B

Fig 7 Phosphorylation of calpain B in vivo S2 Drosophila cells

were incubated in the presence of 10 n M EGF and 100 n M calyculin

A (+EGF) or in the absence of these additions (–EGF) Calpain B

was partially purified from the treated and the untreated cells and

was analyzed by SDS ⁄ PAGE (A) Two hundred nanograms of

recombinant calpain B was used as a control (C) The mass of the

standards (St) is given in kDa on the left-hand side The excessive

band at around 52 kDa indicates the presence of immunoglobulins.

Arrows indicate a single band (–EGF) or a doublet of bands (+EGF)

that corresponds to the molecular mass of calpain B These bands

were excised, digested with trypsin and analyzed by MS The

extracted ion chromatograms of calpain B phosphopeptides after

TiO2 enrichment are shown in (B) m ⁄ z 690 corresponds to the

IAPSLPPPT(p)PKEEDDPQR peptide, whereas m ⁄ z 765 represents

the qNS(p)VSKGDFQSLR peptide The retention time (RT) is

given (min) above the peaks.