Báo cáo khoa học: Proteolytic degradation of nitric oxide synthase isoforms by calpain is modulated by the expression levels of HSP90 potx

In this respect, we have initially observed that, in Jurkat and in bovine aorta endothelium BAE-1 cells, containing different amounts of HSP90, the extent of degradation of nNOS and eNOS

by calpain is modulated by the expression levels of HSP90 Monica Averna, Roberto Stifanese, Roberta De Tullio, Franca Salamino, Mara Bertuccio,

Sandro Pontremoli and Edon Melloni

Department of Experimental Medicine (DIMES) ) Biochemistry Section and Centre of Excellence for Biomedical Research (CEBR),

University of Genoa, Italy

Nitric oxide (NO) is a gaseous free radical promoting

many biological effects, including the control of

micro-vascular tone, the renin and eicosanoic systems and

other modulators of inflammation [1–4] Due to its

high chemical reactivity, NO can be harmful through

the nitrosylation of many proteins [1,5] NO is

gener-ated exclusively by three NO synthase (NOS) isoforms

[3] Two of them constitutively expressed in cells have

been identified as neuronal NOS (nNOS) and

endo-thelial NOS (eNOS) on the basis of their preferential expression in neuronal or in endothelial cells, respec-tively The expression of the third form, inducible NOS (iNOS), is induced by various cytokines [1] All three isozymes catalyze the formation of NO from arginine, oxygen and NADPH [1–4] A number of co-factors are required for their catalytic activity, includ-ing tetrahydrobiopterin, FAD and FMN, in addition

to a heme prosthetic group To acquire the active state

Keywords

Ca 2+ homeostasis; calpain; calpastatin;

HSP90; NOS

Correspondence

S Pontremoli, University of Genoa,

DIMES ) Bicohemistry Section,

Viale Benedetto XV 1, 16132 Genoa, Italy

Fax: +39 010 518343

Tel: +39 010 3538162

E-mail: pontremoli@unige.it

(Received 31 July 2007, revised 12

Septem-ber 2007, accepted 8 OctoSeptem-ber 2007)

doi:10.1111/j.1742-4658.2007.06133.x

Ca2+ loading of Jurkat and bovine aorta endothelium cells induces the degradation of the neuronal and endothelial nitric oxide synthases that are selectively expressed in these cell lines For neuronal nitric oxide synthase, this process involves a conservative limited proteolysis without appreciable loss of catalytic activity By contrast, endothelial nitic oxide synthase diges-tion proceeds through a parallel loss of protein and catalytic activity The chaperone heat shock protein 90 (HSP90) is present in a large amount in Jurkat cells and at significantly lower levels in bovine aorta endothelium cells The differing ratios of HSP90⁄ nitric oxide synthase (NOS) occurring

in the two cell types are responsible for the conservative or nonconservative digestion of NOS isozymes Consistently, we demonstrate that, in the absence of Ca2+, HSP90 forms binary complexes with NOS isozymes or with calpain When Ca2+is present, a ternary complex containing the three proteins is produced In this associated state, HSP90 and NOS forms are almost completely resistant to calpain digestion, probably due to a struc-tural hindrance and a reduction in the catalytic efficiency of the protease Thus, the recruitment of calpain in the HSP90–NOS complexes reduces the extent of the proteolysis of these two proteins We have also observed that calpastatin competes with HSP90 for the binding of calpain in recon-structed systems Digestion of the proteins present in the complexes can occur only when free active calpain is present in the system This process can be visualized as a novel mechanism involving the association of NOS with HSP90 and the concomitant recruitment of active calpain in ternary complexes in which the proteolysis of both NOS isozymes and HSP90 is significantly reduced

Abbreviations

AEBSF, 4-(2-aminoethyl)benzenesulfonylfluoride; BAE-1, bovine aorta endothelium; CaM, calmodulin; eNOS, endothelial nitric oxide

synthase; HSP90, heat shock protein 90; iNOS, inducible nitric oxide synthase; nNOS, neuronal nitric oxide synthase; NO, nitric oxide; NOS, nitric oxide synthase.

nNOS and eNOS isoforms also requires calmodulin

(CaM) and Ca2+ions, indicating that NO synthesis is

triggered, in target tissues or cells, by an elevation of

free [Ca2+]i

A number of structural differences characterize the

three NOS isoforms At the N-terminal region, nNOS

contains a PDZ domain that addresses the enzyme to

specific synaptic sites in brain and muscle [2,3,6] This

segment is absent in eNOS and iNOS forms and the

latter is insensitive to Ca2+ions due to its high affinity

for the CaM binding site also at basal Ca2+levels [7]

NO production is highly regulated by many factors

and mechanisms, including protein–protein interactions

involved in allosteric regulation, scaffolding and

traf-ficking [3,8–13] Of particular interest is the interaction

with the chaperone heat shock protein 90 (HSP90) that

causes the release of the synthase from its association

with caveolin-3, a protein that maintains eNOS in an

inhibited state [14,15] HSP90 can also favour the

insertion of the heme group in the synthase at its

natu-ral site, promoting the dimerization and thus the

acquirement of the active conformation through the

association with CaM and Ca2+ ions [16,17] The loss

of HSP90 or its inhibition by geldanamycin prevents

the onset of the active NOS form and increases its

deg-radation by the ATP–ubiquitin–proteasome pathway

[18–21]

Degradation of NOS has also been proposed as a

regulatory mechanism in conditions of high NO

pro-duction in order to prevent the toxic effects of this

compound [20–22] It has been suggested that calpain

is the protease involved [23–27] because its activation

occurs in conditions that also cause the production of

NO, and because NOS and HSP90 have been

identi-fied as target substrates of the thiol protease [28–30]

In the present study, we further explored the

involvement of calpain in the regulation of nNOS and

eNOS activity taking into account a possible role of

HSP90 in this process

In this respect, we have initially observed that, in

Jurkat and in bovine aorta endothelium (BAE-1) cells,

containing different amounts of HSP90, the extent of

degradation of nNOS and eNOS by calpain was

directly related to the level of the chaperone protein

In reconstructed systems, we have demonstrated that

HSP90 significantly reduces the extent of NOS

proteo-lysis by calpain through the formation of selective

binary and ternary heterocomplexes containing the

synthases and the protease Accordingly, we propose

that the protective effect exerted by HSP90 is due to

the recruitment of calpain in a complex form in which

the chaperone protein becomes resistant to proteolysis,

and also due to a concomitant decrease in the Ca2+

binding capacity of calpain The physiological rele-vance of this novel property of HSP90 is also dis-cussed

Results

Degradation of NOS isozymes in Ca2+loaded Jurkat and BAE-1 cells

Degradation of NOS isozymes was studied in intact Jurkat and BAE-1 cells, containing nNOS and eNOS, respectively Calpain was activated [31] by increasing [Ca2+]i following exposure of cells to the Ca2+ -iono-phore A23187 As shown in Fig 1, in Jurkat cells, the nNOS native 160 kDa band was progressively reduced and converted into two new bands showing an approx-imate molecular mass of 140 and 130 kDa Because, in these conditions, more than 50–60% of the 160 kDa band disappeared, whereas the total catalytic activity

A

B

Fig 1 Digestion of NOS isozymes in cells loaded with Ca 2+ Jurkat and BAE-1 cells (2 · 10 6 for each experiment) were incubated in

2 mL of 10 m M Hepes, pH 7.4, containing 0.14 M NaCl, 5 m M KCl, and 2 gÆL)1glucose for 30 min at 37 C in the absence (control) or

in the presence of 1 l M ionophore A23187 and 1 m M CaCl2(Ca 2+ ion) Alternatively, cells were preincubated for 30 min at 37 C with

50 l M PD151746 (PD) Cells were then collected by centrifugation and lysed by freezing and thawing, followed by sonication (A) Aliquots (15 lg protein) of cell extracts were submitted to 7% SDS ⁄ PAGE and immunoblotting NOS isozymes and b-actin were detected with the specific mAbs (B) NOS activity was assayed as described in the Experimental procedures using an aliquot (200 lL)

of each cell extract The values reported are the arithmetical mean ± SD of three different experiments.

of nNOS was only reduced by 15–20%, it was assumed

that the digestion products still retained the ability to

produce NO In cells pretreated with the synthetic

cal-pain inhibitor PD151746 [22,32], the native 160 kDa

band, as well as the catalytic activity, were completely

preserved whereas no low molecular mass nNOS forms

were accumulated Identical results were obtained

fol-lowing preincubation of the cells with calpain

inhibi-tor-1 [33,34], known to be highly effective on calpain

although not completely specific (data not shown)

Conversely, in Ca2+ loaded BAE-1 cells,

approxi-mately 80% of the native eNOS 130 kDa band and

total catalytic activity disappeared, without the

appear-ance of detectable active intermediates Pre-treatment

with the synthetic calpain inhibitor PD151746 [22,32]

completely prevented the loss of both eNOS protein

and catalytic activity

Thus, whereas Ca2+ loading in Jurkat cells

pro-motes a limited proteolysis of nNOS, in BAE-1, cells

Ca2+-enrichement induces digestion and inactivation

of eNOS to a large extent

Expression of HSP90 and NOS isozymes in Jurkat

and in BAE-1 cells

To explain the different vulnerability of the two NOS

isozymes present in Jurkat and BAE-1 cells, and to

explore their relationship with the level of HSP90, the

three proteins were isolated and quantified By ion

exchange chromatography (Fig 2A), we separated

NOS from HSP90 and directly measured the amount

of these proteins expressed in both cells As shown in

Fig 2B, in Jurkat cells, a large amount of nNOS was

found to be accompanied by an even greater quantity

of HSP90 By contrast, in BAE-1 cells, the levels of

both eNOS and HSP90 were lower than those of

nNOS and HSP90 in Jurkat cells The quantification

of each protein shown in Fig 2C established that the

ratio nNOS⁄ HSP90 in Jurkat cells was largely in

favour of HSP90 whereas, in BAE-1 cells, eNOS

slightly exceeds the level of the chaperone protein

(Fig 2C) These findings indicate that HSP90 could

exert a protective effect in the digestion of NOS by

calpain

Interaction between HSP90, NOS isozymes and

l-calpain

To explore this process, we first examined the ability

of HSP90 to associate with NOS isozymes As

previ-ously reported [8,11,35], it was found that both NOS

isozymes immunoprecipitated from cell lysates with

an antibody-immobilized HSP90 (Fig 3A) Identical

results were obtained using purified protein prepara-tions, indicating that the association between NOS and HSP90 did not require specific factors present in the crude cell extracts We further characterized this

A

B

C

Fig 2 Separation of NOS isozymes from HSP90 in Jurkat and BAE-1 cells by ion exchange chromatography (A) Cell extracts from Jurkat (60 · 10 6

) and BAE-1 cells (10 · 10 6

), obtained as described

in the Experimental procedures, were submitted to ion exchange chromatography following the procedure described in the Experi-mental procedures NOS activity was evaluated on the eluted frac-tions (100 lL) as reported in the Experimental procedures s, nNOS activity; d, eNOS activity The arrow indicates the position

of HSP90 elution (B) Aliquots (30 lL) of the fractions eluted from the ion exchange chromatography described in (A) were submitted

to 7% SDS ⁄ PAGE and Immunoblotting NOS isozymes and HSP90 were detected with the specific mAbs as reported in the Experi-mental procedures (C) The immunoreactive bands of nNOS, eNOS and HSP90 were scanned and quantified as described in the Exper-imental procedures The areas of the peaks were normalized on the basis of the amount of protein loaded on the column The val-ues reported are the arithmetical mean ± SD of three different experiments.

association by the gel penetration technique (Fig 3B)

and established that HSP90 could form a one-to-one

discrete complex with both NO isozymes with a mass

of approximately 500–550 kDa, corresponding to the

association of the two native dimeric proteins

(Fig 3C)

Digestion of NOS isozymes and HSP90

by l-calpain

On the basis of the results so far described, the

suscep-tibility to digestion by human erythrocyte l-calpain of

purified nNOS, eNOS and HSP90 as single proteins or

in the associated forms was then evaluated As shown

in Fig 4A, in the presence of l-calpain, the digestion

of the nNOS native 160 kDa protein band was

pre-ceded by the transient accumulation of a 130 kDa

band The catalytic activity of nNOS also progressively

disappeared, in parallel with the digestion of the

130 kDa protein By contrast, eNOS was

concomi-tantly digested and inactivated by calpain (Fig 4B)

without the appearance of intermediate active

frag-ments

These digestion patterns are consistent with the removal of the N-terminal PDZ domain from the nNOS molecule that converts this enzyme in a molecu-lar form simimolecu-lar to eNOS [2,4] Both synthases are then cleaved in a position close to the CaM binding site that leads to the loss of catalytic activity in both iso-forms [2,4] The addition of CaM has no effect on the pattern on digestion (data not shown) Our findings appear to indicate that the digestion process of both NOS proceeds through the hydrolysis of a very limited number of peptide bonds; specifically, in the case of nNOS, degradation can occur with the cleavage of two peptide bonds and, in the case of eNOS, with the cleavage of a single bond HSP90 isolated from Jurkat cells was also digested by calpain with the transient accumulation of an 85–86 kDa band that replaces the native one (Fig 4C) However, the calpain requirement for HSP90 digestion was found to be five- to ten-fold higher than that required for digestion of NOS In addition, HSP90 from BAE-1 cells was digested by

A

B

C

Fig 3 NOS isozyme–HSP90 interaction in Jurkat and BAE-1 cells.

(A) Immunoprecipitation of nNOS and eNOS–HSP90 complexes.

Aliquots (500 lg protein) of Jurkat and BAE-1 cell extracts (C, Ex.)

obtained as described in the Experimental procedures, were

incu-bated overnight at 4 C with monoclonal anti-HSP90 serum as

pre-viously reported [8,11,35] The mixtures were then incubated for

1 h at 4 C with Protein G-Sepharose (30 lL) The particles were

collected, washed three times with the immunoprecipitation buffer,

resuspended in SDS ⁄ PAGE loading solution (30 lL) and submitted

to 7% SDS ⁄ PAGE The presence of NOS isozymes together with

HSP90 in the solubilized material was established using the specific

mAbs Alternatively, cell extracts were replaced with purified (Pur.)

NOS isozymes (1 lg) and HSP90 (1 lg) For experimental details,

see the Experimental procedures (B) Changes in molecular size of

NOS isozymes in the presence of HSP90 detected by gel

penetra-tion technique Equal amounts (0.5 lg) of nNOS or eNOS isolated

from Jurkat cells and BAE-1 cells, respectively, were diluted alone

or with the indicated amounts of HSP90, isolated from the

corre-sponding cell lines and added to packed Sephacryl S-300 The

distri-bution coefficient of NOS isozymes between the aqueous phase

and the gel fraction was determined as described previously

[36,37] The values reported are the arithmetical mean ± SD of

three different experiments (C) The molecular mass of NOS

iso-zymes, HSP90 and NOS isozyme–HSP90 complexes was evaluated

from the distribution coefficient of aldolase (molecular mass ¼

160 kDa) and ferritin (molecular mass ¼ 450 kDa) used as standard

proteins The distribution coefficient of these proteins between the

aqueous phase and the gel fraction was determined as described

previously [36,37] The values reported are the arithmetical

mean ± SD of three different experiments.

human erythrocyte l-calpain in an identical manner (data not shown)

Comparing the amount of l-calpain required to reduce the native bands of nNOS, eNOS and HSP90,

it was thereby established that nNOS was the most sensitive substrate; eNOS was slightly more resistant, whereas HSP90, independently from its source, was five- to ten-fold less susceptible (Fig 4D) A similar degradation pattern was obtained with m-calpain iso-lated from rat brain (data not shown)

When HSP90 was added to the nNOS digestion mix-ture (Fig 5A), the catalytic activity of the synthase was almost completely preserved in spite of the disap-pearance of the native band, which was completely converted into the 130 kDa protein species These results confirm the previous assumption that the

130 kDa nNOS form retained full catalytic activity In the case of eNOS, the addition of HSP90 prevented its calpain-mediated degradation as well as its inactivation (Fig 5B) Because HSP90 produces an identical pro-tective effect regardless of whether it is isolated from Jurkat or BAE-1 cells, it can be assumed that the effect of the chaperone is not restricted to a single cell type

Thus, the present findings suggest that HSP90 can prevent the degradation by calpain of both NOS by protecting the cleavage of the peptide bond close to the CaM binding site This explains why both NOS are inactivated by calpain in the absence of HSP90 The removal of the PDZ domain by calpain, which occurs without loss of catalytic activity in NOS even in the presence of HSP90, provides additional evidence that the function of this domain is probably related to changes in intracellular localization of the active syn-thase [6] This novel protective effect exerted by HSP90, as well as its higher resistance to calpain pro-teolysis, was then further explored utilizing purified

A

B

C

D

Fig 4 Susceptibility of nNOS, eNOS and HSP90 to calpain diges-tion nNOS (A), eNOS (B) and HSP90 (C) were incubated (1 lg each) with increasing amounts of human erythrocyte calpain as described in the Experimental procedures The insets in (A), (B) and (C) are representative of the immunoblots carried out to detect NOS isozymes or HSP90 digested by calpain NOS isozymes activity was evaluated on aliquots of each incubation (50 lL) as described in the Experimental procedures and is reported as a per-centage of NOS activity assayed in the absence of calpain The amount of native band of HSP90 shown in (C) was quantified as described in the Experimental procedures and is reported as a per-centage of the protein level measured in the absence of calpain.(D) Summary of the susceptibility of the three proteins to calpain diges-tion shown in (A), (B) and (C) The values reported are the arithmet-ical mean ± SD of three separate experiments.

immunocomplexes (see Experimental procedures) to

avoid the contamination by free proteins and possible

artefacts

Susceptibility of the ternary HSP90⁄ NOS ⁄ calpain

complex to proteolysis

For the first time, we were able to show that the

anti-body-immobilized HSP90 was capable of binding NOS

isozymes or calpain in the absence of Ca2+ ions, thus

forming alternative binary complexes (Fig 6A, lanes 1,

2 and 4) When eNOS or nNOS isozymes were

sepa-rately added to the HSP90⁄ calpain immunoprecipitated,

no ternary complexes were formed because calpain was completely displaced by NOS (Fig 6A, lanes 3 and 5) However, in the presence of Ca2+ions, each NOS iso-zyme and the protease could still be recruited and

A

B

Fig 6 Isolation of HSP90 ⁄ NOS ⁄ calpain complexes (A) HSP90 (5 lg) was immobilized to Protein G-Sepharose resin using mono-clonal anti-HSP90 serum as reported in the Experimental proce-dures The immunoprecipitated material was then incubated in the presence of: human erythrocyte calpain (lane 1), nNOS (lane 2), nNOS together with human erythrocyte calpain (lane 3), eNOS (lane 4), eNOS together with human erythrocyte calpain (lane 5) in the conditions described in the Experimental procedures Equal amounts (30 lL) of each sample were submitted to SDS ⁄ PAGE fol-lowed by immunoblot analysis (see Experimental procedures) The formed immunocomplexes were revealed with the specific mAbs against each of the proteins added to the samples containing the immobilized HSP90 (B) The same experiments described in (A) were carried out replacing EDTA with 1 m M CaCl 2 The immunopre-cipitated material was then incubated in the presence of human erythrocyte calpain (lane 1), nNOS (lane 2), nNOS together with human erythrocyte calpain (lane 3), eNOS (lane 5), eNOS together with human erythrocyte calpain (lane 6), in the conditions described

in the Experimental procedures The same experiments reported

in lanes 3 and 6 were also performed with the addition to the immunoprecipitated material of RNCAST600 (0.1 lg) [38] and are reported in lanes 4 and 7, respectively Equal amounts (30 lL) of each sample were submitted to SDS ⁄ PAGE followed by blot analysis (see Experimental procedures) The formed immuno-complexes were revealed with the specific mAbs against each of the proteins added to the samples containing the immobilized HSP90.

A

B

Fig 5 Calpain digestion of NOS isozymes in the presence of

HSP90 (A) nNOS (1 lg) or (B) eNOS (1 lg) were mixed with 1 lg

of HSP90 and incubated in the presence of the indicated amounts

of human erythrocyte calpain in the conditions reported in the

Experimental procedures At the end of the incubation, aliquots of

each sample (50 lL) were utilized to assay NOS activity, which is

expressed as a percentage of the activity detected in the absence

of calpain Insets in (A) and (B) represent the immunoblotting

car-ried out on aliquots (30 lL) of the same incubations, to detect

nNOS and eNOS, respectively The values reported are the

arith-metical mean ± SD of three separate experiments.

remained associated with HSP90, resulting in the

formation of ternary complexes (Fig 6B) These

protein–protein interactions were highly specific, as

indicated by the finding that the addition of calpastatin

removed calpain from the ternary complex (Fig 6B,

lanes 4 and 7) These results may be of physiological

relevance because they indicate that the formation of

the ternary complex is correlated with the level of

cal-pastatin present in the cytosol By quantification

anal-ysis of the immunoblots, it has been established that

the amount of calpain retained by immobilized HSP90

is almost equimolar to that of NOS isozymes Thus,

the ternary complexes may contain a copy of each

enzyme protein Altogether, these results can fit into

Scheme 1, which summarizes the type of protein–

protein interaction that can occur and their

intercon-version

We then tested whether calpain could still express

catalytic activity when inserted in these binary or

ternary complexes It was found that calpain, once

associated with HSP90, was unable to digest the

chaperone protein (Fig 7, upper panel), probably

because, following interaction, the susceptible peptide

bonds in HSP90 are no longer accessible to the

protease This hypothesis was confirmed by the

observation that calpain in its HSP90-associated

form can still digest exogenous substrates, such as

human denatured globin (Table 1) However, the

cat-alytic properties of the associated calpain differ from

that of the native enzyme because its efficiency is

reduced by 50%, probably due to a lower Ca2+

-binding capacity This explains why higher amounts

of calpain are required for digestion of HSP90

(Fig 4D) The changes in catalytic properties of

cal-pain provide an explanation of the mechanism by

which, in the ternary complex form, both HSP90

and NOS isozymes are protected from calpain

diges-tion This protection, however, was complete for

eNOS, whereas the nNOS native 160 kDa band was

still partially converted into the 130 kDa band

(Fig 7) Degradation of the chaperone protein and

the NOS isozymes became detectable only when the calpain concentration exceeded that of HSP90, a condition in which free active calpain molecules are now present (Fig 7)

Taken together, these findings strongly support the idea that the protective effect of HSP90 can represent

a novel mechanism allowing the production of NO even in conditions in which isolated forms of NOS could be rapidly degraded by calpain

This protection is mediated by HSP90, on the basis

of a dual mechanism: binding to NOS, which favours

Fig 7 Calpain digestion of isolated HSP90 ⁄ NOS ⁄ calpain complex-esThe ternary complexes containing HSP90, NOS and calpain were purified as described in the Experimental procedures and incubated with 1 m M CaCl2in the absence or in the presence of 2 lg of exo-genous human erythrocyte calpain Equal amounts (30 lL) of each incubation were then submitted to SDS ⁄ PAGE electrophoresis fol-lowed by immunoblot analysis (see Experimental procedures) The immunoreactive material was revealed using the specific mAbs.

Scheme 1 Binary and ternary complexes generated by HSP90,

NOS and calpain CLP, calpain; CST, calpastatin.

Table 1 Effect of HSP90 on the catalytic efficiency of human erythrocyte l-calpain Calpain was purified from human erythro-cytes as reported previously [39] HSP90 was isolated from Jurkat

or BAE-1 cells as described in the Experimental procedures Calpain activity was assayed using human denatured globin as a substrate,

as previously reported [39], in the absence or in the presence of equimolar amounts of HSP90 chaperone protein Inhibition is expressed as a percentage of the total calpain activity The activity measured in the absence of HSP90 was taken as 100% K 0.5 refers

to the [Ca 2+ ] ions required by calpain to express 1 ⁄ 2 V max The val-ues reported are the arithmetical means of three different experi-ments ± SD.

Addition

Calpain catalytic properties

Vmax(unitsÆmg)1) Inhibition (%) K0.5(l M )

the acquirement of the active conformation and the

fully functional state of the synthases [14–17], and

recruitment of active calpain, which prevents the

inac-tivation of NOS

Discussion

In the present study, we describe the digestion pattern

of NOS isozymes and HSP90 both in reconstructed

systems and in intact cells Our data consistently

indi-cate that the degradation of NOS is a highly regulated

process under the control of different mechanisms

and factors [2,18,20,26,28,35,40] Activation of NOS

requires an increase in [Ca2+]I, a condition promoting

also the activation of calpain If we consider the

sus-ceptibility of the isolated NOS forms to calpain

diges-tion, NO production in cells and organisms should be

very limited both in extent and time However, in

stim-ulated cells, activation of NOS isozymes is sustained

for too long a time period compared to the resistance

of these proteins to calpain proteolysis Thus, the

digestion must be controlled and degradation should

occur only when NO becomes a possible toxic agent

[41–44]

The findings reported in the present study represent

an answer to this question We have demonstrated

that, in addition to the well-known calpain

modula-tors, HSP90 is directly involved in this regulatory

pro-cess This chaperone binds to calpain and, when

associated with the protease, becomes resistant to

digestion In addition, the calpain present in the

com-plex maintains the ability to degrade exogenous

sub-strates, but with a reduced capacity that is also due to

a decreased ability to bind Ca2+ions When nNOS or

eNOS associates to the binary HSP90–calpain

com-plex, they are also protected from digestion by the

endogenous calpain Moreover, the amount of calpain

present in such a ternary complex is under the control

of calpastatin If the active calpain species increases

over the binding capacity of HSP90, the complex and

obviously the isolated proteins are degraded by the

protease On the basis of these observations, we can

explain not only the protective effect exerted by

HSP90 on the digestion of NOS isozymes, but also the

requirement of high levels of active calpain for the

chaperone proteolysis The inhibitory effect of HSP90

could derive from structural constraints on the

flexibil-ity of the calpain molecule, which reduces its

proteo-lytic efficiency but still allows the removal of the

PDZ domain from nNOS This proteolytic step does

not modify the overall catalytic efficiency of nNOS,

but it might produce a change in intracellular

localiza-tion of the synthase

The physiological relevance of these findings becomes particularly evident on the basis of the results observed in both cell models utilized in the present study (Fig 1) Thus, when Jurkat and BAE-1 cells were stimulated in identical conditions, the different patterns of nNOS and eNOS digestion can be attrib-uted to the different amounts of HSP90 expressed in the two cell lines Thus, the limited digestion of nNOS observed in Jurkat cells is due to the high levels of HSP90, which can trap part of the active calpain on one side and protect nNOS on the other In BAE-1 cells, this process is less efficient due to the low level

of HSP90, resulting in a high degree of eNOS diges-tion This new function of HSP90 in the control of

NO production might also be relevant in nervous and vascular tissues in which this free radical plays a par-ticularly important role in vessel relaxation

Experimental procedures

Materials Leupeptin, calpain inhibitor 1 [33,34], NADPH, Ca2+ -iono-phore A23187, calmodulin, FAD, FMN,

tetrahydrobiopter-in, l-arginine l-[14C]arginine (25 nCi; specific activity

308 CiÆmol)1), Source 15Q Sephacryl S-300, phenyl sepha-rose, Sephadex G-200 resins and protein G-Sepharose were obtained from GE Healthcare (Milan, Italy) Dowex 50W8

Na+form resin was obtained from Bio-Rad (Milan, Italy) Monoclonal antibodies against nNOS (catalogue number 611852), eNOS (catalogue number 610427) and HSP90 (cat-alogue number 610419) were purchased from BD Trans-duction Laboratories (Milan, Italy) Monoclonal b-actin antibody (catalogue number A-5441) was obtained from Sigma Aldrich (Milan, Italy) 4-(2-Aminoethyl)benzene-sulfonylfluoride (AEBSF) and calpain inhibitor 3-(5-fluoro-3-indoyl)-2-mercapto-(Z)-2-propenoic acid (PD151746) [22,32] were obtained from Calbiochem (Mississauga, Can-ada) Monoclonal anti-l-calpain serum (mAb 56.3) was produced as indicated previously [45] Human erythrocyte calpain was purified as reported previously [39] Rat brain m-calpain was purified as described previously [46] The ECL Detection System was obtained from GE Health-care

Cell culture BAE-1 cells were purchased from cell bank Interlab Cell Line Collection (accession no ICLCAL 00004) and main-tained in continuous culture at 37C (5% CO2) with DMEM (Sigma Aldrich) growth medium containing 10% fetal bovine serum and 2 mm l-glutamine; Jurkat (T cell leukaemia) cells were kindly provided by C Mingari (DIMES, University of Genoa, Italy) and maintained in

continuous culture at 37C (5% CO2) with RPMI 1640

(Sigma Aldrich) growth medium containing 10% foetal

bovine serum, 10 UÆmL)1 penicillin (Sigma Aldrich),

100 lgÆmL)1streptomycin (Sigma Aldrich) and 4 mm

l-glu-tamine (Sigma Aldrich)

Purification of nNOS, eNOS and HSP90 from

different sources

BAE-1 cells (10· 106

) were collected, lysed by sonication

in three volumes of ice-cold 50 mm sodium borate buffer,

pH 7.5, containing 1 mm EDTA, 0.5 mm

2-mercaptoetha-nol, 0.1 mgÆmL)1leupeptin and 2 mm AEBSF The

particu-late material was discarded by centrifugation (100 000 g for

10 min) and the soluble fraction (cell extract) was collected

and 1 mg protein was loaded onto a ion-exchange Source

15Q column (1.5· 3 cm) previously equilibrated in 50 mm

sodium borate buffer, pH 7.5, containing 0.1 mm EDTA

and 0.5 mm 2-mercaptoethanol (buffer A) The protein

con-centration was determined with the Bradford method [47],

using purified BSA as standard The adsorbed proteins

were eluted with a linear gradient (20 mL) 0–0.6 m NaCl

and collected in 1 mL fractions Aliquots of each eluted

fraction (30 lL) were resuspended in Laemmli loading

buffer [48] and submitted to 7% SDS⁄ polyacrylamide gel

electrophoresis followed by immunoblotting performed as

described in the section ‘Immunoprecipitation and

immuno-blotting’ The immunoreactive material was detected using

the specific mAbs directed against HSP90 and eNOS The

eluted fractions containing the two proteins were separately

collected, concentrated by ultrafiltration and loaded onto

Phenyl Sepharose column (1.5· 3 cm) equilibrated in

buffer A containing 0.3 m NaCl HSP90 and eNOS were

retained by the resin whereas unabsorbed proteins were

washed out The adsorbed proteins were eluted with

buffer A, collected, concentrated and loaded onto Sephadex

G-200 column (1.8· 160 cm) equilibrated in the same

buffer A containing 0.15 m NaCl Proteins were collected in

1 mL fractions and 30 lL of each fraction were utilized to

detect HSP90 and eNOS by immunoblot analysis HSP90

was now separated from eNOS and the two proteins were

separately collected and concentrated by ultrafiltration The

same procedure was applied to obtain HSP90 and nNOS

from Jurkat cells (60· 106

)

Alternatively, aliquots of eluted fractions (100 lL) from

the ion exchange Source 15Q chromatography previously

described were utilized to assay NOS isozymes activity as

described above

Assay of NOS activity

NOS activity was measured by detecting the production

of citrulline from l-[14C]arginine as previously reported

[26] with the following modifications: aliquots of Jurkat

and BAE-1 cell extracts (100 lg) or of the fractions

(100 lL) eluted from the ion exchange chromatography described above, were incubated in buffer A (250 lL) containing 1 mm NADPH, 200 mm calmodulin, 20 lm tetrahydrobiopterin, 1 lm FAD, 1 lm FMN and 5 lm

l-arginine and 25 nCi of l-[14C]arginine (specific radio activity 308 CiÆmol)1) at 37C After 30 min, the reaction was stopped with ice-cold 50 mm Hepes, pH 5.5, contain-ing 5 mm EDTA (2 mL) The samples were then submit-ted to anion exchange chromatography using 2 mL of packed Dowex 50W8 Na+ form resin pre-equilibrated with stop buffer l-citrulline was eluted by washing the resin with 3 mL of stop buffer and the radioactivity present was counted in a liquid scintillation counter One unit of NOS activity is defined as the amount of enzyme producing 1 pmol citrullineÆmin)1 in the specified conditions

Immunoprecipitation and immunoblotting Jurkat (50· 106

) or BAE-1 cells (5· 106

) were lysed in ice-cold 20 mm Tris⁄ HCl, pH 7.4, containing 2.5 mm EDTA, 2.5 mm EGTA, 0.14 m NaCl, 1% Triton X-100,

10 lgÆmL)1 aprotinin, 20 lgÆmL)1 leupeptin, 10 lgÆmL)1 AEBSF and 10 lgÆmL)1 phosphatases inhibitor cocktail I and II, followed by brief sonication Cell lysates were cen-trifuged (12 000 g for 15 min at 4C) and protein quantifi-cation of the supernatants was performed using the Lowry assay The immunoprecipitation was performed as previ-ously described using 500 lg of detergent-soluble protein (cell extract) and 3 lg of monoclonal anti-HSP90 serum [8,11,35]

Alternatively, nNOS (1 lg) isolated from Jurkat cells

or eNOS (1 lg) isolated from BAE-1 cells, as previously described, were incubated with HSP90 (1 lg) isolated from the corresponding cell types The mixtures were immobi-lized to Protein G-Sepharose resin using monoclonal anti-HSP90 serum (1 lg) in 300 lL (final volume) of 50 mm sodium borate buffer, pH 7.5, containing 1 mm EDTA (buffer B), following a previously reported procedure [38] After incubation, the different samples were centrifuged and the pellet was resuspended in 30 lL SDS⁄ PAGE load-ing buffer [48], heated for 5 min and submitted to 7% poly-acrylamide gel electrophoresis in the presence of SDS Proteins were blotted to a nitrocellulose membrane (Bio-Rad) and probed with specific mAbs, followed by a peroxi-dase-conjugated secondary antibody as described [49] The immunoreactive bands were developed with an ECL detec-tion system, detected with a Bio-Rad Chemi Doc XRS apparatus and quantified using the Quantity One software, release 4.6.1 (Bio-Rad) To quantify proteins from bands revealed by western blotting, known amounts of protein were submitted to SDS⁄ PAGE and stained with the appro-priate antibody The bands were then scanned and the area

of the peaks obtained was used to create a calibration curve

Equilibrium distribution experiments

in sephacryl S-300

Equilibrium gel distribution (gel penetration) experiments

with samples containing different mixtures of NOS and

HSP90 were carried out as previously described [38]

Briefly, nNOS (0.5 lg) or eNOS (0.5 lg) isolated from

Jur-kat cells or BAE-1 cells, respectively, were diluted alone or

with increasing amounts (0–1 lg) of HSP90 isolated from

the corresponding cell types in buffer B (0.5 mL) The

solu-tions were added to 0.5 mL of packed Sephacryl S-300

pre-viously equilibrated with buffer B and rotated end-over-end

for 2 h at 4C The resin was packed for 15–20 min at

4C and NOS activity was assayed as previously described

in this section using aliquots (0.2 mL) of the clear aqueous

phase

Isolation of the NOS⁄ HSP90 ⁄ calpain complexes

Isolated HSP90 (5 lg) from Jurkat or BAE-1 cells was

incubated with monoclonal anti-HSP90 serum at 4C for

2 h, in buffer B (300 lL final volume) Protein G-Sepharose

(30 lL) was then added to the samples and the mixtures

were rotated end-over-end for 2 h at 4C Sepharose beads

were collected, washed three times with buffer B (500 lL)

to discard proteins not specifically bound Human

erythro-cyte calpain (0.5 lg in 300 lL of buffer B) was added to

the pellet and incubated for 2 h at 4C Sepharose

immu-noprecipitated material was collected, washed three times

with buffer B (500 lL) and exposed to nNOS (1 lg) or

eNOS (1 lg) isolated from Jurkat or BAE-1 cells,

respec-tively Alternatively, the EDTA present in buffer B was

replaced with 1 mm CaCl2 (final concentration) In these

conditions, human erythrocyte calpain was maintained in

its inactive state by the addition in these mixtures of

0.1 mgÆmL)1(final concentration) leupeptin

In vitro digestion of NOS isozymes and HSP90

with human erythrocyte calpain

HSP90, nNOS, eNOS (1 lg each) isolated from Jurkat or

BAE-1 cells as reported above, were incubated (100 lL

final volume) with human erythrocyte calpain [39] in

buf-fer B for 1 h at 37C, in the presence of 1 mm CaCl2

Digestion of the ternary complex

HSP90⁄ NOS ⁄ calpain

nNOS, eNOS, HSP90 and calpain, coimmunoprecipitated

as previously described, were incubated (30 lL final

vol-ume) in 50 mm sodium borate buffer, pH 7.5, containing

1 mm CaCl2for 1 h at 37C in the absence or in the

pres-ence of exogenous purified calpain (0.1 lg) Reactions were

stopped with 0.1 m EDTA (2 lL) The samples were

sub-mitted to 7% SDS⁄ PAGE and, following blotting, the

nitrocellulose membrane was probed with monoclonal anti-HSP90, nNOS and eNOS sera

Assay of calpain activity Calpain activity was assayed as previously described [39] One unit was defined as the amount of enzyme causing the release from the substrate of 1 nmol of free NH2 groups The specific activity of human erythrocyte calpain and of rat brain m-calpain was 1075 unitsÆmg)1 and 655 unitsÆ

mg)1, respectively

Acknowledgements

This work was supported in part by grants from MIUR, FIRB and PRIN projects, and from the Uni-versity of Genoa

References

1 Guix FX, Uribesalgo I, Coma M & Munoz FJ (2005) The physiology and pathophysiology of nitric oxide in the brain Prog Neurobiol 76, 126–152

2 Kone BC, Kuncewicz T, Zhang W & Z (2003) Protein interaction with nitric oxide synthases: controlling the right time, the right place, and the right amount of nitric oxide Am J Physiol Renal Physiol 285, 178–190

3 Alderton WK, Cooper CE & Knowles RG (2001) Nitric oxide synthases: structure, function and inhibition Biochem J 357, 593–615

4 Govers R & Rabelink TJ (2001) Cellular regulation of endothelial nitric oxide synthase Am J Physiol Renal Physiol 280, 193–206

5 Stamler JS, Jia L, Eu JP & McMahon TJ, Demchenko

IT, Bonaventura J, Gernert K & Piantadosi CA (1997) Blood flow regulation by S-nitrosohemoglobin in the physiological oxygen gradient Science 276, 2034–2037

6 Cao J, Viholainen JI, Dart C, Warwick HK, Leyland

ML & Courtney MJ (2005) The PSD95–nNOS inter-face: a target for inhibition of excitotoxic p38 stress-activated protein kinase activation and cell death J Cell Biol 168, 117–126

7 Cho HJ, Xie QW, Calaycay J, Mumford RA, Swiderek

KM, Lee TD & Nathan C (1992) Calmodulin is a sub-unit of nitric oxide synthase from macrophages J Exp Med 176, 599–604

8 Papapetropoulos A, Fulton D, Lin MI, Fontana J, McCabe TJ, Zoellner S, Garcia-Cardena G, Zhou Z, Gratton J & Sessa WC (2004) Vanadate is a potent acti-vator of endothelial nitric-oxide synthase: evidence for the role of the serine⁄ threonine kinase akt and the 90 kDa heat shock protein Mol Pharmacol 65, 407–415

9 Su Y, Edwards-Bennet S, Bubb MR & Block ER (2003) Regulation of endothelial nitric oxide synthase by the