321

Biochimica et Biophysica Acta. 1076(1991) 321-329

© 1991 ElsevierSciencePublishersB.V. 0167-4838/91/$03.50 ADONIS 01674838001011 BBAPRO 33818

An analysis of the topography of molecular forms I and 3 of rat liver biliverdin reductase using polyclonal antibodies Judith Frydman, Mafia L. Tomaro and Rosalia B. Frydman Facultad de Farmacia y Bioqulmica, Universidad de Buenos 4ires, Buenos Aires (Argentina)

(Received20 September 199~,)

Key words: Heinecatabolism;Biliverdinreductase;Proteinantigenicity;Thiol-disulfideregulation;Oxidativestress; (Rat liver) Rat liver biliverdin reductase exists in two molecular forms. The major one (molecular form !) is t t m ~ o r m e d , under conditions of oxidative stress into another molecular form (molecular form 3) which is an S-S bridged dimer of form 1. The chemical modifications of the thiol, arginine and lysine residues of molecular form I which resulted in an inhibition of its catalytic activity did not affect the activity of molecular form 3. Rabbit polydonal antibodies raised against form 1 did not recognize form 3. This lack of recognition persisted even when the dimer (form 3) was denatured with S D S or urea under non-reductive conditions. Reduction of form 3 with reduced thioredoxin gave the monomeric form 1, which, was fully recognized by the antibodies. The latter recognized the biliverdin reductases from rat spleen and kidney to the same extent as they did with form !. Molecular form 1 was completely inhibited by the addition of the antibodies. This inhibition was prevented by preincuhation of the enzyme with either the substrate (biliverdin) or the cosnhstrnte (NADPH). Preiacnhafion with the latter or with NADP + (but not with bilirubin) strongly impaired the recognition of form ! by the antibodies. Modification of the lysine or arglnine residues of form 1 which were involved in substrate binding, impaired the inter,~tion of the enzyme with the antibodies. The antisera blocked the enzymatic conversion of form 1 to form 3, but alkyl~tion of the thiol residue involved in this dirnerization had no effect on the interaction of form ! with the antibodies. The lack of recognition of form 3 by the antibodies suggest that the antigenic site of the former becomes buried upon dimerization.

Introduction Bifiverdin reductase is the cytosolic enzyme responsible for the NADPH-dependent reduction of biliverdin IXa to bilirubin IXa, and is therefore an important enzyme for the process of heine catabolism [1]. In rat liver it is present in two molecular forms: a major molecular form 1 (MF1) and a minor molecular form 2 (MF2). When hemoprotein degradation is strongly enhanced by administration of CoCl2 or phenylhydrazine a change in the pattern of the molecular forms takes place: MF1 disappears and a new molecular form 3 (MF3), appears. The latter is a dimer of MF1 and

Abbreviations: MFI, MF2 and MF3, molecular form 1, 2 and 3, respectively; NEM, N-ethylmaleimide;NPhM, N-phenylmaleimide; PLP, pyridoxalphosphate; STi, soybean trypsin inhibitor; DTNB, 5,5'-dithiobis(2-nltrobenzoate). Correspondence: R.B. Frydman, Facultad de Farmaciay Bioquimica, Junin 956, Universidadde BuenosAires, BuenosAires,Argentina.

results form the oxidation of a thiol residue of MF1 to form a disulfide bridge in MF3 [2]. This oxidation is an enzyme mediated process when CoCl 2 is used, and can be reverted by reduction of MF3 with reduced thioredoxin [3]. We have recently established that of the three thiol residues of MF1, one is involved in the aforementioned oxidation, a second one (at the catalytic site), is involved in the reduction of biliverdin to bilirubin and the third one is not involved in either of the aforementioned processes [4]. The dimerizatinn of MF1 to MF3 which occurs in the rat liver under conditions of oxidative stress, results in an enzyme form which is much less sensitive to thiol reagents [5]. Hence, the formation of a disulfide bridge in MF3 introduces molecular changes in the enzyme which affect the thiol residue involved in the catalytic activity of MF3. It is well documented that a thiol-disulfide exchange plays a fundamental role in many metabolic and physiological functions, such as protein synthesis, protein degradation, reversible activation and inactivation of enzymes, synthesis of deoxyribose, and so on [6]. It has been recently shown that a 'sulfhydryl-switch' might regulate

322 RNA-protein interactions [7]. There is also mounting evidence that upon oxidative stre~ many proteins (often those with a nucleotide binding domain), have their cysteine, histidine and methionine residues damaged [8]. The dimerization of MF1 under conditions of oxidative stress could produce a conformational change where the catalytic site will be less exposed to oxidizing species. Such a possibility was explored with the help of polyclonal antibodies raised against the native purified rat liver MF1 of biliverdin reductase. The rational behind this approach was based on the fact that antibodies elicited by a native protein recognize epitopes localized on its surface. If the dimerization of MF1 will conceal previously exposed regions of the protein, it could be expected that this change in the protein conformation will be reflected in the masking of the antigenic determinants recognized by the anti-MF1 antibodies. We report below that rabbit anti-MF1 antibodies were able to inhibit the catalytic activity of MF1, but not of MF3, which was not recognized at all by the antibodies. The inhibition of the catalytic activity of MF1 by the latter could be prevented by addition of either the substrate or the cosubstrate. In addition, data are presented which show that chemical modifications of the lysine and arginine groups of biliverdin reductase which completely abolish the catalytic activity of MF1 [4], affected only to a small extent the catalytic activity of MF3. Materials and Methods Materials Biliverdin IXa and IXfl were prepared by the chemical oxidation of heroin IX and separated as described elsewhere [9]. NADPH, NADP +, NADH, NAD +, Nethylmaleimide (HEM), N-phenylmaleimide (NPhM), pyridoxalphosphate (PLP), phenylgiyoxal, butanedione, DEAE-cellulose, CM-cellulose, 5,5'-dithiobis(2-nitrobenzoate) and soybean trypsin inhibitor (STI) were from Sigma. Sephadex (3-25 and Agarose-hexanoylNADP + affinity gel were from Pharmacia. Trypsin (TCPK treated) and thermolysin were from Worthington. All the other chemicals used were of the highest purity available. The solvents used were from Merck. Methods Obtention and purification of the molecular forms of biliverdin reductase. The molecular forms of rat liver biliverdin reductase were obtained from normal and CoCI, treated female albino Wistar rats (180-200 g) and were separated as described elsewhere [2]. They were then purified to homogeneity as described for molecular form 1 [10]. The kidney and spleen biliverdin reductases were also prepared and purified as described

[101.

Preparation and purification of the converting enzyme. The converting enzyme was prepared from the livers of female Wistar rats (180-200 g) which were treated with CoC! 2 (160 mg/kg) for 2.25 h. The converting enzyme was obtained by solubilization with 0.5~ DOCA from the peroxisomes obtained as described [3]. The solubilized enzyme was ultracentrifuged at 150000 × g and further purified by a sequential use of DEAE-cellulose, CM-ceilulose, Sephacryl S-200 and FPLC. Enzyme assays. Biliverdin reductase activity was assayed at 37°C for 10 rain. The incubation mixture contained, in a final volume of 100 /~1:10 mM of potassium phosphate buffer (pH 7.4), 500 /~M of NADPH, 13 /~M of biliverdin IXa or biliverdin IXI] and enzyme (200 ng, specific activity: 3500 u n i t s / m r protein) unless otherwise indicated. Activities were determined by measuring the formation of bilirubin which was estimated from the d(A455nm -A520nm). This difference was found to be proportional to the concentration of bilirubin. An • = 50 raM- i . c m - i (ViSm~ = 455 nm) was used. The conversion of MF1 to FM3 was measured by preincubation of MF1 with the converting enzyme. The reduction rate of biliverdin IXa is twice as fast as the reduction rate of bifiverdin IXfl by MF1, while MF3 reduces both biliverdin isomers at the same rate. The transformation of MF1 to MF3 by the converting enzyme could therefore be followed by measuring the relative reduction rate of both biliverdin isomers. The incubation mixtures contained, in a final volume of 100 /~1:10/~mol of potassium phosphate buffer (pH 7.4), MF1 (200 ng), NAD + (100 nmol) and converting enzyme (10-50 ng). The mixture was preincubated for 10 min at 37°C after which NADPH (500/tM) and either biliverdin IXa or IXfl (13 /~M) were added and the activity was measured in a final volume of 200 /tl as described above. The increase in the reduction rate of biliverdin IXfl over that of a control (where either the converting enzyme or NAD + were omitted) was a measure of the presence of MF3. Analytical procedures. Protein concentrations were determined by the method of Bradford [11] using serum albumin as standard. Succinylation of the enzymes was performed by addition of suecinic anhydride (a 1000 molar excess) to 0.1 nmol (3.5 pC) of the enzyme (MF1 or MF3) in 0.2 ml of phosphate buffer (pH 7.4). The pH of the solution was maintained by addition of 1 M sodium hydroxide. After stirring for 30 rain, the succinylated enzymes were passed through a Sephadex (3-25 equifibrated with 100 mM phosphate buffer (pH 7.4). Afiquots of this enzyme were then used for activity and immunological determinations. Hydroxylamine was used to probe., so that the succinylation acylated the e-NH, of the lysine residues and did not form O- or S-succinyl-derivatives [12]. The incubations were performed with a 0.1 M final concentration of the hydrox-

323 ylamine (pH 9.0) at 0 - 4 ° C for 2 h and were then filtered through a Sephadex G-25. Modification of the enzymes with 2,3-butanedione was performed by preincubating the former (7.0 Fg) in 50 mM borate buffer (pH 8.5) at 370C for 5 rain in the dark in the presence or absence of the dione. Aliquots of the control and treated enzyme mixtures were then added to the biliverdin reduetase assay mixture at a 5-fold dilution or were used for immunological determinations. When the enzymes were modified with phenylglyoxai, a similar procedure as that described for butanedione was used except that the buffer was Hepes (pH 8.0). Controls were run under identical conditions, but in the absence of the modifying reagents. The thiol groups of biliverdin reductase were measured by titration with DTNB as described by DeLuca and McElroy [13]. The thiol content was calculated assuming an c = 13.6 m M -1 . c m -I at 412 n m for the reaction product. The reaction was complete after 30 min. Polyacrylamide gel electrophoresis (PAGE). Native gel electrophoresis was performed on 7.5~ polyacrylamide gels using the neutral discontinuous buffer system of Williams and Reisfeld [14]. The gel system consisted of a 3.7~ stacking gel in 100 raM Tris-PO4H 3 buffer (pH 6.0) and a 7.5~ resolving gel in 70 m M Tris-HCi buffer (pH 7.5). The running buffer was 8 m M Tris-barbital (pH 7.4). Denaturing SDS-PAGE was performed basically according to the method of Laemmli [15]. Samples were electrophoresed on 10 or 15~ polyacrylamide resolving gel in 0.37 M Tris-HCl buffer (pH 8.8), 0.1~ SDS. A 3~ stacking gel in 0.125 M Tris-HCl buffer (pH 6.8) plus 0.1~ SDS was used. The ~'unning buffer used was 50 m M Tris, 192 m M glycine (pH 8.3) plus 0.1~ SDS. The protein samples were usually separated on 10~ polyacrylamide gels except in the case of proteolytic digestions where 15~ polyacrylamide gels were used. Routinely, 75 ng of pure enzymes were run when silver staining of the gels was used, while 185 ng of the enzyme were used for Western blotting. In the case when denaturing (SDS containing) gels were run, the samples were boiled for 10 rain in cracking buffer (125 m M Tris-HCI, pH 6.8, 2~ SDS, 5~ 2-mercaptoethanol and 10~ glycerol) and loaded onto the gels using Bromophenol blue as a tracking dye. When the reduction of the disulfide bridge was not desired, the 2-mercaptoethanol was omitted from the cracking buffer. For protein detection, gels were stained with either Coomassie brilliant blue R-250 (0.25~ in 50~ methanol 10~ acetic acid), or with silver nitrate according to the method described by Merril et al. [16]. Antibodies. Antibodies against the rat liver biliverdin reductase MF1 were raised in a New Zealand female rabbit by injecting underneath the skin, parallel to the underlying muscle, a single dose of a highly purified molecular form 1 of bifiverdin reductase from rat liver -

(2.0 mg) in complete Freund's adjuvant. The sera were collected 1 month after injection. Preimmune serum was taken before the primary injection of the rabbit.

Immunodetection methods Western blotting. The proteins were transferred from polyacrylamide gels to nitrocellulose filters at 4 ° C at a constant current of 200 mA in an LKB Trans-Blot cell for 12 h of 1200 mA for 1 h as described [17]. The efficiency of the transfer was checked and confirmed by staining the polyacrylamide gels with silver nitrate after the transfer. The nitrocellulose filters were then washed with Tris-buffered saline (TBS) (50 m M Tris-HCl, p H 7.4, 150 m M NaCi) for 30 rain and blocked with 3~ BSA, 2~ glycine and 1 m M PMSF in TBS for 3 h to overnight at room temperature. The filters were then probed with anti-biliverdin reductase MF1 rabbit serum diluted 1 : 200 to 1 : 700 in 0.2~ BSA, 2 ~ glycine, 1 m M PMSF, 0.02~ sodium azide and TBS, p H 7.4 (dilution buffer). The membranes were then washed thrice with TBS containing 0.05~ Nonidet P-dO for 30 rain each to remove free and nonspecifically bound antibodies. Immunoreactive bands were then detected using a horseradish peroxidase finked Vectastain ABC kit or 125I-labelled Protein A. Treatment of the filters with antirabbit Vectastain ABC kit (Vector Laboratories, Inc-Burlingame, CA) was performed as recommended by the manufacturer using as peroxidase substrate a solution of 100 m M citrate buffer (pH 4.5), 20 m M imidazol, 0.5 rag/rat diaminebenzidine, 0.4 mg/ml NiCl 2 and 0 0 2 ~ H202. For dot blot analysis, the protein samples were spotted onto nitrocellulose using a BRL Hybridot apparatus and the same procedure as described for the immunodetection of the immnnoreactive bands in Western blots was used. Radioiodination of Protein A was performed as described [18]. The filters were incubated with (2-~)- 10 e d p m / m l of 125I-Protein A in dilution buffer for 2 - 4 h at room temperature, washed as described above and exposed to Kodak X-AR film at - 7 0 ° C overnight. Immunoprecipitation. When the inhibition of biliverdin reductase activity by the normal or immune serum was assayed, the latter was added to an enzyme sample to the final dilution indicated in each case and incubated for 30 rain at 4 ° C in a final volume of 100 FI. When the antigen-antibody complexes were collected, 50 FI of 10~ formalin-fixed S. aureus, strain Cowan I (Pansorbin), were added to the reaction mixture and incubated for 1 h at 4 ° C , after which the mixtures were spun for 10 min at 12000 rpm at 4oC. Activity and conversion were then measured in the supematant. Antigen-antibody complexes were also collected without addition of the Pansorbin. In this case the antibody was added to the enzyme in a dilution of 1:1000 and 1:2000 and was incubated for 60 min at

324 4 ° C in a final vol. of 200 pl. The complex was spun for 15 rain at 12000 rpm at 4 ° C . Activity was then measured in the supernatant and in the precipitate. Proteolytic

treatment

of MFI

biliverdin

reductase.

Peptide fragments were prepared by incubating the enzyme (3.5 pg) with either trypsin (20/tl, 1 m g / m l ) or thermolysin (20/tl, 1 mg/ml). Incubations were run in parallel at 3 7 ° C for times ranging between 5 to 180 rain. Aliquots were used for detection of catalytic activity and conversion to MF3. In the case of trypsin, STI (20 pl, 2 m g / m l ) was added before the assays. Parallel treated samples were used for PAGE and Western blot analysis. Results and Discussion Specificity o f t h e polycional antibodies

The sera of rabbits immunized against the native MF1 of rat liver biliverdin reductase was initially tested by Ouchterlony diffusion assays for production of antibodies that could recognize the three molecular forms of the liver enzyme. A single precipitin line was detected (after staining with Amido Black) only with the MF1 of the reductase, up to an enzyme dilution of 1 : 130. MF3 did not show any reaction with the antisera. Since MF3 is a dimer of MF1, the specificity of the antibodies was tested with the more sensitive Western immunoblotting technique, after electrophoresis under non-denaturing

A B*

B C

A"



C"

r

.~

.6

Fig. 1. Native gel electrophoresisof (A,A') biliverdinreductaseMFI; (B,B') MF3; and (C.C') MF3 treated with reduced thioredoxin. Biliverdinreductase(75 ns) were used for protein detectionwith silver nitrate (A, B and C). while 180 ng of the enzymeswere used for the Western blot analyses (A', B' and C'). MFI, used at the same concentrations as MF3, gave less intense bands with silver nitrate. Experimentaldetails are describedin Materialsand Methods.

a~d denaturing conditions. As can be seen in Fig. 1A when both MF1 (lane A) and MF3 (lane B) were run on PAGE under non-denaturing conditions and then stained with silver nitrate MF3 showed three bands of different intensities; the upper two were catalytically inactive dimers and trimers of this molecular form [19]; MF1 run as a single band. Upon elution of the main band of MF3 (lane B) and realectrophoresis, the three aforementioned bands were again obtained. The immunoblots of the native MF1 gave a purplish precipitation band (Fig. 1B, lane A ' ) which was coincident with the single band visualized by silver staining of the same gels (Fig. 1A, lane A). The antibodies recognized equally well MFI when the latter was electrophoresed and transferred to the membrane under denaturing conditions (data not shown). MF3 however, failed to show a precipitation band on blots corresponding to gels run under nondenaturing conditions (Fig. 1B, lane B'). Since MF3 can be transformed into MF1 by incubation with reduced thioredoxin (a reaction which implies the reduction of a disulfide bridge [2]), it was analyzed by the immunoblotting technique after this reduction step. A precipitation band identical to the one found with MF1 was then obtained (Fig. 1B, lane C'). When analyzed by silver staining it was found that the major part of MF3 was transformed into MF1 (Fig. 1A, lane C), although a small amount of the dimeric MF3 still remained (see arrow). Minor immunoreactive degradation products of MF1 can also be seen; they were, however, absent in the MF3 runs. Similar results, namely, that MF1 was recognized by anti-MF1 antibodies while MF3 was not, were obtained when both molecular forms were electrophoresed under non-reductive denaturing conditions. However, when the SDS-PAGE of both molecular forms were performed under reductive denaturing conditions (see Materials and Methods), it was found that MF3 became immunoreactive and was transformed into MF1 due to the reduction of the disulfide bridge. On the other hand, when MF1 was converted in vitro into MF3 (by incubation with the NAD+-dependent converting enzyme [3]), it lost its ability to be recognized by the antibodies. In order to further characterize the antibodies generated against the native MF1, their effect on the enzymatic activity of the different molecular forms of rat liver bifiverdin reductase was assayed. As can be seen in Fig. 2, rabbit antisera, at different dilutions, inhibited the activity of MF1 while it had no effect on the activities of MF3 and of MF2 (data not shown). Precipitation of the enzyme-antibody complex was performed by using a dilution of 1 : 1000 and 1 : 2000 of the rabbit antisera. The activity was assayed in the supernatant after separation of the enzyme-antibody precipitate. It was found that at a 1 : 2000 dilution no catalytic activity could be detected in the precipitate, while in the supernatant 2 0 - 2 3 ~ of the initial activity could be

325 the homogeneous kidney and liver MF1 reductases (Fig. 3). These results confirm our previous ones which showed that the purified liver MFI and the kidney enzyme were very similar [10]. Although the spleen and kidney reductases lack the ability of the liver MFI to be

10~ "/PX

~ 8c

oxidized to a disulfide-bridged dimer, they contain comm o n antigenic determinants. Hence, biliverdin re~ 6c

ductases from the different tissues appear to have a similar antigenic domain which becomes inaccessible upon dimerization of the liver enzyme.

f

g -g 2(

Q:

o 4;" ...... 10 20

2.5

30

35

40

-Log ~erum dilution

Fig. 2. Effect of increasing dilutions of anti-MF1 antisera on the conversion of MF1 to MF3 (xl and on the activities of MF3 (x) and MFi in the absence (o) and presence (e) of formalin-fixed Staphylo£OgC/tf a/4re/aff.

detected. This activity found in the supematant

was

similar to that found when the enzyme-antibody complex had not been separated by centrifugation (Fig. 2). This result suggests that the antibody very likely new tralizes the active site of the MF1 of biliverdin reductase. Immunoprecipimtion of the enzyme-antibody complexes with formalin-f~ed Staphylococcus aureus (Pansorbin) only slightly enhanced the inhibitory effect on the activity of MF1 when dilutions higher than 1:1000 were used. It had however no effect on the activity of the other molecular forms (Fig. 2). It was therefore concluded that the antibodies raised against MF1 did not recognize MF3. Hence the antigenic domain which is exposed in MF1 became inaccessible to the antibodies upon dimerization of the enzyme to MF3. The latter was not recognized by the anti-MF1 antisera when analyzed by Western immunoblotting and its cataiytic activity -vas not affected even at high antibody concentrations (dilutions up to 1 : 10). As was reporte5 else~,.:~ere [10], biliverdin reductase from rat spleen and kidney appears in a single molecular form which is very similar to MF1 from liver, but is not converted to an MF3-1ike dimer. By using the Western blotting techniques, it was found that the spleen and kidney enzyme cross-reacted strongly with the antisera generated against the liver MF1. Similar results were obtained when whole tissue extracts were assayed. When equal activities of purified liver MF1 and crude reductase from spleen or kidney were reacted with the antiserum, the inhibition of the enzymatic activities of these reductases showed similar kinetics, suggesting that the enzyme from the three tissues share an immunospecific antigenic domain similar to that of the hepatic MF1. lmmunoblotting of the crude spleen enzyme revealed a polypeptide with identical mobility as that of

Addition of substrate and cosubstrate as well as chemical modifications of the lysine and arginine residues impair the interaction of molecular form 1 with its antibodies We have recently shown that thiol, lysine and arglnine residues are essential for the catalytic activity of biliverdin reduetase, that the former two are part of the active site and that the thiol group covalently binds to biliverdin forming a bilirubinoid intermediate [4]. Since the activity of MF1 was completely abolished in the enzyme-antisera complex; the antibodies could either bind near or at the active site or alter the structure of the reductase in such a way that the former is no longer functional. Such an interaction should therefore be prevented by the addition of either the substrate or the cosubstrate or by the chemical modifications of the aforementioned amino acids. The effects of four different concentrations of the substrate, the cosubstrate or both of them on the neutrafization of the catalytic activity of MF1 by the antibodies are summarized in Table I. In all cases the addition of the substrate or the cosubstrate protected the enzymatic activity against its total inactivation by A A

B

n C

A'

B'

C'

i Fig. 3. Cross reactivity of rat biliverdin reduetases from different organs with the anti-MPl antisera. Non-denaturing PAGE was performed with: (A,A') spleen crude extract; (B,B*) purified kidney enzyme; anti (C,C') purified liver MF1 of bifiverdin reductase. The amounts used for silver staining were: A = 100 ng and B and C, 50 ng. For Western blotting (A'. B' and C ' ) twice the former amounts were used.

326 TABLE 1

C

2

3

4

Inhibition of biliverdin reductase activity by the rabbat anti-MFI antibodies: blocking effect of the substrates







~

The molecular form I of biliverdin reductase (1.1 pM, 7 pg protein) was preincubated for 15 rain at 4 ° C with the indicated substrate and cosubstrate concentrations. The anti-MFl rabbit antisera (l : 700) was added after this preincubation and the mixture w':~spreincubated for an additional 5 rain at 37°C, after which the incubation mixture was completed to the concenL',ations indicated in Materials and Methods. Controls of the enzyme preincubated in the absence of the substrate or cosubstrate were also performed. Preincubated with

(pM)

Biliverdin

Biliverdin

NADPH NADPH



5

6

I

!

7

8

GU HCI

Antibody (1/700)

Activity nmol bilirubin formed/15 min

70 control

+

6.6 + 0.30 none

100 0

l.l 5.5 11.0

-

i .1

+

3.3 5.5 ILO

* + +

6.5±0.30 6.4±0.25 6.6±0.30 1.3 + 0.10 2.34-0.10 4.3 ± 0.20 6.7+0.20

100 100 100 20 35 65 100

+ + + +

6.5 + 0.30 6.3 ± 0.25 3.2+0.15 5.6±0.30 6.7±0.25 6.6:t:0.30

100 100 50 85 100 100

20 250 20 50 100 250

e,e'

the antibodies. The effect of the substrate and the cosubstrate on the immunoreactivity of form 1 was analyzed by using dot-blot techniques and it was found

Fig. 4. Effect o f the substrates and products on the intmunoreactivlty

of the biliverdin reductase MF1. The enzyme (185 ng) was preincubated for 10 rain at 4 ° C with: biliverdin (dot 2); NADP:i (dot 3); biliverdin plus NADPH (dot 4); NADP + (dot 7); and bilirubin (dot 8). Samples of the dots 5 and 6 were preincubated as were the samples in the dots 2 and 3. respectively, but guanidinm hydrocldoride (5 M) was added after preincubation and the samples were incubated for additional 10 rain.

to be impaired by about 70% (Fig. 4, dots 2, 3 and 4). When the incubation mixtures were treated with 5 M guanidinium hydrochlofide the enzyme which was preincubated with NADPH was fully recognized (Fig. 4, dot 6), while that preincubated with biliverdin did not recover its ability to be fully recognized by the antibodies (Fig. 4, dot 5). Since biliverdin was shown to form a covalent adduct with the enzyme [4], while the binding of NADPH was of a non-covalent nature, the former could withstand the nitrocellulose washings and theror fore still impair the interaction with the antibodies. When the reaction products were tested, it was found

A

o

10C

8C

Western blot

S i l v e r stoin

9C

.~.\

A

"~--....

B,

C

A'

B'

C'

113

136

7C

6

6C

o g

o ""---o

5C 4C 3C

c

2c

,

1C m

0

~.~, 200

, 400

, ,0o,,oo, (500

Esulfhyery~ re=aent~ , mon/tool enzyme

NEM

0

113

136

0

Fi8. 5. Alkylation of liver biliverdin reductases with N-phenyl (O) and N-ethyl (o) malalmide. (A) Differential inhibition of the catalytic activities of MF1 ( ) and MF3 (. . . . . . ) and of the enzymatic interconversion of MFI to MF3 (A). Inset, effect of the increasing molar excess of N-ethylmaleimide (NEM) on the interaction of: the purified liver MF1 (L) and kidney (K) biliverdin reductase with the anti-MFl antibodies. The reaction was detected by using t2Sl-protein A as described in Materials and Methods. (B) Native gel elecLrophoresis of liver MF1 treated with the indicated molar excesses of NEM. The amounts of enzyme used for silver staining and Wastem blots were 50 and 185 ng, respectively.

327 that NADP + completely inhibited the antigen-antibody complex formation (dot 7) while bifirubin did not affect it at all (dot 8). The fact that the latter did not impair the interaction of the reductase with the antibodies while the effect of NADP + was very marked could be explained by the known sequential mechanism of the NAD(P)-dependent dehydrogenases, where the enzyme dinucleotid¢ complex is the first to form and the last to dissociate [20]. To estabfish the possible involvement of thiol residues in the formation of the enzyme-antibodies complex, MF1 was alkylated with N-ethyimaleimide (NEM) and with N-phenyimaleimide (NPhM). We have shown elsewhere [4] that by using a 90-fold molar excess of NEM over that of the enzyme, a single thiol group - the one involved in the dimerization of MF1 to MF3 - was alkylated. The thiol group involved in the catalytic activity was completely alkylated only when a 250-fold excess of the reagent was used, and the catalytic activity was then totally abolished. Fig. 5 shows the effect of the increasing concentrations of these alkylating reagents on the activities of both, MFI and MF3. As can be seen, the latter was considerably less sensitive to the thiol reagents suggesting that the thiol group essential for the catalytic activity is less exposed and less accessible to the reagents. When a 113-fold molar excess of NEM (Fig. 5A, 1st arrow) was used and only one thiol residue was alkylated, no change in the immunoreactivity of MF1 could be detected after Western blot analysis (Fig. 5B). By rising the NEM concentration to a 136-fold molar excess no additional thiol groups were found to be alkylated [4] (Fig. 5A, 2nd arrow); the immunoreactivity of MF1 was however entirely abolished. Similar results were obtained when the antibodies, bound to the alkylated enzyme were measured by using t2SI-Protein A. The same was true for the kidney enzyme (Fig. 5A, inset). Hence, the loss of recognition of the NEM-treated MF1 by the antibodies should be attributed to the alkylation of amino acid residues other than thiols which are essential for the enzymatic activity. The only other NEM-reactive amino acid which was identified at the catalytic site was lysine. Therefore pyridoxal posphate was assayed as an effectot of the biliverdin reductase activity and it was found to inhibit 65~ of MF1 activity at a 10 mM concentration while it did not affect the activity of the dimeric form MF3 (data not shown). Similar results were obtained when both molecular forms were succinylated with succinic anhydride. When a 1000-fold molar excess of the reagent was used, the activity of the MF1 was completely lost while the activity of MF3 was not affected (data not shown). The determination of the effects of these chemical modifications of the enzyme's lysyl groups on antibody binding was performed using dot blot t.,:cLniques. As can be seen, the recognition of the succinylated enzyme was strongly impaired (Fig. 6,

C

2

3

4

5

6

7

8

Fig. 6. Effect of the specific modification of lysine and arginine residues on the immunoreactivity of MFL The enzyme (185 rig) was modified as indicated and its immunoreactivity was investigated by the dot-blot technique. The experimental details are described in Methods. (C), control; (2), succinylated enzyme; (3), enzyme plus pyridoxal-phosphate; (4), the same as 3 but treated with NaBH4; (5) phenylglyoxal treated enzyme; (6) hntanedione treated enzyme; (7) same as 5 but preineabated with pyridoxal phosphate; and (8) enzyme treated with NaBH4.

dot 2). Treatment of MF1 with PLP did not affect the immunoreactivity to any appreciable degree (Fig. 6, dot 3), very likely due to the reversibility of the PLP-lysine Schiff base. However, when the latter was reduced with sodium borohydride the recognition was entirely abolished (Fig. 6, dot 4). Modification of the argi,;.ne residues by phenylglyoxal or butanedione also impaired the recognition of MF1 by the antibodies (Fig. 6, dots 5 and 6). To discard the possibility that the impairment of the recognition of the enzyme after its treatment with phenylglyoxal was due to the modification of lysine residues, the latter were protected with PLP prior to the phenylglyoxal treatment. We mentioned above that PLP formed a reversible bond which did not impair by intself the immunoreactivity of the reductase (Fig. 6, dot 3). Therefore, the lack of recognition found by the treatment of the PLP-proteeted enzyme with phenylglyoxal must be attributed to the modification of the arglnyl groups. Modification with butanedione gave similar results (Fig. 6, dot 6). When the effects of these two arginyl modifying reagents were assayed on the catalytic activities of MF1 and MF3, the results indicated that the antigenic enzyme was much more sensitive to inhibition while the activity of MF3 was less affected. At a 10 mM concentration, phenylglyoxal and butanedione inhibited the catalytic activity of MF1 by 60 and 75~, respectively; while it only inhibited 18 and 25~ of the activity of MF3. At a 20 mM concentration, both reagents inhibited more than 90~ of the activity of MF1, while MF3 lost 25 and 35~g of its activity (data not shown). This is to be expected if the catalytic site of the latter is less accessible to the modifying reagents. Lysine and arglnine residues are therefore involved in the antigen-antibody interaction. This fits into the well known fact that in dinucleotide binding enzymes the tertiary structure has a well conserved feature. ('the Rossman fold') where the nucleotide interacts with the protein [21,22]. This loop contains arglnine and lysine residues [21,23] which are known to be particularly immunogenic because of their size and polar nature [24], and were shown to interact with acid residues in the antibody's paratope [25]. We have recently shown that the active site of biliverdin reductase strongly re-

328 sembles that of a dinucleotide dependent dehydrogenase [4], a fact which also lends support to the above mentioned antigenic properties. Neutralization of the enzymatic conversion of MFI to MF3 by the antisera produced against MFI In the presence of highly diluted antisera (1 : 5600) the enzymatic conversion of MF1 into MF3 was completely inhibited, while the catalytic activity of the former was inhibited by only 50% (Fig. 2). Since the conversion of MF1 to MF3 results from the oxidation of a single thiol residue which is not involved in the catalytic activity, and since the antisera are probably directed against a domain which becomes hidden upon dimerization, it is conceivable that the thiol involved in the dimerization lies within the region of the major antigenic determinant of the enzyme. It could therefore become inaccessible to the action of the converting enzyme upon antibody binding. The fact that the conversion was inhibited to a higher degree than the catalytic activity could be due to the fact that once half of the antigen molecules react with the antibodies leading to a 50~ loss in the catalytic activity, the dimerization reaction will be impaired by at least 75~. In order to elucidate if the thiol residue involved in the dimerization is also part of the antibody recognition site of MF1, the latter was alkylated with a 90-fold molar excess of N E M which completely abolished the conversion of MF1 to MF3, but not its catalytic activity [4]. The alkylated enzyme was then treated with the converting enzyme, subjected to gel electrophoresis and analyzed both by silver staining and by Western immunoblotting. When MF1 was transformed enzymaticaily into the MF3 by the action of the converting enzyme plus N A D + [3], the latter form did not react with the antibodies. A fain~ band detected by Western blot was due to the residual MF1 which was not transformed into MF3 (data not shown). When MF1 was treated with a 90-fold NEM excess, silver staining showed that it was not transformed enzymaticany into MF3 any longer, it was, however, fully recognized by the antibodies. Hence, despite the bulky NEM group the alkylated MF1 was recognized by the antibodies. At higher N E M concentrations (a 136-molar excess of the reagent over the enzyme), the recognition of the alkylated MF1 by the antibodies was completely lost. These results show that the thiol residue involved in the formation of the disulfide bridged dimer (MF3) is not an essential part of the antigenic determinant of the protein. It is conceivable that the antibodies interfere with the conversion (Fig. 2) simply by steric hindrance. Effect of proteolytic digestion of M F I on its reaction with the antibodies It has been shown that lysine and arginine residues were essential for the activity of MF1 [4] and for its

recognition by the antibodies (see above). We therefore investigated the effect of trypsin on the reductase, since about 10~ of its total amino acid content are basic residues. Upon incubation of MF1 with trypsin for as little as 5 min, all the activity was lost. However, when the digestion was performed with thermolysin, which cleaves at the hydrophobic Val, Leu or lie residues (25% of the amino acid composition of MF1), a 27 kDa fragment was obtained which retained all the enzymatic activity, was recognized by the antibodies and could be converted to a dimer with similar catalytic properties as those of MF3 (data not shown). The transformation of MF1 into the 27 kDa polypeptide was complete after 30 min and no smaller fragments were detected by silver staining. No further changes of the polypeptide occurred even if the incubation with the proteolytic enzyme was prolonged for 180 min. Conclusions This paper reports that it was possible to raise antibodies against molecular form 1 of rat liver biliverdin reductase, which also recognized molecular form 1 of the enzyme from spleen m~d kidney (Fig. 3). The antibodies did not recognize however the S-S bridged molecular form 3 of the liver enzyme, which is formed by the in vivo or in vitro oxidation of molecular form 1 (Fig. 1). This lack of recognition of molecular form 3 by the antibodies must be due to a conformational change which buries the antigenic domain of molecular form 1 upon dimerization. Such a conformational change was also evident when the thiol, arginine and lysine residues of molecular forms 1 and 3 were chemically modified (Fig. 5). The antibodies, even at high dilutions, neutralize the enzymatic activity of molecular form 1. The fact that this inhibition was impaired by the presence of the substrate or the cosubstrate (Table I) suggests that the antigenic domain may be localized near the active site or is part of the same. We have shown elsewhere [4] that lysine and arginine residues were essential for the catalytic activity of the enzyme. It was therefore of interest to find that chemical modifications of the aforementioned amino acids, which led to the inactivation of the enzymatic activity of form 1, also impaired the formation of the enzyme-antibody complex. The possibility however always remains, that the loss of antigenicity in molecular form 1 may be caused by a conformational alteration of the enzyme which takes place upon addition of the substrates [26] or by the chemical modifications of the amino acids. In this case the loss of antigenicity will only be indirectly related to the modifications of the catalytic site, which could alter an antigeaic sequence localized at a different domain of the protein. In the absence of data on the primary sequence of biliverdin reductase or of X-ray diffraction data on

329

its tridimensional structure, both alternatives remain an open question. Acknowledgments This work was m a d e possibly by a g r a n t (GM-11973) f r o m the National Institutes of Health (PHS). Support from C O N I C E T (Argentina) is also acknowledged. W e thank Dr. J. Rosenfeld for his help in p e r f o r m i n g the electrophoretic runs. W e also t h a n k Dr. E. Malchiodi for his help in obtaining the i m m u n e sera.

References 1 Frydman, R.B. and Frydman, B. (1987) Accounts Chem. Res. 20, 250-256. 2 Frydman, R.B., Tomaro, M.L.. Awruch, J. and Frydman, B. (1983) Biochim. Biophys. Acta 759. 257-263. 3 Frydman. ILB. Tomato, M.L. Awruch, J. and Frydman, B. (1984) Biochem. Biophys. Res. Commun. 121. 249-254. 4 Frydman, J., Tomarn" M.L.. Rosenfeld, J. and Ft~jdman, R.B. (1990) Biochim. Biophys. Aeta 1040, 119-129. 5 Frydman, ILB., Tomato, M.L., Awruch. J. and Frydman" B. (1982) Biochim. Biophys. Res. Commun. 105. 752-758. 6 Meister, A. and Anderson, M.E. (1983) Annu. Rev. Biochem. 52, 711-760. 7 Hantze, M.W., Rouault, T.A., Harford, J.B. and Klansner, R.B. (1989) Science 244, 357-359. |SiStadtman" E.R. (1986) TIBS !1.11-12. 9 Frydman, R.B.. Awruch, J.. Tomato, M.L. and Frydman, B. (1979) Biochim. Biophys. Res. Commun. 87, 928-935.

10 Cascone. O., Frydman, R.B., Ferrara, P., Tomato, M.L. and Rosenfeld0 J. (1989) Eur. J. Biochem. 179, 123-130. 11 Bradford, M. (1976) Anal. Biechem. 72, 248-254. 12 Loverde, A. and Striumauer, P. (1968) J. Biol. Chem. 243, 57795787. 13 DeLuea, M. and MeEIroy, W.D. (1966) Arch. Biechem. Biophys. 116, 103-107. 14 Williams, D.E. and Feisfeld, R.A. (1964) Ann. N.Y. Acad. Sci. 121,373-381. 15 Laemmli, U.K. (1970) Nature 227, 680-685. 16 Merril, C.R., Goldman. D. and Van Keuven (1984) Methods Enzymol. 104, 443-446. 17 Towbin, H., Straehelin, T. and Gordon, J. (1979) Proc. Natl. Aead. Sci. USA 76. 4350-4354. 18 McConahey. P.J. and Dixon, FJ. (1980) Methods Enzymol. 70, 210-213. 19 Frydman" R.B., Tomaro, M.L., Rosenfeld, J., Awruch, J., Sambrotta, L., Valasinas, A. and Frydman, B. (1987) Biochim. Biophys. Acta 916, 500-511. 20 Walsh, C. (1979) Enzyme Reaction Mechanism, W.H. Freeman, San Francisco, pp. 311-357. 21 Richardson, J.S. (1977) Nature 268, 495-500. 22 Rossmann" M.G., Moras, D. and Olsen" K.W. (1974) Nature 250, 194-199. 23 Rossmann" M.G. and Gran" V. (1982) The Pyridine Nucleotide Coenzymes, Academic Press, New York, pp. 135-187. 24 Nowotny, J., Handschumacher, M. and Bruccorleri, R.E. (1987) Immunology Today 8. 26-31. 25 Davies, D.R., Sheriff, S. and Padlan" E.A. (1989) Abstracts of papers presented at the LIV/CSH Symposium on Quantitative Biology, Cold Spring Harbonr, N.Y., p. 39. 26 Ladner, J.E., Kitchell, J.P., Honzatko, R.B., Ke, H.M., Volz, K.W., Kalb. AJ., Ladner, R.C. and Lipscomb, W.N. (1982) Proc. Natl. Acad. Sci. USA 79, 3125.

An analysis of the topography of molecular forms 1 and 3 of rat liver biliverdin reductase using polyclonal antibodies.

Rat liver biliverdin reductase exists in two molecular forms. The major one (molecular form 1) is transformed, under conditions of oxidative stress in...
714KB Sizes 0 Downloads 0 Views