ELSEVIER Biochimica et Biophysica Acta 1246 (1995) 167-177

B/t Biochi~ic~a et Biophysica ~ta

Rat liver nucleoside diphosphosugar or diphosphoalcohol pyrophosphatases different from nucleotide pyrophosphatase or

phosphodiesterase I: substrate specificities of Mg 2÷- and/or Mn2÷-dependent hydrolases acting on ADP-ribose

Jos6 Canales, Rosa Maria Pinto, Maria Jesfis Costas, Maria Teresa Hern~ndez 1, Asunci6n Mir6 2, Diego Bernet, Ascensi6n Fernandez, Jos6 Carlos Cameselle *

Departamento de Bioqulraic~ y Biolog{a Molecular y Gen~tica, Facultad de Medicina, Universidad de Extremadura, E06080 Badajoz, Spain

Received 19 July 1994; accepted 19 September 1994

Abstract

Three rat liver nucleoside(5') diphosphosugar (NDP-sugar) or nucleoside(5') diphosphoalcohol pyrophosphatases are described: two were previously identified in experiments measuring Mg2+-dependent ADP-ribose pyrophosphatase activity (Mir6 et al. (1989) FEBS Lett. 244, 123-126), and the other is a new, Mn2÷-dependent ADP-ribose pyrophosphatase. They are resolved by ion-exchange chromatography, and differ by their substrate and cation specificities, K M values for ADP-ribose, pH-activity profiles, molecular weights and isoelectric points. The enzymes were tested for activity towards: reducing (ADP-ribose, IDP-ribose) and non-reducing NDP-sugars (ADP-glucose, ADP-mannose, GDP-mannose, UDP-mannose, UDP-glucose, UDP-xylose, CDP-glucose), CDP-alcohols (CDP-glycerol, CDP-ethanolamine, CDP-choline), dinucleotides (diadenosine pyrophosphate, NADH, NAD +, FAD), nucleoside(5') mono- and diphos- phates (AMP, CMP, GMP, AD]?, CDP) and dTMP p-nitrophenyl ester. Since the enzymes have not been purified to homogeneity, more than three pyrophosphatases may be present, but the co-purification of activities, thermal co-inactivation, and inhibition experiments give support to: (i) an ADP-ribose pyrophosphatase highly specific for ADP(IDP)-ribose in the presence of Mg 2÷, but active also on non-reducing ADP-hexoses and dinucleotides (not on NAD +) when Mg 2+ was replaced with Mn 2 +; (ii) a Mn2+-dependent pyrophos- phatase active on ADP(IDP)-ribose, dinucleotides and CDP-alcohols; (iii) a rather unspecific pyrophosphatase that, with Mg 2÷, was active on AMP(IMP)-containing NDP-sugars and dinucleotides (not on NAD+), and with Mn 2÷, was also active on non-adenine NDP-sugars and CDP-alcohols. The enzymes differ from nucleotide pyrophosphatase/phosphodiesterase-I (NPPase/PDEaseI) by their substrate specificities and by their cytosolic location and solubility in the absence of detergents. Although NPPase/PDEaseI is much more active in rat liver, its known location in the non-cytoplasmic sides of plasma and endoplasmic reticulum membranes, together with the known cytoplasmic synthesis of NDP-sugars and CDP-alcohols, permit the speculation that the pyrophosphatases studied in this work may have a cellular role.

Keywords: Nucleoside diphosphosugar; ADP-ribose; CDP-choline; Pyrophosphohydrolase; Nucleotide pyrophosphatase; Phosphodiesterase-I; Manganese (II); Magnesium (II); (Rat liver)

Abbreviations: ADP-glucose, adenosine(5') diphospho(1)-ct-D-glucose; ADP-mannose, adenosine(5') diphospho(1)-a-D-mannose; ADPRibase, each one of the ADP-ribose pyrophosphatases studied in this work; ADP-ribose, adenosine(5') diphospho(5)-D-ribose; AppA, adenosine(5') diphospho(5')adenosine; CDP-alco]~ol, cytidine(5') diphosphoalcohol; CDP-choline, cytidine(5') diphosphocholine; CDP-ethanolamine, cytidine(5') diphosphoethanolamine; CDP-ghicose, cytidine(5') diphospho(1)-a-D-glucose; CDP-glycerol, cytidine(5') diphospho(sn-3)glycerol; GDP-mannose, gnano- sine(5') diphospho(1)-Ot-D-mannose; IDP-ribose, inosine(5') diphospho-(5)-D-ribose; NDP-sugar, nucleoside(5') diphosphosugar; NPPase/PDEaseI, nu- cleotide pyrophosphatase/phosphodiesterase I; UDP-glucose, uridine(5') diphospho(1)-a-D-glucose; UDP-mannose, uridine(5')-diphospho(1)-Ot-D-man- nose; UDP-xylose, uridine(5')dipho:~pho(1)-ot-o-xylose.

* The first three authors contributed equally to this work. * Corresponding author. Fax: + 34 24 271304. 1 Present address: Departamento de Tecnologla Agroalimentaria, Servicio de Investigaci6n y Desarrollo Tecnol6gico de la Junta de Extremadura,

Badajoz, Spain. 2 Present address: Departamento de Patologla y Cllnicas Humanas (M6dicas), Facultad de Medicina, Universidad de Extremadura, Badajoz, Spain.

0167-4838/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0167-4838(94)00191-X

168 J. Canales et al. / Biochimica et Biophysica Acta 1246 (1995) 167-177

1. Introduction

NPPase /PDEase I is a single, unspecific enzyme hydro- lyzing 5'-nucleotide derivatives with pyrophosphate or phosphodiester linkages; i.e., its specificity corresponds to the sum of the Enzyme Commission entries nucleotide pyrophosphatase (EC 3.6.1.9) and phosphodiesterase-I (EC 3.1.4.1). In mammals, it is a glycoprotein located in the external surface of the plasma membrane and in membrane ves ic les o f in t race l lu la r or ig in [1 -5 ] . Recen t ly , NPPase /PDEase I has been identified with the plasma cell differentiation antigen PC-1 [6-8]. Its biological role is unknown. F rom the e n z y m a t i c po in t o f v iew, NPPase /PDEase I is responsible for the major hydrolytic activities on NDP-sugars in homogenates from liver and other tissues [9-13], and it has been suggested that it may have a role in controlling mammalian glycosylations. For example, in regenerating rat liver, increased glycosylation capaci ty is para l le led by decreased act ivi t ies of NPPase /PDEase I enzymes [14,15]; and plasmacytoma cells over-expressing NPPase /PDEase I activity are defi- cient in the pathway leading to N-linked oligosaccharides and protein glycosylation [16,17]. Interestingly, these NPPase /PDEaseI - r i ch plasmacytoma cells have normal w h o l e cel l U D P - g l u c o s e levels , sugges t ing that NPPase /PDEase I does not have access to the major cyto- plasmic pool of NDP-sugars and, therefore, the enzyme

may affect glycosylation by acting on susceptible precur- sors at non-cytoplasmic compartments.

As discussed by others, insufficiently characterized ac- tivities reported as NDP-sugar pyrophosphatases may cor- respond to NPPase /PDEaseI ( - l ike) enzyme(s) [11,15,18]. In fact, there is a remarkable absence of reports showing animal NDP-sugar /a l coho l pyrophosphatase activities free from NPPase /PDEase I activity (typically assayed with dTMP p-ni trophenyl ester as the substrate). Thus far, it is not yet clear whether all or some NDP-sugars (or NDP-al- cohols) may be attacked by pyrophosphatases other than NPPase /PDEaseI . In an earlier study in rat liver, two Mg2+-dependent enzymes with ADP-ribose pyrophos- phatase (EC 3.6.1.13) activity were partially purified from membrane-free supernatants and were named as ADPRib- ase-I and ADPRibase-II . A limited study of substrate specificity in the presence of Mg 2+ showed that ADPRib- ase-I could be a specific ADP-ribose pyrophosphatase, whereas ADPRibase-II showed activity on other substrates and could be a non-specific pyrophosphatase [19]. This article is aimed to: (i) demonstrate the occurrence of a new Mn2+-dependent ADP-r ibose pyrophosphatase (AD- PRibase-Mn); (ii) report new substrate and cation speci- ficity data of ADPRibase-I and ADPRibase-II ; (iii) claim attention to the occurrence of more or less specific, cyto- p l a s m i c p y r o p h o s p h a t a s e s d i f f e r e n t f r o m NPPase /PDEase I , active towards NDP-sugars a n d / o r

Table 1 Purification of ADPRibase-I, ADPRibase-Mn and ADPRibase-II from rat liver

Step Activity Protein Specific activity Purification Recovery (mU) (mg) (mU/mg) factor (%)

Mg 2+ Mn 2+ Mg 2+ Mn 2+ Mg 2+ Mn 2+ Mg 2+ Mn 2+

Common initial steps (1) 100000 X g supernatant 3715 2136 1190 3.1 1.8 1 1 100 100 (2) (NH4)2SO 4 30-60% fraction 2745 1465 684 4.0 2.1 1.3 1.2 74 69 (3) Sephadex G-100 2037 920 251 8.1 3.7 2.6 2.1 55 43

ADPRibase-I (4) DEAE-cellulose 1202 n.d. 22 55 n.d. 18 n.d. 32 n.d. (5) Hydroxyapatite 566 n.d. 4.5 126 n.d. 41 n.d. 15 n.d. (6) Chelating Sepharose 413 n.d. 0.9 439 n.d. 142 n.d. 11 n.d.

ADPRibase-Mn (4) DEAE-ceUulose n.a. 528 6 n.a. 88 n.a. 49 n.a. 25 (7) Reactive blue 2-Sepharose CL-6B n.a. 382 0.056 n.a. 6821 n.a. 3789 n.a. 18 (8) Q-Sepharose n.a. 111 0.006 n.a. 18 197 n.a. 10 109 n.a. 5

ADPRibase-ll (4) DEAE-cellulose 491 n.d. 0.92 534 n.d. 172 n.d. 13 n.d. (9) Hydroxyapatite 452 n.d. 0.34 1329 n.d. 430 n.d. 12 n.d. (10) Reactive green 19-Agarose 269 n.d. 0.021 12 810 n.d. 4132 n.d. 7 n.d.

The results summarize one purification run performed starting from 29 g of rat liver. The purification procedure for the three enzymes followed a common path up to the DEAE-cellulose chromatography, when they became separated each other in a gradient elution (Fig. 1B). Afterwards, the three enzymes were purified independently. Experimental details are described in Section 2. The Mg2+-dependent and/or the Mn2+-dependent hydrolytic activities towards ADP-ribose were assayed in the presence of 5 mM MgC12 or 5 mM MnC12, respectively, as indicated in the table. n.d., Not determined. n.a., Not active.

J. Canales et al. /Biochimica et Biophysica Acta 1246 (1995) 167-177 169

CDP-alcohols, and that may have been disregarded by the quantitative predominance of NPPase/PDEaseI, and/or escaped observation due to specific cation requirements.

2. Materials and methods

2.1. Materials

Female Wistar rats (200-250 g) fed ad libitum were used throughout this work. Alkaline phosphatase (EC 3.1.3.1) and adenosine deaminase (EC 3.5.4.4) were from Boehringer (cat. No. 108146 and 102091); substrates (ex- cept NAD + which was from Boehringer), phenylmethyl- sulfonyl fluoride, AMP deaminase (EC 3.5.4.6; cat. No. A 8384) and dye-ligand affinity chromatography media were purchased from Sigma; Tris, EDTA, ( N H 4 ) 2 5 0 4 , KC1, MgC1 e and MnCI 2 were from Merck; Sephadex, Q-Sep- harose and Chelating Sepharose gels, as well as Polybuffer exchanger PBE 94 and ]?olybuffer 74 for chromato- focusing, were from Phannacia; DEAE-cellulose DE52 was from Whatman; hydroxyapatite Bio-Gel HTP was from Bio-Rad.

2.2. Enzyme assays

In the standard assay for phosphohydrolytic activities on substrates without terminal phosphoryl groups, the for- mation of products with terminal phosphoryl groups was coupled to alkaline phosphatase, and the resulting Pi was measured with a dodecyl sulfate-ascorbate-molybdate reagent [20,21]. The incubations were carried out at 37°C, in assay mixtures of 0.1 ml or 0.2 ml that, unless otherwise stated, contained 50 mM Tris-HC1 (pH 7.5), 0.35 mM substrate, 0.5 /zg alkaline phosphatase, and either 5 mM MgC12 or 5 mM MnCI 2. The assay mixtures of all the experiments carried out with materials from the DEAE-cel- lulose and later steps (steps 4-10, Table 1) were supple- mented with 1 mg /ml bovine serum albumin (Boehringer, Fraction V). The reactions were stopped with 0.75 ml or 1.45 ml of the molybdate reagent and A820 was measured after 20 min at 45°C. Blanks and Pi standards were pro- cessed in parallel. The standard assay for phosphohydro- lytic activity on substrates with terminal phosphates was as above except that alkaline phosphatase was omitted. After purification on hydroxyapatite columns, ADPRibase-I (step 5, Table 1) and ADPRibase..II (step 9, Table 1) are accom- panied by high Pi concentrations which impeded the use of the above assays. In those cases, to measure the elution profile of ADP-ribose pyrophosphatase activities before removing Pi by dialysis, a spectrophotometric method cou- pled to AMP-deaminase was used [22]. In some cases, to study the saturation kinetics of ADPRibase-I, the hydroly- sis of ADP-ribose was coupled to alkaline phosphatase and adenosine deaminase, and t]ae reaction was followed spec- trophotometrically at 265 nm and 37°C [22]. In some

occasions with ADP-ribose, NAD +, NADH, or ADP as the substrate, the hydrolytic reaction was assayed by HPLC, in incubation mixtures without auxiliary enzymes, follow- ing the accumulation of AMP, oxidized or reduced nicotin- amide nucleotide, or AMP, respectively. One unit enzyme activity represents always the hydrolysis of 1 ~mol sub- strate per min under the conditions of the assay. The results were linear with both time and added enzyme under all the conditions used. Blanks without enzyme and/or substrate were used throughout to correct measurements and obtain true enzyme activity results. This is particularly relevant because Nunez and Barker described the metal-ion catalyzed decomposition of some NDP-sugars, Mn 2+ be- ing much better catalyst than Mg 2+ in this respect [23].

2.3. Purification of ADP-ribose pyrophosphatases

The first part of the purification, up to the DEAE-cel- lulose step inclusive (steps 1-4, Table 1), was identical for ADPRibase-I, ADPRibase-Mn and ADPRibase-II. Rat liv- ers were homogenized in freshly prepared 50 mM Tris-HCl (pH 7.5), 0.5 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride, in the proportion of 2 ml of buffer per gram of fresh tissue. Homogenates were centrifuged for 15 min at 27 000 × g and, after discarding the pellet, a second cen- trifugation was carried out for 1 h at 100000 X g. The supernatants were fractionated by ammonium sulfate pre-

60 A

4 0

20

20

15

10

0

I~ 50 150 250 350 [_.,

. /74 0.4 ~ 3 0 ~ D / I z

< 2 0 ~ ~ / M n II 4 2 O 0.2~

10 0.1

o o o.o

0 2 0 0 4 0 0 6 0 0

ELUTION VOLUME, m L

Fig. 1. Elution profiles of Mg z+- and Mn2+-dependent ADP-ribose pyrophosphatase activities upon (A) gel filtration in Sephadex G-100 and (B) ion-exchange chromatography in DEAE-eellulose. Experimental de- tails in Section 2. ([]) Mg 2 +-dependent and ( • ) Mn2+-dependent phos- phohydrolytie activity on ADP-ribose; (O) absorbanee at 280 nm; ( - - - - - - ) KC1. I, ADPRibase-I; II, ADPRibase-II; Mn, ADPRibase-Mn.

170 J. Canales et aL /Biochimica et Biophysica Acta 1246 (1995) 167-177

cipitation (30-60% saturation) and the pellets were dis- solved in fresh buffer as above, after which they were loaded onto a 2 cm X 100 cm Sephadex G-100 column equilibrated in and eluted with 20 mM Tris-HC1 (pH 7.5), 0.5 mM EDTA at a rate of about 0.5 ml/min. Fractions containing the major part of ADP-ribose pyrophosphatase activity (Fig. 1A, elution volume 155-210 ml) were pooled and loaded onto a 1.6 cm X 39 cm DEAE-cellulose col- umn equilibrated in 20 mM Tris-HCl (pH 7.5). The col- umn was first washed with 20 mM Tris-HC1 (pH 7.5), 0.5 mM EDTA, 50 mM KC1 and then eluted at about 0.5 ml/min with a 50-400 mM KC1 linear gradient in the same buffer. The three ADP-ribose pyrophosphatases were recovered, well separated from each other, in the fractions collected during that gradient elution (Fig. 1B). After this, the three ADP-ribose pyrophosphatases were separately purified as follows.

ADPRibase-I from the DEAE-cellulose step was di- rectly applied (step 5, Table 1) to a hydroxyapatite column (2 cm X 4.5 cm) equilibrated in 1 mM NaCI and immedi- ately washed with 40 mM sodium phosphate (pH 7), 0.5 mM EDTA, at a rate of 0.2 ml/min, conditions under which most of ADPRibase-I activity remained bound to the gel. The enzyme was eluted by washing with 140 mM sodium phosphate (pH 7), and the fractions containing activity were pooled and dialyzed, first for 2 h against 50 volumes of 50 mM Tris-acetic acid (pH 7.5), and, after- wards, overnight against 50 volumes of 50 mM Tris/acetic acid (pH 7.5), 0.5 M NaCI. The dialyzed enzyme was applied (step 6, Table 1) to a chelating Sepharose column (1.5 cm X 4.5 cm) equilibrated and washed, at 0.25 ml/min, first with the latter buffer, followed by three successive washes with the same buffer adjusted to, in this order, pH 7, pH 6.6 and pH 6 (about 30 ml each wash). ADPRibase-I was recovered in the fractions eluting at pH 6.

ADPRibase-Mn from the DEAE-cellulose step was ap- plied (step 7, Table 1) onto a Reactive blue-2 Sepharose CL-6B column (1.4 cm X 5 cm) equilibrated in 20 mM Tris-HC1 (pH 7.5), 0.5 mM EDTA, followed by two washes with the same buffer, the first one of a volume of 10 ml supplemented with 1 M KC1, and the second one of 20 ml without KC1, everything at 0.8 ml/min, conditions under which ADPRibase-Mn remained adsorbed to the column. The enzyme was recovered by washing with 50 mM Tris-HCl (pH 9.7), 0.5 mM EDTA. The fractions eluting at pH 9.7 (1.7 ml each) were collected in tubes containing 0.15 ml of 0.5 M Tris-HCl (pH 6.8). Fractions with activity were pooled and applied (step 8, Table 1) to a Q-Sepharose column (1.1 cm X 10 cm) equilibrated in 20 mM Tris (pH 7.5), followed by a linear 0-0.5 M KC1 gradient in the same buffer supplemented also with 0.5 mM EDTA, at a rate of 0.35 ml/min. ADPRibase-Mn was recovered in the fractions corresponding to 0.2-0.3 M KCI.

ADPRibase-II from the DEAE-cellulose step was di-

rectly applied (step 9, Table 1) to a hydroxyapatite column (1.4 cm X 3.4 cm) equilibrated in 0.4 M KCI and immedi- ately washed with 1 M KC1 at 0.1 ml/min, conditions under which ADPRibase-II remained bound to the gel. The enzyme was eluted by washing with 1 mM potassium phosphate (pH 7), 0.5 mM EDTA, and the fractions con- taining activity were pooled and dialyzed overnight against 15 vols. of 20 mM Tris-acetic acid (pH 6.4), 0.5 mM EDTA. The dialyzed enzyme was applied (step 10, Table 1) to a Reactive green 19-Agarose column (1.5 cm X 3 cm) equilibrated in dialysis buffer. After a wash with the same buffer at 0.1 ml/min, the enzyme was eluted with 50 mM Tris-HCl (pH 7.5), 0.5 mM EDTA.

All the purification scheme was carried out at 4°C. Protein was assayed by the method of Bradford [24].

2.4. Analytical gel-filtration chromatography

Chromatography on a Sephadex G-75 (superfine de- gree) column (0.9 cm x 135 cm) was performed to esti- mate relative molecular masses (M r) of the three ADP- ribose pyrophosphatases. The column was equilibrated in 20 mM Tris-HC1 (pH 7.5), 0.5 mM EDTA and 0.1 M KC1; samples of 1 ml (in the case of ADPRibase-Mn, supple- mented with 2.5 mg of bovine serum albumin) were applied to it, and the chromatography was developed with the same buffer at 1 ml/h. The system was calibrated with bovine serum albumin (66000), ovalbumin (45000), trypsinogen (24000) and cytochrome c (12500) as M r standards.

2.5. Chromatofocusing

A Polybuffer exchanger PBE 94 column (1.4 cm X 5.3 cm) was equilibrated in 25 mM imidazole-HCl (pH 7.5). Samples of ADPRibase-I and ADPRibase-II to be ana- lyzed were buffer-exchanged by gel filtration in a Sephadex G-25 column equilibrated in the same buffer; samples of ADPRibase-Mn were used in the same purification buffer in which they were obtained. In every case, after loading the sample onto the PBE 94 column, a pH 7-to-pH 4 gradient was obtained by elution with Polybuffer 74 (9-fold diluted in water, adjusted to pH 4 with HC1, and degassed) at 0.4 ml/min. Fractions were collected and, immediately, pH was measured and adjusted to pH 7 by addition of 1 M Tris base. In the case of ADPRibase-Mn, the collected fractions were also adjusted to 0.5 mM EDTA by addition of aliquots from a 0.1 M stock solution.

2.6. HPLC

Ion-pair, reverse-phase analyses were carried out with a Hewlett-Packard HP 1090 chromatograph, in a NovaPak C18 column (3.9 mm X 150 mm; Waters) equilibrated in 5 mM sodium phosphate (pH 7), 20 mM tetrabutylammo- nium bromide, 10% (v /v ) methanol. The elution of sam-

J. Canales et al. / Biochimica et Biophysica Acta 1246 (1995) 167-177 171

pies was accompl i shed at 0.5 m l / m i n either wi th a 20

min, 5 - 1 0 0 m M Pi gradient in the same buffer or, w h e n

measur ing by H P L C the hydrolyt ic act ivi ty on N A D +,

wi th a 10 min isocrat ic wash with equi l ibrat ing buffer

fo l lowed by a Pi gradient its before. The ch romatograms

were recorded wi th an abscrbance detector set at 250 nm

(339 nm for assaying the hydrolys is o f N A D H by H P L C )

and analyzed with an HP 3"¢93 A integrator.

3. Results and discussion

3.1. Identification, purification and properties o f three ADP-ribose pyrophosphatases from rat liver supernatants

Mir6 and coworkers had used the Mg2+-dependen t

hydrolys is of ADP- r ibose as the reporter act ivi ty for the

partial pur i f icat ion o f rat l iver A D P R i b a s e - I and A D P R i b -

ase-II f rom 100000 × g supematants by a m m o n i u m sul-

fate precipitat ion, gel-f i l t ra t ion and D E A E - c e l l u l o s e chro-

matography. A D P R i b a s e - I and ADPRibase - I I co-pur i f ied

up to the latter step, where they became separated by

elut ion wi th a KC1 gradient [19]. Fo l lowing the same

purif icat ion protocol , but assaying the hydrolys is o f A D P -

r ibose in the presence o f 5 m M MnC12 rather than 5 m M

MgC12, we did not observe s ignif icant changes in the

purif icat ion and recovery o f the 3 0 - 6 0 % a m m o n i u m sul-

fate fract ion and gel f i l tration chromatography (steps 2 and

3, Table 1), or in the shape o f the elut ion prof i le o f

ADP- r ibose pyrophosphatase act ivi ty upon gel fi l tration

chromatography (Fig. 1A). On the contrary, the elut ion

profi le o f the DEAE-ce l l u lo se column, emp loyed for pu-

r if icat ion after the gel fi l tration step, was strongly affected

by MnC12 (Fig. 1B). With M g 2+ as the act ivat ing cation,

A D P R i b a s e - I and ADPRibase - I I appeared in fract ions cor-

responding to, respect ively , 150 m M and 350 m M KC1,

wi th in a l inear 5 0 - 4 0 0 m M KC1 gradient, in agreement

wi th previous results [19]. With Mn 2÷, three components

Table 2 'Summary of ADP-ribose pyrophosphatases in rat liver

Parameter ADPRibase-I ADPRibase-Mn ADPRibase-II ADPRibase-m

Activating cation Mg 2-" or Mn 2+ (6) Mn 2+ K M 0.5-2 /.tM (6) 15-30/xM pH-activity profile broad profile (pH 7-9) (6) max. at pH 9 a (pH 7-10) (range studied)

M r (gel filtration) b 40 000 (4) 32 000 Isoelectric point 4.9-.';.1 (6) 4.3-4.4 (chromatofocusing)

Reaction products AMP ( + Rib5P) (4) AMP ( + Rib5P) (HPLC) d

Substrate specificity ~ in the presence of Mg 2+ ADP(IDP)-ribose (6) not active

in the presence of Mn 2+ AMP(IMP)-containing ADP(IDP)-ribose; NDP-.sugars; dinucleotides; dinucleotides CDP-alcohols (not NAD + )

(7) Mg 2+ or Mn 2+ (10) Mg 2+ or Mn 2÷ (7) 50-100/zM (10) 2-3 btM (7) broad profile (pH 7-9) (10) broad profile (pH 7-9)

(7) 47000 (10) 35000 (7) < 4.1 c (10) not determined

(8) AMP ( + Rib5P) (10) AMP (+ Rib5P)

(8) AMP(IMP)- (10) ADP(IDP)-ribose containing NDP-sugars; dinucleotides (not NAD)

NDP-sugars; ADP(IDP)-ribose dinucleotides AppA (not NAD + ); CDP-alcohols

The table summarizes the results of this work and includes, for purposes of comparison, published data and a few new results for an ADP-ribose pyrophosphatase located in the mitochondrial matrix of rat liver (ADPRibase-m; [25] and this work). Except when indicated or when strictly required by the experimental design, the results were obtained studying the hydrolysis of ADP-ribose at pH 7.5, in the presence of 5 mM MgC12 or, in the case of ADPRibase-Mn, 5 mM MnC12. NtLmbers in parentheses correspond to the purification step (Table 1) after which data were obtained. a The pH-activity profile of ADPRibase-Mn was determined with 1 mM MnC12 to avoid the relatively high blank values observed at 5 mM MnC12 and above pH 8.

Assuming an error of + 1 ml in elution volumes of standards and samples (actual values 40-70 ml), the errors in the M r values given would be + 2000 to + 3000. c In the chromatofocusing experiment, ADPRibase-II eluted at the end of the pH 7-to-pH 4 gradient; henceforth, its isoelectric point may be equal or lower than 4.1. d Reaction products were assayed by HPLC with an ultraviolet absorbance detector, which does not allow detection of D-ribose 5-phosphate (Rib5P) if formed. Since AMP (5'-nucleotide) was detected as a product of the hydrolysis of ADP-ribose (several NDP-sugars or CDP-alcobols), the other product was assumed to be in each case the corresponding sugar-phosphate. With ADP and dinucleotides such as NAD +, NADH and AppA, only 5'-nucleotides were observed as products, thereby confirming the hydrolytic pattern typical of pyrophosphatases. e Substrate specificity was defined m the light of Table 3 data, except in the case of ADPRibase-m, for what much less evidence was available, as the only substrates tested were: ADP-ribose, IDP-ribose, AppA, ADP-glucose and CDP-choline, all both in the presence of either 5 mM MgC12 or 5 mM MnCI 2 ([25] and results not shown).

172 J. Canales et al. / Biochimica et Biophysica Acta 1246 (1995) 167-177

of ADP-ribose pyrophosphatase activity were observed. Two of them co-eluted, respectively, with the Mg2+-de - pendent ADPRibase-I and ADPRibase-II, but were less prominent, particularly in the first case, in which very little Mn2+-dependent activity was observed. The third peak (named as ADPRibase-Mn) eluted at 270 mM KC1, be- tween and well resolved from ADPRibase-I and -II, and free from Mg2÷-dependent activity (Fig. 1B).

Concerning the Mg2+-dependent activities of ADPRib- ase-I and ADPRibase-II, up to the DEAE-cellulose step, the results were in agreement with previous work [19]. Those two enzymes, as well as the newly described AD- PRibase-Mn, have been now submitted to the additional purification steps described in Section 2, with the results summarized in Table 1.

ADPRibase-I specific activity (step 6, Table 1) was increased 8-fold with respect to the preparation previously reported (Ref. [19] and step 4, Table 1), but it is still rather impure as many bands were seen in denaturing SDS gels stained with silver or with Coomassie brilliant blue.

ADPRibase-Mn was highly purified after a very effi- cient affinity chromatography step, as the enzyme bound strongly to the immobilized dye Reactive blue 2 at pH 7.5,

remained bound in the presence of 1 M KC1, and was recovered by a wash at pH 9.7 with little loss of activity and a near 80-fold enrichment. A subsequent ion-exchange step on Q-Sepharose gave a further 2.5-fold increase of specific activity, up to a final 18 units/mg with a 5% overall recovery (step 8, Table 1). Purified ADPRibase-Mn was not homogeneous when analyzed by denaturing gel electrophoresis, but a limited set of bands was observed. A 150 ng sample stained with silver displayed two major bands of about 64 kDa and 43 kDa, as well as several faint bands ranging from 40 kDa to higher than 66 kDa, and one of 31 kDa. Coomassie brilliant blue staining revealed only the two major bands. Given the M r found for the native enzyme by gel filtration (Table 2), a precise identification of ADPRibase-Mn band(s) after electrophoresis was not possible.

ADPRibase-II was also highly purified by hydroxy- apatite (step 9) and dye-ligand affinity chromatography (step 10, Table 1), with a final specific activity near 13 units/mg, i.e., a 24-fold increase with respect to the DEAE-cellulose step (see Ref. [19] and step 4, Table 1) and a 7% recovery, which is a clear underestimate due to partial co-purification with ADPRibase-I. Purified AD-

Table 3 Substrate specificity of ADPRibase-I, ADPRibase-Mn and ADPRibase-II

Substrate (0.35 mM) ADPRibase-I activities (% of Mg 2 +-dependent ADP-ribose pyrophosphatase) with 5 mM

ADPRibase-Mn activities 2+ (% of Mn -dependent ADP-ribose

pyrophosphatase) with 5 mM

ADPRibase-II activities (% of Mg 2 +.dependent ADP-ribose pyrophosphatase) with 5 mM

MgCI 2 MnC12 MgC12 MnCI 2 MgC12 2h8MnC12

ADP-ribose 100 + 14 (4) 22 ___ 1 (4) < 3 100 ___ 23 (8) 100 ___ 7 (8) 38 + 5 (4) IDP-ribose 52 ___ 6 (4) 17 + 2 (4) < 3 105 + 24 (8) 31 + 5 (4) 43 ___ 4 (4) AppA < 1 8 + 1 (4) < 3 45 + 11 (8) 31 + 1 (4) 31 + 2 (4) NADH < 1 14 ___ 1 (4) < 3 25 (2) 30 (2) 25 (2) NAD + < 2 < 2 < 3 23 (2) < 2 < 2 FAD < 2 2 + 0.4 (4) < 3 39 + 14 (4) 32 + 1 (4) 22 + 1 (4) ADP-glucose < 1 2 + 0.2 (4) < 3 < 3 25 + 1 (4) 33 + 3 (4) ADP-mannose < 1 6 + 1 (4) < 3 < 3 36 + 4 (4) 61 + 3 (4) GDP-mannose < 1 < 1 < 3 < 3 < 2 34 + 1 (4) UDP-mannose < 1 < 1 < 3 < 3 < 2 53 + 4 (4) UDP-glucose < 1 < 2 < 3 < 3 < 2 27 + 6 (4) UDP-xylose < 2 < 1 < 3 < 3 < 2 29 + 1 (4) CDP-glucose < 2 < 1 < 3 < 3 < 2 15 + 1 (4) CDP-glycerol < 1 1 + 0.2 (4) < 3 89 + 23 (8) 2 + 0.3 (4) 104 + 5 (4) CDP-ethanolamine < 1 < 2 < 3 89 + 28 (8) < 2 33 ___ 4 (4) CDP-choline < 1 < 1 < 3 140 + 28 (8) < 2 25 + 1 (4) ADP < 1 < 1 < 3 2 7 + 8 ( 8 ) < 2 < 2 CDP < 2 < 2 < 5 < 2 < 2 < 2 AMP < 2 < 2 < 3 < 3 < 1 < 2 CMP < 2 < 3 < 3 < 3 < 2 < 2 GMP < 2 < 3 < 3 < 3 < 2 < 2 dTMP p-nitrophenyl ester < 3 < 5 < 3 < 4 < 2 < 2

Positive results are means + S.D. with the number of measurements in parenthesis. They were obtained using the enzyme samples with the highest purity available in each case (Table 1). Negative results are shown as 'lesser than' the sensitivity limit and were obtained using either also the purest samples or the DEAE-cellulose step (Table 1). All the activities were assayed (under conditions of linearity with time and enzyme sample and after correction with blanks without substrate and/or enzyme) by measuring the Pi formed in the presence of alkaline phosphatase, with the following exceptions: (i) alkaline phosphatase was omitted when assaying activities on substrates with terminal phosphoryl groups; (ii) the activity or absence of activity towards NAD + and NADH were measured or confirmed by HPLC to avoid interferences by the presence of ADP-ribose as a potential contaminant in commercial nicotinamide coenzymes; in the case of ADPRibase-Mn, the Mn2+-dependent activity on ADP was confirmed by HPLC.

J. Canales et a l . / Biochimica et Biophysica Acta 1246 (1995) 167-177 173

J ADPRibase-I

lOOi S o

5 0 ~ ~ ~.

0 i i i

lADPRibase-Nn 1 0 0 - ~ .t •

100

50 5

I v~- , 0.;1 ' 0.;2 ' 0.;3 0 -IIl~ , j --o , , o ,

l o o ~ ~

0 2 4 [Mgc~t or [MnCt2l, mM

Fig. 2. Requirements of Mg 2+ or Mn 2+ for ADPRibase-l, ADPRibase- Mn and ADPRibase-II. The assays were performed using enzyme sam- ples exhaustively dialyzed against 20 mM Tris-HCl (pH 7.5), to remove EDTA present in purification buffer. Activities were measured in the presence of either ( n ) MgC12 or ( A ) MnC12 at the concentrations indicated, with 0.35 mM ADP-ribose as the substrate. Rates are expressed as percentages of the maximal act~tvity measured for each enzyme. The insert in the central panel shows Mn2+-dependent activity at low salt concentrations in an expanded x-axis.

PRibase-II was also not homogeneous, as silver staining of a 700 ng sample analyzed by denaturing gel electrophore- sis revealed major 31 kDa and 27 kDa bands in addition to faint ones of 63 kDa, 50 kDa and 18 kDa, whereas Coomassie brilliant blue staining revealed only the two most intense bands. Considering the native M r measured by gel filtration chromatography (Table 2), a precise iden- tification of ADPRibase-II bands was also not feasible.

The three enzymes were characterized concerning opti- mal conditions for reaction, kinetics and some molecular properties. ADPRibase-I and ADPRibase-II were activated by Mg 2+ and by Mn 2÷, but at different concentration ranges, as at concentratiorts higher than 0.3-0.5 mM, MgC1 z was more effective than MnC12, whereas the con- verse was true at lower concentrations. ADPRibase-Mn was activated only by Mn 2÷, half-maximal activation be- ing obtained with 5-10 /zM MnCI 2 (Fig. 2). Data relative to K M for ADP-ribose, ptI-activity profiles, native M r, and isoelectric points are not shown in detail, but the results are summarized in Table 2.

3.2. Reaction specificities

An extensive study of substrate specificity was per- formed with 22 substrates either in the presence of Mg 2+ or Mn 2÷. The series of substrates included: reducing

(ADP-ribose and IDP-ribose) and non-reducing NDP- sugars, CDP-alcohols, dinucleotides, nucleoside diphos- phates, nucleotides, and the artificial NPPase/PDEaseI substrate dTMP p-nitrophenyl-ester. The activities were measured in the preparations resulting from the DEAE-cel- lulose chromatography step (Table 1, step 4) and, for the substrates on which significant activity was detected, also after the final purification steps (Table 1, steps 6, 8 or 10). The activities measured in the presence of 5 mM MgCI 2 or MnC12 are all shown in Table 3.

ADPRibase-I showed Mg 2÷- and Mn2+-dependent ac- tivities towards different substrates. With 5 mM MgCI2, ADPRibase-I was significantly active only towards reduc- ing NDP-sugars: ADP-ribose and its close analog IDP- ribose, therefore, as far as Mg 2+ is the activating cation, ADPRibase-I is a very specific enzyme. With 5 mM MnC12 those activities were smaller and several other substrates, namely the AMP(IMP)-containing NDP-sugars and dinucleotides, were split at 2-15% of the Mg2+-de - pendent ADP-ribose pyrophosphatase rate, except for NAD +, which was not hydrolyzed (Table 3). As checked by HPLC, the hydrolysis of Mg/ADP-ribose, Mg/IDP- ribose and Mn/AppA yielded the corresponding 5'- nucleotide(s) as products (not shown).

ADPRibase-Mn, in the presence of 5 mM MgCI 2, did not appreciably hydrolyzed any of the substrates tested, but in the presence of 5 mM MnC12, several substrates were hydrolyzed at 20-140% of the rate measured with Mn/ADP-ribose, namely: the close analog IDP-ribose, dinucleotides and CDP-alcohols (Table 3). A minor activ- ity on ADP (not on CDP) was also measured. By HPLC assays in the cases of ADP-ribose, IDP-ribose, AppA, NADH, NAD +, CDP-choline, CDP-ethanolamine and ADP, it was found that the corresponding 5'-nucleotide(s) were formed as products of ADPRibase-Mn (not shown). It is worth noting that non-reducing NDP-sugars were not hydrolyzed by ADPRibase-Mn.

ADPRibase-II showed activity towards all the NDP- sugars, CDP-alcohols and dinucleotides tested, with the exception of NAD ÷, probably due to the presence of the positive charge in the nicotinamide ring. Also not hydro- lyzed were all the substrates with terminal phosphate(s) and dTMP p-nitrophenyl ester. Interestingly, the enzyme was more selective with Mg 2+ than with Mn 2+ as the activating cation. In this concern, ADPRibase-II substrates can be classified into three groups: (i) ADP-ribose, hydro- lyzed 3 fold faster with 5 mM MgCI 2 than with 5 mM MnCI2; (ii)ADP-hexoses, dinucleotides and IDP-ribose, hydrolyzed at similar rates with 5 mM MgC12 or MnCI2; (iii) non-adenine(inosine) NDP-sugars and CDP-alcohol compounds, hydrolyzed in the presence of 5 mM MnC12 but not in the presence of 5 mM MgCI 2 (Table 3). The hydrolysis of ADP-ribose and NADH, both in the presence of MgCI 2 and MnC12, and the hydrolysis of CDP-glycerol in the presence of MnC12 yielded 5'-nucleotide products, as checked by HPLC (not shown).

174 J. Canales et al./Biochimica et Biophysica Acta 1246 (1995) 167-177

3.3. Number o f pyrophosphatases involved

A major question in this work is how many enzymes are responsible for the activities studied. ADPRibase-I, ADPRibase-Mn and ADPRibase-II show several discrimi- nating properties (see Table 2): substrate and cation speci- ficities, K M values for ADP-ribose, response to pH, native molecular masses, isoelectric points and, of course, chro- matographic behavior in ion-exchange columns (Fig. 1B). This indicates that there are at least three different NDP- sugar/alcohol pyrophosphatases in our experiments. On the other hand, it can be asked whether all the co-purifying activities which are attributed either to ADPRibase-I, AD- PRibase-Mn or ADPRibase-II may in fact correspond to more than three enzymes. Although such a possibility cannot be ruled out, we suggest that only three enzymes may be responsible for the activities shown in Table 3. As described below, activities towards different substrates of each enzyme co-eluted in a variety of chromatographic systems, showed similar thermal inactivation profiles, and in two selected cross-inhibition experiments behaved as expected after our suggestion.

Chromatographic co-elution tests. All the Mg 2÷- and Mne+-dependent activities shown by the more purified preparation of ADPRibase-I (Table 3) were submitted to co-elution tests in several chromatographic systems; namely in DEAE-cellulose and Chelating Sepharose, which were part of the purification scheme (Table 1), and in analytical experiments carried out by gel filtration on a Sephadex G-75 column (Section 2; the sample was ADPRibase-I

from step 4, Table 1) and by chromatofocusing (Section 2; ADPRibase-I sample was from step 6, Table 1). The elution profile of activity on Mn/IDP-ribose was not assayed after chromatofocusing.

The Mn2+-dependent activities of ADPRibase-Mn on reducing NDP-sugars, dinucleotides and CDP-alcohols (Table 3) were all tested for co-elution in DEAE-cellulose and Q-Sepharose ion-exchange chromatographies em- ployed for ADPRibase-Mn purification (Table 1), as well as in analytical gel filtration in a Sephadex G-75 column (Section 2; ADPRibase-Mn sample was from step 7, Table 1) and by chromatofocusing (Section 2; ADPRibase-Mn sample was from step 8, Table 1).

Concerning ADPRibase-II, all the Mg 2+- and Mn2÷-de - pendent activities (except the minor one on Mg/CDP- glycerol, Table 3) were submitted to co-elution tests in the DEAE-cellulose chromatography which makes part of the purification (step 4, Table 1), and in Sephadex G-75 and chromatofocusing columns (Section 2; ADPRibase-II sam- ples from step 10, Table 1).

In all cases, with the three enzymes, the elution profiles of the activities were assayed in the chromatographic fractions having ADP-ribose pyrophosphatase activity and showed coincident single peaks, i.e., all the co-elution tests mentioned above yielded positive results (not shown).

Thermal co-inactivation tests. Two inactivation experi- ments were carried out with ADPRibase-I, using enzyme from the DEAE-cellulose or from the Chelating Sepharose columns (Table 1, steps 4 and 6), respectively. The inacti- vation profile of the ADP-ribose pyrophosphatase activity

125

5O

0 40 80

i

i i

o 4'0 go TEMPERATURE, *C

o 4'0 sb

Fig. 3. Thermal inactivation profiles of ADPRibase-I, ADPRibase-Mn and ADPRibase-II. Samples of the indicated enzyme preparation were heated for 2 min at the temperatures shown and immediately cooled down on ice. Thereafter, activities were assayed on the following cation-substrate combinations. ADPRibase-I (solid lines, enzyme from step 4; dashed lines, enzyme from step 6, Table 1): (D) Mg/ADP-ribose, (~) Mg/IDP-ribose, (©) Mn/ADP-ribose, (zx) Mn/IDP-ribose, (v) Mn/AppA, (11) Mn/NADH, (0) Mn/FAD, (O) Mn/ADP-glucose, (A) Mn/ADP-mannose, (v) Mn/CDP-glycerol. ADPRibase-Mn (from step 7, Table 1): (O) Mn/ADP-ribose, (~) Mn/IDP-ribose, (zx) Mn/CDP-choline, ( • ) Mn/CDP-glycerol, (A) Mn/CDP-ethanolamine, (0) Mn/NAD +, (v) Mn/AppA, (11) Mn/NADH, (×) Mn/FAD. ADPRibase-II (from step 10, Table 1): (O) Mg/ADP-ribose, (~) Mg/NADH, (×) Mg/AppA, (Q) Mg/ADP-glucose, (©) Mn/ADP-ribose, (v) Mn/AppA, (11) Mn/NADH, (0) Mn/GDP- mannose, ( • ) Mn/UDP-xylose, ( • ) Mn/CDP-glycerol, ( ,x ) Mn/CDP-choline. Activities are expressed as percentages of untreated samples kept on ice.

J. Canales et a l . / Biochimica et Biophysica Acta 1246 (1995) 167-177 175

100 ~ ' ~

0 2 4 6 20 60 100

INHIBITORY CATION, mM

Fig. 4. Inhibition of ADPRibase-I activities by Mn 2+ or Mg 2+. Assays were carried out at a constant, 5 mM concentration of an activating cation (Mg 2+ or Mn 2+ ) and varying concentrations as indicated of the in- hibitory cation (Mn 2+ or Mg 2+, respectively). (D) Inhibition by Mn 2+ of the Mg2+-dependent activity on ADP-ribose. ( • ) Inhibition by Mg 2÷ of the Mn2+-dependent activity on AppA.

mined) should equal their respective K M values (15-30 /zM for ADP-ribose and 250/xM for CDP-choline; results not shown).

In summary, although definitive conclusion on the num- ber of pyrophosphatases involved should await until all of them are purified to homogeneity, the results of this work are compatible with the occurrence of only three of those enzymes in the preparations studied, with the substrate specificities tentatively defined in Table 2 and Section 3.2.

3.4. Cytosolic location and comparison of ADPRibase-I, ADPRibase-Mn and ADPRibase-II to mitochondrial ADP- ribose pyrophosphatase

was different in each case, as ADPRibase-I from step 4 was more resistant than ADPRibase-I from step 6. Interest- ingly, all the Mg 2+- and Mn2+-dependent activities of the ADRibase-I preparations (Table 3) behaved in the same way (Fig. 3).

ADPRibase-Mn (from step 7, Table 1) and ADP- Ribase-II (from step 10, Table 1) were submitted also to thermal inactivation experiments. In the case of ADPRib- ase-Mn, the MnE+-dependent activities towards all the NDP-sugars, CDP-alcohols and dinucleotides attacked by the enzyme (Table 3) were followed and showed similar inactivation profiles (Fig. 3). In the case of ADPRibase-II, the inactivation was followed by assay of the activities towards a selection of substrate-cation (Mg 2+ or Mn 2+) combinations chosen among those that in Table 3 gave positive results; all the inactJivation profiles obtained were alike (Fig. 3).

Cross-inhibition tests. Two selected cross inhibition experiments were carried oat to test salient features of ADPRibase-I and ADPRibase-Mn specificities. In the case of ADPRibase-I, it was reasoned that if only one enzyme was responsible for the MgE+-dependent hydrolysis of ADP-ribose and Mn2+-dependent hydrolysis of AppA (Ta- ble 3), then each of those activities should be inhibited by the other cation. Positive results in that concern suggest that both cations may compete for binding to the same enzyme (Fig. 4).

In the case of ADPRibase-Mn, the salient feature tested through cross-inhibition measurements was the ability to sustain the hydrolysis of the relatively dissimilar substrates ADP-ribose and CDP-choline, while e.g. CDP-glucose was not attacked (Table 3). In fact, ADP-ribose inhibited the hydrolysis of CDP-choline lay ADPRibase-Mn and vice versa: the initial rates of hydrolysis of 0.53 mM ADP-ribose and of 5 mM CDP-choline, assayed by HPLC, were inhibited by, respectively, 52 + 2% and 35 _+ 10% (n = 3) when the activity of ADPRibase-Mn was assayed on a mixture of both substrates at the indicated concentrations. The inhibition data agreed reasonably well with the results expected (about 40-60% inhibition in each case) for com- peting substrates, whose K i values as inhibitors (not deter-

Recently it has been shown that rat liver mitochondrial matrix contains an ADP-ribose pyrophosphatase with M r 35 000 by gel filtration, a 2-3 /xM K~a for ADP-ribose, active with Mg 2+ and, to a lesser extent, with Mn 2+, and which, in the presence of Mg 2+, does not hydrolyze AppA or ADP-glucose [25]. These results, compared to those obtained in the present study, suggest that mitochondrial ADP-ribose pyrophosphatase is related to ADPRibase-I (Table 2). Further characterization of the mitochondrial enzyme, partly purified from a mitochondrial extract by gel filtration and DEAE-cellulose chromatography as de- scribed [25], confirmed that view, since when Mn 2+ sub- stituted for Mg 2+ as the activating cation, its substrate specificity underwent a change similar to ADPRibase-I: AppA was thus also hydrolyzed while CDP-choline was not (results not shown). The possibility that ADPRibase-I could appear in liver 100 000 × g supernatants by release from the matrix of damaged mitochondria was considered and unambiguously discarded [25]. On the other hand, neither ADPRibase-Mn not ADPRibase-II have been found in rat liver mitochondria although they have been searched for (results not shown and [25]). In addition, although the starting point for the purification of ADPRibase-I, AD- PRibase-Mn and ADPRibase-II was a liver homogenate prepared in a hypotonic buffer (Table 1), we have checked that the three enzymes are recovered in similar amounts from 100000 × g supernatants of isotonic homogenates (results not shown). Therefore, ADPRibase-I, ADPRibase- Mn and ADPRibase-II can be ascribed to the cytosolic fraction of rat liver.

3.5. Possible biological relevance of the activities studied

Comparison to NPPase /PDEaseI. The enzymes de- scribed in this work are NDP-sugar/alcohol pyrophos- phatases distinguishable from NPPase/PDEaseI by their substrate specificities, including the absence of activity on dTMP p-nitrophenyl ester, and because they are easily obtainable as soluble (cytosolic) proteins without any spe- cial treatment, and do not show aggregatory behavior in the absence of detergents. NPPase/PDEaseI(-like) en- zyme(s) in liver attack dTMP p-nitrophenyl ester with high

176 J. Canales et al. / Biochimica et Biophysica A cta 1246 (1995) 167-177

activity and are membrane proteins that can be solubilized with detergents [3,12,15].

With ADP-ribose as the substrate, the NDP-sugar/al- cohol pyrophosphatases described in this work, as well as the mitochondrial ADP-ribose pyrophosphatase recently described [25], represent 15-30 milliunits/g liver each, whereas the activity of liver NPPase/PDEaseI on ADP- ribose is about 3 units/g [12]. With other NDP-sugar/al- cohol substrates the picture is similar. However, this does not rule out a role in the control of processes that involve NDP-sugar/alcohol as intermediates. The enzymes that synthesize these compounds in rat liver are either cytosolic [26,27] or bound to the cytosolic surface of microsomes [28]. NPPase/PDEaseI, as a glycoprotein located in the external surface of the plasma membrane (see Section 1), should be also located in the luminal face of endoplasmic reticulum, Golgi apparatus and endocytic vesicles. In fact, there is evidence that from its intracellular location(s) NPPase/PDEaseI does not have access to the cytoplasmic pool of NDP-sugars [16]. Therefore, it will be worth to explore whether cytosolic NDP-sugar/alcohol pyrophos- phatases like those studied in this work could be relevant to the turnover of substrate pools that cannot be reached by NPPase/PDEaseI.

Turnover o f free ADP-ribose. ADPRibase-I is a clear candidate to have a relevant role in the metabolism of this reducing NDP-sugar. Free ADP-ribose may be formed by NAD + glycohydrolase (EC 3.2.2.5), that hydrolyzes the N-glycosidic bond of NAD + yielding nicotinamide and ADP-ribose, either as an independent enzyme or as an abortive activity of poly(ADP-ribose) polymerase (EC 2.4.2.30) [29-32], and by poly(ADP-ribose) glycohydro- lase (EC 3.2.1.X), which hydrolyzes the (1"-2') O-glyco- sidic bond between ADP-ribose chain monomers in pro- tein-bound and protein-free poly(ADP-ribose) [33-36]. Therefore, ADP-ribose is a by-product of the biosynthesis of poly(ADP-ribose) and/or a major product of the degra- dation of the polymer. In addition, it has been reported that it may be formed also by an ADP-ribosyl protein hydro- lase that cleaves a linkage between mono(ADP-ribose) and a cysteine residue in the inhibitory GTP-binding protein (G i) of adenylate cyclase [37]. More recently, it has been found that ADP-ribose is also the turnover product of the novel Ca 2+ regulator cyclic ADP-ribose, which is pro- duced from NAD + and hydrolyzed to ADP-ribose by bifunctional enzymes identified as NAD + glycohydrolases either intracellular or from the cell surface [38-43]. Free ADP-ribose contains a reactive aldehyde function and there are reports of non-enzymic reactions of this metabo- lite leading to the covalent modification of proteins [44- 47]. There is some evidence that such reactions may be target-specific [45,46] and may occur at micromolar con- centration of ADP-ribose [47]. Irrespective of whether they represent a regulatory mechanism or just a source of chemical damage to macromolecules, it should be advanta- geous for cells to control the levels and disposal of free

ADP-ribose through an specific enzyme pathway, such as e.g., the one that can be supported by the Mg:+-dependent activity of ADPRibase-I. In this concern, it is interesting also that the K M value of ADPRibase-I for free ADP- ribose (0.5-1 /~M) is much lower than the K M shown by ADPRibase-II or ADPRibase-Mn and it is quite close to the concentration of free ADP-ribose (0.5 /xM) recently reported for the first time in human red cells [48].

Erythrocyte lysates contain also ADP-ribose pyrophos- phatase activity [48,49]. A Mg2+-dependent enzyme has been purified 1000-fold and has been reported to cleave also NAD + and NADH but not ADP-glucose, and to bind to DEAE-cellulose at pH 8.3 in the presence of 200 mM KC1 but not 300 mM KCI [48]. With only this evidence available for comparison to our results, it appears that erythrocyte ADP-ribose pyrophosphatase differs from rat liver enzymes, as neither of them fit the above description of cation-substrate preferences: ADPRibase-I did not show Mg2+-dependent activity on nicotinamide coenzymes, AD- PRibase-Mn was devoid of Mg2+-dependent activities, and ADPRibase-II was active on ADP-glucose and NADH but not on NAD + (Table 3).

Processes requiring CDP-alcohols. CDP-alcohols are intermediates in the de novo biosynthesis of membrane phospholipids [50], the major lung surfactant component dipalmitoylphosphatidylcholine [51], and the platelet- activating factor 1-O-alkyl-2-acetyl-sn-glycero-3-phospho- choline [52]. The supply of CDP-choline or CDP-ethanol- amine is considered to be the rate-limiting step in those pathways [50,53-56]. Therefore, the Mn2+-dependent CDP-alcohol pyrophosphatase activities of ADPRibase-Mn and ADPRibase-II could provide important drainage mech- anisms for key intermediates.

Glycosylation reactions. Out of the three enzymes stud- ied in this work, ADPRibase-II is the only one acting on NDP-sugars involved in mammalian cell glycosylations, e.g., UDP-glucose, GDP-mannose or UDP-xylose. The hydrolyses of those precursors required Mn 2+ (Table 3). If in fact this is a cytosolic enzyme (see Section 3.4), the Mn2+-dependent activity of ADPRibase-II may be relevant to the control of the cytosolic pool of glycosylation precur- sors. It is also worth to recall here the known Mn 2+ requirement of glycosyltransferases [57,58] for two rea- sons. For one thing, this requirement may be related to the cation-substrate specificity of ADPRibase-II, which de- pends on Mn 2+ for the hydrolysis of GDP-mannose, UDP-glucose and UDP-xylose (Table 3). For another, the Mn 2 +-dependent assays of glycosyltransferases are strongly hampered in tissues rich in NPPase/PDEaseI due to the in vitro depletion of NDP-sugars. Since NPPase/PDEaseI has a strong requirement for Zn 2+, which is not satisfied by Mn 2+, it has been proposed that selective chelation of Zn 2+ versus Mn 2+ can help avoid interferences without impeding the action of glycosyltransferases [57,58]. The Mn2+-dependent activities of ADPRibase-II (Table 3) are much less prominent than NPPase/PDEaseI in rat liver

J. Canales et al. / B iochimica et B iophysica A cta 1246 (1995) 167-177 177

(see above) and, possibly, they are not the cause of major trouble in transferase assays,. Nevertheless, one should be aware that different and/or more abundant Mn2+-depen - dent activities in the same or other biological materials might be sources of interference when using selective Zn 2÷ chelation [57,58].

Acknowledgements

We thank J.M. Ribeiro for reading the manuscript. D.B. was a predoctoral fellow supported by the Junta de Ex- tremadura. The research was supported by Grant PB87- 1029 from the Direcci6n General de Investigaci6n Cientifica y T6cnica, and Grant SAL91-1081 from the Comisi6n Interministerial de Ciencia y Tecnologia (Spain).

References

[1] Evans, W.H. (1974) Nature 2_';0, 391-394. [2] Bischoff, E., Tran-Thi, T.-A. and Decker, K.F.A. (1975) Eur. J.

Biochem. 51, 353-361. [3] Elovson, J. (1980) J. Biol. Ch(~m. 255, 5807-5815. [4] Smith, G.D. and Peters, T.J, (1982) Biochlm. Biophys. Acta 716,

24-30. [5] Thyberg, J. (1982) Eur. J. Cell Biol. 26, 265-269. [6] Rebbe, N.F., Tong, B.D., Fialey, E.M. and Hickman, S. (1991)

Proc. Natl. Acad. Sci. USA 88, 5192-5196. [7] Funakoshi, I., Kato, H., Horie, K., Yano, T., Hofi, Y., Kobayashi,

H., Inoue, T., Suzuki, H., Fukui, S., Tsukahara, M., Kajii, T. and Yamashina. I. (1992) Arch. Biochem. Biophys. 295, 180-187.

[8] Rebbe, N.F., Tong, B.D. and I-[ickman, S. (1993) Mol. Immunol. 30, 87-93.

[9] Schliselfeld, L.H., van Eys, J, and Touster, O. (1965) J. Biol. Chem. 240, 811-818.

[10] Corder, C.N. and Lowry, O.H. (1969) Biochim. Biophys. Acta 191, 579-587.

[11] Bachorik, P.S. and Dietrich, L.S. (1972) J. Biol. Chem. 247, 5071- 5078.

[12] Cameseile, J.C., Costas, M.J., Sillero, M.A.G. and Sillero, A. (1984) J. Biol. Chem. 259, 2879-2885.

[13] Yano, T., Horie, K., Kanamoto, R., Kitagawa, H., Funakoshi, I. and Yamashina, I. (1987) Biochem. Biophys. Res. Commun. 147, 1061- 1069.

[14] Van Dijk, W., Lasthuis, A.-M., Trippelvitz, L.A.W. and Muilerman, H.G. (1983) Biochem. J. 214, 1003-1006.

[15] Muilerman, H.G., Lasthuis, A-M., Hooghwinkel, G.J.M. and Van Dijk, W. (1984) Biochem. J. 220, 95-103.

[16] Hickman, S., Wong-Yip, Y.P., Rebbe, N.F. and Greco, J.M. (1985) J. Biol. Chem. 260, 6098-6105.

[17] Rebbe, N.F. and Hickman, S. 111991) Biochem. Biophys. Res. Com- mun. 175, 637-644.

[18] Puhakainen, E. and H~inninen, O. (1976) Eur. J. Biochem. 61, 165-169.

[19] Mir6, A., Costa,s, M.J., Gar(:ia-Diaz, M., Hemfindez, M.T. and Cameselle, J.C. (1989) FEBS ]Sett. 244, 123-126.

[20] Ames, B.N. (1966) Methods Enzymol. 8, 115-118. [21] Costas, M.J., Montero, J.M., Cameselle, J.C., Sillero, M.A.G. and

Sillero, A. (1984) Int. J. Biochem. 16, 757-762. [22] Mir6, A., Hem~ndez, M.T., C~stas, M.J. and Cameselle, J.C. (1991)

J. Biochem. Biophys. Methods 22, 177-184. [23] Nunez, H.A. and Barker, R. (1976) Biochemistry 15, 3843-3847.

[24] Bradford, M.M, (1976) Anal. Biochem. 72, 248-254. [25] Bernet, D., Pinto, R.M., Costas, M.J., Canales, J. and Cameselle,

J.C. (1994) Biochem. J. 299, 679-682. [26] Choy, P.C., Lim, P.H. and Vance, D.E. (1977) J. Biol. Chem. 252,

7673-7677. [27] Coates, S.W., Gumey, T., Sommers, L.W., Yeh, M. and Hirschberg,

C.B. (1980) J. Biol. Chem. 255, 9225-9229. [28] Vance, D.E., Choy, P.C., Farren, S.B., Lim, P.H. and Schneider,

W.J. (1977) Nature 270, 268-269. [29] Kawaichi, M., Ueda, K. and Hayaishi, O. (1981) J. Biol. Chem. 256,

9483-9489. [30] Masmoudi, A. and Mandel, P. (1987) Biochemistry 26, 1965-1969. [31] Kirsten, E., Bauer, P.I. and Kun, E. (1991) Exp. Cell Res. 194, 1-8. [32] Desmarais, Y., M6nard, L., Lagueux, J. and Poirier, G.G. (1991)

Biochim. Biophys. Acta 1078, 179-186. [33] Miwa, M. and Sugimura, T. (1971) J. Biol. Chem. 246, 6362-6364. [34] Ueda, K., Oka, J., Narumiya, S., Miyakawa, N. and Hayaishi, O.

(1972) Biochem. Biophys. Res. Commun. 46, 516-523. [35] Tanuma, S. and Endo, H. (1990) Eur. J. Biochem. 191, 57-63. [36] Braun, S.A., Panzeter, P.L., Collinge, M.A. and Althans, F.R.

(1994) Eur. J. Biochem. 220, 369-375. [37] Tanuma, S. and Endo, H. (1990) FEBS Lett. 261, 381-384. [38] Lee H.C. and Aarhus, R. (1993) Biochim. Biophys. Acta 1164,

68-74. [39] Kim, H., Jacobson, E.L. and Jacobson, M.K. (1993) Science 261,

1330-1333. [40] Howard, M., Grimaldi, J.C., Bazan, J.F., Lund, F.E., Santos-

Argumedo, L., Parkhouse, R.M.E., Walseth, T.F. and Lee, H.C. (1993) Science 262, 1056-1059.

[41] Zocchi, E., Franco, L., Guida, L., Benatti, U., Bargellesi, A., Malavasi, F., Lee, H.C. and De Flora, A. (1993) Biochem. Biophys. Res. Commun. 196, 1459-1465.

[42] Takasawa, S., Tohgo, A., Noguchi, N., Koguma, T., Nata. K., Sugimoto, T., Yonekura, H. and Okamoto, H. (1993) J. Biol. Chem. 268, 26052-26054.

[43] Summerhill, R.J., Jackson, D.G. and Galione, A. (1993) FEBS Lett. 335, 231-233.

[44] Kun, E., Chang, A.C.Y., Sharma, M.L., Ferro, A.M. and Nitecki, D. (1976) Proc. Natl. Acad. Sci. USA 73, 3131-3135.

[45] Hilz, H., Koch, R., Faniek, W., Klapproth, K. and Adamietz, P. (1984) Proc. Nail. Acad. Sci. USA 81, 3929-3933.

[46] Frei, B. and Richter, C. (1988) Biochemistry 27, 529-535. [47] Zocchi, E., Guida, L., Franco, L., Silvestro, L., Guerrini, M.,

Benatti, U. and De Flora, A. (1993) Biochem. J. 295, 121-130. [48] Guida, L., Zocchi, E., Franco, L., Benatti, U. and De Flora, A.

(1992) Biochem. Biophys. Res. Commun. 188, 402-408. [49] Kim, U.-H., Han, M.-K., Park, B.-H., Kim, H.-R. and An, N.-H.

(1993) Biochim. Biophys. Acta 1178, 121-126. [50] Vance, D.E. (1985) in Biochemistry of Lipids and Membranes

(Vance, D.E. and Vance J.E., eds.), pp. 242-270, Benjamin Cum- mins, Menlo Park, CA.

[51] Infante, J.P. and Huszagh, V.A. (1987) Trends Biochem. Sci. 12, 131-133.

[52] Prescott, S.M., Zimmerman, G.A. and Mclntyre, T.M. (1990) J. Biol. Chem. 265, 17381-17384.

[53] Sundler, R. and .~,kesson, B. (1975) J. Biol. Chem. 250, 3359-3367. [54] Mansbach, C.M. and Parthasarathy, S. (1979) J. Biol. Chem. 254,

9688-9694. [55] Vance, D.E. and Choy, P.C. (1979) Trends Biochem. Sci. 4, 145-

148. [56] Blank, M.L., Lee, Y.J., Cress, E.A. and Snyder, F. (1988) J. Biol.

Chem. 263, 5656-5661. [57] Faltynek, C.R., Silbert, J.E. and Hof, L. (1981) J. Biol. Chem. 256,

7139-7141. [58] Lau, J.T.Y. and Carlson, D.M. (1981) J. Biol. Chem. 256, 7142-

7145.