The role of iron in the activation of mannoni and altronic acid ...

Eur. J. Biochem. 166,623-630 (1987) 0FEBS 1987

The role of iron in the activation of mannonic and altronic acidYiydiatases, two Fe-requiring hydro-lyases Jean-Luc DREYER Department of Biochemistry, University of Fribourg (Received December 17, 1986/March 20,1987)

-

EJB 86 1356

D-Altronate hydratase and D-mannonate hydratase belong to a class of Fe'+-requiring enzymes, but the function of iron in these enzymes is largely unknown. Methods are described for the convenient preparation of both these hydratases from Escherichiu coli and studies related to metal activation are presented. The enzymes are inactive in the absence of a bivalent metal and a reducing agent such as dithiothreitol. Fe2+ at low concentrations activates the enzymes efficiently, but inhibits them over 2mM. Furthermore Mn2+ is also capable of activating aldonic acid hydratases and appears to be a constituent of the enzyme active center. A marked synergistic activation is observed in the presence of both ions, raising the possibility that the enzyme has two binding sites for ions. Upon activation, the two aldonic acid hydratases incorporate a single Fe atom and contain no Fe-S core, in contrast to other characterized Fe-hydratases, such as aconitase or maleic acid hydratase. The incorporated iron is losely bound (with Kd about 4.5 mM and 20 mM for mannonate and altronate hydratase, respectively) and can be readily removed with EDTA. The enzymes exhibit no requirement for sulphide ions and are insensitive to thiol reagents. A first-order inhibition is observed with iron chelators and can be removed by competition with excess metal ions. No change in the absorption spectra is observed upon oxidation-reduction or activation with metals. The activated enzymes exhibit no electron paramagnetic (EPR) spectrum under anaerobic conditions; in the presence of oxygen, an intense EPR spectrum develops in Fe2+-activated samples with signal at g = 1.98, which upon reaction of the enzyme with the substrate moves into a species with signals at g = 4.1 5 and g = 9.07, with EPR parameters very similar to those of oxidized rubredoxins. A number of hydro-lyases readily loose their native activity upon purification. In order to regain activity, ferrous ions and a reducing agent such as a thiol are specifically required. The role of iron in this class of enzymes has received poor attention, except for aconitase (EC 4.2.1.3) [l -61, the longknown Krebs-cycle enzyme, and, to a lesser extent, for maleicacid hydratase (EC 4.2.1.31) [7]. In these two enzymes the iron required for activation becomes incorporated into an [Fe-S] cluster to form an active but relatively labile component required for catalysis. In aconitase, the enzyme isolated under aerobic conditions contains a [3Fe-4S] cluster which is converted by incorporation of Fez + into a diamagnetic [4Fe4S] cluster. The inactive enzyme bearing the [3Fe-4S] cluster does not bind to the substrate, whereas the addition of substrate to the active form of the enzyme, containing the [4Fe4S] cluster, causes a major change in the EPR spectra [6]; t h s change is confined to Fe,, the site generated upon activation by incorporation of Fe2+ into the [3Fe-4S] cluster of the inactive enzyme. The other representative of Fe-hydratases, namely maleic acid hydratase, has not been extensively characterized and spectroscopic data are scanty. Nevertheless Correspondence to J.-L. Dreyer, Dkpartement de Biochimie, Universitt. de Fribourg, Rue du MusCe 5 , CH-1700 Fribourg, Suisse Enzymes. D-Altronate hydratase (EC 4.2.1.7); D-mannonate hydratase (EC4.2.1.8); aconitase, aconitate hydro-lyase (EC4.2.1.3); malease, D-malate hydro-lyase (EC 4.2.1.31); D-tartrate hydratase (EC 4.2.1.81). Note. The term 'aldonic acid hydratases' is used to refer to both altronate and mannonate hydratases when the two enzymes are discussed collectively.

the specific requirements for both Fez+ and acid-labile S2has been established. On activation under appropriate conditions, an active and essential Fe-S cluster (probably a [4Fe-4S] cluster) is also built up, which, in contrast to that of aconitase, is readily degraded upon handling [7]. Based on these data, it seemed plausible that the build-up of an Fe-S cluster upon activation would be a general mechanism of action for iron in Fe-hydratases and that observations made on aconitase (and maleic acid hydratase) could serve as a valuable model for other related enzymes with similar metal requirements. In this paper, we describe observations made on two other iron-requiring hydro-lyases, namely D-altronate and D-mannonate hydratases (D-altronate hydro-lyase, EC 4.2.1.7, and D-mannonate hydro-lyase EC 4.2.1.8, respectively), both active in the Ashwell pathway for hexuronic acid [8, 91. This pathway in Escherichiu coli involves the conversion of glucuronic or galacturonic acid to their corresponding aldonic acid forms, i.e. altronic or mannonic acid. These in turn are degraded by the two genetically distinct enzymes described in this paper, both enzymes yielding a common intermediate, 3deoxy-2-~-glucu~osonate [9 - 151. Our results run counter to the notion that the re-building of an Fe-S cluster is involved in all circumstances in the activation by iron of Fe-hydratases. As will be shown in this work, a single iron atom is required for activity in the two aldonic acid hydratases discussed here and becomes incorporated upon activation. The enzymes bear no Fe-S cluster. In further contrast to other Fe-hydratases, e.g. aconitase, other metals (Co and particularly Mn) can substitute for iron, yielding active preparations. Biochemical

624 properties and spectroscopic data together with an improved method for enzyme purification are presented. MATERIALS AND METHODS Chemicals All chemicals were purchased from Sigma, gels for chromatography were from L K B or Pharmacia Fine Chemicals.

by centrifugation. The supernatant was assayed for 3-deoxy2-D-gIUCUlOSOnateby incubating a 200-p1 aliquot with 250 pl 0.25 M periodic acid in 0.125 M H2S04 for 20 min. The oxidation was stopped by the addition of 0.5 mi 2% (w/v) sodium arsenite in 0.5 M HCI. After incubating the solutions for 30 s at 100°C followed by 20 min at room temperature, 2 mlO.3% (w/v) thiobarbituric acid was added and the mixture is placed in a boiling water bath for 10 min, then removed, centrifuged and cooled at room temperature and the absorbance of the pink color formed was measured at 549 nm.

Cultivation and hurvest of bacteria Escherichiu coli K12 wild-type cells were grown aerobically according to [I41 on a minimum salt medium containing Dglucuronic acid as the sole carbon source. The growth medium contained 1.32% potassium dihydrogen phosphate, 0.2% ammonium sulfate, 0.02% magnesium sulfate, 0.001 YOCaCI2, 0.00005% ferric sulfate and 1YO(w/v) potassium glucuronate. The final pH is adjusted to 7.2. The cells were harvested from a 20-1 fermenter after 18 h at 37°C with a Sharples centrifuge, washed twice with 10 mM imidazole buffer (pH 7.0) containing 10 mM dithiothreitol and suspended in the same buffer (about 100 mg bacteria wet weight/ml). DNase was added to the suspension (1 mg/ml) and the crude extract prepared by sonication at 400 W in a Brown Labsonic 1510 sonicator for four 1-min intervals at 0' C. The cell fragments were removed by centrifugation and discarded. Crude extracts were stored in liquid nitrogen or processed immediately for enzyme purification. Synthesis ofmannonic acid (151 90 g of BaCO, dissolved in 600 ml water saturated with C 0 2 were mixed with 27 g D-mannose, followed by 30 ml Br,. After mixing for 10 min at about 20 'C, the mixture was cooled down at 12 "C and excess bromine was removed by the addition of a mixture of 150 ml benzol and 150 ml linseed oil. The suspension was filtered and the filtrate was treated with 45 g Ag2S04 until no Ba2+ was detected. Excess Ag2S04 was removed by drop-wise addition of a solution of 1% BaCI,. The suspension was filtered and the filtrate was passed through a column of Dowex SOW8 (Bio-Rad), concentrated at 40°C under vacuum in the presence of 2 m12 M HCl and recrystallized in 95% ethanol. Synthesis of rrltronic acid (151 SO g D-ribose dissolved in 350 ml water at 3 "C was mixed with 21.75 g NaCN in 150 ml water for 24 h at 5°C. The mixture was then heated for 2 h in a boiling water bath and further boiled for 6 h. The warm mixture was passed through a column of Amberlite IR-120 which has been previously equilibrated in the Ca2+form by passing 10% CaClz followed by thorough washing with water. The flow-through was concentrated under vacuum and recrystallized in 95% ethanol.

Assay of D-mannonate and D-altronate hydratases The procedure followed the outlines of [15] with some modifications. Extract or water (200 pl) was mixed with 200 pl 0.4 M Tris/HCl buffer pH 7.5 containing 340 mM 2mercaptoethanol, 1.6 mM FeS04 and 30 mM substrate. The mixture was incubated for 1.5 - 3 h at 37°C and the reaction was stopped by the addition of 600 pi 10% trichloroacetic acid containing 20 mM HgC12. The precipitate was removed

Chemical and biochemical analyses For activation studies, all vessels were boiled for 1-2 h in 1 M Suprapure HCl (Merck) to remove traces of heavy metals and the vessels were thoroughly rinsed with glassdistilled water. Reagents and buffers were freed from contaminating metals by treatment with Chelex-100 (BioRad). Iron was determined either by the method described in [I61 or according to [17]; acid-labile sulfide was determined according to [18]. Protein concentration was determined by the biuret method [19] with bovine serum albumin (Sigma) as standard. Polyacrylamide gel electrophoresis in the presence of SDS was performed on slab gels (100 x 150 x 1 mm) as outlined in [20]. Anaerobic techniques followed the outlines of [21] or [7]. Light absorption spectroscopy was performed on a Cary 2200 spectrometer interfaced with an Apple IIe microcomputer. EPR spectra at X band and at < 77 K were obtained on a Varian 109 spectrometer interfaced with a HewlettPackard 9825A microcomputer for data acquisition and equipped with a flow-through liquid-helium cryostat from Oxford Instruments Inc. Instrument settings were as given in the figure legends. All spectra used for quantitative work were recorded at non-saturating power levels. The g values of the various components were identified from the microwave frequency and the magnetic field [calibration with diphenylpicrylhydrazyl and Mn(I1) in MgO] measured by a Systron Donner 1017 frequency counter and SD 1292 converter. Evaluation was made from the position of the free radical and from the nominal scan of the field dial. At a given temperature, suitable power levels were determined by plotting the height of a prominent spectral feature versus l/P,where P is the microwave power. RESULTS Purification of'~-altronatrand L-mannonate hydratases All procedures were carried on at 0 - 4°C. About 25 ml crude extract was absorbed on a column (2.5 x 40 cm) of DEAE-Trisacryl (LKB) equilibrated in 25 mM imidazole buffer (pH 7.4). The flow rate was 5 ml xcm-' x h-'. The column was washed with 60ml buffer and developed by means of a linear gradient (600 mi) of 0.15 -0.35 M NaCl in the initial buffer. Active fractions were pooled, and concentrated to 5 - 6 ml in an Amicon microfiltration cell using a PM-10 membrane. The concentrate was applied on a column (2.5 x 100 cm) of AcA-44 (LKB) equilibrated in 25 mM imidazole buffer (pH 7.4) containing 0.3 M NaCl. The enzyme was eluted in that same buffer at a flow rate of 2 mi x c m p 2x h-' and the eluate was collected in fractions of 5 mi. At this stage mannonate hydratase could be separated from altronate hydratase. The fractions with the highest specific activities were pooled separately for each enzyme and

625 Table 1. Purification of L-altronate and L-mannonate hydratases One unit, U, is the amount which catalyzes the production of 1 pmol3-deoxy-2-~-g~ucu~osonate/min under the experimental conditions given (see Methods) Fraction

Crude extract 1st DEAE-Trisacryl AcA-44 2nd DEAE-Trisacryl 3rd DEAE-Trisacryl

D-Altronate hydratase total

specific

activity

activity

U

U/mg

63285 52 600 46 400 42 300 36080

1.6 3.8 12.2 58.3 221 .o

D-Mannonate hydratase yield

purification

total activity

specific activity

yield

purification

Yo

-fold

U 93218

U/mg

Yo

-fold

2.3 15.8 19.7 49.3 203.1

66.5 53.3 47.6

83 73 66

51

each sample was desalted by passage through a column of Sephadex G-25 equilibrated in 25 mM imidazole buffer (pH 7.4). Each batch of enzyme was re-applied individually to the previous column (DEAE-Trisacryl) which had been reequilibrated in the mean time in 25 mM imidazole buffer (pH 7.4). The flow rate was kept at 5 ml x cm-2 x h - l as in the first run. After washing off unbound protein with about 100 ml of initial buffer, the enzymes were eluted by means of a linear gradient (500 ml) of 0.1 - 0.5 M NaCl in the equilibration buffer. Pooled active fractions were diluted twice and once more applied to a column (2.5 x 40 cm) of DEAETrisacryl (LKB) which however had been equilibrated in 25 mM imidazole buffer (pH 7.4) containing 0.18 M NaCl. The column was developed by means of a linear gradient (600 ml) of 0.18-0.25 M NaCl in the same buffer. Active fractions were frozen drop-wise and stored in liquid nitrogen as 20 - 40-p1 pellets until use. The purification scheme outlined in this paper is presented in Table 1 which represents a typical run. The enzymes purified according to this procedure usually yield preparations of >95% purity. Both enzymes exhibited a single peak on gel electrophoresis in the presence of SDS. They showed an apparent molecular mass of 53 1 kDa (altronate hydratase) and 40 f 2 kDa (mannonate hydratase) respectively, in accordance with data obtained in vitro by means of gene expression [22, 231. Occasionally minor contaminants were observed, particularly in preparations of D-altronate hydratase. Both enzymes are relatively stable under the experimental conditions used, at least up to the AcA-44 step. Subsequent chromatographic steps (the second and third DEAE-Trisacryl steps) should be performed rapidly to avoid irreversible loss of enzyme activity, particularly in the case of D-altronate hydro-lyase. The material used throughout this study was devoid of other Fe-binding proteins and exhibited no contamination with other enzyme activities tested, particularly other Fedependent hydro-lyases. The EPR studies reported in a later section of this paper were also performed on the pure enzymes. Activation studies

The activation pattern of aldonic acid hydratases is a complex process because of the numerous parameters involved. No systematic study has been attempted to resolve this complexity, but the major observations will be summarized in this section. Efect of p H , buffers and salts. Aldonic acid hydratases exhibit a sharp pH optimum around pH 7.5 in 50-200 mM

78 100 62 000 49 670 44 320

2.4 7.6 36.5 138.0

83.7

6.9 8.6 21.1 88.3

Tris/HCl. The activity at pH 6.5 and 8.6 is only about 20% of the maximal achievable activity at optimum pH. Concentrated buffers or salts (above about 50 mM) linearly inhibit the activity. Anaerobicity. Both D-altronate and D-mannOnate hydratases are inactive in the absence of a bivalent metal and a reducing agent, even in crude extracts. Our past experience on Fe-hydratases has shown that in most instances the requirements for reducing agents such as cysteine, dithiothreitol, ascorbate or 2-mercaptoethanol and ferrous iron implies anaerobicity in the activation procedure, to investigate the effects of reducing agents as such. However aldonic acids undergo coordination bonding with Fe2+,the stability of the complexes being pH-dependent [24]. The order of stability of the iron-aldonic acid complexes is much greater than, for example, citrate or tartrate [25]. Upon polarimetric titrations the complexes exhibit reversible oxidation-reduction waves of Fe3+-Fe2+ whose Eli2 varies with pH [26]. Under our experimental conditions, the complexes stabilize ferrous ions against oxidation. These data obtained by other authors explain our early findings that, under normal experimental conditions, i.e. in the presence of 30mM substrate in the medium, no significant change is observed whether the activation is performed aerobically or under Ar2. Thus no efforts were usually undertaken to exclude oxygen, unless specified. Nevertheless in experiments where the enzyme were activated in the absence of substrate, then transferred to the assay medium containing the substrate, anaerobicity was required during activation and transfer to achieve reproducible results. Fe and Mn. Table 2 shows the effects of some metals on the activation process. The activation by 0.8 mM Fe2+ in the presence of dithiothreitol has been used throughout this study as a general reference, but it is apparent that Mn2+ also induces high specific activities. This effect is not due to contamination by adventitious iron, since our reagents contained less than 0.001% iron as contaminant (from chemical analysis). When both Mn2+and Fe2+ ions are present, a synergystic activation is observed (Tables 2 and 3). Concentration-dependence studies (Fig. 1) show that Fe2+ activates aldonic acid hydratases at low concentrations but is inhibitory above about 2 mM, while the activation by Mn2+ alone requires higher concentrations of the metal. Other metals. No activity was generated with Fe3' ions, even in the presence of excess dithiothreitol. Furthermore, various bivalent metals such as Co2+,Ni2+,Zn2+,Mg2+ and Ru2 have been tested under different experimental conditions for their ability to activate the aldonic acid hydratases (Table 2) but no significant activity was observed +

626 with these metals, except for Co2+ ions. Chemical analysis gave less then 0.01 % iron contaminant in C o 2 +reagents used, excluding the possibility that the activation by cobalt could be attributed to adventitions iron. The activation pattern induced by Co2+ions on the enzymes is depicted in Fig. 1. Thiols. The activation by metal ions varies very much with the concentration of thiols in the buffer. Activation by iron exhibits a very sharp maximum at 170 mM dithiothreitol or 2-mercaptoethanol, whereas only 20% optimal activity is developed in the presence of 120 mM or 240 mM thiol. On the other hand, Mn-activated samples exhibit a similar effect but with a much broader pattern with an optimum at 230 mM thiol, about 30% maximal activity being reached with 120 mM or 700 mM thiol, respectively. Sulphide. Sulphide ions exhibit a slight activating effect in the concentration range studied (up to 10 mM), but the

observed differences (Table 2 ) are minimal when compared to observations performed under similar experimental conditions on maleic acid hydratase [7], an Fe-hydratase with a strong requirement for sulphide ions. Various influences. The activation patterns for both Daltronate and D-mannonate hydratases are very similar (Tables 1- 3 and Fig. 1). Higher specific activities are reached for the latter enzyme upon activation with Fe and/or Mn: this is probably related to the greater instability of D-altronate hydratase, which readily loses activity (and iron, see next section) upon storage or handling.

Table 2. Activation ofaldonic acid h.ydratase.7 The purified enzymes were rapidly desalted by centrifugation through a small column of Sephadex G-25 according to [7,22] and incubated at 37°C for 3 h in 0.2 M Tris/HCl pH 7.5. Added reagents were 170 mM dithiothreitol, 15 mM substrate and metal(s) at optimal concentrations, i.e. Fe2+, 0.8 mM as 0.05 M Fe(NH4)2S04; C 0 2 + , 1 mM, Mn2+ 50mM (both as 1 M Me2+C12 salts); Niz+, Zn2+, MgZ and Fe3', 5 mM (also as 1 M chloride salts all in 0.01 M Suprapure HCl). All solutions (except metals) were purified by passage through Chelex-100. 100% activation represents enzyme activity reached according to the standard procedure (see Materials and Methods)

mM

+

Effectors

Activation of altronate hydratase

mannonate hydratase

__--0 01

None Fez+,dithiothreitol Fe2+,dithiothreitol, SzMn2+,dithiothreitol M n Z f ,Fez+,dithiothreitol Co2+,dithiothreitol NiZt, dithiothreitol ZnZ', dithiothreitol Mg2', dithiothrcitol Fe3+

5.0 100.0 122.0 152.0 375.0 80.0 5.0 5.0 5.0 5.0

5.0 100.0 105.0 158.0 700.0 80.0 7.0 7.0 8.0 6.0

01

I

10

100

rnM

Fig. 1. Activation of D-altronate hydratase ( A ) and o-mannonate hydratase ( B ) by Fez+, Coz+ and M n 2 + ions. The enzymes (1 mg/ ml), desalted on Sephadex G-25 as described in the text, were activated in 50mM Tris/HCl pH 7.5 in the presence of 170mM 2mercaptoethanol, 15 mM substrate and varying concentrations of the corresponding metal ion. All solutions, except for the metals were treated with Chelex-100 to remove trace contamination by metal ions. (0---0 Activation with F e z + ; ( 0 .. . . . O ) activation with C o z + ; (0-0) activation with M n 2 + . Activity is expressed in nmol substrate transformed min- ' (mg protein)-

Table 3. Effect of Fe and M n on the activation of aldonic acid hydratases The purified enzymes were activated and assayed as described in Materials and Methods in the presence of variable concentrations of Fe2 or M n Z +as indicated. The activity is expressed with reference to standard conditions of assays (= 100%). All solutions (except metals) were purified by passage over Chelex-I 00, to remove trace metals Enzyme

Fe2 '

Activation by Mnz at +

0

0.01 mM

0.1 mM

1 mM

10 mM

50 mM

Mannonate hydratase

0 0.05 0.50 5.00

4.0 120.0 910.0 11.0

5.0 150.0 1 184.0 11.0

9.0 196.0 1460.0 27.0

11.0 222.0 2172.0 25.0

420.0 1010.0 2238.0 33.0

976.0 1240.0 2105.0 56.0

Altronate hydratase

0 0.05 0.50 5.00

4.0 143.0 400.0 22.0

5.0 196.0 899.0 33.0

9.0 231.0 1190.0 45.0

11.0 258.0 1492.0 48.0

210.0 335.0 445.0 52.0

317.0 396.0 342.0 56.0

627 Table 4. Fe incorporation of aldonic acid hydratases The enzymes were desalted as in Table 1. The desalting was repeated three times. The samples (in triplicates) were then activated with 0.8 mM Fez+ in 0.2 M Tris/HCl pH 7.5 and 170mM 2-mercaptoethanol. Activation proceeded: (A) in the absence of, (B) in the presence of 15 mM substrate or (C) in the presence of 5 mM Na2S. After an activation period of 5 min, the samples were quickly desalted again by three subsequent passages through Penefski columns (according to 1221). The samples were split and tested for activity and iron content. Control blanks consisted of unactivated samples tested also for iron content. 100% activity represents the enzyme activity obtained under standard conditions Mode of activation

Table 5. Inhibition of aldonic acid hydratases by chelators or thiol reagents Purified D-altronic acid or D-mannonic acid hydratases were activated for 10 min in 0.2 M Tris/HCl pH 7.5 containing 0.8 mM Fez+, 170 mM 2-mercaptoethanol and 15 mM substrate. The samples were then incubated for 3 h at 37°C in the presence of the indicated inhibitors (final concentrations as stated below). 100% represents enzyme activity tested in the absence of inhibitor Inhibitor

Concn

Residual activity altronate hydratase

mannonate hydratase

Fe content (residual activity) of mM altronate hydratase

mannonate hydratase

mol/mol protein (YO) A. No effector present B. With substrate C. With NaZS

1.1 (108.0) 1.2 (91.3) 0.9 (72.0)

1.02 (98.0) 1.4 (87.0) 1.1 (78.5)

It is apparent from Fig. 1 that high concentrations (> 2 mM) of Fe, but not Mn, are inhibitory. If however the organisms were grown on high Fe concentrations, adapted enzymes could be obtained which exhibited no inhibition towards Fe (up to 15 mM).

Fe chelators o-Phenanthroline

3 16

18.0 0

25.0 7.8

cc-Picolinate

3 16

24.0 0

37.0 12.3

EDTA

3 16

23.0 0

25.0 13.4

Nitrilotriacetic acid

3 16

31.0 0

29.8 9.8

3 16 3 16

106.0 8.0 100.0 11.8

85.6 63.5 57.2 55.7

Thiol reagents p-Chloromercuribenzoate

Iodoacetate

Iron incorporation

Previous studies on Fe-hydratases [4, 81 have reported on the incorporation of iron upon activation. Under similar experimental conditions iron incorporation has been observed in aldonic acid hydratases as well (Table 4). In this type of experiment, the enzymes were activated in the presence of Fez+ and 2-mercaptoethanol for 5 min and desalted several times through a series of small columns of Sephadex G-25. After each passage the iron content was tested on an aliquot, until no change could be observed over three consecutive passages. This corresponds to the Fe 'tightly bound' to the enzyme. Enzymatic activities were tested also after each passage and were practically constant over the first three filtrations, but they decreased in most cases on further handling. The values indicated in the table correspond to an average of five different experiments. As indicated both enzymes retained about 1 mol iron/mol enzyme (assuming the enzymes are monomeric, with molecular masses of 53 kDa for altronate hydratase and 40 kDa for mannonate hydratase respectively, as mentioned in a previous section). However upon subsequent treatment of the activated and desalted enzymes with EDTA, again followed by several passages through Sephadex G-25 columns, >95% of this 'tightly bound' iron was removed. The iron incorporated into the enzymes upon activation exhibited a rather high Kd,estimated approximately as 4.5 mM and 20 mM for D-mannonic and Daltronic acid hydratases, respectively. In the presence of NazS no significant change in the amount of iron incorporated was observed. A small decrease in the maximal activity achieved after three consecutive gel filtrations was found (in contrast to data from Table2, obtained under different experimental conditions). The effect is minimal however, compared to similar experiments on maleic acid hydratase [7].

Effect of chelators and thiol reagents

The two enzymes described in this work are insensitive towards N-ethylmaleimide but exhibit different inhibition patterns towards other thiol reagents presented in Table 5. D-Mannonate hydratase is relatively insensitive towards pchloromercuribenzoate or iodoacetate, and even at high concentrations of inhibitors (> 16 mM) the enzyme still exhibits 55 - 65% residual activity. By contrast, D-altronate hydratase is insensitive to thiol reagents up to 3 mM, but complete inhibition can nevertheless be achieved with high concentrations of inhibitors (Table 5). From these data, considering the high concentrations required for the latter inhibition, it is doubtful that a thiol group is involved in the enzyme active center, in contrast with observations reported under similar experimental conditions on maleic acid hydratase [7] or on aconitase [3] where the binding site for the substrate consists of an active thiol group, located at the enzyme active center and in the vicinity of the Fe-S cluster [l]. When the inactive enzymes were activated and then treated with iron chelators such as o-phenanthroline, nitrilotriacetic acid, a-picolinate or EDTA [27], a concentration-dependent inhibition of enzyme activity was observed (Table 5 and Fig. 2). The sensitivity towards chelators is more pronounced in D-altronate hydratase, where iron appears to be loosely bound to the enzyme. When adding back iron equimolar to the chelator present in addition to the amounts of iron required for activation, normal activity is reached again. Fig. 3 shows the relationship of the Fe/EDTA ratio and enzyme activity under reducing conditions. At low iron concentration, little activity is developed in the presence of EDTA; as the Fe2+ concentration is increased, the enzyme competes with EDTA for Fe2+ and regains activity. From Fig. 3 it is apparent that about 1 mol iron/mol EDTA is re-

628 loo

1

/,.-

A

,,,I”,

0

05

20

10 Fe/EDTA ratio

Fig. 3. Effect of Fe to EDTA ratio on the activation ofD-I??annonate hydratase. Mannonate hydratase (0.5 mg) was activated in 50 mM Tris/HCl (pH 7.5) containing 170 mM dithiothreitol, 0.5 mM EDTA and varying iron concentration followed by incubation for 1.5 h at 37°C in the presence of 30 mM o-mannonate

B

20 K (Fig. 4B and Fig. 5), an indication that the populations of the Kramer’s doublets are unequal. However, a more quantitative estimation of the conversion from the g = 1.98 to the g = 4.15 species is not possible, without an estimation of the zerofield splitting parameters. The species at g = 1.98 is only observable under aerobic conditions in the presence of a thiol and below about 40 K. The EPR parameters (g-tensor, peak-to-peak width, power and saturation properties, temperature dependence) are not compatible with those of Fe-S clusters. As the g = 1.98 signal in our enzymes arises upon activation with Fez+ and a thiol and can only be observed under aerobic conditions, i.e. under conditions where autoxidation of Fez may occur, the signal can tentatively be attributed to low-spin ferric iron. Theoretical calculations correctly predicting low g values for low-spin Fe(II1) have been described elsewhere [28]. The spectra of samples reacted with substrate, with resonances at g = 4.15 and 9.07, are strikingly similar to those observed on rubredoxin-type proteins and can be interpreted as arising from the transitions within the ground and first excited doublet state (middle Kramer’s doublet) of high-spin Fe3+ ( S = 5/2) with a high degree of rhombic distortion [29].

’

629

."I JI

I"

I

L 15

907 883 671

I

I IC

2 28 2 11 2 02 1 98

Fig. 4. EPR spectra at 13 K ofpure D-r?'UZf?nOnate hydratase activated with Fez+ in thepresence of substrate. Mannonate hydratase (about 6 mg/ ml) in 50 mM imidazole pH 7.6 containing 170 mM 2-mercaptoethanol was activated with 1 mM Fez+ [as 0.05 M Fe(NH4)2S04]for 5 min and split into two EPR tubes, the second one (sample B) containing 5 p10.6 M D-mannOnate. Sample B was mixed as quickly as possible and both tubes were instantly frozen in a cold solution (80 K) of isopentane/methylcyclohexane( 5 : 1).Temperature 13 K ; microwave power 1 mW; modulation frequency 100 kHz; modulation amplitude 0.4 mT, microwave frequency 9.1 13 GHz

A

132

9

229 212198

Fig. 5. EPR spectra at 4.2 K ofpure D-altronate hydratase activated with Fe2+ and reacted with substrate. Sample preparation and instrument setting were as in Fig. 4. Temperature 4.2 K

As the temperature is lowered the g = 4.15 signal decreases in intensity as the middle Kramer's doublet becomes depopulated and the g = 9.07 originating from the ground state increases [29]. The function of the second metal-binding site for Mn has not been investigated and Mn, if present in the native (unactivated) enzyme, is not EPR-detectable at low temperature. If present it may then be in form of diamagnetic Mn(II1). However Mn(1II) is rarely present in nature, except for two enzymes, i.e. plant acid phosphatases and Mn-superoxide dismutase [30].

CONCLUSION As shown in this and other studies [13- 151, aldonic acid hydratases typically require ferrous ions and a thiol as reducing agents for activity and thus they belong to a class of Fe-dependent hydro-lyases acting on carboxylic acids, of which aconitase and maleic acid hydratase, the best characterized representatives, are known to be iron-sulphur proteins. More recently, another enzyme, tartrate hydratase,

induced in Pseudomonasputida has been shown to be an Fe-S enzyme also [31]. Many questions remain concerning the role of iron in hydratases, particularly with a view to rationalizing the fact that some enzymes exhibit an absolute requirement for this metal, whereas others, closely related ones, fail to do so. It has been tempting to look at data on aconitase and maleic acid hydratase as general models for Fe-hydratases, implying the occurrence of an Fe-S cluster in similar enzymes as well. However the observations reported in this paper show striking differences between aldonic acid hydratases and the more thoroughly investigated Fe-hydratases. a) One mole of iron is required for activation and is incorporated into the enzymes (this incorporated iron might bear resemblances with the Fe, of activated aconitase). b) The enzymes present a second metal-binding site for Mn and are synergistically activated by MnZ ions, full activity being gained at high Mn concentration even in the absence of Fe (a similar situation is encountered in the case of tartrate hydratase (EC 4.2.1.81), where metal requirement varies depending upon the source [32, 331 and the mode of enzyme induction in the micro-organism). +

630 c) N o requirement for Na,S has been observed, in contrast to malease (it should be remembered, however, that the requirement for sulphide in malease is itself unusual). d) No essential thiol residue seems to be involved in substrate binding or enzymatic activity. e) Optical and EPR spectroscopic data rule out the presence of an iron-sulphur cluster in aldonic acid hydratases, in strong contrast to aconitase and malease. f ) The interpretation of the g = 1.98 signal is difficult in the absence of data from other spectroscopic methods (e.g. Raman or Mossbauer) and a better knowledge of the metal environment. The signal is present under aerobic conditions, and thus it may originate from Fe3+.Chemical analysis indicate the presence of a single iron atom per molecule of enzyme, excluding binuclear Fe-models (like e.g. in purple acid phosphatases [28] or in Fez& proteins). Alternative explanations could assume a spin-coupled Fe-Mn complex but, to our knowledge, such models are unrecorded in biological systems and highly unlikely to occur in our enzymes, because our EPR samples were activated with Fe alone (no Mn). Since about 60-70% of the endogenous Mn is lost upon purification, no more than 0.4 spin/mol enzyme could be expected from this signal, while we found about 0.9 spin/mol upon double integration of the signal (assuming an S = 1/2 system, with Cu-perchlorate as standard). Another reasonable interpretation assigns this signal to low-spin ferric iron. Low g values for low-spin Fe3+ have been reported [28], but lowspin Fe3+ is uncommon among non-heme iron proteins. g) The binding of substrates to iron-activated samples dramatically affects their spectroscopic properties and the iron exhibits properties similar to rubredoxins, strongly suggesting hig-spin ferric iron in a thiol environment. The present work was supported by a grant (3.060.0.84) from the Swiss National Foundation. I am grateful to Dr Ritzenthaler and Dr Stoeber for helpful discussions and for the supply of bacteria mutants at an early stage of this work.

REFERENCES 1. Beinert, H. & Thomson, A. J. (1983) Arch. Biochem. Biophys. 222,333 - 361. 2. Kennedy, M . C., Emptage, M. H., Dreyer, J.-L. & Beinert, H. (1983)J. Bid. ('hem. 258, 11098-11105. 3. Emptagc, M. H., Dreyer, J.-L., Kennedy, M. C. & Beinert, H. (1983) J . Bzol. ('hem. 258, 11 106-11 111.

4. Beinert, H., Emptage, M. H., Dreyer, J.-L., Scott, R. A., Hahn, J. E., Hodgson, 0. K. & Thompson, A. J. (1983) Proc. Nut1 Acad. Sci. USA 80, 393 - 396. 5. Emptage, M. H., Kent, T. A,, Kennedy, M. C., Beinert, H. & Munck, E. (1983) Proc. Natl Acad. Sci. USA 80,4676-4678. 6. Kennedy, M. C., Kent, T. A,, Emptage, M., Merckle, H., Beinert, H. & Munck, E. (1985) J . Biol. Chem. 259,14463 - 14471. 7. Dreyer, J.-L. (1985) Eur. J. Biochem. 150, 145-154. 8. Smiley, D. S. & Ashwell, G. (1960) J. Biol. Chem. 235, 15711575. 9. Ashwell, G., Wahba, A. J. & Hickman, J. (1960) J . Biol. Chem. 235, 1559-1565. 10. Stoeber, F., Lagarde, A., Nemoz, G., Novel, G., Novel, M., Portalier, R., Pouyssegur, J. & Robert-Baudouy, J. (1974) Biochimie (Paris) 56, 199-213. 11. Robert-Baudouy. J., Portalier, R. C. & Stoeber, F. (1974) Eur. J . Biochem. 43, 1 - 15. 12. Pouyssegur, J. M. & Stoeber, F. (1974) Eur. J . Biochem. 21,363373. 13. Robert-Baudouy, J., Jimeno-Abendano, J. & Stoeber, F. (1975) Biochimie (Paris) 57, 1-8. 14. Robert-Baudouy, J., Jimeno-Abendano, J. & Stoeber, F. (1982) Methods Enzymol. 90,288 -294. 15. Robert-Baudouy, J. (1971) Thesis, Lyon University. 16. Beinert, H. (1978) Methods Enzymol. 54, 435-445. 17. Lauber, K. (1965) J. Clin. Chem. 3, 96-99. 18. Beinert, H. (1983) Anal. Biochem. 131, 373-378. 19. Gornall, A. G., Bardawill, C. J. & David, M. M. (1949) J . Biol. Chem. 177,751 - 766. 20. Weber, K. & Osborn, M. (1969) J . Bid. Chem. 244,4406-4412. 21. Orme-Johnson, W. H. & Beinert, H. (1969) Anal. Biochem. 32, 425 - 435. 22. Ritzenthaler, P., Mata-Gilsinger, M. & Stoeber, F. (1981) J. Bacteriol. 146, 181- 190. 23. Ritzenthaler, P., Mata-Gilsinger, M. & Stoeber, F. (1980) J . Bacteriol. 143, 1116 - 1126. 24. Tanabe, H. & Okamoto, K. (1958) Yakugaku Zasshi 78, 951 957. 25. Gergely, A. & Mjozes, J. (1965) Mag-v. Kem. Foly. 71, 309 - 312. 26. Tanabe, H. & Okada, K. (1962) Takeda Kenkyusho Nenpo 2 1 , l 25. 27. Harris, D. C. & Aisen, P. (1973) Arch. Biochem. Biophys. 329, 156-158. 28. Antanaitis, €3. C. & Aisen, P. (1980) J . Bid. Chem. 257, 18551859. 29. Orme-Johnson, W. H. & Sands, R. H . (1970) in Iron-sulfur proteins, vol. 2 (Lovenberg, w . , ed.) pp. 195-238, Academic Press, New York. 30. Kawabe, H., Suguira, Y., Terauchi, M. & Tanaka, H. (1984) Biochim. Biophys. Acta 784, 81 -89. 31. Kelly, J. M. & Scopes, R. K. (1986) FEBS Lett. 202, 274-276. 32. Rode, H. & Griffhorn, F. (1982) J . Bacteriol. 150, 1061- 1068. 33. Rode, H. & Griffhorn, F. (1982) J . Bacteriol. 151, 1602-1604.