Comparative inhibition of hepatic ...

Can. J. Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by Renmin University of China on 06/07/13 For personal use only.

Comparative inhibition of hepatic hydroxymethylbilane synthase by both hard and soft metal cations D. J. FARMER AND B . R. HQLLEBQNE Department of Chemistry, Carleton University, Ottawa, Ont., Canada K l S 5B6 Received January 31, 1983 Farmer, D. J. & Hollebone, B. R. (1984) Comparative inhibition of hepatic hydroxymethylbilane synthase by both hard and soft metal cations. Can. J . Biochern. Cell Bisl. 62, 49-54 The in vias inhibition of hydroxymethylbilanesynthase (K4.3.1.8, uroporphyrinogen I synthetasej obtained from livers of Sprague-Dawley rats has been studied with a wide range of di- and tri-valent metal ions. After purification by cell lysis, heat treatment, and centrifugation, the stable, soluble enzyme yielded sigmoidal inhibition curves with increasing concentrations of each of the 16 test ions. Using the negative logarithm of metal concentration for 50%inhibition (the pM50 value), the metal ions could be classified according to their Klopman hardness values. Very soft ions including H ~ ~intermediate + , ions including c?', and very hard ions including A13+ all yielded large pMSovalues indicating strong inhibition. In comparison to known metal-ion chemical behaviour, these thee ions could indicate three different types of inhibitory binding sites at or near the active site: ~ g corresponding ~ + to sulfur in cysteine, c?+correspondingto nitrogen in histidine, and A13+ corresponding to oxygen in carboxyl groups. The presence of the first two sites is also indicated by the pH dependence of activity. Farmer, D. J. & Hollebone, B. R. (1984) Comparative inhibition of hepatic hydroxymethylbilane synthase by both hard and soft metal cations. Can. J. Biochern. Cell Biol. 62,49-54 Nous avons ktudik l'activitd inhibitrice in vitro d'un grand n o d = d'ions dtaJliques di- et tri-valents sur l'hydroxymkthylbilane synthase (EC 4.3.1.8, urogorphyrinog&neI synthktase) obtenue des foies de rats Sprague-Dawley. Apres purification par lyse cellulaire, tfaitement thennique et centrifugation, l'enzyme soluble stable donne des courbes d'inhibition sigmoi'de avec des concentrationscroissantes de chacun des 16 ions essaybs. tltilisant le logarithme nkgatif de la concentration mktallique donnant une inhibition de 50% (la vdeur pM50), les ions mktalliques peuvent Ctre classifibs selon leur duretk (valeurs de Klopman). Les ions t d s doux dont FIg2+, les ions intermediaires dont C?' et les ions tres durs dont Al" donnent tous des valeurs klevkes de pM50, preuve d'une forte inhibition. Si nous les comparons au comportement chimique connu des ions mktalliques, ces trois ions gourraient indiquer tfois types diffkrents de site de liaison inhibitrice au site actif ou pr&sde ce site: H ~ ' +correspondraitau soufre dans la cystkine; C?, A l'azote de l'histidine et A13+, 8 l'oxyg&nedes groups carboxyle. La dbpendance de l'activitk au pH montre aussi la prdsence des deux premiers sites. [Traduit par la revue]

or "hard" donor atoms have been called "a" type or "hard", while those binding to polarizable or-"soft" atoms are called "b" or "soft" cations (4-6). Hard cations bind strongly to hard donor atoms and vice versa, while bonds between hard cations and soft donors are relatively weak. The different metals identified in the derangement of heme synthesis (2) clearly fall into both the soft and intermediate categories. Most detailed research on this system, however, has been directed at the action of the soft cations on the membrane-bound ALA-D and ferrochelatase (2, 7, 8). The effects are severe and the blocking of thiol sites in both systems is strongly implicated. Less attention has been paid to derangement of the steps between A M - D and ferrochelatase which proceed through a series of enzymes in the cytosol phase. One of these steps, a combined sequence of condensation and cyclization (9) which converts porphobilinogen to uroporphyrinogen 111 involving HMB-S ABBREVIATIONS: ALA-D, aminolevulinic acid dehydratase; (EC 4.3.1.8), has been examined with a few soft metals. HMB-S, hydroxymethylbilane synthase; MW, molecular Russell and Rockwell tested for inhibition by organomercurials and again concluded that one or more essenweight; max, maximum.

Introduction Hepatotoxicity has been observed with a large number of metals in various oxidation states (1) at a variety of cellular or biochemical targets for metal-ion intoxication. Some metals have been implicated in specific derangements of the hematopoietic system (2). These include not only main group elements such as cadmium, mercury, thallium, and lead, but also transition elements such as cobalt and manganese. The reactions and stabilities of compounds involving these elements are quite diverse. Depending on the oxidation state, these metals are bound preferentially to different types of ligands involving donor atoms which can range from the highly electronegative oxygen or fluoride to the much more polarizable sulfur or iodide (3). This diversity of behaviour has been used as the empirical basis for a classification scheme for metal ions. Ions which bind preferentially to electronegative


50

CAN. J. BIOCHEM. CELL BIOL. VOL. 62, 1984

tial sulfhydryl groups may reside at the active site of the enzyme (10). This inhibition was reversible with dithioerythritol, suggesting that mercurials do not denature the enzyme. Shioi et al. noted the same behaviour in very similar experiments (1 1). By comparison, the action of harder cations on the enzyme has not been well studied. Tephly examined the inhibition of ferrochelatase by co2', but pointed out that this ion could well affect other steps in heme synthesis (2). As an initial experiment, the inhibition of the HMB-S system (12) has been studied with a range of metals. The hydroxymethylbilane synthase fraction has been analyzed and was found to be very low in sulfur but high in acidic amino groups, resulting in an acidic isoelectric point (1 1). No cysteine was recovered and only two equivalents of methionine were found in a molecular weight of 33 000. These results are not consistent with the conclusions concerning essential thiol groups from the inhibition studies with mercurials (lo), but could explain why the observed mercurial inhibition is reversible. In this study, the inhibition of this enzyme system with a spectrum of metal ions has been carried out to examine this question more thoroughly.

Experimental procedures PuriJicationof rat liver hydroxymethylbilane synthase Male Sprague-Dawley rats of weights 200-250 g, obtained from Canadian Breeding Farm and Laboratories, were killed by CU2 asphyxiation. The livers were excised, cooled to 4"C, minced, and homogenized in g0mL of 188mM Tris buffer (pH 7.4) containing 1.15% (w /v) of KCl. The homogenate was centrifuged at 10 0 0 x g for 20 min. This supernatant %-10was recentrifuged at la) 000 x g for 60 min at 4°C to produce the supernatant S- 100. The supernatant S- 100 was heated .at 60°C for 20 min to denature any uroporphyrinogen 111 synthase and cytosolic proteins (12). The sample was recentrifuged at 12 800 x g for 30 min and the remaining supernatant S-100* was frozen at -90°C for further use. In this form the activity was retained up to 12 months. Sephadex G-la) gel fractionation of the S- la)* supernatant (1 1) was performed on a previously swollen, cooled column in a buffer of 10 mM KI-I2FQ4, 50 mM KC1, and 0.1 mM dithioerythritol, pH 8.0. The (3-100 eluate was stable up to 30 days at -90°C. Both S-loo* and G-la) fractions were concentrated when necessary using a 40% (w/w) solution of polyethylene glycol (MW 20 0). The samples were retained in dialysis tubing with a cutoff molecular weight of 5000 and allowed to equilibrate for 10 h or to 90% concentration. Activity of hydroxymethylbilane synfhase A 0.50-mL sample of the fraction to be assayed was added to 1 . 0 mL of 130 mA4 Tris buffer (pH 8.2) and preincubated at 37°C for 5 min with gentle shaking (1 1). A 0.50-mL sample of 35 mM porphobilinogen was then added and incubation was continued for 15-30 min depending on expected activity. The reaction was stopped by addition of 250 p,L of 50% (w/w) trichloroacetic acid. The uroporphyrinogen was oxi-

dized to the corresponding urophorphyrin by addition of 500 pL of 0.01% sublimed iodine in 6 M HCl and exposure to room light for 10 min. The precipitated protein was removed by centrifugation and the supernatant was assayed for uroporphyrin I fluorescence. Sample excitation was optimum at 405.5 nm for emission at 600.5 nm (12). The concentration curve was standardized by comparison with the emission of known concentrations of uroporphyrin I produced by overnight hydrolysis in 1 M HCl of uroporphypinogen I octamethyl ester. It was also verified by measurement of absorbance at 405.5 nm where the extinction coefficient for mporphyrin I is 54 1/nM. Eflects of assay conditions The effects of Tris-HC1 buffer concentrations on the kinetics of the S-loo* or G-100 fractions were assayed by fluorescencein a concentration range from 50 nM to 1.250 M at pH 7.4. The effects of pH on both fractions were also assayed by fluorescence at constant buffer concentration in a pH range from 5.50 to 10.00. At constant buffer concentration and pH, the effects of the 5-min preincubation of the assay mixture were tested. The kinetic data were compared with samples in which porphobilinogen was added immediately. Inhibition by metal ions The effects of metal ions on the kinetics of rat liver HMB-S were studied by replacing the 1.00-mL buffer solution in the fluorescence method with a buffer - metal ion mixture. These mixtures were prepared by serial dilutions of standard metal ion solutions. In each case, six different metal-ion concentrations were prepared over the concentration range from lo4 to M; that is pM values were from 6 to 2. In some cases, low buffer concentrations were required to allow the higher metal-ion concentrations. All tests were done in quadruplicate. Substrate chelation To test for chelation of metal ions by porphobilinogen, thin-layerchromatographyon silica plates was performed with selected substrate to metal mixtures. The solvent was a mixture of n-butsmol, acetic acid, and water in a ratio 8:2:10 and afforded a clear separation of peaks. The test solution was 40 nM in porphobilinogen and from 1 nM to 1 mM in metal nitrate. The plates were run with a chelated control and developed with a modified Erhlich's reagent (13). Data analysis The five data p i n t s obtained in each kinetic experiment were analyzed using both the Eadie-Hofstee (13, 14) and Hanes- Woolf (15, 16) methods. All lines were fit using linear regression programs a d rejection of any data was based on the criteria of Cornish-Bowden and Eisenthal(17) and McCarthy (18). The analysis of inhibition by metal ions, followed Chow et al. (19) and the effects were interpreted as dependance of inhibition factors Fiupon loglo[Mn'], which may be abbreviated as pM. Values of pMWwere obtained from the computed best line for each metal ion.

Results Uninhibited kinetics of hydroqmethylbilane synthase The optimum conditions for kinetics experiments with respect to method of purification, range of protein

51


FARMER AND HOLLEBBNE

and substrate concentrations, preincubation , pH, and buffer concentration were established using the ff uorescence assays. The most convenient purification was the heat treatment used to produce fraction S-l W*. While less active than the G-100 fraction, it was very stable and reproducible from one animal to another. It was used as the best source for the comparative metal inhibition studies because of this reliability. Using protein concentrations in the range 1.O-2.0 mg/mL, the concentrations of the porphobilinogen substrate used were 2,5,10,20, and 40 mM. (Fig. 1). In all cases, preincubation for 5 min at 37'C was necessary, as addition of substrate prior to temperature equilibration yielded very low rates of reaction in the first 2 min. The pH profile yields a maximum of 7.2-7.3. Inflection points are observed at 6.2 and 8.4 (Fig. 2), which correspond closely with pK values observed for histidine at 6.0 and cysteine at 8.3 (20). The Tris-HCl buffer showed a tendency to inhibit the reaction at 50 mM and had a slight but noticeable effect at 100- 150 mM. However, one of the end products of porphobilinogen cyclization nium ion and to maintain a stable pH a buffer concentration of 130 mM was used. Under these conditions the kinetic data all fell on hyperbolic curves and the secondary Wanes-Woolf plots (15, 16) were linear throughout the concentration range (Fig. 1). No substrate inhibition was observed at porphobilinogen concentrations of 30 times Km. Over a series of six rat samples, Vr, ranged from 7 to 12gmol/ mE, per minute and Km from 2.5 to 3 -5 pM. In each sample the correlation of the regression line was 0.995 A 0.003. Over all six samples, the Hill coefficient was 1.OO + 0.005 with a correlation of 0.995 + 0.005.

barium, and cobalt, all are divalent ions except Na'. Magnesium ion is more inhibitory than the others, but Fi does not exceed 0.5 through the concentration range. The second roup includes vanadyl ion (v02'), uranyl ion (U02 + ), manganese, and nickel as divalent ions. The pM50 values were clustered near 2.8 and the behaviour of all four can be fit with one well-defined sigrnoid curve. The third group involves the three trivalent ions aluminum, chromium, and iron, as well as divalent lead. The P M ranges ~ ~ from 3.5 to 4.0 and again all four ions lie approximately on a single sigmoid curve. In the last group, four heavy divalent metal ions appear: copper, zinc, cadmium, and mercury. The pMsQ values are all above 4, but the dependence of Fi on pM varies with the metal ion and cannot be represented by a single, sigmoid curve. These data can be summarized as a correlation between pM50values and the position of the metal ion in the periodic table (Fig. 3). The influence of ionic charge on this correlation requires further consideration (5).

f

Detailed study of mercury inhibition Since mercury salts or organic derivatives have been used to probe active sites on this enzyme system (10, 111, the reversibility of inhibition through dialysis was examined in detail. HMB-S was treated with 5, 10, and 75 p M mercuric chloride to span the concentration range which defines pM for Fi,sO.Inhibition as shown in Figs. 4a and 4b was observed. After overnight dialysis against 100 volumes of buffer, the 25% inhibition observed in the 5 p M sample had been removed and the sample was restored to full activity. The 30% inhibition at 10 p M had been reduced to 25%, but the 100%

Metal-ion inhibition of hydroxymethylbilane synthase The extent of inhibition of the enzyme extract by each of the 16 metal ions was tested over the concentration range from 1 r 6 to 1 r 2M. The ions appear to fall into four categories characterized by pM values for FiVso abbreviated as pMso, which differ by one unit from each other. In the first group which includes sodium, magnesium,

fie;. 1. The Hmes-Woolf plot for uninhibited S-100"

fraction of WMB-S for five concentrations of prphobilinogen.

fie;. 2. (a) pH profile of the activity s f the S-100' fraction of HMB-S. ( b )The first derivative of the pH profile. II md I2 are inflection points at 6.4 md 8.4, respectively.


52

CAN. J. BIOCHEM. CELL BIOL. VOL. 62. 1984

Periodic Group

FIG. 3. Histogram of pfio values against positions of metals in the periodic table.

inhibition at 75 pM was unchanged. Thus, the inhibition may be reversible, but the condition for removal of mercury ion must be more potent than simple dialysis. At the same time as V , was reduced in these experiments, K , was also reduced in proportion. A more complete test of this behaviour confirmed the trend (Figs. 4a and 4b); the pM value for Fi.SOwas 5.0 for both parameters. The rat liver extract prepared in this study closely resembles that reported by Peters et al. (12) from Sprague-Dawley rats. To obtain sufficient stable extract from a single animal with which to test the properties of all the metals, purification was routinely carried only FIG.4. ( a )Plot of V, against pHg2+ for S-100* fraction of through the heat treatment step. As observed by Shioi et HMB-S. ( b ) Plot of K, against pH$+. al. (1 I), this results in three- to four-fold purification with nearly 100% yield of a very stable form of enzyme. that the influence of the extraneous protein in the S-1QO* Subsequent purification steps produce rapidly decreas- fraction on the normal functioning of the enzyme is ing yields of more labile material, making comparison minimal. The extra protein could have an influence on of inhibitory behaviour much more difficult. In spite of the reaction of the- enzyme to metals. This seems this lower level of purification, the observed kinetic unlikely for two reasons. The metal cations, like behaviour (Fig. 1) is very similar to that reported by protons, attack negative sites, and since the pH profile is Peters et al. (12), including the absence of substrate not shifted by the remaining protein, the influence on inhibition. metal cations is also probably small. Secondly, all The pH profile of the S-100* fraction shows an protein in the S-100* fraction is soluble and hydrophilic, optimum at 7.4 as reported by Shioi et al. (11). The suggesting that it carries similar types of binding sites to inflection points observed at 6.4 and 8.4 are indicative of the enzyme protein itself. Hence the pMSovalues may be histidine and cysteine, respectively (20). The latter was shifted as a group to higher values than with the not observed as a constituent of the protein by Shioi et al. homogeneous enzyme because of simultanous absorp(1 1) in HMB-S extracted from Chlorella regularis, but tion on other proteins; the relative order of these values was suggested by the work of Russell and Rockwell in is probably the same. For all metals tested, the inhibition curves display a simple sigmoidal shape similar to the extract from wheat germ (10). Both the substrate and pH behaviour patterns suggest conventional pH titration curves. There is no observable

FARMER AND HBLLEBONE

-

N(fer d ions)

§ ( f o r p ions)

O(for s ions)


cd2+

cu2+

A

FIG. 5. Plot of pM58 values against Klopman hardness parameters for all metal ions tested.

deviation which would suggest more than one binding site affecting the enzyme function. The inhibition pattern of the metal ions examined in this work appears to support the presence of both histidine and cysteine. Jones and Vaughn (6) commented that "there appear to be useful correlations between the toxicities of metal ions and parameters measuring their hardness or softness." This conclusion was drawn from LDS0 values alone, but should apply equally well to isolated enzymes. The basic parameters discovered by Arhland et al. (4) have been refined by Klopman (2 1) in more recent models to permit comparison of behaviour of ions with different formal charges. The Klopman hardness parameters are derived from two types of information, a frontier orbital electronegativity which describes the ability of the metal ion to pol&ze charge from a ligand and a desolvation energy which represents the endothermic term required for removal of solvent before a bond can form to a new ligand molecule or anion. The model does not account for differences between metals which may or may not form double bonds ( n donor or acceptor bonds) with various ligands. This is often a possibility when transition metals are involved. The P M values ~ ~ observed for the HMB-S system can be plotted against these Klopman hardness factors as shown in Fig. 5. The order of hardness is approximately the same as the order of ions in the periodic table (Fig. 3), but there are some important differences. The softest metal ions including cu2+,'Zn2+,cd2+, Hg2+,and pb2+ are all strongly inhibitory and, except for Cu2+, are main group elements. They are usually found bound to soft ligands, in the present case most likely a sulfur donor portion of the enzyme. This

53

observation correlates with the implication of cysteine, from the inflection point at pH 8.4 in the pH profile of enzyme activity. Both pieces of evidence appear to contradict the amine acid analysis of Shioi et al. (1 1) and support the earlier evidence from mercurial inhibition studies. The intermediate metal ions between - I and 4-3 on the Klopman scale fall into two distinct groups. Those with d valence orbitals are strongly inhibitory, while those with s valence orbitals were weakly or moderately inhibitory. The d valence metals are typically bonded most strongly to nitrogen (3). However, there is a clear distinction in pM5()values between those metal ions rich in d electrons (N2+,Co2+ and h4n2+)which are moderately inhibitory and those which are low in d electrons (Fe3+ and c?+) which are strongly inhibitory. This is independent of the charge difference, since the Klopman factor already includes this correlation. The main difference is that I?e3+and C?+ are capable of accepting n charge from ligands, while the others cannot. This type of n charge is available on tertiary aromatic nitrogen as found in histidine, so that this pattern is strongly indicative of histidine at the active site. Trivalent iron similarly binds to imidazole in metheme. By contrast, the s valence group Na+, ~ a ~ ca2+, ' , and Mg2+have no vacant d valence orbitals and cannot accept aromatic nitogen n charge. This is quite consistent with the observed reduction of inhibition by 10' for ions of the same hardness, if the binding site is histidine. This is further confirmed by the moderate inhibition by Vo2+with one d electron and ~ 0with~none. ~ In both ' these ions, there are strong double bonds from oxygen to the metal ion which completely block the d orbitals from accepting further n charge. A third trend is visible in this plot, however, which does involve s valence ions. starting at Na' , through E3a2+, Ca2+, and h4g2+, then including Uo2+ and ~ ' is a growing trend of and ending at ~ 1 there inhibition with hardness. These s orbital ions are almost always bound to oxygen donor ligands (3). The observed increase in inhibition with increasing hardness implies a third important mechanism involving blocking or distortion of the active site through binding of an oxygen functional group. This is not unexpected, since heavy carboxylation is necessary to provide solubility to a molecule of molecular weight of 33 080.However, it does suggest that a free carboxyl group may be important in the structure of the active site, in addition to the cysteine and histidine suggested by the other types of metal ions. Overall, comparison of pMS0 values to Klopman hardness implies that there is no single mechanism for toxicity of metal ions with this enzyme. At least three types of inhibition are suggested by the data. Without further kinetic studies on competitive inhibition there is


54

CAN. J. BIOCHEM. CELL BIOL. VOL. 62, 1984

no direct evidence that the three types of metal-binding sites, cysteine, histidine and free carboxyl are part of the binding site for the porphobilinogen substrate. However, the study does show that ions of very different type can be equally toxic and, in this case, that the chemical toxicity of aluminum, uranium, and vanadium may be of as much importance as that of cadmium, mercury, and lead. 1. Casarett, L. J. & Dsull, J. (1975) Toxicology, the Basic Science ofPoisons, MacMillan, New York 2. Tephly, T. R. (1978) Handb. Exp. Pharmakol. 44, 81-91 3. Cotton, F. A. & Wilkinssn, G. (1972) Advanced Inorganic Chemistry, chap. 21, Wiley Interscience, New York 4. Arhland, S., Chatt, J. & Davies, N. 8 . (1958) Q. Rev. Chem. Soc. 12, 265 5. Williams, M. W. & Turner, J. E. (1981) J. Inorg. Nucl. Chem. 43, 1689-1691 6. Jones, M. M. & Vaughn, W. K. (1978) J. Inorg. Nucl. Chem. 48, 2081-2088 7. Gaertner, R. R. W. & Hollebone, B. R. (1983) Can. J. Biochem. Cell Biol. 61, 214-222

8. Wigfield, D. C. & Farant, J. -P. (1979) J. Anal. Toxicol. 3, 161-168 9. Sassa, S. (1978) Handb. Exp. Pharmakol. 44, 333-361 10. Russell, C. S. & Rockwell, P. (1980) FEBS Lett. 116, 199-202 11. Shioi, Y., Nagarnine, M., Kuroki, M. & Sasa, T. (1980) Biochim. Biophys. Acta 616, 300-309 12. Peters, P. G., Shauma, M. L., Hardwicke, D. M. & Piper, W. N. (1980) Arch. Biochem. Biophys. 281, 88-94 13. Eadie, G. S. (1952) Science (Washington, D . C . ) 116, 688 14. Hofstee, B. M. J. (1952) Science (Washington, D . C . ) 116,329-331 15. Hanes, C. S. (1932) Biochem. J. 26, 1406-1421 16. Woolf, B. (1931) Biochem. J. 25, 342-348 17. Cornish-Bswden, A. & Eisenthal, R. (1978) Biochem. Biophys. Acta. 523, 268-272 18. McCarthy, P. J. (1957) Pntroduction to Statistical Reasoning, McGraw-Hill, New York 19. Chow K, , Jiang, S., Liu, W. & Fee, C. (1979) Sci. Sin. (Engl. Ed.) 22, 341-358 20. Lehninger, A. L. (1975) Biochemistry, 2nd ed., Worth Publishers Inc., New York 21. Klopman (1968) J. Am. Chem. Sci. 90,223-234