@ Abstract 5610 Biocemistry 1995,34, 5610-5617 Galactose- 1 -pospate ridylyltransferase from scericia coli, a Zinc and Iron Metalloenzyme: Frank J. Ruzicka, Josep. Wedekind, Jeongmin Kim, Ivan Rayment, and Perry A. Frey* Institute for nzyme Researc, Te Graduate Scool, and Department of Biocemistry, College of Agricultural and Life Sciences, niversity of Wisconsin-Madison, Madison, Wisconsin 53705 Received October 14, 1994; Revised Manuscript Received January 31, 1995@ ABSTRACT: Galactose- 1 -P uridylyltransferase purified from scericia coli cells grown in enriced medium contains approximately 1.2 mol of tigtly bound zinc/mol of subunits as well as variable amounts of iron, up to 0.7 mol/mol of subunits, and no detectable Ca, Cd, Cu, Mo, Ni, Co, Mn, As, Pb, or Se. Te celators, 1, 1 0-penantroline, 8-ydroxyquinoline, 8-ydroxyquinoline sulfonate, and 2,2'-bipyridyl remove metal ions from te enzyme and allow te importance of zinc and iron to be evaluated. Dialysis of tis enzyme against 2 mm 1, 10-penantroline, 8-ydroxyquinoline sulfonate, and 2,2'-bipyridyl at millimolar concentrations slowly removes bot zinc and iron from te enzyme (t112 = 4 days at 24 "C) wit concomitant loss of enzymatic activity. In celation experiments utilizing 1,lO-penantroline, residual enzymatic activity was found to be proportional to te zinc content, to te iron content, and to te sum of zinc and iron. DP-glucose (0.35 mm) protects te enzyme against loss of metal ions and activity in te presence of 1,lO-penantroline, wereas glucose-1-p at 70 mm (400 x K,) fails to protect. Te enzyme purified from cells grown on a minimal medium containing inorganic salts and glucose supplemented wit eiter ZnS04 or FeS04 sows approximately te same level of enzymatic activity as te enzyme from cells grown on enriced medium. Tese experiments sowed tat enzymatic activity is supported by eiter iron or zinc associated wit two sites in te enzyme. nzyme depleted of metal ions by celators can be partially reactivated by addition of ZnS04. ridylyltransferase dialyzed against 5 M urea and DTA is devoid of metal ions and enzymatic activity but can be reconstituted as an active metalloenzyme containing Zn(II), Fe(II), Co(II), Cd(II), or Mn(I1) by furter dialysis against ZnS04, FeS04, CoC12, Cd(OAc)2, or MnC12. Te simplest interpretation of available information is tat uridylyltransferase contains two metal ion binding sites, at least one of wic must be occupied to support enzymatic activity. Te metal ions appear to be structural components rater tan catalytic components of te enzyme. Galactose- 1 -pospate uridylyltransferase (exose- 1 -P uridylyltransferase, C 2.7.7.12) catalyzes te interconversion of galactose- 1-P and uridine 5'-dipospate glucose (DPglucose)' wit glucose- 1-P and DP-galactose: galactose-1-p + DP-glucose * DP-galactose + glucose-1-p (1) Galactosemia in umans arises from any of several defects in tis enzyme. Te defects are inerited as autosomal recessive traits (Kalckar, 1960; Levy & Hammersen, 1978; Reicardt & Woo, 1991). Most mecanistic studies ave been carried out on te uridylyltransferase from scericia coli, a dimeric enzyme composed of identical subunits. Key aspects of te mecanism ave been confiied for te uman and yeast enzymes (Frey et al., 1982; Hester & Rausel, 1987). Te. coli enzyme catalyzes eq 1 troug a double-displacement mecanism and ping-pong kinetics according to eqs 2 and 3, in wic His166 is te nucleopilic catalyst at te active + Supported by Grant GM38480 from te National Institute of General Medical Sciences (P.A.F.) and Grant AR35186 from te National Institute of Artritis and Musculoskeletal and Skin Diseases (I.R.). publised in Advance ACS Abstracts, April 1, 1995. Abbreviations: DP, uridine S'dipospate; DP-glucose, uridine 5'-dipospate glucose; DP-galactose, uridine 5'dpospate galactose; DTA, etylenediaminetetraacetic acid; IPTG, isopropyl P-Dtiogalactopyranoside; BICIN, N,N-bis(2-ydroxyetyl)glycine; HPS, N-(2-ydroxyetyl)piperazine-N'-2-etanesulfonic acid. 0006-2960/95/0434-56 10$09.00/0 -His'66 + DP-glucose == -His'66-MP + glucose-1-p (2) -His'66-MP + galactose-1-p == -His'66 + DP-galactose (3) site (Wong & Frey, 1974a,b; Wong et al., 1977; Yang & Frey 1979; Field et al., 1989; Kim et al., 1990). xperiments by site-directed mutagenesis also led to te identification of Hisla as an essential residue, te function of wic remains unknown (Kim et al., 1990). Te amino acid sequence encompassing te two essential istidine residues is FN- KGAAMGCSNPHPHGQ; te sequence HPH is conserved in te uman, yeast, and Streptomyces enzymes (Reicardt & Berg, 1988). Tis segment of te protein contains Cys160 and G~'~~, bot of wic are conserved in te consensus sequence for four species. Inasmuc as te side cains of istidine, cysteine, and glutamate are often te ligands for Zn2+ in proteins (Vallee & Auld, 1990a), te presence of tese conserved residues raised te question of te possible presence of a metal ion suc as Zn2+. We sow in tis paper tat galactose-1-p uridylyltransferase from. coli is a metalloprotein tat contains two divalent metal ions per subunit, zinc and iron, wen te cells are grown on enriced medium. Te metal ion composition of uridylyltransferase can be varied by growt of cells on minimal media containing variable amounts of divalent metal ions or by celation and reconstitution. Tese experiments indicate tat te enzyme contains two divalent metal ion binding sites 0 1995 American Cemical Society
Metal Ion Content of Galactose- 1 -P ridylyltransferase per subunit and tat one or bot must be occupied to support enzymatic activity. MATRIALS AND MTHODS Materials. ],IO-Penantroline monoydrate was obtained from GFS Cemicals. 8-Hydroxyquinoline, 8-ydroxyquinoline sulfonate, and 2,2'-bipyridyl were obtained from Aldric Cemical Co. BICIN and HPS (ltrol grade) were obtained from Sigma and Calbiocem, respectively. Te following celators were furter purified by recrystallization (2x): 1,lO penantroline from 95% aqueous etanol and 8-ydroxyquinoline from etanol. Cell Culture.. coli BL21 cells transformed wit plasmid ptlc5800 (Field et al., 1989) were grown at 37 "C from seed stocks, eiter in 2 x YT medium (16 g of Bacto tryptone, 10 g of Bacto yeast extract, and 5 g of NaCl per liter of distilled water) or in minimal medium, M9 (3 g of KH2PO4, 6 g of Na2HP04, 0.5 g of NaC1, 1.0 g of NH4C1, 0.25 g of MgS04.7H20, 0.018 g of CaC12.2H20, and 4 g of glucose per liter of distilled water), bot containing 100 pg/ml ampicillin. Cells were allowed to grow to a density of approximately 1 OD at 600 nm after wic te inducer IPTG was added to a concentration of 1 mm. After an additional 3 of culture, cells were arvested by centrifugation at 12 OOOg, frozen in liquid nitrogen, and stored at -135 "C. Purification of Galactose-I -P ridylyltransferase. Galactose-l-P uridylyltransferase from. coli was purified by te metod of Arabsai et al. (1986) wit te following canges. Te enzyme was overexpressed in. coli BL21 cells transformed wit plasmid ptlc5800 (Field et al., 1989). All buffers were scrubbed of metal ions eiter by ditizone extraction or by treatment wit Celex 100 (Holmquist, 1988) and contained 10 mm P-mercaptoetanol. Desalting steps utilized eiter Sepadex G25 column cromatograpy or two dialyses of 2.5 duration. In te last cromatograpic step, Q-Separose Fast Flow was substituted for DA-Sepadex A-50 and a linear salt gradient was instituted consisting of 1 L eac of 0.05 M NaCl and 0.3 M NaCl in 0.01 M HPS buffer at ph 7.5. In addition, te purified enzyme was concentrated in an Amicon ltrafiltration stirred cell prior to drop-freezing in liquid nitrogen. Te frozen enzyme was stored at -135 "C. Protein omogeneity was establised by amino acid analysis wic agreed witin f 5% wit te expected analysis based on DNA sequence data of te galt gene (Lemaire & Muller- Hill, 1986; Comwell et al., 1987). Assay of Galactose-I -P ridylyltransferase. nzymatic activity was measured by use of te standard coupled assay described by Wong and Frey (1974b). Te specific activity of preparations used for tese studies ranged from 170 to 190 units/mg of protein. Metal Analysis. Preliminary metal analyses were conducted by inductively coupled plasma emission spectroscopy, te results of wic indicated te presence of bot zinc and iron, but no oter metals, in all samples of te enzyme. Subsequently, samples were analyzed for zinc and iron by grapite furnace atomic absorption spectrometry. Protein Analysis. Protein concentrations of analyzed samples were measured by amino acid analysis using te Pic0 Tag system of Waters Associates and were based on te expected amino acid composition of isoleucine, leucine, and penylalanine (I + L + F = 52) expressed as residues per subunit of enzyme, wic is known from te nucleotide Biocemistry, Vol. 34, No. 16, 1995 5611 sequence of te galt gene in. coli (Lemaire & Muller- Hill, 1986; Comwell et al., 1987). Te extinction coefficient at 280 nm for galactose- 1-P uridylyltransferase monomer was calculated from te A280 measurements on samples of four enzyme preparations, te molar concentrations of wic ad been establised by amino acid analysis. Te value of 280 was found to be (7.24 f 0.24) x lo4 M-' cm-'. Metal Celation. Metal celation studies were conducted by dialysis of te enzyme (40-50 pm subunits) at eiter 4 or 24 "C against 2000 volumes of buffer (0.05 M HPS at ph 7.5 or 8.0, 0.5 M NaC1, and 10 mm P-mercaptoetanol) containing various celators. Samples evaluated for metal content were subsequently dialyzed against metal-scrubbed buffer for 2 days at 4 "C (Auld, 1988a). nzyme treated wit salts of eiter Zn(I1) or Fe(I1) or bot were dialyzed against 1000 volumes of buffer containing 100 pm of eac metal salt for 2 days at 24 OC and ten dialyzed against buffer alone for an additional 2 days at 4 "C. Metal ion reconstitution was conducted inside a Coy anaerobic camber wic contained no detectable dioxygen (< 1 ppm). All dialyses were conducted by use of 1 cm diameter dialysis tubing (Spec trapor-2). RSLTS Metal Analyses. Initial analyses of four igly purified but not omogeneous preparations of galactose- 1 -P uridylyltransferase by inductively coupled plasma emission spectroscopy indicated te presence of approximately equivalent amounts of zinc and iron. Te average zinc content was 0.72 f 0.24 (SD) mol/mol of dimer, and te average iron content was 0.73 f 0.24 (SD) mol/mol of dimer. Te concentration of enzyme in tese samples was estimated from measurements of A280 assuming an extinction coefficient of 72 400 M-' cm-'. Tese initial preparations exibited a major band upon gel electroporesis, wic migrated at a rate corresponding to galactose- 1 -P uridylyltransferase, and a number of trace impurities. Tese initial analyses by inductively coupled plasma emission spectroscopy ruled out te presence of Cd, Cu, Mo, Ni, Co, Mn, As, Pb, and Se in galactose- 1 -P uridylyltransferase. Te consistent presence of bot iron and zinc in substantial and approximately equimolar amounts in te initial analyses suggested tat te uridylyltransferase may be a zinc-iron metalloprotein. Te metals were unlikely to be associated only wit trace impurities because of te amounts found and te fact tat tere were no major impurities. To obtain accurate values for te zinc and iron content of galactose- 1 -P uridylyltransferase, we purified six additional samples under carefully controlled conditions by malung use of buffers freed of metal ions. We verified te purities of tese preparations by quantitative amino acid analyses and obtained te results sown in Table 1. Te amino acid content corresponded closely wit te amino acid composition of tis protein, as deduced from te amino acid sequence derived by translation of te nucleotide sequence of te gene. Te analysis is also similar to tat reported by Saito et al. (1967). Te purity of four samples was estimated to be '95% based on teir amino acid content. Te amounts of zinc and iron associated wit te uridylyltransferase in six preparations tat ad been purified by use of buffers freed of divalent metal ions are given in Table 2. Te enzyme contains more tan 1 mol of zinc and less tan 1 mol of iron per mole of subunits, and te combined
Metal Ion Content of Galactose- 1 -P ridylyltransferase Biocemistry, Vol. 34, No. 16, 1995 5613 Table 3: ffect of Metal Content of Cell Culture Media on te Zinc and Iron Contents of Galactose-1-P ridvlvltransferase growt medium enzyme Zn (mol/mol)" enzyme Fe (mol/moly Zn i- Fe (movmol)" specific activityb ZnbM) medium analysis (1) enriced medium (2x YT) 1.19 zt 0.13' 0.59 f 0.04 1.78 168 f 2 22 21 (2) M9 + 20 ym ZnS04 1.09 f 0.07 0.23 f 0.01 1.32 174 f 6 21 3 (3) M9 + 100 ym ZnSO4 2.07 f 0.12 0.21 f 0.04 2.28 175 f 7 96 3 (4) M9 + 20 pm FeS04 0.85 f 0.04 1.47 f 0.05 2.32 149 f 3 0.8 22 (5) M9 + 200 pm FeS04 0.12 f 0.01 2.01 f 0.11 2.13 197 f 12 1 196 (6) M9 + 10 ym ZnS04 1.21 f 0.05 0.08 f 0.01 1.29 150 f 3 8 2 Metal content per mole of enzyme subunits. nits per milligram of protein (fsd in quadruplicate assays). MAN f SD. FebM) Q) c) PI % ; m c) 2 c) P Y 50-5 0 1 2 3 4 FIGR 2: Instability of iron-containing uridylyltransferase during dialysis. Samples of iron-uridylyltransferase (40 pm) from experiment 5 of Table 3, wic ad been purified at 4 "C, were dialyzed against 0.05 M HPS buffer at ph 7.5 containing 0.05 M NaCl and 10 mm P-mercaptoetanol at 24 "C inside te anaerobic camber in te presence and absence of 0.35 mm DP-glucose. Aliquots were removed at te indicated times, quenced by diluting into ice-cold buffer, and assayed for enzymatic activity. Symbols: (M) enzyme plus DP-glucose; (+) enzyme alone. In te latter experiment, te iron content decreased to 0.85 mol/mol in te course of dialysis, and te activity decreased to 30% of its initial value. c) I $ 1.00 I/) - 3 0.75 0 1 v 5 c) 0.50 6 0.25 3 200 150 100 * I 0 2 4 6 8 on weter te metal ions are zinc or iron, and te amounts of enzyme purified from te cells are comparable. However, te iron-containing enzyme in experiment 5 of Table 3 is less stable toward dialysis at ambient temperature tan te zinc-containing enzyme, as sown by te experiment of Figure 2. Bot enzymes can be purified at low temperature by te usual procedure and are equally active, but te enzyme containing two irons per subunit and very little zinc quickly loses its activity upon dialysis at 24 OC, wereas te zinccontaining enzyme retains its activity for days wen subjected to dialysis. Inactivation of te iron enzyme cannot ave resulted from air-oxidation of Fe(I1) because te dialysis was carried out under anaerobic conditions. Metal Ion Celation Studies. Celating agents can often remove divalent metal ions from proteins (Wagner, 1988; Auld, 1988b). Te celation experiments in Figures 3-5 demonstrate a dependence of enzymatic activity on te metal ion content of galactose- 1-P uridylyltransferase. 1,lO- Penantroline removes zinc (Figure 3A) and inactivates te enzyme (Figure 3B) in a first-order and celator concentration-dependent process in te course of 1 week. Te alftimes for te loss of enzyme activity in te presence of 1,lOpenantroline at 24 "C are 4.1 f 1 days at 5 mm and 15 f 1 days at 2 mm 1,lO-penantroline, respectively. Te alf-times for te loss of zinc are 4.0 f 1 days at 5 mm and 14 & 1 days at 2 mm 1,lO-penantroline, respectively. Te 50 0 2 4 6 8 FIGR 3: Concentration-dependent loss of Zn and galactose-1-p uridylyltransferase activity by 1,lO-penantroline. Galactose- 1 -P uridylyltransferase purified from cells grown on enriced medium contained bot iron and zinc. Te enzyme (40 pm) was dialyzed against 0, 0.5, 2, or 5 mm 1,lO-penantroline (500 volumes) in HPS buffer (0.05 M, ph 8.0) containing 0.05 M NaCl and 10 mm P-mercaptoetanol24 "C. At selected times, enzyme samples were removed and dialyzed against HPS buffer for 48 at 4 "C to remove te celator. Aliquots were assayed for enzymatic activity and for zinc by grapite furnace atomic absorption spectropotometry. Part A: Zn content versus time. Part B: nzyme activity versus time. Symbols: (0) no celator; (A) 0.5 mm 1,lOpenantroline; (+) 2 mm 1,lO-penantroline; (W) 5 mm 1,lOpenantroline. uridylyl-donor substrate DP-glucose prevents te loss of zinc and iron and enzymatic activity, as sown in Figure 4A,B, wic also sows tat te alf-times for te extraction of iron and zinc by 1,lO-penantroline in te absence of DP-glucose are te same at 33 "C. Glucose-1-P at a
5614 Biocemistry, Vol. 34, No. 16, 1995 Ruzicka et ai..-- I M w 5 2 0.25 s 150 + B f. Q c *z 200 e a % z 150 m z 250 b 100 Y a s v1-0.4 ; 5 1 w $ 0.2 s 0 2 4 6 0 2 4 6 0 150 100 FIGR 4: Protection by DP-glucose against te extraction of metal ions and inactivation of galactose- 1-P uridylyltransferase activity by 1,lO-penantroline. Galactose- 1 -P uridylyltransferase purified from cells grown on enriced medium contained bot iron and zinc. nzyme (40 pm) was treated wit 5 mm 1,lOpenantroline at ph 7.5 (see Figure 3 for procedure) in te presence and absence of 0.35 mm DP-glucose at 33 "C. Part A: (0) zinc content; (W) enzymatic activity; (+) zinc content in te presence of DP-glucose; (A), enzymatic activity in te presence of DP-glucose. Part B: (0) iron content; (M) enzymatic activity; (6) iron content in te presence of DP-glucose; (A) enzymatic activity in te presence of DP-glucose. concentration of 70 mm, over 400 times te K, and 3 times te substrate inibition constant (Wong & Frey, 1974b), does not affect te rate of zinc removal or loss of enzymatic activity (data not sown). In Figure 5, data on te zinc content of various uridylyltransferase samples from te experiments in Table 3 and Figure 3A,B are plotted versus teir enzymatic activities. Te enzymatic activity is directly proportional to te zinc content and increases to a maximum of about 200 unitdmg of protein, as reported previously for tis enzyme (Wong & Frey, 197413). Maximum activity is attained at about 1.1 mol of zinc/mol of enzyme subunits. In addition, incubation of te untreated enzyme at 40 pm wit ZnS04 at 100 pm for 2 days at 24 "C increases enzymatic activity by about 15% to te maximum of 200 unitdmg, and leads to a zinc content of 1.6 mol/mol of subunit and an iron content of 0.6 mol/mol of subunit after extended dialysis to remove excess ZnS04. Te correlation in Figure 5 documents te propor- 50 0 0.0 0.5 1.0 1.5 2.0 Zn Content (g-at/mol Subunit) FIGR 5: Relationsip between Zn2+ content and galactose-1-p uridylyltransferase activity. Galactose- 1 -P uridylyltransferase purified from cells grown on enriced medium contained bot iron and zinc. Samples of tis enzyme were treated in various ways and assayed for zinc content and enzymatic activity. Symbols: (W) enzyme was treated wit 1,lO-penantroline (5 mm) at 4, 24, or 33 "C for varying times by dialysis in HPS buffer and analyzed for zinc and enzymatic activity (see Figure 3 for procedure); (0) uridylyltransferase supplemented wit Zn(I1) by dialysis against HPS buffer containing 0.05 M NaC1, 10 mm P-mercaptwtanol, and 100 pm ZnS04 for 48 at 24 "C and ten against HPS buffer at 4 "C to remove excess ZnS04; (x) uridylyltransferase purified from cells grown in (2x YT) culture medium (experiment 1 in Table 3); (A) uridylyltransferase purified from cells grown in medium M9 supplemented wit 20 pm ZnSO4 (experiment 2, Table 3); (0) uridylyltransferase purified from cells grown in medium M9 supplemented wit 100 pm ZnSOd (experiment 3, Table 3); (0) uridylyltransferase purified from cells grown in medium M9 supplemented wit 20 pm Fe (experiment 4, Table 3). tionality of zinc content wit enzymatic activity and tat at least one zinc per subunit is required for activity. However, analogous plots of iron plus zinc are similarly linear (data not sown), and te maximum activity is attained at a stoiciometry of about two metal ions per subunit. Terefore, Figure 5 does not mean tat only one divalent metal ion per subunit is essential for activity because all of te samples preceding te plateau in Figure 5 contained iron as well as zinc. Te open square in te plateau corresponds to a sample tat contained two zinc ions per subunit and a decreased iron content. Te important point from Figure 5 is tat te enzymatic activity is dependent on te presence of zinc in one or bot sites. We will sow in a later section tat activity is supported by any of several divalent metal ions in bot sites. xcess, adventitiously bound zinc resulting from incubation of te enzyme wit ZnS04 neiter increases nor inibits enzymatic activity; owever, excess adventitiously bound iron resulting from incubation of te enzyme wit FeS04 inibits enzymatic activity (data not sown). Celators oter tan 1,lO-penantroline also remove divalent metal ions from te uridylyltransferase. Alternative celators are 2 mm 8-ydroxyquinoline sulfonate and 2 mm 2,2'-bipyridyl, wic extract zinc and iron commensurate wit loss of enzymatic activity similar to 1,lO-penantroline. In contrast, 8-ydroxyquinoline preferentially extracts iron. Dialysis of 40 pm enzyme against 5 mm 8-ydroxyquinoline in a buffer (50 mm HPS at ph 7.5, 50 mm NaCl, and 10 mm P-mercaptoetanol) for 2 days at 24 "C inside an anaerobic camber reduced te iron content of te enzyme
Metal Ion Content of Galactose- 1 -P ridylyltransferase Biocemistry, Vol. 34, No. 16, 1995 5615 Galactose-1-P uridylyltransferas Dialysis {Buffer 1 I Dialysis ;%k; urea 1 I 0.31 f 0.03 /mg Dialysis I( Buffer + 0.5 mm ZnSO,] 1 Dialysis Buffer + 2 mm ZnSOp ] 1.52 f 0.08 Znbunit FlodtdimglI-( I I -- FIGR 6: Flow cart for reconstitution of enzymatic activity in metal-depleted galactose-l-p uridylyltransferase by ZnS04. Tree samples of galactose-l-p uridylyltransferase (40 pm, 1 ml) were initially dialyzed against solutions containing buffer (0.5 L, 48 at 24 "C), buffer plus 100 pm ZnS04 (0.5 L, 48 at 24 "C and ten against buffer alone at 4 "C for 48 ), or buffer containing 250 pm DTA and 5 M urea (50 ml, 16 at 24 "C). Te buffer consisted of 0.05 M HPS buffer at ph 8.0 containing 0.05 M NaCl and 10 mm P-mercaptoetanol. Te samples were assayed for enzymatic activity and zinc content wit te results indicated. Te zinc-depleted sample was again dialyzed, first against 50 ml of buffer containing 5 M urea and 0.5, 1, or 2 mm ZnS04 for 24 at 24 "C, ten against 50 ml of buffer containing 0.5, 1, or 2 mm ZnS04 for 24 at 24 "C, and finally against 0.5 L (one cange) of buffer alone for 48 at 4 "C. from 0.51 to 0.03 mol/mol of subunit and te zinc from 1.16 to 0.99 moymol of subunit. Altoug 94% of te iron was removed, only 15% of te zinc was removed, and te activity decreased to 55% of its original value. It is clear from tis experiment tat te enzyme does not specifically require iron and significant activity is supported by te zinc remaining after treatment wit 8-ydroxyquinoline. Te decreased activity may represent partial occupancy of bot sites by zinc or tat 8-ydroxyquinoline reacts in some manner oter tan celation to inibit its activity. xperiments To Reconstitute Divalent Metal Ions in ridylyltransferase. Preliminary reconstitution experiments sowed tat simple dialysis of ZnS04 into solutions of uridylyltransferase tat ad been depleted of divalent metal ions by dialysis against a celator invariably increases te residual enzymatic activity by 30-50%, wereas dialysis of te same samples against FeS04 invariably fails to increase te activity and actually decreases it. In no case did dialysis against ZnS04 fully reconstitute activity; moreover, tis treatment led to te binding of excess zinc by te enzyme. Dialysis of metal-depleted enzyme against a mixture of ZnS04 and FeS04 restored activity to levels no greater tan tat attained by dialysis against ZnS04 alone. In order to test oter divalent metal ions for teir ability to substitute for zinc or iron, we undertook to devise an improved reconstitution procedure tat would remove essentially all of te metal ions and ten allow tem to be replaced wit substantial recovery of enzymatic activity. Te procedures described below resulted from tis effort. DTA is not an effective celator of te metal ions bound to galactose-l-p uridylyltransferase in te absence of a denaturant (data not sown). However, at increasing concentrations of urea in te presence of DTA, significant enzymatic activity is lost witin 4 at 24 "C. Te activity loss corresponds to decreased zinc content. In several experiments, te urea concentration was varied at 0.1 mm DTA (buffer: 50 mm HPS, ph 8.0, 10 mm P-mercaptoetanol, 24 "C). Te solutions were placed at 24 "C for 4, dialyzed against te buffer for 16 (fo.l mm ZnS04) to remove te urea, and assayed for zinc and activity. Te following zinc contents (moles per mol of subunit, mean f SD) were measured for eac urea concentration: no urea, 0.87 f 0.03; 1 M urea, 0.86 f 0.07; 2 M urea, 0.75 f 0.02; 3 M urea, 0.70 f 0.06; 5 M urea, 0.35 f 0.01. Te loss of enzymatic activity was substantial at 5 M urea ('95%) but te activity was significantly protected in all te experiments by te presence of 0.1 mm ZnS04 in te urea and dialysis buffers. ven at 5 M urea, approximately 50% of te activity of te untreated control was recovered in te enzyme treated wit bot urea and ZnS04. Te flow cart in Figure 6 summarizes te results of a study designed to find reconstitution conditions for uridylyltransferase tat ad been depleted of zinc by treatment wit DTA in te presence of 5 M urea. Dialysis of te uridylyltransferase against 1 mm DTA in 5 M urea removed essentially all of te zinc and reduced its activity to 0.3% of normal. Continued dialysis, first against 0.5 mm ZnS04 and 5 M urea and ten against 0.5 mm ZnSO4, restored te enzymatic activity to 73% of its original value. Reconstitution at iger concentrations of ZnS04 restored activity to a lesser extent. Te zinc reconstitution procedure of Figure 6 also allows oter divalent metal ions to be tested for teir ability to activate te uridylyltransferase. Table 4 sows tat te zincand iron-depleted enzyme is activated by CoC12, Cadmium acetate, FeS04, and MnClz to 40-50% of te activity of te Zn-reconsituted enzyme. DISCSSION In te most straigtforward interpretation of te present results, we can conclude tat galactose- l-p uridylytransferase from. coli is a metalloprotein containing two divalent metal ions per subunit. Altoug its metal ion composition can be influenced by te metal ion composition of te cellular growt medium, one site tends to be occupied preferentially by zinc and te oter by iron, altoug bot sites can be occupied by eiter metal ion. Tat te metal ions exibit
5616 Biocemistry, Vol. 34, No. 16, 1995 Ruzicka et al. Table 4: Reconstitution of Metal-Depleted Galactose-1-P ridylyltransferase wit Divalent Metal Ions form of enzyme specific activity (unitdmg of protein) zinc (movmo1)" Me(I1) (mol/mol)" iron (mol/mol)" native (Zn/Fe) ureadta treated Zn(I1) reconstitutedb Co(I1) reconstituted Cd(I1) reconstituted Fe(I1) reconstituted Mn(I1) reconstituted 168 f 2 0.028 f 0.004 95.2 f 4.8 65.7 f 2.3 51.5 f 3.5 49.2 f 2.2 44.2 f 2.1 1.19 f 0.13 0.011 f 0.002 1.64 f 0.08 0.012 f 0.003 0.035 f 0.006 0.070 f 0.02 0.075 f 0.013 1.64 f 0.08 1.44 f 0.05 1.11 f 0.02 1.68 f 0.07 1.32 f 0.13 0.59 f 0.04 0.026 f 0.001 0.011 f 0.001 0.052 f 0.012 0.020 f 0.004 1.68 f 0.07 0.012 f 0.004 Moles per mole of subunits. Galactose-1-P uridylyltransferase (10 ml, 40 pm) was dialyzed against 1 L of 5 M urea in HPS buffer at ph 7.5 containing 0.05 M NaC1, 0.25 mm DTA, and 10 mm j3-mercaptoetanol for 24 at 24 "C. Aliquots (1 ml) were dialyzed at 24 "C for 24 against 150 ml of buffer minus DTA plus te following salts at 500 pm: Zn(I1) sulfate, Cd(I1) acetate, FeS04, CoC12, and MnC12. Te solutions were dialyzed against 150 ml of te metal-buffer minus urea and ten against 1 L of 0.05 M HPS at ph 7.5 containing 0.05 M NaCl, 0.25 M DTA, and 10 mm B-mercaptoetanol at 4 "C for 24 (one buffer cange). All dialyses were carried out inside te anaerobic camber. preferences for binding to different sites is supported by te fact tat part of te iron is easily removed by dialysis in te absence of a celator from enzyme tat contains iron in bot sites (Figure 2) and by te difficulty of obtaining enzyme tat lacks zinc. To substantially deplete te enzyme of zinc, we must grow bacteria in a minimal medium supplemented wit 200 pm FeS04, wereas to obtain iron-depleted enzyme we need only 20 pm ZnS04 wit minimal medium (Table 3). Neiter metal ion is easily removed from its preferential binding site by ordinary dialysis, but bot can be removed by dialysis for a week against any of several celators or against DTA in te presence of 5 M urea. Te metal ion depleted enzyme is inactive, and te residual activity is proportional to te amount of metal tat remains, wic sows tat at least one of te divalent metal ions is essential for activity. Activity can be substantially reconstituted in te metal ion depleted enzyme by dialysis against divalent metal-ion salts, suc as ZnS04, FeS04, CoC12, MnC12, and cadmium acetate. Terefore, altoug enzymatic activity requires a divalent metal ion and Zn(I1) is preferred, te enzyme is not absolutely specific for Zn(I1) in its preferential site. At least one divalent metal ion is required for enzymatic activity, and bot metal ions may be required. Prolonged dialysis against 8-ydroxyquinoline removes nearly all of te iron from its binding site and very little zinc from its preferential site, and te enzyme retains 55% of its activity. Te residual activity may represent partial occupancy of bot sites by zinc. Altoug 8-ydroxyquinoline selectively removes iron from te enzyme, te stability constants for te Zn(I1) and Fe(I1) celates of 8-ydroxyquinoline are nearly identical (Auld, 1988b). Inasmuc as 1,lo-penantroline extracts te two metal ions at te same rate (Figure 4), it seems tat 8-ydroxyquinoline may extract iron selectively by a mecanism tat entails its interaction wit te enzyme. Te enzyme must bind zinc and iron very tigtly for tem to be present at nearly stoiciometric levels after several days of cromatograpic purification in metal-scrubbed buffers. Studies wit several celators support tis conclusion. Typical alf-times for removal of zinc and iron are approximately 4 days at 24 OC and 16 days at 4 "C (data not sown) in te presence of 5 m4 1,lO-penantroline, indicating tat te metal ions do not freely diffuse into te medium. Galactose- 1 -P uridylyltransferase differs in tis respect from several zinc-containing proteins, wic relinquis teir metal ions to tis celator wit alf-times of minutes to ours (Haeggstrom et al., 1990; Stocker et al., 1988). Te presence of DP-glucose, wic converts galactose- 1-P into its uridylyl-enzyme form (eq 2), prevents te removal of zinc by 1,lO-penantroline and maintains enzymatic activity. Terefore, te uridylyl-enzyme must bind zinc more tigtly tan te free enzyme. Tis may be because te MP group is a ligand for Zn(I1) in te uridylylenzyme or because te uridylyl-enzyme is more tigtly folded and does not relinquis Zn(I1) to a celator as readily as te free enzyme. Te fact tat bot iron and zinc are removed by 1,lO-penantroline at te same rate suggests tat te latter may be te correct explanation. Te simplest interpretation of equal rates is tat te rate of metal ion extraction is governed by some process oter tan celation per se. If te metal ions were released from te enzyme as a result of a reversible conformational cange, te celator would capture tem as free metal ions at te same rate. Tis is described by eqs 4 and 5, in wic te equilibrium constant slow Zn-Fe * + Zn + Fe (4) fast penantroline + Zn + Fe - penantroline-zn + penantroline-fe (5) for te conformational cange of into * in eq 4 is unfavorable but allows bot zinc and iron to escape from te enzyme at a very slow rate. In te absence of a celator oter tan te enzyme, te metal ions would be recaptured by reversal of eq 4. 1,lO-Penantroline can capture te free metal ions according to eq 5. DP-glucose would inibit tis process by transforming te enzyme into te uridylyl-enzyme according to eq 2, wic presumably does not undergo te conformational cange of eq 4. Tis mecanism of celation does not exclude te DP-glucose or te MP-moiety of te uridylyl-enzyme as metal ion ligands. However, altoug we know tat enzymatic ligands bind te metal ions, we ave no evidence for coordination of te substrate wit metal ions. Te present results do not provide information regarding te locations of metal ions in te protein. Amino acid residues in te vicinity of te active site offer good candidates for zinc ligands, including te side cains of Hisla and His166, CysI6O and G~'~~. In oter Zn(I1) proteins, tree or four ligands are amino acid side cains. Wen Zn(I1) is bound by four protein ligands, tey are generally eiter all cysteine side cains or two or tree cysteines and one or two istidine side cains. In tese cases, Zn(I1) is generally tougt to stabilize te active protein structure as distinguised from participating directly in a catalytic or oter functional role. Wen Zn(I1) as tree protein ligands, tey are generally
Metal Ion Content of Galactose- I-P ridylyltransferase selected from te side cains of istidine, cysteine, and glutamate, wit water as te fourt ligand (Vallee & Auld, 1990b). In tese cases, te protein is generally an enzyme, and Zn(I1) is tougt to participate in catalysis, altoug it is no doubt also important in maintaining structure in te active site. Several enzymes ave Zn(I1) ligands donated by two istidines tat are separated by one or tree oter amino acids in te sequence (Vallee & Auld, 1990a). For example, in carbonic anydrase, two istidines in te active site are separated by penylalanine, and in termolysins, tey are separated by glutamate and two oter amino acids. Galactose- 1-P uridylyltransferase from. coli contains two sequences of HPH, te active site-residues H1&1&166 and te nonessential istidines in te sequence H9PloH11. Te latter sequence is flanked by an aspartate at position eigt. iter or bot sequences may be considered as possible sources of side cain ligands for metal ions. Moreover, te protein contains six cysteine residues, including Cys5* and CyP, in addition to CYS'~O, tat could participate in binding metal ions. Wit tese residues, and a cluster of acidic residues in te sequence ~l,*,ardrlqkyfa~19*~, tere are many possibilities for binding metal ions to tis protein. Te finding of Zn in uridylyltransferase places it among te Zn-containing class I1 nucleotidyl transferase enzymes, wic include DNA polymerase (Slater et al., 1971; Springgate et al., 1973), RNA-dependent DNA polymerase (Coleman, 1974; Auld et al., 1974), and RNA polymerase (Scrutton et al., 1971). Altoug stoiciometric amounts of Zn ave been found, it is not clear wat role it plays in tese enzymes. In fact, in te case of DNA polymerase I from. coli and T7 RNA polymerase, enzymes fully active for DNA polymerization or transcription but devoid of Zn ave been purified (Walton et al., 1982; King et al., 1986). DNA polymerase from. coli is a multifunctional enzyme tat catalyzes not only DNA polymerization but also 3'- and 5'- exonuclease activity on distinct substrate binding and catalytic sites. Te 3'- exonuclease activity appears to be dependent on divalent cations suc as Zn(I1) (Beese & Steitz, 1991). A structural role as been suggested for Zn(I1) in aspartate carbamoyltransferase (Honzatko et al., 1982), anoter class I1 enzyme. Te present studies do not establis a specific role for metal ions in galactose- 1 -P uridylyltransferase, apart from being required for enzymatic activity. Te fact tat Co(II), Cd(II), and Mn(I1) support activity 40-50% as well as Zn- (11) or Fe(I1) indicates tat te metal ions may be more important in a structural role tan as direct participants in catalysis. Te X-ray crystallograpic studies of galactose- I-P uridylyltransferase in our laboratory are at an advanced stage and will resolve tis isssue (Wedekind et al., 1994). ACKNOWLDGMNT We are pleased to acknowledge wit tanks te kind assistance given to us by Dr. Bert Vallee, Dr. James Riordan, and Dr. Robert Sapiro during a visit to teir laboratory in te critical early stage of tis researc. We appreciate te access to te use of teir facilities in generous measure. Te tecniques for accurately measuring te metal ion content of uridylyltransferase were acquired under teir guidance. We also tank Dr. Vallee for is valuable comments on te manuscript. RFRNCS Biocemistry, Vol. 34, No. 16, 1995 5617 Arabsai, A., Brody, R. S., Smallwood, A., Tsai, T.-C., & Frey, P. A. (1986) Biocemistry 25, 5583-5589. Auld, D. S. (1988a) Metods nzymol. 158, 13-14. Auld, D. S. (1988b) Metods nzymol. 158, 110-114. Auld, D. S., Kawaguci, H., Livingston, D. M., & Vallee, B. L. (1974) Proc. Natl. Acad. Sci..S.A. 71, 2091-2095. Beese, L. S., & Steitz, T. A. (1991) MBO J. 10, 25-33. Coleman, J.. (1974) Biocem. Biopys. Res. Commun. 60, 641-648. Comwell, T. L., Adya, S. L., Reznikoff, W. S., & Frey, P. A. (1987) Nucleic Acids Res. 15, 8116. Field, T. L., Reznikoff, W. S., & Frey, P. A. (1989) Biocemistry 28, 2094-2099. Frey, P. A., Wong, L.-J., Seu, K.-F., & Yang, S.-L. (1982) Metods nzymol. 87, 20-36. Haeggstrom, J. Z., Wetterolm, A., Sapiro, R., Vallee, B. L., & Samuelsson, B. (1990) Biocem. Biopys. Res. Commun. 172, 965-970. Hester, L. S., & Rausel, F. M. (1987) J. Biol. Cem. 262, 12092-12095. Holmquist, B. (1988) Metods nzymol. 158, 6-12. Honzatko, R. B., Crawford, J. L., Monaco, H. L., Ladner, J.., dwards, B. F. P., vans, D. R., Warren, S. G., Wiley, D. C., Ladner, R. C., & Lipscomb, W. N. (1982) J. Mol. Biol. 160, 219-263. Kalckar, H. M. (1960) Fed. Proc., Fed. Am. SOC. xp. Biol. 19, 984-990. Kim, J., Ruzicka, F., & Frey, P. A. (1990) Biocemistry 29, 10590-10593. King, G. C., Martin, C. T., Pam, T. T., & Coleman, J.. (1986) Biocemistry 25, 36-40. Layne,. (1957) Metods nzymol. 3, 4477454. Lemaire, H. G., & Muller-Hill, B. (1986) Nucleic Acids Res. 14, 7705-777 1. Levy, H. L., & Hammersen, G. (1978) J. Pediatr. 92, 871-877. Reicardt, J. K., & Woo, S. L. C. (1991) Proc. Natl. Acad. Sci. SA 88, 2633-2637. Reicardt, J. K. V., & Berg, P. (1988) Nucleic Acids Res. 16, 9017-9026. Saito, S., Ozutsumi, M., & Kuraasi, K. (1967) J. Biol. Cem. 242, 2362-2368. Scrutton, M. C., Wu, C. W., & Goldtwait, D. A. (1971) Proc. Natl. Acad. Sci..S.A. 68, 2497-2501. Slater, J. P., Mildvan, A. S., & Loeb, L. A. (1971) Biocem. Biopys. Res. Commun. 44, 37-43. Springgate, C. F., Mildvan, A. S., Abramson, R., ngle, J. L., & Loeb, L. A. (1973) J. Biol. Cem. 248, 5987-5993. Stocker, W., Wolz, R. L., Zwilling, R., Strydom, D. J., & Auld, D. S. (1988) Biocemistry 27, 5026-5032. Vallee, B. L., & Auld, D. S. (1990a) Proc. Natl. Acad. Sci..S.A. 87, 220-224. Vallee, B. L., & Auld, D. S. (1990b) Biocemistry 29, 5647-5659. Wagner, F. W. (1988) Metods nzymol. 158, 21-32. Walton, K.., FitzGerald, P. C., Hemnann, M. S., & Benke, W. D. (1982) Biocem. Biopys. Res. Commun. 108, 1353-1361. Wedekind, J.., Frey, P. A., & Rayment. I. (1994) Acta Crystallogr., Sect. D D50, 329-331. Wong, L.-J., & Frey, P. A. (1974a) J. Biol. Cem. 249, 2322-2324. Wong, L.-J., & Frey, P. A. (1974b) Biocemistry 13, 3889-3894. Wong, L.-J., Seu, K.-F., Lee, S.-L., & Frey, P. A. (1977) Biocemistry 16, 1010-1016. Yang, S.-L. L., & Frey, P. A. (1979) Biocemistry 18, 2980-2984. BI942392Q