Properties of a purified nuclear topoisomerase from L1210 cells

230

Biochimica et Biophysica Acta, 741 (1983) 230-236 Elsevier

BBA 91282

PROPERTIES OF A PURIFIED NUCLEAR TOPOISOMERASE FROM LI210 CELLS CAROL F. ROSS a, MICHAEL J. BROUGHAM c, WILLIAM K. HOLLOMAN c and WARREN E. ROSS a.b.,

Departments of ~ Pharmacology, b Medicine, and " Medical Microbiology and Immunology, Universi(v of Florida, J. 11. Miller Health Center, Gainseville, FL 32610 (U.S.A.) (Received March 7th, 1983) (Revised manuscript received July 11th, 1983)

Key words: Topoisomerase; DNA strand scission; (Mouse leukemia cell)

A nuclear type I topoisomerase from mouse leukemia LI210 cells has been partially purified and characterized. The sedimentation coefficient of the enzyme by velocity sedimentation is 4.3 S, consistent with a globular protein of 68 kDa. Enzyme activity is stimulated 20-fold in the presence of magnesium over that achieved in KCI alone. The enzyme is completely inhibited in the presence of the berenil congeners HOE 13548 and 15030 while berenil itself caused only partial inhibition at concentrations below 200 p g / m l . An acid soluble protein of 30 kDa (by SDS-polyacrylamide gel electrophoresis) co-purified with the topoisomerase but could be separated by precipitation in a low salt buffer. This protein, as well as a protein of similar characteristics, histone HI, stimulated topoisomerase activity over a narrow concentration range. The role of topoisomerase in the DNA strand scission observed in LI210 cells following exposure to intercalating agents remains conjectural as the purified enzyme did not produce nicks in plasmid DNA in the presence of adriamycin.

Introduction Topoisomerases are a group of enzymes which alter the topological conformation of DNA by nicking and resealing the sugar-phosphate backbone of DNA, thereby allowing changes in linking number [1]. Generally two types of topoisomerases have been found in mammalian cells. Type I topoisomerases act by creating transient single stand breaks in the DNA. They require no additional energy source, having conserved the energy of the phosphodiester bond by forming a covalently joined enzyme-DNA intermediate during the reaction. The resealing action is accompanied by displacement of the enzyme from the nick site. Type II topoisomerases are distinguished by the * To whom correspondence should be addressed. Abbreviation: PMSF, phenylmethylsulfonyl fluoride. 0167-4781/83/$03.00 © 1983 Elsevier Science Publishers B.V.

fact that both strands of the double helix are broken. Our interest in topoisomerase stems from a series of observations which suggested the possibility that such an enzyme may be responsible for the DNA strand scission observed when mammalian cells are exposed to intercalating agents such as adriamycin and ellipticine [2,3]. The intercalative mode of binding is particularly noteworthy in that it causes topologically constrained DNA, such as that found in the nucleus, to become progressively more unwound as the drug : basepair ratio increases. We have previously shown that exposing mouse leukemia L1210 cells to intercalating agents results in DNA strand breaks which are spatially associated with a protein which is tightly, if not covalently, bound to the DNA at or near the break site. The frequencies of strand breaks and DNA-protein crosslinks have

231 been quantitated and appear to be identical [3]. Upon drug removal both strand breaks and DNA-protein crosslinks are removed simultaneously [4]. Drugs that bind to DNA by intercalation are the only compounds found to date which cause the protein-associated DNA strand breaks. In order to further our studies regarding the possible role of topoisomerases in intercalator-induced DNA strand scission, we have isolated and partially purified a nuclear type I topoisomerase from L1210 cells. In this paper we describe some of the characteristics of this enzyme as well as our initial efforts to demonstrate the drug-induced DNA lesions using a reconstituted system. A serendipitous product of these investigations was the discovery of a protein which co-purifies with topoisomerase and is able to stirrtulate or inhibit topoisomerase activity depending on its concentration relative to the enzyme. Interestingly, this protein exhibits important similarities to and may be identical with histone H1. Materials and Methods Materials

Plasmid pBR322 DNA was prepared as previously described [5]. Poly(ethyleneglycol) (PEG 6000) was purchased from Baker. Hydroxyapatite (Gio-Gel HTP) and agarose were from Bio-Rad. Phosphocellulose P11 came from Whatman. Phenylmethylsulfonyl fluoride (PMSF) was from Eastman Kodak. Histone H1 was purchased from Sigma. The following were kind gifts: berenil (Squibb Pharmaceutical); HOE 13548 and HOE 15030 (Hoechst-Roussel); amicarbilide and imidocarb (Burroughs Wellcorrie Company); and methylglyoxal-bis(guanylhydrazone) (Drug Synthesis and Chemistry Branch, NCI, NIH).

various amounts of enzyme. The reaction mix was incubated at 37°C for 20 min following addition of enzyme. The reaction was stopped by adding 5 /~1 of a solution containing 2% SDS/0.05% Bromophenol blue/50% glycerol. When the reaction mixture included adriamycin, proteinase K (0.5 m g / m l ) was included in the S D S / Bromophenol blue/glycerol solution and the drug was extracted with phenol prior to electrophoresis. The reaction mix was loaded onto a 1.2% agarose gel made with a buffer of 40 mM Tris-HCl/1 mM E D T A / 5 mM sodium acetate, pH 7.5. Electrophoresis was performed at 3 V / c m for 15-20 h. The gel was stained with ethidium bromide (5 # g / m l ) and photographed under ultraviolet light. SDS-polyacrylamide gel electrophoresis was performed essentially according to Laemmli [6]. Molecular weight determination by velocity sedimentation was accomplished using 20-40% glycerol gradients, centrifuged at 40000 rev./min for 24 h in an SW 50.1 rotor. Each gradient sample contained 1 rag bovine serum albumin as a reference. Protein concentration was determined using the assay described by Bradford [7]. The acid solubility of certain proteins was determined by adding perchloric acid (final concentration 5%, w/v) to the solution containing protein. After standing on ice for 15 min the mix was centrifuged in a microcentrifuge for 5 min~ The supernatant was transferred to another tube to which was added trichloroacetic acid (final concentration 35%). The pellet resulting from this step was washed once with acid acetone (0.1% conc. HC1 in acetone) and then twice with ice-cold acetone. The pellet was then redissolved into 0.5 M KC1. Aliquots of the original preparation as well as one of each pellet and supernatant from the above procedures were then compared by SDSpolyacrylamide gel electrophoresis.

Methods

Mouse L1210 cells were maintained in log phase in RPMI 1630 medium containing 10% fetal calf serum, penicillin and streptomycin. Topoisomerase activity was assayed by following the conversion of form I pBR322 DNA to form IV and other topoisomers. Unless otherwise specified, the typical reaction mixture (50 #1) contained 10 mM Tris-HC1 (pH 7.5)/1 mM E D T A / 100 mM KC1/20 mM MgC1/40 #g/ml DNA and

Results

Isolated nuclei from mouse leukemia L1210 cells were prepared by a modification of the procedure described by Tang [8]. Approximately 5. l0 s cells were incubated in ice-cold 1 mM CaCI 2 for 10 min, then Dounce-homogenized in 30-ml portions with 8-10 strokes of a loose fitting pestle. The

232 presence of nuclei was confirmed by phase microscopy. To each 30 ml portion of nuclear suspension, 10 ml of a buffer containing 1 M sucrose/10 mM T r i s / 1 mM E D T A / 1 mM PMSF was added and the samples were then centrifuged at 2500 r e v . / m i n for 5 min at 4°C. The nuclear pellets were resuspended into 10 ml of 20 mM potassium p h o s p h a t e / 5 0 mM KC1/1 mM dithiothreitol/1 m M PMSF, pooled, and made up to 1 M KCI by the addition of the solid salt. The nuclei were crushed using a French pressure cell at 22000 l b / i n c h 2 and the homogenate was centrifuged for 90 min at 15000 rev./min. The supernatant was removed, stirred in a refrigerator with 5% ( w / v ) poly(ethyleneglycol) for 2 h and then centrifuged for 1 h at 10000 rev./min. This supernatant was then frozen in 10% glycerol. Further purification was achieved by hydroxyapatite column chromatography of the poly(ethyleneglycol) supernatant. Fractions containing topoisomerase activity were pooled and diluted 1 : 3 with a buffer containing 20 mM potassium phosphate/1 mM 2-mercaptoethanol/5% glycerol/1 mM PMSF and loaded onto a phosphocellulose column equilibrated with 0.2 M K C I / 20 m M K H 2 P O 4 / 5 % g l y c e r o l / 1 mM 2mercaptoethanol/1 mM PMSF and eluted with three 2 ml fractions each of buffer containing 0.2, 0.4, 0.6, 0.8 and 1.0 M KC1. Topoisomerase activity eluted in two partially overlapping peaks at KC1 concentrations of 0.8 and 1.0 M. Samples of fractions containing topoisomerase from the phosphocellulose column were examined by SDS-polyacrylamide gel electrophoresis. The material eluting in the 0.8 M KC1 fractions exhibited no distinct bands in the gel when stained by Coomassie blue. However, the 1.0 M fractions demonstrated a distinct single band with a molecular weight of 30000. The 0.8 M fractions were pooled, as were the 1.0 M fractions, and both were dialyzed against a buffer containing 50 mM KC1/ 20 m M K P O 4 / 5 % g l y c e r o l / 1 m M 2merceptoethanol/1 mM PMSF. While the 0.8 M material remained clear throughout the 1 h dialysis, the 1.0 M fraction developed a cloudy precipitate which was removed by a 30 min centrifugation at 15 000 rev./min. The resulting pellet was dissolved in 0.5 M K C I / 2 0 mM KH2PO4/5% glycerol/1 mM 2-mercaptoethanol/1 mM PMSF

and was the subject of further investigations described below. In all other studies of topoisomerase activity the dialysed 0.8 M fraction was employed. Optimal salt concentrations for topoisomerase activity were determined. In the absence of Mg 2÷, enzyme activity was greatest at 200 mM KC1. However, enzyme activity was increased approx. 5-10 fold in the presence of 10-20 mM MgCl 2 and 100 mM KC1 (Fig. 1). Because the protein concentration of our preparation was too low for accurate determination, we were unable to define precisely the specific activity of our enzyme. However, assuming the protein concentration was less than 5.0/~g/ml, based on the absence of a band on the SDS-polyacrylamide electrophoresis gel, we estimate the specific activity to be at least 2- 105 u n i t s / m g protein. The sedimentation coefficient of the topoisomerase activity in a glycerol gradient was 4.3 S, consistent with a globular protein with a molecular weight of 68 000. Inhibition of nuclear type I topoisomerase activity by the trypanocidal diamidine berenil has been observed by some [9], but not all [5] investigators. We examined berenil and a number of other diamidines for their ability to inhibit type I topoisomerase activity from L1210 nuclei. Berenil (1-200/~g/ml) gave variable results. Most enzyme preparations were at least partially inhibited at the lowest berenil concentrations but complete inhibition required anywhere from 10-200 /~g/ml de-

Form ~rm

Fig. 1. Stimulation of topoisomerase activity by magnesium. Reaction mixtures (50 p-l) contained 10 mM Tris-HC1/6 mM KHzPO4/1 mM EDTA/2 p-g form I pBR322 DNA. Lanes b and c also contained 20 mM MgC12 and 100 mM KCl while lanes d-h contained 200 mM KC1 only. A topoisomerase preparation from pooled 0.8 M KC1 fractions of the phosphocelhilose column was added in the following volumes: a, none; b, 0.3 p-l; C, 0.7 p-l; d, 3 ~1; e, 7 p-l; f, 10 p-l; g, 12.5 p.l; h, 15 p-l.

233

N

N

L~I>-N /

N~ C~ \ N

H -15050

N

N

N/~N CH3

~\N CH3 H-13548

N

N

/N v

W

~ N ~ BERENIL

Fig. 2. Structures of berertil and its congeners HOE 13548 and HOE 15030.

Fig. 3. Inhibition of topoisomerase activity by HOE 13548, Each reaction mixture contained 10 mM Tris-HCl/6 mM KH2PO4/1 mM EDTA/20 mM MgC12/2 #g form I pBR322 DNA and enzyme. Lanes a, b - f and g-k contained 0, 1 and 15 #1 of the topoisomerase preparation, respectively. Concentrations of HOE 13548 were as follows: a, b, g, none; c, h, 100 # g / m l ; d, i, 50 # g / m l ; e, j, 30 #g/ml; and f, k, 10 #g/ml. Reactions were incubated at 37°C for 20 rain and then were terminated.

pending on the enzyme preparation. The basis for this variability could not be ascertained but did not appear to be related to drug purity or instability or to activity of the enzyme preparation. Interestingly, a congener of berenil, HOE 13548 (Fig. 2) consistently inhibited topoisomerase activity (Fig. 3). Complete inhibition was observed at 30 #g/ml. Increasing the enzyme concentration 15-fold appeared to partially reverse the inhibition seen at 50 # g / m l HOE 13548 but not at 100 #g/rnl. Another berenil congener, HOE 15030 gave similar results. Other diamide compounds, including pentamidine, methylglyoxal-bis(guanylhydrazone), amicarbilide and imidocarb were without effect. The protein which precipitated during dialysis of the 1.0 M KC1 fractions of the phosphocellulose column was studied further. This protein entered solution readily in a buffer containing 0.5 M KC1/ 20 mM KH2PO4/1 mM 2-mercaptoethanol/5~g glycerol/1 mM PMSF and by SDS-polyacrylamide gel electrophoresis appeared to be a single protein with a molecular weight of 30000. It is henceforth referred to as the 30 kDa protein. When this 30 kDa protein was incubated with pBR322 DNA under standard reaction conditions topoisomerase activity was observed. We hypothesized that the 30 kDa protein was not itself a topoisomerase but rather a protein which was associated with and perhaps modulates the activity of the actual topoisomerase. To test this hypothesis, the 30 kDa protein was heated to 60°C for 15 rain. The topoisomerase activity of the preparation was destroyed by this step but when the heat-inactivated 30 kDa protein was added to a reaction mix containing topoisomerase from the 0.8 M KCI fraction, a marked stimulation of activity was observed (Fig. 4). Interestingly, this stimulation occurred only when the incubation mixture contained 200 mM KC1 and no magnesium. Under the optimal conditions for topoisomerase activity, i.e. 100 mM KC1 and 20 mM MgCI2, the 30 kDa protein was without effect on topoisomerase activity. Rowe et al. [10] have previously demonstrated that H1 histone stimulated the activity of topoisomerase obtained from the fungus Ustilagomaydis [10]. Since our 30 kDa protein migrates during SDS-polyacrylamide gel electrophoresis to a position similar to that of H1, we suspected that the 30

234

Fig. 5. Effect of histone H1 on topoisomerase activity. Each reaction mixture contained 10 mM Tris-HCl/1 mM EDTA/6 mM KH 2 PO4/2 #g form I pBR322 D N A / 1 ~1 of the topoisomerase preparation (except lane a which contained no topoisomerase). Lane b contained 20 mM MgCI 2 and 100 mM KCI while lanes c-i contained 200 mM KCI only. Histone H1 concentrations were 0.1, 0.3, 0.6, 1.0, 3.0 and 20 # g / m l in lanes d-i, respectively.

Fig. 4. Stimulation of topoisomerase activity by the 30 kDa protein from L1210 nuclei. Each reaction mixture contained 10 mM Tris/1 mM E D T A / 6 mM KH2PO4/2/~g form I pBR322 D N A / 1 #1 of the topoisomerase preparation (except lane a which contained no topoisomerase). Lane b contained 20 mM MgCI 2 and 100 mM KCI while lanes c-g contained 200 mM KCI and no magnesium. In addition to the above lanes d - g contained 33, 66, 330 and 660 #g/ml, respectively, of the 30 kDa protein isolated from the 1.0 M KCI fraction of the phosphocellulose column.

kDa protein was indeed H1. When L1210 topoisomerase was incubated with pBR322 in the presence of histone H1 obtained from the Sigma Co., stimulation of topoisomerase activity was observed at H1 concentrations of 0.3-3.0 # g / m l ; no stimulation was seen at 10 # g / m l or higher (Fig. 5). To assess further the similarity between our 30 kDa protein and H1 we tested each for their solubility in 5% perchloric acid as described in Methods. The 30 kDa protein was first dialysed to replace the KC1 with NaC1. Both proteins were soluble under these conditions (data not shown). Our interest in the nuclear topoisomerase of

L1210 cells arose from previous work in this laboratory suggesting that this enzyme may mediate the formation of protein-associated DNA strand breaks which occur when L1210 cells are exposed to drugs such as adriamycin which bind to DNA by intercalation between base pairs. To test this hypothesis plasmid DNA was incubated with the purified type I topoisomerase in the presence of various concentrations of adriamycin (Fig. 6). No evidence of DNA strand scission was observed. In other experiments the electrophoresis

Fig. 6. Effect of adriamycin on topoisomerase activity. Standard reaction mixtures containing 1 #1 of the topoisomerase preparation (except lane a which had none) were incubated for 20 min at 37°C in the presence of the following concentrations of adriamycin: lanes a, b, none; lanes c-j 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4, and 0.5 # g / m l respectively. The reaction was terminated by addition of 5 #1 of 2% SDS/0.05% Bromophenol blue/50% glycerol/0.5 # g / m l Proteinase K. The DNA was extracted with phenol and loaded onto an agarose gel.

235

was performed with ethidium bromide, 0.05 # g / ml, in the agarose and electrophoresis buffer in order to clearly distinguish form II (nicked circular) and form IV (relaxed covalently closed circular) DNA. Other reaction conditions were also studied. The enzyme concentration was increased up to 25 fold relative to DNA and drug and the adriamycin concentration was increased up to 100 # g / m l . None of these modifications result in the demonstration of nicked DNA. Discussion

In order to directly examine the possibility that topoisomerase may mediate the DNA strand scissions observed in L1210 cells follo~wingexposure to intercalating agents, it was necessary to purify and characterize the enzyme from these cells. The nuclear topoisomerase from L1210 mouse lymphoblasts share certain characteristics with other mammalian Type I topoisomerases. The molecular weight of 68000, estimated from its sedimentation coefficient in a glycerol gradient, is consistent with that found in human KB cells (M r 60000, [7]), rat liver (M r 66000, [3]) and mouse L-cells (M r 68000, [17]). Liu and Miller [11] found that while the majority of topoisomerase activity was associated with a monomeric peptide of 100 kDa, a second protein with enzyme activity had a molecular weight of 67000. They hypothesized that the latter was a proteolytic product of the larger protein. Like the preparation of Liu and Miller [11], our enzyme was made from fresh, unfrozen cells and the serine protease inhibitor PMSF was included in each step of the process. Interestingly, the enzyme from HeLa cells exhibited a sedimentation coefficient in a sucrose gradient of 4.3 S, identical to that of the L1210 topoisomerase. Their estimation of the M r 100000 was based on SDS-polyacrylamide gel electrophoresis. Since there was insufficient material to visualize our enzyme by this method we cannot state with certainty that it is not the same size as that found in HeLa cells. Magnesium markedly increases L1210 topoisomerase activity (see Fig. 1) over that achieved in the presence of KC1 only. This is contrary to the previous observations of several other investigators studying eukaryotic type I topoisomerases

[8,12,13]. As discussed by Liu [11], the inclusion of the protease inhibitor PMSF in all solutions and the use of fresh, not frozen, cells may result in an enzyme less subject to proteolytic degradation. This may have important consequences with respect to in vitro activity. During the isolation of the L1210 topoisomerase we identified a protein which had copurified with the topoisomerase and which was able to stimulate the activity of the enzyme. The mechanism by which the 30 kDa protein exerts its effect may be similar to that of magnesium since its activity is not additive to that of the divalent cation. In addition, this protein shares certain features with the polycationic histone H1 in that it is soluble in acid and it appears to have a molecular weight of 30000 by SDS-polyacrylamide gel electrophoresis. Indeed, we have also found that, at the appropriate concentration, calf thymus histone H1 can stimulate L1210 topoisomerase, an observation previously described only for the topoisomerase of the fungus Ustilago maydis [10]. As in the latter study, the effect of H1 on enzyme activity was biphasic; stimulatory at low concentrations but not as the concentration is increased. Interestingly, the concentration of 30 kDa protein required for topoisomerase stimulation (33 # g / m l ) is considerably greater than that required for H1 (0.3 #g/ml). The meaning of this discrepancy, however, is unclear since the two proteins were isolated from vastly different sources and purified by different methods. Heterogeneity of H1 histone structure [14] and affinity for DNA [15] has been previously demonstrated even for single cell types, and this heterogeneity appears to correlate with, among other things, terminally differentiated phenotype [16]. We believe that this heterogeneity, as well as the effects of post-translational modifications of the protein, could account for the differences in potency we have observed. In any case, the L1210 30 kDa protein is worthy of further investigation as it may play a role in modulating activity of the topoisomerase. The effect of the diamidine berenil on type I topoisomerase is somewhat controversial. Fairfield et al. [17] found that berenil did not inhibit the nuclear enzyme from rat liver although a mitochondrial topoisomerase was inhibited at concentrations as low as 4 #M (2 #g/ml). On the

236 other hand, B r u n e t al. [9] found that both nuclear and mitochondrial topoisomerase from Xenopus laeois oocytes was partially inhibited at a berenil concentration of 0.25/~g/ml. The effect of berenil on L1210 topoisomerase was inhibitory but the degree of potency was inconsistent. Interestingly, examination of a number of other diamidines revealed two that were more active in this regard than berenil. H O E 13548 and H O E 15030 are berenil congeners which have previously been shown to inhibit mitochondrial D N A synthesis and induce petite mutants in yeast [18]. The fact that these two compounds are potent inhibitors of topoisomerase has at least two major consequences. First, the drugs may prove to be useful as probes of topoisomerase function in the cell. Second, the data suggest that further studies of structure-activity relationships among the diamidines may turn up even more potent inhibitors. It is of interest that the inhibition of topoisomerase activity by the H O E compounds is partially reversed by increasing the enzyme concentrations. This is similar to the findings of B r u n e t al. [9] using berenil and suggests the possibilities that the drug may act either by competing with the enzyme for D N A binding sites or it may actually inhibit the enzyme directly. Finally, do our data weaken the hypothesis that topoisomerase is involved in the intercalator-induced D N A strand breaks observed in L1210 cells? Clearly under the conditions studied, we found no evidence that protein-bound nicked D N A forms were generated. Rather the data are most consistent with the nicking-closing action of the enzyme on D N A which has been partially unwound by the intercalating agent. Upon removal of the drug, the D N A regains negative supercoiling to a degree that depends on the amount of drug bound. This is similar to the effects of ethidium bromide on the activity of D N A relaxing protein from human KB ceils as reported by Keller [19]. However, it is likely that pBR322 D N A is not an optimal substrate for this study because it is much smaller than chromosomal DNA, thereby making infrequent lesions difficult to detect. It also lacks

the higher order structural characteristics of chromatin, such as the presence of nucleosomes, nuclear matrix and a number of histone and non-histone proteins. Thus, we believe the hypothesis remains viable but its proof will require the development of a better experimental model.

Acknowledgements The authors wish to thank Dr. Tom Rowe for the many invaluable discussions concerning the work reported herein. We also thank Pat Strickland for her assistance in typing this manuscript. This work was supported by N I H research grant Ca-24586 and R C D A CA-00537 (W.E.R.).

References 1 Gellert, M. (1981) Annu. Rev. Biochem. 50, 879-910 2 Ross, W.E., Glaubiger, D.L. and Kohn, K.W. (1978) Bioehim. Biophys. Acta 519, 23-30 3 Ross, W.E., Glaubiger, D. and Kohn, K.W. (1979) Biochim. Biophys. Acta 562, 41-50 4 Ross, W.E. and Smith, M. (1982) Biochem. Pharm. 31, 1931-1935 5 Clewell, D.B, (1972) J. Bacteriol. 110, 667-676 6 Laemmli, U.K. (1970) Nature 227, 680-685 7 Bradford, M, (1976) Anal. Biochem. 72, 248-254 8 Tang, D. (1978) Nucleic Acids Res. 5, 2861-2875 9 Brun, G., Vanner, P., Scovassi, I. and Callen, J.C. (1981) Eur. J. Biochem. 118, 407-415 10 Rowe, T.C., Rusche, J.R., Brougham, M.J. and Holloman, W.K. (1981) J. Biol. Chem. 256, 10354-10361 11 Liu, L.F. and Miller, K.G. (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 3487-3491 21 Champoux,J.J. and McConaughy,B.L. (1976) Biochemistry 15, 4638-4642 13 Keller, W. (1975) Proe. Natl. Acad. Sci. U.S.A. 72, 2550-2554 14 Rall, S.C. and Cole, R.D. (1971) J. Biol. Chem. 246, 7175-7178 15 Welch, S.L. and Cole, R.D. (1980) J. Biol. Chem. 255, 4516-4518 16 Kinkade, J.M. (1969) J. Biol. Chem. 244, 3375-3380 17 Fairfield, F.R., Bauer, W.R. and Simpson, M.V. (1979) J. Biol. Chem. 254, 9352-9354 18 Vaughn, P.R., Loewe, H. and Nagley, P. 91979) Mol. Gen. Genet. 172, 259-269 19 Keller, W. (1975) Proc. Natl. Acad. Sci. U.S.A. 72, 4876-4880