Purification and Characterization of ...

6 downloads 0 Views 4MB Size Report
their excellent technical assistance, Dr. Ken Baughman for assistance in metal ... A,, Beemer, F. A,, Weits-Binnerts, J. J., Penders, T. J., and van der. 16.
THEJOURNAL OF

Vol. 267, No. 24, Issue of August 25, pp. 17102-17109.1992 Printed in U.S. A.

BIOLOGICAL CHEMISTRY

(0 1992 by The American Society for Biochemistry and Molecular Biology, Inc

Purification and Characterization of Dihydropyrimidine Dehydrogenase fromHuman Liver* (Received for publication, March 26, 1992)

Zhi-Hong Lu, Ruiwen Zhang, and Robert B. DiasioS From the Department of Pharmacology, University of Alabama, Birmingham,Alabama 35294

chemotherapy. Studies from our laboratory and others have demonstrated that more than 85% of administered 5‘-fluorouracil (FUra), one of the most frequently used anticancer drugs, is catabolized by this enzyme (6). It has also been demonstrated that the anticancerefficacy of FUra is related to DPD activity (7). Experimental and clinical studies have demonstrated DPD activity to havea circadian variation (810). This circadian pattern may have an important role in FUra chemotherapy, since FUra plasma levels have a corresponding inverse circadian pattern in patients receiving FUra chemotherapy (9). Additional studies with competitive DPD inhibitors (11-13) havealsoshown the importance of this enzyme in cancer chemotherapy. More recently, the significance of this enzyme has been investigated in patients with DPD deficiency (14-19). Since we demonstrated genetic deficiency of DPD in humans (14), several clinical studies (16-19) have shown that patients with DPD deficiency experience severe toxicity during 5’-flUOrOpyrimidine administrationwhich requires cessation of chemotherapy. Pharmacogenetic and pharmacoepidemiologic studies have suggested that thefrequency of this genetic defect is greater thanpreviously recognized (19). Since themajor site of pyrimidine catabolism is in theliver (20), mostof the studies with DPDwere carried out with the liver tissue. In the last threedecades, DPD has been purified to varying degrees from liver of several species, includingcow (21, 22), rat (23-25), mouse (26), and pig (27-29). However, homogeneity in purificationwas not obtained inmost of these preparation, and very little is known about the human liver Dihydropyrimidine dehydrogenase (EC 1.3.1.2, DPD),’ the enzyme. Studies have suggested that DPD might be speciesinitial rate-limiting enzyme in pyrimidine catabolism, cata- specific since the antiserum to rat liver DPD did not precipilyzes the following reaction. tate dog or guinea pig liver DPD activity (25). Species differences in this enzyme were also shown in a recent report of Pyrimidine + NADPH + dihydropyrimidine + NADP+ pig liver DPD (28) compared with rat liver DPD (24, 25). It The biological significance of the enzyme has been demon- is unlikely that DPD obtained from animal tissues can be strated inseveral previous studies.This enzyme catalyzes the used to further investigate the biochemistry and molecular first reaction in the three-stepcatabolic pathway which con- biology of the humanenzyme, particularly thebasis of genetic verts uracil into @-alanine (1); this is the only pathway in deficiency. Therefore,itis highlydesirable to purify the biosynthesis of @-alanine in mammalian tissues (2). p-Alanine human enzyme and characterize it in more detail. has been suggested to be involved in several metabolic and In initial attempts to purify DPD from human liver in our neurotransmitter functions (3-5). laboratory, we used the methods reported inprevious studies Most importantly, this enzyme has a critical role in cancer (24, 28), but were unable to successfully purify this enzyme to homogeneity. This prompted us todevelop a new purifica* Supported by United States Public Health Services Grant CA- tion procedure for the humanenzyme describedin the present 40530. The costs of publication of this article were defrayed in part manuscript.Theexperimental designused inthepresent by the paymentof page charges.This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 study had several advantages over most previous studies of purification and characterization of this enzyme from other solely to indicate this fact. To whom correspondence shouldbeaddressed Box 600, 101 species. First, by introducing two new methods, chromatofoVolker Hall, University of Alabama, Birmingham, AL 35294. Tel.: cusing and HPLC gel filtration, high purity and yield of the 205-934-4578; Fax: 205-934-8240. human enzyme were obtained. Second,a specific reversedThe abbreviations used are: DPD, dihydropyrimidine dehydrophase HPLCmethod was used todeterminethe enzyme genase; FUra, 5-fluorouracil; SDS-PAGE, sodium dodecyl sulfatepurification andinkinetic studies. This polyacrylamide gel electrophoresis; FMN, riboflavin 5’-phosphate; activityduring method is a direct measure of product formation and overPBS, phosphate-bufferedsaline; BSA, bovine serum albumin.

17102

Downloaded from www.jbc.org at MAYO CLINIC MULTI-SITE on April 27, 2007

Although dihydropyrimidine dehydrogenase has been purified to varying degreesfrom several species, very little is known about the human enzyme. The importance of this enzyme has recently been shown with cancer chemotherapy, particularly in patients with genetic deficiency of this enzyme. In the present study, this enzyme was purified 7800-fold to homogeneity from human liver by introducing several novel methods including chromatofocusing, HPLC gel filtration, reversed-phase HPLC for the enzyme assay. Purified human enzyme has a molecular mass of 210 f 5 kDa and appears to be composed of two subunits. The apparent PI is pH 4.6 (2 0.2). The human enzyme contains approximately four flavin nucleotide molecules (two each of FAD and FMN) and 33 iron atoms per molecule of enzyme. Kinetic studies with uracil, thymine, 5-fluorouracil, and NADPH were carriedout. Amino acid composition and the N-terminal amino acid sequence of this enzyme were reported. A rabbit polyclonal antibody was raised and shown to be specific for the human liver enzyme. In conclusion, in the present manuscript, we report not only a novel procedure for purification of dihydropyrimidine dehydrogenase from human liver but also new data on its properties compared to other species, which will provide a basis for further biochemical and molecular studies of this enzyme.

17103

Dihydropyrimidine Liver Dehydrogenase Human TABLE I Purification of dihydropyrimidine dehydrogenase from human liver

ic activity“ Total

protein

Total

Step

Crude supernatant pH 4.85 treatment Ammonium sulfate fractionation” Chromatofocusing 2’,5’-ADP-Sepharose affinity Gel filtration

w

nmollrnin

nrnolfrninlrng

24229 15770 5719 193 1.97 0.63

491 1 4173 4 158 3271 1447 999

0.2027 0.2646 0.7271 16.981 734.50 1585.9

3624

29.5

%

85.0 84.7 66.6 20.3

-fold

1.3 3.6 83.8 7824

All values calculated using 5-fluorouracil as a substrate. *After dialysis and centrifugation. “

B

MATERIALS A N D M E T H O D S ~ RESULTS

Enzyme Purification-In the present study, DPD activity was purified from the soluble fraction of homogenized frozen human liver. Initially, the 100,000 X g human liver supernatant fraction was precipitated by addition of acetic acid to pH 4.85 followed by ammonium sulfate fractionation. After 55% ammonium sulfate precipitation, the pellet was resuspended FIG. 5. Native polyacrylamide gel electrophoresis of puriand dialyzed against 25 mM histidine-HC1buffer, pH 5.7, fied human liver DPD. Lane A contains 10 pg of the purified overnight and then loaded onto a PBE-94 chromatofocusing enzyme stained by Coomassie Blue R-250. Lane R also contains 10 column equilibrated with the same buffer. The column was pg of the purified enzymeand was stained using a silver stain eluted by polybuffer 74, creating a pH gradient from pH 5.6 technique as described under “Materials and Methods” (Miniprint). t o 4.0. DPD activity was subsequently eluted at pH 4.6 (f 0.2) (Fig.1, see Miniprint). Fractions containing DPD activitynique. Following electroelution from the gel, DPD activity were pooled, concentrated,and loaded onto a 2’,5’-ADP- was recovered from the single band. No enzyme activity was Sepharose 4B affinity column; proteins which did not bind detected from otherfractions of the gel. Thedenatured, and those loosely bound to the affinity matrix were sequen- reduced enzyme gave two sharp protein bands with apparent tially eluted with buffer A (see Miniprint) and an increasing molecular masses of 105 and 90 f 3 kDa on a 7% SDSstep gradientof buffer A containing 50 mM, 100 mM, and 200 polyacrylamide gel (Fig. 6). To further characterize this enmMKC1, respectively. DPD activity was recovered from the zyme under denaturing conditions, purified human liver DPD affinity column by elution with 0.1 mM NADPH (Fig. 2, see was examined using a 4-20% gradient SDS gel. With silver Miniprint). Concentrated, affinity-purified DPD activitywas staining, three different bandswith molecular masses of 105, thenchromatographedonanHPLC gel filtrationcolumn 90, and 15 kDa were observed (Fig. 7). The binding capacity (Fig. 3, see Miniprint) which separated DPD activity from of the 15-kDa band for Coomassie Blue R-250 was very low, a typical preparation, the final but this bandwas readily detected by silver staining. other protein contaminants. In product had a 7800-fold enrichment of enzyme activity, with Determination of the Isoelectric Point of DPD-Elution a n overall recovery of 20% (Table I). from the chromatofocusingcolumn demonstrated an apparent Molecular Weight Determination-Purified enzyme was ho- isoelectric point (PI) of4.6 (& 0.2) (Table 11). The elution mogeneous as judged by HPLC gel filtration on a TSK 250 pattern was symmetrical, further suggesting that thepurified column (calibrated with known standards) showing a single, human liver DPD was homogeneous. symmetrical peak corresponding to a molecular mass of 210 Flavin Determination-Purified human liver DPD had an f 5 kDa (Fig. 4, see Miniprint), which was not influenced by amber color (in buffer A) and showed the characteristic abthe presence of 2-mercaptoethanol. The homogeneity of pu- sorption spectrumof a reduced flavoprotein(data notshown). rified human liver DPD was also determined by native gel The nature of the flavin cofactor in the enzyme molecule was electrophoresis. Undernondenaturingconditions, asingle shown by HPLC tobe FAD and FMN.No conversion of FAD band (Fig. 5 ) was obtained from the native gel by staining to FMN was detectable under these experimentalconditions. with either Coomassie Blue R-250 or a silver-staining tech- FAD and FMN were quantitated by a simultaneous fluorometric assay. As illustrated in Table11, human DPD contains ’ Portions of this paper (“Materials and Methods,” Table111, and approximately 2 mol each of FAD andFMNper mol of Figs. 1-4, 8, and 9) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying enzyme. Metal andSulfide Determination-To determine the metal glass. Full size photocopies are included in the microfilm edition of the Journal thatis available from Waverly Press. content in this enzyme, purified human liver DPD was sub-

Downloaded from www.jbc.org at MAYO CLINIC MULTI-SITE on April 27, 2007

comes the problems of the previous DPD assay (22, 24, 28, 29), which was limited both by sensitivity andspecificity (25, 41). Third, using the purified human enzyme, a polyclonal antibody was for the first time generated and shown to be specific for human liver DPD. Finally, the N-terminal amino acid sequence of the human enzyme was determined. These new data and the availability of pure human DPD will provide a basis for further biochemical and molecular studies of the human enzyme.

H u m a n Liver Dihydropyrimidine Dehydrogenase

17104

A

B

C

TABLE I1 Comparison of hepatic dihydropyrimidine dehydrogenase from human. Dit. and rat Liver from Parameter

116.3kD 97.4 kD 66.2 kD

42.7 kD Lane R contains 10 pg of the purified enzyme stained by Coomassie Blue R-250. Lane C also contains 10 pg of the purified enzyme and was stained using a silver stain technique as described under “Materials and Methods” (Miniprint). kD, kilodaltons.

B

A 205.0 kD

,

49.5 kD

32.5 kD 27.5 kD 18.5 kD

FIG. 7. SDS-polyacrylamide gel electrophoresis (4-20% gradient) of purified human liverDPD. Lane A contains molecular mass markers. Lane B contains 10 pg of the purified enzyme stained using a silver stain technique as described under “Materials and Methods” (miniprint). kD, kilodaltons.

Pig ( 2 8 )

Rat (24,25)

220 (207)” 5.25 3.0 (14.0) NA NA (0.7) 3.75 (0.76)

of human liver DPD with rat and pig liver enzymes. N-terminal Amino Sequence-The N-terminal amino residues of the 105- and 90-kDa peptides, following separation on a 7% SDS-PAGE, were identical and the same as thatof native enzyme (Table IV). Optimization of pH and Temperature Conditions-In a series of 100 mM potassium phosphate buffers covering a pH range between 4.0 and 9.0 with FUra as a substrate, the highest DPD activity wasobserved at pH 7.4 (Fig. 8, see Miniprint). Similarly,when incubated at temperatures over a range between 4.0 and 70.0 “C, the highest DPD activitywas observed a t 37 “C (Fig. 9, see Miniprint). Kinetic Properties-Table V summarizes the kinetic studies of purified human liver DPD, with comparison to rat and pig liver enzymes. Using standard assay conditions at pH7.4 and 37 “C,in the presence of 200 PM NADPH, enzyme kinetic studies revealed apparent K,,, values for uracil, thymine, and FUra of 4.9, 4.8 and 3.3 PM, with corresponding V,,, values of 0.6, 0.7, and 0.9 pmol/min/mg protein, respectively. Under the above conditions, substrate inhibition was observed for all substrates examined in the study. In the presence of20 PM pyrimidine substrate, apparent K, values for NADPH were 9.6 PM with uracil, 15.8 PM with thymine, and 10.1 pM with FUra, respectively. Under these conditions, no significant inhibition by NADPH was observed. Immunological Characterization-In the present study, rabbit polyclonal antibody was generated againstpurified human liver DPD. Using thisantiserum,immunoblotanalysis of proteins in 100,000 X g human liver supernatant, after separation on SDS-PAGE(4-20% gradient), revealed a single 105kDa band (Fig. lOA, lane 2). Western blot analysisof purified human liver DPD showed three sharp bands with molecular masses of 105,90, and 15 kDa (Fig. lOA, lane 3 ) . Preimmune serum from the same rabbit did not detect any band under the same conditions (Fig. 10B, lanes 1 and 2). DISCUSSION

This present manuscript is the first report on purificathe jected to atomic absorption spectrometry. Approximately 33 tion and characterization of DPD from human liver. Commol of iron per mol of enzyme were detected; no iron was pared toprevious studies on purificationof this enzyme from detectable in the buffers used in the purification procedure. other species, the present study utilized a novel procedure and No other metal ionswere found inpurified enzyme. The acid- represents a 5-fold improvementon previous methods of labile sulfide content of purified human liver DPD was ana- purification of this enzymefrom rat liver (24) and%fold from lyzed to determine the binding mode of the iron atoms. As pig liver (28). A specific polyclonal antibody has been raised shown in Table 11, the acid-labile sulfide content was almost for the first time against human liver DPD. equal to the iron content suggesting the presence of Fe-S Purification of human liver DPD to homogeneity was accenters inpurified DPD; nosulfide wasdetected in the buffers complished by a combination of acid precipitation, ammoused in this study. nium sulfate fractionation, chromatofocusing, affinity chroAmino Acid Composition-The amino acid composition of matography, and HPLC gel filtration. The final product, when carboxymethylated DPD is listed in Table I11 (Miniprint). analyzed by native PAGE, consistedof a single protein band. These data represent the meanof four separate DPD prepa- Following electroelution from the nondenaturinggel, the band rations. Table I11 also compares the aminoacid composition was shown to have DPD activity. The degree of homogeneity

Downloaded from www.jbc.org at MAYO CLINIC MULTI-SITE on April 27, 2007

FIG. 6. SDS-polyacrylamide gel electrophoresis (7%)of purified human liver DPD. Lane A contains molecular mass markers.

Human

210 206 Molecular mass (kDa) 4.65 PI 4.60 53.4 33.2 Iron (mol/mol enzyme) 31.6 31.3 Inorganic sulfur (mol/mol enzyme) 1.7 1.50 FMN (mol/mol enzyme) 1.51 1.6, 1.9 FAD (mol/mol enzyme) ” Values in parentheses fromRef. 25. NA, data not available.

Dihydropyrimidine Liver Dehydrogenase Human

17105

TABLEIV Amino-terminal amino acid sequences of dih.ydropyrimidine dehydrogenase from humanliver Residue

Sample 1

2

Native enzvme” Val Leu 105-kDa bandh Val Leu 90-kDa bandh Val Leu “After HPLC gel filtration. After separation on SDS-PAGE (7%).

3

4

5

6

7

8

9

10

Ser Ser Ser

Lvs Lis LYs

Asr, Asp ASP

Ser Ser Ser

Ala Ala Ala

Asr,

Ile

Glu

TABLE V Comparison of kinetics for hepatic dihydropyrimidine dehydrogenase from human, pig, and rat

A

Downloaded from www.jbc.org at MAYO CLINIC MULTI-SITE on April 27, 2007

spondedtoDPDactivity; 2) DPD activity was recovered following electroelution from the single band of the nativegel (no other proteins and enzyme no activity were detected from Liver from other fractions of the native gel); 3) fractions from chromaSubstrate Parameter tofocusing, affinity, and HPLC gel filtration columns which Human Pie (28) Rat (24) had DPD activity were shown on SDS-PAGE to contain the Uracil K m ( W ) 4.9 1.98 1.80 105-, 90-, and 15-kDa polypeptides (other fractions without 0.6 0.33 0.69 Vmax(rmol/min/mg) DPD activitydid not contain any oneof these threepolypepK,,, ( p ~ ) 2.66 Thymine 4.8 2.6 tides); 4) N-terminal amino residues from native DPD (210 0.7 0.25 0.49 Vmax (rmol/min/mg) kDa) and from 105- and 90-kDa polypeptides were identical; 5-Fluorouracil K,,, ( p ~ ) 3.3 5.50 NA” 5) immunoblot analysis using the rabbit polyclonal antiand 0.9 0.4 NA V,,, (rrnollminlmg) body detected a single 105-kDa protein band with the crude K,,, ( p M ) with uracil NADPH 9.6 11.36 11 human livercytosol,whereas threebandswith molecular K,,, ( p ~with ) thymine 15.8 NA 15 15 kDa were detected with purified masses of 105, 90, and 10.1 NA K,,, ( p M ) with 5-fluoNA DPD. rouracil Determination of the isoelectric point (PI) of purified en’NA, data not available. zyme revealed alower PI forhuman liver DPD compared with rat liver DPD (24) (PI5.25). In thisrespect, human liver DPD is more like pig liver enzyme (28) with a similar PI (4.60 us. 4.65). 1 2 3 1 2 As illustrated in Table I11 (Miniprint), comparison of the amino acid composition of DPD from three mammalian spe205.0 kD cies (human, rat, and pig) demonstrated that asignificant deviationin compositionoccurredforacidic amino acids (moreabundantinhuman liver DPD).HumanDPDhas approximately twice as many histidineresidues as rat and pig 49.5 kD DPD. 32.5 kD The amber color (in buffer A) and the characteristic absorption spectrum of human liver DPD suggest it is a flavo18.5 kD protein. Equal amounts of FMN and FAD were detected in purified enzyme. Similar results were reported for pig liver DPD (28), as shown in Table 111. In contrast, Shiotani and Weber (24) found only FAD in rat liver DPD (4 mol per mol FIG.10. Immunoblot analysis of human liver DPD. In A , of enzyme). More recently, Fujimoto et al. (25) reported both liver DPD (1mol of each flavin per mol polyclonal antibody was used as theprimary antibody: lane I contains FAD and FMN in rat pre-stained molecular weight standards, lane 2 contains 200 pg of of enzyme). The role of flavins in this enzyme is unclear. It crude human liver cytosol, and lane 3 contains 0.5 pg of purified has been suggested that flavin may regulate the enzyme halfDPD. InR, preimmune serum was used as theprimary antibody: lane life or synthesis (25). I contains 200 pg of crude human liver cytosol, and lane 2 contains Determination of metal and acid-labile sulfide contents of 0.5 pg of purified DPD. Bound antibody in both A and B was detected human liver DPD revealed similar amounts per mol of enwith alkaline phosphatase-labeled goat anti-rabbit IgG as described zyme, suggesting the presence of Fe-S centers. Purified human under “Materials and Methods” (Miniprint). liver DPD contained8 mol each of iron and acid-labile sulfide per mol flavin nucleotide. These results were in agreement of the native enzyme was demonstrated by the symmetry of pig liver enzyme (28).However, the results the single peak (absorbance and DPD activity)by HPLC gel with the report on filtration. Further confirmation of the homogeneous nature of irondetermination for rat liverenzyme from different of purified human liver DPD was obtained usinga polyclonal preparations varied. Shiotani and Weber(24) reported only 3 mol of iron permol of enzyme, but Fujimoto et al. (25) reported antibody raised in rabbits againstpurified enzyme. sulfide contents When purified DPD was resolved by SDS-PAGE on a 4- 14 mol of iron permol of enzyme. The data on 20% gradient gel, three polypeptide bands, with molecular of rat liver DPD were not reported (24, 25). Inmost of the previous DPD purifications from other masses of 105, 90, and 15 kDa, were observed. Based on the following data, we suggest thatnativehuman liver DPD species (24, 28, 29), enzyme activity was determined by the abconsists of two 105-kDa subunits with the 90- and 15-kDa decrease in NADPH assessed by measuring changes in limited inboth sensitivity polypeptides representingdegradation products: 1) under sorbance a t 340 nm. This method is and specificity, particularly in the first several steps of purinondenaturingconditions purifiedenzyme elutedduring than one enzymeconsumesNADPH. HPLC gel filtration as one symmetrical peak which corre- ficationwheremore

17106

Dihydropyrimidine Liver Human

Acknowledgments-We thank Mark Lusco and Xiao-Ying Liu for their excellent technical assistance, Dr. Ken Baughman for assistance in metal determination, Martin Johnson and Dr. Denise Shaw for their advice in the production of the polyclonal antibody, and Dr. Stephen Barnes, Dr. Charles Falany, Dr.Mahmoud el Kouni, and Dr. Fardos Naguib for their critical review of the manuscript. REFERENCES 1. Traut, T. W., and Loechel, S. (1984) Biochemistry 2 3 , 2533-2539 2. Wasternack, C. (1978) Biochem. Physiol. Pflanr. 173,371-403 3. Bauer, K., Hallermeyer, K., Salnikow, J., Kleinkauf, H., and Hamprecht, B. (1982) J. Biol. Chem. 257,3593-3597 4. Zafra, F., Aragon, M. C., Valdiviesco, F., and Gimenez, C. (1984) Neurochem. Res. 9,695-707 5. Holopainen, I., and Konto, P. (1986) Neurochem. Res. 1.1,207-215 6. Diasio, R.B., and Harris, B. E. (1989) Clm. Pharmacokmet. 1 6 , 215-237

7. Iigo, M., Nishikata, K., Nakajima, Y., Hoshi, A,, Okudaira, N., Odagiri, H., and De Clereq, E. (1989) Biochem. Pharmacol. 3 8 , 1885-1889 8. Harris, B. E., Song, R., He, Y.-J., Soong, S.-J., and Diasio, R. B. (1988) Biochem. Pharmacol. 37,4759-4762 9. Harris, B. E., Song, R., Soong, S.-J., and Diasio, R. B. (1990) Cancer Res. 50,197-201 10. Daher. G. C.. Zhane. R.. Soone. S.-J..and Diasio. R. B. (1991) Drug Metab. Dispos. 19,285-287 ’ ’ 11. Wasternack, C. (1980) Pharmacol. & Ther. 8,629-651 12. Martin, D. S., Nayak,P.,Sawyer, R. C., Stolfi, R. L., Young, C. W., Woodcock, T., and Spiegelman, S. (1978) Cancer Bull. 30,219-224 13. Daher. G . C.. Naeuib. F. N. M.. el Kouni. M. H.. Zhane. R.. Soone. S.-J.. and Diasio, R. B. (i991) Bi&hemr Pha;macol. 41,18&1893 ’ 14. Diasio, R. B., Beavers, T. L., and Carpenter, J. T. (1988) J. Clin. Inuest. 81,47-51 15. Wadman, S. K., Berger, R., Duran, M., de Bree, P. K., Stoker-de Vries, S. A,, Beemer, F. A,, Weits-Binnerts, J. J., Penders, T. J., and van der Woude J. K. (1985) J. Inherited Metab. Dis. 8, (Suppl. 2). 113-114 16. Brocksedt, M., Jakobs, C., Smit, L. M. E., van Gennip, A. H., and Berger, R. (1990) J. Inherited Metab. Dis. 1 3 , 121-124 17. Braakhekke, J. P., Renier, W. O., Gabreels,F. J. M., deAbreu, R. A,, Bekkern, J. A. J. M., and Sengers, R. C. A. (1987) J. Neurol. Sci. 7 8 , 71-77 18. van Gennip, A. H., Abeling, N. G., Elzinga-Zoetekouw, L., Scholten, L. G., Van Cruchten, A,, and Bakker, H.D. (1989)Adu. Exp. Med. Biol. 253A, 111-118 19. Harris, B. E., Carpenter, J. T., and Diasio, R. B., (1991) Cancer 6 8 , 499501 20. Ho, D. H., Townsend, L., Luma, M. A,, and Bodely, G. P. (1986)Anticancer Res. 6 , 781-784 21. Grisolia, S., and Cardoso, S. S. (1957) Biochim. Biophys. Acta 25,430-431 22. Porter, D. J. T., Chestnut, W. G., Taylor, L. C.E., Merrill, B. M., and Spector, T. (1991) J. Biol. Chem. 2 6 6 , 19988-19994 23. Fritzson, P. (1960) J. Biol. Chem. 235,719-725 24. Shiotani, T., and Weber, G. (1981) J. Biol. Chem. 2 5 6 , 219-224 25. Fujimoto, S., Matsuda, K., Kikugama, M., Kaneko, M., and Tamak, N. (1990) J. Nutr. Sci. Vitaminol. 37,89-98 26. Sanno, Y., Holzer, M., and Schimke, R. T. (1970) J. Bid. Chem. 2 4 5 , 5668-5676 27. Goedde, H. W., Aganval, D. P., and Eickhoff, K. (1970) Hoppe-Seyler’s Z. Physiol. Chem. 351,945-951 28. Podschun, B., Wahler, G., and Schnackerz, K. D. (1989) Eur. J. Biochem. 185,219-224 29. Podschun, B. Cook, P. F., and Schnackerz, K. D. (1990) J. Biol. Chem. 26R. - - , -12966-1 -- - - --2972 ..30. Sommadossi, J.-P., Gewirtz, D. A,, Diasio, R. B., Aubert, C., Cano, J.-P., and Goldman, I. D. (1982) J. Biol. Chem. 257,8171-8176 31. Merril. C. R.. Goldman., S.., Sedman., S. A.., and Ebert. M. H. (1981) Science 211,1437-1438 32. Faeder, E. J., and Siegel, L. M. (1973) Anal. Biochem. 53,332-336 33. Massey, V., and Swoboda, B. E. P. (1963) Biochem. Z. 338,474-484 34. Rabinowitz, J. C. (1978) Methods Enzymol. 5 3 , 275-277 35. Cleland, W. W. (1979) Methods Enzymol. 63,103-138 36. Allen, G. (1989) in Laboratory Techniques in Biochemistry and Molecular Biology, Volume on Sequencing of Protein and Peptides (Burdon, R. H., and van Knippenber, P. H., eds) pp. 58-59, Elsevier Science Publishing Co., Inc., New York 37. Matsudaira, P. (1987) J. Biol. Chem. 2 6 2 , 10035-10038 38. Gaastra, W. (1984) in Methods in Molecular Biology,Volume 1, Proteins (Walker, J. M., ed) pp. 349-355, Humana Press, Clifton, NJ 39. Towbin, H., Staehelin, T., and Gorden, J. (1979) Proc. Natl. Acad. Sci U. S. A. 76,4350-4354 40. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 1 9 3 , 265-275 41. Naguib, F. N. M., el Kouni, M. H., and Cha, S. (1985) Cancer Res. 4 5 , 5405-5412 ”

~~

~

~



Downloaded from www.jbc.org at MAYO CLINIC MULTI-SITE on April 27, 2007

Therefore, we developed HPLC procedures to separate pyrimidines and their catabolites, modifying an earlier approach (30).The enzyme activity in the present studywas quantitated by measuring specific product formation. Using thisHPLC method, kineticstudies have demonstrated similar kinetic properties for the natural substrates, uracil and thymine. Significant substrate inhibition was observed for uracil, thymine, and FUra at 100 PM or higher. Substrate inhibition was reported with purified pig liver DPD (28, 29) and crude extracts of some humantissues (41). However, no substrate inhibition was reported with rat liver DPD (24). Inthe presence of 20 p~ of each pyrimidine substrate, saturation of enzyme activity was detected at 30 I.IM NADPH, but significant inhibition by NADPH was not observed until 1000 p~ NADPH. In the present study, FUra was the preferred substrate for human liver DPD compared to uracil and thymine.Similarresults were reported with crude human liver extracts (41). The variations in estimated kinetic parameters for different species may result from several factors, including species differences, varying methods in determination of enzyme activity, and varying degrees of purification. In summary, in the present manuscript, we report not only a novel purification procedure for human liver DPD, butalso new information on its properties. Furthermore, the availability of pure human DPD, apolyclonal antibody againstthis enzyme, and new data on amino acid composition and sequence will provide a basis for further biochemical and molecular studies with this enzyme, particularlyrelevant to human DPD deficiency.

Dehydrogenase

17107

Human Liver Dihydropyrimidine Dehydrogenase SupplementalMaterlal To Purification and Characterization of Dihydropyrlmldlna dehydrogenase fromHumanLiver by Zhl-Hang Lu, Ruwen ZhangandRoben

Materials and Methods

m: Polybufferexchangergel(PEE94),polybuffer

74. molecular welght markers, 2 , 8'-ADP-Sepharose48,were obtalned from Pharmacla(Plscataway. NJ). Coomassle brtlllant blue R-250, acrylamlde. and pre-stalned molecular welght markerswerepurchased tram Blo-Rad (Rtchmond.CA)Alkallnephosphatase-labeled goat anti-rabblt antibody. nmoblue tetraroilurn and 5-bromo-4.~hloro-3-1ndolyl from SouthernBlotechnology(Blrmlngham, phosphaten-toluldlnesaltwereobtaned AL) NADPH, FMN, andFADwerepurchased from Srgma (St LOUE MO) L-Hcstrdme was obtalned from Aldrlch (Mrlwaukee, WI) pH]-FUra (25 ClImmol) was obtalned from New EnglandNuclearCorp (Boston, MA) [6-14Cl-Urac~l(55 mClImmol)and[2-'4CIthymlne (52 rnCllmmol) were obtalned from Moravek Btochemlcals (Brea, CA) Radiochemcalswerepurlfledby HPLC andthelr purlty wasdetermlnedby HPLC to be >99% All other solvents and reagents were purchased in the hlghestgrade available The major buffer (buffer A) used In the preparation of thls enzyme contalned 35 mM potasslum phosphate, 2 5 mM magnesium chiortde. 10 mM 2-mercaptoethanol, pH 7 4Theequllibratlon buffer lor thechromatofocuslng Column (hlstldlnebuffer)contamed25 mM L-hlStldlne-HCI, 10 mM 2-mercaptoethanol. pH 5 7 Elution buffers lor alllnltycolumnandgel f8ltratlon column were prepared from butlerA

M o l e c u lW a re i a h l Determination: The molecular welght of native DPD was determlnedbyHPLCgelflitratlonA 2 15 x 60 cm TSK-250gelflltratlonHPLCcolumn (Blo-Rad) was equlllbratedwlthbuffer A, pH 7 4, ata flow rate of 2 5 mllmm The column was calibrated uslngknownmolecularwelghtstandardsandthe retention tlme of lndlvldualprotelnsdetermlnedbythelrpeaks of absorbanceat280nm The retentlon tlme 01 purlfled enzyme was then compared to those of the molecular weightstandards The molecularwelght of reduced,denaturedDPDwasdetermlnedby SDS-polyacrylamide gel electrophores,s. usmg standardprotelos. of known molecular werqhts F l a v iD neterminarion: The purified enzyme dissolved ~n 35 mM In a water phosphate pH 74.2 5 mM MgC12 and 5 mM2-mercaptoethanolwasbolied bath for 10 mln In the dark to release ( l a m After removlng the preclpttate by centrlfugatlon. aliquots of supernatant were analyzed qual#tatlvely for llavln composition byHPLCseparatlononareversephase Cq8 column wltha h e a r gradlent (0-66%methanol) In 20 mM potasslumphosphate.pH 5.6. at a flow rate of 1 mllmln at 25 'C Flavlns Were detected by their absorbance at 230 nm The FADIFMN composctlon of the supernatant was analyzed quantltatlve fiuorescence by purlf8ed on measurements at different pH values(32)wlthFADandFMNstandards DEAE-cellulose(33) M e taanlSdu l f i dDee t e r midetermlned atomlc by absorption measuredbythemethylenebluemethod(34)

The metal content of ourlfled DPD was spectrophotometry Acld-labile sulflde

was

7 Separatmn of HP pyrmcdlnesand their catabollteswasperformedbyreverse-phaseHPLC uscng a Hewtett-Packard 1050 HPLCsystemequlppedwith a fclter spectrometric detector and chromatographic termlnal (HP 3396 S e w s I1 Integrator) Two Hypersli 5 Irm columns (JonesChromatography,Llttleton, CO) were used ~n tandemasthestationary phaseAnalysls of FUraand Its cataboiltes was carrledoutat flow rate of 1 0 mllmln wlth the mobliephasecontalnlng t 5 mMpotasslumphosphate.pH 8 0, wlth 5 mM tetrabutylammonum hydrogen sulfate (30) Under these Conditions. typical retentlon times lor d~hydrofluorourac~l andFUrawere9and 21 mtn respectively Uslng the same statlonary phase as above. analysts of thymlne and (1s calaboltteswascamed out at a tlow rate of 0 5 milmln wlth the moblle phase contalmng 1 5 mM potasslum phosphate, pH 8 wlth 4 5 mMtetrabutylammonlum hydrogensullateUnderthesecondmons.typlcalretentlontlmes for dlhydrothymlne andthymmewere22and27 mm. respectively Analyss of uracIIand 81s cataboilles wasais0carrled out usrng thesameHPLCsystem lor analysts of thymcne and cts catabolrtes. with typical retentton times of 13 and 19 mln lor dthydrourac81 and uracil respectwely eplvacrvlamide SDS Gelelectrorrhoresls:SDSPAGE wascarrledout In a1 0 mm thlck. 7% (wlv) polyacrylam~degelcontalmng 0 375 M T~Is-HCI(pH 8 8) and 0 1% SDS Sampleswerepreparedby mlmng themwlth an equal volume of samplebuffer (0 0625 MTrts-HCI. pH 6 8 . 10% glycerol. 0 2% SDS ( w l v ) 80 mM 2-mercaptoethanol)and bolllng for 5 minutes Electrophoresis was conducted at a constant current of 30 mA for 30 mln at 25% Gradtent SDS-PAGEwas carrled out r a t 0 mm thlck. 4.20% gradlentgel (BIoRad Mm-Protean 11) Samples wereprepared by mlxlng them wlth lour volumes of the abovesample b u l k and b o h g for 5 mcnutes Theeiectrophorests was Conducted tollowlng the manufacturer's Instruction. ata Coflstant voltage 01 200 V lor 60 mln at 25'C Natlve Polvacrvlamide Gel EleytLPohoresip: Natlve gel electrophoress was cawed out In a t 0 mm thick 9% (wIv)polyacrylamldegelcontalnmg 0 06 MTrls-HCI (pH 8 8 ) . wtth 0 0025 % (wlv)rlboflavlnphosphateSampleswerepreparedby mwng them wlth an equalvolume of sample buffer (40%sucrose. 10 mM 2-mercaptoethanol) The electrophoresiswas Conducted at a Consfant current of 30 mA lor 30 mln at 4 'C I The gel was flxed in a 5% methanol I 7% Stainino P r o c e d w 1 1 acettc acid SOlUtlon tor 30 mln andstamedovernlghtusing 0 01% (wlv) Coomass8e brcilrant blue R-250 In a 5% trlchtoroacetc acrd I 2 5% methanol I 3 5% acetrcacld solution 2 ) m e r Sta !me The gel was flxed 10 40%methanol I 10% acetlcacld for 40 mln andthenstalnedusingthe 010-Rad sllver stain asderlved from themethod 01 Merrll et a1 (31) Br8efly. Iollow8ng flxatlon. the gel was Incubated In oxldcer SOlUtlOn for 20 mln The gel was then washed Wlth dlstlileddelonued water and incubatedwlth Silver S o I ~ t 1 0 nfor 30 mln Thegelwasagamwashedwlth d8st111ed delonlzed water and Incubated with the developing solution supplled by the manufacturer

n F r o-m N : Gel electrophoresis was carrled out on 200pgpurlfled DPD under non-denaturing condmons m a 9% ( W I V ) poiyacrylamtde gelas descrcbedaboveA 0 5 cm slnp from !he gelwascutand statned according to themethodsdencrlbedabove Thls strtp washnedup with the unstatned gel and the smgle correspondmgband cut out of the unstained gel The gel was mlnced and electroeluted 8n a Blo-Rad Model 422 eleclro-eluter rn 25 mM TrSsI192 mMglyclnebuffer.pH 8 3, Contain 5 % glycerol, 5 mM 2-mercaptoefhanol for f o u r hours at I O mA (Constant current) at 4 'C The Sample was then dialyzed overnlghtat 4 % In 1 Mer of bufferA, p H7 4 , beforebelngassayedOther fractions from the gelweretreated I " the Same way

m a r a t i o n ot Polvclo n a l Anttbodv;MaleNewZealandrabbltswere Immunized wlth subcutaneous qeCtlOnS ofpurlfledDPD The first lnlectlon Consisted of 50 vg O f purlfled antlgenmlxedwlthanequal Volume 01 Freundscompleteadjuvant TWO weekslater.theserabbltswerelnlectedwlththeantlgen (50119) mlxed I" anequal volume of Freund'slncompleteadjuvant.threeweeks following thesecond In,ectlon, this lnlectlonwasrepeatedAllquots of serum Samples from earnlckswerescreened l o r antibody tormatlonuslngenzyme-lmkedlmmunosorbantassay (38) andWestern blot analysis (39) Two weeks tollowlng the thlrd mlection, the rabblts Were sacnftced by cardtac Puncture. andtheubloodcollected To allowthe blood to clot. thesample was rocubated at 37 "C for 60 m m leftat room temperature for 4 h, and then kept at 4 "C overnight The clot was gently removed, and the serum was centrifuged at2000RPMfor 15 mln The serum wasloaded on a t x t o cm protein A~ Sepharose Fast 4 Flow column (Slgma Chemlcal Co , St LOUIS. MO). prevlourly sallne Thecolumnwaswashedwlth 4 column equlllbratedwlthphosphate-buffered the IgG antibodies wereelutedwlthan acld voiumes of phosphate-bufferedsallne.and washconststing of 0 2Mglyclne-HCIcontalnmg 0 075 MNaCI, p H 2 5 Immediately uponeiutlon from thecolumnthe fractions wereneutraltredwlth1 0 M Trls.HC1, pH 10 SDS-PAGEona4-20 % gradlent gel wasperformedusmg freshlyprepared 100.000 x g humanilversupernatantandpurifredhuman lkver DPD The pratems Were transferred from thegel to a n~troceliulose fclter fallowtng the method of Towbin(39) The nltroceIiuloseIllterwasIncubatedovernlghtat4"C wlth thepolyclonalantlbody(lgG) puilfled by prote8n A column (dlluted t 2000) ~n a 120 mM borate-salme Solution contalnlng 1% (wlv) BSA, p H 8 5 Then~troceIIulose fslter waswashedwlthborate-salmecontalnlng 0 1% Tween20 (vlv) andlncubatedwltha secondary. alkaljne phosphatase-labeled goat ant,-rabblt anttbody The location 01 lmmunoreactlve PrOtelnS O n the n~trOCellUloselllterwasdeveloped ln a 0 1Msodlum carbonatebuffer (100 ml pH 9 5) Contalnlng 30 mg nmo blue tetrazolium (addedasa 1 ml SOlUtlOn dlssolved ~n 70%dlmethyl-formamlde)and15mg5-bromo-4-~hloro-3lndolyl phosphate p-toluldlne salt (added as 1a ml solution dlssolved ~n too% dmethyltormam~de) P r o t e i n Datermlnarlon. . . The amount of proteln ln the sample was determmed by themethod Of Lowry et a1 (40) uslng810-Radprotelndetetmlnatlonreagent BSA asa Standard

usmg

Downloaded from www.jbc.org at MAYO CLINIC MULTI-SITE on April 27, 2007

OPQ Enzvme A s s w The enzyme actlvity durlng purlflcatlon was determlned by measuring the catabolltes of FUrausngreverse-phase HPLC(18.30)Thereactlon mfxture contalned35mMpotass!umphosphate, p H 74.2 5 mM magnesturn chlarrde. 10 mM 2-mercaptoethanol.200pMNADPH,20 vM[IHI-FUraand enzyme solut~onIn flnalvolume of 2 ml The mixture was Incubatedat37°C.and350 pI ofthereactlon sample wastaken out at various llmes andadded Into thesamevolume of ethanol to stop the reactlon The sample was kept In freezer a (-20°C) for 30 mln, flltered through 0 2 pM Acro lllter(GelmanSciences. Ann Arbor. MI)andthenseparated by reverse-phase HPLC

0 Dlasto

17108

i"-jJ

Human Liver Dihydropyrimidine Dehydrogenase nooo

E n z v mPeu r i f i c a t i o n Allprocedureswereperformedat4

'C

6000

F r a c t i o n 1: PreDaration of Crude Extract: Human lher (recewed from the Natlonal Dlsease ResearchInterchangethroughanmstltulionallyapprovedprotocol) soon as posslbleafter cessation 01 cardlac wasremovedfromtransplantdonorsas functlon The 11ssue was cut Into 250grampleces.perfusedwjthcoldsallne,and frozen at -70%Twenty-fourhoursprior lo use, ltver was placed ~n apaper-llned Ice set ~n a 4 "C cold room The partially thawedliverwasmlncedand bucket and homogemred ~n four volumesofbuffer A, m the presence 01 0.25 M sucrose,1 mM benramldme. 1 mM am~noethyl~soth,ouroniumbromlde. and 5 mM EDTA The homogenatewascentrlfugedat100.000 x g for 60 min In order to obtainacytosolic fractlon

5

2

4000

n

. .

D r e c l' w: Fraction 2: B c i d Acetic acld was added ti h oe resulting supernatant 01 homogenate(Fracllon1) lo adlust pH to 4 85 wllhconstantstlrrlng for 15 mln atthls pH Theenzyme Solullon wasthencenlrlfuged at 30,000 x g for 30 mm The supernatant was removed and adpsted wlth 0 5 N KOH io pH 7 4.

.

.

Fraction 3: @ ' ' Sohd ammonlum sulfate was l o Fractlon2untlla33% Saturation was obtalnedThemlxturewas slowlyadded starred lor 30 m l r and then centrlfuged at 30,000 x g lor 30 mln. Addltlonal ammoniumsulfatewasadded l o thesupernatantuntlla55% Saturation wasobtalned mln The enzyme soIut1on was then centrlfuged at wllh constant stirrlng for 30 30,000 x gfor 30 mln Thepreclpllatewasdlssolved In 25 mM hlstldine-HCIbuffer. 10 hter of thesame buffer pH5 7. anddlalyredovernlghtagainst

Fraction 5: 2 ' 5 ' ADP-SeDharose 4 8A f l i n f t vC h r o m a t e c l r a n h r The pooled from thechromalofocuslngcolumnwereconcentratedby lract~onswllh DPDactlwly Amlconcentrlprep10concentratorandloadedontoaZ'.S"ADP-Sepharose 4 8 afflnlty x 40cm)prevlouslyequilibrated w t h bufferAThecolumnwaswashed column(1 01 buffer A, 10columnvolumesof 50 mM KCI-bulfer A, O I wlth20columnvolumes columnvolumes100 mM KCI-buffer A, 2column volumes of 200 mM KCI-buffer A Enzyme actlwtywaselutedwlth 0 1 mM NADPH In buffer AFractions conlamng DPD actwly werepooledandconcentrated In anAmlconcentrlcon10concentrator

0

3 100

0

300

200

Fraction DPD a c l i v i t yf r o m a PBE-94 E l u t i o np a t t e r no f chromaloforusing column. Enzyme actlvtty Wdb eluted wtth polybuffer 74 diluted 1:8 wllh d!snlled delonlzed wdter, p l l 4 0 Fractm m e was 7.5 ml wtth a flow rate nf 75 mlhr.

L

FIGURE

40 .................................

H n

2000 0

........ N,$DVH 0.1 m M

20

0

no

60

40

100

Fraction

Elution pattern of DI'D activity from a Z'S-ADP-Sepharose column. Enzyme acrwlty was eluted by 0.1 mM N A D P I i . Fractlon stze wa\ 1.5 rnl utth a flow rale of IS rnlhr

4 8 affinity

.

.

Fraction 6: ' The pooled, concentratedfractions wllhDPDactlvlly from thealflnltycolumnwerelnlected onto aEloradTSK-250gel filtration column (2 1 5 x 60 cm). previously equlllbrated wlth buffer Enzyme A actlwtywaselutedbybufferA In a flowrate of 2 5 mlimln Fractions contalnmg ~n an Amlcon cenlrlcon 10concentrator. DPDacllwtywerepooledandconcentrated

Amlno acldcompositions

Asp + Asn Glu + Gln Ser G~Y HIS Arg Thr Ala Pro Tyr Val Met CYS Ile Leu Phe LYS Trp

NO

=

of

Table 111 hepatlcdlhydropyrlmldmedehydrogenasefrom human, plg, andrat

163 5 189 8 108 7 160 5 38 7 92 6 130.3 185 1 147.3 66 0 140 4 51 9 16 2 1157 199 7 92 2 94 3

177 7 196 4 1194 185 4 22 0 72 7 109 6 171 5 126 2 30 3 108 4 94 0 30 7 106 9 167 4 74 3 124 1

NO

NO

185 1 1412 121 9 121 1 22 3 63 0 102 7 121 9 113 3 35 9 86 2 41 8 27 9 99 9 134 1 W.1 92 0

NO

Hn

50

70

90

110

no

150

170

190

Fraction ELWELL Elution pattern

of DPD activity from a HPLC gcl fillration A. p H 7.4 c o l u m n . Enzyme actwrty was eluted tn 46minwlthbuffer Fractmn s u e u a s I 25 mi wlth a conttant flaw rate o f 2 5 mllmln.

1000

m

2 2 .-

M

100

Not Determmed

10

4 30

I

I 40

50

60

Retention Time (min) Molcrular wcighl delerminalion of native purified human liver DPD by HPLC gel filtralion chromatography. The determination was conducted on a 2 IS x 60 cm TSK-250 gel flltratmn column uslng buffer A as a mablle phase tn a constant flaw rate o f 2.5 mllmln. Mnlecular welght standards used arc' fcrrltm (440 kDa). catalase (272 kDa): aldolase (158 kDa): bavm serum albumm (66 2 kDa).

Downloaded from www.jbc.org at MAYO CLINIC MULTI-SITE on April 27, 2007

dlalyred sample from Fractlon 3 was Fraction 4: ~ h r o m a t o f o c u s i n a : The centrlfuged at30,000 x g lor 30 mln andthenloaded onto achromatofocusmgcolumn (1 6 x 100cm)packedwlthPEE-94prevlouslyequthbratedwlth25 mM hlstbdlne-HCI The column was re-equlllbrated wlth 5 column volumes of the buffer. p H 5.7 equlllbratlonbufferThecolumnwasthenelutedbyapolybuffer74dlluted1 8 wlth d1stMed delonlredwater(flnalpHadjusted to 4 0 wlth HCI). In thepresence of 10 mM 2-mercaptoethanol

4

',

2000

Human Liver Dihydropyrimidine Dehydrogenase 120

17109

120

. A

-

LOO

5 .Z .-* h

h

h

2

.-

80-

60 40-

P

g

-

20

20

04 3

. , 4

,

.

5

. , . , . , 6

7

8

.

9

, 1

. I

0, 0

.

0

PH

.

.

20

,

.

40

-

60

80

Temperature ("C) 4

B

Optimization of pH of human liver DPD. Enzyme actwny was evaluated in a series of 1 0 0 mM potasslum phosphate buffers covering a pH range between 4.0 and 9.0 wlth J-fiuarouracll as a substrate. Reaction mmtures were incubated at 37 0 "C for 20 min. Acuwty 1s expressed as a percentage ofthe value obtained at pH 7 4.

3.0

3.5

4.0

1/T("K) X 10 .' EU.W& Optimization of temperature of human liver DPD. Enzyme actlvtty was evaluated ~n 1 0 0 m M potasalumphosphatebuffers. pH 7 4, over a temperature range between 4 and 70 'C. wlth 5-fluorourac1l as a substrate.Reaction mixtures wereIncubatedfor 20 mm. (A).Acuvity i s

expressed as a percentage of the value obtamed shown by Arrhenua plot

at

37 T . (B).The data are

Downloaded from www.jbc.org at MAYO CLINIC MULTI-SITE on April 27, 2007

2.5