Proc. Natl. Acad. Sd. USA Vol. 73, No. 5, pp. 1591-1593, May 1976 Biochemistry

Hepatic phenylalanine 4-monooxygenase is a phosphoprotein (adenosine 3':5'-cyclic monophosphate/protein phosphorylation/enzyme activation) SHELDON MILSTIEN, JEAN-PIERRE ABITA*, NORMAN CHANG, AND SEYMOUR KAUFMAN Laboratory of Neurochemistry, National Institute of Mental Health, Bethesda, Maryland 20014 Communicated by E. R. Stadtman, March 8,1976

ABSTRACT Phenylalanine hydroxylase [phenylalanine system have been described. Partial proteolysis of phenylalanine 4-monooxygenase; EC 1.14.16.1; L-phenylalanine, tetrahy- hydroxylase was carried out as described previously (1). dropteridine:oxygen (4-hydroxylating)J isolated Protein kinase (EC 2.7.1.37; ATP:protein phosphotransferase) from rat liver is a phosphoprotein containing approximately 0.31 from bovine heart muscle was ,umol of protein-bound phosphate per Asmol of subunit (50,000 purified and assayed as described molecular weight). When the enzyme is further phosphorylated by Gilman (5). in the presence of ATP and a 3':5'-cyclic-AMP-dependent protein Protein-bound and free phosphate were measured by the kinase (EC 2.7.1.37; ATP:protein phosphotransferase), an ad- method of Ames (6). ditional 0.7 Mmol of phohate perpmol of subunit is introduced, The standard reaction mixture for the phosphorylation of bringing the total phosphate content up to about 1 ;tmol/jsmol of subunit. This phosphorylation of the enzyme in vitro is ac- purified phenylalanine hydroxylase contained (unless otherwise companied by a 2.6-fold increase in hydroxylase activity when specified) in a total volume of 0.1 ml at 300:10 gmol of potas- the activity is assayed in the presence of . sium phosphate, pH 6.8; 1 gmol of MgCl2; 0.1 gmol of ATP; Partial proteolytic digestion of phenylalanine hydroxylase, 0.001 gmol of 3':5'-cyclic AMP; 20-30 ,ug of phenylalanine which previously had been shown to activate the enzyme 20- hydroxylase; 1-2 ,gg of protein kinase. When [,yJ3P]ATP (New to50-fold [Fisher, D. B. & Kaufman, S. (1973)J. BioL Chem. 248, England Nuclear, Boston, Mass.) was used in the phosphoryl- 4345-4353], removes almost all of the phosphate from the en- ation reaction, protein-bound 32p was determined by adding zyme. aliquots to 10% trichloroacetic acid and collecting the precip- Rat liver phenylalanine hydroxylase [phenylalanine 4-mo- itate by centrifugation. The precipitate was dissolved in 0.1 M nooxygenase; EC 1.14.16.1; L-phenylalanine, tetrahydropter- NaOH and immediately reprecipitated by the addition of tri- idine:oxygen oxidoreductase (4-hydroxylating)] can be acti- chloroacetic acid. This was repeated once more and the pellet vated 20- to 50-fold when treated with phospholipids, like ly- was dissolved in 0.1 M NaOH. The radioactivity of an aliquot solecithin, or by partial proteolysis with chymotrypsin (1). These was determined in 10 ml of Aquasol (New England Nuclear) activations are only seen when the hydroxylase activity is in a Beckman liquid scintillation spectrometer. measured with the natural cofactor, tetrahydrobiopterin. With a synthetic cofactor, 6,7-dimethyltetrahydropterin, there is only RESULTS AND DISCUSSION about a 15% increase in activity due to these treatments. Purified rat liver phenylalanine hydroxylase is activated when In the present paper we describe results of experiments on preincubated under conditions that lead to phosphorylation. another method of increasing phenylalanine hydroxylase ac- This increase in activity is seen only when phenylalanine hy- tivity that is also seen only in the presence of tetrahydrobiopt- droxylase is assayed with the natural cofactor, tetrahydro- erin. The purified enzyme is activated 2- to 3-fold by treatment biopterin, as shown in Table 1. When the phenylalanine hy- under phosphorylating conditions with a 3':5'-cyclic-AMP- droxylase activity is measured with the synthetic cofactor, stimulated protein kinase and ATP. Activation is accompanied 6,7-dimethyltetrahydropterin, there is no change after phos- by a transfer of 32P from ['y-32P]ATP to the hydroxylase. Our phorylation. Phenylalanine hydroxylase activity in rat liver present results also show that purified phenylalanine hydrox- extracts is also increased when incubated under phosphorylating ylase contains protein-bound phosphate, but the amount is less conditions.t than 1 mole/mole of hydroxylase subunit. Phosphorylation of In Fig. 1 are shown the results of treating purified phenyl- the enzyme in vitro raises the amount of phosphate bound to alanine hydroxylase under phosphorylating conditions with the enzyme to a total of one mole per mole of hydroxylase ['y-32P]ATP. The phenylalanine hydroxylase activity, when subunit. measured on aliquots after various times of incubation, reached a maximum increase in activity of about 2.6-fold after 50 min, MATERIALS AND METHODS while incorporation of 32P into protein was still increasing. Rat liver phenylalanine hydroxylase was purified through the Control incubations (minus ATP) show a 20-30% decline in Sephadex G-200 step as previously described. In some of the activity after 1 hr of preincubation at 300 (data not shown). The experiments, this enzyme fraction was then purified further maximum incorporation of 32P into phenylalanine hydroxylase by using peak fractions from a 5 to 20% sucrose gradient (2). was always less than 1 mole/mole of 50,000 molecular weight Phenylalanine hydroxylase, purified on a sucrose gradient, gave subunit, and typically was about 0.7 (see below). a single band on sodium dodecyl sulfate electrophoresis. Phe- One possible explanation for the lack of stoichiometry in the nylalanine hydroxylase activity was measured either spectro- phosphate incorporation experiments is that phenylalanine photometrically (3) or by following the conversion of [14C]- hydroxylase, as isolated and purified, is already partially phenylalanine to ['4C]tyrosine (4). The components of the assay phosphorylated. In order to investigate this possibility, the amount of endogenous protein-bound phosphate was deter- * Present address: Institut National de la Sante et de la Recherche Medicale, Facult6 de Medecine (Pasteur), Chemin de Vallombrose, t J. P. Abita, S. Milstien, N. Chang, and S. Kaufman (1976), submitted 06034 Nice Cedex, France. for publication. 1591 Downloaded by guest on September 27, 2021 1592 Biochemistry: Milstien et al. Proc. Natl. Acad. Sci. USA 73 (1976)

Table 1. Requirements for activation of purified rat liver 'v phenylalanine hydroxylase under m phosphorylating conditions z cn Z 5; k Phenylalanine hydroxylase E Z _X activity (nmol of DPNH x 'z m 3 Z oxidized/minm mg of oE *sO0 protein) II ZZn-X M-- Z I tetra- C0 With With 6,7- 0I hydro- dimethyltetra- I Additions biopterin hydropterin

None 78 750 0 10 20 30 40 50 60 70 80 90 100 Complete phosphorylating 159 750 MINUTES system FIG. 1. Phosphorylation and activation of purified phenylalanine minus ATP 78 748 hydroxylase. Phenylalanine hydroxylase, purified on a sucrose gra- dient, was incubated at 300 with [-y-32PJATP as described in Materials minus 3: 5'-cyclic AMP 78 and Methods. Aliquots were removed at the times indicated, and minus MgCl2 80 trichloroacetic acid-precipitable 32P radioactivity (-) and phenylal- minus protein kinase 81 751 anine hydroxylase activities (0) were measured in the presence of tetrahydrobiopterin (14 gM). Purified phenylalanine hydroxylase (123 Mg) was incubated for 30 min at 300 in a total volume of 0.2 ml containing 0.1 M po- tassium phosphate (pH 6.8). The complete phosphorylating system located on this peptide chain. When total protein-bound contained in addition: 10-2 M MgCl2, 2.5 x 10-4 M ATP, 5 x 10-6 M cyclic AMP, and 5 gg of beef heart muscle protein kinase. The phosphate was measured on the chymotrypsin-activated, phenylalanine hydroxylase activity was then determined with phosphorylated hydroxylase (isolated from the peak fractions either tetrahydrobiopterin or 6,7-dimethyltetrahydropterin as a of a sucrose gradient) it was found to contain 0.08 + 0.03 (mean cofactor. ± standard error of the mean, five determinations, 6-30 ,ug of protein per assay) mole of phosphate per mole of hydroxylase mined in five separate preparations of purified phenylalanine subunit. Since the control enzyme used in this experiment had hydroxylase. 0.36 mole of protein-bound phosphate per mole of hydroxylase The average phosphate content per 50,000 molecular weight subunit, it is evident that at least 78% of the endogenous phos- subunit was 0.31 mole, with a range of values for the five phate is also located on the peptide chain(s) that is cleaved from preparations of 0.23-0.42 mole/mole of subunit. The variability the hydroxylase during activation by limited proteolysis. in the phosphate content could be due to variable losses of The above indication that there is only a single phosphoryl- phosphate during the purification procedure.t ation site on the hydroxylase subunit that is involved in this The kinetic properties of phenylalanine hydroxylase acti- modulation of the catalytic activity, together with the finding vated by phosphorylation are similar to those of the "native" that a 3-fold increase in the amount of phosphate bound to the enzyme. The apparent Km values for phenylalanine and te- enzyme (from 0.31 to 1.0 mole/mole of enzyme) is accompa- trahydrobiopterin are essentially unchanged after phospho- nied by a roughly proportionate increase in hydroxylase activity rylation.t (2.6-fold, see Fig. 1), suggests that the hydroxylase activity of Partial proteolysis and lysolecithin activations of phenylal- the completely dephosphorylated enzyme is very low. Whether anine hydroxylase involve changes in conformation of the en- this form of the enzyme is completely devoid of hydroxylase zyme (J. P. Abita and S. Kaufman, manuscript in preparation). activity, however, remains to be determined. It was of interest to determine if these activation effects and The present results suggest that the activity of rat liver phe- activation by phosphorylation involved similar changes in the nylalanine hydroxylase, like that of pig liver pyruvate kinase enzyme. Phenylalanine hydroxylase, phosphorylated and ac- (7), rabbit muscle phosphorylase a (8), rabbit muscle phos- tivated as described in the legend to Fig. 1, was stimulated phorylase kinase (9), yeast glucose-6-phosphate-dependent 8.5-fold when assayed in the presence of 0.5 mM lysolecithin, glycogen synthetase (10), and rabbit muscle glycogen synthetase whereas the control enzyme (minus ATP) was stimulated 17- (11), is regulated by specific phosphorylation at a site that is fold. These results show that the two activating effects are not readily cleaved by proteolytic enzymes without causing inac- additive and suggest that phosphorylation and lysolecithin may tivation. Indeed, with most of these enzymes, partial proteolysis be affecting the same region of the enzyme. When phenylala- results in dramatic increases in activity. nine hydroxylase was phosphorylated with ['y-32P]ATP under It has been reported by Barrenger et al. that rat liver phe- the conditions described in Table 1 and then further activated nylalanine hydroxylase can be separated into three different by partial proteolysis with chymotrypsin, more than 90% of the isozymes by chromatography on calcium phosphate columns 32P was removed from the enzyme. In addition, the maximum (12). It is possible that these different forms of the enzyme are activity reached after chymotrypsin treatment was the same separated on the basis of different degrees of phosphorylation for the control as for the phosphorylated enzyme. This result since our purified phenylalanine hydroxylase contains an av- shows that the site on the hydroxylase that is phosphorylated erage of 0.31 mole of phosphate bound per mole of enzyme under these in vitro conditions is located on the peptide chain(s) subunit. Since purified phenylalanine hydroxylase exists as a that is cleaved by treatment with chymotrypsin. It was im- dimer of two slightly different 50,000 molecular weight sub- portant to determine whether the endogenous phosphate is also units (13), partial phosphorylation could give rise to a hetero- geneous population of enzyme molecules. * Rat liver extracts also contain phenylalanine hydroxylase phosphatase It is of interest that purified bovine caudate tyrosine hy- activity. S. Milstien, E. Jedlicki, and S. Kaufman, unpublished ob- droxylase, an enzyme that catalyzes a similar tetrahydropt- servations. erin-dependent hydroxylation reaction,is also activated when Downloaded by guest on September 27, 2021 Biochemistry: Milstien et al. Proc. Nati. Acad. Sci. USA 73 (1976) 1593

it is treated under phosphorylating conditions with a cyclic- 5. Gilman, A. G. (1970) Proc. Natl. Acad. Sci. USA 67,305-312. AMP-dependent protein kinase (14). In contrast to the present 6. Ames, B. N. (1966) in Methods in Enzymology, eds. Neufeld, case with phenylalanine hydroxylase, however, the activation E. F. & Ginsburg, V. (Academic Press, New York), Vol. 8, pp. of is not accompanied by the incorporation 115-118. 7. Bergstrom, G., Ekman, P., Dahlkvist, U., Humble, E. & Eng- of detectable amounts of 32P from [y-32P]ATP into the hy- strom, L. (1975) FEBS Lett. 56,288-291. droxylase (14). The mechanism of its activation, therefore, re- 8. Fisher, E. H., Graves, D. J., Crittendon, E. R. S. & Krebs, E. G. mains obscure. (1959) J. Biol. Chem. 234, 1698-1704. 1. Fisher, D. B. & Kaufman, S. (1973) J. Biol. Chem. 248, 45- 9. Huston, R. B. & Krebs, E. G. (1968) Biochemistry 7,2116-2122. 4353. 10. Huang, K. P. & Cabib, E. (1974) J. Biol. Chem. 249,3858-3861. 2. Kaufman, S. & Fisher, D. B. (1970) J. Biol. Chem. 245,4745- 11. Nimmo, H. G. & Cohen, P. (1974) FEBS Lett. 47, 162-166. 4750. 12. Barrenger, J. A., Geiger, P. J., Huzino, A. & Bessman, S. P. (1972) 3. Kaufman, S. (1970) in Methods in Enzyrnology, eds. Tabor, H. Science 175,903-905. & Tabor, C. W. (Academic Press, New York), Vol. 17A, pp. 13. Kaufman, S. (1971) Adv. Enzymol. 35,245-319. 603-609. 14. Lloyd, T. & Kaufman, S. (1975) Biochem. Biophys. Res. Com- 4. Milstien, S. & Kaufman, S. (1975) J. Biol. Chem. 250,4777-4781. mun. 66,907-913. Downloaded by guest on September 27, 2021