[6] Bovine retinal nucleoside diphosphate kinase: Biochemistry and molecular cloning

[6]

BOVINERETINALNDP KINASE

87

[6] B o v i n e R e t i n a l N u c l e o s i d e D i p h o s p h a t e K i n a s e : Biochemistry and Molecular Cloning

By NAJMOUTIN G. ABDULAEV,DMITRI L. KAKUEV, and KEVIN D. RIDGE Introduction Signal transduction in vertebrate photoreceptor cells begins with absorption of light by the visual pigment rhodopsin and culminates in the closure of ion channels on the plasma membrane. Consecutive binding and hydrolysis of several guanine nucleotides support the flow of information between these two events. The high levels of GTP required for G-protein activation and cGMP synthesis in the retina are likely to be supported by nucleoside diphosphate kinase (NDP kinase). The activity of the enzyme has been measured in retinal layers and outer segments, 1'2 but the enzyme itself has not been biochemically or structurally characterized. NDP kinase presumably constitutes an integral part of the cGMP cycle (cGMP ~ 5'-GMP --~ GDP ~ GTP --->cGMP) and catalyzes the phosphorylation of nucleoside diphosphates to nucleoside triphosphates by a ping-pong mechanism involving a high-energy phosphorylated enzyme intermediate. The highenergy phosphate is usually supplied by ATP. 3 Although the key role of guanine nucleotides in visual transduction is firmly established, what remains unclear is how the phototransduction cascade blends with the biochemical pathways of nucleotide metabolism. The purpose of this chapter is to highlight the role of NDP kinase in visual transduction and to provide systematic approaches for the study of its biochemical properties. 4 The focus is on the purification of the enzyme from bovine retina, its functional characterization, the cloning of two distinct isoforms from a bovine retinal library, and their heterologous expression in Escherichia coli.

1 S. J. Berger, G. W. DeVries, J. G. Carter, D. W. Schulz, P. N. Passonneau, O. H. Lowry, and J. A. Ferrendelli, J. Biol. Chem. 255, 3128 (1980). 2 S. W. Hall and H. Kuhn, E u r J. Biochem. 161, 551 (1986). 3 R. P. Agarwal, B. Robison, and R. E. Parks, Jr., Methods Enzymol. 51, 376 (1978). n N. G. Abdulaev, G. N. Karaschuk, J. E. Ladner, D. L. Kakuev, A. V. Yakhyaev, M. Tordova, I. O. Gaidarov, V. I. Popov, J. H. Fujiwara, D. Chinchilla, E. Eisenstein, G. L. Gilliland, and K. D. Ridge, Biochemistry 3"/, 13958 (1998).

METHODS IN ENZYMOLOGY, VOL. 316

Copyright © 2000 by Academic Press All rights of reproduction in any form reserved. 0076-6879/00 $30.00

88

PHOTORECEPTOR PROTEINS

[6]

Purification and Characterization of Bovine Retinal Nucleoside Diphosphate Kinase

Purification of Nucleoside Diphosphate Kinase from Bovine Retina The first stage of the NDP kinase purification procedure represents a modified guanylate kinase isolation procedure from Hall and Kuhn. z Two hundred frozen bovine retina (W. L. Lawson, Lincoln, NE) are thawed and suspended in 170 ml of 10 mM (Na,K,H)PO4, pH 7.6, containing 0.2 mM MgC12, 0.2 mM EGTA, 0.2 mM phenylmethylsulfonyl fluoride (PMSF), and 0.02% (w/v) NAN3. After stirring for 30 rain at 4° in the light, the concentration of NaC1 and MgCI2 is adjusted to 150 and 4 raM, respectively. The suspension is stirred for another 30 min at 4° and the insoluble material is removed by centrifugation at 30,000g for 1 hr at 4°. The supernatant is centrifuged again at 100,000g for 30 min at 4° to remove trace amounts of membranes. An equal volume of (NH4)2SO4 solution saturated at 60 ° is added to the supernatant and stirred for 2 hr at 4°. The precipitate is removed by centrifugation at 40,000g for 40 min at 4° and the supernatant is brought to 75% saturation with (NH4)2SO4.After stirring overnight at 4 °, the solution is centrifuged and the pellet, which contains NDP kinase, is resuspended in 10 mM Tris-HC1, pH 7.3, containing 2 mM MgC12, 0.1 mM EDTA, 1 mM dithiothreitol (DTT), and 300 mM NaC1 (TMED buffer). Any insoluble material is removed by centrifugation at 100,000g for 30 min at 4°. The clear supernatant is applied to a Blue Sepharose CL-6B (Pharmacia Biotech, Piscataway, NJ) column (15 x 1.5 cm) equilibrated with TMED buffer. The column is extensively washed with TMED and the NDP kinase is eluted with TMED containing 2 mM GTP. The yield of purified enzyme from this procedure is typically ~1 mg and the specific activity is -763 units/mg (Table I). No appreciable difference is observed in the yield of NDP kinase or in its functional properties after purification is performed from light- or dark-adapted retina. Importantly, purified NDP kinase is free of both guanylate kinase and adenylate kinase activities. The purified enzyme exhibits a doublet on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), with the molecular masses of the two polypeptides -17.5 and 18.5 kDa (Fig. 1A, lane 3). Both of these polypeptides are barely visible in the crude retinal extract or in the pellet obtained after 75% saturation with (NI--I4)2804(Fig. 1A, lanes 1 and 2, respectively), further emphasizing their relatively low abundance compared with other soluble retinal proteins. Immunoblot analysis of the purified enzyme using anti-NDP kinase polyclonal antibodies also shows the two polypeptides (Fig. 1B). The isoelectric focusing pattern of NDP

[6]

BOVINE RETINALNDP KINASE

89

TABLE I PURIFICATION OF

Fraction Crude extract Supernatantc Pelletd Blue Sepharose

NDP KINASE FROM BOVINE

Total activity" Protein (units) (mg) 6909.6 6828.4 5316 839.5

3840 1015 595 1.1

RETINA

Specificactivity Purification Recoverb (units/mg) (-fold) (%) 1.80 6.73 8.93 763.18

1 3.7 5 424

100 98.8 76.9 12.1

a One unit of activity is defined as the amount of enzyme required to convert 1 txmol of CDP and ATP to CTP and ADP per minute at 25° in a coupled pyruvate kinase/lactate dehydrogenase assay. b Recovery relates the total activity and extent of purification to the specificactivity. c The values shown correspond to the supernatant obtained after 50% (NH4)2804precipitation. dThe values shown correspond to the pellet obtained at 75% (NH4)zSO4precipitation. kinase shows seven or eight protein bands with a pI range from 7.4 to 8.2 (Fig. 1C). Such a plurality of protein bands can be explained by the existence of N D P kinase oligomers composed of different proportions of the constitutive subunits and/or by varying degrees of phosphorylation within these oligomers.

Quaternary Structure Analysis The molecular mass of N D P kinase in solution can be determined by comparing its migration with known protein standards by analytical gelfiltration chromatography (Fig. 2). Gel filtration of native (or recombinant) N D P kinase is performed on a Superose-12 H R (Pharmacia Biotech) column (30 × 1 cm) attached to a B i o C A D Sprint perfusion chromatography workstation (PerSeptive Biosystems, Framingham, MA) or a fast protein liquid chromatography (FPLC) system (Pharmacia Biotech). The column is equilibrated with T M E D buffer containing bovine serum albumin (1 mg/ml) at a flow rate of 0.2 ml/min to block nonspecific binding. The column is subsequently washed with T M E D buffer and calibrated several times with 1-mg/ml solutions of the following molecular mass protein standards (Sigma, St. Louis, MO): rabbit muscle aldolase (158 kDa; Fig. 2a), rabbit muscle lactate dehydrogenase (140 kDa; Fig. 2b), bovine serum albumin (66 kDa; Fig. 2c), ovalbumin (45 kDa; Fig. 2d), chymotrypsinogen A (25 kDa; Fig. 2e), and cytochrome c (12.5 kDa; Fig. 2f). A standard curve is constructed by plotting Kay versus log Mr. Kav is calculated from the elution volumes Ve, from the total column volume Vt, and from the void volume V0 using Eq. (1):

90

PHOTORECEPTOR PROTEINS

A

[6]

B

kDa

C pl 7.3 5

43.0 29.2 18.8 16.5 6.4

-

~

8.15 1 2

3

FIo. I. Electrophoretic analysis of NDP kinase. (A) Coomassie blue stain of proteins in crude retinal extract (lane 1) and after 75% saturation with (NH4)2SO4 (lane 2), and purified NDP kinase (lane 3) separated by SDS-PAGE. (B) Immunoblot analysis of purified NDP kinase separated by SDS-PAGE. The polypeptides were detected with the polyclonal antiNDP kinase antibody, and visualized by chemiluminescence. (C) Coomassie blue stain of purified NDP kinase separated by isoelectric focusing. Positions of molecular size standards for (A) and (B) and pI standards for (C) are shown on the left.

Kay = Ve - Vo/Vt - Vo

(1)

V0 is determined by measuring the eluted volume of blue dextran, and Vt is determined by measuring the eluted volume of tryptophan. This analysis yields a molecular mass of 96 _+ 2 kDa for NDP kinase (Fig. 2), which closely approaches that of a hexamer. Notably, this subunit stoichiometry is characteristic of virtually all eukaryotic NDP kinases? Carbohydrate A n a l y s i s

The rather small difference in apparent molecular mass between the two NDP kinase polypeptides (~1 kDa) raises the possibility that they differ in some posttranslational modification(s). In fact, both NDP kinase 5 A.-M. Gilles, E. Presecan, A. Vonica, and I. Lascu, J. Biol. Chem. 266, 8784 (1991).

[6]

BOVINe RETINALNDP KINASE

91

5.4

5.2

L/ 4.8

o

d 4,6

4.4

4.2

4 I

0,15

I

I

I

I

I

0.2

0.25

0.3

0.35

0.4

\

I

0.45

Kay FIG. 2. Size-exclusion chromatography of retinal NDP kinase. The arrow indicates the position of NDP kinase relative to the molecular weight protein standards (details in text).

92

PHOTORECEPTOR PROTEINS

[6]

polypeptides appear to contain equivalent amounts of Gal, Man, GIcNAc, Fuc, and GalNAc saccharides, which are characteristic of O-linked glycosylation. The carbohydrate content is analyzed by separating the purified NDP kinase subunits by SDS-PAGE, electroblotting them onto poly(vinyl difluoride) membranes (Millipore, Bedford, MA), staining the proteins with Ponceau S, and excising the individual polypeptides with a razor blade. The membrane strips are destained with 1% acetic acid, washed extensively with deionized water, and dried. Acid hydrolysis of oligosaccharides in the separated NDP kinase polypeptides is performed as described. 6 The oligosaccharides are coupled with 7-amino-4-methylcoumarin, 7 and the monosaccharide fluorescent derivatives are separated on an Ultrasphere ODS (Beckman-Coulter, Fullerton, CA) column (22 × 0.21 cm) with a mobile phase consisting of 6.8% (v/v) 2-propanol, 3.5% (v/v) acetonitrile, and 0.01% (v/v) trifluoroacetic acid at a flow rate of 0.2 ml/min. The results of this analysis also show that the total content of carbohydrate in each polypeptide accounts for 2-3% of protein weight. Consequently, retinal NDP kinase can be referred to as a glycoprotein with a low content of oligosaccharides. Although it is possible that some properties of this enzyme could be explained by the presence of carbohydrates, it is evident that the difference in molecular mass of the NDP kinase subunits may not be attributed to a distinction in oligosaccharide content.

Functional Assay of Nucleoside Diphosphate Kinase NDP kinase is assayed spectrophotometrically using a coupled pyruvate kinase-lactate dehydrogenase enzyme system. 2'8 The assays are done in a 1-ml reaction mixture containing 50 mM Tris-HC1 (pH 7.6), 5 mM MgC12, 0.05 mM KC1, 0.1 mM phosphoenolpyruvate, 0.5 mM ATP, 0.1 mM TDP, N A D H (0.1 mg/ml), 2 units of pyruvate kinase, and 2.5 units of lacate dehydrogenase. The reaction is initiated by addition of 0.5-5/zg of NDP kinase at 25 °. Monitoring the decrease in absorbance at 334 nm follows the oxidation of NADH, which reflects ADP formation by NDP kinase. The specific activity of 1 unit of enzyme is defined as the turnover of 1 tzmol of substrate in 1 min per milligram of protein. To measure the apparent Michaelis constant (Kin) and maximum velocity (Vmax) for the diphosphate nucleotides CDP and TDP, a constant ATP concentration equal to 2 mM is used. For determination of the Km and Vmax for the 6 H. Takemoto, S. Hase, and T. Ikenaka, A n a l Biochem. 145, 245 (1985). 7 A. Ya. Khorlin, S. D. Shiyan, V. A. Markin, V. V. Nasonov, and M. N. Mirzayanova, Bioorg. Khirn. 12, 1203 (1986). 8 R. E. Parks and R. P. Agarwal, in "The Enzymes" (P. D. Boyer, ed.), Vol. 8, p. 307. Academic Press, New York, 1973.

[6]

BOVINERETINALNDP KINASE

93

triphosphate nucleotides GTP and ATP, 1.4 mM CDP is used as the fixed substrate. The Km and Vmax values are estimated from double-reciprocal plots by the method of Florini and Vestling. 9 This method allows the determination of kinetic constants for two-substrate enzyme systems. Like other eukaryotic NDP kinases, retinal NDP kinase also shows broad specificity for the nucleotide substrate. The Km of the enzyme for ATP (104/xM) is the lowest among the nucleotides tested, while that for GTP is only about three times higher. Unfortunately, because of this relatively small difference in the Km values, it is not possible to obtain accurate and reproducible kinetic measurements using ATP and GDP as the donor and acceptor substrates, respectively.

Molecular Cloning a n d F u n c t i o n a l E x p r e s s i o n of Bovine Retinal Nucleoside D i p h o s p h a t e Kinase

Cloning of Nucleoside Diphosphate Kinase Isoforms from Bovine Retinal Library Direct sequence analysis of NDP kinase does not show cleavage of any amino acids, suggesting that the NH2-terminal residue in each of the polypeptides is blocked. The nature of this blocking group remains to be determined. However, digestion of NDP kinase with trypsin after heat denaturation at elevated pH allows exhaustive cleavage of the protein and provides facile peptide purification and sequencing. Using this sequence information and taking into account the codon usage of several known retinal proteins, oligonucleotides are synthesized and used to screen a bovine retinal cDNA library. The tryptic peptides are prepared by heat denaturing purified NDP kinase (1 mg) in 100 mM NaHCO3, pH 9.0, at 95 ° for 10 min and then subjecting them to trypsin digestion (1:50, trypsin:NDP kinase) at 37° for 6 hr. The digest is applied to a reversedphase high-performance liquid chromatography (HPLC) Zorbax ClS column (25 × 0.5 cm; Waters, Milford, MA) and eluted with a linear gradient of 0-70% (v/v) acetonitrile containing 0.1% (v/v) trifluoroacetic acid. Several peak fractions are collected and further purified on the same column. The peptides are sequenced by Edman degradation on a liquid-phase sequencer (Applied Biosystems, Foster City, CA). Two oligonucleotides based on the sequences of the tryptic peptides, as well as two additional probes corresponding to the most conserved sequences of NDP kinases

9j. R. Florini and C. S. Vestling, Biochim. Biophys. Acta 25, 575 (1957).

94

PHOTORECEPTORPROTEINS

[61

from different sources, are synthesized. The oligonucleotide probes are 5'-end labeled with 32p using T4 polynucleotide kinaseJ ° A bovine retinal cDNA library in bacteriophage A-ZAP (provided by M. Applebury, Massachusetts Eye and Ear Infirmary, Boston, MA) is used for the screening. To screen the library, 4 × 106 bacteriophage are plated, and filter replicas are prepared using nitrocellulose filters (Millipore). The hybridization protocol is adapted from Sambrook et al. lo Briefly, the procedure is carried out in the minimal volume of the hybridization buffer with addition of purified radioactively labeled oligonucleotide probe for 1 hr at 67° and then incubated overnight with slow cooling and shaking. The filters are washed in 4× SSC solution ( l x SSC is 0.15 M NaC1 plus 0.015 sodium citrate) containing 0.1% SDS. Autoradiography is performed using Hyperfilm-flmax film (Kodak, Rochester, NY) with an intensifying screen for 48 hr at - 7 0 °. Isolation of phage DNA and subcloning of cDNA inserts in the pBluescript M13 vector (Stratagene, La Jolla, CA) are done as described. 11 The plasmids are amplified in E. coli XL-1 Blue, the DNA is purified, and the inserts are sequenced. The results of this screening yield two NDP kinase clones, termed NBR-A and NBR-B (they are designated as A and B in analogy with NDP kinases from other sources)J 2 The clone containing NBR-A has a 696-bp-long insert while that containing NBR-B has an 863-bp insert (Fig. 3). Sequence analysis of NBR-A and NBR-B shows considerable base differences in both the 5' and 3' noncoding regions, suggesting that they are indeed isoforms of the enzyme. Further, NBR-A and NBR-B differ from each other in four codons. Two nucleotide substitutions are responsible for the amino acid diversity of the NDP kinase isoforms: Ile-21 and His-135 in NBR-A are replaced by methionine and arginine residues, respectively, in NBR-B (Fig. 3). Two additional codon changes, at Pro-72 and Gly-106, do not result in amino acid substitutions. The deduced amino acid sequences of the NDP kinase isoforms each consist of 152 residues and the calculated molecular masses are 17,262 kDa for NBR-A and 17,299 kDa for NBR-B. Expression and Purification of NBR-A and NBR-B Nucleoside Diphosphate Kinases

The cDNAs are ligated into the pALTER-Ex2 vector (Promega, Madison, WI) at the unique EcoRI restriction site in the multiple cloning region. 10j. Sambrook, E. F. Fritsch, and T. Maniatis, "Molecular Cloning: A Laboratory Manual," 2rid Ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1989. 11 j. M. Short, J. A. Sorge, and W. D. Huse, Nucleic Acids Res. 16, 7583 (1988). 12The nucleotide sequence data have been deposited with the EMBL Sequence Data Bank and are available under accession number X92956 for N B R - A and X92957 for NBR-B.

[6]

BOVINE RETINALNDP KINASE

A

95

NBR-A ATT (I) CCA (P) GGG (G) CAC (H)

I

I

I

I

1 54

513

696

NBR-B ATG (M) CCC (P) GGT (G) CGC OR)

I

1

I

I

I

208

667

863

B ERTF I/M AMKF

Y

I 1

I

I

VKTG

H

H/R

Y

I

I

I 152

FIG. 3. Schematic representation of retinal NDP kinase nucleotide and amino acid sequences. (A) The codons and corresponding amino acids (single-letter code) in NBR-A that differ from those in NBR-B are highlighted. (B) The positions of amino acid diversity between NBR-A and NBR-B, the two potential O-glycosylation sites, the proposed ATP-binding region of the cyclicnucleotide-dependent protein kinases, and the active site histidine residue are highlighted. P r o p e r insertion and orientation of N B R - A and N B R - B in the vector are confirmed by restriction enzyme analysis. Escherichia coli JM109 (DE3) is transformed with the vector p A L T E R - E x 2 harboring the N B R - A and N B R - B cDNAs. T h e transformation mixtures are spread on L B - a g a r plates containing tetracycline and after 16-18 hr at 37 °, colonies are isolated and cultured overnight at 37 ° in LB broth containing tetracycline. T h e overnight cultures are used to inoculate 500 ml of LB broth containing tetracycline

96

PHOTORECEPTORPROTEINS

161

in a 1 liter-capacity shake flask. After further growth at 37 ° for 16 hr, the cells are harvested by centrifugation at 1500g for 5 min, and disrupted in a French pressure cell (800-900 lb/in2). The supernatant is used for purification of the recombinant enzymes. The purification protocol for retinal NDP kinase can be easily adapted for preparation of NBR-A and NBR-B expressed in E. coll. Typically, the yields of purified NBR-A and NBR-B are 18-20 mg from a 500-ml shake flask culture, and the specific activities are in the range of - 8 5 7 and ~1063 units/mg, respectively. Similar to retinal NDP kinase, the Km of the expressed enzymes for ATP is the lowest among the nucleotides tested, with GTP showing a threefold higher difference. NBR-A and NBR-B each show a single polypeptide chain of - 1 7 kDa on S D S - P A G E followed by immunoblot analysis using anti-NDP kinase polyclonal antibodies raised against the retinal enzyme (Fig. 4, lanes 2 and 3). Similar to retinal NDP kinase, the isoelectric focusing patterns of NBR-A and NBR-B show six or seven major protein bands. However, the pI range for NBR-A is 8.0-8.5 while that for NBR-B is 6.5-7.5. Analytical gel filtration of the NBR-A and NBR-B isoforms shows molecular masses of 97 and 95 kDa, respectively. Like retinal NDP kinase, this is suggestive of a hexameric arrangement for the expressed enzymes.

Comments To understand the biochemical events controlling the synthesis of guanine nucleotides involved in visual transduction, the isolation and characterization of retinal enzymes involved in nucleotide metabolism are desirable. Bovine retinal NDP kinase, like those from other sources, 13-15 shows two distinct protein bands with apparent molecular masses of 17.5 and 18.5 kDa when analyzed by S D S - P A G E . In contrast, both NBR-A and NBR-B show a single polypeptide chain of - 1 7 kDa. It is interesting that the correspondence between separate polypeptides and the products of gene isoforms has been shown only for human erythrocyte NDP kinase. 5 Both the S D S - P A G E and cDNA cloning results indicate that there exists at least two NDP kinase isoforms in bovine retina, as is the case for other higher eukaryotes. 16,17 However, NBR-A and NBR-B 13E. Presecan, A. Vonica, and I. Lascu, FEBS Lett. 250, 629 (1989). 14N. Kimura and N. Shimada,J. Biol. Chem. 263, 4647 (1988). 15j. A. Nickerson and W. W. Well, J. Biol. Chem. 259, 11297 (1984). 16N. Shimada, N. Ishikawa, Y. Mukanata, T. Toda, K. Watanabe, and N. Kimura, J. Biol. Chem. 268, 2583 (1993). 17T. Urano, K. Takamiya, K. Furukawa, and H. Shiku, FEBS Lett. 3tl9, 358 (1992).

[61

BOVINERETINALNDP KINASE

97

kDA

43.0

29.2

18.8 m 16.5 m

6°4 1

2

3

FIG. 4. Immunoblot analysis of purified bovine retinal NDP kinase (lane 1), purified NBR-A (lane 2), and purified NBR-B (lane 3). Positions of molecular size standards are shown on the left.

are remarkably similar in their coding nucleotide and amino acid sequences, with identities of 99.1 and 98.8%, respectively. Therefore, the -1-kDa difference in apparent molecular weight between the two retinal NDP kinase polypeptides suggests that they are either subject to different, as yet unknown, posttranslational modifications or bovine retina contains still another NDP kinase isoform(s). Analytical gel filtration of the retinal and expressed NDP kinases suggests that the enzymes exist as a hexamer. This is further substantiated by the three-dimensional structures of NBR-A and NBR-B. 4J8 18j. E. Ladner, N. G. Abdulaev, D. L. Kakuev, M. Tordova, K. D. Ridge, and G. L. Gilliland, Acta Crystallogr. D 55, 1127 (1999).

98

PHOTORECEPTORPROTEINS

[61

TABLE II AMINO ACID SEQUENCEIDENTITYBETWEENNDP KINASESFROM VARIOUSSOURCESa Source

1. awd, Drosophila melanogaster NDP kinase, Dictyostelium discoideum 3. NDP kinase, Myxococcus xanthus 2.

4. 5. 6. 7. 8. 9. 10.

NBR-A, bovine retina NBR-B, bovine retina NDP kinase o~, rat liver NDP kinase/3, rat liver p18, rat mucosal mast cells nm23-H1, human nm23-H2, human a

1

2

3

4

5

47.2 47.2 44.1 44.1 47.2 46.0 46.0

98.8 85.8 88.9 87.0 90.1 87.0

85.8 88.9 87.0 89.5 87.0

6

7

8

9

60.0 48.8

43.1

74.4 47.4 47.4 73.8 76.9 78.1 78.1

58.1 58.1 55.6 58.1 58.1 61.9 59.4

90.1 9 1 . 9 84.5 86.4 92.0 87.0 93.2 86.4 9 5 . 0 89.9

Amino acid sequences within the coding region are compared and the sequence pair distances determined by the CLUSTAL method.

Both subunits of the enzyme from bovine retina contain - 2 - 3 % (w/w) carbohydrate that consists of Gal, Man, GIcNAc, Fuc, and GalNAc saccharides. All eukaryotic NDP kinases contain a potential O-glycosylation site (ERTF, amino acid residues 5-8) and another consensus O-glycosylation motif, VKTG (amino acid residues 84-87), is preserved in proteins from mammals and Drosophila melanogaster. These potential glycosylation sites are also on the outside of the molecule and readily accessible for modification. In fact, these sites are adjacent to one another; the distance between the C~ of Glu-5 and Val-84 is 7 m. 4'18 Differences in the extent of amino acid residue phosphorylation are implied from the isoelectric focusing experiments on both the retinal and expressed enzymes. The His-ll8 residue, a phosphorylation site on NDP kinase, is present in all known NDP kinases. NBR-A and NBR-B both contain an AMKF sequence (amino acid residues 37-40) that is nearly conserved as the ATP-binding region in cyclic nucleotide-dependent protein kinases. Lys-39 in this region is invariant in nearly all NDP kinases and is essential for catalytic activity. 19 The results of comparative sequence analysis (Table II) show that both the NBR-A and NBR-B sequences can be aligned rather well with those of human nm23-H1 protein (90.1 and 89.5% homology, respectively) and the rat/3

19N. Kimura, N. Shimada, K. Nomura, and K. Watanabe, J. Biol. Chem. 265, 15744 (1990).

[61

BOVINERETINALNDP KINASE

99

isoform of NDP kinase (88.9% homology for both isoforms). Almost all sequences are homologous to the sequence of NDP kinase in question, without any internal gaps. The exceptions are the enzymes from Dictyostelium discoideum and Myxococcus xanthus. Kinetic studies of retinal and expressed NDP kinases showed that the enzymes display low substrate specificity, as illustrated for other NDP kinases. However, NDP kinase can function in a low ATP-containing environment (Kin of retinal NDP kinase for ATP, -104/zM) with high velocity, supplying a sufficient level of nucleoside triphosphates in visual cells. The specific activity of retinal NDP kinase, ~763 units/mg, is considerably high when compared with other guanine nucleotide pathway enzymes. For example, the specific activity of retinal guanylate kinase is - 3 4 0 units/mg. 2 The turnover number of NDP kinase is 229 mol of CTP produced per mole of enzyme per second or 1374 mol of CTP produced per mole of enzyme per second for a hexamer. This is considerably greater than the turnover number of guanylate kinase ( - 1 3 0 mol of GDP produced per mole of enzyme per second). These findings suggest that even relatively small amounts of guanylate kinase and NDP kinase in the rod outer segment (ROS) are greater in terms of absolute activity than guanylate cyclase. 2°-22 Therefore, these two enzymes may support local cGMP requirements at the expense of ATP, which is produced abundantly in the ROS as a result of anaerobic glycolysis,23 without the need for shuttling GMP from the ROS to the rod inner segment for conversion to GTP. The biochemical and structural characterization of NDP kinase from bovine retina will allow us to investigate further its role in the regulation of the GTP supply to G proteins and other components involved in visual phototransduction. The availability of the cDNA sequences for retinal NDP kinase should allow for the overexpression of mutant forms of the enzyme in sufficient quantities for three-dimensional structure determination.

Acknowledgments We recognize the contributions of John Fujiwara and Tony Ngo in some aspects of these studies. This work was supported by the National Institute of Standards and Technology. Certain commercial materials, instruments, and equipment are identified in this manuscript

20 I. O. Gaidarov, O. N. Suslov, and N. G. Abdulaev, FEBS Lett. 335, 81 (1993). 21 I. O. Gaidarov, O. N. Suslov, T. V. Ovchinnikova, and N. G. Abdualev, Biorg. Khim. 20, 367 (1994). 22 A. Ames, T. F. Walseth, R. A. Heyman, M. Barad, R. M. Graeff, and N. D. Goldberg, J. Biol. Chem. 261, 13034 (1986). 23 S.-C. Hsu and R. S. Molday, J. Biol. Chem. 266, 21745 (1991).

100

PHOTORECEPTOR PROTEINS

[61

in order to specify the experimental procedure as completely as possible. In no case does such identification imply a recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the materials, instruments, or equipment identified is necessarily the best available for the purpose.