Protein Inhibitors of Crystal Growth

0022-534 7/89/1413-0750$02.00/0 THE JOURNAL OF UROLOGY

Vol. 141, March Printed in U.S.A.

Copyright© 1989 by The Williams & Wilkins Co.

PROTEIN INHIBITORS OF CRYSTAL GROWTH EMIL THOMAS KAISER*

AND

SUSAN CLARK BOCKt

From the Rockefeller University, New York, New York

ABSTRACT

Nephrocalcin is a urinary glycopeptide that may be a physiological inhibitor of nephrolithiasis. Monomeric nephrocalcin purified from ethylenediaminetetracetic acid-treated urine is 14,000 daltons. Compositional analyses indicate that nephrocalcin is 10 per cent carbohydrate by weight and that 25 per cent of the amino acid residues are acidic (glutamic acid, aspartic acid and 'Ycarboxyglutamic acid). Nephrocalcin binds reversibly to calcium oxalate crystals with a dissociation constant of about 0.5 µM. The high collapse pressure of nephrocalcin, 41.5 dynes per cm., measured for a monolayer at the air-water interface, suggests a highly organized structure in which hydrophilic and hydrophobic regions occupy separate regions on the surface of the inhibitor. Nephrocalcin contains the unusual amino acid, 'Y-carboxyglutamic acid. Nephrocalcin isolated from urine of stone formers and from kidney stones does not contain 'Y-carboxyglutamic acid and it has altered surface properties compared to normal nephrocalcin. The presence of the 'Y-carboxyglutamic acid modification and the ability to form stable films with high collapse pressures may be important factors enabling nephrocalcin to prevent stone formation in vivo. The blood of cold water fishes contains antifreeze glycopeptides and/or peptides to prevent it from freezing. The structure of one such antifreeze peptide and its interactions with the crystal lattice of hexagonal ice are discussed as a model for how nephrocalcin might interact with calcium oxalate-crystals and arrest their growth in urine. (J. Ural., part 2, 141: 750-752, 1989) We review the properties of nephrocalcin, a urinary glycopeptide that inhibits calcium oxalate crystal growth in vitro and presumably serves as a physiological defense against nephrolithiasis in vivo. We also address the case of another biological inhibitor of crystal growth, that of antifreeze peptide 3 from the winter flounder Pseudopleuronectes americanus. Examination of the interaction between this antifreeze peptide and the ice crystal lattice may provide a useful model for understanding how nephrocalcin inhibits calcium oxalate crystal growth in the kidneys. Calcium oxalate is the most common component of kidney stones. The growth of crystals in an in vitro calcium oxalateseeded crystal growth assay is inhibited by nondialyzable, protease-sensitive substances from urine. Nakagawa and associates isolated an acidic glycoprotein that inhibits calcium oxalate crystal growth from human urine. 1 This glycoprotein crystal growth inhibitor was originally called GCI but it is now referred to as nephrocalcin. Nephrocalcin is distinct from Tamm-Horsfall protein and albumin, which are 2 other abundant urinary proteins that bind calcium. 2 Evidence from immunochemical staining of human kidney tissue and from the study of medium conditioned by primary renal cell cultures suggests that the proximal convoluted tubule cells produce nephrocalcin.'3 PURIFICATION OF NEPHROCALCIN

Nephrocalcin is present in urine at about 16 mg.fl. and it represents a considerable fraction of the crystal growth inhibitory activity. In the purification scheme developed by Nakagawa and associates urine was· first dialyzed against water. 1 Only 10 per cent of the inhibitory activity was lost, indicating that dialyzable, small molecular weight substances, such as pyrophosphate and citrate, are not major contributors to crystal growth inhibition. After removal of small molecules, the inhibitor was concentrated and chromatographed on DEAE-cellulose. Several peaks of activity that were similar to

* Deceased.

t Current address: Department of Microbiology, Temple University Medical School, 3400 N. Broad St., Philadelphia, Pennsylvania 19140.

each other in amino acid composition were obtained. This material was colored due to tight but noncovalent binding of urobilirubin. The chromophore was removed by formamide treatment and Biogel PIO chromatography. Gel filtration on Sephacryl S-200 generated 3 peaks with molecular weights of 64,000, 27,000 and 14,000 daltons. Material from the higher molecular weight peaks dissociated into a 14,000 dalton species upon treatment with ethylenediaminetetracetic acid. Thus, the 14,000 dalton nephrocalcin monomer occurs in aggregated, oligomeric forms in urine. PHYSICAL CHEMICAL CHARACTERIZATION OF NEPHROCALCIN

Purified monomeric nephrocalcin was subjected to chemical analysis. 1 It is 10 per cent carbohydrate by weight and the carbohydrate composition is shown in table 1. Table 2 shows the amino acid composition of nephrocalcin. Nephrocalcin is extremely rich in acidic amino acids; 25 per cent of the residues are aspartic acid, glutamic acid and ')"-carboxyglutamic acid. The binding of nephrocalcin to calcium oxalate crystals displays Langmuir-type behavior, suggesting that it reversibly adsorbs to and blocks growth sites on the crystal surface. The dissociation constant for the crystal-nephrocalcin complex is about 0.5 µM. Normal adult urine contains about 16 mg./1. nephrocalcin, which is equivalent to an inhibitor concentration of about 1 µM. Since this value is above that of the dissociation constant, physiological concentrations of nephrocalcin should be able to repress calcium oxalate crystal growth efficiently. 1 Surface properties of nephrocalcin were studied with a film balance. Nephrocalcin readily forms an insoluble monomolecular layer at the surface of an aqueous buffer solution. This monolayer remained stable when it was compressed between 0 and 30 dynes per cm. and it did not collapse until a pressure of about 41.5 dynes per cm., 1 which is quite unusual for a protein. These observations suggest that the inhibitor has a highly organized structure at the air-water interface and that hydrophilic and hydrophobic regions occupy separate regions of its surface. 750

CRYSTAL GROWTH INHIBITORS TABLE 1.

Carbohydrate composition of human nephrocalcin 1

Carbohydrate

Residues/Molecule

0.4 1.2

Fucose Mannose Galactose Glucose Galactosamine

G lucosamine N-acetylneuraminic acid Total

TABLE 2.

Weight%

1.6

2

1.1 1.6 4.0 0.4 10.3

1

2 4

Amino acid composition of nephrocalcin 1

Amino Acid Lysine Histidine Arginine Aspartic acid Glutamic acid -y-carboxyglutamic Threonine Serine Praline Glycine Alanine Isoleucine Leucine Valine Tyrosine Phenylalanine Tryptophan Methionine Cysteine Total

Nearest Integer 4 2

5 12 13 2 10 11 6 12 8 3 7 7 1 3

27/110 = 25%

2

llO

NEPHROCALCIN CONTAINS -y-CARBOXYGLUTAMIC ACID

The modified amino acid -y-carboxyglutamic acid is posttranslationally derived from glutamic acid through a vitamin K dependent carboxylation reaction. 4 -y-Carboxyglutamic acid is present in a number of calcium-binding proteins, including osteocalcin," prothrombin" and several other plasma proteins that are involved in blood coagulation. As indicated previously nephrocalcin is heterogeneous on anion exchange chromatography, resolving into 4 peaks that are similar to each other in amino acid and carbohydrate composition. The material in 3 of these peaks (A to C) is posttranslationally modified to contain 2 to 3 residues of -y-carboxyglutamic acid, whereas that in the last peak (D) is not modified.7 The ')'-carboxyglutamic modification may contribute to the surface properties of nephrocalcin. When the DEAE peaks A to D nephrocalcin subfractions were separately analyzed in film balance studies, the A to C fractions, which contain 'Ycarboxyglutamic acid, formed stable monolayers at the airwater interface, whereas the D peak, which lacks 'Y-carboxyglutamic acid, exhibited a lower collapse pressure. Thus, segregation of hydrophilic and hydrophobic domains in nephrocalcin may be related to the presence of the 'Y-carboxyglutamic acid modification. Studies of nephrocalcin isolated from urine of stone formers and from kidney stones also support the hypothesis that the -y-carboxyglutamic acid modification may be a critical requirement for crystal growth inhibitory activity. NEPHROCALCIN FROM STONE FORMER URINE AND KIDNEY STONES

Studies on nephrocalcin isolated from stone former urine and pulverized kidney stones suggest that the presence of the 7-carboxyglutamic acid modification and the ability to form stable films with high collapse pressures may be important factors enabling nephrocalcin to prevent stone formation in vivo. Nephrocalcin was purified from pulverized kidney stones' and pooled stone former urine.' These materials lacked -ycarboxyglutamic acid and they had altered surface properties compared to normal nephrocalcin. Monolayers formed with

751

kidney stone nephrocalcin and stone former nephrocalcin were less stable than those formed by nephrocalcin from normal urine. ANTIFREEZE PEPTIDE MODEL FOR PROTEIN INHIBITION OF CRYSTAL GROWTH

Examination and analysis of the amino acid sequence of nephrocalcin will enable us to understand its structure and how it functions as a crystal growth inhibitor more fully. Since such information currently is not available, however, we will discuss instead another biological inhibitor of crystal growth, antifreeze peptide 3 from the winter flounder, Pseudopleuronectes americanus. Understanding the interaction of this antifreeze peptide with the ice lattice may provide us with a useful conceptual model for thinking about nephrocalcin-calcium oxalate crystal interactions. Water molecules in hexagonal ice are arranged in roughly hexagonal rings with an oxygen atom at every corner. These oxygens are separated by 4.5 angstroms in the lattice face parallel to the a axis. The predicted structure of winter flounder antifreeze peptide 3 allows it to recognize and interact specifically with this 4.5 angstrom periodicity in the ice lattice and, thus, to block crystal growth. Figure 1, a shows the sequence of antifreeze peptide 3 from the winter flounder."· 10 It contains 37 amino acids, of which 24 (66 per cent) are alanines. The sequence of the peptide can be divided into 3 segments of 11 amino acids in length, each of which begins with a threonine, and has a polar residue (aspartic acid or asparagine) in the fourth position and alanines (with a few exceptions) in positions 5 through 11. Circular dichroism11 12 and viscosityu studies suggest that this peptide assumes an a-helical structure at low temperatures. The a-helical form of peptide 3 is especially well suited to binding to the surface of hexagonal ice crystals and to preventing their enlargment because a 4.5 angstrom periodicity is generated in the peptide as a result of the a-helical conformation. Figure 1, b shows a helical net projection of the a-helical peptide generated from the antifreeze peptide 3 sequence. Polar and nonpolar residues are concentrated on opposite sides of the helix surface, and pairs of polar residues (gray) are separated by 4.5 angstroms along the length of the rod-shaped molecule. This 4.5 angstrom separation in the peptide matches the 4.5 angstrom repeat distance between oxygens along the a axis of hexagonal ice. Such a match suggests that the binding of the antifreeze peptide to ice could occur by means of hydrogen bonding between water molecules in the ice lattice, and the hydroxyls of threonines and the carbonyls of aspartate or asparagine residues in the antifreeze peptide. 11 • 13 The proposed relationship is illustrated in figure 2. Computer graphics was used to model the interaction between antifreeze peptide 3 and hexagonal ice. The illustration shows water molecules in the ice lattice interacting with antifreeze peptide 3. Bonds have been drawn for water molecules in the background of the lattice, and van der Waals radii have been filled in for some water molecules in the foreground of the lattice. The a carbon chain of the peptide 3 a-helix is on the right side and van der Waals radii for threonine-13 and asparagine-16 have been drawn. The arrows indicate the hydroxyl group of threonine-13 and the carbonyl group of the asparagine-16, which can precisely displace water molecules in the growing ice lattice. Thus, water molecules in the liquid surrounding an ice crystal are prevented from joining it by bound peptide 3. On the side facing the ice crystal, polar side chains of the peptide a- helix precisely hydrogen bond with specific oxygen atoms in the ice lattice; on the opposite, outward facing side of the helix the density of hydrophobic amino acids (mainly alanines) prevents further ice water interactions. The architecture of a calcium oxalate crystal is obviously different from the architecture of a water crystal; however, we believe that this example may prove instructive and the struc-

752

KAISER AND BOCK D

a

T T T T

A A A A

s A D A

D N N

A A A

A A A

A A A

A A A

A A A

A K A

L L A

FIG. 2. Molecular graphics illustration of antifreeze peptide 3-hexagonal ice lattice interaction.

different from the described interaction between antifreeze peptide 3 and ice, it is believed that the latter represents a useful conceptual model for investigating the structure and function of kidney stone crystal growth inhibitors. Drs. John Kuriyan and John Taylor provided helpful discussions and assistance with graphics. REFERENCES

b FIG. 1. a, amino acid sequence of Pseudopleuronectus americanus antifreeze peptide 3.9 •10 b, helical net diagram of antifreeze peptide 3. Distribution of amino acid side chains on surface of regular a-helix is shown. Gray areas represent threonines and aspartic acids or asparagines that are separated by 4.5 angstroms. D, aspartic acid. T, threonine. A, alanine. S, serine. L, leucine. N, asparagine. K, lysine.

ture assumed by nephrocalcin may be similarly optimized for interacting with the calcium oxalate lattice and preventing propagation of the crystal. The force-area studies presented earlier already indicate that hydrophobic and hydrophilic regions are separated on the surface of nephrocalcin, as also occurs in antifreeze peptide 3. The polar face of nephrocalcin may specifically interact with the surface of the calcium oxalate crystal, causing the nonpolar face to form a growth-arresting hydrophobic coating around the crystal. CONCLUSION

Nephrocalcin has been purified from urine. The physical and chemical properties of nephrocalcin are consistent with its proposed function as a physiological inhibitor of calcium oxalate crystal growth in vivo. Studies on nephrocalcin isolated from stone former urine and pulverized kidney stones, and on the subfraction of nephrocalcin that fails to be -y-carboxylated in normal urine suggest that the presence of the -y-carboxyglutamic acid modification and the ability to form stable films with high collapse pressures may be important factors enabling nephrocalcin to prevent stone formation. Antifreeze peptide 3 from the winter flounder is a well characterized biological inhibitor of crystal growth. Sequence and circular dichroism analyses predict that this peptide assumes an a-helical conformation in which pairs of polar residues on one face of the helix are optimally positioned for hydrogen bonding with specific oxygens in the lattice of hexagonal ice. The opposite, exposed face of helical antifreeze peptide then presents a nonpolar surface to the liquid milieu, preventing addition of further water molecules and arresting growth of the ice crystal. Although the specifics of nephrocalcin adsorption to calcium oxalate crystals will undoubtedly be

1. Nakagawa, Y., Abram, V., Kezdy, F. J., Kaiser, E. T. and Coe, F. L.: Purification and characterization of the principal inhibitor of calcium oxalate monohydrate crystal growth in human urine. J. Biol. Chem., 258: 12594, 1983. 2. Worcester, E., Kumar, S., Nakagawa, Y., Hunt, J. and Coe, F. L.: Immunochemical and functional differentiation of urinary glycoprotein crystal growth inhibitor (CGI), Tamm-Horsfall protein (THP) and albumin (ALB). Kidney Int., 29: 295, abstract 72, 1986. 3. Coe, F. L.: Unpublished data. 4. Suttie, J. W.: Vitamin K-dependent carboxylase. Ann. Rev. Biochem., 54: 459, 1985. 5. Price, P.A., Otsuka, A. S., Poser, J. W., Kristaponis, J. and Raman, N .: Characterization of a -y-carboxyglutamic acid-containing protein from bone. Proc. Natl. Acad. Sci., 73: 1147, 1976. 6. Stenflo, J., Ferlund, P., Egan, W. and Roepstorff: Vitamin K dependent modifications of glutamic acid residues in prothrombin. Proc. Natl. Acad. Sci., 71: 2730, 1974. 7. Nakagawa, Y., Abram, V., Parks, J. H., Lau, H.-S., Kawooya, J. K. and Coe, F. L.: Urine glycoprotein crystal growth inhibitors: evidence for a molecular abnormality in calcium oxalate nephrolithiasis. J. Clin. Invest., 76: 1455, 1985. 8. Nakagawa, Y., Ahmed, M., Hall, S. L., Deganello, S. and Coe, F. L.: Isolation from human calcium oxalate renal stones of nephrocalcin, a glycoprotein inhibitor of calcium oxalate crystal growth: evidence that nephrocalcin from patients with calcium oxalate nephrolithiasis is deficient in gamma-carboxyglutamic acid. J. Clin. Invest., 79: 1782, 1987. 9. De Vries, A. L. and Lin, Y.: Structure of a peptide antifreeze and mechanism of adsorption to ice. Biochim. Biophys. Acta, 495: 388, 1977. 10. Lin, Y. and Gross, J. K.: Molecular cloning and characterization of winter flounder antifreeze cDNA. Proc. Natl. Acad. Sci., 78: 2825, 1981. 11. Raymond, J. A., Radding, W. and DeVries, A. L.: Circular dichroism of protein and glycoprotein fish antifreezes. Letter to the Editor. Biopolymers, 16: 2575, 1977. 12. Ananthanarayanan, V. S. and Hew, C. L.: Structural studies of the freezing-point-depressing protein of the winter flounder Pseudopleuronectes americanus. Biochem. Biophys. Res. Comm., 74: 685, 1977. 13. DeVries, A. L.: Antifreeze peptides and glycopeptides in cold-water fishes. Ann. Rev. Physiol., 45: 245, 1983.