Modulation of Calcium Oxalate Crystallization by Linear Aspartic Acid

Langmuir 2006, 22, 7279-7285

7279

Modulation of Calcium Oxalate Crystallization by Linear Aspartic Acid-Rich Peptides Lijun Wang,†,| S. Roger Qiu,‡,| William Zachowicz,† Xiangying Guan,† James J. DeYoreo,‡ George H. Nancollas,† and John R. Hoyer*,§,⊥ Department of Chemistry, Natural Sciences Complex, State UniVersity of New York at Buffalo, Buffalo, New York 14260, Department of Chemistry and Materials Science, Lawrence LiVermore National Laboratory, LiVermore, California 94551, and The Children’s Hospital of Philadelphia and UniVersity of PennsylVania School of Medicine, Philadelphia, PennsylVania 19104 ReceiVed April 4, 2006. In Final Form: June 8, 2006 Calcium oxalate monohydrate (COM) kidney stone formation is prevented in most humans by urinary crystallization inhibitors. Urinary osteopontin (OPN) is a prototype of the aspartic acid-rich proteins (AARP) that modulate biomineralization. Synthetic poly(aspartic acids) that resemble functional domains of AARPs provide surrogate molecules for exploring the role of AARPs in biomineralization. Effects of linear aspartic acid-rich peptides on COM growth kinetics and morphology were evaluated by the combination of constant composition (CC) analysis and atomic force microscopy (AFM). A spacer amino acid (either glycine or serine) was incorporated during synthesis after each group of 3 aspartic acids (DDD) in the 27-mer peptide sequences. Kinetic CC studies revealed that the DDD peptide with serine spacers (DDDS) was more than 30 times more effective in inhibiting COM crystal growth than the DDD peptide with glycine spacers (DDDG). AFM revealed changes in morphology on (010) and (-101) COM faces that were generally similar to those previously described for OPN and citrate, respectively. At comparable peptide levels, the effects of step pinning and reduced growth rate caused by DDDS were remarkably greater. In CC nucleation studies, DDDS caused a greater prolongation of induction periods than DDDG. Thus, nucleation studies link changes in interfacial energy caused by peptide adsorption to COM to the CC growth and AFM results. These studies indicate that, in addition to the number of acidic residues, the contributions of other amino acids to the conformation of DDD peptides are also important determinants of the inhibition of COM nucleation and growth.

Introduction The primary mineral constituent of the majority of human kidney stones is calcium oxalate monohydrate, CaC2O4‚H2O (COM), the most thermodynamically stable form of calcium oxalate (CaOx).1 Despite the frequent supersaturation of normal urine with respect to calcium oxalate, stone formation in most humans is prevented by biological control mechanisms. These mechanisms include the effects of the presence of inhibitors of crystallization in urine2,3 and the excretion of crystals before they attain a size sufficient for retention within the kidney. Osteopontin (OPN) is one of the most potent of these urinary inhibitors; levels of OPN below those prevailing in normal human urine (>100 nM)4,5 inhibit the nucleation, growth, and aggregation of COM crystals5,6 as well as the binding of COM crystals to renal cells.7 A growing body of evidence has linked a molecular feature of the proteins associated with mineralization, an abundance of carboxylate groups, to the modulation of biomineralization.8,9 Osteopontin (OPN) is a prototype of a group of molecules that have been identified as inhibitors of CaOx crystallization. OPN * Corresponding author. E-mail: [email protected]. Tel: (302)831-6348. † State University of New York at Buffalo. ‡ Lawrence Livermore National Laboratory. § The Children’s Hospital of Philadelphia and University of Pennsylvania. | These authors contributed equally. ⊥ Current address: Department of Biological Sciences, University of Delaware, Newark, DE 19716. (1) Coe, F. L.; Parks, J. H.; Asplin, J. R. N. Engl. J. Med. 1992, 327, 11411152. (2) Edyvane, K. A.; Hibberd, C. M.; Harnett, R. M.; Marshall, V. R.; Ryall, R. L. Clin. Chim. Acta 1987, 167, 329-338. (3) Nakagawa, Y.; Abram, V.; Kezdy, F. J.; Kaiser, E. T.; Coe, F. L. J. Biol. Chem. 1983, 258, 12594-12600.

contains several sequence domains that are rich in the dicarboxylic acid, aspartic acid (HOOC-CH2-CH(NH2)-COOH). A total of 48 of the 298 amino acids in the peptide sequence of human OPN are aspartic acid. Many other proteins that modulate the mineralization of calcium crystals in a variety of organisms also have an abundance of aspartic acid residues.9-12 It is noteworthy that monomers of aspartic acid are able to inhibit the growth of calcium carbonate (calcite) crystals,13 albeit at molar concentrations considerable higher than those required to produce a comparable inhibition of crystallization by aspartic acid-rich proteins (AARP) such as OPN. Synthetic poly(aspartic acids) (PAAs) provide surrogate molecules for exploring the role of aspartic acid-rich domains within AARPs in the process of biomineralization. For example, the bulk crystallization studies of Wesson and Worchester showed that the presence of PAA preferentially caused the formation of calcium oxalate dihydrate (COD) rather than COM.14 Furthermore, scanning electron microscopy (SEM) and atomic force microscopy (AFM) studies (4) Min, W.; Shiraga, H.; Chalko, C.; Goldfarb, S.; Krishna, G. G.; Hoyer, J. R. Kidney Int. 1998, 53, 189-193. (5) Asplin, J. R.; Arsenault, D.; Parks, J. H.; Coe, F. L.; Hoyer, J. R. Kidney Int. 1998, 53, 194-199. (6) Worcester, E. M.; Beshensky, A. M. Ann. N.Y. Acad. Sci. 1995, 760, 375377. (7) Lieske, J. C.; Leonard, R.; Toback, F. G. Am. J. Physiol. 1995, 268, F604F612. (8) Addadi, L.; Weiner, S. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 4110-4114. (9) Shiraga, H.; Min, W.; VanDusen, W. J.; Clayman, M. D.; Miner, D.; Terrell, C. H.; Sherbotie, J. R.; Foreman, J. W.; Przysiecki, C.; Neilson, E. G.; Hoyer, J. R. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 426-430. (10) Berman, A.; Addadi, L.; Weiner, S. Nature 1988, 331, 546-548. (11) Belcher, A. M.; Wu, X. H.; Christensen, R. J.; Hansma, P. K.; Stucky, G. D.; Morse, D. E. Nature 1996, 381, 56-58. (12) Walters, D. A.; Smith, B. L.; Belcher, A. M.; Paloczi, G. T.; Stucky, G. D.; Morse, D. E.; Hansma, P. K. Biophys. J. 1997, 72, 1425-1433. (13) Orme, C. A.; Noy, A.; Wierzbicki, A.; McBride, M. Y.; Grantham, M.; Dove, P. M.; DeYoreo, J. J. Nature 2001, 411, 775-779. (14) Wesson, J. A.; Worcester, E. Scanning Microsc. 1996, 10, 415-424.

10.1021/la060897z CCC: $33.50 © 2006 American Chemical Society Published on Web 07/22/2006

7280 Langmuir, Vol. 22, No. 17, 2006

Wang et al. Table 1. Conditions in COM Growth Experiments

Figure 1. Sequences of linear aspartic acid-rich peptides with serine (S) spacers (MW ) 2955) and glycine (G) spacers (MW ) 2775).

supersaturation WCaCl2 ) WK2C2O4 VCaCl2 ) VK2C2O4 WNaCl VNaCl σ /10-4 mol L-1 /10-3 mol L-1 /mol L-1 /mol L-1 0.266

2.500

5.500

0.149

0.288

showed that PAA specifically inhibited COM growth in the 〈101〉 direction on the (-101) face. This was postulated to be due to the preferential binding of poly(aspartic acid) to the {120} and {-101} surfaces.15 (Direction and face designations used in refs 15 and 30 were translated to match the convention used in our AFM studies.17,18) The objective of the present in vitro studies was to define the effects of synthetic linear aspartic acid-rich peptides on COM crystal growth using constant composition (CC)16 and AFM analysis.17,18 The peptides were designed to be comparable in size to previously studied aspartic acid-rich sequence domains within OPN19 such as amino acids 70-105, a domain that contains 53% aspartic acids. To avoid blocked synthetic sites and the unpredictable branching present in PAA peptides used in earlier studies, spacer amino acids (either glycine or serine) were periodically incorporated into the peptide sequence. Both the CC and AFM studies revealed that the peptide with serine spacers (DDDS) had remarkably greater effects on COM crystal growth than the peptide with glycine spacers (DDDG).

Constant-Composition Analysis of COM Growth and Nucleation. Supersaturated solutions (ionic strength I ) 0.15 mol L-1, pH 7.0) were prepared by slowly mixing filtered (0.22 µm Millipore filter) calcium chloride, potassium oxalate, and sodium chloride. To exclude carbon dioxide, the reaction vessels were purged with nitrogen saturated with water vapor at 37 °C. Crystallization experiments, initiated by the introduction of known amounts of COM seed crystals, were conducted in magnetically stirred (450 rpm) double-jacketed vessels thermostatically controlled at 37.0 ( 0.1 °C. In the CC method, lattice ions are simultaneously added to the reaction solution to compensate for changes due to crystal growth. When growth is initiated, changes in electrode potential trigger the addition of two titrant solutions, which were designed to maintain a constant activity of all ionic species in the reaction solution. The titrants having the following COM stoichiometries were prepared:

Materials and Methods

W and V are the total concentrations in the reaction solutions and titrants, respectively, and Ceff is the effective titrant concentration (5.00 × 10-3 mol L-1 in this study). The experimental conditions are summarized in Table 1. In these COM growth experiments, peptides were added to supersaturated solutions to achieve concentrations ranging from 6.8 to 84.6 nM for DDDS and from 288 to 3640 nM for DDDG. The reaction solution was maintained at pH 7.0 using 0.010 mol L-1 tris (hydroxymethyl) aminomethane buffer. Nucleation experiments were performed at a higher relative supersaturation, σCOM ) 1.53, pH ) 7.00, T ) 37.0 ( 0.1 °C, and an ionic strength of 0.15 mol L-1, adjusted by the addition of sodium chloride. These nonseeded experiments were initiated by the slow mixing of solutions to make the reaction solution supersaturated. The peptide concentration used in nucleation studies was 6.7 or 20.1 nM of either DDDS or DDDG. Induction times are reported as mean ( SD. Titrant addition was triggered by a potentiometer (Orion 720A, U.K.) incorporating a calcium ion selective electrode (Orion 93-20) and a reference electrode (Orion 900100, U.K.). During the experiment, the output of the potentiometer was constantly compared with a preset value, and a difference in output (error signal) activated motor-driven titrant burets, thereby maintaining a constant thermodynamic driving force. Chemical analysis of solution samples periodically withdrawn and filtered (0.22 µm Millipore filter) showed that the total calcium (atomic absorption) concentration remained constant to within (1.5% during the experiments. The driving force for the crystal growth of COM in supersaturated solutions is due to the change in Gibbs free energy from the supersaturated solutions to equilibrium. The relative supersaturation, σ, is defined by eq 4

Peptide Synthesis. The aspartic acid-rich peptides with serine or glycine spacers were synthesized by the sequential addition of Fmoc amino acids according to standard procedures.20 These 27-mer peptides were purified, and molecular weights were verified by mass spectrometry as previously described.19 The condensed formula for the peptide with serine spacers is (DDDS)6 DDD, and that for the peptide with glycine spacers is (DDDG)6 DDD with D ) aspartic acid, S ) serine, and G ) glycine. The linear sequences and molecular weights of these DDD peptides are shown in Figure 1. Constant-Composition Analysis. Solution Preparation. All solutions were prepared using reagent-grade chemicals and triply distilled deionized (TDW) water. They were filtered through 0.22 µm Millipore filters before use. Calcium chloride solutions were standardized with EDTA using Erichrome Black T as an indicator. Potassium oxalate solutions were passed through an anion-exchange resin in hydroxide form, and the eluted base was titrated with hydrochloric acid solution using methyl red as an indicator. COM Seed Preparation. COM crystals were prepared by the dropwise addition of 500 mL aliquots of 0.40 mol L-1 calcium chloride and potassium oxalate into 1 L of triply distilled deionized water (TDW) at 70.0 ( 0.1 °C. The suspension was stirred continuously at this temperature for 6 h to allow for complete conversion of the precipitate to calcium oxalate monohydrate. The solid was then filtered and washed with TDW until free from residual chloride ion, suspended in TDW, and aged at 37.0 °C for at least a month prior to use. It has been shown that COM crystallites may be stored for long periods of time without the conversion of the monohydrate to other hydrates. The specific surface area (SSA) of the COM crystals was determined to be 3.3 m2/g by a single-point Brunauer-Emmett-Teller (BET) method using a Quantasorb continuous 20/80 nitrogen/helium gas mixture (Quantachrome Corporation). (15) Sheng, X.; Ward, M. D.; Wesson, J. A. J. Am. Chem. Soc. 2003, 125, 2854-2855. (16) Tomson, M. B.; Nancollas, G. H. Science 1978, 200, 1059-1060. (17) Qiu, S. R.; Wierzbicki, A.; Orme, C. A.; Cody, A. M.; Hoyer, J. R.; Nancollas, G. H.; Zepeda, S.; De Yoreo, J. J. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 1811-1815. (18) Qiu, S. R.; Wierzbicki, A.; Zepeda, S.; Orme, C. A.; Hoyer, J. R.; Nancollas, G. H.; Cody, A. M.; De Yoreo, J. J. J. Am. Chem. Soc. 2005, 127, 9036-9044. (19) Hoyer, J. R.; Asplin, J. R.; Otvos, L, Jr. Kidney Int. 2001, 60, 77-82. (20) Fields, G. B.; Noble, R. L. Int. J. Pept. Protein Res. 1990, 35, 161-214.

titrant buret no. 1: VCaCl2 ) 2WCaCl2 + Ceff

(1)

VNaCl ) 2WNaCl - 2Ceff

(2)

titrant buret no. 2: VK2C2O4 ) 2WK2C2O4 + Ceff

(3)

σ)S-1)

[

]

(Ca2+)(C2O42-) Ksp

1/2

-1

(4)

where (Ca2+) and (C2O42-) are the ionic activities calculated by successive approximation for the ionic strength using the Davies extended form of the Debye-Hu¨ckel equation and mass balance expressions for total calcium and total oxalate with appropriate equilibrium constants. The solubility activity product, Ksp, of COM at 37.0 °C is 2.20 × 10-9 mol2 L-2 21 and at room temperature (AFM studies) is 1.66 × 10-9 mol2 L-2.21 (21) Tomazic, B. B.; Nancollas, G. H. InVest. Urol. 1979, 16, 329-335.

Modulation of Calcium Oxalate Crystallization

Langmuir, Vol. 22, No. 17, 2006 7281

The crystal growth rates were determined from the slopes of the plots of titrant volume added as a function of time during the first 15 min of reaction. The overall growth rate, R, is defined by eq 5 R)

Ceff dV SAms dt

(5)

where dV/dt is the titrant curve gradient, SA is the specific surface area of the seed crystals, and ms is the initial seed mass. Growth rates are expressed as the mean ( SD for a series of seven experiments at each peptide concentration. Atomic Force Microscopy. All in situ images were collected at room temperature (24 °C) in contact mode (Digital Instruments Nanoscope III, Santa Barbara, CA) on surfaces of the COM crystals (∼500 µm in length) that were anchored inside the enclosed fluid cell. The images were obtained while supersaturated solutions (σ ) 0.82 for COM) were flowing through the system. To ensure that growth was limited by surface kinetics and not by diffusion, the flow rate was adjusted to 2 mL/min.18 At this supersaturation, the step speed did not change when the flow rate was further increased. To ensure that the AFM images were authentic representations of the surface morphology, they were obtained while observing several precautions as previously described in detail.18,22,23 These precautions include the following: (1) Minimal imaging force was used while determining step speeds. (2) The step speed was estimated from images collected on both large and small scales to avoid the effects of imaging. (3) Images were collected at different scan angles, and trace and retrace images were regularly compared to eliminate potential imaging artifacts induced by contamination or sticking of the tip. (4) The inevitable distortion of step orientation in in-situ images caused by step movement during the finite imaging time was taken into account. Thus, the recorded step orientation differed from the true orientation, and as previously described,23 this change in angle is used to extract the step velocities under different growth conditions.

Figure 2. Plots of titrant volume against time for COM crystal growth in the absence and presence of the DDDS peptide. The curves were normalized to the same seed mass (10.0 mg), and growth rates were determined from the slope of the plots of titrant volume added as a function of time during the first 15 min of reaction.

Results CC Investigations of COM Growth. The plots of titrant volume against time in the CC experiments determining COM crystal growth rates that were performed in the absence and presence of a series of concentrations of the linear aspartic acidrich peptides, DDDS and DDDG, are shown in Figures 2 and 3, respectively. Increasing the concentration of each of these 27-mer peptides progressively decreased the crystal growth rates derived from measurements of added titrant volumes versus time. As shown in Figure 4, DDDS is remarkably more potent as an inhibitor of COM growth than DDDG. The presence of 27.1 nM DDDS caused the COM growth rate in pure supersaturated solutions, 14.7 ( 0.1 µmol/m-2/min (n ) 7), to decrease to 6.36 ( 0.06 µmol/m-2/min (n)7) (∼56% inhibition) with a further progressive decrease to 1.22 ( 0.03 µmol/m-2/min (n ) 7) at 84.6 nM DDDS (∼92% inhibition). By contrast, the presence of 865 nM DDDG was required to decrease the COM growth rate from 14.7 ( 0.1 µmol/m-2/min (n ) 10) in pure supersaturated solutions to 6.64 ( 0.07 µmol/m-2 /min (n ) 7) (55% inhibition); 3640 nM DDDG caused a further decrease to 1.12 ( 0.04 µmol/ m-2/min (n)7) (∼92% inhibition). Thus, when compared to growth inhibition by DDDG, the potency of growth inhibition by DDDS is more than 30 times greater at all peptide levels that were studied. CC Investigations of COM Nucleation. The basis for the marked differences in inhibitory effects of DDDS and DDDG on COM growth was investigated in CC nucleation experiments (22) Land, T. A.; De Yoreo, J. J.; Lee, J. D. Surf. Sci. 1997, 384, 136-155. (23) DeYoreo, J. J.; Orme, C. A.; Land, T. A. In AdVances in Crystal Growth Research; Sato, K., Nakajima, K., Furukawa, Y., Eds.; Elsevier Science, Amsterdam, 2001; pp 361-380.

Figure 3. Plots of titrant volume against time for COM crystal growth in the absence and presence of the DDDG peptide. The curves were normalized to the same seed mass (10.0 mg), and growth rates were determined from the slopes of the plots of titrant volume added as a function of time during the first 15 min of reaction.

to determine the effects of these peptides on induction times for nucleation. These studies showed significantly longer induction times in the presence of DDDS than with DDDG (Figure 5). COM nucleation was retarded in a concentration-dependent fashion by both peptides. The induction time for pure supersaturated solution was 38 ( 4 min (n ) 4). The presence of a 6.7 nM concentration of either DDDG or DDDS, respectively, increased the induction times to 50 ( 3 (n ) 4) and 58 ( 2 min (n ) 3), respectively. The presence of a 20.1 nM concentration of either DDDG or DDDS, respectively, increased the induction times further to 65 ( 6 (n ) 4) and 138 ( 5 min (n ) 3), respectively. Atomic Force Microscopy. AFM images were taken on the (010) and (-101) faces of COM crystals. The DDDS peptide strongly altered both the morphology and kinetics of step movement on the growing COM crystals faces. As the DDDS concentration was increased from 0.54 to 134 nM, step pinning occurred much more rapidly. At σ ) 0.82, step movement ceased completely at DDDS levels above 1.34 nM, and the time for the step speed to become zero decreased as the levels of DDDS increased. Figure 6 shows the temporal evolution of steps on the

7282 Langmuir, Vol. 22, No. 17, 2006

Wang et al.

Figure 4. Inhibition of the COM growth rate by a series of concentrations of DDDS (b) and DDDG (4). The concentration of DDDG needed to achieve either 50% inhibition or 90% inhibition of COM growth is more than 30 times larger than for DDDS. Figure 6. AFM images showing the surface evolution of dislocation hillocks on the COM (010) face during growth in a supersaturated (σ ) 0.82) solution containing 6.7 nM DDDS peptide (serine spacers). (a) t ) 0 min, (b) t ) 20 min, and (c) t ) 76 min. Image horizontal dimension: 2 µm.

Figure 5. Nucleation of COM at σCOM ) 1.53 in pure solution (a) and in the presence of 6.7 nM DDDG (b) or 6.7 nM DDDS (c) and 20.1 nM DDDG (d) or 20.1 nM DDDS (e) showing dosedependent inhibition by both peptides. The induction times (mean ( SD) and the number of experiments are presented in the Results section.

(010) face in the presence of 6.67 nM DDDS peptide. Images of this face collected at other scan angles displayed the same morphological features. Within 20 min or less after exposure to the DDDS peptide, step edges in both the 〈021〉 and 〈100〉 directions were markedly roughened, indicating that these steps were highly pinned. Morphological changes in the 〈100〉 direction were more prominent than those in the 〈021〉 direction, particularly in a relative reduction of terrace width. Ultimately, the rectangularly shaped hillocks on the (010) face were transformed into a series of elongated step segments along the 〈021〉 direction (Figure 6c). The formation of similar fingerlike step bunches on crystals of KH2PO4 grown in the presence of Fe3+ was shown to be intrinsically related to the dynamic competition between the pinning of steps by Fe3+ adsorption and step-step interactions within the step bunches24,25 The inherent tendency of steps on (24) Land, T. A.; Martin, T. L.; Potapenko, S.; Palmore, G. T.; De Yoreo, J. J. Nature 1999, 399, 442-445. (25) Land, T. A.; DeYoreo, J. J.; Martin, T. L.; Palmore, G. T. Crystallogr. Rep. 1999, 44, 655-666.

the (010) face to bunch into quadruple-height steps is likely to be a source of the same behavior in the COM system. The peptide markedly altered the step kinetics in both directions. Step speed was reduced by 50% after only 20 min of exposure to DDDS (6.67 nM). Subsequently, step speed was difficult to estimate because of the extreme serration of step edges. All growth had ceased after ∼1 h, and no additional changes in surface morphology were detected in images obtained at later times. In contrast to the extensive changes on the (010) face induced by DDDS described above, the morphology and step kinetics on this face were much less severely affected by the DDD peptide with glycine spacers (DDDG). Morphological features of growth hillocks on the (010) face induced over time by the DDDG peptide (8 nM) are shown in Figure 7. Step edges, though somewhat rougher, remained clearly delineated throughout the experiment. Although the overall shape of the hillocks was slightly changed, strong pinning of step edges in either direction was not detected. The major alteration of the shape of dislocation hillocks was elongation along the crystallographically equivalent corners between the [-100] and [021] directions. This was due to a relative slowing of the step speed in the direction of the other two corners of the hillock, which led to a flattening of the hillock in that direction. Because the difference between these two sets of corners is related to differences in the structure of left-facing and right-facing kink sites along the steps, this result indicates a stronger interaction with those kinks facing the slow corners. Overall, there was less effect on step speed than observed with the DDDS peptide with only a 40% reduction in step speed after an hour of exposure to the DDDG peptide and no further reduction at later times. Both peptides caused changes on the (-101) face of COM. As on the (010) face, the effects on both morphology and step kinetics exerted by the DDDS peptide were much greater than those exerted by the DDDG peptide. The evolution of surface features of the (-101) face in the presence of DDDS at 1.34 nM is shown in Figure 8. Both the [101] and 〈120〉 steps were strongly

Modulation of Calcium Oxalate Crystallization

Figure 7. AFM images showing the surface evolution of COM dislocation hillocks on the COM (010) face during growth in a supersaturated (σ ) 0.82) solution containing 8.0 nM DDDG peptide (glycine spacers). (a) t ) 0 min, (b) t ) 20 min, (c) t ) 40 min, and (d) 70 min. Image horizontal dimension: 2 µm.

affected. The step edges along both directions were pinned and eventually became undefined. It is noteworthy that a new step is expressed along the [120] direction (Figure 8c). The emergence of the [120] step suggests that this step is slowed by the presence of DDDS because it was not detected during its rapid growth in pure solution17,18 (Figure 8a). Step speeds along the [101] and the 〈120〉 directions were also significantly reduced. After 90 min of exposure to a low level (1.34 nM) of DDDS, growth in all step directions ceased. In contrast, effects of the same level of DDDG on the step kinetics and step morphology were very minimal (Figure 9). The presence of [120] steps was not detected. DDDG reduced step speed by only 20%, and the maximal morphologic change was similar to that shown in Figure 8b.

Discussion The present study clearly demonstrates that the number of acidic amino acids in a peptide is not the sole determinant of the inhibition of COM nucleation and growth. The presence of a serine spacer after each triplet of aspartic acids in DDDS, a linear 27-mer aspartic acid-rich peptide, caused remarkably greater inhibition of COM growth than presence of glycine spacers in DDDG, the corresponding 27-mer peptide (Figure 4). The role of differences in interfacial energy in the genesis of these differences in growth inhibition was investigated by nucleation studies (Figure 5). The dependence, on a macroscopic scale, of the energy of formation of nucleated droplets of minimum radius r* (assuming spherical shape) and surface tension γSL is expressed

Langmuir, Vol. 22, No. 17, 2006 7283

Figure 8. AFM images showing the surface evolution of dislocation hillocks on the (-101) face of COM during growth in a supersaturated (σ ) 0.82) solution containing 1.4 nM DDDS peptide (serine spacers). (a) t ) 0 min, (b) t ) 20 min, (c) t ) 41 min, and (d) t ) 95 min. Image horizontal dimension: 5 µm.

by eq 6 (the Kelvin-Gibbs equation) .26,27

r* )

2γSLΩ kT ln S

(6)

In this equation, Ω is the volume occupied by each growth unit. The equation implies that spontaneous crystallization does not occur in the bulk until critical conditions are reached or the driving force S is sufficiently high. Rather, a metastable equilibrium condition persists during an “induction period”, τ, prior to crystal formation. If the simplifying assumption is made that τ is essentially concerned with classical nucleation, then we can use eq 7

[

]

γSL3 ln τ ∝ C1 + C2 3 3 k T (ln S)2

(7)

in which C1 and C2 are independent constants. Equation 7 shows that the increase in the interfacial tension γSL following the introduction of peptide results in the inhibition of COM growth and an increase in the induction time. Figure 5 shows that the relationship between the induction time and surface tension is a function of peptide concentration. Because C2 contains only geometric parameters and C1 is controlled by entropy changes, the only variable relating τ and S that can be altered by the peptides is the interfacial tension.26 Thus, the results indicate (26) Wu, W.; Nancollas, G. H. AdV. Colloid Interface Sci. 1999, 79, 229-279. (27) Tang, R. K.; Darragh, M.; Orme, C. A.; Guan, X.; Hoyer, J. R.; Nancollas, G. H. Angew. Chem., Int. Ed. 2005, 44, 3698-3702.

7284 Langmuir, Vol. 22, No. 17, 2006

Figure 9. AFM images showing the surface evolution of dislocation hillocks on the (-101) face of COM during growth in a supersaturated (σ ) 0.82) solution containing 1.4 nM DDDG peptide (glycine spacers). (a) t ) 0 min, (b) t ) 20 min, (c) t ) 40 min, and (d) t ) 70 min. Image horizontal dimension: 5 µm.

that COM has a higher interfacial energy, γSL, after the adsorption of DDDS than the γSL of COM after the adsorption of DDDG in bulk crystallization. The inhibition of COM crystal growth by aspartic acid-rich peptides containing 11 and 13 aspartic acids and 3 serines each in synthetic sequences of 25 amino acids matching those in a human OPN sequence domain was previously studied using endpoint kinetic analysis.19 Both peptides inhibited COM growth by ∼50% at 1000 nM.19 Although these results are not strictly comparable to the present CC studies of 27-mer peptides with 21 aspartic acids in their sequence, less than 30 nM DDDS inhibited COM growth by more than 50%, suggesting that increasing the number of aspartic acid residues in the sequence was responsible for the increased growth inhibition. However, this is not the sole determinant of inhibition. DDDG has an equal number of aspartic acids but required at least ∼30 times higher levels than DDDS to achieve comparable COM growth inhibition. Thus, the amino acid used as a spacer between the triplets of aspartic acid in these linear sequences is a crucial determinant of the interaction with COM crystals. Amino acids adjacent to aspartic acids in the sequence may exert such an effect by influencing the pattern of charge density presented to the steps of COM crystals. The sites of interactions of these DDD peptides identified by our AFM studies were generally similar for the two peptides; the morphological changes showed that the inhibition was due to the peptide pinning of step motion. Consistent with the results of our CC studies, the AFM results showed that the

Wang et al.

peptide with serine spacers exerted much greater effects than the same level of the peptide with glycine spacers. The direct implication is that the binding energy of the peptides to the steps is significantly higher for DDDS than for DDDG. Because these two peptides differ only in the spacer amino between the triplets of aspartic acid, these differences in binding energies that are the source of the observed differences in effects exerted on step morphology and kinetics must be related to differences in peptide structure. All of the amino acids in the sequence of DDDS are hydrophilic. Thus, a linear coil structure of the peptide while in solution is most likely. With a linear structure of DDDS, nearly all of the carboxylic groups would participate in the interactions when this peptide reaches step edges. Moreover, the hydroxyl groups of serine molecules can also contribute to the chemical bonding to steps as well. Hydroxyl groups can interact by directly binding to Ca ions in the steps at both risers and basal planes and can also form hydrogen bonds with oxygen atoms in the oxalate groups.18 The importance of the latter is that hydrogen bonding increases the overall interaction strength and may play a pivotal role in stabilizing molecular structures bound to step edges. Hydrogen bonding has been shown to play such a role in the interaction between citrate and COM steps.18 In contrast, the glycines in DDDG are hydrophobic and may confer high flexibility.28 The glycine spacers may change peptide properties in two ways. Because glycine is not hydrophilic, the presence of glycine as a spacer may increase the tendency toward peptide folding. It may also enable high mobility, particularly at the ends of the molecule. Folding of DDDG would make its 3D structure more compact. Thus, the number of carboxylic groups available for interaction with steps may be reduced if they have become oriented away from the step edges. If there is no folding, then DDDG is probably highly mobile and flexible,28 which makes the peptide structure less stable. The binding strength between the peptide and crystal steps depends on many factors, including enthalpic and entropic contributions. The entropic contribution is usually negative.29 Thus, by making the peptide structure less stable, the absolute value of the entropic term will be increased; this weakens the interaction between DDDG and the steps. Therefore, both features conferred by glycine would be expected to reduce the inhibition of COM growth. Jung et al.30 proposed that local binding of anionic side chains to crystal surface sites (rather than any secondary polymer structure) governed the inhibition of COM growth when additivesspoly(acrylic acid), poly(aspartic acid), and poly(glutamic acid)swere used in AFM studies. The magnitude of this effect depended on the macromolecule structure, macromolecule concentration, and identity of the step site. Poly(aspartic acid) inhibited growth on the (120) step of the (-101) hillocks to a greater extent than poly(glutamic acid), but the opposite was found for the same step on the (010) hillocks. This suggests that growth inhibition is due to macromolecule binding to both planes of the step (i.e., the terrace and the step riser) or the pinning of the steps due to binding to the (-101) and (010) faces alone.30 The first option would be consistent with conclusions based on a combination of AFM and molecular modeling concerning step binding in numerous systems, including amino acids and polyaspartate peptides on calcite13,31 and citrate on COM17,18 (28) Branden, C.; Tooze, J. Introduction to Protein Structure, 2nd ed.; Garland Publishing: New York, 1999. (29) Noy, A.; Zepeda, S.; Orme, C. A.; Yeh, Y.; De Yoreo, J. J. J. Am. Chem. Soc. 2003, 125, 1356-1362. (30) Jung, T.; Sheng, X.; Choi, C. K.; Kim, W. S.; Wesson, J. A.; Ward, M. D. Langmuir 2004, 20, 8587-96. (31) Elhadj, S.; Salter, A.; Wierzbicki, A.; De Yoreo, J. J.; Han, N.; Dove, P. M. Cryst. Growth Des. 2006, 6, 197-201.

Modulation of Calcium Oxalate Crystallization

The simplest explanation for the fact that DDDS and DDDG interact with steps on both faces is that the size of the DDD peptides, being intermediate between that of citrate and OPN, allows strong interactions with steps that are otherwise precluded by either the smaller size of citrate or the much larger size of OPN. However, size does not explain why the potency of the two peptides is so different. Computer-based molecular modeling will be required to determine the precise basis for the differences in the strength of interaction shown in our nucleation studies. Such studies are now in progress. The insights derived from molecular modeling investigations of DDDS and DDDG may

Langmuir, Vol. 22, No. 17, 2006 7285

facilitate the design of potent new inhibitors of COM crystallization that will ultimately be useful in the therapy of kidney stone disease. Acknowledgment. This work was performed under the auspices of the U.S. Department of Energy by the University of California, Lawrence Livermore National Laboratory under contract no. W-7405-Eng-48. This work was supported by research grants from the National Institutes of Health (DK33501, DE03223, and DK61673). LA060897Z