Interaction of Anionic Compounds with Gelatin. 11: Effect on Some Physicochemical Properties of Gelatin

JAVA GAUTAM'~ AND HANS SCHOTT'~ Received June 22, 1992, from the 'School of Pharmacy, Temple University, Philadelphia, PA 79140. publication May 27, 1993@. 'Present address: Alcon Laboratories, Inc., Fort Worth, TX 76134.

Accepted for

Abstract 0 The interaction of gelatin with four monosulfonated or monocarboxylated azo dyes was investigated by measuring the surface tension and intrinsic viscosity of gelatin solutions containing the dyes at different concentrations, the rigidity and melting point of their gels, and the moisture regain of their films. The dyes, which were used as models for anionic drugs, differed in the size of their aromatic hydrocarbon moieties. Surface tension measurements showed that the gelatin did not affect the critical micelle concentration of the free dyes and that the bound dyes increased the surface activity of the gelatin. The dyes reduced the intrinsic viscosity of gelatin by as much as 2/3. They also lowered the rigidity and the melting point of dilute gelatin gels and reduced the moisture regain of dry gelatin films. These changes became more pronounced with increasing dye concentrations. The effectiveness of the dyes in producing these changes increased with the size of their hydrocarbon moieties because, as had been shown in a previous study, increasingly larger hydrocarbon moieties increased the binding of the dyes to gelatin. At the pH of the measurements, which was 1.9 units below the isoelectric point of the gelatin, the gelatin was a cationic polyelectrolyte. Binding of the dye anions by ion pairing, hydrogen bonds, and other secondary valence forces rendered the gelatin less ionic and less hydrophilic, which accounts for the present observations.

The interaction of gelatin with drugs or excipients encapsu- lated in hard or soft gelatin capsules may lead to changes i n the physicochemical properties of the gelatin wall and affect the bioavailability of the drugs. Even though many of the more recently approved drugs are anionic or acidic and some of them are dispensed in capsules or microcapsules where gelatin is a par t of the wall material, only scant attention has been paid t o the interaction of such drugs with gelatin. The interaction of gelatin with surfactants and dyes has been reviewed recently.' The interaction of gelatin with ionic and nonionic drugs and excipienta has been reviewed and investigated through their effect on the swelling of gelatin films. The only anionic compound included in these swelling studies was dicloxacillin sodium.*

The purpose of the present work was to investigate the effect of the gelatin binding of four acid azo dyes on some properties of gelatin in the solution, gel, and solid states. The dyes were used as model compounds for anionic drugs. This work is a sequel to our binding studies of t h e four dyes by gelatin.'

Experimental Section Materials-The gelatin was Type A (lot #1, Type K&K T-5542),

manufactured by acid treatment of pork skins by Kind & Knox, Sioux City, IA. It had a weight-average molecular weight of 114 000 determined by size exclusion chromatography, a bloom strength of 301 g, and a viscosity of 4.96 cp at 6.67% solids content and 60 0C.3 The isoelectric point of the gelatin was 8.93. One gram of dry gelatin contained 0.876 mmol of acid amino acid residues and 0.925 mmol of basic residues.'

The dyes used, their abbreviations and structures are listed in Table 1. HABA-H (lot #165496 780, Fluka Chemicals, Buchs, Switzerland) was neutralized to the sodium salt, HABA, with NaOH to pH 7.00. HASA (lot #MANOl, TCl, Tokyo, Japan) was 98% pure and used as supplied.

e Abstract published in Advance ACS Abstracts, December 1, 1993.

0-11 (lot #237189 287, Fluka Chemicals) and AR-88 (Ref N13693, Atlantic Industries Inc., Nutley, NJ) were recrystallized from anhydrous methanol and dried over phosphorus pentoxide to constant weight. A buffer, 0.15 or 0.075 M ammonium acetate, was used to maintain the pH at 7.00. All chemicals were ACS reagent grade. The water was double distilled in an all-glass still.

Surface Tension-Because both the gelatin and the dyes are surface active in aqueous solution, surface tension measurements were made to study the effect of binding on their respective surface properties. All surface tension measurements were made with solutions containing 0.15 M ammonium acetate a t 40.0 "C. The tensiometer was a Wilhelmy balance (Rosano Surface Tensiometer, V. W.R. Scientific) equipped with a thin rectangular sandblasted platinum blade, which was cleaned by rinsing and flaming. The tensiometer was calibrated with double-distilled water at 40.0 "C, which has a surface tension of 69.6 dyne/cm. Before measuring surface tensions, all solutions were stored overnight in crystallizing dishes at 40.0 "C. Their surfaces were cleaned by suction with fine glass capillaries. The surface tension of dye solutions without gelatin was measured immediately after cleaning their surface. The surface tension of solutions containing gelatin was measured 2 h after cleaning their surface because the macromolecular gelatin diffuses slowly, so that 1-1.5 h was required to attain constant surface tension. The reproducibility of the measurements was within f O . l dyne/cm.

Viscosity-The effect of dye binding on the viscosity of gelatin solutions was studied because gelatin is used as a viscosity builder in liquid and semisolid pharmaceutical preparations. The flow times of the solutions were measured with size 50 glass capillary viscometers (Cannon Instrument Company, State College, PA) at 45.0 "C. The viscometers were calibrated with water and 5.0 % (w/w) sucrose solutions at 45.0 f 0.2 "C, where their viscosities were 0.596 and 0.668 cp, respectively.4

The density of the solutions was measured with 25-mL pycnometers at the same temperature. The pycnometers were calibrated with double- distilledwater at45.0 "C, where itsdensityis0.9902g/mL. Themeasured density of 1.0096 g/mL for a 5.0% sucrose solution was identical with the published value.4

The viscosity, q, of all solutions was calculated from the average flow time, t , of three consecutive concordant measurements and the density, p , based on the following equation:

(1)

In eq 1, K , the instrument constant, equals q$p8ts, the subscript u refers to the unknown solution whose viscosity is being measured, and s refers to the standards; namely, water and a 5.0% sucrose solution.

Gel Rigidity-Especially in food applications, gelatin is used to form gels. Because such gels are rated according to their rigidity or strength, it became of interest to investigate the effect of dye binding on the rigidity of gelatin gels. The measurements were made with a pene- trometer (Micrometer Adjustment Penetrometer, Thomas Scientific, Swedesboro, NJ) equipped with a custom-made conical aluminum plunger. The cone, which was attached to a 6.7-cm stem, had a base diameter of 3.5 cm and a height of 4.1 cm. The length of the edge from the tip to the plane of the base was 4.5 cm, and the solid angle was 45". The combined weight of the cone and the stem was 40.5 g.6 The depth of penetration of the cone, measured 30 s after its release into the gels, was expressed in millimeters.

The gelatin gels with and without dye contained 3.0% (w/v) gelatin. They were prepared by heating 75-mL aliquots of 3.0% gelatin solutions in covered cylindrical jars of 5.5-cm inside diameter and 6.0-cm height at 50 "C for 10 min to erase their memory. The samples were then cooled to the test temperature of 10 "C at the rate of 5 "C/h and stored at this temperature for 18 h before measuring their rigidity. Each sample was used for a single measurement of penetration depth, and three

vu = ~ & , / p , t , = KPutu

316 /Journal of Pharmaceutical Sciences Vol. 83, No. 3. March 1994

0022-3549/94/ 1200-3 16$04.50/0 @ 1994, American Chemical Society and American Pharmaceutical Association

Table 1-Dyes and Their Structure

Molecular Dye Structure Weight Symbol

2-(4-Hydroxyphenylazo)benzoic acid, sodium saR 264.24 HABA

COONa 4-(4-Hydroxyphenylazo)sulfonlc acid, sodium salt

4-(2-Hydroxy-l-naphthyiazo)benzenesulfonic acid, sodium salt (C.I. 155 10)

e(2-Hydroxy- 1-naphthylaz0)- 1-naphthalenesulfonic acid, sodium salt (C.I.

350.39 0-11 OH

400.39 AR-00

samples were measured for each composition. The reproducibility of individual penetration measurements was within &2%.

Gel Melting Point-The effect of dyes on the gel melting point was studied as a function of dye binding. The gel melting point was taken as the temperature a t which a neoprene sphere placed on the top of a gel fell through it.6 Twenty-milliliter aliquots of 2.0% (w/v) gelatin solutions were placed in stoppered 25-mL test tubes and heated at 50 "C for 10 min. They were then cooled to 0 "C at the rate of 5 "C/h, and the solidified gels were stored at that temperature for 18 h before measuring their melting point.

A ball of filled neoprene rubber 1.00 cm in diameter and weighing 0.98 g was placed on top of each gel, and the gels were heated at the rate of 5 OC/h. The temperature a t which the neoprene ball fell through the gel and touched the bottom of the test tube was taken as the melting point. The reproducibility of the gel melting point was within &O.l "C.

Moisture Sorption-The moisture content is an important char- acteristic of solid gelatin. Therefore, the effect of dye binding on the moisture sorption of gelatin films was studied at 25 "C. Blank gelatin films and gelatin films containing different concentrations of 0-11 or AR-88 were investigated. HABA and HASA were not included in these studies because, owing to their smaller aromatic hydrocarbon moieties, they were less extensively bound to gelatin.1

The films were prepared by casting 20% (w/v) gelatin solutions in 0.15 M ammonium acetate, which had been heated to 40 OC, onto a Teflon-coated tray with a doctors knife with a 0.020-inch clearance. The cast gelatin layers set to gels on cooling and were air dried at room temperature. The thickness of the air-dried films varied between 0.38 and0.41 mm. Stripsofthese films wereequilibratedatroomtemperature over phosphorus pentoxide (-0% relative humidity) or over saturated solutions of sodium bromide (58% relative humidity).

Some of the air-dried films that had been equilibrated at 58% relative humidity over saturated sodium bromide solution were re-equilibrated at -0% relative humidity over phosphorus pentoxide. Likewise, some of theair-dried films that had been equilibrated at -0% relative humidity were re-equilibrated at 58% relative humidity. Moisture contents were determined by heating the films at 108 "C to constant weight.

Results and Discussion Surface Tension-The surface tension data of the ternary

system gelatin-dye-buffer solution provided information on the effect of gelatin binding of the dyes AR-88 and 0-11 on the critical micelle concentration (cmc of the dyes as well as on the surface properties of gelatin.

Effect of Gelatin on the cmc of Dyes-The surface tensions of solutions of 0-11 and AR-88 were measured in the presence and absence of 1.0% (w/v) gelatin as a function of dye concentration. The plots in Figure 1 at first decreased linearly with increasing dye concentration and then leveled off abruptly.

75 i

0 0.04 0.08 0.12

Free dye concentration, mol/L

Figure 1-Effect of 1.0% (w/v) gelatin on the surface tension of dye solutions in 0.15 M ammonium acetate at 40.0 "C. The dye concentrations refer to free dye. Key: (A) 0-11 in 0.15 M ammonium acetate; (A) 0-11 in 0.15 M ammonium acetate containing 1 .O % gelatin; (0) AR-88 in 0.15 M ammonium acetate; (0) AR-88 in 0.15 M ammonium acetate containing 1.0% gelatin.

The breaks indicate that the dyes formed micelles. The cmc values of the blank dye solutions without gelatin were 5.81 X 10-2 and 8.88 X 103 M, respectively, for 0-11 and AR-88. The corresponding cmc values of the free dyes in the presence of gelatin were 5.69 X 10-2 and 8.99 X 10-3 M. Thus, gelatin had a negligible effect on the micellization of the two dyes.

Effect of Dye Binding on the Surface Activity of Gelatin-To study the effect of the two dyes with the largest hydrocarbon moieties on the behavior of gelatin monolayers at the air-water interface, the surface tension of 1.0% (w/v) gelatin solutions in 0.15 M ammonium acetate at 40.0 "C was measured as the function of the concentration of added dyes. The dyes AR-88 and 0-11 lowered the surface tension of the gelatin solutions just as the gelatin had lowered the surface tension of the dye solutions.

The net effect of the bound dye on the surface tension of gelatin, Au, was calculated by subtracting the surface tension lowering caused by the free dye from the surface tension lowering caused by the total dye:

(2)

In eq 2, u represents surface tension and the subscripts B, G, D, and F refer to the 0.15 M ammonium acetate buffer, gelatin, bound dye, and free dye, respectively. The term in the first parentheses in eq 2 represents the combined effect of bound and

= ( U B ~ - UBGDF) - (UB - UBF)

Journal of phsrmaceutical Sciences / 3 17 Vol. 83, No. 3, March 1994

Table 2-Effect of Dye Concentratlon on the Surface Tension of 1% (w/v) Gelatln Solutlons In 0.15 M Ammonlum Acetate

Total Dye Conc., Bound Dye Conc., (UBG - haBCI)Fh (ha6 - UBFh A ha,

0-11 2.88 0.18 0.0 0.0 O.Od

0-11 5.70 4.50 2.3 0.2 2.1 0-11 14.0 4.50 4.1 2.0 2.1

Dye moi/L X lo3 mol/g gelatin X lo4 dynelcmaVb dynelcma*c dynelcma

AR-88 3.07 0.28 2.4 0.1 2.3 AR-88 5.64 5.60 4.4 0.2 4.2 AR-88 7.13 6.90 5.9 0.3 5.6 AR-88 8.57 7.50 7.8 1.7 6.1 AR-88 9.12 7.50 8.7 2.6 6.1

a Terms defined by eq 2 and explained in text. haffi = 51.6 dynelcm. uB = 71.0 dynelcm. dThe observed surface tension lowering of 0.03 dynelcm is not significant.

free dye on the surface tension of gelatin. The term in the second parentheses corrects the first term for the lowering of the surface tension of the ammonium acetate solution by the free dye. The values of the three terms of eq 2 at different dye concentrations are listed in Table 2. The surface tension of gelatin solutions decreased with increasing concentration of the bound dyes. At the maximum binding of 7.5 X 104 mol of AR-88/g of gelatin and 4.5 x 104 mol of 0-II/g of gelatin at 40.0 O C in 0.15 M ammonium acetate,l the bound dyes lowered the surface tension of the gelatin solution by 6.1 and 2.1 dynelcm, respectively. Beyond these binding limits, further increases in the total dye concentrations did not reduce the surface tension any further. The net reduction of the surface tension of gelatin, Aha, indicates that dye binding rendered the gelatin more surface active and, hence, less hydrophilic.

Similar values can be obtained from Figure 1. At a concen- trationof 1.0% (w/v), gelatin lowered the plateausurface tensions (i.e., the surface tensions at free dye concentrations > cmc) of buffered 0-11 and AR-88 solutions by 21.4 and 25.5 dynelcm, respectively, whereas it lowered the surface tension of the dye- free buffer solution by only 19.4 dynelcm. For 0-11, Aha = 21.4- 19.4 = 2.0 = 2.1 dynelcm, and for AR-88, Aha = 25.5-19.4 = 6.1 dyne/cm. This estimate of the increase in the surface activity of gelatin by the two chemisorbed dyes is equivalent to the use of eq 2.

Viscosity-The effect of dye binding on the intrinsic viscosity of gelatin solutions was studied as a function of the bound dye concentration for the four dyes. The viscosity of solutions containing 0.05, 0.1, 0.2, and 0.4% (w/v) gelatin in 0.075 M ammonium acetate and constant concentrations of bound dyes (expressed in mollg of gelatin) was measured at 45.0 "C. The concentration of the buffer was halved to minimize its swamping salt effect: Simple electrolytes tend to reduce the ionic nature of polyelectrolytes. Lowering the buffer concentration to values of <0.075 M impaired its buffer capacity. Furthermore, enough sodium chloride was added to the 0.075 M ammonium acetate buffer to maintain the combined concentrations of free dye plus NaCl constant at 3.75 mM. It was important to maintain a constant ionic strength because the dimensions of randomly coiled polyelectrolyte chains depend strongly on the concen- tration of any added simple electrolyte. The medium containing 3.75 mM of combined free dye plus NaCl in 0.075 M ammonium acetate at 45.0 "C was used in all viscosity measurements; it is subsequently referred to as the base solution.

The relative viscosity, orel, is the ratio of the viscosity of the solutions containing gelatin in the base solution with or without dye to the viscosity of the base solution containing the same amount of free dye. The reduced viscosity, vred, is defined as follows:

(3) In eq 3, C is the gelatin concentration, expressed in g1mL. The intrinsic viscosity, [TI, is the reduced viscosity at infinite dilution:

2501 P -I E I 'l

50 4 0.0 0.2 0.4

Gelatin concentration, glmL x 10' Flgure 2-Reduced viscosity of gelatin solutions containing constant levels of AR-88 versus gelatin concentration. Key: The numbers ahead of the parentheses represent bound dye concentrations, expressed in mollg gelatin X 1 04. The percentage values between parentheses represent percent of saturation binding: (0) blank (no dye added); (0) 2.04 (22%); (a) 4.03 (44%); (A) 6.59 (71%); (A) 8.85 (96%).

(4)

The units of 77red and [v] are mL/g of gelatin. For dilute solutions of nonionic polymers (and of polyelec-

trolytes in the presence of swamping concentrations of simple electrolytes), plots of reduced viscosity versus polymer concen- tration are usually straight lines with positive slopes. However, in the case of solutions of cationic polyelectrolytes containing little or no simple electrolytes, increasing dilution makes available increasing volumes to the anionic counterions, which can distribute themselves at larger distances from the cationic protonated amine and guanidine groups attached to the polymer backbone. As a consequence, the shielding of the fixed positive charges by the counterions is reduced, and the randomly coiled polyions expand. This process increases the reduced viscosity with increasing d i l~ t ion .~ This behavior is exemplified in Figure 2. Similar plots, with somewhat shallower curves, were obtained for the other three dyes.

Because of the difficulty in determining intrinsic viscosities by extrapolating the curves of Figure 2 to zero gelatin concen- tration, the Fuoss equation for polyelectrolyte solutions was employed:8

(5) In eq 5, the constant A is the intrinsic viscosity, and the empirical constant B is a measure of the efficiency of shielding within the coils.8 To yield linear plots, the Fuoss equation is inverted. For polyelectrolyte systems that obey the Fuoss equation (including the present one), plots of the reciprocal of the reduced viscosity versus the square root of the gelatin concentration are straight lines with slopes equal to B/A and intercepts of 1/A equal to the

3 18 / Journal of Pharmaceutical Sciences Vol. 83, No. 3, March 1994

X -1 1.2 E rn . .- 5 1.0

8 ln

ln

0 u

‘f 0.8

3 0.6

!? T-

0.4 4 2 3 4 5 6 7

dGelatin concentration, (g/mLf”x 10’

Flgure 3--Reciprocal of the reduced viscosity of gelatin solutions containing constant levels of AR-88 versus square root of the gelatin concentration. Key: same as In Figure 2.

Table 3-Effect of Bound Dye Concentratlon on Intrlnslc Vlscoslty of Gelatln and Constant B of Fuoss Equatlon

Intrinsic Bound Dye Conc., Viscosity, B,

Dye moi/g gelatin X lo4 mL/g (mL/g)l’* a rb

None 0 303 25 0.9995 HABA 2.05 156 9 0.9987 HABA 3.95 121 8 0.9986 HABA 6.61 115 7 1 .oooo HASA 2.70 147 10 0.9988 HASA 4.26 123 8 1 .oooo HASA 7.30 109 7 0.9998 6.11 0.33 296 22 1 .oooo 0-11 2.16 157 12 0.9998 0-1 I 5.56 126 7 1.0000 0-11 8.28 112 8 0.9997 AR-88 2.04 228 18 1 .oooo AR-88 4.03 172 13 0.9987 AR-88 6.59 141 12 1 .oooo AR-88 8.85 131 11 1 .oooo

a Defined by eq 5. Linear regressbn coefficients refer to the straight lines according to the reciprocal Fuoss equation (Figure 3) and are based on the four gelatin concentrations of 0.05,0.1,0.2, and 0.4 % (w/v).

reciprocal of the intrinsic viscosity (see Figure 3). The values of these constants are listed in Table 3.

Gelatin at pH 1.9 units below its isoelectric point is a positively charged polyelectrolyte. The dimensions of its randomly coiled chains are larger than those of a nonionic polymer of comparable molecular weight and chain stiffness (or of the gelatin in the presence of swamping concentrations of simple electrolytes) because of the electrostatic repulsion between the ionized arginine and lysine repeat units. Ion pairing of these cationic functional groups with dye anions reduces the net positive charge of the randomly coiled gelatin chains and causes them to contract. Consequently the intrinsic viscosity, which is a measure of the dimensions of the randomly coiled macromolecule or of its hydrodynamic volume, is also reduced.

The intrinsic viscosity decreased with increasing concentra- tions of bound dyes (Figure 4). At a comparable degree of binding, the largest dye molecule (AR-88, with two naphthalene rings) resulted in the highest intrinsic viscosity, whereas binding of the two smallest dye molecules (HASA and HABA, with two benzene rings) resulted in the lowest intrinsic viscosity. Binding of 0-11, which has an intermediate molecular size (one benzene and one naphthalene ring), resulted in intermediate values. To explain this pattern, it is assumed that the dye anions (which

4001

300

200 -

100’ . ’ . ‘ 0 2 4 6 8 1 0

4 Mol bound dyelg gelatin x 10

Flgure 4-Effect of binding of four dyes on the intrinsic viscosity of gelatin. Key: (0) AR-88; (0) 0-11; (B) HASA; (0) HABA.

4001 380-

360 -

340 -

280

2604 . - . . 0 2 4 6 8

4 Mol bound dydg gelatin x 10

Flgure 5-Effect of binding of four dyes on the penetration depth of 3.0% (w/v) gelatin gels. Key: (0) AR-88 (0) 0-11; (m) HASA; (0) HABA.

are flat molecules) are chemisorbed in a coplanar fashion onto the polypeptide chains (which may be in an extended planar zig-zag conformation). Such orientation allows for maximum dye-gelatin attraction via dispersion forces and hydrogen bonds. The larger adsorbed dye molecules were more effective than the smaller ones in rendering the segments of the gelatin chain on which they lie more rigid and more extended. This resulted in increased chain dimensions and, hence, in a higher intrinsic viscosity. Geometric considerations indicated that, at saturation binding, - 28 % of one of the two flat surfaces of the extended planar gelatin chains is covered with adsorbed dye molecules.’ In Figure 4, the bound dye concentration is expressed in mol of bound dyelg of gelatin. If the units were g of bound dyetg of gelatin instead, the curves would be drawn slightly closer together because the molecular weight of the AR-88 anion exceeds that of the 0-11 anion by 13%, and the molecular weight of the 0-11 anion exceeds that of the HASA anion by 15%.

Gel Rigidity-To study the effect of binding of the four dyes on the rigidity of gelatin gels, the penetration depth of the penetrometer probe into 3.0% (w/v) gelatin gels was measured at 10 O C as a function of the bound dye concentration. All penetration depths were measured in triplicate. The average values f S D were 285 f 8, 293 f 6, and 291 f 7 mm for blank gelatin gels and for gels containing HABA and HASA, reapec- tively. According to t and F tests, the differences between the three sets of values are not significant at the 5% probability level.

However, the penetration depths into gels containing AR-88 and 0-11 were significantly greater than into gels without dye, and increased with increasing dye concentration (Figure 5). Because the penetration depth is inversely related to the gel rigidity, biding of the dyes AR-88 and 0-11 lowered the rigidity of the gelatin gels. A t comparable binding, expressed in mol of

Journal of Pharmaceutical Sciences / 319 Vol. 83, No. 3, March 1994

Table 4-Film Composltlons, Bound Dye Concentrations, and Moisture Contents of Gelatin In Films Stored at 0% and 58% Relative Humidity at Room Temperature

7'0 Moisture Content ( W / W ) ~ ~

After Re-equilibrating at Relative Humidity o f Bound Dye Conc., Before

A. Air-Dried Films

Film Composition, ( w h y mol/g gelatin X lo4 %-equilibratingd 0% 58 %

100 YO Gelatin 0 86.4% Gelatin + 13.6% 0-11 84.7% Gelatin + 15.3% AR-88 77.0% Gelatin + 23.0% AR-88

4.5 4.5 7.5

8.6 1.9 7.2 (0.16)' 1 .O (0.47) 6.2 (0.28) 0.7 (0.63) 4.2 (0.51) 0.5 (0.74)

17.1 14.1 (0.18) 11.6 (0.32) 7.9 (0.54)

B. Films Transferred from 0 % to 58 % Relative Humidity 100% Gelatin 0 1.9 17.2 86.4% Gelatin + 13.6% 0-11 4.5 1 .o 14.1 84.7% Gelatin + 15.3% AR-88 4.5 0.7 11.7 77.0% Gelatin + 23.0% AR-80 7.5 0.5 7.9

100% Gelatin 0 17.2 1.9 86.4% Gelatin + 13.6% 0-11 4.5 4.1 0.9

C. Films Transferred from 58 % to 0 % Relative Humidity

84.7% Gelatin + 15.3% AR-88 4.5 11.6 0.7 77.0% Gelatin + 23.0% AR-88 7.5 7.9 0.5

a Omitting the ammonium acetate buffer, which amounts to 4.6-6.4% of the film weight. In warm casting solutions before cooling and drying. Based on the weight of gelatin; the values are averages of four strips; the SDs are not listed because they amounted to <O. 1 % .d Refers to conditions from which film strips were transferred, listed in the three subheadings (A, B, and C) of this table. ' Numbers in parentheses represent fractional decrease in moisture content due to dye binding,

401

204 ' ' " " '

0 2 4 6 4 a Mol bound dyelg gelatin x 10

Flgure 6-Effect of binding of four dyes on the gel melting point of 2.0 % (wlv) gelatin gels. Key: (0) AR-88; (0) 0-11; (U) HASA; (0) HABA.

bound dye/g of gelatin, AR-88 was more effective in lowering the gelrigidity than 0-11. If the bound dye concentration is expressed in g of bound dye/g of gelatin, the three curves are drawn slightly closer together than those of Figure 5.

Gel Melting Pointg-The melting points of 2.0 % (w/v) gelatin gels were measured as a function of the bound dye concentration. Because that concentration may vary with temperature, i t was measured at the temperature at which the gel melted. The concentrations of the four bound dyes at 10 "C were similar to those at 30 "C. However, between 30 and 45 O C , they decreased with increasing temperature and followed the van't Hoff equation.' For each of the four dyes, the gel melting points decreased with increasing concentration of bound dye. The dyes with the larger hydrocarbon moieties were more effective in lowering the gel melting point (Figure 6). If the bound dye concentrations are expressed in g of bound dye/g of gelatin instead of mol of bound dye/g of gelatin, the curves representing the gel melting points are drawn somewhat closer together than those of Figure 6.

Effect of Dye Binding on Gel Rigidity and Gel Melting Point-Gelatin gels are formed by interchain hydrogen bonds between amide NH and CO groups and by other secondary

calculated with eq 6.

valence forces between adjacent chains at points where segments of such chains come into contact.10 Moreover, when gelatin solutions are cooled below 30 O C , some chain segments crystallize to form the triple-stranded collagen helix, which is also held together by interchain hydrogen bonds and other secondary valence f0rces.l'

Bound dye anions are attached to gelatin chains by electro- valences leading to ion pairing plus secondary valence forces, including hydrogen bonds between the phenolic hydroxyl groups of the dyes and amide groups of the gelatin. The chemisorbed dye anions cover a portion of the polypeptide chains with their aromatic rings, which act as wedges between adjacent chains at points of interchain contact, prevent interchain attraction by hydrogen bonds, and may obstruct the crystallization of triple- stranded helical sequences. This action weakens the gelatin gels and lowers their melting points. Larger amounts of bound dyes (and, at comparable concentrations of bound dyes expressed as mol dye/g gelatin, dye molecules with larger aromatic moieties) cover larger fractions of the surface area of the polypeptide chains. Both factors bring about greater reductions in gel rigidity and melting point. Whether bound dyes actually interfere with the crystallization process in gelatin gels is the subject of current research.

Moisture Sorption-The composition of the gelatin film strips, the concentration of the bound dyes in the casting solutions, and the moisture contents of film strips equilibrated at different relative humidities at room temperature are listed in Table 4. The listed bound dye concentrations refer to the warm gelatin solutions before casting into films, cooling, and air drying. The ratios bound dyelfree dye are presumably higher in the air-dried films than in the corresponding casting solutions. After drying over phosphorus pentoxide, the pure dye powders 0-11 and AR-88 and solid ammonium acetate were conditioned at 58% relative humidity to constant weight. The amounts of moisture absorbed, namely, 0.002, 0.001, and 0.008%, respec- tively, were negligible. Therefore, the moisture contents of the films listed in Table 4 were not corrected for the moisture absorbed by the free or bound dyes nor by the ammonium acetate. The percentages of moisture content are based on the weight of gelatin taken as 100%.

320 /Journal of Pharmaceutical Sciences Vvl. 83, No. 3, March 1994

Film strips re-equilibrated at 58% relative humidity over saturated sodium bromide had the same moisture content regardless of whether they were air-dried or dried over phos- phorus pentoxide before re-equilibration over saturated sodium bromide (compare the last column in Sections A and B of Table 4). Likewise, film strips re-equilibrated at -0% relative humidity over phosphorus pentoxide had the same moisture content regardless of whether they were air-dried or equilibrated over saturated sodium bromide before re-equilibration over phosphorus pentoxide (compare Column IV in Sections A and C of Table 4). Because the moisture contents of the gelatin in the film strips were independent of the prior conditioning treatment, they represent equilibrium values.

The fractional decreases in moisture content due to dye binding, listed in parentheses in Table 4, were calculated as follows:

(MBG - MBGDF)/MBG (6) In eq 6, M represents percent moisture content based on the weight of gelatin in the films; for instance (1st and 2"d row in Section A of Table 4)) air-dried film strips without dye had a moisture content of 8.6% compared with 7.2% for strips containing 13.6% 0-11. Both percentage values are based on the weight of the gelatin in the films. The fractional decrease in the moisture content of gelatin due to dye binding, which is shown in parentheses in Table 4, was (8.6-7.2V8.6 = 0.16.

Higher dye concentrations resulted in lower moisture contents; for instance, the moisture content of air-dried films containing 15.3% of AR-88 (4.5 X 104 mol of bound dye/g of gelatin) was 6.2 % , compared with 8.6 96 for the films containing gelatin alone (Column 111, Rows 3 and 1 in Section A of Table 4). An increase in the dye concentration to 23.0% (w/w) (7.5 x 104mol of bound dye/g of gelatin) lowered the moisture content further to 4.2% (ColumnIII,Row4inSectionAofTable4). Thus, the fractional decrease in moisture content increased with increasing dye concentrations. The reduction in the moisture content of gelatin due to chemisorption of dyes indicates that the dye binding rendered the gelatin less hydrophilic. At comparable levels of dye binding the dye AR-88, which has a larger hydrocarbon moiety than the dye 0-11, was also more effective in lowering the moisture content of gelatin. This is true even when the concentration of bound dye is expressed as g of dye/g of gelatin rather than as mol of dye/g of gelatin.

Conclusions The dyes were bound to gelatin by ion pairing, hydrogen bonds,

and dispersion forces. Dye binding due to ion pairing reduced

the ionic nature of the gelatin, as shown by progressively lower intrinsic viscosities with increasing concentrations of bound dyes. The attachment of the aromatic hydrocarbon rings of the chemisorbed dye molecules to gelatin by secondary valence forces further reduced its hydrophilic nature. The reduced ionic character and the increased hydrophobicity of the gelatin due to dye binding brought about the reductions in the surface tension of its solutions and in the moisture content of its solid films. The same factors, plus the fact that the cbemisorbed dye molecules acted as wedges between adjacent gelatin chain segments and sterically blocked their interchain attraction via hydrogen bonds and other secondary valence forces, impaired the gel formation in the gelatin solutions. This was shown by reductions in the rigidity of the gels and by decreases in their melting points.

References and Notes 1. Gautam, J.; Schott, H., submitted for publication in J. Pharm. Sci. 2. Ofner, C. M., 111; Schott, H. J. Pharm. Sci. 1987, 76, 715-723. 3. Keenan, T. R., Kind and Knox, Division of Knox Gelatine, Inc.,

4. Bingham, E. C., Jackson, R. F. Bur. Standards Bull. 1918,14,59-

5. Schott, H.; Ofner, C. M., 111, Drug Cosmetic Znd. 1975.11 7,42-48.

Sioux City, IA, personal communication, Oct. 11, 1988.

86.

6. Bello, J.; Riese, H. C. A.; Vinograd, J. R. J. Phys. Chem. 1956,60, 1299-1306.

7. Morawetz, H. In Macromolecules in Solution; 2nd ed.; Wiley-

8. Fuoss, R. M. J. Polym. Sci. 1948, 3, 602-603. 9. The gel melting point is a gel-sol transition. The term refers to

the temperature of collapse of the gelatin network. There is no evidence that it coincides with the melting point of the triple helical crystalline regions (see Stainsby, G. In The Science and Technology of Gelatin; Ward, A. G.; Courts, A., Eds.; Academic: London, 1977; pp 189-205).

10. Bello, J.; Bello, H. C. A.; Vinograd, J. R. Biochim. Biophys. Acta

11. Djabourov, M.; Leblond, J.; Papon, P. J. Phys. Paris 1988, 49,

Interscience: New York, 1975; p 360.

1962,57,222-229.

319-332.

Acknowledgments

The gift of the gelatin sample and its characterization by Kind and Knox is gratefully acknowledged. Adapted in part from a thesis submitted by J. Gautam to Temple University in partial fulfillment of the Doctor of Philosophy Degree requirements. Presented at the American Association of Pharmaceutical Scientists Sixth Annual Meeting, Washington, D.C., Nov. 17-22, 1991.

Journal of Pharmaceutical Sclences / 321 Vol. 83, No. 3, March 1994