Anal. Chem. 1986, 58,631-636
responds to changes in solvent conditions and structure of the alkyl ligands.
ACKNOWLEDGMENT The authors acknowledge the generous help of S. Saavedra who assisted with the frontal elution experiments to confirm surface concentration control, J. Mullaney who packed the monomeric C18 column for k'measurements, and J. M. Kime who carried out comparison fluorescence studies of the mobile phase environment; a separate manuscript describing the solution phase studies is in preparation. Registry No. Pyrene, 129-00-0. LITERATURE CITED (1) Horvath, C.; Melander, W.; Molnar, I.J . Chromatogr. 1976, 125, 129. (2) Karger, 6.;Grant, J. R.; Hartkopf, A,; Welner, P. J . Chromatogr. 1978, 138 .-- , 65 - -. (3) Horvath, C.; Melander, W. J . Chromatogr. Sci. 1977, 15, 393. (4) Hemetsberger, H.; Maasfeld, W.; Ricken, H. Chromatographia 1976, 9 , 303. 15) Lochmuller, C. H.; Wilder, D. R. J . Chromatogr. Scl. 1979, 17, 574. I Scott, R. P. W.; Simpson, C. F. J . Chromatogr. 1980, 197, 11. 1 Spacek, P.; Kubin, M.; Vozka, S.; Porsch, 6. J . Liq. Chromatogr. 1980. 3. 1465. McCormick, R . M.; Karger, 6.L. J . Chromatogr. 1980, 199, 259. Gilpln, R. K.; Squires, J . A. J . Chromatogr. Scl. 1981, 19, 195. Gilpin, R. K.; Gangoda, M. E.; Krlshen, A. E. J . Chromatogr. Sc;. 1982,2 0 , 345. Wise, S.A.; May, W. E. Anal. Chem. l983,55,1479. Snyder, L. R.; Ward, J. W. J . Phys. Chem. 1988, 70, 3941. Hair, M. L.; Hertl, W. J . Phys. Chem. 1969, 73, 2372. Hertl, W.; Halr, M. L. J . Phys. Chem. 1971, 7 5 , 2181. Scott, R. P. W.; Kucera, P. J . Chromatogr. 1979, 171, 37. Sander, L. C.; Callis, J. B.; Field, L. R. Anal. Chem. 1983,55, 1068. Pickett, J. H.; Lochmuller, C. H.; Rogers, L. 6.S e p . Sci. 1970,5 , 23. Stotfeldt-Ellingsen, D.; Resing, H. A. J . Phys. Chern. 1980,8 4 , 2204. Gilpin, R. K.; Gangoda, M. E. J . Chromatogr. Sci. 1983, 2 1 , 352.
631
Gangoda, M. E.;Gilpin, R. K. J . Magn. Reson. 1983,5 3 , 140. Sindorf, D. W.; Maciel, G. E. J . Am. Chem. SOC. 1983, 105, 1848. Gilpin, R. K.; Gangoda, M. E. Anal. Chem. 1984,5 6 , 1470. Marshall, D. 6.;McKenna, W. P. Anal. Chem. 1984,5 6 , 2090. Lochmuller, C. H.; Marshall, D. B.; Wilder, D. R. Anal. Chim. Acta 1981, 130, 31. (25) Lochmuller, C. H.; Marshall, D. 6.;Harris, J. M . Anal. Chim. Acta 1981, 131, 263. (26) Lochmuller, C. H.; Colborn, A. S.; Hunnicutt, M. L.; Harris, J. M. Anal. Chem. 1983,55, 1344. (27) Lochmuller, C. H.; Colborn, A. S.; Hunnicutt, M. L.; Harris, J. M. J . A m . Chem. SOC. 1984, 106, 4077. (28) Dowling, S. D.; Seitz, W. R. Anal. Chem. 1985,5 7 , 602. (29) Bogar, R. G.; Thomas, J. C.; Callis, J. B. Anal. Chem. 1984, 5 6 , 1080. (30) Rice, M. R.; Gold, H. S. Anal. Chem. 1984,5 6 , 1436. (31) Dong, D. C.; Winnik, M. W. Photochem. Photobiol. 1984,3 5 , 17. (32) Langkiide, F. W.; Thulstrup, E. W.; Michl, J. J . Chem. Phys. 1983, 78, 3372. (33) Ham, J. S.J . Chem. Phys. 1953,2 1 , 756. (34) Platt, J. R. J . Mol. Spectrosc. 1962,9 , 288. (35) Stahlberg, J.; Almgren, M. Anal. Chem. 1984,5 7 , 817. (36) Perrin, D. D.; Marego, W. F. A.; Perrin, D. R. "Purification of Laboratory Chemicals", 2nd ed.; Permagon Press: New York, 1980; p 426. (37) Ward, R. L. J . Am. Chem. SOC. 1981, 8 3 , 1296. (38) McCormick, R. M.; Karger, 6.L. Anal. Chem. 1980,5 2 , 2249. (39) Huber, J. F. K.; Gerritse, R. G. J . Chromatogr. 1971,5 8 , 137. (40) Knox, J. H.; Hartwick, R. A. J . Chromatogr. 1981,204, 3. (41) Hunnicutt, M. L.; Harris, J. M. Lochmuller, C. H. J . Phys. Chem. 1985, 89,5246. (42) Synder, L. R.; Kirkland, J. J. "Introduction to Modern Liquid Chromatography", 2nd ed,; Wiley: New York, 1979. (20) (21) (22) (23) (24)
RECEIVED for review August 7, 1985. Accepted October 28, 1985. Support for this research was provided by grants from 3M Corp. and the Office of Naval Research, by the donors for the Petroleum Research Fund, administered by the American Chemical Society, and by fellowship support (to J.M.H.) from the Alfred P. Sloan Foundation.
Determination of Reactive Aldehyde Groups Immobilized on Silica Using ( p-Nitrophenyl)hydrazine as a Chromophoric Probe Matthew C. Gosnell and Horacio A. Mottola"
Department of Chemistry, Oklahoma State University, Stillwater, Oklahoma 74078
One of the slmplest, most gentle, and rapld methods for covalent attachment of enzyme molecules to polar surfaces Involves the use of glutaraldehyde as a blfunctlonal reagent. Dlfferent aspects of this, as well as a method for the determination of reactive aldehyde groups after attachment of the aldehyde to an amlnosllylated silica surface, are presented here. The method of determlnatlon involves a Schiff-base type of coupling with ( p -nltrophenyl)hydrazlne used as a probe at a pH of about 5, hydrolytlc removal of the coupled probe at about pH 7, and photometrlc measurement of the released hydrazlne. The method, being nondestructive, Is of special Interest In the determination of reactlve aldehyde groups on controlled-pore glass. The method is less successful when applied to the determinatlon of aldehyde attached to borosilicate glass etched with hydrogen fluorlde.
The covalent attachment of enzymes (via amino acid residues not essential to the catalytic function of the enzyme) to functionalized supports is the prevalent method for the immobilization of enzymes to be used as analytical reagents.
One of the simplest, most gentle, rapid, and efficient methods for covalent attachment is the so-called "glutaraldehyde attachment" ( I , 2). The detailed chemistry of this procedure is not well understood but the individual steps include the reaction of the silica framework with an aminosilane, modification of the product of this reaction with glutaraldehyde, and finally immobilization of the enzyme, as shown below -0-Si-OH
-
+ (C2H50)3Si(CH2)3NH2
-
-O-Si-O-Si(CH2)3NH2
-NH2
+ OHC-(CH,),-CHO
-N=CH-(
-CHO
+ H2N-E
CH2)3-CHO
-CH=N-E
(1) (2)
(3) An understanding of these processes as well as optimization of the immobilization process and reactor design and performance requires an analytical accounting of the reactive groups attached to the surface after each step has been completed. The step illustrated by eq 2 is critical to enzyme immobilization, since the availability of reactive aldehyde groups after it has been completed largely dictates the amount of protein material amenable to immobilization. The need
0003-2700/86/0358-0631$01.50/00 1986 American Chemical Society
632 ANALYTICAL CHEMISTRY, VOL. 58, NO. 3, MARCH 1986
for a quantitative method to determine the “free” aldehyde groups on the surface has recently attracted attention and a kinetic determination has been proposed ( 3 ) . The method is selective and affords low limits of detection (2-5 pg of aldehyde, or 0.02-0.05 pmol if the weight reported corresponds to an entire glutaraldehyde molecule, per gram of supporting material) and has been termed a “catalytic” method based on the effect of the aldehyde on the oxidation of p-phenylenediamine by hydrogen peroxide. The claimed catalytic action of the aldehyde groups represented a very attractive feature of the method because it would qualify as nondestructive, leaving the aldehyde groups unaffected and ready for protein immobilization. Attempts to use this fixed-time catalytic determination to determine reactive aldehyde groups in open-tubular coil reactors and in single-bead string reactors (1,4), however, resulted in apparent contradictory behavior in a continuous-flow/stopped-flow system. The catalytic response was observed to decrease with repetitive determinations using the same coil with attached glutaraldehyde, and the response for the same coil showed a decreasing catalytic activity from day to day. The same was observed with aldehyde groups immobilized on controlled-pore glass by using the synthetic route illustrated by eq 1-3. A reexamination of the overall chemistry involved pointed out that even in homogeneous aqueous medium the effect of aldehyde was not catalytic but instead was that of a promoter ( 5 ) ,with the aldehyde group being deactivated or destroyed during the course of the reaction (6). An empirical reaction model describing the experimental observations of the promoting effect has been proposed ( 5 ) . Most spectroscopic techniques that could be used for the aldehyde determination either lack sensitivity (e.g., infrared or NMR spectroscopy) or else the topology of controlled-pore glass and tubular reactors and the nondestructive requirement rule out their application (e.g., photoacoustic spectroscopy). An indirect chemical method is left as the only viable alternative. A search was made for such a method using a chromophoric chemical probe that could be stoichiometrically attached to the aldehyde and then removed from the treated surface without impairing the subsequent use of the aldehyde groups for protein immobilization. Imine bond formation from the Schiff-base reaction illustrated by eq 4 and the subsequent release of the aminic material, by hydrolysis at the imine bond (reversal of eq 4),was explored. Results of such a search are
-CHO
+ H,N-chromophore
-
-CH=N-chromophore
+ HzO (4)
presented here and resulted in the development of a satisfactorily reproducible, rather simple and sensitive method of a nondestructive nature applicable to controlled-pore glass and based on the use of (p-nitropheny1)hydrazine as the chromophoric probe. Attempts to utilize the same approach for the determination of aldehyde groups immobilized on borosilicate glass etched with hydrogen fluoride were, however, unsuccessful. EXPERIMENTAL S E C T I O N Apparatus. A Perkin-Elmer Lambda 3840 W-vis linear diode array spectrophotometer operated by a Perkin-Elmer 7300 Professional Computer (Perkin-Elmer, Inc., Norwalk, CT) and a Integral Data Systems P-132printer (Integral Data Systems, Inc., Milford, NH) was used for collection, manipulation, and output of all spectra. Temperature studies were performed with the aid of a Lauda K-2/R constant-temperature bath (Brinkman Instruments, Inc., Westbury, NY) with a glass circulating water bath. Experiments involving controlled-pore glass (CPG) used a mechanical stirrer to avoid destruction of the glass particles by the grinding action of a magnetic stirring bar. The experimentalsetup used for determination on open tubular reactors (OTR) is shown in Figure 1. A Gilson Minipuls 2
1-w
Figure 1. Experimental setup used for determinations in OTRs: (NPH) @-nitrophenyl)hydrazine, pH 5.00 buffer solution; (B) pH 7.10 buffer solution; (P) peristaltic pump; (V) four-way valve; (CTB) constant temperature bath; (OTR) open tubular reactor: (W) waste receiver; (F) flow cell; (S) spectrophotometer: (C) computer; (P) printer.
peristaltic pump (Gilson Medical Electronics, Middleton, WI) was used to pump first the probe solution and then the hydrolysis buffer solution. Flow was switched with a Rheodyne Model 5041 four-way Teflon rotary valve (Rheodyne, Inc., Cotati, CA). All tubing was made of Teflon (Cole-Parmer,Chicago, IL) to prevent adsorption of the probe reagent onto the surface of the tubing. An all-quartz flow cell was used for absorbance measurements. Adjustment of pH was made with an Orion Research Model 601A pH meter (Orion Research, Cambridge, MA) equipped with an epoxy-body combination electrode (Sensorex, Westminster, CA). Reagents and Solutions. All chemicals used were of AR grade. The water used for solution preparation was deionized and further purified by distillation in an all-borosilicate-glassstill with a quartz immersion heater (Wheaton Instruments, Millville, NJ). The hydrochloride of (p-nitropheny1)hydrazine (NPH) was prepared by the reaction between HC1 and NPH. Phoshate buffer solutions were prepared by mixing appropriate volumes of 0.10 M solutions of NaH2P04and Na2HP04(Fisher Scientific, Fair Lawn, NJ) until the desired pH was obtained. The ionic strength of buffer solutions was adjusted by the addition of an appropriate amount of NaC10, (GFS Chemicals, Columbus, OH). (3Aminopropy1)triethoxysilane (Petrarch Systems, Inc., Bristol, PA) was used to silanize the OTRs with the exception of one series of experiments when (aminopheny1)triethoxysilane (Petrarch Systems, Inc., Bristol, PA) was used. The CPG used for glutaraldehyde treatment was purchased with the aminopropyl group already immobilized (Electro-Nucleonics, Inc., Fairfield, NJ). Borosilicate glass beads (1.0 mm, Propper Mfg. Co., Inc., Long Island City 1, NY) were used as an alternative support for glutarladehyde. A 25% w/w stock solution of glutaraldehyde (Aldrich Chemical, Milwaukee, WI) was used for most of the aldehyde immobilizations. (3-[Bis(2-hydroxyethyl)amino]propyl}triethoxysilane (Petrarch Systems, Inc., Bristol, PA) was immobilized in an attempt to provide alcoholic functional groups that could be oxidized to aldehyde groups with chromic acid (J. T. Baker Chemical Co., Phillipsburg, NJ). Sodium cyanoborohydride (Aldrich Chemical, Milwaukee, WI) was used to attempt the reduction of the glutaraldehyde-silane imine bond. Procedure for Surface Preparation of Open Tubular Reactors. A piece of Pyrex glass tubing (8 mm 0.d. and 5 mm id.) was drawn on a capillary drawing machine (Hewlett-Packard, Model 1045A) to produce a coiled glass capillary of 0.7 mm i.d. Whiskers were then grown inside the capillary by hydrogen fluoride treatment (7). The OTR was then filled with concentrated HC1 and heated at 80 “C for 12 h. This acts to convert unreactive siloxane bonds to reactive silanol groups ( 3 ) . Procedure for Surface Preparation of Glass Beads. Glass beads were placed inside a glass tube so that a saturated solution of ammonium bifluoride in methanol (J. T. Baker Chemical Co., Phillipsburg, NJ) could be made to flow through. The beads were dried with a stream of nitrogen. The ends of the tube were sealed with a torch, and it was heated at 450 OC for 3 h. Aldehyde Immobilization Procedure. One milliliter of the appropriate silane was added to 9 mL of 95% ethanol. The solution was pumped through the OTR at room temperature for
ANALYTICAL CHEMISTRY, VOL. 58, NO. 3, MARCH 1986
1 h. The silane was washed from the coil by pumping a few milliliters of 95% ethanol through it followed by a few milliliters of distilled water. For immobilization on glass beads, the beads were added to the silane solution and shaken for 1 h. Excess silanol groups were converted to siloxane bonds by heating at 100 “C for 1h (8). This curing process increases the stability of the silane coating (9). Glutaraldehyde was immobilized by placing about 1g of aminopropyl CPG or glass beads in a vial and adding 20 mL of 2.5% (w/w) aqueous solution of glutaraldehyde. The mixture was shaken for 1h at room temperature. The glass was then filtered out by suction, washed with distilled water and acetone, and allowed to dry. The aldehyde-treated CPG was stored in a desiccator. Immobilization onto OTRs differed only in that the glutaraldehyde solution was pumped through the coil for 1 h. Procedure for Reduction of Imine Bond. The reduction of the glutaraldehyde-silane imine bond was attempted by dissolving 0.5 g of NaBH3CN in pH 5.0 buffer, and pumping this solution into a glutaraldehyde-treated OTR for 1 h. Alternative for Producing ImmobilizedAldehyde Groups Procedure. An alternative method of immobilizing aldehyde groups was attempted by immobilizing (3-[bis(2-hydroxyethyl)amino]propyl)triethoxysilanein an OTR by the same procedure outlined for the immobilization of the other silanes. An oxidizing solution was made by dissolving 5 g of NazCrz07in 10 mL of water and adding 5 g of concentrated HZSO4and just enough water to dissolve any remaining precipitate. The solution was pumped into the OTR and allowed to react for 1 h. The OTR was then washed with water and acetone. Procedure for Determining Blank Readings. Aminopropyl CPG, with no aldehyde, was used to determine a blank reading by using the same experimental conditions given. This was done to determine the limit of detection. Ten blank runs were averaged and the amount of aldehyde that represented the average plus three standard deviations was taken as the limit of detection. Blanks were run, as well, to demonstrate that the probe was not interacting with any species other than the aldehyde. Procedure for Optimization of Conditions. The aldehyde-immobilized CPG was used for initial experiments and optimization of conditions. I t was found somewhat more convenient to use because of its very high surface area (139 m2/g) and its physical form. For the attachment of the probe about 0.10 g of CPG was added to a vial containing a 20-mL solution of 0.030g of NF” in an apropriate buffer. For temperature studies the vial was placed in a recirculating water bath at an appropriate temperature. The mixture was stirred for a fixed amount of time and immediately suction filtered. The CPG was washed with about 5 mL of water, then 5 mL of acetone, and allowed to dry. At this point it was weighed because all comparisons were made on a signal per gram basis. Hydrolysis was performed by adding the CPG to a vial containing 20 mL of buffer of the selected pH. After the mixtue had been stirred for a fixed amount of time, a portion of the solution was added to a quartz cuvette and the spectrum obtained. The progress of each reaction was not directly followed, but the effect a particular parameter had on each reaction was inferred by an absorbance measurement made after the hydrolysis step. The amount of aldehyde present was calculated from the absorbance of the solution at 400 nm. The heterogeneous nature of the CPG gave rise to different results from different portions of the same CPG sample. In order to overcome this problem and permit comparison of different samples several experiments were performed on the same portion of CPG. One of those experiments performed utilized the same conditions for all portions used for the optimization of a given operational parameter. In this manner an evaluation of each condition could be made by first determining the amount of immobilized aldehyde present and then comparing the amount of signal produced per micromole of aldehyde. Procedure for Determinations in Open Tubular Reactors. A continuous-flow/stopped-flow method was employed for experiments with OTRs. Determinations were made by using the same operational parameters found with CPG. The NPH-buffer solution was pumped into the coil and then the flow was stopped. At the end of the reaction time the solution was pumped out of the OTR with the hydrolysis buffer at a high flow rate. Once all the NPH had been pumped out of the coil, the flow was reduced
0.25r
I
633
\
e\
e
0.05
I
0.00‘
I
I
/
I
2
I
3
4
I
I
I
5
6
7
RUN
Flgure 2. Absorbance vs. run for repetitive determinations using Shapilov’s method on CPG.
e‘
0.00 I
1
/
I
I
2 3 INJECTION
e‘
I
I
I
4
5
6
NUMBER
Figure 3. Absorbance vs. injection number for repetitive determinations using Shapilov’s method modified for use with OTRs.
to 0.190 mL/min. During the reaction period a flow cell was used immediately downstream from the OTR so that the absorbance at 400 nm could be monitored with time. The area under the absorbance vs. time plot was computer-integratedby a trapezoidal approach.
RESULTS AND DISCUSSION Considerations on Shapilov’s “Catalytic”Fixed-Time Kinetic Method. Shapilov (3) reported a kinetic method for the determination of aldehyde groups immobilized on glass beads. The procedure utilized the “catalytic” effect aldehyde groups have on the reaction between p-phenylenediamine and hydrogen peroxide to form Bandrowski’s base (IO). Attempts to utilize Shapilov’s “catalytic” method to determine reactive aldehyde groups on CPG and in OTRs resulted in contradictory behavior in batch studies and in a continuous-flow/ stopped-flow system. The signal response (absorbance) was observed to decrease with successive determinations using the same CPG sample or OTR. Implementation of the method, for CPG, reproduced the reported procedure as closely as possible. Figure 2 shows the absorbance obtained by repeatedly performing the procedure on a single sample of CPG. The first run produced an absorbance of 0.230,but by the time the solution had been in contact with the aldehyde-immobilized CPG for 60 min (run 4) the absorbance obtained had dropped to a level about 50% of its original value. Results obtained with this method in OTRs were similar (Figure 3): after a contact time of 60 min (injection 2) the absorbance had dropped, as well, to a value
634
ANALYTICAL CHEMISTRY, VOL. 58, NO. 3, MARCH 1986 /.om
Table I. Results of Determinations Made on Different Portions of a Single Controlled-Pore Glass Sample Using Optimum Experimental Conditions
portion no.
amt of aldehyde, wmol/g
1 2 3 4
3.53 2.58 2.02 3.86 2.81
av st WAVELENGTH
C n m)
Flgure 4. Spectra of (A) 0.10 M phosphate buffer, 1.0 M NaCIO,, (B) 8.58 X M aniline, and (C) 8.32 X M @-nitropheny1)hydrazine.
about 50% of the initial value. This observed behavior of the absorbance dropping to lower values with successive experimenta indicates that the aldehyde group is somehow becoming deactivated in the course of the reaction. These observations led to a kinetic study of the accelerating effect of nonimmobilized glutaraldehyde on the reaction between p-phenylenediamine and hydrogen peroxide (5). This study suggested that glutaraldehyde forms a complex with hydrogen peroxide and that this complex is responsible for the reaction with p-phenylenediamine to form Bandrowski's base. Instead of being first order, as would be expected if a true catalyst were operative, the reaction was found to have an experimental order of 2/3 with respect to the complex. This finding suggests a much more complicated step in the mechanism than a catalytic one. Other evidence for noncatalytic behavior reported includes the fact that the initial reaction rate was not directly related to the glutaraldehyde concentration and that the equilibrium absorbance is not the same for different amounts of glutaraldehyde. Glutaraldehyde, it was concluded, has a promoting effect on the reaction, and both the immobilized and nonimmobilized species are inactivated in the course of the determination. Rationale for a Schiff-Base Type of Chemistry. Schiff-base reactions are used for the coupling of glutaraldehyde as well as enzymes to the surface of silica. Under normal operating conditions (0-25 "C and pH 4-8) the carbon-nitrogen double bond is stable and is not broken by hydrolysis. This has been shown by the retention of enzymatic activity for immobilized uricase for months at a time ( I ) . However, it has also been shown that, under the proper circumstances, hydrolysis of some imines can take place (11). The criterion for reversibility is the nature of the amine and the aldehyde involved. The presence of an aromatic ring in either the amine or the aldehyde facilitates the formation of a four-center nitrogen intermediate and makes hydrolysis more likely. The actual rate and specific conditions necessary for reversibility are dependent on the specific nature of the side chain itself. Choice of Probe. A suitable chromophoric probe of amine type has three structural requirements. First, it must have only one amino group. More than one amino group presents the possibility of cross-linking with neighboring aldehyde groups. Second, the amine must be a primary one in order to undergo the Schiff-base reaction. And third, the amino group should be in the vicinity of an aromatic ring. Several aromatic amines were considered for use as probes. Aniline is a logical first choice. However, aniline absorbs radiation well into the ultraviolet region of the spectrum,,A,( = 234 nm, 6 = 7.9 x lo3M-l cm-I). The absorbance in the ultraviolet region due to NaC10, and the phosphate buffer was too high to permit obtaining quantitative results for small amounts of = aniline in solution (Figure 4). p-Benzeneazoaniline,,A,( 384 nm, = 2.5 x lo4 M-l cm-I) and (2,4-dinitrophenyl)-
dev
n fii
hydrazine (Ama = 349 nm, E = 1.2 x IO4 M-l cm-I) both absorb well into the visible region; unfortunately their hydrochlorides are not sufficiently soluble in water to allow their use as probes. (p-Nitropheny1)hydraine hydrochloride (which forms a hydrazone) is sufficiently soluble in water and it absorbs in the visible region (Figure 4). A molar absorptivity of 1.04 x lo4 M-' cm-I was determined a t a wavelength of 400 nm. Optimization of Operational Parameters. As stated above, the heterogeneous nature of a solid support such as CPG produced different results for different portions of the same aldehyde-immobilized CPG sample. Table I shows the results of determinations made on different portions of a single sample. The inability to directly compare results of different experiments made the implementation of the simplex technique of parameter optimization (12, 13) successful only to a limited extent. Only the pH used for the probe attachment and removal could be narrowed down. The best pH for the attachment reaction was found to be in the range 4-6 and that for hydrolysis was found to be 6-8. Instead of several parameters being optimized a t once, as simplex optimization makes possible, one parameter was optimized at a time by using the method discussed in the Experimental Section. Hydrogen ion concentration has a marked effect on both the attachment and the release of NPH. For the hydrazone formation reaction between pH 2.7 and 5.4 a maximum was observed at pH about 5 . For the hydrolysis step in the pH range 5-8, a maximum was obtained at a pH of about 7. At each of these pH values a maximum signal of 0.5 Alpmol was obtained. Both the hydrazone and hydrolysis reactions can occur, to some extent, at a pH between about 5 and 6. At this pH the hydrolysis step gave a signal of O.lC-O.20 AlpmoI. In this same range the hydrazone formation step gave a final signal of 0.30-0.50 A/pmol. This is probably due to the effect that a low equilibrium constant has on the reaction. During the probe attachment the N P H concentration is high, so the hydrazone formation reaction is forced in the forward direction. The concentration of NPH in the hydrolysis buffer is comparatively low, so the reaction proceeds in the reverse direction. The optimum ionic strength was found to be near 1.0 M for both steps in the reaction. A maximum signal of 0.5 A/bmol was obtained. In general, Schiff-base formations, which are similar to hydrazone formations, and hydrolysis are relatively fast reactions (11),so the rate of reaction with CPG immobilized aldehyde would be greatly affected by diffusion. The two final parameters, time and temperature of reaction, have large influences on a diffusion-controlledprocess. For the hydrazone formation the time was varied on 15-min intervals between 15 and 75 min. The maximum signal was obtained after 30 min, which corresponds with a maximum signal of 0.50 A/ pmol. Instead of reaching a plateau at longer times, the hydrolysis reaction has a maximum of 0.60 Alpmol at 80 min. At times less than 80 min, not ail the NPH has been liberated from the aldehyde. A decrease in absorbance was observed at times greater than 80 min. This is due to the destruction
ANALYTICAL CHEMISTRY, VOL. 58, NO. 3, MARCH 1986
Table IV. Results of Several Determinations on a Single Sample of CPG Treated with Ammonium Bifluoride Using Optimum Experimental Conditions
Table 11. Results of Several Determinations Made on a Single Controlled-Pore Glass Sample Uaing Optimum Experimental Conditions determination no.
amt of aldehyde, pmol/g
run
amt of aldehyde, pmol
1 2 3 4
2.58 2.47 2.86 2.61 2.63 0.14
1 2 3
5.84 4.34 3.59
4
2.73
av std dev
Table 111. Results of Several Determinations on Open Tubular Reactors Using the Same Conditions as with Batch Samples amt of aldehyde, wmol OTR no.
run 1
run 2
1 2
0.062 0.059
3
0.046
0.018 0.023 0.013
of NPH in solution with time. As evidence of this, a plot of absorbance vs. time for NPH in buffer (pH 7.0) decreases with time. In order to optimize temperature of the probe attachment, the temperature was varied between 15 and 70 "C. The hydrazone formation had a maximum value a t 25 "C, which Corresponds to a maximum signal of 0.50 Alpmol. At both higher and lower temperatures the signal was observed to decrease. This can be explained as the result of two competing processes. At temperatures less than 25 "C the signal increases with temperature because such increase improves the ability of NPH to diffuse into the pores of the glass. When the temperature exceeds 25 "C, another process becomes predominant. The equilibrium constant for the hydrazone formation decreases, at higher temperatures, so that the reverse reaction (hydrolysis) is favored. This is supported by the plot of signal vs. temperature for the hydrolysis step in the determination. Between 5 and 70 "C, the intensity of the signal continuously increased with temperature. Operational characteristics set the maximum temperature at 70 "C, which corresponds to a signal of 2.2 Alpmol. Application of Procedure to Aldehyde Immobilized on Controlled-Pore Glass. Because of the need to immobilize protein following a determination, reproducibility is of great importance. Table I1 shows the results of a series of determinations made on a single sample of CPG. These results exemplify an important point. The immobilized aldehyde groups are not significantly destroyed in the course of the determination. Implementation of this method on CPG gave the following equation for the calibration plot:
W = 1.917A - 0.458
635
(5)
where W is the amount of immobilized aldehyde per gram of CPG (pmol/g) and A is the absorbance of the solution per gram of CPG. The limit of detection was found to be 0.30 fimol/g. Application of Procedure on Aldehyde Immobilized in Open Tubular Reactors. The use of OTRs instead of CPG did not produce reproducible results. Single determinations gave values for the amount of glutaraldehyde immobilized ranging between 0.040 and 0.060 pmol (Table 111). However, the next determination on the same OTR produced aldehyde amounts in the range 0.010-0.020 pmol. This behavior indicates that the aldehyde was being lost in the course of the reaction.
It is believed that the cause of the aldehyde loss is reversal of reaction 2, i.e., detachment of the glutaraldehyde from the modified OTR. When only a pH 7 buffer solution is pumped through the OTR at 70 "C, spot tests (10)indicate the presence of aldehyde groups. Also glutaraldehyde can be reimmobilized to an amount in the same range as was detected by the first run of that OTR. This behavior is observed only a t high temperatures. At lower temperatures there is no hydrolysis. This explains previous observations that when enzymes are immobilized there is no loss in enzymatic activity with time. In an attempt to explain this behavior the same procedure used for aldehyde CPG was employed on borosilicate glass beads as well as CPG that had been treated with NH,HF; this is a treatment similar to that to produce whisker growth. A surface effect that is present with borosilicate glass but absent with CPG is the most likely cause of the imine hydrolysis. It should be indicated here that CPG is produced from a borosilicate base material heated to cause phase separation of the borates and silicates. The borate groups are subsequently leached from the material, leaving the porous silica structure. No phase separation or borate leaching is expected to take place in whisker growth procedures. Glass beads also showed the same decrease upon successive determinations of aldehyde. CPG treated with ammonium bifluoride showed a decrease in signal of about 50% after four determinations (Table IV). Although there is some reduction of aldehyde even with CPG, there is much more aldehyde left after a single run than with borosilicate glass. Attempts To Produce a More Stable Immobilized Aldehyde. Sodium cyanoborohydride is a highly selective reducing agent capable of reducing carbon nitrogen double bonds but leaving carbonyl groups unaffected (14). Determinations made after the use of this reagent showed that the imine was not reduced because the amount of aldehyde decreased with each run. It has been shown that commercial (aminopheny1)triethoxysilane contains 70% meta and para isomers of the compound (15). On the reasoning that the rigidity of the silane would keep the imine group away from the silica surface, (aminopheny1)triethoxysilane was immobilized instead of (aminopropy1)triethoxysilane.However the presence of the aromatic ring facilitated the hydrolysis, so that the first determination showed 0.038 kmol but the second showed no aldehyde at all. Finally, an attempt was made to indirectly immobilize aldehyde groups by immobilizing the alcohol, (3-[bis(2hydroxyethyl)amino]propyl)triethoxysilaneon the OTR. Then chromic acid (16) was used to oxidize the alcoholic groups. Experiments performed showed that aldehyde groups has not been generated by this procedure, or at least had not survived it.
LITERATURE CITED (1) Iob, A.; Mottola, H. A. Anal. Chern. 1980, 52,2332-2336. (2) Mottola, H. A. Anal. Chirn. Acta 1983, 745,27-39. (3) Shapilov. 0.D. Zh. Anal. Khim. 1980, 35, 2199-2202; J . Anal. Chern. USSR (Engl. Trans/.) 1980, 35, 1429-1431. (4) Gnanasekaran, R.; Mottola, H. A. Anal. Chern. 1985, 57, 1005-1009. (5) Thornpsen, J. C.; Mottola, H. A. Anal. Chern. 1984, 5 6 , 2834-2836. (6) Eswara Dutt, V. V. S.; Mottola, H. A. Anal. Chern. 1974, 46, 1090-1094.
636
Anal. Chem. 1986, 58,636-638
(7) Iob, A.; Mottola, H. A. Clin. Chem. (Winston-Salem, N . C . ) 1981, 27, 195. (8) Ishlda, H.; Koenig, J. L. J. Colloid Interface Sci. 1978, 64, 565-576. (9) Waddell, T. G.; Leyden, D. E.; DeBello, M. T. J . Am. Chem. SOC. 1981, 103, 5305-5307. (10) Feigl, F.; "Spot Tests In Organic Analysis", 7th ed.; Elsevier: New York, 1966; pp 198-199. (11) Cordes, E. H.;Jencks, W. P. J. A m . Chem. soc. 1963, 85, 2843-2848. (12) Deming, S. N.; Parker, L. R., Jr. CRC Crit. Rev. Anal. Chem. 1978, 7, 187-202.
(13) Shavers, C. L.; Parsons, M. L.; Deming, S. N. J. Chem. €doc. 1979, 56,307-309. (14) Lane, C. F. Synthesis 1975, 135-146. (15) Marshall, M. A.; Mottola, H. A. Anal. Chem. 1985, 57,375-376. (16) Brown, H. C.; Garg, C. P.; Liu, K. J. Org. Chem. 1971, 3 6 , 387-390.
RECEIVED for review August 19, 1985. Accepted October 30, lg85*This work was supported by a grant from the National Science Foundation (Grant No. CHE-8312494).
Subambient Temperature High-Performance Liquid Chromatographic Determination of the Enantiomers of (f)-((6,7-Dichloro-2,3-dihydro-2-methyl- I-oxo-2-phenyl4 Hinden-5-yl)oxy)acetic Acid David J. Mazzo,* C a r l J. Lindemann, a n d Gerald S. B r e n n e r Department of Pharmaceutical Research and Development, Merck Sharp and Dohme Research Laboratories, West Point, Pennsylvania 19486
The enantiomers comprising the antihypertensive compound MK-286 were determined by subambient temperature hlghperformance liquid chromatography. A Pirkie covalent Lphenylglycine chirai stationary phase was used in conjunction with a predominantly (90% ) organic buffered ternary reversed-phase mobile phase to effect the separation. Subambient column temperatures to an optimum of 0 O C improved enantiomer resolutlon and provided reproducible chromatography. Under the assay conditions the separation was accompllshed in less than 15 mln. Both enantiomers were detectable to the low pg/mL concentratlon range with an assay accuracy within f2% and an assay precision (relative standard deviation) of
