Derivatization of Steroids with Dansylhydrazine Using

Anal. Chem. 1997, 69, 4905-4911

Derivatization of Steroids with Dansylhydrazine Using Trifluoromethanesulfonic Acid as Catalyst Patrik Appelblad,†,‡ Einar Ponte´n,† Hans Jaegfeldt,§ Torbjo 1 rn Ba 1 ckstro 1 m,‡ and Knut Irgum*,†

Department of Analytical Chemistry, University of Umea˚, S-901 87 Umea˚, Sweden, Department of Obstetrics and Gynecology, University Hospital, S-901 85 Umea˚, Sweden, and Bioanalytical Chemistry, Astra Draco AB, Box 34, S-221 00 Lund, Sweden

A new dansylation reaction, where trifluoromethanesulfonic acid is used as catalyst, has been characterized for six ketosteroids by employing experimental design followed by multivariate data analysis. The molar ratio between the steroid and the derivatization reagent was found to be the factor most strongly affecting the reaction. Faster reaction kinetics was achieved when the molar ratio between dansylhydrazine and the steroid was increased. Mass spectroscopic analysis showed that the dual peaks observed when derivatized progesterone was separated on an octadecyl silica stationary phase were due to the syn and anti hydrazones formed. We furthermore conclude that the dansylation reaction is subject to alkyl catalysis rather than acid catalysis, since methyl trifluoromethanesulfonate showed a strong catalytic action, while the catalytic action of trifluoromethanesulfonic acid was lower when diluted in other alcohols and disappeared in aprotic solvents. A sensitivity to water in the reaction mixture strengthens the evidence for alkyl catalysis. When optimal experimental conditions were used, derivatization of picomole amounts of ketosteroids could be accomplished in 25 min. Analysis of spiked plasma containing 0.2-2.0 nmol each of progesterone and 3rhydroxy-5β-pregnan-20-one showed overall recoveries of 69-76% and 40-55%, respectively. The corresponding 3σ detection limits estimated from calibration curve data were 12 and 15 pmol (n ) 4, 500 µL injected). Dansylhydrazine1,2 is a fluorescent reagent frequently used for precolumn derivatization of carbonyl compounds. By using dansylhydrazine in the presence of a strong acid, it is possible to introduce a fluorescent label onto a ketosteroid.3-5 The determination of ketosteroids as their corresponding dansylhydrazones is a rather simple and sensitive technique that can be used for qualitative as well as quantitative analysis. However, the dansylation reaction of carbonyl functionalities has several limitations. Formation of multiple hydrazones may take place, since the ketosteroids often carry more then one carbonyl group and usually * To whom correspondence should be addressed. E-mail: [email protected]. † University of Umea ˚. ‡ University Hospital, Umea ˚. § Astra Draco AB. (1) Chayen, R.; Dvir, R.; Gould, S.; Harell, A. Isr. J. Chem. 1970, 8, 157. (2) Chayen, R.; Dvir, R.; Gould, S.; Harell, A. Anal. Biochem. 1971, 42, 28385. (3) Kawasaki, T.; Maeda, M.; Tsuji, T., J. Chromatogr. 1981, 226, 1-12. (4) Kawasaki, T.; Maeda, M.; Tsuji, T. J. Chromatogr. 1982, 232, 1-11. (5) Weinberger, R.; Koziol, T.; Millington, G. Chromatographia 1984, 19, 4526. S0003-2700(97)00295-3 CCC: $14.00

© 1997 American Chemical Society

have multiple stereocenters in the steroid backbone.5 Therefore, it is important to ascertain that the derivatization reaction is groupspecific. Moreover, when derivatization of low steroid concentration levels is carried out, overnight reactions are common, due to the slow reaction rates encountered with most catalysts. However, implementation of a derivatization technique in routine analysis requires faster reaction kinetics. In a recently published paper, it has been shown that it is possible to introduce dansylhydrazine onto the 3-ketosteroid budesonide (I) when using trifluoromethanesulfonic acid (TFMSA) as a catalyst.6 In this work, we have used experimental design7 coupled with multivariate analysis8 to further investigate the parameters that affect the dansylation of budesonide (I) and some other ketosteroids. The parameters studied are the molar ratio between the steroid and the derivatization reagent, the catalyst concentration, and the reaction time and temperature. EXPERIMENTAL SECTION Reagents and Solutions. Structures of the steroids used in this study are shown in Figure 1. Budesonide (11β,21-dihydroxy16R,17R-[22R,S]-propylmethylenedioxypregna-1,4-diene-3,20-dione) was obtained from Astra Draco AB (Lund, Sweden). Progesterone (4-pregnene-3,20-dione) [57-83-0], 5β-androstane3R,11R,17β-triol [32212-61-6], 5R-pregnane-3,20-dione [566-65-4], 5β-pregnane-3,20-dione [128-23-4], 3R-hydroxy-5β-pregnan-20-one [128-20-1], and 3R,21-dihydroxy-5β-pregnane-20-one [567-03-3] were purchased from Sigma (St. Louis, MO). Dansylhydrazine (5-[dimethylamino]-1-naphthalenesulfonic hydrazide, 98%) [3300806-9], trifluoromethanesulfonic acid (>98%) [1493-13-6], methyl trifluoromethanesulfonate (>99%) [333-27-7], and styrene (99%) [100-42-5] were purchased from Aldrich (Steinheim, Germany). Ammonium acetate (>98%), 2-propanol (>99.7%), and diethyl ether (>99%) were obtained from Merck (Darmstadt, Germany), while acetonitrile (HPLC, Ultragrade) and methanol (HPLC grade) were purchased from J. T. Baker (Deventer, The Netherlands). Absolute ethanol (>99.5%) was obtained from Kemetyl (Haninge, Sweden). Divinylbenzene (technical grade, 70-85%, mixtures of isomers) [1321-74-0], dimethyl sulfoxide (>99.5%) [67-68-5], 1-dodecanol (99.5%) [112-53-8], and 2,2′-azobis(isobutyronitrile) (AIBN, 95%) [673-66-5] were purchased from Fluka (Buchs, Switzerland). (Numbers given above in brackets are CAS Registry (6) Hyytia¨inen, M.; Appelblad, P.; Ponte´n, E.; Stigbrand, M.; Irgum, K.; Jaegfeldt, H. J. Chromatogr. 1996, 740, 279-283. (7) Box, G. E. P.; Hunter, W. G.; Hunter, J. S. Statistics for Experimenters; Wiley: New York, 1978. (8) Wold, S.; Johansson, E.; Cocchi, M. In 3D QSAR in Drug Design, Theory, Methods and Applications, Kubinyi, H., Ed.; ESCOM Science Publishers: Leiden, The Netherlands, 1993; pp 523-550.

Analytical Chemistry, Vol. 69, No. 23, December 1, 1997 4905

Figure 1. Structures of the steroids investigated: budesonide (I), progesterone (II), 5β-pregnane-3,20-dione (III), 5R-pregnane-3,20-dione (IV), 3R-hydroxy-5β-pregnan-20-one (V), 3R-21-dihydroxy-5β-pregnan-20-one (VI), and 5β-androstane-3R,11R,17β-triol (VII).

Numbers, provided by the author.) Dry ice was obtained from AGA Gas AB (Sundbyberg, Sweden). All water used was purified by Super-Q (Millipore, Bedford, MA) equipment and had an electrolytic conductivity less than 60 nS cm-1. Dansylhydrazine solutions were prepared daily and stored in glass bottles protected from light with aluminum foil. Standard solutions of steroids, trifluoromethanesulfonic acid, methyl trifluoromethanesulfonate, and dansylhydrazine were all prepared in methanol. Derivatization of Steroids. Derivatization was performed in 1.5 mL crimp-top glass vials using PTFE-lined septa (Skandinaviska Genetec AB, Kungsbacka, Sweden), with the reaction mixture protected from direct light with aluminum foil. The steroid solutions (1000 µL) were initially mixed with trifluoromethanesulfonic acid (250 µL), whereafter dansylhydrazine (250 µL) was added to the solution. The concentrations of the steroids, the acid, and the dansylhydrazine varied between experiments, and the experimental levels are, therefore, given in connection with each figure and table. Following reagent addition, the reaction mixture was allowed to stand on an orbital shaker (IKA Janke and Kunkel, Staufen, Germany) for varying periods of time. Blanks were prepared in a similar manner as the samples but contained only methanol instead of steroid solution. Extraction and Cleanup. Plasma (follicle and postmenopausal phase) was pooled and stored at -8 °C. Prior to use, the plasma was thawed at room temperature and centrifuged for 10 min at 6000 rpm. Next, 0.2 mL plasma was pipetted into an extraction vial, followed by the addition of 0.5 mL of water and 3.0 mL of diethyl ether. The samples were then allowed to stand on an orbital shaker for 20 min. After the liquid-liquid extraction, the vials were transferred into an ethanol/dry ice bath for 10 min. The ether phase was then decanted and evaporated at room temperature. The residue was finally redissolved in 1.0 mL of methanol and transferred into a 1.5 mL crimp-top glass vial prior to analysis. Spiking of plasma was performed by adding a 1.0 mL steroid standard solution, which was evaporated to dryness at room temperature before addition of the plasma. Preparation of the Polymer-Based Enrichment Column. A 2.01 g quantity of a stock solution containing divinylbenzene and styrene (1:1, 4 parts w/w) and 1-dodecanol (6 parts w/w) was mixed with 8.1 mg of AIBN. The mixture was purged with helium for 10 min and transferred into a 50 mm long × 3 mm i.d. PEEK column blank (Upchurch Scientific Inc., Oak Harbor, WA). 4906 Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

The polymerization was carried out at 75 °C for 4 h. The resulting monolithic column was equipped with porous polypropylene frits. A MeOH/H2O eluent (85:15 v/v) with a flow rate of 1.0 mL/min was used to remove the porogen and unreacted monomers. No significant back-pressure was observed during the washing procedure. Chromatographic System. The configuration of the coupled column chromatographic system has previously been presented.6 It consists of two blocks, one for enrichment and the other for separation. The enrichment block consisted of an LKB 2150 LCpump (pump 1, Pharmacia, Uppsala, Sweden). The flow rate was set to 1.0 mL/min. Standard solutions were injected by an LKB 2153 autoinjector (Pharmacia & UpJohn) into the enrichment column (column 1), which was mounted in the loop position of an LC injector (Model 7000, Rheodyne, Cotati, CA) equipped with a Rheodyne 5701 pneumatic actuator. A Crouzet Top 948 counter (ELFA, Stockholm, Sweden) was used for timing the collection and transfer of injected solutions from the enrichment column into the separation column. A dwell time was used before the enrichment column was connected to the separation flow system. The separation block consisted of an LKB 2150 LC-pump (pump 2, Pharmacia), set to a flow rate of 1.0 mL/min, and a separation column (column 2). Detection was carried out with a Merck Hitachi F1000 fluorescence detector (Darmstadt, Germany) equipped with a 12 µL flow cell. The excitation and emission wavelengths used were 350 and 520 nm, respectively. The detector output was recorded with an electronic integrator (Model 3395, Hewlett Packard, Palo Alto, CA). In the budesonide (I) experiments, the enrichment column and the separation column were a Nucleosil C18 3 µm, 50 mm long × 4.6 mm i.d., and a Nucleosil C18 5 µm, 150 mm long × 4.6 mm i.d. (Skandinaviska Genetec AB, Kungsbacka, Sweden), respectively. The derivatized solution was injected (100 µL), and a dwell time of 5 min was used before the enrichment column was connected to the separation flow system. The mobile phases used for enrichment and separation were based on mixtures between acetonitrile and an ammonium acetate buffer (50 mM, pH 6.8). The mobile phase used for the enrichment contained acetonitrile-ammonium acetate buffer (30:70 v/v), whereas the mobile phase used for the separation contained acetonitrileammonium acetate buffer (65:35 v/v).

Table 1. Results from the 24 Full Factorial Designa responses expt no.

MR

variables CC TE

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

5 50 5 50 5 50 5 50 5 50 5 50 5 50 5 50 25 25 25 25

3 3 15 15 3 3 15 15 3 3 15 15 3 3 15 15 9 9 9 9

21.3 21.3 21.3 21.3 50.0 50.0 50.0 50.0 21.3 21.3 21.3 21.3 50.0 50.0 50.0 50.0 35.0 35.0 35.0 35.0

TI

peak 1 area (units/106)

peak 2 area (units/106)

peak 1 height (units/104)

peak 2 height (units/104)

10.0 10.3 10.0 10.0 10.3 10.0 10.7 10.0 30.0 33.0 31.0 30.0 30.0 30.0 32.3 30.0 20.0 20.0 20.0 20.0

0 0.75 0.11 0.93 0.47 2.39 0.36 2.86 0.19 2.43 0.14 1.96 0.98 5.42 na 3.23 1.52 1.45 1.59 1.46

0 0.64 0.12 0.92 0.22 1.81 0.32 2.16 0.16 1.63 0.12 1.71 0.77 4.41 na 3.09 1.36 1.24 1.22 1.25

0 1.56 0.29 1.96 0.81 5.83 0.96 6.42 0.45 4.91 0.49 3.93 1.90 9.75 2.04 7.80 3.85 3.81 3.75 3.84

0 1.06 0.21 1.40 0.51 3.39 0.65 4.15 0.31 3.12 0.34 2.80 1.08 5.26 1.41 4.98 2.50 2.44 2.39 2.44

a The following factors were studied: molar ratio between derivatization reagent and steroid (MR), catalyst concentration (CC) in %, reaction temperature (TE) in °C, and reaction time (TI) in minutes. A standard solution containing 2.0 µM budesonide (I) and dansylhydrazine standard solutions containing 40, 200, and 400 µM were used. na, Not available. Area and height are in integrator units.

In the progesterone metabolites experiments, the enrichment column was the 50 mm long × 3 mm i.d. PEEK column, containing the styrene-divinylbenzene monolith, the preparation of which was described above. An R-Chrom C18 5 µm, 150 mm long × 3.0 mm i.d. (Upchurch Scientific, Oak Harbor, WA) separation column was used. Derivatized solution was injected (500 µL), and a dwell time of 1 min was used before the enrichment column was connected to the separation flow system. The mobile phases used for enrichment and separation were based on mixtures between acetonitrile, methanol, and an ammonium acetate buffer (50 mM, pH 4.7). The mobile phase used for the enrichment contained methanol and ammonium acetate (15:85 v/v), whereas the mobile phase used for the separation contained acetonitrile and ammonium acetate (55:45 v/v). Mobile phases were filtered through a 0.47 µm PVDF filter (Millipore) and degassed with helium for 15 min prior to use. Mass Spectrometry. The mass spectrometric analyses were performed on an API I single-quadrupole mass spectrometer (Sciex, Thornhill, ON, Canada). The samples were infused at 2 µL min-1 with a Harward Apparatus (Southnatick, MA) syringe pump into the ion spray interface via a 50 µm i.d. fused-silica capillary. The spray chamber was maintained at 54 °C during the analysis, and the orifice voltage varied between 200 and 249 V. The spectra were acquired in the positive mode with a step size of 0.5 Da and a dwell time of 1.0 ms/step. Molecular masses of the HPLC fractionated derivatives were determined with MacSpec 3.3 (Sciex software). In most cases, the saved spectra were obtained by summing five spectra in the multichannel analyzer mode. Experimental Design and Multivariate Analysis. Screening and optimization of the derivatization reaction was performed using experimental design and multivariate analysis in Modde 3.0 software (Umetri, Umea˚, Sweden). Repeatability, Reproducibility, Recovery, and Sensitivity Evaluation. A data set, two separate series containing a total of

83 unique samples, was used to determine the repeatability, reproducibility, recovery, and sensitivity of the technique. The first series consisted of 38 standard solutions, whereas the other series contained 45 samples, both standard solutions and spiked plasma. Each series of analyses was performed over 4 days. The evaluated steroid standard solutions contained a mixture of five progesterone metabolites ranging from 0.05 to 5 µM (0.05-5 nmol of each steroid in the reaction mixture). Retention time repeatability of the coupled column system was calculated for the whole data set (n ) 83), the variability within series (n ) 45), and the within-day variability (n ) 11). Reproducibility for steroid standard solutions was evaluated by several injections of standards at the 0.2, 1.0, and 5.0 µM levels. The relative standard deviation (RSD) was calculated for each steroid on the basis of peak height measurements. Plasma aliquots spiked with known amounts of steroids (0.2-2.0 nmol) were analyzed to determine the recovery and the detection limit. RESULTS AND DISCUSSION Multilinear Regression Analysis. Multilinear regression analysis (MLR) on a 24 full factorial design (see Table 1) has been utilized for exploration of the dansylation of ketosteroids using trifluoromethanesulfonic acid as catalyst. Experimental results show that the molar ratio between the steroid and dansylhydrazine has the strongest influence on the derivatization yield (see Figure 2). The catalyst concentration as well as the interaction terms including the catalyst did not appear to have significant effects on the reaction in the experimental domain tested. The main effects on the reaction could thus be ranked in a descending order as molar ratio between steroid and dansylhydrazine (MR) > temperature (TE) > reaction time (TI) > interaction term molar ratio between steroid and dansylhydrazine and temperature (MR*TE) > interaction term molar ratio between steroid and dansylhydrazine and reaction time (MR*TI). After removing nonsignificant factors and inserting quadratic terms (MR*MR), Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

4907

Figure 2. Diagram showing the effects of the investigated factors and the interactions between these factors in a 24 full factorial design (see Table 1 for abbreviations).

R2 ) 0.99 and Q2 ) 0.97 were obtained when evaluating the peak height for one of the budesonide (I) derivatives (see Table 1). These results show that the projected model is valid and that the predictive power is reliable. The model shows that the derivatization yield increased as the reaction temperature rose, but temperatures over 50 °C could not be implemented due to evaporative loss of solvent. MLR analysis also demonstrates that faster reaction kinetics were achieved when the molar ratio between the steroid and the derivatization reagent was increased (see Figure 3). Using the equation for pseudofirst-order kinetics and plotting the amount of formed derivatives (area percentage) versus the reaction time, a straight line with R2 ) 0.97 was obtained. Therefore, we conclude that the dansylation reaction follows pseudo-first-order kinetics; i.e., the reaction rate is predominantly dependent on the concentration of the derivatization reagent. In the 24 full factorial design, a molar ratio of 50 between the steroid and the derivatization reagent was set as a maximum. When this ratio was increased, the detectability of the formed derivatives was impaired by the reagent tailing peak. To overcome the shortcomings of the commercially available enrichment column, a custom polymer-based column exhibiting a high capacity at low back-pressure was prepared.9 By using this styrene-divinylbenzene-based column, the injected sample volume could be increased from 0.1 to 0.5 mL. It was also possible to increase the molar ratio between the steroid and the derivatization reagent due to the improved enrichment performance with the divinylbenzene-styrene column. The chromatographic system was further improved by introducing a narrow-bore column in the separation block. The new chromatographic setup was optimized by a central composite circumscribed (CCC) experimental design on progesterone (II) (see Table 2). MLR analysis shows that the results obtained from the CCC design coincide with the full factorial design. It can thus be concluded that the molar ratio between the steroid and the derivatization reagent is the most important variable also for progesterone (see Figure 4). The model also revealed that a high yield of derivatization could be obtained within 25 min and that the derivatization strategy is applicable to other 3-ketosteroids in addition to budesonide (I). Using peak area measurements of the first-eluting derivative, R2 ) 0.95 and Q2 ) 0.76 were obtained (9) Viklund, C.; Svec, F.; Fre´chet, J. M. J.; Irgum, K. Chem. Mater. 1996, 8, 744-50.

4908 Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

Figure 3. Response surface plot showing the derivatization yield dependency on molar ratio and reaction temperature (upper plot), and molar ratio and reaction time (lower plot). The model was based on peak height measurements on budesonide (I) derivative 2.

for the model. This fit is not as good as in the full factorial design for budesonide (I), but it still shows that the projected model is valid. A plausible explanation for the lower validity and predictability of the model is that the experimental domain was too wide for the CCC design approach.

Table 2. Central Composite Circumscribed Design Used in the Multivariate Analysis of the Reaction Conditions for Progesteronea responses variables

expt no.

MR

1 2 3 4 5 6 7 8 9 10 11

26.38 79.14 26.38 79.14 15.83 88.47 52.76 52.76 52.76 52.76 52.76

TI

peak 1 area (units/106)

peak 2 area (units/106)

peak 1 height (units/104)

peak 2 height (units/104)

10.0 10.0 30.0 30.0 20.0 20.0 5.9 34.1 20.0 20.3 21.0

1.15 2.44 1.50 2.34 0.90 2.72 1.44 2.00 2.29 1.94 2.37

1.93 3.89 2.16 3.33 1.38 4.35 2.34 3.22 2.89 3.41 3.43

2.21 4.61 2.69 3.69 1.45 4.36 2.46 3.44 3.68 3.82 3.53

2.80 5.75 3.20 4.65 1.74 5.42 3.03 4.17 4.11 4.88 4.26

a The following factors were studied: molar ratio between derivatization reagent and steroid (MR) and reaction time (TI) in minutes. A standard solution containing 2.0 µM progesterone and dansylhydrazine solutions containing 126, 211, 422, 633, and 708 µM were used. Reaction temperature and catalyst concentration were set to 50 °C and 3%, respectively. Area and height are in integrator units.

Table 3. Central Composite Circumscribed Design Used in the Multivariate Analysis of the Reaction Conditions for 3r-Hydroxy-5β-pregnan-20-onea responses expt no. 1 2 3 4 5 6 7 8 9 10 11 Figure 4. Response surface plot showing the derivatization yield dependency on molar ratio and reaction time. The model was based on peak area measurements for progesterone derivative 1.

General Derivatization Applicability. To determine whether the derivatization strategy has a general applicability for ketosteroids, a CCC design was carried out on 3R-hydroxy-5β-pregnan20-one (V), which has a keto functionality in the 20-position only, i.e., adjacent to the five-carbon ring (see Figure 1 and Table 3). In that design, MLR analysis showed that a molar ratio above 60 had a low influence on the derivatization yield. The model also revealed that, when working in the experimental domain tested, the highest derivatization yield was obtained after 25 min (see Figure 5). One of the center points, experiment no. 10, proved to be an outlier and was consequently removed from the data set. By removing a center point, the pure error of the data set became significantly smaller than the model error. After removing nonsignificant factors, R2 ) 0.97 and Q2 ) 0.91 were obtained. Thus, the model proved to be valid, and the predictive power was adequate. Both progesterone and 3R-hydroxy-5β-pregnan-20-one showed the same response in the CCC design approach; namely, the highest derivatization yield was obtained after a reaction time of 25 min. A longer reaction time resulted in a decrease in yield, and a plausible explanation could be that the formed derivatives degrade or rearrange. Another cause of the decreased yield may

variables MR TI 61.25 89.13 61.25 89.13 50.00 97.50 72.50 72.50 72.50 72.50 72.50

10.00 10.50 30.00 30.00 20.00 20.00 5.86 34.14 20.00 20.00 20.00

peak area (units/105)

peak height (units/104)

3.23 3.86 5.83 7.27 4.38 6.37 1.06 5.33 4.95 2.79 4.96

0.52 0.64 0.97 1.25 0.80 1.08 0.13 0.92 0.96 0.49 0.95

a The following factors were studied: molar ratio between steroid and derivatization reagent (MR) and reaction time (TI) in minutes. A standard solution containing 2.0 µM 3R-hydroxy-5β-pregnan-20-one and dansylhydrazine standard solutions containing 400, 490, 580, 713, and 780 µM were used. Reaction temperature and catalyst concentration was set to 50 °C and 3%, respectively, when the experiment was performed. Area and height are in integrator units.

be multiple functionalization, since several steroids carry more than one carbonyl group in the steroid backbone.5 The reaction yield was monitored as a function of time by collecting fractions containing relatively high concentrations of progesterone derivatives from the coupled column chromatographic system and analyzing these by mass spectrometry. In Figure 6, it is shown that a reaction time of 30 min generates only two derivatives, each with a m/z of 562, which corresponds to the MH+ for monofunctionalized progesterone. From these results, we conclude that syn and anti bonding to the formed hydrazones occur when a steroid containing an unsaturated carbonyl is derivatized with dansylhydrazine.5 Using a reaction time of 60 min, four peaks were observed in the chromatogram. The two main peaks appeared at the retention times of the monofunctionalized progesterone derivatives, while the other two were from late-eluting compounds with a retention time of >75 min. The chemical nature of these two compounds was not possible to determine from the mass spectrometry experiments. The Effect of Water. Experiments with deliberate addition of water to the reaction mixture showed that even small amounts Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

4909

Table 4. Water Dependency of the Derivatization Reactiona water added (µL)

(%)

yield %

0 0.5 1 2 5 7 10

0.00 0.03 0.07 0.13 0.33 0.47 0.67

100 62 56 38 24 18 12

a Standard solutions containing 2.0 µM budesonide (I), 400 µM dansylhydrazine, and 3% trifluoromethanesulfonic acid were used. The influence of water in the reaction was studied by successively decreasing the TFMSA volume (250-240 µL) and adding water (010 µL). The yield when no water was added to the reaction mixture was set to 100%.

Figure 5. Response surface plot showing the derivatization yield dependency on molar ratio and reaction time using peak height measurements on 3R-hydroxy-5β-pregnan-20-one.

Figure 6. Chromatogram showing the dual peaks obtained from the syn and anti hydrazones when a 400 µM progesterone standard solution was derivatized with danzylhydrazine. The corresponding electrospray mass spectra both show an m/z of 562, which corresponds to the protonated molecular ion (MH+) of progesterone.

(