Solid Reagent for sampling and Derivatization of Carbonyl Compounds The present invention relates to the use of a new procedure for the formation of chemical derivatives of chemical compounds containing an aldehyde or a keto group as part of their structure for the purpose of carrying out a determination of said compounds. To achieve this means, we have discovered that a heterogeneous reac- tion can be advantageously utilized for the formation of hydrazones in pre-column labeling of carbonyl containing compounds, using ketosteroids and aldehydes as model compounds. This procedure has been reduced to practice through the con- struction of a solid phase enrichment and enhanced derivatization (SPEED) device, which has been prepared from porous polyethylene that has been coated with solu- tions of a polymer carrying perfluoroalkylsulfonic acid groups and 5- (dimethylamino) naphthalene-1-sulfonhydrazide (dansylhydrazine). With this device it is possible to simultaneously perform sampling, enrichment, and derivatization of analytes containing carbonyl groups, from both liquids and gases.

Description of Prior Art It is well known that an acid-catalyzed condensation reaction between a carbonyl containing compound, such as an aldehyde or a ketone, and a derivatization reagent with a hydrazine or hydrazide functionality results in the formation of a new chemi- cal compound, a hydrazone derivative, that is more suitable than the parent com- pound regarding detection and/or separation in chemical analysis. One group of compounds amenable to this technique is the physiologically significant C-21 keto- steroids from the pregnenolone and progesterone metabolic pathways; Figure l. la b c.

Considering the vast number of steroid metabolites naturally occurring at very low levels, high flexibility is required in both the sample pre-treatment and detection procedures of this group of compounds. Several of the metabolites lack functional groups with good detectability properties in their molecular backbone, wherefore only few analytical techniques are applicable. At present, mass spectrometric detec-

tion coupled with either gas or liquid chromatographic separation are considered as the state-of-the-art techniques. Although these instrumental systems provide high efficiency and excellent sensitivity, they are expensive and complicated. Other tech- niques with sensitivities sufficient for trace determination are therefore also needed and an alternative is to produce dansylhydrazone derivatives, by using 5- dimethylamino-1-naphthalenesulfonic hydrazide (dansylhydrazine or DNSH), that can be separated with liquid chromatography and determined with either fluores- cence or chemiluminescence detection. Numerous ketosteroids carry multiple car- bonyl groups and usually have multiple stereocenters in their steroid backbone. The formation of multiple derivatives is thus another problem associated with this pre- column labeling reaction, as illustrated in Figure 2. l2 l4-l6 The model compound pro- gesterone is a C-21 ketosteroid that carries two keto groups in the molecular back- bone; at the 3-and the 20-positions. There is also a double bond conjugated to the 3- keto group. Formation of the hydrazones is also subject to syn and anti addition, yielding products that tend to separate on achiral HPLC columns. When progester- one is derivatized with 5- (dimethylamino) naphthalene-1-sulfonhydrazide (DNSH), we could therefore expect up to four hydrazone peaks in reversed phase liquid chromatographic runs on regular C18 columns. l2 l4 l5 It has been shown in the litera- ture for different ketosteroids, that the 3-keto group is more reactive towards 2,4- dinitrophenylhydrazine than the keto groups in the 17-and 20-positions. It has also been demonstrated that selective derivatization is possible to perform at the 3- position, leaving an unreacted 20-keto group, simply by carrying out the reaction at room temperature, instead of at elevated temperatures. l5 Since it has been found that different acids influence both the rate and yield of the hydrazone formation, several acids have been studied as catalysts. While trifluoroacetic acid (TFA) has proven to be efficient in catalyzing the dansylation reaction, it is not very useful at room tem- perature, and it is therefore important to search for new selective and efficient cata- lysts for the dansylation reaction.

Trifluoromethanesulfonic acid (TFMSA) is among the strongest known monoprotic Brönsted acids and possesses several unique properties, such as extreme thermal

stability and a high resistance towards both reductive and oxidative cleavage. Al- though this suggest that TFMSA should be an efficient catalyst, it has only recently been employed in analytical derivatization reactions. In an on-going project the in- ventors thus have introducedll and optimizedl2 the use of trifluoromethanesulfonic acid (TFMSA) solutions as liquid phase catalyst in the pre-column dansylation reac- tion of a number of ketosteroids and merely by using TFMSA as catalyst it is possi- ble to control the hydrazones formed to the syn and anti mono-derivatives only. 12 This derivatization procedure is however carried out through homogenous acid ca- talysis, where the reaction kinetics primarily depends on the concentration of the analyte. Derivatization and quantitation of trace levels of analyte in complex ma- trixes is thus not feasible. Unfortunately, these conditions are prevailing and the only way of circumventing this problem is to use a larger excess of reagent or a more efficient catalyst in order to enhance the reaction kinetics and thereby possibly enable labeling of minute amounts. Another alternative to improve the sensitivity of this technique would be to attach multiply dansyl moieties to the molecular back- bone, and thus yield bis-derivatives for the model compound progesterone, as can be seen in Figure 2. With two labels attached to the molecular backbone, the deriva- tives will become more lipophilic and a mobile phase with high eluting strength can be used, thereby reducing the risk of co-elution with impurities, or with other la- beled metabolites. Provided internal quenching does not take place, the total quan- tum yield would also be higher from a bis-derivative compared to a mono- derivative. Correspondingly, Weinberger et al. have shown that it is possible to yield both mono-and bis-derivatives from progesterone by using an"evaporative- derivatization technique". 15 With their scheme, it is possible to complete the reac- tion within 10 minutes and to control the reaction in favor of the bis-derivatives, but the experimental conditions that have to be applied are rather harsh.

However, because of the low sample concentration involved, the derivatization re- action has been plagued by slow kinetics with the acid catalysts traditionally used, leading to overnight reaction times in order to reach reasonable yields. This hampers the applicability of the reaction scheme to routine analyses. While pursuing a faster

catalyst for the dansylation reaction, one important aspect is the"cleanliness"of the reaction, i. e., the catalyst used for dansylation of ketosteroids should also be selec- tive in the reaction it promotes and result in minimal sample breakdown.

Another complication with the use of TFMSA is that it is highly toxic, and it spon- taneously forms carcinogenic alkyl triflates when diluted in anhydrous alcohols which is necessary for the derivatization scheme. Derivatized samples must there- fore be handled as risk waste and TFMSA-catalyzed derivatization is also problem- atic to implement in routine laboratories. l3 Consequently, there are both environ- mental and economical incentives to rid these methods in favor of new and more ef- fective derivatization schemes.

Nevertheless, it has previously been suggested by Waller et al 18 that Nafioni, could be used instead of traditional homogeneous acids for a number of synthetic organic reactions. That article describes various processes that are catalyzed by Naf1on but the reaction between carbonyl containing compounds and hydrazine or hydrazide reagents has not been described. Waller et al 18 do however state that Nation may be a substitute for TFMSA and that it may be supported on carriers.

Terms As disclosed herein, the term"analytical column"means a tubular container fit- ted with retaining end pieces and containing any suitable chromatographic sta- tionary phase usually on supports such as particles or monolithic structures, but also other supports, e. g., capillaries and the like, useful in performing liquid chromatographic separations (HPLC).

As disclosed herein, the term"eluent"means a liquid mobile phase suitable for use with the analytical column for performing liquid chromatographic separa- tions.

As disclosed herein, the term"carrier"means an inert organic or inorganic solid or gel-like material that acts as a support for a bound, attached, or adsorbed functional group, e. g., chromatography supports, or porous discs manufactured

from, e. g., polypropene, polyethene, polystyrene, or a fluorinated polymer. These are only examples of suitable carriers, and many other exist.

As disclosed herein, the term"polymers carrying a plurality of pendant (poly- fluoroalkylene) sulfonic acid moieties"relates to polymers comprising several perfluoroalkylene sulfonic acid moieties. Examples of such polymers can be found in US, A, 3,041,317, US, A, 3,282,875, US, A, 4,337,137, US, 4,358,545, US, A, 4,417,969, US, A, 4,478,695 and US, A, 4,940,525. A typical preferred example is Nafion (D, which is provided by DuPont de Nemours, Inc., USA.

A"reagent"generally means a chemical compound or a combination of chemical compounds that by its/their presence takes part in a chemical reaction and trans- forms an analyte of interest into a new compound, the derivative, which has al- tered separation and/or detection properties. As disclosed herein, the term "reagent"relates to compounds comprising a hydrazine functionality and/or a hydrazide functionality. Reagents according to the present invention are able to take part in derivatization reactions wherein carbonyl-containing molecules are derivatized to corresponding hydrazone derivatives. Preferably, reagents ac- cording to the present invention are labeled in such a way that the absorptive, fluorescent, electrochemical and/or chemiluminescent properties of the resulting derivative is/are affected, thereby facilitating the subsequent analytical detection.

The skilled person is well aware of commonly used labels and how to include them in a derivatization reaction according to the present invention. Typical rea- gents according to the present invention are 5-(dimethylamino) naphthalene-1- sulfonhydrazide (dansylhydrazine, DNSH) and 2-nitrophenylhydrazides, tri- azinehydrazines, 2,4-dinitrophenylhydrazine (DNPH), 4,4-difluoro-5,7- dimethyl-4-bora-3a, 4a-diaza-s-indacene-3-propionohydrazide, 4-hydrazino-7- nitrobenzofurazan (NBD-H), 4'-hydrazino-2-stilbazole, and 9-fluorenylmethoxy- carbonyl hydrazide.

As disclosed herein, the term"solid phase/catalyst combination"refers to a com- position comprising a reagent according to the invention as defined above, to-

gether with at least one polymer carrying a plurality of pendant (perfluoroallcyl) sulfonic acid moieties as a catalytic element. Preferably, the solid phase/catalyst combination is coated on an inert carrier.

As disclosed herein, the term"derivatization device"relates to a container en- closing a solid phase/catalyst combination. A suitable container can be a column, a capillary or a syringe. The sample containing the molecule that is going to be derivatized is injected into the derivatization device, where it is exposed to the solid phase/catalyst combination resulting in a derivatization reaction.

Summary of the Invention It has now turned out that the above mentioned problems obtained for the homoge- nous derivatization reaction using trifluoromethanesulfonic acid (TFMSA) may be obviated by using a solid reagent/catalyst combination.

The preferred embodiment of the invention concerns a solid reagent/catalyst combi- nation for simultaneous enrichment and derivatization of carbonyl compounds to hydrazone compounds, comprising Nation or a similar polymer carrying perfluoro- alkylsulfonic acid groups, such as the material defined above, and a compound ca- pable of acting as derivatization reagent, having as its coupling moiety/moieties hy- drazine and/or hydrazide group (s). By the expression"comprising"we understand "including but not limited to". Thus, other non-mentioned carriers, reagents, or ad- ditives may be present. Using the solid phase enrichment and enhanced derivatiza- tion (SPEED) device (derivatization device), prepared from porous polyethylene that has been coated with both Nation and 5-(dimethylamino) naphthalene-1- sulfonhydrazide, the exposure to hazardous chemicals is minimized compared to the TFMSA catalyzed homogenous solution reaction. Moreover, Nation has ion- exchange properties and the new solid reagent can act as a solid phase extraction material where sample clean-up and enrichment could be accomplished simultane- ously with the derivatization.

The SPEED devices have been optimized using experimental design and character- ized for dansylation of C-21 ketosteroids by multivariate data analysis, using pro- gesterone as the model compound. The reaction temperature and the molar ratio between the steroid and the derivatization reagent coated onto the carrier were found to be the factors most strongly affecting the reaction. Using optimal reaction condi- tions, the derivatives mainly constitute of the syn-and anti-conformers of bis- derivatives, and picomole amounts of ketosteroids could be derivatized in 10 min- utes at room temperature. A sample volume of 400 uL could be loaded and derivat- ized and finally be eluted in 300 pL, demonstrating the sample enrichment property of the device. In contrast to homogenous solution-based acid catalysis, the SPEED device was remarkably insensitive to the presence of water in the reaction mixture, i. e., the sample.

Analysis of spiked serum samples containing 0.4 to 2.0 nanomoles of progesterone showed overall recoveries of 52-63 %. The corresponding 3a detection limit was 1.3 pmole (n = 4, 100 pL injected on the column ; corresponds to 3.9 pmole derivat- ized), as estimated from calibration curve data.

In a similar way aldehydes such as formaldehyde and n-nonyl aldehyde were both trapped and derivatized at room temperature using the new solid reagent.

Detailed Description of The Invention The discovery of the invention was based on the hypothesis that a polymer carrying perfluoroalkylsulfonic acid groups could act as a catalyst and reagent carrier for the derivatization of ketosteroids, a concept that was envisaged shortly after the inven- tors made the discovery that trifluoromethanesulfonic acid has an extraordinary catalytic effect on the reaction of 5- (dimethylamino) naphthalene-I-sulfonhydrazide with ketosteroids in solution.

A preferred poly(perfluoroalkylene)sulfonic acid according to the invention is Nafion#, a family of polymers with a ydrophobic perfluoroalkylene backbone and pendant groups terminating in the catalytic perfluoroalkylsulfonic moiety. The polymers have ion-exchange properties and are permselective to allow the passage

of cations but not of anions in a manner common to cation exchange materials, and they have been developed for an entirely different purpose, i. e., to act as ionically conducting separator membranes in industrial electrodialysis units and fuel cells.

Thus, the inventors carried out an initial experiment where pieces were chopped from a narrow diameter tubiform Nafion membrane (811X) and placed in a vial, to which a standard solution of the ketosteroid budesonide and a DNSH solution were added; see Experiment 1. When solution eluted from this polymeric catalyst reagent was subjected to separation by reversed phase liquid chromatography, peaks ap- peared at the same retention time as the derivatives formed in homogeneous solu- tion. Peaks did not appear at these retention times in liquid chromatograms of solu- tions prepared in the absence of the polymeric catalyst. This clearly demonstrated that hydrazones were formed from the steroid, and that this formation took place be- cause of the presence of the perfluorosulfonic acid carrying polymer.

An embodiment of the invention concerns a solid reagent for simultaneous enrich- ment and derivatization of carbonyl compounds to hydrazone derivatives, compris- ing Nation and a compound having hydrazine and/or hydrazide functionality. By the expression"comprising"we understand including but not limited to. Thus, other non-mentioned carriers (see Terms) or additives (see below) may be present.

Although solid Nafiono is commercially available in several confections, the in- ventors found the Nafion solution prepared by dissolving the polymer in lower aqueous alcohols9 to be most suitable. The liquid form makes it possible to prepare coatings of varying thickness on carriers with different surface area and porosity. It is thereby possible to prepare different types of devices-based on particles to be packed in columns, or on discs incorporated in solid phase extraction (SPE) reser- voirs. The latter proved to be the most feasible approach, as the familiar format of the SPE reservoir allows its direct incorporation in existing sample preparation ro- bots.

The compound having a hydrazine or hydrazide groups may be any substance that could form a hydrazone bond with a compound of interest having one or more car- bonyl groups. Preferably a substance that can be used as a marker substance is used

such as, for example, DNSH, 2-nitrophenylhydrazides, triazinehydrazines, 2,4- dinitrophenylhydrazine (DNPH), 4,4-difluoro-5,7-dimethyl-4-bora-3a, 4a-diaza-s- indacene-3-propionohydrazide, 4-hydrazino-7-nitrobenzofurazan (NBD-H), 4'- hydrazino-2-stilbazole, and 9-fluorenylmethoxycarbonyl hydrazide. All these rea- gents are highly conjugated and thus amenable to detection techniques based on electronic excitation by photons, i. e., UV-Vis spectrophotometry, fluorescence spectrometry, or chemiluminescence.

The invention may be used for enrichment, sorption, and determination of com- pounds comprising one or more carbonyl groups from both gases and solvents, and it may be thereby used for the sampling and analysis of aldehydes and ketones.

Thus, irritating carbonyl compounds that are used in industry, whose volatility and reactivity makes it difficult to capture and keep in a sample, may be sampled, de- rivatized and analyzed from e. g. air samples. Examples of such compounds are for- maldehyde, glutaraldehyde, and acrolein. The disclosed scheme may thus be used for the formation of derivatives that capture and stabilize analytes, or to form de- rivatives with increased hydrophobicity, which provides for enhanced separation properties in reversed phase HPLC compared to the native analytes.

The formation of hydrazones is a reversible reaction and the invention may therefore be used for reactive solid phase extraction in organic chemical reaction schemes. such as in the clean-up and characterization of combinatorial syntheses and other parallel synthesis schemes, and in biochemical diagnostical reactions.

The final use envisioned for the disclosed process is for entrapment of analytes from a liquid or a gas phase, with concurrent derivatization of the analyte to impart prop- erties characteristic of a labelled analyte. This is done to enable or facilitate separa- tion and/or detection in a chemical analysis system. This invention in particular ad- dresses the needs in ultra-trace analysis, where enrichment and an acid-catalyzed chemical transformation can be carried out in a single step, reducing the risks of sample loss or contamination. The technique has several significant advantages over solution phase reactions, as the derivatization may be done in an environmentally

acceptable way, at room temperature, with a controlled reaction outcome, at lower concentrations, and automatically much faster than by earlier technologies.

Derivatization of Progesterone Using the SPEED Device. Porous polyethylene (PE), a material which is chemically inert and available with controlled porosity was considered the most suitable carrier for Nation coating. One milliliter PP reservoirs were thus pre-packed with porous PE discs, loaded with solutions of Nafion and DNSH according to Figure and Table footnotes, and thereafter dried. The final ver- sion of SPEED utilized commercially available PP syringes equipped with 1.5 mm thick PE discs of nominal pore size 10 urn (IST ; part no. 120-1061-A). These discs were modified by adding 30 ptL of Nafion solution, followed by 50 VL of a 3.0 uM methanolic solution of 5-(dimethylamino) naphthalene-1-sulfonhydrazide. The SPEED devices thus produced were heat sealed for storage in aluminum foil/PE laminate with a silica gel sachet.

Attempts at labeling progesterone with this device using a reaction temperature of 50°C, conditions previously found optimal for TFMSA catalysis, 2 demonstrated that four hydrazones, corresponding to the syn-and anti-derivatives of both mono- and bis-derivatives, were formed, see Example 2.

Derivatization Yield as a Function of Catalyst Concentration. Amounts of Nation ranging from 0.4 to 2.2 gg were added as 5 wt-% aqueous alcoholic solu- tion (10 to 50 pL) to a number of PE discs that were used for derivatization. The relative derivatization yield based on the combined peak height of both bis- derivatives was studied, see Table 1 and Example 3. From these experiments it be- came evident that there was no significant correlation between the derivatization yield and the amount of Nation loaded on the PE-discs in the range tested. At- tempts at derivatizing other ketosteroids in addition to progesterone were thereafter done, using compounds containing either one or two keto-groups in the molecular backbone as models. Based on the number of peaks appearing in the chroma- tograms, it was concluded that 5a-pregnane-3, 20-dione (IV), 5-pregnane-3, 20- dione (V) formed bis-derivatives while 20a-hydroxy-4-pregnane-3-one (VI), 3a- hydroxy-5p-pregnane-20-one (IX) and 3p-hydroxy-5a-pregnane-20-one (VIII)

formed mono-derivatives. Consequently, the inventors believe that the derivatization procedure is generally applicable to steroids containing a carbonyl moiety.

Experimental Design and Multivariate Data Analysis. The effect of reaction time and temperature on the derivatization yield for the bis-derivatives was evaluated using a centrally circumscribed composite (CCC) experimental design and partial least squares (PLS) analysis in the Modde software package, see Table 2 and Exam- ple 4. Modeling of data revealed that the reaction time (TI) had no significant effect while the reaction temperature (TE) had a strong influence on the reaction yield for formation of progesterone bis-derivatives, in the tested experimental domain. Thus when using the SPEED device for labeling of new compounds, an optimization of the reaction temperature must be carried out.

Automated Pre-column Derivatization. Having established room temperature as the optimal condition for formation of progesterone bis-derivatives, the derivatiza- tion procedure was then automated in order to improve the reproducibility. Nineteen SPEED devices with varying amounts of Nafioe and DNSH were prepared and used for derivatization of progesterone using an Gilson Aspec Xli autoinjector ac- cording to a 24 full factorial experimental design; Table 3 and Example 5. The effect of the DNSH (DN) and Nation volume (NA), reaction temperature (TE), and reac- tion time (TI) were investigated using PLS analysis, as above. Modeling of experi- mental data revealed that the derivatization yield increased with the DNSH concen- tration, see Figure 4, and the highest yield was achieved after ten minutes at RT us- ing the lower amount of Nafion as catalyst. These data correlate well with the above CCC design, and it thus can be concluded that RT and a short reaction time are the best conditions for obtaining progesterone bis-derivatives.

Comparison of Porous Polyethylene Carriers for the Solid Phase Derivatization Reaction. Four different PE discs with different thickness and porosity were evalu- ated as carriers, using derivatization yield as optimization criterion. The data ob- tained, presented in Table 4, show that there were no significant differences be- tween SPEED devices prepared from carriers A, B, and D, whereas material C pro- duced a device with a significantly lower efficiency. This cannot be explained by

differences in the nominal porosities of the carriers, but since the same amount of desorption solution was used on all devices, it may have been an insufficient volume to quantitatively elute the derivative from carrier C, which was the thickest carrier tested. Since carrier D is commercially available pre-packed in one milliliter PP res- ervoirs, it was selected for the continued studies.

Final Optimization of the SPEED Derivatization. As a provocation test of the de- rivatization reaction a final CCC design was carried out, where the experimental domain was expanded around the provisional optimum found in the 24 full factorial design described above, see Example 7. The derivatization yields obtained in these experiments are presented in Table 5. Interestingly, as visualized in the response plot in Figure 5, the derivatization yield remained constant over a large part of the experimental domain tested. The reaction time thus played a minor role and the yield was nearly constant, as long as the amount of reagent was sufficient.

Influence of Water on Derivatization Yield. It has been shown in a previous study, where solution phase TFMSA catalysis was used, that pre-column dansylation of progesterone is sensitive towards water in the reaction mixture.'2 Experiments with water deliberately added to the reaction mixture were therefore carried out to investigate the sensitivity towards water in the samples with the solid phase reagent, and to compare these results with the heterogeneous TFMSA catalyzed solution re- action, see Example 8. A remarkable relationship was found, as shown in Figure 6, where a maximum in reaction yield was obtained when the reaction mixture con- tained 10-20 % methanol. Sample breakthrough is a plausible explanation for the inferior recovery at high methanol concentrations. Experiments with a sample size of 40 iL instead of 80 uL resulted in better yield at higher methanol concentrations (data not shown), supporting breakthrough due to self-elution as a probable cause of the low yields at higher methanol concentrations. Maximum sample loadability was thereafter investigated and a linear relationship was seen over the entire tested range, see Example 9 and Figure 7.

Application to Human Serum Samples. The repeatability, the reproducibility, and the analytical recovery were evaluated using standard solutions and spiked human

blood serum, see Example 10. The analytical system was capable of a retention time repeatability of < 0. 8 % R. S. D. within one day of analysis (n = 10), and < 1.0 % R. S. D. for all samples (n = 45). The reproducibility was calculated for several repli- cated derivatizations of 0.5 to 5.0 uM standards over two days of analysis. At the 0.5 and 5.0 uM levels, the R. S. D. ranged between 10 tol7 %, while it varied be- tween 20 to 27 % at the 1.0 and 2.5 uM levels (n = 33). Using spiked human serum samples containing 0.4 or 2.0 nmoles of progesterone, the recoveries were 52 10 % and 63 17 %, respectively. The limit of detection for progesterone was estimated from calibration curve data to 1.3 pmole (100 tL injected), based on three times the standard deviation of four 0. 5 pM standard solutions. The recoveries were in the same range as in published studies, where also liquid-liquid extraction was applied prior to heterogeneous TFMSA catalyzed solution dansylation of various C- 21 ketosteroids. 6 2 We hence conclude that the non-quantitative recovery of the pre- sented technique may be an effect of the sample pre-treatment, rather than an infe- rior function of the SPEED device.

Application to volatile aldehydes. Sampling and labeling of volatile aldehydes such as formaldehyde and pelargonaldehyde (n-nonanal) have also been carried out with the SPEED device, see Examples 11 and 12. Hydrazone derivatives were rea- dily formed from both aldehydes when using a reaction time of ten minutes, see Fi- gure 9.

The invention will now be described by reference to the enclosed figures describing the following: Figure 1. Structures of the steroids investigated. Progesterone (4-pregnene-3,20- dione), 5ß-pregnane-3, 20-dione (I), 5a-pregnane-3, 20-dione (II), 3a-hydroxy-5p- pregnan-20-one (III), 3p-hydroxy-5a-pregnan-20-one (IV), 20a-hydroxy-4- pregnen-3-one (V).

Figure 2. Schematic drawing showing the formation of mono-and bis-derivatives of progesterone

Figure 3. Response surface plot showing the dependence of the relative derivatiza- tion yield on temperature and time of the reaction. The model was based on combi- ned peak areas for the progesterone bis-derivatives.

Figure 4. Response surface plot showing the dependence of the relative derivatiza- tion yield on the amount of catalyst and the concentration of the derivatization rea- gent. The model was based on peak area for the first eluting progesterone bis- derivative.

Figure 5. Response surface plot showing the dependence of the relative derivatiza- tion yield on the reaction time and the concentration of the derivatization reagent.

The model was based on peak height measurements of the first eluting progesterone bis-derivative.

Figure 6. The influence of water on derivatization yield. Progesterone (80 uL of a 50 M solution) was derivatized and thereafter eluted with 300 I1L of eluent, of which 100 uL was injected in the HPLC system. Error bars indicate standard error of the mean (n = 3 at each level).

Figure 7. Derivatization response as a function of sample volume applied to the SPEED device. Varying volumes of a 5 uM progesterone standard solution in 10 % methanol were applied to SPEED devices to determine maximum loadability (n = 4 at each volume). Error lines indicate the standard error of estimation.

Figure 8. Chromatogram showing a derivatized progesterone standard solution (5.0 uM ; solid line) superimposed on a blank run (dotted line) Figure 9. Chromatogram showing derivatized formaldehyde and n-nonal standard solutions superimposed on a blank run.

The invention will now be described more in detail with reference to the examples below which however do not limit the scope of the claims. All references are hereby incorporated by references.

Examples Reagents and Materials. The ketosteroids 4-pregnene-3,20-dione, 5a-pregnane- 3,20-dione, 5p-pregnane-3, 20-dione, 3a-hydroxy-5p-pregnan-20-one, 3 3-hydroxy-

5a-pregnan-20-one, 20a-hydroxy-4-pregnen-3-one were purchased from Sigma (St.

Louis, MO) and used as received. The Nafion perfluorinated ion-exchange resin in HF form was purchased as a 5 % (w/w) solution in lower aliphatic alcohols and wa- ter. Other reagents and solvents were of p. a. grade or similar. The porous polyethene materials evaluated were all commercially available products with pore sizes ranges and thicknesses of 15-45 um and 3 mm (carrier A), 9 um (average) and 3.2 mm (carrier B), 16 um (average) and 4.75 mm (carrier C), and 10 um (average) and 1.5 mm (carrier D).

General Procedures Preparation of SPEED Devices. Initial experiments were carried out with 6.2 mm 0 frits manually punched from porous PE discs obtained from several manufactu- rers. These discs were forcibly inserted into 1 mL PP syringe barrels. The porous PE discs were then washed with 2 mL of water followed by 2 mL of methanol, and the- reafter dried at 85 °C for ten minutes. Solutions containing Nafion and 5- (dimethylamino) naphthalene-1-sulfonhydrazide were then added to the porous PE so that the material was thoroughly and evenly soaked by the catalyst and reagent solutions (exact experimental conditions are given in Table and Figure footnotes).

The devices were then allowed to dry at 85 °C for ten minutes, or at room tempera- ture for 48 hours, wrapped in aluminum foil for protection against light. The prefer- red version of the SPEED device utilized commercially available PP syringes equip- ped with 1.5 mm thick PE discs of nominal pore size 10 um. These discs were mo- dified by adding 30 uL of Nation solution, followed by 50 uL of a 3.0 aM metha- nolic 5-(dimethylamino) naphthalene-1-sulfonhydrazide solution.

Reversed Phase HPLC Separations. The chromatographic system consisted of a high pressure reciprocating dual pistong pump made from stainless steel, operated at 0.5 mL/min, that was interconnected to a high pressure injector valve equipped with loops of either 20,50, or 100 pL volume. Separations were carried out on a 150 mm high performance reversed phase separation column having an inner diameter of 2 mm and packed with 5 urn particles. A fluorescence detector with the excitation and

emission wavelengths set at 350 and 520 nm, respectively, was used for detection of the separated compounds and data were collected and analyzed by a computer sof- ware. In the fully automated derivatization procedure the sample handling was ac- complished using an automatic sample handling system equipped with either a 100 or 168 uL loop and designed for application, mixing and washing of sample onto syringe-like solid phase cartridges.

Isocratic separation was carried out with an eluent consisting of acetonitrile purified for the use in HPLC and a 50 mM ammonium acetate buffer, pH 4.7 (80: 20 v/v). All eluents were filtered through a 0.47 urn inert filter and degassed by purging with helium for 15 minutes prior to use Example 1-Preliminary Experiments using Tubular Nafions Membrane as Catalyst.

This experiment was designed to verify the action of Nation as solid catalyst for the reaction between carbonyl compounds and hydrazinic reagents. Chopped pieces of 811X tubular Nation membrane (approx. 1 mm long) were placed in a 1.5 mL polypropylene vial, whereafter 40 p1L of 1.2 mM budesonide in acetonitrile and 100 RL of 2.3 mM DNSH in methanol were added. The reaction was allowed to proceed for one hour and finally 60 uL of 0.4 M NaOH was used to strip the formed derivatives off the membrane catalyst. The sample recovered by this extraction was diluted 100 times with eluent and analyzed by reversed phase liquid chromatography using fluorescence detection. Peaks appearing at the same retention time as deriva- tives formed in homogeneous solution clearly showed that hydrazones were formed from the steroid, and that this formation took place because of the presence of the perfluorosulfonic acid polymer, since peaks did not appear in the absence of this polymeric catalyst.

Example 2-Derivatization of Progesterone Using the SPEED Device.

In this experiment, the intention was to determine the identity of the compounds formed in the derivatization reaction using the catalyst/reagent combination as a solid phase reaction vehicle. In this experiment, the intention was to determine the

identity of the compounds formed in the derivatization reaction using the cata- lyst/reagent combination as a solid phase reaction vehicle. A high concentration standard of progesterone was used to certify formation of sufficient amount of pro- gesterone syn-/anti-mono and bis derivatives. The formed derivatives were there- after qualitatively fractionated after being injected on the chromatographic column, and the retrieved fractions were stored in test tubes sealed with aluminium foil until MS verification of each fraction was carried out.

Attempts at labeling 250 p1 of 517 tM progesterone standard (10/90 MeOH/HzO, v/v) with the SPEED device using a reaction temperature of 50°C, which previously was found optimal for TFMSA catalysis, l2 demonstrated that four hydrazones, cor- responding to the syn-and anti-derivatives of both mono-and bis-derivatives, were formed. Mass spectrometric analyses were performed on an LCQ DUO mass spec- trometer using an ESI source (Finnigan/Thermoquest, San Jose, CA, USA). The samples were infused into the ion spray interface at 2.5 pL min~l via a 50 um i. d. fused silica capillary, using an electronically controlled syringe pump integrated in the instrument. An ESI spray voltage of 4.5 kV was applied and the nitrogen sheath gas and auxiliary gas flows were set to 59 and 38 (instrument settings in arbitrary units), respectively. The heated capillary was operated at a temperature of 200 °C and a voltage of 30.4 V. The tube lens offset voltage was set to 5 V. A maximum injection time of 500 ms and 3 microscans (scan range 150-1000 Da) were used.

Mass spectrometric analysis on fractions collected from derivatives from the HPLC separation verified that the two first eluting derivatives had a N4w of 562.3 and the later eluting derivatives both had MHF of 809.6.

Example 3-Derivatization Yield as a Function of Catalyst Concentration.

This experiment was designed to investigate the yield in the solid phase derivatiza- tion reaction as a function of the amount of catalyst deposited onto the porous car- rier in the SPEED device. Amounts of Nafion ranging from 0.4-2.2 llg were added as 5 % (w/w) aqueous alcoholic solution (10 to 50 pL) to the PE discs and the rela- tive derivatization yield based on the combined peak height of both bis-derivatives

was studied, see Table 1. From these experiments it became evident that, within the tested range, there was no significant correlation between the derivatization yield and the amount of Nation loaded on the PE discs. The variation among all replica- tes was found to be 10.4 % R. S. D. (n=12) while it varied between 2 to 12 % at the individual concentration levels. The difference is mainly due to small variations in reaction time and the fact that the derivatization and sample injection were carried out manually in these experiments. It can thus be concluded that the catalytic action is established already with half a microgram of Nafion coated onto the disc. At- tempts at derivatizing other ketosteroids in addition to progesterone were thereafter done, using compounds containing either one or two keto-groups in the molecular backbone as model substances. Based on the number of peaks appearing in the chromatograms, it was concluded that 5c-pregnane-3, 20-dione (IV) and 5ß- pregnane-3,20-dione (V) formed bis-derivatives while 20a-hydroxy-4-pregnane-3- one (VI), 3a-hydroxy-5p-pregnane-20-one (IX) and 3i-hydroxy-5a-pregnane-20- one (VIII) formed mono-derivatives. The successful derivatization of these five model substances thus suggests that the derivatization procedure is generally appli- cable to steroids containing a carbonyl moiety.

Example 4-Experimental Design and Multivariate Data Analysis.

This experiment was set up to study the effect of reaction time and temperature on the derivatization yield for the bis-derivatives. These experimental variables were included in a centrally circumscribed composite (CCC) experimental design, and the combined areas for the peaks appearing at the characteristic retention times for the derivatives in a reversed phase chromatogram were used as responses in a partial least squares (PLS) analysis of the design, using the Umetrics (Umea, Sweden) Modde software package, see Table 2. Further experimental details are found in the footnote of Table 2. Modeling of data revealed that the reaction time (TI) had no significant effect while the reaction temperature (TE) had a strong influence on the reaction yield for formation of progesterone bis-derivatives, within the tested ex- perimental domain. After removing non-significant factors and inserting a quadratic term (TE*TE), an W = 0.97 and a Q2 = 0. 96 were obtained, showing that the pro-

jected model was valid and that the predictive power was reliable. From Figure 3 it could moreover be concluded that the best derivatization yield was obtained at room temperature (RT), and it should be noted that the lowest yield for progesterone bis- derivatives was obtained at 50 °C. This observation correlates well with the previ- ous findings on solution phase catalysis using TFMSA, where this temperature was found to give the highest derivatization yield for the mono derivatives of progester- one.' Example 5-Automated Pre-column Derivatization.

In this experiment, nineteen SPEED devices with varying amounts of Nation and DNSH were prepared and used for derivatization of progesterone using an Gilson Aspec Xli autoinjector according to a 24 full factorial experimental design; Table 3.

The purpose was to investigate the effect of the DNSH (DN) and Nation volume (NA), reaction temperature (TE), and reaction time (TI), and the evaluation was aided by the PLS analysis, as described in Example 4. Removing non-significant factors and inserting two quadratic terms (NA*NA) and (DN*DN) produced a model with an R2 = 0.91 and a Q2 = 0.80, which show that the projected model was valid and that the predictive power was adequate. Modeling of experimental data using this model revealed that the derivatization yield increased with the DNSH concentration, see Figure 4, and the highest yield was achieved after ten minutes at RT using a low amount of Nation as catalyst. These data correlate well with the CCC design described in Example 4, and it thus can be concluded that RT and a short reaction time are the best conditions for obtaining progesterone bis- derivatives.

Example 6-Comparison of Porous PE Carriers for the Solid Phase Derivati- zation Reaction.

In this experiment, the differences between different porous polyethylene carriers was sought. Disks were punched from four porous polyethene sheets with different properties and inserted into the syringe barrels, as described in the General Proce- dure for the Preparation of SPEED Devices above. The materials evaluated were all

commercially available products with pore sizes ranges and thicknesses of 15-45 nm and 3 mm (carrier A), 9 pm (average) and 3.2 mm (carrier B), 16 um (average) and 4.75 mm (carrier C), and 10 um (average) and 1.5 mm (carrier D), respec- tively. The derivatization yield was used as optimization criterion. The data ob- tained, presented in Table 4, show that there were no significant differences be- tween SPEED devices prepared from carriers A, B, and D, whereas carrier C pro- duced a device with a significantly lower catalytic ability. This cannot be explained by differences in the nominal porosities of the carriers. From the carriers producing similar results, Carrier D was selected for the continued studies, a choise mainly based on practical considerations.

Example 7-Final Optimization of the SPEED Derivatization.

This experiment was planned to chart the sensitivities to intentional variations in experimental variabilities in the domain provisionally determined as optimal. As a provocation test of the derivatization reaction a final CCC design was carried out, where the experimental domain was expanded around the provisional optimum found in the 24 full factorial design described in Example 4. The derivatization yields obtained in these experiments are presented in Table 5. As visualized in the response plot in Figure 5, the derivatization yield remained constant over a large part of the experimental domain tested. The reaction time thus played a minor role and the yield was nearly constant, as long as the amount of reagent was sufficient.

PLS analysis on the experimental data furthermore revealed that an optimum had been reached since no statistically significant factors were found. This is augmented by the low regression prediction coefficients (R2 = 0.82, Q2 = 0.66), indicating that the experimental noise was high in comparison to the systematic variations in de- rivatization efficiency resulting from the provocations. Since the parameters appear to be situated on a plateau, excellent derivatization reproducibility was obtained among replicates (2.3 % R. S. D ; n = 3) when evaluating the area response for the first eluting bis-derivative.

Example 8-Influence of Water on Derivatization Yield.

Presence of even small amounts of water in samples is known to inhibit the forma- tion of derivatives in the solution phase reaction using TFMSA as catalyst. An ex- periment was therefore set up to determine the sensitivity to the presence of water with the SPEED device. Seven standard solutions, each containing 50 uM proges- terone in varying amounts of water (0-99 %) deliberately added to the reaction mixture were prepared from progesterone stock solutions A (5104 uM) or B (505.4 MM) in methanol. The 99 % aqueous solution was prepared by weighing one mL of A and adding to a 100 mL volumetric flask, whereafter water was added to the mark while weighing. For all other standard solutions, a 5 mL quantity of B was taken to a 50 mL volumetric flask, whereafter appropriate volumes of methanol (5- 45 mL) and water (0-40 mL) were added while weighing the solution. In all experi- ments, 80 liL of standard solution (4 nmoles) was added to a SPEED device, then allowed to react for ten minutes, and finally eluted from the SPEED device with 300 pL of eluent. The eluate was then homogenized by applying three aspira- tion/discharge cycles and finally 100 uL (1.33 nmole) was injected into the reversed phase HPLC system. Triplicate analyses were carried out at each concentration level. A remarkable relationship was found, as shown in Figure 6, where the maxi- mum in reaction yield was obtained when the reaction mixture contained 10-20 % methanol.

Experiment 9-Loadability Maximum sample loadability was investigated by applying increasing volumes (80- 500 pL) of a 5.0 u. M progesterone solution (0.4-2.5 mnole) containing 10 % methanol to a SPEED device. A linear relationship between the amount of proges- terone added and the peak areas of bis derivative one in the HPLC determination of the derivatives eluted from the SPEED device was seen over the entire tested range, see Figure 7. From the chromatograms it was also apparent that the formed deriva- tives were mainly the syn-and anti-conformers of bis-derivatives, as visualized in Figure 8.

Example 10-Application to Human Serum Samples.

Extraction and Clean-up was performed as follows. Human serum (0.50 mL) was pipetted into a cylindrical flat bottom glass vial of 10 mL volume, to which water (0.5 mL) and diethyl ether (3.0 mL) were subsequently added. The samples were then allowed to swirl gently on an orbital shaker for ten minutes. Following the liq- uid-liquid extraction, the vials were transferred into an ethanol/dry ice bath, where the aqueous phase was frozen. The ether phase was then decanted and evaporated under a stream of helium. The residue was finally re-dissolved in400 uL of a wa- ter/methanol mixture (90: 10 v/v) prior to analysis. To determine the analytical re- covery of the technique, known amounts of progesterone (0.2 and 2.0 nmoles) were added to an extraction vial and the solvent was evaporated to dryness at room tem- perature prior the addition of serum. The repeatability, the reproducibility, and the analytical recovery were evaluated using standard solutions ranging from 0.5 to 5 uM (0.2 to 2.0 mnole in the reaction mixture) and spiked human serum containing 0.4 to 2.0 nmole of progesterone. The analytical system was capable of a retention time repeatability of < 0. 8 % R. S. D. within one day of analysis (n = 10), and < 1.0 % R. S. D. for all samples (n = 45). The reproducibility was calculated for sev- eral replicated derivatizations of 0.5 to 5.0 uM standards over two days of analysis.

At the 0.5 and 5.0 u. M levels, the R. S. D. ranged between 10 to 17 %, while it varied between 20 to 27 % at the 1.0 and 2.5 u. M levels (n = 33). Using spiked human se- rum samples containing 0.4 or 2.0 nmoles of progesterone, the recoveries were 52 10 % and 63 17 %, respectively. The limit of detection for progesterone was estimated from calibration curve data to 1.3 pmole (100 RL injected), based on three times the standard deviation of four 0.5 tM standard solutions. The recoveries were in the same range as in previous studies, where also liquid-liquid extraction was ap- plied prior to heterogeneous TFMSA catalyzed solution dansylation of various C-21 ketosteroids. '' We hence conclude that the recovery deficiency of the presented technique may be an effect of the sample pre-treatment, rather than an inferior func- tion of the SPEED device, and that the SPEED device performs sufficiently well to

be incoprorated as part of an analytical analysis scheme for the investigated com- pounds.

Example 11-Derivatization of formaldehyde.

This experiment was designed to demonstrate the applicability of the SPEED device for the capture, stabilization, and derivatization of a small and volatile aldehyde.

Formaldehyde (100 uL) was added to a SPEED device and the reaction was allowed to proceed for ten minutes at room temperature. Thereafter the formed derivatives were eluted with 300 iL of mobile phase, of which 100 ut was injected in a re- versed phase high-perfonnance liquid chromatographic system using a mixture of acetonitrile and 12 mM phosphate buffer at pH 6 50: 50 v/v as mobile phase and a flow rate of 0.5 mL/min. A peak identified as the derivative of formaldehyde and DNSH was detected in the chromatogram, see Figure 9.

Example 12-Derivatizationofpelarg ; onaldehyde.

This experiment complements Example 11 and shows the derivatization of a less volatile and less reactive aliphatic aldehyde. Pelargonaldehyde (n-nonanal; 100 u. L) was added to a SPEED device and the reaction was allowed to proceed for ten min- utes at room temperature according to the procedure previously presented. Thereaf- ter the formed derivatives were eluted with 300 u. L of mobile phase and 100 iL was injected in a reversed phase high-performance liquid chromatographic system and separated as described in Example 11. A peak identified as the derivative of pelar- gonaldehyde and DNSH was detected in the chromatogram, see Figure 9.

Although the invention has been described by way of specific embodiments, it is not intended to be limited thereto. As will be apparent to those skilled in the art, numer- ous embodiments can be made without departing from the spirit of the invention or the scope of the following claims.

Table 1. Derivatization yield as a function of the amount of catalyst coated onto the porous disc'\ Experiment Nafion# Amount Total Peak Heightb) R.S. D.

µ (units/105) (%) 1 0. 44 2.13 A 0.23 10.8 2 0. 88 2. 57 ~ 0. 06 2.2 3 1.31 2.20 0.26 12.0 4 2.18 2. 47 ~ 0. 06 2.3 a) In each experiment, three 15 nL aliquots of 5 uM progesterone standard were derivatized using 10 µL of a 2 mM DNSH solution. Formed derivatives were eluted with 1 mL of HPLC eluent and the volume injected in the chromatographic system was 100 liL. b) Total peak height corresponds to the integrator response for both syn and anti bis-derivatives of progesterone.

Table 2. The effect of temperature and time on dansylation of progesterone according to the CCC designa). Experiment Variables Responses TE TI (units/105) (units/105) (%) (units/106) (units/106) (%) 1 32 6 0.74 0.46 71 1.68 1.0 71 2 30 30 0.96 0.55 89 2.17 1.79 88 3 28 60 1.03 0.64 99 2.32 2.13 99 4 28 37.5 1.00 0.54 91 2.25 1.67 87 5 40 37.5 0.62 0.39 60 1.41 1.43 63 6 40 37.5 0.67 0.40 63 1.52 1.34 64 7 50 15 0.52 0.31 49 1.19 1.00 49 8 50 60 0.56 0.34 53 1.38 1.28 59 9 25 25 1.02 0.55 92 2.32 1.82 92 10 25 60 1.09 0.60 100 2.50 1.99 100 11 85 10 0.76 0.54 76 1.83 2.15 88 a) The following factors were studied: Temperature (TE) in °C and reaction time (TI) in minutes. Ten microliters of 5 µM progesterone standard, 50 µL of<BR> 2 mM DNSH, and 30 µL of Nafion# was used in all experiments. The sorbent was then dried with nitrogen and the derivatized samples were eluted with 1 mL<BR> of eluent. The injected sample volume was 100 µL. Experiment number 10 showed the highest response; the total area and total height were consequently<BR> normalized to this experiment.

Table 3. Results from the 2f Ful Factorial designa). Experiment Variables Responses CC DN TE TI Peakheight 1 Peakheight 2 Total Height Peakarea 1 Peakarea 2 Total Area µg mM °C Min (units/105) (units/105) (units/105) (units/107) (units/107) (units/107) 1 0.87 1 25 10 1.49 0.79 2.28 0.30 0.18 0.47 2 2.18 1 25 10 1.48 0.81 2.28 0.30 0.19 0.50 3 0.87 3 25 10 4.15 2.30 6.45 0.81 0.51 1.32 4 2.18 3 25 10 3.05 1.61 4.66 0.59 0.35 0.94 5 0.871 40 10 1.69 1.05 2.73 0.32 0.24 0.56 6 2.181 40 10 1.49 0.93 2.42 0.28 0.20 0.48 7 0.873 40 10 3.88 2.50 6.38 0.61 0.55 1.16 8 2.18 3 40 10 2.21 1.20 3.40 0.43 0.27 0.70 9 0.871 25 50 1.81 1.05 2.86 0.34 0.24 0.57 10 2.18 1 25 50 0.90 0.56 1.46 0.17 0.13 0.30 11 0.873 25 50 3.15 1.64 4.79 0.60 0.37 0.97 12 2.18 3 25 50 2.88 1.51 4.39 0.55 0.32 0.88 13 0.87 1 40 50 1.81 1.14 2.95 0.35 0.28 0.63 14 2.18 1 40 50 1.63 1.03 2.66 0.29 0.24 0.53 15 0.87 1 40 50 2.27 1.35 3.62 0.45 0.31 0.76 16 2.18 3 40 50 1.77 1.08 2.86 0.35 0.26 0.62 17 1.53k 2 30 30 3.04 1.64 4.68 0.56 0.36 0.93 18 1.53 2 25 30 3.70 2.17 5.87 0.62 0.481.10 19 1.53 2 25 30 3.69 2.07 5.76 0.68 0.46 1.15 a) Catalyst amount in µg (CC), concentration of the derivatization reagent in mM (DN), reaction temperature in °C (TE) and reaction time in minutes (TI). Progesterone<BR> (35 µL of a 100 µM solution in 100 % methanol) was used as standard and derivatized in all experiments Table 4. Comparison of PE discs used as carriers for the reagents.

PE Material Material characteristics Responses Pore size Thickness Total peak area R. S. D n uM mm Units/107 (%) A 15-45b) 3 1.05 + 0. 10 9.3 4 B 9 3.2 1.03 ~ 0.04 3.9 3c) C 16 4.75 0.70 0.02 3.6 4 D 10 1.5 1. 06 0. 08 7.4 4 a) Four PE discs with different thickness and porosity were studied: In all experiments were 30 IAL of Nation@, 50 µL of a 3 mM DNSH and 15 IlL of a 100, uM Progesterone standard solution used. The reaction time was set to 10 minutes at RT. The formed derivatives were eluted with 1 mL eluent and 100 aL was injected in the HPLC system. b) Only pore size distribution was given by the vendor. c) One of the replicates for carrier B was found to be an outlier and consequently excluded. However, as the experiment was at the boundary of being excluded we report all the individual values for this carrier (1.02,1. 39,0.99 and 1.07).

Table 5. Results from the CCC design using commercially packed syringes. a) Variables Responses Experiment DN TI Peak 1 height Peak 2 height Peak 1 area Peak 2 area Total area mM minutes (Units/104) (Units/104) (Units/107) (Units/107) (Units/107) 1 2. 46 8 2.26 1.59 0.64 0.49 1.13 2 3.55 8 3.68 2.50 1.02 0.80 1.83 3 2.46 12 1. 93 1.38 0.54 0.40 0. 93 4 3.55 12 3.56 2.40 0.96 0.75 1.71 5 2.30 10 2.24 1. 77 0.69 0. 59 1.28 6 3.63 10 3.37 2.26 0.96 0.74 1.71 7 3 7.17 3.18 2.20 0.83 0.63 1.46 8 3 12.83 3.76 2.45 0.96 0.70 1.66 9 3 10 3.40 2.25 0.93 0.67 1.60 10 3 10 3.15 2.24 0.90 0.74 1.63 11 3 10 3.25 2.27 0.89 0.72 1.61 a) The following factors were studied: Concentration of the derivatization reagent in mM (DN) and reaction time in minutes (TI). Progesterone (40 aL of a 49.1 uM standard solution) was derivatized throughout all experiments. The formed derivatives were eluted with 500 ttL of eluent and 168 uL was injected into the HPLC system.

LITERATURE CITED la Baulieu, E-E.; Robel, P., J. Steroid Biochem. Mol. Biol. 1990,37 (3), 395-403. lb Baulieu, E-E., Psychoneuroendocrinology. 1998,23 (8), 963-987. lc Compagnone, N. A.; Mellon, S. H.; Frontiers in Neuroendocrinology, 2000,21, 1-56.

2 Ma, Y-C.; Kim, H-Y., J Am. Soc. Mass Spectrom. 1997,8,1010-1020.

3 Liere, P.; Akwa, Y.; Weill-Engerer, S.; Eychenne, A.; Pianos, B.; Robel, P.; Sjövall, J.; Schumacher, M.; Baulieu, E-E., J. Chromatogr. B 2000,739,301.

4 Wolthers, B. G.; Kraan, G. P., J. Chromatogr. A 1999,843,247-.

5 Higashidate, S.; Hibi, K. ; Senda, M.; Kanda, S.; Imai, K., J. Chromatogr. 1990, 515, 577-584.

6 Appelblad, P.; Jonsson, T.; Bäckström, T.; Irgum, K., Anal. Chem. 1998,70, 5002-5009.

7 Koziol, T.; Grayeski, M. L.; Weinberger, R. J., J. Chromatogr. 1984,317, 355 8 Kawasaki, T.; Maeda, M. ; Tsuji, A., J. Chromatogr. 1979,163,143 9 Iwata, J.; Suga, T. J., J. Chromatogr. 1989,474,363 10 Kawasaki, T.; Maeda, M.; Tsuji, A., J. Chromatogr. 1983, 272, 261 11 Hyytiäinen, M.; Appelblad, P.; Pontén, E.; Stigbrand, M.; Jaegfeldt, H.; Irgum, K. J. Chromatogr. 1996,740,279-283.

12 Appelblad, P.; Pontén, E.; Jaegfeldt, H.; Bäckström, T.; Irgum, K. Anal. Chenu.

1997,69,4905-4911.

13 Howells, R. D., McCown, J. D., Chem. Rev. 1977,77,72

14 Goehl, T. J.; Sundaresan, G. M.; Prasad, V. K., J. Pharm. Sci. 1979,68 (li), 1374-1376 15 Weinberger, R.; Koziol, T.; Millington, G., Chromatographia 1984, 19, 452-456 16 Herraez-Hernandez, R.; Campins-Falc6, P.; Sevillano-Cabeza, A., Anal. Chem.

1996,68,734-739 17 Vogel, M.; Karst, U.; poster presented at the International conference on HPLC'99 in Granada, Spain, May 1999.

18 Waller, F. J.; Van Scoyoc, R. W., Chemtech 1987,17 (7), 438-441 19 United States Patent US 4,433,082