Analysis of intact steroid conjugates by secondary ion mass spectrometry (including FABMS) and by gas chromatography

0022-4731/83 $3.00 + 0.00 Copyright © 1983 Pergamon Press Ltd

J. steroid Biochem. Vol. 19. No. I, pp. 209-217. 1983 Printed in Great Britain. All rights reserved

ANAL YSIS OF INTACT STEROID CONJUGATES BY SECONDARY ION MASS SPECTROMETRY (INCLUDING FABMS) AND BY GAS CHROMATOGRAPHY C. H. L. SHACKLETON*§, v. R. MATToxt and 1. W. HONouRt "Departments of Laboratory Medicine and Pharmaceutical Chemistry, S1124, University of California, San Francisco, CA, 94143, U.S.A. tThe Mayo Clinic, Rochester, MN, U.S.A. tClinical Research Centre, Watford Road, Harrow, Middlesex, England SUMMARY

The analysis of intact steroid conjugates by two different methods is described. One method employed secondary ion mass spectrometry (SIMS) using a Cs" beam for ionisation, although comparable data were obtained by fast atom bombardment (F AB) using a Xe" beam. In both of these mass spectrometric techniques the samples were analysed in a liquid matrix (glycerol). Positive and negative ion spectra have been obtained, the latter being most useful for steroid sulphate and glucuronide analysis. The negative ion spectrum of each steroid is dominated by a pseudomolecular ion at mlz [M - H] " (M of free acid) and a lackof marked fragmentation. Mixtures of steroids can be resolved in a single spectrum, providing the individual steroids differ in mass. Thesecond method was gaschromatography. The carboxylic acid moieties of the steroid glucuronides were derivatised with diazomethane and the remaining functional groups in the steroids were thermally protected by methyloxime formation (for carbonyls) and trimethylsilylation (for all steroidal and glucuronicacid hydroxyls). Satisfactory analysis of steroid glucuronides was achieved through the use of glass or fused silica columns stable at high temperature (330°C). Conveniently, trimethylsilylation resulted in exchange of the sulphate in 3p-hydroxy-5-ene steroid sulphates for a trimethylsilyl group so these could effectively be analysed as "free" steroids. Two available methods will be discussed. The first is a new mass spectrometric technique which allows Gas chromatography on capillary columns has mass spectra to be obtained with ease for polar become the method of choice for analysing the major charged molecules. The technique was originally human urinary steroids [I , 2, 3]. Classical method- termed FAB (fast atom bombardment) since ionisaology relies on the hydrolysis of the steroid conju- tion was achieved by bombarding the sample with a gates prior to analysis. The most frequently employed beam of neutral atoms (e.g. Ar? or XeD) [5]. Later methods use the enzymes present in Helix pomatia studies have shown that a neutral beam is not essendigestive juice which hydrolyse steroid glucuronides tial and in our laboratory we use either a XeD atom or and most steroid sulphates. Some methods depend on Cs" ion beam [6]. The initial success of the method the separation of individual conjugate groups (e.g. sul- was not due to the type of primary beam but due to a phates, glucuronides) by ion-exchange lipophilic viscous liquid matrix (e.g. glycerol) in which the anaSephadex prior to enzymatic hydrolysis [3]. Where Iyte was suspended. Positive and negative ion mass only sulphates are of interest, solvolysis in ethyl acet- spectra are readily obtained using this technique (to ate or tetrahydrofuran can be employed [3,4]. There be referred to as secondary ion mass spectrometry, or is no doubt that at the present time the above metho- SIMS) and for most molecules pseudomolecular ions dologies olTer the most reliable way of obtaining com- predominate in the spectra and minimal fragmentaprehensive information of steroid excretion. tion is apparent [7]. However, it is the purpose of this communication The second method for analysis of steroid conjuto discuss the development of methods for analysing gates involves gas chromatographic analysis of steroid steroid conjugates directly without hydrolysis. If this glucuronides as methylester and trimethylsilyl ether were possible the time required for profile analysis derivatives. This approach was first applied in 1967 could be shortened considerably and results would be by Jaakomaki et al.[8] but only glucuronides of less likely to reflect losses occurring during the pro- simple steroids were analysed (androsterone, etiochocedure. lanolone and pregnanediol glucuronides). It was recognised at that time that special problems existed § Please address all correspondence to C. H. L. Shackle- for the analysis of corticosteroid glucuronides which ton, Department Pharmaceutical Chemistry, University of contain many additional functional groups. The mass California SanFrancisco, San Francisco, CA 94143, U.S.A. of the methylester methyloxime-trimethylsilyl ether of INTRODUCTION

209

210

C. H. L.

SHACKLETON

tetrahydrocortisol glucuronide, for example, is lOI7 which has rendered volatilisation and chromatography extremely difficult. Also, we realised that for individual corticosteroids resolutions may be inadequate since column efficiency tends to decrease with increasing temperature. Fortunately, glass capillary columns can now be prepared of sufficient stability to ,allow operation at temperatures in excess of 300° [9J which permits elution of relatively large molecules. Direct derivatisation of steroid sulphates has been demonstrated by Touchstone and Dobbins[IOJ and Murray and Baillie[ll]. These studies demonstrated that phenolic and 3fJ-hydroxy-S-ene steroids could be converted to a perfluorester by reaction with perfluoracid anhydrides. Replacement of the sulphate group by a trimethylsilyl group has also been demonstrated for phenol sulphates, although not for steroids [12]. Based on these experiences, we considered direct silylation of 3fJ-hydroxy-S-ene steroid sulphates before profile analysis. EXPERIMENTAL

Steroids Many of the steroid sulphates and glucuronides used in this study were obtained from the MRC Steroid Reference Collection (D. N. Kirk, Curator), the Chemistry Department, Westfield College, London. Glucuronides of corticosteroid metabolites were prepared by one of us (V.R.M.) by a method previously described [13].

Steroid abbreviations The following steroid abbreviations have been used-AlloTHF, (SIX-THF) allotetrahydrocortisol: 31X,11fJ, 171X,21-tetrahydroxy-SIX-pregnan- 20-one; alloTHB, (SIX-THB) allotetrahydrocorticosterone: 31X,lIfJ,21-trihydroxy-SIX-pregnan-20-one; HHB, hexahydrocorticosterone: SIX(p)-pregnane, 31X, II fJ,201X,21tetrol; THA, tetrahydro Compound A: 31X,21-dihydroxy-SfJ-pregnane-Il,20-dione; HHA, hexahydroCompound A: 31X,201X,21-trihydroxy-SIX(p)-pregnan-llone; THS, tetrahydrosubstance S: 31X,17,21-trihydroxySfJ-pregnan-20-one; 16-oxo-androstenediol: 3fJ,17fJdihydroxy-S-androsten-16-one; d 5AT, androstenetriol: S-androstene-3fJ, I61X, I7fJ-triol; 161X-hydroxypregnenolone: 3fJ, 16-dihydroxy-S-pregnen-20-one; pregnanolone: 31X-hydroxy-SfJ-pregnan-20-one; pregnanolone: 31X-hydroxy-SIX-pregnan-20-one: d 5pD, pregnenediol: S-pregnene-3fJ,201X-diol; d 5PT, pregnenetriol: S-pregnene-3fJ,171X,201X-triol.

Urine and plasma processing Extraction-r-Urine. Extraction of steroid conjugates (sulphates and glucuronides) was carried out according to the method of Shackleton and Whitney[14J employing Sep-pak C 18 cartridges. Briefly, up to SO ml urine was passed through the cartridge from a syringe. Conjugates were retained by the cartridge, letting water and salts pass through. The cartridge

et al.

was washed with S ml distilled water and the conjugates were recovered by elution with 4 ml methanol. Extraction-plasma. Plasma (2-S ml) was dripped into 2S ml acetone-ethanol (I: I, v/v) present in 40 ml centrifuge tubes. The tubes were continuously sonicated for 30 min and were then kept at - 20° overnight. The tubes were centrifuged and the supernatant decanted. The pellets were sonicated with a further IS ml of the extracting solvent and after a brief cooling period (30 min, - 20°) the samples were centrifuged and the solvent decanted and combined with the first extract. The solvent was removed under reduced pressure by rotary evaporation.

Separation of steroid sulphates and qlucuronides Separate steroid conjugate fractions were prepared by two methods: (I) Sephadex LH-20 chromatography [IS], or (2) DEAP-LH-20 chromatography [16]. (1) The sample was separated on a gravity feed column of Sephadex LH-20 (4 g) using the solvent system methanol-chloroform (I: I, vl», salt saturated). Free steroids and steroid glucuronides were eluted in the first 30 ml. Steroid monosulphates were eluted between 30 and 6S ml and recovery of steroid dilsulphates was achieved by elution of the column with SO ml methanol. The steroid fractions from LH-20 chromatography were desalted by dissolving the samples in water and extracting the conjugates by Sep-pak cartridge according to a previously described method [14]. (2) The sample was dissolved in 20 ml 72% ethanol and was passed through an Amberlyst A-IS column previously washed with 20 ml 0.1 M HCI in 72% ethanol and until neutral with 72% ethanol. The dimensions of the bed were lOx SO mm. The initial 20 ml and a further SO ml of 72% ethanol were collected and dried. A DEAP-LH-20 column (S x 130 mm) was prepared in 72% ethanol. The sample was applied in S ml of the solvent and the following fractions were collected: (I) 20ml 72% ethanol (neutral steroids), (2) 20 ml O.lS M acetic acid (steroid acids), (3) 20 ml O.2S M formic acid (glucuronides), (4) 20 ml acetic acid, potassium hydroxide, OJ M, pH 6J (mono sulphates), and (S) acetic acid, potassium hydroxide, O.S M, pH lO (disulphates). The samples were desalted by Sep-pak extraction by the method employed after LH-20 chromatography. Derioatisation (for gas chromatography only). The C 22 , C 36 , and C 4 4 n-alkanes were added to the samples as internal standards. Methylation of glucuronic acid carboxyl. Ethereal diazomethane was prepared in a Wheaton generator according to the method desccribed by Fales et al.[I7]. One ml of the ethereal diazomethane was added to the sample in So-IOO,u1 methanol. After 20 min the solvent was removed under nitrogen. Methyloxime formation. The methylester derivative

211

Intact steroid conjugate analysis was dissolved in 100 J.ll methox yamine hydrochloride in pyridine (2% w/v) and condensation was allowed to proceed at 60° for I h. The pyridine was blown off under nitrogen. Trimethylsilylation. The methyl ester, methyloxime derivative was heated at 100° with 100 J.lI trimethylsilylimidazole for 16 h. Derivative purification. Excess reagents and other involatiles were remo ved from the sample by Lipidex chromatography using an adaptation of the method of Axelson and Sjovall]'18].

Classical hydrolysis methods Results obtained by the 'direct der ivatis ation' method were compared to those obt ained by solvolysis [4] and enzymatic hydrolysis. Solvolysis was carried out by dissolving the extract in 3 ml of acidified ethyl acetate and allowing the reaction to proceed at 40° for 16 h. The solvol ysate was washed with I ml of 1 N NaOH and 1 ml of water before being dried . For enzymatic hydrol ysis, the urin e was brought to pH 4.6 by add ition of 0.1 vol.% 5 M acetic buffer [I]. The Helix pomatia preparat ion was added and hydrol ysis was allowed to proceed fo r 48 h. The free steroids were extracted by the Sep-pak method. Products from the solvolysis and the enzyme hydrolysis of the conjugates were converted into methyloxime trimethyl silyl ether derivatives in the usual manner.

Gas chromatography and gas chromatography/mass spectrometry. Gas chromatography was carried out on glass or fused silica columns coated with O Y-1. These columns were prepared in the laboratory. The empty columns were deacti vated by the silanisation technique of Grob et al.[9] prior to static coating with the stationary phase. The columns were housed in a Packard 430 gas chromatograph equipped with carousel automatic solid sampler [1]. The samples were injected with oven temperature at 60° (injector temperature 300°). Following an initial period of I min, the oven temperature was quickly raised to 180° for a second isothermal period of 3 min. The temperature was then programmed to 330° at a rate of 3°/min.

Secondary ion mass spectrometry Mass analysis was performed on a Kr atos MS-50S mass spectrometer equipped with a 23 KG magnet and means for negative ion switching. Typical operating conditions were : scan rate , 30 s/decade ; dynamic resolution, M/L\M 3000). The spectrometer was interfaced to our laboratory's LOGOS-II data system, permitting real time assignment of masses. A standard Kratos FAB gun and ion source chamber were used to obtain the neutral (XeD) primary beam FAB spectra. Details of operation of the SIMS Cs ' source have been described in a separate communication [6]. The primary Cs + ion beam gun (located within the ion source housing of a standard Kratos FAB ion source)

pr oduced a 6 keY Cs ' beam focussed to a spot size between 1 and 2 mm dia at the target surface. A Cs + current of ab out 1.0 J.lA was used. Samples were prepared by either dissolving or making a slurry of the solid in glycerol. Alternatively, I J.l1 of a solution of the sample (1-10/lg/J.l1) dissolved in water was added to about 3 J.lI of glycerol on the copper probe tip and the excess water pumped away in the vacuum lock of the mass spectrometer. Calibration of the mass spectrometer was achieved using Ultra mark 1621 (PCR, Gainesville, FL). RESULTS

Secondary iOIl mass spectrometry (Sl MS) The data presented here has been obtained using either the Cs + ion or XeD beam (FAB). Spectra produced by these techn iques are indistinguishable so the primar y beam used in a particular case will not be reported. Steroid qlucurontdes. The SIMS spectra of all stero id glucuronides are essenti ally similar and a typical example, pregnanediol-3-glucuronide is illustrated in Fig. I (upper spectrum). In this negative io n spectrum the base peak w]z 495 has a mass equi valent to [M - H] - (M equivalent to free acid) with a smaller peak at an mlz value corresponding to [MNa - Hr . Several fragment ions oflow mass are also seen. The positive ion spectrum (not illustrated) is dominated by an ion of mass equivalent to [M + 2Na] + although ions of mass corresponding to [M + Na] + are also significant. Positive ion spectra are generally more complex and are less suitable for analysis of steroid conjugate mixtures. Estrogen glucuronides beha ve in similar fashion to neutral steroid glucuronides. In the negative ion spectrum of oestriol-16-glucuronide (Fig. 1), the pseudomolecular ion [M - H] ~ is at I11/Z 461 Negative ion mass spectra of corticosteroid glucuronides (tetra hydrocortisone, tetrah ydrocortisol and «-corto lo ne] have been published in a previous commun icat ion [7]. Steroid monosulphates. The pseudomolecular ion in the positive ion spectra of steroid mon osulphates is at m]z equi valent to [M + 2Na] + although ions at mlz [M + Na] + are also significant. In the negative ion spectra the pseudomolecular ion is [M - H] - which for pregnenolone sulphate is at miz 395 (Fig. 2). Steroid disulphates. In a previous communication the SIMS spectrum of pregnanediol-3,20-disulphate was illustrated [7]. A useful positive ion spectrum was not obtained and the pseudomolecular ion of the negat ive ion spectru m was at [MNA - H] - . Table 1 lists the pseudomolecular ions (negati ve ion SIMS ) for man y of the most important steroid sulphates and glucuronide present in plasma and urine. Urine samples. The relative simplicity of negative ion SIMS spectra encouraged us to attempt analysis of steroid conjugate mixtures. In previous communications [7, 19] characteristic profiles of urinary

212

C. H . L. SHACKLETON et al.

495

Pregnanediol-3-Gluc.

II

~

~

I

II

Estriol-16-Gluc.

I

463

II

.It

.L

Fig. I. Negative ion SIMS spectra of pregnanediol-3-g1ucuronide and oestriol-16-g1ucuronide.

steroid sulphates and glucuronides were obtained, for steroids in urine of patients with enzymatic defects in the biosynthetic pathway leading to cortisol, ie, 17hydroxylase deficiency, 3p-hydroxy-steroid dehydrogenase deficiency, 21-hydroxylase deficiency and liP-hydroxylase deficiency could be clearly differentiated by this technique [19]. In the current study the SIMS method was used to examine the sulphate and glucuronide conjugated steroids present in late pregnancy urine and the results obtained were compared to those obtained from urine of a patient with placental sulphatase deficiency (PSD). Deficiency of this enzyme results in the inability of the placenta to convert the fetal 3P-hydroxy-5-ene steroid sulphates to estriol. In the "normal pregnancy" spectrum (Fig. 3) the major ions represent pregnanediol glucuronide (mlz 495, cf. Fig. 1) and related compounds. The pseudomolecular ion of estriol glucuronide is clearly seen at mlz 463 (cf. Fig. I). All steroid conjugates below mass mlz 450 are sulphate conjugated and above this mass are steroid glucuronides although steroid disulphates, if present, would also be in the mass range of steroid

In the spectrum produced from urine of a patient with PSD the major components are steroid monosulphates (Fig. 3) although pregnanediol glucuronide and related compounds are present. As expected the ion corresponding to estriol glucuronide is small (mlz 463). Deutero-glycerol was used as matrix for this sample as a complex of unlabelled glycerol gives an

glucuronides,

Fig. 2. Negative SIMS spectrum of pregnenolone sulphate.

Pregnenolone Sulfate

.l.L

395

.~".

21 3

Intact steroid conjugate analysis Table l. Pseudomolecular ions (negative ion SIMS) of human urinary and plasma steroid sulphates and glucuronides

mlz Glucuronides Oestriol Androsterone and etiocholanolone l l-Oxo-etiocholanolone IIp.Rydroxyandrosterone and IIp-hydroxyetiochololanolone Pregnanolone Pregnanediol J l-Oxo-pregnanediol l7a-Rydroxypregnanolone Pregnanetriol Tetrahydro-ll-dehydrocorticosterone 11-0xo-pregnanetriol Tetrahydrosubstance S 5a-Tetrahydrocorticosterone and tetrahydrocorticosterone

463 465 279 481 493 495 509 509 511 523 525 525 525

Hexahydrosubstance S

(5p-pregnane-3a,17a,20a(Pl-tetrol) Tetrahydrocortisone Tetrahydrocortisol and 5a-tetrahydrocortisol a and P-Cortolone a and p-Cortol

527 539 541 541 543

Sulphates Dehydroepiandrosterone sulphate (DHAS) Ifux-Hydroxy DRA Pregnenolone 5-Pregnene-3p,20a-diol 17a-Hydroxypregnenolone 5-Pregnene-3p,17a,20-triol

367 383 395 397 411 413

Gas chromatography All steroid glucuronides ranging in polar ity and molecular weight from androsterone through e-cortol were eluted between the C 3 6 and C 4 4 alkanes, with elution temperatures between 3000 and 3300 under the conditions used. Table 2 lists the relati ve retention times (methylene units) for both reference and urinary glucuronides. Figure 5 illustrates a gas chromatogram of corticosteroid metabolite glucuronides. Response factors (peak height relative to peak height equivalent amounts of intern al standards) var ied considerably for each compound. Several steroids (e.g. androsterone glucuronide, pregnanediol glucuronide p-cortol and f1-cortolone glucuronides gave a response of near unity but steroids with glycol or dihydroxyacetone sidecha ins (i.e. the major corticosteroid metabolites) gave response of about 25%. However, even when the response was relatively poor, only single chromatographic peaks were given when each corticosteroid glucuronide was analysed alone . Figure 6 illustrates a profile of steroids from a normal adult male. All the normal constituents can be identified but it is plain that the peaks of tetrahydrocortiso ne and tetrahydrocortisol are small compared

17·0H·P'one

Pregnancy urine. Normal PO

PT I

ion at mlz 367 which would contribute to the DHAsulphate pseudomolecular ion. Plasma sample. The negative ion spectrum shown in Fig. 4 represents the steroids present in the monosulphate fraction of umbilical cord blood. All significant ions present are consistent with the known steroids present in this fluid. The ion at m/:!. 367 represents DHA-sulphate ; mjz 383, 16C(-hydroxy-DHA-sulphate ; mlz 393, pregnenolone-sulphate ; mjz 395, 5-pregnene3p,2()c(-diol sulphate; mlz 397, pregnanediol sulphate; mjz 211,5-pregnenetriols; and mlz 213, pregnanetriols. The relative intensities of ions in the spectra are not necessarily proportional to the amounts of steroids present. Hence , although 16C(-hydroxy DHA sulphate is the major component in the sulphate fraction of umbilical cord blood and PSD urin e it appears not to be one of the most intense ions in the SIMS spectra shown in Figs 3 and 4. This is almost certainly due to relatively large differences in "response factors " between different steroid conjugates. This is also evident among the glucuronides, e.g. it is unlikely that there is a greater concentration of l7-hydroxy pregnanolone glucuronide than pregnanediol glucuronide in a pregnancy urine sample although the spectrum illustrated in Fig. 3 suggests this is the case. Clearly, the "relative response factors" of all steroid conjugates of interest needs to be determined.

E3 \

And, Etio, /

OHA Placental Sulfet..e Deficiency

16·0H·

OHA /

PO 17.0H.P 'one

I

~

J

aI

~

I I §

~ ~

! , , I ! , , • "'Iz

~

I ! I I I

~

Fig. 3. Negative ion SIMS spectra of glucuronide and sulphate conjugated steroids from urines of a normal pregnant woman and a patient with Placental Sulphate Deficiency. Abbreviations not previously noted-E) : oestriol; PD : pregnanediol; /l5PD: 5-pregnene-3p,20a-diol; /l5 PT: 5-pregnene-3p, 16a (or l7a) 20a-triol; PT: pregnanetriol; l7-0H-P' one: l71X-hydroxy-pregnanolone ; And: androsterone and Etio: aetiocholanolone.

214

C. H. L.

SHACKLETON

fob neg

Umbi Iica I plasma monosulfate

DHA

J6-ond J7-0HtrP'one 16-0H-

DHA

/

360

370

380

PT

/

+,65 PT

390 400 410

I

420 430

Fig. 4. Negative ion SIMS spectrum of steroid sulphates from umbilical cord blood. Most abbreviations are listed in legend to Fig. 3 8 5 P'one: pregnenolone; 16- and 17-0H,1,5P'one: 16il( and 17il(-hydroxypregnenolone.

to androsterone and etiocholanolone. These steroids are all excreted in similar amounts so the discrepancy is due to the lower response factors of the cortisol metabolites. The GC method was used to examine the steroid excretion by patients with disorders of steroid synthesis or metabolism. In Fig. 7 separation of steroid glucuronides from a patient with 17-hydroxylase deficiency is illustrated. The upper chromatogram illustrates analysis of derivatives of hydrolysed steroid glucuronides and shows that the major steroid excreted was 5cHetrahydrocorticosterone. The identity of all steroids in this sample was verified by GC-MS analysis. Derivatisation of the intact glucuronides resulted in a marked increase in retention time of all components compared to the derivatised "free" steroids (Fig. 7, lower chromatogram) but 5ex-tetrahydrocorticosterone glucuronide was eluted successfully at 315°.

et al.

Direct analysis of steroid sulphates For this study, urine containing predominantly steroid sulphates was analysed. In this urine of a patient with 3fJ-hydroxy-steroid dehydrogenase deficiency (3fJ-OHSD) approximately 80% of the identifiable urinary metabolites are 3fJ-hydroxy-5-ene steroids [21] excreted almost exclusively as sulphates. The urine profiles are illustrated in Fig. 8. The upper chromatogram was obtained by derivatisation of an enzyme hydrolysed urine and the lower chromatogram was obtained by first direct derivatisation of a total conjugate extract, without prior hydrolysis. The profile of steroid sulfates produced by direct derivatisation was almost indistinguishable from a profile produced following solvolysis or enzyme hyTable 2. Relative retention times (methylene units, MUl for derivatives of glucuronide conjugated steroids MU conjugated

Steroid Androsterone Aetiocholanolone Il-Oxo-aetiocholanolone ll{l-Hydroxyandrosterone 1I {I-Hydroxyaetiocholanolone Pregnanediol 17il(-Hydroxypregnanolone Pregnanetriol TetrahydroDOC TetrahydroSubstance S TetrahydroCompound A Tetrah ydrocorticosterone 5il(-Tetrahydrocorticosterone Tetrahydrocortisone Tetrahydrocortisol 5il(-Tetrahydrocortisol «-Cortolone il(-Cortol {I-Cortol {I-Cortolone Oestrone Oestradiol Oestriol-3-glucuronide Oestriol-16-glucuronide

36.84 37.l6 37.89 37.61 37.89 39.50 39.80 40.87 40.70 41.29 41.47 41.29 40.46 42.14 42.33 41.25 43.00 43.00 42.70 43.26 39.59 39.76

40.88 40.33

THF

~Cortol

Corticosteroid Metabolite Glucuronides

~/ '

THS

Cortolona

\ Alkanes 25 ng Tatrahydro compels 100ng 20 Dlhydro" 50 ng

Fig. 5. Gas chromatography of a mixture of reference corticosteroid metabolite glucuronides as methylester, methyloxime, trimethylsilylethers. The conditions of analysis are described in the text.

Intact steroid conjugate analysis

Normal Male Adult

2 15

An ~

SOng

Et

Glucuronides

Ii C44 25ng

DHAS I

Fig. 6. Gas chromatographic profile of steroid glucuronides and sulphates from adult urine . The glucuronides are methylester, methyloxime, trimethylsilyl derivatives. Steroid sulphates (DHAS and AS triol S) have their sulphate group replaced by a silyl gro up during the derivatisation.

drol ysis of a sulphate fraction. In comparing the two methods quantitatively, intern al standards (5a-and rostane-3a,17a-diol, stigmasterol and cholesteryl but yrate) were added to plasma or urine steroid fractions prior to division of the sample. When the results were compared direct derivatisation only gave

peak height s 50-6 8% of the height s of peaks produced from steroids released from sulphate conjugation by solvolysis. We therefore investigated the time course for hydrolysis by direct derivatisation . After adding intern al standards, a sulphate fraction from 3{J-OHSD urine alioTHB

17o-HYDRQXYLASE DEFICIENCY Partial Condit ions 20m ov i WCOT Prog.180 ·3 30 3'Ymin. Solid Injection

Hydrolysis - der ivat izat ion

alloTHF

I

HHA / /

HB

.lIoTHB-G "

Conjugate and Free Steroid Derivatization

THE-\

THFtHHA-G

~(;aIlO~~~:G ~J

\~

Fig. 7. Gas chromatographic analysis of steroids excreted in urine of a patient with l71I-hydroxylase deficiency. The upper chromatogram was obtained following enzymatic hydrolysis of urine . The lower chromatogram was obtained following direct derivatisation of steroid glucuronides.

216

C. H. L.

SHACKLETON

pStrioi En~yme Hydrolysis derivetlz.lion

OHA Conditions

PT

20m (1011 WCOT

"'ag.I80 ·330 3'rmln. Solid tniec:tion

Et An

16-oJ-oHA pStriol

ConjugateDerivalization

PTG

OHA

PDG

1&Ol+OHA

AnG \ \ EtG

I

I

\ 17-QH,PG Fig. 8. Urinary steroids from a patient with 3p-hydroxy steroid dehydrogenase (3P-OHSD) deficiency. The upper chromatogram was obtained following enzymatic hydrolysis of urine. The lower chromatogram was obtained following direct derivatisation of the steroid sulphates and glucuronides. The sulphate group of the steroid sulphate is subject to exchange with a silyl groupduring derivatisation so this group of compounds are analysed "free" in each case.

et al.

little sample preparation and no derivatisation is required. In fact, where large amounts of steroid are present (for example in urine during pregnancy and from patients with congenital adrenal hyperplasia), diagnosis of steroid disorders can be achieved from concentrated urine extracts without conjugate fractionation. The major disadvantage of SIMS in comparison to gas chromatography is the lack of specificity since isomers cannot be distinguished. Development of tandem mass spectrometric techniques (MSjMS, mass spectrometry-mass spectrometry) will probably allow separate analysis of isomers by SIMS through specific fragmentation of molecular ions [22]. SIMS techniques have not been ful1y evaluated for quantitative analysis and this information will clearly be necessary for the technique to be generally useful in clinical laboratories. However, Gaskel1 et al. have reported experience in the use of FAB-MS for quantifying plasma DHA-sulphate using a deuterated internal standard [23]. Gas chromatography of steroid glucuronides has become more attractive with improvements in column technology and derivatisation techniques. We have found that fused-silica or glass columns made in our laboratory are stable enough to last about two years even when repeatedly taken to high temperature. However, we have no data on the longevity of commercial1y coated columns so caution must be exercised when using these expensive items. Direct silylation of steroid sulphates with spontaneous replacement of the sulphate group is a useful ~5 PT

17

52 %

....................... f··············

15 jJg in

sample

10

was derivatised by addition of reagents and placing this reaction mixture in a 100 oven. Portions of the derivative were removed after 2 h, 8 h, 24 h, 48 hand 72 h. These portions were purified by Lipidex and were analyzed by gas chromatography. The results are presented in Fig. 9 which shows that the maximum recovery of free steroids occurs at around 16h. Little improvement in recovery occurs after this length of time. The recoveries still only represent 51-68% of the recoveries achieved by solvolysis. No explanation for this discrepancy has yet been found.

DHA

0

59%

.................................... .../ ~~~~:~:; .......•

5

//~../

51%

If/..r=::~.2:::::~~:: ~5AT

f...? ..:1::''''''''''

..

:.~::::::

o

2

8

24

68%

48

72

time (hI

DISCUSSION

Fig. 9. Time course of "sulphate-silyl group" exchange The results presented here have shown that both during direct derivatisation of steroid sulphate. The maximum recovery of free steroids is obtained at about 16 h. techniques described can give useful information on Thepercentage figures refer to the maximum recovery relasteroid excretion and circulation. Developments in tive to the amounts of free steroids determined after ethyl the SIMS techniques are particularly exciting since acetate solvolysis.

intact steroid conjugate anal ysis

217

nar y steroid glucuronides. Biochim. biophys. Acta 137 (1967) 216-219. . 9. Grob K., Grob G . and Grob K. Jr .: Deactivation of glass capillary columns by silylation : Part i-Principles and Basic Technique. J. high res. chromat. chromat. Commun. 2 (1979) 31-35. 10. Murray S. and Baillie T. A.: Direct derivatisation of sulphate esters for analysis by gas chromatography mass spectrometry. Biomed. mass Spectrom. 6 (1979) 82- 89. I I. Touchstone 1. C. and Dobbins M. F.: Direct determination of steroidal sulphates. J . steroid Biochem. 6 (1975) 1389-1392. 12. Yeh S. Y., Gorodetsky C. W. and Krebo H. A.: isolaAcknowledgements-These studies were made possible tion and identification of morphine 3- and 6-g1ucurothrough the support of Dr . A. L. Burlingame and other nides , morphine 3,6-diglucuronide, morphine 3-ethercolleagues at the Biomedical Mass Spectrometry Resource. eal sulfate, normorphine, and normorphine 6-glucuroMajor funding for this study was provided by the NIH nide as morphine metabolites in humans. J. Pharm. (Grant 00719 and 01614). Purchase of the negative ion unit Sci. 66, (1977) 1288-1293. for the Kratos MS-50S was made possible by a grant from 13. Mattox V. R. and Vrieze W. D.: Glucosiduronates of the Academic Senate, University of California, San Fran3a,21-dihydroxy-5t/-pregnane-ll,20-dione. Synthesis of cisco, 198I. C-3, C-2! and C-3,21 derivatives. J. org. Chem. 37 (1972) 3990--3996. 14. Shackleton C. H. L. and Whitney 1. L.: Use of Sep-pak cartridge for urinary steroid extraction : Evaluation of REFERENCES the method for use prior to gas chromatographic I. Shackleton C. H. L. and Honour 1. W.: Simultaneous analysis . Clinica chim. Acta 107 (1980) 231-243. estimation of urinary steroids by semi-automated gas 15. Janne 0 ., Vihko R., Sjovall 1. and Sjovall K.: Determichromatography: in vestigation of neo-natal infants nation of steroid mono- and di-sulphates in human and children with abnormal steroid synthesis. Clin. plasma . Clinica. chirn. Acta 23 (1969) 405-412. chim. Acta 69 (1976) 267-283 . 16. Setchell K. D. R. and Shackleton C. H. L. : The group 2. Shackleton C. H. L., Taylor N. F. and Honour 1. W. : separation of plasma and urinary steroids by column In An Atlas of Gas Chromatographic Profiles of Neutral chromatography on Sephadex LH-20. Clin. chim. Acta Urinary Steroids in Health and Disease. Packard47 (1973) 381-388. Becker, Delft, The Netherlands (1980). 17. Fales H. N., Jaouni T. 1. and Babashak 1. F.: Simple 3. Axelson M., Sahlberg B.-L. and Sjovall 1.: Analysis of device for preparing ethereal diazomethane without profiles of conjugated steroids in urine by ionresorting to codistillation. Anal. Chern. 45 (1973) exchange separation and gas chromatography-mass 2302-2303 . spectrometry. J. Chromatogr. 224 (1981) 355-370 . 18. Axelson M. and Sjovall 1.: Separation and computer4. Burstein S. and Lieberman S.: Hydrolysis of ketoised gas chromatography of unconjugated neutral steroid hydrogen sulfates by solvolysis procedures. steroids in plasma . J. steroid Biochem. 5 (1974) J. bioi. Chem. 233 (1958) 331-335 . 733-738. 5. Barber M., Bordoli R. S., Sedgwick R. C. and Tyler 19. Shackleton C. H. L.: Inborn errors of steroid biosynA. N. : Fast atom bombardment of solids as an ion thesis : Detection by a new mass spectrometric method. source in mass spectrometry. NalUre 293 (J 981) Clin. Chern. 29 (!983) 246-249 . 270-275. 20. Rosenfield R. L., Rich B. H., Wolfsdorf 1. 1., Casso ria 6. Aberth W., Straub K. M., and Burlingame A. L.: F., Parks 1. S., Bongiovanni A. M., Wu C. H. and Secondary ion mass spectrometry using a cesium ion Shackleton C. H. L.: Pubertal presentation of congeniprimary beam and a liquid target matrix for analysis of tal A5-3p-hydroxysteroid dehydrogenase ("3t/-o!") defibioorganic compounds. Analyt. Chem. 54 (1982) ciency. J. elin. Endoc. Metab. 51 (1980) 345-353. 2029-2034 . 22. McLafferty F. W.: Tandem mass spectrometry. Science 7. Shackleton C. H. L. and Straub K. M.: Direct analysis 214 (1981),280--297. of steroid conjugates : The use of secondary ion mass 23. Gaskell S. J., Brownsey B. G ., Brooks P. W. and Green spectrometry. Steroids 40 (1982) 35-5 I. B. N.: Fast atom bombardment mass spectrometry of 8. Jaakomaki P. 1., Yarger K. A. and Horning E. c.. steroid sulphates : qualitative and quantitative Gas-liquid chromatographic separation of human urianal yses. Biomed. mass Spectrom. 10 (1983) 215-219 .

technique. Even though the procedure apparently does not give quantitative recovery, recoveries are consistent for individual steroid sulphates providing the length of derivatisation remains constant. Results can readily be corrected by appropriate factors. This method offers considerable advantage over conventional solvolysis since there are fewer manipulations and is time saving. Replacement of a sulphate group by a silyl group is only effective for 3-sulphates of 3p-hydroxy-5-ene steroids or oestrogens.