Steroids excreted in urine by neonates with 21-hydroxylase deficiency: Characterization, using GC-MS and GC-MS/MS, of the D-ring and side chain structure of pregnanes and pregnenes

Steroids 75 (2010) 34–52

Contents lists available at ScienceDirect

Steroids

journal homepage: w ww.elsevier.com/locate/steroids

Steroids excreted in urine by neonates with 21-hydroxylase deﬁciency:

Characterization, using GC–MS and GC–MS/MS, of the D-ring and side

chain structure of pregnanes and pregnenes^ଝ

Soﬁa Christakoudi^a,∗, David A. Cowan^b, Norman F. Taylor^a

^aDepartment of Clinical Biochemistry, King’s College Hospital, Denmark Hill, London SE5 9RS, UK

^bDepartment of Forensic Science and Drug Monitoring, King’s College London, 150 Stamford Street, London SE1 9NH, UK

a r t i c l e i n f o

a b s t r a c t

Article history:

Received 13 April 2009

Received in revised form 12 August 2009

Accepted 22 September 2009

Available online 30 September 2009

Steroid metabolites in urine from neonates with 21-hydroxylase deﬁciency are predominantly poly-

hydroxylated 17-hydroxyprogesterone and androgen metabolites, and most have incompletely deﬁned

structure. This study forms part of a comprehensive project to characterize and identify these in order to

enhance diagnosis and to further elucidate neonatal types of steroid metabolism.

Steroids were analyzed, after extraction and enzymatic conjugate hydrolysis, as methyloxime-

trimethylsilyl ether derivatives on gas-chromatographs coupled to quadrupole and ion-trap

mass-spectrometers. GC–MS and GC–MS/MS spectra, obtained with constant excitation conditions, were

used together to determine the structure of the D-ring and the side chain of 20-oxo and 20-hydroxy

pregnane(ene)s without oxo groups on the A-, B-, and C-ring.

All possible combinations of D-ring and side chain conﬁguration were considered. Most fragmenta-

tions could be interpreted as partial or complete D-ring cleavages with loss of the side chain, aided by

comparison with spectra of deuterated derivatives and of borohydride reduced metabolites. Possible

rearrangement ions are also discussed. More than 140 endogenous metabolites were characterized.

GC–MS/MS was especially beneﬁcial for characterization of compounds with 16,17-dihydroxy-20-oxo

structure, interpreted as markers of intra-uterine enzyme induction. It also assisted the differentia-

tion of 16-hydroxy-20-oxo metabolites, present in urine of non-affected neonates, from the diagnostic

17-hydroxy-20-oxosteroids and enabled the detection of 15,17-dihydroxy-20-oxo compounds in low

concentrations. The presence of 17,21-dihydroxylated pregnane(ene)s despite the deﬁcit in CYP21A2 is

discussed.

Keywords:

CYP21A2

D-ring

GC–MS/MS

Neonate

Pregnane

Urine

We conclude that GC–MS combined with GC–MS/MS allows reliable identiﬁcation of the structure of

the D-ring and side chain of pregnane(ene)s without prior isolation, even when in low concentrations in

urine.

1. Introduction

ﬁrst days of life [2,4]. Neonates have a low capacity of hepatic

glucuronyl-transferases [5] and preferentially carry out additional

The most common cause of congenital adrenal hyperplasia

is 21-hydroxylase deﬁciency. It is characterized by increased

serum 17-hydroxyprogesterone and 21-deoxycortisol, which are

excreted in urine as numerous metabolites. Urinary steroid pro-

ﬁling is a valuable tool for diagnosis. Interpretation is relatively

straightforward in adults and children [1–3], where there are only

three major metabolites. These are relatively minor in neonates,

where diagnosis often presents a challenge, particularly in the

hydroxylation prior to excretion in urine, producing a very complex

proﬁle. In the ﬁrst months after birth, polyhydroxylated A-ring sat-

urated steroid metabolites are gradually replaced by glucuronides.

For 3␤-hydroxy-5-ene steroids, which are quantitatively major but

less important for diagnosis, sulfation predominates and additional

hydroxylations also occur in the ﬁrst weeks of life. A practical

complication is that samples may have been collected during

hydrocortisone treatment without this being disclosed, or during

the ﬁrst days of life, when steroids derived from the maternal com-

partment are still being cleared. A ﬁnding of cortisol metabolites

then requires a judgement as to whether either of these conditions

is responsible, or the defect is incomplete.

ଝ

This work originated at the Department of Clinical Biochemistry, King’s College

Hospital, Denmark Hill, London SE5 9RS, UK.

A comprehensive, systematic study of the metabolic proﬁle

in 21-hydroxylase deﬁciency in the neonate is lacking. Published

∗

Corresponding author. Tel.: +44 20 3299 4131; fax: +44 20 7737 7434.

E-mail address: hristakudi@doctor.bg (S. Christakoudi).

doi:10.1016/j.steroids.2009.09.011

S. Christakoudi et al. / Steroids 75 (2010) 34–52

35

studies either encompass only a limited number of cases [4,6,7]

2. Experimental

or are focused mainly on the quantiﬁcation of metabolites that are

major in adults and children [1]. Following development of the con-

cept of metabolomics, the term steroidomics has been suggested

for the detailed characterization and quantiﬁcation of metabolic

proﬁles of steroids under different conditions [8]. A detailed inves-

tigation of the urinary steroidome in 21-hydroxylase deﬁciency

would contribute to the establishment of early diagnostic mark-

ers that should reliably distinguish it from other steroid disorders.

Since the enzymes involved in hydroxylation of steroid metabo-

lites are non-speciﬁc, such a study could also contribute to the

general knowledge of induction of the phase one drug metaboliz-

ing enzymes in the neonate. Identiﬁcation of unknown metabolites

is also potentially useful for characterization of steroid secreting-

tumors, where unusual metabolites are frequent and have potential

as tumor markers.

This paper presents the ﬁrst part of a systematic characterization

of urinary steroid metabolites in 21-hydroxylase deﬁciency in the

neonate. It is focused on the identiﬁcation of the structure of the D-

ring and side chain in pregnane(ene)s and presents structural data

on over 140 steroids in this class that do not have additional oxo

groups on the A-, B-, and C-rings and were found to be present in

samples collected from a total of 70 neonates with 21-hydroxylase

deﬁciency. For some of the presented structures commercial stan-

dards were not available and endogenous compounds were used for

reference. Borohydrid e r eductio nwas used as necessary to generate

20-hydroxy from 20-oxo precursors.

2.1. Materials

All solvents were purchased from Rathburn, UK. Other reagents

were supplied as follows: Helix pomatia juice (SHP) – Biosepra,

Cergy, France and also kindly given by Dr. A.O.K. Chan; sodium

ascorbate – Sigma–Aldrich, UK; sodium borohydride – VWR Inter-

national, UK; methoxyamine hydrochloride (MO) – Fluka, via

Sigma–Aldrich, UK; trimethylsilylimidazole (TSIM) – Perbio Sci-

ence, UK.

Sep-Pak cartridges originated from Waters, UK, the capillary col-

umn for GC–MS from Varian, UK and for GC–MS/MS from Agilent,

J&W Scientiﬁc, UK.

Steroid standards were obtained from: Sigma–Aldrich, UK;

Steraloids, Rhode Island, USA; the MRC Steroid Reference Col-

lection (now administered by Sigma–Aldrich) or kindly provided

as personal gifts by Dr. R.A. Harkness, Prof. H. Makin, and

Dr. C.H.L. Shackleton. These were used to obtain reference

fragmentation data (Tables 1 and 2) and to identify urinary

steroids (Table 5, marked with an S). Ethanol solutions in

concentrations up to 1 mg/ml were prepared. All deuterated

materials (d₃MO, d₉TSIM, and 17␣-hydroxy-pregn-4-ene-3,20-

dione-2,2,4,6,6,21,21,21-d₈(d₈17Prog)) were purchased from CDN

Isotopes via QM_xLaboratories, UK.

2.2. Urine samples

The most comprehensively discussed additional position of

hydroxylation of pregnanes in 21-hydroxylase deﬁciency is that

at C_15␤[6,9]. In neonates there is also a signiﬁcant hepatic 16␣-

hydroxylase activity due to the expression of the fetal cytochrome

3A7. This contributes to the generation of 16␣-hydroxy DHA,

which is further transformed in utero to estrogens, but it is

probably involved in 16 ␣-hydroxylation of pregnanes as well.

It is also known that the adrenal 17␣-hydroxylase (CYP17)

can perform 16␣-hydroxylation that parallels 17␣-hydroxylating

activity when the enzyme is induced [10]. Intriguingly, there are

indications of the presence of an extra-adrenal 21-hydroxylase

activity [11–14]. Thus, regarding the D-ring and side chain

of 20-oxo pregnanes, all possible positions for hydroxylation,

i.e. at C₁₅, C₁₆, C₁₇, and C₂₁, are present in 21-hydroxylase

deﬁciency.

The identiﬁcation of steroid compounds is classically con-

sidered to be most reliable when based on a combination of

GC–MS, microchemistry, and NMR [15] but NMR still remains

fairly inaccessible and insensitive. Whilst steroid fragmentations

during electron impact ionization GC–MS have been reviewed

[15,16], all possible structures have not been covered, and the ori-

gin of some described fragments is not explained. In this paper

we summarize the literature on GC–MS ions derived from D-

ring and side chain fragmentation and discuss some useful ions

that have not been previously emphasized, aiming for a compre-

hensive description of all known and novel fragments. We also

take the identiﬁcation process a step further and present for the

ﬁrst time data on GC–MS/MS fragmentation patterns and out-

line the origin of the ions, assisted by the use of derivatization

with deuterated reagents. GC–MS/MS is becoming more accessi-

ble and widely used in the steroid ﬁeld but its application so far

has been for quantitative purposes, aiming for a higher sensitiv-

ity [17–19]. Whilst its usefulness for structure identiﬁcation has

been foreseen [20] it has not to our knowledge been exploited

for derivatized steroids. We demonstrate the usefulness of a com-

bined GC–MS and GC–MS/MS approach without recourse to NMR

for structure elucidation of unknown pregnanes and especially for

structures for which no commercial standards are currently avail-

able.

Urine samples from 70 newborns with 21-hydroxylase deﬁ-

ciency between birth and 3 months of age were sent by various

centres to our referral service as a routine part of their medical

investigation. They were obtained using polythene collecting bags,

so fecal contamination was avoided. Most were from untreated

neonates, but some had recently received hydrocortisone (cortisol)

treatment, either disclosed to us or evident because considerable

amounts of free cortisol and its metabolites were present.

2.3. Extraction, separation and hydrolysis of urinary steroids

Steroid metabolites in urine were analyzed by a modiﬁcation of

the well-established method of urinary steroid proﬁling [21], The

chromatographic separation of glucuronide and sulfate conjugates

prior to hydrolysis was omitted, thus producing only one deriva-

tive per sample. Steroid conjugates were ﬁrst extracted from urine

(2 ml per sample) using C18 silica-containing Sep-Pak cartridges

[22] and eluted with 4 ml methanol into tubes containing 2 ␮g of

the internal standard 5␣-androstane-3␣,17␣-diol. The methanol

was evaporated under nitrogen. Hydrolysis was carried out in 1 ml

0.5 M acetate buffer at pH 4.6 with 40 ␮l of H. pomatia juice (equiva-

lent to 4000 Fishman units of ␤-glucuronidase and 40,000 Roy units

of aryl-sulfatase) and 40 mg of sodium ascorbate for 72 h in a water

bath at 37 ^◦C. The freed steroids were then extracted on Sep-Pak

columns as above and eluted into stoppered glass tubes, containing

2 ␮g each of three further internal standards: 5␣-androstane-3␣-

ol, stigmasterol, and cholesterol butyrate.

2.4. Derivatization

Urine extracts as above or mixtures of standards were deriva-

tized after evaporation of solvent. This comprised oximation of the

oxo groups with 7 drops (125 ␮l) of 2% MO in pyridine for 90 min

at 60 ^◦C, followed by per-silylation of the hydroxyl groups with 7

drops (85 ␮l) of TSIM overnight at 100 ^◦C. The pyridine was then

evaporated under nitrogen and 1 ml of cyclohexane was added. The

remaining derivatization reagents were washed out by vortexing

twice with 0.5 ml deionized water.

36

S. Christakoudi et al. / Steroids 75 (2010) 34–52

Derivatized d₈17Prog and deuterated derivatives, prepared

2.9. Abbreviations

either with d₃MO or with d₉TSIM were used to help establish the

origin of the fragmentation ions.

For chemical derivatives, ‘MO’ represents a methyloxime and

‘TMS’ represents a trimethylsilyl ether. For the results and discus-

sion sections, GC–MS and GC–MS/MS are further abbreviated as

‘MS’ and ‘MS/MS’, respectively (with the exception of the subhead-

ings).

2.5. Borohydride reduction

Borohydride reduction was used to convert the 20-oxo into a

20-hydroxy group, leaving intact the C₅–C₆or C₄–C₅double bonds

(where present). This was performed according to a modiﬁcation of

a published method [23]. A portion of hydrolyzed urine extract or

standard solution was evaporated. The dry residue was dissolved

in 0.5 ml methanol, previously cooled to 4 ^◦C for 10 min. Sodium

borohydride was added (5 mg), the tube was stoppered after the

generation of hydrogen had ceased and incubated for 2 h at 4 ^◦C.

The reaction was stopped by adding 1.5 ml 10% (v/v) of acetic acid

and the mixture was diluted with 20 ml deionized water. Steroids

were extracted on Sep-Pak columns, eluted with 4 ml methanol,

and the dried residue was derivatized after the addition of the same

internal standards as for the urine extracts.

The names of all steroids we have identiﬁed in urine and

standards are abbreviated as hydroxylated variants of the fol-

lowing compounds: for 20-oxosteroids – ‘P’ for pregnanolone

(3␣-hydroxy-5␤-pregnan-20-one) and ‘ꢀP’ for pregnenolone (3␤-

hydroxy-pregn-5-en-20-one); for 20-hydroxysteroids – ‘P2’ for

pregnanediol (5␤-pregnane-3 ␣,20 ␣-diol), ‘P3’ for pregnanetriol

(5␤-pregnane-3␣,17␣,20␣-triol), ‘ꢀ P2’ for pregnenediol (pregn-

5-ene-3␤,20␣-diol) and ‘ꢀP3’ for pregnenetriol (pregn-5-ene-

3␤,17␣,20␣-triol). These same designations are used when the

orientation at C₃and at C₅is unknown: the assigned orientations

can be found in Table 5, which lists all compounds with their

systematic names. The position of the additional hydroxyls is indi-

cated in front of the respective abbreviations and X designates

an unknown position. The preﬁx ‘allo’ indicates a 5␣-epimer. Any

epimers at C₃or C₂₀other than the above are indicated in brackets

at the end, where there are sufﬁcient reasons to be fairly conﬁ-

dent about the orientation. The hydroxyl at C₁₇is always in the

␣-position and the ␣- is omitted.

2.6. GC–MS

GC–MS was performed on a Perkin Elmer Clarus 500 sin-

gle quadrupole mass spectrometer with a 20 m capillary column,

0.32 mm internal diameter, wall-coated with VF-1ms (100%

dimethylpolysiloxane), ﬁlm thickness 0.10 ␮m. The temperature

program was: 100 ^◦C (3 ␮l injection volume), hold for 1 min, then

20 ^◦C/min to 220 ^◦C, then 4 ^◦C/min to 270 ^◦C, hold for 4 min. Elec-

tron ionization mode was used at 70 eV. The mass range was

100–900 amu. The temperature of the ion source was 220 ^◦C.

Epimers of incompletely characterized structures are sufﬁxed

with a Roman numeral in retention time sequence (see Table 5).

3. Results

All urinary steroids referred to below are designated by a trivial

name followed by a serial number in round brackets, e.g. 17P (comp.

3). These are listed along with the full systematic name in Table 5.

2.7. GC–MS/MS

GC–MS/MS was performed on a Finnigan MAT GCQ ion-trap sys-

tem ﬁtted with a 25 m capillary column 0.2 mm internal diameter,

wall-coated with HP-1 (100% dimethylpolysiloxane), ﬁlm thickness

0.11 ␮m. The temperature program was as for the GC–MS above

and 1 ␮l was injected. Electron ionization mode was used. The

temperature of the ion source was 250 ^◦C. After initial adjustment

of the excitation voltage (see Section 3.1), the MS/MS parameters

were kept constant for all compounds as follows: isolation width

3 amu, isolation time 12 ms, collision time 15 ms, excitation voltage

1.2 V. The precursor ions chosen are termed P⁺and P₁⁺. For 20-oxo

compounds: P⁺= [M-31]⁺(loss of methoxyl radical from a methox-

imated 20-oxo group) and P₁⁺= [M-31-90]⁺(an additional loss of

a trimethylsilylated hydroxyl group). For 20-hydroxy compounds:

P⁺= [M-117]⁺in 17,20-dihydroxylated compounds (loss of the side

chain with a derivatized hydroxyl group at C₂₀) or P⁺= [M-205]⁺in

17,20,21-trihydroxylated compounds (loss of the side chain with

two derivatized hydroxyls at C₂₀and C₂₁).

3.1. GC–MS/MS excitation voltage optimization

The dependence of the percentage of the P⁺= [M-31]⁺ion cho-

sen for MS/MS of 20-oxo pregnane(ene)s on the excitation voltage

was investigated using examples of the pregnanetriol-20-ones, -

pentol-20-ones, and -hexol-20-ones observed in 21-hydroxylase

deﬁciency (Fig. 1). A constant excitation voltage approach was

found to provide additional structural information since com-

Throughout the text, all ion derivations are represented in

square brackets with a plus sign to indicate that positive ion mode

was used. It is not intended to indicate whether an odd or even

electron species is formed.

2.8. Retention times

These are based on chromatography on the GC–MS system and

are expressed as methylene units (MU), calculated relative to a

mixture of even and odd numbered hydrocarbons containing all

members in the series between C₂₆and C₃₆and based on a linear

interpolation between the retention times of neighboring hydro-

carbons. The retention times of the hydrocarbons were determined

on an extracted ion chromatogram of the ion at m/z 113, while

those of the metabolites in urine were determined on extracted

ion chromatograms of ions speciﬁc for each structure.

Fig. 1. Inﬂuence of the excitation voltage on the percentage of parent ion P⁺= [M-

31]⁺remaining in the GC–MS/MS spectrum of some major pregnane-polyol-20-

oxo metabolites found in 21-hydroxylase deﬁciency. Abbreviations are described in

Section 2.9.

Table 1

Characteristic ions (m/z) in the GC–MS and GC–MS/MS spectra of 20-oxo and 20-hydroxy pregnane(ene)s with different structures of the D-ring and side chain. For 20-oxo compounds, P⁺= [M-31]⁺and P₁⁺= [M-31-90]⁺. For

20-hydroxy compounds, P⁺= [M-117]⁺for 17,20-dihydroxylated compounds and P⁺= [M-205]⁺for 17,20,21-trihydroxylated compounds. Series are derived from further losses of 90 amu for each derivatized hydroxyl group.

All spectra of 20-oxo pregnanes contain in addition ions of the respective series [M-31-90]⁺, [P-90]⁺or [P₁-90]⁺with little further diagnostic value. Crystalline standards were used where available. Compounds present in

urine of neonates affected with 21-hydroxylase deﬁciency were used for structures lacking commercially synthesized standards. For them the respective number (comp. No) according to Table 5 (where their systematic names

are presented) is indicated. Compounds described in the literature are marked with ‘•’. Data for compounds of the 15-hydroxy and 15,21-dihydroxy type are based only on published spectra. For references see Section 3.3.

Compounds labeled with ‘*’ in the 20-hydroxy series were derived by borohydride reduction of the respective 20-oxo compounds (on the left of the table). Those not published have the following structures: pregn-5-ene-

3␤,15␰,16␣,20␰-tetrol for 15,16␣ꢀP2 (this has a different retention time to the endogenous comp. 49); pregn-5-ene-3␤,16␣,17␣,20␰-tetrol for 16␣ꢀP3 (standard); 5␰-pregnane-3␤,X,15␤,17␣,20␰,21-hexol for X,15␤,21P3 and

5␤-pregnane-1␤,3␣,6␣,16␣,17␣,20␰,21-heptol for 1␤,6␣,16␣,21P3 (does not correspond to the endogenous comp. 142).

D-ring/side chain

20-Oxo compounds

20-Hydroxy compounds

Standards

GC-MS ions

GC-MS-MS ions

+

Standards

P2; alloP2

–

GC-MS ions

Exclusively 117

–

GC-MS/MS ions

+

-

allo6␣P

70, 87, 100

[P-28] series, [P-107] series

–

-

15-OH

15␣/␤ꢀP •; allo 15␣P •; allo

15␣P(3␤) •

16␣ꢀP; 16␣P; allo16␣P;

1␤,16␣P comp. 82) •

87,100, 170, 201

–

+

16-OH

16-oxo

17-OH

156, 188, [M -72] series

156, prominent P , [P-43]

16␣ꢀP2

117 base peak, [M-90]

,

–

+

series, [P-57] series

[M-90-15]

+

–

16oxoꢀP2 (comp. 92) •

P3; 6␣P3; 11␤P3

117 base peak, 201, [M-31] ,

+

[M-116] , [M-131]

+

17P; allo17P; 6␣,17P; 11␤17P

156, 158, 170, 172, 188, 200, [M

[P-16] , [P-17] series, [P-28]

series

231, 245, [M-117] series

Low intensity [P-132] series

+

-143]

+

[M -72] series

-

+

158,175,188

103, [M -103]

21-OH

21P

21ꢀP2

[P-16] series, [P-17]

[M-103] series

+

series, [P-28] series

+

15,16-diOH

15,16␣ꢀP (comp. 77)

188,191,212,217,258,

[M-72-90] series, [M-31-90]

[P-16] , [P -41] , [P -57]

15,16␣ꢀP2 *

117 base peak, [M-190],

[M-190-15]

–

1

+

base peak

15-OH-16-oxo

15,17-diOH

–

16oxo15ꢀP2 (comp. 115)

117 base peak, 289,

[M-117] series

–

+

170, 258, 188, 200

231, [M-231]

15␤,17P (comp. 4) •;

15␤P3 (comp. 17) •;

dominant [P-90]

[P-107] , [P-114]

,

333, [M-247] , [M-117]

[P-130] (approx 20%)

+

15␤,17ꢀP (comp. 39) •

15␤P3(20␤) (comp. 14)

•; 15␤ꢀP3 (comp. 67) •

series

+

15,21-diOH

16,17-diOH

allo 11␤,15␣,21P •

16␣,17ꢀP

175, 188, 199, 258, 289

–

allo 11␤,15␣,21P2 * •

16␣ꢀP3 *

231, [M-103]

–

+

200, 217, 275, dominant [M-31]

[M -72] series

,

[P-116] series, [P-190] series,

+

217, 319, [M-247] , [M-117]

series

[P-16] , [P-116] , low [P-130]

+

[P -41] series

1

+

16,21-diOH

17,21-diOH

16␣,21ꢀP (comp. 114) •

276, 103, [M -103]

168, [P-16] , [P-131] , [P-107]

series

16␣,21ꢀP2 (comp. 129) •

[M-103] series

–

+

244, 246, 276

17,21P; 6␣,17,21P;

11␤,17,21P; allo

11␤,17,21P

6␣,21P3; 11␤,21P3;

11␤,21P3 (20␤)

[P-131] series

243, [M-205] series,

Low intensity [P-132]

series

+

103, [M -103]

[M-103] series

+

[M-160] series

–

+

258, 346

103, [M -103] , [M-160] series

15,17,21-triOH

16,17,21-triOH

346, [P-107]+

X,15␤,17,21P (comp. 94)

X,15␤,21P3 *

421, low [M-335]

[M-205] series

,

+

1␤,6␣,16␣,17,21P (comp. 126)

217, 288, 363, 103, [M -103]

,

256, [P-204] series, [P-190]

1␤,6␣,16␣,21P3 *

217, 305, 407, [M-335]

,

–

+

[M-160] series

series, [P -129] series

[M-205] series

1

Table 2

Intensities (% relative to the base peak) of the ions belonging to the GC–MS/MS spectra of reference 20-oxo pregnanes. For abbreviations see Section 2.9. P⁺= [M-31]⁺. The m/z of the particular ions is in bold. Endogenous compounds

(systematic names are given in Table 5) were used for structures lacking commercial standards (a – comp. 39; b – comp. 82; c – comp. 4).

+

Series

P

[P-90]

[P-16]

[P-17]

[P-28]

[P-43]

+

Ions (m/z)

[P-90] [P-2x90] [P-3x90] [P-4x90] [P-16] [P-16-90] [P-17] [P-17-90] [P-17-2x90] [P-17-3x90] [P-17-4x90] [P-28] [P-28-90] [P-28-2x90] [P-43] [P-43-90] [P-43-2x90] [P-43-3x90]

Compounds

m/z of ions 474

384

40

294

17

458

15

457

5

367

7

277

7

431

65

341

5

251

13

16␣ꢀP

100

m/z of ions 476

386

100

63

296

48

24

13

55

40

85

382

10

18

460

370

459

369

48

12

6

279

54

7

448

358

38

1

10

11

17

12

268

19

3

433

343

253

allo 6aP

21

16␣P

allo 16␣P

17P

100

50

7

5

7

3

37

456

8

12

2

4

11

30

60

46

14

7

10

12

59

15

22

28

22

546

3

100

88

472

100

35

8

17

20

100

365

1

12

5

23

allo 17P

21P

40

12

4

3

18

29

455

12

m/z of ions 562

292

4

5

a

15␤,17ꢀP

16␣,17ꢀP

1

8

1

5

m/z of ions 564

474

78

384

63

65

60

13

40

472

47

62

61

294

38

51

35

10

13

382

30

49

32

548

25

17

8

458

547

457

367

277

13

12

536

446

356

8

521

95

431

34

341

45

251

13

b

1␤,16␣P

6␣,17P

100

11

1

100

562

100

7

4

7

15

2

10

2

12

11␤,17P

10

c

15␤,17P

1

13

17,21P

7

3

4

365

3

534

m/z of ions 652

292

13

22

455

275

624

444

9

6␣,17,21P

11␤,17,21P

allo11␤,17,21P

1

5

7

13

7

14

6

10

6

11

10

12

+

Series

[P-57]

[P-114]

[P-116]

[P-131]

[P-190]

156

56

+

Ions (m/z)

[P-57-90]

[P-57-2x90]

[P-57-3x90]

[P-116]

[P-116-90]

[P-116-2x90]

[P-131]

[P-131-90]

[P-131-2x90]

[P-131-3x90]

[P-190]

[P-190-90]

Compounds

m/z of ions

16␣ꢀP

417

2

327

7

237

17

m/z of ions

allo 6aP

16␣P

allo 16␣P

17P

419

329

239

255

4

8

2

10

15

68

83

11

5

59

allo 17P

21P

m/z of ions

15␤,17ꢀP

448

8

446

356

266

372

282

a

16␣,17ꢀP

100

18

10

31

23

m/z of ions

1␤,16␣P

507

1

417

6

327

6

237

15

450

433

343

253

b

50

6␣,17P

14

11␤,17P

15␤,17P

c

7

4

17,21P

14

521

9

2

1

14

431

10

6

3

m/z of ions

6␣,17,21P

11␤,17,21P

allo11␤,17,21P

341

18

22

11

251

5

10

5

S. Christakoudi et al. / Steroids 75 (2010) 34–52

39

pounds with different structure evidently have different stability

3.3. Steroids representative of structures without available

commercial standards

of the P⁺ion. The same voltage was used for MS/MS of 20-hydroxy

structures.

Commercially synthesized standards were not available for

many of the presented structures. To overcome this, compounds

that were invariably present in urine of newborns affected with

21-hydroxylase deﬁciency were used for reference. Some of these

are described in the literature: 15␤,17␣-dihydroxy pregnane(ene)s

(both 20-oxo and 20-hydroxy) [6,9] in 21-hydroxylase deﬁciency;

1␤,16␣P (comp. 82 ) in the ﬁrst days of life [24]; 16␣,21ꢀP (comp.

114) in urine from neonates [24]; 16␣,21ꢀP2 (comp. 129) in urine

[24] and feces [25] from neonates; 16oxoꢀP2 (comp. 92) in feces

from neonates [25]. For others, identiﬁcation was based on features

in their spectra shared with compounds of known structure (see

Section 3.5 and Fig. 2). Borohydride reduction of either endogenous

compounds or standards was used to generate missing structures

in the 20-hydroxy series (MS spectra are presented in Fig. 3).

Two steroids with 20-hydroxy structures also containing an oxo

group on the D-ring were noted (comp. 92, Fig. 3C; comp. 116,

3.2. Characteristic GC–MS and GC–MS/MS ions

The ions in the MS and MS/MS spectra characteristic of preg-

nane(ene)s, with all observed combinations of hydroxyls on the

D-ring and side chain, are summarized in Table 1. Only refer-

ence compounds without oxo groups on the A-, B-, and C-ring are

included, because they illustrate best the D-ring and side chain

fragmentation pattern. Presence of oxo groups other than that at

C₂₀have signiﬁcant inﬂuence on the MS and especially on the

MS/MS spectra and such compounds will be discussed in a sepa-

rate communication. Table 2 presents the MS/MS spectra of 20-oxo

pregnanes (pure standards and published endogenous compounds)

on which Table 1 was based. It can be seen that orientation at C₅

does not inﬂuence signiﬁcantly the MS/MS fragmentation pattern.

Fig. 2. (A–D) GC–MS (left) and GC–MS/MS (right) spectra of the endogenous 20-oxo metabolites observed in urine of neonates affected with 21-hydroxylase deﬁciency and

used for reference. For structure of the D-ring and side chain see Table 1. P⁺= [M-31]⁺and P₁⁺= [M-31−90]⁺(the latter shown for 15,16␣ꢀP (comp. 77) only (A)).

40

S. Christakoudi et al. / Steroids 75 (2010) 34–52

Fig. 3. GC–MS spectra of the 20-hydroxy compounds used as reference structures. Compounds (C) and (D) are endogenous. The others were obtained as borohydride

reduction products as follows: (A) pregn-5-ene-3␰,15␰,16␣,20␰-tetrol from 15,16␣ꢀP (comp. 77); (B) pregn-5-ene-3␰,16␣,17␣,20␰-tetrol from 16␣,17ꢀP standard; (E)

5x-pregnane-3␰,X,15␤,17␣,20␰,21-hexol from X,15␤,17,21P (comp. 94) and (F) 5x-pregnane-1␤,3␣,6␣,16␣,17␣,20␰,21-heptol from 1␤,6␣,16␣,17,21P (comp. 126).

Fig. 3D). We considered in both that it was positioned at C₁₆rather

than C₁₅

These ions appear to be the result of an intramolecular rearrange-

ment in which the side chain is retained on the steroid nucleus but

the entire D-ring, including all the hydroxyls attached to it, as well

as an additional methyl group, are lost. The presence of the com-

plementary ions at m/z 143 and 231 respectively in the MS spectra

indicates this rearrangement is followed by a cleavage, with the

charge residing on either side of the cleavage line. We observed

this type of fragmentation only in 17-hydroxylated compounds,

irrespective of the presence or absence of a C₁₅-hydroxyl or the

.

Two combinations were not observed in urine from new-

borns with 21-hydroxylase deﬁciency and the data in Table 1 are

based entirely on published data. These are: 15-hydroxy, described

respectively in urine of newborns, rat feces, and pregnancy urine

[26–28] and 15,21-dihydroxy, described in rat feces [27].

Three of all the theoretically possible combinations of hydrox-

yls are not included (15,16,21-trihydroxy, 15,16,17-trihydroxy,

and 15,16,17,21-tetrahydroxy) because no representatives of these

groups were found and there are no published data for such struc-

tures.

3.4. Composition of ions and mechanism of formation based on

deuterated derivatives

Deuterated derivatives d₃MO-TMS or MO-d₉TMS were used to

help establish the composition of the characteristic ions of m/z

greater than 100 amu (Table 3). Most MS fragments could be ratio-

nalized as derived from partial (X) or complete (Y) cleavages of the

D-ring and/or losses of the side chain (Z) (Fig. 4 and Table 4). Mech-

anisms previously described in the literature are cited. Below are

described the formation of the rearrangement ions and of those

that may be derived by several mechanisms.

3.4.1. Formation of the [M-143]⁺and [M-231]⁺ions

In the [M-143]⁺ion (17-hydroxy-20-oxosteroids) and the

equivalent [M-231]⁺ion (15,17-dihydroxy-20-oxosteroids), all the

deuterium atoms are retained, including those at C₂₁in d₈17Prog

and at the methoxy group at C₂₀in d₃MO-TMS derivatives (Table 3).

Fig. 4. Partial (X) and complete (Y) D-ring cleavages and side chain loss (Z) involved

in the formation of ions in the GC–MS and GC–MS/MS spectra of 20-oxo and 20-

hydroxy steroids.

S. Christakoudi et al. / Steroids 75 (2010) 34–52

41

Table 3

Ion m/z shifts in deuterated derivatives and in 17␣-hydroxy-pregn-4-ene-3,20-dione-2,2,4,6,6,21,21,21-d₈(d₈17Prog) or in its borohydride reduction product (*). The

interpretation of the number of hydroxyl and methoxyl groups present or absent in the recorded ion and the presence of C₂₁is given.

m/z of ions in different derivatives

m/z in d₈17Prog

Structure of ions

Present

MO-TMS

d₃MO-TMS

shift

3

MO-d₉TMS

shift

Absent

100

103

100

–

OCH₃

–

103

–

112

9

–

O-TMS

–

117

156

158

170

172

188

200

201

212

217

231

244

245

246

258

275

276

288

289

305

319

333

346

363

–

156

158

170

175

191

203

204

215

217

–

244

–

246

258

278

279

291

292

–

346

366

126

165

167

179

178

197

209

210

221

235

249

262

263

264

276

293

294

306

307

332

346

360

373

390

9

6

9

120*

159

161

173

175

191

203

–

234

–

248

–

3

O-TMS + C₂₁

O-Si(CH₃)₂+ OCH₃+C₂₁

O-TMS + OCH₃+ C₂₁

O-TMS + OCH₃

2 (O-TMS)

2 (O-TMS) + C₂₁

2 (O-TMS)

2 (O-TMS) + C₂₁

2 (O-TMS)

2 (O-TMS) + OCH₃

3 (O-TMS)

3 (O-TMS) + OCH₃

4 (O-TMS)

–

OCH₃

–

OCH₃

OCH₃+ C₂₁

–

3

–

OCH₃

–

OCH₃

–

OCH₃

–

OCH₃

–

9

2x9

3x9

4x9

–

9

2x9

3x9

2x9

3

–

3

–

3

–

3

OCH₃

–

407

421

–

443

457

–

[M-72]⁺

[M-103]⁺

[M-116]⁺

[M-117]⁺

[M-131]⁺

[M-143]⁺

[M-160]⁺

[M-190]⁺

[M-205]⁺

[M-231]⁺

[M-247]⁺

[M-335]⁺

256

[M-75]⁺

3

–

M⁺-72

[M-112]⁺

[M-125]⁺

[M-126]⁺

[M-140]⁺

[M-152]⁺

[M-169]⁺

[M-208]⁺

[M-223]⁺

[M-249]⁺

[M-265]⁺

[M-362]⁺

274

[M-75]⁺

3

OCH₃+ C₂₁

O-TMS

O-TMS + C₂₁

O-TMS

O-TMS + OCH₃

2 (O-TMS)

3 (O-TMS)

–

CH₃from O-TMS

primary CH₃

CH₂=CH₂

–

primary CH₃

CH₃from O-TMS

O-TMS

2 (O-TMS)

–

[M-116]⁺

–

[M-120]⁺*

3

–

[M-131]⁺

–

3

–

[M-143]⁺

[M-163]⁺

–

[M-231]⁺

–

OCH₃

–

2 (O-TMS)

[P-16]⁺

[P-17]⁺

[P-28]⁺

[P-43]⁺

[P-57]⁺

[P-107]⁺

[P-114]⁺

[P-116]⁺

[P-131]⁺

[P-190]⁺

[P-204]⁺

[P₁-41]⁺

[P-19]⁺

[P-17]⁺

[P-28]⁺

[P-43]⁺

[P-57]⁺

[P-60]⁺

[P-114]⁺

[P-123]⁺

[P-125]⁺

[P-140]⁺

[P-208]⁺

[P-222]⁺

[P₁-41]⁺

[P-16]⁺

–

3

9

2x9

–

derivatization of the hydroxyl at C₁₇: in partially silylated deriva-

tives (with non-silylated C₁₇hydroxyl) the respective ions were

[M-71]⁺and [M-159]⁺(72 amu units less due to the lack of a TMS

group at C₁₇). It was not inﬂuenced by substituents on the A- and

B-rings but was hindered by a derivatized 11␤-hydroxyl group and

absent in the spectra of compounds additionally hydroxylated at

3.4.2. Formation of the [M-247]⁺ion

The [M-247]⁺ion is present in both 15,17,20-trihydroxy and

16,17,20-trihydroxy steroids. A mechanism for this fragmentation

has been suggested for the former [9], involving a partial D-ring

cleavage (X) and a loss of C₁₉. This interpretation would not explain

the formation of the [M-247]⁺ion in the latter but it is clear that the

side chain is lost in both cases since a low intensity ion correspond-

ing to [M-335]⁺([M-247-88]⁺) is found in the 21-hydroxylated

versions of the above structures (see Fig. 3E and F).

C₁₂, C₁₆or C₂₁.

The [M-143]⁺ion in 17-hydroxy-20-oxosteroids has been

recorded [16] but not previously explained. The [M-231]⁺ion has

been noted before for 15,17-dihydroxy-20-oxo compounds and

was suggested to arise from a partial D-ring cleavage with loss of

3.4.3. Formation of the ion at m/z 217

C

₁₆, C₁₇, the 17-hydroxyl, the side chain, and a further loss of 44 amu

A mechanism of formation of the ion at m/z 217 has been

suggested for 15,17-dihydroxy androgens [34] involving a com-

plete D-ring cleavage with generation of an ion of the type

of unidentiﬁed origin [9]. However, the data above do not support

such a mechanism.

42

S. Christakoudi et al. / Steroids 75 (2010) 34–52

Table 4

Suggested mechanism of formation of the ions in the GC–MS and GC–MS/MS spectra of 20-oxo and 20-hydroxy pregnanes(enes) based where possible on data from deuterated

derivatives (Table 3) and on the two major types of D-ring cleavage (X and Y) and the side chain loss (Z) observed in pregnane(ene)s (Fig. 4). P⁺= [M-31]⁺, P₁⁺= [M-31-90]⁺,

‘C₁₅’ to ‘C₂₁’ – conﬁguration at the speciﬁed carbon number; ‘Ref’ column gives the number of the reference for published fragmentations; ‘O’ – derivatized hydroxyl group;

‘oxo’ – derivatized oxo group; ‘ ’ – present or absent hydroxyl group; (C₁₅–C₁₆) – the part of the D-ring containing C₁₅and C₁₆with any derivatized hydroxyls or hydrogen

atoms attached to them.

Ions

C 15 C 16 C 17 C 20 C 21 Mechanism

Ref

Ions

C 15 C 16 C 17 C 20 C 21 Mechanism

Ref

GC-MS

70

87

100

103

117

156

158 (70+88)

170

–

O

–

oxo

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

X–OCH₃+ 2H

Z + C¹⁷H + 2H

X + H

[15] 288 (200+88)

[30] 289 (201+88)

[27] 289

–

O

–

O

–

oxo

O

–

O

–

oxo

O

–

O

–

Y–O-TMS

Y

–

[27]

–

[9]

–

[30] 305 (217+88)

[27] 319 (231+88)

[16] 333 (245+88)

[16] 346 (170+2x88)

[15] 363 (275+88)

[29] 407 (231+2x88)

[31] 421 (245+2x88)

[29] [M-72]⁺

O

–

Y–(C₁₅-C₁₆)-H

O

X + H

–

X–H

Y–H

Y–OCH₃

X

X–H

Y–H

C²¹H₂-O-TMS

Z

O

–

O

–

O

–

oxo

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

oxo

O

oxo

O

X–OCH₃

X–OCH₃+ 2H

Y–OCH₃

X–CH₃

O

–

oxo

O

M

^+•–Z

–

[31]

–

[M-72]⁺

O

–

[27] [M-72]⁺

O

–

170

172

O

–

O

–

O

–

[29] [M-72]⁺

–

[29] [M-103]⁺

O

–

^+•–C²¹H₂-O-TMS [16]

175 (87+88)

188 (100+88)

200

201

217

231

243

244 (156+88)

245

246 (70+2x88)

258 (170+88)

275

276 (100+2x88)

Z + C¹⁷H + 2H

X + H

Y–H

Y–O-TMS

Y

Y–Z

[15] [M-103]⁺

[27] [M-117]⁺

O

–

O

^+•–Z

[15]

–

[29] [M-143]⁺

–

O

–

O

–

oxo

O

oxo

O

–

Rearrangement

[31] [M-160]⁺([M-72-88]⁺)

[15] [M-160]⁺([M-72-88]⁺)

O

–

O

–

O

–

M

^+•–Z

–

[M-160]⁺([M-72-88]⁺)

[M-190]⁺

O

Rearrangement

–

[27] [M-205]⁺([M-117-88]⁺)

M

^+•–Z

[15]

[9]

–

[9]

–

O

–

[M-231]⁺([M-143-88]⁺)

[M-231]⁺

O

–

O

Rearrangement

M

?

^+•–X + H

[M-247]⁺

O

^+•–X–CH₃

–

[27] [M-247]⁺

oxo

O

–

O

–

[25]

–

O

–

O

–

O

–

O

–

Y–Z

Y–(C₁₅-C₁₆)-H

Y–Z

–

GC-MS/MS

168

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

oxo

O

–

O

–

O

Y–HO-TMS

P₁⁺–Z

–

256 (168+88)

Y–Z

[P₁-41]⁺

Y–CH₂–O-TMS [27] [P₁-41]⁺

P₁⁺–Z

X–H

–

[P–43]⁺

P⁺–Z–2H

Y–HO-TMS–H

X–OCH₃

Y–H

X–OCH₃+ 2H

Y–OCH₃

X

[33] [P-57]⁺

[15] [P₁-57]⁺

[33] [P-114]⁺

[15] [P-116]⁺

–

P⁺–Z–CH₃–H

P⁺–Z–TMS

P⁺–Z–TMS–2H

P₁⁺–Z

oxo

O

–

oxo

[9]

–

[P₁-129]⁺([P₁-41-88]⁺)

[P-131]⁺([P-43-88]⁺)

[P-190]⁺

P⁺–Z–2H

Rearrangement

P⁺–Z–TMS–2H

X + H

[15] [P-190]⁺

O

[32] [P-204]⁺([P-116-88]⁺)

(CH₃)₃SiO–CH CH–CH O⁺Si(CH₃)₃. A similar ion is formed in

17,18-dihydroxy androgens where the two hydroxyls are also sep-

arated by three carbon atoms [35] and in a 16,20,21-trihydroxy

pregnene (16␣,21ꢀP2 (comp. 114, Fig. 2B) [25]). Although no clear

mechanism has been suggested for C₂₁steroids, the above struc-

mechanism would require a further loss of a hydrogen atom and

could also be responsible for the formation of the ion at m/z 217 in

16,20,21-trihydroxysteroids.

3.4.4. Formation of the [P-114]⁺, [P-116]⁺, and [P-204]⁺ions

The MS/MS ions [P-114]⁺(in 15,17-dihydroxy-20-oxosteroids,

Table 2), [P-116]⁺(in 16,17-dihydroxy-20-oxosteroids, Table 2),

and [P-204]⁺(in 16,17,21-trihydroxy-20-oxosteroids, Fig. 2 D) must

involve a loss of the D-ring or the side chain, because no such frag-

mentation is observed in compounds lacking the double hydroxyls

on the D-ring. The ﬁrst two ions are derived from losses contain-

ing a single hydroxyl group. The last is found in 1␤,6␣,16␣,17,21P

(comp. 126, Fig. 2D), discussed in Section 3.5.4, which contains an

additional derivatized hydroxyl at C₂₁. The m/z of the ion [P-204]⁺

is shifted by 18 (2x9) amu in the MO-d₉TMS derivative (contain-

ing two derivatized hydroxyls) and by 88 amu relative to [P-116]⁺,

so the additional derivatized hydroxyl group at C₂₁is included.

This suggests that in the above fragments the side chain is lost.

One possible mechanism is a migration of the TMS group at C₁₇

to the nitrogen attached to C₂₀in an analogous manner to the

ture could be achieved if the carbon chain includes C₁₆–C₁₇–C₂₀

.

However, an ion at m /z 217 is also observed in 16,17-dihydroxy

and 16,17,21-trihydroxy-20-oxosteroids, where two hydroxyls

are attached to neighboring carbon atoms, and also in the 20-

hydroxylated variants of those compounds. We suggest there is a

complete D-ring cleavage (Y) followed by a loss of the side chain

(Z, Table 4). Interestingly, an ion at m/z 305 is present in addition

to that at m/z 217 in 16,17,20,21-tetrahydroxy compounds, which

contains three derivatized hydroxyls (Table 3) and most likely cor-

responds to the ion at m/z 217 in 16,17,20-trihydroxy steroids,

shifted by 88 amu for one additional derivatized hydroxyl. This sug-

gests that both ions contain the C₂₁atom and may be formed by a

complete D-ring cleavage (Y) followed by a cleavage between C₁₆

and C₁₇and a loss of the C₁₅and C₁₆carbon atoms with the hydro-

gens attached to them and the derivatized hydroxyl at C₁₆. This

S. Christakoudi et al. / Steroids 75 (2010) 34–52

43

reciprocal exchange of TMS groups between C₁₇and C₂₀observed

in partially (non-silylated at C₁₇) and fully silylated derivatives

of 17,20-dihydroxy pregnanes [33], followed by a cleavage of the

C₁₇–C₂₀bond. Alternatively, the side chain and the TMS at C₁₇could

be lost consecutively. This is accompanied by an additional loss of

two hydrogenatoms whenahydroxylis presentat C₁₆. The [P-204]⁺

series has a much lower intensity than the [P-116]⁺series, so the

21-hydroxy l m ay have asigniﬁcant inﬂuence onthisfragmentation.

structure is further supported by the presence of the ion at m/z

168 in MS/MS that is absent in MS/MS of 17,21-dihydroxy-20-

oxosteroids. It is 88 amu less than the ion at m/z 256 observed in

16,17,21-trihydroxy-20-oxosteroids (Fig. 2D), thus indicating the

absence of one of these hydroxyl groups.

3.5.4. 15,17,21- and 16,17,21-trihydroxy-20-oxo structures

These are represented respectively by X,15␤,17,21P (comp. 94,

Fig. 2C) and 1␤,6␣,16␣,17,21P (comp. 126, Fig. 2D). Both contain

MS features of C₂₁-hydroxylation, i.e. ions at m/z 103, [M-103]⁺, and

the [M-160]⁺series. Their borohydride reduction products (Fig. 3E

and F) share the [M-205]⁺series, derived from loss of the side

chain in 20,21-hydroxy pregnanes. In addition, 1␤,6␣,16␣,17,21P

contains ions for 16,17-dihydroxy-20-oxosteroids, some of which

are shifted by 88 amu, corresponding to one derivatized hydroxyl

group. Hence, in MS the ion at m/z 288 is equivalent to m/z 200,

363 is equivalent to 275 while that at m/z 217 is unchanged. On

MS/MS, the [P-204]⁺loss is equivalent to [P-116]⁺while [P-190]⁺

is unchanged. In MS, 16,17,20-trihydroxysteroids contain an ion

at m/z 319 (see the borohydride reduction product of 16␣,17ꢀP;

Fig. 3B), most likely derived from partial cleavage of the D-ring. This

is matched by the ion shifted by 88 amu at m/z 407 in the spec-

trum of the borohydride reduction product of 1␤,6␣,16␣,17,21P

(Fig. 3F). All the above strongly supports a 16,17,21-trihydroxy-20-

oxo structure for comp. 126. In contrast, X,15␤,17,21P contains the

ion at m/z 346 in both MS and MS/MS, corresponding to the ion at

m/z 258 of 15,17-dihydroxy-20-oxosteroids shifted by 88 amu.

3.4.5. Formation of the [P-190]⁺ion

The MS/MS ion [P-190]⁺in 16,17-dihydroxy-20-oxosteroids

(Table 2) can be rationalized as a loss of both hydroxyl groups

(at C₁₆and C₁₇) together with an additional carbon, most likely

the one at C₁₆, in a manner analogous to the formation of the

ion at m/z 191 in polyhydroxylated androgens [34], where vicinal

hydroxyls are present. The 21-hydroxyl in 16,17,21-trihydroxy-20-

oxosteroids (Fig. 2D) probably hinders this fragmentation, resulting

in a lower intensity of this series. A similar analogy concerning vic-

inal hydroxyls should explain the formation of the [M-190]⁺ion in

15,16,20-trihydroxy compounds (Fig. 3A).

3.5. Tentative identiﬁcation of endogenous metabolites used for

reference

3.5.1. 15,16-hydroxy-20-oxo structure

This is represented by 15,16␣ꢀP (comp. 77, Fig. 2A). It

contains the ion at m/z 258, similarly to the 15,17-dihydroxy-

20-oxosteroids. Both contain two hydroxyl groups, as shown by

increments of 18 amu as deuterated derivatives (Table 3), but have

different MS/MS spectra (data not shown), indicating a differ-

ence in their structure. The ion at m/z 188, observed in 16, 17

or 21-monohydroxy-20-oxosteroids is present in 15,16␣ꢀP but

there is no indication of a primary hydroxyl, thus excluding 21-

hydroxylation. If 15-hydroxylation determines the formation of the

ion at m/z 258 and the compound has a different structure than

15,17-dihydroxy-20-oxo, the only remaining position for the sec-

ond hydroxyl is at C₁₆. In MS the ions at m/z 191 and 217 are also

present, indicative of neighboring hydroxyls.

3.6. Steroids characterized in urine from neonates with

21-hydroxylase deﬁciency

Applying the structural criteria summarized in Table 1, we have

characterized the D-ring and side chain of as many 20-oxo and 20-

hydroxy pregnane(ene)s (with no oxo groups on the A-, B- or C-

ring) as we could observe in separate samples of urine from a group

of 70 neonates affected with 21-hydroxylase deﬁciency (Table 5).

This was based on systematic plotting of the [M-31]⁺ions for 20-

oxo pregnane(ene)s and the [M-117]⁺and [M-205]⁺ions for 20-

hydroxy pregnane(ene)s. Representative plots are shown for one

affected infant in Fig. 5.

3.5.2. 15,20-Dihydroxy-16-oxo structure

This is represented by 16oxo,15ꢀP2 (comp. 116 , Fig. 3D). The MS

spectrum resembles that of 16oxoDP2 (comp. 92 , Fig. 3C), identiﬁed

by comparison with a published spectrum [25]. The presence of

an additional derivatized hydroxyl on the D-ring or side chain is

indicated by the 88 amu shift of the ion at m/z 201, rationalized

as a complete D-ring cleavage (Table 3), becoming the ion at m/z

289. Absence of an ion at m/z 103 excludes a C₂₁position, and the

high proportion of the ion at m/z 117 relative to the other ions of

the spectrum excludes a C₁₇position. Thus the additional hydroxyl

Detailed information on the concentration, diagnostic impor-

tance, source and evolution with age (both postconceptional and

postnatal) of the steroids we have deﬁned, and the effect of hydro-

cortisone treatment on these, are beyond the scope of the current

paper and will be the subject of a separate communication.

3.6.1. Assigning orientations of epimers and positions of

hydroxyls

All pregn-5-enes are considered to be 3␤-hydroxy as the most

likely position except for those eluting very early, which are con-

sidered to be 3␣ (3␣-hydroxy-androst-5-en-17-one standard has

an earlier retention time than 3␤-hydroxy-androst-5-en-17-one

standard). The 3␣-5-enes include 16␣ꢀP (3␣) (comp. 1; MU 26.76,

compared to MU 29.65 of 16␣ꢀP (comp. 51)), 6␣ꢀP2 (3␣) (comp.

2; MU 27.17), ꢀP3 (3␣) (comp. 5; MU 27.46), 16␣,17,21ꢀP (3␣)

(comp. 80; MU 30.32, compared to MU 32.41 of 16␣,17,21ꢀP

(comp. 135)) and 16␣,17ꢀP (3␣) (comp.81; MU 30.33, compared to

MU 31.25 of 16␣,17ꢀP (comp. 117)). The major pregnane metabo-

lites are considered to be 3␣-hydroxy, because they are mostly

found as glucuronides and 5␤-reduced, because for known com-

pounds the 5␣-epimers are found in lower concentrations (e.g.

17P (comp. 3) and allo 17P (comp. 6), 15␤,17P (comp. 4) and allo

15␤,17P (comp. 7)). Exceptions are the two 15,17,21-hydroxy-20-

oxo pregnanes (15␤,17,21P (comp. 32) and X,15␤,17,21P (comp.

94)). These are found only as sulfates, which we thus believe to be

3␤-epimers. All 15,17-dihydroxy compounds are considered to be

must be at C₁₅

.

3.5.3. 16,21-Dihydroxy-20-oxo structure

This is represented by 16 ␣,21 ꢀ P (comp. 114, Fig. 2B). Only a

spectrum of the TMS derivative has been published [24]. The par-

tial D-ring cleavage ion at m/z 276 in MS indicates the presence

of two derivatized hydroxyls that could be at C₁₆, C₁₇or C₂₁. A

16 and 17 combination is excluded, because in 16,17-dihydroxy-

20-oxosteroids (Table 1), the ion at m/z 275 corresponds to this

fragmentation. The presence of ions at m/z 103 and [M-103]⁺indi-

cates that one of the hydroxyls is at C₂₁. The ion at m/z 276

is more prominent than that in 17,21-dihydroxy-20-oxosteroids,

while no ion at m/z 244 is observed. Similar features were observed

in MS of tentatively identiﬁed 16␣-hydroxydeoxycorticosterone

and 16␣,18-dihydroxydeoxycorticosterone, produced from 16␣-

hydroxyprogesterone in newborn rat adrenal cell culture [32], thus

placing the second hydroxyl at C₁₆. A 16,21-dihydroxy-20-oxo

44

S. Christakoudi et al. / Steroids 75 (2010) 34–52

Table 5

Structure of the D-ring and side chain of 20-oxo and 20-hydroxy pregnane(ene)s with no oxo groups on the A-, B- or C-ring, observed in urine of neonates affected with

ꢀ

21-hydroxylase deﬁciency. Column titles: ‘No’ – compound number in order of elution; ‘MU’ – methylene units; ‘C₁₅to ‘C₂₁– conﬁguration at the speciﬁed carbon number;

‘MW’ – molecular weight of the respective MO-TMS derivative; ‘Ref’ – number of the reference, where the relevant compound has been noted or discussed in relation to

urine from neonates; ‘S’ – compound identiﬁcation was based on comparison to pure standard; ‘O’ – hydroxyl group present, ‘oxo’ – oxo group present.

No

MU

Abbreviation

Systematic name

C 15

C 16

C 17

C 20

C 21

MW

Ref

1

2

3

4

5

6

7

8

26.76

27.17

27.27

27.39

27.46

27.54

27.73

27.76

27.86

27.97

28.12

28.21

28.22

28.23

28.31

28.36

28.38

28.42

28.49

28.58

28.61

28.64

28.66

28.70

28.71

28.77

28.78

28.79

28.80

28.84

28.87

28.89

28.93

29.06

29.10

29.11

29.16

29.17

29.22

29.25

29.32

29.34

29.41

29.48

29.54

29.58

29.65

29.66

29.70

29.72

29.73

29.74

29.80

29.81

29.83

29.88

29.92

29.97

30.02

30.04

30.07

30.10

30.11

30.15

16␣ꢀP (3␣)

6␣ꢀP2 (3␣)

17P

15␤,17P

ꢀP3 (3␣)

allo 17P

allo 15␤,17P

X,X,ꢀP2 (3␣)

P3 (20␤)

P2

3␣,16␣-dihydroxy-pregn-5-en-20-one

pregn-5-ene-3␣,6␣,20␰-triol

3␣,17␣-dihydroxy-5␤-pregnan-20-one

3␣,15␤,17␣-trihydroxy-5␤-pregnan-20-one

pregn-5-ene-3␣,17␣,20␰-triol

3␣,17␣-dihydroxy-5␣-pregnan-20-one

3␣,15␤,17␣-trihydroxy-5␣-pregnan-20-one

pregn-5-ene-3␣,X,X,20␰-tetrol

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

oxo

O

oxo

O

oxo

O

oxo

O

oxo

O

oxo

O

oxo

O

oxo

O

oxo

O

oxo

O

–

O

–

O

–

O

–

O

–

505

550

507

595

550

507

595

638

552

464

595

552

595

640

552

683

640

595

640

683

462

505

683

593

507

595

640

683

681

507

683

552

595

640

552

683

593

683

640

595

640

681

728

638

683

505

683

681

769

507

595

507

683

640

683

726

728

683

640

683

638

552

640

681

640

683

816

–

[7] S

[9]

–

S

[6]

–

S

[4] S

–

[6]

S

–

[6]

[7]

[6]

–

S

–

O

–

9

5␤-pregnane-3␣,17␣,20␤-triol

5␤-pregnane-3␣,20␣-diol

O

–

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

6␣,17P

P3

3␣,6␣,17␣-trihydroxy-5␤-pregnan-20-one

5␤-pregnane-3␣,17␣,20␣-triol

3␣,6␣,17␣-trihydroxy-5␣-pregnan-20-one

5␤-pregnane-3␣,15␤,17␣,20␤-tetrol

5␣-pregnane-3␣,17␣,20␣-triol

O

–

allo 6␣,17P

15␤P3 (20␤)

allo P3

X,15␤,17P I

15␤P3

3␰,X,15␤,17␣-tetrahydroxy-5␰-pregnan-20-one

5␤-pregnane-3␣,15␤,17␣,20␣-tetrol

3␣,1␤,17␣-trihydroxy-5␤-pregnan-20-one

5␣-pregnane-3␣,15␤,17␣,20b-tetrol

3␰,X,15␤,17␣-tetrahydroxy-5␰-pregnan-20-one

pregn-5-ene-3␤,20␣-diol

3␰,16␰-dihydroxy-pregn-5-en-20-one

3␰,X,15␤,17␣-tetrahydroxy-5␰-pregnan-20-one

3␤,X,17␣-trihydroxy-pregn-5-en-20-one

3␣,6␣-dihydroxy-5␤-pregnan-20-one

3␣,16␣-dihydroxy-5␤-pregnan-20-one

3␣,17␣,21-trihydroxy-5␤-pregnan-20-one

5␰-pregnane-3␰,X,17␣,20␰-tetrol

3␰,X,15␤,17␣-tetrahydroxy-5␰-pregnan-20-one

3␤,X,15␤,17␣-tetrahydroxy-pregn-5-en-20-one

3␣,16␣-dihydroxy-5␣-pregnan-20-one

3␤,15␤,17␣,21-tetrahydroxy-5␰-pregnan-20-one

5␰-pregnane-3␰,16␰,20␰-triol

3␰,16␰,17␣-trihydroxy-5␰-pregnan-20-one

3␰,X,17␣-trihydroxy-5␰-pregnan-20-one

5␤-pregnane-3␣,6␣,17␣,20␣-tetrol

1␤,17P

allo15␤P3 (20␤)

X,15␤,17P II

ꢀP2

16ꢀP

–

X,15␤,17P III

X,17ꢀP I

6␣P

16␣P

17,21P

O

–

[24]

[24] S

S

–

S

–

S

–

[9]

–

[6]

–

[24] S

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

XP3 I

X,15␤,17P IV

X,15␤,17ꢀP I

allo 16␣P

15␤,17,21P

16P2 I

16,17P I

X,17P

6␣P3

6␣P2

X,15␤,17P V

15␤,17ꢀP

X,X,17P I

X,16,17P I

XP3 II

allo15␤17P (3␤)

16,17P II

X,16P I

XP3 III

X,X,17ꢀP I

X,XP3 I

15,16ꢀP2

X,16,17P II

16␣ꢀP

X,17,21P I

X,15␤,17ꢀP II

X,X,X,17ꢀP

XP

oxo

O

5␤-pregnane-3␣,6␣,20␣-triol

O

3␰,X,15␤,17␣-tetrahydroxy-5␰-pregnan-20-one

3␤,15␤,17␣-trihydroxy-pregn-5-en-20-one

3␰,X,X,17␣-tetrahydroxy-5␰-pregnan-20-one

3␰,X,16␰,17␣-tetrahydroxy-5␰-pregnan-20-one

5␰-pregnane-3␰,X,17␣,20␰-tetrol

3␤,15␤,17␣-trihydroxy-5␣-pregnan-20-one

3␰,16␰,17␣-trihydroxy-5␰-pregnan-20-one

3␰,X,16␰-trihydroxy-5␰-pregnan-20-one

5␰-pregnane-3␰,X,17␣,20␰-tetrol

oxo

O

oxo

O

3␤,X,X,17␣-tetrahydroxy-pregn-5-en-20-one

5␰-pregnane-3␰,X,X,17␣,20␰-pentol

pregn-5-ene-3␤,15␰,16␰,20␰-tetrol

oxo

O

3␰,X,16␰,17␣-tetrahydroxy-5␰-pregnan-20-one

3␤,16␣-dihydroxy-pregn-5-en-20-one

3␰,X,17␣,21-tetrahydroxy-5␰-pregnan-20-one

3␤,X,15␤,17␣-tetrahydroxy-pregn-5-en-20-one

3␤,X,X,X,17␣-pentahydroxy-pregn-5-en-20-one

3␰,X-dihydroxy-5␰-pregnan-20-one

3␰,X,16␰-trihydroxy-5␰-pregnan-20-one

3␰,16␰-dihydroxy-5␰-pregnan-20-one

3␰,X,X,17␣-tetrahydroxy-5␰-pregnan-20-one

5␰-pregnane-3␰,X,17␣,20␰-tetrol

3␰,X,X,17␣-tetrahydroxy-5␰-pregnan-20-one

3␰,X,16␰,17␣-tetrahydroxy-5␰-pregnan-20-one

pregn-5-ene-3␤,X,X,17␣,20␰-pentol

5␰-pregnane-3␰,X,X,17␣,20␰-pentol

3␰,X,16␰,17␣-tetrahydroxy-5␰-pregnan-20-one

5␣-pregnane-3␤,15␤,17␣,20␤-tetrol

3␰,X,X,17␣-tetrahydroxy-5␰-pregnan-20-one

pregn-5-ene-3␤,15␤,17␣,20␣-tetrol

5␰-pregnane-3␰,16␰,20␰-triol

oxo

O

oxo

O

oxo

O

X,16P II

16P

–

X,X,17P II

XP3 IV

X,X,17P III

X,16,17P III

X,XꢀP3

O

–

O

–

O

–

[24]

–

X,XP3 II

X,16,17P IV

allo15␤P3(3␤,20␤)

X,X,17P IV

15␤ꢀP3

16P2 II

allo15␤P3(3␤)

X,X,16ꢀP

16P3 I

oxo

O

oxo

O

5␣-pregnane-3␤,15␤,17␣,20␣-tetrol

3␤,X,X,16␰-tetrahydroxy-pregn-5-en-20-one

5␰-pregnane-3␰,16␰,17␣,20␰-tetrol

3␰,X,17␣,21-tetrahydroxy-5␰-pregnan-20-one

5␰-pregnane-3␰,X,X,15␤,17␣,20␰-hexol

X,17,21P II

X,X,15␤P3

oxo

O

–

S. Christakoudi et al. / Steroids 75 (2010) 34–52

45

Table 5 (Continued )

No

MU

Abbreviation

Systematic name

C 15

C 16

C 17

C 20

C 21

MW

Ref

–

[4] S

–

[36]

–

[24]

–

[25]

–

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

30.18

30.23

30.24

30.28

30.31

30.32

30.33

30.35

30.40

30.41

30.44

30.47

30.48

30.53

30.58

30.59

30.60

30.61

30.65

30.66

30.69

30.72

30.79

30.81

30.83

30.92

30.97

31.05

31.20

31.21

31.23

31.25

31.30

31.32

31.34

31.35

31.42

31.43

31.51

31.64

31.69

31.79

31.81

31.82

31.83

31.84

32.12

32.41

32.46

32.56

32.85

32.93

32.97

33.12

33.17

33.56

34.48

X,16P III

21ꢀP

X,16P3 I

15,16␣ꢀP

X,X,16P I

3␰,X,16␰-trihydroxy-5␰-pregnan-20-one

3␤,21-dihydroxy-pregn-5-en-20-one

5␰-pregnane-3␰,X,16␰,17␣,20␰-pentol

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

oxo

O

oxo

O

oxo

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

–

O

595

505

728

593

683

640

681

593

595

683

640

728

771

683

595

640

593

505

595

771

728

593

683

640

681

593

769

771

769

683

640

771

550

771

593

771

593

816

593

638

816

769

681

769

728

683

593

859

857

681

638

681

640

681

816

681

683

638

683

681

859

904

726

771

726

771

O

3␤,15␰,16␣-trihydroxy-pregn-5-en-20-one

3␰,X,X,16␰-tetrahydroxy-5␰-pregnan-20-one

5␤-pregnane-3␣,16␣,17␣,20␣-tetrol

3␣,16␣,17␣,21-tetrahydroxy-pregn-5-en-20-one

3␣,16␣,17␣-trihydroxy-pregn-5-en-20-one

1␤,3␣,16␣-trihydroxy-5␰-pregnane-20-one

3␰,X,16␰,17␣-tetrahydroxy-5␰-pregnan-20-one

3␰,16␰,17␣,21-tetrahydroxy-5␰-pregnan-20-one

5␰-pregnane-3␰,X,17␣,20␰-tetrol

–

16␣P3

O

–

O

–

16␣,17,21ꢀP (3␣)

16,17ꢀP (3␣)

1␤,16␣P

X,16,17P V

16,17,21P I

XP3 V

X,16P3 II

5␰-pregnane-3␰,X,16␰,17␣,20␰-pentol

O

oxo

O

–

O

–

O

X,X,16,17P I

X,16,17P VI

2␣,16␣P

1␤,16␣P2

X,16ꢀP I

16oxoꢀP2

X,16P IV

X,15␤,17,21P

X,XP3 III

17,21ꢀP

X,15␤,17P VI

X,16P2

X,15␤,17ꢀP III

X,17ꢀP II

X,X,16,17ꢀP I

1␤,6␣,17,21P

X,X,17,21ꢀP I

X,X,16P II

X,16,17P VII

XP3 VI

6␣,11␤,17,21P

21ꢀP2

1␤,6␣,16␣,17P

X,16,17,21P I

X,16ꢀP II

X,X,16,17P II

X,X,17,21P

16␣21ꢀP

X,X,16P3 I

16oxo15ꢀP2

16␣17ꢀP

3␰,X,X,16␰,17␣-pentahydroxy-5␰-pregnane-20-one

3␰,X,16␰,17␣-tetrahydroxy-5␰-pregnan-20-one

2␣,3␣,16␣-trihydroxy-5␤-pregnane-20-one

5␤-pregnane-1␤,3␣,16␣,20␣-tetrol

3␤,X,16␰-trihydroxy-pregn-5-en-20-one

3␤,20␣-dihydroxy-pregn-5-en-16-one

3␰,X,16␰-trihydroxy-5␰-pregnan-20-one

3␤,X,15␤,17␣,21-pentahydroxy-5␰-pregnane-20-one

5␰-pregnane-3␰,X,X,17␣,20x-pentol

3␤,17␣,21-trihydroxy-pregn-5-en-20-one

3␰,X,15␤,17␣-tetrahydroxy-5␰-pregnan-20-one

5␰-pregnane-3␰,X,16␰,20␰-tetrol

3␤,X,15␤,17␣-tetrahydroxy-pregn-5-en-20-one

3␤,X,17␣-trihydroxy-pregn-5-en-20-one

3␤,X,X,16␰,17␣-pentahydroxy-pregn-5-en-20-one

1␤,3␣,6␣,17␣,21-pentahydroxy-5␤-pregnane-20-one

3␤,X,X,17␣,21-pentahydroxy-pregn-5-en-20-one

3␰,X,X,16␣-tetrahydroxy-5␰-pregnan-20-one

3␰,X,16␰,17␣-tetrahydroxy-5␰-pregnan-20-one

5␰-pregnane-3␰,X,17␣,20␰-tetrol

oxo

O

oxo

O

oxo

O

oxo

O

oxo

O

–

O

–

O

–

O

–

O

–

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

–

O

–

O

–

O

–

O

oxo

O

–

3␣,6␣,11␤,17␣,21-pentahydroxy-5␤-pregnane-20-one

pregn-5-ene-3␤,20␣,21-triol

oxo

O

[6]

[4] S

–

[24]

–

S

–

[24]

–

1␤,3␣,6␣,16␣,17␣-pentahydroxy-5␤-pregnane-20-one

3␰,X,16␰,17␣,21-pentahydroxy-5␰-pregnane-20-one

3␤,X,16␰-trihydroxy-pregn-5-en-20-one

3␰,X,X,16␰,17␣-pentahydroxy-5␰-pregnane-20-one

3␰,X,X,17␣,21-pentahydroxy-5␰-pregnane-20-one

3␤,16␣,21-trihydroxy-pregn-5-en-20-one

5␰-pregnane-3␰,X,X,16␰,17␣,20␰-hexol

3␤,15␰,20␣-trihydroxy-pregn-5-en-16-one

3␤,16␣,17␣-trihydroxy-pregn-5-en-20-one

pregn-5-ene-3␤,17␣,20␣,21-tetrol

oxo

O

oxo

O

–

O

–

O

–

O

–

O

–

O

21ꢀP3

X,X,21P3

5␰-pregnane-3␰,X,X,17␣,20␰,21-hexol

–

O

X,X,17,21ꢀP II

X,X,17ꢀP II

X,X,16,17ꢀP II

X,16P3 III

X,17,21P III

X,16ꢀP III

1␤,6␣,16␣,17,21P

X,X,X,16,17ꢀP

X,16,17ꢀP I

16␣,21ꢀP2

16ꢀP3

3␤,X,X,17␣,21-pentahydroxy-pregn-5-en-20-one

3␤,X,X,17␣-tetrahydroxy-pregn-5-en-20-one

3␤,X,X,16␰,17␣-pentahydroxy-pregn-5-en-20-one

5␰-pregnane-3␰,X,16␰,17␣,20␰-pentol

3␰,X,17␣,21-tetrahydroxy-5␰-pregnan-20-one

3␤,X,16␰-trihydroxy-pregn-5-en-20-one

1␤,3␣,6␣,16␣,17␣,21-hexahydroxy-5␤-pregnan-20-one

3␤,X,X,X,16␰,17␣-hexahydroxy-pregn-5-en-20-one

3␤,X,16␰,17␣-tetrahydroxy-pregn-5-en-20-one

pregn-5-ene-3␤,16␣,20␣,21-tetrol

pregn-5-ene-3␤,16␰,17␣,20␰-tetrol

3␤,X,17␣,21-tetrahydroxy-pregn-5-en-20-one

5␰-pregnane-3␰,16␰,17␣,20␰-tetrol

3␤,X,16␰,17␣-tetrahydroxy-pregn-5-en-20-one

5␰-pregnane-3␰,X,X,16␰,17␣,20␰-hexol

3␤,16␣,17␣,21-tetrahydroxy-pregn-5-en-20-one

3␰,X,16␰,17␣-tetrahydroxy-5␰-pregnan-20-one

3␰,16␰,17␣,21-tetrahydroxy-5␰-pregnan-20-one

pregn-5-ene-3␤,16␰,20␰,21-tetrol

3␰,X,17␣,21-tetrahydroxy-5␰-pregnan-20-one

3␤,X,16␰,17a-tetrahydroxy-pregn-5-en-20-one

3␤,X,X,X,16␰,17␣-hexahydroxy-pregn-5-en-20-one

5␰-pregnane-3␰,X,X,16␰,17␣,20␰,21-heptol

pregn-5-ene-3␤,16␰,17␣,20␰,21-pentol

3␰,X,X,16␰,17␣-pentahydroxy-5␰-pregnane-20-one

pregn-5-ene-3␤,16␰,17␣,20␰,21-pentol

3␰,X,16␰,17␣,21-pentahydroxy-5␰-pregnane-20-one

oxo

O

oxo

O

oxo

O

oxo

O

oxo

O

–

O

–

O

–

O

X,17,21ꢀP

16P3 II

X,16,17ꢀP II

X,X,16P3 II

16␣,17,21ꢀP

X,16,17P VIII

16,17,21P II

16,21ꢀP2

X,17,21P IV

X,16,17ꢀP III

X,X,X,16,17P

X,X,16,21P3

16,21ꢀP3 I

X,X,16,17P III

16,21ꢀP3 II

X,16,17,21P II

oxo

O

oxo

O

–

oxo

–

46

S. Christakoudi et al. / Steroids 75 (2010) 34–52

Fig. 5. GC–MS ion plots of steroids in urine from a 5-day-old baby with 21-hydroxylase deﬁciency. Total ion current chromatograms (top row) are shown on each side

for convenience, followed by selected ion plots for pregnanes (left) and pregnenes (right). For 20-oxo compounds (top ﬁve rows), the [M-31]⁺ions are plotted in the

sequence -diolone, -triolone, -tetrolone, pentolone and -hexolone. For 20-hydroxy compounds (bottom three rows), the [M-117]⁺ions for 17,20-dihydroxylated pregnanes

and [M-205]⁺ions for 17,20,21-trihydroxylated pregnanes (both side chain losses) are plotted on the left in the sequence -triol, -tetrol and -pentol/hexol. On the right,

the ion at m/z 117 is plotted for various 20-hydroxy compounds, followed by the ion at m/z 445 ([M-90-103]⁺for 16,20,21-trihydroxy pregnenes) and the ion at m/z 521

([M-117]⁺for a 17,20-dihydroxylated pregnene and [M-205]⁺for 17,20,21-trihydroxylated pregnenes). Peaks are labeled with compound numbers according to Table 5,

where systematic names are given. Alphabetically labeled peaks, included as useful markers (top row) are: (a) 3␣,17␣,20␣-trihydroxy-5␤-pregnan-11-one (MU 29.35);

(b) 3␤,16␣-dihydroxy-androst-5-en-17-one (MU 27.63/27.72); (c) 3␤,17␤-dihydroxy-androst-5-en-16-one (MU 28.51); (d) androst-5-ene-3␣,16␣,17␤-triol (MU 28.65).

Unlabeled peaks correspond to compounds with structures outside the scope of this paper, or to isotope ions, or to ions derived from fragmentations other than those

described above.

15␤-hydroxy epimers, in common with all published compounds

of this type [6,9]. Those compounds that are 15-hydroxylated but

lack a 17-hydroxyl (15,16ꢀP2 (comp. 49), 15,16␣ꢀP (comp. 77),

and 16oxo15ꢀP2 (comp. 116)) are considered to be of unknown

orientation at C₁₅. The major 16-hydroxylated compounds are

considered to be 16␣-epimers, although because of the multiple

epi/isomers observed, 16␤-epimers should also be present. The

major 20-hydroxylated compounds are considered to be of 20␣-

orientation, in analogy to pregnanetriol and 11-oxo-pregnanetriol,

which are the major urine metabolites in children with 21-

hydroxylase deﬁciency.

niﬁcant presence of endogenous 11␤-hydroxylated compounds in

21-hydroxylase deﬁciency (see Sections 3.6.3 and 3.6.4 ).

3.6.2. 16,17-Dihydroxy pregnane(ene)s

The MS and MS/MS spectra of the quantitatively major 16,17-

dihydroxy pregnanes are presented in Fig. 6. Whilst 16␣P3 (comp.

79) (Fig. 6E) has been previously discussed [36] the MS spectrum

presented was that of the partially silylated derivative. The partially

silylated product of the compound we observe matches in retention

time and spectrum the published compound.

Considering previously uncharacterized metabolites that are

present in signiﬁcant amounts in all affected neonates (comp.

18, comp. 102, comp. 109 and comp. 126), hydroxylations else-

where than in the D-ring and side chain and at C₃are most

likely to be at positions 1␤ and 6␣. These are major corticosteroid

hydroxylations in unaffected neonates [37,38] and we found no sig-

3.6.3. Absence of 11ˇ,17P in 21-hydroxylase deﬁciency

Two pregnanetriol-20-ones, consistently present in urine from

neonates with 21-hydroxylase deﬁciency, elute near P3 (comp.

12). The ﬁrst (comp. 11), matches in retention time and MS char-

acteristics 6␣17P standard. The second is identiﬁed as 1␤,17P

(comp. 18, Fig. 7A). In this region some authors have located

S. Christakoudi et al. / Steroids 75 (2010) 34–52

47

Fig. 6. GC–MS (left) and GC–MS/MS (right) spectra of the quantitatively major 16,17-dihydroxylated 20-oxo (A–D) and 20-hydroxy (E) pregnanes observed in urine of

neonates affected with 21-hydroxylase deﬁciency.

11␤,17P (21-deoxytetrahydrocortisol) [6,7] in single neonates with

21-hydroxylase deﬁciency. We believe that this identiﬁcation is

erroneous. We ﬁnd that 1␤,17P elutes slightly earlier than 11␤,17P

standard (MU 28.45) in our system but it is possible that in the dif-

ferent chromatographic conditions of others they co-elute. More

importantly, the intensity of the [M-143]⁺ion of 1␤,17P is higher

(reaching about 50%) than it is in 11␤,17P standard (Fig. 7B). The

MS/MS spectra also show differences. Thus 11␤,17P is not present

in urine from patients with 21-hydroxylase deﬁciency. We have,

however, observed a compound matching the standard in the urine

of some patients with adrenal tumors.

3.6.4. Absence of 6˛11ˇ,17,21P in untreated 21-hydroxylase

deﬁciency

There are two major pregnanepentol-20-ones in untreated

21-hydroxylase deﬁciency 1␤,6␣,17,21P (comp. 102, Fig. 8A)

and 1␤,6␣,16␣,17P (comp. 109, Fig. 6 D). An additional peak

appears between them (comp. 107, Fig. 8 B) when hydrocor-

tisone treatment is given. We identify this as 6␣11␤,17,21P

(6␣-hydroxytetrahydrocortisol). Although we could not obtain a

standard for direct comparison, the characteristics of this steroid

have been well described [23,39]. Two pregnanepentol-20-ones

have been noted before in this region in a 4-month-old girl with

48

S. Christakoudi et al. / Steroids 75 (2010) 34–52

Fig. 7. GC–MS (left) and GC–MS/MS (right) spectra of endogenous 1␤,17P (comp. 18) (A) and 11␤,17P standard (3␣,11␤,17␣-trihydroxy-5␤-pregnan-20-one, MU 28.45) (B).

P⁺= [M-31]⁺.

21-hydroxylase deﬁciency [6]. Both were assigned a 21-hydroxy

structure and the ﬁrst of them was identiﬁed as 6␣,11␤,17,21P.

Cortisol metabolites were also reported in this urine specimen but

no indication was given as to whether those were endogenous or

might have arisen from treatment. Similarities of 1␤,6␣,17,21P and

6␣,11␤,17,21P include that they are clearly 21-hydroxylated com-

pounds, based on the presence of ions at m/z 244, 246, and 276 on

MS, the [P-131]⁺series on MS/MS (Fig. 8A and B), and on similar

MU values. However, the spectrum of the borohydride reduction

product of 1␤,6␣,17,21P (Fig. 8C) differs from that published for

Fig. 8. GC–MS (left) and GC–MS/MS (right) spectra of the two major 21-hydroxylated pregnanepentolones observed in urine from neonates affected with 21-hydroxylase

deﬁciency – 1␤,6␣,17,21P (comp. 102) (A) and 6␣,11␤,17,21P (comp. 107) (B), present only when treatment is given and GC–MS spectra of their borohydride reduction

products. These are respectively 5␤-pregnane-1␤,3␣,6␣,17␣,20␰,21-hexol (C) and (D) 5␤-pregnane-3␣,6␣,11␤,17␣,20␰,21-hexol. P⁺= [M-31]⁺.

S. Christakoudi et al. / Steroids 75 (2010) 34–52

49

the respective product of 6␣,11␤,17,21P (Fig. 8D). It contains ions

that belong only to the [M-205]⁺series and lacks the [M-90]⁺

series present in the product of 6␣,11␤,17,21P and has only traces

of the [M-103]⁺series. Thus we conclude that 6␣,11␤,17,21P is

not present in urine of untreated neonates with 21-hydroxylase

deﬁciency and 1␤,6␣,17,21P is a different 21-hydroxylated com-

pound. Intriguingly, both of the above compounds, together with

1␤,6␣,16␣,17P (comp. 109) are present in late pregnancy urine.

Three other, previously unreported pregnanepentolones (comp.

110, comp. 113 and comp. 146) are also only observed when hydro-

cortisone treatment is given.

intensity than in the 20␤-epimers. For example, the ion at m/z 243

is lower in ␣- relative to ␤-cortol and the ion with m/z 333 is lower

in 15␤P3 (comp. 17) relative to 15␤P3 (20␤) (comp. 14). We have

found that the ions at m/z 231 and 245, which we consider to be

diagnostic of a 17,20-dihydroxy structure, are at low intensity or

even absent in the spectra of some examples in urine, which thus

may have a 20␣-conﬁguration.

4.2. Contribution of GC–MS/MS to structure characterization

Whilst for high intensity compounds in relatively simple mix-

tures GC–MS is often sufﬁcient for structure characterization,

identifying those ions in a complex mixture and for low intensity

compounds is challenging. The combined MS and MS/MS approach

we present allows a reliable positioning of hydroxyls on the D-ring

and side chain of unknown pregnane(ene)s in complex mixtures,

without in most cases the need to alter the analytical conditions,

e.g. by including an additional Sephadex LH-20 separation step or

by changing the GC parameters, which may alter the order of elu-

tion. In urine from neonates with 21-hydroxylase deﬁciency, most

types of compounds are present as multiple epimers and isomers

which are often clustered in the same part of the chromatogram or

they co-elute. In such circumstances, MS/MS proves to be especially

useful, particularly for deconvoluting 20-oxo isomers. Sometimes,

the ﬁnding that a certain chromatographic peak has an MS/MS

spectrum containing features of different structures, with variable

proportions between patients, may be the only indication that more

than one compound is present.

It is, however, worth noting that the ion trap is relatively

unfavorable for identiﬁcation purposes in MS mode because the

medium size ions at m/z between 100 and 300 have much lower

intensity compared to those in spectra obtained using a quadrupole

instrument, thus rendering the high intensity D-ring fragmenta-

tion ions hardly usable. Unfortunately, in MS/MS mode, the ion

trap limits the bottom range of masses. Thus ions such as that

at m/z 168 in 16␣,21-dihydroxy-20-oxosteroids or at m/z 156 in

16␣-hydroxy-20-oxosteroids remain below the allowed range for

tetrahydroxy-20-ones and higher molecular weight compounds.

Thus one needs to bear in mind the particular ionization conditions

of the instrument used and be aware of any limitations.

4. Discussion

4.1. Role of D-ring and side chain fragmentations in GC–MS for

structure characterization

Many characteristic GC–MS fragmentations have been previ-

ously discussed and reviewed [9,16,29,31]. Among these, ions

derived from D-ring and side chain fragmentations are promi-

nent and enable hydroxyls positioned in this region, i.e. at C_15–17

and C₂₁to be located. Fragmentations of this type are described

in non-derivatized [40–42], as well as in derivatized pregnanes

[9,27,29,31] and these are also observed in the presence of other

substituents in this region, such as methyl or acetyl groups [43]. It

is thus not surprising that the majority of the new fragmentations

presented in this paper comfortably ﬁt partial or complete D-ring

cleavage or a side chain loss. This is true for 20-hydroxy as well as

20-oxo structures, e.g. the loss of the side chain that gives rise to

the main series in the MS spectra of 20-hydroxy pregnane(ene)s is

also present in 20-oxosteroids, although the equivalent series has

a lower intensity relative to the other ions in the spectra. Similarly,

in 16,17-dihydroxylated pregnane(ene)s the ion at m/z 275 in the

20-oxo variants, derived from a partial D-ring (X) cleavage (Table 4)

is matched by the ion at m/z 319 in 20-hydroxy variants (Table 4).

Our ﬁndings thus demonstrate a continuity in the D-ring and side

chain fragmentation of pregnane(ene)s and we have shown how

most major ions in the MS spectra of less common and unfamiliar

compounds can be inferred and the m/z of the ions calculated by

applying these basic fragmentation principles.

It is likely that both partial and complete D-ring fragmenta-

tions and the loss of the side chain initially occur in all compounds,

irrespective of the positions of the hydroxyls in this region. How-

ever, the stability of the resulting ions and the necessity of further

hydrogen losses or gains are inﬂuenced by the hydroxyls, thus

determining the relative abundance (which may range down to

zero) of the respective ions in the GC–MS spectra. For example,

a Y–OCH₃type of fragmentation is observed in 17-hydroxy-20-

oxosteroids irrespective of the absence (ion at m/z 170) or presence

(ion at m/z 258) of a C₁₅-hydroxyl, but it is not observed when a

The MS/MS approach is less useful for compounds lacking an

oxo group. Such compounds tend to only lose derivatized hydrox-

yls in steps of 90 amu. For 20-hydroxy pregnane(ene)s it may not

provide more clues for characterization than MS alone. Similarly,

it contributes little to the positioning of hydroxyls on the A-, B- or

C-rings.

4.3. Combining GC–MS and GC–MS/MS diagnostic ions for

structure characterization

In the following discussion, the compound structures are

arranged in order of decreasing importance of MS/MS for struc-

ture characterization. Both 20-oxo and 20-hydroxy variants are

considered in each section, with the most useful ions emphasized.

C₁₆-hydroxyl alone is present.

Although we aim to discuss in a separate communication the

MS and especially the MS/MS spectra of pregnane(ene)s with oxo

groups on the A-, B- or C-rings, it is worth noting here that these

often inﬂuence the stability of the D-ring fragmentation ions. For

example, ions of the type [M-X]⁺have high intensity in steroids

with an oxo group at C₃but are absent or with very low intensity

when a hydroxyl group is present at the same position. Thus, in

the MS spectrum of progesterone the [M-99]⁺ion has over 80%

intensity but is absent in pregnanolone and in the MS spectrum of

17-hydroxyprogesterone the [M-187]⁺ion has over 60% intensity

but is absent in 17P.

4.3.1. Characterizing 16,17-dihydroxy pregnane(ene)s

From the triad of ions in MS at m/z 200, 217, and 275, character-

istic of the 20-oxo forms of these compounds, only the last is unique

to this structure but usually has a fairly low intensity. In contrast, in

MS/MS the [P-116]⁺and [P-190]⁺series have high intensity in the

spectra and one of them is the base peak (Fig. 6A–D). Thus, com-

pounds with such a structure stand out in a mixture, even when

in low concentrations. For 20-hydroxy variants, the partial D-ring

fragmentation ion at m/z 319 in MS has very good diagnostic value

(Fig. 6E).

In addition, in 20-hydroxylated pregnane(ene)s the stability of

the speciﬁc D-ring fragmentation ions depends on the orientation

of the hydroxyl at C₂₀. The ions in the 20␣-epimers have lower

50

S. Christakoudi et al. / Steroids 75 (2010) 34–52

4.3.2. Differentiating 16- and 17-monohydroxy pregnane(ene)s

Differentiation of 16␣- and 17-hydroxy compounds is impor-

tant for 21-hydroxylase deﬁciency, because compounds with a

17-hydroxyl have potential diagnostic value while those that are

hydroxylated only at C₁₆on the D-ring are present in the ﬁrst

week of life in unaffected as well as affected neonates. Proges-

terone metabolites with a non-substituted D-ring are observed in

pregnancy and in the ﬁrst days of life but their fragmentation ions

have different m/z values and thus are easy to distinguish. The dif-

ferentiation of C₁₆and C₁₇hydroxylation positions in 20-hydroxy

structures is unambiguous even in MS, because the [M-117]⁺ion

is apparently stable only when a hydroxyl is present at C₁₇. It is

absent when there are no substituents on the D-ring or when only

a hydroxyl at C₁₆is present (see Table 1). The presence of a 21-

hydroxyl does not complicate the distinction, because only the

[M-103]⁺and not the [M-205]⁺series is stable when a 17-hydroxyl

is absent. This is not, however, the case with 20-oxo structures.

MS ions formed by cleavage of the D-ring (the classic fragments

at m/z 156 and 188) are shared by both types of compound, are

independent of the structure of the A-, B-, and C-rings, and thus

are fairly abundant in a complex urine mixture. In contrast, the [M-

143]⁺ion is present only in 17-hydroxy-20-oxosteroids, varies with

the molecular mass of the analyte and appears in a region of the

MS spectrum which has relatively few other ions. The ions at m/z

170 and 172 are also present only in 17-hydroxy-20-oxosteroids

and thus are useful for differentiation in MS. MS/MS gives a fur-

ther distinction. The spectra of the 16-hydroxy-20-oxosteroids are

characterized by the speciﬁc [P-43]⁺series, the ion at m/z 156 and

the fairly robust precursor ion P⁺.

C₁₇and an hydrogen from C₁₆in 17-hydroxyl-20-oxosteroids. The

loss of the side chain in 17,21-dihydroxy-20-oxosteroids comprises

129 amu but the two additional hydrogens must have a different

origin.

4.3.5. Characterizing 16,21-dihydroxylated compounds

Small intensity ions in MS/MS can also be quite valuable when

present in fairly simple spectra, where they stand out. Thus, the ions

at m/z 256 in MS/MS of 16,17,21-trihydroxy-20-oxosteroids and the

analogous ion at m/z 168 in 16,21-dihydroxy-20-oxosteroids have

a very good diagnostic value.

4.4. Some reﬂections on steroid biochemistry in 21-hydroxylase

deﬁciency

4.4.1. Structural variety in 21-hydroxylase deﬁciency

Fetuses affected with congenital adrenal hyperplasia have

to clear large amounts of steroid precursors and this leads to

intra-uterine induction of hydroxylating enzymes that gener-

ate polyhydroxylated metabolites. More than 140 pregnane(ene)

metabolites are listed in Table 5 as detectable in separate sam-

ples of urine from neonates with 21-hydroxylase deﬁciency. We

have yet to report steroids containing oxo groups in rings A–C

and androgens. The extraordinary number of steroids that can

be distinguished is far greater than we ﬁnd in other forms of

congenital adrenal hyperplasia. This may reﬂect the pluripotent

nature of 17-hydroxyprogesterone in the center of the pathway of

cortisol synthesis. A 3␤-hydroxysteroid dehydrogenase (3␤-HSD)

defect results in only pregnenes as major components (although the

presence of pregnanes presents a challenge for distinction from 21-

hydroxylase deﬁciency) while for 17␣-hydroxylase deﬁciency, only

17-deoxy metabolites are formed. Metabolism of 11-deoxycortisol

in 11-hydroxylase deﬁciency appears to be far more limited than

for 17-hydroxyprogesterone. The presence of a 21-hydroxyl group

must make it a substrate for fewer hydroxylating enzymes.

We did not ﬁnd any steroids with 15,16,21-trihydroxy,

15,16,17-trihydroxy, and 15,16,17,21-tetrahydroxy structures.

Hydroxylation of all carbons on the D-ring may be impossible

due to steric hindrance. Even if such compounds did exist, it is

likely that their spectra would be dominated entirely by common,

non-speciﬁc ions derived from vicinal hydroxyls such as the ions

at m/z 191 and 217. The diagnostically useful [M-31]⁺series in

20-oxosteroids diminishes in intensity with increasing number of

hydroxyls in the molecule and it is possible that it would become

undetectable in D-ring per-hydroxylated structures.

Pregnanes included in Table 5 that lack a 17-hydroxyl would

be derived from progesterone, which may originate both from the

mother and the neonate. While some compounds such as 6␣P

(comp. 25) disappear rapidly after birth in all neonates, clearly sug-

gesting a maternal origin, others, such as 16␣P (comp. 26) and allo

16␣P (comp. 31), are similarly eliminated in the ﬁrst few days of

life in unaffected neonates but persist for months in 21-hydroxylase

deﬁciency, clearly indicating a dual origin.

An absence of major 11␤-hydroxylated compounds in untreated

21-hydroxylase deﬁciency is unsurprising, given that neonatal

metabolism of cortisol is shifted towards the 11-oxo form, cortisone

[37]. Only in samples in which there was clear evidence of cortisol

treatment were substantial amounts of 11-hydroxylated steroids

such as tetrahydrocortisol found (11␤,17,21P, used as a standard for

structural analysis – see Tables 1 and 2) and these were accordingly

excluded from Table 5. There are six minor 17,21-dihydroxylated

compounds with one or two hydroxyls in an unknown positions on

the A–C-rings that could be at C_11␤(comp. 119, comp. 142, comp.

52, comp. 72, comp. 124 and comp. 139). These were observed in

neonates older than a week, thus suggesting their neonatal ori-

gin. Thus we believe the 21-hydroxylated compounds reported in

4.3.3. Characterizing 15,17-dihydroxylated compounds

Although the ion at m/z 258 is a very good marker of the 20-oxo

forms of these compounds, in low concentrations MS/MS proves

superior due to the simplicity and clarity of the spectra, with the

[P-114]⁺and [P-107]⁺([P-17-90]⁺) ions being found in constant

proportions in spite of being at much lower intensity than the

base peak [P-90]⁺ion. The 15,17-dihydroxy-20-oxosteroids that we

report in addition to those already published were ﬁrst noticed on

MS/MS.

Regarding the 15,17,20-trihydroxy compounds, the ion at m/z

333 is a good diagnostic marker, but it has lower intensity in

20␣-epimers. This is the only type of structure of 20-hydroxy

compounds where additional use of MS/MS has some merit. The

[P-130]⁺ion has the same m/z as the [M-247]⁺ion and thus prob-

ably is derived by the same fragmentation mechanism. Although

it is present in low intensity in MS/MS of compounds with other

structures (see Table 1) it has the highest intensity in 15,17,20-

trihydroxy pregnane(ene)s.

4.3.4. Characterizing 17,21-dihydroxylated compounds

The ions at m/z 244, 246, and 276 are good markers for the

20-oxo variants but they have a fairly low intensity, belong to

a busy region of the spectrum and could easily be missed in

the complex urinary mixture. The [P-131]⁺series in the MS/MS

spectra assists identiﬁcation. It has a similar intensity for all its

members. The [P-131]⁺series is also observed in some 17-hydroxy-

20-oxosteroids, but the member with lowest m/z (all hydroxyls

lost) has the highest intensity (see Table 2). Although both fragmen-

tations include a derivatized hydroxyl (Table 3) and the side chain

is most likely lost, the mechanism appears to be different. The frag-

mentation in 17-hydroxy-20-oxo compounds is probably the same

as in16-hydroxy-20-oxosteroids, wherea[P-43]⁺series is observed

and probably involves a loss of the non-hydroxylated side chain

(41 amu) and the generation of a double bond between C₁₆and C₁₇

This would require the loss of two hydrogens (one from C₁₆and C₁₇

in 16-hydroxylated compounds and the derivatized hydroxyl at

.

)

S. Christakoudi et al. / Steroids 75 (2010) 34–52

51

Table 5, with the exception of 4 pregnanepentolones which are

metabolites of administered cortisol (Section 3.6.4), to be endoge-

nous.

By comparison of well-preserved samples over a range of post-

natal ages, we believe that we have excluded products of bacterial

degradation and fecal contamination. We also avoided artifact

formation during conjugate hydrolysis by addition of sodium ascor-

bate [44].

deﬁciency as well as in children with this disorder up to the age

of at least 2.5 years. It is thus possible that CYP2D6, which is

capable of hydroxylating pregnanes at C₁₆as well as at C₂₁, is

responsible for the generation of both compounds. This enzyme is

normally present at low levels in the fetus and in the newborn and

increases rapidly after birth but it may be induced in utero by high

concentrations of 17-hydroxyprogesterone. The later decline of the

two pregnanepentolones as well as all the other polyhydroxylated

steroid metabolites, despite the continuing presence of CYP2D6,

could be related to the maturation of glucuronyl-transferases after

birth, which would decrease the necessity of polyhydroxylation as

a route for steroid metabolite excretion.

4.4.2. 16,17-Dihydroxysteroids as potential markers of adrenal

induction in utero

Compounds of this type are very prominent in 21-hydroxylase

deﬁciency, including some with additional oxo groups on the A-, B-

or C-ring, not presented here. Only one example, 16␣P3 (comp.

79), has been previously reported [36]. These may result from

action of CYP17, the enzyme responsible for the generation of 17-

hydroxyprogesterone which also has a parallel 16␣-hydroxylating

activity [10] and so might carry out hydroxylations at C₁₆and

at C₁₇on the same molecule. The gene for this enzyme would

be up-regulated in 21-hydroxylase deﬁciency as part of a general

induction of the adrenal cortisol synthesis pathway. Another pos-

sibility is that 17-hydroxyprogesterone generated in the adrenal is

16␣-hydroxylated in the liver. Although no hepatic cytochromes

have so far been tested for 16␣-hydroxylating activity towards

17-hydroxyprogesterone, likely candidates include CYP3A7, which

has a pivotal role in the metabolism of endogenous and exoge-

nous substances in the fetus and has a well-established steroid

16␣-hydroxylating activity towards dehydroepiandrosterone [45]

and CYP2C19 and CYP2D6. These are present in the later stages

of pregnancy, increase rapidly after birth and are known to 16␣-

hydroxylate progesterone [46,47].

Compounds of the 16,17-dihydroxy type are not present in our

experience in urine from normal newborns but we have observed

them in a 2-week-old boy later diagnosed with 3␤-HSD deﬁciency.

It is thus possible that they are common to any condition asso-

ciated with a block in the cortisol synthesis pathway that leads to

the accumulation of 17-hydroxyprogesterone and so could serve as

general CAH markers. Currently, we ﬁnd the positive diagnosis of

3␤-HSD deﬁciency by urinary steroid proﬁling can reliably be made

only after the age of 3 months, when 3␤-hydroxy-5-ene steroids are

shown to persist and pregnenetriol becomes relatively important.

The 16,17-dihydroxy compounds may enable earlier diagnosis.

The evolution of polyhydroxylated metabolites with postnatal

age and their merit as markers of 21-hydroxylase deﬁciency and

other forms of congenital adrenal hyperplasia will be at the centre

of our further investigations.

5. Conclusion

Based on the presented data, we conclude that GC–MS com-

bined with GC–MS/MS allows reliable identiﬁcation of the structure

of the D-ring and the side chain of pregnane(ene)s. Thanks to

the sensitivity and speciﬁcity of GC–MS/MS, this approach enables

characterization of compounds in low concentrations without prior

isolation from complex mixtures, such as are found in urine from

neonates with 21-hydroxylase deﬁciency and other forms of con-

genital adrenal hyperplasia.

Acknowledgments

We thank Dr. Roy Sherwood for permission to carry out most

of this work within the Department of Clinical Biochemistry, King’s

College Hospital, London, wholly supported from internal resources

and Mr. Christopher Walker and Mr. Richard Caldwell of the Depart-

ment of Forensic Science and Drug Monitoring, King’s College

London, for help with preparing the GC–MS/MS instrument.

References

[1] Caulﬁeld MP, Lynn T, Gottschalk ME, Jones KL, Taylor NF, Malunowicz EM, et al.

The diagnosis of congenital adrenal hyperplasia in the newborn by gas chro-

matography/mass spectrometry analysis of random urine specimens. J Clin

Endocrinol Metab 2002;87:3682–90.

[2] Honour JW, Brook CG. Clinical indications for the use of urinary steroid proﬁles

in neonates and children. Ann Clin Biochem 1997;34:45–54.

[3] Shackleton CHL. Proﬁling steroid hormones and urinary steroids. J Chromatogr

1986;379:91–156.

[4] Shackleton CHL. Congenital adrenal hyperplasia caused by defect in steroid

21-hydroxylase. Establishment of deﬁnitive urinary steroid excretion pattern

during ﬁrst weeks of life. Clin Chim Acta 1976;67:287–98.

[5] McCarver DG, Hines RN. The ontogeny of human drug-metabolizing enzymes:

phase II conjugation enzymes and regulatory mechanisms. J Pharmacol Exp

Ther 2002;300:361–6.

[6] Kraan GPB, Wolthers BG, van der Molen JC, Nagel GT, Drayer NM, Joannou GE.

New identiﬁed 15␤-hydroxylated 21-deoxy-pregnanes in congenital adrenal

hyperplasia due to 21-hydroxylase deﬁciency. J Steroid Biochem Mol Biol

1993;45:421–34.

[7] Steen G, Tas AC, de Brauw MCT, Drayer NM, Wolthers BG. The early recognition

of the 21-hydroxylase deﬁciency variety of congenital adrenal hyperplasia. Clin

Chim Acta 1980;105:213–24.

[8] Sjovall J. Fifty years with bile acids and steroids in health and disease. Lipids

2004;39:703–22.

[9] Joannou GE. Identiﬁcation of 15␤-hydroxylated C₂₁steroids in the neo-natal

period: the role of 3␣,15␤,17␣-trihydroxy-5␤-pregnan-20-one in the perina-

tal diagnosis of congenital adrenal hyperplasia (CAH) due to 21-hydroxylase

deﬁciency. J Steroid Biochem 1981;14:901–12.

[10] Swart P, Swart AC, Waterman MR, Estabrook RW, Mason JI. Progesterone 16␣-

hydroxylase activity is catalyzed by human cytochrome P450 17␣-hydroxylase.

J Clin Endocrinol Metab 1993;77:98–102.

4.4.3. 21-Hydroxylation of 17-hydroxyprogesterone

The paradoxical presence of 21-hydroxylated steroids that are

also 17-hydroxylated (17,21-dihydroxy, 15,17,21- and 16,17,21-

trihydroxy structures, see Table 1) requires explanation. They

are unlikely to be products of residual adrenal 21-hydroxylase

(CYP21A2) since 1␤,6␣,17,21P (comp. 102) provides an excel-

lent marker of 21-hydroxylase deﬁciency and is not diminished

in those with the more severe salt wasting form of the dis-

order. Extra-adrenal 21-hydroxylation of 17-deoxy substrates in

21-hydroxylase deﬁciency is well established [13]. A recent study

has conﬁrmed that the cytochromes CYP2C19 and CYP3A4 will 21-

hydroxylate progesterone but not 17-hydroxyprogesterone in vitro

[14] but a more likely candidate for extra-adrenal 21-hydroxylation

is CYP2D6. It has been shown in brain studies to be capable of 21-

hydroxylating 17-hydroxyprogesterone [48]. The likely substrate

for the putative enzyme, apart from being 17␣-hydroxylated, must

lack an 11␤-hydroxyl.

The structures we have suggested for the two major

pregnanepentol-20-ones 1␤,6␣,17,21P (comp.

1␤,6␣,16␣,17P (comp. 109) share four hydroxyls at C₁, C₃,

C₆, and C₁₇and differ by having a hydroxyl at C₂₁in the former

and at C₁₆in the latter. We consistently ﬁnd these in similar pro-

portions in urine from untreated neonates with 21-hydroxylase

102) and

[11] Winkel CA, Casey ML, Worley RJ, Madden JD, MacDonald PC. Extraadrenal

steroid 21-hydroxylase activity in a woman with congenital adrenal hyper-

plasia due to steroid 21-hydroxylase deﬁciency.

1983;56:104–7.

J Clin Endocrinol Metab

52

S. Christakoudi et al. / Steroids 75 (2010) 34–52

[12] Shackleton CHL, Homoki J, Taylor NF.

A

paradox: elevated 21-hydroxy-

[30] Brooks CJW, Harvey DJ. Gas chromatographic and mass spectrometric studies

of oximes derived from 20-oxosteroids. Steroids 1970;15:283–301.

[31] Gustafsson BE, Gustafsson JA, Sjovall J. Steroids in germfree and conventional

rats. 2. Identiﬁcation of 3␣,16␣-dihydroxy-5␣-pregnan-20-one and related

compounds in faeces from germfree rats. Eur J Biochem 1968;4:568–73.

[32] Ramirez LC, Bournot P, Maume BF. Bioconversion of 16␣-hydroxyprogesterone

to 16␣-hydroxylated corticosteroids by newborn rat adrenal cells in primary

culture. Biochim Biophys Acta 1986;886:275–85.

pregnenolone production in newborns with 21-hydroxylase deﬁciency.

Steroids 1987;49:295–311.

[13] Mellon SH, Miller WL. Extraadrenal steroid 21-hydroxylation is not mediated

by P450c21. J Clin Invest 1989;84:1497–502.

[14] Gomes LG, Huang N, Agrawal V, Mendonc¸ a BB, Bachega TASS, Miller WL. Extraa-

drenal 21-hydroxylation by CYP2C19 and CYP3A4: effect on 21-hydroxylase

deﬁciency. J Clin Endocrinol Metab 2009;94:89–95.

[15] Sjovall J, Axelson M. Newer approaches to the isolation, identiﬁcation and

quantiﬁcation of steroids in biological materials. Vitam Horm 1982;39:31.

[16] Shackleton CH. Mass spectrometry: application to steroid and peptide research.

Endocr Rev 1985;6:441–86.

[17] Gambelunghe C, Sommavilla M, Ferranti C, Rossi R, Aroni K, Manes N, et

al. Analysis of anabolic steroids in hair by GC/MS/MS. Biomed Chromatogr

2007;21:369–75.

[33] Paul V, Harvey DJ. The electron impact induced cleavage of the C₁₇–C₂₀bond

and D-ring in trimethylsilyl derivatives of C₂₁steroids. Reciprocal exchange of

trimethylsilyl groups. Biomed Mass Spectrom 1980;7:217–25.

[34] Gustafsson JA, Ryhage R, Sjovall J, Moriarty RM. Migrations of the trimethylsilyl

group upon electron impact in steroids. J Am Chem Soc 1969;91:1234–6.

[35] Lisboa BP, Gustafsson JA. Biosynthesis of 18-hydroxytestosterone in the human

foetal liver. Eur J Biochem 1969;9:402–5.

[18] Marcos V, Perogordo E, Espinosa P, De Pozuelo MM, Hooghuis H. Multiresidue

analysis of anabolic compounds in bovine hair by gas chromatography–tandem

mass spectrometry. Anal Chim Acta 2004;507:219–27.

[19] Nielen MWF, Lasaroms JJP, Essers ML, Sanders MB, Heskamp HH, Bovee TFH, et

al. The ultimate veal calf reference experiment: hormone residue analysis data

obtained by gas and liquid chromatography tandem mass spectrometry. Anal

Chim Acta 2007;586:30–4.

[20] Cheng MT, Barbalas MP, Pegues RF, McLafferty FW. Tandem mass spectrome-

try: structural and stereochemical information from steroids. J Am Chem Soc

1983;105:1510–3.

[21] Taylor NF. Urinary steroid proﬁling. In: Wheeler MJ, Hutchinson JSM, editors.

Hormone assays in biological ﬂuids. Totowa, New Jersey: Humana Press; 2006.

p. 159–75.

[22] Shackleton CHL, Whitney JO. Use of Sep-Pak (R) cartridges for urinary steroid

extraction: evaluation of the method for use prior to gas chromatographic

analysis. Clin Chim Acta 1980;107:231–43.

[23] Setchell KDR, Gontscharow NP, Axelson M, Sjovall J. The identiﬁcation

of 3␣,6␤,11␤,17,21-pentahydroxy-5␤-pregnan-20-one (6␤-hydroxy-THF) –

the major urinary steroid of the baboon (Papio papio). J Steroid Biochem

1976;7:801–8.

[24] Shackleton CHL, Gustafsson JA, Sjovall J. Steroids in newborns and infants.

Identiﬁcation of steroids in urine from newborn infants. Steroids 1971;17:

265–80.

[25] Gustafsson JA, Shackleton CH, Sjovall J. Steroids in newborns and infants. C₁₉

and C₂₁steroids in faeces from infants. Eur J Biochem 1969;10:302–11.

[36] van der Ploeg KR, Wolthers BG, Nagel GT, Volmer M, Drayer NM. The diagnosis

of 21-hydroxylase deﬁciency in a prematurely born infant on the basis of the

urinary steroid excretion pattern. Clin Chim Acta 1982;120:341–53.

[37] Shackleton CHL. Steroid synthesis and catabolism in the fetus and neonate.

In: Makin HLJ, editor. Biochemistry of steroid hormones. Oxford: Blackwell

Scientiﬁc Publications; 1984. p. 441–77.

[38] Taylor NF, Curnow DH, Shackleton CHL. Analysis of glucocorticoid metabolites

in the neonatal period: catabolism of cortisone acetate by an infant with 21-

hydroxylase deﬁciency. Clin Chim Acta 1978;85:219–29.

[39] Setchell KDR, Axelson M, Sjovall J, Kirk DN, Morgan RE. The identiﬁca-

tion of 6␣-hydroxytetrahydrocortisol (3␣,6␣,11␤,17␣,21-pentahydroxy-5␤-

pregnan-20-one) in urine. FEBS Lett 1978;88:215–8.

[40] Hammerum S, Djerassi C. Mass spectrometry in structural and stereochemical

problems CCXLV. [1] The electron impact induced fragmentation reactions of

17-oxygenated progesterones. Steroids 1975;25:817–26.

[41] Tokes L, LaLonde RT, Djerassi C. Mass spectrometry in structural and

stereochemical problems. CXXVII. Synthesis and fragmentation behavior of

deuterium-labeled 20-keto steroids. J Org Chem 1967;32:1020–9.

[42] Tokes L, Jones G, Djerassi C. Mass spectrometry in structural and stereochemical

problems. CLXI. Elucidation of the course of the characteristic ring D fragmen-

tation of steroids. J Am Chem Soc 1968;90:5465–77.

[43] Hammerum S, Djerassi C. Mass spectrometry in structural and stereochemical

problems. CCXLIV. The inﬂuence of substituents and stereochemistry on the

mass spectral fragmentation of progesterone. Tetrahedron 1975;31:2391–400.

[44] Christakoudi S, Cowan DA, Taylor NF. Sodium ascorbate improves yield

of urinary steroids during hydrolysis with Helix pomatia juice. Steroids

2008;73:309–19.

[26] Gachancard-Bouya J-LJ, Begue R-J. Urinary steroids from

a newborn

human infant. Identiﬁcation of 2␣-hydroxy-4-pregnene-3,20-dione, 3␤,15␤-

dihydroxy-5-pregnen-20-one and 3␤,15␣-dihydroxy-5-pregnen-20-one.

Steroid Biochem Mol Biol 1994;49:213–26.

J

[45] Kitada M, Kamataki T, Itahashi K, Rikihisa T, Kanakubo Y. P-450 HFLa, a form of

cytochrome P-450 puriﬁed from human fetal livers, is the 16␣-hydroxylase of

dehydroepiandrosterone 3-sulfate. J Biol Chem 1987;262:13534–7.

[46] Niwa T, Yabusaki Y, Honma K, Matsuo N, Tatsuta K, Ishibashi F, et al.

Contribution of human hepatic cytochrome P450 isoforms to regioselective

hydroxylation of steroid hormones. Xenobiotica 1998;28:539–47.

[47] Yamazaki H, Shimada T. Progesterone and testosterone hydroxylation by

cytochromes P450 2C19, 2C9, and 3A4 in human liver microsomes. Arch

Biochem Biophys 1997;346:161–9.

[27] Gustafsson JA, Sjovall J. Steroids in germfree and conventional rats. 6. Identiﬁ-

cation of 15␣- and 21-hydroxylated C₂₁steroids in faeces from germfree rats.

Eur J Biochem 1968;6:236–47.

[28] Janne OA, Vihko RK. Neutral steroids in urine during pregnancy. Identiﬁca-

tion of sulphate-conjugated 3␤,15␣-dihydroxy-5␣-pregnan-20-one, 3-␤,15␣-

dihydroxy-5␤-pregnan-20-one and 5-pregnene-3␤,15␤,17␣,20␣-tetrol. Eur J

Biochem 1970;17:134–40.

[29] Vouros P, Harvey DJ. Method for selective introduction of trimethylsilyl and

perdeuterotrimethylsilyl groups in hydroxy steroids and its utility in mass

spectrometric interpretations. Anal Chem 1973;45:7–12.

[48] Kishimoto W, Hiroi T, Shiraishi M, Osada M, Imaoka S, Kominami S, et

al. Cytochrome P450 2D catalyze steroid 21-hydroxylation in the brain.

Endocrinology 2004;145:699–705.

Article Doi

Steroids excreted in urine by neonates with 21-hydroxylase deficiency: Characterization, using GC-MS and GC-MS/MS, of the D-ring and side chain structure of pregnanes and pregnenes

DOI: 10.1016/j.steroids.2009.09.011

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.steroids.2009.09.011

Full text of DOI:10.1016/j.steroids.2009.09.011

Products guided by the article

R&D Labs maybe for 33287-42-2

Relevant to this article

Hot Product