Supplemental Data
CHEMICALS AND REAGENTS
3,3-dimethylglutaric, methoxyamine hydrochloride, sodium chloride, pyridine and acetyl chloride were from Sigma Aldrich. N,O-bis(trimethylsilyl)trifluoroacetamide containing 1% trimethylchlorosilane was from Regis Technologies. Ethoxyamine hydrochloride was from Fluka and ethyl acetate, iso-octane, 25% ammonia and n-butanol were from Merck.
URINE ORGANIC ACID ANALYSIS
Creatinine was used to correct for urine dilution and samples were diluted to a creatinine concentration of 1 mmol/L. Diluted urine samples (1 mL) were mixed with 100 mL of 1 mol/L methoxyamine hydrochloride containing 1 mmol/L 3,3-dimethylglutaric acid (internal standard) and incubated at 60oC for 30 minutes. Fifty mL of 6 mol/L hydrochloric acid was added, followed by solid sodium chloride and solvent extraction with 5 mL ethyl acetate. The ethyl acetate phase was removed and 10 mL of 25% ammonia was added to minimize loss of volatile organic acids. Ethyl acetate was removed by evaporation under an air stream and 20 mL of pyridine and 80 mL of N,O-bis(trimethylsilyl)trifluoroacetamide containing 1% trimethylchlorosilane was added, the mixture incubated at 80oC for 30 minutes, cooled and 900 mL of iso-octane added. This process is referred to as “standard derivatization”.
Experiments were performed in which alternative derivatives were prepared and analysed by GC-MS to assist in the structural characterization of some of the novel organic acids. GC-MS conditions were identical. Control urine was analysed in each experiment and the chromatograms were checked for: a) presence of a new derivative and absence of the standard derivative in affected samples using the alternative derivatization, b) absence of the new derivative in the control sample using the alternative derivatization, c) absence of the new derivative in affected samples using standard derivatization.
Ethoxyamime/trimethylsilyl derivatives were prepared by substitution of ethoxyamine hydrochloride for methoxyamine hydrochloride in the above analytical process. All other conditions were the same. Butyl/trimethylsilyl derivatives were prepared by heating the residue after urine solvent extraction with n-butanol:acetyl chloride (9:1 v/v) for 30 minutes at 65oC. Excess reagents were then removed under an air stream and the residue trimethylsilylated as described above.
GAS CHROMATOGRAPHY-MASS SPECTROMETRY
Two GC-MS systems were used to analyse the derivatized extracts. System 1 was an Agilent 5973 GC-MS system fitted with a 30 m HP-5MS column, internal diameter 0.2 mm, film thickness 0.25 mm. One µL of derivatized extract was injected using splitless injection, the column was held at 80oC for 1 minute followed by a temperature programmed ramp of 8oC/minute to 300oC. The mass spectrometer was scanned from 50 to 550 m/z every second. Semi-quantitation of organic acids was performed assuming that TIC peak areas gave the same molar response as the internal standard. System 2 was a Perkin Elmer Turbomass GC-MS fitted with a J&W 20 m BP5-MS column, internal diameter 0.18 mm, film thickness 0.18 mm.
SYNTHESIS OF METABOLITES
Metabolites B, C, E and G were prepared from commercially available 4-hydroxy-6-methyl-2-pyrone (D) by the following reactions: Catalytic hydrogenation of D in ethanol solution with 10% palladium on carbon afforded 3,5-dihydroxyhexanoic 1,5 lactone (B) (Huck et al. 2001). Oxidation of B with Jones reagent gave 5-hydroxy-3-ketohexanoic 1,5-lactone (J). The methoxyamine derivative of J afforded two peaks by GC-MS analysis with identical mass spectra (54 (100%), 59 (68%), 68 (38%), 113 (29%), 157 (M+, 8%)). Hydrolysis of J with 1 mol/L sodium hydroxide yielded two peaks of 5-hydroxy-3-ketohexanoic acid (E) with identical mass spectra due to oxime syn/anti isomerism). Dehydration of B with phosphorous pentoxide in diethyl ether gave 6-methyl-5,6-dihydro-2-pyrone (K). Hydrolysis of K with 1 mol/L sodium hydroxide yielded cis-5-hydroxyhex-2-enoic acid (C). Hydrolysis of B with 1 mol/L sodium hydroxide gave metabolite G.
IDENTIFICATION OF ORGANIC ACIDS
Peaks B, C, D, E and G in patient samples were identified as the methoxyamine/TMS derivatives of 3,5-dihydroxyhexanoic 1,5 lactone, cis-5-hydroxyhex-2-enoic acid, 4-hydroxy-6-methyl-2-pyrone, 5-hydroxy-3-ketohexanoic acid and 3,5-dihydroxyhexanoic acid respectively by spectral comparisons with the authentic compounds and co-elution of the endogenous and authentic compounds. The spectra are shown in Fig. 1. Double chromatographic peaks for 3,5-dihydroxyhexanoic 1,5 lactone and 3,5-dihydroxyhexanoic acid were observed in the subject urine samples. Both compounds contain two chiral centers and this observation indicates the formation of diastereoisomeric pairs in vivo. Single chromatographic peaks were obtained for the synthetic compounds indicating stereo-specificity in the synthesis. Synthetic derivatized 3,5-dihydroxyhexanoic 1,5 lactone corresponded to the slower eluting peak and synthetic derivatized 3,5-dihydroxyhexanoic acid corresponded to the faster eluting peak of the pairs seen in subject urine samples. The remaining two compounds were provisionally identified by comparisons with above compounds and the formation of alternative derivatives.
The spectrum of peak A, shown in Fig. 1A, has several features in common with that of cis-5-hydroxyhex-2-enoic acid (Fig. 1C). M-15 ions are prominent in trimethylsilyl derivatives (Halket 1993) and the ion of 259 m/z and the less abundant ion at 274 m/z indicated the same molecular weight of 274 Da. This molecular weight and the abundant ion at 143 m/z, from cleavage α to a trimethylsilyl group (Rontani et al. 2008), suggested a 3-hydroxy unsaturated hexanoic acid. Formation of a butyl/trimethylsilyl derivative produced a mass spectrum (73(100%), 143(81%), 157(26%), 159(28%), 243(14%)) consistent with this structure. Because of the inter-relationship with 4-hydroxy-6-methyl-2-pyrone we propose a trans-configuration about the double bond and provisionally identify peak A as trans-3-hydroxyhex-4-enoic acid.
The spectrum of peak F, shown in Fig. 1F, has several features in common with that of 5-hydroxy-3-keto-hexanoic acid (Fig. 1E). The M-15 ion at 304 m/z indicated the same molecular weight of 319 Da and the abundant ion at 233 m/z is typical of trimethylsilyl derivatives of 3-hydroxy acids (Rontani et al. 2008). The odd molecular weight, in combination with an M-31 ion at 288 m/z and an 89 m/z fragment suggested a methoxyamine derivative of a keto acid. This was confirmed by the formation of an ethoxyamine/trimethylsilyl derivative which gave a homologous spectrum (73(100%), 103(11%), 147(13%), 189(10%), 233(16%), 288(1%), 318(3%)) with several 14 m/z shifts. On the basis of these results, peak F was provisionally identified as 3-hydroxy-5-ketohexanoic acid.
Peaks A and C were not detected when derivatized metabolite G was analysed as a pure standard, indicating that the former two unsaturated endogenous compounds did not arise from thermal dehydration of the latter in the gas chromatography injection system. In contrast to 3,5-dihydroxyhexanoic acid 1,5 lactone, 5-hydroxy-3-ketohexanoic acid 1,5 lactone (peak J) was not detected in any urine samples. The methoxyamine of J was synthesized to confirm this.
References
Halket JM. Derivatives for Gas Chromatography-Mass Spectrometry. In: Blau K, Halket JM, eds. Handbook of Derivatives for Chromatography. 2 Ed. Chichester: John Wiley & Sons 1993:297-325.
Huck W-R, Burgi T, Mallat T, Baiker A (2001) Asymmetric hydrogenation of 4-hydroxy-6-methyl-2-pyrone: role of acid-base interactions in the mechanism of enantiodifferentiation. Journal of Catalysis 200171-80
Rontani JF, Aubert C (2008) Hydrogen and trimethylsilyl transfers during EI mass spectral fragmentation of hydroxycarboxylic and oxocarboxylic acid trimethylsilyl derivatives. J Am Soc Mass Spectrom 19(1):66-75
Fig. 1 Mass spectra of peaks A to G as methoxyamine/trimethylsilyl (TMS) derivatives. Chiral centers are indicated by asterisks. A, trans-3-hydroxyhex-4-enoic acid (provisional identification); B, 3,5-dihydroxyhexanoic 1,5 lactone; C, cis-5-hydroxyhex-2-enoic acid; D, 4-hydroxy-6-methyl-2-pyrone; E, 5-hydroxy-3-ketohexanoic acid; F, 3-hydroxy-5-ketohexanoic acid (provisional identification); G, 3,5-dihydroxyhexanoic acid.
Fig. 2 Proposed metabolic pathways for the formation of the novel metabolites
Page 1 of 11
Table 1 Clinical and mutation details of biochemically characterized individuals with HMCS2 deficiency (cohort A)
Biochemical measurements were performed during the first presentation. Reference ranges are indicated in brackets.
Family:Ethnic origin
Consanguinity / Allele 1 mutation / Allele 2 mutation / Family-patient / Clinical features at presentation / Symptomatic presentations1 / Ammonium µmol/L (<50) / Lactate mmol/L (0.5 – 2.2) / D-3-hydroxybutyrate mmol/L (<0.6) / Free fatty acids mmol/L (<0.7) / Current age (years)
A1:
Lebanese
Distantly related / Exon 1 deletion / Exon 1 deletion / A1-1
A1-2
A1-3
A1-4 / hypoglycemia, hepatomegaly, metabolic acidosis / 8m; 23m
9m
6m
10m; 11m / -
194
29
51 / -
0.8
5.0
0.8 / 0.3
0.1
-
0.2 / 3.1
2.3
-
0.7 / 19
17
10
8
A2:
Greek / p.L266S
(c.797T>C) / p.I407T
(c.1220T>C) / A2-1 / hypoglycemia / 3y1m / 52 / 0.7 / 0.6 / 2.5 / 9
A3:
Egyptian
Consanguineous / p.G388R
(c.1162G>A) / p.G388R
(c.1162G>A) / A3-1
A3-2 / hypoglycemia, metabolic acidosis, hepatomegaly / 6m
Asymptomatic / -
- / 1.1
- / 0.2
- / 2.7
- / 7
5
A4:
Indian / p.G169D
(c.506G>A) / p.R505Q
(c.1514G>A) / A4-1 / hypoglycemia / 2y5m / Normal / - / 0.2 / - / 8
1 y=year, m=month
Page 1 of 11