Chem. Res. Toxicol. 1992,5, 553-560
553
Characterization of Urinary Metabolites from [ 1,2,metho~y-~~C]-2-Methoxyethanol in Mice Using 13C Nuclear Magnetic Resonance Spectroscopy' Susan C. J. Sumner,. Donald B. Stedman, David 0. Clarke,2 Frank Welsch, and T i m o t h y R. Fennel1 Chemical Industry Institute of Toxicology, P.O.Box 12137, Research Triangle Park, North Carolina 27709 Received March 10,1992
2-Methoxyethanol (2-ME) is an industrial solvent that induces developmental and testicular toxicity in laboratory animals. Oxidation of 2-ME to 2-methoxyacetic acid (2-MAA) is required for the generation of these adverse effects. The urinary metabolites of 2-ME were investigated to characterize the fate of 2-ME and 2-MAA. 13C NMR spectroscopy was used to detect and assign metabolites in the urine of pregnant CD-1 mice following administration of 250 mg/kg of [1,2,methoxy-W]-2-ME. Two-dimensional NMR methods were used to correlate signals from the labeled carbons in each 2-ME metabolite and to determine the number of hydrogens attached to each carbon. Structures were assigned from the NMR data together with calculated values of shift for biochemically feasible metabolites and by comparison to standards. Pathways involved in forming metabolites assigned in this study include transformation of 2-ME via ethylene glycol, conjugation with glucuronide or sulfate, and oxidation to 2-MAA. Additional metabolites were assigned that can be formed from further conversion of 2-MAA to glycine and glucuronide conjugates, as well as metabolites derived from the incorporation of 2-methoxyacetyl CoA derivatives into intermediary metabolism. Elucidation of the further metabolism of 2-MAA may be important for understanding the mechanisms by which 2-ME induces adverse effects. Introduction
2-Methoxyethanol is an industrial solvent used tomanufacture protective coatings such as paints, lacquers, and epoxy resins. Testicular ( I , 2) and developmental toxicities occur in laboratory animals (3-6) and subhuman primates (7)exposed to relatively high doses of 2-ME. The adverse effects of 2-methoxyacetic acid (2-MAA) are similar to those observed for 2-ME, indicating that bioactivation of 2-ME to the putative proximate toxicant, 2-MAA, is essential for adverse biological responses (1, 6-10). The conversion of 2-ME to 2-MAA is catalyzed by alcohol dehydrogenase (ADH) (1, 111, and inhibition of ADH reduces the adverse effects of 2-ME (1,6, 10). 2MAA is excreted in urine from rats (1,11)and mice (10) following the administration of 2-ME. The presence of N-(methosyacety1)glycinein urine from rats (1)and mice (12) that were administered 2-ME indicates that 2-MAA is also converted to a coenzyme A thioester (2-methoxyacetyl-CoA), which is then conjugated with glycine. Ratsexposedto [W1-2-ME (11)andmiceexposedto [ W I 2-ME labeled in any carbon position or given [l4C1-2MAA exhale 14C02(10,12), indicating the involvement of other routes of metabolism. Inhibition of 14C02evolution by fluoroacetate and sodium acetate in mouse embryos exposed to [Wl-2-MAA in whole embryo culture suggests
* Author to whom correspondence ahould be addressed.
1 Presented in part at the 1991 Annual Meeting of the Society of Toxicology (Dallas, TX). * Currentaddreea: LillyResearchLaboratories,Eli Lilly and Company, Greenfield, IN 46140. 3 Abbreviations: 2-ME, 2-methoxyethanol;2-MAA, 2-methoxyacetic acid; ADH, alcohol dehydrogenase; gd, gestation day; INADEQUATE, incredible natural abundance double quantum coherence transfer; HETSDJ, heteronuclear 2D J-resolved.
that 2-methoxyacetyl4oA may enter the Krebs cycle (12).Alternative pathways in the metabolism of ethoxy-, butoxy-, and methoxyethanol in rats have been suggested that involve conversion of the glycol ethers to ethylene glycol or conjugation to form glucuronides (13). Evolution of 14C02after administration of [l4C1-2-MEis consistent with the oxidation of ethylene glycol (14,15). The mechanisms by which 2-MAA exerts its adverse effects have not been elucidated. These effects may be directly mediated by 2-MAA or may result from other metabolites. The objectives of this study were to identify the urinary metabolites of 2-ME in CD-1 mice and to determine the pathways involved in their formation. l3C NMR spectroscopywas used to detect metabolites directly in the urine collected from pregnant CD-1 mice that received [1,2,metho~y-~%] -2-ME. Structures were characterized using two dimensional NMR methods for correlating the labeled carbon signalsof each 2-ME metabolite and for determining the number of hydrogens attached to each labeled carbon. Assignments were made by comparing calculated values of shift for biochemically feasible metabolites with chemical shifts for the labeled carbons. As described previously (16, 17), this approach provides structural information for the characterization of metaboliteswithout the need for authentic standardsand without the possibility of generating structural artifacts that may arise during extraction, chromatography, and mass spectrometry. Materials and Methods Chemicals. [1,2,metho~y-~3C1-2-ME (99% enriched) was obtained from Isotec, Inc. (Miamisburg, OH). A sample of [1,2,methoxyJW1-2-MEin DzO gave the expected ['HI-decou-
0893-228x/92/2705-0553$03.00/00 1992 American Chemical Society
664 Chem. Res. Toxicol., Vol. 5, No. 4, 1992
pled 13CNMR spectrum consisting of a doublet (73 ppm, 41 Hz, CHzOR),a doublet (58ppm, 3 Hz,CH3), and a doublet of doublets (60 ppm, 41 and 3 Hz, CHzOH). DzO and CDCls were obtained from Merck & Co., Inc. (St. Louis, MO). 2-Hydroxy-3-methoxy1,2,3-propanetricarboxylic acid triethyl ester (triethyl methoxycitrate) and 4-methoxy-3-butenoic acid methyl ester were prepared by Dr. V. Amarnath, Department of Pathology, Duke University, Durham, NC, and analyzed by lH and NMR spectroscopy. Animals. Male and female Cr1:CD-l(1CR)BR (CD-1) mice, approximately 10 weeks old, were purchased from Charles River Breeding Laboratories (Raleigh, NC) and housed as described (18). Nulliparous females of 25-35 g body weight were paired 1:lin the home cage of males of proven fertility during the morning hours for the last 2 h of the dark cycle. Females with vaginal plugs were held in groups of four per cage. The next 24 h were designated as gestation day (gd) 0. Administration of 2-ME.On gd 11,three pregnant mice were administered [1,2,neth0xy-~~C1-2-ME (250 mg/kg) by gavage. The 2-ME was diluted in distilled water to yield a constant dosing volume of 5.2 pL/g body weight. Embryos are maximally susceptible to 2-ME-induced digit malformations on gd 11 (5, 10). The animals were placed in all glass metabolism cages designed to collect urine and feces separately. They were supplied food (NIH 07 diet, Zeigler Bros., Gardner, PA) and deionized water ad libitum and maintained on a 12-h light-dark cycle at a temperature of 22 A 2 OC and relative humidity of 55 f 5%. Urine excreted in the first 24 h following gavage was collected. Control urine samples were collected from three additional mice. Sample Preparation. All urine samples were centrifuged at lO00g for 10 min and either analyzed immediately or stored at -20 O C until analysis. Samples were prepared for NMR studies by adding 75 pL of D20 to 400 pL of the centrifuged urine. One to 475 sample was prepared by adding [1,2,meth0xy-~~C1-2-ME pL of control urine containing 16% DzO for a concentration of -100 mM. A concentrated sample of urine was prepared for two-dimensional NMR studies by combining 400-pL aliquots of the urine collected individually from three mice that received 250 mg/kg of [1,2,methoxy-W]-2-ME. Methanol (1.5 mL) was added to the pooled urines, and the sample was centrifuged for 10 min. The volume of the supernatant was reduced (to -100 pL) under a stream of nitrogen gas, and 400 pL of DzO was added. NMR Spectroscopy. NMR spectra were acquired with a 5-mm dual proton-multinuclear probe on a Varian VXR-300 spectrometer (Palo Alto, CA). Carbon-13 NMR spectra were acquired in the double precision mode with an acquisition time of 0.9 s,30K data points, a relaxation delay of 5 s, and a 60' pulse width. Spectra were obtained with approximately 10 OOO transients and referenced to urea at 162.45 ppm. Two-dimensional incredible natural abundance double quantum transfer (INADEQUATE) spectra (19) were acquired on the sample of concentrated mouse urine using the INADQTprogram from the Varian pulse sequence library. A relaxation delay of 5 s and T values corresponding to coupling constants between 40 and 80 Hz were used to obtain data over the entire spectral window. Broadband decoupling was employed throughout the pulse sequence, and spectra were acquired in the phase-sensitive mode with 2048 points in tz and 32 points in tl. Two-dimensional heteronuclear J-resolved (20) spectra were acquired using the HFIWXTprogram from the Varian pulse sequence library, with gated decoupling applied to the proton spin system. NMR spectra were obtained with 2048 points in tz and 64 points in tl. Values of shift for carbons of feasible metabolites were calculated using incremental substituent effects for alkanes and alkenes (21,22)and using the Carbon-13 NMR Database (STN International, Columbus, OH).
Results Metabolitesin MouseUrine. The l3C NMR spectrum of control mouse urine (Figure 1A) shows an intense sin-
Sumnet et al.
t I
ll
7a 8a
/IIF
Figure 1. 13C NMR spectra of control urine (A) and urine collected for 24 h after administration of 250 mg/kgof [1,2,meth0xy-~~Cl-2-ME (B) to pregnant CD-1 mice. Signals for metabolites are labeled according to metabolite number (see Table I and Scheme I) and the letter of the carbon derived from 2-ME (a, b, or c for the 1, 2, and methoxy carbons, respectively). Spectrum C is plotted at 40 times the vertical intensity of spectrum B, showing signals for metabolites in lowest concentrations. glet at 162.5 ppm that is assigned to urea. The less intense signals in the control spectrum are consistent with sugars, hippurate, citrate, and creatinine (23). The 13C NMR spectrum of the mouse urine collected for 24 h following oral administration of 250 mgikg of [1,2,metho~y-~W1-2-ME (Figure 1B,C)contains the signals from endogenous compounds and also contains signals not present in the control spectrum. Signals that are not present in the control spectrum and that have coupling patterns associated with adjacent 13Cnuclei are assigned to metabolites of 2-ME. The coupling patterns are produced by interactions between the labeled carbons in the portion of each metabolite derived from 2-ME. The carbons in normal urinary compounds (1.1%natural abundance of I3C) give rise to signals which appear as singletsbecause of the low incidence of adjacent 13Cnuclei. T w o metabolites of 2-ME give rise to the intense signals (labeled 6 and 7) in Figure lB, while many metabolites that are formed in lower concentrations give rise to the less intense signals shown in Figure 1C (40 X vertical scale of Figure 1B). The chemicalshifts (recorded for the center of the multiplet patterns) for signals derived from the labeled carbons of 2-ME and the carbon-carbon coupling constants are listed in Table I. Additional signals (having low intensities) derived from the labeled carbons of 2-ME are also present in the urine spectra but have not been assigned. INADEQUATE spectroscopy was used to correlate carbon signals in the urine samples, allowing for the de-
Urinary Metabolites of [1,2,metho~y-~~C]-2-Methoxyethanol
Chem. Res. Toxicol., Vol. 5, No. 4,1992 555
expanded region of the HET2DJ spectrum is shown in Figure 3, and the 1D 13CNMR spectrum is plotted on the F2 axis. The number of hydrogens attached to a carbon is equal to the number of contours at the carbon’s chemical shift position minus one. For example,each of the intense signals between 56 and 59 ppm (F2 axis) has four contours on the F1 axis, indicating methyl carbons (CH3). Each of the two signals for carbon 2b (61 ppm) has three contours on the F1 axis, indicating a methylene carbon (CH2). The proton multiplicity for each signal derived from 2-ME is listed in Table I. In the HET2DJ experiment, signals from carbons that exhibit direct carbon-13splitting produce contour patterns that are not symmetrical around 0 Hz on the F1 axis. Signals from endogenous or exogenous compounds not exhibiting carbon-13splitting give rise to contour patterns that are symmetrical around 0 Hz. This phenomenon was used to help assign metabolite structures and to distinguish endogenous from exogenous signals. The intense signals from methyl carbons (56 to 59 ppm), not in the control spectrum, produce symmetrical contour patterns due to the absence of 13C coupling and indicate the presence of methyl ether groups (OCH3). The signal at 62.5 ppm (not present in the control urine) also produces a symmetrical contour pattern. This signal (assigned below) could be derived from an endogenous compound that is excreted in a higher concentration after administration of 2-ME, Metabolite Structures: 1, HOZCCH~NH~; 2, HOzCCHzOH; 3, or it could be derived from an exogenous metabolite that HOCHzCHzOH; 4, H3COCHzCHzOR 5, H3COCHzCOzW; 6, HaCOCHzCONHR 7, HsCOCH2COzH; 8, H3COCH(C02H)R 9, either retains only one of the 2-ME carbons or does not HsCOCH(C02H)R 10, H 3 C O C H 4 H R 11,H~COCHZCH~OSO~-. exhibit carbon-13 splitting. The signal at 63.2 ppm that Structures (1-11)are proposed from the connectivity, multiplicity, produces a symmetrical pattern is also present in the and chemical shift data. Assignments were based on this data and control spectrum, indicating its derivation from an encomparison with calculated and chemical shift values for biochemically feasible metabolites. b Carbon-carbon coupling constants were dogenous compound. Two signals (labeled lb, Figure 1C) measured from the 1D spectrum. Carbon connectivities were in the noise level of the spectrum near 41 ppm produce obtained using INADEQUATE spectroscopy. The carbon atom is nonsymmetrical contour patterns in the HET2DJ specassigned a number corresponding to the metabolite and a letter that trum (data not shown), indicating their derivation from designates its derivation from methoxyethanol (a, b, or c correspondthe labeled carbons of 2-ME. ingtothe 1,2,ormethoxycarboneofmethoxyethanol).d Thenumber of hydrogens attached to each carbon was determined using HET2DJ Assignment of Metabolite Structures. Since the spectroscopy. e Values of shiftwere obtained through the Carbon-13 NMR data describe above are obtained for the portion of NMFt Database. f Signalsfor metabolite 5 overlap more intense signals each metabolite derived from the labeled carbons of 2from metabolites 6 and 7 and were observed only in an HPLC isolate. ME, the remainder of each structure was assigned by comparing the experimental shifts with calculated values termination of connectivity for carbon signals that arise of shift for mechanistically feasible metabolites or with from each metabolite of 2-ME. The INADEQUATE shifts obtained for synthetic standards. The rationale spectrum of a concentrated sample of mouse urine (see behind the assignment of each metabolite structure is Methods) is shown in Figure 2. The 1D 13CNMR spectrum presented below, where metabolites are grouped in catof the concentrated urine (plotted on the F2 axis) is nearly egories related to mechanism of formation and amount identical to the spectrum obtained before concentration, excreted. Although the 13CNMR spectra acquired in this indicating that metabolite structures were not altered. In study were not quantitative, approximate relative quanthe INADEQUATE spectrum, Signals arising from adjatities can be obtained by comparing signal intensities. The cent carbons are indicated by contour peaks with the same numbers and letters assigned to metabolites designate the position on the F1 axis. For example, the alignment of compound number and the letter of the carbon derived contours at 85 ppm (carbon 8b) and 175 ppm (carbon 8a) from 2-ME (a, b, or c for the 1, 2 or methoxy carbons, shows that metabolite 8 contains two adjacent carbons respectively). The proposed metabolism of 2-ME is shown derived from 2-ME, where the structural environment for in Scheme I. the two carbons gives rise to signals at these particular (A) 2-Methoxyacetic acid and N-(Methoxyacety1)shift positions. The INADEQUATE spectrum in Figure glycine. The two metabolites in highest concentration 2 was plotted to show specificallythe connectivityof signals (metabolites 6 and 7) give rise to the most intense signals that have low intensities, resulting in lines of noise for the centered at 58, 59, 71, 172, and 178 ppm (Figure 1B). intense signals (71, 172, and 178 ppm) for carbons of Chemical shift values near 60 and 175 ppm are indicative metabolites 6 and 7. The INADEQUATE connectivities of methyl ether (OCH3) and carbonyl (CO) carbons, for signals derived from the labeled carbons of 2-ME are respectively. The signals at 58 and 59 ppm give rise to summarized in Table I. quartets in the HET2DJ spectrum, confirming the presThe number of hydrogens attached to each labeled ence of methyl groups. HET2DJ spectroscopyshows that the peaks near 71 ppm arise from CH2 carbons, and the carbon was determined by HET2DJ spectroscopy. An Table I. Chemical Shifts, CarbonCarbon Coupling Constants, CarbonCarbon Correlations,and Proton Multiplicities for Signals Derived from 2-Methoxyethanol in the Urine of Mice Administered PO [lf~aetboxy-W]-2-Met hosyethanola chemical carbon proton shift coupling connectivity multiplicity carbon-13 (Hz)* (1NADEQUATE)c (HET2DJ)d databasee (ppm) lb CHz 43 42.68 54 CH3 57 8c 56.61 s 7c CH3 59 57.91 s CH3 59 6c 58.90 s CHZ 61 2b 61.12 54 CHz 64 3a,b 62.48 s 69 lla,b 67.26 41,3 4a 68 67.48 42,4 5b 70 69.55f 38,3 6b CHz 70 70.65 53 CHz 69 7b 70.94 55 CHz 72 75.52 42 4b CH 81 8b 84.79 55 85.80 55 9b 80 9b 80 86.10 54 10a 107 96.00 83,5 10a 107 97.91 81,4 149.01 81 10b 143 149.89 83, 2 10b 143 172.14 54,2 6a 168 174.89 51 8a 171 177.66 56,2 7a 174 179.48 55 2a 177
556 Chem. Res. Toxicol., Vol. 5, No.4,1992
Sumner et al.
F1 IH 2000 3000 4000 5000 6000
7000 8000
9000 10000 11000 12000 13000 160
170
160
150
140
130
120
110 F2
100
93
80
70
60
50
40
IPPM)
Figure 2. INADEQUATE spectrum on mouse urine collected for 24 h after administration of 250 mg/kg of [1,2,nethoxy-l3C]-2-ME. The sample was concentrated as described in the Materials and Methods section. The spectrum was plotted to show carbon-carbon connectivities for compounds with low signal intensities and therefore contains three lines of noise for the intense signals from carbons in metabolites 6 and 7. The correlation of signals for each metabolite is determined by tracing the connectivities for signals at different chemical shifts (along the x-axis) which have the same double quantum frequency along the y-axis.
200
150
100
50
0 F l
-50 IHZI
-100
-150
-200
-2 0
Figure 3. Expanded region of the HET2DJ spectrum of mouse urine collected for 24 h after PO administration of 250 mg/kg of
[1,2,methoxy-W]-2-ME. The urine was concentrated as described in the Materials and Methods section. The number of hydrogens attached to a carbon is equal to the number of contours at the carbon’s chemical shift minus one.
INADEQUATE data show direct connectivity between the CH2 and CO carbon signals. These data indicate H3COCH2CORstructures for both metabolites 6 and 7. Given the previous identifications of 2-methoxyacetic acid (2MAA) and N-(methoxyacety1)glycine as major metabolites of 2-ME (I, 10-121, the R groups are assigned to OH (metabolite 7) and NHCH2C02H (metabolite 61, respectively. Calculated values of shift (Table I) for 2-MAA and N-(methoxyacety1)glycine are consistent with the experimental shifts obtained directly in the urine. (B)Metabolites with HaCOCHRC02H Structures. Metabolite 8 has three signals derived from the labeled carbons of 2-ME with chemical shifts near 57,85, and 175 ppm. The peaks at 57 and 175 ppm are indicative of OCH3 and CO carbons, respectively. The INADEQUATE spectrum shows connectivity between the signals at 85 and 175 ppm, and HET2DJ spectroscopy shows CH and CH3
carbonsfor the peaks at 85 and 57 ppm, respectively. These data suggest a H3COCHRC02H structure for metabolite 8. Several other signals are present near 85 ppm (e.g., 9b in Figure 1C)with less than one-tenth the intensity of the CH signal from metabolite 8. These signals indicate the presence of additional metabolites that have structural environmentsfor the 2-ME-derived carbonssimilar to that of metabolite 8 and occur at less than one-tenth the concentration of metabolite 8. Metabolites that contain all three carbons from 2-ME and have structures consistent with the NMR data (H3COCHRCOzH) can be derived from incorporation of 2methoxyacetyl-CoA into the Krebs cycle (Scheme 11). The first product expected from entry into the Krebs cycle is 2-hydroxy-3-methoxy-1,2,3-propanetricarboxylic acid (methoxycitrate, I). Here, carbon b of 2-methoxyacetyl-coA is converted to a methine carbon (CH), and
Chem. Res. Toxicol., Vol. 5, No.4,1992 557
Urinary Metabolites of [1J2Jmethoxy-~3C]-2-Methoxyethanol
Scheme I. Proposed Metabolism Scheme for 2-Methoxyethanol in Pregnant CD-1 Mice H&O-CHz-CHz-OH * H&-O-CH&O+i meltmxyelhanol
H&O-CHp-CI++R
methoxyaceticacid 7
H,C-O-CHz-CHz-O-GI~c methoxyethyl P-Dglucuronide 4
H&-O-CHz-CO-Gly methoxy-N-acetyl-
H3CQCHz-CO-Gluc methoxyacetyl B-D-glucuronide 5 Methoxyacetyl-CoA
HOCHZ-CHz-OH ethylene glycol
H,C-O-CHZ-CH~-OSO,' methoxyethylsulfate
3
,
11
COzH I
H3C-O-CH-C-CH2-COzH I
Methoxycitrate (Krebs cycle)
t
yJ2.
I
F-O FH2 cop-
(
CHIO-LH.\
co, I
FH q -co,
CO,
I
HF-OH
'0,C-CH CH,O-CH
co, CH CH30-C
11
111
I I
co,
VI1
I
co,
GTP
',
methoxybutenoic acid (fatty acid synthesis) 10
methoxyoxabacetate (Krebs cycle)
Scheme 11. Proposed Scheme for the Incorporation of 2-Methoxyacetyl-CoA into the Krebs Cycle, Producing Structures Consistent with Metabolites Detected in the Urine of Mice Administered 2-ME SCoA
,
C'OZH H3C-O-CH-CO-COzH
HZN-CHZ-COpH glycine 1
c.0
H3C-O-CH-CH-CH~-COzH
I
H02C OH
HOCH2-COzH glycolic acid 2
K
GDP
+ PI
carbon a is converted to a carboxyl carbon (C02H), consistent with the structure derived from the NMR data. The calculated value of shift (Table I) for the CH carbon in methoxycitrate and the shift detected for a CH carbon in urine are in good agreement. Additionally, a synthetic standard of triethyl methoxycitrate (Table 11)gives a shift for the CH carbon nearly identical to the CH carbon detected in mouse urine. Further metabolism of methoxycitrate in the Krebs cycle (Scheme 11)may result in the formation of l-hydroxy-3methoxy-l,2,3-propanetricarboxylicacid (III),2-hydroxy-
3-methoxybutanedioic acid (VIII), and 2-methoxy-3oxobutanedioic acid (IX), which would be analogous to isocitrate, malate, and oxalacetate, respectively. These potential Krebs cycle metabolites have structural features and calculated values of shift (Table 111) that are also consistent with the NMR data. Carbon-13 NMR spectra for molecules having two chiral centers generally contain two signals for each carbon, while the spectra for molecules with only one chiral center show one signal for each carbon. In the urine spectrum (Figure 10,the most intense doublet at 85 ppm arises from a single carbon. Since methoxyoxalacetate (1x1is the only potential metabolite from the Krebs cycle that has a single chiral center, it is likely that the most intense doublet at 85 ppm is derived from this metabolite. The synthetic standard of triethyl methoxycitrate has two signals (at similar chemical shifts) for each carbon, which are assigned to the diastereomers resulting from the two chiral centers in the molecule (Table 11). Four of the small signals near 85 ppm (Figure lC, labeled 9b) may be attributed to two doublets, arising from the CH carbon of the methoxycitrate diastereomers. The other twoKrebs cycle metabolites, 1-hydroxy-3-methoxy-1,2,3-propanetricarboxylicacid (111)and 2-hydroxy3-methoxybutanedioicacid (VIII), may also be present in the urine and may give rise to other small signals near 85 PPm. (C)Glucuronides. Metabolite 4 has signals connected in the INADEQUATE spectrum between 67 and 76 ppm, which are defined as CH2 carbons in the HET2DJ spectrum. The peaks at 67 ppm have a long-range coupling of 3 Hz, indicating the presence of the OCH3 group. The structure H~COCHZCH~OR is consistent with this data. Calculated values of shift (Table I) for the CH2 carbons in methoxyethyl p-D-glucuronide are consistent with the experimental shifts detected in urine. This metabolite could be formed by conjugation of the parent compound with glucuronic acid. Metabolite 5 has signals near 70 ppm which overlap with the intense signals detected for the 2-MAA and its glycine conjugate (metabolites 6 and 7). The signals for metabolite 5 are resolved in the I3C NMR spectrum obtained on a fraction collected after chromatographic separation of the urine (12). Calculated values of shift for
558 Chem. Res. Toxicol., Vol. 5, No. 4, 1992
Sumner et al.
Table 11. Carbon Chemical Shifts of Standard Compounds 72.84 (CHzOR) 59.92 (CHzOH) 57.73 (CH3) methoxyethanola 41 Hz 41,3 Hz 3 HZ glycolic acid" 179.64 (C02H) 61.27 (CHzOH) [1,2-W]ethyleneglycolb 62.53 (CH20H) [1,2,methoxy-W]-2-
singlet
[2-W]glycineb
41.43 (CH2)
singlet triethyl methoxycitratec 83.87 (CH) 83.57
methoxybutenoic acid methyl ester
39.53 (CHz) 39.46
97.59 (CHOR) 149.44 (CH) 94.59 148.00
14.08, 13.96 (CHI) 76.79 (COH) 171.47,169.86 (COS) 59.20 (OCHs) 62.16,62.12 (OCHz) 13.91, 13.84 76.44 171.27,169.37 58.85 61.33,61.15 168.22, 168.08 60.65,60.62 33.07 (CH2) 172.66 (COS) 59.52,55.72 (OCH3) 29.33 172.39 51.64, 51.57
0 Samples were dissolved in DzO and referenced using the same values (RFL and RFP on Varian VXR-300)that place urea at 162.45 ppm in the spectra of urine samples. b A sample of [Wlethylene glycol or [2-l3C1glycinewas placed in the control urine and referenced to urea at 162.45 ppm. c Samples were dissolved in CDCl3 and referenced at 77 ppm.
Table 111. Calculated Values of Shift for Carbons Derived from 2-Methoxuethanol in Krebs Cycle Metabolites. Krebs cycle structures calculated shift values ~
I I1 I11 IV VI VI1 VI11 IX 01
2-hydroxy-3-methoxy-1,2,3-propanetricarboxylic acid 3-methoxy-l,2,3-propenetricarboxylic acid l-hydroxy-3-methoxy-1,2,3-propanetricarboxylic acid l-oxo-3-methoxy-1,3-propanedicarbosylic acid 2-methoxybuGedioic acid 2-methosy-2-butenedioic acid 2-hydroxy-3-methoxybutanedioic acid 2-methosy-3-oxobutanedioicacid
80 (CH) 69 (CH) 80 (CH) 69 (CH) 69 (CH) 155 (C) 80 (CH) 81 (CH)
~
~~~~
173 (CO) 172 (CO) 171 (CO) 175 (CO) 175 (CO) 163 (CO) 169 (CO) 171 (CO)
52 (OCH3) 57 (OCH3) 52 (OCHa) 59 (OCH;) 59 (OCH3) 59 (OCH3) 52 (OCH3) 57 (OCH3)
Values of shift were calculated using the 13C NMR Database.
methoxyacetyl j3-D-glucuronideare consistent with shifts for one of the metabolites in this fraction. This metabolite could be formed by glucuronic acid conjugation with 2methoxyacetic acid. Glucuronide conjugates have been identified in the urine of rats administered 2-butoxyethan01 (13). (D) Sulfate Conjugate. An additional set of signals (metabolite 11) arise at 67 ppm, nearly overlapping one of the CH2 signals assigned to methoxyethyl j3-D-glucuronide (metabolite 4). The signals for metabolite 11and for metabolite 4 have nearly the same intensity, showing similar concentrations. The 13CNMR spectrum for urine collected from CD-1 mice for 8 h after administration of [1,2,methoxy-l3C1-2-MEtogether with serine contains the signal for metabolite 11 a t approximately 10 times the intensity of the signals assigned to metabolite 4.4 In contrast, the 13C NMR spectrum for urine collected between 8 and 24 h contains the signals for metabolite 4 at approximately 10 times the intensity of the signal for metabolite 11. This data confirms that these signals are derived from two independent metabolites and shows a time dependence of elimination. Calculations for 2-methoxyethyl sulfate (H~COCH~CHZOSO~-) give shift values (Table I) near 69 ppm for both CH2 carbons, indicating that metabolite 11could be derived from 2-MEconjugation with sulfate. Sulfate conjugates have been identified in the urine of rats administered propylene glycol monomethyl ether (11). (E) Metabolites Containing Vinyl Carbons. The INADEQUATE spectrum shows connectivity between signals at 97 and 149 ppm, indicating the presence of a metabolite with two vinyl carbons (C-C) derived from the labeled 2-ME. Two sets of doublets (ratio 2:l) appear at both 97 and 149ppm, suggestingcidtrans isomers. Longrange coupling to the vinyl carbons shows the presence of the OCH3 group. These data reveal a H&OC=CR structure. A metabolite may be formed after incorporation 4
~
Sumner et al., unpublished results, 1991.
Scheme 111. Proposed Scheme for the Incorporation of 2-Methoxyacetyl-CoA in the Synthesis of Fatty Acids, Producing a Structure Consistent with a Metabolite Detected in the Urine of Mice Administered 2-ME H&-O-CH,GO-SCoA HS-ACP
lic
CoA-SH
H&-O-CH&O-SACP HOpC-CH&O-SACP
I C i=
HS-ACP + C02
H,C-O-CHp-CO-CH,CO.SACP NADPH NADP+
+
H+
HSC-O-CH~-CH(OH)-CH,.CO-SACP
h,
H20
HsC-O-CH-CH-CH2CO-SACP
L
HS-ACP
H&-O-CH-CH-CH&02H Cmethoxy-3-butenobacid
of 2-methoxyacetyl-CoA in the synthesis of fatty acids (Scheme 111), where the a and b carbons of 2-ME are converted to vinyl carbons eventually forming 4-methoxy3-butenoic acid. Chemical shift values for a synthetic standard of 4-methoxy-3-butenoicacid methyl ester (Table 11)and calculated values of shift for 4-methoxy-3-butenoic acid (Table I) are consistent with those detected in urine. (F) Ethylene Glycol. A singlet at 62.5 ppm (assigned as a CHZcarbon in the HET2DJ spectrum) in the spectrum of urine collected after administration of [1,2,methoryl3C1-2-ME(Figure lB, metabolite 3) is not present in the
Urinary Metabolites of [1,2,methoxy-*~C]-2-Methoxyethanol
Chem. Res. Toxicol., Vol. 5,No.4,1992 559
each metabolite, it does allow for extensive investigation of important structural aspects of each metabolite. Once metabolites and pathways of formation are identified with such an approach, further characterization can be made using other analytical methods. This approach has been used for the analysis of metabolites produced from [1,2,313Clacrylonitrile(16)and [ 1,2,3-l3C1acrylamide and the results have been compared with previously described metabolites or synthetic standards for new metabolites. The primary objective of the present study was to provide a more complete characterization of 2-ME metabolism, with a specific emphasis on determining the fate of 2-MAA. The ability to characterize metabolites directly in the urine using NMR spectroscopy enabled the elucidation of structures for many 2-ME metabolites. Metabolites were assigned without using chromatographic techniques that can result in the generation of structural artifacts, and assignmentswere basedon specificstructural data rather than trial and error with synthetic standards. Carbon-13 NMR also provides the resolution needed to uniquely define metabolites at high and low concentrations, whereas chromatographic techniques are often unable to assign metabolites associated with small percentages of the total radioactivity administered. Although quantitative 13C NMR spectra were not acquired in this study, the relative intensities of signals in the reported (14). (G) Excretion of Unchanged 2-ME.The possible spectra were used to estimate the percent contribution of each metabolite to the total excreted. Metabolites were excretion in mouse urine of unchanged 2-ME was examined by adding [1,2,methoxy-W]-2-ME to control urine (see identified that account for less than 1%of the total Methods) and comparing the signal positions of 2-MEmetabolites excreted in the urine. derived carbons with signals detected after administration The most intense signals (indicating metabolites in of [1,2,nethoxy-13C]-2-ME. The most intense signals (at highest concentrations) are assigned to 2-MAA and N58,60, and 73 ppm) in the spectrum of control urine with (methosyacetyl)glycine, which have been previously de[W]-2-ME addition corresponded to the parent comtected in urine following administration of 2-ME a t 250 pound. Less intense signals derived from the labeled mg/kgip torats (1)and250mg/kgpo tomice (10,121.The carbons of 2-ME were also present with chemical shifts glycine conjugates was not identified in the urine of rats (at 64 and 71 ppm) consistent with a H~COCH~CHZOR that ingested 2-ME (180 to 1620 ppm) for 24 h via the structure, indicating 2-ME reaction with a urinary comdrinking water (13) or were administered a low oral dose pound. Signalsfor both the parent compound and possible (76 mg/kg) of 2-ME (111, and it was suggested that the 2-ME conjugation producta are present dust above the conversion of 2-ME to 2-methoxyacetyl- CoA and further noise level) in the spectrum obtained after administration conjugation with glycine are route- and/or dose-related. of 250 mg/kg of [1,2,methoxy-l3C]-2-ME to CD-1 mice Preliminary results4 show that rats excrete N-(methoxy(Figure 1C). These data suggest that a small amount of acety1)glycine following an oral dose of 25 or 250 mg/kg 2-ME accumulates in the 24-h urine and that a portion of of [1,2,methoxy-W]-2-ME, where the ratio of 2-MAA to the excreted 2-ME reacts with a normal urinary compound. the glycine conjugate is much higher in rat urine compared The parent compound is excreted in the 24-h urine of rats with mouse urine. administered 250 mg/kg of 2-ME at much higher levels Ethylene glycol, glycolic acid, and glycine were excreted than detected for mice.4 as metabolites of 2-ME in the urine. Ethylene glycol has been previously detected as a major metabolite in rats Discussion administered low doses of glycol ethers, and it was suggested that low-dose-exposure situations favor the Carbon-13 NMR was used to detect and assign metabformation of ethylene glycol over alkoxyacetic acids (13). olites directly in the urine collected from pregnant CD-1 The detection of glycolic acid and glycine in the present mice that received an oral teratogenic dose of 250 mg/kg study indicates further metabolism of ethylene glycol. Rats of [1,2,nethoxy-l3C]-2-ME. One- and two-dimensional administered [14C]ethylene glycol excrete glycolic acid in NMR experiments were used to characterize the carbontheir urine and exhale 14C02(14,15). Further metabolism carbon correlations and proton multiplicity for the 13Cof ethylene glycol in CD-1 mice administered [l4C1-2-ME labeled carbons in each metabolite. The structural data may in part account for the W02that is expired following for the 13C-labeled portion of each metabolite were administration of [l4C1-2-ME (10). compared to metabolites derived from mechanistically Signals were also detected that are consistent with feasible pathways of metabolism. Structures were assigned products formed from 2-ME and 2-MAA conjugation with using this information and by comparing chemical shifts glucuronic acid. Glucuronide conjugates have been sugwith previously assigned metabolites and with calculated gested as metabolites formed after exposure to rats of low values of shift or shifts obtained for synthetic standards doses of butoxyethanol (13). Additionally, signals that of potential compounds. While this method does not may be attributed to the excretion of 2-ME conjugation enable direct identification of the unlabeled portion of
control spectrum. This shift value is consistent with that obtained for a standard of [l3C]ethylene glycol in control urine (Table 11). In addition, the synthetic standard of [Wlethylene glycol gives rise to a singlet (as seen for the peak at 62.5 ppm in the urine samples), indicating the absence of carbon-13coupling due to symmetrical carbons. The detection of ethylene glycol in mice is consistent with the previous identification of ethylene glycol in the urine of rats administered glycol ethers (13). Metabolite 2 has a CH2 carbon at 61 ppm which has INADEQUATE connectivity to a CO carbon at 180 ppm, indicating a RCH2COR' structure. Calculated values of shift (Table I) and experimental values for glycolic acid (R = R' = OH) in D2O (Table 11)are consistent with those detected in the urine. Glycolic acid can be formed from further metabolism of ethylene glycol and has been detected in rats administered ethylene glycol (15). Metabolite 1has a doublet positioned at 41 ppm, close to the noise level of the spectrum, that is assigned to a CH2 carbon using HET2DJ spectroscopy (see above for details). Chemical shift values for a standard of glycine in control urine (Table 11) and calculated values of shift (Table I) are consistent with the signal detected in the urine. Glycine has been assigned as a urinary metabolite of ethylene glycol in rats coadministered sodium benzoate
(In,
560 Chem. Res. Toxicol., Vol. 5, No. 4, 1992
with sulfate were detected, consistent with the identification of sulfate conjugates in the urine of rats administered propylene glycol monomethyl ether (11). Several signals present in the spectra of urine have been assigned to metabolites produced after incorporation of 2-methoxyacetyl-CoA into the Krebs cycle. The inhibition of 14C02evolution by fluoroacetate or sodium acetate from mouse embryos exposed to 11- or 2-l4C1-2-MAA in whole embryo culture suggested that 2-methoxyacetyl-coA entered the Krebs cycle (12). The metabolites assigned in this study have retained all three 13Ccarbons from the administered [ 1,2,methoxy-l3C1-2-ME. Since these metabolites contain all three carbons from the 13Clabeled 2-ME, they cannot account for the exhaled 14C02 measured in analogous experiments using [l4C1-2-ME. Thus, either two-carbon metabolites are produced after the incorporation of 2-methoxyacetyl- CoA into the Krebs cycle that are not assigned in this study or another route of disposition is responsible for the formation of 14C02.A metabolite containing two vinyl carbons derived from the labeled 2-ME was also detected in the urine. 4-Methoxy3-butenoic acid has a structure consistent with the NMR data and may be formed after incorporation of 2methoxyacetyl-CoA in the synthesis of fatty acids. This type of incorporation has not been previously suggested for glycol ethers. Little is known about the biochemical mechanism that induces the embryotoxicity of 2-ME and the possible involvement of metabolites of 2-MAA. A number of compounds such as acetate, serine, formate, glycine, or D-glucose ameliorate the adverse effects of 2-ME in rats and mice (18, 24). It has therefore been suggested that 2-ME may alter one-carbon metabolism and impair purine biosynthesis. However,these attenuatingagents may enhance repair of, rather than prevent, cell damage. The role of the new metabolites described here in the adverse effects of 2-ME is currently unknown. These agents may impair normal cellular metabolism by acting as inhibitors of enzymes involved in the Krebs cycle, fatty acid synthesis, and/or oxidation. The alkene metabolite detected in this study could induce toxic effects by direct interaction or from biotransformation to electrophilic agents that are capable of reacting with cellular informational macromolecules. Its structural analogy to the proximate toxicant (2-n-propyl-4-pentenoic acid) of valproic acid (an anticonvulsant that induces toxic and teratogenic effects) (25-27) suggests that further metabolism of 2-MAA may contribute to the adverse effects of 2-ME. The assignment of metabolites which appear to be produced after incorporation of 2-methoxyacetyl- CoA into pathways of intermediary metabolism may have important implications in understanding the adverse effects of 2-ME. We are currently investigating the metabolism of 2-MEin pregnant CD-1 mice administered toxic, nontoxic, and attenuated doses of the chemical in an attempt to correlate levels of metabolites from particular pathways with the degree of teratogenic effects.4
References Moss, E. J., Thomas, L. V., Cook, M. W., Walters, D. G., Foster, P. M.D.,Creasy,D.M.,andGray,T.J.B. (1985)Theroleofmetabolism in 2-methoxyethanol-inducedtesticular toxicity. Toxicol. Appl. Pharmacol. 79, 480-489. Gray, T. J. B., Moss, E. J., Creasy, D. M., and Gangolli, S. D. (1985) Studies on the toxicity of some glycol ethers and alkoxyacetic acids in primary testicular cell cultures. Toxicol. Appl. Pharmacol. 79, 490-501.
Sumner et al. Nagano, K., Nakayama, E., Oobayashi, H., Yamada,T., Adachi, H., Nishizawa, T., Ozawa, H., Nakaichi, M., Okuda, H., Minami, K., and Yamazaki, K. (1981) Embryotoxic effects of ethylene glycol monomethyl ether in mice. Toxicology 20, 336-343. Hanely, T. R., Jr., Young, J. T., John, J. A., and Rao, K. S. (1984) Ethylene glycol monomethyl ether (EGME) and propylene glycol monomethyl ether (PGME): inhalation fertility and teratogenicity studies in rata, mice, and rabbits. Enuiron.Health Perspect. 57, 7-12. Horton, V. L., Sleet, R. B., John-Greene,J. A., and Welsch, F. (1985) Developmental phase-specific and dose-related teratogenic effects of ethylene glycol monomethyl ether in CD-1 mice. Toxicol. Appl. Pharmacol. 80,108-118. Ritter, E. J., Scott, W. J., Randall, J. L., and Ritter, J. M. (1985) Teratogenicity of dimethoxyethyl phthalate and ita metabolites methoxyethanol and methoxyacetic acid in the rat. Teratology 32, 25-31. Scott, W. J., Fradkin, R., Wittfoht, W., and Nau, H. (1989) Teratologic potential of 2-methoxyethanoland transplacental distribution of ita metabolite, 2-methoxyacetic acid, in non-human primates. Teratology 39, 363-373. Brown, N. A., Holt, D., and Webb, M. (1984) The teratogenicity of methoxyacetic acid in the rat. Toxicol. Lett. 22, 93-100. Miller, R. R., Carreon, R. E., Young, J. T., and McKenna, M. J. (1982)Toxicity of methoxyacetic acid in rats. Fundam. Appl. Toxicol. 2, 158-160. Sleet, R. B., Greene, J. A., and Welsch, F. (1988) The relationship of embryotoxicityto disposition of 2-methoxyethanol. Toxicol. Appl. Pharmacol. 93, 195-207. Miller, R. R., Hermann, E. A., Langvardt, P. W., McKenna, M. J., and Schwetz, B. A. (1983) Comparativemetabolism and disposition of ethylene glycol monomethyl ether and propylene glycol monomethyl ether in male rats. Toxicol. Appl. Pharmacol. 67,229-237. Mebus, C. A., Clarke, D. O., Stedman, D. B., and Welech, F. (1992) 2-Methoxyethanol metabolism in the pregnant CD-1 mouse and embryos. Toxicol. Appl. Pharmacol. 112,87-94. Medinsky, M. A., Singh, G., Bechthold, W. E., Bond, J. A., Sabourin,P. J.,Birnbaum,L. S.,andHendereon,R. F. (1990)Disposition of three glycol ethers administered in drinking water to male F344/ N rata. Toxicol. Appl. Pharmacol. 102,443-455. Gessner, P. K., Parke, D. V., and Williams, R. T. (1961) The metabolism of 14Clabeled ethylene glycol. Biochem. J. 79,482-489. Marshall, T. C. (1982) Dose-dependent disposition of ethylene glycol in the rat after intravenous administration. Toxicol.Enuiron.Health 10,397-409. Fennell, T. R., Kedderis, G. L., and Sumner, S. C. J. (1991)Urinary metabolites of [1,2,3-W]acrylonitrile in rata and mice detected by l3C nuclear magnetic resonance spectroscopy. Chem. Res. Toxicol. 4, 678-687. Sumner, S. C. J., McNeela, J. P., and Fennell, T. R. (1992) Characterization and quantitation of the urinary metabolites of [1,2,3-13Clacrylamide detected by l3C nuclear magnetic resonance spectroscopy. Chem. Res. Toxicol. 5, 81-89. Mebus, C. A,, and Welech, F. (1989) The possible role of one-carbon moieties in 2-methoxyethanol and 2-methoxyacetic acid-induced developmental toxicity. Toxicol. Appl. Pharmacol. 99, 98-109. Bas, A., Freeman, R., and Frenkiel, T. A. (1981) An NMRtechnique for tracing out the carbon skeleton of an organic molecule. J.Am. 103, 2102-2104. Chem. SOC. Muller, L., Kumar, A,, and Emst, R. R. (1975) Two-dimensional carbon-13 NMR spectroscopy. J. Chem. Phys. 63, 5490-5491. Wehrli, F. W., and Wirthlin, T. (1976) Interpretation of carbon-13 NMR spectra, Heyden & Son, New York. Breitmaier,E.,and Voelter,W. (1987) Carbon-13NMRspectroscopy: high resolution methodsand applicationsin organic chemistryand biochemistry, VCH, New York. Nicholson, J. K., and Wilson, I. H. (1987) High resolution nuclear magnetic resonance spectroscopy of biological samples as an aid to drug development. Prog. Drug Res. 31,427-479. Mebus, C. A., Welsch, F., and Working, P. K. (1989) Attenuation of 2-methoxyethanol-induced testicular toxicity in the rat by simple physiological compounds. Toxicol. Appl. Phurmacol. 99,110-121. Granneman, G. R., Wang, S. I., Machinist, J. M., and Kesteraon, J. W. (1984)Aspects of the metabolism of valproic acid. Xenobiotica 14, 357-387. Nau, H., and Lijscher, W. (1986) Pharmacological evaluation of various metabolites and analogues of valproic acidteratogenic potencies in mice. Fundam. Appl. Pharmacol. 6, 669-676. Rettenmeier,A. W., Prickett, S.,Gordon, W. P., Bjorge, S. M., Chang, S.-L., Levy, R. H. and Baillie, T. A. (1985) Studies on the biotransformation in the perfused rat liver of 2-n-propyl-4-pentenoic acid, a metabolite of the antiepileptic drug valproic acid. Drug Metab. Dispos. 13, 81-96.
