Energetics of the metabolic production of (E,E)-muconaldehdye from

Arthur Greenberg, Charles W. Bock, Philip George, and Jenny P. Glusker. Chem. Res. Toxicol. , 1993, 6 (5), pp 701–710. DOI: 10.1021/tx00035a016...
0 downloads 0 Views 1MB Size
Chem. Res. Toxicol. 1993,6, 701-710

701

Energetics of the Metabolic Production of (&E) -Muconaldehyde from Benzene via the Intermediates 2,3-Epoxyoxepin and (2,Z) - and (E,Z) -Muconaldehyde: Ab Initio Molecular Orbital Calculations Arthur Greenberg,+ Charles W. Bock,$ Philip George,*'$ and J e n n y P. Gluskerg Department of Environmental Sciences, Cook College, Rutgers University, New Brunswick, New Jersey 08903,Department of Chemistry, Philadelphia College of Textiles and Science, Philadelphia, Pennsylvania 19061, and The Institute for Cancer Research, The Fox Chase Cancer Center, Philadelphia, Pennsylvania 19111 Received May 4,1993@ T o assess the feasibility of the pathway set out in the title, we have carried out ab initio molecular orbital calculations of the energies of cisoid- and transoid-2,3-epoxyoxepinand the eZzZz-, eZeZz-, eEeZz-, eEeEz-, and eEeEe-conformers of muconaldehyde at the MP2/631G* (frozen core, valence orbitals active) level with full geometry optimization using the splitvalence RHF/6-31G* basis set. Including thermal energies, derived from vibrational frequencies obtained at the RHF/6-31G*//RHF/6-31G* level, reaction energies (that would correspond to gas phase data at 298 K) have been evaluated. The muconaldehyde conformers are lower in energy than the epoxyoxepin, i.e., are formed exothermically, and in accord with experiment, the Z,Z-conformers are less stable than the E,& and the E,Z less stable than the E,E. In addition, we have characterized the transition state for the fission of the C-0 bond in the epoxy ring and the C-0 bond in the oxepin ring of transoid-2,3-epoxyoxepin,thereby generating (eZzZz)-muconaldehyde. With an activation energy of 16.5 kcal mol-' the half-life of the epoxyoxepin is very short a t the temperatures employed in the experiments, less than 1min, which explains why it has not been detected. T o gain some insight into the isomerization about the carbon-carbon double bonds in the muconaldehyde conformers, we have examined the possible as intermediates. In involvement of 2-formyl-2H-pyran and cyclobutene-3,4-dicarboxaldehyde model reactions simulating the microsomal monooxygenase system, the formation of benzene oxide and of the epoxyoxepin are found to be very favorable exothermic processes.

Introduction The elucidation of the pathways of benzene metabolism and their relation to benzene toxicity is the subject of extensive research (1-11). (E,E)-Muconic acid, 1 in Scheme I, has long been recognized as one of the metabolic products (l),and ita presence in urine is widely used as a marker of benzene exposure (12). The corresponding aldehyde, (E,E)-muconaldehyde (2a), has been identified as a metabolite by Goldstein, Witz, and co-workers following incubation of benzene with mouse liver microsomes (5,7),and shown to be hematotoxic (5,6). This aldehyde, a logical precursor of the diacid, was subsequently shown to be metabolized to the partial oxidation product OHCCH=CHCH=CHCOOH (9). A possible pathway of benzene metabolism involving the benzene oxide-oxepin tautomeric pair (13) as intermediates, 5/4 in Scheme I, was suggested by Davies and Whitham (4). An equilibrium mixture of the oxide and oxepin was found to react readily with peroxy acids to yield (Z,Z)-muconaldehyde (2c), which was thermally unstable and slowly isomerized, first to an E,Z-isomer, 2b, and finally to the E,E-isomer, 2a. Comparative studies on indane 3a77a-oxideand 2,'7-dimethyloxepinled to the To whom correspondence should be addressed. Rutgers University. Philadelphia Collegeof Textiles and Science. I The Fox Chase Cancer Center. * Abstract published in Advance ACS Abstracts, September 1,1993. t

t

conclusion that the oxidative ring opening involves the oxepin tautomer. On this basis they proposed that 2,3epoxyoxepin (31, otherwiseknown as 2,Bdioxa[5.1.0locta3,5-diene, was the immediate precursor of (2,Z)-muconaldehyde (2c) via the ring fission process illustrated in Scheme IIa. In this respect 2,3-epoxyoxepin is unique among epoxy-oxepin isomers since neither 4,5-epoxyoxepin (6) norsyn-benzenediepoxide(7) is capable of yielding muconaldehyde. In the former, quite apart from the disposition of the oxygen atoms, the carbon-carbon bond in the epoxide ring is immobilized by the positioning of the double bonds in the oxepin ring (see Scheme IIb). In the latter, the tautomeric shift of the epoxide bonding gives 1,4-dioxocin (141, (81, with the oxygen atoms separated by two (and four) carbon atoms, (see Scheme IIc). The preferential fission of the (23-08 bond in the epoxide ring of 2,3-epoxyoxepin, which is nearer to the C4=C5 double bond than (72-08, is in keeping with calculations on the fission of C-0 epoxide bonds in unsymmetrical epoxides of benzene, naphthalene, and phenanthrene. These calculations showed that fission of the C-0 bond nearest to the shorter carbon-carbon bond in the adjacent ring was in accord with experimentalfindings (15). Further evidence supporting the pathway in Scheme I comes from the work of Mello et al. (10). Dioxiranes are versatile epoxidation reagents which function by oxygen atom transfer, and although benzene does not react with dimethyldioxirane, it was found to react with the stronger

0893-228x/93/2706-0701$04.00/00 1993 American Chemical Society

702 Chem. Res. Toxicol., Vol. 6, No. 5, 1993

Greenberg et al.

Scheme I COOH

I

COOH

CHO

t

CHO

I

t 2b

conformer formed by the fission of the C3-Oe and CZ-07 bonds in the epoxyoxepin, Scheme IIa, is therefore eZzZz. Undoubtedly the corresponding conformer, eZeZz, with the central CrC3 bond in the e configuration, would be more stable because steric hindrance between the terminal CHO groups is far less. We have therefore studied this structure as well, and, taking the eEeZz conformer to be representative of the “E,Z” isomer identified in the experimental investigations (4,191, we have completed the conversion of the epoxyoxepin into the “E,E” isomer with calculations on the eEeEz and eEeEe conformers. In addition, we have characterized the transition state for the formation of (eZzZz)-muconaldehyde from 2,3epoxyoxepin-a critical step in the reaction sequence-and have explored the likelihood that the degenerate Cope rearrangement of 2,3-epoxyoxepin, Scheme IIe, could provide experimental evidence for its participation. To gain some insight into the isomerization about the carboncarbon double bonds in the muconaldehyde isomers, we have examined the possible involvement of 2-formyl-2Hpyran (19) (10) and cyclobutene-3,4-dicarboxaldehyde(20) (11) as intermediates (Scheme IIf,g). Finally, we have investigated energetic aspects of the alternative reaction pathway via dioxetane 9 (Schemes IId and 111) for comparison with the 2,3-epoxyoxepin pathway.

CHO

Computational Methods I

+cHo2c

epoxidizingagent (trifluoromethy1)methyldioxiraneunder neutral conditions, giving muconaldehyde. As yet, however, 2,3-epoxyoxepin has not been observed experimentally. Other pathways which have been advanced to explain the formation of muconaldehyde are depicted in Scheme I11 (2,3, 5, 7). Dioxetane (9) has been suggested as an intermediate in the photooxidation of benzene, in which, among other products, (E,E)-muconaldehyde is formed (5, 7), ring fission occurring as shown in Scheme IId. A possible mechanism involves attack by singlet oxygen (16), as in the cleavage of unsaturated fatty acids in which dialdehydes including muconaldehyde have been isolated (17). But there is no direct evidence for the formation of 9 from benzene or substituted dioxetanes from benzene derivatives (18). To assess the feasibility of the sequence of reactions in Scheme I, we have carried out ab i n i t i o computational molecular orbital calculations of the energies of 2,3epoxyoxepin (3) and the Z,Z-, E,Z-, and E,E-isomers of muconaldehyde, 2c, 2b, and 2a, with full geometry optimization. In addition to the specification of the configurations about the carbon-carbon double bonds in the aldehyde structures, it is necessary to specify the configurations about the carbon-carbon single bonds-for these we shall use the lower-case notations, e and z. The

The calculations were performed at the Advanced Scientific Computing Laboratory, NCI-FCRF, using the GAUSSIAN 90 series of programs (21) on a Cray YMP computer. All structures were optimized using the split-valence RHF/6-31G* basis set (22). The effects of electron correlation were included by performing single-pointMaller-Pleaset perturbation calculations (23-25) a t the MP2/6-31G*//RHF/6-31G* (frozen core, valence orbitals active) level. Vibrational frequencies were calculated a t the RHF/6-31G*//RHF/6-31G* level to determine whether the computed structures correspond to local minima on the potential energy surface or to transition states (26-28)) and to evaluate total thermal energies and entropies a t 298 K. The total molecular energies, MP2/6-31G*//RHF/6-31G*, plus the total thermal energies are listed in Table I, together with the entropies. Unless otherwise stated, the energy changes and activation barriers for the reactions discussed below have been calculated from these values and would thus correspond to experimental gas phase data (29). The thermal energy terms cancel to a large extent, and compared to the differences in total molecular energy, their inclusion alters the reaction energies by less than 2 kcal mol-’ and decreases the activation energies by less than 4 kcal mol-‘. Judging from calculations on the benzene oxide-oxepin valence tautomerism (30), solvent interactions would modify these calculated gas phase values by no more than 3 kcal mol-’. Z-matrix orientations for all the organic structures in Table I are available as supplementary material.

Results 2,3-Epoxyoxepin. Two stable conformers have been identified (see Figure lA,B), which, by analogy with the hydrocarbon structures having CH2 groups in place of the 0 atoms (31),we shall refer to as the “cisoid”, or “open”, structure and the “ t r u m o i d ” , or “closed”, structure, respectively. The parent oxepin (4) likewise has a marked nonplanar structure with the C107C2 plane some 130°to the C1C&&2 plane (32). So the cisoid conformer can be envisaged as the result of an attack by the epoxidizing agent on the same side of the ring as the oxepin oxygen, and the t r u n s o i d conformer as the result of an attack on

Muconuldehyde Production from Benzene

Chem. Res. Toxicol., Vol. 6, No. 5, 1993 703 Scheme I1

cn

zc

10

(B)

Scheme I11

I

11

Table 1. &E, the Sum of the MP2/6-3lG*//RHF/6-31G* Electronic Energy and the Thermal Energy of 298 K,in Hartrees, and Sm,the Entropy at 298 K,in cal mol-' K-1

I

\'s

2c

5

molecule

Em

benzene benzene oxide (5) oxepin (4) cisoid-2,3-epoxyoxepin(3, Figure 1A) transoid-2,3-epoxyoxepin(3, Figure 1B) Cope TS (transoid-2,3-epoxyoxepin) TS [transoid-2,3-epoxyoxepin (e2zZ.z)-muconaldehyde](12, Figure 2) (eZzZz)-muconaldehyde(Figure 1C) (e2eZz)-muconaldehyde (Figure 1D) (eEeZz)-muconaldehyde (Figure 1E) (eEeEz)-muconaldehyde (Figure 1F) (eEeEe)-muconaldehyde (Figure 1G) 2-formyl-W-pyran, equatorial CHO (10) 2-formyl-W-pyran, axial CHO (10) trans-cyclobutens-3,4-dicarboxaldehyde (11) cie-cyclobutene-3,4-dicarboxaldehyde(11) dioxetane (9, Figure 1H)

-231.34474 -306.30076 -306.29575 -381.30901 -381.30715 -381.24666 -381.28088

Sm 63.17 71.91 73.92 77.59 79.24 75.66 77.74

-381.33430 -381.34052 -381.34663 -381.34864 -381.34992 -381.34115 -381.33923 -381.32374 -381.32189 -381.22993 -149.93440 -149.87964 - 76.17014 - 1.13115

87.46 89.54 88.74 88.96 88.71 82.51 86.66 86.47 85.53 80.62 48.85 46.66 44.99 31.06

-

02(%

Oz(lA)

H2O Hz 2b

1

Is

the opposite side of the ring. The cisoid conformer is 1.2 kcal mol-' more stable than the trunsoid. A notable structural feature is the striking dissymmetry in the C-0 bond lengths in both conformers, which may be attributed to the germinal bis alkoxy linkage at CZ. A corresponding dissymmetry has been found in monofluoro oxirane, where the length of the C-0 bond nearer the fluorine is 1.349 A compared to 1.424 A for the farther C-0 bond (33). (eZzZz)-Muconaldehyde. This isomer is 17.0 kcal mol-' lower in energy than 2,3-epoxyoxepin, with the bond lengths shown in the schematic structure (Figure 1C). The ZzZ configuration of C&4C&6 brings CS and CS close

together like C1 and C4 in z(cis)-1,3-butadiene (34). As a consequence, there is severe steric hindrance between the two formyl groups, C1H07 and CzHOs, and the structure is far from planar, with the carbon chain twisted into a small helical segment. The configurations about the single bonds CrC6 and c6-c1 are thus more correctly described as guuche(skew) rather than 2. Transition State: traasoid-2,3- Epoxyoxepin (eZzZz)-Muconaldehyde. A search along the reaction coordinate leading from the trunsoid conformer of 2,3epoxyoxepin to (eZzZz)-muconaldehyde, (Scheme IIa) resulted in the identification of a transition state, 12,having the nonplanar structure depicted in Figure 2, 16.5 kcal mol-1 higher in energy than the epoxyoxepin. No similar transition state was found starting with the cisoid conformer. The incipient fission of C3-06 in the epoxide ring and CZ-07 in the oxepin ring is very evident in Figure 2a.

-

704 Chem. Res. Toxicol., Vol. 6, No. 5, 1993 A

Greenberg et al. B

C

8

E

F

e 1.461

1.483 $ 5

Ha

G

Hb

179.7

-

4 @ 3

e 11.461

(B)transoid- (closed) 2,3-epoxyoxepin; (C, D, E, F, and G) (eZzZz)-, (eZeZz)-, (eEeZz)-, (eEeEz)-,and (eEeEe)-muconaldehyde; and (Ha), dioxetane. (Hb) Schematic cross section of the molecule in the plane of symmetry passing through the midpoints of Cd-Cb,CZ-CI,and 08-07. Figure 1. Bond length diagrams for (A) cisoid- (open) and

a)

1.378

0 1.

I

I

respectively. On the other hand, C2-08and CI-07 decrease in length from 1.371 to 1.274 A, and from 1.353 to 1.281 A, respectively. In the aldehyde these bond lengths are 1.190 and 1.191 A: thus in the transition state there is about 50 5% conversion of the carbon-oxygen single bonds of the epoxyoxepin into the carbonyl double bonds of the aldehyde. However, the cyclic structure of the oxepin ring in the epoxyoxepin is retained to a large extent. The internuclear C1---C2 and CI---C~ distances are respectively 2.485 and 3.084 A in the transition state compared to 2.412 and 3.084 A in the epoxyoxepin, whereas in the aldehyde these distances are significantly larger, 3.535 and 3.579 A, respectively. The cyclic nature of the transition state is further supported by the sign and magnitude of the entropy of activation: ASSz9, = S0z98(TS)- So,,,(epoxyoxepin) =

Figure 2. Transition state for the conversion of transoid-2,3-

77.7 - 79.2 = -1.5 eu For the Overall reaction the entropy change

epoxyoxepin into (eZz2z)-muconaldehyde. (a) Diagram showing bond lengths. (b) Schematic 3-D representation.

ASo,,

Compared to the transoid conformer, these bond distances increase from 1.434 to 2.016 A, and from 1.375 to 1.508 A,

is quite large and positive, in keeping with other reactions

12

= Soz,,[(eZzZz)-muconaldehydel So,(epoxyoxepin) = 87.5 - 79.2 = +8.3 eu

Chem. Res. Toxicol., Vol. 6, No. 5,1993 705

Muconaldehyde Production from Benzene

20.0

i

+B.O TS

cyclobutene3,4dicarboxaIdehyde

Cls trans

-

+300

k-

+20.0

i

17.8

18.4

I

ENERGY

ENERGY

kcal mol''

I5.0

c

ezzzz

2-formyl-2H -pyran

2L

I

-

muconaldehyde

0.0

ex. CHO

5.9 eZeZz

eq. CHO

2.1

8.7

5.5

0.8

eEeEz

eEeEe

Figure 3. Energy profile of the muconaldehyde conformers, 2-formyl-W-pyran (equatorial and axial CHO groups) and cyclobutene-3,4-dicarboxaldehyde(cis and trans CHO groups), relative to the energy of (eEeEe)-muconaldehydetaken as zero. in which vibrational modes in a ring structure become internal rotations in a chain structure. Typical values range from +5 to +10 eu (35). The small entropy of activation is a clear indication that scarcely any alteration in the type of structure has occurred. It may be noted that the close similarity between the structures of the reactant and the transition state, coupled with the substantial exothermicity of the overall reaction, is in accord with Hammond's postulate' (36). Other Muconaldehyde Structures. The structures of the eZeZz-, eEeZz-, eEeEz-, and eEeEe-conformers are depicted in Figure ld-g, and their energies relative to that of the most stable conformer, eEeEe, are plotted in Figure 3. As expected, the eZeZz conformer in which the central c4-C~bond has the e configuration is more stable than the eZzZz conformer, to the extent of 3.9 kcal mol-'. 2-Formyl-2H-pyran and Cyclobutene-3,4-dicarboxaldehyde. Two conformers of 2-formyl-2H-pyran have been identified, one with the formyl group equatorial with respect to the nominal pyran ring plane and the other with the group in the axial position, 5.5 and 6.7 kcal mol-', respectively, higher in energy than (eEeEe)-muconaldehyde. Trans and cis isomers of cyclobutene-3,Cdicarboxaldehyde have been characterized,but these structures are significantly less stable, their energies lying 16.4 and 17.6 kcal mol-', respectively above that of (eEeEe)muconaldehyde, see (Figure 3). Transition State: Degenerate CopeRearrangement of 2,$-Epoxyoxepin. This rearrangement involves a switch of the carbon-carbon bond from C3-cZ to Cz-Cl, so that 08,formerly in the epoxide ring, is now in the ~~

~

1 Hammond's postulate suggests that if two states occur consecutively

during a reaction process and have nearly the same energy, their interconversionwill involve only a small reorganizationof the molecular structure.Then, in a highlyexothermicstep the transitionstate will closely resemble the reactant.

1

muconaldehyde

Figure 4. Energy profile of cisoid- and transoid-2,3-epoxyoxepin, the transition statesfor the degenerate Cope rearrangement of the transoid conformer and for ita conversion into (eZz2z)muconaldehyde,and (eZeZz)-,(eEe2z)-,(eEeEz)-,and (eEeEe)muconaldehyde. oxepin ring, and 07,formerly in the oxepin ring, is now in the epoxide ring (see Scheme IIe). In the transition state the bond lengths C4-C5 and c5-c6 are equal, 1.403 A, and likewise the bond lengths C3-C4 and C&, 1.366 A, with a plane of symmetry passing through C5 and C1, perpendicular to the c3c4c6c1plane. Thus, compared to epoxyoxepin, the long bonds are shorter and the short bonds longer. These features are indicative of homoaromatic conjugation (37, 38) in the six-membered C&c4c5c6c1 ring-a stabilizing factor. Nevertheless, this transition state is far higher in energy than epoxyoxepin by some 38.0 kcal mol-' (see Figure 41, and there is no likelihood of the rearrangement being observable under experimental conditions-for example by the formation of muconaldehyde labeled OCH-CDO as well as OCDCDO from the 1,a-dideuterio derivative of benzene oxideoxepin, unless the transition state has considerable polar character in aqueous solution. Reaction Path from 2,3-Epoxyoxepin to (eEeEe)Muconaldehyde. A schematic energy profile linking the cisoid and transoid conformers of 2,3-epoxyoxepin via the transition state, 12,to (e2zZ.z)-muconaldehyde,and thence to the most stable conformer, eEeEe, is shown in Figure 4. The production of (eEeEe)-muconaldehyde is a very favored reaction, with an exothermicity AHOzga of -26.8 kcal mol-'. The reaction steps will now be taken in turn. First, even if the cisoid conformer of the epoxy-oxepin were to be formed initially, its conversion into the transoid conformer would probably be rapid since only ring inversion is involved, and the two conformers would be in equilibrium. With M 0 2 9 8 = +1.2 kcal mol-' and As0298 = 79.2 - 77.6 = +1.6 eu, AGO298 = +0.6 kcal mol-', giving an equilibrium constant of 0.36. About 25% of the equilibrium mixture would thus be the transoid conformer. The activation energy for the formation of (e2zZz)muconaldehyde from the transoid conformer is 16.5 kcal mol-l, and the entropy of activation -1.5 eu (see above). Calculations of the velocity constant (39) at 80,15, and 0 "c, the temperatures employed in the experimentalstudies, give half-reaction times of 0.04,10, and 50 s, respectively; hence it is not surprising that the epoxyepoxin was not

706 Chem. Res. Toxicol., Vol. 6, No. 5, 1993

observed. Further calculations, however, show that at a considerably lower temperature, say -40 oC, the halfreaction time would be about 3 h; so under these conditions detection by NMR measurements is quite a possibility. In the final sequence of reactions leading from (eZzZz)to (eEeEe)-muconaldehyde there are changes in configuration about both the carbon-carbon single bonds and the carbon-carbon double bonds. The former are likely to be rapid processes with each pair of conformers in equilibrium. From the data at hand it can be shown, for example, that in the equilibrium eZzZz + eZeZz more than 99.9% of eZeZz would be present, and in the equilibrium eEeEz + eEeEe there would be about 75% of eEeEe. The changes in configuration about the carboncarbon double bonds, however, are clearly rate-determining steps since both the Z,Z-isomer and the E,Z-isomer are isolable prior to the formation of the E,E-isomer. Rotation about a carbon-carbon double bond is an exceedingly unfavorable process with a barrier of some 30-50 kcal mol-' (40,411, so each step must proceed via an intermediate that allows the double bond in question to become a single bond about which rotation can then readily occur. For the first isomerization,Z,Z- E,Z, 2-formyl-2H-pyran(19) (10) and cyclobutene-3,4-dicarboxaldehyde(201211 are such intermediates. Following ring closure in the eZzZzconformer, across 07".C3 and Cg-C3, respectively, ring fission could in principle lead to the other isomer with CpC4 having the E instead of the 2-configuration (see Scheme IIf,g). As shown in Figure 3, the energies of the equatorial-CH0 and axial-CHO conformers of 2-formyl2H-pyran lie within the range of values for the muconaldehyde conformers. Therefore, provided the activation energyfor their formation from the more stable conformer of the Z,Z-isomer, eZeZz, does not exceed 20 kcal mol-' (which is a reasonable value), the pyran structures could well serve as intermediates for the Z,Z- E,Z isomerization. On the other hand, the energies of the cyclobutene derivatives lie well above those for the aldehyde conformers-by at least 6 kcal mol-' (see Figure 3). Moreover, judging from the values for cyclobutene with various substituents (42,43),the activation energy for ring opening is unlikely to be less than 25 kcal mol-'. Hence the activation energy for the formation of these derivatives from muconaldehyde would probably exceed 30 kcal mol-', and so they are far less likely than the pyran structures to serve as intermediates in the Z,Z E,Z isomerization. For steric reasons the E configuration in the E,Z-isomer prevents the formation of another ring structure that could serve as an intermediate for the E,Z E,E isomerization, and we conclude that in the gas phase this final isomerization would not occur-except, possibly at very high temperatures. Under experimental conditions, Le., in solution at 0, 15, and 80 OC (4, 191, the formation of the E,E-isomer can be attributed to acid catalysis as illustrated in Scheme IV (44),3to base catalysis as exemplified by the addition of a tertiary amine giving a resonance-stabilized Michaelis intermediate (19) (see Scheme V), or to less specific solvent effects. Under experimental conditions these other mechanisms could also, of course, be responE,Z. sible for the first isomerization, Z,Z

-

-

* In attempted syntheses of cyclobutene-3,4-dicarboxaldehyde,although it may have been generated, it was not observed, and the E,Zisomer of muconaldehyde was formed instead (20). 3 This mechanism is baaed on the acid-catalyzed isomerization of dimethyl maleate to dimethyl fumarate, as depicted in Figure 12-23 on p 344 of reference 44.

Greenberg et al. Scheme IV eEeZz

f0

l\O7

t -H+

I +H+ ."fO

Ht/f3 1

-/5

6

1 k 0 +

17

H

I

4 /3 .i/3

Dioxetane Pathway. The structure of dioxetane (9) is depicted in Figure 1Ha. There is a plane of symmetry passing through the midpoints of CrC5, C2-C1, and OS07,and, as shown in the schematic cross section of the molecule in Figure lHb, the six-membered carbon ring system is almost flat and the folding angle with respect to the bonded oxygens,i.e., the angle between the C ~ C S C ~ C ~ and C2C1O7O8 planes, is about 116O. This structure, however, is 48.5 kcal mol-' higher in energy than the transoid conformer of the epoxyoxepin. Furthermore, since typical activation energies for ring cleavage in 1,2dioxetanes giving carbonyl derivatives are about 25 kcal mol-' (45,46), it is very unlikely that dioxetane (9) is a precursor of muconaldehyde under metabolic conditions. Energetic Aspects of t h e Formation of Transoid2,3-Epoxyoxepin and Dioxetane from Benzene. In epoxidation by the microsomal monooxygenase system the second of the two atoms of the 0 2 molecule is reduced to water by the cofactor NADPH (47):

-

benzene + 0, + NADPH + H+ benzene oxide + H 2 0 + NADP' The size and complexity of NADPH and NADP+ make it impracticable to calculate LiH0298 for this reaction as such; however, an approximate value can be arrived at by the summation of the AH0298 values for two component redox

Muconaldehyde Production from Benzene Scheme V

Chem. Res. Toxicol., Vol. 6, No. 5, 1993 707 Table 11. A P m a Values, in koa1 mol-', for the Reactions Involved in the Formation of tramso~d-2,3-Epoxyoxepin and Dioxetane from Benzene, Calculated from the Energies in Table I and Additional Data Given in the Footnotes

-

reactions 1.benzene + 02 + H2 benzene oxide + HzO 2. NADPH H+ NADP+ Hz 3. benzene + 0 2 NADPH + H+ benzene oxide HzO + NADP+ 4. benzene oxide oxepin 5. oxepin 02 Hz transoid-2,3epoxyoxepin + HzO 6. oxepin 02 NADPH H+ transoid-2,3-epoxyoxepin HzO NADP+ 7. benzene 202 2Hz transoid-2,3epoxyoxepin + 2Hz0 8. benzene 202 2NADPH 2H+ transoid-2,3-epoxyoxepin + 2H20 2NADP+ 9. benzene + 02 dioxetane

+ + + + + + + + + + + + + +

'k0 I -R3N

+

+ +

+

-38.6" +3.7b -34.9 +3.1 -73.4' -69.7"

-108.9

-101.5f +30.3O

Including the pressurevolume work term, A(PV) = AnRT, where An is the number of product molecules minus the number of reactant molecules (29). Calculated from an experimental determination of the temperature dependence of the oxidation-reduction potential for the NADH/"+ couple (49), and adoption of the same value for the NADPH/NADP+ couple (see text). Sum of reactions 1and 2. d Sum of reactions 5 and 2. e Sum of reactions 1,4, and 5. f Sum of reaction 7 and twice reaction 2. a

*

processes. In the first Hz serves as the reducing agent in a Ymodel"epoxidation reaction: benzene + 0, + H, benzene oxide + H,O and in the second, since the oxidation-reduction potentials of NADH/NAD+and NADPH/NADP+ are essentially the same, to within 1mV (481,AHo~9s for the cell reaction (49)

-

+

-

+

NADH H+ NAD' H, +3.7 kcal mol-' can be adopted for the analogous cell reaction

-

NADPH + H+ NADP' + H, The results of these calculations are listed in Table 11. The epoxidation of benzene to give benzene oxide, reaction 1,and oxepin to give transoid-2,3-epoxyoxepin, reaction 5, are both highly exothermic processes-driven in large part by the reduction of one of the two atoms of the 02 molecule to water. The formation of benzene oxide is rather less exothermicdue to the disruption of the aromatic stabilization in the benzene. Taking the NADPH/NADP+ couple into account, the exothermicity is reduced but not significantly. For the overall reaction in which transoid2,3-epoxyoxepin is formed from benzene, reaction 8, the exothermicity exceeds 100 kcal mol-1, in marked contrast to the formation of dioxetane from benzene which is endothermic to the extent of 30 kcal mol-'-a further reason for doubting whether it plays any part in the metabolicoxidation. In the photochemicaloxidation with singlet oxygen the reaction is slightly exothermic to the extent of 4 kcal mol-'.

Discussion Ring cleavage leading to compounds analogous to muconaldehyde has not been observed in the oxidative metabolism of polycyclic aromatic hydrocarbons (PAHs)4 4 Abbreviations: PAHs stands for polycyclic aromatic hydrocarbons; and N, naphthalene (as in N-1,2-oxide).

(50, 51). Likewise, methyl(trifluoromethyl)dioxirane, which, going beyond the epoxidestage, reacts with benzene to form muconaldehyde, gives only a variety of epoxides with PAH's such as anthracene, phenanthrene, chrysene, and pyrene (10, 52-54). Several factors may be noted which undoubtedly contribute to this striking difference in behavior. (a) In the case of benzene, aromatic stabilization is no longer an issue once the initial epoxidationhas taken place. The derivatives have either cyclic or acyclic aliphatic structures; and, with no specific stabilizing feature, the oxide and oxepin forms have very similar energies (13). (b) With PAH's on the other hand, depending on the reaction site and the particular PAH, the initial epoxidation leads to the conversion of one or more benzenoid rings into cyclic aliphatic rings (55,561. Either the oxide or the oxepin form thus retains the greater number of benzenoid rings. In view of these differences in structure, it was postulated at the time of their discovery that "resonance stabilization demands that either the oxide or the oxepin form be the stable species" (57,58). Recent ab initio calculations at the RHF/6-31G//RHF/6-31G level have fully substantiated these expectations (56). Naphthalene-1,2-oxide (N-l,a-oxide),in which the unsubstituted ring is benzenoid, is some 20 kcal mol-' more stable than ita oxepin counterpart, in which the ring is orthoquinonoid (see Scheme VI-C, a b). (c) The microsomal monooxygenase system is, however, remarkably site specific, and N-1,2-oxide is formed exclusively (59). Furthermore the energetics also favor its formation (56). As in the case of benzene oxide, ring cleavage to give a dicarbonyl derivative analogous to muconaldehyde can be envisaged as proceeding via corresponding steps: oxepin formation, Scheme VIC, a b; a second epoxidation to give N-2,3-epoxy-1,2-oxepin, Scheme VIC, b c; followed by ring cleavage giving the C1=Og/C2=Olo dialdehyde, Scheme VIC, c d. But the endothermicity associated with the conversion of N-1,2oxide into N-l,Zoxepin, about 20 kcal mol-', quite apart from any additional activation that might be required, imposes a formidable barrier to ring cleavage by this

-

-

-

-

708 Chem. Res. Toxicol., Vol. 6, No. 5, 1993

Greenberg et al.

Scheme VI

N .2.3ax1de

N .2.3.ox@n

N-1.2-x*

a

paC.~ay.~The metabolic fate of N-1,2-oxideevidently lies in more favorable competing reactions, such as isomerization giving a-naphthol, hydration giving 1,2-dihydroxy1,2-dihydronaphthalene catalyzed by epoxide hydratase, or the formation of conjugates with, for example, glutathione (50, 59). (d) A structural feature that would otherwise favor ring cleavage by this pathway is the generation of a benzenoid ring in the dialdehyde in place of the orthoquinonoid rings that are present in N-1,2-oxepin and N-2,3-epoxy-1,2oxepin (see Scheme VIC, b, c, and d). On the other hand, formation of dialdehydes by both the N-8a,l-oxepin and N-2,3-oxepin pathways involves the replacement of a benzenoid ring by an orthoquinonoid ring (see Scheme VIA, b, c, and d, and Scheme VIB, b, c and d). The adverse energy change associated with this replacement would thus affect the favorable energy change associated with the ring fission, and judging from the data for the benzene pathway via 2,3-epoxyoxepin, there might be very little difference in energy between these epoxyoxepins and the dialdehydes.

Summary and Conclusions In broad outline the present calculations fully corroborate the metabolic pathway for the production of muconaldehyde from benzene via 2,3-epoxyoxepin, as suggested by Davies and Whithan in 1977 (4). (i) In model reactions simulating the microsomal monooxygenase system, the formation of benzene oxide and the epoxyoxepin are very favorableexothermic processes, with A H o ~values s of -38.6 and -73.4 kcal mol-', respectively, the smaller exothermicity of the former reaction being due in large part to the disruption of the aromatic stabilization present in benzene. (ii) The epoxyoxepin is predicted to have a very short half-life, less than 1min, at the temperatures The isolationof optically active naphthalene l,?-oxide (60),completely stable for 8 h in CHaOH at 20 OC (and longer still in CHCla), enables an estimate to be made of the activation energy for conversion into the oxepin-assuming that the oxepin is the transition-state structure for the racemization.Taking 2 X 8hours to be the half-life for the racemization, the rate constant, k = (In 2)/t1,2, would be 1.2 X 1od 8-1, and, with A in the rate constant expression k = Ae-E'RT the same as that for benzene oxide, i.e. 2.5 X 1014 8-1 (I3),E is found to be about 26 kcal mol-'.

d

C

L a t were employed in the experimental studies, 80, 15, and 0 OC, which explains why it escaped detection. On the other hand, at a considerably lower temperature, say -40 OC, the half-life is increased to several hours, and detection, for example, by NMR measurements, becomes a distinct possibility. (iii) The eZzZz-, eZeZz-, eEeZz-, eEeEz-, and eEeEe-conformers of muconaldehyde are found to be lower in energy than the epoxyoxepin by 17.0, 20.9, 24.8,26.0, and 26.8 kcal mol-', respectively, and are thus formed exothermically in successive steps with the Z,Zisomer less stable than the E&, and the E,Z- less stable than the E,E-, in accord with experiment. (iv) Two conformers of 2-formyl- W-pyran, formed by ring closure in the Z,Z-isomer, have been identified, almost equal in energy, one with the CHO group equatorial, the other with the CHO group urial. Moreover, both conformers are at about the same energy level as the Z,Z-conformers and hence could serve as intermediates for the conversion of a Z,Z into an E,Z structure. (v) For steric reasons the E configuration in the E,Z-isomer prevents the formation of another ring structure that could serve as an intermediate for the E,Z E,E isomerization, and we conclude that in the gas phase this final isomerization would not occur. Under experimental conditions in solution, the formation of the E,E-isomer can be attributed to acid catalysis, base catalysis, or less specific solvent effects. (vi) But the mechanism by which the 2-formyl-W-pyran conformers could bring about the Z,Z E,Z conversion remains unresolved. What is at issue is the characterization of both the transition state that links a particular Z,Z-conformer with one of the pyran conformers, and the transition state that links the other pyran conformer with a particular E,Z-conformer. To be certain regarding the particular Z,Z- and E,Z-conformers involved, it might be necessary to determine the structure of some of the Z,Zand E,Z-conformers which have not yet been studied. Calculations to establish this mechanism are currently under way. (vii)The degenerate Cope-likerearrangement of 2,3-epoxyoxepinappears to be too slowto permit a labelscrambling experiment to assess its intermediacy as a metabolite. However, in solution the transition state may have considerable dipolar (or zwitterionic) character which

-

-

Muconaldehyde Production from Benzene

could considerably lower the activation barrier. This possibility is also being studied.

Acknowledgment. We thank the Advanced Scientific Computing Laboratory, NCI-FCRF, for providing time on the CRAY YMP supercomputer. This work was also supported by Grants CA-10925 and CA-06927 from the National Institutes of Health and CN-10 from the American Cancer Society and by an appropriation from the Commonwealth of Pennsylvania. SupplementaryMaterial Available: Tables IS-IXS, giving Z-matrix orientations for all the organic structures in Table I (9 pages). (This material is also contained in libraries on microfiche and immediately following this article in the microfilm version of the journal.) Ordering information is given on any current masthead page.

References (1) Williams, R. T. (1959) Detoxification Mechanisms, 2nd ed., p 189,

Wiley-Interscience, New York. (2) Nakajima, M., Tomida, I., and Takei, S. (1959) The chemistry of 3,5-cyclohexadiene-1,2-diol. IV. Preparation of cis-3,5-cyclohexadiene-1,2-diol, the stereoisomeric muconic dialdehydes, and conduritol F. Chem. Ber. 92, 163-172. (3) Tomida, I.. and Nakaiima. M. (1960) The chemistrv of 3.5cyclohexadiene-l,2-diol.~VI. Metabolismof the glycols ani muconic dialdehyde. 2.Physiol. Chem. 318, 171-178. Davies, S. G., and Whitham, G. H. (1977) Benzene oxide-oxepin. Oxidation to muconaldehyde. J.Chem. SOC.,Perkin Trans. I, 13461347. Goldstein, B. D., Witz, G., Javid, J., Amoruso, M., Rossman, T., and Wolder, B. (1982) Muconaldehyde, a potential toxic intermediate of benzene metabolism. In Biological Reactive Intermediates. (Snyder,R., Parke,V. V., Kocsis, J. J., Jallow,D., Gibson, G . G., and Witmer, C. M., Eds.) Vol. 2, Part A, pp 331-339, Plenum, New York. Latriano, L., Zaccaria, A., Goldstein, B. D., and Witz, G. (1986) Muconaldehyde formation from I'C-benzene in a hydroxyl radical generating system. J. Free Radicals Biol. Med. 1,363-371. Latriano, L., Goldstein, B. D., and Witz, G. (1986) Formation of muconaldehyde, an open-ring metabolite of benzene, in mouse liver microsomes an additional pathway for toxicmetabolites.Proc. Natl. Acad. Sci. U.S.A. 83,8356-8360. (1989)DHHS Publication No. (NIH)89-218 Symposiumon Benzene Metabolism, Toxicity and Carcinogenesis. Environmental Health Perspectiues, Vol. 82, National Institute of Environmental Health Science, Research Triangle Park, NC 27709. Kirley, T. A., Goldstein, B. D., Maniara, W. M., and Witz, G. (1989) Metabolism of trans, trans-muconaldehyde, a microsomal hematotoxic metabolite of benzene, by purified yeaat aldehyde dehydrogenase and a mouse liver soluble fraction. Toxicol. Appl. Pharmacol. 100, 360-367. Mello, R., Ciminale, F., Fiorentino, M., Fusco, C., Prencipe, T., and Curci, R. (1990) Oxidation by methyl(trifluoromethy1)dioxirane. 4. Oxyfunctionalizationsof aromatic hydrocarbons. Tetrahedron Lett. 31,6097-6100. Goldstein, B. D., and Witz, G. (1991) Benzene. In Critical Reviews of Enuironmental Toxicants-Human Exposure and Their Health Effects (Lippmann, M., Ed.) Chapter 3, Van Nostrand, New York. Bechtold, W. E., Lucier, G., Bimbaum, L. S., Yin, S. N., and Henderson, R. F. (1991) Muconic acid determinations in urine aa a biological exposure index for workers occupationally exposed to benzene. Am. Ind. Hyg. Assoc. J. 52, 473-478. Vogel, E., and Giinther, H. (1967) Benzene oxideoxepin valence tautomerism.. Angew. Chem., Znt. Ed Engl. 6,385-401. Altenbach, H.-F., and Vogel, E. (1972) Valence tautomerism of 1,4dioxocin with syn-benzene dioxide. Angew. Chem., Int. Ed. Engl. 11,937-939. Bock, C. W., George, P., and Glusker, J. P. (1990) The lengths of C-O epoxide bonds and intramolecular hydrogen bonding in typical oxidative metabolites of benzene, naphthalene and phenanthrene: A computational molecular orbital study. 13th Austin Symposium on Molecular Structure,p 90, Universityof Texaaat Austin, Austin, TX, Abstract 54. Wei, K., Mani, J. C., and Pitts, J. N., Jr. (1967) The formation of polyenic dialdehydes in the photooxidation of pure liquid benzene. J. Am. Chem. SOC. 89,4225-4227.

Chem. Res. Toxicol., Vol. 6, No. 5, 1993 709 (17) Frankel, E. N., Neff, W. E., and Selke, E. (1984) Analysis of autoxidized fats by gas chromatography-mass spectrometry. IX. Homolytic vs heterolytic cleavageof primary and secondaryoxidation products. Lipids 19,790-800. (18) Kaplan, M. L., and Trozzolo, A. M. (1979) Role of singlet oxygen in the degradation of polymers. Insinglet Oxygen (Waeserman, H. H., and Murray, R. W., Ekls.) pp 584-585, Academic Press, New York. (19) Golding, B. T., Kennedy, G., and Watson, W. P. (1988) Simple syntheses of isomersof muconaldehydeand 2-methylmuconaldehyde. Tetrahedron Lett. 29, 5991-5994. (20) Hinshaw, J. C. (1974) Attempted synthesis of cis-cyclobutene-3,4dicarboxaldehyde. J. Org. Chem., 39,3951-3953. Frisch, M. J., Head-Gordon, M., Trucks, G. W., Foresman, J. B., Schlegel, H. B., Raghavachari, K.,Robb, M. A., Binkley, J. S., Gonzalez, C., Defrees, D. J., Fox, D. J., Whiteside, R. A., Seeger, R., Melius, C. F., Baker, J., Martin, R. L., Kahn,L. R., Stewart, J. J. P., Topiol, S., and Pople, J. A. (1990) GAUSSIAN 90,CarnegieMellon, Quantum Chemistry Publishing Unit, Pittsburgh, PA. Hariharan,P. C., and Pople, J. A. (1973)The influence of polarization functions on molecular orbital hydrogenation energies. Theor. Chim. Acta. 28, 213-222. Meller, C., and Plesset, M. S. (1934) Approximation treatment for many-electron systems. Phys. Rev. 46, 618422. Binkley, J. S., and Pople, J. A. (1976) Meller-Plesset theory for atomic ground-state energies. Int. J. Quantum Chem. 9,22*236. Pople, J. A,, Binkley, J. S., and Seeger, R. (1976)Theoretical models incorporating electron correlation. Znt. J.Quantum Chem., Symp. 10,l-19. McIver, J. W., Jr., andKormomicki,A. (1972)Structureoftransition states in organic reactions: general theory and an application to the cyclobutene-butadiene isomerization using a semiempirical molecular orbital method. J. Am. Chem. SOC. 94,2625-2633. Pople, J. A., Kriehnan, R., Schlegel, H. B., and Binkley, J. S. (1979) Derivative studies in Hartree-Fock, and Maller-Plesset theories. Int. J. Quantum Chem.: Quantum Chem. Symp. 13, 225-241. Schlegel, H. B. (1989) Some practical suggestions for optimizing geometries and locating transition states. In New Theoretical Concepts for Understanding Organic Reactions (Bertran, J., and Csizmadia, I. G., Eds.) pp 33-53, Kluwer Academic Publishers, Dordrecht. Del Bene, J. E. (1986) Quantum mechanical reaction enthalpies. In Molecular Structure and Energetics (Liebman, J.F., and Greenberg, A,, Eds.) Chapter 9, pp 319-349, VCH Publishers, Deerfield Beach, FL. Thieme, R., and Weiss, C. (1983) Quantum-chemical calculations on the valence isomerization of benzene oxide. Stud. Biophys. 93, 273-274. Grimme, W., and Doering, W. v. E. (1973)A degenerate butadienylcyclopropane rearrangement in bicyclo[5.1.0locta-2,4-diene.Chem. Ber. 106, 1765-1780. Bock, C. W., George, P., Stezowski, J. J., and Glusker, J. P. (1990) Theoretical studies of the benzene oxide-oxepin valence tautomerism. Stmct. Chem. 1, 33-39. George, P., Bock, C. W., and Glusker, J. P. (1992) Protonation of monosubstituted oxiranes: a computational molecular orbital study of oxonium ion versus carbonium ion formation. J.Phys. Chem. 96, 3702-3708. Bock, C. W., George, P., and Trachtman, M. (1984) A molecular orbital study of the rotation about the C C bond in 1,bbutadiene. Theor. Chim. Acta (Berlin) 64, 293-311. Moore, J. W., and Pearson, R. G. (1981) Kinetics and Mechanism, 3rd ed., p 214, Table 6.6, Wiley, New York. Hammond, G. S. (1955)A correlation of reactionrates. J.Am. Chem. SOC.77, 334-338. Cremer, D., Kraka, E., Slee, T. S., Bader, R. F. W., Lau, C. H. D., Nguyen-Dang, T. T., and MacDougall, P. J. (1983) Description of homoaromaticity in terms of electron distributions. J.Am. Chem. SOC. 105,5069-5075. Dewar, M. J. S. (1988) Localization and delocalization. In Modern Models of Bonding and Delocalization (Liebman, J. F., and Greenberg, A., Eds.) Chapter 1,pp 1-59, VCH Publishers, New York. Kondratiev, V. N., and Nikitin, E. E. (1981) Gas-Phase Reactions, pp 60 and 104, Springer-Verlag, New York. Crombie, L. (1952) Geometrical isomerism about carbon-carbon double bonds. Q. Rev. (London) 6, 101-140. Magee, J. L., Shand, W., Jr., and Evrina, H. (1941) Non-adiabatic readions: rotation about the doubie bind. J. Am. Chem. SOC. 63, 677-688 (see Table 11, p 684). (42) Rondan, N. G., and Houk, K. N. (1985) Theory of stereo selection in conrotatory electrocyclic reactions of substituted cyclobutenes. J. Am. Chem. SOC. 107, 2099-2111. (43) Spellmeyer, D. C., and Houk, K. N. (1988) Transition structures of pericyclic reactions: electron correlation and basis set effects on the transition structure and activation energy for the electrocyclization of cyclobutene to butadiene. J. Am. Chem. SOC. 110, 3412-3416.

710 Chem. Res. Toxicol., Vol. 6,No. 5, 1993 (44) Eliel, E. L. (1962) Mechanism based on the acid-catalyzed isomerization ofdimethylmaleatetodimethylfumarate.InStereochemistry of Carbon Compounds, p 344, Figure 12-23, McGraw-Hill, New York. (45) Bartlett, P. D., and Landis, M. E. (1979) The 1,2-dioxetanes. In Singlet Oxygen (Waaserman, H. H., and Murray, R. W., Eds.) pp 243-286, Academic Press, New York. (46) Baumstark, A. L., and Wilson, C. F. (1981) Thermolysis of 3,4-dialkyl-1,2-dioxetanes: effect of cyclicsubstituents. Tetrahedron Lett. 22,4363-4364. (47) West, C. A. (1980) Hydroxylases, monooxygenasesand cytochrome P-450. In The Biochemistry of Plants (Davies, D. D., Ed.) Vol. 2, Chapter 8, pp 317-359, Academic Press, New York. (48) Rodkey, F. L., and Donovan, J. A,, Jr. (1959) Oxidation-reduction potentials of the triphosphopyridine nucleotide system. J. Biol. Chem. 234, 677680. (49) Rodkey, F. L. (1959) The effect of temperature on the oxidationreduction potential of the diphosphopyridine nucleotide system. J. Biol. Chem. 234, 188-190. (50) Sims, P., and Grover, P. L. (1974) Epoxides in polycyclic aromatic hydrocarbon metabolism and carcinogenesis. Adu. Cancer Res. 20, 165-274. (51) Harvey, R. G. (1981) Activated metabolites of carcinogenic hydrocarbons. Acc. Chem. Res. 14,218-226. (52) Jeyaraman, R., and Murray, R. W. (1984)Production of arene oxides

Greenberg et al. by the caroate-acetone system (dimethyldioxirane). J.Am. Chem. SOC. 106, 2462-2463. (53) Murray, R. W., and Jeyaraman, R. (1985) Dioxiranes: synthesis and reactions of methyldioxiranes. J. Org. Chem. SO, 2847-2853. (54) Murray, R. W. (1989) Dioxiranes. Chem. Rev. 89, 1187-1201 and references therein. (55) Boyd, D. R., and Stubbs, M. E. (1983) Arene oxide-oxepin isomerization: theoretical predictions and experimental evidence. J. Am. Chem. SOC.105, 2554-2559. (56) Bock, C. W., George, P., and Glusker, J. P. (1991) A computational molecular orbital study of the oxide and oxepin valence tautomers of naphthalene. J. Mol. Struct. (THEOCHEM) 234, 227-246. (57) Vogel, E., and Klirner, F.-G. (1968) 1,2-Naphthalene Oxide. Angew. Chem., Int. Ed. Engl. 7, 374-375. (58) Jerina,D. M., Yagi,H., and Daly, J. W. (1976)Areneoxides-oxepins. Heterocycles 1, 267-326 and references therein. (59) Jerina, D. M., Daly, J. W., Witkop, B., Zaltsman-Nirenberg, P., and Udenfriend, S. (1968) The role of arene oxide-oxepin systems in the metabolism ofaromaticsubstratea. 111.Formationof 1,2-naphthalene oxide from naphthalene by liver microsomes.J. Am. Chem. SOC. 90, 6525-6527. (60) Boyd, D. R., Jerina, D. M., and Daly, J. W. (1970) Optically active 1,a-naphthalene oxide. J. Org. Chem. 35, 3170-3172.