Conformational Analysis. 50. C-Methyl-1,2,3,4-tetrahydroisoquinolines

Article abstract of DOI:10.1021/jo971257r

Conformational equilibria in 1-, 3-, and 4-methyl-1,2,3,4-tetrahydroisoquinolines (THIQs) and the diastereomeric pairs of their 1,3- and 1,4-dimethyl homologs have been determined by measurement of H₃/H₄(trans) coupling constants and have been confirmed by molecular mechanics [MMP2(85)] calculations. The experimental -ΔG ° values (a → e) for the monomethyl compounds (computed values in parentheses) in kcal mol_-1 are Me-1, 0.56 (0.46); Me-3, 1.63 (1.53); and Me-4, -0.32 (-0.22). Agreement of experimental and calculated values is very good as is the additivity of values for the dimethyl compounds (Table 1). Values for the corresponding hydrochlorides are Me-1, 0.19 (-0.34); Me-3, 1.15 (1.46); and Me-4, 0.35 (0.10) kcal mol_-1. The less than satisfactory agreement of experimental with computed data here is probably due to neglect of solvation. The very small or negative ΔG ° values for Me-1 and Me-4 were ascribed not only to the pseudoaxial (rather than axial) nature of Me(ax) and the absence of a syn-axial hydrogen on the side of the benzene ring but also to a peri interaction with H(8) and H(5), respectively, destabilizing equatorial methyl at positions 1 and 4. This was confirmed by comparing computed conformational energy values with values at corresponding positions in Δ^3,4-tetrahydropyridines (THPs). While ΔG ° in the two series is the same for Me-3 (THIQ numbering), that for Me-1 and Me-4 is considerably smaller in the THIQ than in the THP series which latter is devoid of peri hydrogens.

Full text of DOI:10.1021/jo971257r

9
154
J . Org. Chem. 1997, 62, 9154-9158
Con for m a tion a l An a lysis. 50.
C-Meth yl-1,2,3,4-tetr a h yd r oisoqu in olin es†
‡
Edward M. Olefirowicz and Ernest L. Eliel*
W. R. Kenan J r. Laboratories, Department of Chemistry, University of North Carolina,
Chapel Hill, North Carolina 27599-3290
Received J uly 11, 1997X
Conformational equilibria in 1-, 3-, and 4-methyl-1,2,3,4-tetrahydroisoquinolines (THIQs) and the
diastereomeric pairs of their 1,3- and 1,4-dimethyl homologs have been determined by measurement
of H
3 4
/H (trans) coupling constants and have been confirmed by molecular mechanics [MMP2(85)]
calculations. The experimental -∆G° values (a f e) for the monomethyl compounds (computed
-1
values in parentheses) in kcal mol are Me-1, 0.56 (0.46); Me-3, 1.63 (1.53); and Me-4, -0.32 (-0.22).
Agreement of experimental and calculated values is very good as is the additivity of values for the
dimethyl compounds (Table 1). Values for the corresponding hydrochlorides are Me-1, 0.19 (-0.34);
-
1
Me-3, l.15 (1.46); and Me-4, 0.35 (0.10) kcal mol
. The less than satisfactory agreement of
experimental with computed data here is probably due to neglect of solvation. The very small or
negative ∆G° values for Me-1 and Me-4 were ascribed not only to the pseudoaxial (rather than
axial) nature of Me(ax) and the absence of a syn-axial hydrogen on the side of the benzene ring but
also to a peri interaction with H(8) and H(5), respectively, destabilizing equatorial methyl at
positions 1 and 4. This was confirmed by comparing computed conformational energy values with
values at corresponding positions in ∆3 -tetrahydropyridines (THPs). While ∆G° in the two series
is the same for Me-3 (THIQ numbering), that for Me-1 and Me-4 is considerably smaller in the
THIQ than in the THP series which latter is devoid of peri hydrogens.
,4
In tr od u ction
The 1,2,3,4-tetrahydroisoquinoline (THIQ) skeleton (1)
1
is found in a variety of alkaloids such as laudanosine
1-[(3′,4′-dimethoxyphenyl)methyl]-2-methyl-6,7-dimethoxy-
[
THIQ], salsoline (1-methyl-6-hydroxy-7-methoxy-THIQ),
and anhalonidine (1-methyl-6,7-dimethoxy-8-hydroxy-
THIQ). It also occurs in the recently commercially
2
available THIQ-3-carboxylic acid and the 1-substituted
bis(1,2,3,4-tetrahydroisoquinolinium) quaternary salts3
such as atracurium used in general anesthesia (2).
Despite its wide occurrence, no systematic study has been
made of the conformational equilibrium (axial f equato-
rial) of substituents on the heterocyclic ring of THIQ. In
an early X-ray crystallographic study of quaternary salts4
it was found that this ring has the shape of a half-chair
In the present work we have investigated the confor-
mational equilibria and associated conformational ener-
gies (-∆G°) of the 1-, 3-, and 4-methyl and cis- and trans-
1,3- and 1,4-disubstituted THIQs, both experimentally
and computationally. The corresponding energy differ-
ences in the saturated system, methyl- and dimethyl-
substituted piperidine, have been studied in the past both
(
1
3), a conclusion which was confirmed in later studies of
- and 3-substituted THIQs.
5
6
+
Dedicated to Professor Norman L Allinger on the occasion of his
7
oth birthday.
13
7
by low-temperature C NMR spectroscopy and by mo-
lecular mechanics calculation,8 with good agreement
between the two methods. Unfortunately the low-tem-
perature NMR method is not applicable to the THIQs,
since the heterocyclic ring, being an analog of cyclohex-
‡
From the Ph.D. Dissertation of Edward M. Olefirowicz. Present
address: Amgen, Inc., MS 14-1-A, 1840 DeHavilland Drive, Thousand
Oaks, CA 91320-1789.
X
Abstract published in Advance ACS Abstracts, December 1, 1997.
(
1) Hesse, M. Alkaloid Chemistry; Wiley: New York, 1981, pp 32-
4
3. Shamma, M. The Isoquinoline Alkaloids; Academic Press: New
York, 1972, Chapters 1 and 2. Bentley, K. W. The Isoquinoline
Alkaloids; Pergamon: New York, 1965. Bringmann, G.; Pokorny, F.
In The Alkaloids; ed. G. A., Cordell, Ed.; Vol. 46, Academic Press: 1995;
Chapter 4.
9
ene, has an inversion barrier so low that we were not
able to “freeze out” individual conformers even at -120
1
°C. We therefore chose the high-resolution H NMR
(2) Cf. Chem. Eng. News 1997, March 10, 13. The compound is an
measurement of the coupling constants of protons located
N-methyltransferase inhibitor: (Grunewald, G. L.; Sall, D. J .; Monn,
J . A. J . Med. Chem. 1988, 31, 824-830) and is also used as a synthetic
R-amino acid in synthetic peptides; e.g., Wilkes, B. C.; Schiller, P. W.
Biopolymers 1994, 34, 1213-1219.
(5) Charifson, P. S.; Bowen, J . P.; Wyrick, S. D.; Hoffman, A. J .;
Cory, M.; McPhail, A. T.; Mailman, R. B. J . Med. Chem. 1989, 32,
2050-2058.
(3) (a) Stenlake, J . B.; Waigh, R. D.; Dewar, G. H.; Urwin, J .; Dhar,
N. C. Patentoffenlegung DE2655883, 1975. (b) Stenlake, J . B.; Waigh,
R. D.; Dewar, G. H.; Dhar, N. C.; Hughes, R.; Chapple, D. J .; Lindon,
J . C.; Ferige, A. G.; Cobb, P. H. Eur. J . Med. Chem. 1984, 19, 441-
(6) Iba n˜ ez, A. F.; Moltrasio Iglesias, G. Y.; Delfino, J . M. J .
Heterocycl. Chem. 1996, 33, 265-270.
(7) Eliel, E. L.; Kandasamy, D.; Yen, C.-Y.; Hargrave, K. D. J . Am.
Chem. Soc. 1980, 102, 3698-3707.
4
50. (c) Dhar, N. C.; Maehr, R. B.; Masterson, L. A.; Midgley, J . M.;
Stenlake, J . B.; Wastila, W. B. J . Med. Chem. 1996, 39, 556-561 and
references there cited.
(8) Profeta, S., J r.; Allinger, N. L. J . Am. Chem. Soc. 1985, 107,
1907-1918.
(4) El-Sayad, H. A.; Swaringen, R. A.; Yeowell, D. A.; Crouch, R.
(9) The barrier in cyclohexene is 5.3 kcal mol-1: Anet, F. A. L.; Haq,
M. Z. J . Am. Chem. Soc. 1965, 87, 3147-3150. J ensen, F. R.;
Bushweller, C. H. J . Am. Chem. Soc. 1969, 91, 5774-5782.
C.; Hurlburt, S.; Miller, R. W.; McPhail, A. T. J . Chem. Soc., Perkin
Trans. 1 1982, 2067-2077.
S0022-3263(97)01257-7 CCC: $14.00 © 1997 American Chemical Society

Conformational Analysis
J . Org. Chem., Vol. 62, No. 26, 1997 9155
trans to each other at positions 3 and 4 as the experi-
mental method.10 Molecular mechanics calculations were
effected with the MMP2(85)11 program.
Sch em e 1
Syn th esis
2
-Phenylethylamine and its 1- and 2-methyl-substi-
tuted homologs were either commercially available or
prepared by standard procedures. Formylation of the
appropriate amine and Bischler-Naperialski ring closure
followed by sodium borohydride reduction provided 1,2,3,4-
tetrahydroisoquinoline or its 3- or 4-methyl substituted
homologs, respectively. Similarly, acetylation of the
amine, ring closure and reduction provided the 1-methyl,
1
,3-dimethyl, and 1,4-dimethyl homologs. In the latter
two cases the stereoisomers were separated by chroma-
tography: HPLC in the case of the 1,3-, and GC in the
case of the 1,4-isomers.
Meth od ology
To obtain credible results for the conformational equilibria
of methyl- and dimethyl-substituted THIQs, we chose to
employ two entirely independent approaches as indicated
above, one experimental, the other computational. While the
computational approach had proved very successful with
a
Lowest-energy conformer is on right. b Methyl substituents are
at positions a, a′, e, and e′.
Ta ble 1. Con for m a tion a l Equ ilibr ia a n d F r ee En er gy
Differ en ces for Meth yl- a n d Dim eth yl-Su bstitu ted
1
2
methylisochromanes, there was no a priori assurance that
1
,2,3,4-Tetr a h yd r oisoqu in olin es
it would work equally well with the nitrogen analogs, although
8
conformer ratioa -δG°, kcal mol-1
the piperidine calculations mentioned earlier were a good
omen in this regard. On the other hand, while calculations
on the position of conformational equilibria from averaged
coupling constants have been used frequently and often with
substituent(s)
exptl
calcdb
exptl
calcdb
1-Me
3-Me
72/28
94/6
68/32
93/7
0.56
1.63
-0.32
large
1.18
0.45
1.53
-0.22
3.68
1.03
0.65
0.19
1
0
good outcome, the problem of the model constants to be used
always looms. In view of our inability to freeze out individual
conformers, coupling constants had to be inferred indirectly
from conformationally averaged spectra. And while it was
4
-Me
37/63
(100/0)c
88/12
78/22
59/41
41/59
99.8:0.2
85/15
75/25
58/42
cis-1,3-Me2
d
trans-1,3-Me2
cis-1,4-Me2e
0.75
0.22
1
3
similarly impossible to see individual conformers in the
C
trans-1,4-Me2e
1
3
NMR spectrum, the positions of C signals in the averaged
spectra were used to confirm qualitatively the more precise
results obtained from proton coupling constants.
a
Axial to equatorial conformer. b By MMP2(85) assuming ∆S
) 0. Separate calculations were performed for structures with
equatorial and axial NH; the ratios shown are, in each case, for
Calculated and experimental results of the free energy
differences between the two conformers (a f e) of 1-, 3-, and
c
the sums of the populations of these two structures. In the case
of the experimental result, this is an assumed value, based on the
4
-methyl-1,2,3,4-THIQ and the corresponding cis and trans
calculated one. d 3a f 3e. e 1a f 1e.
isomers of the 1,3- and 1,4-dimethyl compounds (Scheme 1)
are shown in Table 1. The model coupling constants were
obtained as follows: It was assumed that the cis-1,3-dimethyl
derivative would exist exclusively in the conformation with
both methyl groups equatorial. The large coupling constant
derived thus. [This treatment assumes that the methyl
substituents do not affect the magnitude of J H3/H4 (trans).]
in this compound, 11.00 Hz, was thus assumed to be J H3a/H4a
.
Resu lts
Unfortunately there is no H3e in this compound, so J H3e/H4e had
to be computed indirectly as follows: The 1-methyl compound
displays both J 3a*/4a′* (8.70 Hz) and J 3e*/4e′* (5.14 Hz); the starred
subscripts indicate that both coupling constants, while pre-
dominantly axial/axial or equatorial/equatorial, are in fact
averaged by the equilibration of the two conformers of the
Before discussing the results a few checks on accuracy
are in order. The agreement between experimental and
calculated results (Table 1) is very satisfactory, consider-
ing that they were obtained by totally different methods.
1
-methyl compound. Now, it can be readily shown13 that the
sum of the trans coupling constant is constant; i.e., J 3a/4a
(
13) Take, for example, the conformational equilibrium with n
e
and
+
n
a
representing mole fractions of the two conformers.
J
5
3e/4e ) J 3a*/4a′* + J 3e*/4e′*. Numerically, 11.00 + J 3e/4e ) 8.70 +
.14 when J 3e/4e ) 2.84 Hz.1 Given thus the values of the two
4
trans coupling constants, one can calculate conformational
equilibria for all other compounds using the equation15 K )
(
J - J ee)/(J aa - J ) where J ee ) 2.84 Hz, J aa ) 11.00 Hz, and J
Let us call the coupling constant for diaxially disposed protons J aa and
that for diequatorially disposed protons J ee. The coupling constant for
the unmarked set is J ) n J + n J (i) and that for the starred set
is the observed H3,4-trans shown in Table 2. The equilibrium
compositions and free energy differences shown in Table 1 are
e
aa
a
ee
is J * ) n
(n + n
constant.
e
J
ee + n
a
J
aa. Adding the two equations, J + J * ) (n
e a
+ n )J ee
+
e
a
)J aa. But n
e
+ n ) 1; hence, J + J * ) J ee + J aa which is a
a
(10) The earliest application of this method we could find is by
LaBlanche-Combier, A.; Levisalles, J .; Pete, J .-P.; Rudler, H. Bull.
Soc. Chim. Fr. 1963, 1689-1701. See also: Eliel, E. L.; Wilen, S, H.
Stereochemistry of Organic Componds; Wiley: New York, 1994; pp 641.
(14) Regarding the unusually low J ee coupling constants of protons
antiperiplanar to electronegative atoms (in this case N), see: Booth,
H. Tetrahedron Lett. 1965, 411-416.
(
11) Sprague, J . T.; Tai, J . C.; Yuh, Y.; Allinger, N. L. J . Comput.
Chem. 1987, 8, 581-603.
12) Olefirowicz, E. M.; Eliel, E. L. J . Comput. Chem. 1989, 10, 407-
12.
(15) From (i) in ref 13 above and n
- n )J aa + n ee when J - J aa ) n (J ee - J aa) or n
aa); hence 1 - n ) (J ee - J )/(J ee - J aa) and K ) n
(J ee - J )/(J - J aa) ) (J - J ee)/(J aa - J ).
e
) 1 - n
a
it follows that J ) (1
a
a
J
a
a
) (J - J aa)/(J ee
/n ) (1 - n )/n
-
)
(
J
a
e
a
a
a
4

9
156 J . Org. Chem., Vol. 62, No. 26, 1997
Olefirowicz and Eliel
ones; a similar situation is seen in 1-methyltetralins.1
6
Ta ble 2. P r oton P a r a m eter s for C3, C3-Me (P a r en th eses),
a n d Vicin a l H3-H4 Cou p lin g Con sta n ts in th e
In any case, qualitatively speaking the order of 13C
chemical shifts is as predicted with only one (very minor)
inversion in the case of the 1-Me and cis-1,4-Me com-
2
Meth yl-Su bstitu ted 1,2,3,4-Tetr a h yd r oisoqu in olin esa
compound
δa
δe
J 3e3a J 3a4a′ J 3e4e′ J 3a4e′ J 3e4a′ J H-Me
1
3
4
-Me
-Me
-Me
2.93 3.18 12.4 8.70 5.14 4.66 5.14
pounds.
2.94 (1.17)
2.75 3.13 12.3
2.98 (1.18)
10.52
3.88
5.91 4.84
3.56
(6.3)
Discu ssion
cis-1,3-Me2
trans-1,3-Me2 3.22 (1.15)
cis-1,4-Me2 2.88 3.05 12.6
11.00
10.03
(6.2)
(6.2)
3.89
We start with a comparison with C-methylpiperidines
cf. 4).7 The conformational energy (-∆G°Me) for 4-methyl
4.66 3.93
(
trans-1,4-Me2 2.63 3.24 12.5 7.62
5.25
-
1
in piperidine is 1.9 kcal mol , similar to that in cyclo-
a
At room temperature in CD2Cl2 + 15% CHCldCCl2.
17
-1
hexane, 1.74 kcal mol . The value for 3-methyl is
-
1
Ta ble 3. 13C NMR Ch em ica l Sh ifts of Alip h a tic Ca r bon s
smaller, 1.6 kcal mol , presumably because one of the
syn-axial Me/H interactions is replaced by Me/lone pair,
the lone pair being less “space consuming” than a
for C-Meth yl-1,2,3,4-tetr a h yd r oisoqu in olin es a t Room
a
Tem p er a tu r e
1
8
hydrogen atom. In contrast, the value for 2-methyl, 2.5
compound
C1
1-Me
22.94
C3
3-Me
22.60
C4
4-Me
20.84
-
1
kcal mol , is substantially larger, presumably because
the axial methyl group is considerably closer to the syn-
axial hydrogen atom [at C(6)] than in cyclohexane and
thus sterically more destabilized. (This results from the
shorter C-N distance, 147 pm, as compared to C-C, 153
pm.) However, the conformational energy of the 1-methyl
group in 1-MeTHIQ, 0.56 kcal mol (Table 1), is much
less than that in 2-methylpiperidine. In part, this is
presumably due to the absence of the second syn-axial
H of the piperidine (the benzene ring in THIQ has no
axial hydrogens).
1
3
4
-Me
-Me
-Me
51.99
48.98
49.19
53.03
51.33
52.32
52.43
42.19
49.67
51.58
49.50
43.16
48.97
49.60
30.52
37.69
32.63
38.82
38.11
33.39
33.36
cis-1,3-Me2
trans-1,3-Me2
cis-1,4-Me2
22.59
24.46
23.01
23.11
22.81
22.58
21.88
20.08
trans-1,4-Me2
-
1
a
In CD2Cl2 + 15% CHCldCCl2.
In most instances the difference is about 0.10 kcal mol-1
.
Another check is additivity of the data. In the confor-
mational analysis of saturated six-membered rings it is
generally assumed that conformational energies are
additive or subtractive, so that the conformational energy
of a disubstituted compound issdepending on whether
the substituents are cis or transseither the sum or the
difference of the conformational energies of the individual
substituents. While it cannot be assumed, a priori, that
this principle also holds for cyclohexenes or heteracyclo-
hexenes (since these compounds are more flexible and
lie in a less deep conformational well than their saturated
analogs), it does in fact apply quite satisfactorily to the
present data set. Thus, based on experimental data, the
additively calculated values for the disubstituted com-
However, if one assumes that a syn-axial Me/H inter-
action across a C-C-C juncture amounts to one-half of
the conformational energy of (axial) methyl in methyl-
cyclohexane19 or about 0.87 (1.74/2) kcal mol , the
experimental value is still much smaller than the ex-
pected 2.5-0.87 or ca. 1.6 kcal mol . There is clearly
another cause for the low conformational energy of Me(1),
and it would seem to be the peri effect;2 i.e., the
unfavorable steric interaction of the equatorial Me(1)
with the peri hydrogen at C(8) in the benzene ring.
To obtain a quantitative measure of the peri effect, we
carried out calculations on the conformational energies
-1
-
1
-1
pounds in kcal mol are as follows (experimental values
repeated in parentheses for convenience): cis-1,3, 2.19
0
(large); trans-1,3, 1.07 (1.18); cis-1,4, 0.88 (0.75); trans-
1
,4, 0.24 (0.22). The agreement strengthens one’s con-
fidence in the data set.
Other checks are only qualitative. The 13C chemical
shifts of the eight compounds studied are given in Table
3
,4
of C-methyl groups at the 2-, 5-, and 6-positions of ∆
tetrahydropyridine (5) (corresponding to the 1-, 4-, and
-
3
. The data are not extensive enough to calculate a
3
-positions in THIQ). The results are shown in Table 4.
Again, the values for the disubstituted compounds in
reliable parameter set for the ring carbons, and the
methyl shifts span too small a range to be quantitatively
useful. The cis-1,3-Me
the last column are properly additive, with a slightly
2
compound is all equatorial (Table
1
) and provides values for equatorial Me(1) and Me(3).
(16) Morin, F. G.; Horton, W. J .; Grant, D. M.; Dalling, D. K.;
The corresponding trans compound has largely but not
entirely equatorial Me(3); its slight upfield shift is
undoubtedly caused by the small contribution (12%) of
the Me(3a) and a similar upfield shift is seen in the 3-Me
compound (6% axial). A different situation arises, how-
ever, with Me(4): it is very largely axial in the cis-1,4-
Pugmire, R. J . J . Am. Chem. Soc. 1983, 105, 3992-3998.
(17) Booth, H.; Everett, J . R. J . Chem. Soc., Perkin Trans. 2 1980,
2
3
55-259.
(
18) Eliel, E. L.; Knoeber, M. C. J . Am. Chem. Soc. 1968, 90, 3444-
458.
(19) Actually the proper offset may be less, since the interaction
energy of axial Me(2) with C(4) is not only with the syn-axial H at
C(4) but also with the carbon atom itself. The latter interaction does
not go away in axial 1- or 3-methyl-THIQ. By the argument used in
the text, the conformational energy of methyl in 3-methyl-THIQ should
also be 1.6 kcal mol- . The good agreement of this value with the
2
Me compound (78%) which yet has the most downfield
Me; the 4-Me compound (63% axial) is more upfield and
the trans-1,4-compound (41% axial) even more so. A
similar seemingly anomalous situation is found in the
1
-1
experimental 1.63 kcal mol is not as pleasing as one might think,
since the most important steric interactionsthat between Me(3) and
axial H(1), across the short C-N distancessshould be diminished when
H(1) is pseudoaxial rather than axial [as is the corresponding H(6) in
piperidine (4)]. The conclusion from both findings is that the energy
advantage of methyl groups at positions 1a′, 3a, and 4a′ due to the
1
-methyl compounds where successive downfield shifts
of Me(1) are seen from the cis-1,3-Me compound (0%
axial) to the cis-1,4-Me (22% axial) and 1-Me (28% axial)
to the trans-1,4-Me (41% axial) to the trans-1,3-Me (88%
2
2
absence of a hydrogen substituent in positions 4a and 8a of THIQ
2
2
amounts to less than 0.87 kcal mol-
1
.
axial) compounds. It is evident that in these cases the
axial methyl groups resonate downfield of the equatorial
(20) See, for example: J ameson, M. B.; Penfold, B. R. J . Chem. Soc.
1965, 528-536.

Conformational Analysis
J . Org. Chem., Vol. 62, No. 26, 1997 9157
Ta ble 4. Com p u ted Con for m a tion a l Equ ilibr ia a n d F r ee
En er gy Differ en ces for Meth yl- a n d
5.1 Hz), 2.93 (ddd, 1H, J gem ) 12.4 Hz, J vic ) 8.7, 4.7 Hz), 2.80
(dddd, 1H, J gem ) 16.2 Hz, J vic ) 8.7, 4.7 Hz, J long range ) 1.1
Hz), 2.67 (br dt, 1H, J gem ) 16.2 Hz, J vic ) 4.7 Hz), 1.68 (s,
Dim eth yl-Su bstitu ted Tetr a h yd r op yr id in es
1
3
substituents(s)a
conformer ratiob
-∆G°, kcal mol-1
1H), 1.38 (d, 3H, J ) 6.7 Hz). C NMR (CD
δ 141.3 , 135.4 , 129.4 , 126.1 , 126.0 , 126.0
0.5 , 22.9 H NMR (D O + DCl): δ 7.37-7.26 (m, 4H), 4.67
quartet, 1H, J ) 6.8 Hz), 3.62 (ddd, 1H, J gem ) 13.0 Hz, J vic
6.6, 5.9 Hz), 3.44 (ddd, 1H, J gem ) 13.0 Hz, J vic ) 7.5, 5.7
Hz), 3.19 (br dt, 1H, J gem ) 17.6 Hz, J vic ≈ 6.7 Hz), 3.11 (dt,
2
Cl
2
+ CHCldCCl
2
):
3
9
5
5
6
1
, 51.9 , 42.1
9
9
,
1
3
4
-Me
-Me
-Me
87/13
93/7
1.13
1.53
0.72
large
0.53
0.42
1.88
1
3
2
4
.
2
(
)
77/23
100/0
71/29
67/33
96/4
c
cis-1,3-Me2
trans-1,3-Me2
cis-1,4-Me2
trans-1,4-Me2
c
1
H, J gem ) 17.6 Hz, J vic ) 6.2 Hz), 1.72 (d, 3H, J ) 6.8 Hz).
d
1
3
d
C NMR (D
28.6 , 53.7
2
O + DCl): δ 135.5 , 129.7
4
, 133.6
4
, 131.5
5
, 130.5
7
5
,
1
13
1
9
2
, 41.6 , 27.3 , 21.2 . The H and C NMR data
4
3
9
a
23
For easier comparison with Table 1, the THIQ numbering
system is used here. Properly the locants should be 2, 6, and 5.
are in excellent agreement with those reported.
-Met h yl-1,2,3,4-t et r a h yd r oisoq u in olin e. The corre-
sponding crude 3,4-dihydroisoquinoline (25.54 g, 0.176 mol)
was reduced with NaBH as described above: yield 21.32 g,
2.3%; bp 85-87 °C (0.5 mm), lit. bp 105-107 °C (7 mm).
3
b
d
c
e/a as calculated by MMP2(85) assuming ∆S ) 0. Me(3) a f e.
Me(1) a f e.
4
2
4
8
1
H NMR (CD
2
Cl
.96 (m, 1H), 4.02 (d, 1H, J gem ) 15.9 Hz), 3.96 (d, 1H, J gem
5.9 Hz), 2.94 (dd of quartet, 1H, J Me ) 6.3 Hz, J vic ) 10.3, 3.9
2
+ CHCldCCl ): δ 7.09-7.01 (m, 3H), 6.99-
2
6
1
)
Hz), 2.71 (dd, 1H, J gem ) 16.3 Hz, J vic ) 3.9 Hz), 2.43 (dd, 1H,
gem ) 16.3 Hz, J vic ) 10.3 Hz), 1.58 (br s, 1H), 1.17 (d, 3H, J
6.3 Hz). 13C NMR (CD
Cl + CHCldCCl ): δ 136.1 , 135.6
129.3 , 126.2 , 126.1 , 125.7 , 48.9 , 49.6 , 37.6
(D O + DCl): δ 7.39-7.25 (m, 4H), 4.44 (d, 1H, J gem ) 16.0
Hz), 4.39 (d, 1H, J gem ) 16.0 Hz), 3.59 (dd of quartet, 1H, J )
.5 Hz, J vic ) 10.7, 4.7 Hz), 3.14 (dd, 1H, J gem ) 17.5 Hz, J vic
4.7 Hz), 2.96 (dd, 1H, J gem ) 17.5 Hz, J vic ) 10.7 Hz), 1.53
d, 3H, J ) 6.5 Hz). 13C NMR (D
O + DCl): δ 133.7 , 131.4
, 52.2 , 46.5 , 34.8 , 20.2 . The
J
)
2
2
2
9
0
,
1
larger than desirable deviation for the trans-1,3 com-
pound where a relatively small difference between rela-
tively large numbers is involved. More to the point of
the discussion is the fact that, while the value for Me(3)
in Table 4 is the same as the calculated value for
5
5
4
8
8
7
9 0
, 22.6 . H NMR
2
6
)
(
2
2
0
,
3
-MeTHIQ in Table 1, the values for Me(1) and Me(4)
1
30.4
8
, 129.5
8
, 129.4
5
, 128.9
4
9
7
3
7
are much smaller for the THIQ derivatives shown in
1
H NMR spectrum is in very good agreement with that
-1
25
Table 1sby 0.68 and 0.94 kcal mol , respectively. These
numbers thus express the magnitude of the peri effect
in positions 1 and 4, respectively. [That the values differ
is presumably due to differences in Me-C-C-Hperi torsion
angles C(1) and C(4).] It is of particular note that Me(4)
in 4-MeTHIQ prefers the axial position; in this case the
axial Me has no syn-axial Me/H interaction but the
equatorial Me has a peri interaction. In contrast, for the
corresponding tetrahydropyridine, which lacks the peri
interaction, the equatorial conformer is yet preferred,
thus confirming the known fact21 that the syn-axial Me/H
interaction is not the only factor destabilizing axial
methyl.
reported except that the reported spectrum was not resolved
in the 2.4-3.5 ppm region.
4-Met h yl-1,2,3,4-t et r a h yd r oisoq u in olin e. The corre-
sponding crude 3,4-dihydroisoquinoline (0.41 g, 2.8 mmol) was
reduced with NaBH
4
as described above: yield 0.32 g, 77%;
2
3
bp 76-78 °C (airbath temperature) (2 mm), lit. bp 55-60 °C
1
(
0.1 mm). H NMR (CD
2
Cl
2
+ CHCldCCl ): δ 7.17 (d, 1H, J
2
)
7.3 Hz), 7.12 (td, 1H, J ) 7.4 Hz, J ) 1.6 Hz), 7.07 (td, 1H,
J ) 7.3, 1.6 Hz), 6.96 (d, 1H, J ) 6.6 Hz), 3.94 (d, 1H, J gem
16.2 Hz), 3.90 (d, 1H, J ) 16.2 Hz), 3.13 (dd, 1H, J gem
)
)
gem
12.3 Hz, J vic ) 4.8 Hz), 2.82 (br sextet, 1H, J vic ) 5.9 Hz), 2.74
(dd, 1H, J gem ) 12.3 Hz, J ) 5.9 Hz), 1.92 (s, 1H), 1.24 (d,
vic
1
3
3
1
H, J ) 6.9 Hz). C NMR (CD
2
Cl
2
+ CHCldCCl
2
): δ 140.8
, 20.8
2
,
.
36.3 , 128.5 , 126.3 , 126.2 , 125.8
7
0
9
7
6
, 51.5 , 49.1 , 32.6
8
9
3
4
1
H NMR (D
7.6 Hz), 4.42 (d, 1H, J gem ) 15.8 Hz), 4.38 (d, 1H, J gem
15.8 Hz), 3.64 (dd, 1H, J ) 12.5 Hz, J ) 5.6 Hz), 3.34
2
O + DCl): δ 7.46-7.31 (m, 3H), 7.26 (br d, 1H, J
)
)
Exp er im en ta l Section
gem
vic
(
sextet, 1H, J ≈7.0 Hz), 3.17 (dd, 1H, J gem ) 12.5 Hz, J vic ) 8.3
Gen er a l P r oced u r es. 1H NMR and 13C NMR spectra were
recorded at 250.13 or 399.92 MHz and 62.89 or 100.57 MHz,
13
Hz), 1.41 (dd, 3H, J ) 6.9 Hz, J long range ) 0.6 Hz). C NMR
O + DCl): δ 139.2 , 130.6 , 129.8 , 129.5 , 129.4 , 129.1
9.9 , 47.1 , 31.3 , 21.0
. The 1H and C NMR data are in
excellent agreement with those reported.
(
4
D
2
3
8
8
0
2
2
,
respectively, using TMS or (in D
2
O) DSS [3-(trimethylsilyl)-
13
2
7
9
2
1
-propanesulfonic acid, sodium salt] as internal standards.
2
3
Abbreviations used are s, singlet; d, doublet; t, triplet; dd,
doublet of doublets; m, multiplet; and br, broad. All 1,2,3,4-
tetrahydroisoquinolines were purified by preparative GC on
a 20% Carbowax 20M plus 10% KOH on Chromosorb A, 60/
cis- a n d tr a n s-1,3-Dim eth yl-1,2,3,4-tetr a h yd r oisoqu in -
olin e.26 The corresponding crude 3,4-dihydroisoquinoline
(
11.04 g, 0.069 mol) was reduced with NaBH
above to yield a diastereomeric mixture (6.86 g, 61.4%, c:t ≈
:1 by NMR); bp 119-123 °C (20 mm), lit.26 bp 125-130 °C
23 mm). The two isomers were separated by HPLC using
ethyl acetate/acetone (9/1) as solvent. The cis isomer was the
4
as described
8
0 mesh. Melting points are uncorrected.
-Meth yl-1,2,3,4-tetr a h yd r oisoqu in olin e. A solution of
the crude 3,4-dihydroisoquinoline (1.51 g, 10.4 mmol) and 0.55
g of NaBH in 50 mL methanol was refluxed for 1 h and
allowed to cool to room temperature.
removed under reduced pressure, and the product was taken
up in H O (20 mL) and extracted with Et
O (3 × 50 mL). The
organic layers were combined, dried (MgSO ), and concen-
trated. Kugelrohr distillation yielded the pure product (0.92
3
(
1
4
1
2
2
first product to be eluted from the column. H NMR (CD
CHCldCCl ): δ 7.17-7.05 (m, 3H), 7.03-7.00 (m, 1H), 4.09
quartet, 1H, J ) 6.5 Hz), 2.98 (dd of quartet, 1H, J ) 6.2 Hz,
vic ) 11.0, 3.6 Hz), 2.68 (dd, 1H, J gem ) 15.9 Hz, J vic ) 3.6
Hz), 2.49 (dd, 1H, J gem ) 15.9 Hz, J vic ) 11.0 Hz), 1.49 (br s,
2 2
Cl
The methanol was
+
(
J
2
2
2
4
1
3
2
3
1H), 1.40 (d, 3H, J ) 6.5 Hz), 1.18 (d, 3H, J ) 6.2 Hz).
NMR (CD
126.2 , 125.5
C
g, 60%): bp 76-79 °C (airbath temperature) (0.5 mm), lit.
2
Cl
2
+ CHCldCCl
2
): δ 140.8
, 22.8
3
, 135.8
, 22.5
7
, 129.3
H NMR (D O
1
, 126.2
5
,
bp 78-80 °C (0.06 mm).
1
1
0
7
, 53.0 , 49.5 , 38.8
3
0
2
1
9
.
2
H NMR (CD
2
Cl
2
+ CHCldCCl ): δ 7.11-7.02 (m, 4H), 4.03
quartet, 1H, J ) 6.7 Hz), 3.18 (dt, 1H, J gem ) 12.4 Hz, J vic )
2
+
DCl): δ 7.41-7.24 (m, 4H), 4.64 (quartet, 1H, J ) 6.8 Hz),
(
3
.59 (septet, 1H, J vic ) 5.8 Hz), 3.12 (dd, 1H, J gem ) 17.3 Hz,
(
21) Cf. Eliel, E. L.; Allinger, N. L.; Angyal, S. J .; Morrison, G. A.
Conformational Analysis; Wiley: New York, 1965; p 456.
22) Awe, W.; Wichmann, H.; Buerhop, R. Chem. Ber. 1957, 90,
997-2003.
23) Grunewald, G. L.; Sall, D. J .; Monn, J . A. J . Med. Chem. 1988,
1, 433-444.
(24) Nose, A.; Kudo, T. Chem. Pharm. Bull. 1984, 32, 2421-2425.
(25) Beugelmans, R.; Chastanet, J .; Roussi, G. Tetrahedron 1984,
(
1
40, 311-314.
(
(26) Bailey, D. M.; DeGrazia, C. G.; Lape, H. E.; Frering, R.; Fort,
D.; Skulan, T. J . Med. Chem. 1973, 16, 151-156.
3

9
158 J . Org. Chem., Vol. 62, No. 26, 1997
Olefirowicz and Eliel
J
1
vic ) 4.5 Hz), 2.98 (dd, 1H, J gem ) 17.3 Hz, J vic ) 12.0 Hz),
Ta ble 5. P r oton P a r a m eter s for C3 a n d C3-Me
2
7
13
.71 (d, 3H, J ) 6.8 Hz), 1.49 (d, 3H, J ) 6.4 Hz).
C NMR
9
, 128.1 ,
(P a r en th eses) In clu d in g Vicin a l H3-H4 Cou p lin g in th e
Meth yl-Su bstitu ted THIQ Hyd r och lor id e Sa lts a t Room
Tem p er a tu r e in D2O
(D
2
O + DCl): δ 135.1
3
, 134.1
, 20.6
6
, 131.5
.
3
, 130.7
5
, 130.0
0
5
5.3 , 52.8 , 35.8 , 20.9
5
7
6
4
2
The trans isomer was the second product to be eluted from
compound
δ3a
δ3e
J 3a3e J 3a4a′ J 3e4e′ J 3a4e′ J 3e4a′ J H-Me
the column but, due to the large amount of the cis isomer
present and the poor separation between the two compounds,
could not be obtained completely free of cis. The fractions
which contained the largest amounts of the trans isomer were
combined and reseparated by HPLC using the above condi-
1
3
4
-Me
-Me
-Me
3.44 3.62 13.0 7.55 5.85 5.89 6.6
3.59 (1.53)
3.17 3.64 12.5 8.25
3.59 (1.49) 12.05
3.27 3.58 12.0 8.33
10.70
4.65
5.58
4.49
5.08
5.73
(6.5)
(6.6)
cis-13-Me2
cis-14-Me2
trans-14-Me2 3.13 3.70 12.7 8.73
tions. This second separation afforded a 7:1 mixture of trans:
1
cis, suitable for NMR analysis. H NMR (CD
2
Cl
2
+ CHCld
CCl
2
): δ 7.12-7.01 (m, 4H), 4.17 (quartet, 1H, J ) 6.9 Hz),
Ta ble 6. Con for m a tion a l Equ ilibr ia a n d F r ee En er gy
Differ en ces for Meth yl- a n d Dim eth yl-Su bstitu ted
3
(
.22 (dd of quartet, 1H, J ) 6.2 Hz, J vic ) 10.0, 3.9 Hz), 2.73
a
dd, 1H, J gem ) 16.2 Hz, J vic ) 3.9 Hz), 2.39 (dd, 1H, J gem
)
1,2,3,4-Tetr a h yd r oisoqu in olin iu m Sa lts
1
6.2 Hz, J vic ) 10.0 Hz), 1.49 (br s, 1H), 1.39 (d, 3H, J ) 6.9
conformer ratiob -∆G°, kcal mol-1
1
3
Hz), 1.15 (d, 3H, J ) 6.2 Hz). C NMR (CD
2
Cl
, 125.9
2
+ CHCldCCl
2
):
substituent(s) exptl
calcdc
exptl
calcdc (free amined)
δ 140.9
8.1 , 24.4
cis- a n d tr a n s-1,4-Dim eth yl-1,2,3,4-tetr a h yd r oisoqu in -
3
, 135.2
5
, 129.4
5
, 127.0
6
, 126.1
5
4
, 53.3 , 43.1
3
6
,
3
1
6
, 22.5
8
.
1
3
-Me
-Me
1.37
6.92
1.82
large 499
f
0.56
11.8
1.18
0.19
1.15
0.35
large
f
-0.34
1.46
0.10
3.68
1.05
0.33
0.27
0.56
1.63
2
8
olin e. The corresponding crude 3,4-dihydroisoquinoline was
reduced with NaBH as described above to yield a diastereo-
meric mixture (13.75 g, 73.3%, c:t ≈ 2:1 by NMR); bp 127-
31 °C (21 mm). The two isomers were separated by prepara-
tive GC. The cis isomer was the first product to come off the
4-Me
-0.32
large
1.18
4
cis-1,3-Me2
trans-1,3-Me2
cis-1,4-Me2g
e
5.90
1.75
1.57
1
1.87
0.37
0.47
0.75
0.22
trans-1,4-Me2g 2.22
1
column. H NMR (CD
H), 4.00 (quartet, 1H, J ) 6.6 Hz), 3.05 (dd, 1H, J gem ) 12.6
Hz, J vic ) 4.5 Hz), 2.88 (dd, 1H, J gem ) 12.6 Hz, J vic ) 3.9 Hz),
2
Cl
2
+ CHCldCCl
2
): δ 7.12-7.07 (m,
a
Hydrochlorides (or deuteriochlorides) in D2O. b e/a. c From
4
MMP2(85) using the N ) 8 (NH2) parameter. Use of the N ) 39
+
(NH2 ) parameter gives about the same result for 4-Me and a
2
.77 (dd of quartet, 1H, J ) 7.1 Hz, J vic ) 4.5, 3.9 Hz), 1.79
-1
somewhat larger value (1.78 kcal mol ) for the 3-Me, but the
(
br s, 1H), 1.40 (d, 3H, J ) 6.6 Hz), 1.27 (d, 3H, J ) 7.1 Hz).
-1
result for 1-Me (+0.85 kcal mol ) reverses the e/a ratio; the
1
3
C NMR (CD
26.2 , 126.1 , 126.0
O + DCl): δ 7.45-7.31 (m, 4H), 4.70 (quartet, 1H, J ) 6.9
2
Cl
2
+ CHCldCCl
2
): δ 141.0
1
, 140.9
, 21.8
7
, 128.9
0
,
experimental value lies between the two calculated values. d Ex-
. 1H NMR
e
1
9
4
6
, 52.3 , 48.9 , 33.3
2
7
9
, 23.0
1
8
perimental values for free amine from Table 1 for comparison. 3a
f 3e. f Not determined, coupling was not resolved. g 1a f 1e.
(D
2
Hz), 3.58 (dd, 1H, J gem ) 12.0 Hz, J vic ) 5.1 Hz), 3.32 (br sextet,
obtains the conformer ratios and conformational energy
values shown in Table 6. Also shown in this table are
the values calculated by MMP2(85).
The following observations may be made (1) Additivity
for the 1,4-dimethyl-substituted compounds is only mod-
1
1
H, J vic ) 7.0 Hz), 3.27 (dd, 1H, J gem ) 12.0 Hz, J vic ) 8.4 Hz),
13
.76 (d, 3H, J ) 6.9 Hz), 1.45 (d, 3H, J ) 6.7 Hz). C NMR
(
5
D
2
O + DCl): δ 138.9
4
, 135.0
, 20.9
7
, 130.8
.
3 1 1 5
, 129.7 , 129.7 , 128.6 ,
3.5 , 46.7 , 31.5 , 21.4
4
2
0
6
2
The trans isomer was the second product to come off the
1
-1
column. H NMR (CD
H), 7.14-7.08 (m, 3H), 4.05 (quartet, 1H, J ) 6.7 Hz), 3.24
dd, 1H, J gem ) 12.5 Hz, J vic ) 5.2 Hz), 2.87 (dd of quartet,
H, J ) 7.0 Hz, J vic ) 7.6, 5.2 Hz), 2.63 (dd, 1H, J gem ) 12.5
2
Cl
2
+ CHCldCCl
2
): δ 7.22-7.18 (m,
erately good: 0.54 vs 0.47 kcal mol observed for the
1
(
1
trans, 0.16 vs 0.37 for the cis. The situation is much
worse for the computed values: for N ) 8 (see footnote c
to Table 6), for example, in the cis compound, the 1a,4e
conformer should predominate whereas the 1e,4a is the
calculated preferred conformation. We believe that lack
of specific consideration of solvation of the axial hydrogen
of the charged NH
unreliable. (2) The experimentally determined prefer-
ence of 1-Me and 3-Me for the equatorial position is
considerably less in the salt than in the free amine. A
similar observation was made in N,2-dimethylpiperidine
Hz, J vic ) 7.6 Hz), 1.38 (d, 3H, J ) 6.7 Hz), 1.22 (d, 3H, J )
1
3
7
.0 Hz). C NMR (CD
2
Cl
2
+ CHCldCCl
, 52.4 , 49.6
2
): δ 141.0
1
, 140.7
, 20.0
3
,
.
1
27.9 , 126.3 , 126.1 , 125.9
8
2
2
3
3
0
, 33.3 , 23.1
6
1
8
1
H NMR (D O + DCl): δ 7.44-7.31 (m, 4H), 4.71 (quartet,
2
+
1
3
H, J ) 6.8 Hz), 3.70 (dd, 1H, J gem ) 12.7 Hz, J vic ) 5.7 Hz),
.39 (br sextet, 1H, J vic ) 7.2 Hz), 3.13 (dd, 1H, J gem ) 12.7
2
moiety makes the calculated values
Hz, J vic ) 8.7 Hz), 1.72 (d, 3H, J ) 6.7 Hz), 1.39 (d, 3H, J )
1
3
7
1
.1 Hz).
C NMR (D
2
O + DCl): δ 138.8
2
, 134.9
0 3
, 130.7 ,
8
29.7 , 128.4
0
6
, 130.0 , 54.3
6
8
, 48.2 , 31.6 , 21.4
9
2
5
, 21.3 .
7
and its salt and was ascribed to the longer C-N bond
Ap p en d ix: 1,2,3,4-Tetr a h yd r oisoqu in olin iu m
Sa lts
in the salts vs the free amines, which increases the
distance between an axial methyl group and the axial
hydrogen on the other side of the nitrogen. (The alter-
We add here some data on methyl- and dimethyl-
7
nate solvation argument made for the N-methyl com-
1
,2,3,4-tetrahydroisoquinolinium salts. The proton-
pounds probably does not apply to the NH species.) (3)
In the case of 4-MeTHIQ, whereas the axial conformer
is preferred for the free amine because of the peri effect,
the equatorial conformer is preferred for the salt. This
is perhaps as expected, since where the free amine has
a syn-axial lone pair on nitrogen, the salt, instead, has a
hydrogen atom probably further swelled by solvation.
Thus the N/4 syn-axial interaction is much enhanced in
the salt, to the extent that it outweighs the peri effect.
proton coupling constants [J H3/H4(trans)] are listed in
Table 5. As discussed earlier, the sum of J H3a/H4a and
J
1
H3e/H4e is constant and, from the proton spectrum of the
-methyl compound, is found to be 13.40 Hz. If it is
2
assumed that the cis-1,3-Me compound exists exclusively
in the conformation having both methyl groups equato-
rial, J H3/H4(trans) for this compound is J H3a/H4a or 12.05
Hz when J H3e/H4e is 13.40-12.05 or 1.35 Hz. With these
value and the averaging calculation described above, one
Ack n ow led gm en t. This work was supported by
NSF Grants CHE-8314160 and CHE-8703060.
(
27) Gray, N. M.; Cheng, B. K.; Mick, S. J .; Lair, C. M.; Contreras,
P. C. J . Med. Chem. 1989, 32, 1242-1248.
28) Govindan, C. K.; Taylor, G. J . Org. Chem. 1983, 48, 5348-5354.
(
J O971257R