Substituent effects on 15N and 13C NMR chemical shifts of 3-phenylisoxazoles: A theoretical and spectroscopic study

Article abstract of DOI:10.1002/mrc.1860

The synthesis and assignment of ¹⁵N and ¹³C NMR signals of the isoxazole ring in a series of parasubstituted 3-phenyl derivatives are reported. DFT calculations of ¹⁵N and ¹³C chemical shifts are presented and c

Full text of DOI:10.1002/mrc.1860

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
MAGNETIC RESONANCE IN CHEMISTRY
Magn. Reson. Chem. 2006; 44: 851–855
Published online 27 June 2006 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/mrc.1860
Substituent effects on 15N and 13C NMR chemical shifts
of 3-phenylisoxazoles: a theoretical and spectroscopic
study
Mark H. Schofield,∗ Marie-Adele Sorel, Ryan J. Manalansan, David P. Richardson∗ and
J. Hodge Markgraf
Department of Chemistry, Williams College, Williamstown, MA 01267-2692, USA
Received 25 January 2006; Revised 27 April 2006; Accepted 2 May 2006
The synthesis and assignment of 15N and 13C NMR signals of the isoxazole ring in a series of para-
substituted 3-phenyl derivatives are reported. DFT calculations of 15N and 13C chemical shifts are
presented and compared to observed values. Substituent effects are interpreted in terms of the Hammett
correlation and calculated bond orders. Copyright  2006 John Wiley & Sons, Ltd.
KEYWORDS: 13C NMR; 15N NMR; chemical shifts; HMBC; DFT calculations; natural resonance theory; electronic effects;
Hammett correlation; 3-phenylisoxazoles
methods using the B3LYP functional and 6–311CCG(2d,2p)
basis set, which have been shown to be effective for
the prediction of 13C and 15N chemical shifts,6 were also
employed. The theoretical component of this analysis also
permitted commentary on bond orders of the isoxazole
ring, a topic related to prior reports of electron densities
of isoxazoles.7
INTRODUCTION
The 3-phenylisoxazole (1,2-oxazole) unit constitutes an
important pharmacophore in a wide range of biologically
and agriculturally active compounds.1 Most of the reports
assessed the impact of a series of substituents on the desired
activity. Although substituent effects on 13C NMR spectral
assignments for 3-phenylisoxazoles (1) are unknown, such
correlations have been reported for related heterocycles: 3,5-
diphenylisoxazoles,2 2-phenyl-1,3,4-oxadiazoles,3 3-phenyl-
1,2,4-oxadiazoles,4 and 3-phenyl-1,2,5-oxadiazoles.5 In par-
ticular, the extensive studies of Baumstark and coworkers
on 3,5-diphenylisoxazoles established that para substituents
on the 5-phenyl ring (2a) significantly affected the chemical
shifts of isoxazole C-4 and C-5, while the same substituents
on the 3-phenyl ring (2b) exerted minimal effect on either of
these isoxazole ring carbons.2 Those studies did note, how-
ever, that C-3 of 2b was shielded by electron-withdrawing
substituents. The same effect was observed for C-2 of
oxadiazole 3 (see Fig. 1).3
In the NMR studies of the various oxazoles cited above,
only for 3 were 15N spectral assignments made.3 Those data
established a significant correlation for N-3 with electron-
releasing substituents. Given the comparable electronic
relationship between Y and N-2 and N-3 in 1 and 3,
respectively, and between X and C-4 in 2a, an NMR study of
a series of para-substituted 3-phenylisoxazoles (1a–1i) was
undertaken to probe the transmission of substituent effects.
To compare theory with experiment, density functional
RESULTS AND DISCUSSION
Syntheses
Conversion of para-substituted benzaldehyde oximes (4a–4i)
to 3-phenylisoxazoles (1a–1i) was carried out by a one-
pot procedure (Scheme 1). The 1,3-dipolar cycloaddition
of benzonitrile N-oxides (5) to alkynes or to substituted
alkenes (as alkyne equivalents) is well documented.8 The
present method used phenyl vinyl sulfoxide and a solvent
at the boiling point of which intramolecular elimination of
phenylsulfenic acid occurred.9 This sequence constituted
a concise route to 3-phenylisoxazoles. The efficiency of
measuring 15N chemical shifts was increased by enhancing
the 15N content of 1a–1i to ca 20%.
15N NMR chemical shift analysis
The calculated and experimentally determined 15N chemical
shifts for our series of compounds, 1a–1i, are listed in Table 1,
arranged according to ꢀp values.10 As shown, the chemi-
cal shifts span a fairly narrow range (calculated: 11.0 ppm;
experimental: 13.3 ppm), with a trend in shifts that is con-
sistent with the electron-releasing or electron-withdrawing
ability of the phenyl para substituent. As expected, electron-
withdrawing groups cause a downfield shift (i.e., to less
negative values, since CH3NO2 is the chemical shift ref-
erence) while electron-releasing groups lead to an upfield
ŁCorrespondence to: Mark H. Schofield, Department of Chemistry,
Haverford College, 370 Lancaster Avenue, Haverford, PA
19041-1336, USA. E-mail: mschofie@haverford.edu
David P. Richardson, Department of Chemistry, Williams College,
47 Lab Campus Drive, Williamstown, MA 01267-2692, USA.
E-mail: david.p.richardson@williams.edu
Copyright  2006 John Wiley & Sons, Ltd.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
852
M. H. Schofield et al.
Figure 1. Phenylisoxazoles (1, 2) and 2-phenyl-1,3,4-oxadiazole.
Scheme 1. Synthesis of 3-phenylisoxazoles (1a–1i). Reagents: (a) N-chlorosuccinimide, pyridine, 1,1,2-trichloroethane (TCE); (b) Et3N,
TCE; (c) Ph–SO–CH CH2, TCE.
Table 1. Experimental and calculated 15N NMR chemical shift
data (υ, relative to CH3NO2 D 0.00 ppm) Hammett substituent
constants and natural atomic charges of N-2 for isoxazoles
1a–1i
shift. Similar effects have been observed for both N-3 and
N-4 in 2-phenyl-1,3,4-oxadiazoles.3 Although the calculated
chemical shifts are consistently higher than the experimental
values (υave D 13.6), their variation (υst.dev. D 1.3 ppm)
is small. This small systematic deviation, which is related
to the calculated chemical shift of the reference compound
(CH3NO2), is in agreement with the study by Cheeseman
et al.6 and does not compromise the excellent correlation
between theory and experiment. Moreover, the calculated
natural charge of the isoxazole ring nitrogen (N-2) is entirely
consistent with the increasing electron donating ability of
the phenyl ring substituent. Thus, the calculated charge on
N-2 mirrors the observed (and calculated) chemical shifts
nicely.
The data in Table 1 also display the expected response
of the 15N chemical shifts to ꢀp Hammett values. Graphical
analysis of the experimental shifts gave a trend line with a
positive slope (ꢁ D 9.18) and excellent linear correlation
(r2 D 0.9853). The strongly positive ꢁ value is entirely
consistent with our observation that 15N chemical shifts
are highly sensitive to the electronic nature of the phenyl
Natural
charge
a
Compd.
Y
ꢀp
Calcd.
Expt.
1a
1b
1c
1d
1e
1f
1g
1h
1i
NMe2
OMe
Me
H
F
Cl
Br
CF3
NO2
ꢀ0.83
ꢀ0.27
ꢀ0.17
0.00
0.06
0.23
0.23
0.54
0.78
0.43
2.74
5.52
6.50
4.86
6.47
6.54
9.21
11.47
ꢀ14.50
ꢀ10.28
ꢀ10.05
ꢀ9.07
ꢀ7.79
ꢀ6.68
ꢀ6.54
ꢀ3.05
ꢀ1.24
ꢀ0.164
ꢀ0.161
ꢀ0.154
ꢀ0.147
ꢀ0.150
ꢀ0.146
ꢀ0.146
ꢀ0.144
ꢀ0.134
a Ref. 10.
para substituent and they were congruent with analogous
measurements reported for 3.3
Copyright  2006 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2006; 44: 851–855
DOI: 10.1002/mrc

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Substituent effects on 15N and 13C NMR chemical shifts of 3-phenylisoxazoles 853
Table 2. Experimental and calculated 13C NMR chemical shift data (υ, relative to CDCl3 D 77.00 ppm) and
Hammett substituent constants of C-3, C-4, and C-5
C-3
C-4
C-5
a
Compd.
Y
ꢀp
Calcd.
Expt.
Calcd.
Expt.
Calcd.
Expt.
1a
1b
1c
1d
1e
1f
1g
1h
1i
NMe2
OMe
Me
H
F
Cl
Br
CF3
NO2
ꢀ0.83
ꢀ0.27
ꢀ0.17
0.00
0.06
0.23
0.23
0.54
0.78
168.03
167.81
168.15
168.56
167.42
167.35
167.33
167.39
166.93
161.46
161.09
161.42
161.48
163.82
160.45
160.63
160.39
159.82
103.94
104.28
104.49
104.87
104.52
104.58
104.56
104.98
105.18
101.93
102.18
102.33
102.42
102.32
102.26
102.31
102.5
164.18
164.84
164.89
165.08
165.35
165.51
165.54
165.95
166.36
158.19
158.62
158.68
158.85
159.02
159.07
159.17
159.43
159.77
102.66
a Ref. 10.
13C NMR chemical shift analysis
increases for phenyl ring substituents that are either strongly
releasing or withdrawing (e.g. NMe2, NO2), but the bond
order decreases for less-polarizing substituents (e.g. H).
Consistent with this model, the N-2/C-3 bond order follows
a different trajectory with the lowest bond order observed
for the most polarizing phenyl substituent (1a). Finally,
the O-1/N-2 bond order increases monotonically with the
electron-withdrawing ability of the phenyl substituent. In
other words, the isoxazole ring has the capacity to attenuate
strongly polarizing substituents bound at the isoxazole C-3
position. On the basis of these results, the isoxazole C-3 atom
should be (and is) relatively insensitive to the nature of the
phenyl para substituent, while the isoxazole N-2 atom should
be (and is) sensitive to the nature of the phenyl substituent.
The corresponding 13C chemical shifts for compounds 1a–1i
are presented in Table 2. We found that the calculated
and experimental ranges of chemical shifts agreed closely.
However, the correlations with Hammett ꢀp constants for
phenyl para substituents were weak (r2: C-3, 0.1883; C-4,
0.8432; C-5, 0.7194), which we attribute to the fact that the
narrow range of 13C chemical shifts observed is similar to
the variance in the calculations. At first glance, it may seem
unusual that the 15N chemical shift is sensitive to the phenyl
para substituent, while the 13C chemical shifts, particularly
at C-3, are not. In an effort to understand the chemical shift
data, the Wiberg bond indices for the isoxazole ring were
calculated (Table 3).
For the series of compounds examined, the O-1/N-2
bond order increases with the electron-withdrawing ability
of the phenyl para substituent. This trend is in agreement
with the resonance model shown in Fig. 2, in which excess
electron density provided by electron-releasing groups is
localized on the isoxazole N-2 atom (structure I), whereas
electron density removed by electron-withdrawing groups
is offset by the O-1 lone pair (structure III). Consistent with
this model, the calculated natural charge of the isoxazole
N-2 atom decreases with increasing electron release of
the phenyl para substituent (Table 1). Furthermore, this
model also anticipates the results of the Wiberg bond index
calculations. As shown in Table 3, the C-3/C-10 bond order
EXPERIMENTAL
Melting points were determined on a MelTemp apparatus
and were uncorrected. Extracts were dried over Na2SO4, and
solvents were removed by rotary evaporation at reduced
pressure. Silica gel 60 (230–400 mesh) was used for flash
liquid chromatography; solvent systems for elution were
hexane-ethyl acetate (9 : 1 to 4 : 1 v/v). Product purities
were determined by gas chromatography-mass spectrometry
analysis on a Hewlett-Packard HP 6890 system equipped
with a HP-5MS cross-linked diphenyl (5%) dimethyl (95%)
polysiloxane capillary column (30 m ð 0.25 mm ð 0.25 µm
Table 3. Calculated Wiberg bond orders of isoxazoles 1a–1i
a
Compd.
Y
ꢀp
O-1/N-2
N-2/C-3
C-3/C-4
C-4/C-5
C-5/O-1
C-3/C-10
1a
1b
1c
1d
1e
1f
1g
1h
1i
NMe2
OMe
Me
H
F
Cl
Br
CF3
NO2
ꢀ0.83
ꢀ0.27
ꢀ0.17
0.00
0.06
0.23
0.23
0.54
0.78
1.043
1.048
1.052
1.054
1.054
1.056
1.057
1.061
1.067
1.569
1.572
1.575
1.581
1.579
1.579
1.578
1.573
1.574
1.207
1.212
1.213
1.217
1.217
1.218
1.218
1.218
1.221
1.639
1.636
1.636
1.635
1.634
1.634
1.634
1.633
1.632
1.104
1.105
1.104
1.104
1.105
1.104
1.104
1.104
1.103
1.064
1.059
1.056
1.054
1.056
1.056
1.056
1.053
1.057
a Ref. 10.
Copyright  2006 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2006; 44: 851–855
DOI: 10.1002/mrc

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
854
M. H. Schofield et al.
Figure 2. Resonance effects in 3-phenylisoxazoles.
film), a 5973 mass selective detector, and a HP Kayak XA
computer. HRMS analyses were performed by the Nebraska
Center for Mass Spectrometry at the University of Nebraska-
Lincoln.
chemical shifts, some spectra were measured prior to the
chromatographic removal of 7.
NMR spectroscopy
NMR spectra were measured at 298 K with a Bruker Avance
DRX 500 MHz NMR spectrometer operating at frequencies
of 500.630 (1H), 125.884 (13C), and 50.748 (15N) using a 5-mm
Compounds
Preparation of oximes (4a–4i)
Oximes were prepared by a standard procedure employ-
ing condensation of a para-substituted benzaldehyde with
hydroxylamine hydrochloride (6).11 Oximes 4a–4i have been
reported previously; their physical constants and spectral
data matched literature values.12–17
inverse gradient (TXI) probehead (90 pulse widths: 1H,
°
7.4 µs; 13C, 12.5 µs; and 15N, 40.0 µs).
Chemical shifts for 1H and 13C (ppm) were measured
relative to internal Me4Si; 15N chemical shifts were measured
relative to an external solution of CH3NO2/CDCl3 (1 : 2,
v : v) in a 1-mm diameter coaxial insert tube. Samples for
Preparation of 3-phenylisoxazoles (1a–1i). General
procedure
To a solution of N-chlorosuccinimide (0.267 g, 2.00 mmol)
and pyridine (0.010 ml, 0.124 mmol) in TCE (2.1 ml) was
added 4 (2.00 mmol) and the solution was stirred at 50 C
1
measurement of H and 13C chemical shifts were prepared
in CDCl3; 15N NMR samples, which were ca 20% enriched in
15N, were prepared with concentrations ranging from 12 to
64 mg per 400 µl of CDCl3 as solvent.
°
Two-dimensional 1H,15N gradient selected (heteronu-
clear multiple-bond coherence) HMBC experiments were
performed to observe 15N chemical shifts using a stan-
dard pulse sequence (hmbcgpndqf) from the Bruker
pulse sequence library. The 1H,15N correlation observed
for 30 min. A solution of phenyl vinyl sulfoxide (0.335 g,
2.20 mmol) and triethylamine (0.29 ml, 2.01 mmol) in TCE
(0.35 ml) was added dropwise via Pasteur pipet over 3 min,
and the solution was stirred at reflux for 30 min. The reaction
mixture was evaporated to dryness and the residue was
treated with 2M NaOH (10 ml). The mixture was stirred
at reflux for 30 min, neutralized with saturated NH4Cl
solution, and extracted with CH2Cl2. The extract was washed
with water, dried, and evaporated to give crude product.
Purification by flash liquid chromatography removed Ph2S2
(7) coproduct and afforded 1 in >99% purity. Isoxazoles
1b–1g, 1i have been reported previously; their physical
constants and spectral data matched literature values.18–25
1
(3Jꢂ1H,15Nꢃ) was that between the isoxazole N-2 and the H
at C-4 (chemical shift υ 6.6–6.8). For optimal simultaneous
2
observation of both the Jꢂ1H,15Nꢃ correlations in CH3NO2,
and the 3Jꢂ1H,15Nꢃ correlations in our analytes, a value of 8 Hz
was used for the parameter CNST13, which sets the pulse
sequence delay for evolution of multiple-bond couplings.
1
Spectral windows were set at 6 ppm in H and 100 ppm in
15N. A total of 16 scans of 1024 data points were collected
along the t2-axis and then zero-filled to 2048 points prior to
Fourier transformation (FT). In the indirect (t1) dimension,
128 time increments were collected and zero-filled to 1024
prior to FT. Total acquisition time per sample was ca 1 h.
Chemical shift measurement accuracy was estimated to be
better than š0.1 ppm.
°
3-(4-Dimethylaminophenyl)isoxazole (1a). m.p. 120.5–121.5 C.
HRMS for C11H12N2O [M]C requires 188.0950. Found:
188.0944.
°
3-(4-Trifluoromethylphenyl)isoxazole (1h). m.p. 98.0–98.55 C.
HRMS for C10H6F3NO [M C H]C requires 214.0480. Found:
Computational methods
214.0474.
For the preparation of 15N-enriched 1, natural-abundance
6 was mixed (4 : 1 w/w) with 15N-enriched 6 (Cambridge
Isotope Laboratories, Inc.; 98.6% 15N) to give ca 20% 15N
enrichment. To maximize the efficiency of obtaining 15N
All computations were carried out using Gaussian 03
program26 and employed the B3LYP functional.27 Substi-
tuted 3-phenylisoxazoles were optimized using the very tight
convergence criteria in the gas phase (the 6–311CCg(2d,2p)
Copyright  2006 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2006; 44: 851–855
DOI: 10.1002/mrc