Synthesis and conformational analysis of saturated 3,1,2- benzoxazaphosphinine 2-oxides

Article abstract of DOI:10.1002/ejoc.200400611

N-Unsubstituted, N-methyl and N-benzyl cis- and trans-2-(hydroxymethyl) cyclohexylamines were subjected to ring closure with phenylphosphonic dichloride, phenyl dichlorophosphate and bis(2-chloroethyl)phosphoramidic dichloride in order to synthesize P-epi

Full text of DOI:10.1002/ejoc.200400611

FULL PAPER
Synthesis and Conformational Analysis of Saturated
3,1,2-Benzoxazaphosphinine 2-Oxides
Henri Kivelä,[a] Zita Zalán,[a,b] Petri Tähtinen,[a] Reijo Sillanpää,[c] Ferenc Fülöp,*[b] and
Kalevi Pihlaja*[a]
Keywords: Phosphorus heterocycles / Fused-ring systems / Structure elucidation / Configuration determination / Confor-
mation analysis
N-Unsubstituted, N-methyl and N-benzyl cis- and trans-2-
(hydroxymethyl)cyclohexylamines were subjected to ring
closure with phenylphosphonic dichloride, phenyl dichlo-
rophosphate and bis(2-chloroethyl)phosphoramidic dichlo-
ride in order to synthesize P-epimeric diastereomers of the
corresponding unsubsituted and N-substituted 2-phenyl-, 2-
phenoxy- and 2-[bis(2-chloroethyl)amino]octahydro-2H-
3,1,2-benzoxazaphosphinine 2-oxides. The stereochemistry
and conformations of the prepared compounds were ana-
lyzed mainly by variable-temperature1H, 13C and 31P NMR
spectroscopy. Geometry optimizations were performed for
some trans-fused molecules by utilizing the B3LYP DFT
ling constant calculations at the UB3LYP/cc-pVTZ level of
theory. The population of the N-out conformation of the cis-
fused compounds was found to increase with increasing ste-
ric size of the N substituent (H Ͻ Me Ͻ CH2Ph), and also to
be strongly influenced by the configuration at the phospho-
rus atom. The heteroring of the trans-fused compounds pre-
dominantly adopts a chair conformation with the notable ex-
ception of the (2R*,4aS*,8aS*) epimers of the 2-phenoxy-
substituted compounds, for which the axial tendency of OPh
causes the heteroring to adopt mainly B2,4a- and/or 1S4a-type
nonchair conformations.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
method and a locally dense basis set, followed by JH,P coup- Germany, 2005)
ane moiety have recently been reported to possess matrix
Introduction
metalloproteinase-inhibitory[5] and pesticide activities.[6]
Conformational analyses of monocyclic 1,3,2-oxazaphos-
phinanes have revealed numerous interesting details con-
cerning the conformational properties of the heteroring.[7]
However, the synthesis and conformations of cycloalkane-
condensed 1,3,2-oxazaphosphinanes have been studied less
extensively. Their conformational behavior may differ from
that of cycloalkane-fused 1,3-heterocycles which often
adopt a biased chair–chair conformation in trans-fused de-
rivatives and two interconvertible chair–chair conforma-
tions in the cis isomers.[8] As with bicyclic heterocycles,[8]
the orientation of the substituents in the 2-position, and
also the substituent on the nitrogen atom when it is next to
the ring annulation position, liekly has a significant influ-
ence on the conformers that predominate.
Conformational analyses of various 2-aryloxy- and 2-
bis(2-chloroethyl)amino-substituted cis- and trans-perhy-
dro-1,3,2-benzoxazaphosphinine 2-oxides have revealed
that the heteroring in these compounds can exist not only
in a chair, but also in some unusual skew conforma-
tions.[9,10] The chair/skew and O-in/O-out conformational
equilibria in these compounds are strongly influenced by
the nature of the ring’s N substituent.[10]
The synthesis and transformations of 1,3,2-oxazaphos-
phinane 2-oxide derivatives have been thoroughly studied
in recent decades as a consequence of their potential wide-
ranging synthetic applications and the therapeutic impor-
tance of this type of compounds. Alkylation and condensa-
tion reactions of P-benzyl-1,3,2-oxazaphosphinane 2-oxides
and nucleophilic addition to vinylphosphonate derivatives
that contain a 1,3,2-oxazaphosphinane ring are effective
methods for the stereoselective formation of single carbon–
carbon bonds.[1,2]
The 1,3,2-oxazaphosphinane ring system is found in al-
kylating anticancer drugs (e.g. cyclophosphamide and ifos-
famide),[3] numerous derivatives of which have been synthe-
sized in order to determine their structure–activity relation-
ships.[4] Compounds that contain the 1,3,2-oxazaphosphin-
[a] Department of Chemistry, University of Turku,
20014 Turku, Finland
Fax: +358-2-333-6750
E-mail: kpihlaja@utu.fi
[b] Institute of Pharmaceutical Chemistry, University of Szeged,
P. O. Box 121, 6701 Szeged, Hungary
E-mail: fulop@pharma.szote.u-szeged.hu
[c] Department of Chemistry, University of Jyväskylä,
P. O. Box 35, 40351 Jyväskylä, Finland
E-mail: resillan@jyu.fi
Of the regioisomeric cis- and trans-perhydro-3,1,2-benz-
oxazaphosphinine 2-oxides, only the 1-benzyl-2-phenyl-sub-
stituted derivatives have been reported in the literature,[11]
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2005, 1189–1200
DOI: 10.1002/ejoc.200400611
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1189

H. Kivelä, Z. Zalán, P. Tähtinen, R. Sillanpää, F. Fülöp, K. Pihlaja
FULL PAPER
and their solution-state conformations have not been ana- prepared in order to allow a comparison of the conforma-
lyzed. As a continuation of our previous studies on cycloal- tional effects of the different N substituents.
kane-condensed 1,3,2-oxazaphosphinane 2-oxides,[10] the
effect of P and N substituents on the stereochemistry of
ring-closure reactions and the conformations of the title
compounds are discussed in this paper.
Characterization of Structures
The structures formed in the cyclization reactions were
characterized by solution-state multinuclear (1H, 13C and
31P) NMR methods and were supported by DFT calcula-
tions on compounds 8, 10, 14 and 16 and the X-ray crystal
structure determined for 16a. The 1H and 13C chemical
shifts were assigned with the aid of 1D NOESY or NOE
Results and Discussion
Synthesis
The amino alcohols 1–6 were prepared according to the difference and 2D homonuclear (dqf-COSY) and heteronu-
procedures described in the literature.[12] Compounds 1–6 clear (HETCOR, COLOC) correlation experiments. The
were cyclized with phenylphosphonic dichloride, phenyl proton chemical shifts and coupling constants were deter-
dichlorophosphate and bis(2-chloroethyl)phosphoramidic mined by an iterative analysis with PERCH NMR soft-
1
dichloride at ambient temperatures by using a procedure ware.[13] For the cis-fused compounds, the H spin systems
similar to that described earlier[10a] to give 1,4,4a,5,6,7,8,8a- could be solved completely, whereas the fine structure of
octahydro-2H-3,1,2-benzoxazaphosphinine 2-oxides 7–20 the -(CH2)4- proton resonances of the trans-fused com-
(Scheme 1 and Scheme 2). In most cases, two diastereomers pounds was often poorly reproduced by the simulations be-
(a and b), which differ in their configuration at the phos- cause of the difficulties arising from overlapping signals and
phorus atom, were formed and were separated by column the complexity of the coupling network. Comprehensive
chromatography. The ring-closure reactions of amino NMR data are presented in the Supporting Information.
alcohols 1 and 4 with phenylphosphonic dichloride pro-
In the configurational and conformational analysis of
ceeded with complete diastereoselectivity to give compound certain 1,3,2-benzoxaza- and -dioxaphosphinine derivatives
7a and 10b, respectively, as single diastereomers; no traces it has been possible to use the 31P chemical shifts.[9,10] How-
of the corresponding P epimers (7b or 10a) were detected, ever, for the 2-Ph- and 2-OPh-substituted compounds
even in the crude products. Compound 20 was obtained as studied here, the differences in δP between epimers were al-
a 45:55 mixture of 20a and 20b. The known 1-benzyl-2- ways less than 3 ppm. Only in case of the 2-N(CH2CH2Cl)2
phenyl-substituted derivatives 11a,b and 12a,b[11] were also derivatives were distinctive ∆δP values found; epimers with
Scheme 1. Reagents: (i) Cl2POPh/Et3N in THF, column chromatography, 8–45%; (ii) Cl2PO(OPh)/Et3N in THF, column chromatography,
14–37%; (iii) Cl2PO[N(CH2CH2Cl)2]/Et3N in THF, column chromatography, 30–35%. R = H: 1, 7, 13; R = Me: 3, 9, 15; R = CH2Ph:
5, 11, 17.
1190
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1189–1200

Saturated 3,1,2-Benzoxazaphosphinine 2-Oxides
FULL PAPER
Scheme 2. Reagents: (i) Cl2POPh/Et3N in THF, column chromatography, 10–37%; (ii) Cl2PO(OPh)/Et3N in THF, column chromatog-
raphy, 16–42%; (iii) Cl2PO[N(CH2CH2Cl)2]/Et3N in THF, column chromatography, 63%. R = H: 2, 8, 14; R = Me: 4, 10, 16; R =
CH2Ph: 6, 12, 18.
1
Table 1. Selected H NMR chemical shifts [ppm] of the trans-fused compounds.[a]
Proton
8a
8b
10b
12a
12b
14a
14b
16a
16b
18a
18b
20a[b]
20b[b]
4-Hax
4-Heq
4a-H
8a-H
4.22
4.08
1.78
3.27
3.64
4.17
1.76
2.71
3.79
4.07
2.03
3.00
4.12
4.04
1.87
2.85
3.81
4.15
2.04
3.14
4.07
4.22
1.85
3.10
4.07
4.20
1.71
2.99
4.06
4.13
2.04
2.77
4.03
4.09
1.90
2.90
3.93
4.11
2.24
2.66
4.08
4.12
1.90
3.18
3.87
3.97
1.43
2.90
3.77
4.03
1.51
2.74
[a] In CDCl3 at 25 °C, δTMS = 0.00 ppm. [b] Measured on a ca. 45:55 mixture of 20a and 20b; solvent [D6]DMSO.
an axial P=O bond had deshielded 31P resonances. This is
in agreement with the results for similarly 2-substituted re-
gioisomeric 1,3,2-benzoxazaphosphinines.[10] In this work
the configurational assignments were not based on the 31P
chemical shifts (see the Supporting Information for δP val-
ues).
Figure 1. Diastereomers a and b of the trans-fused 1-R-2-Z-perhy-
dro-3,1,2-benzoxazaphosphinine 2-oxides [R = H, Me, CH2Ph; Z
= Ph, OPh, N(CH2CH2Cl)2]. Only chair–chair conformers are
shown.
Assignment of the P Configuration; trans-Fused Compounds
The diastereomers a and b of the P-phenyl-substituted
compounds 8, 10 and 12 were easily identified: the 4-Hax
protons are deshielded (owing to the syn-diaxial P=O bond)
In the P-phenoxy-substituted compounds 14, 16 and 18,
with respect to the 4-Heq protons in epimers a, whereas the neither NOEs nor deshielding effects on 4-H could be used
opposite is true in the corresponding epimers b (Table 1, to differentiate reliably between the P epimers. Instead, the
Figure 1). The coupling constant 1JP,i-C is also larger for the configuration at the phosphorus atom was observed to have
3
2-Pheq (a) than for the 2-Phax (b) epimers (Table 2).[14] In profound effects on the J4-H,P values. Configuration a was
3
the P-bis(2-chloroethyl)amino-substituted compound 20b, a assigned to the diastereomers that had J4-Heq,P values in
cis relationship was established between 8a-H and the P the range of 14–17 Hz, and b to those for which this coup-
substituent on the basis of the observed NOE in the 2-chlo- ling was approximately 24 Hz (Table 3). These differences
roethyl protons upon irradiation of 8a-H. No such NOE between the diastereomeric pairs originate from the con-
was observed in epimer 20a.
siderable differences in the preferred conformations of the
Eur. J. Org. Chem. 2005, 1189–1200
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1191

H. Kivelä, Z. Zalán, P. Tähtinen, R. Sillanpää, F. Fülöp, K. Pihlaja
Table 2. Selected JP,C coupling constants [Hz] of the trans-fused compounds.[a]
FULL PAPER
Carbon
8a
8b
10b
12a
12b
14a
14b
16a
16b
18a
18b
20a[b]
20b[b]
C4a
C8
2-i-C6H5
4.1
9.6
180.6
6.8
9.6
165.9
6.1
5.3
162.9
7.3
4.1
182.0
5.9
5.5
164.5
11.0
6.8
7.3
5.5
11.5
7.3
11.5
2.7
7.3
4.1
9.2
9.2
12.4
1.4
6.9
3.6
8.2
8.2
3.7
11.9
–
7.7
9.7
–
[a] In CDCl3 at 25 °C. [b] Measured on a ca. 45:55 mixture of 20a and 20b; solvent [D6]DMSO.
Table 3. Selected JH,H and JH,P coupling constants [Hz] of the trans-fused compounds.[a]
Coup-
ling
8a
8b
10b
12a
12b
14a
14b
16a
16b
18a
18b
20a[b]
20b[b]
4ax,4a
4ax,P
4eq,4a
4eq,P
4a,8a
8a,P
11.5
2.7
4.1
20.6
10.3
2.0
11.4
2.2
4.2
21.7
9.9
n.d.
11.4
2.4
4.1
21.6
9.9
n.d.
11.4
2.7
3.5
19.7
10.7
3.1
11.2
3.1
4.3
21.1
10.1
2.2
11.6
6.3
4.4
14.5
10.6
4.1
11.6
1.2
4.2
23.7
10.1
1.2
11.7
3.9
3.8
16.9
10.6
6.3
11.6
1.7
4.2
24.3
9.9
n.d.
11.7
4.3
3.7
15.8
10.8
7.3
11.6
1.7
4.2
24.4
9.9
n.d.
11.2
1.8
4.0
22.6
9.9
3.4
11.5
3.6
4.4
19.2
9.1
4.7
[a] In CDCl3 at 25 °C. [b] Measured on a ca. 45:55 mixture of 20a and 20b; solvent [D6]DMSO.
heteroring, as discussed below. The correct assignment of epimeric pair: the epimer in which its 2-Z substituent
the relative P configuration was confirmed in the case of adopts the position preferred by the N-in conformation will
16a by X-ray crystallography (Figure 2).
have a larger x(N-in) value. Since Ph and N(CH2CH2Cl)2
prefer an equatorial position and OPh prefers an axial
position,[15] it can be concluded that the “N-in-favoring”
epimer is a when Z = OPh (13, 15 and 17) and b when Z =
Ph and N(CH2CH2Cl)2 (9, 11 and 19). For 7, only one epi-
mer was obtained, which was assigned as 7a on the basis of
the NOE observed between the aromatic ortho protons and
8a-H since this enhancement is possible only in the N-in
conformation of epimer a (Figure 3). For compounds 7a,
9a, 11a, 11b and 19b, which are predominantly single in-
vertomers, P=O-induced deshielding of one of the 4-H
atoms is observed whenever P=O is axial (Table 4) thereby
confirming the above diastereomeric assignments.
Figure 2. X-ray crystal structure of 16a.
Assignment of the P Configuration; cis-Fused Compounds
In the cis-fused compounds, the possibility of ring inver-
sion can lead to equilibria between the N-in and N-out
forms (Figure 3). The equilibrium populations of these
forms can be reliably deduced for each compound (vide in-
fra, Table 7). The P configuration can then be inferred by
Figure 3. N-in/N-out ring inversion of epimers a and b of the cis-
fused 1-R-2-Z-perhydro-3,1,2-benzoxazaphosphinine 2-oxides [R =
H, Me, CH2Ph; Z = Ph, OPh, N(CH2CH2Cl)2]. Geminal protons
are labeled with subscripts x and y, Hx atoms being axial in the N-
in invertomer. Upon ring inversion, axial protons become equato-
comparison of the N-in mole fractions x(N-in) within each rial and vice versa. Only chair–chair conformers are shown.
1192 © 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.eurjoc.org
Eur. J. Org. Chem. 2005, 1189–1200

Saturated 3,1,2-Benzoxazaphosphinine 2-Oxides
FULL PAPER
1
Table 4. Selected H NMR chemical shifts [ppm] of the cis-fused compounds.[a]
Proton[b]
7a
9a
9b
11a
11b
13a
13b
15a
15b
17a
17b
19a
19b
4-Hx
4-Hy
4a-H
8a-H
4.06
4.15
1.55
3.54
4.04
4.61
2.42
3.21
4.49
4.12
2.22
3.43
4.06
4.83
2.52
3.01
4.30
4.34
2.54
3.07
4.49
4.14
1.60
3.77
4.22
4.56
2.34
3.44
4.37
4.12
1.91
3.42
4.09
4.59
2.54
3.08
4.31
4.42
2.44
3.28
4.10
4.66
2.38
2.98
4.16
4.34
1.92
3.48
4.53
3.98
1.62
3.87
[a] In CDCl3 at 25 °C, δTMS = 0.00 ppm. [b] For the meaning of subscripts x and y, see Figure 3.
Conformations of the trans-Fused Compounds
five of the last-mentioned compounds, the calculated JH,P
values for the optimized chair–chair conformations agree
remarkably well with the observed ones. The view that no
noteworthy equilibrium exists between the different hetero-
ring conformers of compounds 14b, 16b and 18b is sup-
ported by the nondependence of the above-mentioned
coupling constants on the solvent or temperature (see the
Supporting Information for more details).
Variable-temperature NMR measurements were carried
out on epimers 16a and 16b in CD2Cl2, the temperature
gradually being lowered to –90 °C, but no signal splitting
1
was observed in the H or 31P NMR spectra. This is to be
expected from our earlier studies on closely related struc-
tures,[10a] for which distinct heteroring conformers were ob-
served only below 160 K in freon solutions. Such measure-
ments, however, are plagued by experimental difficulties,
such as reduced solubility at low temperatures, vanishing
population of the minor conformer(s), line-broadening due
to temperature instabilities and increased solvent viscosity,
as well as the general difficulties involved in cooling the
sample down to such low temperatures with our experimen-
tal set-up. In this work, therefore, this route was not pur-
sued.
The P-Ph- and P-N(CH2CH2Cl)2-substituted com-
pounds 8, 10, 12 and 20 exist almost exclusively in the
chair–chair conformation, regardless of the P configura-
tion. This conformation is also predominant in epimers b
of the P-OPh-substituted compounds (14b, 16b and 18b)
since the OPh group can then occupy its preferred axial
position (Figure 1). The chair conformation of the hetero-
For the epimers 14a, 16a and 18a, the heteroring cannot
adopt a chair conformation and the OPh group an axial
position simultaneously. The anomeric effect of 2-OR in cy-
clophosphamide-type heterocycles is often strong enough to
drive the heteroring into a skew conformation if ring inver-
sion is restricted.[9] That is the case here. While the coupling
constants J4-Hax,4a-H, J4-Heq,4a-H and J4a-H,8a-H remain ne-
arly the same upon changing from P epimer b to P epimer
a, the JH,P values change significantly (Table 3). From the
3
3
large values of J4-Hax,4a-H and J4a-H,8a-H and the relatively
3
3
small values of J4-Hax,P and JC8,P, it is evident that any
nonchair conformation populated to an appreciable extent
by these 2-OPh-substituted epimers a should lie between or
close to the B2,4a and 1S4a conformations in the pseudorota-
tional cycle depicted in Figure 4. In these conformations,
the phenoxy group can occupy the preferred pseudoaxial
position. The DFT-calculated geometries and energies sup-
port these conclusions; low-energy structures are predicted
for compounds 14a and 16a, as presented in Table 5 and
Figure 5. The calculations indicate that the skew conforma-
tions are relatively flexible as the conformational search for
compounds 14a and 16a resulted in several local minimum-
energy nonchair conformations within about 3.5 kJ/mol
(∆G°) of each other and distributed along the above-speci-
fied sector of the pseudorotational circuit. The DFT-opti-
mized chair–chair conformation was in both cases less
stable than the most favorable chair–skew conformation by
more than 5.8 kJ/mol.
ring of these compounds is characterized by large values of
3
3J4-Heq,P
,
3J4-Hax,4a-H and J4a-H,8a-H (19–25, 11–12 and
3
9–11 Hz, respectively), and small values of J4-Hax,P and
3J8a-H,P (1–3.5 Hz) (Table 3), as expected from the Karplus-
like dependence on the torsion angles. The values for 20b
are actually on the border of or slightly beyond these limits,
suggesting that small amounts of this compound may adopt
a nonchair conformation. The DFT-calculated energies
(Table 5) are in good agreement with the observations for
compounds in which the 2-Z substituent favors the chair
conformation (8a, 14b and 16b), the lowest-energy chair–
chair conformations being more than 13 kJ/mol (∆G°) more
stable than the most favorable nonchair conformations of
The usual approach to estimating chair/skew ratios in
the heteroring. On the other hand, the relative energies cal- analogous structures from fast-exchange NMR data is to
3
culated for different conformations of 8b and 10b predict assume values for the relevant JP,H coupling constants for
that nonchair conformers make significant contributions to the limiting conformers and to assign a population to each
the equilibrium, which is not supported by the observed (usually two) conformer in such a way that the population-
values of the coupling constants. Thus, the calculations weighted average of these couplings best agrees with the
seem to overestimate the destabilizing effect of an axial P- observed one. Naturally, the problem is that of how to ob-
Ph group in the chair conformation. As observed recently tain these limiting values. Here, the flexibility of the non-
for similar 1,3,2-benzoxazaphosphinines, the DFT calcula- chair conformation of the heteroring and the uncertainties
tions fail to predict the relative energies of the conforma- in the calculated coupling constants for these conforma-
tions with good accuracy at the chosen level of theory.[16] tions necessarily leave rather large margins of error in the
However, coupling constants are calculated with satisfac- calculated conformer populations. On the basis of the quali-
tory accuracy at the selected level of theory. Indeed, in all tative dependence of the calculated coupling constant
Eur. J. Org. Chem. 2005, 1189–1200
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1193

H. Kivelä, Z. Zalán, P. Tähtinen, R. Sillanpää, F. Fülöp, K. Pihlaja
FULL PAPER
Table 5. ∆G° [kJ/mol] values, selected torsion angles ω [°] and 3JH,P coupling constants [Hz] for representative DFT-optimized conforma-
tions of compounds 8a, 8b, 10b, 14a, 14b, 16a and 16b.
P,4-Hax
P,8a-H
Conformation[a] ∆G°
P,4-Heq
4-Hax,4a-H 4-Heq,4a-H
P,C8
ω
ω
3J
ω
3J
ω
3J
ω
ω
8a
chair (1)
chair (2)
3,8aB
0.0
8.7
15.3
37.1
1.2
0.0
0.0
1.4
1.6
7.7
5.8
1.2
1.5
1.7
2.2
0.0
3.4
0.0
2.1
16.4
13.5
9.0
21.8
0.0
2.1
0.4
1.1
0.0
8.6
25.8
24.8
–61.7
–64.6
–167.6
–177.7
–69.7
–154.5
–64.3
–72.8
174.2
–63.1
–64.2
–101.8
–102.4
–103.5
–105.9
–109.6
–112.1
–67.9
–70.4
–150.2
–165.3
–60.7
–65.6
–93.8
–93.9
–94.4
–95.5
–66.8
–73.0
–87.7
178.1
4.2
–178.0
179.4
77.6
67.4
174.3
90.1
179.5
171.5
59.0
–179.4
179.6
142.6
141.9
140.8
138.5
134.9
132.4
176.0
173.7
94.5
22.0
19.8
2.0
5.1
21.5
0.9
22.6
21.2
9.0
24.2
23.1
14.5
14.0
13.4
12.6
10.3
9.5
62.1
62.1
55.2
47.6
71.5
85.3
83.9
61.4
50.9
3.7
–179.9
–59.0
–56.5
–12.4
8.4
–59.4
–34.2
–63.3
–55.4
10.0
–59.3
–56.8
–61.5
–61.7
–61.2
–59.6
–57.7
–56.6
–58.6
–58.9
–34.8
–19.4
–60.1
–51.5
–65.4
–64.8
–65.1
–64.7
–58.6
–55.7
–68.5
2.1
–178.7
179.3
175.0
3.4
25.3
27.6
2.6
22.0
3.1
1.7
24.5
2.8
2.8
3.1
2.8
3.1
2.8
6.2
6.4
0.9
0.1
–0.2
2.6
3.4
2.0
2.0
4.8
5.9
5.6
3.9
3.4
2.6
2.6
0.3
–0.2
0.9
1.2
2.8
10.3
8.7
10.1
9.5
2.3
1.4
17.8
4.2
–176.8
–132.3
–111.2
179.2
–154.6
176.0
–174.6
–110.6
179.4
–177.5
178.1
3S1
165.0
8b
chair (2)
–171.3
–157.1
–157.9
178.0
8aS2
10b
chair (1)
chair (2)
3S1
167.8
14a
chair (1)
chair (2)
B2,4a (1)
B2,4a (2)
B2,4a (3)
B2,4a (4)
B2,4a (5)
B2,4a (6)
chair (1)
chair (2)
8aS2
65.3
63.6
–175.7
–179.3
–134.3
–132.7
–133.3
–136.7
–139.1
–140.2
–174.7
–171.1
–164.7
–172.9
–171.9
173.1
–120.5
–122.2
–121.0
–121.5
–171.6
178.7
106.5
107.9
107.4
104.1
101.7
100.6
65.9
71.7
77.7
69.4
69.5
177.9
178.5
4.2
5.3
6.7
2.8
–179.9
–177.7
–176.7
–179.6
–179.3
–155.2
–140.0
178.4
–171.9
173.6
174.1
173.9
174.5
14b
16a
24.8
23.1
0.1
2.3
23.7
28.4
3.4
2.8
1.2
1.2
1.2
1.4
3.1
3,8aB
79.5
1.7
chair (1)
chair (2)
1S4a (1)
1S4a (2)
1S4a (3)
1S4a (4)
chair (1)
chair (2)
1S4a
–177.4
178.2
150.2
150.2
149.6
148.5
176.9
171.3
155.7
63.0
24.2
22.3
18.1
18.7
17.8
17.3
24.2
23.4
20.9
6.7
56.7
121.0
119.6
120.6
120.1
69.5
62.1
132.7
55.3
16b
–179.7
–175.5
169.3
2.0
–0.2
29.8
–108.7
172.6
3S1
–118.9
[a] See Figure 4 for crude pictorial representations of nonchair conformations: chair (1) has an equatorial N substituent, chair (2) an
axial N substituent.
Figure 5. The nonchair conformers found for compounds 14a and
16a by a computational conformational search. The numbering of
these conformers refers to that in Table 5. These closely related
conformers arise mainly from the various rotameric states of the
OPh substituent. The magnitude of the calculated torsion angle
P2–O3–C4–Hax was used as the pseudorotational coordinate.
3J4-Hax,P on the pseudorotational coordinate (Table 5, Fig-
ure 5), it is expected that the predominant nonchair confor-
mations of 14a are closer to B2,4a than to 1S4a [3J4-Hax,P(obs)
= 6.3 Hz]. By arbitrarily choosing conformation B2,4a(6) as
the model nonchair conformation for 14a, and with the use
Figure 4. The idealized pseudorotational cycle for the epimers a of
of Equation (1), the mole fraction of the nonchair compo-
the trans-fused 1-R-2-OPh-perhydro-3,1,2-benzoxazaphosphinine
nent is estimated to lie between 0.6 and 0.9. This depends
2-oxides. For clarity, the fused cyclohexane ring, N substituent and
on which of the two optimized chair–chair conformations
is chosen as the model chair structure and which coupling
2-oxo group have been omitted. The shaded area is inaccessible
owing to conformational locking by the trans-fused carbocycle. The
cycle is sketched for the (2R,4aS,8aS) enantiomers.
3
constant is used (3J4-Heq,P or J4-Hax,P). A more precise de-
1194
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1189–1200

Saturated 3,1,2-Benzoxazaphosphinine 2-Oxides
FULL PAPER
termination of this mole fraction from the time-averaged
Although the interconversion between the N-in and N-
NMR data does not seem plausible owing to the above- out conformers is fast on the NMR timescale, resulting in
mentioned uncertainties.
time-averaged spectra at 25 °C, it is possible to estimate
their equilibrium mole fractions quite accurately for the fol-
lowing reasons. (1) In the ring reversal process large diaxial
vicinal JH,H couplings are exchanged for small diequatorial
ones. (2) The values of these coupling constants for protons
in the cyclohexane ring are relatively independent of the
substitution on the heteroring. (3) Compounds 13a and 17b
populate the N-in and N-out conformations, respectively, al-
most completely and hence provide reference axial–axial
and equatorial–equatorial coupling constants for these con-
formers (Table 6). Accordingly, Equation (1), modified for
the N-in/N-out equilibrium, may be utilized to calculate the
equilibrium mole fractions of the ring invertomers for all of
the cis-fused compounds 7–19a,b (Table 7). In the follow-
ing, geminal protons will be denoted as Hx or Hy according
to the convention adopted in Figure 3. The N-in mole frac-
tions presented in Table 7 are averages of eight calculations
x(nonchair) = [J(obs) – J(chair)]/[J(nonchair) – J(chair)]
(1)
For 16a, all the nonchair DFT-calculated structures
found were similar to that of S4a, and as such seemed to
exhibit slightly too-large J4-Heq,P values (17.3–18.7 Hz vs.
16.9 Hz observed). It is therefore possible that in CDCl3 the
predominant nonchair conformations of 16a are somewhat
more biased towards the B2,4a structures than is predicted
by the DFT calculations. The torsion angles measured from
the X-ray crystal structure of 16a are also consistent with
this view [ω(P,4-Hax) = 102.5°, ω(P,4-Heq) = –141.3° (cf.
Table 5)]. In any case, the 3J4-Heq,P(obs) value together with
1
3
3
the small value observed for JC8,P (2.7 Hz) and the inter-
3
mediate J8a-H,P (6.3 Hz) suggest that the heteroring of 16a
1
as there are eight suitable H,1H coupling constants in the
does not populate the chair conformation in an appreciable
3
carbocycle (3J4a-H,5-Hx, ... , J8-Hy,8a-H in Table 6). The mole
3
3
amount (the values of JC8,P and J8a-H,P for 16b are 9.2
and Ͻ1 Hz, respectively). This holds even more so for 18a
(1.4 and 7.3 Hz, as opposed to 8.2 and Ͻ1 Hz for 18b).
fractions calculated from the different coupling constants
are in good agreement (standard deviations Ͻ0.04), which
supports the assumption that the cyclohexane moiety is al-
ways in the chair conformation. It can be seen that the pref-
erence for the N-out form (the P substituents and their con-
figurations being kept fixed) increases with increasing steric
size of the N substituent in the sequence H Ͻ Me Ͻ
CH2Ph. This is to be expected since in the N-in form the
nitrogen atom is in a sterically hindered axial position with
respect to the cyclohexane ring. From the x(N-in) values it
is also evident that OPh is strongly axial-seeking, whereas
the other two substituents prefer an equatorial position in
the sequence Ph Ͻ N(CH2CH2Cl)2. It can now be seen why
13a and 17b should stay in their sole dominant conforma-
tions assumed under point (3) above: in 13a, both the rela-
tive configuration of the P substituent (OPh) and the steric
size of the N substituent (H) favor the N-in conformation,
whereas in 17b these factors favor the N-out form (Z = OPh,
R = CH2Ph). The exclusivity of the N-out conformation for
17b is supported by the observations made at low tempera-
3
Thus, from the systematic decrease in JC8,P and the in-
crease in 3JC4a,P in the sequence 14a Ǟ 16a Ǟ 18a, it would
be tempting to suggest that the nonchair heteroring popula-
tion increases with the increasing steric size of the N sub-
stituent (H Ͻ Me ϽCH2Ph) from perhaps 80% in 14a to
close to 100% in 18a. Consistently with this, the DFT cal-
culations predict an increased preference for nonchair con-
formations of the heteroring in 16a as compared with 14a.
The relevant coupling constants of 14a, 16a and 18a display
more pronounced solvent and temperature dependencies
than those of the corresponding epimers b (see the Support-
ing Information for data), which confirms that the epimers
a exist in a conformational equilibrium whereas the epimers
b remain exclusively in the chair–chair conformation. By
assuming that the equilibrium in epimers a exists between
the skew and chair conformers, the data show that increas-
ing solvent polarity and temperature favor the skew heteror-
ing over the chair.
1
tures in CD2Cl2. In the H NMR spectrum of 17a, distinct
signals for both conformers become visible between approx-
imately –50 and –70 °C, with an approximate N-in/N-out
ratio of 45:55 at –90 °C, whereas only a single set of signals
Conformations of the cis-Fused Compounds
In the cis-fused compounds, it is expected that the fused was observed for 17b even at –90 °C.
cyclohexane moiety adopts the chair conformation. Owing
In order to evaluate the conformation of the heteroring
to ring inversion, the heteroring’s nitrogen atom can there- in the N-in and N-out forms it is assumed that the hetero-
fore be either axial (N-in) or equatorial (N-out) with respect ring adopts the chair conformation in the invertomer in
to the cyclohexane ring (Figure 3). The equilibrium between which the P substituent can thus occupy its preferred posi-
these conformers is strongly dependent on the nature and tion [axial for OPh; equatorial for Ph and N(CH2CH2Cl)2].
configuration of the P substituents since the ring reversal The other ring invertomer is then expected to populate
permits their preferred positions to be adopted [axial for nonchair conformations. Thus, it is assumed that the
OPh, equatorial for Ph and N(CH2CH2Cl)2] without the heteroring adopts a chair conformation in the N-in forms
need for the heteroring to adopt the nonchair conforma- of 9b, 11b, 13a, 15a, 17a and 19b, and in the N-out forms
tions. However, the steric size of the N substituent also has of 7a, 9a, 11a, 13b, 15b, 17b and 19a. This assumption is
a marked effect on the equilibrium, as will be discussed be- supported by the observation that for the compounds that
low.
almost exclusively favor a single invertomer (11a, 13a, 15b,
Eur. J. Org. Chem. 2005, 1189–1200
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1195

H. Kivelä, Z. Zalán, P. Tähtinen, R. Sillanpää, F. Fülöp, K. Pihlaja
Table 6. Selected JH,H and JH,P coupling constants [Hz] of the cis-fused compounds.[a]
FULL PAPER
Coupling[b]
7a
9a
4.0
16.9
10.4
6.7
3.7
4.8
12.1
4.5
11.4
5.1
9b
3.3
11a
4.5
20.4
12.2
3.2
2.4
2.4
14.0
2.9
13.3
3.3
11b
5.0
13.1
11.1
9.0
3.4
4.2
12.7
3.5
12.1
3.9
13a
2.5
1.8
1.7
23.7
12.6
13.2
3.2
13.0
3.0
13.6
3.3
2.8
7.8
1.0
13b
4.0
18.9
10.4
5.8
4.3
4.7
12.0
4.9
11.2
5.3
15a
4.0
2.8
3.7
21.2
10.8
11.3
5.1
11.0
4.8
11.7
5.2
5.0
3.7
3.1
15b
4.2
22.9
12.0
2.8
2.5
3.0
13.7
3.5
12.8
3.6
17a
4.8
7.1
9.1
15.5
5.2
5.9
10.4
6.0
10.2
6.6
10.2
9.3
1.7
17b
4.4
23.5
12.4
2.0
2.1
2.6
14.0
3.1
13.3
3.5
19a
3.6
19b
2.8
2.5
1.8
23.0
12.5
13.2
3.2
13.0
3.0
13.3
4.0
2.6
7.1
1.9
4x,4a
4x,P
4y,4a
4y,P
4a,5x
5x,6x
5y,6y
6x,7x
6y,7y
7x,8x
7y,8y
8y,8a
8x,P
2.7
4.2
2.7
20.8
11.7
12.4
4.0
12.1
3.9
12.7
3.9
3.8
5.9
2.3
7.6
5.9
15.7
8.6
9.2
7.2
9.3
7.2
9.7
7.2
6.8
2.1
3.8
12.2
6.6
11.9
7.8
8.5
8.1
8.4
7.9
9.0
7.6
7.1
3.7
11.2
11.7
10.3
n.d.
14.5
13.2
12.1
n.d.
20.0
12.2
11.3
n.d.
12.0
11.4
10.6
n.d.
20.9
12.9
12.0
n.d.
22.4
13.1
12.4
n.d.
22.9
8a,P
8.4
[a] In CDCl3 at 25 °C. [b] For the meaning of subscripts x and y, see Figure 3.
3
3
Table 7. N-in mole fractions and approximate values of the coupling constants J4-Hx,P and J4-Hy,P [Hz] in the limiting invertomers (N-
in and N-out) of the cis-fused compounds. The heteroring of the invertomer A[b] is assumed to adopt the chair conformation; the coupling
constants for the invertomer B[b] are back-calculated from Equation (2).
x(N-in)
∆x[a]
Invertomer A[b]
3J(obs)
3J(N-in)
3J(N-out)
4-Hx,P
4-Hx,P
4-Hy,P
4-Hx,P
4-Hy,P
4-Hy,P
7a
0.92
0.17
0.61
0.00
0.10
1.00
0.19
0.81
0.03
0.31
0.00
0.55
0.99
0.01
0.03
0.02
0.02
0.04
–
0.01
0.02
0.01
0.01
–
N-out
N-out
N-in
N-out
N-in
N-in
N-out
N-in
N-out
N-in
N-out
N-out
N-in
4.2
16.9
7.6
20.4
13.1
1.8
18.9
2.8
22.9
7.1
20.8
6.7
15.7
3.2
9.0
23.7
5.8
21.2
2.8
15.5
2.0
11.9
23.0
2.7
0.0
22.3
23.1
22.3[d]
–
20.4[c]
20.4[c]
15.3
20.4
14.2
–
3.2[c]
3.2[c]
5.5
3.2
7.5
9a
9b
2.7[d]
–
11a
11b
13a
13b
15a
15b
17a
17b
19a
19b
2.7[d]
1.8
22.3[d]
23.7
22.1
23.7[f]
26.0
23.7[f]
–
–
–1.0
1.8[f]
4.6
23.5[e]
7.1
2.0[e]
10.6
2.0[e]
11.8
2.0
23.5[e]
9.4
23.5
23.0[g]
–
1.8[f]
–
23.5
12.2
2.5
0.01
0.03
3.4
2.5
19.6
23.0
2.5[g]
–
[a] Standard deviation of eight calculations of x(N-in), each using a different coupling constant. [b] In invertomer A the P substituent is
in a preferred position if a chair–chair conformation is adopted; B is the other invertomer. [c] From 11a. [d] From 7a. [e] From 17b. [f]
From 13a. [g] From 19b, assuming reversal of coupling constants upon ring inversion.
17b and 19b) the heteroring is in the chair conformation, as uration. The results of the calculations are presented in
3
deduced from the J4-Hx,4a-H
,
3J4-Hx,P
,
3J4-Hy,4a-H
,
3J4-Hy,P Table 7. Although this approach is necessarily prone to
3
and J8a-H,P values. In these compounds, the P substituent large errors, some conclusions can be drawn. (1) In the N-
is in its preferred position. By choosing suitable coupling in conformers, the heteroring always seems to populate
constant model values for the N-in and N-out invertomers chair-like conformations even when the P configuration is
in which the heteroring adopts the chair conformation, it “unfavorable” (e.g. 7a, 9a, 13b and 19a). (2) In the N-out
is possible to back-calculate the coupling constants for the conformers, the heteroring noticeably populates nonchair
invertomers with the “nonpreferred” P configuration by conformation(s) if the P configuration is unfavorable (9b,
using Equation (2)
11b, 15a and 17a). These results can be rationalized by ste-
ric factors: in the chair–chair conformation the N-out in-
vertomer has its axially orientated hydrogen atoms at C4
and C6, as well as the axial P substituent and 8-H, facing
each other, thus making the nonchair conformers more
likely than in the case of the N-in form in which these repul-
sive interactions are missing. Unfortunately, the nature of
these heteroring conformations and their populations can-
J(B) = [J(obs) – x(A)×J(A)]/x(B),
(2)
where A is the ring invertomer with the preferred P configu-
ration (assuming a chair conformation for the heteroring)
and B the other invertomer with the nonpreferred P config-
1196
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1189–1200

Saturated 3,1,2-Benzoxazaphosphinine 2-Oxides
FULL PAPER
to occupy a sufficiently large region of space. For the remaining
atoms, the 6-31G(d,p) basis set was applied, except for the four
carbon and hydrogen atoms bound to them in the outmost cyclo-
hexane ring, which were treated with the 3-21G basis set. Vi-
brational analysis (1 bar, 298.15 K, scaling factor 0.9804)[19] was
performed on the optimized structures to ensure that they corre-
sponded to true energy minima and to yield the ∆G° values. Se-
lected coupling constants JH,P (Fermi contact contribution only)
were calculated by the finite perturbation theory (FPT) method at
the spin-unrestricted UB3LYP/cc-pVTZ level by placing the pertur-
bation on the phosphorus atom. Subsequently, the calculated coup-
lings were scaled by using the calibration given by Equation (3).
not be reliably assigned owing to the uncertainties in the
data.
Conclusions
We have synthesized a series of cis- and trans-fused per-
hydro-3,1,2-benzoxazaphosphinine 2-oxides and analyzed
their P configurations and solution-state conformations,
mainly by means of NMR spectroscopy. As expected, the
2-OPh substituent strongly favors an axial position,
whereas the 2-Ph and 2-N(CH2CH2Cl)2 substituents prefer
an equatorial position, the latter more so than the former.
In the trans-fused compounds, the nonpreferred position of
the 2-OPh substituent in the chair–chair conformation can
cause the heteroring to predominantly populate nonchair
conformations around the B2,4a ... 1S4a sector in the pseudo-
rotational cycle. On the other hand, in the 2-Ph- and 2-
N(CH2CH2Cl)2-substituted trans-fused compounds, the
heteroring predominantly adopts the chair conformation,
regardless of the P configuration. In the cis-fused com-
pounds, the axial or equatorial preference of the P substitu-
ents shifts the N-in/N-out equilibrium in a direction such
that the substituents can occupy their preferred positions
in the chair conformation of the heteroring. This tendency
increases in the sequence Ph Ͻ N(CH2CH2Cl)2 Ͻ OPh. The
preference for the N-out form increases with increasing ste-
ric size of the N substituent (H Ͻ Me Ͻ CH2Ph), which can
result in either cooperation or competition with the above-
mentioned effects of the P substituents. Thus, compounds
13a and 17b, for instance, populate almost exclusively a sin-
gle invertomer (N-in and N-out, respectively), whereas in 9b
and 19a both invertomers are noticeably populated.
More details about the applied methods can be found in ref.[16]
.
JH,P = 0.9057 × JH,P(FC) + 0.3349 Hz
(3)
Cartesian coordinates and energies of the DFT-optimized geome-
tries are provided as Supporting Information.
X-ray Crystallography: Suitable crystals of 16a for X-ray measure-
ments were grown from n-hexane solution. Crystal data for com-
pound 16a along with other experimental details are summarized
in the Supporting Information. The crystallographic data were col-
lected at 173 K on
a Nonius-KappaCCD area-detector dif-
fractometer using graphite-monochromatized Mo-Kα radiation (λ
= 0.71073 Å). The data were collected by using φ and ω scans and
processed by using DENZO-SMN v0.93.0.[20] The structures were
solved by direct methods by using the SHELXS-97 program[21] and
full-matrix least-squares refinements on F2 were performed by
using the SHELXL-97 program.[21] For all compounds the heavy
atoms were refined anisotropically. The aromatic CH hydrogen
atoms were included at fixed distances with fixed displacement
parameters from their host atoms and the rest of the hydrogen
atoms were refined isotropically. Figures were drawn with Ortep-3
for Windows.[22]
CCDC-247963 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
Experimental Section
NMR Measurements: The 1H, 1H{31P}, 31P{1H}, 13C{1H}, NOE
difference, dqf-COSY, CH shift correlation and COLOC NMR
spectra were recorded with a JEOL JNM-LA400 (1H: 399.78 MHz,
31P: 161.84 MHz and 13C: 100.54 MHz) or JEOL JNM-A500 (1H:
500.16 MHz, 31P: 202.47 MHz and 13C: 125.78 MHz) instrument.
Samples were typically dissolved in CDCl3 and the NMR spectra
were measured at 25 °C in 5-mm diameter NMR tubes. Owing to
the poor solubility of compound 20 in CDCl3, its NMR spectra
were recorded in [D6]DMSO at 30 °C. A few 1H NMR experiments
were also performed with [D6]acetone as solvent (at 25 °C) in order
to check solvent effects, and for low-temperature measurements
General Procedures: Melting points were recorded with a Kofler
hot plate microscope apparatus and are uncorrected. Elemental
analyses were performed with a Perkin–Elmer 2400 CHNS elemen-
tal analyzer. Chemicals were generally of the highest purity. For
column chromatography, silica gel 60 (0.040–0.063 mm) was used.
Merck Kieselgel 60F254 plates were used for TLC. The starting
amino alcohols (1–6) were prepared according to the literature pro-
cedures.[12] Amino alcohols 1 and 2 were used for the ring-closure
reactions as their hydrochlorides.
(+20 to –90 °C) CD2Cl2 was used. H and 13C spectra were refer-
1
General Method for Ring-Closure Reactions: A solution of the ap-
propriate P-containing reagent [phenylphosphonic dichloride,
phenyl dichlorophosphate or bis(2-chloroethyl)phosphoramidic di-
chloride, 1 equiv.] in dry THF (50 mL) was added dropwise to a
stirred solution of the appropriate amino alcohol (1–6, 6 mmol)
and triethylamine (2 equiv.; 3 equiv. in the case of amino alcohol
hydrochlorides 1 and 2) in dry THF (100 mL) at room temperature
. The reaction mixture was stirred for 48 h and then filtered to
remove triethylamine hydrochloride. The filtrate was evaporated to
dryness.
enced to internal TMS (δ = 0.00 ppm), whereas in 31P NMR experi-
ments, external 90% H3PO4 in D2O was used as a reference (δ =
0.00 ppm). 1H NMR spectra were analyzed by using PERCH
NMR software.[13]
Computations: The initial conformational search for compounds 8,
10, 14 and 16 was made at the molecular mechanics (MM+) level
by using a utility implemented in the HyperChem. 7.0 molecular
modeling software package. The geometries of the starting struc-
tures were optimized with Gaussian 98[17] utilizing the B3LYP DFT
method[18] and a locally dense basis set, as follows. For the phos-
phorus, nitrogen and oxygen atoms, which have lone-pair electrons,
the 6-31+G(d,p) basis set was applied in order to allow the orbitals
(2S*,4aS*,8aR*)-2-Phenyl-1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-
benzoxazaphosphinine 2-Oxide (7a): The crude product (in which
1
only the isomer 7a was detected by H NMR) was purified by col-
Eur. J. Org. Chem. 2005, 1189–1200
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1197

H. Kivelä, Z. Zalán, P. Tähtinen, R. Sillanpää, F. Fülöp, K. Pihlaja
FULL PAPER
umn chromatography with ethyl acetate as eluent. Diastereomer 7a
(0.33 g, 22%) was crystallized by evaporation, filtered from n-hex-
ane and recrystallized from diisopropyl ether/ethyl acetate: m.p.
135–137 °C. 1H, 13C and 31P NMR spectroscopic data are given in
the Supporting Information. C13H18NO2P (251.26): calcd. C 62.14,
H 7.22, N 5.57; found: C 62.07, H 7.25, N 5.62.
m.p. 107–108 °C). 1H, 13C and 31P NMR spectroscopic data are
given in the Supporting Information. C20H24NO2P (341.38): calcd.
C 70.37, H 7.09, N 4.10; found: C 70.44, H 7.13, N 4.07.
(2S*,4aS*,8aS*)-
and
(2R*,4aS*,8aS*)-1-Benzyl-2-phenyl-
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-Ox-
ides (12a and 12b): The crude product (ratio of the isomers 55:45,
based on the 31P NMR spectrum) was purified by column
chromatography with ethyl acetate as eluent. The more mobile dia-
stereomer was crystallized by evaporation and was filtered from n-
hexane to yield isomer 12b (0.31 g, 15%), which was recrystallized
from diisopropyl ether/ethyl acetate: m.p. 145–150 °C (lit.[11] m.p.
148–149 °C). 1H, 13C and 31P NMR spectroscopic data are given
in the Supporting Information. C20H24NO2P (341.38): calcd. C
70.37, H 7.09, N 4.10; found: C 70.28, H 7.14, N 4.07.
(2S*,4aS*,8aS*)- and (2R*,4aS*,8aS*)-2-Phenyl-1,4,4a,5,6,7,8,8a-
octahydro-2H-3,1,2-benzoxazaphosphinine 2-Oxides (8a and 8b):
The crude product (in which only isomer 8b was detected by 1H
and 31P NMR spectroscopy, with the signals of 8a possibly hidden
beneath more intense bands) was purified by column chromatog-
raphy with ethyl acetate/methanol (15:1) as eluent. The more mo-
bile diastereomer was crystallized by evaporation and was filtered
from n-hexane to yield isomer 8b (0.56 g, 37%), which was recrys-
1
tallized from diisopropyl ether/ethyl acetate: m.p. 149–155 °C. H,
The less mobile diastereomer was identified as isomer 12a (0.37 g,
18%, colorless, transparent oil) with 12b as a minor impurity (ca.
13C and 31P NMR spectroscopic data are given in the Supporting
Information. C13H18NO2P (251.26): calcd. C 62.14, H 7.22, N 5.57;
found: C 62.20, H 7.17, N 5.54.
1
1
2:100 by H NMR). H, 13C and 31P NMR spectroscopic data are
given in the Supporting Information. C20H24NO2P (341.38): calcd.
C 70.37, H 7.09, N 4.10; found: C 70.41, H 7.06, N 4.04.
The less mobile diastereomer was crystallized by evaporation and
was filtered from n-hexane to yield isomer 8a (0.15 g, 10%, m.p.
1
(2R*,4aS*,8aR*)-
and
(2S*,4aS*,8aR*)-2-Phenoxy-
200–205 °C) with 8b as a minor impurity (ca. 3:100 by H NMR).
1H, 13C and 31P NMR spectroscopic data are given in the Support-
ing Information. C13H18NO2P (251.26): calcd. C 62.14, H 7.22, N
5.57; found: C 62.08, H 7.30, N 5.63.
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-Ox-
ides (13a and 13b): The crude product (ratio of the isomers 73:27,
based on the 1H NMR spectrum) was purified by column
chromatography with ethyl acetate as eluent. The more mobile dia-
stereomer was crystallized by evaporation and was filtered from
diethyl ether to yield isomer 13a (0.59 g, 37%), which was recrys-
(2S*,4aS*,8aR*)-
and
(2R*,4aS*,8aR*)-1-Methyl-2-phenyl-
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-Ox-
ides (9a and 9b): The crude product (ratio of the isomers 18:82,
based on the 31P NMR spectrum) was purified by column
chromatography with ethyl acetate/methanol (10:1) as eluent. The
more mobile diastereomer was crystallized by evaporation and was
filtered from n-hexane to yield isomer 9a (0.21 g, 13%, m.p. 104–
108 °C) in the pure form. 1H, 13C and 31P NMR spectroscopic data
are given in the Supporting Information. C14H20NO2P (265.29):
calcd. C 63.38, H 7.60, N 5.28; found: C 63.29, H 7.57, N 5.33.
1
tallized from diisopropyl ether/ethyl acetate: m.p. 149–152 °C. H,
13C and 31P NMR spectroscopic data are given in the Supporting
Information. C13H18NO3P (267.26): calcd. C 58.42, H 6.79, N 5.24;
found: C 58.50, H 6.73, N 5.27.
The less mobile diastereomer was crystallized by evaporation and
was filtered from diethyl ether to yield diastereomer 13b (0.18 g,
11%, m.p. 100–104 °C) with 13a as a minor impurity (ca. 2:100 by
1
1H NMR). H, 13C and 31P NMR spectroscopic data are given in
The less mobile diastereomer was crystallized by evaporation and
was filtered from n-hexane to yield isomer 9b (0.30 g, 19%), which
was recrystallized from diisopropyl ether/ethyl acetate: m.p. 88–
90 °C. 1H, 13C and 31P NMR spectroscopic data are given in the
Supporting Information. C14H20NO2P (265.29): calcd. C 63.38, H
7.60, N 5.28; found: C 63.41, H 7.54, N 5.30.
the Supporting Information. C13H18NO3P (267.26): calcd. C 58.42,
H 6.79, N 5.24; found: C 58.38, H 6.83, N 5.26.
(2R*,4aS*,8aS*)-
and
(2S*,4aS*,8aS*)-2-Phenoxy-
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-ox-
ides (14a and 14b): The crude product (ratio of the isomers 48:52,
based on the 31P NMR spectrum) was purified by column
chromatography with ethyl acetate as eluent. The more mobile dia-
stereomer 14b (0.59 g, 37%) was obtained as a crystalline com-
pound, which was filtered from n-hexane and recrystallized from
ethyl acetate: m.p. 160–162 °C. 1H, 13C and 31P NMR spectro-
scopic data are given in the Supporting Information. C13H18NO3P
(267.26): calcd. C 58.42, H 6.79, N 5.24; found: C 58.50, H 6.68,
N 5.28.
(2R*,4aS*,8aS*)-1-Methyl-2-phenyl-1,4,4a,5,6,7,8a-octahydro-2H-
3,1,2-benzoxazaphosphinine 2-Oxide (10b): The crude product (in
which only isomer 10b was observed by 1H NMR) was purified
by column chromatography with ethyl acetate/methanol (15:1) as
eluent. Diastereomer 10b (0.38 g, 24%, m.p. 58–60 °C) could be
obtained by column chromatography as a pure isomer. 1H, 13C and
31P NMR spectroscopic data are given in the Supporting Infor-
mation. C14H20NO2P (265.29): calcd. C 63.38, H 7.60, N 5.28;
found: C 63.32, H 7.63, N 5.26.
The less mobile diastereomer 14a (0.40 g, 25%) was crystallized by
evaporation and was filtered from n-hexane and recrystallized from
EtOAc: m.p. 103–104 °C. 1H, 13C and 31P NMR spectroscopic data
are given in the Supporting Information. C13H18NO3P (267.26):
calcd. C 58.42, H 6.79, N 5.24; found: C 58.35, H 6.84, N 5.21.
(2S*,4aS*,8aR*)-
and
(2R*,4aS*,8aR*)-1-Benzyl-2-phenyl-
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-Ox-
ides (11a and 11b): The crude product (ratio of the isomers 56:44,
based on the 1H NMR spectrum) was purified by column
chromatography with ethyl acetate as eluent. The more mobile dia-
stereomer was identified as pure isomer 11b (0.16 g, 8%, colorless,
transparent oil). 1H, 13C and 31P NMR spectroscopic data are given
in the Supporting Information. C20H24NO2P (341.38): calcd. C
70.37, H 7.09, N 4.10; found: C 70.47, H 7.04, N 4.14.
(2R*,4aS*,8aR*)-
and
(2S*,4aS*,8aR*)-1-Methyl-2-phenoxy-
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-Ox-
ides (15a and 15b): The crude product (ratio of the isomers 49:51,
based on the 1H NMR spectrum) was purified by column
chromatography with ethyl acetate as eluent. The more mobile dia-
stereomer was identified as diastereomerically pure 15a (0.35 g,
The less mobile diastereomer was crystallized by evaporation and
was filtered from n-hexane to yield isomer 11a (0.92 g, 45%), which
was recrystallized from diisopropyl ether: m.p. 99–101.5 °C (lit.[11]
1
21%, pale yellow, transparent oil). H, 13C and 31P NMR spectro-
scopic data are given in the Supporting Information. C14H20NO3P
1198
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1189–1200

Saturated 3,1,2-Benzoxazaphosphinine 2-Oxides
FULL PAPER
(281.29): calcd. C 59.78, H 7.17, N 4.98; found: C 59.66, H 7.20,
N 4.93.
more mobile diastereomer was crystallized by evaporation and was
filtered from n-hexane to yield isomer 19a (0.66 g, 35%), which
was recrystallized from diisopropyl ether/ethyl acetate: m.p. 125–
The less mobile diastereomer was identified as pure isomer 15b
1
127 °C. H, 13C and 31P NMR spectroscopic data are given in the
1
(0.39 g, 23%, pale yellow, transparent oil). H, 13C and 31P NMR
Supporting Information. C11H21Cl2N2O2P (315.18): calcd. C 41.92,
H 6.72, N 8.89; found: C 41.81, H 6.66, N 8.93.
spectroscopic data are given in the Supporting Information.
C14H20NO3P (281.29): calcd. C 59.78, H 7.17, N 4.98; found: C
59.84, H 7.13, N 5.05.
The less mobile diastereomer was crystallized by evaporation and
was filtered from n-hexane to yield isomer 19b (0.57 g, 30%), which
was recrystallized from diisopropyl ether/ethyl acetate: m.p. 180–
(2R*,4aS*,8aS*)-
and
(2S*,4aS*,8aS*)-1-Methyl-2-phenoxy-
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-Ox-
ides (16a and 16b): The crude product (ratio of the isomers 50:50,
based on the 31P NMR spectrum) was purified by column
chromatography with toluene/ethyl acetate (3:2) as eluent. The
more mobile diastereomer was crystallized by evaporation and was
filtered from n-hexane to yield isomer 16b (0.62 g, 37%), which was
recrystallized from n-hexane/diethyl ether: m.p. 101–103 °C. 1H,
13C and 31P NMR spectroscopic data are given in the Supporting
Information. C14H20NO3P (281.29): calcd. C 59.78, H 7.17, N 4.98;
found: C 59.60, H 7.22, N 5.02.
1
183 °C. H, 13C and 31P NMR spectroscopic data are given in the
Supporting Information. C11H21Cl2N2O2P (315.18): calcd. C 41.92,
H 6.72, N 8.89; found: C 42.11, H 6.80, N 8.96.
(2R*,4aS*,8aS*)- and (2S*,4aS*,8aS*)-2-[Bis(2-chloroethyl)amino]-
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-Ox-
ides (20a and 20b): The crude product (ratio of the isomers 50:50,
based on the 31P NMR spectrum) was purified by column
chromatography with ethyl acetate/methanol (7:1) as eluent. The
compound was then obtained as an approximately 45:55 mixture
of epimers (1.14 g, 60%). 1H, 13C and 31P NMR spectroscopic data
are given in the Supporting Information. C11H21Cl2N2O2P
(315.18): calcd. C 41.92, H 6.72, N 8.89; found: C 41.79, H 6.65,
N 8.94.
The less mobile diastereomer was crystallized from n-hexane to yi-
eld isomer 16a (0.39 g, 23%), which was recrystallized from n-hex-
ane/diethyl ether: m.p. 73–74 °C. 1H, 13C and 31P NMR spectro-
scopic data are given in the Supporting Information. C14H20NO3P
(281.29): calcd. C 59.78, H 7.17, N 4.98; found: C 59.88, H 7.13,
N 4.92.
Supporting Information (see also footnote on the first page of this
1
article): H, 13C and 31P NMR chemical shifts, and JH,H, JH,P and
JP,C coupling constants for compounds 7–20a,b. Solvent and tem-
perature dependencies of coupling constants for selected trans-
fused compounds. X-ray data for compound 16a. Cartesian coordi-
nates and energies of the DFT-optimized geometries.
(2R*,4aS*,8aR*)-
and
(2S*,4aS*,8aR*)-1-Benzyl-2-phenoxy-
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-Ox-
ides (17a and 17b): The crude product (ratio of the isomers 49:51,
based on the 1H NMR spectrum) was purified by column
chromatography with n-hexane/ethyl acetate (2:1) as eluent. The
more mobile diastereomer was identified as pure isomer 17b
(0.58 g, 27%, colorless, transparent oil). 1H, 13C and 31P NMR
spectroscopic data are given in the Supporting Information.
C20H24NO3P (357.38): calcd. C 67.22, H 6.77, N 3.92; found: C
67.30, H 6.69, N 4.03.
Acknowledgments
Z. Z. is grateful for a grant from the Center for International Mo-
bility (CIMO) in Finland. The authors thank Dr. Karel Klika for
help with some NMR experiments. The Finnish IT Center for Sci-
ence (CSC, Espoo, Finland) is thanked for the allocated computing
time.
The less mobile diastereomer was crystallized after evaporation and
was filtered from n-hexane to yield the pure isomer 17a (0.30 g,
14%, m.p. 99.5–101 °C). 1H, 13C and 31P NMR spectroscopic data
are given in the Supporting Information. C20H24NO3P (357.38):
calcd. C 67.22, H 6.77, N 3.92; found: C 67.15, H 6.72, N 3.88.
[1] a) S. E. Denmark, R. L. Dorow, J. Org. Chem. 1990, 55, 5926;
b) S. E. Denmark, C.-T. Chen, J. Org. Chem. 1994, 59, 2922;
c) N. J. Gordon, S. A. Evans, Jr., J. Org. Chem. 1993, 58, 5293;
d) S. E. Denmark, J.-H. Kim, J. Org. Chem. 1995, 60, 7535.
[2] a) K. Afarinkia, H. M. Binch, E. De Pascale, Synlett 2000,
1769; b) K. Afarinkia, H. M. Binch, I. Forristal, Synlett 2000,
1771; c) K. Afarinkia, E. De Pascale, Synlett 2002, 990; d) K.
Afarinkia, E. De Pascale, S. Amara, Arkivoc 2002, 6, 205.
[3] N. Brock, Cancer Res. 1989, 49, 1.
(2R*,4aS*,8aS*)-
and
(2S*,4aS*,8aS*)-1-Benzyl-2-phenoxy-
1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine 2-Ox-
ides (18a and 18b): The crude product (ratio of the isomers 41:59,
based on the 1H NMR spectrum) was purified by column
chromatography with toluene/ethyl acetate (3:2) as eluent. The
more mobile diastereomer was crystallized by evaporation and was
filtered from n-hexane to yield isomer 18b (0.90 g, 42%), which was
[4] a) S. M. Ludeman, G. Zon, J. Med. Chem. 1975, 18, 1251; b)
S. M. Ludeman, V. L. Boyd, J. B. Regan, K. A. Gallo, G. Zon,
K. Ishii, J. Med. Chem. 1986, 29, 716; c) R. F. Borch, G. W.
Canute, J. Med. Chem. 1991, 34, 3044; d) V. Gilard, R. Mar-
tino, M. Malet-Martino, B. Kutscher, A. Müller, U. Niemeyer,
J. Pohl, E. E. Polymeropoulos, J. Med. Chem. 1994, 37, 3986;
e) A. Paci, D. Guillaume, H.-Ph. Husson, J. Heterocycl. Chem.
2001, 38, 1131; f) A. Paci, T. Martens, J. Royer, Bioorg. Med.
Chem. Lett. 2001, 11, 1347; g) L. Hu, Ch. Yu, Y. Jiang, J. Han,
Zh. Li, P. Browne, P. R. Race, R. J. Knox, P. F. Searle, E. I.
Hyde, J. Med. Chem. 2003, 46, 4818.
recrystallized from ethyl acetate: m.p. 136–139 °C. 1H, 13C and 31
NMR spectroscopic data are given in the Supporting Information.
C20H24NO3P (357.38): calcd. C 67.22, H 6.77, N 3.92; found: C
67.30, H 6.81, N 4.00.
P
The less mobile diastereomer was identified as isomer 18a (0.34 g,
16%, colorless, transparent oil) with 18b as a minor impurity (ca.
1
1
9:100 by H NMR). H, 13C and 31P NMR spectroscopic data are
given in the Supporting Information. C20H24NO3P (357.38): calcd.
C 67.22, H 6.77, N 3.92; found: C 67.09, H 6.72, N 3.97.
[5] Int. Patent Appl. WO 02/06293-A1; [Chem. Abstr. 2002, 136,
118577].
(2R*,4aS*,8aR*)-
and
(2S*,4aS*,8aR*)-2-[Bis(2-chloroethyl)
[6] A. E. Shipov, G. K. Genkina, O. I. Artyushin, Z. O.
Mndzhoyan, B. E. Gushchin, E. I. Chumakova, S. A. Ros-
lavtseva, O. Y. Eremina, E. I. Bakanova, Y. S. Kagan, E. A. Er-
shova, T. A. Mastryukova, M. I. Kabachnik, Izv. Akad. Nauk,
Ser. Khim. 1995, 2241 [Chem. Abstr. 1996, 124, 317319].
amino]-1,4,4a,5,6,7,8,8a-octahydro-2H-3,1,2-benzoxazaphosphinine
2-Oxides (19a and 19b): The crude product (ratio of the isomers
50:50, based on the 31P NMR spectrum) was purified by column
chromatography with ethyl acetate/methanol (10:1) as eluent. The
Eur. J. Org. Chem. 2005, 1189–1200
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1199

H. Kivelä, Z. Zalán, P. Tähtinen, R. Sillanpää, F. Fülöp, K. Pihlaja
[7] a) G. S. Bajwa, S. Chandrasekaran, J. H. Hargis, A. E. Sop- [15] W. G. Bentrude in Phosphorus-31 NMR Spectral Properties in
FULL PAPER
chik, D. Blatter, W. G. Bentrude, J. Am. Chem. Soc. 1982, 104,
6385; b) W. G. Bentrude, R. O. Day, J. M. Holmes, Gy.
Szakal Quin, W. N. Setzer, A. E. Sopchik, R. R. Holmes, J.
Am. Chem. Soc. 1984, 106, 106; c) R. R. Holmes, R. O. Day,
W. N. Setzer, A. E. Sopchik, W. G. Bentrude, J. Am. Chem. Soc.
1984, 106, 2353; d) W. N. Setzer, A. E. Sopchik, W. G. Ben-
trude, J. Am. Chem. Soc. 1985, 107, 2083; e) W. G. Bentrude,
W. N. Setzer, A. E. Sopchik, G. S. Bajwa, D. D. Burright, J. P.
Hutchinson, J. Am. Chem. Soc. 1986, 108, 6669; f) W. G. Ben-
trude, W. N. Setzer, A. E. Sopchik, S. Chandrasekaran, M. T.
Ashby, J. Am. Chem. Soc. 1988, 110, 7119; g) K. Afarinkia, R.
Angell, C. L. Jones, J. Lowman, Tetrahedron Lett. 2001, 42,
743; h) S. L. Spassov, M. J. Lyapova, M. E. Ivanova, Phospho-
rus Sulfur 1988, 37, 199.
Compound Characterization and Structural Analysis (Eds.:
L. D. Quin, J. G. Verkade), VCH Publishers, New York, 1994,
pp. 45–46.
[16] P. Tähtinen, A. Bagno, A. Koch, K. Pihlaja, Eur. J. Org. Chem.
2004, 4921.
[17] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Mont-
gomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M.
Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas,
J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C.
Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson,
P. Y. Ayala, Q. Cui, K. Morokuma, P. Salvador, J. J. Dannen-
berg, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B.
Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul, B. B. Ste-
fanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gom-
perts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. G.
Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M.
Head-Gordon, E. S. Replogle, J. A. Pople, Gaussian 98, revision
A.11, Gaussian, Inc., Pittsburgh, PA, 2001.
[8] F. Fülöp, G. Bernáth, K. Pihlaja, Adv. Heterocycl. Chem. 1998,
69, 349.
[9] a) D. G. Gorenstein, R. Rowell, J. Am. Chem. Soc. 1979, 101,
4925; b) D. G. Gorenstein, R. Rowell, J. Findlay, J. Am. Chem.
Soc. 1980, 102, 5077; c) J.-C. Yang, D. O. Shah, N. U. M. Rao,
W. A. Freeman, G. Sosnovsky, D. G. Gorenstein, Tetrahedron
1988, 44, 6305.
[10] a) T. Viljanen, P. Tähtinen, K. Pihlaja, F. Fülöp, J. Org. Chem.
1998, 63, 618; b) T. Viljanen, K. D. Klika, F. Fülöp, K. Pihlaja,
J. Chem. Soc., Perkin Trans. 2 1998, 1479.
[18] a) A. D. Becke, Phys. Rev. A 1988, 38, 3098; b) A. D. Becke, J.
Chem. Phys. 1993, 98, 5648; c) C. Lee, W. Yang, R. G. Parr,
Phys. Rev. B 1988, 37, 785; d) B. Miehlich, A. Savin, H. Stoll,
H. Preuss, Chem. Phys. Lett. 1989, 157, 200.
[11] R. J. Goodridge, T. W. Hambley, D. D. Ridley, Aust. J. Chem.
1986, 39, 591.
[19] J. B. Foresman, Æ. Frisch, Exploring Chemistry with Electronic
Structure Methods, 2nd ed., Gaussian Inc., Pittsburgh, PA,
1996, p. 64.
[12] a) G. Bernáth, K. Kovács, K. L. Láng, Acta Chim. Acad. Sci.
Hung. 1970, 64, 183; b) L. Gera, G. Bernáth, P. Sohár, Acta [20] Z. Otwinowski, W. Minor in Methods in Enzymology (Eds.:
Chim. Acad. Sci. Hung. 1980, 105, 293; c) F. Fülöp, G. Bernáth,
Gy. Argay, A. Kálmán, P. Sohár, Tetrahedron 1984, 40, 2053.
[13] See for example: R. Laatikainen, M. Niemitz, U. Weber, J. Sun-
delin, T. Hassinen, J. Vepsäläinen, J. Magn. Reson. Ser. A 1996,
120, 1; http://www.perchsolutions.com.
[14] a) G. S. Bajwa, W. G. Bentrude, Tetrahedron Lett. 1980, 21,
4683; b) W. J. Stec, W. S. Zielinski, Tetrahedron Lett. 1980, 21,
1361.
C. W. Carter, R. M. Sweet), Academic Press, London, 1997,
vol. 276, Part A, pp. 307–326.
[21] G. M. Sheldrick, SHELXS97 and SHELXL97, University of
Göttingen, Germany, 1997.
[22] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837.
Received: August 30, 2004
1200
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1189–1200