From (E)- and (Z)-ketoximes to N-sulfenylimines, ketimines or ketones at will. Application to erythromycin derivatives

Article abstract of DOI:10.1016/j.tetlet.2004.06.002

Reactions of (E)- and (Z)-ketoximes with trialkylphosphines and diphenyl disulfide (PhSSPh) have been compared to gain insight into the mechanisms involved and their potential applications. N-Sulfenylimine isomers and ketimines have been spectroscopically characterised. Both the E and Z isomers of erythromycin A oxime, when treated with Bu₃P and PhSSPh (1:4:8 ratio), give the same N-phenylsulfenyl ketimine (of configuration E) as the major compound, whereas with Bu₃P or Me₃P and PySeSePy (1:8:4 ratio) afford the imine in good yield. Clarithromycin oxime behaves similarly.

Full text of DOI:10.1016/j.tetlet.2004.06.002

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 5563–5567
From (E)- and (Z)-ketoximes to N-sulfenylimines, ketimines or
ketones at will. Application to erythromycin derivatives
*
*
ꢀ
Jorge Esteban, Anna M. Costa, Felix Urpı and Jaume Vilarrasa
ꢁ
Departament de Quꢁımica Orgꢀanica, Facultat de Quꢁımica, Universitat de Barcelona, 08028 Barcelona, Catalonia, Spain
Received 31 March 2004; revised 24 May 2004; accepted 1 June 2004
Abstract—Reactions of (E)- and (Z)-ketoximes with trialkylphosphines and diphenyl disulfide (PhSSPh) have been compared to
gain insight into the mechanisms involved and their potential applications. N-Sulfenylimine isomers and ketimines have been
spectroscopically characterised. Both the E and Z isomers of erythromycin A oxime, when treated with Bu3P and PhSSPh (1:4:8
ratio), give the same N-phenylsulfenyl ketimine (of configuration E) as the major compound, whereas with Bu3P or Me3P and
PySeSePy (1:8:4 ratio) afford the imine in good yield. Clarithromycin oxime behaves similarly.
Ó 2004 Elsevier Ltd. All rights reserved.
Oximes are at the crossroad of the chemistry of carbonyl
groups and nitrogen functional groups.1 Recovery of
carbonyl compounds from the corresponding oximes,
when the latter have been used to protect2 or purify the
former, or when oximes have been obtained by other
routes, is a subject of continuous interest.3 Whereas
N–O bonds of hydroxylamines are easily cleaved by
reducing agents and whereas C@N bonds of imines are
readily hydrolysed, the C@N–OH group is a robust,
relatively stable substructure. In this connection, the
work carried out by Zard and co-workers4 on the use of
tributylphosphine and diphenyl disulfide is remarkable:
ketoximes react at rt with Bu3P/PhSSPh,5 to afford,
besides Bu3P@O, ketimines and/or N-phenylthio keti-
mines (usually called sulfenylimines or sulfenimines);6
these imines or imine derivatives may be then hydroly-
sed to ketones during the workup or converted to other
compounds.4 Much more recently Lukin and Naraya-
When a similar reaction was performed8 with equimolar
amounts of cyclopentadecanone oxime 1a, Me3P and
PhSSPh, in THF at rt, and the final mixture was directly
separated by column chromatography on alumina or
was first quenched with water buffered at pH 10 and
extracted, N-(phenylsulfenyl)cyclopentadecanimine (N-
SPh derivative 1b)9 was the major product; imine 1c was
not detected but ketone 1d was isolated as a minor
compound. On the other hand, using an excess of Me3P
and quenching the reaction with neutral water, 1d was
mainly obtained.8 Thus, the cleavage of the N–S bond of
1b (mediated by remaining Me3P) was slow in the
reaction medium or during the basic workup, but it took
place rapidly in THF–H2O.10 Moreover, a pure sample
of 1b, treated in a NMR tube (CDCl3) with a THF
solution of Me3P, allowed us to detect immediately
imine 1c,11 which was only fully converted into ketone
1d when a drop of water was added. All these facts may
be summarised as in Scheme 1 (where HX means any
proton source present or added, either PhSH, remaining
oxime or water),12 which complements previous re-
sults.4;7;8 We describe here how to drive the process to-
wards one or another compound as well as
stereochemical details of the first step.
0
00
nan7 have demonstrated, on O2,O4 -dibenzoyl erythro-
mycin A oxime, that its sulfenylimine is the first
long-lived intermediate and that it is slowly cleaved
in situ by Bu3P (in the presence of a proton source, such
as benzenethiol) to give the imine.
To favour the percentage of 1b––to stop the sequence at
the stage of 1b––we repeated the reaction with a less
active phosphine (Bu3P, 1.2 equiv)13 and an excess of
PhSSPh (2.4 equiv), which afforded 1b almost quanti-
tatively. The four anhydrous solvents checked (THF,
CH3CN, CH2Cl2 or CHCl3) gave the same result.
Keywords: Oximes; Sulfenylimines; Ketone protecting groups; Macro-
lide antibiotics; Erythromycin A; Clarithromycin.
* Corresponding authors. Tel.: +34-934021258; fax: +34-933397878;
e-mail: jvilarrasa@ub.edu
0040-4039/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.06.002

5564
J. Esteban et al. / Tetrahedron Letters 45 (2004) 5563–5567
Hex-5-en-2-one oxime 2a, an E–Z mixture (65:35 in
SPh
OH
N
N
CDCl3, 55:45 in THF-d8, 70:30 in CD3CN), when sulf-
enylated in THF at 0 °C with Bu3P and excess of
PhSSPh for 1 h, afforded (E)-2b plus (Z)-2b (75:25 ratio
in CDCl3, 80:20 in THF-d8, 80:20 in CD3CN) in excel-
lent yield; the isomer ratios just indicated were kept after
heating 2a as well as 2b overnight at 50 °C and regis-
tering again the spectra at 18 °C. Thus, they must be the
equilibrium ratios at rt between the respective isomers.14
When the NMR spectra of 1a, (E)-2a, (Z)-2a and their
sulfenylimines were compared, regular trends could be
observed (see Scheme 2), among which we should
highlight the downfield shifts of several ppm (7–10 ppm)
that undergo the relevant dC values of the sulfenyl-
imines as well as the expected fact that the dC values of
methyl and methylene groups anti to the hydroxy and
phenylthio groups are always higher than those of the
corresponding syn groups, as known (steric compres-
sion).
δ
δ
H 2.19
C 34.3
δ
δ
H 2.34
C 27.9
δ
δ
H 2.41
C 40.3
δ
δ
H 2.45
C 36.3
PhSSPh
Me3P
δ
C 172.8
δ
C 162.8
THF
1b
1a
Me3P HX
NH
O
δ
H 2.25
δ
δ
H 2.42
C 42.0
δ
C 39.1
H2O
δ
C 212.6
δ
C 185.0
1d
1c
Scheme 1. Preliminary results. NMR spectral data in CDCl3.12
To check the scope of these experiments, a few non-
symmetric ketones were transformed, by known proce-
dures, to oximes 5a–4a. The reactivity of their E and Z
isomers with Bu3P plus PhSSPh was compared to gain
insight into the first steps of these reactions. The main
results are summarised in Scheme 2.
Acetophenone oxime, (E)-3a, when treated with Bu3P
(1.2 equiv) and PhSSPh (2.4 equiv) at 0 °C or at rt for
several hours, gave only a trace of sulfenylimine; most of
the starting oxime was recovered after the workup. By
sharp contrast, its less stable isomer, (Z)-3a,15 quickly
afforded (E)-3b under the same conditions (Scheme 2).
To force the conversion of (E)-3a to sulfenylimine a
much larger reagent concentration was required: only
when (E)-3a was stirred with 4 equiv of Bu3P and 8 equiv
of PhSSPh at 20 °C did we obtain (E)-3b in excellent
yield, after 1 h of reaction. This sulfenylimine was
known16 and our spectral data were coincident with
those reported. Indeed, only the E isomer of 3b was seen
by NMR in all trials (the methyl carbon of the sulfe-
nylimine E, at 20.0 ppm, can be related, as explained in
the preceding paragraph, to the chemical shift of
12.4 ppm for the methyl carbon of oxime E, whereas for
the methyl carbon of sulfenylimine Z, if it had been
obtained, a chemical shift around 28 ppm would be ex-
pected). We confirmed the reactivity differences by
mixing samples of (E)-3a and (Z)-3a and subjecting the
mixture to sulfenylation (under the mild conditions of
Scheme 2); we noted that oxime Z disappeared, to give
(E)-3b, while (E)-3a remained. In this case the equilib-
rium between the oxime isomers is very slow (and, if
established on heating and/or under basic catalysis, it is
largely shifted towards isomer E). The reactivity differ-
ence between the two isomers may be explained by the
different steric hindrance around the nitrogen atom lone
pairs of the plausible intermediates (as drawn in Scheme
2) and/or by the higher or lower strength of the N–O
bonds depending on the electronic conjugation. We
propose that a configuration inversion occurs, either via
an intramolecular SPh transfer or by an intermolecular
process involving the relative excess of sulfenylating
agent. Probably such a configuration inversion takes
place in most cases, although it cannot be obviously
noted if only one compound is possible (from 1a to 1b)
and it is very difficult to monitor and prove when the
isomeric equilibria are more rapid than the sulfenylation
reaction (as in the conversion of 2a to 2b) or when the
equilibrium between the two sulfenylimines is com-
pletely shifted towards isomer E (as in the case of 3b).
PhS
HO
N
N
35.0
41.9
1.89
13.5
2.07
22.7
167.8
157.7
(E)-2b
(Z)-2b
(E)-2a
(Z)-2a
SPh
OH
N
N
36.9
29.3
1.88
19.8
2.13
27.3
158.1
168.9
PhS
PhSSPh
Bu3P
PBu3
O
OH
N
.
.
N
slow
2.31
12.4
[(Z)-3b]
Ph
PhSH
155.8
(E)-3a
quick
Bu3P
PhSSPh
Bu3P
SPh
HO
Ph
SPh
N
N
O
.
.
quick
2.21
21.4
N
2.47
20.0
Ph
PhSH
154.0
161.6
Ph
(Z)-3a
Bu3PO
(E)-3b
OH
SPh
N
N
quick
37.2
slow
44.2
2.02
18.5
1.88
10.0
[(Z)-4b]
164.2
173.3
(E)-4b
(E)-4a
:
:
16.9
17.0
N
N
37.1
42.8
OH
SPh
163.0
173.8
H
H
34.2
3.10
2.80
H
2.40
2.32 H
1.90
24.8
(E)-5a
(E)-5b
H
H
2.16
16.4
26.7
3.60
15.4
37.2
H
2.38
2.22
H
H
3.10
175.1
H
163.6
35.5
H
2.52
28.4
N
SPh
H
2.30
N
OH
..
..
(Z)-5b
(Z)-5a
Scheme 2. Reagents and conditions: Bu3P (1.2 equiv) and PhSSPh
(2.4 equiv), 0.1–0.2 M THF solutions, 0–20 °C. Relevant 1H and 13C
chemical shifts in CDCl3 are indicated.12

J. Esteban et al. / Tetrahedron Letters 45 (2004) 5563–5567
5565
The relative significance of the steric effects was checked
with pinacolone oxime [(E)-3,3-dimethylbutan-2-one
oxime, (E)-4a].17 Under the mild conditions of Scheme
2, it hardly reacted. As in the case of (E)-3a, we had to
increase four times the concentration of both Bu3P and
PhSSPh to observe reaction rates similar to those of
oximes 1a and 2a. Also, we only detected and isolated
the E sulfenylimine, (E)-4b. We suggest in Scheme 2 that
it is due to the expected fact that the E–Z equilibrium is
more shifted than ever towards the less congested iso-
mer. Thus, the close parallelism between the reactions of
pinacolone and acetophenone derivatives indicates, at
least, that there is one rate-limiting step in which the
steric effects are important.
dissolving the precipitate in EtOAc, followed by chro-
matography on alumina, a N-SPh derivative, 6b, could
be isolated in 40–50% yields, while imine 6c (see below)
was later eluted and obtained in 20–30% yields. That is
to say, the sulfenylimine had been formed in high yields,
but it was partially cleaved and/or hydrolysed to the
imine during the isolation procedure. The well-known
stability towards hydrolysis of erythromycin-related
imines20 explains why 6c survived during the isolation
procedure.
The Z oxime, (Z)-6a,21 treated separately under identical
conditions gave the same N-SPh derivative, 6b, and
imine 6c as a minor compound. Comparison of the
NMR parameters of the product 6b with those of the
two oximes22 (Scheme 3), as well as with those of (E)-5a
and (E)-5b (Scheme 2), suggested that, as expected, it
was (E)-6b, bearing in mind: (i) the 13C chemical shifts of
the a carbon atoms (for the a methyne groups of the
unknown Z isomer of 6b we would have expected dC
values of 38–40 ppm and ca. 45 ppm, respectively,
2-Methylcyclohexanone oxime 5a, an E–Z mixture (ca.
90:10 in CDCl3), when treated at 0 °C with Bu3P/
PhSSPh for 1 h, gave an E–Z mixture of 5b (75:25 in
CDCl3) in ca. 80% yield. In agreement with a confor-
mational study of (Z)-5a,18 we noted that (Z)-5b prefers
the 4C1 conformation (with an axial methyl group), thus
avoiding the steric hindrance that would result in the
other chair conformation (with an equatorial methyl
group and a syn NOH). Secondly, the rigid chair con-
formations allowed us, in addition, to use as relevant
data also the dH of the axial and equatorial hydrogen
atoms, more or less shifted downfield depending on their
proximity to the heteroatoms (when in the preceding
examples we only saw mean chemical shifts for the
methylene hydrogens and no significant changes from
oximes to sulfenylimines); in the present case, the
equatorial hydrogen atoms near the OH group showed
the highest dH values. Thirdly, the a methylene carbon
atoms follow the general trend already mentioned in
previous examples for the a carbon atoms (a downfield
shift of several ppm––between 5.7 and 10.5 ppm until
now––from each oxime to its analogue sulfenylimine).
Fourthly, it is clear that the 2-Me carbon atoms do not
show significant changes among them and with respect
to the Me carbon atom of the parent compound 2-
methylcyclohexanone (ca. 15 ppm), probably as no
important steric compression effects do exist in the four
structures depicted.
OH
N
9
8
10
OR
HO
NMe2
OH
HO
O
O
O
SPh
O
N
MeO
O
O
OH
(E)-6a, R = H
OR
OH
(E)-7a, R = Me
HO
ODes
O
HO
N
O
OCla
(E)-6b, R = H
(E)-7b, R = Me
OR
OH
O
HO
ODes
NH
O
OCla
OR
(Z)-6a, R = H
(Z)-7a, R = Me
OH
HO
ODes
OCla
O
O
With these results in hand, we could tackle the more
complex substrates that really attracted our attention,
the oximes of macrolide antibiotics erythromycin A and
clarithromycin.19 Erythromycin A (E)-oxime, (E)-6a, the
common EA oxime, treated in THF at 0 °C for 1 h with
an excess of Bu3P and PhSSPh (1:4:8 ratio), gave a
mixture in which no oxime remained; the main and less
polar spot on TLC was a N-SPh derivative, 6b, and the
more polar and very minor one could be the imine, 6c,
according to the MALDI MS spectra. With lower con-
centrations of reagents the reaction was not complete––
with only 200 mol % of Bu3P, the N-SPh derivative was
hardly detected––as in the previous cases of oximes with
sterically crowded nitrogen atoms and/or as if the free
hydroxy groups of 6a had consumed some equivalents
of the reagent mixture to give rise to the corresponding
phosphonium alkoxides (fortunately without further
reactions). Separation of the reaction mixture by pour-
ing it into an excess of aqueous K2CO3, filtering and
6c, R = H
7c, R = Me
:
N
14.3
32.6
10
OH
HO
171.3
:
H
3.74
8
N
14.7
18.7
H
2.65
25.4
36.4
SPh
HO
(E)-6a
OH
182.5
18.7
H
3.20
37.9
H
2.87
10.8
N
(E)-6b
:
34.1
HO
168.1
H
2.93
H
2.79
35.7
(Z)-6a
19.8
Scheme 3. Reactions of erythromycin A and clarithromycin oximes.
Relevant chemical shifts in CDCl3 are given for the erythromycin
series. For clarithromycins the chemical shifts are very close to those
indicated.

5566
J. Esteban et al. / Tetrahedron Letters 45 (2004) 5563–5567
according to the trends mentioned above); (ii) the
chemical shift of the 10-Me group (equatorial), which is
almost identical for (E)-6a and (E)-6b, as it was the case
for the equatorial Me of (E)-5a and (E)-5b, whereas the
congested 10-Me of (Z)-6a appears at 10.8 ppm; and (iii)
the dH values for equatorial hydrogens, which show a
parallelism with those observed in the case of (E)-5a and
(E)-5b (see above). We should remark that no signals
attributable to sulfenylimine Z were detected in the
crude product or in equilibrium with the E isomer.
Apparently, the less stable Z isomer is in the case of 6a
much more disfavoured than in the case of 5,23 a fact
that can be accounted for on steric grounds and by the
lack of an alternative chair conformation because of the
restrictions imposed by the macrolide ring.
ylsulfenyl derivatives (substrate/Bu3P/PhSSPh in the
1:n:2n ratio), imines (substrate/Me3P/PySeSePy in the
1:2n:n ratio) or ketones (1:2n:n, followed by addition of
water). Applications of these reactions to the protection
and modification of erythromycin-related antibiotics are
in progress.
Acknowledgements
Thanks are due to the Ministerio de Ciencia y Tecno-
ꢁ
logıa and the Generalitat de Catalunya for support, and
ꢁ
ꢁ
to G. Martınez and Dr. M. Martın for preliminary
experiments. A generous gift of clarithromycin oxime
from Fyse–Ercros (Aranjuez) deserves to be mentioned.
Comments of a referee regarding the possible causes of
the lower reactivity of some E oximes are also
acknowledged.
Clarithromycin oxime,24 when treated with an excess of
Bu3P and PhSSPh (1:4:8 ratio as above) in 1:1 THF–
dioxane (to solubilise the oxime), gave a high percentage
of conversion to (E)-7b,25 even though as in the previous
case after separation by column chromatography on
alumina we obtained only a 48% yield (and later a 25%
yield of imine 7c was eluted). Again, only the more
stable E isomer was formed or detected by NMR (see
Scheme 3).
References and notes
ꢁ
ꢁ
ꢁ
1. Martın, M.; Martınez, G.; Urpı, F.; Vilarrasa, J. Tetra-
hedron Lett. 2004, 45, see preceding paper. doi:10.1016/
j.tetlet.2004.06.001.
Having established that, in all the reactions of 1a–7a
with Bu3P and PhSSPh, the 1:n:2n ratio favours the
formation of the N-SPh derivatives (1b–7b), we focused
our attention on finding suitable conditions for obtain-
ing the ketimines as major compounds. It was achieved
with a more active reagent combination.1 Thus, imine 1c
was the unique cyclopentadecanone derivative observed
by carrying out the reaction of oxime 1a with Me3P
(2.4 equiv) and PySeSePy (1.2 equiv) in a NMR tube in
CDCl3 (confirming the data shown in Scheme 1). The
success of this method is based on the fact that in a
preceding work1 we had not observed or detected any N-
SePy intermediate. Our working hypothesis was that this
intermediate and/or a previous one are very quickly
cleaved by the phosphine present in the medium. Iso-
lation of 1c as a pure compound was not possible due to
its easy hydrolysis, but the sample was stable protecting
it from the moisture. This imine can be trapped,4 but we
preferred to convert it to its parent ketone 1d by pouring
the crude product into aqueous ethanol. The hydrolysis
was instantaneous and complete.
2. (a) Greene, T. W.; Wuts, P. G. M. Protective Groups in
Organic Synthesis; Wiley: New York, 1999; p 355; (b)
Robertson, G. M. In Comprehensive Organic Functional
Group Transformations; Katritzky, A. R., Met-Cohen, O.,
Rees, C. W., Pattenden, G., Eds.; Pergamon: Oxford,
1995; Vol. 3, p 425.
ꢀ
3. For a review, see: Corsaro, A.; Chiacchio, U.; Pistara, V.
Synthesis 2001, 1903, Also see Ref. 2 of the preceding
paper.
4. (a) Barton, D. H. R.; Motherwell, W. B.; Simon, E. S.;
Zard, S. Z. J. Chem. Soc., Perkin Trans. 1 1986, 2243;
Preliminary communications: (b) Barton, D. H. R.;
Motherwell, W. B.; Simon, E. S.; Zard, S. Z. J. Chem.
Soc., Chem. Commun. 1984, 337; (c) Barton, D. H. R.;
Motherwell, W. B.; Zard, S. Z. Tetrahedron Lett. 1984, 25,
3707.
5. The reactive species may be a charge transfer complex (P–
S interaction), and/or a P-tetracoordinated ionic species,
R3PþSPhÀSPh, and/or
a P-pentacoordinated species,
R3P(SPh)2, depending on the medium polarity, tempera-
ture, etc. For an entry to the subject, see: Godfrey, S. M.;
Kelly, D. G.; McAuliffe, C. A.; Mackie, A. G.; Pritchard,
R. G.; Watson, S. M. J. Chem. Soc., Chem. Commun.
1991, 1163. Generally speaking, in R3P/XY equilibrium
mixtures several P-tetravalent species (R3PXþYÀ and
R3PYþXÀ, but also R3PXþXÀ and R3PYþYÀ, arising
from exchange reactions) and/or P-pentavalent species
(R3PXY, as well as R3PXX and R3PYY) may be present.
6. Craine, L.; Raban, M. Chem. Rev. 1989, 89, 689.
7. (a) Lukin, K. A.; Narayanan, B. A. Tetrahedron 2002, 58,
215; Also see: (b) Lukin, K. A. PCT Int. Appl., WO 2002,
050093A2, 27 June 2002.
When oximes (E)-6a and (E)-7a were independently
treated with PySeSePy (4 equiv) and a larger excess of
Me3P or Bu3P (8 equiv) for 10 min at 0 °C, only imines
6c and 7c were observed (MALDI, NMR). They were
isolated, after a cold aqueous workup at pH 10 and
extraction with dichloromethane, in ca. 80% yields.26
Thus, in these exceptional cases, owing to the relatively
high kinetic stability of these macrocyclic, congested
imines, their isolation and full characterisation is feasi-
ble. To hydrolyse these imines to the parent macrolides
we dissolved them into aqueous ethanol and added some
drops of dilute HCl until pH 5–6.
ꢁ
8. Martınez, G. Master Thesis (Reducciꢁo d’oximes amb
trimetilfosfina i disulfurs o diselenurs); Universitat de
Barcelona, 1997.
9. N-(Phenylsulfenyl)cyclopentadecanimine (1b): colourless
oil; Rf 0.75 (50:50 EtOAc/hexane); IR (film) 1711, 1610,
1584, 1478, 1441 cmÀ1
;
1H NMR (CDCl3, 400 MHz)
In conclusion, under appropriate conditions, one can
stop or drive the conversion of ketoximes to N-phen-
d 7:50 (pseudo dd, J ¼ 7:2 Hz, J ¼ 1:2 Hz, 2H), 7.33
(pseudo t, J ꢀ 7:2 Hz, 2H), 7.19 (tt, J ¼ 7:2 Hz,

J. Esteban et al. / Tetrahedron Letters 45 (2004) 5563–5567
5567
J ¼ 1:2 Hz, 1H), 2.45 (t, J ¼ 7:6 Hz, 2H), 2.41 (t,
J ¼ 7:6 Hz, 2H), 1.76–1.64 (m, 4H), 1.48–1.30 (m, 20H);
13C NMR (CDCl3, 100.6 MHz) d 172.8, 139.6, 128.7,
125.5, 124.8, 40.3, 36.3, 27.8, 27.3, 26.7, 26.5, 26.4, 26.3,
26.2, 26.1, 25.0, 23.8; MS (CI) m=z 334 (Mþ3þ, 7), 333
(Mþ2þ, 25) 332 (Mþ1þ, 100); HRMS (þFAB) calcd for
[MþH]þ C21H34NS 332.2412, found 332.2421.
Costa, A. M.; Vilarrasa, J. Tetrahedron Lett. 2000, 41,
3371.
20. Davies, J. S.; Hunt, E.; Zomaya, I. I. J. Chem. Soc., Perkin
Trans. 1 1990, 1409.
21. Prepared according to: Wilkening, R. R.; Ratcliffe, R. W.;
Doss, G. A.; Bartizal, K. F.; Graham, A. C.; Herbert, C.
M. Bioorg. Med. Chem. Lett. 1993, 3, 1287.
22. For a conformational analysis of (E)-6a, see: McGill,
J. M.; Johnson, R. Magn. Res. Chem. 1993, 31,
273.
10. The quick cleavage of the N–S bond by phosphines and
water was confirmed by treating a pure sample of
sulfenimine 1b with an equivalent amount of Me3P in
THF–H2O at rt (Ref. 8), in agreement with the results of
Lukin and Narayanan with a relevant erythromycin
derivative (Ref. 7). Ketone 1d was detected immediately
by TLC and was quantitatively isolated after half an hour.
Hydrolysis of 1b in THF–H2O, in the absence of the
phosphine, either in acidic media or in the presence of
cyanide ion, required an overnight stirring (Ref. 8).
Phenylsulfenylimines are much more resistant to conven-
tional hydrolyses than the corresponding imines. See:
Davis, F. A.; Slegeir, W. A. R.; Evans, S.; Schwartz, A.;
Goff, D. L.; Palmer, R. J. Org. Chem. 1973, 38, 2809.
11. This imine was also detected as an intermediate when all
23. It would be interesting to investigate whether both
pathways to (E)-6b from (E)-6a, summarised by means
of arrows at the top of Scheme 3, are involved or not. In
other words, if (Z)-6b is formed first and it immediately
isomerises to (E)-6b, or if (E)-6b comes mainly from the
small percentage of (Z)-6a that may be in equilibrium with
(E)-6a under the reaction conditions. However, such a
study is outside the scope of this communication.
24. The samples we have utilised were 90:10 E/Z. cf. Waddell,
S. T.; Santorelli, G. M.; Blizzard, T. A.; Graham, A.; Occi,
J. Bioorg. Med. Chem. Lett. 1998, 8, 1321.
25. Spectral data of (E)-7b: 1H NMR (CDCl3, 300 MHz) d
7.43(pseudo dd, J ¼ 8:4 Hz J ¼ 1:2 Hz, 2H, Ho), 7.34
(pseudo dd, J ꢀ 8:4 Hz, J ꢀ 7:2 Hz, 2H, Hm), 7.18 (tt,
J ¼ 7:2 Hz, J ¼ 1:4 Hz, 1H, Hp), 5.11 (dd, J ¼ 11:1 Hz,
J ¼ 2:1 Hz, 1H, H13), 4.94 (d, J ¼ 4:5 Hz, 1H, H100), 4.91
(s, 1H, 11-OH), 4.49 (d, J ¼ 7:2 Hz, 1H, H10), 4.04 (dq,
J ¼ 9:0 Hz, J ¼ 6:5 Hz, 1H, H500), 3.78–3.71 (m, 3H, H3,
H5, H11), 3.58–3.44 (m, 1H, H50), 3.33 (s, 3H, 300-OMe),
3.30 (d, J ¼ 11:7 Hz, 1H, H20), 3.18 (s, 3H, 6-OMe), 3.24–
3.10 (m, 1H, H8), 3.02 (t, J ¼ 9:6 Hz, 1H, H400), 3.00–2.86
(m, 1H, H2), 2.77 (q, J ¼ 7:2 Hz, 1H, H10), 2.48–2.40 (m,
1H, H30), 2.36 (d, J ¼ 15:3Hz, 1H, H2 00b), 2.29 (s, 6H,
NMe2), 2.18 (d, J ¼ 8:7 Hz, 1H, 400-OH), 2.03–1.84 (m,
2H, H14b, H4), 1.80–1.41 (m, 8H, 6-Me, 0H0 40b, H7a, H7b,
H200a, H14a), 1.30 (d, J ¼ 6:3Hz, 3H, 5 -Me), 1.25–1.19
(m, 13H, 10-Me, 2-Me, 50-Me, 300-Me, H40a), 1.13(s, 3H,
12-Me), 1.11 (d, J ¼ 7:5 Hz, 4-Me), 1.09 (d, J ¼ 6:6 Hz,
3H, 8-Me), 0.83 (t, J ¼ 7:5 Hz, 3H, H15) (assignments
confirmed by COSY and HSQC); 13C NMR (CDCl3,
75.3MHz) d 182:2 (C9), 175.2 (C1), 138.9 (Cipso), 128.9
(Cm), 126.0 (Cp), 125.5 (Co), 102.5 (C10), 96.1 (C100), 80.1
(C5), 78.6 (C3), 78.3 (C400), 77.9 (C6), 76.7 (C13), 74.1
(C12), 72.7 (C300), 71.0 (C20), 70.2 (C11), 68.6 (C50), 65.7
(C500), 65.5 (C30), 52.1 (6-OMe), 49.4 (300-OMe), 44.9 (C2),
1
the process was monitored by H NMR in CD3CN. The
rapid proton exchange makes both sides of 1c magneti-
cally equivalent. For NMR spectra of relevant imines, see:
(a) Shoppee, C. W.; Henderson, G. N. J. Chem. Soc.,
Perkin Trans. 1 1977, 1028; (b) Findeisen, K.; Heitzer, H.;
Dehnicke, K. Synthesis 1981, 702; (c) Richey, H. G.;
Erickson, W. F. J. Org. Chem. 1983, 48, 4349.
1
12. The NMR assignments shown were corroborated by H–
13C 2D experiments (HSQC).
13. No reaction occurs with Ph3P under the same conditions.
14. The E–Z isomerisation energy barriers of N-sulfenyl
ketimines are, in general, relatively low (ca. 15–20
kcal/mol, coalescence temperatures around 50–70 °C).
See: (a) Brown, C.; Grayson, B. T.; Hudson, R. F.
Tetrahedron Lett. 1970, 4925; (b) Davis, F. A.; Slegeir, W.
A. R.; Kaminski, J. M. J. Chem. Soc., Chem. Commun.
1972, 634; (c) Brown, C.; Hudson, R. F.; Grayson, B. T. J.
Chem. Soc., Chem. Commun. 1978, 156; Also see: (d)
Bharatam, P. V.; Amita; Kaur, D. J. Phys. Org. Chem.
2003, 16, 183.
15. Smith, J. H.; Kaiser, E. T. J. Org. Chem. 1974, 39, 728.
16. Morimoto, T.; Nezu, Y.; Achiwa, K.; Sekiya, M. J. Chem.
Soc., Chem. Commun. 1985, 1584; Also see: Ref. 10.
17. (a) Karabatsos, G. J.; Taller, R. A. Tetrahedron 1968, 24,
3347; (b) Hawkes, G. E.; Herwig, K.; Roberts, J. D. J.
Org. Chem. 1974, 39, 1017.
40.3(NMe ), 38.5 (C4), 38.3 (C7), 37.8 (C8), 36.3 (C10),
2
34.9 (C200), 28.6 (C40), 21.5 (50-Me), 21.5 (300-Me), 21.1
(C14), 20.5 (6-Me), 18.6 (500-Me), 18.6 (8-Me), 16.1 (2-Me
or 10-Me), 15.9 (12-Me), 15.0 (10-Me or 2-Me), 10.6
(C15), 9.2 (4-Me); MS (MALDI) m=z 855.6 (Mþ1þ).
26. Compounds 6c and 7c thus obtained were identical (TLC,
NMR) to those prepared by us from the TiCl3/HCl/
NH4OAc reduction of (E)-6a and (E)-7a, respectively. See:
Timms, G. H.; Wildsmith, E. Tetrahedron Lett. 1971, 12,
195.
18. (a) Ribeiro, D. S.; Abraham, R. J. Magn. Res. Chem. 2002,
40, 49, and references cited therein; Also see: (b) Geneste,
P.; Durand, R.; Kamenka, J.-M.; Beierbeck, H.; Martino,
R.; Saunders, J. K. Can. J. Chem. 1978, 56, 1940.
19. For reports from our lab on chemical modifications of
ꢁ
erythromycin-like macrolides, see: (a) Bartra, M.; Urpı,
F.; Vilarrasa, J. Tetrahedron Lett. 1992, 33, 3669; (b)