A tuneable method for N-debenzylation of benzylamino alcohols

Article abstract of DOI:10.1021/ol050624f

(Chemical Equation Presented) N-lodosuccinmide provides a mild, convenient, and tuneable reagent for the selective mono- or didebenzylation in representative, multifunctionalized carbohydrate and amino acid derived N-dibenzylamines with neighboring O-func

Full text of DOI:10.1021/ol050624f

ORGANIC
LETTERS
2005
Vol. 7, No. 12
2361-2364
A Tuneable Method for N-Debenzylation
of Benzylamino Alcohols
Elizabeth J. Grayson† and Benjamin G. Davis*
Department of Chemistry, UniVersity of Oxford, Chemistry Research Laboratory,
South Parks Road, Oxford OX1 3TA, UK
ben.daVis@chem.ox.ac.uk
Received March 23, 2005
ABSTRACT
N-Iodosuccinmide provides
a mild, convenient, and tuneable reagent for the selective mono- or didebenzylation in representative,
multifunctionalized carbohydrate and amino acid derived N-dibenzylamines with neighboring O-functionality.
The best-established method for the debenzylation of diben-
zylamino groups involves hydrogenolysis over a heteroge-
neous palladium catalyst;1 accompanying selectivity for
mono- or di-N-debenzylation is rare. Incompatibility exists
with other common hydrogenolyzable groups, particularly
in sugars, such as Bn, Bz, or benzylidene. Davies et al. have
described the monodebenzylation of differentially protected
benzyl(p-nitrobenzyl)amines with ceric ammonium nitrate.2
Debenzylation of benzylamines can also be achieved using
other oxidizing agents.3 However, a flexible homogeneous
system with a wider range of potential outcomes starting from
simple dibenzylamine substrates would be useful.
compatibility of other protecting groups and functionality
(e.g., benzylidene/methyl acetals, silyl ethers).8
In the course of our ongoing studies into the generation
of imines through N-halogenation9 we considered that
sequential halogenation, elimination, and hydrolysis might
provide a route to ready N-debenzylation. Initial studies
revealed that although N-chlorosuccinimide was poorly
effective, N-iodosuccinimide (NIS) could indeed effect
debenzylation of dibenzylamines, and we show here its utility
and selectivity in a variety of biomolecule component
substrates, namely, carbohydrates and amino acids. Various
Access to differentially protected carbohydrate derivatives,4
especially hexosamine derivatives,5 is attractive because of
their use in carbohydrate-scaffold libraries.6 In particular the
N-2 dibenzylamine substituent is a useful participatory trans-
directing group in glycosyl donors.7 Selective manipulations
of complex hexosamine substrates such as 1 or 7 not only
provide access to such systems but also provide a test of the
(5) (a) Jain, R.; Kamau, M.; Wang, C.; Ippolito, R.; Wang, H.; Dulina,
R.; Anderson, J.; Gange, D.; Sofia, M. J. Bioorg. Med. Chem. Lett. 2003,
13, 2185. (b) Nair, L. G.; Fraser-Reid, B.; Szardenings, A. K. Org. Lett.
2001, 3, 317. (c) Abrous, L.; Hynes, J.; Friedrich, S. R.; Smith, A. B.;
Hirschmann, R. Org. Lett. 2001, 3, 1089. (d) Ghosh, M.; Dulina, R. G.;
Kakarla, R.; Sofia, M. J. J. Org. Chem. 2000, 65, 8387. (e) Liang, R.; Yan,
L.; Loebach, J.; Ge, M.; Uozumi, Y.; Sekanina, K.; Horan, N.; Gildersleeve,
J.; Thompson, C.; Smith, A.; Biswas, K.; Still, W. C.; Kahne, D. Science
1996, 274, 1520.
(6) (a) Le, G. T.; Abbenante, G.; Becker, B.; Grathwohl, M.; Halliday,
J.; Tometzki, G.; Zuegg, J.; Meutermans, W. Drug DiscoVery Today 2003,
8, 701. (b) Peri, F.; Cipolla, L.; Forni, E.; Nicotra, F. Monatsh. Chem. 2002,
133, 369. (c) Sofia, M. J.; Silva, D. J. Curr. Opin. Drug DiscoVery DeV.
1999, 2, 365. (d) Sofia, M. J.; Hunter. R.; Chan, T. Y.; Vaughan, A.; Dulina,
R.; Wang, H.; Gange, D. J. Org. Chem. 1998, 63, 2802.
† Dept. of Chemistry, Durham University, South Rd, Durham DH1 3LE,
UK.
(1) Greene, T. W.; Wuts, P. G. M. Protecting Groups in Organic
Synthesis, 3rd ed.; Wiley: New York, 1999.
(2) Bull, S. D.; Davies, S. G.; Fenton, G.; Mulvaney, A. W.; Prasad, R.
S.; Smith, A. D. J. Chem. Soc., Perkin Trans. 1 2000, 3765.
(3) Talukdar, S.; Nayak, S. K.; Banerji, A. J. Org. Chem. 1998, 63, 4925.
(4) (a) Zhu, T.; Boons, G-J. Tetrahedron: Asymmetry 2000, 11, 199.
(b) Prabhu, A.; Venot, A.; Boons, G.-J. Org. Lett. 2003, 5, 4975. (c) Tanaka,
H.; Amaya, T.; Takahashi, T. Tetrahedron Lett. 2003, 44, 3053. (d) Peri,
F.; Nicotra, F.; Leslie, C. P.; Micheli, F.; Seneci, P.; Marchioro, C. J.
Carbohydr. Chem. 2003, 22, 57.
(7) Jiao, H.; Hindsgaul, O. Angew. Chem., Int. Ed. 1999, 38, 346.
(8) Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups in Organic
Synthesis, 3rd. ed.; John Wiley & Sons Inc.: New York, 1999.
(9) (a) Davis, B. G.; Maughan, M. A. T.; Chapman, T. M.; Villard, R.;
Courtney, S. Org. Lett. 2002, 4, 103. (b) Chapman, T. M.; Courtney, S.;
Hay, P.; Davis, B. G. Chem. Eur. J. 2003, 9, 3397. (c) Maughan, M. A. T.;
Davies, I. G.; Claridge, T. D. W.; Courtney, S.; Hay, P.; Davis, B. G. Angew.
Chem., Int. Ed. 2003, 42, 3788.
10.1021/ol050624f CCC: $30.25
© 2005 American Chemical Society
Published on Web 05/14/2005

parameters were investigated: (a) equivalents of NIS, (b)
Lewis acid additives, (c) base additives, (d) the role of water,
and (e) reaction time (Table 1 and SI (p 36) for entries) to
Scheme 1
Table 1. Debenzylations with NIS
with secondary Zemple´n deprotection, 2 was obtained in an
excellent 98% overall yield from 1 (entry 8). Monobenzoyl-
monobenzyl sugar 3 could also be obtained through separa-
tion in 50% yield using this route.
Onset of NIS-mediated reaction was manifested by the
development of a maroon color typical of I2. To establish
whether I2 was effecting debenzylation, 1 was treated directly
with I2 but showed no reaction (even up to 3 equiv);
repetition with base TTBP showed only poor debenzylation
(54% 2 after 48 h, Entry 10).
Formation of benzamide 6 after prolonged reaction (20 h
rather than 4) of 1 with 3 equiv of NIS (entry 9) interestingly
indicated that, as well as allowing efficient monodebenzyla-
tion, other products might result from the NIS deprotection
system through suitable tuning of conditions. Indeed, prolong-
ed reaction (20 h) of 1 with 6 equiv of NIS and 1 equiv of
water (entry 11) gave 6 in 46% and for the first time the
didebenzylated product 5 (21%). Indication of didebenzyla-
tion was a highly promising addition to the NIS deprotection
protocol and was improved. Greater excess of NIS (10 equiv)
increased formation of 5 (58% 6 and 33% 5, entry 12; 52%
5 and 39% 6 with added water, entry 13). Prolonged reaction
time (1 week, entry 14) gave solely 6 (50%) at the expense
of 5. Additives again proved key: 10 equiv of NIS, 0.5 equiv
of TESOTf, 1 equiv of lutidine, and shorter reaction time of
4 h gave 68% didebenzylated sugar 5 and 30% 6 from 1
(entry 15). Sequential coupling with Zemple´n deprotection
gave a method for didebenzylation of 1 in a workable overall
yield of 4 of 64% (entry 16). Added water gave more 6;
68% of 6 was obtained from 1 with 10 equiv of NIS, 0.5
equiv of TESOTf, 1 equiv of lutidine under moist conditions
and in a shortened reaction time of only 2 h (entry 17).
Having established conditions for the preparation of five
differentially protected products from 1 simply by tuning con-
ditions (Scheme 2), we next explored other substrates. For-
mation of monobenzoyl-monobenzyl 3 from 1 clearly neces-
sitated participation of neighbouring OH-3. To investigate
the need for a neighboring OH group, as is the case for some
O-debenzylations,10 1 was converted to 3-O-silyl ether 7
(97%). Treatment of 7 with the monodebenzylation condi-
a Entries not listed can be found in the full table in Supporting
Information. Reaction conditions: (A) “dry” - 4 Å sieve, dry DCM, room
temperature, under dry N2; (B) “wet” - DCM, room temperature, open to
atmosphere or addition of 1 equiv of H2O. b Parentheses indicate yield based
on recovered starting material.
reveal a convenient debenzylation system that allows tune-
able levels of deprotection (Scheme 1).
Dibenzylglucosamine 1 was chosen as a representative
complex substrate containing multiple protecting groups.
Reaction with 1 equiv of NIS in the presence of sieves did
not proceed to completion (50% monodebenzylated product
2 after 4 d, entry 1). Two equivalents of NIS (entry 2) gave
59% 2 with concomitant formation of monodebenzylated
3-O-benzoyl 3 (37% after 20 h) in a combined monodeben-
zylation yield of 96%. Prolonged treatment, Lewis acid
(TESOTf), or base (TTBP) reduced yield (entries 3-5).
Increasing NIS levels further (3 equiv) increased yield and
rate (entry 7, 2 and 3 formed in 50% after 4 h), allowing
quantitative overall monodebenzylation of 1. When coupled
(10) Madsen, J.; Viuf, C.; Bols, M. Chem. Eur. J. 2000, 6, 1140.
Org. Lett., Vol. 7, No. 12, 2005
2362

Scheme 2
Scheme 3
Applicability to sugars having been demonstrated, non-
carbohydrate substrates were investigated (Scheme 4). L-
Scheme 4
tions determined for 1 (3 equiv of NIS, 4 h) smoothly
effected monodebenzylation to 8 (87%, entry 18). Clean
deprotection of 8 with TBAF gave 2 in 95%; the three-step
conversion of 1 f 7 f 8 f 2 represents an effective (80%
overall) alternative to direct monodebenzyation of 1.
Treatment of 7 under conditions for didebenzylation (10
equiv of NIS) gave 9 in 63-72% (entries 19 and 20).11
Moreover, debenzylation of monobenzylamine 8 successfully
gave fully debenzylated 9 in 79% yield after 24 h (entry
23). Deprotection of 9 with TBAF gave 4 in 73%; three-
step conversion 1 f 7 f 9 f 4 represents an effective (62%
overall) alternative to the direct monodebenzyation of 1
(64%). Attempts to debenzylate 10 with free OH-6 were
sluggish, and the reaction was less clean than for monohy-
droxyl 1; 3 equiv of NIS gave a poor 23% of 11 (entry 24).
However, treatment of 10 with didebenzylation conditions
gave 50% of 12 (entry 25).
Because NIS is also an activator of thioglycoside glycosyl
donors in glycosylation reactions12 we next explored the intri-
guing potential for one-pot glycosylation-deprotection reac-
tions; initial analyses suggested faster glycosylation than de-
benzylation at room temperature. Pleasingly, reaction of ethyl
thioglucoside donor13 with dibenzylamine acceptors 1 and 10
using slightly modified monodebenzylation conditions (3
equiv of NIS, 0.5 equiv of TESOTf, 1 equiv of TTBP) suc-
cessfully gave monodebenzylated coupling products 14 and 15
(Scheme 3)), respectively.13 Although yields are moderate
(35%-42%, 2 steps), access to unusual monobenzylamine
disaccharides is useful and highlights the potential for
cascade glycosylations using both dibenzylamine donors7 and
acceptors.
Phenylalanine-derived 20 was monodebenzylated to 21 in
88% (1.5 h, entry 26) and didebenzylated to 22 in 65% yield
(entry 28). Various L-serine derivatives were examined next.
Methyl ester 25 and benzyl ester 27 both monodebenzylated
smoothly (5 h) to give 26 (83%) and 28 (85%), respectively
(entries 29 and 30). A second course of monodebenzylation
converted 28 in fair yield (51%) to fully debenzylated Ser
benzyl ester 29 (entry 31). Substrates containing free
hydroxyl groups â and γ to the NBn2, dibenzylphenylalaninol
16 and γ-dibenzyl alcohol 30, monodebenzylated with
simultaneous benzoylation of OH giving monobenzoyl-
monobenzyl amino acids 18 (69%) and 31 (53%), respec-
tively (entries 32 and 33).14
(11) Interestingly, treatment of 7 with 5 equiv of NIS, followed by 1
equiv of water (2 h when TLC indicated no 7) gave 9 in improved 87%
yield (entry 21).
(12) Davis, B. G. J. Chem. Soc., Perkin Trans. 1 2000, 2137
(13) Low nucleophilicity acceptors gave 47-49% glycosylsuccinimide.
(14) Lower yield of 31 may be from volatility.
(15) Three equivalents of NIS gave 60% 24 with 20% monodebenzylated-
monobenzoylated 18. O-Benzyldecyl ether did not debenzylate with 3 equiv
of NIS, and control of NIS (2 equiv) gave more 24 from 23.
Usefully, selectivity over and hence compatibility with
O-benzyl ethers10 was possible: tribenzylated 23 was
chemoselectively monodebenzylated on nitrogen in 78%
(entry 35).15 Compatibility with O-benzoyl protection was
also demonstrated: 17 with 3 equiv of NIS gave 77% 18
Org. Lett., Vol. 7, No. 12, 2005
2363

(18 h). Interestingly, L-Phe-derived substrates 17, 20, and
23, which differ only in O-protection, showed ordered
reactivities (OBn > OTBS > OBz) perhaps suggesting a
Lewis base type role for O-1 with a neighboring amino
substituent during critical rate-limiting step(s).
TBS, Bz). Treatment of substrates 32-36 did not result in
efficient debenzylation, e.g., 30 successfully debenzylates,
but deoxygenated analogue 36 does not. Neighboring O-
substituents may act as intramolecular Lewis bases/nucleo-
philic catalysts. Cyclic acetal intermediates have been
proposed in the O-debenzylation of sugars10 with neighboring
OH groups; migration products observed here for substrates
with free alcohol substituents (e.g., 3 from 1) may support
a similar mechanism.
In summary, use of NIS allows tuneable N-debenzylation
of benzylamine substrates that also contain O-functionality;
3 equiv of NIS under dry conditions monodebenzylates,
whereas >3 equiv in the presence of adventitious water
didebenzylates.
The ability of NIS to provide tuneable debenzylation opens
intriguing mechanistic possibilities. Monitoring of the reac-
tion of 1 with 3 equiv of NIS for 20 h (entry 9) by TLC16
showed sequential formation of monobenzyl 2 (5-20 min);
2 and monobenzylmonobenzoyl 3 (45 min to 2.5 h); and 2,
3, and monobenzylmonobenzamide 6 (20 h). In many crude
1H NMR spectra, benzaldehyde was detected, suggesting
hydrolytic formation upon workup. The most plausible
mechanism involves benzyliminium formation before hy-
drolysis or intramolecular nucleophilic trapping by neighbor-
ing O-substituents prior to collapse to debenzylated products
Acknowledgment. Oxford University and Daphne Jack-
son Trust for a returner’s fellowship funded by the Royal
Commission of the Exhibition of 1851; R. Stafford, D.
Emmerson, T. Royston for technical support and EPSRC for
access to mass spectrometry and database services.
in the presence of adventitious water or upon workup.17
speculative mechanistic outline is shown in Scheme 5.
A
Supporting Information Available: General experimen-
tal details, characterization data, and spectra. This material
is available free of charge via the Internet at http://pubs.acs.org.
Scheme 5
OL050624F
(16) Products detected by TLC not necessarily present in the reaction
mixture and may be the result of hydrolysis of intermediates.
(17) Interestingly, reaction of 7 with 3 equiv of NIS for 3 h gave a weak,
transient EPR signal showing a nitrogen 1:1:1 triplet signal (g ) 2.0053,
a(N) ) 1.55 mT at 9.414 GHz) consistent with a nitrogen centered radical/
radical cation (a) [Landoh-Bornstein, A.: Fischer, H., Hellwege, I.-H., Eds.;
Springer-Verlag: Berlin, 1985; Vol. 9, Part d. (b) Martin Goez, M.;
Sartorius, I. J. Am. Chem. Soc. 1993, 115, 11123. (c) Shaffer, S. A.; Martin
Sad´ılek, M.; Turecˇek, F. J. Org. Chem. 1996, 61, 5234] and might suggest
alternative pathways. However, identical reaction outcomes were found in
the absence of light, and attempts to trap putative radical intermediates,
e.g., 2.5 equiv of methyl methacrylate, failed. Further mechanistic details
are currently under investigation.
All successful debenzylations involved substrates with
neighboring oxygenated O-substituents -OR (R ) H, Bn,
2364
Org. Lett., Vol. 7, No. 12, 2005