The conversion of pentoses to 3,4-dihydroxyprolines

Article abstract of DOI:10.1016/j.tet.2005.07.072

The synthesis of two naturally-occurring isomers of 3,4-dihydroxyproline is reported. L-2,3-cis-3,4-trans-3,4-Dihydroxyproline was synthesized from L-arabinose in 10 steps and 31% overall yield. The same series of reactions was employed to convert L-xylos

Full text of DOI:10.1016/j.tet.2005.07.072

Tetrahedron 61 (2005) 9611–9617
The conversion of pentoses to 3,4-dihydroxyprolines
*
Carol M. Taylor, Chantelle E. Jones and Kristin Bopp
Institute of Fundamental Sciences, Massey University, Private Bag 11-222, Palmerston North 5301, New Zealand
Received 1 June 2005; revised 11 July 2005; accepted 21 July 2005
Available online 10 August 2005
Abstract—The synthesis of two naturally-occurring isomers of 3,4-dihydroxyproline is reported. L-2,3-cis-3,4-trans-3,4-Dihydroxyproline
was synthesized from L-arabinose in 10 steps and 31% overall yield. The same series of reactions was employed to convert L-xylose to L-2,3-
trans-3,4-trans-3,4-dihydroxyproline. Orthogonally protected versions of these amino acids were produced on gram scale, en route to the
free amino acids, and these will serve as versatile intermediates in peptide synthesis. This synthetic strategy involved Na-Fmoc protection
and protection of the C3 and C4 secondary alcohols as methoxyethoxymethyl (MEM) ethers.
q 2005 Elsevier Ltd. All rights reserved.
1. Introduction
delivering any of the eight stereoisomers. Most other
approaches are limited in this regard to the preparation of
only a subset of these target molecules. For example, syn-
dihydroxylation of a 3,4-dehydroproline can lead only to
DHPs with 3,4-cis relative stereochemistry.9,10 Conversely,
the opening of an epoxide has been a useful tool in the
synthesis of DHPs with a 3,4-trans relative stereo-
chemistry.11,12
3,4-Dihydroxyproline (DHP)1 contains three stereogenic
centers: C2, C3 and C4 and so there are eight possible
stereoisomers. Three members of the L-series have been
isolated from natural sources (Fig. 1). The L-2,3-cis-3,4-
trans isomer (1) was isolated from the cell wall of the diatom
Navicula pelliculosa more than 30 years ago.2 In 1980, the
L-2,3-trans-3,4-trans isomer (2) was isolated from the acid
hydrolysates of the toxic mushroom Amanita virosa3 and
identified in the virotoxin cyclic heptapeptides.4 In 1994, the
L-2,3-trans-3,4-cis isomer 3 was identified as the sixth
residue in the repeating decapeptide sequence of Mepf1, an
adhesive protein produced by the marine mussel Mytilus
edulis.5
Our goal over the past several years has been to develop a
synthesis of 3,4-dihydroxyprolines, which should be
amenable to producing useful quantities of any stereo-
isomer, in an efficient and stereochemically predictable
manner. We reasoned that the eight stereoisomers of 3,4-
dihydroxyproline ought to be accessible from the eight
pentose sugars, utilizing Fleet’s double displacement
chemistry, which we had adopted previously to good
effect.13 The configuration at C3 and C4 would be derived
directly from the sugar and that at C2 inverted during the
sequence. The overall retrosynthetic analysis is embodied in
Table 1.
Figure 1. Naturally occurring 3,4-dihydroxyprolines.
There are many syntheses of dihydroxyprolines in the
literature.6 When we began our bid to synthesize the Mefp1
decapeptide,7 we utilized the approach of Fleet and co-
workers8 to prepare a suitably protected derivative of 3 from
D-gulonolactone. Our long term goal, however, was to
investigate the role of 3,4-dihydroxyprolines in nature and
as such we required a synthesis, which was capable of
2. Results and discussion
In our previous endeavors, we utilized tert-butyldimethyl-
silyl (TBDMS) ethers for protection of the secondary
alcohols at C3 and C4 (DHP numbering).14 Thus, we
demonstrated the proof of concept by converting
D-ribonolactone (4) to amino acid building block 7.14a We
later reported full experimental details for the application of
this chemistry to the synthesis of two other stereoisomers of
compound 7.14b We felt compelled to publish this work,
given the amount of effort that had been expended.
Keywords: Dihydroxyprolines; Pentose sugars.
*
Corresponding author. Tel.: C64 63569099; fax: C64 63505682;
e-mail: c.m.taylor@massey.ac.nz
0040–4020/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2005.07.072

9612
C. M. Taylor et al. / Tetrahedron 61 (2005) 9611–9617
Table 1. Retrosynthetic scheme of 3,4-dihydroxyprolines
another, requiring close monitoring of the reaction progress
and fine-tuning of reaction times to optimize the yield of
desired products.
We proposed that substitution of the MEM
(methoxyethoxymethyl) group for the TBDMS group
might improve the viability of the synthesis. Indeed, the
use of the MEM protecting group has been described
previously for trans-4-hydroxyproline building blocks
during the synthesis of collagen mimetics.15
DHP isomer
Pentose precursor
L-2,3-cis-3,4-cis-
L-ribose
L-2,3-cis-3,4-trans-
L-2,3-trans-3,4-cis-
L-2,3-trans-3,4-trans-
D-2,3-cis-3,4-cis-
L-arabinose
L-lyxose
L-xylose
D-ribose
Herein, we report the synthesis of two isomers of 3,4-
dihydroxyproline in this manner, on reasonable scale and
with greatly improved overall yields. While our original aim
was to produce building blocks for peptide synthesis, we are
becoming increasingly aware of the need for the free amino
acids as authentic samples for comparison in amino acid
analyses where the occurrence of 3,4-dihydroxyprolines is
suspected in peptides and proteins.16 Thus, we also describe
conditions for producing the free amino acids.
D-2,3-cis-3,4-trans-
D-2,3-trans-3,4-cis-
D-2,3-trans-3,4-trans-
D-arabinose
D-lyxose
D-xylose
Unfortunately, the overall efficiency of the synthesis was
disappointing and our goal of producing useful quantities of
DHP building blocks for peptide synthesis was not realized.
An objective assessment of our previous work (Scheme 1)
led us to the conclusion that the problems encountered in the
syntheses were directly, or indirectly, associated with the
TBDMS protecting group. For example, the reductive
opening of the lactone ring in compound 8 was anticipated
to give 9 (by analogy to the reduction of two diastereo-
isomers of 8). Unfortunately, the major product 10 arose
from migration of the silyl group to the primary alcohol.14b
Another problem in the synthesis involved the chemoselec-
tive hydrolysis of the trityl ether from compound 5; partial
removal of one TBDMS group was also observed. The
extent of this problem varied from one stereoisomer to
The conversion of compound 11 (derived from L-arabinose)
to 2,3-cis-3,4-trans-3,4-dihydroxy-L-proline (1) is summar-
ized in Scheme 2. In our 2002 paper,14b we described some
of the trials and tribulations associated with producing the
5-O-triphenylmethyl ethers typified by 11. Some further
observations and recommendations are given in the Section
4. Standard conditions for the protection of alcohols as
MEM ethers (MEMCl, iPr2NEt, CH2Cl2, rt)17 were
ineffective for the conversion of 11 to 12. Optimized
conditions involved heating at reflux in chloroform with
5 equiv of alkylating agent. A drawback of the MEM
protecting group, which is perhaps not widely appreciated,
1
is the complexity that it introduces into H NMR spectra.
Each MEM group introduces a singlet at w3 ppm; in
addition there are three –CH2– groups, which give rise to six
signals, since the two protons of each –CH2– unit are
diastereotopic. 1H NMR spectra of intermediates in
Schemes 2 and 3 were thus difficult to fully assign.
However, 13C NMR spectra were less complex and
provided evidence for the identity and purity of compounds.
Reduction of the fully protected lactone 12 gave diol 13,
which was readily converted to the bis-mesylate 14.
Cyclization to form the pyrrolidine 15 gave an essentially
quantitative yield if the benzylamine was distilled
immediately prior to the reaction. Benzylamine, which
had been distilled in recent weeks and stored over KOH
Scheme 1. Important issues from previous work.
Scheme 2. Synthesis of L-2,3-cis-3,4-trans-3,4-dihydroxyproline.

C. M. Taylor et al. / Tetrahedron 61 (2005) 9611–9617
9613
Scheme 3. Synthesis of L-2,3-trans-3,4-trans-3,4-dihydroxyproline.
gave yields in the order of 60–70%. Due to the MEM ethers,
pyrrolidine 15 is much more polar than analogous
compounds with other protecting groups at C3 and C4
(acetonide,13 TBDMS14). This facilitated separation of the
pyrrolidine from benzylamine-derived byproducts, includ-
ing benzaldehyde. The benzyl group was removed
hydrogenolytically from pyrrolidine 15 and the amine
protected as its Fmoc derivative. The trityl ether was
removed cleanly in 2 h. The differential in acid lability of
the two ethers (trityl vs. MEM) removed all problems
associated with the chemoselectivity in this step (vide supra,
conversion of 5/6, Scheme 1).
For clarity, the reaction sequence, as applied to the
conversion of L-xylose to amino acid building block 26, is
depicted in Scheme 2. The intermediates behaved similarly
and yields were comparable.
3. Conclusion
In summary, two isomers of 3,4-dihydroxyproline have
been synthesized, via orthogonally protected derivatives.
Building blocks 18 and 26 ought to prove useful in peptide
synthesis. We believe that the synthetic route should be
applicable to the conversion of the appropriate pentose
sugar to any of the eight stereoisomers of 3,4-dihydroxy-
proline. This route has been arrived at via the exploration of
several protecting groups and reaction conditions for each
step. While the bromine oxidation of the pentose sugars,
followed by 5-O-trityl ether formation remains a somewhat
capricious undertaking, the subsequent steps are highly
reproducible. The overall yield for the production of 1 from
11 is 44% (eight steps); this reduces to 31% starting from
L-arabinose (10 steps). Likewise, isomer 2 was produced
from lactone 19 in 36% yield (eight steps); or 22% yield (ten
steps) from L-xylose.
As in our previous work, we intially utilized a two step
oxidation of 17: Swern oxidation to the aldehyde, followed
by sodium chlorite oxidation to acid 18.14b The NMR
spectra of compound 18 featured very broad peaks, which
were doubled up in the 13C NMR. We attributed this to
restricted rotation about the carbamate C–N bond. Unfortu-
nately, removal of the protecting groups gave a 3:1 mixture
(determined by integration of 1H NMR signals) of two
dihydroxyprolines, which appeared to be diastereoisomers.
HPLC analysis of compound 18 (from Swern/NaClO2
oxidation of 17) revealed the same ratio of two similar
compounds. Our suspicion was that the a-amino aldehyde
derived via Swern oxidation of 17 was undergoing
epimerization. This was not observed for the analogous
compound bearing TBDMS ethers.14b Moreover, others
have reported the successful Swern oxidation of similar
prolinol compounds.18
4. Experimental
4.1. General details
All reactions were conducted under a dry nitrogen
atmosphere unless otherwise noted. Reagents were obtained
from commercial suppliers and used directly with the
following exceptions. Tetrahydrofuran was dried over
sodium/benzophenone and distilled prior to use. Diiso-
propylethylamine, triethylamine and pyridine were dried
and distilled from CaH2 and stored over KOH pellets.
Acetonitrile and benzylamine were freshly distilled from
CaH2. Methanesulfonyl chloride was best distilled from
P2O5 immediately prior to use. Flash chromatography was
performed using Scharlau 60 silica gel (230–400 mesh) with
the indicated solvents. Thin-layer chromatography (TLC)
was carried out on precoated silica plates (Merck Kieselgel
60F254) and compounds were visualized by UV fluorescence
Ruthenium (III) oxidation of 17 under Sharpless con-
ditions19 gave acid 18 directly; deprotection yielded a
stereoisomerically pure sample of 1. Our hunch was
confirmed, but as we had found earlier,14a the yield of this
oxidation (36%) was unsatisfactory. A number of recent
reports suggested that TEMPO may be the best reagent for
this oxidation.20 Indeed, oxidation of 17 to 18 was achieved
in good yield, with no loss of stereochemical integrity.
The deprotection of the amino acid was attempted in a
number of ways. Reactions employing TFA/dichloro-
methane, 95% TFA/H2O or 1 N HCl in TFA were slow
and not clean. The best results were obtained using a
solution of HBr in acetic acid. This acid treatment removed
the MEM protecting groups. The Fmoc group was removed
using Tesser’s base21 and the free amino acid was purified
by ion exchange chromatography.
1
or by staining with anisaldehyde or ninhydrin. H and 13C
NMR spectra were obtained using either a JEOL JNM-
GX270W or a Bruker Avance 400 spectrometer. Chemical
shifts for spectra in CDCl3 are given in parts per million

9614
C. M. Taylor et al. / Tetrahedron 61 (2005) 9611–9617
1
(ppm) downfield from tetramethylsilane as internal standard
(1H) or relative to residual solvent (13C). Spectra of the free
amino acids in D2O were referenced to DSS as an external
standard. High resolution mass spectra were recorded using
a VG7070 mass spectrometer operating at nominal
accelerating voltage of 70 eV.
EtOAc/hexane); H NMR (CDCl3, 400 MHz) d 3.26 (dd,
JZ10.9, 4.6 Hz, 1H), 3.29 (s, 3H), 3.34 (t, JZ4.6 Hz, 2H),
3.38 (s, 3H), 3.39–3.61 (m, 5H), 3.71–3.87 (m, 2H), 4.35–
4.38 (m, 1H), 4.47 (t, JZ7.1 Hz, 1H), 4.58 (d, JZ7.3 Hz,
1H), 4.64 (d, JZ6.9 Hz, 1H), 4.73 (d, JZ5.9 Hz, 1H), 4.89
(d, JZ6.9 Hz, 1H), 5.12 (d, JZ6.9 Hz, 1H), 7.22–7.32 (m,
9H), 7.43–7.46 (m, 6H); 13C NMR (CDCl3, 100 MHz) d
58.9, 59.0, 62.0, 67.4, 67.7, 71.4, 71.6, 76.4, 77.6, 79.8,
86.8, 95.0, 95.2, 127.1, 127.9, 128.6, 143.3, 172.1; HRMS
(EIC) calcd for C32H38O9 (MC): 566.25158; obsd:
566.25163.
4.1.1. 5-O-Trityl-L-arabinono-g-lactone (11). Potassium
carbonate (2.26 g, 16.3 mmol, 1.22 equiv) was added in
portions, over 2 h, to a solution of L-arabinose (2.00 g,
13.3 mmol, 1.00 equiv) in Milli-Q water (6 mL) at 0 8C.
Bromine (0.8 mL, 2.48 g, 15.5 mmol, 1.16 equiv) was then
added dropwise, from a dropping funnel (constructed of
glass and teflon only) over 2 h. The solution was warmed to
rt and left to stir overnight. The solution was still yellow/
orange (if it was a not, a couple more drops of bromine were
added and the mixture left to stir another 12 h) and the
excess bromine was quenched by the addition of neat formic
acid (two drops from a Pasteur pipette). Decolorization was
not immediate, but was complete within 10 min. The
mixture was concentrated on a rotary evaporator, with the
water bath at 60 8C. When the volume of the mixture
reached w10 mL, glacial acetic acid (1 mL) was added.
Rotary evaporation at 60 8C was continued for at least 2 h
and then under high vacuum. The dry residue was
suspended in pyridine (25 mL) under N2. DMAP (325 mg,
2.7 mmol, 0.2 equiv) was added, followed by trityl chloride
(4.45 g, 16.0 mmol, 1.2 equiv) and the mixture heated at
reflux for 12 h. The brown solution was cooled, diluted with
CH2Cl2 (300 mL) and washed successively with water
(250 mL), 1 M HCl (250 mL), satd aq NaHCO3 (250 mL)
and brine (250 mL). The organic layer was filtered through
MgSO4 and concentrated. The residue was purified by flash
chromatography, eluting with 1:1 EtOAc/hexane to give 11
as a colorless foam (3.71 g, 71%). Data reported
elsewhere.14b
4.1.4. 2,3-Di-O-(methoxyethoxymethyl)-5-O-trityl-L-
xylono-g-lactone (20). By analogy to the procedure in
Section 4.1.3, on a scale of 10.1 mmol of lactone 19,
affording 20 as a colorless oil (4.70 g, 82%). Rf 0.46 (2:1
EtOAc/hexane); 1H NMR (CDCl3, 400 MHz) d 3.29 (s, 3H),
3.35 (s, 3H), 3.26–3.44 (m, 8H), 3.45–3.82 (m, 2H), 4.44
(app. t, JZ7.1 Hz, 1H), 4.59–4.70 (m, 3H), 4.87 (d, JZ
6.6 Hz, 1H), 4.90 (d, JZ7.3 Hz, 1H), 5.04 (d, JZ6.8 Hz,
1H), 7.20–7.32 (m, 9H), 7.42–7.44 (m, 6H); 13C NMR
(CDCl3, 67.5 MHz) d 58.9 (2C), 61.0, 67.5, 67.6, 71.4, 71.5,
74.6, 77.9, 78.5, 87.5, 95.0, 95.9, 127.0, 127.7, 128.6, 143.1,
172.5; HRMS (FABC, NBA, CH2Cl2) calcd for C32H39O9
(MHC): 567.259408; obsd: 567.259347.
4.1.5. 2,3-Di-O-(methoxyethoxymethyl)-5-O-trityl-L-ara-
binitol (13). A solution of LiBH4 (2 M in THF, 6.87 mL,
13.7 mmol, 2.0 equiv) was added dropwise over 1 h to a
solution of lactone 12 (3.89 g, 2.20 mmol, 1.0 equiv) in
THF (40 mL) at rt under N2. The mixture was stirred for
1.5 h after the addition was complete, then quenched by the
cautious, dropwise addition of satd aq NH4Cl (5 mL). The
mixture was stirred 10 min then partitioned between EtOAc
(300 mL) and brine (300 mL). The brine was extracted with
a further portion of EtOAc (300 mL). The organic layers
were filtered through MgSO4 and concentrated. The residue
was purified by flash chromatography, eluting with 95:5
CH2Cl2/MeOH, to give 13 as a colorless oil (3.46 g, 88%).
Rf 0.15 (95:5 CH2Cl2/MeOH); 1H NMR (CDCl3, 270 MHz)
d 3.16 (dd, JZ9.7, 5.1 Hz, 1H), 3.33–3.58 (m, 7H), 3.35 (s,
3H), 3.38 (s, 3H), 3.65–3.77 (m, 4H), 3.86 (dd, JZ7.6,
2.7 Hz, 1H), 3.92–4.00 (m, 2H), 4.43 (d, JZ6.8 Hz, 1H),
4.58 (d, JZ6.8 Hz, 1H), 4.70–4.83 (m, 2H), 7.20–7.32 (m,
9H), 7.42–7.48 (m, 6H); 13C NMR (CDCl3, 67.5 MHz) d
59.0 (2C), 61.5, 64.4, 67.6, 67.7, 69.8, 71.4, 71.6, 78.5, 80.0,
86.6, 96.2, 96.7, 126.9, 127.7, 128.6, 143.6; HRMS (FABC,
NBA) calcd for C32H43O9 (MHC): 571.29163; obsd:
571.29163.
4.1.2. 5-O-Trityl-L-xylono-g-lactone (19)22. By analogy to
the procedure in Section 4.1.1, on a scale of 13.3 mmol,
affording 5-O-trityl-L-xylono-g-lactone (19) as a colorless
1
oil (3.25 g, 63%). Rf 0.19 (1:1 EtOAc/hexane); H NMR
(CDCl3, 400 MHz) d 3.34 (dd, JZ11.1, 2.8 Hz, 1H), 3.38
(d, JZ6.7 Hz, 1H), 3.64 (dd, JZ11.1, 2.8 Hz, 1H), 4.29 (br
s, 1H), 4.50 (q, JZ7.5 Hz, 1H), 4.57 (dt, JZ7.5, 2.8 Hz,
1H), 4.82 (d, JZ7.5 Hz, 1H), 7.18–7.36 (m, 9H), 7.50–7.57
(m, 6H); 13C NMR (CDCl3, 67.5 MHz) d 60.9, 73.5, 74.2,
78.3, 88.0, 127.3, 128.0, 128.4, 142.8, 175.3; HRMS
(FABC, NBA, CH2Cl2) calcd for C24H22O5 (MC):
390.146088; obsd: 390.1461.
4.1.3. 2,3-Di-O-(methoxyethoxymethyl)-5-O-trityl-L-ara-
binono-g-lactone (12). MEMCl (4.12 mL, 4.50 g,
36.1 mmol, 5.0 equiv) was added dropwise to a solution
of 5-O-trityl-L-arabinono-g-lactone (11) (2.82 g, 7.2 mmol,
1.0 equiv) and diisopropylethylamine (6.29 mL, 4.67 g,
36.1 mmol, 5.0 equiv) in AR-grade chloroform (45 mL) at
rt under N2. The mixture was heated at reflux for 11 h,
cooled, diluted with chloroform (200 mL), washed with
10% aq citric acid (200 mL), satd aq NaHCO3 (200 mL) and
brine (200 mL). The organic layer was filtered through
MgSO4 and concentrated. The orange residue was purified
by flash chromatography, eluting with 2:1 hexanes/EtOAc
to give 12 as a colorless oil (3.89 g, 95%). Rf 0.31 (1:1
4.1.6. 2,3-Di-O-(methoxyethoxymethyl)-5-O-trityl-L-
xylitol (21). By analogy to the procedure in Section 4.1.5,
on a scale of 8.29 mmol of lactone 20, affording 21 as a
colorless oil (4.20 g, 89%). Rf 0.22 (95:5 CH2Cl2/MeOH);
1H NMR (CDCl3, 270 MHz) d 1.80 (br s, 2H), 3.14 (dd, JZ
9.2, 6.2 Hz, 1H), 3.25 (dd, JZ9.2, 6.0 Hz, 1H), 3.32 (s, 3H),
3.35 (s, 3H), 3.42–3.89 (m, 13H), 4.05 (td, JZ5.9, 2.6 Hz,
1H), 4.61–4.67 (m, 2H), 4.77 (d, JZ7.3 Hz, 1H), 7.17–7.30
(m, 9H), 7.40–7.43 (m, 6H); 13C NMR (CDCl3, 67.5 MHz)
d 58.9, 59.0, 61.0, 64.4, 67.6, 68.9, 71.5, 77.9, 80.9, 86.6,
96.0, 96.9, 126.9, 127.7, 127.9, 143.6; HRMS (FABC,
NBA, CH2Cl2) calcd for C32H43O9 (MHC): 571.2907; obsd:
571.2898.

C. M. Taylor et al. / Tetrahedron 61 (2005) 9611–9617
9615
4.1.7. 1,4-Bis-O-(methanesulfonyl)-2,3-di-O-(methoxy-
ethoxymethyl)-5-O-trityl-L-arabinitol (14). N,N-
Dimethylaminopyridine (75 mg, 0.62 mmol, 0.2 equiv)
was added to neat methanesulfonyl chloride (0.95 mL,
1.41 g, 12.3 mmol, 4.0 equiv) at 0 8C under N2. A solution
of diol 13 (1.76 g, 3.08 mmol, 1.0 equiv) in pyridine (8,
2 mL rinse) was added dropwise over 40 min. The mixture
was gradually warmed to rt and stirred 5 h. The mixture was
concentrated and the residue partitioned between chloro-
form (150 mL) and water (150 mL). The aqueous layer was
extracted further with chloroform (2!80 mL). The organic
extracts were filtered through MgSO4 and concentrated. The
residue was purified by flash chromatography, eluting with
2:1 EtOAc/hexanes, to give 14 as a slightly yellow oil
(2.01 g, 90%). Rf 0.26 (2:1 EtOAc/hexanes); 1H NMR
(CDCl3, 400 MHz) d 3.02 (s, 3H), 3.11 (s, 2H), 3.34 (s, 3H),
3.36 (s, 3H), 3.36–3.74 (m, 10H), 3.98 (app. q, JZ5.0 Hz,
1H), 4.05 (t, JZ4.3 Hz, 1H), 4.28 (dd, JZ10.5, 5.8 Hz, 1H),
4.34 (dd, JZ10.5, 5.2 Hz, 1H), 4.61 (d, JZ7.0 Hz, 1H),
4.67–4.75 (m, 3H), 5.00–5.03 (m, 1H), 7.23–7.27 (m, 3H),
7.29–7.34 (m, 6H), 7.41–7.44 (m, 6H); 13C NMR (CDCl3,
100 MHz) d 37.2, 28.8, 58.9, 62.6, 67.6, 68.0 (2C), 71.5
(2C), 75.4, 76.0, 80.9, 87.4, 96.4, 96.7, 127.3, 128.0, 128.6,
143.1; HRMS (FABC, NBA) calcd for C35H47O13S2
(MHC): 727.24581; obsd: 727.24494.
HRMS (FABC, NBA) calcd for C39H48NO7 (MHC):
642.34294; obsd: 642.34294.
4.1.10. (2R,3R,4R)-1-Benzyl-3,4-di-O-(methoxyethoxy-
methyl)-2-triphenylmethoxymethyl-pyrrolidine (23). By
analogy to the procedure in 4.1.9, on a scale of 5.47 mmol of
bis-mesylate 22, to give pyrrolidine 23 as a yellow oil
(3.23 g, 93%). Rf 0.32 (1:1 EtOAc/hexanes); 1H NMR
(CDCl3, 400 MHz) d 2.55 (dd, JZ10.8, 4.8 Hz, 1H), 2.80
(dd, JZ10.0, 5.7 Hz, 1H), 2.94 (d, JZ10.8 Hz, 1H), 3.23
(ddd, JZ16.0, 10.0, 5.7 Hz, 2H), 3.31 (s, 3H), 3.32 (s, 3H),
3.36–3.46 (m, 2H), 3.54–3.61 (m, 2H), 4.03 (m, 1H), 4.11
(d, JZ9.4 Hz, 1H), 4.58 (d, JZ7.2 Hz, 1H), 4.65 (d, JZ
7.2 Hz, 1H), 4.76 (d, JZ2.8 Hz, 2H), 7.12–7.28 (m, 14H),
7.43–7.45 (m, 6H); 13C NMR (CDCl3, 100 MHz) d 57.7,
58.8, 58.9, 59.5, 64.8, 66.8, 67.0, 69.5, 71.5, 71.6, 79.4,
83.4, 86.7, 94.0, 94.3, 126.8, 127.0, 127.7, 128.1, 128.7,
128.8, 138.8, 144.0; HRMS (FABC, NBA) calcd for
C39H48NO7 (MHC): 642.34294; obsd: 642.34294.
4.1.11. (2R,3S,4S)-1-Fluorenylmethoxycarbonyl-3,4-di-
O-(methoxyethoxymethyl)-2-triphenylmethoxy-methyl-
pyrrolidine (16). Pd/C (10%, 600 mg) was added to a
solution of the pyrrolidine 15 (2.43 g, 3.79 mmol) in
absolute ethanol (30 mL). The flask was evacuated and
then opened up to an atmosphere of H2 and stirred for 16 h.
The mixture was filtered through a pad of Celite, washing
well with ethanol. The filtrate was concentrated and
evaporated down from toluene. A solution of the pyrrolidine
in toluene (15, 4 mL rinse) was added dropwise to a solution
of fluorenylmethyl chloroformate (1.08 g, 4.17 mmol,
1.1 equiv) in toluene (8 mL) at 0 8C. Triethylamine
(380 mL, 276 mg, 2.73 mmol, 1.1 equiv) was added over
10 min, the mixture warmed to rt and stirred for 2 h. The
suspension was filtered through a sintered glass funnel,
washing well with toluene. The filtrate was concentrated
and the residue purified by flash chromatography, eluting
with 1:1 EtOAc/hexanes to give 16 (2.56 g, 87%). Rf 0.27
(1:1 EtOAc/hexanes); 1H NMR (CDCl3, 400 MHz) d
3.22–3.55 (m, 1H), 3.37–3.48 (m, 4H), 3.52–3.62 (m, 4H),
3.71–3.82 (m, 3H), 3.95–4.38 (m, 5H), 4.53–4.86 (m, 5H),
7.17–7.30 (m, 11H), 7.35–7.48 (m, 9H), 7.59 (t, JZ7.6 Hz,
1H), 7.71–7.77 (m, 2H); 13C NMR (CDCl3, 100 MHz) d
47.0 and 47.1, 49.8 and 49.9, 57.9 and 58.2, 59.0 (2C), 60.0
and 60.3, 67.3, 71.5, 71.6, 77.2, 78.5 and 79.0, 80.1 and
80.7, 87.1 and 87.2, 95.3 and 95.5, 96.1, 119.9, 124.8, 124.9,
125.0, 126.9, 127.0, 127.5, 127.6, 127.7, 128.6, 128.7,
141.2, 143.8, 143.9, 154.8 and 154.9; HRMS (FABC, NBA)
calcd for C47H52NO9 (MHC): 774.364208; obsd:
774.362996.
4.1.8. 1,4-Bis-O-(methanesulfonyl)-2,3-di-O-(methoxy-
ethoxymethyl)-5-O-trityl-L-xylitol (22). By analogy to
the procedure in Section 4.1.7, on a scale of 5.14 mmol of
diol 21, affording 22 (2.776 g, 74%). Rf 0.19 (2:1 EtOAc/
hexanes); 1H NMR (CDCl3, 270 MHz) d 2.99 (s, 3H), 3.06
(s, 3H), 3.32 (s, 3H), 3.34 (s, 3H), 3.42–3.75 (m, 7H), 3.65
(dd, JZ10.0, 4.7 Hz, 1H), 3.68 (app. t, JZ4.7 Hz, 1H),
4.37–4.42 (m, 3H), 4.54–4.60 (m, 3H), 4.54–4.60 (m, 3H),
4.75–4.81 (m, 3H), 5.02–5.03 (m, 1H), 7.22–7.35 (m, 9H),
7.41–7.45 (m, 6H); 13C NMR (CDCl3, 100 MHz) d 37.1,
38.8, 58.8, 58.9, 62.8, 67.7, 68.2, 68.4, 71.4, 71.5, 75.3,
75.9, 80.7, 87.3, 96.3, 97.5, 127.3, 128.0, 128.5, 143.0;
HRMS (FABC, NBA) calcd for C35H47O13S2 (MHC):
727.24581; obsd: 727.243268.
4.1.9. (2R,3S,4S)-1-Benzyl-3,4-di-O-(methoxyethoxy-
methyl)-2-triphenylmethoxymethyl-pyrrolidine (15). A
solution of bis-mesylate 14 (1.80 g, 2.48 mmol) in freshly
distilled benzylamine (10 mL) was stirred at 90 8C under N2
for 4 days. The mixture was cooled and partitioned between
chloroform (50 mL) and brine (50 mL). The aqueous layer
was extracted further with chloroform (2!50 mL). The
combined organic extracts were washed with water
(100 mL), filtered through MgSO4 and concentrated. The
residue was applied directly to a flash column, eluting with
5:1 hexanes/EtOAc until the benzaldehyde had eluted (Rf
0.63, 1:1 EtOAc/hexanes). The eluant was changed to
1.5:1.0 EtOAc/hexanes to elute pyrrolidine 15 as a yellow
4.1.12. (2R,3R,4R-1-Fluorenylmethoxycarbonyl-3,4-
di-O-(methoxyethoxymethyl)-2-triphenyl-methoxy-
methyl)-pyrrolidine (24). By analogy to the procedure in
Section 4.1.11, on a scale of 3.41 mmol of pyrrolidine 23, to
give compound 24 (2.247 g, 85%). Rf 0.28 (1:1 EtOAc/
hexanes); 1H NMR (CDCl3, 400 MHz) d 3.15 (t, JZ8.7 Hz,
1H), 3.27 (t, JZ8.7 Hz, 1H), 3.36 (s, 3H), 3.37 and 3.38 (2s,
3H), 3.42–3.62 (m, 7H), 3.70–3.75 (m, 2H), 3.84 (ddd, JZ
23.0, 11.8, 5.5 Hz, 1H), 4.02–4.33 (m, 5H), 4.45–4.65 (m,
3H), 4.79–4.85 (m, 2H), 7.14–7.29 (m, 11H), 7.36–7.45 (m,
8H), 7.47 (d, JZ7.6 Hz, 1H), 7.57 (d, JZ7.6 Hz, 1H), 7.75
(d, JZ7.6 Hz, 1H); 13C NMR (CDCl3, 100 MHz) d 47.1 and
1
oil (1.590 g, 99%). Rf 0.28 (1:1 EtOAc/hexanes); H NMR
(CDCl3, 270 MHz) d 2.31 (dd, JZ10.8, 4.4 Hz, 1H), 3.10–
3.68 (m, 13H), 3.34 (s, 3H), 3.32 (s, 3H), 3.98 (d, JZ
13.4 Hz, 1H), 4.06–4.10 (m, 1H), 4.26 (dd, JZ5.0, 1.9 Hz,
1H), 4.59 (d, JZ7.2 Hz, 1H), 4.68–4.71 (m, 2H), 4.77 (d,
JZ7.2 Hz, 1H), 7.15–7.30 (m, 14H), 7.41–7.47 (m, 6H);
13C NMR (CDCl3, 67.5 MHz) d 57.7, 58.9 (2C), 59.5, 62.2,
65.5, 66.9, 67.0, 71.5, 71.6, 79.7, 81.5, 86.9, 94.4, 95.1,
126.6, 126.8, 127.6, 127.9, 128.5, 128.7, 138.6, 143.9;

9616
C. M. Taylor et al. / Tetrahedron 61 (2005) 9611–9617
47.2, 51.1 and 51.3, 59.0 (2C), 61.3 and 61.8, 62.8 and 63.1,
67.1, 67.2, 67.5, 71.5, 71.6, 77.2, 78.2 and 79.1, 79.6 and
80.8, 86.6 and 86.7, 94.4, 94.5, 119.9, 124.9, 125.0, 125.2,
127.0, 127.7, 128.6, 141.2, 143.6, 143.8, 144.0, 144.2,
154.8; HRMS (FABC, NBA) calcd for C47H52NO9 (MHC):
774.364208; obsd: 774.362996.
CH2Cl2/MeOH); [a]2D0 K23.9 (c 1.00, CHCl3); nmax/cmK1
(CHCl3) 3500–2384 (O–H), 1708 (C]O), 1122 (C–O–C);
1H NMR (CDCl3, 400 MHz) d 3.34–3.37 (m, 6H), 3.50–
3.81 (m, 10H), 4.14–4.82 (m, 10H), 7.24–7.38 (m, 4H),
7.52–7.57 (m, 2H), 7.67–7.73 (m, 2H); 13C NMR (CDCl3,
100 MHz) d 47.0, 50.0 and 50.8, 58.8, 59.0, 67.2, 67.6 and
67.8, 71.4 and 71.5, 77.2 and 77.8, 79.4 and 80.4, 94.7, 95.3,
119.8, 125.0, 125.1, 126.9, 127.5, 141.1, 143.5, 143.7,
143.9, 154.7 and 155.3, 172.9; HRMS (EIC) calcd for
C28H36NO10 (MHC): 546.23392; obsd: 546.23345.
4.1.13. (2R,3S,4S)-1-Fluorenylmethoxycarbonyl-3,4-di-
O-(methoxyethoxymethyl)-2-hydroxymethyl-pyrroli-
dine (17). A mixture of formic acid (6.5 mL) and
acetonitrile (45 mL) was added to the pyrrolidine 16
(1.637 g, 1.90 mmol) and stirred at rt under N2 for 2.5 h.
The mixture was partitioned between EtOAc (200 mL) and
satd aq NaHCO3 (200 mL). The organic layer was washed
further with brine (200 mL), filtered through MgSO4, and
concentrated. The residue was purified by flash chromato-
graphy, eluting with 2% MeOH in EtOAc to give compound
17 (1.049 g, 93%). Rf 0.18 (100% EtOAc); 1H NMR
(CDCl3, 400 MHz) d 3.38 (s, 3H), 3.40 (s, 3H), 3.54–3.57
(m, 5H), 3.64–3.91 (m, 7H), 4.07–4.10 (m, 1H), 4.19–4.29
(m, 3H), 4.38–4.56 (m, 2H), 4.73–4.85 (m, 4H), 7.32 (t, JZ
7.5 Hz, 2H), 7.41 (t, JZ7.5 Hz, 2H), 7.59 (d, JZ7.4 Hz,
2H), 7.77 (d, JZ7.5 Hz, 2H); 13C NMR (CDCl3, 67.5 MHz)
d 47.2, 50.7, 59.0, 62.1 and 62.3, 67.3, 67.5, 67.7, 71.6, 77.3,
80.6, 94.6, 95.3, 119.9, 124.9, 126.9, 127.6, 141.2, 143.7,
156.3; HRMS (FABC, NBA) calcd for C28H38NO9 (MHC):
532.25466; obsd: 532.25662.
4.1.16. (2R,3R,4R)-1-Fluorenylmethoxycarbonyl-3,4-di-
O-(methoxyethoxymethoxy)-L-proline (26). By analogy
to the procedure in Section 4.1.15 on a scale of 0.63 mmol
of alcohol 25, to afford acid 26 (302 mg, 88%). Rf 0.22 (9:1
CH2Cl2/MeOH); [a]2D0 K13.6 (c 1.04, CHCl3); nmax/cmK1
(CHCl3), 3502–2600 (O–H), 1703 (C]O), 1160 (O–C–O);
1H NMR (CDCl3, 400 MHz) d 3.31 (s, 6H), 3.30–3.80 (m,
10H), 4.12–4.18 (m, 10H), 7.19–7.66 (m, 8H); 13C NMR
(CDCl3, 100 MHz) d 46.9 and 47.0, 50.8, 58.8, 64.3 and
64.9, 67.2, 67.8 and 68.0, 71.4, 71.5, 77.3 and 78.1, 71.5 and
83.0, 94.4, 94.7, 119.8, 125.1, 127.0, 127.6, 141.1, 143.6,
143.7, 144.0, 154.8 and 156.0, 172.9; HRMS (EIC) calcd
for C28H36NO10 (MHC): 546.23392; obsd: 546.23412.
4.1.17. 2,3-cis-3,4-trans-3,4-Dihydroxy-L-proline (1). A
solution of acid 18 (96 mg, 0.18 mmol) in a solution of HBr
in glacial acetic acid (33 wt%; 3 mL) was stirred for 18 h at
rt. The mixture was concentrated and then dissolved in
Tesser’s base.21 The pH was adjusted to about 10 by the
addition of a few drops of 2 M aq NaOH. The mixture was
stirred for 3 h and then concentrated. The residue was
partitioned between water (5 mL) and EtOAc (5 mL). The
aqueous layer was extracted further with EtOAc (4!5 mL).
The aqueous layer was added to the top of a short column of
Dowex HC resin and eluted with two column volumes of
water. The eluant was changed to 0.5 M NH4OH and the
fractions checked by TLC, staining with ninhydrin.
Relevant fractions were lyophilized to give a colorless
solid, which was dissolved in water, filtered through a 0.2 m
nylon filter and lyophilized again to give compound 1
(25 mg, quant.). Rf 0.26 (3:3:3:1 nBuOH, EtOH, NH3, H2O);
[a]2D0 K48.1 (c 1.00, H2O) lit.2a [a]D20 K61.2 (c 0.5, H2O)
lit.11 [a]2D7 K56 (c 0.62, H2O) lit.23 [a]2D7 K63.2 (c 0.5,
H2O) lit.24 [a]D27 K63.0 (c 0.8, H2O); 1H NMR (D2O,
400 MHz) d 3.16 (d, JZ12.8 Hz, 1H), 4.56 (dd, JZ12.8,
3.7 Hz, 1H), 4.21 (d, JZ4.0 Hz, 1H), 4.27 (d, JZ3.7 Hz,
1H), 4.32 (d, JZ4.0 Hz, 1H); 13C NMR (D2O, 100 MHz) d
50.9, 65.2, 75.0, 75.4, 171.0; HRMS (FABC, glycerol)
calcd for C5H10NO4 (MHC): 148.06098; obsd: 148.06063.
4.1.14. (2R,3R,4R)-1-Fluorenylmethoxycarbonyl-3,4-di-
O-(methoxyethoxymethyl)-2-hydroxymethyl-pyrroli-
dine (25). By analogy to the procedure in Section 4.1.13, on
a scale of 3.26 mmol of pyrrolidine 24, to give compound 25
1
(1.66 g, 96%). Rf 0.19 (100% EtOAc); H NMR (CDCl3,
400 MHz) d 3.38 (s, 6H), 3.49–3.56 (m, 5H), 3.65–3.89 (m,
7H), 3.97 (d, JZ5.8 Hz, 1H), 4.17 (br s, 2H), 4.23 (t, JZ
7.0 Hz, 1H), 4.32–4.50 (m, 2H), 4.75–4.87 (m, 4H), 7.31
(td, JZ7.4, 0.9 Hz, 2H), 7.39 (t, JZ7.4 Hz, 2H), 7.59 (dd,
JZ6.9, 4.2 Hz, 2H), 7.75 (d, JZ7.5 Hz, 2H); 13C NMR
(CDCl3, 100 MHz) d 47.1, 51.0 and 51.4, 58.9, 61.2 and
63.2, 64.6 and 65.6, 66.9 and 67.1, 67.3, 67.5, 71.5, 78.9,
80.4 and 81.1, 94.4, 94.9 and 95.0, 119.8, 124.9, 125.0,
126.9, 127.6, 141.2, 143.7, 143.8, 154.9 and 156.2; HRMS
(FABC, NBA) calcd for C28H38NO9 (MHC): 532.25466;
obsd: 532.25514.
4.1.15. (2R,3S,4S)-1-Fluorenylmethoxycarbonyl-3,4-di-
O-(methoxyethoxymethoxy)-L-proline (18). Sodium
chlorite (78 mg, 0.87 mmol, 2.0 equiv) and TEMPO
(4 mg, cat.) were added to a solution of alcohol 17
(229 mg, 0.43 mmol, 1.0 equiv) in a mixture of acetonitrile
(0.85 mL) and 0.67 M aq NaH2PO4 (0.75 mL). This mixture
was heated to 40 8C and bleach (12 mL) added, resulting in a
deep rose color. The reaction mixture was checked by TLC
and bleach added periodically to maintain the deep rose
color, and until the conversion to the acid was complete.
The mixture was poured onto ice-water (20 mL) containing
Na2SO3 (100 mg). This led to immediate decolorizaton; the
pH was 5–6 and 2 M HCl (w0.5 mL) was added to give a
pH of 2 and considerable precipitation. The mixture was
extracted with EtOAc (3!20 mL). The combined extracts
were filtered through MgSO4 and concentrated. The residue
was purified by flash chromatography, eluting with 10–20%
MeOH in CH2Cl2 to isolate 18 (207 mg, 88%). Rf 0.36 (9:1
4.1.18. 2,3-trans-3,4-trans-3,4-Dihydroxy-L-proline (2).
By analogy to the procedure in Section 4.1.17, starting
with acid 26 (80 mg) and giving rise to 2 (21 mg, quant.). Rf
n
0.39 (3:3:3:1 BuOH, EtOH, NH3, H2O); [a]2D0 K21.4 (c
0.28, H2O) lit.11 [a]D25 K19 (c 0.4, H2O) lit.24 [a]2D2 K12.6
(c 0.53, H2O);22 1H NMR (D2O, 400 MHz) d 3.36 (d, JZ
12.4 Hz, 1H), 3.33 (dd, JZ12.7, 3.7 Hz, 1H), 3.89 (m, 1H),
4.16 (m, 1H), 4.37 (m, 1H); 13C NMR (D2O, 100 MHz) d
50.9, 67.6, 74.1, 78.5, 171.6; HRMS (FABC, glycerol)
calcd for C5H10NO4 (MHC): 148.06098; obsd: 148.06062.

C. M. Taylor et al. / Tetrahedron 61 (2005) 9611–9617
9617
Cenci di Bello, I.; Winchester, B. Tetrahedron 1988, 44,
2649–2655.
Acknowledgements
9. Hudson, C. B.; Robertson, A. B.; Simpson, W. R. J. Aust.
J. Chem. 1968, 21, 769–782.
´
10. Schumacher, K. K.; Jiang, J.; Joullie, M. M. Tetrahedron:
We thank the Marsden Fund, administered by the Royal
Society of New Zealand, for their support of this work under
the auspices of grants 96-UOA-617 and 00-MAU-018.
Asymmetry 1998, 9, 47–53.
11. Kahl, J.-U.; Wieland, T. Liebigs Ann. Chem. 1981, 1445–1450.
¨
12. Herdeis, C.; Aschenbrenner, A.; Kirfel, A.; Schwabenlander,
Supplementary data
F. Tetrahedron: Asymmetry 1997, 8, 2421–2432.
13. Weir, C. A.; Taylor, C. M. J. Org. Chem. 1999, 64,
1554–1558.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.tet.2005.07.
072.13C NMR spectra for compounds 1, 2 and 12–26.
14. (a) Weir, C. A.; Taylor, C. M. Org. Lett. 1999, 1, 787–789. (b)
Taylor, C. M.; Barker, W. D.; Weir, C. A.; Park, J. H. J. Org.
Chem. 2002, 67, 4466–4474.
15. Vadolas, D.; Germann, H. P.; Thakur, S.; Keller, W.;
Hedemann, E. Int. J. Pept. Protein Res. 1985, 25, 554–559.
¨
16. Poulsen, N.; Kroger, N. J. Biol. Chem. 2004, 279,
42993–42999.
References and notes
17. Corey, E. J.; Gras, J.-L.; Ulrich, P. Tetrahedron Lett. 1976, 11,
809–812.
1. (a) Mauger, A. B. J. Nat. Prod. 1996, 59, 1205–1211. (b)
Taylor, S. W.; Waite, J. H. In Prolyl Hydroxylase, Protein
Disulfide Isomerase, and Other Structurally Related Proteins;
Guzman, N. A., Ed.; Marcel Dekker: New York, 1998;
pp 97–107.
18. (a) Mulzer, J.; Becker, R.; Brunner, E. J. Am. Chem. Soc. 1989,
111, 7500–7504. (b) Zhang, R.; Mamai, A.; Madalengoitia,
J. S. J. Org. Chem. 1999, 64, 547–555.
19. Oxidation of prolinols via this protocol: (a) Gryko, D.;
Prokopowicz, P.; Jurczak, J. Tetrahedron: Asymmetry 2002,
13, 1103–1113. (b) Langlois, N.; Rakotondradany, F.
Tetrahedron 2000, 56, 2437–2448.
2. (a) Nakajima, T.; Volcani, B. E. Science 1969, 164,
1400–1401. (b) Karle, I. L.; Daly, J. W.; Witkop, B. Science
1969, 164, 1401–1402. (c) Karle, I. L. Acta Crystallogr., Sect.
B 1970, 26, 765–770.
20. (a) Mamai, A.; Madalengoitia, J. S. Org. Lett. 2001, 3,
561–564. (b) Mamai, A.; Hughes, N. E.; Wurthmann, A.;
Madalengoitia, J. S. J. Org. Chem. 2001, 66, 6483–6486. (c)
Del Valle, J. R.; Goodman, M. J. Org. Chem. 2003, 68,
3923–3931. (d) Raghavan, S.; Reddy, S. R. Tetrahedron Lett.
2003, 44, 7459–7462.
3. Buku, A.; Faulstich, H.; Wieland, T.; Dabrowski, J. Proc. Natl.
Acad. Sci. U.S.A. 1980, 77, 2370–2371.
4. Faulstich, H.; Buku, A.; Bodenmueller, H.; Wieland, T.
Biochemistry 1980, 19, 3334–3343.
5. Taylor, S. W.; Waite, J. H.; Ross, M. M.; Shabanowitz, J.;
Hunt, D. F. J. Am. Chem. Soc. 1994, 116, 10803–10804.
6. For some recent examples, see: (a) Garcia, A. L. L.; Correia,
C. R. D. Tetrahedron Lett. 2003, 44, 1553–1557. (b) Kim,
J. H.; Long, M. J. C.; Kim, J. Y.; Park, K. H. Org. Lett. 2004, 6,
2273–2276. (c) Angle, S. R.; Belanger, D. S. J. Org. Chem.
2004, 69, 4361–4368.
21. A mixture of dioxane:methanol:4 N NaOH in a ratio of 30:9:1;
referred to in: Chang, C.-D.; Waki, M.; Ahmad, M.;
Meinenhofer, J.; Lundell, E. O.; Haug, J. D. Int. J. Pept.
Protein Res. 1980, 15, 59–66.
22. For the oxidation and peracetylation of D-xylose, see: Sun,
R. C.; Okabe, M. Org. Synth. 1995, 72, 48–56.
23. Arakawa, Y.; Yoshifuji, S. Chem. Pharm. Bull. 1991, 39,
2219–2224.
7. Taylor, C. M.; Weir, C. A. J. Org. Chem. 2000, 65,
1414–1421.
24. Ohfune, Y.; Kurokawa, N. Tetrahedron Lett. 1985, 26,
5307–5308.
8. (a) Fleet, G. W. J.; Son, J. C. Tetrahedron 1988, 44,
2637–2647. (b) Fleet, G. W. J.; Son, J. C.; Green, D. St. C.;