A simple modular synthesis of pyridinoline a collagen cross-link of biochemical interest

Article abstract of DOI:10.1016/S0957-4166(03)00400-2

A simple convergent synthesis of the collagen cross-link pyridinoline starting from glycine is reported.

Full text of DOI:10.1016/S0957-4166(03)00400-2

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 14 (2003) 2005–2012
A simple modular synthesis of pyridinoline a collagen cross-link
of biochemical interest
Pietro Allevia,* and Mario Anastasiab
aDipartimento di Medicina, Chirurgia e Odontoiatria, Universita` di Milano, via A. Di Rudin`ı 8, I-20142 Milano, Italy
bDipartimento di Chimica, Biochimica e Biotecnologie per la Medicina, Universita` di Milano, via Saldini 50,
I-20133 Milano, Italy
Received 1 April 2003; revised 7 May 2003; accepted 13 May 2003
Abstract—A simple convergent synthesis of the collagen cross-link pyridinoline starting from glycine is reported. © 2003 Elsevier
Ltd. All rights reserved.
1. Introduction
these compounds, various synthetic protocols for their
synthesis have recently been reported.4 Moreover, while
the routes to obtain dPyd 2, starting from natural
amino acids, have been widely exploited and well set-
up, those affording Pyd 1 are far less advanced, due to
the difficulties connected with the presence of an addi-
tional stereogenic centre in the chain bonded to the
heterocyclic nitrogen of the pyridine portion of the
molecule. Thus, we started this work with the aim of
solving the pending troubles connected with the modu-
lar preparation4a,g,i of Pyd 1, from the protected pyridi-
nes 3a4b or 3b, and regarding the availability of the
stereomerically pure amino acidic halohydrins 4a–e.
Pyridinoline (Pyd; 1; Fig. 1) and deoxypyridinoline
(dPyd; 2) are two pyridinium cross-links formed in the
mature form of collagen from lysine and hydroxylysine
residues.1 Currently the urinary levels of these two
collagen cross-links are used in the clinic as reliable
biomarkers for the non-invasive assessment of bone
degradation in various pathologies, such as osteoporo-
sis, bone cancer and alteration of bone metabolism.2
Thus it is particularly important to have available
suitable amounts of Pyd 1 and dPyd 2 in pure form for
use as reference standards in diagnostics. As a conse-
quence of both, the difficulties involved in the procure-
ment of Pyd 1 and of dPyd 2 from natural sources3 and
of the chemical interest for the structural features of
In fact, the amino acidic iodohydrin 4a was obtained4g
from the corresponding bromohydrin 4b. This derived
Figure 1.
* Corresponding author. Fax: +39 0250316040; e-mail: pietro.allevi@unimi.it
0957-4166/03/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0957-4166(03)00400-2

2006
P. Alle6i, M. Anastasia / Tetrahedron: Asymmetry 14 (2003) 2005–2012
from glutamic acid as a 1:1 mixture with the 2S,5S
stereoisomer from which it can not be directly resolved
by common methods, not even qualitatively. Moreover,
both free and the derivative esterified at the alcoholic
group, even as Mosher esters, behave similarly on TLC
In fact, for the construction of the stereogenic centre at
C-5 of the iodohydrin 4c, one available option was to
perform a stereoselective iodolactonization of the
unsaturated acid 5 (Scheme 3) hoping to obtain the
(2S,5R)-d-valerolactone 6 and modest quantities of the
undesired isomeric (2S,5S)-d-valerolactone 7. The lac-
tone 6, by successive alcoholysis, could afford the iodo-
hydrin 4c.
and
reverse
phase
HPLC.4g
Similarly,
the
diastereomeric mixture of the parent epoxides is unsep-
arable under several different conditions. As a conse-
quence, the 2S,5R diastereomer 4b useful for the
preparation of the iodohydrin 4a and of the Pyd 1,
needed be isolated by esterification of the alcoholic
group with a thyroxin acid derivative, followed by
separation of the esters through preparative HPLC and
subsequent selective hydrolysis thereof.4g Similar
difficulties were observed in the preparation of the
iodohydrin 4c which was only recently obtained in our
laboratory from the corresponding bromohydrin 4d
which was separated by a relatively simple artifice4a
from its (2S,5S)-diastereomer. Moreover, in no case
was the possibility of constructing the stereogenic cen-
tres of the halohydrins 4 considered.
With this goal in mind, we first obtained the unreported
L
-amino acid 5 by asymmetric alkylation of a chiral
derivative of glycine (Scheme 2), according to the
method suggest by Hoarau et al.5 for the synthesis of
the (R)-isomer of 5, used in the synthesis of pipecolic
acid. The enantiomeric excess of the amino acid 5 was
shown to be higher than 98% by chiral GLC of the
L
-glutammic acid diethyl ester, obtained via initial oxi-
dation of the double bond of 5 (see Section 3).
The iodolactonization of the amino acid 5 could be
performed under two sets of conditions: under thermo-
dynamic control, involving the equilibration of a cyclic
intermediate by action of hydrogen iodide formed in
the reaction, or under kinetic control, preventing the
action of the acid with a base.6 In a preliminary step we
performed MM+ and AM1 calculation of the desired
lactone 6 and of its diastereomer 7, in order to have an
estimate of the relative thermodynamic stability to help
us choosing the most favourable conditions to run the
iodolactonization reaction. Using both computational
methods, we found very low energy differences between
the trans- and cis-lactones 6 and 7 showing that the
2. Results and discussion
Here we report the modular synthesis of Pyd 1 by
alkylation of the substituted pyridine 3a with the iodo-
hydrin 4c (Scheme 1). Both the pyridine nucleus and
the aminoacidic iodohydrin are obtained in a short
stereoselective way, starting from the achiral amino
acid glycine, which is the starting material for the
preparation of the key (S)-2-benzyloxycarbonylamino-
5-hexenoic acid 5 (Scheme 2).
Scheme 1.
Scheme 2. Reagents and conditions: (i) BF3·Et2O, toluene, reflux, 2.5 h, 72%; (ii) LiHMDS, 4-bromo-1-butene, THF, −100“
−30°C, 20 h, 85%; (iii) citric acid, H2O/THF, room temp., 84 h, 85%; (iv) CbzCl, dioxane/H2O, room temp., 2.5 h, 89%; (v) TFA,
room temp., 1 h, 98%.

P. Alle6i, M. Anastasia / Tetrahedron: Asymmetry 14 (2003) 2005–2012
2007
Scheme 3. Reagents and conditions: (i) I2, KI, NaHCO3/H2O, 0°C, 24 h, 6 (39%) and 7 (18%); (ii) NaN3, DMF, room temp., 12
h, 96%; (iii) Cs2CO3, CH2ꢀCHCH2OH, reflux, 8 min, 82%; (iv) MeCN, reflux, 16 h, 65%; (v) (Ph3)4Pd, N-methylaniline, THF,
rt, 76 h, 60%; (vi) H2, Pd/C, EtOH–H2O (80:20), room temp., 95%; (vii) CF3CO2H–H2O (95:5), room temp., 40 min, 65%.
However, under thermodynamic conditions, the reac-
tion does not occur at all, and surprisingly our initial
conditions were found to be the best ones so far to
obtain the iodolactone 6. Fortunately, the undesired
(2S,5S)-lactone 7 could be recycled and reutilised for
the synthesis of Pyd 1 and so the obtained diastereose-
lectivity was considered acceptable.
cis-lactone is a little more stable (Dcis“trans E=1.0 kJ
mol−1, for MM+, and Dcis“trans E=1.1 kJ mol−1, for
AM1). This corresponds to a very close equilibrium
mixture of these lactones (trans/cis ratio=0.6–0.7:1).7
Moreover, this result prompted us to study first the
iodolactonization reaction under kinetic control, treat-
ing the amino acid 5 with a mixture of iodine and
potassium iodide in dichloromethane–water saturated
with sodium hydrogen carbonate. A mixture of the
iodolactones 6 and 7 was then obtained in satisfactory
yields and in a 7:3 molar ratio.
The iodolactone 6 was then treated with allyl alcohol
(Scheme 3), in the presence of caesium carbonate, to
afford by transesterification the iodohydrin 4c. These
controlled conditions allow the loss of iodine and the
formation of the corresponding epoxide to be avoided
which smoothly forms in the presence of anhydrous
sodium carbonate.6a,9 The iodocompound 4c was then
reacted with the trisubstituted pyridine 3a in boiling
acetonitrile to afford the fully protected Pyd 10 and
finally Pyd 1, after gradual regeneration of the
aminoacidic functions via compounds 11 and 12.
The iodolactones 6 and 7 were easily separated by rapid
chromatography and their stereochemistry was assigned
by transformation into the corresponding azido lac-
tones 8 and 9 to which we had recently assigned a
complete structure.8 To confirm that under the reaction
conditions no equilibration of the isomeric iodolactones
6 and 7 occurs, each iodolactone was isolated and
stirred in the presence of a new mixture of the solvent
and reagents used in the iodolactonization. In each
case, after 3 h, the same pure starting lactone was
detected by HPLC.
With the synthesis of Pyd 1 in hand, we improved it by
recycling the (2S,5S)-lactone 7. This was transesterified
under the same conditions studied for its stereomer 6 to
afford the iodohydrin 13 (Scheme 4). This was then
oxidized to the iodoketone 14 which afforded addi-
tional pure iodohydrin 4c in three steps: sodium boro-
hydride reduction of the keto group to afford a
diastereomeric mixture of the two iodohydrines 4c and
In order to improve the diastereoselectivity observed in
the lactonization, we tried to optimise the reaction
conditions (varying solvent, temperature etc.) and also
performed the reaction under thermodynamic control.6a

2008
P. Alle6i, M. Anastasia / Tetrahedron: Asymmetry 14 (2003) 2005–2012
Scheme 4. Reagents and conditions: (i) Cs2CO3, CH2ꢀCHCH2OH, reflux, 8 min, 80%; (ii) pyridinium chlorochromate, CH2Cl2,
room temp., 5 h, 81%; (iii) NaBH4, MeOH, 0°C, 20 min, 90%; (iv) CF3CO2H, room temp., 40 min, 6 (32%) and 7 (36%); (v)
benzylamine, K2CO3, MeCN, room temp., 8 h, then TMG, O2, MeOH, room temp., 6 h, 46%.
13, lactonization of the obtained iodohydrin mixture to
the separable iodolactones 6 and 7, separation and
transesterification of 6 with allylic alcohol. Moreover,
the iodoketone 14 was also a good starting material for
the preparation of the substituted pyridine 3c since it
reacts with benzylamine to afford the pyridinium salt
15, the proximate precursor of 3c.10 The ‘one pot’
construction of the pyridinium nucleus, starting with
the iodoketone 14, proceeds even better than using the
analogue bromoketone studied in our original
protocol.4b
trin (Lipodex E)12 capillary column (25 m, 0.25 mm ID,
purchased from Macherey-Nagel); carrier gas was He
set at 88 kPa column head pressure and the column
temperature was set at 220°C. HPLC analyses were
carried out on a silica direct phase column (superspher
Si-60, 12.5 cm, 4 mm ID, purchased from Merck); the
mobile phase was hexane/2-propanol, 93:7, v:v; the flow
rate was 1 mL/min and the detection was performed at
221 nm. UV spectra were obtained using a Perkin–
Elmer Lambda11 UV/VIS spectrometer. Mass spectra
were obtained using a Finnigan LCQdeca (Thermo-
Quest) ion trap mass spectrometer fitted with an elec-
trospray source (ESI). All reactions were monitored by
thin-layer chromatography (TLC) carried out on 0.25
mm E. Merck silica gel plates (60 F254) using UV light,
50% sulphuric acid or 0.2% ninhydrin in ethanol and
heat as developing agent. E. Merck 230–400 mesh silica
gel was used for flash column chromatography.13 Usual
work-up refers to washing the organic layer with water,
drying over Na2SO4, and evaporating the solvent under
reduced pressure.
Thus, we have set-up a concise and efficient protocol
for the synthesis of pyridinoline 1 starting from glycine.
The results allow a more direct and less expensive
synthesis of this precious collagen cross-link.
3. Experimental
3.1. General
Nuclear magnetic resonance spectra were recorded at
298 K on Bruker AM-500 spectrometer operating at
3.2. (S)-2-Benzyloxycarbonylaminohex-5-enoic acid 5
1
500.13 MHz for H and 125.76 MHz for 13C. Chemical
The acid 5 was obtained using the same asymmetric
shifts are reported in parts per million (ppm, l units)
relative to solvent signal (residual proton signal for
proton spectra or carbon signal for carbon spectra).11
Proton and carbon assignments were established, if
necessary, with homonuclear and heteronuclear 2D J-
alkylation reported by Hoarau et al.5 for the prepara-
tion of the tert-butyl ester of the
D
-isomer of 5. In our
preparation we started with the tert-butyl glycinate
Schiff base of (1R,2R,5R)-2-hydroxypinan-3-one and
obtained the tert-butyl (2S)-2-aminohex-5-enoate (75%
yield from Schiff base) as an oil: [h]D=+26.6 (c 1,
MeOH) (lit.5 −25.2 for its enantiomer); 1H NMR
(CDCl3): l 7.34–7.28 (5H, aromatics-H), 5.78 (1H, m,
5-H), 5.26 (1H, d, J=6.3, NH), 5.09 (2H, s, OCH2Ph),
5.02 (1H, dd, J=16.8 and <1, 6-Ha), 4.97 (1H, dd,
J=9.8 and <1, 6-Hb), 4.26 (1H, m, 2-H), 2.08 (2H, m,
4-H2), 1.89 (1H, m, 3-Ha), 1.71 (1H, m, 3-Hb), 1.44
[9H, s, C(CH3)3]. Anal. calcd for C10H19NO2: C, 64.83;
H, 10.34; N, 7.56. Found: C, 64.67; H, 10.41; N, 7.48%.
1
revolved experiments. H NMR data are tabulated in
the following order: number of protons, multiplicity (s,
singlet; d, doublet; bs, broad singlet; m, multiplet),
coupling constant(s) in hertz, assignment of proton(s).
Optical rotations were taken at 24°C on a Perkin–
Elmer 241 polarimeter and [h]D values are given in 10−1
deg cm2 g−1. Chiral GLC analyses were carried out on
a Hewlett-Packard 5890 gas chromatography equipped
with a octakis(3-O-butyryl-2,6-di-O-pentyl)-g-cyclodex-

P. Alle6i, M. Anastasia / Tetrahedron: Asymmetry 14 (2003) 2005–2012
2009
1H NMR is in agreement with that of the reported
-isomer by Hoarau et al.5 After carbobenzylation of
oxycarbonylamino-5-iodomethyl-l-valerolactone
7
D
(1.33 g, 18%): mp 142–143°C (from CH2Cl2–diiso-
propyl ether); HPLC: Rt=7.7 min; [h]D=+41.5 (c 1,
the amino group,5 the tert-butyl (2S)-2-benzyloxycar-
bonylaminohex-5-enoate (85% yield) was obtained: an
oil; [h]D=+1.04 (c 1, CHCl3) [lit.15 −15 (c 1, MeOH)];
1H NMR (CDCl3): l 7.35–7.30 (5H, aromatics-H),
5.77 (1H, m, 5-H), 5.28 (1H, d, J=7.7, NH), 5.08
(2H, s, OCH2Ph), 5.02 (1H, dd, J=17.0 and <1, 6-
Ha), 4.97 (1H, dd, J=10.1 and <1, 6-Hb), 4.26 (1H,
m, 2-H), 2.08 (2H, m, 4-H2), 1.89 (1H, m, 3-Ha), 1.74
(1H, m, 3-Hb), 1.44 [9H, s, C(CH3)3]. Anal. calcd for
C18H25NO4: C, 67.69; H, 7.89; N, 4.39. Found: C,
67.80; H, 7.81; N, 4.43%. 1H NMR is in agreement
1
CHCl3); H NMR (CDCl3): l 7.36–7.28 (5H, aromat-
ics-H), 5.59 (1H, d, J=4.5, NH), 5.11 (2H, s,
OCH2Ph), 4.48–4.40 (2H, overlapping, 2-H and 5-H),
3.34 (1H, dd, J=10.5 and 5.1, CHHI), 3.27 (1H, dd,
J=10.5 and 6.2, CHHI), 2.64 (1H, m, 3-Ha), 2.28
(1H, m, 4-Ha), 1.80 (1H, m, 4-Hb), 1.63 (1H, m,
3-Hb). Anal. calcd for C14H16INO4: C, 43.21; H, 4.14;
N, 3.60. Found: C, 43.13; H, 4.22; N, 3.69%.
Further elution gave the (2S,5R)-2-benzyloxycarbonyl-
amino-5-iodomethyl-d-valerolactone 6 (2.88 g, 39%):
mp 112–114°C (from CH2Cl2–diisopropyl ether);
HPLC: Rt=13.2 min; [h]D=7.0 (c 1, CHCl3); 1H
NMR (CDCl3): l 7.36–7.28 (5H, aromatics-H), 5.50
(1H, d, J=3.0, NH), 5.10 (2H, s, OCH2Ph), 4.30 (1H,
m, 5-H), 4.15 (1H, m, 2-H), 3.34 (2H, br s, CH2I),
2.47 (1H, m, 3-Ha), 2.24 (1H, m, 4-Ha), 1.94–1.85
(2H, overlapping, 3-Hb and 4-Hb). Anal. calcd for
C14H16INO4: C, 43.21; H, 4.14; N, 3.60. Found: C,
43.32; H, 4.10; N, 3.57%.
with that of the
D
-isomer reported by Hoarau et al.5
The tert-butyl
(2S)-2-benzyloxycarbonylamino-5-
hexenoate (1.0 g; 3.1 mmol) was dissolved in aqueous
trifluoracetic acid (8 ml; 95%) and the resulting solu-
tion was stirred at room temperature for 1 h. The
solvent was then removed under reduced pressure to
afford the pure acid 5 (807 mg, 98%): mp 69–71 (from
CH2Cl2–diisopropylether); [h]D=+21.6 (c 1, CHCl3);
1H NMR (CDCl3): l 7.36–7.26 (5H, aromatics), 6.36
(1H, m, 4-H), 5.27 (1H, d, J=7.9, NH), 5.16–4.99
(4H, overlapping, 6-Ha, 6-Hb and CH2Ph), 4.40 (1H,
m, 2-H), 2.16–2.12 (2H, overlapping, 4-Ha and 4-Hb),
1.98 (1H, m, 3-Ha), 1.78 (1H, m, 3-Hb). Anal. calcd
for C14H17NO4: C, 63.87; H, 6.51; N, 5.32. Found: C,
63.60; H, 6.40; N, 5.45%.
3.4. Separate transformation of the (2S,5R)- and
(2S,5S)-2-benzyloxycarbonylamino-5-iodomethyl-d-
valerolactones 6 and 7 into the corresponding azido-
lactones 8 and 9
The enantiomeric excess (>98%) of the amino acid 5
was detected by chiral GLC (see Section 3.1 for condi-
tions) after transformation of 5 into glutammic acid
diethyl ester N-trifluoroacetate. The transformation
requires: an initial esterification of 5 (20 mg) with
EtOH (in 1 M HCl at 25°C for 24 h) followed by: an
oxidation with NaIO4–KMnO4 (tBuOH–H2O at 25°C
for 12 h);14 a second esterification with EtOH of the
obtained acid; the decarbobenzylation of the amino
group (by hydrogenolysis in the presence of Pd/C);
and a trifluoracetilation (with trifluoroacetic anhy-
dride).
Each lactone 6 and 7 (200 mg, 0.51 mmol), dissolved
in DMF (2 mL), was treated with NaN3 (66 mg, 1.02
mmol) at room temperature for 12 h. Then the reac-
tion mixture was poured in ice cold water and
extracted with AcOEt to afford, after usual work-up,
the appropriate crystalline azidolactone. In particular,
starting from the lactone 6, the (2S,5R)-5-azidomethyl-
2-benzyloxycarbonylamino-d-valerolactone
8
(96%
yield) was obtained: mp 86–87°C (from CH2Cl2–diiso-
propyl ether); identical in all respects with that previ-
ously described.8 Starting from the (2S,5S)-lactone 7,
the pure (2S,5S)-5-azidomethyl-2-benzyloxycarbonyl-
amino-l-valerolactone 9 (97% yield) was obtained:
mp 90–92°C (from CH2Cl2–diisopropyl ether); identi-
cal in all respects with that previously described.8
The retention time of the obtained trifluoroacetate was
identical to that of an authentic sample obtained from
L
-glutamic acid (31.87 min) and differed from that of
the -isomer derivative (29.26 min).
D
3.5. Attempts of separate equilibration of the (2S,5R)-
and (2S,5S)-2-benzyloxycarbonylamino-5-iodomethyl-d-
valerolactones 6 and 7 under the conditions used for
their preparation by iodolactonization
3.3. Iodolactonization of (S)-2-benzyloxycarbonylamino-
hex-5-enoic acid 5
A saturated aqueous solution of NaHCO3 (200 mL)
containing iodine (14.48 g, 57.00 mmol) and KI (28.96
g; 174.44 mmol) was added to a solution of the acid 5
(5.00 g, 19.0 mmol) in CH2Cl2 (200 mL), at 0°C under
vigorous stirring. After 48 h, a saturated aqueous solu-
tion of Na2S2O4 was added to the reaction mixture
which was extracted with dichloromethane and
worked up. The obtained residue, after separation of
the organic phase and usual work-up, was quickly
chromatographed on column (eluting with CH2Cl2–
AcOEt; 100:10 v/v) to afford first the (2S,5S)-2-benzyl-
Each lactone 6 and 7 (74 mg, 0.19 mmol) was sepa-
rately dissolved in CH2Cl2 (2 mL) and treated at 0°C
for 48 h with a saturated solution of NaHCO3 con-
taining iodine (145 mg, 0.57 mmol) and KI (290 mg;
1.75 mmol). The mixture was then diluted with a
saturated aqueous solution of Na2S2O4 and worked up
to afford a crude residue which was directly subjected
to HPLC analysis. This showed that each starting
lactone was uncontaminated by the respective
diastereomer (the lactone 6 showed Rt=13.2 min; the
lactone 7 showed: Rt=7.7 min).

2010
P. Alle6i, M. Anastasia / Tetrahedron: Asymmetry 14 (2003) 2005–2012
3.6. Allyl (2S,5R)-2-benzyloxycarbonylamino-5-
hydroxy-6-iodohexanoate 4c
2-hydroxypentyl]-3-pyridiniumolate 11 (508 mg, 60%)
as a resinous material.
The (2S,5R)-lactone 6 (2.00 g, 5.1 mmol), dissolved in
allylic alcohol (60 mL) was treated with Cs2CO3 (1.20
g) at reflux for 7–8 min. The time control is very
important since, at longer reflux time, the progressive
formation of the corresponding epoxide is observed.
The reaction mixture was then diluted with AcOEt (20
mL), washed with water and worked up. Column chro-
matography (eluting with hexane–AcOEt; 80:20 v/v)
afforded the pure iodohydrin 4c4a (1.87 g; 82%): mp
72–73°C (from CH2Cl2–diisopropyl ether); [h]D=+7.0
(c 1, CHCl3); 1H NMR (CDCl3): l 7.36–7.30 (5H,
aromatics-H), 5.88 (1H, m, CH2ꢀCH-CH2O), 5.42 (1H,
d, J=6.5, NH), 5.32 (1H, dd, J=16.8 and <1,
CHHꢀCH-CH2O), 5.25 (1H, dd, J=10.3 and <1,
CHHꢀCH-CH2O), 5.12–5.07 (2H, AB system,
OCH2Ph), 4.62 (2H, d, J=4.0, CH2ꢀCH-CH2O), 4.45
(1H, m, 2-H), 3.54 (1H, m, 5-H), 3.30 (1H, dd, J=9.9
and 3.5, 6-Ha), 3.18 (1H, dd, J=9.9 and 6.6, 6-Hb),
2.04 (1H, m, 3-Ha), 1.75 (1H, m, 3-Hb), 1.64 (1H, m,
4-Ha), 1.57 (1H, m, 4-Hb). Anal. calcd for C17H22INO5:
C, 45.65; H, 4.96; N, 3.13. Found: C, 45.52; H, 4.88; N,
3.21%.
The crude acid (450 mg, 0.48 mmol), dissolved in
aqueous EtOH (180 mL, 80%), was then hydrogenated
in the presence of 10% Pd/C (80 mg) at room tempera-
ture and atmospheric pressure. Filtration of the catalyst
and evaporation of the solvent give a residue, which
was chromatographed (eluting with MeOH–NH3; 100:5
v/v) to afford the title compound 12 (244 mg, 95%) as
a glass. The compound was identical (1H and 13C
NMR, mass spectra) to that previously obtained.4a
3.9. Pyridinoline 1
The tert-butyl derivative 12 (230 mg, 0.43 mmol) was
dissolved in aqueous CF3CO2H (8 mL, 95%: v/v) and
the resulting solution was stirred at room temperature
for 40 min. The solvent was then removed under
reduced pressure and the residue was purified by
column chromatography (eluting with MeOH–NH3;
80:20 v/v). The obtained residue was then dissolved in
aqueous HCl (7.0 mL, 1.0 M) and the solution was
lyophilised to afford the pyridinoline 1 as tetrachloride
dihydrate (167 mg, 65%): umax (HCl 0.1 M)/nm 242
(m/dm3 mol−1 cm−1 3850), 294 (6520); umax (50 mM
phosphate buffer, pH 7.5)/nm 252 (m/dm3 mol−1 cm−1
3710), 324 (6150); ESI/MS m/z 429 (M+). Anal. calcd
for C18H36N4Cl4O10: C, 35.42; H, 5.95; N, 9.18; Cl,
23.24. Found: C, 35.48; H, 6.00; N, 9.22; Cl, 23.36%.
Other physico-chemical properties (1H and 13C NMR)
confirmed those reported.4a
3.7. 4-[(S)-2-Benzyloxycarbonylamino-2-(tert-butyloxy-
carbonyl)ethyl]-5-[(S)-3-benzyloxy carbonylamino-3-
(tert-butyloxycarbonyl)propyl]-1-[(2R,5S)-5-benzyloxycar
bonylamino-5-(allyloxycarbonyl)-2-hydroxypentyl]-3-
pyridiniumolate 10
The 4-[(S)-2-benzyloxycarbonylamino-2-(tert-butyloxy-
carbonyl)ethyl] - 5 - [(S) - 3 - benzyloxycarbonylamino - 3-
(tert-butyloxycarbonyl)propyl]-3-hydroxypyridine 3a4b
(795 mg, 1.2 mmol) and the allyl (2S,5R)-2-benzyl-
3.10. Allyl (2S,5S)-2-benzyloxycarbonylamino-5-
hydroxy-6-iodohexanoate 13
oxycarbonylamino-5-hydroxy-6-iodohexanoate
4c
(2.142 g, 4.8 mmol) were dissolved in CH3CN (15
mL) and refluxed under argon atmosphere for 16 h.
The solvent was then evaporated under reduced pres-
The (2S,5S)-lactone 7 (2.00 g, 5.1 mmol), dissolved in
allylic alcohol (60 mL) was treated with NaHCO3 (1.20
g) at reflux for 7–8 min. At longer reflux time (15–20
min) the formation of the corresponding epoxide starts.
Dilution with AcOEt (20 mL) and usual workup
afforded a residue which, after column chromatography
(eluting with hexane–AcOEt; 80:20 v/v), afforded the
pure iodohydrin 13 (1.83 g; 80%): an oil; [h]D=+3.1 (c
sure to give
a crude product which was chro-
matographed (eluting with AcOEt–MeOH; 100:7 v/v)
to afford the title compound 10 (764 mg, 65%): a
resinous material; [h]D=−2.6 (c 1, CHCl3). The com-
pound was identical (1H and 13C NMR, mass spectra)
to that previously obtained.4a
1
1, CHCl3); H NMR (CDCl3): l 7.36–7.29 (5H, aro-
3.8. 4-[(S)-2-Amino-2-(tert-butyloxycarbonyl)ethyl]-5-
[(S)-3-amino-3-(tert-butyloxycarbonyl)propyl]-1-
[(2R,5S)-5-amino-5-carboxy-2-hydroxypentyl]-3-
pyridiniumolate 12
matics-H), 5.88 (1H, m, CH2ꢀCH-CH2O), 5.36 (1H, d,
J=6.3, NH), 5.31 (1H, dd, J=17.0 and <1, CHHꢀCH-
CH2O), 5.24 (1H, dd, J=10.2 and <1, CHHꢀCH-
CH2O), 5.12–5.06 (2H, AB system, OCH2Ph), 4.62 (2H,
d, J=4.2, CH2ꢀCH-CH2O), 4.41 (1H, m, 2-H), 3.56
(1H, m, 5-H), 3.31 (1H, dd, J=9.8 and 3.0, 6-Ha), 3.17
(1H, dd, J=9.8 and 7.0, 6-Hb), 2.09 (1H, m, 3-Ha),
1.97 (1H, m, 3-Hb), 1.84 (1H, m, 4-Ha), 1.66 (1H, m,
4-Hb). Anal. calcd for C17H22INO5: C, 45.65; H, 4.96;
N, 3.13. Found: C, 45.78; H, 5.05; N, 3.08%.
A solution containing the allylic ester 10 (885 mg, 0.9
mmol), N-methylaniline (2.89 g, 27.00 mmol) and tet-
rakis(triphenylphosphine) palladium(0) (104 mg, 0.09
mmol) in anhydrous THF (15 mL) was stirred, under
argon, at room temperature for 76 h. After evapora-
tion of the solvent under reduced pressure, the residue
was chromatographed (eluting with CH2Cl2–
MeOH; 9:1 v/v) to afford the 4-[(S)-2-benzyloxycar-
bonylamino-2-(tert-butyloxycarbonyl)ethyl]-5-[(S)-3-
benzyloxycarbonylamino-3-(tert-butyloxycarbonyl)pro-
pyl]-1-[(2R,5S)-5-benzyloxycarbonylamino-5-carboxy-
3.11. Allyl (2S)-2-benzyloxycarbonylamino-6-iodo-5-oxo-
hexanoate 14
The iodohydrin 13 (1.0 g, 2.24 mmol), dissolved
in CH2Cl2 (90 mL), was treated with pyridinium

P. Alle6i, M. Anastasia / Tetrahedron: Asymmetry 14 (2003) 2005–2012
2011
chlorochromate (966 mg, 4.48 mmol) at room tempera-
ture for 5 h. Then isopropanol was added, followed by
ice cold water and the solution was worked up to afford
a residue which, after column chromatography on silica
(eluting with hexane–AcOEt; 80:20 v/v), afforded the
pure iodo ketone 14 (806 mg, 81%): mp 79–81°C (from
temperature for 40 min. The solvent was then removed
under reduced pressure (under 40°C) and the residue (492
mg), was quickly chromatographed on column (eluting
with CH2Cl2/AcOEt; 100:10; v:v) to afford first the
lactone 7 (154 mg; 36%) and then the lactone 6 (137 mg;
32%) described above.
1
CH2Cl2–diisopropyl ether); [h]D=−3.0 (c 1, CHCl3). H
NMR (CDCl3): l 7.36–7.29 (5H, aromatici), 5.88 (1H,
m, CH2ꢀCH-CH2O), 5.36 (1H, d, J=4.0, NH), 5.32
(1H, dd, J=17.4 e <1, CHHꢀCH-CH2O),), 5.25 (1H,
dd, J=10.7 e<1, CHHꢀCH-CH2O), 5.12–5.05 (2H, sis-
tema AB, CH2Ph), 4.62 (2H, d, J=5.5, CH2ꢀCH-
CH2O), 4.36 (1H, m, 2-H), 3.75 (2H, s, 6-H2), 2.80 (2H,
m, 4-H2), 2.22 (1H, m, 3-Ha), 1.94 (1H, m, 3-Hb).
Anal. calcd for C14H16INO5: C, 41.50; H, 3.98; N, 3.46.
Found: C, 41.62; H, 4.08; N, 3.37%.
Acknowledgements
This paper is dedicated to Professor Paolo Edgardo
Todesco on the occasion of his 70th birthday. This work
was supported financially by Italian MIUR (Ministero
dell’Istruzione, dell’Universita` e della Ricerca).
3.12.
(S,S)-1-Benzyl-4-[2-benzyloxycarbonylamino-2-
References
(allyloxycarbonyl)ethyl]-5-[3-benzyloxycarbonylamino-3-
(allyloxycarbonyl)propyl]-3-pyridiniumolate 15 starting
from benzylamine and iodo ketone 14
1. For recent reviews on the collagen cross-links, see: (a)
Watts, N. B. Clin. Chem. 1999, 45, 1359–1368; (b) Knott,
L.; Baily, A. J. Bone 1998, 22, 181–187; (c) James, I. T.;
Walne, A. J.; Perret, D. Ann. Clin. Biochem. 1996, 33,
397–420 and references cited therein.
2. (a) Coleman, R. E. Cancer 2002, 94, 2521–2533; (b)
Seregni, E.; Martinetti, A.; Ferrari, L.; Bombardieri, E.
Q. J. Nucl. Med. 2001, 45, 7–17; (c) Editorial, Lancet
1992, 340, 278.
To a solution containing benzylamine (0.109 mL, 1 mmol)
and the iodo ketone 14 (979 mg, 2.2 mmol) in CH3CN
(20mL), anhydrousK2CO3 (460mg, 3.3mmol)wasadded
and the mixture was stirred under nitrogen for 8 h. At
this time the disappearance of the starting iodo ketone
was monitored (TLC), the solid K2CO3 was filtered on
a pad of Celite and the solvent was evaporated under
reduced pressure. The residue was then dissolved in
MeOH (20 mL) and tetramethylguanidine (0.375 mL, 3
mmol) was added to the mixture, which was vigorously
stirred at 25°C under oxygen atmosphere for 6 h. The
solvent was then removed under reduced pressure and the
residue was purified by rapid chromatography (eluting
with CH2Cl2–MeOH; 9:1 v/v) to afford the title com-
3. Robins, S. P.; Duncan, A.; Wilson, N.; Evans, B. J. Clin.
Chem. 1996, 42, 1621–1626.
4. (a) Allevi, P.; Galligani, M.; Anastasia, M. Tetrahedron:
Asymmetry 2002, 13, 1901–1910; (b) Anastasia, L.; Anas-
tasia, M.; Allevi, P. J. Chem. Soc., Perkin Trans. 1 2001,
2404–2408; (c) Adamczyk, M.; Johnson, D. D.; Reddy,
R. E. Tetrahedron: Asymmetry 2000, 11, 2289–2298 and
Corrigendum: Tetrahedron: Asymmetry 2000, 11, 5017;
(d) Adamczyk, M.; Johnson, D. D.; Reddy, R. E. Tetra-
hedron Lett. 1999, 40, 8993–8994 and Corrigendum: Tet-
rahedron Lett. 2001, 42, 767; (e) Allevi, P.; Longo, A.;
Anastasia, M. J. Chem. Soc., Perkin Trans. 1 1999,
2867–2868; (f) Allevi, P.; Longo, A.; Anastasia, M. Chem.
Commun. 1999, 559–560; (g) Adamczyk, M.; Johnson, D.
D.; Reddy, R. E. Tetrahedron 1999, 55, 63–88; (h) R. P.
Hatch, USP 5 723 619/1998 (Chem. Abstr. 1998, 128,
205137); (i) Waelchli, R.; Beerli, C. H.; Meigel, H.;
Re´ve´sz, L. Bioorg. Med. Chem. Lett. 1997, 7, 2831–2836.
5. Hoarau, S.; Fauchere, J. L.; Pappalardo, L.; Roumestant,
M. L.; Viallefont, P. Tetrahedron: Asymmetry 1996, 7,
2585–2593.
6. (a) Bartlett, P. A.; Myerson, J. J. Am. Chem. Soc. 1978,
100, 3950–3952; (b) Corey, E. J.; Hase, T. Tetrahedron
Lett. 1979, 335–338; (c) Bartlett, P. A. Tetrahedron 1980,
36, 2–72.
7. The gas phase conformational search of each lactone 6
and 7 was performed using molecular mechanic force
field (MM+; implemented within the hyperchem software
package) and varying all the significative dihedral angles.
The 100 lowest energy conformations founded for each
1
pound 15 (332 mg, 46%) as a glass: H NMR (CDCl3):
l 7.50 (1H, br s, pyridinium-H), 7.39–7.22 (15H, aromat-
ics-H), 7.04 (1H, br s, pyridinium-H), 5.83 (2H, m,
2×CH2ꢀCH-CH2O), 5.59 (2H, m, 2×NH), 5.31–4.94
(10H, overlapping, 2×OCH2Ph, N+CH2Ph, 2×CH2ꢀCH-
CH2O), 4.62 (2H, m, CH2ꢀCH-CH2O), 4.56 (2H, m,
CH2ꢀCH-CH2O),
4.34
(2H,
m,
2×CH(N-
HCbz)CO2Allyl), 3.29 (1H, m), 2.98 (1H, m), 2.64 (2H,
m), 2.01 (1H, m), 1.75 (1H, m). Anal. calcd for
C41H43N3O9: C, 68.22; H, 6.00; N, 5.82. Found: C, 68.06;
H, 5.85; N, 5.73%.
3.13. Transformation of the iodo ketone 14 into a
diastereoisomeric mixture of iodo lactones 6 and 7
NaBH4 (65 mg; 1.7 mmol) was gradually added to a
solution of the iodo ketone 14 (700 g; 1.57 mmol) in
MeOH (100 mL), at −5°C. The mixture was stirred at
−5°C for 20 min, then was poured into an ice cold aqueous
solution of HCl (2 M) and extracted with AcOEt. Usual
work-up afforded a chromatographically unseparable
mixture of diastereoisomeric iodohydrines 4c and 13 (633
mg; 90%, in a 1:1 ratio).
isomer were successively optimised with a more stringent
minimization criterion (10−2 kcal mol−1
A
of the deriva-
−1
,
The obtained mixture of iodohydrines was dissolved in
CF3CO2H (1.5 mL) and the solution was stirred at room
tive) using both molecular mechanic (MM+) and quan-
tum mechanical (AM1) methods.

2012
P. Alle6i, M. Anastasia / Tetrahedron: Asymmetry 14 (2003) 2005–2012
8. Allevi, P.; Anastasia, M. Tetrahedron: Asymmetry 2000,
11, 3151–3160.
12. Konig, W. A. J. High Resolut. Chromatogr. 1993, 16,
569–586.
13. Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43,
2923–2925.
9. Bartlett, P. A.; Myerson, J. J. Org. Chem. 1979, 44,
1625–1627.
14. Douat, C.; Heitz, A.; Martinez, J.; Fehrentz, J. A. Tetra-
hedron Lett. 2001, 42, 3319–3321.
10. The possible selective debenzylation of 15 is now under
study.
15. Expo`sito, A.; Ferna`ndez-Sua`rez, M.; Iglesias, T.; Munoz,
11. Gottlieb, H. E.; Kotlyar, V.; Nudelman, A. J. Org. Chem.
1997, 62, 7512–7515.
L.; Riguera, R. J. Org. Chem. 2001, 66, 4206–4213.