The stereo structures of some mycophenolic acid derivatives

Article abstract of DOI:10.1071/CH06481

X-Ray Crystallographic evidences for the stereochemistry of three iodo compounds derived from mycophenolic acid, including the crystallographic structure of a tertiary iodide, are provided. Mycophenolic acid, a compound of considerable medicinal interest, exhibits excellent antimitotic properties. Mycophenolic acid, when shaken in bicarbonate solution with iodine in ethyl acetate, produces almost equal amounts of an acid and the dilactone. These compound, both crystalline, can be easily isolated from the bicarbonate and ethyl acetate layers respectively. The acid compound obtained by the mycophenolic acid in bicarbonate solution, was converted into its methyl ester, to provide material suitable for X-ray examination.

Full text of DOI:10.1071/CH06481

Full Paper
CSIRO PUBLISHING
Aust. J. Chem. 2007, 60, 354–357
www.publish.csiro.au/journals/ajc
The Stereo Structures of Some MycophenolicAcid Derivatives
Paul V. Bernhardt,A Raymond M. Carman,A,B and Tri T. LeA
ADepartment of Chemistry, University of Queensland, Brisbane, QLD 4072, Australia.
BCorresponding author. Email: carman@chemistry.uq.edu.au
X-Ray crystallography affords the stereochemistry of three iodo compounds derived from mycophenolic acid. These
included the crystallographic structure of a tertiary iodide.
Manuscript received: 18 December 2006.
Final version: 25 March 2007.
Introduction
Mycophenolic acid 1 remains a compound of considerable
medicinal interest.[1] It shows exceptional antimitotic properties,
but is rapidly eliminated from the human body as the glucuronate
derived from the phenolic hydroxyl group.Attempts to block this
hydroxylgroupleadtolossofbiologicalactivity, butvariousiodo
compounds have been explored[2] with the aim to temporarily
block the phenolic group to give derivatives which might then
revert back to mycophenolic acid in vivo. These compounds,
such as 2–4, all have two new stereogenic centres (Scheme 1).
The relative stereochemistry at these centres was previously
assigned only on mechanistic grounds; and we now provide
X-ray crystallographic evidence for the correct structures.
Tertiary iodides are normally unstable and unsuited to X-ray
examination. The crystallographic structures of only 38 tertiary
iodides are to be found in the Cambridge Structural Database,[3]
and the large majority of these are bridgehead iodides which owe
their relative stability to the protection that the bridgehead posi-
tion provides against SN2, SN1, E2, or E1 loss of the iodine atom.
It was hoped that the electronegative oxygen on the carbon atom
vicinal to the iodine in compound 2 would provide a stabiliz-
ing influence, and that this tertiary iodide would be sufficiently
stable for X-ray crystallographic examination.
16
COOR
15
Me
OH
Me
14
O
17
Me
13
I
H
HOOC
O
O
7
12
11
O
MeO
6
5
8
3
1
O
2
1
MeO
10
4
Me
9
2a R ϭ H
2b R ϭ Me
15
17
14
16
COOMe
18
14 16
O
Me
17
Me
I
13
H
I
15
O
13
o4
H OH
O
12
11
O
1
o1
12
11
O
7
6
5
8
3
7
6
8
3
1
O
2
O
2
MeO
10
4
5
4
MeO
10
Me
9
Me
9
3
4
15
Results and Discussion
O
15
17
Me
14 16
O
16 14
As previously reported,[2] mycophenolic acid, when shaken in
bicarbonate solution with iodine in ethyl acetate, provides almost
equal amounts of the acid 2a and the dilactone 3. These two
compounds, both crystalline, are easily isolated from the bicar-
bonate and ethyl acetate layers respectively. Both compounds are
racemic as befits the achiral nature of the starting mycophenolic
acid 1. Common crystallographic numbering is used for the car-
bon skeletons in the following compounds as it enables a direct
comparison of atoms to be made within the variety of chemical
structures.
Compound 3 afforded crystals suitable for X-ray work. The
structure comprises two independent molecules in the asym-
metric unit. There are no significant differences between their
respective bond lengths, angles, or dihedral angles. One of these
molecules is shown in Fig. 1. Of most relevance to this study is
that the stereochemistry 12RS,13SR is defined, which is consis-
tent with the trans addition of iodine and the carboxyl group to
17
Me
13
13
O
O
O
7
12
11
O
O
12
O
11
7
6
8
3
6
8
1
H
1
2
H
O
O
2
5
MeO
10
5
3
4
MeO
10
4
Me
9
Me
9
5
6
12
MeOOC
OH
7
O
11
O
4
6
5
8
3
1
1
5
Me
O
3
2
2
COOH
MeO
4
10
8
Me
9
7
Scheme 1. Structures 1–8.
© CSIRO 2007
10.1071/CH06481
0004-9425/07/050354

Structures of Mycophenolic Acid Derivatives
355
Fig. 1. Compound 3.
Fig. 3. Compound 4.
Fig. 2. Compound 2b.
the starting double bond. The hydroxyl groups of each molecule
participate in both intramolecular (O4A–H4A· · ·O1A 2.38 Å,
136.5◦; O4B–H4B· · ·O1B 2.52 Å, 124.2◦) and intermolecu-
lar H-bonds (O4A–H4A· · ·O5A [−x + 2, y, z + 1/2] 2.27 Å,
138.7◦; O4B–H4B· · ·O5B [−x + 1, y, z + 1/2] 2.20 Å, 135.1◦),
which generate two independent and racemic H-bonded chains
that each follow a glide plane along the c-axis.
Acid 2a crystallized as fine feathers from several solvents,
and did not provide material suitable for X-ray examination.
The compound was therefore converted into its methyl ester 2b,
either with diazomethane or through the mixed anhydride route.
This product 2b gave needles from acetone/water at room tem-
perature, and provided the crystallographic structure shown in
Fig. 2. Again the 12RS,13SR configuration is consistent with the
trans addition of iodine and the phenolic oxygen to the starting
mycophenolic double bond.
Perhaps surprisingly, given the presence of the tertiary iodide,
compound 2a cleanly dissolves in concentrated sulfuric acid.
When this solution is poured into methanol the rearranged
methyl ester 4 is formed[2] in near quantitative yield. X-Ray
crystallography of this compound provided the stereochem-
istry depicted in Fig. 3. The configuration is again 12RS,13SR,
consistent with inversion at both these carbon atoms during the
rearrangement 2a to 4. In the most stable conformer, the methyl
group in compound 4 is quasi axial to the pyran ring, while the
iodine is equatorial.
Fig. 4. Compound 7.
triethylamine, followed by methanol), the reaction temperature
of the first step was accidentally allowed to become too high.
The crystalline product failed to provide suitable crystals for X-
ray examination, but gave excellent NMR data that indicated
that it had lost iodine to form a trisubstituted double bond.
The compound is neutral and possesses no methyl ester group,
and is formulated as either 5 or 6. The carbon connectivites
in these two compounds are identical, and NMR examination
did not allow a final distinction between the two possibilities.
A least-squares analysis of the chemical shift differences, for
both the proton and carbon data, from those shifts predicted by
CHEMDRAW also did not allow a clear distinction, although
structure 5 was slightly favoured. However, the magnitude of
the vicinal coupling constants J14,15 (10, 9.8, 9.5, 2.1 Hz) might
suggest that these protons are on a six-membered ring rather than
a five, favouring structure 6. We note that structure 5 has been
tentatively proposed[2] for a compound isolated from a consid-
erable mixture when acid 2a was treated with hot pyridine. The
product now obtained (mp 148º) is different from that previously
described[2] (mp 192–194º).
Ironically, given the time we had spent in attempting to grow
crystals of compound 5/6, the NMR sample deposited very large
crystals as colourless needles following the slow evaporation
(about one week) of CDCl3/methanol. X-Ray crystallographic
examination of these crystals provided the decomposed struc-
ture 7 (Fig. 4). In contrast to the H-bonded polymer seen
During one attempt to convert acid 2a into its methyl ester
2b by the mixed anhydride route (ethyl chloroformate and

356
P. V. Bernhardt, R. M. Carman, and T. T. Le
Table 1. 1H NMR chemical shifts and multiplicities
Compound
2b
3
4
5/6
7
H2 (2)
H9 (3)
H10 (3)
H11a
H11b
H12
H14a
H14b
H15a
5.10 (s)
2.04 (s)
3.96 (s)
5.19 (s)
2.13 (s)
3.78 (s)
5.05 (s)
2.11 (s)
3.79 (s)
5.23 (s)
2.16 (s)
4.12 (s)
6.30 (s)
–
5.21 (s)
2.15 (s)
3.76 (s)
3.71 (s)
3.59 (dd)
3.42 (dd)
4.71 (dd)
2.29 (ddd)
2.07 (ddd)
2.71 (ddd)
2.59 (ddd)
1.85 (s)
3.36 (dd)
3.25 (dd)
4.66 (dd)
2.36 (ddd)
2.19 (ddd)
2.63 (m)
2.63 (m)
1.70 (s)
3.50 (dd)
3.33 (dd)
4.34 (dd)
2.36 (ddd)
2.17 (ddd)
2.67 (ddd)
2.56 (ddd)
1.46 (s)
3.71 (s)
–
–
–
–
–
–
–
2.70 (ddd)
2.35 (ddd)
2.82 (ddd)
2.51 (ddd)
1.80 (s)
–
H15b
H17 (3)
Other
3.69 (OMe)
7.75 (OH)
3.61 (OMe)
7.70 (OH)
3.70 (OMe)
Table 2. 13C NMR chemical shifts in CDCl3
in the (acentric) structure 3, the compound 7 forms cen-
trosymmetric H-bonded dimers involving both intramolecular
(O4–H4· · ·O12.38 Å, 139.5◦)andintermolecular(O4–H4· · ·O1
[−x, −y + 1, −z + 2] 2.17 Å, 141.1◦) interactions. The com-
pound provided excellent NMR data consistent with structure 7.
The compound 7 as formed was contaminated with an equimo-
lar mixture of 4-ketopentanoic acid (levulinic acid) 8, identified
through NMR spectroscopy and by comparison with an authentic
sample.
The mechanism for the conversion of 5/6 in CDCl3/methanol
into a mixture of 7 and 8 is unclear.All the carbons are accounted
for, and the products are consistent with an ozonolysis-type
cleavage, but of material with a C12–C13 double bond. We
note that levulinic acid 8 has been previously reported[4] as an
ozonolysis product of mycophenolic acid 1, but compound 7
now appears to be new.
Compound
2b
3
4
5/6
7
C1
C2
C3
C4
C5
C6
C7
C8
168.6
69.0
172.7
70.1
168.3
68.0
168.4
69.4
171.7
70.1
145.6
116.7
163.8
115.7
153.6
106.5
11.6
61.0
28.9
172.6
–
147.3
115.4A
159.6
115.3A
157.4
102.6
11.0
145.4
116.9
164.0
120.0
153.5
106.5
11.6
147.9
116.2
161.0
115.3
151.0
109.1
11.1
143.9
116.0
144.7
111.5
156.6
103.6
11.2
C9
C10
C11
C12
C13
C14
C15
C16
C17
Me ester
59.2
34.3
91.2
56.8
60.9
60.3
31.9
28.1
78.4
59.9
29.07A
40.4
119.1
155.9
86.8
87.5
38.7
32.7
173.0
27.6
35.1
35.1
28.0
173.6
20.3
34.2
28.0
175.1
23.3
–
–
–
–
29.11A
175.8
21.9
Experimental
1H and 13C NMR spectra in CDCl3 solution at ambient tem-
perature were recorded using Bruker 500 MHz spectrometers.
Assignments were made using DEPT, HSQC, HMBC, COSY,
and double quantum filtered COSY pulse sequences.
51.9
–
51.8
–
53.3
AValues bearing the same superscript within a column may be interchanged.
Reaction of Mycophenolic Acid 1 with Iodine
Ester 2b
Following the literature procedure,[2] iodine in ethyl acetate was
added with shaking to mycophenolic acid dissolved in aque-
ous sodium bicarbonate. When the iodine colour persisted, the
aqueous layer was acidified to give the (12RS,13SR) iodo acid
2a, mp 168◦C, identical with the literature,[2] in approx. 50%
Acid 2a in acetone was treated overnight at 0◦C with dia-
zomethane in ether. The solution was taken to dryness to provide
colourless crystals of the (12RS,13SR) ester 2b from acetone
at room temperature. mp 126◦C (slow dec. ∼108◦C) (lit.[2]
133–135◦C). δH see Table 1, with J11a,11b −16.0, J11a,12 9.6,
J11b,12 7.0, J14a,14b −15.0, J14a,15a 4.9, J14a,15b 11.2, J14b,15a 11.0,
J14b,15b 5.2, J15a,15b −16.2 Hz, consistent with the literature,[2]
1
yield. The H NMR spectrum ((CD3)2SO) showed peaks con-
sistent with the literature,[2] but the compound decomposed too
fast for elaborate pulse sequencing. Crystals suitable for X-ray
crystallography could not be obtained.
1
but now with better resolution. δC see Table 2, with the com-
pound >98% pure. Crystals for X-ray structure analysis were
grown from acetone by slow evaporation at room temperature.
The same ester 2b was available from acid 2a by the
mixed anhydride route (tetrahydrofuran, triethylamine, ethyl
chloroformate, and methanol; all at ice temperature).[2]
The ethyl acetate layer from the above reaction was washed
with sodium thiosulfate solution, with brine, and then taken
to dryness to yield crystals of the (12RS,13SR) dilactone 3
(∼50%). mp 168◦C (dec.; lit.[2] 170◦C). δH see Table 1, with
J11a,11b −14.5, J11a,12 11.3, J11b,12 3.1, J14a,14b −13.2, J14a,15a
9.0, J14a,15b 9.0, J14b,15a 6.8, J14b,15b 8.4, J15a,15b ≈ −13 Hz, con-
sistent with the literature,[2] but now with better resolution. δC
see Table 2, with the compound >98% pure. Crystals for X-ray
structure analysis were grown from acetone.
Ester 4
Acid 2a was dissolved in conc. H2SO4 at room temperature.
The solution was poured onto methanol at 0◦C. Ice was added

Structures of Mycophenolic Acid Derivatives
357
to give a gummy semi-crystalline precipitate. The liquid was
decanted off and the residue taken into ethyl acetate, washed with
bicarbonate solution, and with brine to provide crystals of the
(12RS,13SR)ester 4 fromethylacetate. mp 168◦C (lit.[2] 170◦C).
δH see Table 1, with J11a,11b −17.0, J11a,12 5.5, J11b,12 10.5,
J14a,14b −14.0, J14a,15a 10.5, J14a,15b 5.5, J14b,15a 5.5, J14b,15b
10.5, J15a,15b −16.0 Hz, consistent with the literature,[2] but now
with better resolution. δC seeTable 2, with the compound >98%
pure. Crystals for X-ray structure analysis were grown from ethyl
acetate by slow evaporation at room temperature.
Compound 3 (Fig. 1)
C17H19IO6: M 446.22. T 293(2) K, orthorhombic, space
group Pc21b, a 8.932(2), b 19.743(4), c 20.210(3) Å, V
3564(1) Å3, Z 8, F(000) 1776, Dc 1.663 g cm−3, µ 18.25 cm−1
,
3237 unique data (2θmax 50◦), R 0.0408 (for 1857 reflections
with I >2σ(I)), wR2 0.0826 (all data).
Compound 4 (Fig. 3)
C18H21IO6: M 460.25, T 293(2) K, monoclinic, space group
P21/a, a 13.624(3), b 8.752(2), c 15.957(8) Å, β 104.66(3)◦, V
1841(1) Å3, Z 4, F(000) 920, Dc 1.661 g cm−3, µ 17.70 cm−1
,
3226 unique data (2θmax 50◦), R 0.0410 (for 1814 reflections
with I >2σ(I)), wR2 0.1191 (all data).
Compound 5/6
Acid 2a in ether was treated with triethylamine. Ethyl chlo-
roformate was added with inadequate cooling, when the ether
boiled. Methanol was added followed by iced water to the clear
solution. The product was extracted into ethyl acetate, washed
with bicarbonate solution and brine, and was taken to dryness.
The crystalline product went somewhat orange over the weekend
(loss of iodine). Recrystallization (ethyl acetate) gave colourless
feathery needles. mp 148◦C. Found: C 64.4, H 5.2. C17H16O6
requires C 64.6, H 5.1%. δH see Table 1, with H11 as a sharp
singlet, J14a,14b −13.5, J14a,15a 9.5, J14a,15b 2.1, J14b,15a 10.0,
J14b,15b 9.8, J15a,15b −17.7 Hz. δC see Table 2, with the com-
pound >98% pure. Crystals suitable for X-ray structure analysis
could not be obtained.
Compound 7 (Fig. 4)
¯
C13H14O6: M 266.24. T 293(2) K, triclinic, space group P1,
a 7.910(1), b 8.393(1), c 10.536(1) Å, α 66.07(1), β 83.32(1),
γ 81.96(1)◦, V 631.7(1) Å3, Z 2, F(000) 280, Dc 1.400 g cm−3
,
µ 1.12 cm−1, 2224 unique data (2θmax 50◦), R 0.0448 (for 1144
reflections with I > 2σ(I)), wR2 0.1452 (all data).
IntensitydatawerecollectedonanEnraf–NoniusCAD4four-
circle diffractometer using graphite monochromatized MoKα
radiation (λ 0.71073 Å) in the ω–2θ scan mode. Data were mea-
sured at room temperature. Lattice dimensions were determined
by a least-squares fit of the setting parameters of 25 independent
reflections. Data reduction and empirical absorption corrections
(ψ-scans) were performed with the WINGX package.[6] Struc-
tures were solved by direct methods with SHELXS and refined
by full matrix least-squares analysis with SHELXL97.[7] All non-
H atoms were refined with anisotropic thermal parameters, and
H-atoms were constrained at estimated positions using a riding
model. The atomic nomenclature is defined in Figs 1, 2, 3, and
4 drawn with ORTEP3.[8] Crystallographic data in CIF format
are available from the Cambridge Crystallographic Data Base
(CCDC deposition nos 629744, 629742, 629741, and 629743
for compounds 2b, 3, 4, and 7, respectively).
Ester 7
The above compound 5/6 by slow evaporation (∼1 week)
of a CDCl3/methanol solution produced large colourless
needles in a residual oily base. 1H and 13C NMR analysis
showed an approximately equimolar mixture of compounds
7 and 8. Methyl 2-(4ꢀ-hydroxy-6ꢀ-methoxy-7ꢀ-methyl-3ꢀ-oxo-
1ꢀ,3ꢀ-dihydroisobenzofuran-5ꢀ-yl)ethanoate 7 had mp 132◦C
(from chloroform). Found: C 58.4, H 5.5. C13H14O6 requires
1
C 58.7, H 5.3%. δH see Table 1, with no coupled protons. δC
see Table 2, with the compound >98% pure apart from traces of
levulinic acid 8. Large needles for X-ray crystallography were
grown from chloroform.
Acknowledgments
R.M.C. thanks James DeVoss for providing office and laboratory space, and
Patricia Hayes for technical support.
Levulinic acid (4-oxopentanoic acid) 8 had δH 2.72 (2H t,
H3), 2.59 (2H t, H2), 2.16 (3H s, H5), with J2,3 6.5 Hz. δC
206.6 (C4), 178.4 (C1), 37.7 (C3), 29.7 (C5), 27.7 (C2). These
assignmentswereconfirmed by two-dimensional andlong-range
coupling experiments. The spectra were identical to those of an
authentic sample of 4-oxopentanoic acid, and are consistent with
the literature[5] provided that the assignments to H2 and H3, and
to C2 and C3, are reversed in this database.[5]
References
[1] Over 2950 hits for ‘mycophenolic acid’ in Scifinder Scholar since
2000.
[2] R. M. Carman, Aust. J. Chem. 1978, 31, 353.
[3] F. H. Allen, Acta Crystallogr. 2002, B58, 380.
[4] J. H. Birkinshaw, A. Bracken, E. N. Morgan, H. Raistrick, Biochem.
J. 1948, 43, 216.
[5] Spectral Database for Organic Compounds; www.aist.go.jp/RIODB/
SDBS/cgi-bin/ENTRANCE.cgi, SDBS#1418[accessed13November
2006].
[6] L. J. Farrugia, J.Appl. Crystallogr. 1999, 32, 837. doi:10.1107/S00218
89899006020
[7] G. M. Sheldrick, SHELX97, Programs for Crystal Structure Analysis
(Release 97–2) 1998 (Universität Göttingen: Göttingen).
[8] L. J. Farrugia, J.Appl. Crystallogr. 1997, 30, 565. doi:10.1107/S00218
89897003117
Crystallography
Compound 2b (Fig. 2)
C18H21IO6: M 460.25. T 293(2) K, monoclinic, space group
P21/c, a 6.437(1), b 23.369(2), c 12.520(3) Å, β 101.17(2)◦, V
1847.7(4) Å3, Z 4, F(000) 920, Dc 1.655 g cm−3, µ 17.63 cm−1
,
3246 unique data (2θmax 50◦), R 0.0429 (for 1846 reflections
with I > 2σ(I)), wR2 0.1179 (all data).