´
G. M. Salamonczyk / Tetrahedron Letters xxx (2015) xxx–xxx
2
S
dendrimers composed of both thiophosphate and 1,3,5-benzenet-
ricarboxylate units are reported herein.
S
P
P
O
O
O
OH
2
3
Although the surface unit precursor, 3,5-bis[(dimethoxyphos-
phonyl)methyl]benzoic acid (4) is a known compound,17 its syn-
thesis required slight improvements. The synthesis commenced
from a convenient starting material, methyl 3,5-bis(hydroxyl-
methyl)benzoate (1), which can be easily obtained from the parent
trimesic acid.16
8
EDC
DMAP
4
88%
CH2P(O)(OR)2
S
S
P
P
Bromination of diol 1 with phosphorus tribromide18 in carbon
CH2P(O)(OR)2
O
O
O
O
2
3
O
tetrachloride provided dibromide
2 in 90% isolated yield.
9: R=Me
10: R=H
(Scheme 1) An Arbusov19 reaction (90 °C, overnight) using an
excess of trimethyl phosphite (neat) afforded the corresponding
bisphosphonate 3 in high yield (84%). It is worth noting that bis-
phosphonate 3 cannot be obtained directly from diol 1 using recent
methodology,20 which does not work in the case of methyl phos-
phonates. The carboxylic ester functionality in bisphosphonate 3
was hydrolyzed using aqueous lithium hydroxide (1 M) to provide
the key monomer, 3,5-bis[(dimethoxyphosphonyl)methyl]benzoic
acid (4) in 86% yield. This compound represents an AB2-type mono-
mer. The A group (carboxyl) is active and the B groups (phospho-
nate) are protected such that the A group solely reacts with the B
(active) groups in the prior generation of the dendrimer. Another
monomer, this time, A2B, 7 (in concordance to the above descrip-
tion) was obtained from trimesic acid dimethyl ester (5).18,21 In
this case the free carboxylic acid of compound 5 was reacted with
di-tert-butyl-dicarbonate to yield the corresponding trimesic acid
tert-butyl-dimethyl ester (89%).22 Selective hydrolysis of both
methyl groups in the triester was effectively achieved using aque-
ous lithium hydroxide to provide compound 6 in high yield (88%).
Another highly chemoselective reaction was the reduction of the
two carboxylate groups in diacid monoester 6 using borane-
dimethyl sulfide complex, which furnished 3,5-bis(hydrox-
ymethyl)benzoic acid tert-butyl ester (7) in a high isolated yield
(83%).
TMSBr/MeOH
Scheme 2. Synthesis of dendrimer 10 with 12 phosphonate groups.
corresponding trimethylsilyl phosphonate ester intermediate that
was swiftly hydrolyzed in situ to give phosphonic acids 10 in a pro-
tic solvent (MeOH). All of the deprotection stages were unambigu-
ously evidenced by 31P {1H} NMR. The presence of subsequent
intermediates, bis-trimethylsilyl phosphonate esters (8 ppm) and
even very unstable phosphonic acids trimethylsilyl esters
(ꢀ17 ppm, broad singlet, most likely due to the existence of the
P-epimeric isomers) were detected. These rather mild conditions
did not cause cleavage of the other ester bonds present in the
macromolecule and the integrity of the dendritic structure was
preserved. Dodecaphosphonic acid 10 was further transformed
into its dodecaanion (sodium hydrogen phosphonates) using aque-
ous sodium bicarbonate.
To demonstrate the usefulness of the presented synthetic
approach, another, much larger phosphonate polyester dendrimer
13, (theoretical mol. wt. 12096 D) was synthesized from den-
drimeric substrate 11.4b During this synthesis, the fully protected
intermediate 12 was isolated, which was used to form novel
polyanionic compound 13 (Fig. 1). Despite the fact that dendrimer
13 (generation four, the largest unprotected compound described
here, overall yield 50%, from 11) was composed of a lipophilic, thio-
phosphate-based interior (22 P = S functions), possessing 24 4-car-
bon chains, 21 3-carbon chains, and 24 aromatic rings, the effect of
the surface polar phosphonate groups seemed to predominate.
Free acid 13 was sparingly soluble in water, and in order to achieve
reasonable solubility (50 mg/1 mL), it was necessary to convert a
few (statistically three/four) of the 48 P(O)(OH)2 groups into the
corresponding sodium salts.
The phosphorus-based dendrimers, used in this project as sub-
strates were synthesized in a divergent manner from readily avail-
able, inexpensive chemicals, via an amidophosphite method. The
detailed syntheses of thiophosphate dendrimers 8 and 11 have
been described previously.4b,15 It is worth noting that thiophos-
phate dendrimers turned out to be well tolerated by biological sys-
tems.23 Hydroxy-terminated, first generation, thiophosphate
dendrimer 8 was reacted with an excess of acid 4, (Scheme 2) in
the presence of the water-soluble carbodiimide, 1-ethyl-3-(3-
dimethylaminopropyl)carbodiimide24 (EDC), and 4-dimethy-
laminopyridine (DMAP) to afford the second generation dendrimer
9 in 88% isolated yield. Cleavage of the terminal methyl phospho-
The presented methodology also enabled the convergent syn-
thesis of phosphonate-capped polyester dendrimers. Moreover,
depending on the nature of the substituents connected to both
sides of the starting monomer or focal point, the interior of the
macromolecule can be tuned accordingly. The synthesis of the
nate esters in
method,25 by means of a large excess of trimethylsilyl bromide
(TMSBr) which transformed methyl phosphonate into the
9 was accomplished by applying, McKenna’s
9
O
O
HOH2C
MeOOC
CH2OH
BrH2C
CH2Br
(MeO)2PH2C
CH2P(OMe)2
a
90%
b
84%
COOMe
COOMe
COOR
3
4
: R=Me
c
86%
1
2
: R=H
HOOC
COOH
HOH2C
CH2OH
COOMe
d then c
78%
e
83%
COOH
COOtBu
COOtBu
5
7
6
Scheme 1. Reagents and conditions: (a) PBr3, CCl4, rt; (b) P(OMe)3, 90 °C; (c) LiOH aq, MeOH, 4 °C; (d) di-tert-butyl-dicarbonate, DMAP, CH2Cl2, rt; (e) B2H6Á(CH3)2S, THF, rt.
´