Synthesis of new dendrimers - Trimesic acid derivatives

Article abstract of DOI:10.1016/j.tetlet.2010.11.016

The efficient synthesis of new dendrimeric polyesters up to generation 3 that consist of 1,3,5-benzenetricarboxylic acid building blocks with potential applications in drug delivery is described. The dendrimers possess hydroxy or allyl functional groups o

Full text of DOI:10.1016/j.tetlet.2010.11.016

Tetrahedron Letters 52 (2011) 155–158
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Synthesis of new dendrimers—trimesic acid derivatives
⇑
´
Grzegorz M. Salamonczyk
´
Centre of Molecular and Macromolecular Studies, The Polish Academy of Sciences, Sienkiewicza 112, 90-363 Łódz, Poland
a r t i c l e i n f o
a b s t r a c t
Article history:
The efficient synthesis of new dendrimeric polyesters up to generation 3 that consist of 1,3,5-benzenetri-
carboxylic acid building blocks with potential applications in drug delivery is described. The dendrimers
possess hydroxy or allyl functional groups on the surface and were prepared through a divergent
approach using readily available 2-(hydroxymethyl)-2-ethylpropan-1,3-diol and 1,3,5-benzenetrimetha-
nol as central cores, with 3,5-bis[(allyloxy)methyl]benzoic acid being an essential unit of the dendrimer.
The latter compound was synthesized, in high yield, from 1,3,5-benzenetricarboxylic acid, applying selec-
tive hydrolysis of the corresponding triester as the key step.
Received 28 September 2010
Revised 26 October 2010
Accepted 5 November 2010
Available online 11 November 2010
Keywords:
Dendrimer
Polyester
Ó 2010 Elsevier Ltd. All rights reserved.
Selectivity
1,3,5-Benzenetricarboxylic acid
Deallylation
Dendrimers are highly branched, three dimensional monodi-
spersed polymers made up of repeating cascade structures ema-
nating from a central core to a polyfunctional surface. These
synthetic macromolecules offer numerous applications,1 especially
in biomedical sciences as drugs in their own right,2 or as drug and
gene carriers.3 While several types of dendrimers have been de-
scribed as in vivo biomolecule delivery agents, carboxylate polyes-
ter dendrimers have proved to be the most important4 due to their
biodegradability, which causes the macromolecules, in the pres-
ence of enzymes, to hydrolyze into small molecules which can
leave the body. Although the chemical synthesis of dendrimers is
well-founded, the synthesis of carboxylate polyester-based dendri-
mers is still very much warranted.
On the other hand, trimesic acid (1,3,5-benzenetricarboxylic
acid), a planar and highly symmetrical trifunctional compound,
seems to be an appropriate candidate as a structural monomer of
the dendrimer framework. Previously, 1,3,5-benzenetricarboxylic
acid has been used as the central core in both dendrimeric polyary-
lesters5 and polyarylamides.6 A few years ago, we developed a
method for the synthesis of dendrimeric polyphosphates and their
analogs.7 In continuation of our efforts, an efficient and divergent
synthesis of new polyester dendrimers possessing the trimesic acid
skeleton is reported herein.
spectrally pure after crystallization from ethyl acetate (79% yield
from 1).
Selective formation of the monoester can be explained due to
very slow and disfavored nucleophilic attack of the hydroxy anion
on the remaining ester carbonyl group in the double-negatively
charged molecule. Another highly chemoselective reaction was
reduction (rt, 24 h) of the two carboxylate groups in diacid mono-
ester 2 using borane–dimethyl sulfide complex, which provided
3,5-bis(hydroxymethyl)benzoic acid methyl ester (3) in a high iso-
lated yield (80% after crystallization from CH2Cl2–MeOH–cyclohex-
ane). At this point, difficulties were encountered in the selection of
suitable protecting group for the hydroxy functions. The first
choice was the acetyl group, which worked perfectly for the syn-
thesis of phosphoric acid polyester dendrimers.7 Unfortunately,
this actual dendrimer was not stable enough for hydrolysis and/
or alcoholysis of the peripheral acetates without affecting the ben-
zenecarboxylic acid polyester scaffold. On the other hand, the den-
drimer showed an important feature—it was relatively easy to
hydrolyze. After further experimentation, it turned out that allyl
and p-methoxybenzyl (a benzylic diol cannot be protected and
deprotected successfully with a benzyl group) groups were the
most suitable for this synthesis. Hence, the opportunity of further
functionalization, the size, and particularly the atom economy
strongly supported the allyl group. Therefore, diol 3 was reacted
with an excess (3 equiv) of allyl bromide [NaOH, benzene–DMF
(6:1), 70 °C, 7 h]. This reaction produced a mixture of the expected
diallyl ether methyl ester 4 and diallyl ether allyl ester 5 in a 7:1
ratio and 92% yield, after passing the crude mixture through a short
pad of silica gel using dichloromethane as the eluent. Obviously,
separation of the above mixture was not necessary as basic hydro-
lysis and subsequent acidification afforded the target dendrimer
The synthesis of the key monomer proceeds from 1,3,5-benzen-
etricarboxylic acid, which was transformed into trimethyl ester 1.8
Careful basic hydrolysis with dilute NaOH (2.07 equiv) in a metha-
nol–water mixture (Scheme 1) produced almost exclusively9 (90%,
according to NMR) the corresponding monoester 2,10 which was
⇑
Tel.: +48 42 684 7113; fax: +48 42 684 7126.
E-mail address: gmsalamo@cbmm.lodz.pl
0040-4039/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2010.11.016

´
G. M. Salamonczyk / Tetrahedron Letters 52 (2011) 155–158
156
propyl)carbodiimide12 (EDCI) (3.3 equiv), and 4-dimethylamino-
pyridine (DMAP) (0.35 equiv) to furnish the first generation
dendrimer 713 in 93% yield.
CH2OH
MeOOC
COOMe HOOC
COOH HOH2C
a
79%
b
80%
COOMe
COOMe
COOMe
1
3
2
AllOH2C
CH2OAll
OH
O
O
AllOH2C
CH2OAll
HO
O
O
4
c
92%
d
94%
COOMe
OH
OH
OH
+
All = allyl
AllOH2C
CH2OAll
COOH
HO
6
5
O
O
O
:
= 7:1
4 5
COOAll
O
O
O
Scheme 1. Reagents and conditions: (a) NaOH, MeOH, H2O, then HCl (aq); (b)
B2H6Á(CH3)2S, THF; (c) NaOH, allyl bromide, benzene, DMF; (d) NaOH, MeOH, H2O,
then HCl (aq).
O
O
O
O
OH
HO
O
O
building block 611 in 94% yield. This compound represents an AB2-
type monomer. The A group (carboxyl) is active and the B groups
(hydroxy) are protected such that the A group reacts solely with
the B (active) groups in the prior generation of the dendrimer.
The core triol, 2-(hydroxymethyl)-2-ethylpropane-1,3-diol,
commonly known as trimethylolpropane (TMP) reacted readily
(Scheme 2) with acid 6 (3.2 equiv, 12 h, CH2Cl2, rt), in the presence
of the water-soluble carbodiimide, 1-ethyl-3-(3-dimethylamino-
OH
O
HO
O
OH
9OH
OH
Figure 1. Structure of the hydroxy-terminated 2nd generation dendrimer 9OH
possessing a 1,3,5-benzenetrimethanol core.
OAll
OH
AllO
HO
OAll
AllO
OAll
OAll
OH
OH
O
O
O
O
OH
6
H
O
OH
O
O
O
O
O
a
b
OH
O
O
O
O
OH
AllO
OAll
OH
7
7OH
AllO
OAll
AllO
AllO
AllO
AllO
O
O
OAll
O
AllO
O
O
O
O
O
O
O
O
O
O
O
OAll
O
AllO
AllO
O
O
O
OAll
O
O
O
O
O
O
O
AllO
O
O
O
O
O
AllO
O
OAll
O
O
O
O
OAll
OAll
O
O
AllO
O
O
OAll
AllO
O
O
OAll
All = allyl
O
OAll
AllO
AllO
8
Scheme 2. Reagents and conditions: (a) EDCI, DMAP, CH2Cl2; (b) Pd/C, AcOH–H2O–MeOH, TsOH.

´
157
G. M. Salamonczyk / Tetrahedron Letters 52 (2011) 155–158
Figure 2. MALDI TOF mass spectrum of 3rd generation dendrimer 8.
Cleavage of the terminal allyl ethers was carried out under
Acknowledgment
milder (than original) conditions using a slight modification of
the procedure disclosed by Boss and Scheffold.14 Thus, hexaallyl
This work was supported by the Ministry of Science and Higher
Education, Grant No. N N204 030836.
ether
7 was stirred in acetic acid–water–methanol mixture
(7:2:1), at 50 °C, in the presence of catalytic amounts of 10% Pd/C
(0.4 equiv) and p-toluenesulfonic acid (ꢀ0.2 equiv). The reaction
was monitored by 1H NMR; after 40 h, no trace of vinylic protons
was observed and the desired hexol 7OH was obtained in 90%
yield.15 It is noteworthy that the only by-product, propionalde-
hyde, as a volatile compound, was continuously removed from
the reaction mixture to avoid any undesired side reactions. Reiter-
ation of growth and deprotection provided the 2nd (75% yield), 3rd
(8, 45%), and 4th (<10%, not isolated) generation dendrimers.16 An-
other set of polyester dendrimers17 (Fig. 1) was prepared, starting
from 1,3,5-benzenetrimethanol18 by repeating the synthetic proce-
dures described above.
In this case, the yields of the dendrimers were somewhat lower:
88% (1st), 63% (2nd), and 31% (3rd generation dendrimer), respec-
tively. All allyl terminated compounds were stable, colorless oils
the viscosity of which increased with increasing molecular weight.
Hydroxy terminated dendrimers were white solids. The high purity
of the final products was confirmed by NMR and MALDI TOF mass
spectrometry. For example, Figure 2 shows the MALDI TOF mass
References and notes
1. For a recent book and reviews, see: (a) Vögtle, F.; Richardt, G.; Werner, N.;
Rackstraw, A. J. Dendrimer Chemistry: Concepts, Syntheses, Properties,
Applications; Wiley-VCH: Weinheim, 2009; (b) Newkome, G. R.; Shreiner, C.
Chem. Rev. 2010, 110, 6338–6442; (c) Lo, S.-C.; Burn, P. L. Chem. Rev. 2007, 107,
1097–1116.
2. McCarthy, T. D.; Karellas, P.; Henderson, S. A.; Giannis, M.; O’Keefe, D. F.; Heery,
G.; Paull, J. R. A.; Matthews, B. R.; Holan, G. Mol. Pharmaceutics 2005, 2, 312–318.
3. (a) Medina, S. H.; El-Sayed, M. E. H. Chem. Rev. 2009, 109, 3141–3157; (b)
Tekade, R. K.; Kumar, P. V.; Jain, N. K. Chem. Rev. 2009, 109, 49–87; (c) Wolinsky,
J. B.; Grinstaff, M. W. Adv. Drug Delivery Rev. 2008, 60, 1037–1055; (d) Gingras,
M.; Raimundo, J.-M.; Chabre, Y. M. Angew. Chem., Int. Ed. 2007, 46, 1010–1017;
(e) Kim, S. H.; Katzenellenbogen, J. A. Angew. Chem., Int. Ed. 2006, 45, 7243–
7248.
4. For example, see: (a) Ma, X.; Tang, J.; Shen, Y.; Fan, M.; Tang, H.; Radosz, M. J.
Am. Chem. Soc. 2009, 131, 14795–14803; (b) Parrott, M. C.; Benhabbour, S. R.;
Saab, C.; Lemon, J. A.; Parker, S.; Valliant, J. F.; Adronov, A. J. Am. Chem. Soc.
2009, 131, 2906–2916; (c) Almutairi, A.; Akers, W. J.; Berezin, M. Y.; Achilefu,
S.; Fréchet, J. M. J. Mol. Pharmaceutics 2008, 5, 1103–1110; (d) Guillaudeu, S. J.;
Fox, M. E.; Haidar, Y. M.; Dy, E. E.; Szoka, F. C.; Fréchet, J. M. J. Bioconjugate
Chem. 2008, 19, 461–469; (e) Gillies, E. R.; Fréchet, J. M. J. J. Am. Chem. Soc. 2002,
124, 14137–14146.
5. Miller, T. M.; Kwock, E. W.; Neenan, T. X. Macromolecules 1992, 25, 3143–3148.
6. Jikei, M.; Kakimoto, M.-A. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 1293–
1309.
spectrum of 3rd generation dendrimer 8 (calcd for C267H278O66
,
M = 4539.84), where the signal at 4564.4 is attributed to the
molecular ion (M+Na). All three peaks differ in mass by exactly
42 amu, most probably due to minor fragmentation via loss of
one or two allyl cations.
7. For example, see: (a) Salamon´ czyk, G. M.; Kuz´nikowski, M.; Poniatowska, E.
´
Chem. Commun. 2001, 2202–2203; (b) Poniatowska, E.; Salamonczyk, G. M.
Tetrahedron Lett. 2003, 44, 4315–4317; (c) Salamon´ czyk, G. M. Tetrahedron Lett.
2003, 44, 7449–7453.
In conclusion, new carboxylate polyester-based dendrimers
have been prepared as potential candidates for drug delivery. The
mild conditions of both the coupling and deprotection reactions
provided highly pure macromolecular material in good overall
yields. The dendrimers obtained are terminated with carbon–car-
bon double bonds or benzylic hydroxy groups (in contrast to the
much less reactive, frequently reported phenolic hydroxy
groups5,19). Therefore, these reactive peripheral residues could be
transformed into other functions via addition, oxidation, nucleo-
philic substitution reactions, etc. Consequently, these hydrolysable
dendrimers also possess good properties for grafting various com-
pounds, including biomolecules, onto their surfaces.
8. Dimick, S. M.; Powell, S. C.; McMahon, S. A.; Moothoo, D. N.; Naismith, J. H.;
Toone, E. J. J. Am. Chem. Soc. 1999, 121, 10286–10296.
9. To a suspension of 1,3,5-benzenetricarboxylic acid trimethyl ester (1) (2.2 g,
8.8 mmol) in MeOH (100 mL), aq NaOH (22.8 mL, 0.8 M, 2.07 equiv) was added.
The resulting cloudy mixture was stirred vigorously (rt) and slowly formed a
solution over ꢀ5 h. After 15 h, MeOH was removed in vacuo. H2O (50 mL) was
added and the solution acidified (pH ꢀ2.0) with 6 M HCl, yielding a white
dispersion. The organic material was extracted with EtOAc (50 mL). The solvent
was removed to give 2 (impure with 7–8% of 1,3,5-benzenetricarboxylic acid
dimethyl ester, according to NMR) as a white powder. Crystallization from
EtOAc provided pure monoester
2
(1.6 g, 79%): mp 220–221 °C; 1H NMR
(200 MHz, CD3OD/CDCl3 1:1) d = 3.95 (s, 3H), 8.81 [d, 4J(H,H) = 1.6 Hz, 2H], 8.83
[t, 4J(H,H) = 1.6 Hz, 1H] ppm. 13C NMR (50 MHz, CD3OD/CDCl3 1:1) d = 53.01
(CH3), 131.8 (Ar), 132.6 [2C, (Ar)], 134.9 [2C, (ipso Ar)], 135.4 (ipso Ar), 166.4
(C@O), 167.4 [2C, (C@O)] ppm.

´
G. M. Salamonczyk / Tetrahedron Letters 52 (2011) 155–158
158
10. Roosma, J.; Mes, T.; Leclère, P.; Palmans, A. R. A.; Meijer, E. W. J. Am. Chem. Soc.
2008, 130, 1120–1121.
4:1 (5 mL) and the sodium salts removed by filtration to yield the polyol 7OH,
130 mg (90%).
11. Compound 6: 1H NMR (200 MHz, CDCl3) d = 4.04 [d, 3J(H,H) = 5.7 Hz, 4H,
(OCH2CH)], 4.56 (s, 4H, ArCH2O), 5.22 [dd, 3Jcis(H,H) = 10.3 Hz, 2J(H,H) = 1.5 Hz,
2H, (CH@CH2)], 5.29 [dd, 3Jtrans(H,H) = 15.7 Hz, 2J(H,H) = 1.5 Hz, 2H, (CH@CH2)],
16. Allyl protected dendrimers (after aqueous work-up) were purified through a
short plug of silica gel (CH2Cl2/acetone from 50:1 to 10:1, depending on
generation). Data for compound 8: 1H NMR (500 MHz, CDCl3) (superscripts
refer to generation number) d = 0.98 [br t, 3H, (CCH2CH3)], 1.71 [br q, 2H,
(CCH2CH3)], 4.01 [d, 3J(H,H) = 5.2 Hz, 48H, (OCH2CH)], 4.50 [s, 48H, (Ar3CH2O)],
3
3
5.96 [ddd, Jtrans(H,H) = 15.7 Hz, Jcis(H,H) = 10.3 Hz, 3J(H,H) = 5.7 Hz, 2H,
(CH2CH@CH2)], 7.56 (s, 1H, Ar), 7.94 [d, 4J(H,H) = 0.7 Hz, 2H, Ar], 8.81–9.53
(br s, 1H, COOH) ppm. 13C NMR (50 MHz, CDCl3) d = 70.69 [2C, (OCH2CH)],
71.29 [2C, (ArCH2O)], 117.4 [2C, (CH@CH2)], 128.5 [2C, (Ar)], 129.6 [2C, (ipso
Ar)], 132.0 [1C, (Ar)], 133.8 [2C, (CH2CH@CH2)], 139.0 [1C, (ipso Ar)], 171.5 [1C,
(C@O)] ppm.
3
4.52 [s, 6H, (CCH2O)], 5.17 [d, Jcis(H,H) = 9.8 Hz, 24H, (CH@CH2)], 5.27 [d,
3Jtrans(H,H) = 17.1 Hz, 24H, (CH@CH2)], 5.37 [s, 24H, (Ar2CH2O)], 5.39 [s, 12H,
(Ar1CH2O)], 5.89 [m, 24H, (CH2CH@CH2)], 7.55 [s, 12H, (Ar3)], 7.73 [s, 6H, (Ar2)],
7.76 [(s, 3H, (Ar1)], 7.92 [s, 24H, (Ar3)], 8.05 [s, 6H, (Ar1)], 8.11 [s, 12H, (Ar2)]
ppm. 13C NMR (125 MHz, CDCl3) d = 7.60 (CCH2CH3), 29.42 (CCH2CH3), 41.88
[CH2(CH2)CCH2], 64.73 (3C, CCH2O), 66.01 [6C, (Ar1CH2O)], 66.33 [12C,
(Ar2CH2O)], 71.31 [24C, (OCH2CH@)], 71.35 [24C, (Ar3CH2O)], 117.0 [24C,
(CH@CH2)], 128.0 (24C, Ar3), 129.2 (12C, Ar2), 129.4 (12C, Ar3), 129.9 (6C, Ar1),
130.5 (6C, Ar2), 131.5 (3C, Ar1), 134.4 [24C, CH2CH@CH2)], 137.1 [21C, (ipso
12. The use of more traditional dicyclohexyl carbodiimide (DCC) instead of EDCI
resulted in a significantly lower (ꢀ63%) yield in the coupling reaction. Also,
purification of the product from dicyclohexyl urea was very tedious.
13. Compound 7: 1H NMR (200 MHz, CDCl3) d = 1.07 [t, 3J(H,H) = 6.4 Hz, 3H,
(CCH2CH3)], 1.82 [q, 3J(H,H) = 6.4 Hz, 2H, (CCH2CH3)], 4.05 [d, 3J(H,H) = 5.5 Hz,
12H, (OCH2CH)], 4.51 [s, 6H, (CCH2O)], 4.55 [s, 12H, (ArCH2O)], 5.22 [dd,
Ar1,2,3)], 139.1 [42C, ipso Ar1,2,3)], 165.5 [3C, (C@O)1], 166.0 [18C, (C@O)2,3
]
3
3Jcis(H,H) = 10.2 Hz, 2J(H,H) = 0.7 Hz, 6H, (CH@CH2)], 5.32 [dd, Jtrans(H,H) =
ppm. MALDI TOF MS calcd for C267H278O66, M = 4539.84. Found m/z = 4564.4
3
16.4 Hz, 2J(H,H) = 0.7 Hz, 6H, (CH@CH2)], 5.95 [ddd, Jtrans(H,H) = 16.4 Hz,
(M+Na), 4522.8 [MÀ(allyl+)+Na], 4480.3 [MÀ2Á(allyl+)+Na].
3Jcis(H,H) = 10.2 Hz, 3J(H,H) = 5.5 Hz, 6H, (CH2CH@CH2)], 7.60 (s, 3H, Ar), 7.90
(s, 6H, Ar) ppm. 13C NMR (50 MHz, CDCl3) d = 7.36 [1C, (CCH2CH3)], 23.40 [1C,
(CCH2CH3)], 41.40 {1C, [CH2(CH2)CCH2]}, 64.53 [3C, CCH2O)], 71.11 [12C,
(OCH2CH@) and (ArCH2O)], 117.0 [3C, (CH@CH2)], 127.6 (6C, Ar), 129.8 (3C, Ar),
131.2 (3C, ipso Ar), 134.3 [6C, (CH2CH@CH2)], 139.0 (6C, ipso Ar), 165.8 [3C,
17. Data for the 2nd generation dendrimer 9OH (superscripts refer to generation
number): 1H NMR (200 MHz, CD3OD) d = 4.62 [s, 24H, (Ar2CH2OH)], 5.23 [s,
12H, (Ar1CH2OC@O)], 5.42 [s, 6H (Ar0CH2OC@O)], 7.55 [s, 6H (Ar2)], 7.57 [s, 3H
(Ar0)], 7.72 [s, 3H (Ar1)], 7.93 [s, 18H, (Ar1, Ar2)] ppm. 13C NMR (50 MHz,
CD3OD) d = 63.38 (12C, Ar2CH2OH), 66.14 (3C, Ar0CH2O), 66.33 (6C, Ar1CH2O),
127.7 (3C, Ar0), 127.9 (6C, Ar1), 129.2 (12C, Ar2), 129.4 (6C, Ar2), 129.7 (3C, Ar1),
130.2 (3C, ipso Ar1), 131.3 (6C, ipso Ar2), 136.8 (3C, ipso Ar0), 137.7 (6C ipso Ar1),
(C@O)] ppm. MALDI TOF MS calcd for
z = 889.45 (M+Na) (100%).
C51H62O12, M = 866.4. Found m/
14. Boss, R.; Scheffold, R. Angew. Chem., Int. Ed. Engl. 1976, 15, 558–559.
15. Deallylation procedure: First generation dendrimer (200 mg) was stirred
139.9 (12 C, ipso Ar2), 167.2 (9C, C@O) ppm. MALDI TOF MS calcd for C90H84O30
,
M = 1644. Found m/z, fragmentation: 530.6, 574.6, 727.7, 811.7, 833.6, 920.6,
1006.5, 1168.0, 1368.8, 1542.4.
7
(open flask) in AcOH–H2O–MeOH (7:2:1) mixture (5 mL) with 10% Pd/C
(100 mg) and TsOH (ꢀ10 mg) at 50 °C for 40 h. The catalyst was filtered off and
all the liquids removed. The residue was washed with Et2O (2 mL) and
dissolved in saturated NaHCO3 solution in MeOH (2 mL) and again
concentrated in vacuo. Finally, the residue was redissolved in acetone–MeOH
18. Cochrane, W. P.; Pauson, P. L.; Stevens, T. S. J. Chem. Soc., C 1968, 630–632.
19. Potluri, S. K.; Ramulu, A. R.; Pardhasaradhi, M. Tetrahedron 2004, 60, 3739–
3744.