Enabling ScFvs as multi-drug carriers: A dendritic approach

Article abstract of DOI:10.1016/S0968-0896(03)00057-9

To enable scFvs as multi-drug carriers, we designed and synthesized dendritic linker molecules bearing up to nine chlorambucil residues at the branch ends. A maleimide group was used at the focal point of the dendron for easy linkage to the scFv. Original

Full text of DOI:10.1016/S0968-0896(03)00057-9

Bioorganic & Medicinal Chemistry 11 (2003) 1761–1768
Enabling ScFvs as Multi-Drug Carriers: A Dendritic Approach
Chengzao Sun, Peter Wirsching* and Kim D. Janda*
Department of Chemistry, The Scripps Research Institute and the Skaggs Institute for Chemical Biology, 10550 N.Torrey Pines Road,
La Jolla, CA 92037, USA
Received 12 November 2002; accepted 18 December 2002
Abstract—To enable scFvs as multi-drug carriers, we designed and synthesized dendritic linker molecules bearing up to nine
chlorambucil residues at the branch ends. A maleimide group was used at the focal point of the dendron for easy linkage to the
scFv. Originally designed molecules showed poor water solubility. To address this problem, a lysine residue with an unprotected
carboxylic acid group was inserted into the dendron branches. The new molecules showed excellent water solubility and are now
suitable for conjugation. Such dendritic molecules will allow studies to understand the relationship between the drug/antibody ratio
and the potency of the immunoconjugates. The dendritic approach could also be applied to drugs other than chlorambucil and
carriers other than scFvs to greatly increase the drug/carrier ratio.
# 2003 Elsevier Science Ltd. All rights reserved.
Introduction
tumor-associated antigens have improved efficacy and
reduced systemic toxicity due to the high specificity of
the antibody.7,8 To improve the potency of such immu-
noconjugate drugs, efforts have been expended to select
better antibodies as well as designing more efficient
drug-conjugation and releasing strategies.7À12 With
our novel human scFvs, we are investigating new
and more efficient methods to conjugate anti-cancer
drugs.
The successes of antibody-based therapeutics have
opened a new era for the treatment of cancer. Some
monoclonal antibody (mAb)-based drugs such as
Rituxan1, Herceptin1, and Mylotarg1 have been
approved by the FDA and more are in clinical trials.
However, current generation mAb-based drugs can be
limited by the physical properties of the mAbs which are
intact immunoglobulin G (IgG) molecules with a mole-
cular weight of 150kDa. In order to address the limi-
tations of large IgG molecules, smaller engineered
single-chain Fv (scFv) fragments were developed.1,2
These scFvs are composed of a variable region of the
light chain (VL) and a variable region of the heavy chain
(VH) linked by a peptide spacer and are expected to
have improved tumor penetration and pharmacokinetic
properties.3,4
It has been observed that the potency of immuno-
conjugates can be improved by increasing the quantity
of drug delivered per mAb molecule.13,14 Therefore,
one critical aspect of investigation is to improve the
loading of a conjugated drug with respect to the pro-
tein carrier. Typically, mAbs have multiple sites at
which a drug molecule could be covalently attached
without affecting the antigen binding of the mAb. In
contrast, the small size of scFvs (25 kDa) limits the
possible number of sites for drug conjugation. To
address this problem, a linker unit is necessary that will
enable multiple drug molecules to be attached onto one
residue of the scFv. Reports have shown the utility of
doubly branched linkers to carry two drug molecules
onto one site of a mAb.14 Hence, this increased the
molar ratio (MR) of the drug/antibody by 2-fold. Since
scFvs have far fewer sites for the attachment of the
drugs, an alternative and robust way of increasing the
MR is needed. We hypothesized that dendritic mole-
cules could be excellent candidates.
Recently, we successfully developed high-affinity human
scFvs to tumor-associated carbohydrate antigens sialyl
Lewisx, Lewisx and GM3 using phage-display technol-
ogy.5,6 Prospective applications of these scFvs are sig-
nificant and therefore we have examined the possibility
of attaching anti-cancer drugs to form potent immuno-
conjugates. The use of mAb–drug conjugates targeted at
*Corresponding author. Fax: +1-858-784-2595; e-mail: kdjanda@
scripps.edu
0968-0896/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00057-9

1762
C.Sun et al./ Bioorg.Med.Chem.11 (2003) 1761–1768
Dendrimers are hyper-branched macromolecules char-
acterized by defined cascade structures and branch-end
functionalities.15,16 The unique properties of dendrimers
enable their applications as both model compounds for
fundamental research as well as applications in biome-
dical sciences.17 In particular, a properly designed den-
drimer with a well defined internal cavity can be used to
host various small guest molecules and by adjusting the
steric density of the dendrimer periphery allows the
controlled release rate of the guest molecule.18À20 In this
respect, dendrimers have been extensively studied and
applied in the drug delivery field.21À24 However, for our
purposes, we can not use the ‘host–guest’ or encapsula-
tion concept because of the large dendrimeric size which
would be required. A dendrimer designed for drug
release would require a molecular weight of 10–20 kDa
or even more and attachment of such a large molecule
to an scFv could impair its specificity, tumor penetra-
tion and other pharmacokinetic advantages. Another
way in which a dendrimer could be used for drug delivery
would be the attachment of drug molecules covalently
onto the dendrimer periphery. The sizes of a dendrimer
required for this purpose would vary depending on the
loading desired and the mechanism of tumor-cell tar-
geting. Frechet and others have reported a drug-
attached dendrimer, used for both drug delivery and
tumor targeting, that preferentially accumulated in the
solid tumor tissue;25,26 the dendrimer used was
23.5 kDa. Our goal is to apply our newly developed
scFvs as the specificity device while the dendritic lin-
ker functions only as a drug carrier, thus we require
only small dendritic molecules with multiple surface
groups. In addition, we specifically require a dendritic
unit or dendron, which will have two functionalizable
termini denoted as the focal point/core and the surface
groups.
Results and Discussion
Original design and synthesis
Dendritic molecules with a polyamidoether backbone
were selected as our desired multivalent drug carriers.
The syntheses of these three molecules have been pre-
viously discussed,27 and thus are just outlined in Scheme
1. As shown, M-CBL1 carries one, M-G1-CBL3 three
and M-G2-CBL9 nine CBL moieties. The M-G2-CBL9
is the primary compound that we hoped to study while
the other two were prepared for comparison purposes.
These molecules were designed to have CBL at one end
and a maleimide group at the other for covalent linkage
with a cysteine residue site-specifically engineered into the
scFv.28 Notably, the dendritic nonamer is only 5.5 kDa,
which is about 20% of the scFv weight. Unfortunately,
the water solubility of these compounds were quite dis-
appointing. None of the three molecules had a solubility
greater than 200mM in phosphate buffer with 10%
DMSO as cosolvent. Therefore, water-soluble versions of
these three molecules were designed and synthesized.
Water-soluble dendritic molecules
There are a number of routes to make a water-soluble
dendrimer. One example would be to install hydrophilic
peripheral groups like hydroxyl, amino or carboxyl.
Since our compounds will have CBL on the peripheral
termini and we would like to keep the already estab-
lished convergent synthetic methodology to create
homogenous drug carriers, the only feasible way is to
insert a water-soluble unit into the branches. Poly-
ethylene glycol (PEG) units are well known for their
hydrophilic property.29 However, insertion of a hex-
aethylene PEG into the dendrimers failed to give soluble
molecules. Inserting longer PEG chains might be bene-
ficial, but this would also increase the molecular weight
of the dendrons. Therefore, we decided to insert amino
acid residues with a hydrophilic side chain. Both lysine
and glutamic acid were good candidates and have been
incorporated in dendrimer scaffolds.30À33 We chose
lysine for synthetic ease. Two points were taken into
consideration for the arrangement of the three func-
tional groups on lysine. First, it is known that amide
bond linkage to CBL does not impair its alkylating
activity while ester conjugates of CBL may be not as
active as the free drug. Hence one of the amino groups
must be linked to the carboxylic group on CBL. Second,
since CBL itself is in the acid form, it would be better if
the final compound also were in the acid form. Conse-
quently, the carboxylic group must be protected and the
other lysine amino group connected with the dendron.
At the final stage, the carboxylic acid would be depro-
tected to afford a water soluble molecule. According to
these considerations, a set of three new compounds CBL1-
Lys(M)-OH, CBL3-Lys(M-G1)-OH, and CBL9-Lys(M-
G2)-OH were designed as shown in Figure 2. The com-
pound CBL9-Lys(M-G2)-OH is composed of four build-
ing blocks, namely the maleimide focal terminus, a
dendritic linker, a lysine residue and a CBL moiety.
Our dendritic approach to enable scFvs as multi-drug
carriers is depicted in Figure 1. At first, we designed and
synthesized two model compounds containing small
dendritic linkers with three and nine chlorambucil
(CBL) moieties on the surface and a maleimide group at
the focal terminus for attachment to an scFv. Syntheti-
cally we have disclosed this approach.27 However, when
attempting to conjugate these molecules to an scFv, we
discovered that the solubility in a buffer medium was
insufficient to allow reaction. Hence, we now report our
newly designed and synthesized water-soluble version of
these molecules as well as the experimental details for all
compounds.
Figure 1. Cartoon depicting the preparation of an antibody–multidrug
immunoconjugate. Here an scFv is equipped with a cysteine residue
and a dendrimer carrying multiple drug residues is functionalized for
conjugation at the core with a maleimide group.
Syntheses of the compounds were based on our previous
routes with modifications. Thus, di-protected lysine 10

C.Sun et al./ Bioorg.Med.Chem.11 (2003) 1761–1768
1763
Scheme 1. (a) HBTU, NMM, DMF; (b) (1) 50% TFA, CH2Cl2; (2) 1, NMM, CH2Cl2; (c) acrylonitrile, KOH, 1,4-dioxane; (d) CbzCl, aq NaHCO3,
.
1,4-dioxane; (e) (1) NaBH4, CoCl2 6H2O; (2) CBL, HBTU, NMM, DMF; (f) H2, 10% Pd/C, MeOH; (g) 1, NMM, CH2Cl2; (h) (1) H2SO4, EtOH,
reflux; (2) Boc2O, TEA, or CbzCl, aq NaHCO3, 1,4-dioxane; (3) 3 N aq NaOH; (i) 7, HBTU, NMM, DMF; (j) (1) 50% TFA, CH2Cl2; (2) 2, HBTU,
NMM, DMF.
Figure 2. Dendritic drug compounds with improved water solubility as a result of lysine insertion.
was coupled first with CBL to generate 11 in 96% yield.
Deprotection of the e-Z group by Pd/C hydrogenation,
HBTU coupling with maleimido-propionic acid 2 fol-
lowed by TFA deprotection of Boc gave CBL1-Lys(M)-
OH in 88% yield. The e-deprotected amine from 11 was
coupled with tri-acid 8b to give the trimer 12 in 81%
yield. Repeating the above procedures of hydrogena-
tion, HBTU coupling and TFA deprotection gave the
final trimer CBL3-Lys(M-G1)-OH in 91% yield. The Z
group on 12 was then removed by hydrogenation and

1764
C.Sun et al./ Bioorg.Med.Chem.11 (2003) 1761–1768
Scheme 2. (a) HBTU, NMM, DMF; (b) (1) H2, 10% Pd/C, MeOH; (2) 8b, HBTU, NMM, DMF; (c) (1) H2, 10% Pd/C, MeOH; (2) 2, HBTU,
NMM, DMF; (d) 50% TFA, CH2Cl2.
the free amine was coupled to 4b to provide the non-
amer 15 in a convergent fashion. The nonamer 15 was
again hydrogenated, coupled with 2 and treated with
TFA to reveal the final nonameric CBL dendron CBL9-
Lys(M-G2)-OH. It is worth noting two synthetic aspects
in the preparation of these compounds. First, it was
previously observed that the introduction of the mal-
eimide moiety by coupling with the active ester 1 suf-
fered from lower yields as the compounds increased in
size. Therefore, the HBTU coupling method of the free
acid was adopted throughout and provided consistently
good yields. Second, the convergent methodology gave
us control of the homogeneity of the final nonamer.
Since this method worked well, we did not invoke the
divergent method as described for the original com-
pounds. We then examined the solubility of the new
dendrons and found all to be soluble to at least 2 mM in
phosphate buffer (pH 7.4)/10% DMSO, or >10-fold
more soluble than the original compounds (Scheme 2).
lysine greatly increased the water solubility of the com-
pounds.34 The method of using a dendritic linker mole-
cule to connect multiple drug molecules with scFvs has
not been previously reported. Yet, the strategy is not
limited to scFvs, since any antibody, including Fab
fragments and whole IgG, or other protein carriers
could be utilized which would effectively increase the
drug loading and the drug/carrier ratio. Furthermore,
we chose to attach chlorambucil since this drug is com-
mercially available, inexpensive and robust and allowed
facile demonstration of the concept as part of our anti-
cancer investigations. However, other drugs could be
used. Future studies will incorporate more potent anti-
cancer drugs, such as novel enediynes, and will be
reported in due course.
Experimental
General methods
1H and 13C NMR spectra were recorded on a Brucker
AMX-500 spectrometer at 500 MHz and 125 MHz, or
Bruker AMX-600 at 600 and 150 MHz, respectively.
Chemical shifts (ppm) were reported on the d scale
relative to internal CDCl3 (1H, 7.26 ppm and 13C,
77.0ppm) and acetone- d6 (1H, 2.04 ppm and 13C,
29.9 ppm). MS spectra were recorded using electrospray
ionization (ESI) or MALDI techniques. Glassware and
solvents were dried by standard methods. Flash chro-
matography was performed on Merck silica gel 60(230–
400 mesh) and thin-layer chromatography on glass
plates coated with a 0.25 mm (analytical) or 0.50 mm
(preparative) layer of silica gel 60F 254. Chlorambucil
and other chemical reagents and solvents were from
Aldrich Chem. Co. or Sigma, unless otherwise noted,
Conclusion
Recently, we reported the design and synthesis of three
compounds containing one, three and nine CBL drug
molecules at the branching termini and a maleimide
moiety at the core terminus of a dendron. With these
compounds we attempted to link them to our scFvs to
form drug/scFv immunoconjugates for use in cancer
therapy. Such conjugates would enable for the first time
the loading of an scFv with nine drug molecules. How-
ever, solubility problems with the dendrons precluded
conjugate formation. To solve this problem, we
designed and synthesized three new compounds with
lysine residues added at interior sites in the dendron
branches. The free carboxylic acid functionality on

C.Sun et al./ Bioorg.Med.Chem.11 (2003) 1761–1768
1765
and used without further purification. Compounds 4,35
5,36 8a,37 8b38 were synthesized according to literature
procedures. Reactions were carried out at room tem-
peratures unless otherwise noted.
resulting tri-amine (0.92 g, 94% yield) is a thick yellow
liquid, which was used without further purifications. To
a cooled (0 ꢀC) solution of chlorambucil (0.304 g,
1.0mmol) and 4-methylmorpholine (0.22 mL, 2.0mmol)
in 5 mL DMF was added HBTU (0.40 g, 1.05 mmol).
After 10min, the above crude tri-amine (0.128 g,
0.3 mmol) was added and the mixture was stirred over-
night. The solution was diluted with 10mL of EtOAc
and washed with water. The organic layer was then
dried over MgSO4 and evaporated. After chromato-
graphy over a silica gel column eluted with CHCl3/
MeOH/NH4OH (96/4/0.3), the product was obtained in
Preparation of compound 3. To a cooled (0 ꢀC) solution
of chlorambucil (0.61 g, 2.0 mmol) and 4-methylmor-
pholine (0.44 mL, 4.0 mmol) in 5 mL DMF was added
HBTU (0.80 g, 2.1 mmol). After 10 min, N-Boc-1,4-dia-
minobutane (0.38 g, 2.0 mmol) was added and the mix-
ture was stirred for 3 h. The solution was diluted with
20mL of EtOAc and washed with water. The organic
layer was then dried over MgSO4 and evaporated. After
chromatography over a silica gel column eluted with
CHCl3/MeOH (97/3–96/4), the product was obtained in
1
a yield of 0.28 g (72%) as a slight yellow powder. H
NMR (600 MHz, CDCl3) d 7.34–7.29 (m, 5H), 7.04 (d,
6H, J=8.3 Hz), 6.60(d, 6H, J=8.3 Hz), 6.09 (t, 3H,
NH, J=5.3 Hz), 5.00 (s, 2H), 3.69–3.59 (m, 30H), 3.45
(t, 6H, J=5.7 Hz), 3.28 (td, 6H, J=6.2, 5.3 Hz), 2.52 (t,
6H, J=7.5 Hz), 2.15 (t, 6H, J=7.5 Hz), 1.88 (tt, 6H,
J=7.5, 7.5 Hz), 1.71 (tt, 6H, J=5.7, 6.2 Hz); 13C NMR
(150MHz, CDCl 3) d 172.87, 155.25, 144.22, 136.37,
130.76, 129.60, 128.50, 128.15, 127.96, 112.16, 70.02,
69.26, 66.32, 58.63, 53.56, 40.49, 36.92, 35.94, 34.13,
1
a yield of 0.82 g (87%) as a white powder. H NMR
(500 MHz, CDCl3) d 7.05 (d, 2H, J=8.8 Hz), 6.60(d,
2H, J=8.8 Hz), 5.78 (s, 1H, NH), 4.67 (s, 1H, NH), 3.65
(m, 8H), 3.24 (dd, 2H, J=6.4, 12.3 Hz), 3.11 (m, 2H),
2.54 (t, 2H, J=7.5 Hz), 2.16 (t, 2H, J=7.5 Hz), 1.90(tt,
2H, J=7.5, 7.5 Hz), 1.49 (m, 4H), 1.42 (s, 9H); 13C
NMR (125 MHz, CDCl3) d 172.8, 156.1, 144.2, 130.6,
129.6, 112.1, 79.1, 53.5, 40.5, 40.0, 39.0, 35.9, 34.0, 28.3,
27.6, 27.4, 26.7; HRMS (MALDI) calcd for
C23H37N3O3Cl2 [M+Na]+ 496.2104, found 496.2116.
29.25,
27.46;
ESI-MS
(positive)
calcd
for
C63H89N7O8Cl6 [M+H]+ 1282.5, found 1282.3.
Preparation of compound 7. Compound 6 (0.25 g,
0.195 mmol) in MeOH (10 mL) was hydrogenated with
10% Pd/C (98 mg) using a hydrogen balloon. After
45 min, the reaction mixture was filtered through Celite
and the filter cake was washed with MeOH. The filtrate
was evaporated and the product (224 mg, 99% yield)
was used without further purifications. 1H NMR
(600 MHz, CDCl3) d 8.48 (m, NH, 3H), 7.04 (d, 6H,
J=8.3 Hz), 6.89 (s, NH, 2H), 6.59 (d, 6H, 8.3 Hz), 3.71–
3.39 (m, 42H), 2.52 (m, 6H), 2.24 (m, 6H), 1.87 (m, 6H),
1.71 (m, 6H); 13C NMR (150MHz, CDCl 3) d 173.70,
144.27, 130.85, 129.63, 112.19, 69.48, 68.22, 53.59,
40.61, 36.08, 35.70, 34.25, 29.47, 27.73; ESI-MS (posi-
tive) calcd for C55H83N7O6Cl6 [M+H]+ 1148.5, found
1148.7.
Preparation of M-CBL1
A solution of 3 (0.10 g, 0.21 mmol) in CH2Cl2 (2.0mL)
was treated with TFA (2 mL) at 0 ꢀC. After completion,
the mixture was evaporated with toluene (three times).
The residue was dissolved in 5 mL CH2Cl2 and 4-
methylmorpholine (55 mL, 0.50 mmol) was added fol-
lowed by 3-maleimidopropionic acid N-hydro-
xysuccinimide ester (compound 1, 80mg, 0.30mmol).
After 2 h stirring, the reaction mixture was evaporated
and the residue was purified by chromatography over a
silica gel column eluted with CHCl3/MeOH (97/3 to 95/
5), the product was obtained in a yield of 88 mg (80%)
1
as a white powder. H NMR (500 MHz, CDCl3) d 7.06
(d, 2H, J=8.8 Hz), 6.68 (s, 2H), 6.61 (d, 2H, J=8.8 Hz),
6.16 (s, 1H, NH), 5.80(s, 1H, NH), 3.82 (t, 2H,
J=7.0Hz), 3.65 (m, 8H), 3.24 (m, 4H), 2.52 (m, 4H),
2.17 (t, 2H, J=7.5 Hz), 1.90(tt, 2H, J=7.5, 7.5 Hz),
1.50(m, 4H); 13C NMR (133 MHz, CDCl3) d 173.15,
170.50, 169.81, 144.29, 134.19, 130.61, 129.64, 112.12,
53.54, 40.51, 39.04, 38.87, 35.96, 34.70, 34.31, 34.05,
27.41, 26.98, 26.50; HRMS (MALDI) calcd for
C25H34N4O4Cl2 [M+Na]+ 547.1849, found 547.1869.
Preparation of compound M-G1-CBL3. This compound
was prepared from 7 (60mg, 52 mmol) and 1 (42 mg,
156 mmol) following analogous procedures as in the
synthesis of M-CBL1. The product was purified by
chromatography over a silica gel column eluted with
CHCl3/MeOH (97/3–95/5) as a white powder in a yield
1
of 30mg (45%). H NMR (400 MHz, CDCl3) d 7.05 (d,
6H, J=8.8 Hz), 6.65 (s, 2H), 6.61 (d, 6H, J=8.8 Hz),
6.03 (t, 3H, NH, J=5.9 Hz), 5.30(s, 1H), 3.78 (t, 2H,
J=7.3 Hz), 3.71–3.59 (m, 30H), 3.44 (t, 6H, J=5.6 Hz),
3.31 (td, 6H, J=6.0, 6.0 Hz), 2.57–2.52 (m, 8H), 2.18 (t,
6H, J=7.5 Hz), 1.90(tt, 6H, J=7.5, 7.5 Hz); 13C NMR
(100 MHz, CDCl3) d 173.00, 170.43 (2), 144.30, 134.12,
130.73, 129.64, 112.16, 69.89, 68.72, 59.72, 53.58, 40.55,
36.57, 35.98, 35.08, 34.40, 34.14, 29.42, 27.50; HRMS
(MALDI) calcd for C62H88N8O9Cl6 [M+Na]+
1321.4697, found 1321.4714.
Preparation of compound 6. To a stirred methanol solu-
tion of the tri-nitrile compound 5 (0.95 g, 2.3 mmol) and
cobalt (II) chloride hexahydrate (3.28 g, 13.8 mmol) was
added NaBH4 (5.2 g, 137 mmol) in small portions (vio-
lent gas evolution). The mixture was stirred for 2 h and
then acidified with concentrated hydrochloric acid
(15 mL). The solvents were removed under vacuum and
the resulting deep blue residue was taken up in con-
centrated ammonia solution (20mL) and CHCl 3
(20mL). The insoluble solid was filtered off and the
aqueous phase was extracted with CHCl3 (3 Â 20mL).
The combined organic layer was dried over MgSO4,
evaporated and then co-evaporated with toluene. The
Preparation of compound 9a. To a cooled (0 ꢀC) solution
of 8a (34 mg, 78 mmol) and 4-methylmorpholine (51 mL,
0.47 mmol) in 1.5 mL DMF was added HBTU (99 mg,
0.26 mmol). After 30 min, 7 (0.30 g, 0.26 mmol) was

1766
C.Sun et al./ Bioorg.Med.Chem.11 (2003) 1761–1768
added and the mixture was stirred for 16 h. The solution
was diluted with 10mL of EtOAc and washed with
water. The organic layer was then dried over MgSO4
and evaporated. After chromatography over a silica gel
column eluted with CHCl3/MeOH (97/3 to 96/4), the
crude product was purified again on P-TLC developed
with CHCl3/MeOH (94/6). The product was obtained in
10% Pd/C (155 mg) using a hydrogen balloon. After 2 h,
the reaction mixture was filtered through Celite and the
filter cake was washed with MeOH. The filtrate was
evaporated and the product was dissolved in 2 mL
DMF and added to a DMF solution containing 8b
(0.10 g, 0.21 mmol), NMM (97 mL, 0.88 mmol) and
HBTU (0.33 g, 0.88 mmol). The mixture was stirred for
2 h, diluted with 50mL of EtOAc and washed with
water. The organic layer was then dried over MgSO4
and evaporated. After chromatography over a silica gel
column eluted with CHCl3/MeOH (97/3–96/4), the pro-
duct was obtained in a yield of 0.32 g (81%, two steps)
as a white powder. 1H NMR (600 MHz, CDCl3) d 7.35–
7.29 (m, 5H), 7.05 (d, 6H, J=8.8 Hz), 6.64 (t, 3H,
J=5.7 Hz, NH), 6.61 (d, 6H, J=8.8 Hz), 6.25 (d, 3H,
J=7.5 Hz), 5.40(s, 1H, NH), 5.01 (s, 2H), 4.42 (m, 3H),
3.70–3.60 (m, 36H), 3.19 (m, 6H), 2.54 (t, 6H,
J=7.5 Hz), 2.35 (t, 6H, J=5.9 Hz), 2.21 (t, 6H,
J=7.6 Hz), 1.89 (m, 6H), 1.75 (m, 3H), 1.63 (m, 3H),
1.51 (m, 6H), 1.45 (s, 27H), 1.32 (m, 6H); 13C NMR
(150MHz, CDCl 3) d 172.78, 171.71, 171.33, 163.11,
144.26, 130.55, 129.64, 128.49, 128.13, 127.96, 112.10,
82.01, 69.36, 67.35, 58.69, 53.54, 52.30, 40.51, 38.89,
36.63, 35.72, 34.04, 32.20, 28.94, 27.98, 27.38, 22.45;
HRMS (MALDI) calcd for C93H140N10O17Cl6
[M+Na]+ 1901.8421, found 1901.8363.
1
a yield of 0.19 g (64%) as a slightly yellow powder. H
NMR (400 MHz, CDCl3) d 7.04 (d, 18H, J=8.5 Hz),
6.59 (d, 18H, J=8.5 Hz), 3.67–3.58 (m, 102H), 3.44 (br-
t, 18H, J=5.3 Hz), 3.28 (m, 18H), 2.52 (br-t, 18H,
J=7.5 Hz), 2.40(br-t, 6H), 2.18 (t, 18H, J=7.4 Hz),
1.89 (m, 18H), 1.71 (m, 18H), 1.40(s, 9H); 13C NMR
(100 MHz, CDCl3) d 173.09, 172.94, 171.32, 154.89,
144.26, 130.70, 129.59, 112.14, 70.01, 69.71, 68.89,
67.57, 59.79, 58.62, 53.54, 40.53, 37.43, 36.64, 35.96,
34.20, 29.41, 28.44, 27.58; ESI-MS (positive) calcd for
C173H258N22O22Cl18 Mave 3836.4, found 3835.7.
Preparation of compound M-G2-CBL9. A solution of 9a
(15 mg, 3.9 mmol) in CH2Cl2 (0.5 mL) was treated with
TFA (0.5 mL) at 0 ꢀC. After completion, the mixture
was evaporated with toluene (three times). The residue
was dissolved in 0.5 mL DMF and added to a DMF
solution containing 2 (1 mg, 5.9 mmol), NMM (1 mL,
10 mmol) and HBTU (2.2 mg, 5.9 mmol). The mixture
was stirred for 3 h, diluted with 10mL of EtOAc and
washed with water. The organic layer was then dried
over MgSO4 and evaporated. After P-TLC developed
with CHCl3/MeOH (94/6), the product was obtained in
Preparation of compound 13. This compound was pre-
pared from 11 (0.25 g, 0.40 mmol) and 2 (74 mg,
0.44 mmol) following analogous procedures as in the
synthesis of 12. The product was purified by chromato-
graphy over a silica gel column eluted with CHCl3/
MeOH (98/2) as a white powder in a yield of 227 mg
1
a yield of 13 mg (85%) as a slightly yellow powder. H
NMR (600 MHz, CDCl3) d 7.04 (d, 18H, J=8.3 Hz),
6.66 (s, 2H), 6.61 (d, 18H, J=8.3 Hz), 6.45 (s, 9H, NH),
3.75 (t, 2H, J=7.5 Hz), 3.69–3.59 (m, 102H), 3.43 (t,
18H, J=5.7 Hz), 3.28 (td, 18H, J=6.1, 5.7 Hz), 2.52 (m,
20H), 2.40 (br-t, 6H), 2.18 (t, 18H, J=7.5 Hz), 1.89 (tt,
18H, J=7.5, 7.5 Hz), 1.71 (m, 18H); ESI-MS (positive)
calcd for C185H271N23O27Cl18 Mave 3887.4, found
3887.3.
1
(89%, two steps). H NMR (500 MHz, CDCl3) d 7.05
(d, 2H, J=8.4 Hz), 6.68 (s, 2H), 6.61 (d, 2H, J=8.4 Hz),
6.11 (d, 1H, J=8.1 Hz, NH), 5.98 (br-t, 1H, J=5.9 Hz,
NH), 4.45 (m, 1H), 3.80(t, 2H, J=7.0Hz), 3.71–3.60
(m, 8H), 3.19 (m, 2H), 2.55 (t, 2H, J=7.5 Hz), 2.49 (t,
2H, J=7.0Hz), 2.22 (t, 2H, J=7.5 Hz), 1.91 (tt, 2H,
J=7.5, 7.5 Hz), 1.88 (m, 1H), 1.61 (m, 1H), 1.51 (m,
2H), 1.45 (s, 9H), 1.33 (m, 2H); 13C NMR (125MHz,
CDCl3) d 172.73, 171.72, 170.46, 169.67, 144.24, 134.15,
130.54, 129.63, 112.09, 82.16, 77.25, 77.19, 76.99, 76.74,
53.53, 52.00, 40.48, 39.06, 35.79, 34.58, 34.25, 33.97, 32.51,
28.52, 27.96, 27.32, 22.33; HRMS (MALDI) calcd for
C31H44N4O6Cl2 [M+Na]+ 661.2533, found 661.2539.
Preparation of compound 11. This compound was pre-
pared from CBL (1.0g, 3.3 mmol) and H-Lys(Z)-O t-Bu
hydrochloride (1.1 g, 3.3 mmol) following analogous
procedures as in the synthesis of 3. The product was
purified by chromatography over a silica gel column
eluted with CHCl3/MeOH (98/2) as a white powder in a
1
yield of 1.95 g (96%). H NMR (600 MHz, CDCl3) d
7.99 (s, 1H, NH), 7.34–7.27 (m, 5H), 7.04 (d, 2H,
J=8.3 Hz), 6.59 (d, 2H, J=8.3 Hz), 6.10(d, 1H,
J=7.9 Hz, NH), 5.04 (s, 2H), 4.46 (m, 1H), 3.68–3.58
(m, 8H), 3.16 (td, 2H, J=6.6, 6.6 Hz), 2.52 (t, 2H,
J=7.5 Hz), 2.19 (t, 2H, J=7.5 Hz), 1.89 (tt, 2H, J=7.5,
7.5 Hz), 1.79 (m, 1H), 1.60(m, 1H), 1.51 (m, 2H), 1.44
(s, 9H), 1.35 (m, 2H); 13C NMR (150MHz, CDCl 3) d
172.47, 171.69, 162.43, 156.41, 144.16, 136.52, 130.54,
129.58, 128.39, 127.95, 127.94, 112.02, 81.97, 66.41,
53.47, 52.12, 40.49, 40.43, 38.49, 36.37, 35.66, 33.92,
32.18, 31.31, 29.28, 27.89, 27.86, 27.22, 22.12; HRMS
(MALDI) calcd for C32H45N3O5Cl2 [M+Na]+
644.2628, found 644.2637.
Preparation of CBL1-Lys(M)-OH. A solution of 13
(50mg, 78 mmol) in CH2Cl2 (2.0mL) was treated with
TFA (2 mL) at 0 ꢀC. After completion, the mixture was
evaporated with toluene (three times). The product is a
slightly yellow powder, which is used without further
1
purifications. H NMR (500 MHz, acetone-d6) d 7.32–
7.28 (m, 2H, NH), 7.07 (d, 2H, J=8.8 Hz), 6.83 (s, 2H),
6.70(d, 2H, J=8.8 Hz), 4.41 (m, 1H), 3.78–3.70(m,
10H), 3.15 (m, 2H), 2.53 (t, 2H, J=7.5 Hz), 2.43 (t, 2H,
J=7.3 Hz), 2.25 (t, 2H, J=7.5 Hz), 1.89–1.80(m, 3H),
1.69 (m, 1H), 1.51–1.36 (m, 4H); 13C NMR (125 MHz,
acetone-d6) d 173.84, 173.53, 171.35, 170.49, 145.43,
135.17, 131.59, 130.35, 113.04, 53.88, 52.73, 41.57,
39.30, 35.75, 35.19, 35.08, 34.72, 31.92, 23.64; HRMS
(MALDI) calcd for C27H36N4O6Cl2 [M+Na]+
605.1904, found 605.1908.
Preparation of compound 12. Compound 11 (0.50 g,
0.80 mmol) in MeOH (15 mL) was hydrogenated with

C.Sun et al./ Bioorg.Med.Chem.11 (2003) 1761–1768
1767
Preparation of compound 14. This compound was pre-
pared from 12 (0.14 g, 74 mmol) and (19 mg,
silica gel column eluted with CHCl3/MeOH (95/5) as a
slightly yellow powder in a yield of 122 mg (74%, two
steps). 1H NMR (600 MHz, CDCl3) d 7.13 (s, 6H, NH),
7.04 (d, 18H, J=8.3 Hz), 6.68 (s, 2H), 6.60(d, 18H,
J=8.3 Hz), 6.50(m, 9H, NH), 4.39 (m, 9H), 3.77 (t, 2H,
J=7.0Hz), 3.69–3.59 (m, 120H), 3.18 (m, 18H), 2.54–
2.51 (m, 20H), 2.36 (m, 24H), 2.22 (t, 18H, J=7.5 Hz),
1.88 (m, 18H), 1.76 (m, 9H), 1.64 (m, 9H), 1.51 (m,
18H), 1.44 (s, 81H), 1.32 (m, 18H); 13C NMR
(150MHz, CDCl 3) d 172.89, 171.75, 171.50, 171.45,
170.52, 170.48, 144.26, 134.24, 130.56, 129.63, 112.10,
81.93, 69.05, 67.51, 67.39, 59.94, 59.77, 53.52, 52.51,
40.53, 38.93, 36.50, 35.70, 35.02, 34.30, 34.07, 32.08,
29.09, 27.98,27.42, 22.60; ESI-MS (positive) calcd for
C275H424N32O54Cl18 Mave 5680.6, found 5681.9.
2
0.11 mmol) following analogous procedures as in the
synthesis of 12. The product was purified by chromato-
graphy over a silica gel column eluted with CHCl3/
MeOH (97/3 to 96/4) as a white powder in a yield of
1
0.13 g (92%, two steps). H NMR (600 MHz, CDCl3) d
7.05 (d, 6H, J=7.9 Hz), 6.68 (s, 2H), 6.64 (t, 3H,
J=5.7 Hz, NH), 6.61 (d, 6H, J=7.9 Hz), 6.52 (s, 1H,
NH), 6.30(d, 3H, J=7.9 Hz, NH), 4.42 (m, 3H), 3.76 (t,
2H, J=7.0Hz), 3.70–3.60(m, 36H), 3.21 (m, 6H), 2.55–
2.50(m, 8H), 2.36 (t, 6H, J=5.7 Hz), 2.22 (t, 6H,
J=7.5 Hz), 1.90(m, 6H), 1.79 (m, 3H), 1.64 (m, 3H),
1.52 (m, 6H), 1.45 (s, 27H), 1.36 (m, 6H); 13C NMR
(150MHz, CDCl 3) d 172.79, 171.72, 171.46, 170.50
170.25, 144.27, 134.18, 130.54, 129.64, 112.09, 82.00,
69.35, 67.36, 59.63, 53.53, 52.33, 40.52, 38.97, 36.49,
35.74, 35.03, 34.33, 34.04, 32.21, 28.98, 27.98, 27.38,
22.49; HRMS (MALDI) calcd for C92H139N11O18Cl6
[M+Na]+ 1918.8322, found 1918.8321.
Preparation of CBL9-Lys(M-G2)-OH. This compound
was prepared from 16 (50mg, 8.8 mmol) following analo-
gous TFA deprotection procedures as in the synthesis of
CBL1-Lys(M)-OH. The product is a slightly yellow pow-
der, which is used without further purifications. 1H NMR
(500 MHz, acetone-d6) d 10 .34 (br-s, 33H, COOH), 7.0 6 (d,
18H, J=8.8 Hz), 6.85 (s, 2H), 6.69 (d, 18H, J=8.8 Hz),
4.46 (m, 9H), 3.74–3.69 (m, 122H), 3.24 (m, 18H), 2.54–
2.47 (m, 44H), 2.31 (t, 18H, J=7.5 Hz), 1.89–1.86 (m,
27H), 1.76 (m, 9H), 1.58–1.45 (m, 36H); 13C NMR
(125 MHz, acetone-d6) d 174.51, 174.08, 173.05, 172.98,
172.91,171.59, 158.99, 158.66, 158.34, 158.02, 145.42,
135.32, 131.55, 130.41, 119.31, 117.04, 114.77, 113.12,
112.50, 69.84, 68.54, 68.39, 61.11, 61.01, 53.94, 52.99,
41.64, 39.74, 37.78, 37.17, 35.90, 34.84, 31.98, 31.96,
28.66, 27.54, 23.80; HRMS (MALDI) calcd for
C239H352N32O54Cl18 Mave 5175.7, found 5175.2.
Preparation of CBL3-Lys(M-G1)-OH. This compound
was prepared from 14 (50mg, 26 mmol) following analo-
gous TFA deprotection procedures as in the synthesis of
CBL1-Lys(M)-OH. The product is a slightly yellow pow-
der, which is used without further purifications. 1H NMR
(500MHz, acetone-d6) d 11.45 (br-s, 13H, COOH), 7.81–
7.73 (m, 6H, NH), 7.07 (d, 6H, J=8.6 Hz), 6.82 (s, 2H),
6.72 (d, 6H, J=8.6 Hz), 4.46 (m, 3H), 3.77–3.66 (m,
38H), 3.26 (m, 6H), 2.55–2.46 (m, 14H), 2.33 (t, 6H,
J=7.5 Hz), 1.91–1.87 (m, 9H), 1.74 (m, 3H), 1.59–1.42
(m, 12H); 13C NMR (125 MHz, acetone-d6) d 175.07,
173.73, 173.60, 171.81, 171.43, 159.18, 158.86, 158.54,
158.22, 145.25, 135.19, 131.67, 130.38, 119.38, 117.10,
114.83, 113.26, 112.55, 69.93, 68.16, 64.68, 60.98, 54.01,
53.07, 41.51, 39.92, 36.94, 35.84, 35.75, 35.07, 34.77,
31.86, 30.54, 28.67, 23.67; MS (MALDI-TOF) calcd for
C23H37N3O3Cl2 [M+H]+ 1729, found 1729.
Acknowledgements
This work was supported in part by grants from the
NIH NIAID-AI47127, California Cancer Research
Program (00-00757V-20012) and the Skaggs Institute
for Chemical Biology.
Preparation of compound 15. This compound was pre-
pared from 12 (0.88 g, 0.47 mmol) and 8b (61 mg,
0.13 mmol) following analogous procedures as in the
synthesis of 12. The product was purified by chromato-
graphy over a silica gel column eluted with CHCl3/
MeOH (97/3 to 95/5) as a white powder in a yield of
References and Notes
1
1. Huston, J. S.; Levinson, D.; Mudgett-Hunter, M.; Tai,
M. S.; Novotny, J.; Margolies, M. N.; Ridge, R. J.; Bruccoleri,
R. E.; Haber, E.; Crea, R.; Oppermann, H. Proc.Natl.Acad.
Sci. U.S.A. 1988, 85, 5879.
2. Bird, R. E.; Hardman, K. D.; Jacobson, J. W.; Johnson, S.;
Kaufman, B. M.; Lee, S. M.; Lee, T.; Pope, S. H.; Riordan,
G. S.; Whitlow, M. Science 1988, 242, 423.
3. Jain, R. K. Cancer Metastasis Rev. 1990, 9, 253.
4. Milenic, D. E.; Yokota, T.; Filpula, D. R.; Finkelman,
M. A.; Dodd, S. W.; Wood, J. F.; Whitlow, M.; Snoy, P.;
Schlom, J. Cancer Res. 1991, 51, 6363.
5. Mao, S. L.; Gao, C. S.; Lo, C. H. L.; Wirsching, P.; Wong,
C. H.; Janda, K. D. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
6953.
6. Lee, K. J.; Mao, S.; Sun, C.; Gao, C.; Blixt, O.; Arrues, S.;
Hom, L. G.; Kaufmann, G. F.; Hoffman, T. Z.; Coyle, A. R.;
Paulson, J.; Habermann, B. F.; Janda, K. D. J.Am.Chem.
Soc. 2002, 124, 12439.
7. Trail, P. A.; Bianchi, A. B. Curr.Opin.Immunol. 1999, 11,
584.
0.43 g (58%, two steps). H NMR (600 MHz, CDCl3) d
7.32–7.30(m, 5H), 7.04 (d, 18H, J=8.5 Hz), 6.60(d,
18H, J=8.5 Hz), 6.48 (d, 6H, J=7.9 Hz), 5.01 (s, 2H),
4.40(m, 9H), 3.69–3.59 (m, 120H), 3.18 (m, 18H), 2.52
(t, 18H, J=7.5 Hz), 2.35 (m, 24H), 2.22 (t, 18H,
J=7.5 Hz), 1.88 (m, 18H), 1.76 (m, 9H), 1.64 (m, 9H),
1.51 (m, 18H), 1.44 (s, 81H), 1.32 (m, 18H); 13C NMR
(150MHz, CDCl 3) d 172.76, 171.62, 171.27, 162.38,
144.15, 130.48, 129.50, 128.35, 127.69, 111.99, 81.71,
68.97, 67.41, 67.27, 59.62, 58.82, 53.40, 52.41, 40.44,
38.82, 36.34, 35.55, 33.96, 31.91, 31.26, 29.00, 27.87,
27.32,
C276H425N31O53Cl18 Mave 5663.7, found 5664.7.
22.51;
ESI-MS
(positive)
calcd
for
Preparation of compound 16. This compound was pre-
pared from 15 (0.16 g, 28 mmol) and 2 (7.3 mg, 43 mmol)
following analogous procedures as in the synthesis of
12. The product was purified by chromatography over a

1768
C.Sun et al./ Bioorg.Med.Chem.11 (2003) 1761–1768
8. Garnett, M. C. Adv.Drug Deliv.Rev. 2001, 53, 171.
9. Chari, R. V. J. Adv.Drug Deliv.Rev. 1998, 31, 89.
10. Funaro, A.; Horenstein, A. L.; Santoro, P.; Cinti, C.;
Gregorini, A.; Malavasi, F. Biotechnol.Adv. 2000, 18, 385.
11. Senter, P. D.; Springer, C. J. Adv.Drug Deliv.Rev. 2001,
53, 247.
12. Denny, W. A. Eur.J.Med.Chem. 2001, 36, 577.
13. Trail, P. A.; Willner, D.; Lasch, S. J.; Henderson, A. J.;
Greenfield, R. S.; King, D.; Zoeckler, M. E.; Braslawsky, G. R.
Cancer Res. 1992, 52, 5693.
K.; Frey, H.; Weener, J. W.; Meijer, E. W.; Paulus, W.; Dun-
can, R. J.Controlled Release 2000, 65, 133.
25. De Jesus, O. L. P.; Ihre, H. R.; Gagne, L.; Frechet, J. M. J.;
Szoka, F. C., Jr. Bioconjugate Chem. 2002, 13, 453.
26. Ihre, H. R.; De Jesus, O. L. P.; Szoka, F. C., Jr.; Frechet,
J. M. J. Bioconjugate Chem. 2002, 13, 443.
27. Sun, C.; Wirsching, P.; Janda, K. D. Bioorg.Med.Chem.
Lett. 2002, 12, 2213.
28. Unpublished results from our laboratory.
29. Zalipsky, S., Harris, J. M. In Poly(ethylene glycol)
Chemistry and Biological Applications; J. M. Harris, S.
Zalipsky, Eds.; American Chemical Society: Washington, DC,
1997; p 1.
14. King, H. D.; Yurgaitis, D.; Willner, D.; Firestone, R. A.;
Yang, M. B.; Lasch, S. J.; Hellstrom, K. E.; Trail, P. A. Bio-
conjugate Chem. 1999, 10, 279.
15. Newkome, G. R.; Moorefield, C. N.; Vogtle, F. Dendritic
Molecules: Concepts, Syntheses, Perspectives; VCH: Wein-
heim, Germany, 1986.
30. Baek, M. G.; Roy, R. Bioorg.Med.Chem. 2001, 9, 3005.
31. Tam, J. P.; Lu, Y. A.; Yang, J. L. Eur.J.Biochem. 2002,
269, 923.
16. Frechet, J. M. J.; Tomalia, D. A. Dendrimers and Other
Dendritic Polymers; Wiley: Chichester, UK, 2001.
17. Esfand, R.; Tomalia, D. A. Drug Discov.Today 2001, 6, 427.
18. Jansen, J. F. G. A.; de Brabander van den Berg, E.; Mei-
jer, E. W. Science 1994, 266, 1226.
19. Hawker, C. J.; Wooley, K. L.; Frechet, J. M. J. J.Chem.
Soc., Perkin Trans. 1 1993, 12, 1287.
20. Hecht, S.; Frechet, J. M. J. Angew.Chem,. Int.Ed.Engl.
2001, 40, 74.
21. Patri, A. K.; Majoros, I. J.; Baker, J. R. Curr.Opin.Chem.
Biol. 2002, 6, 466.
22. Liu, M. J.; Frechet, J. M. J. Pharm.Sci.Technol.Today
1999, 2, 393.
32. Higashi, N.; Koga, T.; Niwa, M. Chembiochemistry 2002,
3, 448.
33. Twyman, L.; Beezer, A. E.; Mitchell, J. C. Tetrahedron
Lett. 1994, 35, 4423.
34. Our preliminary results indicate that conjugation with
scFvs is successful.
35. Newkome, G. R.; Lin, X. F. Macromolecules 1991, 24,
1443.
36. Lebreton, S.; Newcombe, N.; Bradley, M. Tetrahedron
Lett. 2002, 43, 2479.
37. Takahashi, M.; Hara, Y.; Aoshima, K.; Kurihara, H.;
Oshikawa, T.; Yamashita, M. Tetrahedron Lett. 2000, 41,
8485.
23. Liu, M. J.; Kono, K.; Frechet, J. M. J. J.Controlled
Release 2000, 65, 121.
24. Malik, N.; Wiwattanapatapee, R.; Klopsch, R.; Lorenz,
38. Nierengarten, J. F.; Habicher, T.; Kessinger, R.; Cardullo,
F.; Diederich, F.; Gramlich, V.; Gisselbrecht, J. P.; Boudon,
C.; Gross, M. Helv.Chim.Acta 1997, 80, 2238.