Carbohydrate-based synthetic approach to control toxicity profiles of folate-drug conjugates

Article abstract of DOI:10.1021/jo100448q

Figure presented To better regulate the biodistribution of the vinblastine-folate conjugate, EC145, a new folate-spacer that incorporates 1-amino-1-deoxy-d-glucitol-γ-glutamate subunits into a peptidic backbone, was synthesized. Synthesis of Fmoc-3,4;5,6-di-O-isopropylidene-1-amino-1-deoxy- d-glucitol-γ-glutamate 20, suitable for Fmoc-strategy solid-phase peptide synthesis (SPPS), was achieved in four steps from δ-gluconolactone. Addition of alternating glutamic acid and 20 moieties onto a cysteine-loaded resin, followed by the addition of folate, deprotection, and cleavage, resulted in the isolation of the new folate-spacer: Pte-γGlu-(Glu(1-amino-1-deoxy- d-glucitol)-Glu)₂-Glu(1-amino-1-deoxy-d-glucitol)-Cys-OH (21). The addition of 21 to an appropriately modified desacetylvinblastine hydrazide (DAVLBH) resulted in a conjugate (25) with an improved therapeutic index. Treatment of 25 with DTT in neutral buffer at room temperature demonstrated that free DAVLBH would be released under the reductive environment of the internalized endosome.

Full text of DOI:10.1021/jo100448q

pubs.acs.org/joc
Carbohydrate-Based Synthetic Approach to Control Toxicity Profiles
of Folate-Drug Conjugates
Iontcho R. Vlahov,* Hari Krishna R. Santhapuram, Fei You, Yu Wang, Paul J. Kleindl,
Spencer J. Hahn, Jeremy F. Vaughn, Daniel S. Reno, and Christopher P. Leamon
Endocyte Inc., 3000 Kent Avenue, Suite A1-100, West Lafayette, Indiana 47906
ivlahov@endocyte.com
Received March 10, 2010
To better regulate the biodistribution of the vinblastine-folate conjugate, EC145, a new folate-
spacer that incorporates 1-amino-1-deoxy-D-glucitol-γ-glutamate subunits into a peptidic back-
bone, was synthesized. Synthesis of Fmoc-3,4;5,6-di-O-isopropylidene-1-amino-1-deoxy-D-glucitol-
γ-glutamate 20, suitable for Fmoc-strategy solid-phase peptide synthesis (SPPS), was achieved in
four steps from δ-gluconolactone. Addition of alternating glutamic acid and 20 moieties onto a
cysteine-loaded resin, followed by the addition of folate, deprotection, and cleavage, resulted in the
isolation of the new folate-spacer: Pte-γGlu-(Glu(1-amino-1-deoxy-D-glucitol)-Glu)2-Glu(1-amino-
1-deoxy-D-glucitol)-Cys-OH (21). The addition of 21 toan appropriately modified desacetylvinblastine
hydrazide (DAVLBH) resulted in a conjugate (25) with an improved therapeutic index. Treatment of
25 with DTT in neutral buffer at room temperature demonstrated that free DAVLBH would be
released under the reductive environment of the internalized endosome.
Introduction
toxicity for EC145 was found to be related to ileus, albeit in
heavily pretreated patients.3 Because of the low to undetect-
able levels of the folate receptor (FR) expressed in hepatic
tissue,4 it was suspected that non-FR related liver clearance,
with subsequent metabolic release of free DAVLBH, may
have been (in part) responsible for producing the observed
gastrointestinal toxicity. Support for our hypothesis came
from (i) our ability to block the liver uptake of folate
conjugates with coinjected bromosulfophthalein (an organic
anion transporter inhibitor) and (ii) the detection of free
DAVLBH in the bile of rats that had been intravenously
dosed with EC145.
Receptor-targeted chemotherapy has emerged as one of
the major approaches in modern drug discovery as it can
potentially satisfy the selective delivery criteria for toxic
agents to pathologic cells. In a previous publication,1 we
reported the design and synthesis of a folate-targeted desa-
cetylvinblastine hydrazide (DAVLBH) conjugate, EC145,
which is currently in phase 2 clinical trials. In EC145, the
anticancer drug vinblastine was modified and attached to
folic acid (FA) via a highly charged water-soluble peptide-
based spacer unit (1, Figure 1) and a self-immolative linker
system containing a reducible disulfide bond. Once adminis-
tered, the conjugate EC145 targets cancer cells which over-
express folate receptor (FR) and releases the base drug after
internalization.2 Receptor-targeted delivery reduces collat-
eral toxicity, resulting in improved efficacy. The dose-limiting
Work by Suzuki et al. on alkylglycoside-derivatized arginine
vasopressin (AVP) suggested a method to modify the biodistri-
bution.5 When AVP is modified, particularly by specific glu-
copyranosyl, mannopyranosyl, and 2-deoxyglucopyranosyl
(3) Li, J.; Sausville, E. A.; Klein, P. J.; Morgenstern, D.; Leamon, C. P.;
Messmann, R. A.; LoRusso, P. J. Clin. Pharm. 2009, 49, 1467.
(4) Parker, N.; Turk, M. J.; Westrick, E.; Lewis, J. D.; Low, P. S.;
Leamon, C. P. Anal. Biochem. 2005, 338, 284.
(5) Reviewed in: Molema, G.; Meijer, D. K. F. Drug Targeting: Organ
Specific Strategies; Wiley-VCH: New York, 2001; Vol. 12, p 126. Suzuki, K.;
Susaki, H.; Okuno, S.; Yamada, H.; Watanabe, H. K.; Sugiyama, Y. J.
Pharmacol. Exp. Ther. 1999, 288, 888.
(1) Vlahov, I. R.; Santhapuram, H. K.; Kleindl, P. J.; Howard, S. J.;
Stanford, K. M.; Leamon, C. P. Bioorg. Med. Chem. Lett. 2006, 16, 5093.
(2) Leamon, C. P.; Reddy, J. A. Adv. Drug Delivery Rev. 2004, 56, 1127.
Reddy, J. A.; Dorton, R.; Westrick, E.; Dawson, A.; Smith, T.; Xu, L. C.;
Vetzel, M.; Kleindl, P. J.; Vlahov, I. R.; Leamon, C. P. Cancer Res. 2007, 67,
4434. Yang, J.; Chen, H.; Cheng, J.-X.; Vlahov, I. R.; Low, P. S. Proc. Natl.
Acad. Sci. U.S.A. 2006, 103, 13872.
DOI: 10.1021/jo100448q
r
Published on Web 04/28/2010
J. Org. Chem. 2010, 75, 3685–3691 3685
2010 American Chemical Society

Vlahov et al.
JOCArticle
FIGURE 1. FA-spacer units.
SCHEME 1. Synthesis of 3,4;5,6-Di-O-isopropylidene-2-deoxy-2-(Fmoc-amino)-D-mannonic Acid
moieties, a significant shift in the partitioning of the pro-drug
toward the kidneys is observed. Folate conjugates are known
to have high kidney uptake with little associated toxicity.4 A
folate conjugate with carbohydrate character seemed to be a
promising synthetic target. Such an approach could modulate
the gastrointestinal toxicity.
standard fluorenylmethyloxycarbonyl-based solid-phase pep-
tide synthesis (Fmoc SPPS) techniques. The structural design
of the saccharo-amino acid was of crucial importance to the
successful execution of a simple and high-yielding SPPS pro-
tocol. The selected approach relied on a ring-opened carbohy-
drate chain rather than on cyclic pyranoside or furanoside
units. Pyranosides and furanosides are of limited usefulness for
a simplified SPPS-based protocol because they require inher-
ently tedious protecting/deprotecting strategies, utilize sensi-
tive glycosyl donors, and generate diastereomeric glycosidic
linkages. Therefore, we designed and synthesized uniformly
protected 2-amino-2-deoxyglyconic acid derivatives that are
orthogonally stable to SPPS. Thus, treatment of commercially
available and inexpensive D-glucono-1,5-lactone 3 (Scheme 1)
with 2,2-dimethoxypropane in acetone under mild acidic con-
ditions resulted in the well-known methyl 3,4;5,6-di-O-isopro-
pylidene-D-gluconate 4.6 According to Csuk and Vasella,7
Results and Discussion
In this paper, we report a novel, chemistry-based approach
to substantially decrease the hepatobiliary route of clearance
of free DAVLBH without affecting the conjugate’s targeted
antitumor activity. Thus, selective placement of structurally
optimized carbohydrate segments in the spacer region re-
sulted in conjugates that were equipotent but less toxic than
EC145. Our first generation carbohydrate-containing FA-
spacer unit 2 was designed to be bifunctional, containing
alternately repeating acidic (Asp) and saccharo-amino acids,
thus providing high water-solubility of the final drug conjugate
under physiological conditions. This unit was assembled using
(6) Redeling, H.; de Rouville, E.; Chittenden, G. Recl. Trav. Chim. Pays-
Bas 1987, 106, 461.
3686 J. Org. Chem. Vol. 75, No. 11, 2010

Vlahov et al.
JOCArticle
SCHEME 2. Synthesis of First Generation Carbohydrate-Based FA-Spacer Unit 2a
aReagents and conditions: (i) 20% piperidine, DMF; (ii) 8, PyBop, DIPEA, DMF; (iii) Fmoc-Asp(O-t-Bu)-OH, PyBop, DIPEA, DMF; (iv) Fmoc-Glu-O-t-
Bu, PyBop, DIPEA, DMF; (v) N10-TFA-Pte-OH, PyBop, DIPEA, DMSO, DMF; (vi) 2% NH2NH2 in DMF; (vii) TFA, water, HSCH2CH2SH, i-Pr3SiH.
SCHEME 3. Optimizing the Carbohydrate-Based FA-Spacer Unit
this compound was transformed via triflate into manno-
azide 5. Reduction with Bu3SnH8 followed by hydrolysis
of the methyl ester9 and introduction of an Fmoc group
resulted in protected saccharo-amino acid derivative 8.
(7) Csuk, R.; Vasella, A. Helv. Chem. Acta 1988, 71, 609.
(8) Vlahov, I. R.; Vlahova, P. I.; Schmidt, R. R. Tetrahedron Lett. 1992,
33, 7503.
(9) Lee, S. G.; Park, K. H.; Yoon, Y. J. Heterocycl. Chem. 1998, 35,
711.
J. Org. Chem. Vol. 75, No. 11, 2010 3687

Vlahov et al.
JOCArticle
SCHEME 4. Synthesis of Fmoc-3,4;5,6-di-O-isopropylidene-1-amino-1-deoxy-D-glucitol-γ-glutamate
SCHEME 5. Synthesis of the Second Generation Carbohydrate-Based Folate-Spacer Unita
aReagents and conditions: (i) 20% piperidine, DMF; (ii) 20, PyBop, DIPEA, DMF; (iii) Fmoc-Glu(O-t-Bu)-OH, PyBop, DIPEA, DMF; (iv) Fmoc-Glu-
O-t-Bu, PyBop, DIPEA, DMF; (v) N10-TFA-Pte-OH, PyBop, DIPEA, DMSO, DMF; (vi) 2% NH2NH2 in DMF; (vii) TFA, water, HSCH2CH2SH,
i-Pr3SiH.
The newly designed base- and acid-sensitive compound 8
allowed for the assembly of the FA-spacer unit 2 using
standard Fmoc SPPS on a Wang-resin polymeric support
(Scheme 2). Pteroic acid10 served as the N-terminus, whereas
the thiol group of cysteine (Cys) was selected as the attach-
ment site for the self-immolative linker system.
After cleavage from the solid support, the LC-MS profile
of the reaction mixture displayed a major peak and several
satellite peaks with very close retention times. All possessed
molecular masses identical with that expected for structure 2.
We also observed an additional peak which mass indicated
was the expected structure 2 lacking the C-terminal Cys.
These results were attributed to three major problems related
to the molecular architecture of FA-spacer unit 2. Asparti-
mide formation (Scheme 3, route 1, 9) is a frequently
encountered side reaction affecting Asp residues during
Fmoc SPPS.11 In the presence of a strong base such as
piperidine, used to remove the Fmoc group, the R-amide
nitrogen atom attacks the carbonyl group of the β-carboxy
side chain resulting in ring closure with loss of the ester
protecting group. Aspartimides of type 9 are susceptible to
base-catalyzed epimerization to 10, and both of these forms
easily experience ring-opening with predominant formation
(10) Xu, L.; Vlahov, I. R.; Leamon, C. P.; Santhapuram, H. K. R.; Li, C.
U.S. Patent 009153, 2006.
(11) Nicolas, E.; et al. Tetrahedron Lett. 1989, 30, 497.
3688 J. Org. Chem. Vol. 75, No. 11, 2010

Vlahov et al.
SCHEME 6. Application of Optimized Carbohydrate-Based Spacer for the Synthesis of a Releasable Folate-DAVLBH Conjugate
JOCArticle
of β-aspartyl peptide chain 11 in the presence of water.
Compound 2 and all epimerized and β-rearranged counter
partners possess identical masses and elute closely or coelute
with the desired R-isomer on HPLC. In our first generation
spacer design, the peptide sequence possessed multiple sites
for potential aspartimide formation making this problem
even more serious. Therefore, we exchanged all aspartic acid
residues for glutamic acids. To our surprise, there were still
peaks in the LC-MS with identical masses and close reten-
tion times. That they exhibited different fragmentation
patterns in the mass-spectrum brought us to the indirect
conclusion that, most likely, we were dealing not only with
rotameric forms but also with some other sort of closely
related structural isomers. In compound 2 or in its isopro-
pylidene-protected precursor, the protons attached to C-2-
atoms in the saccharo-amino acid moieties in 2 are suffi-
ciently acidic to readily undergo epimerization (Scheme 3,
route 2, 14) under the strong basic and/or acidic reaction
conditions of SPPS protocols.12 “Peeling” of the C-terminal
Cys (Scheme 3, route 3) is a side reaction initiated by 5- and/
or 6-membered ring closure, resulting in energetically
stabilized lactone forms (e.g., 15).13 Apparently, in our
first-generation design of the saccharo-amino acid we ig-
nored this not so obvious problem and suitably positioned a
hydroxyl group in the carbohydrate chain that attacked the
carboxy amide bond of C-terminal Cys and formed lactones.
To address and avoid epimerization and “peeling” issues,
we redesigned the structure of our saccharo-amino acid
by connecting via an amide bond two epimerization-inert
modules, namely glutamic acid and glucamine. Each possess
C-H bonds that are not inherently prone to changes of con-
figuration. Also, in the new FA-spacer unit 16, all hydroxyl
groups in the glucamine module are more than seven atoms
away from the amide bond connecting the saccharo-amino
acid and the C-terminal Cys, thus suppressing lactone for-
mation and the resulting “peeling”. Diisopropylidene- and
Fmoc-protected saccharo-amino acid 20 is orthogonally
stable to the reaction conditions used in classical Fmoc SPPS
protocols. This compound is easily accessible on a large scale
using simple synthetic manipulations (Scheme 4). In brief,
treatment of 4 with ammonia and reduction of the resulting
amide provided the desired glucamine derivative 18 in high
yield. Coupling the amine group in 18 with R-carboxy-
protected glutamic acid and consecutive allyl ester removal
furnished protected amino acid 20.
(12) Subsequent HPLC chromatography using a Waters Atlantis 3.0 mm ꢀ
50 mm column allowed for the resolution of one of isomers (derived from either
aspartimide formation or epimerization) of 2. The spectrum is shown in the
Supporting Information.
(13) An LC-MS spectrum of a monosaccharo analogue of 2, exemplify-
ing the “peeling” seen with this type of compound, is shown in the Supporting
Information.
Having synthesized 20, a new FA-spacer unit, which was
not saddled with the “peeling” and isomerization issues of 2,
could be assembled. Rather than synthesizing 16, a simplified
J. Org. Chem. Vol. 75, No. 11, 2010 3689

Vlahov et al.
JOCArticle
SCHEME 7. Chemical Release of 22 from 25 via Reduction with
DTT
FIGURE 2. HPLC profile (280 nm) of the treatment of 25 with
DTT at t = 0, 0.25, and 4 h.
1
with the expected structure. In brief, the H NMR spectrum
(800 MHz, DMSO-d6) contained 11 aromatic signals in the
range from 6.4 to 8.8 ppm (five from the folate moiety and six
from the desacetylvinblastine moiety). The signals for the two
olefinic protons in DAVLBH moiety appeared at 5.53 ppm (d)
and 5.69 ppm (m).
Release of the drug from conjugate 25 was also studied
(Scheme 7). A 1 mM solution of 25 in phosphate buffer
(pH = 7.4) was treated with 40 equiv of dithiothreitol (DTT)
at room temperature. The HPLC profile (UV detection at
280 nm) showed complete cleavage of the disulfide bond with
concomitant release of the FA-spacer 21 within 15 min
(Figure 2). At this early time point, the DAVLBH exists as its
free form (22) with the remnants of the linker still attached
(26). Over the next 4 h, the remaining linker is released
leaving only 21 and 22 in solution.
Following positive in vitro and in vivo results and toxico-
logical evaluation, compound 25, also known as EC0489,
was selected as a clinical candidate. The results of the
pharmacological, toxicological, and phase 1 investigations
are being prepared for publication and will be reported soon
in an appropriate scientific journal.
version (21) would be synthesized. The new FA-spacer unit
would preserve the N- and C-termini of 2, while replacing the
components of the core peptidic backbone with alternating
glutamic acid and saccharo-amino acid 20 moieties. FA-spa-
cer unit 21 was prepared using standard Fmoc SPPS on a
Wang-resin polymeric support as shown in Scheme 5. After
cleavage from the resin, LC-MS on the crude reaction mixture
of 21 displayed a single peak of the desired mass. 21 was
purified by preparative RP-HPLC, and its structure was con-
firmed by 1Hand13CNMRandLC-MS[ESI(Mþ H)þ 1679]
analysis.
Our drug, DAVLBH 22, was prepared from commercially
available vinblastine (VLB) sulfate following a literature
procedure.14 Activated carbonate 231 served as heterobi-
functional cross-linker in the assembly of the final conjugate.
Generally, 23 has been found to react under mild conditions
with many N- and O-nucleophiles and in our hands has been
shown to be a convenient and universal tool for the incor-
poration of reductively labile disulfide linkages into a wide
variety of drug conjugates.
As shown in Scheme 6, DAVLBH was treated with the
activated carbonate 23 and diisopropylethylamine (DIPEA)
in dichloromethane to yield 2-(vinblastinyl)hydrazinecar-
boxylic acid 2-pyridyldithioethyl ester 24. After chromato-
graphic purification on silica gel, 24 was isolated in 80%
yield. Treatment of a suspension of FA-spacer 21 in H2O
under argon with 0.1 N NaHCO3 resulted in a clear yellow
solution at pH > 6.8. To this mixture was added at once
under extensive stirring a solution of 24 in THF. According
to the HPLC profile, the reaction was completed in 15 min.
HPLC purification gave pure conjugate 25, and recorded
LC-MS (ESI) and 1Hand13C NMR signals were in agreement
Experimental Section
Synthesis of 3,4;5,6-Di-O-isopropylidene-1-amino-1-deoxy-
(Fmoc-Glu-Oallyl)-D-glucitol 19. Fmoc-Glu-OAll (2.17 g, 1 equiv),
PyBOP (2.88 g, 1 equiv), and DIPEA (1.83 mL, 2 equiv) were
added to a solution of 18 (1.40 g, 5.3 mmol) in dry DMF (6 mL),
and the reaction mixture was stirred at rt under Ar for 2 h. The
solution was diluted with EtOAc (50 mL) and washed with brine
(10 mL ꢀ 3), and the organic layer separated, dried (MgSO4),
filtered, and concentrated to give a residue, which was purified by a
flash column (silica gel, 60% EtOAc/petroleum ether) to afford 19
(1.72 g, 50%) as a solid. 1H NMR (300 MHz, CD3OD): δ 7.79 (d,
J = 7.2 Hz, 2H), 7.67 (t, J = 6.9 Hz, 2H), 7.39 (t, J = 6.9 Hz, 2H),
7.31 (t, J = 7.2 Hz, 2H), 6.02-5.82 (m, 1H), 5.32 (d, J = 17.4 Hz,
1H), 5.21 (m, 1H), 4.62 (d, J = 5.4 Hz, 2H), 4.46-4.20 (m, 2H),
4.15-4.00 (m, 2H), 4.00-3.83 (m, 3H), 3.79 (m, 1H), 3.40-3.33
(m, 2H), 2.32 (t, J = 6.9 Hz, 2H), 2.26-1.88 (m, 2H), 1.38 (s, 3H),
1.36 (s, 3H), 1.34 (s, 3H), 1.27 (s, 3H). 13C NMR (75.5 MHz,
CD3OD): δ 175.2, 173.4, 158.8, 145.5, 145.3, 142.8, 133.4, 129.0,
128.3, 126.5, 126.4, 121.1, 118.8, 111.0, 110.9, 82.4, 78.8, 78.5, 69.7,
68.7, 68.2, 67.0, 55.2, 44.7, 33.4, 28.7, 27.6, 27.2, 27.1, 25.6. HRMS
(ESI): (M þ Na)þ calcd for C35H44N2O10 675.2894, found
675.2899.
(14) Barnett, C. J.; Cullinan, G. J.; Gerzon, K.; Hoying, R. C.; Jones, W.
E.; Newlon, W. M.; Poore, G. A.; Robison, R. L.; Sweeney, M. J.; Todd, G.
C. J. Med. Chem. 1978, 21, 88.
3690 J. Org. Chem. Vol. 75, No. 11, 2010

Vlahov et al.
JOCArticle
Synthesis of 3,4;5,6-Di-O-isopropylidene-1-amino-1-deoxy-
(Fmoc-Glu-OH)-D-glucitol 20. Pd(Ph3)4 (300 mg, 0.1 equiv)
was added to a solution of 19 (1.72 g, 2.81 mmol) in NMM/
AcOH/CHCl3 (2 mL/4 mL/74 mL). The resulting yellow solu-
tion was stirred at rt under Ar for 1 h, to which was added a
second portion of Pd(Ph3)4 (300 mg, 0.1 equiv). After being
stirred for an additional 1 h, the reaction mixuture was washed
with 1 N HCl (50 mL ꢀ 3) and brine (50 mL), and the organic
layer separated, dried (MgSO4), filtered, and concentrated to
give a yellow foamy solid, which was subjected to chromatog-
raphy (silica gel, 1% MeOH/CHCl3 followed by 3.5% MeOH/
CHCl3) to give 20 (1.3 g, 81%) as a solid: 1H NMR (300 MHz,
CD3OD): δ 7.79 (d, J = 7.2 Hz, 2H), 7.66 (t, J = 6.9 Hz, 2H),
7.38 (t, J = 7.2 Hz, 2H), 7.30 (t, J = 7.2 Hz, 2H), 4.45-4.15 (m,
4H), 4.15-4.00 (m, 2H), 4.00-3.85 (m, 3H), 3.73 (m, 1H),
3.40-3.35 (m, 2H), 2.29 (t, J = 7.5 Hz, 2H), 2.25-2.05 (m,
2H), 1.38 (s, 3H), 1.35 (s, 3H), 1.34 (s, 3H), 1.26 (s, 3H); 13C
NMR (75.5 MHz, CD3OD): δ 175.4, 158.8, 145.5, 145.3, 142.7,
128.9, 128.3, 126.5, 126.4, 121.1, 111.0, 110.9, 82.3, 78.8, 78.5,
69.7, 68.7, 68.2, 55.0, 44.7, 33.6, 29.0, 27.6, 27.2, 27.1, 25.6;
HRMS (ESI) (M þ Na)þ calcd for C32H40N2O10 635.2581,
found 635.2573.
Synthesis of Pte-γGlu-(Glu(1-amino-1-deoxy-D-glucitol)-Glu)2-
Glu(1-amino-1-deoxy-D-glucitol)-Cys-OH 21. H-Cys(4-methoxy-
trityl)-2-chlorotrityl-resin (0.17 g, 0.10 mmol) was loaded into a
peptide synthesis vessel and washed with i-PrOH (3 ꢀ 10 mL),
followed by DMF (3 ꢀ 10 mL). To the vessel was then introduced
a solution of 20 (82 mg, 0.13 mmol) in DMF, i-Pr2NEt (2 equiv),
and PyBOP (1 equiv). The resulting solution was bubbled with Ar
for 1 h, the coupling solution was drained, and the resin washed
with DMF (3 ꢀ 10 mL) and i-PrOH (3 ꢀ 10 mL). Kaiser tests
were preformed to assess reaction completion. Fmoc deprotec-
tion was carried out using 20% piperidine in DMF (3 ꢀ 10 mL).
This procedure was repeated to complete all coupling steps
(1.9 equiv of Fmoc-Glu(O-t-Bu)-OH and Fmoc-Glu-O-t-Bu
and 1.6 equiv of N10TFA-pteroic acid were used on each of their
respective coupling steps). After the pteroic acid coupling, the
resin was washed with 2% hydrazinein DMF (3 ꢀ for 5 min each)
to remove the trifluoroacetyl protecting group. The resin was
washed with DMF (3 ꢀ 10 mL) and MeOH (10 mL) and dried
under reduced pressure. The peptide was cleaved from the resin in
the peptide synthesis vessel using a cleavage mixture consisting of
92.5% CF3CO2H, 2.5% H2O, 2.5% triisopropylsilane, and 2.5%
ethanedithiol. Twenty-five milliliters of the cleavage mixture was
added to the peptide synthesis vessel, and the reaction was
bubbled under Ar for 10 min. The resin was treated with two
additional 15 mL quantities of the cleavage mixture for 5 min
each. The cleavage mixture was concentrated to ca. 5 mL, and
ethyl ether was added to induce precipitation. The precipitate
was collected by centrifugation, washed with ethyl ether three
times, and dried under high vacuum, resulting in the recovery
of ca. 100 mg of crude material. One-half of the material was
purified by preparative HPLC (mobile phase: A = 10 mM
ammonium acetate pH = 5, B = ACN; method: 0% B to 20%
B in 25 min at 15 mL/min). The pure fractions were pooled and
freeze-dried, furnishing spacer 21 (43 mg, 51%). 1H NMR (800
MHz, DMSO-d6/D2O): δ 8.6 (s, 1H), 7.6 (d, J = 8 Hz, 2H), 6.62
(d, J = 8 Hz, 2H), 4.47 (s, 2H), 4.26-4.08 (m, 7H), 3.61(m, 3H),
3.53(m, 6H), 3.46(m, 3H), 3.37(m, 6H), 3.21(m, 3H), 3.03(m, 3H),
2.80 (dd, 1H), 2.72 (dd, 1H), 2.3-2.08 (m, 12H), 1.95-1.6 (m,
12H). 13C NMR (125 MHz, DMSO-d6/D2O): δ 175.2, 175.0,
174.9, 173.4, 173.3, 173.2, 173.0, 172.6, 172.1, 172.0, 171.3, 167.2,
162.2, 156.2, 154.3, 151.2, 149.5, 149.3, 129.5, 128.1, 121.8, 112.0,
72.2, 71.82, 71.78, 70.1, 63.6, 56.2, 53.2, 53.0, 52.8, 42.4, 32.3, 30.8,
28.3, 28.2, 27.6, 27.4, 27.2, 26.8. LCMS (ESI): (M þ H)þ = calcd
for C65H98N16O34S, 1679.6, found 1680.0.
Synthesis of Pte-γGlu-(Glu(1-amino-1-deoxy-D-glucitol)-Glu)2-
Glu(1-amino-1-deoxy-D-glucitol)-Cys(S-ethyl-3-(4-desacetylvinbla-
stinyl)hydrazinecarboxylate) 25. In a polypropylene centrifuge
bottle, folate linker 21 (26 mg, 0.015 mmol) was dissolved in
2.5 mL of Ar-sparged water. In another flask, a saturated NaH-
CO3 solution was Ar sparged for 10 min. The pH of the linker
solution was carefully adjusted, with argon bubbling, to 6.9 using
the NaHCO3 solution. Vinblastine hydrazide derivative 24 (15 mg,
1.0 equiv) in 2.5 mL of tetrahydrofuran (THF) was added quickly
to the above solution. The resulting clear solution was stirred
under argon. Progress of the reaction was monitored by analytical
HPLC (2 mM sodium phosphate buffer, pH = 7.0 and acetonitrile).
After 20 min, 2 mM phosphate buffer (pH = 7, 12 mL) was added
to the reaction. The resulting cloudy solution was filtered, and the
filtrate was injected on the prep-HPLC (mobile phase: A = 2 mM
sodium phosphate pH = 7, B = ACN; method: 1% B to 50% B
in 25 min at 26 mL/min). Pure fractions were pooled and freeze-
driedresultinginthe recoveryof 25 as a fluffy yellow powder (27.5
mg, 71%). 1H NMR (800 MHz, DMSO-d6/D2O): δ 8.60(s, 1H),
7.59 (d, J = 8.8 Hz, 2H), 7.36 (d, J = 8 Hz, 1H), 7.21 (d, J = 8 Hz,
1H), 7.0 (t, J = 8 Hz, 1H), 6.93 (t, J = 8 Hz, 1H), 6.77 (d, J = 8.8
Hz, 2H), 6.36 (s, 1H), 6.17 (s, 1H), 5.69 (bd, J = 8 Hz, 1H), 5.52
(d, J = 10.4 Hz, 1H), 4.47 (s, 2H), 4.19 (m, 2H), 4.13-4.05 (m,
6H), 3.96 (m, 1H), 3.76 (s, 1H), 3.68 (s, 3H), 3.64-3.62 (m, 4H),
3.55 (m, 6H), 3.5 (s, 3H), 3.48 (m, 3H), 3.40 (m, 8H), 3.21 (m, 6H),
3.11-3.00 (m, 5H), 2..93 (m, 1H), 2.88 (m, 1H), 2.72 (s, 3H), 2.70
(d, J = 7.2, 1H), 2.60 (m, 1H), 2.41 (bs, 1H), 2.34 (bd, J = 10.4
Hz, 1H), 2.24 (s, 1H), 2.15-2.05 (m, 12H), 1.97-1.78 (m, 14H),
1.53 (m, 2H), 1.31 (d, J = 11.2, 1H), 1.21-1.17 (m, 4H), 0.76 (t,
J = 7.2 Hz, 3H), 0.70 (t, J = 7.2 Hz, 3H), 0.65 (m, 1H). 13C NMR
(125 MHz, DMSO-d6/D2O): δ 177.7, 177.4, 175.9, 175.6, 174.2,
173.5, 173.3, 173.2, 172.7, 172.6, 172.3, 172.2, 172.1, 171.2, 166.5,
158.0, 156.6, 156.1, 155.1, 152.9, 151.0, 149.4, 149.2, 135.6, 131.8,
131.7, 129.2, 129.1, 128.1, 124.0, 123.7, 123.0, 122.4, 122.1, 119.9,
119.0, 118.5, 116.3, 112.1, 111.8, 93.5, 83.2, 80.9, 73.9, 72.2, 71.9,
71.8, 70.0, 69.9, 68.0, 65.8, 63.6, 63.3, 62.9, 56.6, 55.6, 54.5, 54.3,
53.9, 53.7, 53.3, 52.9, 52.6, 50.5, 49.7, 46.9, 46.2, 45.4, 42.5, 38.7,
38.4, 37.8, 35.3, 34.7, 33.6, 33.5, 33.2, 32.3, 32.2, 30.6, 29.7, 28.9,
28.7, 28.0, 27.7, 25.7, 8.7, 7.6. LCMS (ESI): (M þ H)þ = calcdfor
C111H158N22O43S2 2551.1, found 2551.8.
Acknowledgment. We thank the analytical group at
Endocyte, Inc., for LC/MS support.
Supporting Information Available: Experimental proce-
dures and spectral data. This material is available free of charge
via the Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 75, No. 11, 2010 3691