(-)-Sandramycin: Total synthesis and characterization of DNA binding properties

Article abstract of DOI:10.1021/ja952799y

Full details of the total synthesis of the potent antitumor antibiotic (-)-sandramycin (1), a cyclic decadepsipeptide possessing a 2-fold axis of symmetry, is described and constitutes the first total synthesis of a member of the growing class of naturall

Full text of DOI:10.1021/ja952799y

J. Am. Chem. Soc. 1996, 118, 1629-1644
1629
(-)-Sandramycin: Total Synthesis and Characterization of
DNA Binding Properties
Dale L. Boger,* Jyun-Hung Chen, and Kurt W. Saionz
Contribution from the Department of Chemistry, The Scripps Research Institute,
10666 North Torrey Pines Road, La Jolla, California 92037
ReceiVed August 14, 1995X
Abstract: Full details of the total synthesis of the potent antitumor antibiotic (-)-sandramycin (1), a cyclic
decadepsipeptide possessing a 2-fold axis of symmetry, is described and constitutes the first total synthesis of a
member of the growing class of naturally occurring agents now including the luzopeptins and quinaldopeptin. Key
strategic elements of the approach include the late stage introduction of the heteroaromatic chromophore thereby
providing access to analogs possessing altered intercalation capabilities, symmetrical pentadepsipeptide coupling
and 32-membered macrocyclization conducted at the single secondary amide site in superb conversion (90%), and
a convergent assemblage of the precursor pentadepsipeptide in which the potentially labile ester linkage was introduced
in the final key coupling reaction. This approach also provided the cyclic decadepsipeptides 24-26 lacking both
chromophores and was extended to provide 32 lacking one of the two chromophores. The characterization of the
DNA-binding properties of sandramycin vs 25 and 32 is detailed. The largest share of the binding is derived from
the cyclic decadepsipeptide (∆G° ) -6.0 kcal/mol) and the incremental addition of each chromophore increases the
binding approximately 3.2 and 1.0 kcal/mol, respectively. This is consistent with the representation of sandramycin
and the luzopeptins as minor groove binding cyclic decadepsipeptides incrementally stabilized by mono and
bisintercalation. Following the same trends, sandramycin and luzopeptin A were found to be nearly equivalent,
exceptionally potent cytotoxic agents (6-0.02 nM), 500-1000× more potent than the cyclic decadepsipeptide 32
possessing a single chromophore, and g105× more potent than the cyclic decadepsipeptides 24 and 25 lacking both
chromophores. DNase I footprinting studies revealed that sandramycin and luzopeptin A behave comparably and
appear to bind best to regions containing alternating A and T residues. Binding at other and perhaps all sites is
observed at modest agent concentrations with a perceptible preference for 5′-AT dinucleotide sequences many of
which were preceded by a 5′-C, i.e. 5′-CAT. Preliminary studies of the 1:1 complex of sandramycin with 5′-
d(GCATGC)2 revealed that it maintains the 2-fold axis of symmetry of the components with the agent sandwiching
the central two AT base pairs and adopting a compact conformation in which the interchromophore distance is 10.1
Å. The cyclic decadepsipeptide is positioned in the minor groove and the adopted conformation permits a rich array
of complementary hydrophobic contacts extending over much of the interacting surface.
Sandramycin (1), a potent antitumor antibiotic1 isolated from
the culture broth of a Norcardioides sp. (ATCC 39419) and
structurally characterized through extensive spectroscopic and
chemical degradation studies,2 constitutes one of the newest
members of a growing class of cyclic decadepsipeptides
including luzopeptins A-C, E2,3 and quinaldopeptin4 which
possess potent antitumor, antiviral, and antimicrobial activity.3-5
Characteristic of this class of agents, sandramycin possesses a
2-fold axis of symmetry and two pendant heteroaromatic
chromophores that could be anticipated to result in DNA
bifunctional intercalation similar to that detailed for the lu-
zopeptins which span two base pairs preferentially at 5′-AT
sites.6-8 In this respect, the agents are functionally related to
echinomycin and triostin A, bicyclic octadepsipeptides, which
also bind to DNA by bisintercalation but with a different
sequence selectivity (5′-CG versus 5′-AT).9-11
Herein, we provide full details of the total synthesis of (-)-
sandramycin (1) which constitutes the first total synthesis12,13
of a naturally occurring member of this class of agents and that
served to confirm the structural and absolute stereochemical
assignments of 1. Key strategic elements of the approach
(6) (a) Fox, K. R.; Woolley, C. Biochem. Pharmacol. 1990, 39, 941. (b)
Fox, K. R.; Davies, H.; Adams, G. R.; Portugal, J.; Waring, M. J. Nucleic
Acids Res. 1988, 16, 2489.
(7) Huang, C.-H.; Mong, S.; Crooke, S. T. Biochemistry 1980, 19, 5537.
Huang, C.-H.; Prestayko, A. W.; Crooke, S. T. Biochemistry 1982, 21, 3704.
Huang, C.-H.; Crooke, S. T. Cancer Res. 1985, 45, 3768. Huang, C.-H.;
Mirabelli, C. K.; Mong, S.; Crooke, S. T. Cancer Res. 1983, 43, 2718.
(8) Leroy, J. L.; Gao, X.; Misra, V.; Gueron, M.; Patel, D. J. Biochemistry
1992, 31, 1407. Zhang, X.; Patel, D. J. Biochemistry 1991, 30, 4026. Searle,
M. S.; Hall, J. G.; Denny, W. A.; Wakelin, L. P. G. Biochem. J. 1989, 259,
433.
(9) Waring, M. J.; Fox, K. R. In Mol. Aspects of Anticancer Drug Action;
Waring, M. J., Neidle, S., Eds.; Verlag: Weinheim, 1983; p 127. UK-63052
and related agents: Rance, M. J.; Ruddock, J. C.; Pacey, M. S.; Cullen, W.
P.; Huang, L. H.; Jefferson, M. T.; Whipple, E. B.; Maeda, H.; Tone, J. J.
Antibiot. 1989, 42, 206. Fox, K. R. J. Antibiot. 1990, 43, 1307.
(10) Wang, A. H.-J.; Ughetto, G.; Quigley, G. J.; Hakoshima, T.; van
der Marel, G. A.; van Boom, J. H.; Rich, A. Science 1984, 225, 1115.
Ughetto, G.; Wang, A. H.-J.; Quigley, G. J.; van der Marel, G. A.; van
Boom, J. H.; Rich, A. Nucleic Acids Res. 1985, 13, 2305. Van Dyke, M.
M.; Dervan, P. B. Science 1984, 225, 1122.
X Abstract published in AdVance ACS Abstracts, January 15, 1996.
(1) Matson, J. A.; Bush, J. A. J. Antibiot. 1989, 42, 1763.
(2) Matson, J. A.; Colson, K. L.; Belofsky, G. N.; Bleiberg, B. B. J.
Antibiot. 1993, 46, 162.
(3) (a) Ohkuma, H.; Sakai, F.; Nishiyama, Y.; Ohbayashi, M.; Imanishi,
H.; Konishi, M.; Miyaki, T.; Koshiyama, H.; Kawaguchi, H. J. Antibiot.
1980, 33, 1087. Tomita, K.; Hoshino, Y.; Sasahira, T.; Kawaguchi, H. J.
Antibiot. 1980, 33, 1098. Konishi, M.; Ohkuma, H.; Sakai, F.; Tsuno, T.;
Koshiyama, H.; Naito, T.; Kawaguchi, H. J. Antibiot. 1981, 34, 148. (b)
Konishi, M.; Ohkuma, H.; Sakai, F.; Tsuno, T.; Koshiyama, H.; Naito, T.;
Kawaguchi, H. J. Am. Chem. Soc. 1981, 103, 1241. (c) Arnold, E.; Clardy,
J. J. Am. Chem. Soc. 1981, 103, 1243.
(11) Synthesis: Chakravarty, P. K.; Olsen, R. K. Tetrahedron Lett. 1978,
19, 1613. Shin, M.; Inouye, K.; Otsuka, H. Bull. Chem. Soc. Jpn. 1984, 57,
2203. Ciardelli, T. L.; Chakravarty, P. K.; Olsen, R. K. J. Am. Chem. Soc.
1978, 100, 7684. Shin, M.; Inouye, K.; Higuchi, N.; Kyogoku, Y. Bull.
Chem. Soc. Jpn. 1984, 57, 2211.
(4) Toda, S.; Sugawara, K.; Nishiyama, Y.; Ohbayashi, M.; Ohkusa, N.;
Yamamoto, H.; Konishi, M.; Oki, T. J. Antibiot. 1990, 43, 796.
(5) Take, Y.; Inouye, Y.; Nakamura, S.; Allaudeen, H. S.; Kubo, A. J.
Antibiot. 1989, 42, 107.
0002-7863/96/1518-1629$12.00/0 © 1996 American Chemical Society

1630 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Boger et al.
Scheme 1
include the deliberate late stage introduction of the heteroaro-
matic chromophore thereby providing simple access to structural
analogs possessing modified intercalation capabilities, sym-
metrical pentadepsipeptide coupling and macrocyclization of the
32-membered decadepsipeptide conducted at the single second-
ary amide site, and a convergent assemblage of the precursor
pentadepsipeptide in which the potentially labile ester linkage
was introduced in the final key coupling reaction (Scheme 1).
The characterization of the high affinity, bifunctional intercala-
tion of 1 is disclosed.
as did DCC-HOBt and DCC-DMAP (CH2Cl2, 25 °C, 24 h,
42%). Other carboxylate activation procedures including the
use of N-methyl-2-chloropyridinium iodide,20 a pivolyl mixed
anhydride (toluene, 60 °C, 12 h, 19%),21 or DPPA22 (1.8 equiv,
DMF, 25 °C, 20 h, 20-40%) were less successful. Attempts
to employ O-silyl or O-benzyl derivatives of N-SES-D-Ser did
not improve these observations and suggested that the prob-
lematic feature was not competitive reaction of the D-serine
hydroxyl but rather the sluggish reaction of the benzyl L-
pipecolate secondary amine.
Esterification of 7 with 14 provided the pentadepsipeptide
15 and was accomplished through use of DCC-DMAP23 (1
equiv of DCC, 1.0 equiv of DMAP, CH2Cl2, 0 °C, 24 h, 79-
89%). This esterification was anticipated to be problematic
since the carboxylic acid coupling partner contains a N-methyl
amide which is known to decelerate or preclude sensitive
esterifications and increase the propensity for racemization and
this is especially true of a sensitive N-methyl-L-valine center.24
Moreover, the product sensitivity to subsequent â-elimination
further restricted the choice of acylation conditions. Nonethe-
less, the coupling proved remarkably straightforward with the
exception that racemization did prove problematic. Competitive
racemization of the L-valine center was observed if the reaction
was conducted under conventional conditions employing cata-
lytic DMAP but the use of increasing amounts of DMAP was
found to suppress epimerization (Table 1). Alternative esteri-
fication procedures including DCC-HOBt (1 equiv of DCC,
1.2 equiv of HOBt, DMF, 0 °C, 20 h, 54%), BOP-Cl (1.1
equiv, 2.2 equiv of Et3N, CH2Cl2, 0 °C, 24 h, 41%) proved less
successful.
Pentadepsipeptide Synthesis. Coupling of BOC-Gly-Sar-
15
OH (3)14 with L-NMe-Val-OCH3 (5, 1 equiv of DCC, 1.05
equiv of Et3N, 0.1 equiv of DMAP, CH2Cl2, 25 °C, 24 h, 74%)
followed by methyl ester hydrolysis of 6 (3 equiv of LiOH,
3:1:1 THF-CH3OH-H2O, 25 °C, 3 h, 90%) provided 7 and a
key subunit for incorporation into the pentadepsipeptide 15
(Scheme 2). Coupling of L-pipecolic acid benzyl ester (11)
prepared as illustrated in Scheme 2 from L-pipecolic acid16 with
D-N-SES-Ser-OH (13, 1.3-1.4 equiv of BOP-Cl,17 2.6-3.0
equiv of Et3N or i-Pr2NEt, CH2Cl2, 0 °C, 10 h, 82-85%)
provided the dipeptide 14. Notably, this coupling to provide a
tertiary amide could be conducted without deliberate protection
of the D-serine hydroxyl group, competitive racemization,17 or
â-elimination and provided 14 in excellent yield suitably
protected for direct incorporation into 15. Moreover, the
[â-(trimethylsilyl)ethyl]sulfonyl (SES) group18 incorporated into
14-15 served as an admirable orthogonal peptide protecting
group stable to BOC and benzyl ester deprotection yet capable
of selective removal in the presence of the depsipeptide ester.
However, this coupling of the secondary amine of a pipecolic
acid derivative proved more challenging than the results
described above might suggest. A wide range of amide coupling
procedures were examined and provided modest results analo-
gous to the related observations of others.19 Attempts to
promote the coupling with EDCI-HOBt under a range of
reaction conditions (-30 to 25 °C, 12-48 h) with added bases
(NaHCO3, Et3N, i-Pr2NEt) in a range of reaction solvents (DMF,
THF, CH2Cl2) provided 14 in modest conversions (20-40%)
(12) Boger, D. L.; Chen, J.-H. J. Am. Chem. Soc. 1993, 115, 11624.
(13) Olsen, R. K.; Apparao, S.; Bhat, K. L. J. Org. Chem. 1986, 51,
3079. Ciufolini, M. A.; Swaminathan, S. Tetrahedron Lett. 1989, 30, 3027.
Hughes, P.; Clardy, J. J. Org. Chem. 1989, 54, 3260. Ciufolini, M. A.; Xi,
N. J. Chem. Soc., Chem. Commun. 1994, 1867. Greck, C.; Bischoff, L.;
Genet, J. P. Tetrahedron: Asymmetry 1995, 6, 1989.
In initial efforts, the prospect that the racemization of the
L-valine center might be occurring during the conversion of 6
to provide 7 was also a concern especially because of the
propensity for N-methyl amino acids to racemize during basic
hydrolysis.24 Consequently, BOC-NMe-L-Val-OBn (18)25 was
(14) Seebach, D.; Bossler, H.; Gru¨ndler, H.; Shoda, S.; Wenger, R. HelV.
Chim. Acta 1991, 74, 197.
(15) N-BOC-NMe-Val-OCH3 was prepared in one step from N-BOC-
Val by exhaustive methylation (2.5-3 equiv of NaH, 8 equiv of CH3I, 10:1
THF-DMF, reflux, 12-24 h, 78%) following the procedure of Coggins
and Benoiton: Coggins, J. R.; Benoiton; N. L. Can. J. Chem. 1971, 49,
1968.
(16) L-Pipecolic acid benzyl ester was prepared from the L-pipecolic acid
D-tartrate salt obtained by recrystallization and resolution (EtOH-H2O, 3×,
71%): [R]23D -20 (c 10, H2O), lit. [R]25D -20.2 (c 10, H2O); Rodwell, V.
W. Methods Enzymol. 1971, 17, Part B, 174.
(17) Tung, R. D.; Rich, D. H. J. Am. Chem. Soc. 1985, 107, 4342.
(18) Weinreb, S. M.; Demko, D. M.; Lessen, T. A. Tetrahedron Lett.
1986, 27, 2099.
(19) Perlow, D. S.; Erb, J. M.; Gould, N. P.; Tung, R. D.; Freidinger, R.
M.; Williams, P. D.; Veber, D. F. J. Org. Chem. 1992, 57, 4394.
(20) Bald, E.; Saigo, K.; Mukaiyama, T. Chem. Lett. 1975, 1163.
(21) Schreiber, S. L.; Anthony, N. J.; Dorsey, B. D.; Hawley, R. C.
Tetrahedron Lett. 1988, 29, 6577 and references cited therein.
(22) Shioiri, T.; Yamada, S. Chem. Pharm. Bull. 1974, 22, 849. Kitada,
C.; Fujino, C. Chem. Pharm. Bull. 1978, 26, 585.
(23) Hassner, A.; Alexanian, V. Tetrahedron Lett. 1978, 19, 4475.
(24) McDermott, J. R.; Benoiton, N. L. Can. J. Chem. 1973, 51, 2555.
McDermott, J. R.; Benoiton, N. L. Can. J. Chem. 1973, 51, 2562. Ciardelli,
T. L.; Chakravarty, P. K.; Olsen, R. K. J. Am. Chem. Soc. 1978, 110, 7684.
Tung, R. D.; Dhaon, M. K.; Rich, D. H. J. Org. Chem. 1986, 51, 3350.
Dhaon, M. K.; Olsen, R. K.; Ramasamy, K. J. Org. Chem. 1982, 47, 1962.
Zeggaf, C.; Poncet, J.; Jouin, P.; Dufour, M.-N.; Castro, B. Tetrahedron
1989, 45, 5039. Coste, J.; Fre´rot, E.; Jouin, P.; Castro, B. Tetrahedron Lett.
1991, 32, 1967.

(-)-Sandramycin: Total Synthesis and Characterization
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1631
Scheme 2
Table 1. Pentadepsipeptide 15 Synthesis
Scheme 3
DMAP, equiv
% yield 15
epi-15
0.1
0.15
0.2
0.5
1.0
50
59
63
65
32
27
24
11
79-89
2-8
incorporated into the tripeptide 7 by coupling 19 with 3 (1.0
equiv of DCC, 0.5 equiv of DMAP, 1.1 equiv of Et3N, CH2-
Cl2, 25 °C, 24 h, 85%). Deprotection of 20 by catalytic
hydrogenolysis (H2, cat 10% Pd-C, CH3OH, 25 °C, 12 h,
100%) cleanly provided 7 free of concerns of racemization of
the L-valine center (eq 1). The use of material prepared in this
manner provided results comparable but perceptibly better than
those obtained from the methyl ester 6. The minor extent of
racemization obtained in the esterification coupling of 7 with
14 in the presence of 1.0 equiv of DMAP was further diminished
(2-6%) with the material derived from 20, indicating that
racemization also accompanies the hydrolysis of the methyl ester
6. Thus, the approach detailed in eq 1 proceeding through 20
was adopted for our studies. Fortunately, the major (79-89%)
and minor diastereomers (2-8%) were readily separable (SiO2,
Rf ) 0.44 and 0.35, 67% EtOAc-hexane) and stereochemically
homogeneous material was employed in the preparation of 1.
Similarly, concerns arose in initial studies as to whether the
benzyl L-pipecolate center had racemized under the conditions
required for coupling of 11 with 13. Consequently, racemic
11 was coupled with N-SES-D-Ser (13) and a distinguishable
pair of resulting diastereomers was obtained. Their comparison
with authentic 14 ensured that no detectable racemization of
the L-pipecolic acid subunit had occurred. In addition, the
sample of 14 containing the 1:1 mixture of D- and L-pipecolic
acid was coupled with 7 and the resulting diastereomeric mixture
of products 15 was not the same mixture of products obtained
above indicating that the racemization encountered could be
attributed to the NMe-Val center.
sipeptide formation was accomplished by independent depro-
tection of the amine (3 M HCl-EtOAc, 25 °C, 30 min, 100%)
and carboxylic acid terminus (H2, 10% Pd-C, CH3OH, 25 °C,
12 h, 98%) of 15 to provide 16 and 17, respectively, which
were coupled with formation of the single secondary amide (1
equiv of EDCI, 1 equiv of HOBt, 4.0 equiv of NaHCO3, CH2-
Cl2, 25 °C, 24 h, 81%) to provide 21 in excellent conversion
(Scheme 3). This same coupling reaction conducted with Et3N
(2.0 equiv) in place of NaHCO3 provided 21 in substantially
lower conversions (55%) and may reflect the sensitivity of the
Cyclic Decadepsipeptide Formation and Completion of the
Total Synthesis of (-)-Sandramycin (1). Linear decadep-
(25) N-BOC-NMe-Val-OBn was prepared by N-methylation of N-BOC-
Val-OBn (1.5 equiv of NaH, 4.2 equiv of CH3I, 10:1 THF-DMF, reflux,
24 h, 90%). For 18: Wenger, R. M. HelV. Chim. Acta 1983, 66, 2672.

1632 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Boger et al.
pentadesipeptides 16 and 17 or the product decadepsipeptide
to base-catalyzed â-elimination.
Scheme 4
Cyclization of 21 to provide the 32-membered cyclic dec-
adepsipeptide 24, [R]23 -88 (c 0.85, CHCl3), with the ring
D
closure strategically conducted at the single secondary amide
site was accomplished in superb conversion by sequential benzyl
ester (H2, 10% Pd-C, CH3OH, 25 °C, 12 h) and BOC
deprotection (3 M HCl-EtOAc, 25 °C, 30 min) followed by
treatment of 23 with diphenyl phosphorazidate (4 equiv of
DPPA, 10 equiv of NaHCO3, 0.003 M DMF, 0 °C, 48 h, 90%
overall).26 Upon cyclization, the cyclic decadepsipeptide adopts
a single rigid solution conformation comparable to that observed
with 1. Given the facility of the 32-membered ring macrocy-
clization reaction, we also attempted to simply couple the
pentapeptides and effect the ring closure in a single reaction.
However, hydrogenolysis of the benzyl ester 15, acid-catalyzed
deprotection of the resulting 17 (3 M HCl-EtOAc, 25 °C, 30
min, 100%) and treatment of the product amino acid with DPPA
(4 equiv, 10 equiv of NaHCO3, 0.01 M DMF, 0 °C, 48 h)
provided only small amounts of 24 along with a full range of
oligomers and higher order macrocycles.
Removal of the SES amine protecting group was accom-
plished under mild conditions (10 equiv of Bu4NF, 22-30 equiv
of BOC2O, THF, 25 °C, 48 h, 70-73%) but required in situ
trap of the liberated amine as its BOC derivative 25, [R]23
D
-53 (c 0.15, CHCl3), for isolation. Presumably this may be
attributed to the instability of the linking ester in the decadep-
sipeptide to the liberated amine under the anhydrous and basic
reaction conditions. However, this deprotection was accom-
plished under surprisingly mild reaction conditions (25 °C).
Since SES amine deprotections generally required higher
reaction temperatures (50-100 °C), we cannot rule out the
possibility that BOC acylation of the amine precedes and
activates the subsequent SES deprotection. Although this was
not investigated in detail, initial efforts to check our reagents
(CsF or Bu4NF) with the simple substrate N-SES-Phe-OCH3
provided only low conversions to the expected free amine and
afforded substantial or predominant amounts of the correspond-
ing diketopiperazine (eq 2). Conducting this deprotection with
10% Pd-C, EtOAc, 25 °C, 12 h, 78%) to provide (-)-1, [R]23
-153 (c 0.17, CHCl3), which was identical in all respects with
a sample of authentic material (1H NMR, 13C NMR, IR, MS,
UV, mp, [R]D, and chromatographic properties).
D
Preparation of Cyclic Decadepsipeptide 32 Possessing a
Single Chromophore. For comparisons with 1 and the agents
24-26 lacking both of the chromophores, the agent 32 pos-
sessing a single chromophore was also prepared. In initial
studies on the deprotection of 24, exposure to Bu4NF (4 equiv)
for shorter periods of time (24 h, 25 °C) in the presence of
BOC2O (10 equiv) led to partial deprotection to provide 29
(33%) along with recovered 24 (11%) and 25 (27%) (Scheme
4). Without further attempts at optimization, this inadvertent
preparation of 29 provided sufficient material for our synthesis
of 32. Acid-catalyzed BOC deprotection of 29 (3 M HCl-
EtOAc, 25 °C, 30 min) followed by acylation of the liberated
amine hydrochloride salt 30 with 28 (4 equiv of EDCI, 6 equiv
of HOBt, 10 equiv of NaHCO3, 25 °C, 48 h, 63%) and
subsequent catalytic hydrogenolysis of the benzyl ether 31 (H2,
cat. 10% Pd-C, EtOAc, 25 °C, 14 h, 86%) cleanly provided
32, [R]25 -105 (c 0.3, CHCl3).
D
Conformational Properties of 1 and the Related Cyclic
Decadepsipeptide 25. The X-ray structure determination of
2527 revealed a backbone conformation nearly identical to that
of luzopeptin A (Figure 1, rms ) 1.40 Å).3c The most
significant difference in the two structures is the twisted
orientation of the linking esters. The relative placement of the
ring nitrogens (rms ) 0.73 Å) and the backbone conformation
of the pentapeptides excluding the ester atoms (rms ) 0.74 Å)
are even more similar in the two structures. The overall shape
of the agent is rectangular with a 2-fold axis of symmetry. The
long sides of the rectangle consist of antiparallel and twisted
â-extended chains capped on either end by the two decadep-
sipeptide ester linkages. Each of the amides including the three
tertiary amides adopt a trans or extended stereochemistry and
the two decadepsipeptide esters adopt the preferred syn con-
formation. The two symmetrical glycine secondary amide NH’s
are engaged in tight transannular H-bonds (2.08 Å, gly-
NHsOdC-gly) to the glycine carbonyl oxygen across the ring
and cap two reverse peptide turns induced in part by the
incorporation of unnatural D-serine at one corner of each turn.
CsF or Bu4NF in the presence of BOC2O (10 equiv) cleanly
provided N-BOC-Phe-OCH3 and this protocol may serve as an
excellent solution for those who encounter similar difficulties.18
A single-crystal X-ray structure determination of 2527 con-
firmed the structural and stereochemical assignments and further
revealed a rigid cyclic decadepsipeptide conformation essentially
identical to that found in the X-ray structure of luzopeptin A.3c
Completion of the synthesis required BOC deprotection of 25
(3 M HCl-EtOAc, 25 °C, 30 min), coupling of the resulting
bis amine 26 with 3-(benzyloxy)quinoline-2-carboxylic acid
(28,28 4.0 equiv of EDCI, 6.0 equiv of HOBt, 10 equiv of
NaHCO3, DMF, 25 °C, 72 h, 91%) and a final deprotection of
bis-O-benzylsandramycin (27, [R]23D -107 (c 0.3, CHCl3); H2,
(26) Brady, S. F.; Freidinger, R. M.; Paleveda, W. J.; Colton, C. D.;
Homnick, C. F.; Whitter, W. L.; Curley, P.; Nutt, R. F.; Veber, D. F. J.
Org. Chem. 1987, 52, 764.
(27) The author has deposited the atomic coordinates for this structure
with the Cambridge Crystallographic Data Centre. The coordinates may be
obtained upon request from the Director, Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge, CB2, 1EZ, UK.
(28) Boger, D. L.; Chen, J.-H. J. Org. Chem. 1995, 60, 7369.

(-)-Sandramycin: Total Synthesis and Characterization
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1633
and exchange peaks. In all solvents except DMSO-d6, the agent
25 adopted a single, rigid solution conformation comparable to
that observed in the X-ray and comparable to that observed with
1 itself (Figure 6, supporting information). In DMSO-d6, the
1H NMR spectrum was broad and nondescript, indicating
multiple conformations with no single one dominating although
the conformation adopted in other solvents was observable. Clear
from these studies, the agent 25 as well as 1 adopt a single
solution conformation in all solvents except DMSO-d6 that is
analogous to the X-ray conformation in which all amides
including the three tertiary amides are trans.
Several significant NOEs and diagnostic coupling constants
were used to establish the stereochemistry of the amides, their
orientation, and decadepsipeptide backbone conformation. The
presence of pip-ꢀ-CH(eq)/ser-R-CH and ser-â-CH2 NOE cross-
peaks established the trans D-ser-pip amide stereochemistry, the
local orientation of the pip six-membered ring, and the presence
of the D-ser turn. The absence of pip-ꢀ-CH(ax)/ser-R-CH or
ser-â-CH2 NOEs further fixed these orientations analogous to
that seen in the X-ray and the absence of a ser-R-CH/pip-R-
CH NOE excluded the presence of a cis-ser-pip amide. A strong
gly-NH/pip-R-CH NOE established their syn orientation com-
parable to that observed in the X-ray and the trans pip-gly amide
bond. The equally intense and strong NOE crosspeaks between
sar-NMe and both protons of gly-R-CH2 indicated a trans-gly-
sar N-methyl amide stereochemistry. Moreover, the two protons
of gly-R-CH2 and gly-NH exhibit coupling constants of 5.2 and
0 Hz and are consistent with only two combinations of dihedral
angles (-90 and 150° or 90 and -30°). The former corresponds
to that observed in the X-ray with the gly-NH and gly carbonyl
eclipsed and oriented for transannular H-bonding to the corre-
sponding residue on the opposite side of the ring and is
consistent with the observed weak gly-NH/gly-R-CH2 NOE
crosspeaks. The strong NOE to both protons of sar-R-CH2/
val-NMe are diagnostic of a trans-sar-NMe-val amide and
served to establish their proximal orientations. Diagnostic of
the twist in the extended sheet observed in the X-ray, one sar-
NMe/sar-R-CH NOE was observed and one was not. A strong
val-NMe/val-â-CH NOE was observed and is consistent with
their syn orientation. Similarly, the pattern of NOEs for the
val side-chain methyl groups and the large val-â-CH/val-R-CH
coupling constant (J ) 11.1-10.9 Hz) are consistent with a
well-defined solution conformation analogous to that observed
in the X-ray where the val-â-CH and val-R-CH are trans
antiperiplanar to one another. The absence of a ser-â-CH2/val-
NMe NOE established the relative orientations of the sar-NMe-
val amide and decadepsipeptide ester similar to that observed
in the X-ray as do the absence of ser-â-CH2/val-R-CH, â-CH
or γ-CH3 NOEs. The orientation of the BOC-NH-ser is fixed
by the observance of a val-NMe/ser-NH NOE and by the
absence of a ser-NH/ser-â-CH2 NOEs.
Figure 1. ChemDraw 3D representations of the X-ray structures of
luzopeptin A3c (top) and 25 (bottom).
Table 2. 1H NMR (400 MHz) of 25
δ
proton
Gly-NH
Boc-NH
Sar-R-CH
Pip-R-CH
Ser-R-CH
Val-R-CH
Ser-â-CH2
25, CDCl3
8.46 (d, 5.2)
5.85 (d, 6.1)
5.35 (d, 16.8)
5.28 (d, 4.8)
4.82 (d, 6.1)
4.80 (d, 11.0)
4.47 (s)
sandramycin, CDCl3
8.52 (d, 4.4)
2H
2H
2H
2H
2H
2H
4H
5.54 (d, 16.6)
5.57 (d, 6.4)
5.26 (d, 5.0)
4.87 (d, 11.0)
4.99 (d, 11.7)
4.43 (d, 11.7)
4.43 (d, 11.7)
4.06 (m)
Gly-R-CH
Gly-R-CH
Pip-ꢀ-CH (ax) 2H
Pip-ꢀ-CH (eq) 2H
Sar-R-CH
Val-NCH3
Sar-NCH3
Val-â-CH
Pip-(CH2)3
2H
2H
4.41 (dd, 18.0, 5.2)
4.03 (d, 18.0)
3.90 (app t, 12.0)
3.61 (d, 12.0)
3.42 (d, 16.8)
2.95 (s)
4.10 (m)
3.74 (d, 14.5)
3.55 (d, 16.6)
3.12 (s)
2H
6H
6H
2H
2.92 (s)
2.94 (s)
2.13 (dsp, 11.0, 6.5) 2.04 (dsp, 11.0, 6.4)
12H 1.65 (m)
1.73 (m), 1.59 (m),
1.47 (m)
Boc
18H 1.40 (s)
Val-γ-CH3
Val-γ-CH3
6H
6H
0.98 (d, 6.5)
0.84 (d, 6.5)
0.92 (d, 6.4)
0.78 (d, 6.4)
The pipecolic acid residue adopts a chair conformation with
the R-carboxylate adopting an axial position and skewed by
approximately 48° from the optimal carbonyl/CR-H anti
relationship. In this conformation the D-ser-NH/D-ser-NH
distance is 15.1 Å. The comparable luzopeptin A D-ser-NH/
D-ser-NH distance is 14.8 Å and the distance between the centers
of the two chromophores in this X-ray is 17.4-19.9 Å.
The gly-NH exhibited a high chemical shift (δ 8.46-8.35)
and no solvent dependence providing a clear indication that it
is participating in a strong H-bond with the transannular gly
carbonyl. In contrast, the D-ser-NH exhibited a much lower
chemical shift of δ 5.09-6.53 and a much larger solvent
dependence. Finally, the exchange rates for the D-ser-NH and
gly-NH were measured in 5% D2O in DMF-d7. Consistent with
this proposed H-bonding, the solvent accessible D-ser-NH
exchanged rapidly (t1/2 < 2 min) while the H-bonding gly-NH
exchanged much more slowly and required 9-10 h for complete
exchange (t1/2 ) 1.5 h).
The 1D H1 NMR of 24 and 25 indicate that they adopt a
single, rigid solution conformation comparable to that observed
with sandramycin. Consequently, 25 was examined in detail.
Complete proton assignments in the H1 NMR of 25 (Table 2)
were made by both 1H-1H COSY and 1D H1 NMR decoupling
experiments in a range of solvents (CDCl3, THF-d8, CD3OD,
DMF-d7, DMSO-d6, supporting information) and confirmed by
1H-1H NOESY and ROESY NMR experiments (Table 3). The
ROESY spectrum was used to distinguish intermolecular NOE
Similar characteristics to those detailed above were observed
with the solution conformation of 1 in both our studies and those
detailed by Matson2 and are consistent with the solution

1634 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Table 3. 1H-1H NOEs Observed with 25 in CDCl3
Boger et al.
proton
Gly-NH
δ,
observed NOE to
δ
proton
δ,
observed NOE to
δ
8.46
Pip-R-CH
Gly-R-CH
Gly-R-CH
Ser-R-CH
Val-NCH3
Sar-R-CH
Val-NCH3
Gly-NH
5.28
Pip-ꢀ-CH (eq)
3.61
Ser-R-CH
Ser-â-CH2
Pip-ꢀ-CH
Sar-R-CH
Val-NCH3
Sar-NCH3
Boc-NH
4.82
4.47
3.90
5.35
2.95
2.92
5.85 (w)
5.35
4.80 (w)
3.42
2.13
1.40 (w)
0.84
4.41
4.03
3.42
4.80 (w)
2.95
0.98 (w)
0.84 (w)
2.95 (w)
4.80
4.41 (w)
4.03 (w)
4.82
2.95
3.42
2.95
8.46
5.85
4.47
Boc-NH
5.85
5.35
Sar-R-CH2
3.42
2.95
Sar-R-CH
Val-NCH3
Pip-R-CH
Ser-R-CH
5.28
4.82
Sar-R-CH
Val-R-CH
Sar-R-CH
Val-â-CH
BOC-NH
Val-γ-CH3
Gly-R-CH
Gly-R-CH
Sar-R-CH
Val-R-CH
Val-NCH3
Val-γ-CH3
Val-γ-CH3
Val-NCH3
Val-R-CH
Val-â-CH
Val-R-CH
Val-NCH3
Val-â-CH
Boc-NH
Ser-â-CH2
Pip-ꢀ-CH
Val-NCH3
Val-â-CH
Val-γ-CH3
Val-γ-CH3
Ser-R-CH
Pip-ꢀ-CH
Gly-NH
Gly-R-CH
Sar-NCH3
Gly-NH
Gly-R-CH
Sar-NCH3
Pip-ꢀ-CH
3.61
Val-R-CH
4.80
2.95 (w)
2.13 (w)
0.98
0.84
4.82
3.61
8.46 (w)
4.03
2.92
8.46 (w)
4.41
Sar-NCH3
2.92
2.13
Ser-â-CH2
Gly-R-CH
4.47
4.41
Val-â-CH
Gly-R-CH
4.03
3.90
Boc
Val-γ-CH3
1.40
0.98
2.92
3.61
2.13
4.80
2.95
2.13
Pip-ꢀ-CH (ax)
Val-γ-CH3
0.84
Chart 1. Luzopeptin Structures
properties detailed previously for luzopeptin (Chart 1).29 Analo-
gous to studies conducted with luzopeptin,30 the bisintercalation
of 1 has been shown to span two base pairs and requires the
adoption of a conformation in which the two chromophores are
separated by 10.1-10.2 Å. However, the X-ray and related
solution conformation of 1 consists of a more extended structure
in which the interchromophore distance is 17-19.5 Å more
consistent with bisintercalation spanning three base pairs.
Consequently, we elected to examine the conformational
properties of 1 and the related cyclic decadepsipeptide under a
variety of conditions. Since the conformation of cyclic peptides
in water may be substantially different from that observed in
nonpolar or aprotic solvents or in the solid state, we were
interested in assessing the conformational properties of 1 or 25
in water as well. However, the solubility was much too low
for such an evaluation and the progressive addition of D2O to
DMF-d7 solutions of sandramycin (0-40%) or 25 did not result
in perceptible alterations in the 1H NMR spectrum or the
detection of new or altered conformational states. Consequently,
the 1H NMR of 25 was examined in the presence of LiCl. The
addition of lithium salts in THF has been shown to affect
solubility,31 promote deaggregation,32 and alter the conforma-
tional properties of peptides.33-37 Notably, the LiCl complexed
cyclic peptides have been suggested to more accurately reflect
the conformational properties in water and to do so by disrupting
intramolecular hydrogen bonds necessarily adopted in nonpolar
or aprotic solvents.
The progressive addition of LiCl (1-40 equiv) to 25 in THF-
d8 was examined by 1H NMR. Additions of 1, 2, 5 or 10 equiv
of LiCl to the solution resulted in surprisingly little change in
the conformational properties of 25 with the original, still
dominant conformation being observed and several (>3-4)
(31) Senn, H.; Loosli, H.-R.; Sanner, M.; Braun, W. Biopolymers 1990,
29, 1387.
(32) Seebach, D.; Thaler, A.; Beck, A. K. HelV. Chim. Acta 1989, 72,
857.
(33) Thaler, A.; Seebach, D.; Cardinaux, F. HelV. Chim. Acta 1991, 74,
617, 628.
(34) Kessler, H.; Hehlein, W.; Schuck, R. J. Am. Chem. Soc. 1982, 104,
4534.
(35) Kock, M.; Kessler, H.; Seebach, D.; Thaler, A. J. Am. Chem. Soc.
1992, 114, 2676. Seebach, D.; Beck, A. K.; Bossler, H. G.; Gerber, C.;
Ko, S. Y.; Murtiashaw, C. W.; Naef, R.; Shoda, S.; Thaler, A.; Krieger,
M.; Wenger, R. HelV. Chim. Acta 1993, 76, 1564. Seebach, D.; Bossler, H.
G.; Flowers, R.; Arnett, E. M. HelV. Chim. Acta 1994, 77, 291. Review:
Seebach, D.; Beck, A. K.; Studer, A. In Modern Synthetic Methods; Ernst,
B., Leumann, C., Eds.; VCH Publishers: Weinheim, 1995; pp 3-178.
(36) Kofron, J. L.; Kuzmic, P.; Kishore, V.; Gemmecker, G.; Fesik, S.
W.; Rich, D. H. J. Am. Chem. Soc. 1992, 114, 2670.
(37) Boger, D. L.; Patane, M. A.; Zhou, J. J. Am. Chem. Soc. 1995,
117, 7357.
(29) Searle, M. S.; Hall, J. G.; Wakelin, L. P. G. Biochem. J. 1988, 256,
271. Searle, M. S.; Hall, J. G.; Denny, W. A.; Wakelin, L. P. G. Biochem.
J. 1989, 259, 433. Frey, M.-H.; Leupin, W.; Sorensen, D. W.; Denny, W.
A.; Ernst, R. R.; Wuthrich, K. Biopolymers 1985, 24, 2371. Searle, M. S.;
Hall, J. G.; Penny, W. A.; Wakelin, L. P. G. Biochemistry 1988, 27, 4340.
(30) Zhang, X.; Patel, D. J. Biochemistry 1991, 30, 4026.

(-)-Sandramycin: Total Synthesis and Characterization
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1635
Table 4. Comparative DNA Binding Properties
sandramycin luzopeptin
property
KB,a M-1
(1)
A
32
25b
3.4 × 107
1.2 × 107
5.7 × 106 2.4 × 104
(1:6.7)
(1:4.5)
(1:4.8)
(-)-unwinding [c]c 0.022
0.044-0.11
0.22
(+)-winding [c]d
0.044
a Calf thymus DNA, KB ) apparent absolute binding constant
determined by fluorescence quenching. The value in parentheses is
the agent/base pair ratio at saturating high-affinity binding and may be
considered a measure of the selectivity of binding. b Determined
indirectly by competitive binding with 1. c Agent/base pair ratio required
to unwind negatively supercoiled ΦX174 DNA (form I f form II gel
mobility, 0.9% agarose gel). d Agent/base pair ratio required to induce
complete rewinding or positive supercoiling of ΦX174 DNA (form II
f form I gel mobility, 0.9% agarose gel).
trace conformational isomers (<5%) starting to appear. Notably,
the gly-NH experienced no chemical shift change throughout
the addition, indicating its maintained participation in a tight
transannular H-bond while the D-ser-NH exhibited a more
typical 0.4-0.5 ppm downfield chemical shift. Upon addition
of 20-40 equiv of LiCl, the relative proportion of the dominant
conformation diminished and at 40 equiv no unique or discrete
conformations were identifiable. Thus, upon addition of LiCl
to 1 no discrete set of new conformations could be identified
and the agent appears to adopt a large number of additional
accessible conformations.
As detailed in subsequent studies, sandramycin adopts a
DNA-bound conformation that is substantially different than
its native X-ray or solution conformation and this preferred
conformation is inherent in the cyclic decadepsipeptides 24-
26. Only in DMSO or upon addition of LiCl (40 equiv) do
multiple alternative conformations become apparent with 25.
DNA Binding Affinity. Apparent absolute DNA binding
constants and the apparent binding site sizes for 1, luzopeptin
A, and 32 were obtained by measurement of fluorescence
quenching upon titration addition of calf thymus DNA.7 The
excitation and emission spectra for sandramycin (1), luzopeptin
A, and 32 in aqueous buffer were measured (Figure 7,
supporting information) and, for the DNA binding assays which
quantitate the fluorescence quenching, excitation outside the
absorbance range of DNA was necessarily employed (360, 340,
or 400 nm) and the more intense 530, 520, or 510 nm
fluorescence emission monitored, respectively. For assay of
the DNA-induced fluorescence quenching of the agents, a 2 mL
buffer solution of Tris-HCl (pH 7.4) and 75 mM NaCl was
employed. For titration, small aliquots of DNA were added to
solutions of the agents in the Tris-HCl buffer (pH 7.4). The
DNA quenching of fluorescence was nearly 70% with 360 nm
excitation and 530 nm fluorescence for 1, 60% with 340 nm
excitation and 520 nm fluorescence for luzopeptin A, and 40%
with 400 nm excitation and 510 nm fluorescence for 32. The
titrations were carried out with 5 min time intervals between
DNA additions to allow binding equilibration. Notable differ-
ences have not been detected with different time intervals
indicating that tight binding equilibration is rapid and the results
are summarized in Table 4. The DNA titration fluorescence
quenching was analyzed by Scatchard analysis38 with the
following equation: rb/c ) Kn - Krb where rb is the number
of agent molecules bound per DNA nucleotide phosphate, c is
the free drug concentration, K is the apparent association
constant, and n is the number of agent binding sites per
nucleotide phosphate. From a plot of rb/c versus rb as shown
in Figure 2 for 1, association constants (KB) for 1, luzopeptin
A, and 32 were derived from the slopes and the binding site
sizes determined from the intercept values (n) for the number
Figure 2. (a) Sandramycin fluorescence quenching at increasing calf
thymus DNA concentrations, excitation at 360 nm, and emission at
530 nm in 10 mM Tris-HCl (pH 7.4) and 75 mM NaCl buffer solution,
and (b) Scatchard plot of the fluorescence quenching of part a.
of binding sites per nucleotide phosphate. The results are
summarized in Table 4.
The DNA binding constant for 25 could not be established
by direct spectroscopic methods but was indirectly determined
by two complementary and self consistent techniques. The first
relied on competitive binding with 1 and inhibition of its DNA
binding derived fluorescence quenching. This was accom-
plished by titrating small aliquots of 1 (1 mM in DMSO) into
a solution of calf thymus DNA (320 µM base pair) and 25 (320
µM) in 10 mM Tris-HCl, 75 mM NaCl (pH 7.4) buffer.
Scatchard plots of the titrations conducted in the presence or
absence of various concentrations of 25 exhibited a well-defined
competitive binding culminating in a common x intercept
corresponding to the 1:6.7 agent/base pair ratio for saturated
sandramycin binding. The second relied on the displacement
of prebound ethidium bromide from calf thymus DNA and
measurement of the resulting decrease in fluorescence.39 Both
methods provided comparable binding constants of 2.4 × 104
M-1 and 4.0-8.0 × 103 M-1, respectively. The titration of
calf thymus DNA prebound with ethidium bromide relies on a
determination of the amount of 25 required to displace one-
half of the ethidium bromide established by a 50% fluorescence
(39) (a) Boger, D. L.; Invergo, B. J.; Coleman, R. S.; Zarrinmayeh, H.;
Kitos, P. A.; Thompson, S. C.; Leong, T.; McLaughlin, L. W. Chem.-Biol.
Interact. 1990, 73, 29. Boger, D. L.; Sakya, S. M. J. Org. Chem. 1992, 57,
1277. (b) Baguley, B. C.; Denny, W. A.; Atwell, G. J.; Cain, B. F. J. Med.
Chem. 1981, 24, 170. For KB of ethidium bromide: LePecq, J. B.; Paoletti,
C. J. Mol. Biol. 1967, 27, 87.
(38) Scatchard, G. Ann. N. Y. Acad. Sci. 1949, 51, 660.

1636 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Boger et al.
reduction. This titration followed a well-defined linear reduction
in fluorescence with added agent and provided comparable
estimates of the KB for 25 using either a competitive39b (4.0 ×
10-3 M-1) or noncompetitive39a (8.0 × 103 M-1) binding model.
Given the weak binding of 25, these estimates are relatively
insensitive the stoichiometry of the displacement and binding
site sizes of the agents. They are, however, subject to error if
the binding of 25 does not preclude or compete with ethidium
bromide binding. Unlike the observations actually made with
25, such a nonideal behavior would also result in a nonlinear
reduction in the ethidium bromide fluorescence, and the errors
introduced by such nonideal behavior would lead to an
underestimation of the apparent KB. Thus, the KB [(4-8) ×
103 M-1] established using this method may be best represented
as the lower limit of the binding constant of 25 and, as such,
agrees nicely with the results derived from the first method.
Sandramycin was found to exhibit an exceptionally high
affinity for duplex DNA (KB ) 3.4 × 107 M-1, ∆G° ) -10.2
kcal/mol) with a saturating stoichiometry of high affinity binding
at a 1:6.7 agent to base pair ratio. Notably, both the affinity
and apparent selectivity is enhanced with 1 exhibiting saturated
high affinity binding at a 1:6.7 agent/base pair ratio while that
of luzopeptin A or 32 was observed at 1:4.5 and 1:4.8 agent/
base pair ratios. The DNA binding constant of 1 proved to be
Figure 3. Agarose gel electrophoresis: (A) lanes 1-5, luzopeptin-A
treated supercoiled ΦX174 RFI DNA; lane 6, untreated DNA, 95%
form I and 5% form II; lanes 7-12 sandramycin-treated ΦX174 RFI
DNA. The [agent] to [DNA] base pair ratios were 0.022 (lane 1), 0.033
(lane 2), 0.044 (lane 3), 0.11 (lane 4), 0.22 (lane 5), 0 (lane 6), 0.011
(lane 7), 0.022 (lane 8), 0.033 (lane 9), 0.044 (lane 10), 0.066 (lane
11), and 0.11 (lane 12). (B) Lanes 1-3, sandramycin-treated DNA;
lane 4 untreated DNA, 95% form I and 5% form II; lanes 5-8, 32-
treated DNA; lanes 9-12, ethidium bromide-treated DNA. The [agent]
to [DNA] base pair ratios were 0.011 (lane 1), 0.033 (lane 2), 0.066
(lane 3), 0 (lane 4), 0.87 (lane 5), 2.2 (lane 6), 5.0 (lane 7), 11.0 (lane
8), 0.87 (lane 9), 2.2 (lane 10), 5.0 (lane 11), 11.0 (lane 12).
slightly higher than that of luzopeptin A (KB ) 1.2 × 107 M-1
,
∆G° ) -9.6 kcal/mol), substantially more effective than 32
(KB ) 5.7 × 106 M-1, ∆G° ) -9.2 kcal/mol) lacking one
chromophore, and much more effective than 25 (KB ) 2.4 ×
104 M-1, ∆G° ) -6.0 kcal/mol) lacking both chromophores.
Importantly, the largest share of the binding affinity is derived
from the cyclic decadepsipeptide and the addition of the first
and second chromophores incrementally increase binding by
approximately 3.2 and 1.0 kcal/mol, respectively. This is
consistent with a representation of sandramycin and the lu-
zopeptins as minor groove binding cyclic decadepsipeptides
incrementally stabilized by mono and bisintercalation.
Bifunctional Intercalation. Confirmation that 1 binds to
DNA with intercalation was derived from its ability to induce
the unwinding of negatively supercoiled ΦX174 DNA.7 This
was established by its ability to gradually decrease the agarose
gel electrophoresis mobility of supercoiled ΦX174 DNA
(unwinding) at increasing concentrations followed by a return
to normal mobility (rewinding) at even higher agent concentra-
tions. Similar types of changes have been reported for ethidium
bromide under conditions which prevent dissociation of the
bound agent during electrophoresis.40
Under the conditions employed in our study, sandramycin
completely unwound ΦX174 DNA at a 0.022 agent/base pair
ratio and luzopeptin A at a 0.044-0.11 agent/base pair ratio
(Figure 3, Table 4). Complete rewinding of the supercoiled
DNA occurred at agent/base pair ratios of 0.044 and 0.22 for 1
and luzopeptin A, respectively. These comparisons along with
the KB measurements illustrate that sandramycin binds with
either a higher unwinding angle or much slower offrate than
luzopeptin A (unwinding angle ) 43°).7
Similar evaluation of 32 lacking one of the two chromophores
revealed a behavior analogous to ethidium bromide (Figure 3B).
Under conditions where the monointercalator is not additionally
present in the agarose gel, only a slight streaking and retardation
of the electrophoretic mobility of the supercoiled ΦX174 DNA
was observed. In comparable assays where ethidium bromide
is present in the agarose gel which prevents its dissociation from
DNA, ethidium bromide completely unwound supercoiled DNA
at agent/base pair ratios of 0.0907 which is approximately 2-4×
that required for 1 or luzopeptin A. Thus, the extent of
unwinding of negatively supercoiled DNA and the subsequent
positive supercoiling of the DNA by sandramycin, like luzopep-
tin A, was found to be indicative of bisintercalation while that
of 32, like ethidium bromide, was consistent with monointer-
calation.
DNA Binding Selectivity. DNase I41 and FeII-EDTA foot-
printing42 conducted following binding of sandramycin (1) and
luzopeptin A to singly 32P-end labeled w794 and w836 DNA43
revealed that the agents behave comparably. Sandramycin more
effectively protected DNA than luzopeptin A and more clearly
revealed subtle distinctions in relative protection from DNA
cleavage. Like the preceding studies with luzopeptin,6 sandra-
mycin appears to bind best to regions containing alternating A
and T residues, although no consensus di- or trinucleotide
sequence was prominently detected. Binding at other sites is
observed and at moderate agent concentrations the DNA is
almost evenly protected from digestion. Illustrated in Figure 4
is the DNase I footprinting pattern obtained upon binding of
sandramycin to w794 DNA. As indicated in Figure 4, the
alteration of the DNase I digestion is barely perceptible at 2
µM sandramycin, is evenly diminished at the higher agent
concentrations of 10 and 20 µM relative to controls. At these
higher concentrations, some regions exhibit complete or near
complete protection indicating preferential binding at these sites.
Most, but not all such sites surround the dinucleotide sequence
5′-AT and most such sites are preceded by a 5′-C, i.e. 5′-CAT.
The majority of the remaining sites are 5′-TA and both 5′-AT
and 5′-TA appear to be preferred over 5′-AA or 5′-TT.
However, these distinctions are subtle and most all sites are
more evenly protected at even higher agent concentrations. The
footprinting studies conducted with FeII-O2 (HOCH2CH2SH) or
FeIII-H2O2 in the presence of EDTA provided comparable results
but were more difficult to conduct with sufficient control to
(41) Galas, D. J.; Schmitz, A. Nucleic Acids Res. 1978, 5, 3157.
(42) Tullius, T. D.; Dombroski, B. A.; Churchill, M. E. A.; Kam, L.
Methods Enzymol. 1987, 155, 537.
(43) Boger, D. L.; Munk, S. A.; Zarrinmayeh, H.; Ishizaki, T.; Haught,
J.; Bina, M. Tetrahedron 1991, 47, 2661.
(40) Espejo, R. T.; Lebowitz, J. Anal. Biochem. 1976, 72, 95.

(-)-Sandramycin: Total Synthesis and Characterization
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1637
complexed with short deoxyoligonucleotides confirmed minor
groove bisintercalation spanning two base pairs.8,29,30 Detailed
NMR analysis of a luzopeptin complex with 5′-d(CATG)2 has
provided its solution structure with the agent sandwiching the
central two Watson-Crick A-T base pairs and adopting a
compact conformation in which the interchromophore distance
is 10.1-10.2 Å.30 Both the pip-gly secondary amides and the
tertiary gly-sar amides adopt cis vs trans amide stereochemistries
in order to accommodate this shortened distance and the agent
maintains its 2-fold axis of symmetry. The gly-NH’s are
reoriented to form intermolecular H-bonds with the thymine C2
carbonyls and nicely explain the preference for the 5′-AT
sequence.
In efforts which confirm this same binding mode for
29
sandramycin (1), its 1:1 complex with 5′-d(GCATGC)2 was
prepared and examined in preliminary 1H NMR studies.44 Clear
from the initial inspection of the 1D 1H NMR was that complex
formation had occurred cleanly to provide a symmetrical 1:1
complex. This rules out binding at one end of the oligo which
would produce an unsymmetrical 1H NMR pattern. Moreover,
symmetrical binding in a 1:1 complex is restricted to bisinter-
calation spanning the central two or four base pairs and
necessarily ruled out an unsymmetrical three base pair bisin-
tercalation site. Since bisintercalation spanning four base pairs
is structurally unrealistic, this also further restricts the bisin-
tercalation to binding spanning the central two A-T base pairs
of 5′-d(GCATGC)2 analogous to the solution complexes ob-
served with luzopeptin A bound to both 5′-d(GCATGC)2 and
5′-d(CATG)2.8,29,30
The assignments for unbound 1 (Table 2), the free deoxyo-
ligonucleotide29 (Table 7, supporting information) and those for
the complex were obtained from a combination of 1D and 2D
1H NMR (Tables 7 and 8, supporting information) that identify
connectivity and through-space interactions.44 The position and
orientation of the bound drug in the complex were revealed by
perturbations in the 1H NMR chemical shifts and intramolecular
agent or oligonucleotide NOESY contacts and confirmed by
several key intermolecular NOESY contacts. Characteristic of
intercalation, all of the quinoline chromophore chemical shifts
in the complex are shielded relative to those of the free agent
(0.48-1.46 ppm) with the C7 and C8-H exhibiting the largest
upfield shifts of 1.03 and 1.46 ppm, respectively. Further
diagnostic of the intercalation site was the clear NOESY contacts
between the quinoline C5-H and C6-H with the cytosine2 C5-H
and a complementary quinoline C7-H and adenine3 C8-H
NOESY crosspeak. Not only does this identify the intercalation
site but it also serves to orient the chromophore at the
intercalation site and places the quinoline C4-C6 on the major
groove interface and the chromophore C7, C8, and N1 on the
minor groove side. These chemical shift perturbations and the
location of the intermolecular contacts indicate that the chro-
mophore carbocyclic ring stacks principally on the adenine base
diagnostic of intercalation between the 5′-CpA and 5′-TpG steps.
The connectivity of NOE interactions between the C6-H or
C8-H of a particular base and its own sugar H-2′ to make
intranucleotide connections and between the C6-H or C8-H and
the sugar H-2′′ of the nucleotide on the 5′ side of the sequence
to assign the residues sequentially confirmed this site of
intercalation. The characteristic adenine3 C8-H/cytosine2 H-2′′
and thymine4 H-2′′/guanine5 C8-H NOESY contacts were
Figure 4. DNase footprinting of sandramycin (1) bound to w794 DNA.
Lanes 1-4, G, C, A and T Sanger sequencing reactions; lane 5, control
DNA; lanes 6-8, 2, 10, and 20 µM sandramycin with DNase I treatment
(1 min); lanes 9 and 10, DNase I treatment of w794 DNA alone for 1
and 2 min.
detect the subtle selectivities. Under our conditions, little ligand
induced enhancements of DNase I cleavage was observed in
either GC or AT-rich regions although this was reported to be
observed with the luzopeptins.6
Binding to 5′-d(GCATGC)2. Similar to the results sum-
marized above, prior footprinting studies with luzopeptin
established that the agent binds best at regions containing
alternating A-T base pairs although no consensus di- or
trinucleotide sequence was established.6 Further studies of the
luzopeptins with short DNA fragments (15-35 base pairs)
indicated very strong incremental binding of the agent with
saturated binding at about four base pairs per agent, implying
it is capable of binding to all sequences with bisintercalation
spanning two base pairs.6 Although the native X-ray structure
and solution structures of 1, luzopeptin, or the related cyclic
decadepsipeptides place the intercalating chromophores 17-
19 Å apart and provides the potential for bisintercalation
spanning three base pairs, recent NMR studies of luzopeptin
(44) Boger, D. L.; Saionz, K. Unpublished studies. Tables 6 and 7
(supporting information) provide the nucleic acid and sandramycin proton
chemical shift assignments in the d(GCATGC)2-sandramycin complex
along with their comparisons with free agent, free DNA, and the analogous
luzopeptin A complexes taken from refs 29 and 30. Full refinements of
this structure are in progress and will be disclosed in full detail in due
time.

1638 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Boger et al.
Figure 5. Three views of the 5′-d(GCATGC)2-sandramycin complex illustrating the symmetrical minor groove binding of the cyclic decadepsipeptide
(top and bottom) and the bisintercalation sandwiching the central two A-T base pairs (middle).
interrupted and absent in the complex while the intense NOE
linking thymine4 C5-CH3 and adenine3 C8-H was unperturbed.
Similarly, the guanine1 H-2′′/cytosine2 C6-H, adenine3 H-2′′/
thymine4 C6-H, and guanine5 H-2′′/cytosine6 C6-H NOE
connectivity were unperturbed. In addition, the connectivities
between the base protons and the sugar H-1′ could be used to
trace the chain by monitoring the NOEs between the base C8-H
or C6-H protons and their own and the 5′ flanking sugar H-1′
protons. This also clearly established bisintercalation with
chromophore insertion between the 5′CpA and 5′-TpG sites.
Notably absent in the complex were the adenine3 C8-H/cytosine2
H-1′ as well as the guanine5 C8-H/thymine4 H-1′ NOEs while
the interresidue cytosine2 C6-H/guanine1 H-1′ thymine4 C6-H/
adenine3 H-1′, cytosine6 C6-H/guanine5 H-1′ NOEs were
unperturbed in the complex. Additional quinoline-nucleic acid
base NOE contacts and a rich array of cyclic decadepsipeptide-
nucleic acid NOE contacts clearly indicate the minor groove
binding in addition to the intercalation orientation analogous
to the complexes of luzopeptin which have been described in
detail.29,30 The cyclic decadepsipeptide is positioned in the
minor groove of the duplex and, like luzopeptin, most likely
adopts a conformation in which the pip-gly and gly-sar amides

(-)-Sandramycin: Total Synthesis and Characterization
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1639
Table 5. In Vitro Cytotoxic Activitya
were substantially more effective than 32 lacking one chro-
mophore and much more effective than 25 lacking both
chromophores. The largest share of the binding affinity is
derived from the cyclic decadepsipeptide (∆G° ) -6.0 kcal/
mol) and the addition of each chromophore was found to
incrementally increase the affinity by approximately 3.2 and
1.0 kcal/mol, respectively. This is consistent with its repre-
sentation as cyclic decadepsipeptide minor groove binding agent
incrementally stabilized by mono and bisintercalation. Studies
of the unwinding of supercoiled DNA and its subsequent
rewinding confirmed bisintercalation binding. DNase I foot-
printing studies revealed that sandramycin and luzopeptin A
behave similarly and appear to bind best to regions containing
alternating A and T residues. Binding at other and perhaps all
sites is observed at modest agent concentrations although a
perceptible preference for 5′-CAT was noted. Preliminary
studies of the 1:1 complex of sandramycin with 5′-d(GCATGC)2
revealed that it forms a complex analogous to that observed
with luzopeptin A. The agent sandwiches the central two
Watson-Crick A-T base pairs and adopts a compact conforma-
tion in which the interchromophore distance is 10.1 Å (vs 17-
19 Å). This suggests that the relatively low contribution to the
binding affinity that is attributable to the second intercalation
is due to an accompanying destabilizing conformational change
in the cyclic decadepsipeptide that offsets much of the gains
derived from the second intercalation. The cyclic decadep-
sipeptide is positioned in the minor groove and adopts a compact
conformation that permits a rich array of complementary
hydrophobic contacts extending over much of the interacting
surface.
IC50, nM
agent
Molt-4
L1210
0.02
0.02
786-0
0.2
4
120
Ovcar-3
B16
luzopeptin A
sandramycin
0.8
0.8
6
2
60
0.07
0.4
8
nt
nt
nt
nt
nt
27
32
25
24
21
15
4
400
20-2
500
nt
nt
>105
>105
>105
5000
>105
>105
>105
>105
80 000
80 000
50 000
80 000
80 000
80 000
60 000
50 000
a Molt-4 (human T-cell leukemia), L1210 (mouse leukemia), 786-0
(human perirenal cell carcinoma), Ovcar-3 (human ovarian carcinoma),
B16 (melanoma).
are cis to accommodate the bisintercalation sandwiching the
central A-T base pairs.30 Complementary intermolecular hy-
drophobic contacts between the agent and DNA extend over
much of the interacting surface. Although our studies are
preliminary and not yet refined,44 they established that sandra-
mycin and luzopeptin interact with 5′-d(GCATGC)2 in an
analogous fashion. Illustrated in Figure 5 is a model of the
5′-(GCATGC)2 complex with sandramycin constructed on the
basis of the structure of the luzopeptin complex8,30 established
by Patel and Zhang which highlights the minor groove binding
and bisintercalation spanning the central two A-T base pairs.
In Vitro Cytotoxic Activity. Table 5 summarizes the
comparison in vitro cytotoxic activities of luzopeptin A, the most
potent of the naturally occurring luzopeptins, and sandramycin
alongside that of the key partial structures including its bis
benzyl ether 27, the cyclic decadepsipeptide 32 possessing a
single chromophore, the cyclic decadepsipeptides 24 and 25
lacking both chromophores, the linear decadepsipeptide 21, and
the linear pentadepsipeptide 15. Consistent throughout the five
assays, luzopeptin A and sandramycin exhibit comparable and
exceptionally potent cytotoxic activity (6-0.02 nM). The bis
benzyl ether 27 was generally 20-100× less potent than 1 and
the results represent a consistent general observation that
alkylation of the quinoline C3 phenol diminishes biological
potency.45 However, the impact of the introduction of this bulky
benzyl ether is smaller than anticipated but consistent with the
subsequent finding that the removal of the phenol altogether
results in little change in the cytotoxic activity.45 Consequently,
the diminished properties of 27 most likely may be attributed
to the introduction of unfavorable steric interactions which
diminish the agents intercalation capabilities rather than lost
H-bonding capabilities. The agent 32 possessing a single
chromophore proved to be approximately 500-1000× less
potent than 1 and the cyclic decadepsipeptides 24 and 25 lacking
both chromophores were inactive and g105× less potent than
1. Similarly, the linear decadepsipeptide 21 and pentadepsipep-
tide 15 were inactive.
Experimental Section
Boc-Gly-Sar-OMe (2). A solution of Boc-Gly-OH (2.70 g, 15.4
mmol) and the HCl salt of H2N-Sar-OMe (2.15 g, 15.4 mmol) in CH2-
Cl2 (50 mL) was treated sequentially with Et3N (2.2 mL, 15.8 mmol),
DCC (3.20 g, 15.5 mmol), and DMAP (306 mg, 2.5 mmol), and the
reaction mixture was stirred at 25 °C for 20 h. A white precipitate
formed in the first 10 min and was removed by filtration at the end of
the reaction. The filtrate was concentrated in vacuo. Flash chroma-
tography (SiO2, 5 × 16 cm, 40% EtOAc-hexane eluent) afforded 2
(3.21 g, 4.01 g, theoretical, 80%) as a colorless oil which solidified on
standing: mp 72-73 °C (EtOAc-hexane, colorless cubes); Rf ) 0.32
1
(50% EtOAc-hexane); H NMR (CDCl3, 400 MHz) (4:1 mixture of
two conformers, for the major conformer) δ 5.44 (s, 1H), 4.14 (s, 2H),
4.02 (d, 2H, J ) 4.3 Hz), 3.73 (s, 3H), 3.02 (s, 3H), 1.43 (s, 9H); 13C
NMR (CDCl3, 100 MHz) for major conformer: δ 169.3, 168.7, 155.7,
79.6, 52.2, 49.4, 42.2, 35.2, 28.3; IR (KBr) νmax 3419, 2978, 2934,
1754, 1715, 1667, 1488, 1424, 1367, 1249, 1208, 1175, 1120, 1051,
952, 871, 764, 712 cm-1; FABHRMS (NBA-NaI) m/z 283.1259 (M
+ Na+, C11H20N2O5 requires 283.1270).
Anal. Calcd for C11H20N2O5: C, 50.75; H, 7.74; N, 10.76. Found:
C, 50.96; H, 7.62; N, 10.63.
Boc-Gly-Sar-OH (3). Lithium hydroxide monohydrate (598 mg,
14.3 mmol) was added to a solution of 2 (1.22 g, 4.7 mmol) in 20 mL
of THF-CH3OH-H2O (3:1:1) at 25 °C and the resulting reaction
mixture was stirred for 3 h. The reaction mixture was poured onto 3
M aqueous HCl (10 mL) and extracted with EtOAc (3 × 20 mL). The
combined organic phases were dried (Na2SO4), filtered, and concen-
trated in vacuo to give 314 (1.16 g, 1.16 g theoretical, 100%) as a
colorless oil. This acid was identical to authentic material14 and was
used directly in the next step without further purification: 1H NMR
(CDCl3, 400 MHz)14 δ 5.76 and 5.66 (two br s, 1H), 4.11, 4.00, and
3.91 (three s, 4H), 2.99 amd 2.95 (two s, 3H), 1.38 (s, 9H); IR (neat)
νmax 3348, 2979, 2937, 1717, 1654, 1691, 1409, 1368, 1287, 1252,
Conclusions. A concise and efficient total synthesis of
sandramycin amendable to the preparation of analogs was
reported which served to confirm the structure and stereochem-
istry of the natural product and provided key partial structures.
Preliminary studies of the DNA binding properties revealed that
sandramycin possesses a DNA binding constant slightly greater
than that of luzopeptin A, binds to DNA with a higher selectivity
than luzopeptin A (saturated binding at a 1:6.7 vs 1:4.5 agent/
base pair ratio), and induces the unwinding of negatively
supercoiled ΦX174 DNA and its rewinding or positive super-
coiling characteristic of bisintercalation at lower agent concen-
trations than luzopeptin A. Both sandramycin and luzopeptin
bind at least 10× more tightly than echinomycin7 and exhibit
extraordinarily slow off rates.8 Sandramycin and luzopeptin A
1167, 1053, 1030, 954, 866, 782, 736 cm-1
.
Boc-Gly-Sar-NMe-Val-OMe (6). A solution of 314 (1.81 g, 7.4
mmol) and the HCl salt of 515 (1.34 g, 7.4 mmol) in CH2Cl2 (40 mL)
was treated sequentially with Et3N (1.1 mL, 7.9 mmol, 1.05 equiv),
DCC (1.52 g, 7.4 mmol), and DMAP (93 mg, 0.76 mmol, 0.1 equiv),
and the reaction mixture was stirred at 25 °C for 24 h. A white
(45) Boger, D. L.; Chen, J.-H. Unpublished studies.

1640 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Boger et al.
precipitate formed in the first 15 min and was removed by filtration at
the end of the reaction. The filtrate was concentrated in vacuo. Flash
chromatography (SiO2, 4 × 16 cm, 50% EtOAc-hexane eluent)
afforded 6 (2.04 g, 2.75 g theoretical, 74%) as a colorless oil: Rf )
1340, 1246, 1154, 1091, 1045, 1002, 930, 873, 783, 752, 697 cm-1
;
FABHRMS (NBA-NaI) m/z 342.1672 (M + Na+, C18H25NO4 requires
342.1681).
Method B. A solution of N-BOC-Pip-OH16 (9, 1.28 g, 5.6 mmol)
and benzyl alcohol (1.05 g, 9.7 mmol, 1.7 equiv) in CH2Cl2 (20 mL)
was cooled to -30 °C and sequentially treated with DMAP (68.3 mg,
0.56 mmol, 0.1 equiv) and DCC (1.16 g, 5.6 mmol, 1.0 equiv). The
resulting mixture was stirred at -30 °C under Ar for 20 h. The white
precipitate of DCU was removed by filtration, and the filtrate was
concentrated in vacuo. Chromatography (SiO2, 4 × 16 cm, 1:15
EtOAc-hexane eluent) afforded 10 (1.74 g, 1.79 g theoretical, 97%)
0.22 (66% EtOAc-hexane); [R]23 -62 (c 2.6, CHCl3); 1H NMR
D
(CDCl3, 400 MHz) mixture of multiple conformers, δ 5.42 (br s, 1H),
4.82 (d, 0.6H, J ) 10.6 Hz), 4.37-4.00 (m, 4H), 3.79 (d, 0.4 H, J )
10.9 Hz), 3.72-3.67 (three s, 3H), 3.00-2.84 (six s, 6H), 2.30-2.10
(m, 1H), 1.41 and 1.40 (two s, 9H), 0.94 and 0.84 (two d, 6H, J ) 6.6
Hz); IR (neat) νmax 3421, 2969, 2934, 1740, 1712, 1655, 1485, 1404,
1366, 1291, 1251, 1204, 1170, 1051, 1617, 952, 870, 835, 781 cm-1
;
CIHRMS (isobutane) m/z 374.2303 (C17H31N3O6 requires 374.2291).
Boc-Gly-Sar-NMe-Val-OBn (20). A solution of 3 (4.65 g, 18.9
mmol) and the HCl salt of 1925 (4.87 g, 18.9 mmol) in CH2Cl2 (100
mL) was treated sequentially with Et3N (3 mL, 21.5 mmol, 1.1 equiv),
DMAP (1.15 g, 9.4 mmol, 0.5 equiv), and DCC (3.90 g, 18.9 mmol),
and the reaction mixture was stirred at 25 °C for 24 h. A white
precipitate formed during the reaction and was removed by filtration.
The filtrate was concentrated in vacuo. Flash chromatography (SiO2,
6 × 20 cm, 50% EtOAc-hexane eluent) afforded 20 (7.21 g, 8.49 g
theoretical, 85%) as a white crystalline solid which was further
as a white solid: mp 51-53 °C; [R]23 -46 (c 2.7, CHCl3); identical
D
in all respects to the material above.
A sample of 10 (6.73 g, 21.1 mmol) in a 100 mL round-bottom
flask was treated with 3 M HCl-EtOAc (40 mL, 120 mmol, 5.7 equiv).
The resulting mixture was stirred at 25 °C for 30 min. The volatiles
were removed in vacuo. The residual HCl was further removed by
adding Et2O (40 mL) to the hydrochloride salt of 11 followed by its
removal in vacuo. After repeating this procedure three times, 5.38 g
of the hydrochloride salt of 11 (5.39 g theoretical, 100%) was obtained.
The hydrochloride salt of 11 was neutralized with saturated aqueous
NaHCO3 (50 mL) and extracted with EtOAc (3 × 100 mL). The
combined organic layers was dried (Na2SO4), filtered, and concentrated
recrystallized from EtOAc-hexane; mp 97-99 °C; Rf ) 0.21 (50%
1
EtOAc-hexane); [R]23 -63 (c 0.8, CHCl3); H NMR (CDCl3, 400
D
MHz) δ 7.32 (m, 5H), 5.43 (br s, 1H), 5.16, 5.15, 5.13 (3s, 2H), 4.88
(d, 0.7H, J ) 10.4 Hz), 4.41, 4.30, 4.11, 4.05 (4d, 2H, J ) 16 Hz),
4.00 (dd, 1.3H, J ) 1.7, 4.3 Hz), 3.96 (d, 0.7H, J ) 4.3 Hz), 3.84 (d,
0.3H, J ) 10.4 Hz), 2.98, 2.93, 2.89 (3s, 3H), 2.90, 2.85, 2.83 (3s,
3H), 2.33-2.15 (m, 1H), 1.43, 1.42, 1.41 (3s, 9H), 0.95, 0.91, 0.84
(3d, 6H, J ) 6.6 Hz); 13C NMR for the major rotamer (CDCl3, 100
MHz) δ 170.6, 169.1, 168.3, 155.7, 135.5, 128.5, 128.3, 128.2, 79.5,
66.5, 61.9, 49.6, 42.2, 35.3, 30.6, 28.3, 27.5, 19.6, 19.0; IR (KBr) νmax
3337, 2973, 1736, 1706, 1664, 1534, 1473, 1394, 1296, 1249, 1186,
in vacuo to give 11 (4.62 g, 4.61 g theoretical, 100%) as white
1
crystalline plates: mp 146-148 °C; [R]23 -23.3 (c 0.7, CHCl3); H
D
NMR (CDCl3, 400 MHz) δ 7.33 (m, 5H), 5.24 (d, 1H, J ) 12.2 Hz),
5.18 (d, 1H, J ) 12.2 Hz), 3.97 (dd, 1H, J ) 4.0, 10.0 Hz), 3.56 (ddd,
1H, J ) 4.2, 4.5, 12.9 Hz), 3.06 (ddd, 1H, J ) 3.4, 10.1, 12.9 Hz),
2.27-2.21 (m, 1H), 2.15-2.06 (m, 1H), 2.01-1.97 (m, 1H), 1.84-
1.73 (m, 2H), 1.60-1.52 (m, 1H); 13C NMR (CDCl3, 100 MHz) δ
168.2, 134.6, 128.7, 128.6, 128.4, 68.0, 56.3, 43.7, 25.6, 21.6, 21.5;
IR (KBr) νmax 3347, 2934, 2853, 1737, 1453, 1257, 1179, 1126, 1052,
749, 698 cm-1; FABHRMS (NBA-NaI) m/z 220.1345 (M + H+,
C13H17NO2 requires 220.1338).
1052, 955, 742, 703 cm-1
.
Anal. Calcd for C23H35N3O6: C, 61.45; H, 7.85; N, 9.35. Found:
C, 61.44; H, 7.81; N, 9.23.
N-SES-D-Ser-OBn (12). Solution of D-serine benzyl ester (4.38 g,
22.4 mmol) and Et3N (3.2 mL, 23.0 mmol) in 90 mL of degassed
anhydrous DMF at -30 °C was treated slowly with trimethylethane-
sulfonyl chloride (4.50 g, 22.4 mmol). The reaction mixture was stirred
at -30 °C under Ar for 9 h and poured onto 100 mL of H2O and
extracted with EtOAc (3 × 150 mL). The combined organic layers
were washed with saturated aqueous NaCl (150 mL), dried (Na2SO4),
filtered, and concentrated in vacuo. Flash chromatography (SiO2, 5 ×
20 cm, 20-40% EtOAc-hexane gradient) to afford 12 (6.84 g, 8.05 g
Boc-Gly-Sar-NMe-Val-OH (7). From 6. Lithium hydroxide mono-
hydrate (249 mg, 5.9 mmol) was added to a solution of 6 (740 mg,
1.98 mmol) in 15 mL of THF-CH3OH-H2O (3:1:1) at 25 °C and the
resulting reaction mixture was stirred for 3 h. The reaction mixture
was poured onto 3 M aqueous HCl (8 mL) and extracted with EtOAc
(3 × 20 mL). The combined organic phases were dried (Na2SO4),
filtered, and concentrated in vacuo to give 7 (638 mg, 712 mg
theoretical, 90%) as a white solid which was employed directly in the
next reaction without further purification: white foam, mp 57-60 °C;
1H NMR (CDCl3, 400 MHz) δ 5.67 and 5.58 (two s, 1H), 4.63 (d, 1H,
J ) 10.4 Hz), CH), 4.10-3.82 (m, 4H), 3.05, 3.04, 3.01, and 2.88
(four s, 6H), 2.30-2.20 (m, 1H), 1.43 and 1.41 (two s, 9H), 1.04 and
0.88 (two d, 6H, J ) 6.7 Hz); IR (KBr) νmax 3421, 2974, 1706, 1656,
theoretical, 85%) as a colorless oil: [R]23 -2.2 (c 1.5, CHCl3); Rf )
D
0.48 (SiO2, 50% EtOAc-hexane); 1H NMR (CDCl3, 400 MHz) δ 7.35
(m, 5H), 5.41 (d, 1H, J ) 8.5 Hz), 5.22 (s, 2H), 4.24 (dt, 1H, J )
11.2, 3.4 Hz), 4.00 (dd, 1H, J ) 11.2, 3.8 Hz), 3.93 (dd, 1H, J ) 11.2,
3.4 Hz), 3.00-2.90 (m, 2H), 1.10-0.98 (m, 2H), 0.01 (s, 9H); 13C
NMR (CDCl3, 50 MHz) δ 171.0, 135.5, 129.2, 129.1, 128.8, 68.1, 64.4,
58.4, 50.4, 10.5, -1.9; IR (neat) νmax 3504, 3288, 2954, 1742, 1498,
1495, 1419, 1367, 1292, 1250, 1171, 1053, 953, 870, 837, 670 cm-1
.
From 20. A solution of 20 (2.62 g, 5.85 mmol) in 40 mL of CH3-
OH was treated with 10% Pd-C (300 mg) and the resulting black
suspension was stirred at 25 °C under H2 (1 atm) for 16 h. The catalyst
was removed by filtration through Celite, and the filtrate was
concentrated in vacuo to give 7 (2.14 g, 2.10 g theoretical, 100%) as
a white foam: [R]23D -63.3 (c 1.1, CHCl3); identical in all respects to
the material above.
1330, 1252, 1174, 1130, 1070, 966, 894, 862, 842, 738, 698 cm-1
;
FABHRMS (NBA) m/z 359.1220 (C15H25NO5SiS requires 359.1223).
N-SES-D-Ser-OH (13). A solution of 12 (1.05 g, 2.91 mmol) in
CH3OH (20 mL) was treated with 10% Pd-C (100 mg). The resulting
black suspension solution was stirred under H2 (1 atm) at 25 °C for 12
h. The catalyst was removed by filtration through Celite, and the filtrate
was concentrated in vacuo to give 13 (785 mg, 784 mg theoretical,
Benzyl L-Pipecolate (11). Method A. A solution of 916 (2.96 g,
12.9 mmol) in CH2Cl2 (60 mL) was treated sequentially with saturated
aqueous NaHCO3 (40 mL), Bu4NI (4.76 g, 12.9 mmol, 1.0 equiv) and
benzyl bromide (3.31 g, 19.4 mmol, 1.5 equiv). The resulting mixture
was stirred at 25 °C under N2 for 24 h. The reaction mixture was
extracted with CH2Cl2 (3 × 100 mL). The combined organic layers
were dried (Na2SO4), filtered, and concentrated in vacuo. Chroma-
tography (SiO2, 5 × 18 cm, 1:15 EtOAc-hexane eluent) afforded 10
100%) as a white solid: mp 61-63 °C; [R]23 -2.1 (c 2.2, CHCl3);
D
1H NMR (CDCl3, 400 MHz) δ 6.12 (d, 1H, J ) 8.7 Hz), 5.42 (br s,
2H), 4.20 (d, 1H, J ) 8.7 Hz), 4.11 (d, 1H, J ) 10.2 Hz), 3.92 (d, 1H,
J ) 10.2 Hz), 3.05-2.96 (m, 2H), 1.10-0.98 (m, 2H), 0.04 (s, 9H);
13C NMR (CDCl3, 100 MHz) δ 173.3, 64.2, 57.7, 50.2, 10.2; IR (KBr)
ν
max 3416, 3313, 2956, 1740, 1321, 1252, 1177, 1120, 1023, 843, 759,
741, 700 cm-1; FABHRMS (NBA-NaI) m/z 292.0663 (M + Na+,
(3.71 g, 4.12 g theoretical, 90%) as a white solid: mp 51-53 °C; [R]23
D
C8H19NO5SiS requires 292.0651).
-48 (c 3.4, CHCl3); Rf ) 0.49 (10% EtAOc-hexane); 1H NMR revealed
a 1:1 mixture of two conformers, 1H NMR (CDCl3, 400 MHz) δ 7.32
(m, 5H), 5.21 (s, 2H), 4.94 (br s, 0.5H), 4.74 (br s, 0.5H), 4.01 (d,
0.5H, J ) 12.0 Hz), 3.90 (d, 0.5H, J ) 12.0 Hz), 2.91 (m, 1H), 2.22
(m, 1H), 1.70-1.10 (m, 5H), 1.44 (s, 4.5H), 1.36 (s, 4.5H); 13C NMR
(CDCl3, 100 MHz) δ 172.0, 171.8, 156.0, 155.4, 135.8, 128.5, 128.2,
128.1, 127.9, 79.9, 66.6, 54.9, 53.8, 42.1, 41.1, 28.3, 28.2, 26.7, 24.8,
24.5, 20.8, 20.6; IR (KBr) νmax 2941, 2861, 1734, 1700, 1454, 1364,
Anal. Calcd for C8H19NO5SiS: C, 35.67; H, 7.11; N, 5.20; S, 11.90.
Found: C, 35.93; H, 6.96; N, 5.39; S, 12.24.
N-SES-D-Ser-Pip-OBn (14). A solution of 1116 (1.27 g, 5.77 mmol,
1.3 equiv) and 13 (1.23 g, 4.58 mmol) in CH2Cl2 (20 mL) was cooled
to 0 °C and sequentially treated with Et3N (1.90 mL, 13.6 mmol, 3.0
equiv) and bis(2-oxo-3-oxazolidinyl)phosphinic chloride (BOP-Cl, 1.62
g, 6.36 mmol, 1.40 equiv), and the resulting reaction mixture was stirred

(-)-Sandramycin: Total Synthesis and Characterization
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1641
at 0 °C for 10 h. The reaction mixture was diluted with CH2Cl2 (50
mL) and washed sequentially with 10% aqueous HCl (30 mL), H2O
(30 mL), saturated aqueous NaHCO3 (30 mL), and saturated aqueous
NaCl (30 mL). The organic layer was dried (Na2SO4), filtered, and
concentrated in vacuo. Flash chromatography (SiO2, 4 × 16 cm, 40%
EtOAc-hexane eluent) afforded 14 (1.83 g, 2.15 g theoretical, 85%)
and 25 °C (24 h). The reaction mixture was poured onto H2O (20
mL) and extracted with EtOAc (3 × 40 mL). The combined organic
phase was washed with saturated aqueous NaCl (20 mL), dried (Na2-
SO4), filtered and concentrated in vacuo. Flash chromatography (SiO2,
2 × 16 cm, 80-100% EtOAc-CH2Cl2 gradient elution) afforded 21
(2.28 g, 2.84 g theoretical, 80%) as a glassy solid: Rf ) 0.6 (5% CH3-
CN-EtOAc); [R]23D -124 (c 0.9, CHCl3); 1H NMR (CDCl3, 400 MHz)
δ 7.36-7.26 (m, 5H), 5.70-5.50 (m, 2H), 5.30-5.05 (m, 4H), 4.80-
3.60 (m, 16H), 3.40-3.20 (m, 2H), 3.10-2.70 (m, 16H), 2.35-2.10
(m, 4H), 1.85-1.20 (m, 21H), 1.05-0.80 (two d and m, 16H, J ) 6.2
Hz and 6.5 Hz), 0.01 to -0.08 (m, 18H); IR (KBr) νmax 3240, 2954,
1740, 1655, 1482, 1456, 1415, 1318, 1287, 1250, 1169, 1021, 841,
739 cm-1; FABHRMS (NBA-NaI) m/z 1437.6599 (M + Na+,
C62H106N10O19Si2S2 requires 1437.6531).
as a white crystalline solid: mp 105-106 °C; Rf ) 0.35 (50% EtOAc-
1
hexane); [R]23 -72 (c 1.1, CH2Cl2); H NMR (CDCl3, 400 MHz) δ
D
7.39-7.31 (m, 5H), 5.55 (d, 1H, J ) 8.7 Hz), 5.30 (d, 1H, J ) 5.2
Hz), 5.20 (d, 1H, J ) 12.3 Hz), 5.09 (d, 1H, J ) 12.3 Hz), 4.50 (m,
1H), 3.79-3.68 (m, 2H), 3.30 (dt, 1H, J ) 3.0, 13.1 Hz), 2.95-2.87
(m, 2H), 2.66 (m, 1H), 2.31 (d, 1H, J ) 14.2 Hz), 1.76-1.18 (m, 6H),
1.06-0.98 (m, 2H), 0.03 (s, 9H); 13C NMR (CDCl3, 100 MHz) δ 170.1,
169.9, 135.3, 128.6, 128.4, 128.0, 67.1, 64.3, 55.4, 53.0, 49.5, 43.6,
26.3, 25.0, 20.7, 10.1, -2.1; IR (KBr) νmax 3466, 3270, 2950, 2860,
(N-SES-D-Ser-Pip-Gly-Sar-NMe-Val)2 (Serine Hydroxyl) Dilac-
tone (24). A solution of 21 (1.69 g, 1.14 mmol) in CH3OH (20 mL)
was treated with 10% Pd-C (200 mg), and the black suspension was
stirred at 25 °C under an atmosphere of H2 (1 atm) for 16 h. The
catalyst was removed by filtration through Celite, and the filtrate was
concentrated in vacuo to give crude 22 (1.47 g, 1.51 g theoretical, 97%).
Crude 19 was treated with 3 M HCl-EtOAc (10 mL), and the mixture
was stirred at 25 °C for 30 min. The volatiles were removed in vacuo,
and the excess HCl was removed by suspending the hydrochloride salt
in Et2O (30 mL) followed by its removal in vacuo. After this procedure
was repeated three times, 1.41 g (1.40 g theoretical, 100%) of the
hydrochloride salt 23 was obtained and used in the next step without
further purification.
1738, 1643, 1418, 1322, 1250, 1162, 1143, 1017, 843, 738, 698 cm-1
;
FABHRMS (NBA) m/z 471.1985 (M+ + H, C21H34N2O6SiS requires
471.1985).
Anal. Calcd for C21H34N2O6SiS: C, 53.59; H, 7.28; N, 5.95; S,
6.81. Found: C, 53.68; H, 7.19; N, 6.11; S, 6.83.
N-SES-D-Ser[Boc-Gly-Sar-NMe-Val]-Pip-OBn (15). A solution
of 14 (2.76 g, 5.85 mmol) and 7 (2.10 g, 5.86 mmol) in CH2Cl2 (40
mL) was cooled to 0 °C and sequentially treated with DMAP (0.71 g,
5.86 mmol, 1.0 equiv) and DCC (1.21 g, 5.86 mmol, 1.0 equiv), and
the resulting reaction mixture was stirred at 0 °C for 24 h. The white
precipitate that formed was removed by filtration, and the filtrate was
concentrated in vacuo. Flash chromatography (SiO2, 4 × 16 cm, 50%
EtOAc-hexane eluent) afforded 15 which was separated into two
isomers. The major isomer constitutes the desired product 15 (3.75 g,
79%, typically 79-89%) and the minor isomer constitutes the Val R-CH
epimerized product (300 mg, 6%).
For the Major Diastereomer 15: White foam; mp 68-72 °C; Rf
) 0.44 (67% EtOAc-hexane); [R]23D -110 (c 2.0, CHCl3); 1H NMR
(CDCl3, 400 MHz) δ 7.67 (d, 0.4 H, J ) 9.6 Hz), 7.34-7.27 (m, 5H),
5.69-5.49 (m, 1.6H), 5.29 and 5.22 (two d, 1H, J ) 6.7 Hz), 5.21-
4.90 (m, 3H), 4.79 (d, 1H, J ) 10.7 Hz), 4.75-4.35 (m, 2H), 4.15-
3.55 (m, 4H), 3.35-3.20 (m, 1H), 3.05-2.70 (four s and a set of
multiplets, 8H), 2.32-2.15 (m, 2H), 1.75-1.15 (m, 15H), 1.05-0.82
(m and three d, 8H, J ) 6.4 Hz), -0.05 to -0.08 (several s, 9H); IR
(KBr) νmax 3223, 2956, 1740, 1708, 1658, 1485, 1416, 1325, 1250,
1168, 1018, 841 cm-1; FABHRMS (NBA) m/z 812.9362 (M+ + H,
C37H61N5O11SiS requires 812.3936).
A solution of the hydrochloride salt 23 (1.41 g, 1.11 mmol) in
degassed DMF (370 mL) cooled to 0 °C and sequentially treated with
NaHCO3 (933 mg, 11.1 mmol, 10 equiv) and diphenyl phosphorazidate
(DPPA, 0.86 mL, 4.45 mmol, 4.0 equiv), and the reaction mixture was
stirred at 0 °C for 72 h. The mixture was concentrated in vacuo, and
the residue was diluted with EtOAc (100 mL). The organic phase was
washed with 10% aqueous HCl (50 mL), H2O (50 mL), saturated
aqueous NaHCO3 (50 mL) and saturated aqueous NaCl (50 mL), dried
(Na2SO4), filtered, and concentrated in vacuo. Flash chromatography
(SiO2, 2 × 16 cm, 10% CH3CN-EtOAc eluent) afforded 24 (1.21 g,
1.36 g theoretical, 89%, typically 85-90%) as a white powder: mp
185-188 °C dec; Rf ) 0.5 (5% CH3CN-EtOAc); [R]23D -88 (c 0.85,
CHCl3); 1H NMR (CDCl3, 400 MHz) δ 8.38 (d, 2H, J ) 4.5 Hz, Gly-
NH), 5.79 (d, 2H, J ) 6.7 Hz, D-Ser-NH), 5.30 (d, 2H, J ) 16.7 Hz,
Sar-R-CH), 5.25 (d, 2H, J ) 4.6 Hz, Pip-R-CH), 4.78 (d, 2H, J ) 10.9
Hz, Val-R-CH), 4.63 (d, 4H, J ) 8.6 Hz, D-Ser-R-CH and â-CH), 4.40
(d, 2H, J ) 10.4 Hz, D-Ser-â-CH), 4.38 (dd, 2H, J ) 5.6, 18.0 Hz,
Gly-R-CH), 3.99 (d, 2H, J ) 18 Hz, Gly-R-CH), 3.90 (dd, 2H, J )
10.8, 12.6 Hz, Pip-ꢀ-CH), 3.55 (d, 2H, J ) 13.4 Hz, Pip-ꢀ-CH), 3.42
(d, 2H, J ) 16.7 Hz, Sar-R-CH), 2.94 (s, 6H, NCH3), 2.91 (s, 6H,
NCH3), 2.98-2.82 (m, 4H, SO2CH2), 2.16-2.07 (d split septet, 2H, J
) 10.3, 6.7 Hz, Val-â-CH), 1.76-1.36 (m, 12H, Pip-(CH2)3), 1.04-
0.94 (m, 4H, SO2CH2CH2), 0.95 (d, 6H, J ) 6.7 Hz, Val-γ-CH3), 0.84
(d, 6H, J ) 6.7 Hz, Val-γ-CH3), 0.03 (s, 18H, SiMe3); 13C NMR
(CDCl3, 100 MHz) δ 172.2, 169.3, 169.2, 167.7, 166.6, 65.2, 62.3,
53.7, 52.9, 49.5, 49.3, 44.0, 41.9, 35.0, 30.3, 28.4, 26.8, 24.6, 20.0,
19.3, 19.1, 10.2, -2.0; IR (KBr) νmax 3330, 2953, 2871, 1743, 1644,
Anal. Calcd for C37H61N5O11SiS: C, 54.72; H, 7.57; N, 8.62; S,
3.95. Found: C, 55.00; H, 7.65; N, 8.70; S, 4.13.
For the Minor Isomer: White foam; mp 72-76 °C Rf ) 0.35 (67%
1
EtOAc-hexane); [R]23 -36 (c 0.15, CHCl3); H NMR (CDCl3, 400
D
MHz) δ 7.35-7.25 (m, 5H), 5.86 (d, 1H, J ) 9.2 Hz), 5.68-5.61 (m,
1H), 5.28-5.00 (m, 3H), 4.82 (d, 1H, J ) 10.8 Hz), 4.67 (m, 1H),
4.50-3.80 (m, 6H), 3.31 (m, 1H), 3.04-2.79 (two s and a set of
multiplets, 8H), 2.32-2.15 (m, 2H), 1.80-1.35 (m, 15H), 1.05-0.83
(two d and m, 8H, J ) 6.7 Hz), 0.01 (s, 9H); IR (KBr) νmax 3421,
3237, 2962, 1741, 1657, 1325, 1250, 1167, 1051, 1017, 972, 842, 742,
700 cm-1; FABHRMS (NBA-CsI) m/z 944.2922 (M + Cs+,
C37H61N5O11SiS requires 944.2912).
N-SES-D-Ser[N-SES-D-Ser[(Boc-Gly-Sar-NMe-Val)-Pip-Gly-Sar-
NMe-Val]-Pip-OBn (21). A solution of 15 (1.62 g, 2.0 mmol) in CH3-
OH (30 mL) was treated with 10% Pd-C (160 mg) and the resulting
black suspension was stirred at 25 °C under H2 (1 atm) for 12 h. The
catalyst was removed by filtration through Celite and the filtrate was
concentrated in vacuo to give the crude acid 17 (1.45 g, 1.45 g
theoretical, 100%) which was used directly in the next reaction without
further purification.
Another 1.62 g sample of 15 (2.0 mmol) was treated with 10 mL of
3 M HCl-EtOAc and the mixture was stirred at 25 °C for 30 min.
The volatiles were removed in vacuo. The residual HCl was removed
by adding Et2O (15 mL) to the hydrochloride salt 16 followed by its
removal in vacuo. After repeating this procedure three times, 1.50 g
of 16 (1.49 g theoretical, 100%) was obtained and used directly in the
following reaction without further purification.
A solution of 17 (1.45 g, 2.0 mmol) and the hydrochloride salt 16
(1.50 g, 2.0 mmol) in DMF (10 mL) was treated sequentially with
NaHCO3 (675 mg, 8.0 mmol), HOBt (271 mg, 2.0 mmol), and EDCI
(385 mg, 2.0 mmol), and the reaction mixture was stirred at 0 °C (2 h)
1460, 1418, 1288, 1251, 1171, 1136, 1108, 844, 738, 699 cm-1
;
FABHRMS (NBA-CsI) m/z 1207.5535 (M + H+, C50H90N10O16S2Si2
requires 1207.5595).
(N-BOC-D-Ser-Pip-Gly-Sar-NMe-Val)2 (Serine Hydroxyl) Dilac-
tone (25). A solution of 24 (120 mg, 0.10 mmol) in THF (3 mL) was
treated sequentially with (BOC)2O (0.7 mL, 3.05 mmol, 30 equiv) and
1.0 M Bu4NF-THF (1.0 mL, 1.0 mmol, 10 equiv), and the resulting
mixture was stirred at 25 °C for 48 h. The reaction mixture was diluted
with EtOAc (30 mL), washed with H2O (20 mL) and saturated aqueous
NaCl (20 mL), dried (Na2SO4), filtered, and concentrated in vacuo.
Flash chromatography (SiO2, 2 × 16 cm, 5% EtOH-CH2Cl2 eluent)
afforded 25 (78 mg, 107 mg theoretical, 73%, 70-73%) as a white
powder: mp 245-247 °C (EtOAc, plates); Rf ) 0.43 (10% CH3CN-
EtOAc); [R]23 -53 (c 0.l5, CHCl3); 1H NMR (CDCl3, 400 MHz)
D
(Table 2); 13C NMR (CDCl3, 100 MHz) δ 172.7, 169.31, 169.26, 167.7,
167.3, 155.1, 79.8, 63.4, 62.3, 52.6, 51.3, 49.2, 43.8, 41.8, 34.9, 30.4,
28.5, 28.4, 26.7, 24.7, 20.1, 19.5, 19.0; IR (KBr) νmax 3422, 3333, 2964,
2937, 2862, 1742, 1713, 1647, 1491, 1458, 1368, 1290, 1250, 1167,

1642 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Boger et al.
1014, 849, 780 cm-1; FABHRMS (NBA-CsI) m/z 1211.4985 (M +
Cs+, C50H82N10O16 requires 1211.4965).
SO4), filtered, and concentrated in vacuo. Chromatography (SiO2, 2
× 18 cm, 10% CH3CN-EtOAc) gave 29 (18.2 mg, 55.1 mg theoretical,
33%) as a white solid along with recovered 24 (6.2 mg, 11%) and 25
(14 mg, 27%). For 29: Rf 0.6 (30% CH3CN-EtOAc); [R]23D -74 (c
0.4, CHCl3); 1H NMR (CDCl3, 400 MHz) δ 8.43 (d, 1H, J ) 5.7 Hz),
8.42 (d, 1H, J ) 6.0 Hz), 5.84 (d, 1H, J ) 6.1 Hz), 5.80 (d, 1H, J )
7.2 Hz), 5.35 (d, 1H, J ) 16.2 Hz), 5.31 (d, 1H, J ) 16.2 Hz), 5.26
(m, 2H), 4.83 (m, 1H), 4.80 (d, 1H, J ) 11.0 Hz), 4.79 (d, 1H, J )
11.0 Hz), 4.66-4.60 (m, 2H), 4.48-4.30 (m, 5H), 4.02 (d, 1H, J )
17.3 Hz), 4.00 (d, 1H, J ) 17.5 Hz), 3.91 (m, 2H), 3.62 (d, 1H, J )
12.1 Hz), 3.55 (d, 1H, J ) 12.8 Hz), 3.42 (d, 2H, J ) 16.2 Hz), 2.95
(s, 3H), 2.94 (s, 3H), 2.92 (s, 3H), 2.91 (s, 3H), 2.93-2.83 (m, 2H),
2.16-2.10 (m, 2H), 1.70-1.40 (m, 12H), 1.43 (s, 9H), 1.05-0.94 (m,
2H), 0.98 (d, 3H, J ) 6.5 Hz), 0.97 (d, 3H, J ) 6.5 Hz), 0.85 (d, 3H,
J ) 6.5 Hz), 0.84 (d, 3H, J ) 6.5 Hz), 0.03 (s, 9H); 13C NMR (CDCl3,
100 MHz) δ 172.7, 172.1, 169.4, 169.3, 169.2, 169.1, 167.7, 167.6,
167.3, 166.6, 155.0, 79.8, 65.2, 63.9, 62.3, 53.7, 52.9, 52.5, 51.2, 49.5,
49.2, 44.0, 43.8, 41.9, 41.8, 35.0, 34.9, 30.4, 30.3, 29.7, 29.6, 28.5,
28.4, 28.3, 26.8, 26.6, 24.7, 24.6, 20.0, 19.9, 19.4, 19.3, 19.1, 19.0,
10.2, -1.99; IR (KBr) νmax 3324, 2939, 1743, 1672, 1641, 1487, 1456,
1416, 1287, 1251, 1169, 1135, 1016, 849, 732 cm-1; FABHRMS
(NBA-CsI) m/z 1275.4716 (M + Cs+, C50H86N10O16SiS: requires
1275.4768).
The structure of 25 was established unambiguously in a single-crystal
X-ray structure determination conducted on plates grown from EtOAc.27
Sandramycin Bis-O-benzyl Ether (27). A solution 25 (48 mg,
0.044 mmol) in 3 M HCl-EtOAc (2 mL) at 25 °C was stirred for 30
min. The solvent was removed in vacuo to afford the hydrochloride
salt 26 (43.3 mg, 42.3 mg theoretical, 100%) as a white powder which
was used directly in next reaction.
A solution of the hydrochloride salt 26 (43.3 mg, 0.044 mmol) and
2828 (50.0 mg, 0.179 mmol, 4.0 equiv) in DMF (4 mL) was treated
sequentially with NaHCO3 (37.5 mg, 0.45 mmol, 10.2 equiv), HOBt
(36.2 mg, 0.268 mmol, 6.0 equiv), and EDCI (34.3 mg, 0.178 mmol,
4.0 equiv), and the reaction mixture was stirred at 25 °C for 72 h. The
reaction mixture was diluted with EtOAc (20 mL) and washed with
H2O (10 mL) and saturated aqueous NaCl (10 mL), dried (Na2SO4),
filtered, and concentrated in vacuo. Flash chromatography (SiO2, 1 ×
15 cm, 5% EtOH-CH2Cl2 eluent) afforded 27 (56.8 mg, 62.3 mg
theoretical, 91%) as a white powder: mp 270-273 °C; Rf ) 0.42 (30%
CH3CN-EtOAc); [R]23D -107 (c 0.29, CHCl3); 1H NMR (CDCl3, 400
MHz) δ 9.01 (d, 2H, J ) 6.3 Hz), 8.48 (d, 2H, J ) 4.3 Hz), 7.92 (d,
2H, J ) 7.8 Hz), 7.70 (d, 2H, J ) 8.5 Hz), 7.59 (s, 2H), 7.54 (m, 8H),
7.39 (t, 4H, J ) 7.5 Hz), 7.30 (t, 2H, J ) 7.4 Hz), 5.46 (d, 2H, J ) 4.8
Hz), 5.44 (d, 2H, J ) 16.6 Hz), 5.34 (m, 6H), 4.87 (dd, 2H, J ) 2.0,
11.5 Hz), 4.83 (d, 2H, J ) 11 Hz), 4.58 (dd, 2H, J ) 2.0, 11.5 Hz),
4.42 (dd, 2H, J ) 5.7, 17.4 Hz), 4.03 (d, 2H, J ) 17.4 Hz), 4.01 (m,
2H), 3.76 (d, 2H, J ) 13.3 Hz), 3.47 (d, 2H, J ) 16.6 Hz), 3.08 (s,
6H), 2.92 (s, 6H), 2.05 (d split septet, 2H, J ) 11, 6.5 Hz), 1.80-1.40
(m, 12H), 0.95 (d, 6H, J ) 6.5 Hz), 0.81 (d, 6H, J ) 6.5 Hz); 13C
NMR (CDCl3, 100 MHz) δ 172.7, 169.2, 167.8, 167.0, 163.5, 151.7,
142.6, 141.6, 136.0, 130.2, 129.5, 128.7, 128.4, 127.9, 127.5, 126.9,
126.4, 117.2, 70.7, 62.8, 62.3, 52.5, 50.8, 49.3, 43.8, 41.9, 34.9, 30.4,
29.7, 28.7, 26.5, 24.8, 20.2, 19.4, 19.0; IR (KBr) νmax 3366, 2934, 2862,
1744, 1641, 1492, 1456, 1420, 1344, 1323, 1287, 1256, 1215, 1184,
1133, 1092, 1010, 918, 841, 774, 733, 697 cm-1; FABHRMS (NBA-
CsI) m/z 1533.5490 (M + Cs+, C74H88N12O16 requires 1533.5496).
N1-SES-N6-[[3-(benzyloxy)quinolyl]-2-carbonyl]-(D-Ser-Pip-Gly-
Sar-NMe-Val)2 (Serine Hydroxyl) Dilactone (31). A solution of 29
(17.5 mg, 0.015 mmol) in 3 M HCl-EtOAc (1 mL) at 25 °C was stirred
for 30 min. The solvent was removed in vacuo to afford the
hydrochloride salt 30 (16.5 mg, 16.5 mg theoretical, 100%) as a white
powder which was used directly in the next reaction.
A solution of the hydrochloride salt 30 (16.5 mg, 0.015 mmol) and
2828 (17.1 mg, 0.06 mmol, 4 equiv) in DMF (1 mL) was treated
sequentially with NaHCO3 (14.0 mg, 0.16 mmol, 11 equiv), HOBt (13.1
mg, 0.97 mmol, 6.5 equiv), and EDCI (11.7 mg, 0.06 mmol, 4 equiv),
and the reaction mixture was stirred at 25 °C for 48 h. The mixture
was diluted with EtOAc (20 mL) and washed with H2O (10 mL) and
saturated aqueous NaCl (10 mL), dried (Na2SO4), filtered, and
concentrated in vacuo. Flash chromatography (SiO2, 1 × 15 cm, 5%
EtOH-CH2Cl2) afforded 31 (12.6 mg, 20 mg theoretical, 63%) as a
Sandramycin (1). A sample of 10% Pd-C (3 mg) was added to a
solution of 27 (6.2 mg, 0.0044 mmol) in EtOAc (4 mL), and the black
suspension was stirred at 25 °C under an atmosphere of H2 (1 atm) for
12 h. The catalyst was removed by filtration through Celite, and the
filtrate was concentrated in vacuo. Chromatography (SiO2, 0.5 × 6
cm, EtOAc eluent) afforded 1 (4.2 mg, 5.4 mg theoretical, 78%) as a
white powder identical in all respects with a sample of natural
material: white powder, mp 206-209 °C, lit.1 mp 208-212 °C; Rf )
0.4 (SiO2, 5% CH3OH-CHCl3 eluent), lit.1 Rf ) 0.4 (SiO2, 5% CH3-
white powder: Rf 0.51 (20% CH3CN-EtOAc); [R]23 -84 (c 0.3,
D
CHCl3); 1H NMR (CDCl3, 400 MHz) δ 9.00 (d, 1H, J ) 6.3 Hz), 8.45
(d, 1H, J ) 5.7 Hz), 8.43 (d, 1H, J ) 5.7 Hz), 7.93 (d, 1H, J ) 7.5
Hz), 7.69 (d, 1H, J ) 7.6 Hz), 7.60 (s, 1H), 7.58-7.52 (m, 4H), 7.39
(m, 2H), 7.30 (m, 1H), 5.81 (d, 1H, J ) 7.0 Hz), 5.45 (d, 1H, J ) 16.6
Hz), 5.44 (d, 1H, J ) 5.8 Hz), 5.37-5.26 (m, 6H), 4.86 (dd, 1H, J )
2.0, 12.0 Hz), 4.82 (d, 1H, J ) 11.0 Hz), 4.79 (d, 1H, J ) 11.0 Hz),
4.67-4.61 (m, 2H), 4.59 (dd, 1H, J ) 2.7, 12.0 Hz), 4.46-4.35 (m,
3H), 4.04-3.98 (m, 3H), 3.93-3.87 (m, 1H), 3.76 (d, 1H, J ) 13.3
Hz), 3.56 (d, 1H, J ) 14.3 Hz), 3.48 (d, 1H, J ) 16.6 Hz), 3.42 (d,
1H, J ) 16.6 Hz), 3.08 (s, 3H), 2.94 (s, 3H), 2.93 (s, 3H), 2.92 (s,
3H), 2.95-2.83 (m, 2H), 2.15-2.04 (m, 2H), 1.75-1.35 (m, 12H),
0.97 (d, 3H, J ) 6.6 Hz), 0.95 (d, 3H, J ) 6.6 Hz), 0.84 (d, 3H, J )
6.6 Hz), 0.81 (d, 3H, J ) 6.6 Hz), 0.04 (s, 9H); 13C NMR (CDCl3, 100
MHz) δ 172.7, 172.1, 169.4, 169.2, 169.1, 167.8, 167.7, 167.0, 166.6,
163.5, 151.7, 142.6, 141.6, 136.0, 130.2, 129.5, 128.7, 128.4, 128.0,
127.5, 126.9, 126.4, 117.2, 70.9, 65.2, 62.8, 62.3, 62.2, 53.7, 53.0, 52.4,
50.7, 49.5, 49.3, 49.2, 44.0, 43.8, 41.9, 41.8, 35.0, 34.9, 30.4, 30.3,
29.7, 28.7, 28.4, 26.8, 26.5, 24.8, 24.6, 20.1, 20.0, 19.4, 19.3, 19.1,
19.0, 10.2, -2.0; IR (KBr) νmax 3322, 2936, 1742, 1668, 1639, 1491,
1462, 1285, 1255, 1135, 1015, 874, 734 cm-1; FABHRMS (NBA-
CsI) m/z 1436.6084 (M + Cs+, C62H89N11O16SiS requires 1436.5033).
OH-CHCl3); [R]23 -153 (c 0.17, CHCl3); 1H NMR (CDCl3, 400
D
MHz) δ 11.74 (s, 2H, OH), 9.56 (d, 2H, J ) 5.7 Hz, Ser-NH), 8.52 (d,
2H, J ) 4.4 Hz, Gly-NH), 7.81 (m, 2H, C5′-H), 7.71 (dd, 2H, J ) 4.4,
5.4 Hz, C8′-H), 7.63 (s, 2H, C4′-H), 7.50 (dd, 4H, J ) 4.1, 5.3 Hz,
C6′ and C7′-H), 5.57 (d, 2H, J ) 6.4 Hz, Pip-R-CH), 5.54 (d, 2H, J )
16.6 Hz, Sar-R-CH), 5.26 (d, 2H, J ) 5.0 Hz, Ser-R-CH), 4.99 (d, 2H,
J ) 11.7 Hz, Ser-â-CH), 4.87 (d, 2H, J ) 11.0 Hz, Val-R-CH), 4.43
(d, 4H, J ) 11.7 Hz, Ser-â-CH and Gly-R-CH), 4.10 (m, 2H, Pip-ꢀ-
CH), 4.06 (m, 2H, Gly-R-CH), 3.74 (d, J ) 14.5 Hz, Pip-ꢀ-CH), 3.55
(d, 2H, J ) 16.6 Hz, Sar-R-CH), 3.12 (s, 6H, Val-NCH3), 2.94 (s, 6H,
Sar-NCH3), 2.04 (d split septet, 2H, J ) 11.0, 6.4 Hz, Val-â-CH), 1.85-
1.50 (m, 12H, Pip-(CH2)3), 0.92 (d, 6H, J ) 6.4 Hz, Val-γ-CH3), 0.78
(d, J ) 6.4 Hz, Val-γ-CH3); 13C NMR (CDCl3, 100 MHz) δ 172.6,
169.4, 169.2, 167.8, 167.7, 166.2, 153.8, 141.4, 134.6, 132.0, 129.4,
128.5, 127.1, 126.4, 120.3, 62.2, 61.9, 52.5, 50.6, 49.3, 43.9, 41.9, 34.9,
30.3, 28.8, 26.2, 24.9, 20.2, 19.4, 18.7; IR (KBr) νmax 3487, 3329, 2932,
1744, 1662, 1637, 1518, 1466, 1418, 1333, 1285, 1191, 1135, 1016,
887, 734 cm-1; UV (CH3OH) λmax 217 (62 000), 229 (60 000), 300
(8070), 356 nm (7840); lit.1 UV (CH3OH) λmax 217 (63 700), 229
(62 800), 356 nm (8100); FABHRMS (NBA) m/z 1221.5565 (M +
H+, C60H76N12O16 requires 1221.5581).
N1-SES-N6-[(3-hydroxylquinolyl)-2-carbonyl]-(D-Ser-Pip-Gly-Sar-
NMe-Val)2 (Serine Hydroxyl) Dilactone (32). A solution of 31 (10
mg, 0.0077 mmol) in 5 mL of EtOAc was treated with 10% Pd-C (4
mg), and the resulting black suspension was stirred at 25 °C under an
atmosphere of H2 (1 atm) for 14 h. The catalyst was removed by
filtration through Celite, and the filtrate was concentrated in vacuo.
Flash chromatography (SiO2 1 × 10 cm, 10% CH3CN-EtOAc)
afforded 32 (8.0 mg, 9.3 mg theoretical, 86%) as a white powder: Rf
N1-SES-N6-Boc-(D-Ser-Pip-Gly-Sar-NMe-Val)2 (Serine Hydroxyl)
Dilactone (29). A solution of 24 (58.2 mg, 0.048 mmol) in 5 mL of
THF was treated sequentially with (BOC)2O (110 µL, 0.48 mmol, 10
equiv) and 1.0 M Bu4NF in THF (192 µL, 0.192 mmol, 4 equiv). The
mixture was stirred at 25 °C under N2 for 24 h. The reaction mixture
was diluted with 40 mL of EtOAc and washed with H2O (20 mL) and
saturated aqueous NaCl (20 mL). The organic layer was dried (Na2-
0.7 (20% CH3CN-EtOAc); [R]23 -105 (c 0.3, CHCl3); 1H NMR
D
(CDCl3, 400 MHz) δ 11.74 (s, 1H, OH), 9.55 (d, 1H, J ) 6.4 Hz,
Ser6-NH), 8.50 (d, 1H, J ) 5.0 Hz, Gly8-NH), 8.44 (d, 1H, J ) 5.0
Hz, Gly3-NH), 7.81 (m, 1H, C5′-H), 7.70 (m, 1H, C8′-H), 7.63 (s, 1H,
C4′-H), 7.50 (m, 2H, C6′ and C7′-H), 5.81 (d, 1H, J ) 7.0 Hz, Ser1-

(-)-Sandramycin: Total Synthesis and Characterization
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1643
NH), 5.55 (d, 1H, Sar9-R-CH), 5.54 (d, 1H, J ) 5.1 Hz, Pip7-R-CH),
5.30 (d, 1H, J ) 16.6 Hz, Sar4-R-CH), 5.28 (d, 1H, J ) 4.6 Hz, Pip2-
R-CH), 5.25 (d, 1H, J ) 6.4 Hz, Ser6-R-CH), 4.98 (d, 1H, J ) 11.0
Hz, Ser6-â-CH), 4.86 (d, 1H, J ) 11.0 Hz, Val10-R-CH), 4.79 (d, 1H,
J ) 11.0 Hz, Val5-R-CH), 4.64 (m, 2H, Ser1-R and â-CH), 4.45-4.35
(m, 4H, Ser1-â-CH, Gly3-R-CH, Ser6-â-CH, and Gly8-R-CH), 4.10-
3.99 (m, 3H, Gly3-R-CH, Gly8-R-CH, and Pip7-ꢀ-CH), 3.90 (m, 1H,
Pip2-ꢀ-CH), 3.72 (d, 1H, J ) 13.0 Hz, Pip7-ꢀ-CH), 3.56 (d, 1H, J )
13.0 Hz, Pip2-ꢀ-CH), 3.55 (d, 1H, J ) 16.6 Hz, Sar9-R-CH), 3.43 (d,
1H, J ) 16.6 Hz, Sar4-R-CH), 3.11 (s, 3H, Val10-NCH3), 2.95 (s, 3H,
Val5-NCH3), 2.94 (s, 3H, Sar9-NCH3), 2.92 (s, 3H, Sar4-NCH3), 2.89
(m, 2H, SO2CH2), 2.12 (d split septet, 1H, J ) 11.0, 6.5 Hz, Val5-â-
CH), 2.04 (d split septet, J ) 11.0, 6.5 Hz, Val10-â-CH), 1.85-1.45
(m, 12H, Pip2- and Pip7-(CH2)3), 1.01 (m, 2H, CH2TMS), 0.97 (d, 3H,
J ) 6.5 Hz, Val5-γ-CH3), 0.92 (d, 3H, J ) 6.5 Hz, Val10-γ-CH3), 0.85
(d, 3H, J ) 6.5 Hz, Val5-γ-CH3), 0.79 (d, 3H, J ) 6.5 Hz, Val10-γ-
CH3), 0.05 (s, 9H, Si(CH3)3); 13C NMR (CDCl3, 100 MHz) δ 172.6,
172.2, 169.4, 169.3, 169.2, 169.1, 167.8, 167.7, 167.6, 166.5, 166.2,
153.8, 141.5, 134.6, 132.1, 129.4, 128.5, 127.1, 126.4, 120.3, 65.2,
62.3, 62.2, 61.9, 53.7, 53.0, 52.4, 50.6, 49.5, 49.3, 49.2, 44.0, 43.9,
41.9, 41.8, 35.0, 34.9, 30.3, 30.2, 28.7, 28.5, 26.8, 26.2, 24.9, 24.6,
20.1, 20.0, 19.5, 19.3, 19.1, 18.7, 10.2, -2.0; IR (KBr) νmax 3329, 2936,
sandramycin (1), a 2 mL of sample containing 10 µM sandramycin (1)
was titrated with 20 µL of calf thymus DNA (320 µM) solution. The
quenching of fluorescence was measured 5 min after each addition of
DNA to allow binding equilibration with 360 nm excitation and 530
nm fluorescence. The results graphically represented in Figure 2 were
analyzed by Scatchard analysis,38 and the results are summarized in
Table 4.
Similar titrations of solutions of luzopeptin A (10 µM) and 32 (10
µM) with calf thymus DNA (320 µM) were conducted with 340 nm
excitation/520 nm fluorescence and 400 nm excitation/510 nm fluo-
rescence, respectively, and the results are summarized in Table 4.
DNA Binding Constant Determination for 25. Method A. Calf
thymus DNA (1.0 × 10-5 M in base pair) was mixed with ethidium
bromide (5.0 × 10-6 M) resulting in a 2:1 ratio of base pair/ethidium
in a 10 mM Tris-HCl (pH 7.4), 75 mM NaCl buffer solution (2 mL).
The fluorescence was calibrated at 24 °C to 100% F and 0% F with a
DNA-ethidium buffer solution and ethidium buffer solution, respec-
tively. The premixed DNA-ethidium solution was titrated with small
aliquots of 25 (20-40 µL of 3 mM 25 in DMSO) and incubated at 24
°C for 30 min prior to each fluorescence measurement. The fluores-
cence was measured with 545 nm excitation and 595 nm emission with
a slit width of 10 nm. The absolute binding constant from three such
titrations were determined at 50% ethidium bromide displacement as
measured by a drop in fluorescence to 50%. The binding constant of
ethidium bromide employed to calculate the absolute binding constant
with a competitive or noncompetitive binding model39 was 4.5 × 105
M-1, and the results are summarized in Table 4.
1744, 1639, 1518, 1462, 1413, 1287, 1255, 1135, 1015, 843, 754 cm-1
;
UV (CH3OH) λmax 202 (43 000), 229 (30 000), 300 (4000), 356 nm
(3400); FABHRMS (NBA) m/z 1214.5671 (M + H+, C55H83N11O16-
SiS requires 1214.5588).
NMR Measurements. All samples were degassed by six freeze-
pump-thaw cycles, and all spectra were recorded at 296 K. All 2D
spectra were recorded with quadrature detection in both dimensions,
TPPI46 was used in F1. The 2D spectra were processed and analyzed
with the Felix program (version 2.3.0, BIOSYM Technologies) on a
Silicon Graphics Personal IRIS Workstation. The parameters of the
individual NMR experiments are given in the following experimentals.
Method B. A 2 mL of sample containing 400 µL of calf thymus
DNA (320 µM in base pair) with or without the presence of 40 µL of
25 (3.2 × 10-3 M in DMSO) in 10 mM Tris-HCl (pH 7.4), 75 mM
NaCl buffer solution was titrated with small aliquots of sandramycin
(1 µL, 1.0 × 10-3 M in DMSO). The quenching of fluorescence was
measured 5 min after each addition of sandramycin with 360 nm
excitation and 530 nm fluorescence. The results were analyzed by
Scatchard analysis and summarized in Table 4.
1
1. 1D H Spectrum: Pulse length, P1 ) 5.0 µs; relaxation delay,
d1 ) 1.0 s; 128 acquisitions.
2. 1D 1H-1H Decoupling Spectrum (Homodecoupler Mode):
Pulse length, P1 ) 10.0 µs; relaxation delays, D1 ) 1 s, D11 ) 1 ms;
the power set for the decoupled nucleus (DEC), dL0 ) 50 dB; 64
acquisitions.
General Procedure for Agarose Gel Electrophoresis. Due to the
low solubility of the agents in water, all agents were dissolved in DMSO
as stock solutions, stored at -20 °C in the dark, and were diluted to
the working concentrations in DMSO prior to addition to the DNA
solution. A buffered DNA solution containing 0.25 µg of supercoiled
ΦX174 RF I DNA (1.0 × 10-8 M) in 9 µL of 50 mM Tris-HCl buffer
solution (pH 8) was treated with 1 µL of agent in DMSO (the control
DNA was treated with 1 µL of DMSO). The [agent] to [DNA] base
pair ratios were 0.022 (lane 1), 0.033 (2), 0.044 (3), 0.11 (4), 0.22 (5)
for luzopeptin A; 0 (6 control DNA), 0.011 (7), 0.022 (8), 0.033 (9),
0.044 (10), 0.066 (11), 0.11 (12) for sandramycin in gel 5A; 0.011 (1),
0.033 (2), 0.066 (3), 0.11 (4) for sandramycin; and 0 (5, control DNA),
0.022 (6), 0.11 (7), 0.22 (8), 0.44 (9), 0.88 (10), 1.74 (11), 2.2 (12) for
compound 32 in gel 5B. The reactions were incubated at 25 °C for 1
h and 5 h for gel A and B, respectively, and quenched with 5 µL of
loading buffer formed by mixing Keller buffer (0.4 M Tris-HCl, 0.05
M NaOAc, 0.0125 M EDTA, pH 7.9) with glycerol (40%), sodium
dodecyl sulfate (0.4%), and bromophenol blue (0.3%). Electrophoresis
was conducted on a 0.9% agarose gel at 90V for 3 h. The gel was
stained with 0.1 µg/mL ethidium bromide, visualized on a UV
transilluminator, and photographed using Polaroid T667 black and white
instant film and directly recorded on a Millipore BioImage 60S RFLP
system.
DNase I Footprinting. The DNase I footprinting system was
obtained from BRL (Life Technologies, Inc.). The 32P 5′-end-labeled
w794 DNA was prepared as previously described.43 Stock solutions
of sandramycin were prepared in DMSO. The solutions were stored
in the dark at -20 °C and were diluted to working conditions with
buffer (10 mM Tris-HCl, pH 7.0; 10 mM KCl; 10 mM MgCl2; 5 mM
CaCl2) immediately prior to use. The final concentration of DMSO
did not exceed 2%.6b A buffered DNA solution (7 µL) containing the
32P 5′-end-labeled w794 DNA (5000 cpm) in 10 mM Tris-HCl (pH
7.0), 10 mM KCl, 10 mM MgCl2, and 5 mM CaCl2 was treated with
2 µL of a freshly prepared sandramycin solution and H2O (1 µL). The
final concentrations of sandramycin were 2 µM, 10 µM, and 20 µM as
indicated. The DNA reaction solutions were incubated at 25 °C for
30 min. The DNA cleavage reactions were initiated by the addition
of 1 µL of a stock solution of DNase I (0.1 µg/mL) containing 1 mM
1
3. 2D H-1H NOESY Spectrum: Sequence D1 -90° -t1 -90°
-τmix -90° -t2; pulse length (90°), P1 ) 18 µs; delays, d0 ) 3 µs, d1
) 2 s, d8 ) 450 ms; sweep width in F1 and F2, SWH ) 4424.779 Hz;
32 acquisitions; 512 increments.
1
4. 2D H-1H ROESY Spectrum: Sequence D1 -90° -t1 -90°
-τmix -90° -t2; pulse lengths, P1 (90° transmitter high power pulse)
) 18 µs; P15 (CW pulse for ROESY spinlock) ) 400 ms; delays, d0
(incremented delay) ) 3 µs, d1 (relaxation delay) ) 2s, d12 (delay for
power switching) ) 20 µs, d13 (short delay) ) 3 µs; powers, hl1
(ecoupler high power) ) 3 dB, hl4 (ecoupler low power) ) 17 dB;
sweep width in F1 and F2, SWH ) 4424.79 Hz; 32 acquisitions; 512
increments.
DNA Binding Constant Measurements. All fluorescence mea-
surements were conducted on a JASCO FP-777 spectrofluorometer
equipped with a Fisons Haake D8 circulated water cooling system. The
temperature was maintained at 24 °C throughout the experimental work.
A 4 mL quartz cuvette equipped with a Teflon-coated magnetic stir
bar was used in all experiments. Calf thymus DNA (Sigma) was
dissolved in 10 mM Tris-HCl (pH 7.4) buffer solution containing 75
mM NaCl. The DNA concentration (320 µM in base pair) was
determined by UV (ꢀ260 ) 12 824 M-1 in base pair). The excitation
and emission spectra were recorded with a sample (2 mL) containing
10 mM Tris-HCl (pH 7.4), 75 mM NaCl buffer, and 20 µL of a DMSO
stock solution of agent with a 10 nm slit width in excitation and
emission. The final concentration for sandramycin, luzopeptin A, or
32 was 10 µM. For sandramycin (1), the fluorescence emission spectra
exhibited a maximum at 530 nm, and the excitation spectrum showed
a sharp band at 260 nm and two broad bands at 300 nm and 360 nm,
respectively. When excited at 360 nm, only the band at 530 nm was
observed in the emission spectrum, and this excitation wavelength was
chosen so that the absorbance of DNA would not interfere with that of
agent. For the determination of the DNA binding constant of
(46) Marion, D.; Wuthrich, K. Biochem. Biophys. Res. Commun. 1983,
113, 967.

1644 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Boger et al.
sandramycin, for copies of its 1H and 13C NMR, and for
information regarding its absolute stereochemistry in advance
of publication (ref 2).
of dithiothreitol and allowed to proceed for 1 min at 25 °C. The
reactions were stopped by addition of 3 M NH4OAc containing 250
mM EDTA followed by EtOH precipitation and isolation of the DNA.
The DNA was resuspended in 8 µL of TE buffer, and formamide dye
(6 µL) was added to the supernatant. Prior to electrophoresis, the
samples were warmed at 100 °C for 5 min, placed in an ice bath, and
centrifuged and the supernatant was loaded onto the gel. Sanger
dideoxynucleotide sequencing reactions were run as standards adjacent
to the treated DNA. Gel electrophoresis was conducted using a
denaturing 8% sequencing gel (19:1 acrylamide-N,N-methylenebi-
sacrylamide, 8 M urea). Formamide dye contained xylene cyanol FF
(0.03%), bromophenol blue (0.3%), and aqueous Na2EDTA (8.7%, 250
mM). Electrophoresis running buffer (TBE) contained Tris base (100
mM), boric acid (100 mM), and Na2EDTA-H2O (0.2 mM). The gel
was prerun for 30 min with formamide dye prior to loading the samples.
Autoradiography of the dried gel was carried out at -78 °C using Kodak
X-Omat AR film and a Picker Spectra intensifying screen.
Supporting Information Available: The 1D 1H NMR
1
decoupling results that establish the H NMR assignments for
1
14, 24, 25, and 29; a table of H NMR assignments for 25 in
CDCl3, THF-d8, CD3OD, DMF-d7, and DMSO-d8; a figure
1
comparison of the H NMR spectra of 25 and sandramycin
(Figure 6); a figure of the excitation and emission spectra of 1,
luzopeptin A, and 32 (Figure 7); tables of the nucleic acid and
sandramycin proton chemical shift assignments in the d(G-
CATGC)2-sandramycin complex and their comparisons with
free agent, free DNA, and the analogous luzopeptin A com-
plexes29,30 as well as the X-ray crystal structure data for 25 (21
pages). This material is contained in many libraries on
microfiche, immediately follows this article in the microfilm
version of the journal, can be ordered from the ACS, and can
be downloaded from the Internet; see any current masthead page
for ordering information and Internet access instructions.
Acknowledgment. We gratefully acknowledge the financial
support of the National Institutes of Health (CA 41101). We
wish to thank H. Cai and D. S. Johnson for Figure 5 and Dr. J.
A. Matson (Bristol-Myers Squibb) for a sample of authentic
JA952799Y