Definition of the effect and role of the bleomycin A2 valerate substituents: Preorganization of a rigid, compact conformation implicated in sequence-selective DNA cleavage

Article abstract of DOI:10.1021/ja9816640

The synthesis and a comparative study of deglycobleomycin A₂ analogues containing key modifications in the valerate subunit are described in efforts that define the role of the linker length and its substituents. In addition to demonstrating th

Full text of DOI:10.1021/ja9816640

J. Am. Chem. Soc. 1998, 120, 9149-9158
9149
Definition of the Effect and Role of the Bleomycin A2 Valerate
Substituents: Preorganization of a Rigid, Compact Conformation
Implicated in Sequence-Selective DNA Cleavage
Dale L. Boger,*,† Timothy M. Ramsey,† Hui Cai,† Silvia T. Hoehn,‡ and JoAnne Stubbe‡
Contribution from the Department of Chemistry and the Skaggs Institute for Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, and
Departments of Chemistry and Biology, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139
ReceiVed May 13, 1998
Abstract: The synthesis and a comparative study of deglycobleomycin A2 analogues containing key
modifications in the valerate subunit are described in efforts that define the role of the linker length and its
substituents. In addition to demonstrating that the C3-hydroxy group exhibits only a modest effect (1-2×)
that may be attributed to an intermolecular H-bond with DNA and that the length of the linker is important
(C4 > C5 > C3 > C2), the studies illustrate that the substantial impact of the C4-methyl group is linked to
the presence of the C2-methyl group, while the modest impact of the C2-methyl group itself (ca. 2×) is not
coupled to the presence of the C4-methyl group. These effects may be explained by the conformational
properties of the agents resulting from the substitutions and beautifully fit the models of DNA-bound Co(III)-
OOH bleomycin A2 and deglycobleomycin A2. The magnitude of the effects suggests that an important
functional role of the linker region is to facilitate adoption of a rigid, compact conformation productive for
DNA cleavage. The studies also identify a second potential site that may adopt two nearly equivalent
conformations which may constitute a swivel point that permits access to a class of related bound structures
adaptable to the conformational characteristics of the multiple cleavage sites or access to both strands of DNA
from a common intercalation site important for both primary (single strand) and secondary (double strand)
cleavage.
Bleomycin A2 (1, Figure 1) is the major constituent of the
clinical antitumor drug Blenoxane and is thought to derive its
therapeutic effects from the ability to mediate the cleavage of
double-stranded DNA or RNA by a metal ion- and oxygen-
dependent process.1-8 Studies employing derivatives of the
natural product, its degradation products, or semisynthetic
analogues, as well as its analogues, have contributed to an
understanding of the structural features contributing to the
sequence-selective cleavage of DNA.9 In our efforts, this has
entailed single changes in the structure of bleomycin A2 or
deglycobleomycin A2 (2), conducted with the intent of defining
the role of each subunit, functional group, or substituent.9 These
studies, carried out in conjunction with recent structural
studies,10-13 have begun to unravel the subtle structural features
contributing to the properties of the natural product.
Despite these studies, the role of the linker region joining
the metal binding domain with the bithiazole DNA binding
domain and its substituents is poorly defined. In conjunction
with the systematic examination of the L-threonine subunit
detailed in the accompanying article,9 herein we describe a
detailed study of the valerate linker region. Early studies with
a select set of bleomycin A2 analogues disclosed by Umezawa,
Ohno, and co-workers illustrated the importance of the presence
and absolute stereochemistry of the C4-methyl group in the
substituted pentanoic acid (valerate) subunit in the DNA
† The Scripps Research Institute.
‡ Massachusetts Institute of Technology.
(10) (a) Wu, W.; Vanderwall, D. E.; Lui, S. M.; Tang, X.-J.; Turner, C.
J.; Kozarich, J. W.; Stubbe, J. J. Am. Chem. Soc. 1996, 118, 1268. Wu,
W.; Vanderwall, D. E.; Turner, C. J.; Kozarich, J. W.; Stubbe, J. J. Am.
Chem. Soc. 1996, 118, 1281. Wu, W.; Vanderwall, D. E.; Stubbe, J.;
Kozarich, J. W.; Turner, C. J. J. Am. Chem. Soc. 1994, 116, 10843. (b)
Vanderwall, D. E.; Lui, S. M.; Wu, W.; Turner, C. J.; Kozarich, J. W.;
Stubbe, J. Chem. Biol. 1997, 4, 373. Lui, S. M.; Vanderwall, D. E.; Wu,
W.; Tang, X.-J.; Turner, C. J.; Kozarich, J. W.; Stubbe, J. J. Am. Chem.
Soc. 1997, 119, 9603.
(11) Wu, W.; Vanderwall, D. E.; Teramoto, S.; Lui, S. M.; Hoehn, S.;
Tang, X.-J.; Turner, C. J.; Boger, D. L.; Kozarich, J. W.; Stubbe, J. J. Am.
Chem. Soc. 1998, 120, 2239.
(12) Fulmer, P.; Zhao, C.; Li, W.; DeRose, E.; Antholine, W. E.; Petering,
D. H. Biochemistry 1997, 36, 4367. Mao, Q.; Fulmer, P.; Li, W.; DeRose,
E. F.; Petering, D. H. J. Biol. Chem. 1996, 271, 6185. Xu, R. X.; Nettesheim,
D.; Otvos, J. D.; Petering, D. H. Biochemistry 1994, 33, 907.
(13) Manderville, R. A.; Ellena, J. F.; Hecht, S. M. J. Am. Chem. Soc.
1994, 116, 10851. Manderville, R. A.; Ellena, J. F.; Hecht, S. M. J. Am.
Chem. Soc. 1995, 117, 7891.
(1) Natrajan, A.; Hecht, S. M. In Molecular Aspects of Anticancer Drug-
DNA Interactions; Neidle, S., Waring, M. J., Eds.; CRC: Boca Raton, 1994;
Vol. 2, p 197.
(2) Kane, S. A.; Hecht, S. M. In Progress in Nucleic Acids Research
Molecular Biology; Cohn, W. E., Moldave, K., Eds.; Academic: San Diego,
1994; Vol 49, p 313.
(3) Ohno, M.; Otsuka, M. In Recent Progress in the Chemical Synthesis
of Antibiotics; Lukacs, G., Ohno, M., Eds.; Springer-Verlag: New York,
1990; p 387.
(4) Dedon, P. C.; Goldberg, I. H. Chem. Res. Toxicol. 1992, 5, 311.
(5) Petering, D. H.; Byrnes, R. W.; Antholine, W. E. Chem.-Biol. Interact.
1990, 73, 133.
(6) Stubbe, J.; Kozarich, J. W. Chem. ReV. 1987, 87, 1107.
(7) Stubbe, J.; Kozarich, J. W.; Wu, W.; Vanderwall, D. E. Acc. Chem.
Res. 1996, 29, 322.
(8) Hecht, S. M. Acc. Chem. Res. 1986, 19, 383.
(9) Boger, D. L.; Ramsey, T. M.; Cai, H.; Hoehn, S. T.; Stubbe, J. J.
Am. Chem. Soc. 1998, 120, 9139 and references therein.
S0002-7863(98)01664-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/28/1998

9150 J. Am. Chem. Soc., Vol. 120, No. 36, 1998
Boger et al.
substituents led to a >100-fold reduction in DNA cleavage
efficiency. In these studies and prior to the recent structural
studies, we suggested this may be due to their effect on the
adoption of a compact, bent bound conformation productive for
DNA cleavage15 and consistent with the 3.5 base-pair binding
site size of 1 and 217 although disruption of the bleomycin A2
interaction with duplex DNA or altered activation kinetics could
also account for the diminished properties.
Herein, we detail the synthesis and examination of 3-8 that,
in conjunction with recent structural studies, clearly define the
effects of the valerate subunit and the impact of its substituents.
Analogue 3 permitted the assessment of the functional impor-
tance of the C3-hydroxy group, not directly addressed in our
earlier studies. The examination of 3 and 4 relative to our prior
analogues 515 and 715 was anticipated to further clarify the role
of the C2- and C4-methyl substituents. Similarly, the direct
comparison of 6 and 8 with 715 permitted an assessment of the
length of the valerate linker. The results of the studies were
more revealing than anticipated. In addition to demonstrating
that the C3-hydroxy group exhibits only a modest effect (1-
2×) and that the length of the linker is important (C4 > C5 >
C3), the studies illustrate that the significant impact of the C4-
methyl substituent is directly coupled to the presence of the
C2-methyl substituent, while the more modest impact of the
C2-methyl group itself is not coupled to the presence of the
C4-methyl group. These results beautifully fit the expected
effects of the substitutions on the conformational properties of
the analogues and agree remarkably well with the models of
DNA-bound Co(III)OOH bleomycin A210 and deglycobleomycin
A2.11 This suggests that the linker substituent effects can be
ascribed principally to their role in facilitating the adoption of
a rigid, compact DNA-bound conformation that leads to
productive sequence-selective strand cleavage. Moreover, the
magnitude of the effects suggests that the functional role of the
linker region extends well beyond that of simply connecting
the bithiazole DNA binding domain to the pyrimidoblamic acid
metal chelation subunit. Rather, its length and substituents serve
to facilitate preorganization of the metal complexes into a
compact conformation productive for DNA cleavage. The
studies also identify a potential site that may adopt two nearly
equivalent turn conformations. This site may constitute a swivel
point in the structure that permits access to a class of related
bound conformations adaptable to multiple cleavage sites or
perhaps access to both strands of duplex DNA from a single
intercalation site important for both primary single strand (ss)
and secondary double strand (ds) DNA cleavage.
Figure 1.
cleavage efficiency of the resulting analogues.3 Similarly, Hecht
and co-workers have shown that substitution of (gly)n, n ) 0-4,
for the L-threonine subunit caused no significant change in the
characteristic 5′-GC, 5′-GT cleavage selectivity of bleomycin
A2 although the cleavage occurred with altered efficiencies.14
More recently, we disclosed the examination of a number of
linker modifications of deglycobleomycin A2 (2) that, while not
altering the cleavage selectivity of 2, revealed an unusually large
impact of its length and substituents on the cleavage efficiency.15
Notably, replacement of the L-threonine subunit with glycine
substantially reduced the cleavage efficiency and the double
strand/single strand (ds/ss) cleavage ratio, and the studies in
the accompanying paper further clarify the role of this subunit’s
substituents.9 Additionally, the studies revealed an important
role (ca. 7×) for the C4-methyl substituent of the valerate
subunit and less prominent roles for the C3-hydroxy group (1-
2×) and C2-methyl substituent (1-2×). These effects were
cumulative, and removal of all of the valerate and threonine
Synthesis. The analogues were prepared employing a recent
modification16 of our original synthesis of 2.17-20 The coupling
of the pentapeptide S analogues 15a-e with NR-BOC-pyrimido-
blamic acid avoids the occasional dehydration reaction of the
â-hydroxy-L-histidine subunit that accompanies the coupling of
tetrapeptide S with the prelinked pyrimidoblamic acid/erythro-
â-hydroxy-L-histidine. It is accomplished with the sulfonium
salt installed, on substrates incorporating only one protecting
group, without deliberate protection of the imidazole, and
without competitive imidazole acylation. Thus, coupling of the
(16) Boger, D. L.; Ramsey, T. M.; Cai, H.; Hoehn, S.; Kozarich, J. W.;
Stubbe, J. J. Am. Chem. Soc. 1998, 120, 53.
(17) Boger, D. L.; Colletti, S. L.; Honda, T.; Menezes, R. F. J. Am. Chem.
Soc. 1994, 116, 5607.
(18) Boger, D. L.; Honda, T.; Dang, Q. J. Am. Chem. Soc. 1994, 116,
5619.
(19) Boger, D. L.; Honda, T.; Menezes; R. F.; Colletti, S. L. J. Am. Chem.
Soc. 1994, 116, 5631. Boger, D. L.; Menezes, R. F.; Honda, T. Angew.
Chem., Int. Ed. Engl. 1993, 32, 273.
(14) Carter, B. J.; Reddy, K. S.; Hecht, S. M. Tetrahedron 1991, 47,
2463. Carter, B. J.; Murty, V. S.; Reddy, K. S.; Wang, S.-N.; Hecht, S. M.
J. Biol. Chem. 1990, 265, 4193.
(15) Boger, D. L.; Colletti, S. L.; Teramoto, S.; Ramsey, T. R.; Zhou, J.
Bioorg. Med. Chem. 1995, 3, 1281.
(20) Boger, D. L.; Honda, T. J. Am. Chem. Soc. 1994, 116, 5647.

Bleomycin A2 Valerate Substituents
J. Am. Chem. Soc., Vol. 120, No. 36, 1998 9151
Scheme 1
Scheme 2
DMF, 0 °C, 10 h, 60-62%) and acid-catalyzed N-BOC
deprotection of 17a-b (20% TFA-CH2Cl2, 0 °C, 2.5 h, 85%)
provided 3 and 4.
Similarly, the unsubstituted N-BOC amino acids 9c-9e were
converted to the deglycobleomycin A2 analogues 6-8 incor-
porating 3-5 carbon linkers in place of the valerate subunit
(Scheme 2).
In preliminary efforts, the alternative approach of coupling
1819,23 with the tetrapeptide S analogues 12 was explored
(Scheme 3). This coupling generally proved less effective
regardless of the reagents employed and was often accompanied
by competitive elimination of water from the â-hydroxy-L-
histidine subunit. This presumably results from the enhanced
acidity of the CR proton when the L-histidine amine is N-acylated
as an amide (i.e., 18) rather than a carbamate (i.e., 13), and the
approach outlined in Schemes 1-2 typically proved superior.
DNA Cleavage Properties. The assessment of the DNA
cleavage properties of 3-8 were conducted using four different
assays. The initial examination was conducted with the Fe(II)
complexes of 3-8 in the presence of O2 and 2-mercaptoethanol
as an appropriate reducing agent. The assessment was made
through the determination of the relative concentrations of the
Fe(II) complex required to produce comparable ss and ds
cleavage of supercoiled ΦX174 DNA (Form I) to generate
relaxed (Form II) and linear (Form III) DNA, respectively
valerate replacement subunits 9a21 and 9b with the tripeptide S
precursor 9f 17 (1.2 equiv of EDCI,22 1 equiv of HOBt, DMF,
23 °C, 24 h, 85%) followed by S-methylation (50 equiv of CH3I,
CH3OH, 23 °C, 24-26 h, 95%) and acid-catalyzed N-BOC
deprotection (3 N HCl-EtOAc, 23 °C, 1.5 h, 97%) provided
the tetrapeptide S analogues 12a-b (Scheme 1). Subsequent
coupling of 12a-b with N-BOC-erythro-â-hydroxy-L-histidine17
(13, 1.4 equiv of BOP reagent,22 2 equiv of i-Pr2NEt, DMF, 23
°C, 8 h, 62-68%) and acid-catalyzed N-BOC deprotection (20%
TFA-CH2Cl2, 0 °C, 4 h, 83-85%) cleanly afforded the
pentapeptide S analogues 15a-b. Diphenylphosphoryl azide
(DPPA) promoted coupling of 15a-b with NR-BOC-pyrimido-
blamic acid (16,18 1.4 equiv of DPPA, 2 equiv of i-Pr2NEt,
(21) Koskinen, A. M. P.; Pihko, P. M. Tetrahedron Lett. 1994, 35, 7417.
(22) EDCI ) 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydro-
chloride; BOP reagent ) benzotriazol-1-yloxytris(dimethylamino)phos-
phonium hexafluorophosphate.; DCC ) 1,3-dicyclohexylcarbodiimide.
(23) Boger, D. L.; Teramoto, S.; Honda, T.; Zhou, J. J. Am. Chem. Soc.
1995, 117, 7338.

9152 J. Am. Chem. Soc., Vol. 120, No. 36, 1998
Boger et al.
Table 1. Summary of ΦX174 and w794 DNA Cleavage Properties
relative efficiency of DNA cleavage
ratio of double to
agent
ΦX174a
w794b
single strand cleavagec
DNA cleavage selectivityb
1, bleomycin A2
2-5
1.0
0.8
0.20
0.40
0.08
0.13
0.08
0.04
5.8
1.0
1:6
5′-GC, 5′-GT > 5′-GA
5′-GC, 5′-GT > 5′-GA
5′-GC, 5′-GT > 5′-GA
5′-GC, 5′-GT > 5′-GA
5′-GC, 5′-GT > 5′-GA
very weak 5′-GC, 5′-GT > 5′-GA
5′-GC, 5′-GT > 5′-GA
weak 5′-GC, 5′-GT > 5′-GA
none
2, deglycobleomycin A2
3, deshydroxy dBLM A2
4, C4-Me dBLM A2
5, C4-desmethyl dBLM A2
6, C3 dBLM A2
1:12
1:14
1:25
1:15
1:43
1:31
1:38
1:98
0.40
0.20
0.20
0.10
0.13
0.13
0.03
7, C4 dBLM A2
8, C5 dBLM A2
Fea,b
a Relative efficiency of supercoiled ΦX174 DNA cleavage, Fe(II)-O2, 2-mercaptoethanol. The results are the average of eight experiments.
b Examined within [32P]-5′-end-labeled w794, Fe(III)-H2O2. The results are the average of 10 experiments. c Ratio of double- to single-stranded
cleavage of supercoiled ΦX174 DNA calculated as FIII ) n2 exp(-n2), FI ) exp[-(n1 + n2)].
Scheme 3
Figure 2. Agarose gel illustrating the cleavage reactions of supercoiled
ΦX174 DNA by Fe(II)-agents at 25 °C for 1 h in buffer solutions
containing 2-mercaptoethanol. After electrophoresis on a 1% agarose
gel, the gel was stained with 0.1 µg/mL ethidium bromide and visualized
on a UV transilluminator and quantified on a Millipore BioImage 60S
RFLP system. The results are tabulated in Table 1.
(Figure 2). Like Fe(II)-bleomycin A2 and deglycobleomycin
A2, the Fe(II) complexes of 3-8 produced both ss and ds DNA
Figure 3. Representative kinetics of supercoiled ΦX174 DNA cleavage
by Fe(II)-7 (16 µM) in buffer solution containing 2-mercaptoethanol.
The DNA cleavage reactions were run at 25 °C for various lengths of
time, and electrophoresis was conducted on a 1% agarose gel. Direct
fluorescence quantitation of the percentage of forms I-III DNA present
at each time point was conducted using a Millipore BioImage 60S RFLP
system visualized on a UV (312 nm) transilluminator in the presence
of 0.1 µg/mL ethidium bromide, taking into account the relative
fluorescence intensities of forms I-III ΦX174 DNA (forms II and III
have fluorescence intensities that are 0.7 times that of form I).
cleavage, and the results are summarized in Table 1. Removal
of the C3-hydroxyl group gave 3 which was only slightly less
effective (0.8×) than deglycobleomycin A2 (2) at cleaving the
supercoiled DNA. The observation that 5 which lacks the C4-
methyl group of 2 was found to be substantially less effective
than deglycobleomycin A2 (0.4-0.5×) proved analogous to
observations in our prior studies.15 Removal of C2-methyl and
C3-hydroxyl groups with 4 led to a more substantial 5-fold
reduction in the DNA cleavage efficiency, while further removal
of the remaining C4-methyl substituent of 4 to give 7 resulted
in an even greater impact on the cleavage efficiency (0.13×).
Shortening and extending the linker length by one carbon gave
agents 6 and 8 which were less effective than 7, and each proved
to be only 2-3× more efficient than Fe(II) cleavage itself.
the Freifelder-Trumbo equation.24 The results are illustrated
in Figure 3 for 7 and are summarized for the full set of agents
in Table 1. Consistent with the relative efficiencies of DNA
cleavage, the ratio of ds to ss cleavage for 3 and 5 were 1:14
and 1:15, being slightly lower but nearly indistinguishable from
that of deglycobleomycin A2 (1:12). The extent of ds cleavage
was lower for 4 (1:25), and that of 6-8 was considerably lower
(1:43, 1:31, and 1:38, respectively).25
Although both ss and ds DNA lesions result from the
oxidative cleavage of DNA by bleomycin A2, the latter have
been considered to be the more significant biological event. For
3-8, the extent of ds to ss DNA cleavage was established in a
study of the kinetics of supercoiled ΦX174 DNA cleavage to
produce linear and circular DNA, indicating that 2 and its
analogues do not promote ds cleavage to the same extent as
bleomycin A2. We assumed a Poisson distribution for the
formation of ss and ds breaks to calculate the average number
of double and single strand cleavages per DNA molecule using
The sequence selectivity of DNA cleavage was examined with
duplex w794 DNA26 by monitoring strand cleavage of singly
[32P]-5′-end-labeled duplex DNA after exposure to the Fe(III)
19,23,27
complexes of the agents following activation with H2O2
in 1 mM phosphate buffer (pH 7.0).26 This assay is more
(24) Freifelder, D.; Trumbo, B. Biopolymers 1969, 7, 681.

Bleomycin A2 Valerate Substituents
J. Am. Chem. Soc., Vol. 120, No. 36, 1998 9153
Figure 4. Cleavage of double-stranded DNA by Fe(III)-agent (SV40 DNA fragment, 144 base pairs, nucleotide no. 5238-138, clone w794) in
phosphate/KCl buffer containing H2O2. The DNA cleavage reactions were run for 90 min at 4 °C, and electrophoresis was run on an 8% denaturing
PAGE and visualized by autoradiography.
sensitive to differences in the cleavage efficiency than the
supercoiled DNA cleavage assay and dependent upon the
reaction conditions and amount of radiolabeled DNA employed.25b
Thus, while the magnitude of the efficiency differences may
vary, the relative trends in efficiency have not been observed
to be altered (Table 1). An extensive range of conditions were
examined, including variations in reaction temperature, time,
and amount of DNA employed. Under all conditions, 3 was
found to be nearly half as effective as deglycobleomycin A2
and to cleave DNA in the sequence-selective fashion charac-
teristic of 1 and 2 (Figure 4). In a manner similar to that seen
with ΦX174 DNA, 4 and 5 showed more significant reductions
(4-5-fold), and 6-8 exhibited even greater reductions (8-10-
fold, 7 > 8 > 6) in the DNA cleavage efficiency. Notably, the
side-by-side comparison of 5 illustrated once again that the
removal of the C4-methyl group from the valerate linker has a
very substantial impact on the cleavage efficiency (0.2×),
analogous to our prior observation (0.15×).15 The distinctions
between 4 and 5 observed in the ΦX174 DNA cleavage assay
(5 > 4, 2×) essentially disappeared in the w794 assay, and the
two analogues were nearly indistinguishable (5 g 4). The
significance of these subtle distinctions is not yet known but
cannot be attributed to the samples themselves since both assays
were conducted concurrently with the same sample stock
solutions. With the exception of 6, all displayed the charac-
teristic DNA cleavage selectivity of bleomycin A2. However,
the clarity of the sequence-selective cleavage diminished as the
efficiency of DNA cleavage decreased. The nonselective
cleavage of DNA competes more effectively with selective
cleavage with the less effective agents 6-8, and the ability to
detect the selective cleavage nearly disappears with 6. Thus,
in the series of C3-C5 linkers (6-8), not only was 7 (C4) more
efficient than 6 and 8 (7 > 8 > 6, C4 > C5 > C3) but the
DNA cleavage selectivity was more pronounced as well.
The cleavage properties of 3-8 were also investigated using
an internally labeled hairpin oligonucleotide assay as described
previously that is useful for quantitating ds cleavage.28 The
results of this investigation are presented in Table 2. The
cleavage efficiency values are somewhat lower than those
detected in the other assays,25b but qualitatively they suggest
the same trends. These results indicate that removal of the
valerate substituents at the R, â, and γ positions significantly
diminishes the cleavage efficiency. However, the trends in
terms of the general importance of these positions are analogous.
The trends for the length of the linker are also very similar, but
no sequence specificity above background was observed for
longer or shorter linker analogues. The ds cleavage efficiency
(25) (a) A theoretical ratio of approximately 1:100 is required for the
linear DNA to be the result of the random accumulation of single strand
breaks within the 5386 base-pair size of ΦX174 DNA, assuming that
sequential cleavage on the complementary strands within 15 base-pairs is
required to permit formation of linear DNA. (b) The larger increases in the
magnitude of differences observed with the hairpin assay and, to a lesser
extent, with the w794 DNA versus φX174 DNA can be attributed in part
to a corresponding amount of unlabeled carrier DNA which competes for
the consumption of the agent and its oxidizing capability. Additionally, the
hairpin assay is run under single turnover conditions which may affect the
magnitude of the differences. For the discussion of the impact of a
substituent, the φX174 DNA cleavage results which entail no added carrier
DNA are used for the comparisons.
(26) Ambrose, C.; Rajadhyaksha, A.; Lowman, H.; Bina, M. J. Mol. Biol.
1989, 209, 255. Boger, D. L.; Munk, S. A.; Zarrinmayeh, H.; Ishizaki, T.;
Haught, J.; Bina, M. Tetrahedron 1991, 47, 2661.
(27) Natrajan, A.; Hecht, S. M.; van der Marel, G. A.; van Boom, J. H.
J. Am. Chem. Soc. 1990, 112, 3997. Natrajan, A.; Hecht, S. M.; van der
Marel, G. A.; van Boom, J. H. J. Am. Chem. Soc. 1990, 112, 4532.
(28) Absalon, M. J.; Stubbe, J.; Kozarich, J. W. Biochemistry 1995, 34,
2065.

9154 J. Am. Chem. Soc., Vol. 120, No. 36, 1998
Table 2. Hairpin DNA Cleavage Efficiency
Boger et al.
agent
µMa
cleavage efficiencyb
cleavage specificity
2
3
4
5
6
7
8
5
50
235
100
187
100
200
1.0
0.1
5′-GT, GC
5′-GT, GC
5′-GT, GC
5′-GT, GC
nd
5′-GT, GC
nd
<0.02
<0.05
<0.03
<0.05
<0.02
a For assay, see refs 9 and 28. Agent concentration needed to achieve
a comparable amount of total DNA cleavage as deglycobleomycin A2
(2, 5 µM). b Total DNA cleavage relative to deglycobleomycin A2, nd
) none detected.
was not determined for each of the analogues using this assay
since the ss cleavage efficiency was so low. Only analogue 3
showed a significant amount of an oligonucleotide fragment of
appropriate size to be associated with ds cleavage (data not
shown). However, this analogue shows the fragment only at a
high concentration which suggests that this band may also be
caused by two separate instances of ss cleavage since the
activation conditions for the analogue allow multiple turnovers.
Figure 5. DNA-bound conformation of Co(III)OOH BLM A2 taken
from ref 10, highlighting the compact, rigid conformation of the valerate
linker region enlisting two turns, one at C2 and one at C4, and a C3-
OH H-bond to the phosphate spanning the intercalation site. At the C2
and C4 positions, the valerate methyl substituents adopt the extended
conformation, inducing the two turns.
Length of the Valerate Linker and Cumulative Substitu-
ent Effect. Whereas removing the valerate substituents showed
no impact on the DNA cleavage selectivity, changing the chain
length diminished or abolished sequence specificity, depending
on the analogue and assay used. All modifications substantially
altered the cleavage efficiency and the ratio of ds to ss DNA
cleavage. The length of the linker followed a well-defined
relationship of C4 > C5 > C3 where the length of the natural
linker was established to be optimal. When these results are
combined with our prior observation of the further diminished
cleavage efficiency of the analogue containing a D-Ala substitu-
tion (C2) for the valerate subunit,15 the trend is extended (C4
> C5 > C3 > C2), indicating that linkers longer than four
carbons exhibit a diminished DNA cleavage efficiency and those
shorter than four carbons are progressively and more seriously
impacted. This would be consistent with the adoption of a well-
defined compact conformation productive for DNA cleavage
where the agents possessing a chain shorter than four carbons
may have progressively more difficulty in adopting effective
conformations spanning the linker region. In turn, those longer
than four carbons possess greater flexibility and less of an
opportunity to adopt the productive bound conformation required
for selective cleavage. Moreover, the substantially reduced
cleavage efficiency and ds/ss cleavage ratio of even C4 (7, ca.
0.1×) established that the valerate substituents contribute
substantially to properties of 1-2 and, as detailed below, may
do so by facilitating the adoption of a compact bound conforma-
tion implicated in recent structural studies.
E (Figure 6).29 These differ only in the ability of the extended
conformation to benefit from an intramolecular H-bond from
the C3-hydroxyl group to the valerate carbonyl. Despite this
potential, both free and bound Co(III)OOH bleomycin A2 adopt
the compact carbonyl/C2-methyl eclipsed conformation D,
presumably compensated for by the adoption of H-bonds from
the valerate NH (1.8 Å, 155°) and the threonine NH (1.9 Å,
152°) to the metal-complexed hydroperoxide.10 However, it is
this site within the valerate linker that could adopt two near
equivalently accessible conformations. Since there is more than
one energetically favorable conformation, it is possible that this
site constitutes a swivel point within the molecule which
complements one found in the threonine subunit.9 This swivel
point would allow access to several related bound conformations
that could meet the conformational requirements of multiple
cleavage sites as well as those of ds cleavage from a common
intercalation site. Alternative conformations about the C3
(g2.65-3.35 kcal, not shown) or C4 centers (Figure 6 top,
g2.4-2.65 kcal) are much less stable and constitute noncon-
tributing conformations.29
An important ramification of this rigid conformation of the
valerate linker is the defined placement and orientation of the
amide bond linking the valerate and â-hydroxy-L-histidine
subunits. This is due to the presence of the C4-methyl
Conformational Effects of the Valerate Substituents. The
valerate subunit of both free and DNA-bound Co(III)OOH
bleomycin A2 and deglycobleomycin A2 adopts a well-defined,
rigid conformation (Figure 5). Important characteristics of this
conformation are two turns, one at the C2 center and one at the
C4 center. At both sites, the methyl substituents adopt the
extended conformation position, inducing turns in the linker and
producing a compact conformation. Diagnostic of this confor-
mation is the small coupling constant for C2-H/C3-H (J =
1.8 ( 1.2 Hz) and the large coupling constant for C3-H/C4-H
(J ) 9.5 ( 1.2 Hz).10
(29) The following A-values of 0.9 kcal/mol (OH), 1.8 kcal/mol (CH3),
1
and 1.6 kcal/mol (NHCOPh for NHCOR) were employed. Values of
/
2
the A-value were employed to represent a gauche interaction with a methyl
group: OH/CH3 (0.45 kcal/mol), CH3/CH3 (0.9 kcal/mol), and NHCOR/
CH3 (0.8 kcal/mol). The latter was also used to approximate NHCOR/CH2R
and NHCOR/CHR2 gauche interactions. A value of 0.4 kcal/mol was used
to approximate an OH/NHCOR gauche interaction (versus 0.45 kcal/mol
for OH/CH3) and should prove sufficiently accurate for the purposes used
herein. It was approximated by using a value 1/4 of the A-value of NHCOPh
(0.4 versus 1.6 kcal/mol) in analogy to that of CH3 (0.45 versus 1.8 kcal/
mol). The 1,3-diaxial NHCOR/CH3 was estimated to be >3.5 kcal/mol,
slightly smaller than a 1,3-diaxial CH3/CH3 interaction of 3.7 kcal/mol. It
was approximated by using a value that was 0.1 kcal/mol higher than 2×
the A-value of the substituents (1.6 + 1.8 kcal/mol) in analogy to that of a
CH3/CH3 1,3-diaxial interaction (2× 1.8 + 0.1 ) 3.7 kcal/mol). 1,2-
Eclipsing interactions of H/CO and CH3/CO were approximated, enlisting
the corresponding H/CH3 and CH3/CH3 values of 1.3 and 3.7 kcal/mol,
respectively.
As a consequence of the C2 substitution, the subunit is rigidly
held in a compact conformation A but could adopt two
equivalent conformations consisting of a compact carbonyl/ C2-
methyl eclipsed conformation D or an extended conformation

Bleomycin A2 Valerate Substituents
J. Am. Chem. Soc., Vol. 120, No. 36, 1998 9155
Figure 7. Conformational analysis of 3 and impact of the C3-hydroxy
group.
Figure 8.
The presence of the C3-hydroxy group could preferentially
stabilize a nonproductive extended versus compact conformation
through a H-bond to the valerate carbonyl (Figure 8). Its
disruption, observed in free and bound Co(III)OOH bleomcyin
A2, through the use of a compensating threonine NH H-bond
to the metal-bound hydroperoxide frees up this hydroxyl group
which is enlisted to engage in a H-bond with a DNA backbone
phosphate spanning the intercalation site (see Figure 5). This
intermolecular H-bond with DNA may contribute to the affinity
and alignment of the DNA-bound agent as well as compensate
for the lost H-bond resulting from adoption of the compact
versus extended conformation. It is this lost intermolecular
H-bond with the nonbridging oxygen of the phosphate backbone
that may account for the modest decrease in the cleavage
efficiency with 3 (0.8×).
Figure 6. Conformational analysis of the natural valerate subunit.
substituent which, when combined with the presence of the C2-
methyl group, must adopt the extended orientation A on the
valerate chain, inducing a second turn in the valerate linker
(Figure 6, g2.4-2.65 kcal). In the absence of the C4-methyl
group, the C4 amide would preferentially adopt the extended
B versus turn conformation A. In the absence of the C2-methyl
group, either the C4-methyl group or the C4-NHCO could
equivalently adopt the extended orientation. Just as important
as this placement of the amide, the preferential adoption of the
H-eclipsed carbonyl conformation X (∆E ca. 2.4 kcal, Figure
6 bottom) sets a defined orientation for the â-hydroxy-L-histidine
subunit and the entire metal binding domain. In total, this
provides the core rigid structure about which the DNA binding
domain (C-terminus) and the metal chelation domain (N-
terminus) are linked.
Intuitively, the most unexpected effect of the substituents is
the modest contribution of the C3-hydroxy group.10 Its removal
to provide 3 afforded an agent that exhibited a modest reduction
in the DNA cleavage efficiency (1-0.5×), and the change had
almost no impact on the ds/ss cleavage ratio. Consistent with
this, the removal of the C3-hydroxy group does not directly
impact the conformational properties of the valerate subunit.
Its removal has little effect on the conformation about the C3
center (∆E alternatives ) g2.25-2.45 kcal) or the C4 center
(Figure 7, g2.0-2.7 kcal).
As detailed in our earlier study,15 removal of the C4-methyl
group resulted in a larger reduction in the cleavage efficiency
of the resulting agent 5. This impact may be attributed to its
conformational effects on the valerate subunit. Its removal
allows the agent to adopt two nearly equivalent conformations
about the C4 center including both the productive turn confor-
mation A (∆E ) 0.4 kcal) and that where the amide adopts the
now more stable extended conformation B (Figure 9). It
additionally lowers the barrier to adopting alternative C3
conformations (∆E ) g0.45-3.9 kcal, not shown). Even upon
adoption of the productive compact conformation with the
correct positioning of the â-hydroxy-L-histidine linking amide,
it may now adopt either of two orientations (X and Y) with
equivalent energies (Figure 9, bottom). This further increases
the number of accessible, potentially nonproductive conforma-
tions available to the agent, contributing significantly to the
reduced efficiency of DNA cleavage by 5 and constitutes part
of the important role attributed to the C4-methyl substituent.

9156 J. Am. Chem. Soc., Vol. 120, No. 36, 1998
Boger et al.
Figure 11. Conformational analysis of 7.
Table 3
rel DNA-cleavage
efficiency
rel
agent
∆Ea (kcal/mol) ΦX174
w794b
2, dBLM A2
0.0
0.0c
0.4d
0.8
1.6
1.0
1.0 (1.0)
0.5c(nd)
3, deshydroxy dBLM A2
5, C4-desmethyl dBLM A2
4, C4-Me dBLM A2
7, C4 dBLM A2
0.8c
0.4d
0.2
0.2d(0.14)
0.2 (nd)
0.13
0.13 (0.14)
Figure 9. Conformational analysis of 5 and the impact of the C4-
methyl substituent.
a E (low energy conformation) - E (DNA-bound conformation
adopted from Co(III)OOH BLM A2). b Data in parentheses taken from
ref 15, conducted under alternate assay conditions (0.3× DNA, 37 °C,
30 min). nd ) not done. c Lacks capability for intermolecular C3-
OH/phosphate H-bond, see text. d Possesses two equivalent L-histidine
H-eclipsed amide orientations versus one for 2-4, further contributing
to a reduction in the cleavage efficiency, see text.
the potential for adopting a greater ensemble of accessible
conformations of which those derived from the extended C2
conformation E would be preferred over the productive con-
formation D.
The agent 7, lacking all the valerate substituents which
constitutes the further removal of the C4-methyl group from 4,
exhibits an even further diminished DNA cleavage efficiency.
Like 4, 7 preferentially adopts the C2 extended versus compact
conformation (∆E ) 0.8 kcal), preferentially adopts the C4
amide extended versus turn conformation (∆E ) 0.8 kcal), and
has accessible the alternative turn (gauche) conformation (∆E
) 0.8 kcal). It also possesses two more alternative accessible
C3 conformations (∆E ) 0.9 kcal, not shown) and may adopt
either of two orientations for the linking â-hydroxy-L-histidine
amide (∆E ) 0.0 kcal, not shown), Figure 11. This larger
ensemble of accessible conformations and the larger preference
for the fully extended versus compact conformation (∆E ) 1.6
kcal) accounts for the larger reduction in DNA cleavage
efficiency with 7.
Although the most appropriate correlation would be calculated
Boltzmann probabilities for adoption of the productive bound
conformation, the number of potential conformations for agents
such as 7 precludes such an accurate assessment. However, a
rough correlation derived simply by calculation of the relative
energy difference between the low-energy conformation and the
DNA-bound conformation adopted from Co(III)OOH bleomycin
A2 nicely accounts for the differences in the relative DNA
cleavage efficiencies (Table 3). The exception to this is 3 which
lacks the capabilities for formation of the intermolecular H-bond
of the C3-hydroxy group with the nonbridging oxygens of the
DNA backbone phosphate spanning the intercalation site.
Figure 10. Conformational analysis of 4.
Despite the substantial impact of removing the C4-methyl
group and the small impact of removing the C3-hydroxy group,
its presence alone with the agent 4 lacking both the C2-methyl
and the C3-hydroxy group resulted in an even larger reduction
in the cleavage efficiency. The properties of this agent, which
constitutes the further removal of the C2-methyl group from 3,
indicates that a large portion of the impact of the C4-methyl
group is intimately linked to the presence of one or both of the
additional substituents and suggests that its effect is principally
related to the presence of the C2-methyl group (Figure 10).
Agent 4 now preferentially adopts the C2 extended R eclipsed
conformation (Figure 10, conformation E) rather than the
productive compact conformation (∆E ) 0.8 kcal), conformation
D. This agent may also adopt two C4 conformations, constitut-
ing either the productive turn (A) or the extended amide
conformation B (∆E ) 0.1 kcal). The gauche/gauche confor-
mation C (∆E ) 0.9 kcal) and alternative C3 conformations
(∆E ) 1.4-2.7 kcal, not shown) are also accessible. Nonethe-
less, upon adoption of the productive valerate conformation,
the C4-methyl group does ensure the correct orientation of the
â-hydroxy-L-histidine linking amide. The net effect of the
removal of the C2-methyl group and the C3-hydroxy group is
Conclusions
The synthesis and examination of 3-8 containing key
modifications in the valerate subunit of deglycobleomycin A2
revealed that each of the substituents contributes in a significant

Bleomycin A2 Valerate Substituents
J. Am. Chem. Soc., Vol. 120, No. 36, 1998 9157
4.05 (m, 2H), 3.96 (dd, J ) 6.0, 4.0 Hz, 2H), 3.70 (dd, J ) 6.0, 6.0
Hz, 2H), 3.59 (dd, J ) 6.5, 6.5 Hz, 2H), 3.37 (dd, J ) 7.5, 7.5 Hz,
2H), 3.25 (dd, J ) 7.5, 7.5 Hz, 2H), 2.93 (s, 6H), 2.78 (m, 1H), 2.59
(m, 1H), 2.33 (m, 2H), 2.21 (s, 3H), 2.15 (tt, J ) 7.0, 7.0 Hz, 2H),
1.82 (m, 1H), 1.73 (m, 1H), 1.17 (d, J ) 6.5 Hz, 3H), 1.11 (d, J ) 6.5
Hz, 3H); IR (neat) νmax 3313, 2923, 1668, 1550, 1506, 1432, 1195,
1130 cm-1; FABHRMS (NBA-CsI) m/z 1017.3928 (M+, C41H61N16O9S3
requires 1017.3970).
Preparation of 4 from 19b. A solution of 19b (2.0 mg, 0.0014
mmol) in CH2Cl2 (150 µL) at 0 °C was treated with CF3CO2H (50 µL)
under Ar, and the solution was stirred at 0 °C (4 h). The solvents
were evaporated under a stream of N2, and the residue was dissolved
in CH3OH (200 µL). Aqueous NH4OH (28%, 15 µL) was added, and
the colorless solution was stirred at 23 °C (1 h). The solvent was
evaporated in vacuo and the residue was purified by reverse phase
chromatography (C-18, 0-70% CH3OH-H2O to afford 4 (0.88 mg,
1.4 mg theoretical, 62%) as a white film.
or substantial manner to the DNA cleavage efficiency but has
minimal impact on the primary 5′-GPy cleavage selectivity. The
length of the valerate linker appears to be optimal (C4 > C5 >
C3 > C2) although the effect was established in the absence of
linker substituents which have a substantial effect (ca. 10×).
The most modest of the effects is that of the C3-hydroxy group10
(1.2×). Its presence has little or no impact on the conforma-
tional properties of the valerate linker, and its effect may be
attributed to its participation in an intermolecular H-bond with
the nonbridging oxygens of a phosphate spanning the intercala-
tion site that enhances binding affinity and/or alignment of the
DNA-bound agent. The more substantial impact of the C4-
methyl group is linked to the presence of the C2-methyl group,
while the modest impact of the C2-methyl group itself (ca. 2×)
is not coupled to the presence of the C4-methyl group. These
effects are consistent with the expected conformational proper-
ties of the agents resulting from the substitutions and beautifully
fit models of DNA-bound Co(III)OOH bleomycin A2 and
deglycobleomycin A2. In addition to indicating that it is not
an extended conformation, the cumulative magnitude of the
effects is substantial and suggests that an important functional
role of the valerate substituents is to preorganize bleomycin A2
into a rigid, compact conformation productive for DNA cleav-
age. One site, the C2 position possessing a methyl substituent,
may constitute a flexible position capable of adopting two
equivalently accessible conformations that may accommodate
conformational variations among cleavage sites, facilitate con-
formational reorganization to accommodate ds cleavage from
a single intercalation site, or enhance the mechanics of DNA
binding/decomplexation required for sampling multiple sites.
Nâ-[1-Amino-3(S)-[4-amino-6-[[[1(S)-(((2-(((1(S)-(((2-(4′-((((3-di-
methylsulfonio)-1-propyl)amino)carbonyl)-2′,4-bithiazol-2-yl)-1-
ethyl)amino)carbonyl)-2(R)-hydroxy-1-propyl)amino)carbonyl)-1-
ethyl)amino)carbonyl)-2(R)-hydroxy-2-(imidazol-4-yl)-1-
ethyl]amino]carbonyl]-5-methylpyrimidin-2-yl]propion-3-yl]-(S)-â-
aminoalanine amide (6):30 (91%) Rf ) 0.2 (SiO2, 10:9:1 CH3OH-
10% aqueous CH3CO2NH4-10% aqueous NH4OH); [R]23D -33 (c 0.16,
1
CH3OH); H NMR (CD3OD, 400 MHz) δ 8.85 (s, 1H), 8.40 (s, 1H),
8.12 (s, 1H), 7.45 (s, 1H), 5.20 (d, J ) 8.0 Hz, 1H), 4.74 (d, J ) 8.0
Hz, 1H), 4.26 (m, 2H), 4.21 (d, J ) 4.5 Hz, 1H), 4.09 (dq, J ) 6.5, 4.5
Hz, 2H), 3.69 (dd, J ) 6.0, 6.0 Hz, 2H), 3.60 (dd, J ) 6.5, 6.5 Hz,
2H), 3.45 (dd, J ) 7.0, 7.0 Hz, 2H), 3.37 (t, J ) 7.5 Hz, 2H), 3.26 (t,
J ) 7.5 Hz, 2H), 2.92 (s, 6H), 2.81 (m, 1H), 2.56 (dd, J ) 6.5, 6.5 Hz,
2H), 2.21 (s, 3H), 2.14 (tt, J ) 7.0, 7.0 Hz, 2H), 1.10 (d, J ) 6.5 Hz,
3H); IR (neat) νmax 3334, 3018, 2936, 1665, 1548, 1507, 1431, 1202,
1132 cm-1; FABHRMS (NBA-CsI) m/z 989.3699 (M+, C39H57N16O9S3
requires 989.3657).
Experimental Section30
Preparation of 6 from 19c: (52%).
Nâ-[1-Amino-3(S)-[4-amino-6-[[[1(S)-(((3-(((1(S)-(((2-(4′-((((3-di-
methylsulfonio)-1-propyl)amino)carbonyl)-2′,4-bithiazol-2-yl)-1-
ethyl)amino)carbonyl)-2(R)-hydroxy-1-propyl)amino)carbonyl)-1-
propyl)amino)carbonyl)-2(R)-hydroxy-2-(imidazol-4-yl)-1-
ethyl]amino]carbonyl]-5-methylpyrimidin-2-yl]propion-3-yl]-(S)-â-
aminoalanine amide (7):30 (87%) Rf ) 0.2 (SiO2, 10:9:1 CH3OH-
Nâ-[1-Amino-3(S)-[4-amino-6-[[[1(S)-(((4(S)-(((1(S)-(((2-(4′-((((3-
dimethylsulfonio)-1-propyl)amino)carbonyl)-2′,4-bithiazol-2-yl)-1-
ethyl)amino)carbonyl)-2-(R)-hydroxy-1-propyl)amino)carbonyl)-
2(R)-pentyl)amino)carbonyl)-2(R)-hydroxy-2-(imidazol-4-yl)-1-
ethyl]amino]carbonyl]-5-methylpyrimidin-2-yl]propion-3-yl]-(S)-â-
aminoalanine Amide (3).30 A solution of 17a30 (6 mg, 0.005 mmol,
1 equiv) in CH2Cl2 (100 µL) at 0 °C was treated with 20% TFA-
CH2Cl2 (200 µL), and the mixture was stirred at 0 °C (2.5 h). The
solvent was evaporated in vacuo and the product was purified by reverse
phase chromatography (C-18, 1.2 × 4 cm, 0-90% CH3OH-H2O
gradient elution) affording 3 (4.6 mg, 5.5 mg theoretical, 85%) as a
10% aqueous CH3CO2NH4-10% aqueous NH4OH); [R]23D -35 (c 0.13,
1
CH3OH); [R]23 -8 (c 0.05, 0.1 N aqueous HCl); H NMR (CD3OD,
D
400 MHz) δ 8.85 (s, 1H), 8.32 (s, 1H), 8.13 (s, 1H), 7.50 (s, 1H), 5.21
(d, J ) 8.0 Hz, 1H), 4.74 (d, J ) 8.0 Hz, 1H), 4.26 (d, J ) 4.5 Hz,
1H), 4.20 (m, 2H), 4.12 (dq, J ) 6.5, 4.5 Hz, 1H), 3.73 (m, 2H), 3.64
(dd, J ) 6.0, 6.0 Hz, 2H), 3.59 (dd, J ) 6.5, 6.5 Hz, 2H), 3.37 (t, J )
7.5 Hz, 2H), 3.27 (dd, J ) 6.0, 6.0 Hz, 2H), 2.93 (s, 3H), 2.70 (m,
1H), 2.35 (dd, J ) 6.5, 6.5 Hz, 2H), 2.21 (s, 3H), 2.16 (tt, J ) 7.0, 7.0
Hz, 2H), 1.82 (dd, J ) 6.5, 6.5 Hz, 2H), 1.12 (d, J ) 6.5 Hz, 3H); 1H
NMR (D2O, 400 MHz) δ 8.54 (s, 1H), 8.02 (s, 1H), 7.89 (s, 1H), 7.11
(s, 1H), 5.03 (d, J ) 8.0 Hz, 1H), 4.41 (m, 1H), 4.28 (m, 1H), 4.17
(m, 1H), 3.97 (d, J ) 4.5 Hz, 1H), 3.89 (m, 1H), 3.50 (m, 1H), 3.39
(m, 4H), 3.26 (m, 1H), 3.18 (t, J ) 7.5 Hz, 2H), 3.08 (t, J ) 6.5 Hz,
2H), 2.98 (t, J ) 7.0 Hz, 2H), 2.80 (m, 2H), 2.70 (s, 6H), 2.09 (m,
2H), 1.96 (tt, J ) 7.0, 7.0 Hz, 2H), 1.77 (s, 3H), 1.54 (t, J ) 7.0 Hz,
2H), 0.88 (d, J ) 6.5 Hz, 3H); IR (neat) 3325, 2923, 2852, 1669, 1556,
1538, 1515, 1450, 1426, 1201, 1130 cm-1; FABHRMS (NBA-CsI)
m/z 1003.3810 (M+, C40H59N16O9S3 requires 1003.3813).
white solid: Rf ) 0.2 (SiO2, 10:9:1 CH3OH-10% aqueous CH3CO2-
1
NH4-10% aqueous NH4OH); [R]23 -5 (c 0.1, CH3OH); H NMR
D
(CD3OD, 400 MHz) δ 8.27 (s, 1H), 8.22 (s, 1H), 8.13 (s, 1H), 7.31 (s,
1H), 5.16 (d, J ) 7.5 Hz, 1H), 4.73 (d, J ) 7.5 Hz, 1H), 4.30 (d, J )
4.5 Hz, 1H), 4.05 (m, 4H), 3.69 (dd, J ) 6.0, 6.0 Hz, 2H), 3.60 (dd,
J ) 6.5, 6.5 Hz, 2H), 3.38 (dd, J ) 7.0, 7.0 Hz, 2H), 3.28 (dd, J )
7.5, 7.5 Hz, 2H), 2.94 (s, 6H), 2.76 (m, 1H), 2.56 (m, 1H), 2.47 (m,
1H), 2.25 (s, 3H), 2.16 (tt, J ) 7.0, 7.0 Hz, 2H), 1.79 (m, 1H), 1.49
(m, 1H), 1.14 (d, J ) 6.5 Hz, 3H), 1.25 (d, J ) 6.5 Hz, 3H), 1.12 (d,
J ) 6.5 Hz, 3H); IR (neat) νmax 3333, 2943, 1676, 1641, 1548, 1430,
1194, 1133, 1112 cm-1; FABHRMS (NBA-CsI) m/z 1031.4100 (M+,
C42H63N16O9S3 requires 1031.4126).
Nâ-[1-Amino-3(S)-[4-amino-6-[[[1(S)-(((4-(((1(S)-(((2-(4′-((((3-di-
methylsulfonio)-1-propyl)amino)carbonyl)-2′,4-bithiazol-2-yl)-1-
ethyl)amino)carbonyl)-2(R)-hydroxy-1-propyl)amino)carbonyl)-
2(R)-butyl)amino)carbonyl)-2(R)-hydroxy-2-(imidazol-4-yl)-1-
ethyl]amino]carbonyl]-5-methylpyrimidin-2-yl]propion-3-yl]-(S)-â-
aminoalanine amide (4):30 (85%) Rf ) 0.2 (SiO2, 10:9:1 CH3OH-
Preparation of 7 from 19d: (95%).
Nâ-[1-Amino-3(S)-[4-amino-6-[[[1(S)-(((4-(((1(S)-(((2-(4′-((((3-di-
methylsulfonio)-1-propyl)amino)carbonyl)-2′,4-bithiazol-2-yl)-1-
ethyl)amino)carbonyl)-2(R)-hydroxy-1-propyl)amino)carbonyl)-1-
butyl)amino)carbonyl)-2(R)-hydroxy-2-(imidazol-4-yl)-1-
ethyl]amino]carbonyl]-5-methylpyrimidin-2-yl]propion-3-yl]-(S)-â-
aminoalanine amide (8):30 (85%) Rf ) 0.2 (SiO2, 10:9:1 CH3OH-
10% aqueous CH3CO2NH4-10% aqueous NH4OH); [R]23 -13.6 (c
D
1
0.28, CH3OH); H NMR (CD3OD, 400 MHz) δ 8.81 (s, 1H), 8.24 (s,
10% aqueous CH3CO2NH4-10% aqueous NH4OH); [R]23 -3.5 (c
D
1H), 8.11 (s, 1H), 7.48 (s, 1H), 5.20 (d, J ) 8.0 Hz, 1H), 4.71 (d, J )
8.0 Hz, 1H), 4.25 (d, J ) 4.5 Hz, 1H), 4.09 (dq, J ) 6.5, 4.5 Hz, 2H),
1
0.25, CH3OH); H NMR (CD3OD, 400 MHz) δ 8.21 (s, 1H), 8.12 (s,
1H), 8.00 (s, 1H), 7.20 (s, 1H), 5.15 (d, J ) 8.0 Hz, 1H), 4.79 (d, J )
8.0 Hz, 1H), 4.25 (d, J ) 4.5 Hz, 1H), 4.10 (dq, J ) 6.5, 4.5 Hz, 1H),
4.02 (m, 2H), 3.68 (dd, J ) 6.5, 6.5 Hz, 2H), 3.63 (dd, J ) 6.5, 6.5
(30) Full characterization of intermediates 9b-e, 10a-e, 11a-e, 12a-
e, 14a-e, 15a-e, 17a-e, and 19a-e is provided in Supporting Information.

9158 J. Am. Chem. Soc., Vol. 120, No. 36, 1998
Boger et al.
Hz, 2H), 3.59 (dd, J ) 6.5, 6.5 Hz, 2H), 3.37 (t, J ) 7.5 Hz, 2H), 3.26
(dd, J ) 6.0, 6.0 Hz, 2H), 3.21 (dd, J ) 6.5, 6.5 Hz, 2H), 2.93 (s, 6H),
2.73 (m, 1H), 2.53 (m, 1H), 2.26 (m, 2H), 2.23 (s, 3H), 2.14 (tt, J )
7.0, 7.0 Hz, 2H), 1.54 (m, 2H), 1.49 (m, 2H), 1.11 (d, J ) 6.5 Hz,
3H); IR (neat) νmax 3325, 2923, 2864, 1668, 1651, 1556, 1515, 1426,
1201, 1183, 1130 cm-1; FABHRMS (NBA-CsI) m/z 1017.3930 (M+,
C41H61N16O9S3 requires 1017.3970).
GM34454, J.S.) and the Skaggs Institute of Chemical Biology
(D.L.B.).
Supporting Information Available: Experimental proce-
dures and characterization of 9b-e, 10a-e, 11a-e, 12a-e,
14a-e, 15a-e, 17a-e, and 19a-e and experimental details
for the DNA cleavage studies are provided (14 pages, print/
PDF). See any current masthead page for ordering information
and Web access instructions.
Preparation of 8 from 19e: (42%).
Acknowledgment. We gratefully acknowledge the financial
support of the National Institutes of Health (CA42056, D.L.B.;
JA9816640