Cyclic disulfide C8 iminoporfiromycin: nucleophilic activation of a porfiromycin.

Article abstract of DOI:10.1021/ja030577r

The clinical success of mitomycin C (1) and its associated toxicities and resistance have led to efforts to prepare semisynthetic analogues (i.e., KW-2149 (3), BMS-181174 (4)) that have improved pharmacological profiles. In this study, we report the preparation and evaluation of the novel 7-N-(1'-amino-4',5'-dithian-2'-yl)porfiromycin C(8) cyclized imine (6) and its reference compound, 7-N-(1'-aminocyclohex-2'-yl)porfiromycin C(8) cyclized imine (13). Porfiromycin 6 contains a disulfide unit that, upon cleavage, may provide thiol(s) that affect drug reactivity. We demonstrated that phosphines dramatically accelerated 6 activation and solvolysis in methanolic solutions ("pH 7.4") compared with 13. Porfiromycins 6 and 13 efficiently cross-linked EcoRI-linearized pBR322 DNA upon addition of Et3P. We found enhanced levels of interstrand cross-link (ISC) adducts for 6 and 13 compared with porfiromycin (7) and that 6 was more efficient than 13. The large Et3P-mediated rate enhancements for the solvolysis of 6 compared with 13 and a N(7)-substituted analogue of 1, and the increased levels of ISC adducts for 6 compared with 13 and 7 are attributed to a nucleophile-assisted disulfide cleavage process that permits porfiromycin activation and nucleophile (MeOH, DNA) adduction. The in vitro antiproliferative activities of 6 and 13 using the A549 tumor cell line (lung adenocarcinoma) were determined under aerobic and hypoxic conditions and then compared with 7. Both 6 and 13 were more cytotoxic than 7, with 13 being more potent than 6. The C(8) iminoporfiromycins 6 and 13 displayed anticancer profiles similar to 3.

Full text of DOI:10.1021/ja030577r

Published on Web 03/12/2004
Cyclic Disulfide C(8) Iminoporfiromycin: Nucleophilic
Activation of a Porfiromycin
Sang Hyup Lee and Harold Kohn*
Contribution from the DiVision of Medicinal Chemistry and Natural Products,
School of Pharmacy, UniVersity of North Carolina at Chapel Hill,
Chapel Hill, North Carolina 27599-7360
Received October 6, 2003; E-mail: harold_kohn@unc.edu
Abstract: The clinical success of mitomycin C (1) and its associated toxicities and resistance have led to
efforts to prepare semisynthetic analogues (i.e., KW-2149 (3), BMS-181174 (4)) that have improved
pharmacological profiles. In this study, we report the preparation and evaluation of the novel 7-N-(1′-amino-
4′,5′-dithian-2′-yl)porfiromycin C(8) cyclized imine (6) and its reference compound, 7-N-(1′-aminocyclohex-
2′-yl)porfiromycin C(8) cyclized imine (13). Porfiromycin 6 contains a disulfide unit that, upon cleavage,
may provide thiol(s) that affect drug reactivity. We demonstrated that phosphines dramatically accelerated
6 activation and solvolysis in methanolic solutions (“pH 7.4”) compared with 13. Porfiromycins 6 and 13
efficiently cross-linked EcoRI-linearized pBR322 DNA upon addition of Et3P. We found enhanced levels of
interstrand cross-link (ISC) adducts for 6 and 13 compared with porfiromycin (7) and that 6 was more
efficient than 13. The large Et3P-mediated rate enhancements for the solvolysis of 6 compared with 13
and a N(7)-substituted analogue of 1, and the increased levels of ISC adducts for 6 compared with 13 and
7 are attributed to a nucleophile-assisted disulfide cleavage process that permits porfiromycin activation
and nucleophile (MeOH, DNA) adduction. The in vitro antiproliferative activities of 6 and 13 using the A549
tumor cell line (lung adenocarcinoma) were determined under aerobic and hypoxic conditions and then
compared with 7. Both 6 and 13 were more cytotoxic than 7, with 13 being more potent than 6. The C(8)
iminoporfiromycins 6 and 13 displayed anticancer profiles similar to 3.
Mitomycin C (1) is a clinically significant antineoplastic
agent.1 Mechanisms of action have been advanced whereby drug
function is initiated upon quinone ring reduction leading to
aziridine ring and C(10) carbamate cleavage with DNA adduc-
tion.2,3 In the 1960s, 1 was introduced into clinical use and today
is extensively employed in combination therapies for the
treatment of lung, breast, and other cancers.1,4 The clinical
success of 1 and its associated toxicities and resistance have
led to an active drug development program and the preparation
of over 1000 semisynthetic mitomycins.5
Three mitomycin analogues serve as the basis for this study.
In the 1980s, Remers et al.6 and Saito et al.7 showed that the
C(8) iminomitomycin 2 displayed excellent anticancer activities.
Subsequently, Wang and Kohn reported that 2 underwent acid-
mediated aziridine ring cleavage ∼100 times faster than 1.8 The
enhanced reactivity of 2 compared with 1 was attributed to the
diminished delocalization of the indoline N(4) electrons with
the adjacent R,â,γ,δ-unsaturated system permitting N(4)-assisted
expulsion of C(9a) methoxy group and mitomycin activation.
These findings suggested that 2 may react with DNA by both
nonreductive and reductive-mediated pathways. In the 1980s,
mitomycins 3 (KW-2149)9 and 4 (BMS-181174)10 were re-
(1) Carter, S. K.; Crooke, S. T. Mitomycin C. Current Status and New
DeVelopments, Academic Press: New York, 1979.
(2) (a) Iyer, V. N.; Szybalski, W. Science 1964, 145, 45-58. (b) Szybalski,
W.; Iyer, V. N. In Antibiotics. Mechanism of Action; Gottlieb, D., Shaw,
P. D., Eds.; Springer-Verlag: New York, 1967; Vol. 1, pp 211-245. (c)
Moore, H. W.; Czerniak, R. Med. Res. ReV. 1981, 1, 249-280.
(3) (a) Remers, W. A. “The Chemistry of Antitumor Antibiotics,” Vol. 1,
Wiley: New York, 1979; pp. 271-276. (b) Franck, R. W.; Tomasz, M. In
The Chemistry of Antitumor Agents, Wilman, D. F. V., Ed., Blackie and
Sons, Ltd.: Scotland, 1990; pp 379-394. (c) Tomasz, M. In Topics in
Molecular and Structural Biology: Molecular Aspects of Anticancer Drug-
DNA Interactions; Neidle, S., Waring, M., Eds.; Macmillan: New York,
1994; Vol. 2, pp. 312-347. (d) Tomasz, M. Chem. & Biol. 1995, 2, 575-
579.
(7) Saito, Y.; Kasai, M.; Shirahata, K.; Kono, M.; Morimoto, M.; Ashizawa,
T. (Kyowa Hakko Kogyo Co., Ltd.) U.S. Patent 4,853,385, Aug. 1, 1989;
Eur. Pat. Appl. EP 287,855, Oct. 26, 1988; JP Appl. 63246379, March 31,
1987; Chem. Abstr. 1989, 111, 7145v.
(8) Kohn, H.; Wang, S. Tetrahedron Lett. 1996, 37, 2337-2340.
(9) Kono, M.; Saitoh, Y.; Kasai, M.; Sato, A.; Shirahata, K.; Morimoto, M.;
Ashizawa, T. Chem. Pharm. Bull. 1989, 37, 1128-1130.
(10) Vyas, D. M.; Chiang, Y.; Benigni, D.; Rose, W. C.; Brander, W. T. In
Recent AdVances in Chemotherapy. Anticancer Section; Tshigami, J., Ed.;
University of Tokyo Press: Tokyo, 1985; pp. 485-486.
(4) Dollery, C. Therapeutic Drugs, 2nd ed.; Churchill Livingstone: Edinburgh,
1999; Vol. 2, pp. M197-M200.
(5) (a) Arai, H.; Kanda, Y.; Ashizawa, T.; Morimoto, M.; Gomi, K.; Kono,
M.; Kasai, M. J. Med. Chem. 1994, 37, 1794-1804. (b) Bradner, W. T.;
Remers, W. A.; Vyas, D. M. Anticancer Res. 1989, 9, 1095-1099.
(6) Iyengar, B. S.; Sami, S. M.; Remers, W. A.; Bradner, W. T.; Schurig, J. E.
J. Med. Chem. 1983, 26, 16-20.
9
10.1021/ja030577r CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 4281-4292
4281

A R T I C L E S
Lee and Kohn
ported. Both compounds contained a C(7) aminoethylene
disulfide group in place of the C(7) amino unit in 1. Signifi-
cantly, 3 was active in 1-resistant P388 and nonhypoxic cells.11
Both 3 and 4 were entered into clinical trials, and 3 advanced
to phase II testing. Compounds 3 and 4 are members of an
emerging class of multi-sulfur anticancer agents where cyto-
toxicity is associated, in part, with a nucleophile-assisted sulfur-
sulfur cleavage transformation.9,10,12-16 For 3 and 4, thiol-
assisted cleavage (e.g., glutathione (GSH)) of the C(7)
aminodisulfide unit provides 5, and 5 is projected to initiate
drug-DNA adduction.17-19 Thus, 2-4 are semisynthetic mi-
tomycins of pharmacological interest where multiple pathways
likely contribute to their bioactivity.
be initiated by intracellular thiol or serum albumin (CSH)-
mediated disulfide cleavage to give 8a and 8b (Scheme 1) and
Scheme 1. Envisioned Nucleophile-Mediated Activation Pathway
for 6
This study describes the synthesis and evaluation of porfiro-
mycin 6, which contains both a C(8) imino moiety and a
strategically placed cyclic disulfide unit. The parent compound
for 6, porfiromycin3a (7), is the N(1a) methyl derivative of 1
and is less prone to isomerization than 1.20 We envisioned that
6 activation can proceed by reductive, acid, and nucleophile-
mediated pathways. In this study, we document that phosphines
dramatically enhanced the activation rates of 6 compared with
7, and that C(8) iminoporfiromycins, such as 6, effectively cross-
link DNA under nonreductiVe conditions.
Results and Discussion
1. Substrate Design and Choice of Substrates. Porfiromycin
6 was designed to undergo nonreductive and reductive activa-
tion. We envisioned that nonreductive drug activation would
(11) Kobayashi, E.; Okabe, M.; Kono, M.; Arai, H.; Kasai, M.; Gomi, K.; Lee,
J.-H.; Inaba, M.; Tsuruo, T. Cancer Chemother. Pharmacol. 1993, 32, 20-
24.
(12) (a) Norton, R. A.; Finlayson, A. J.; Towers, G. H. N. Phytochemistry 1985,
24, 356-357. (b) Constabel, C. P.; Blaza, F.; Towers, G. H. N. Phytochem-
istry 1988, 27, 3533-3535. (c) Wang, Y.; Koreeda, M. J. Org. Chem. 1998,
63, 8644-8645.
(13) (a) Hara, M.; Takahashi, I.; Yoshida, M.; Kawamoto, I.; Morimoto, M.;
Nakano, H. J. Antibiot. 1989, 42, 333-335. (b) Hara, M.; Asano, K.;
Kawamoto, I.; Takiguchi, T.; Katsumata, S.; Takahashi, K.; Nakano, H. J.
Antibiot. 1989, 42, 1768-1774. (c) Behroozi, S. J.; Kim, W.; Gates, K. S.
J. Org. Chem. 1995, 60, 3964-3966.
(14) (a) Lee, M. D.; Ellestad, G. A.; Borders, D. B. Acc. Chem. Res. 1991, 24,
235-243. (b) Ellestad, G. A.; Hamann, P. R.; Zein, N.; Morton, G. O.;
Siegel, M. M.; Pastel, M.; Borders, D. B.; McGahren, W. J. Tetrahedron
Lett. 1989, 30, 3033-3036. (c) Myers, A. G.; Cohen, S. B.; Kwon, B. M.
J. Am. Chem. Soc. 1994, 116, 1255-1271.
(15) Makarieva, T. N.; Stonik, V. A.; Dmitrenok, A. S.; Grevnev, B. B.; Isakov,
V. V.; Rebachyk, N. M. J. Nat. Prod. 1995, 58, 254-258.
(16) (a) Davidson, B. S.; Molinski, T. F.; Barrows, L. R.; Ireland, C. M. J. Am.
Chem. Soc. 1991, 113, 4709-4710. (b) Searle, P. A.; Molinski, T. F. J.
Org. Chem. 1994, 59, 6600-6605.
(17) He, Q.-Y.; Maruenda, H.; Tomasz, M. J. Am. Chem. Soc. 1994, 116, 9349-
9350.
(18) Wang, S.; Kohn, H. J. Med. Chem. 1999, 42, 788-790.
(19) Na, Y.; Wang, S.; Kohn, H. J. Am. Chem. Soc. 2002, 124, 4666-4677.
(20) For related process, see: (a) Kono, M.; Saitoh, Y.; Kasai, M.; Shirahata,
K. J. Antibiotics 1995, 48, 179-181. (b) Kono, M.; Saitoh, Y.; Shirahata,
K. J. Am. Chem. Soc. 1987, 109, 7224-7225.
that these adducts may permit either increased cellular uptake
of the porfiromycin or permit drug activation to proceed by a
nonreductive pathway.17-19,21-23 Intermediates 8a and 8b are
expected to give the C(8) modified derivatives 9a and 9b or
the isomeric C(7) cyclic adducts (not shown), species likely to
undergo ring activation and DNA adduction.18,19 A similar
pathway has been proposed for 3.18,19,24 Critical for generation
of cyclic 9 was the placement of the disulfide unit three atoms
removed from the porfiromycin C(7) site. We reported that the
(21) Kobayashi, S.; Ushiki, J.; Takai, K.; Okumura, S.; Kono, M.; Kasai, M.;
Gomi, K.; Morimoto, M.; Ueno, H.; Hirata, T. Cancer Chemother.
Pharmacol. 1993, 32, 143-150.
(22) (a) Masters, J. R. W.; Know, R. J.; Hartley, J. A.; Kelland, L. R.; Hendricks,
H. R.; Conners, T. Biochem. Pharmacol. 1997, 53, 279-285. (b) McAdam,
S. R.; Knox, R. J.; Hartley, J. A.; Masters, J. R. W. Biochem. Pharmacol.
1998, 55, 1777-1783.
(23) Yasuzawa, T.; Tomer, K. B. Bioconjugate Chem. 1997, 8, 391-399.
(24) Wang, S.; Kohn, H. J. Org. Chem. 1997, 62, 5404-5412.
9
4282 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Cyclic Disulfide C(8) Iminoporfiromycin
A R T I C L E S
Scheme 2. Synthesis of Cyclic Disulfide 14
acyclic 7-N-(2-mercaptoethyl)porfiromycin (10) underwent cy-
clization, but the corresponding 7-N-(3-mercaptopropyl)por-
firomycin (11) did not.19 Thus, 6 was designed to be activated
by reductases, acids and nucleophiles and permit cellular
compartmentalization. The incorporation of the disulfide unit
within the appended cyclic array distinguishes our disulfide 6
from 3 and 4. This feature permits both thiol units to remain
fixed to the porfiromycin after disulfide cleavage and, thus, be
properly positioned to activate the porfiromycin and influence
its pharmacological properties.
We initially prepared 6 and the mitomycin analogue 12.
Because compound 12 underwent slow change upon isolation20
and storage, we focused our study on 6. We also synthesized
the corresponding cyclohexyl porfiromycin 1325 to serve as a
reference compound for 6.
a Ref 27. b Ref 28. c Ref 26.
diazide 17.27 NMR and TLC analyses for 17 showed only a
single product suggesting that azide displacement proceeded by
a SN2-type mechanism that led to inversion. Deprotection of
the benzyl group of 17 with BCl3/Me2S in CH2Cl2 gave 18.27
Catalytic hydrogenation of the azide groups in 18 using PtO2
under an atmospheric pressure of H2 afforded the key intermedi-
ate 19 ([R]589 ) -5.0 (c ) 0.63, MeOH), lit.27 [R]436
)
25
24
-7.3 (c ) 1.46, MeOH)).27 The amino groups in 19 were
protected with either BOC2O or BOC-ON to give 2028 and
then converted to the di-tosylated derivative 21 using TsCl, Et3N,
and DMAP. The toluenesulfonyl groups were subsequently
replaced by KSAc to provide 22. Compound 22 was also
obtained directly from 20 using the Mitsunobu reaction29 (Ph3P,
DEAD with AcSH). Hydrolysis of the thioacetate groups in 22
using K2CO3 in aqueous MeOH gave dithiol derivative 23,
2. Experimental Design. C(8) Iminoporfiromycins 6 and 13
were prepared and underwent chemical and biochemical tests
to determine the structural parameters that governed reactivity.
First, we monitored the solvolysis rates of 6 and 13 in the
absence and presence of select nucleophiles (thiols, phosphines)
in buffered methanolic solutions (“pH” 7.4). Next, we deter-
mined the extent to which 6 and 13 cross-linked complementary
DNA strands using linearized pBR322 DNA. The in vitro
cytotoxicities of 6 and 13 against human lung adenocarcinoma
cell line A549 were also determined. This information was not
used in our assessment of the structural requirements for
porfiromycin activation since neither the effective concentrations
of 6 and 13 within the A549 cells nor the factors that contributed
to the C(8) iminoporfiromycin inhibition of cell replication were
determined.
3. Synthesis. 3.1. The Cyclic Disulfide Bridge: 4,5-
Diamino-1,2-dithiane (14). Synthesis of 6 and 12 required the
intermediate preparation of 4,5-diamino-1,2-dithiane (14).26 Our
synthesis of 14 (Scheme 2) began with commercially available
(2R,3R)-1,4-dibenzyloxy-2,3-butanediol (15) and took advantage
of the synthetic route reported by Scheurer and co-workers for
(2R,3R)-2,3-diamino-1,4-butanediol (19).27 Treatment of 15 with
MsCl gave 16,27 which was then reacted with NaN3 to afford
1
which was characterized by H NMR (CHCH2SD methylene
signals: δ 2.75 (br s)) when the reaction was run in a deuterated
solvent (CD3OD-D2O). Derivative 23 was oxidized by O2/KOH
to provide 24.26 Compound 24 could also be obtained directly
from 22 by sequential hydrolysis followed by oxidation. Several
methods were attempted to remove the BOC groups in 24. Use
of either HCl/EtOAc30 or TMSI31 gave unsatisfactory results.
With TFA,32 we obtained diamino cyclic disulfide 14 in good
yield (∼100%). The overall yield of (R,R)-14 from (R,R)-15
was 19% (10 steps), and the reaction proceeded with stereo-
chemical control. Compound 14 has been referenced in the
patent literature,26 but no synthetic details, spectroscopic and
physical data for this diamine were provided.
3.2. Synthesis of Porfiromycin Imines 6 and 13. Using
disulfide 14 we prepared imines 6 and 12 (Scheme 3).
(28) Dondoni, A.; Merchan, F. L.; Merino, P.; Tejero, T.; Bertolasi, V. J. Chem.
Soc., Chem. Commun. 1994, 1731-1733.
(29) Mitsunobu, O. Synthesis 1981, 8, 1-28.
(25) Na, Y.; Kohn, H. Heterocycles 2001, 55, 1347-1363.
(26) (a) DuPriest, M. T.; York, B. M., Jr. (Alcon Laboratories, Inc.) U. S. Patent
4,659,733 (April 21, 1987). (b) DuPriest, M. T.; York, B. M., Jr. (Alcon
Laboratories, Inc.) U. S. Patent 4,755,528 (July 5, 1988).
(27) Scheurer, A.; Mosset, P.; Saalfrank, R. W. Tetrahedron: Asymmetry 1997,
8, 1243-1251.
(30) Stahl, G. L.; Walter, R.; Smith, C. W. J. Org. Chem. 1978, 43, 2285-
2886.
(31) Lott, R. S.; Chauhan, V. S.; Stammer, C. H. J. Chem. Soc., Chem. Commun.
1979, 495-496.
(32) Lundt, B. F.; Johansen, N. L.; Volund, A.; Markussen, J. Int. J. Peptide
Protein Res. 1978, 12, 258-268.
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4283

A R T I C L E S
Lee and Kohn
Scheme 3. Synthesis of Porfiromycin Imines 6 and 13
ppm, respectively, while the C(8a) carbon appeared downfield
(4.2-4.6 ppm) from 7. The structures for 6, 12, and 13 were
supported by their mass spectrum (low- and high-resolution),
which showed a molecular ion peak 18 mass units (-H2O) lower
than the expected signal for the corresponding non-cyclized
porfiromycin (mitomycin). These findings paralleled the spectral
results obtained for 2.8
5. Chemical Reactivity of C(8) Iminoporfiromycins: Sol-
volysis Studies. 5.1 Experimental Method. We first determined
the rate of methanolysis of C(8) iminoporfiromycins 6 and 13
in buffered methanolic solutions (“pH” 7.4, 25 °C) in the
absence and presence of nucleophiles to learn if the disulfide
unit in 6 affected ring activation. The reactions were monitored
by either UV-vis spectroscopy (200-600 nm) or HPLC using
UV-vis detection for greater than two half-lives, when possible,
and then the absorbance of the C(8) iminoporfiromycin (365
nm) was plotted against time. We used nonlinear regression
analysis to fit the observed exponential decay of the iminopor-
firomycin, using the SigmaPlot Program (SigmaPlot, 2001) to
provide pseudo-first-order rate constants. The reactions were
conducted in duplicate, and the results averaged. The products
(31 + 32, 33 + 34) were identified by co-injection with
authentic samples using HPLC and cospotting with authentic
samples using TLC.
Authentic samples of 6 and 13 solvolysis products, cis- and
trans-1-methoxy-2-methylaminomitosenes 31 + 32 and 33 +
34,35 respectively, were prepared by maintaining the porfiro-
mycins in acidic methanolic solutions (“pH” 5.5, 0.1 M bis-
Tris‚HCl, 18-21 h). In the HPLC chromatograms, we observed
the expected peaks for the cis- and trans-isomers, and the two
spots were also identified by TLC. The product mixtures from
6 and 13 were purified by PTLC and identified by UV-vis,
mass and 1H NMR spectroscopy. Compounds 31-34 displayed
UV maxima at ∼253 and ∼313 nm characteristic of 2-meth-
ylaminomitosenes.36 The mass spectra showed the expected
molecular ion peaks for the 1-methoxy-2-methylaminomitosenes
indicating that the imino structure remained intact during
solvolysis. The 1H NMR spectra permitted us to distinguish the
cis- and trans-1-methoxy-2-methylaminomitosenes. We found
that for the cis-isomers 31 and 33 the C(1)H-C(2)H and the
C(2)H-C(3)Hâ vicinal coupling constants were moderate
(∼5-7 Hz) and the C(2)H-C(3)HR coupling interaction was
large (∼9 Hz). Correspondingly, for the trans-isomers 32 and
34 the C(1)H-C(2)H vicinal coupling was near 0 Hz, the
C(2)H-C(3)Hâ vicinal coupling was small (∼0-1 Hz), and the
C(2)H-C(3)HR coupling interaction was moderate (∼5 Hz).
These coupling constant patterns have proven to be reliable
indicators of C(1) stereochemistry.37
Compound 14 was reacted with mitomycin F33 (MMF, 25) in
the presence of Et3N to give 6 (72% yield). Various reaction
conditions and solvents (e.g., CH2Cl2, EtOH, MeOH, CHCl3)
were examined for this reaction, and MeOH gave the best
results. The solution color changed from violet to dark red over
the 36-h reaction period. The intermediate 26 was not isolated
and is believed to undergo cyclization rapidly to give porfiro-
mycin imine 6. Using similar reaction conditions we obtained
mitomycin imine 12 in 71% yield from 14 and mitomycin A34
(MMA, 27). Compound 12 gradually underwent change during
PTLC purification (10% MeOH-CHCl3) and storage. We also
prepared the reference porfiromycin imine 1325 (66% yield)
using 29 and 25.
4. Structural Characterization of Porfiromycin (Mitomy-
cin) Imines 6, 12, and 13. Compounds 6, 12, and 13 showed
distinctive UV maxima at ∼224 and ∼366 nm, similar to those
previously observed for other C(8) iminomitomycins.8,25 We
observed the expected upfield shift (∼23.9 ppm) for the C(8)
imine carbon in the 13C NMR spectra for 6, 12, and 13 (152.8-
153.0 ppm) compared with the corresponding signal in 7 (176.8
ppm).8,25 In addition, the C(7) and C(5a) resonances for 6, 12,
and 13 were shifted upfield by 7.7-8.6 ppm and 10.1-10.3
5.2. Methanolysis in the Absence of Nucleophiles. At “pH”
7.4, the rate of 6 solvolysis was similar to that of 13 (Table
1A, kobs (d-1): 6, 0.062; 13, 0.069). HPLC and TLC analyses
indicated that the products were the expected cis- and trans-1-
methoxy-2-methylaminomitosenes (6: 31 + 32; 13: 33 + 34).
Correspondingly, mitomycins such as 1 and 7 did not undergo
appreciable change in buffered methanolic solutions at “pH”
7.4 (10 d), demonstrating that conversion of the C(8) quinone
unit to the imine facilitated mitosene production (data not
(35) Na, Y. Ph.D. Dissertation, University of Houston, Dec, 2000.
(36) Han, I.; Kohn, H. J. Org. Chem. 1991, 56, 4648-4653.
(37) Hornemann, U.; Iguchi, K.; Keller, P. J.; Vu, H. M.; Kozlowski, J. F.;
Kohn, H. J. Org. Chem. 1983, 48, 5026-5033.
(33) Kasai, M.; Kono, M. Synlett. 1992, 778-790.
(34) Hata, T.; Sano, Y.; Sugawara, R.; Matsomae, A.; Kanamori, K.; Shima,
T.; Hosha, T. J. Antibiot. 1956, 9, 141-146.
9
4284 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Cyclic Disulfide C(8) Iminoporfiromycin
A R T I C L E S
Table 1. Methanolysis Rates for 6 and 13 Using Nucleophiles at
5.3. Solvolysis in the Presence of Nucleophiles. We
measured the rate of solvolysis of C(8) iminoporfiromycins 6
and 13 at “pH” 7.4 in the presence of select nucleophiles (Table
1B-F). We chose two thiols (L-DTT, GSH) and three phos-
phines (Et3P, tris(2-carboxyethyl)phosphine hydrochloride39
(TCEP‚HCl), Ph3P). Ph3P was selected because it was found to
trigger the leinamycin activation cascade.40
5.3.1. Thiols. First, we examined the effect of L-DTT on
activation of 6 and 13. L-DTT has previously been used to cleave
the unsymmetrical disulfide units in the KW-2149 (3) and BMS-
181174 (4) analogues 35 and 36.18,24
“pH” 7.4a
reagents
6
13
equiv
kobs (d-1
)
t1/2 (d)
11
9.6
5.9
5.0
2.9
1.8
0.8
8.3
8.2
8.7
5.5
0.042
0.021
0.0083
0.0042
0.0017
5.8
0.13
0.083
3.1
kobs (d-1
)
t1/2 (d)
A. no nucH
B. L-DTT
0.062
0.072
0.120
0.140
0.240
0.390
0.870
0.083
0.085
0.080
0.130
16.6
33.3
83.2
166
415
0.120
0.069
0.073
0.077
0.079
0.390
0.410
0.430
-
-
0.144
10
9.5
9.0
8.8
1.8
1.7
1.6
-
2
5
10
20b,c
50b,c
100b,c
2
C. GSH
D. Et3P
5
-
0.5c
4.8
4.6
5.4
4.7
4.9
-
2c
0.151
0.128
0.147
0.141
5
10
20
50
100
2
5
10
2
-
-
-
-
-
-
-
-
-
E. TCEP (HCl)
F. Ph3P
-
5.50
8.30
0.220
0.260
0.460
-
-
-
5
10
2.7
1.5
-
We observed that the 6 and 13 activation rates gradually
increased (6.3-14-fold) as the number of equivalents of L-DTT
increased from 0 to 100 (Table 1B). Product analyses (HPLC,
TLC) indicated the formation of cis- and trans-1-methoxy-2-
methylaminomitosenes (6: 31 + 32; 13: 33 + 34). Only small
increases in the activation rate were observed over the first 10
equiv of L-DTT, with 6 undergoing a 2.3-fold increase in
activation and 13 a 1.1-fold rate increase. These findings
indicated that if L-DTT-mediated disulfide cleavage had occurred
in 6 then either the newly generated thiol 37 provided only a
slight rate boost or the thiol 37 rapidly regenerated the cyclic
disulfide 6.
-
a Reactions were run in buffered methanolic solution (0.1 M Tris‚HCl,
“pH” 7.4) at 25 °C. The reactions were run in duplicate and the values
averaged. The data were obtained using a Cary 3Bio Varian UV-visible
spectrophotometer and the reactions monitored at 365 ( 3 nm unless
indicated. The concentration of the porfiromycin was 0.03 mM unless
indicated. b The concentration of the porfiromycin was 0.06 mM. c The data
were obtained using HPLC.
Use of higher levels of L-DTT (10-100 equiv) led to further
increases in the methanolysis rates for 6 and 13. Since both
compounds were similarly affected, the high L-DTT levels likely
promoted ring activation by routes not central to the disulfide
unit in 6. These pathways can include L-DTT-mediated reduction
of the C(8) iminoquinone ring, L-DTT addition to the C(8) imine
unit, and L-DTT-initiated removal of the C(9) proton leading
to the loss of methanol at C(9) and C(9a).
When GSH was used as the nucleophile with 6 we obtained
results similar to those with L-DTT (Table 1C). The activation
rate of 6 was not measurably affected by GSH (2-5 equiv)
addition, and the products were the cis- and trans-1-methoxy-
2-methylaminomitosenes 31 + 32 (HPLC, TLC analyses). These
results mirrored those observed for L-DTT, a finding that
suggests that reclosure of the intermediate GSH-cleaved disul-
shown). We also monitored the solvolysis of 6 and 13 at “pH”
5.5 (kobs (d-1): 6, 6.30; 13, 12.8) and compared these values
with the “pH” 7.4 results. The rate of methanolysis increased
by 102-fold for 6 and 186-fold for 13. These results are in
agreement with previous studies and are consistent with the acid-
catalyzed loss of methanol at the C(9) and C(9a) pathways
previously proposed for both mitomycins38 and C(8) iminomi-
tomycins.8 On the basis of these observations we concluded that
the disulfide group exerted little effect on porfiromycin activa-
tion in the absence of nucleophiles between “pH” 5.5 and 7.4.
(38) (a) Stevens, C. L.; Taylor, K. G.; Munk, M. E.; Marshall, W. S.; Noll, K.;
Shah, G. D.; Shah, L. G.; Uzu, K. J. Med. Chem. 1964, 8, 1-10. (b)
Tomasz, M.; Lipman, R. J. Am. Chem. Soc. 1979, 101, 6063-6067. (c)
Iyengar, B. S.; Remers, W. A. J. Med. Chem. 1985, 28, 963-967. (d)
Hornemann, U.; Keller, P. J.; Takeda, K. J. Med. Chem. 1985, 28, 31-36.
(e) Rebek, J.; Shaber, S. H.; Shue, Y.-K.; Gehret, J.-C.; Zimmerman, S. J.
Org. Chem. 1984, 49, 5164-5174. (f) McClelland, R. A.; Lam, K. J. Am.
Chem. Soc. 1985, 107, 5182-5186. (g) Boruah, R. C.; Skibo, E. B. J. Org.
Chem. 1995, 60, 2232-2243.
(39) Burns, J. A.; Butler, J. C.; Moran, J.; Whitesides, G. A. J. Org. Chem.
1991, 56, 2648-2650.
(40) Zang, H.; Breydo, L.; Mitra, K.; Dannaldson, J.; Gates, K. S. Bioorg. Med.
Chem. Lett. 2001, 11, 1511-1515.
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4285

A R T I C L E S
Lee and Kohn
fide to 6 competes with porfiromycin activation. This notion is
also supported by the relative rates of GSH activation of 6 and
3. The half-live for GSH (5 equiv) mediated solvolysis of 6
was ∼8 d, while acyclic disulfide 3 underwent efficient
activation at near neutral pH within 2 h when treated with excess
GSH.17 Compound 3 is less likely to regenerate starting material
upon GSH cleavage than 6. Results similar to 3 were observed
for 35 upon treatment with DTT.18
5.3.2. Phosphines. Next, we determined the effect of alkyl
and aryl phosphines on 6 and 13 activation (Table 1D-F). The
nonbonding electron pair present on phosphorus permit phos-
phines to be classified as both organic bases and nucleo-
philes.41,42 Phosphine nucleophilicity depends on the reaction
type (SN1, SN2), the phosphorus substituents, the electrophile
(hard, soft), and the displaced leaving group. Nevertheless,
phosphines are often more nucleophilic than their sulfur and
nitrogen counterparts42a and have been used in desulfurization
reactions43 and to cleave cyclic and protein disulfide bonds.44-46
In addition, phosphines are known to react with carbonyl groups
and aziridines.47
Use of 5 equiv of Et3P with 6 led to a 270-fold increase in
2-methylaminomitosene product formation, and use of 10 equiv
of Et3P enhanced the rate 540-fold (Table 1D). Only modest
increases (∼2-fold) were observed for the reference compound
13 when Et3P (10-20 equiv) was added. Far greater rate
differences were observed when we compared 6 with a N(7)-
substituted analogue of 1 where no appreciable mitosene
production occurred after 8 d with Et3P (50 equiv).48 The Et3P-
mediated activation rate of 6 was linear with phosphine
concentration (5-100 equiv). Interestingly, the rapid rate
increases for 6 were observed only after addition of 5 or more
Et3P equiv. We suspect that this represents the threshold level
of Et3P needed for porfiromycin nucleophilic activation and
likely reflects the consumption of Et3P by adventitious amounts
of O2 present in reaction medium and that were not purged by
Ar. This is the first example of a phosphine activation of a
porfiromycin (mitomycin).
Our finding that Et3P affected only 6 activation and not 13
indicated that the disulfide group in 6 was involved in the drug
activation process. Accordingly, we suggest that Et3P reacted
with the disulfide unit to generate thiol 38, which in turn
activated the quinone ring permitting 2-methylaminomitosene
formation. The 2-fold rate increase in solvolysis observed for
13 upon Et3P (0.5-20 equiv) addition was similar to the result
obtained with L-DTT and may reflect additional roles for Et3P
(e.g., C(8) iminoquinone reduction, imine addition, base-
catalyzed removal of methanol at C(9) and C(9a)) in the overall
drug activation process.
(Table 1E). Significant increases in the solvolysis rates were
only observed with 5 or more equiv of the phosphine. With 5
equiv, the rate enhancement was 89-fold, and with 10 equiv, it
was 134-fold.
The final phosphine employed was Ph3P. The rate enhance-
ments observed for 6 activation were less than those with Et3P
and TCEP (Table 1F). We detected a 4.2-fold activation with 5
equiv of Ph3P and a 7.4-fold activation with 10 equiv. We have
attributed the diminished level of activation for these transfor-
mations to the decreased nucleophilicity of aromatic (Ph3P)
versus aliphatic (Et3P, TCEP) phosphines.41,49
The reaction rates for most entries in Table 1 were determined
by UV-vis spectroscopy on which we observed a decrease in
the absorption for the starting C(8) iminoporfiromycin (∼365
nm) and a concomitant increase in absorption for the C(8) imino-
2-methylaminomitosene products (∼313 nm). When the 0.5 and
2 equiv Et3P reactions were monitored by HPLC, the loss of
starting porfiromycin was seen to correlate with the UV-vis
spectroscopic data. In these cases, we observed the HPLC peaks
corresponding to cis- (31, tR 31.8 min) and trans- (32, tR 29.7
min) 1-methoxy-2-methylaminomitosenes. However, when HPLC
was used to monitor the methanolysis of 6 with g 10 equiv of
Et3P (data not shown) the progressive loss of 6 was observed,
but we could not monitor the concomitant increase in 31 and
32. For the 10 equiv Et3P reaction, we detected at the outset
(0-1 h) 31 and 32, but these products diminished with time
and no new adducts were recorded by HPLC. When 20 equiv
of Et3P were employed we were only able to follow the loss of
6 and did not observe 31 and 32. These results differed from
those observed for 13. Treatment of 13 with Et3P (10 and 20
equiv) led to the production of 33 and 34 (data not shown).
Correspondingly, with Ph3P (2-10 equiv) we were able to
follow 31 and 32 production on HPLC as 6 was consumed (data
not shown). Thus, we suspect that the excess Et3P present in
the reaction medium reacted with the 1-methoxy-2-methylami-
nomitosene products (31 and 32) leading to adducts that were
either unstable or unable to elute from the HPLC column under
the experimental conditions. Several unsuccessful attempts were
made to identify the products generated under these high Et3P
conditions (data not shown). First, we passed O2 through the
solution (2 d) but observed no change in the HPLC and TLC
profiles. Second, we reduced the temperature for the 6 plus Et3P
(10 equiv) reaction from 25 °C to -50 °C. The HPLC product
profiles were similar to those observed at room temperature.
Finally, we treated the cis-(31) and trans-(32) 1-methoxy-2-
methylaminomitosenes with Et3P (10 equiv, buffered methanol,
“pH” 7.4, 25 °C). After 1 h, we were unable to detect 31 and
32 (HPLC, TLC analyses). We concluded that the excess Et3P
(g10 equiv) in the 6 methanolysis reactions consumed the
initially generated 31 and 32 products.
When TCEP (2-10 equiv) was used in place of Et3P, we
again observed increased rate enhancements for 6 activation
5.4. Reflections on the Mechanism for Phosphine-Medi-
ated 6 Solvolysis. Addition of phosphines to methanolic
solutions of 6 led to rapid porfiromycin consumption and
2-methylaminomitosene production (Table 1D-F). We observed
that C(8) iminoporfiromycin activation depended on the strength
and concentration of the phosphine. Analysis of the correspond-
ing data for 13 showed that the large rate enhancements were
attributable to a phosphine-mediated disulfide cleavage process.
We propose that the rate-limiting step is the phosphine rupture
(41) (a) Overman, L. E.; Matzinger, D.; O’Connor, E. M.; Overman, J. D. J.
Am. Chem. Soc. 1974, 96, 6081-6089. (b) Overman, L. E.; Petty, S. T. J.
Org. Chem. 1975, 40, 2779-2782. (c) Overman, L. E.; O’Connor, E. M.
J. Am. Chem. Soc. 1976, 98, 771-775.
(42) (a) Bartlett, P. D.; Lohaus, G.; Weis, C. D. J. Am. Chem. Soc. 1958, 80,
5064-5069. (b) Hudson, F. Organic Chemistry: Structure and Mechanism
in Organo-Phosphorous Chemistry, Vol 6; Blomquist, A. T., Ed.; Academic
Press: London and New York, 1965; p 16.
(43) Moore, C. G.; Trego, B. R. Tetrahedron 1962, 18, 205-218.
(44) Sweetman, B. J.; Maclaren, J. A. Aust. J. Chem. 1966, 19, 2347-2354.
(45) Maclaren, J. A.; Sweetman, B. J. Aust. J. Chem. 1966, 19, 2355-2360.
(46) Lee, S. H.; Kohn, H. Heterocycles 2003, 60, 47-56.
(47) Fan, R.-H.; Hou, X.-L. J. Org. Chem. 2003, 68, 726-730.
(48) Lee, S. H. Ph.D. Dissertation, 2003, University of North Carolina, Chapel
Hill.
(49) Parker, A. J.; Kharasch, N. J. Am. Chem. Soc. 1960, 82, 3071-3075.
9
4286 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Cyclic Disulfide C(8) Iminoporfiromycin
A R T I C L E S
Scheme 4. Proposed Phosphine-Mediated Activation Pathway for
6 in MeOH
Figure 1. Denaturing 1.2% Alkaline Agarose Gel for 6, 7, and 13 (0.05
mM) using Et3P (5 equiv). DNA cross-linking experiments for 6, 7, and 13
at 0.05 mM concentration using EcoRI-linearized pBR322 plasmid DNA
and Et3P (5 equiv). All reactions were incubated at room temperature
(2 h). Lane 1: λ Hind III DNA molecular weight marker. Lane 2: control
(only linearized pBR322). Lane 3: 7 + Et3P (5 equiv). Lane 4: 13 + Et3P
(5 equiv). Lane 5: 6 + Et3P (5 equiv). Lane 6: only Et3P (5 equiv).
40. Formation of 40 is expected to lead to rapid aziridine ring
opening8 to give 41 followed by attack of the nucleophile
(NucH) to give 42. An alternative activation pathway is
conceivable where thiol 38a gives the C(7) cyclized species,
which then undergoes further nucleophilic attack to give the
leucoporfiromycin. We have previously suggested a comparable
route for 5 upon thiol activation of 3 and 4.19 Both pathways
are supported by the kinetic studies in which the solvolysis
products in methanol were 31 and 32 (HPLC, TLC analyses)
and the reaction rates depended on the nucleophile and its
concentration.
Depicted in Scheme 4 is a proposed pathway for C(10)
activation (42 f 43 f 44). Although we did not detect C(1),
C(10)-disubstituted products (i.e., 44), the disruption of the
N(4)-C(5a)-C(8a)-C(8)-N conjugated system in 42 is ex-
pected to activate the C(10) site toward nucleophilic substitution
(see Sections 6.2 and 6.5).
6. C(8) Iminoporfiromycins-DNA Bonding Profiles. Mi-
tomycin-DNA modifications include both monofunctional and
bifunctional alkylation processes (intrastrand and interstrand
cross-links (ISC)).3 The efficiency of DNA cross-linking
depends on the structure of the mitomycin and the activation
conditions. Compound 6 was designed to undergo activation
under nonreductive (e.g., nucleophilic) and reductive conditions
leading to DNA adduction.
6.1. Experimental Method. The ability of C(8) iminopor-
firomycins 6 and 13 to cross-link complementary EcoRI-
linearized pBR322 DNA was determined using denaturing
alkaline agarose gel electrophoresis as reported by Cech.50a The
size of the DNA product(s) was estimated using λ DNA digested
with HindIII as a molecular weight marker. We adopted the
term DNA ISC for interstrand cross-linked (ISC) DNA adducts.
The extent of DNA ISC formation for 6 and 13 was determined
under nonreductive (Et3P) and reductive (Na2S2O4) conditions.
6.2. Effect of Et3P on DNA ISC. Preliminary experiments
led us to choose 5 equiv of Et3P and 50 µM porfiromycins 6,
7, and 13. The reactions were run at room temperature (2 h).
We found that both 6 and 13 formed DNA ISC (6: 96%, 13:
73%) while 7 exhibited low levels of DNA ISC (5%) (Figure
1). Interestingly, the band for the 6 DNA ISC was more diffuse
and encompassed DNA products with enhanced gel mobilities,
compared with 13. Although the origin of this phenomenon has
of the disulfide bond in 6. In Scheme 4, we postulate that
disulfide cleavage occurred at the sulfur atom closest to the C(8)
imine bond leading to thiol 38a. Alternatively, disulfide cleavage
may have occurred at the other sulfur site to give 38b (pathway
not shown). There is precedent for the proposed phosphine-
mediated disulfide transformation (6 f 38a). Alkyl phosphines
have been utilized to cleave disulfide units in peptides,44,45 and
recently we showed that Et3P and TCEP efficiently cleaved
cyclic disulfides (e.g., DTTox) under neutral to basic conditions.46
A similar transformation has also been reported for leinamycin
and Ph3P.40
We project that thiol 38a undergoes intramolecular cyclization
at the C(8) imine carbon to give 39. Disruption of the N(4)-
C(5a)-C(8a)-C(8)-N conjugated system in 39 unleashes the
N(4) electrons leading to 1-methoxy-2-methylaminomitosene
(50) (a) Cech, T. R. Biochemistry 1981, 20, 1431-1437. (b) Tepe, J. J.; Williams,
R. M. J. Am. Chem. Soc. 1999, 121, 2951-2955.
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4287

A R T I C L E S
Lee and Kohn
Figure 3. Denaturing 1.2% Alkaline Agarose Gel for 6, 7, and 13 (0.2
mM) using Na2S2O4 (1 equiv). DNA cross-linking experiments for 6, 7,
and 13 at 0.2 mM concentration using EcoRI-linearized pBR322 plasmid
DNA and Na2S2O4 (1 equiv). All reactions were incubated at 0 °C (1 h).
Lane 1: λ Hind III DNA molecular weight marker. Lane 2: control (only
linearized pBR322). Lane 3: 7 + Na2S2O4 (1 equiv). Lane 4: 13 + Na2S2O4
(1 equiv). Lane 5: 6 + Na2S2O4 (1 equiv). Lane 6: only Na2S2O4 (1 equiv).
Figure 2. Denaturing 1.2% Alkaline Agarose Gel for 6, 7, and 13 (0.01
mM) using L-DTT (5 equiv). DNA cross-linking experiments for 6, 7, and
13 at 0.01 mM concentration using EcoRI-linearized pBR322 plasmid DNA
and L-DTT (5 equiv). All reactions were incubated at room temperature
(2 h). Lane 1: λ Hind III DNA molecular weight marker. Lane 2: control
(only linearized pBR322). Lane 3: 7 + L-DTT (5 equiv). Lane 4: 13 +
L-DTT (5 equiv). Lane 5: 6 + L-DTT (5 equiv). Lane 6: only L-DTT
(5 equiv).
The modest increase in DNA ISC with L-DTT (5 equiv) for
6 compared with 13 paralleled the differences observed in the
solvolysis rates for these compounds (Table 1B: 6, 0.12 d-1
;
not been determined, we suspect that the diffuse band may result
from extensive DNA ISC adduction,51a multiple ISC adducts
within the linearized DNA fragment, and DNA ISC isomers
that differed in their site of DNA ISC formation.51b,c We expect
that adducts with altered DNA structures will display different
mobilities under denatured agarose gel conditions.51b
We attempted to distinguish the Et3P-mediated activation of
6 and 13 further by using lower porfiromycin concentrations
(10 µM) (Supporting Figure 1). Under these conditions, 6
efficiently generated DNA ISC (89%), but the extent of DNA
ISC for 13 was 35% and for 7 was ∼2%. Again, we observed
that the band for 6 DNA ISC adducts was more diffuse than
that for 13.
The enhanced DNA ISC efficiency of 6 compared with 13
upon Et3P addition is attributed to the role of the disulfide unit
in the drug activation steps. Moreover, the observation of DNA
ISC adducts documents that Et3P activation of 6 leads to
modification of both the C(1) and C(10) sites within the
porfiromycin (Scheme 4). The DNA studies were unable to
differentiate the extent of DNA ISC within any given DNA
fragment. The increased efficiency of 6 compared with 13
paralleled the trend observed in the kinetic studies (Table 1D),
but the large difference (5 equiv Et3P: 130-fold) found in the
kinetic studies were not observed in the DNA experiments. We
believe that only qualitative comparisons of these data sets can
be made and that direct comparisons are not warranted unless
the number of DNA ISC sites are known.
6.3. Effects of Thiols. We next determined the effects of
L-DTT and GSH on porfiromycin 6 and 13 DNA ISC trans-
formations and compared the results with 7. Using the conditions
established for Et3P, we treated aqueous buffered solutions
containing DNA and 6, 7, or 13 (10 µM) with L-DTT (5 equiv)
at room temperature (2 h). Gel analysis (Figure 2) showed that
L-DTT (5 equiv) led to appreciable levels of DNA ISC for 6
(61%) and 13 (48%) and a low level (5%) for 7. Only minor
differences existed in the ISC bands for 6 and 13. A similar
result was obtained for GSH using the same conditions (ISC:
6, 48%; 13, 47%; 7, 3%; see Supporting Figure 2).
13, 0.077 d-1). Surprising to us, however, was the amount of
DNA ISC produced for each compound with L-DTT, which
neared 50%, despite the finding that both compounds underwent
slow solvolysis. We have attributed the high percentage of DNA
ISC to the number of potential adduction sites within the 4361-
bp EcoRI-linearized pBR322 DNA. Mitomycin C ISC processes
have been shown to proceed at 5′CG‚5′CG sites.52 Thus, ISC
adduction at any one of these many sites would lead to a slower
migrating band in the denaturing 1.2% alkaline agarose gel,
despite the fact that only low levels of either 6 or 13 were
activated. In light of this analysis we are uncertain to attribute
the modest difference in the extent of DNA ISC for 6 and 13
with L-DTT (Figure 2) to L-DTT-mediated activation of the 6
disulfide unit.
6.4. Effect of the Chemical Reductant Sodium Dithionite
(Na2S2O4). Mitomycins, such as 1 and 7, are activated in vitro
by chemical reductants (e.g., Na2S2O452,53) and in vivo by
reductases.1,3,54 Thus, previous studies have examined the extent
of DNA ISC for mitomycins (porfiromycins) under reductive
conditions.52 In most studies, chemical reductants were em-
ployed.52,53 We chose to use Na2S2O4 and previously employed
conditions that led only to efficient mitomycin C(1) activation.55
Porfiromycins 6, 7, and 13 served as our test compounds, and
the reactions were conducted at 0 °C (1 h) using 0.2 mM drug
concentration and 1 equiv of Na2S2O4. We found that under
these conditions C(8) iminoporfiromycins 6 and 13 provided
high levels of DNA ISC (6: 94%, 13: 96%) while 7 showed
the expected low amounts (5%) (Figure 3). Therefore, we
concluded that under reductive conditions, imines 6 and 13 were
more effective in forming DNA ISC than was 7. These results
may reflect the ease with which iminoquinones undergo
reduction56 or the ability of Na2S2O4 or its byproduct, SO3
2,53b
to assist porfiromycin activation by either C(8) imine addition
(52) Tomasz, M.; Lipman, R.; Chowdary, D.; Pawlak, J.; Verdine, G. L.;
Nakanishi, K. Science 1987, 235, 1204-1208.
(53) (a) Schiltz, P.; Kohn, H. J. Am. Chem. Soc. 1993, 115, 10 510-10 518.
(b) Schlitz, P.; Kohn, H. J. Am. Chem. Soc. 1993, 115, 10 497-10 509.
(54) Mikami, K.; Naito, M.; Ishuiguro, T.; Yano, H.; Tomida, A.; Yamada, T.;
Tanaka, N.; Shirakusa, T.; Tsuro, T. Jpn. J. Cancer Res. 1998, 89, 910-
915, and references therein.
(51) (a) Johnson, W. S.; He, Q.-Y.; Tomasz, M. Bioorg. Med. Chem. 1995, 3,
851-860. (b) Serwer, P.; Allen, J. L. Biochemistry 1984, 23, 922-927.
(c) Millard, J. T.; Spencer, R. J.; Hopkins, P. B. Biochemistry 1998, 37,
5211-5219.
(55) (a) Li, V.-S.; Kohn, H. J. Am. Chem. Soc. 1991, 113, 275-283. (b) Kohn,
H.; Li, V.-S.; Tang, M.-s. J. Am. Chem. Soc. 1992, 114, 5051-5509. (c)
Li, V.-S.; Tang, M.-s. Bioorg. Med. Chem. 2001, 9, 863-873.
(56) Corbett, J. F. J. Chem. Soc., B, 1969, 207-212.
9
4288 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Cyclic Disulfide C(8) Iminoporfiromycin
A R T I C L E S
Table 2. Antiproliferative Activities for C(8) Iminoporfiromycinsa
or C(9) proton abstraction. Interestingly, in these experiments
the 13 DNA ISC products (Figure 3, lane 4) were more diffuse
than the corresponding 6 adducts and appeared as two, closely
spaced bands. The gel findings provided no evidence that the
disulfide unit in 6 played a role in porfiromycin activation under
reductive conditions, and thus we suspect iminoquinone reduc-
tion superseded thiol-activation pathways (e.g., Scheme 1).
6.5. Reflections Concerning the DNA Bonding Profiles for
6 and 13 and the Differences in the Kinetic and DNA
Adduction Findings. Significant levels of DNA ISC were
obtained for 6 and 13 when phosphine (Et3P) and thiols (L-
DTT, GSH) were added. Careful adjustment of the reaction
condition permitted us to differentiate 6 from 13 with Et3P
(Figure 1 and Supporting Figure 1). With Et3P, we observed
the appearance of a diffuse DNA band corresponding to the
DNA ISC products. We have partially attributed this band to
extensively cross-linked DNAs. If this is the case, then the
calculated percentage cross-linked efficiencies of 6 and 13
underestimated the degree of ISC within the DNA and the role
of the disulfide unit in 6 in promoting these transformations.
Why did we see appreciable levels of DNA ISC with 13 but
not with 7 upon addition of nucleophiles (Et3P, L-DTT, GSH)?
The imino group in 13 (6) is more easily reduced than the
quinone unit in 7,56 and phosphine and thiols can both serve as
reducing agents.44,45 Furthermore, the imino unit in 13 (6) is
expected to undergo nucleophilic addition.57 Generating a C(8)
tetrahedral intermediate deconjugates the N(4)-C(5a)-C(8a)-
C(8)-N unsaturated system and activates the C(8) iminopor-
firomycin toward 2-methylaminomitosene formation. The find-
ing that C(8) iminoporfiromycins undergo extensive DNA ISC
compared with 7 has not been reported.
The HPLC data for the 6 kinetic methanolysis experiments
with L-DTT and Et3P showed no measurable amounts of C(1),
C(10)-dimethoxy adducts 44 (Nuc ) OMe) yet we obtained
clear evidence of 6 C(1), C(10) modification in the presence of
DNA in aqueous buffered solutions. A similar result was
obtained for 13. How do we reconcile these results? Previous
studies have shown that the C(1) site is 10-100 times more
reactive than C(10) under reductive conditions.58 If this is a
generalizable finding, we would expect to see few C(1), C(10)-
dimethoxy products59 in our kinetic studies. This difference in
C(1) and C(10) site reactivity is more difficult to discern with
the EcoRI DNA (4361 bp) where even a single ISC would lead
to the appearance of a slower migrating band in the denaturing
1.2% alkaline agarose gel.
IC50 (µmol/L)
IC50 (hypoxic)
IC50 (aerobic)
compd
aerobic
hypoxic
6
13
1
7
3
0.56
0.11
2.4
13
0.12
2.3
4.1
4.1
8.2
>7.6
3.4
0.47
20
>100
0.41
a A549 cell line was used. The IC50 (µmol/L) value is the concentration
that inhibits cell replication by 50% under the assay condition.
low ratios determined from the IC50 values under hypoxic and
aerobic conditions (IC50 (hypoxic)/IC50 (aerobic)).61 For 3, this
ratio was 3.4 and for 1 and 7 it was 8.2 and >7.6, respectively.
Thus, we were interested to see if either 6 or 13 had ratios of
activities comparable with either 3 or 1 (7). The cells were
precultured at 37 °C (24 h) under either aerobic or hypoxic
conditions and then treated with the drug candidates (72 h).
7.2. Antiproliferative Activities. The in vitro antiproliferative
activity test data for 1, 3, 6, 7, and 13 using the A549 tumor
cell line are listed in Table 2. In general, the cell proliferation
rate under hypoxic conditions was slower than under aerobic
conditions. We observed that 6 was 25-fold more potent than 7
under aerobic and >50-fold more cytotoxic under hypoxic
conditions. The ratio of activity (hypoxic/aerobic) for 6 was
4.1, for 7 was >7.6, and for 3 was 3.4. When we compared 6
with its cyclohexyl equivalent 13, we found it was ∼5-fold less
active than 13 under aerobic and hypoxic conditions, and the
activity ratio (hypoxic/aerobic) remained constant (4.1). KW-
2149 (3) showed potent activity under both aerobic and hypoxic
conditions and was 4.7-5.6-fold more potent than 6. The
porfiromycin 6 and 13 in vitro antiproliferative profiles with
the A549 cell line showed no apparent advantages by incorpo-
rating a disulfide unit in the appended cyclohexyl ring. Thus,
under these conditions we saw no evidence that the nucleophile-
mediated activation pathway (Scheme 1) contributed to cell
growth inhibition.
Conclusions
We report the synthesis and evaluation of the novel C(8)
iminoporfiromycin 6. This compound contained two key
structural elements that were expected to enhance drug activation
and DNA adductionsthe C(8) imino group and the disulfide
unit. Earlier studies documented that when the C(8) carbonyl
group in 1 was replaced with an imino unit to give 2 there was
an approximate 100-fold increase in the rate of ring activation
under acidic conditions.8 Recent findings have documented that
the disulfide unit likely contributes to the enhanced activity of
3 and may be necessary for its efficient activation and usage
under aerobic conditions. Accordingly, we envisioned that 6
would be activated by reductases, acids, and nucleophiles.
Distinguishing 6 from 3 and 4 was accomplished by incorporat-
ing the disulfide unit in a fused six-membered ring system. This
feature permitted both thiol units that were generated upon
disulfide cleavage to remain appended to the porfiromycin and
positioned to activate the porfiromycin and influence its
pharmacological properties.
7. Cytotoxicities. 7.1. Experimental Approach. We tested
whether C(8) iminoporfiromycins 6 and 13 inhibited tumor cell
line replication using an in vitro antiproliferative activity assay60
at the Kyowa Hakko Pharmaceutical Company (Shizuoka,
Japan). The tests were conducted using the human tumor cell
line A549 (lung adenocarcinoma) because of its sensitivity to
KW-2149 (3). The antiproliferative activities of these com-
pounds were determined under aerobic and hypoxic conditions
and then compared with 3. A distinguishing feature previously
found for 3, compared with other mitomycins (e.g., 1), was the
(57) Corbett, J. F. J. Chem. Soc., B, 1969, 213-216.
(58) Hornemann, U.; Keller, P. J.; Kozlowski, J. F. J. Am. Chem. Soc. 1979,
101, 7121-7124, and references therein.
We found that moderate amounts of Et3P dramatically
enhanced 6 activation compared with 13. With 10 equiv of Et3P
(59) Hong, Y. P.; Kohn, H. J. Am. Chem. Soc. 1991, 113, 4634-4644.
(60) Skehan, P.; Storeng, R.; Scudiero, D.; Monks, A.; McMahon, J.; Viotica,
D.; Warren, J. T.; Bokesch, H.; Kenny, S.; Boyd, M. R. J. Natl. Cancer
Inst. 1990, 82, 1107-1112.
(61) Dr. H. Arai, private communication.
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4289

A R T I C L E S
Lee and Kohn
(40 mL) was added to the residue, and the mixture was extracted with
EtOAc (2 × 40 mL). The combined organic layers were dried (MgSO4)
and concentrated in vacuo. Purification by PTLC (1:2 EtOAc/hexanes)
afforded 22 as a white solid: yield, 24 mg (34%); mp 154-163 °C;
1H NMR (CDCl3) δ 1.43 (s, 18 H, OC(CH3)3), 2.35 (s, 6 H, C(O)-
CH3), 3.05-3.12 (m, 4 H, SCH2), 3.75-3.86 (m, 2 H, CHNH), 4.86
(br s, 2 H, NHCO); 13C NMR (CDCl3) δ 28.3 (OC(CH3)3), 30.5 (C(O)-
CH3), 31.7 (SCH2), 54.0 (CHNH), 79.8 (OC(CH3)3), 156.2 (NHCO),
195.7 (COCH3).
the difference in reactivity was 230-fold. Far greater rate
differences have been observed when 6 was compared with a
N(7)-substituted analogue of 1,48 which underwent no ap-
preciable change after 8 d with Et3P (50 equiv). These findings
are consistent with a Et3P-mediated activation pathway (Scheme
4) that involves phosphine attack of the disulfide unit in 6.
Consistent with this pathway, the rates of 6 consumption
increased with increasing phosphine concentrations and with
increasing nucleophilicity of the phosphine. This is the first
example in which phosphines were used to activate mitomycins
(porfiromycins). The efficiency of 6 to generate DNA ISC was
assessed. We found significantly enhanced levels of DNA ISC
for 6 and 13 compared with 7 and that 6 was found more
efficient than 13. These results substantiate the nucleophile-
mediated pathway for mitomycin (porfiromycin) activation and
documents its efficiency for DNA ISC. This route complements
the previously reported acid-catalyzed and reductive pathways
for mitomycin (porfiromycin) activation.
(4R,5R)-trans-4,5-Bis(tert-butyloxycarbonylamino)-1,2-dithiane (24).
To a stirred solution of 22 (31 mg, 0.07 mmol) in MeOH-H2O (5:1,
2 mL) was added K2CO3 (58 mg, 0.42 mmol). After stirring at room
temperature (30 min), KOH (8 mg, 0.15 mmol) was added and O2 was
bubbled through the solution (5 h). The solvent was removed in vacuo
and H2O (20 mL) was added to the residue. The mixture was extracted
with EtOAc (2 × 20 mL) and the combined organic layers were dried
(MgSO4) and concentrated in vacuo. Purification by PTLC (1:2 EtOAc/
25
hexanes) afforded 24 as a sticky solid: yield, 23 mg (92%); [R]589
)
+46 (c ) 0.19, CHCl3); Rf 0.51 (1:2 EtOAc/hexanes); IR (KBr) 3363,
1
2978, 2360, 1682, 1520, 1057 cm-1; H NMR (CDCl3) δ 1.44 (s, 18
H, OC(CH3)3), 2.83 (br s, 2 H, SCHH′), 3.15 (br s, 2 H, SCHH′), 3.71
(br s, 2 H, CHNH), 5.06 (br s, 2 H, NHCO); 13C NMR (CDCl3) δ 28.4
(OC(CH3)3), 40.5 (SCH2), 55.4 (CHNH), 80.1 (OC(CH3)3), 155.7
(NHCO); MS (+CI) m/z 351 [M+1]+; Mr (+CI) 351.140 62 [M+1]+
(calcd for C14H27N2O4S2 351.141 23).
Experimental Section
(General experimental details can be found in the Supporting
Information.)
(2R,3R)-2,3-Bis(tert-butyloxycarbonylamino)-1,4-bis(acetylthio)-
butane (22).
(4R,5R)-trans-4,5-Diamino-1,2-dithiane‚2TFA (14). Compound 24
(21 mg, 0.06 mmol) was dissolved in TFA (2.0 mL) and stirring was
continued at room temperature (30 min). The reaction was concentrated
in vacuo to afford 14 as a viscous oil: yield, 23 mg (∼100%);
Method A. Ph3P (125 mg, 0.48 mmol) was dissolved in THF
(2 mL) and DEAD (74 µL, 0.47 mmol) was added at room temperature.
The solution was cooled to 0 °C and stirring was continued (20 min).
Compound 20 (64 mg, 0.20 mmol) in THF (0.5 mL) and AcSH
(34 µL, 0.48 mmol) were added with stirring. The reaction was allowed
to stir at 0 °C (1 h) and then at room temperature (16 h). The mixture
was concentrated in vacuo and H2O (20 mL) was added to the residue.
The mixture was extracted with EtOAc (2 × 20 mL) and the combined
organic layers were dried (MgSO4) and concentrated in vacuo.
Purification by PTLC (1:2 EtOAc/hexanes) afforded 22 as a white
solid: yield, 38 mg (44%); mp 156-166 °C; [R]58925 ) -31 (c ) 0.26,
CHCl3); Rf 0.40 (1:2 EtOAc/hexanes); IR (KBr) 3363, 2985, 2360, 1689,
25
[R]589 ) +22 (c ) 0.37, MeOH); IR (neat) 3433, 2900, 1689, 1203
cm-1 1H NMR (CD3OD) δ 3.00 (app dd, J ) 15.0, 6.0 Hz, 2 H,
;
SCHH′), 3.50-3.63 (m, 4 H, SCHH′, CHNH2); 13C NMR (CD3OD) δ
34.1 (SCH2), one signal was not detected and is believed to overlap
with one of the solvent peaks; MS (+CI) m/z 151 [M-2TFA+1]+; Mr
(+CI) 151.035 94 [M-2TFA+1]+ (calcd for C4H11N2S2 151.036 37).
7-N-(1′-Amino-4′,5′-dithian-2′-yl)porfiromycin C(8) Cyclized Im-
ine (6) ([(4aR,9aS,10aS,10bR,11S,12aR)-10b-Methoxy-6,10-dimethyl-
7-oxo-1,4,4a,5,7,9,9a,10,10a,10b,11,12a-dodecahydroazireno[2′,3′:
6,7]pyrrolizino[2,3-f][1,2]dithiino[4,5-b]quinoxalin-11-yl]methyl
Carbamate). To an anhydrous methanolic solution (1 mL) of 14 (5.4
mg, 0.014 mmol) and Et3N (8 µL, 0.06 mmol) was added a methanolic
solution (1 mL) of 25 (4.3 mg, 0.012 mmol). The reaction solution
was stirred at room temperature (1.5 d) and then the solvent was
removed in vacuo. Purification by PTLC (10% MeOH-CHCl3)
afforded 6: yield, 4.0 mg (72%); HPLC tR 29.2 min; Rf 0.57 (10%
MeOH-CHCl3); UV-vis (CH3CN-H2O) λmax 224, 276 (sh), 366 nm;
1H NMR (pyridine-d5, 300 MHz) δ 1.86 (s, C(6)CH3), 2.12 (dd, J )
4.5, 1.5 Hz, C(2)H), 2.29 (s, N(1a)CH3), 2.59 (d, J ) 4.5 Hz, C(1)H),
3.13-3.26 (C(2′)H, C(3′)HH′, C(6′)HH′), 3.28 (s, C(9a)OCH3), 3.43-
3.58 (C(1′)H, C(3′)HH′, C(6′)HH′), 3.62 (br d, J ) 12.0 Hz, C(3)-
HH′), 4.01 (dd, J ) 11.7, 4.5 Hz, C(9)H), 4.38 (d, J ) 12.0 Hz,
C(3)HH′), 5.00 (dd, J ) 11.7, 10.2 Hz, C(10)HH′), 5.71 (dd, J ) 10.2,
4.5 Hz, C(10)HH′), 6.73 (br s, C(7)NH), the signal for the C(10)OC-
(O)NH2 protons was not detected and is believed to be beneath the
solvent peak, the 1H NMR assignments were consistent with the COSY
spectrum; 13C NMR (pyridine-d5, 75 MHz) δ 8.2 (C(6)CH3), 38.7
(C(3′)), 40.0 (C(6′)), 43.1 (N(1a)CH3), 43.5 (C(2)), 47.2 (C(9)), 47.7
(C(1)), 49.2 (C(9a)OCH3), 51.1 (C(3)), 57.0 (C(2′)), 63.2 (C(1′)), 63.3
(C(10)), 106.5 (C(6)), 107.1 (C(9a)), 115.1 (C(8a)), 140.8 (C(7)), 148.0
(C(5a)), 152.8 (C(8)), 158.3 (C(10a)), 179.3 (C(5)); MS (+FAB) m/z
464 [M+1]+; Mr (+FAB) 464.142 35 [M+1]+ (calcd for C20H26N5O4S2,
464.142 62).
1
1527, 1173, 625 cm-1; H NMR (acetone-d6) δ 1.41 (s, 18 H, OC-
(CH3)3), 2.31 (s, 6 H, C(O)CH3), 2.93-3.17 (m, 4 H, SCH2), 3.85-
3.96 (m, 2 H, CHNH), 5.91 (app d, J ) 9 Hz, 2 H, NHCO), the H
1
NMR data were in agreement with the HETCOR spectrum; 13C NMR
(acetone-d6) δ 28.7 (OC(CH3)3), 30.6 (C(O)CH3), 32.5 (SCH2), 54.6
(CHNH), 79.3 (OC(CH3)3), 156.9 (NHCO), 195.5 (COCH3), the 13C
NMR data were in agreement with the HETCOR spectrum; MS (+CI)
m/z 437 [M+1]+; Mr (+CI) 437.177 73 [M+1]+ (calcd for C18H33N2O6S2
437.178 01).
Method B. To a cooled (0 °C) solution of 20 (64 mg, 0.20 mmol),
Et3N (86 µL, 0.60 mmol) and DMAP (58 mg, 0.48 mmol) in CH2Cl2
(6 mL) was slowly added a solution of TsCl (116 mg, 0.6 mmol) in
CH2Cl2 (1 mL) for 1 h. The reaction was allowed to stir at 0 °C (30
min) and then at room temperature (24 h). Dichloromethane (20 mL)
was added and then the reaction mixture was successively washed with
aqueous 0.1 N HCl (30 mL) and aqueous 10% NaHCO3 (30 mL)
solutions. The organic layer was dried (MgSO4) and concentrated in
vacuo to afford 21 as a yellow solid: yield, 100 mg (80%); Rf 0.30
1
(1:2 EtOAc/hexanes); H NMR (CDCl3) δ 1.39 (s, 18 H, OC(CH3)3),
2.45 (s, 6 H, PhCH3), 3.87-4.04 (m, 6 H, TsOCH2, CHNH), 5.11 (br
s, 2 H, NHCO), 7.36 (d, J ) 8.4 Hz, 4 H, Ph (ortho)), 7.76 (d, J ) 8.4
Hz, 4 H, Ph (meta)); 13C NMR (CDCl3) δ 21.6 (PhCH3), 28.2 (OC-
(CH3)3), 50.4 (CHNH), 68.5 (TsOCH2), 80.2 (OC(CH3)3), 128.0 (Ph,
meta), 130.1 (Ph, ortho), 132.3 (Ph, para), 145.3 (Ph, ipso), the signal
for the NHCO resonance was not detected.
To a stirred solution DMF (7 mL) of 21 (100 mg, 0.16 mmol) was
added KSAc (46 mg, 0.40 mmol). After warming to 60 °C, stirring
was continued (3 h) and then the solvent was removed in vacuo. H2O
7-N-(1′-Amino-4′,5′-dithian-2′-yl)mitomycin C C(8) Cyclized
Imine (12) ([(4aR,9aS,10aS,10bR)-10b-Methoxy-6-methyl-7-oxo-1,4,-
4a,5,7,9,9a,10,10a,10b,11,12a-dodecahydroazireno[2′,3′:6,7]pyrrolizino-
[2,3-f][1,2]dithiino[4,5-b]quinoxalin-11-yl]methyl Carbamate). To an
9
4290 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004

Cyclic Disulfide C(8) Iminoporfiromycin
A R T I C L E S
anhydrous methanolic solution (1 mL) of 14 (8.7 mg, 0.023 mmol)
and triethylamine (11 µL, 0.08 mmol) was added a methanolic solution
(1 mL) of 27 (5.4 mg, 0.015 mmol). The reaction solution was stirred
at room temperature (2 d) and then the solvent was removed in vacuo.
Purification by PTLC (10% MeOH-CHCl3) afforded 12: yield, 4.8
mg (71%); HPLC tR 27.6 min; Rf 0.45 (10% MeOH-CHCl3); UV-
vis (CH3CN-H2O) λmax 225, 366 nm; 1H NMR (pyridine-d5, 300 MHz)
δ 1.85 (s, C(6)CH3), 2.71 (dd, J ) 3.9, 1.8 Hz, C(2)H), 3.08-3.25 (m,
C(1)H, C(2′)H, C(3′)HH′, C(6′)HH′), 3.30 (s, C(9a)OCH3), 3.42-3.61
(m, C(1′)H, C(3′)HH′, C(6′)HH′), 3.64 (br d, J ) 11.9 Hz, C(3)HH′),
4.05 (dd, J ) 11.4, 4.2 Hz, C(9)H), 4.46 (d, J ) 11.9 Hz, C(3)HH′),
5.26 (dd, J ) 11.4, 10.2 Hz, C(10)HH′), 5.77 (dd, J ) 10.2, 4.2 Hz,
C(10)HH′), 6.70 (br s, C(7)NH), the signals for the N(1a)H and C(10)-
OC(O)NH2 protons were not detected and are believed to be beneath
the solvent peaks; 13C NMR (pyridine-d5, 75 MHz) δ 8.3 (C(6)CH3),
34.0 (C(2)), 38.1 (C(1)), 38.8 (C(3′) or C(6′)), 47.7 (C(9)), 49.4 (C(9a)-
OCH3), 56.6 (C(2′)), 63.4 (C(10)), 63.5 (C(1′)), 106.9 (C(6)), 107.5
(C(9a)), 114.9 (C(8a)), 140.9 (C(7)), 148.9 (C(5a)), 153.1 (C(8)), 158.6
(C(10a)), 178.8 (C(5)), the signals for C(3) and either C(6′) or C(3′)
were not detected and are believed to overlap with the observed peaks;
MS (+ESI) m/z 450 [M+1]+.
Methanolysis of C(8) Iminoporfiromycin 6 To Give cis-(31) and
trans-(32) C(1) 1-Methoxy (8) Imino-2-methylaminomitosenes.
Compound 6 (4.8 mg, 0.01 mmol) was dissolved in a buffered
methanolic solution (0.1 M bis-Tris‚HCl, “pH” 5.5, 2 mL) and then
stirred at room temperature (21 h). The solvent was removed under
reduced pressure and the residue was purified using PTLC (10%
MeOH-CHCl3) to provide the desired compounds.
Compound 31: yield, 1.7 mg (37%); HPLC tR 31.8 min; Rf 0.26
(10% MeOH-CHCl3); UV-vis (CH3CN-H2O) λmax 255, 319 nm; 1H
NMR (pyridine-d5, 300 MHz) δ 2.00 (s, C(6)CH3), 2.48 (s, NHCH3),
3.16-3.34 (m, C(2′)H, C(3′)HH′, C(6′)HH′), 3.51 (s, C(1)OCH3), 3.52-
3.64 (m, C(1′)H, C(3′)HH′, C(6′)HH′), 3.73-3.82 (m, C(2)H), 4.09
(dd, J ) 11.7, 9.3 Hz, C(3)HH′), 4.92 (dd, J ) 11.7, 7.5 Hz, C(3)-
HH′), 5.01 (d, J ) 5.1 Hz, C(1)H), 5.76 (1/2ABq, J ) 12.9 Hz, C(10)-
HH′), 5.94 (1/2ABq, J ) 12.9 Hz, C(10)HH′), 6.61 (br s, C(7)NH),
the signals for the C(2)NH and C(10)OC(O)NH2 protons were not
detected and are believed to be beneath the solvent peaks; MS (+FAB)
m/z 464 [M+1]+; Mr (+FAB) 464.141 49 [M+1]+ (calcd for
C20H26N5O4S2, 464.142 62).
4.88 (dd, J ) 11.5, 7.5 Hz, C(3)HH′), 5.84 (1/2ABq, J ) 13.2 Hz,
C(10)HH′), 6.11 (1/2ABq, J ) 13.2 Hz, C(10)HH′), 6.57 (br s, C(7)-
NH), the signals for the C(1)H, C(2)NH and C(10)OC(O)NH2 protons
were not detected and are believed to be beneath the solvent peaks;
MS (+FAB) m/z 428 [M+1]+; Mr (+FAB) 428.230 34 [M+1]+ (calcd
for C22H30N5O4, 428.229 78).
Compound 34:35 yield, 0.9 mg (45%); HPLC tR 27.9 min; Rf 0.29
(10% MeOH-CHCl3); UV-vis (CH3CN-H2O) λmax 251, 315 nm; 1H
NMR (pyridine-d5, 300 MHz) δ 1.17-1.48 (m, C(4′)HH′, C(5′)HH′,
C(3′)HH′, C(6′)HH′), 1.52-1.56 (m, C(4′)HH′), 1.62-1.66 (m, C(5′)-
HH′), 1.95-1.99 (m, C(3′)HH′), 2.10 (s, C(6)CH3), 2.42 (s, NHCH3),
2.42-2.45 (m, C(6′)HH′), 2.93 (ddd, J ) 12.3, 12.3, 3.0 Hz, C(2′)H),
3.23-3.30 (m, C(1′)H), 3.55 (s, C(1)OCH3), 3.86 (d, J ) 5.1 Hz,
C(2)H), 4.51 (d, J ) 12.9 Hz, C(3)HH′), 4.63 (dd, J ) 12.9, 5.1 Hz,
C(3)HH′), 5.05 (s, C(1)H), 5.90 (1/2ABq, J ) 12.6 Hz, C(10)HH′),
5.98 (1/2ABq, J ) 12.6 Hz, C(10)HH′), 6.45 (br s, C(7)NH), the signals
for the C(2)NH and C(10)OC(O)NH2 protons were not detected and
are believed to be beneath the solvent peaks; MS (+FAB) m/z 428
[M+1]+; Mr (+FAB) 428.229 56 [M+1]+ (calcd for C22H30N5O4,
428.229 78).
General Procedure for the Solvolysis of Porfiromycins (Kinetic
Studies). To a buffered methanolic solution (0.1 M Tris‚HCl, “pH”
7.4; 0.1 M bis-Tris‚HCl, “pH” 5.5) (final volume 1.5 mL) maintained
at 25 °C containing the porfiromycin (10-60 µL of 4 mM methanolic
solution, final concentration 0.03-0.17 mM) was added a methanolic
solution (5-50 µL) of the nucleophile of choice (stock solution: 4-20
mM, final nucleophile concentration 0.015-3.0 mM). The reaction was
monitored by UV-visible spectroscopy (200-600 nm), and typically
followed for greater than two half-lives. The “pH” of the solution was
determined at the conclusion of the reaction and found to be within (
0.1 pH units of the original solution. The reaction products (31 + 32,
33 + 34) were identified by co-injection with authentic samples using
HPLC and cospotting with authentic samples using TLC. The λmax of
porfiromycin (∼365 nm) was plotted versus time and found to decrease
in a first-order decay (exponential decay) process. The nonlinear
regression analysis to fit the observed exponential decay by SigmaPlot
Program (SigmaPlot, 2001) yielded pseudo-first-order rate constants
(kobs) and half-lives (t1/2). The reactions were done in duplicate and the
results averaged.
General Procedure for Alkaline Agarose Gel Electrophoresis.50
The agarose gels were prepared by adding 1.2 g of agarose to 100 mL
of an aqueous 100 mM NaCl and 2 mM EDTA solution (pH 8.0). The
suspension was heated in a microwave oven until all of the agarose
was dissolved (1 min). The gel was poured and was allowed to cool
and solidify at room temperature (1 h). The gel was soaked in an
aqueous alkaline running buffer solution (50 mL) containing 40 mM
NaOH and 1 mM EDTA (1 h) and then the comb was removed. The
buffer solution was refreshed prior to electrophoresis.
Compound 32: yield, 1.7 mg (37%); HPLC tR 29.7 min; Rf 0.27
(10% MeOH-CHCl3); UV-vis (CH3CN-H2O) λmax 255, 316 nm; 1H
NMR (pyridine-d5, 300 MHz) δ 1.99 (s, C(6)CH3), 2.47 (s, NHCH3),
3.17-3.38 (m, C(2′)H, C(3′)HH′, C(6′)HH′), 3.54 (s, C(1)OCH3), 3.55-
3.78 (m, C(1′)H, C(3′)HH′, C(6′)HH′), 3.92 (d, J ) 5.1 Hz, C(2)H),
4.53 (d, J ) 12.6 Hz, C(3)HH′), 4.64 (dd, J ) 12.6, 5.1 Hz, C(3)HH′),
5.11 (s, C(1)H), 5.75 (1/2ABq, J ) 12.9 Hz, C(10)HH′), 5.88 (1/2ABq,
J ) 12.9 Hz, C(10)HH′), 6.51 (br s, C(7)NH), the signals for the C(2)-
NH and C(10)OC(O)NH2 protons were not detected and are believed
to be beneath the solvent peaks; MS (+FAB) m/z 464 [M+1]+; Mr
(+FAB) 464.142 47 [M+1]+ (calcd for C20H26N5O4S2, 464.142 62).
Methanolysis of C(8) Iminoporfiromycin 13 To Give cis-(33) and
trans-(34) 1-Methoxy C(8) Imino-2-methylaminomitosenes.35 Com-
pound 13 (2.0 mg, 0.004 mmol) was dissolved in a buffered methanolic
solution (0.1 M bis-Tris‚HCl, “pH” 5.5, 1 mL) and then stirred at room
temperature (18 h). The solvent was removed under reduced pressure
and the residue was purified using PTLC (10% MeOH-CHCl3) to
provide the desired compounds.
Compound 33:35 yield, 0.6 mg (30%); HPLC tR 29.0 min; Rf 0.26
(10% MeOH-CHCl3); UV-vis (CH3CN-H2O) λmax 251, 315 nm; 1H
NMR (pyridine-d5, 300 MHz) δ 1.19-1.32 (m, C(4′)HH′, C(5′)HH′),
1.40-1.48 (m, C(3′)HH′, C(6′)HH′), 1.54-1.58 (m, C(4′)HH′), 1.64-
1.68 (m, C(5′)HH′), 1.96-2.02 (m, C(3′)HH′), 2.12 (s, C(6)CH3), 2.38
(s, NHCH3), 2.45-2.50 (m, C(6′)HH′), 2.93 (ddd, J ) 11.7, 11.7, 2.7
Hz, C(2′)H), 3.28 (ddd, J ) 11.7, 11.7, 2.7 Hz, C(1′)H), 3.51 (s, C(1)-
OCH3), 3.58-3.62 (m, C(2)H), 4.01 (dd, J ) 11.5, 9.0 Hz, C(3)HH′),
To an aqueous solution of ∼85 µL of H2O (sterile) and 2.5 µL of 1
M Tris‚HCl (pH 7.4) was added a solution of linearized pBR322 (5
µL, 5 µg) in 10 mM Tris solution containing 1 mM EDTA (pH 8.0).
After deaeration with Ar (15 min), porfiromycin (1-5 µL of 1-4 mM
DMSO solution, final concentration 0.01-0.2 mM) and a nucleophile
of choice (1-5 µL of 1-20 mM DMSO solution, final concentration
0.01-1.0 mM) were added and the resulting solution (final volume
100 µL) was incubated at room temperature (2 h). The solution was
washed with 1:1 PhOH/CHCl3 (100 µL) and CHCl3 (2 × 100 µL), and
precipitated (12.1 µL of 3 M NaOAc and 250 µL of EtOH, -70 °C
(10 min)). The mixture was centrifuged at 0 °C (15 min), and the EtOH
was decanted off and evaporated in vacuo. The remaining DNA was
dissolved in 25 µL of 10 mM Tris solution containing 1 mM EDTA
(pH 8.0). Agarose loading dye (5 µL) was added to the sample (5 µL)
and the samples were loaded onto the wells. The gel was run at 75
mA/25 V (30 min) and then at 145 mA/38 V (3-4 h). The gel was
then neutralized for 45 min in an aqueous 100 mM Tris pH 7.0 buffer
9
J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004 4291

A R T I C L E S
Lee and Kohn
solution containing 150 mM NaCl, which was refreshed every 15 min.
The gel was stained with an aqueous 100 mM Tris pH 7.5 buffer
solution (100 mL) containing ethidium bromide (20 µL of an aqueous
ethidium bromide stock solution (10 mg/10 mL)) and 150 mM NaCl
for 20 min. The background staining was then removed by soaking
the gel in an aqueous 50 mM NH4OAc and 10 mM â-mercaptoethanol
solution (3 h). The gel was then analyzed by two methods. In one
method, the gel was visualized by UV and photographed using Polaroid
film 667. In the second method, the gel was analyzed with a Storm
860 phosphorimager operating in the blue fluorescence mode and
ImageQuant 5.0 software (Molecular Dynamics, Sunnyvale, CA).
General Procedure for Antiproliferative Activity Test.60 In vitro
antiproliferative tests were conducted using human tumor cell line A549
(lung adenocarcinoma) by Dr. Hitoshi Arai (Kyowa Hakko Kogyo Co.,
Shizuoka, Japan). The cells (2 × 103 cells/well) were precultured at
37 °C (24 h) in 96-well microtiterplates containing the culture medium
(RPMI-1640 supplemented with 10% (v/v) fetal bovine serum, 100
U/mL of penicillin, and 100 µg/mL of streptomycin) under either
aerobic (5% of CO2 and 95% of air) or hypoxic (5% of CO2 and <2%
of O2) conditions. The cells were then treated with the drug candidates
(1 h), washed twice with the medium and further incubated (71 h) in
the drug-free medium. The antiproliferative activity of drugs against
tumor cells was measured by MTT assay. Cell growth (%) was
calculated by the equation, {[A - Ao]/[Ac - Ao]} × 100 (A: absorbance,
Ao: blank absorbance, Ac: control absorbance), and the activity was
expressed by IC50 values (concentration required for 50% inhibition).
Acknowledgment. The authors gratefully acknowledge the
NIH (CA29756) for support of these studies. We thank Dr. Junji
Kanazawa and Ms. Yoshino Yamada (Kyowa Hakko Kogyo
Co., Shizuoka, Japan) for conducting the in vitro antiproliferative
tests and Drs. Masaji Kasai and Hitoshi Arai (Kyowa Hakko
Co., Ltd., Shizuoka, Japan) for generously supplying mitomycins
A and C.
Supporting Information Available: General methods and
synthetic procedures for the preparation of 20, supporting figure
for DNA ISC experiments for 6, 7, and 13 with Et3P and GSH
(PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
JA030577R
9
4292 J. AM. CHEM. SOC. VOL. 126, NO. 13, 2004