Studies on the synthesis of landomycin A. Synthesis of the originally assigned structure of the aglycone, landomycinone, and revision of structure

Article abstract of DOI:10.1021/jo049426c

The originally proposed structure (2) of landomycinone, the aglycone of landomycin A, has been synthesized and shown to be nonidentical to the naturally derived landomycin A aglycone. The synthesis of 2 features the Doetz benzannulation reaction of chromium carbene 5 and alkyne 6, and the intramolecular Michael-type cyclization reaction of the phenolic naphthoquinone 20. It is proposed that natural landomycinone possesses the alternative structure 3, but attempts to access this structure via the Michael-type cyclization of the isomeric phenolic naphthoquinone 38 have been unsuccessful.

Full text of DOI:10.1021/jo049426c

Stu d ies on th e Syn th esis of La n d om ycin A. Syn th esis of th e
Or igin a lly Assign ed Str u ctu r e of th e Aglycon e, La n d om ycin on e,
a n d Revision of Str u ctu r e
William R. Roush* and R. J effrey Neitz
Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109
roush@umich.edu
Received April 6, 2004
The originally proposed structure (2) of landomycinone, the aglycone of landomycin A, has been
synthesized and shown to be nonidentical to the naturally derived landomycin A aglycone. The
synthesis of 2 features the Do¨tz benzannulation reaction of chromium carbene 5 and alkyne 6, and
the intramolecular Michael-type cyclization reaction of the phenolic naphthoquinone 20. It is
proposed that natural landomycinone possesses the alternative structure 3, but attempts to access
this structure via the Michael-type cyclization of the isomeric phenolic naphthoquinone 38 have
been unsuccessful.
In tr od u ction
Landomycin A1 is a member of the angucycline anti-
biotic family, a group that now numbers over 100
members.2,3 Landomycin A in particular has been studied
as a potential antitumor agent.1,4,5 Although the mode
of action of landomycin A has not been established
unequivocally, it is known that the natural product
interacts with DNA5 and inhibits DNA synthesis and
G1/S cell cycle progression.6
Landomycin A possesses a broad spectrum of activity
against the National Cancer Institute’s panel of 60
cancerous cell lines.5,7 At 1 µM, landomycin A is more
active than bleomycin, vinblastine, and paclitaxel in
inhibition of tumor colony formation in single cell sus-
pensions of freshly obtained human tumor cells.5 Al-
though the mechanism of action has not been determined
at the molecular level, landomycin A is known to inhibit
DNA synthesis by stopping cell cycle progression from
the G0/G1 (resting) phase to S (synthesis) phase.6 We have
speculated that the hexasaccharide chain of landomycin
A could serve as a DNA binding agent, similar to that of
the oligosaccharide portions of the aureolic acids and
calicheamycin, which are known to be critical for DNA
binding and recognition.8-11 Supporting this hypothesis
F IGURE 1. Structures of landomycin A (1) and landomyci-
none (2) proposed by Rohr and the revised structure of
landomycinone (3).
is knowledge that the cytostatic activities of other
members of the landomycin family (e.g., landomycins
B-E) depend on the length of the oligosaccharide chain.4
The initial structural assignment of landomycin A was
made by Rohr in 19907 and was subsequently revised to
the currently accepted structure, 1.1 The gross connectiv-
ity assignments were made primarily through the use of
high-field 2D-NMR analysis, and the absolute configu-
ration of the secondary alcohol was assigned by applica-
tion of Nakanishi’s exciton chirality method.12
(1) Weber, S.; Zolke, C.; Rohr, J .; Beale, J . M. J . Org. Chem. 1994,
59, 4211.
(2) Rohr, J .; Thiericke, R. Nat. Prod. Rep. 1992, 9, 103.
(3) Krohn, K.; Rohr, J . Top. Curr. Chem. 1997, 188, 127.
(4) Rohr, J .; Wohlert, S.-E.; Oelkers, C.; Kirschning, A.; Ries, M.
Chem. Commun. 1997, 973.
Despite their interesting biological properties and
structural relationship to other members of the angucy-
(5) Depenbrock, H.; Bornschlegl, S.; Peter, R.; Rohr, J .; Schmid, P.;
Schweighart, P.; Block, T.; Rastetter, J .; Hanauske, A.-R. Ann. He-
matol. 1996, 73, A80.
(6) Crow, R. T.; Rosenbaum, B.; Smith, R., III; Guo, Y.; Ramos, K.
S.; Sulikowski, G. A. Bioorg. Med. Chem. Lett. 1999, 9, 1663.
(7) Henkel, T.; Rohr, J .; Beale, J . M.; Schwenen, L. J . Antibiot. 1990,
43, 492.
(8) Remers, W. A.; Iyengar, B. S. In Cancer Chemotherapeutic
Agents; Foye, W. O., Ed.; American Chemical Society: Washington,
DC, 1995; pp 578, and references therein.
(9) Kahne, D.; Silva, D.; Walker, S. In Bioorganic Chemistry:
Carbohydrates; Hecht, S. M., Ed.; Oxford University Press: New York,
1999; p 174.
(10) Silva, D. J .; Kraml, C. M.; Kahne, D. Bioorg. Med. Chem. 1994,
2, 1251.
(11) Bifulco, G.; Galeone, A.; Nicolaou, K. C.; Chazin, W. J .; Gomez-
Paloma, L. J . Am. Chem. Soc. 1998, 120, 7183 and references therein.
(12) Harada, N.; Chen, S.-M. L.; Nakanishi, K. J . Am. Chem. Soc.
1975, 97, 5345.
10.1021/jo049426c CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/19/2004
4906
J . Org. Chem. 2004, 69, 4906-4912

Revised Structure of Landomycinone
SCHEME 1
F IGURE 2. Retrosynthetic analysis of 2.
cline group,2,3 relatively little work has appeared on the
synthesis of the landomycins. Syntheses of this hexa-
saccharide fragment have been completed by the
Sulikowski, Yu, and Roush groups,13-15 but thus far no
reports on the synthesis of the landomycine aglycone,
landomycinone, have appeared. A flexible total synthesis
of landomycin A would allow for the synthesis of ana-
logues that could be key in elucidating the details of both
the possible DNA binding properties and global biological
activity.
We report herein an enantioselective synthesis of the
originally assigned structure (2) of the landomycine
aglycone, landomycinone, and revise the structure of the
aglycone to the structure depicted as 3.
Resu lts a n d Discu ssion
halogen exchange (t-BuLi, THF, -78 °C), and the result-
ing aryllithium species was then treated with (R)-glycidyl
ether 8.21 This reaction provided 12 in 64-75% yield.
Protection of the secondary alcohol of 12 as a TBS ether
and then removal of the PMB ether upon treatment with
DDQ22 provided 13 in excellent yield. Swern oxidation23
of 13 provided the corresponding aldehyde, which was
then elaborated to acetylene 14 by application of the
Corey-Fuchs protocol.24 Finally, selective removal25 of
the phenolic TBS ether by using TBAF (1 equiv) in the
presence of acetic acid and then acylation of the phenol
provided the targeted alkyne 6.
Chromium carbene 5 was prepared from 1,4-dihydro-
quinone bismethoxymethyl ether (15) under standard
conditions (Scheme 2).19 Ortho lithiation of 15 by treat-
ment with t-BuLi in THF at -78 °C gave the aryllithium
species that was then treated with solid Cr(CO)6. The
resulting hydroxy carbene was then treated with methyl
triflate, which provided carbene 5 in 75% overall yield.
The benzannulation reaction of 5 and 6 gave best results
when performed in heptane at 55 °C. Under these
conditions, the intermediate hydroquinone mono ether
(devoid of the arene Co(CO)3 unit) was obtained in 35-
40% yield, along with varying amounts of recovered 5.
However, alkyne 6 was completely consumed, and it was
not possible to identify any of the other products of this
reaction. Numerous attempts to improve the efficiency
of this reaction by screening other benzannulation
Syn th esis of Or igin a lly Assign ed La n d om ycin
Aglycon e 2. Our strategy for the synthesis of the original
landomycin aglyone structure 2 called for the C(12a)-
C(12b) bond to be constructed by an intramolecular
Michael-type addition16-18 of the phenolate derived from
4 onto C(12a) of the naphthoquinone residue. We antici-
pated that intermediate 4, in turn, could be prepared by
the Do¨tz reaction of chromium carbene 5 and acetylene
6.19 The synthesis of acetylene 6 involved the coupling
of protected bromophenol 7 and glycidol ether 8 as the
key step (Scheme 1). Aryl bromide 7 was synthesized
following the procedure of Brittain.20 Thus, two-stage
bromination of m-cresol in acetic acid provided tetrabro-
mocyclohexadienone 10 in 52% yield. Exposure of this
material to concentrated sulfuric acid induced a 1,2-
migration of a bromine substituent, thereby providing
tetrabromophenol 11. Reductive removal of the o- and
p-Br substituents by treatment of 11 with aqueous HI
at reflux and then protection of the resulting phenol as
a TBS ether provided 7a contaminated with 5-20% (from
various runs) of the corresponding aryl iodide 7b. The
yield of the 7a /7b mixture was nearly quantitative.
Because the mixture of 7a and 7b could not be separated
conveniently, the mixture was subjected to lithium-
(13) Guo, Y.; Sulikowski, G. A. J . Am. Chem. Soc. 1998, 120, 1392.
(14) Yu, B.; Wang, P. Org. Lett. 2002, 4, 1919.
(15) Roush, W. R.; Bennett, C. E. J . Am. Chem. Soc. 2000, 122, 6124.
(16) Brown, P. M.; Thomson, R. H. J . Chem. Soc., Perkin Trans. 1
1976, 9, 997-1000.
(17) Parker, K. A.; Ding, Q. Tetrahedron 2000, 56, 10249.
(18) Krohn, K.; Bo¨ker, N. J . Prakt. Chem. 1997, 339, 114.
(19) Do¨tz, K. H.; Tomuschat, P. Chem. Soc. Rev. 1999, 28, 187 and
references therein.
(21) Hansen, R. M. Chem. Rev. 1991, 91, 437.
(22) Horita, K.; Yoshioka, T.; Tanaka, T.; Oikawa, Y.; Yonemitsu,
O. Tetrahedron 1986, 42, 3021.
(23) Tidwell, T. T. Org. React. 1990, 39, 297.
(20) Brittain, J . M.; De la Mare, P. B. D.; Newman, P. A. J . Chem
Soc., Perkin Trans. 2 1981, 32.
(24) Corey, E. J .; Fuchs, P. L. Tetrahedron Lett. 1972, 3769.
(25) Nelson, T. D.; Crouch, R. D. Synthesis 1996, 1031.
J . Org. Chem, Vol. 69, No. 15, 2004 4907

Roush and Neitz
SCHEME 2
SCHEME 4
SCHEME 3
was 18 deriving from elimination of the C(6)-OTBS unit.
However, this elimination reaction was effectively sup-
pressed in a small-scale pilot reaction when 16 was
treated with NaOEt in EtOH at 23 °C with warming to
65 °C, which provided the targeted (model) tetracyclic
product 17 in 60% yield. Interesting, no deuterium was
incorporated when the cyclization of 16 was performed
with NaOEt in EtOD, indicating that racemization of the
potentially acidic C(6)-H will not occur under these
conditions.
On the basis of these results, the decision was made
to perform the deacylation and cyclization of 4 in a one-
pot sequence. Treatment of 4 with NaOEt in EtOH
indeed provided the protected landomycin aglycone 20
in yields up to 60%, but with considerable variability in
the yield from run to run. Further examination of the
reaction (Scheme 4), by using the purified phenol 19
(generated in 87% yield by treatment of 4 with NaOEt
in EtOH at 0 °C) as substrate, revealed that under
several sets of conditions, only decomposition of 19 was
observed, without any of 20 being obtained. In other
instances, a nearly 1:1 mixture of 19 and 20 was obtained
(again in variable yield).
Consideration of plausible mechanisms for this Michael-
type cyclization indicates that the initial Michael adduct
23 undergoes a series of enolization reactions leading to
26. An oxidant is required for the final conversion of 26
to naphthoquinone 20. Because it is known that quinones
can oxidize hydroquinones in alkaline solution,30 it
seemed possible that 19 (derived from deacylation of 4)
was serving as the oxidant in this case. Indeed, on several
occasions small amounts of dihydronaphthoquinone 21
were obtained from these reactions. Accordingly, efforts
were made to identify an alternative, stoichiometric
oxidant for use in the cyclization of 19 to 20. Oxidants
such as DDQ, benzoquinone, HgO, or Ag2O were em-
conditions26-29 using 5 (and related carbenes with dif-
ferent phenol protecting groups as substrates) were not
successful. Oxidation of the intermediate hydroquinone
mono ether with ceric ammonium nitrate (CAN) then
provided the targeted naphthoquinone 4.
Our plan at the outset was to effect cyclization of
phenol deriving from 4 under mildly basic conditions.
Initial studies (Scheme 3) were performed by using the
phenolic benzoquinone 16 as substrate (deriving from a
Do¨tz benzannulation sequence analogous to the one
described above). No reaction was observed when 16 was
treated with NaHCO3 in MeOH (23 °C), KHMDS in THF
(-78 to 23 °C), or K2CO3 in acetone (23 to 50 °C), and
only products of decomposition were obtained when 16
was treated with Et3N in CH2Cl2 at ambient tempera-
ture. When 16 was treated with K2CO3 in MeOH at 65
°C, cyclization was observed, but the only product isolated
(26) Chan, K. S.; Peterson, G. A.; Brandvold, T. A.; Faron, K. L.;
Challener, C. A.; Hyldahl, C.; Wulff, W. D. J . Organomet. Chem. 1987,
334, 9.
(27) Anderson, B. A.; Wulff, W. D.; Rheingold, A. L. J . Am. Chem.
Soc. 1990, 112, 8615.
(28) Bos, M. E.; Wulff, W. D.; Miller, R. A.; Chamberlain, S.;
Brandvold, T. A. J . Am. Chem. Soc. 1991, 113, 9293.
(29) Yamashita, A.; Toy, A. Tetrahedron Lett. 1986, 27, 3471.
(30) Campbell, T. W. J . Am. Chem. Soc. 1951, 73, 4190.
4908 J . Org. Chem., Vol. 69, No. 15, 2004

Revised Structure of Landomycinone
SCHEME 5
ployed in various reactions of 19 with stoichiometric
NaOEt, but no more than trace amounts of 20 was
observed under these conditions. Use of O2 as the oxidant
similarly gave poor results (only products of decomposi-
tion) when stoichiometric NaOEt was used, but 30% of
20 (together with 30% of recovered 19) was obtained from
an experiment in which 0.1 equiv of NaOEt was used
under a steady stream of O2. Ultimately, a reproducible
set of reaction conditions was developed in which 19 was
treated with 0.1 equiv of NaOEt in EtOH (0.05 M) under
air, from which the targeted naphthoquinone 20 was
obtained in 45% yield along with 40% of recovered 19,
which could be resubjected to the cyclization conditions
to obtain additional 20.
F IGURE 3. X-ray crystal structure of synthetic 2.
SCHEME 6
With 20 in hand, we initiated studies on the deprotec-
tion sequence that we anticipated would provide syn-
thetic landomycinone, 2 (see Scheme 6). Treatment of 20
with either HCl in ether or MgBr2 in CH2Cl2 smoothly
removed the two phenolic MOM ethers and provided 27
in excellent yield. However, attempted removal of the
C(6)-TBS ether from this intermediate using a variety
of fluoride sources failed to provide any of 2; as a general
rule, 27 was recovered from these attempted deprotec-
tions. Similarly, attempts to remove the TBS ether from
20 by using a variety of fluoride sources were also
unsuccessful. However, successful cleavage of the TBS
ether was accomplished by treatment of 20 with 0.05 M
HCl in MeOH, which provided a ca. 1:1 mixture of 28
and 29, in which one of the two MOM ethers had been
removed,31 in 85% yield. It is known that landomycinone
undergoes dehydration within 1 h when treated with 1
M HCl at 23 °C.7 It was not surprising, therefore, that a
small amount (ca. 5%) of the elimination product 30 was
also obtained from the deprotection of 20 in 0.05 M HCl.
Finally, treatment of 28 and 29 with MgBr2 (prepared
according to Vedejs’ procedure)32 in THF at 0 °C provided
synthetic “landomycinone” 2 in 80% yield. A single crystal
X-ray structure analysis of the fully deprotected synthetic
“landomycinone” verified our structure assignment of the
synthetic material as 2. However, comparison of NMR
data for synthetic 2 with published data for the natural
1
landomycinone and with copies of the actual H NMR
spectra provided by Prof. Rohr, indicated that natural
landomycinone and synthetic 2 are not identical. Among
the more striking differences are the chemical shift data
for H(2), H(4), H(5ax), H(5eq), H(9), and H(10).1
(31) The assignment of 29 as the mono-MOM ether deriving from
the deprotection of 19 is tentative.
(32) Vedejs, E.; Daugulis, O. J . Org. Chem. 1996, 61, 5702.
J . Org. Chem, Vol. 69, No. 15, 2004 4909

Roush and Neitz
F IGURE 5. Minimizations of potential atropisomers of 2.
F IGURE 4. Comparison of spectroscopic data for synthetic 2
and natural landomycinone.
(Figure 5). Atropisomerism in fused ring systems is
known in the literature (e.g., colchicine).36 Monte Carlo
minimizations of 2 were conducted from a number of
different starting points, all of which minimized to
conformer 2a , which has a positive (P ) twist about the
biaryl bond and an axial orientation for the secondary
alcohol.37 The conformer closest in energy with a minus
(M) twist about the biaryl axis is 0.8 kcal/mol higher in
energy. This conformer, 2b, places the secondary alcohol
in the equatorial plane. The computational exclusion of
the low energy conformational diastereomer 2b with an
equatorial alcohol is consistent with vicinal coupling
constants observed in the 1H NMR spectra for both
synthetic 2 and natural landomycinonesboth of which
indicate that the secondary alcohol is axial. None of the
low energy conformations detected by the Monte Carlo
search had both an M configuration about the biaryl bond
and an axial alcohol. When the alcohol was constrained
to adopt such an axial orientation and the dihedral angle
of the biaryl linkage was constrained to avoid computa-
tional relaxation to the P atropisomer, the computed
energy of 2c with M atropisomer stereochemistry is 7.8
kcal/mol higher than that of 2a . 1H NMR spectra of
synthetic 2 were measured at temperatures ranging from
-20 to 80 °C in order to probe the possibility that other
conformations might be accessible, but no significant
changes in the NMR spectra were observed.
These data indicate that synthetic 2 and natural
landomycinone are constitutional and not conformational
isomers. Because 1H NMR data obtained for synthetic
anhydrolandomycinone (31), small quantities of which
were obtained during the deprotection of 19, were in
excellent agreement with data previously reported for
naturally derived anhydrolandomycinone, we propose
that the revised structure for natural landomycinone
must be 3, and therefore also that landomycin A must
be 32. Attempts have been made to synthesize 3 by
appropriate modifications of our synthesis of 2 (Scheme
The poor solubility of synthetic 2 precluded the collec-
tion of a complete set of 13C NMR data, however the
partial 13C spectrum that we obtained was sufficient to
confirm lack of identity of the synthetic and natural
materials. Synthetic 2 exhibited resonances at δ 61.1 and
36.6 for C(5) and C(6), respectively, differing from the
13C chemical shifts for the previously assigned to C(5)
and 33C(6) the natural aglycone (62.0 and 36.3, respec-
tively).1
Analysis of the infrared and ultraviolet spectra of the
natural and synthetic material provided further clues as
to the structural differences (Figure 4). The carbonyl
stretch of the synthetic material at 1623 cm-1 is com-
pletely consistent with the proposed structure of a doubly
hydrogen-bound quinone system, but the 1705 absorption
recorded for the natural aglycone1 is high for even a
parent quinone.33 These data further support the mis-
assignment of the structure of natural landomycinone.
We considered the possibility that the synthetic (2) and
natural landomycinones might be tautomers, with the
quinoide ketones in the A-ring and quinols on the B-ring.
The synthetic material was assigned the tautomeric
structure assigned for 2 on the basis of bond lengths
determined by the X-ray analysis of the synthetic mate-
rial. The 1.35 and 1.25 Å bond lengths observed for the
C-O and CdO bonds, respectively, of synthetic 2 are
consistent with standard values for phenol C-O and
quinone CdO bond lengths of 1.355 and 1.208 Å.34 It is
unlikely (but not impossible)33 that the natural material
is the other tautomer because of the known propensity
of vicinally disubstituted naphthazarins to exist in the
tautomer depicted in structure 2.35
To exclude the possibility that synthetic 2 may be a
conformational isomer of natural landomycinone, a series
of VT-NMR and computational studies were undertaken
(33) Glazunov, V. P.; Tchizhova, A. Y.; Pokhilo, N. D.; Anufriev, V.
P.; Elyakov, G. B. Tetrahedron 2002, 58, 1751.
(34) These bond length data were obtained from Spartan.
(35) Moore, R. E.; Scheuer, P. J . J . Org. Chem. 1966, 31, 3272.
(36) Wildman, W. C.; Pursey, B. A. In The Alkaloids; Manske, R.
H. F., Ed.; Academic Press: New York, 1968; Vol. XI, p 407.
(37) All energies were determined by AM1 calculations (Spartan)
after Monte Carlo minimization of the appropriate starting structures.
4910 J . Org. Chem., Vol. 69, No. 15, 2004

Revised Structure of Landomycinone
SCHEME 7
7). Thus, treatment of aryl bromide 7 with t-BuLi in THF
at -78 °C followed by addition of DMF provided aldehyde
33 in 95% yield following aqueous workup. 1,3-Dilithio-
propyne (34), generated by treatment of propargyl bro-
mide with 2 equiv of n-BuLi, was added to a solution of
33 to provide racemic propargyl alcohol 35.38 This sec-
ondary alcohol was then protected as the TBS ether. The
phenolic TBS ether was then cleaved by treatment of the
bis-silylated derivative with TBAF, and the resulting
phenol was acylated to give 36. The Do¨tz reaction of
alkyne 36 and chromium carbene 5 gave hydroquinone
product 37 in 38% yield.19 This reaction proceeded best
in THF, in contrast to heptane, which was the best
solvent for the Do¨tz reaction of 5 and 6 discussed
previously. Oxidation of hydroquinone 37 to the quinone
followed by deprotection of the pendent phenol provided
38 (70% yield), the substrate for the Michael closure.
Unfortunately, when phenolic quinone 38 was subjected
to the conditions previously used in our synthesis of 19
(vide supra), none of the targeted tetracyclic produce 39
was obtained. Only the elimination product 30 and
recovered phenol 38 were obtained. Extensive modifica-
tion of the reaction temperature, concentration, and base
were studied, but without success. In all cases the only
product isolated was 30 deriving from cyclization and
elimination of the secondary alcohol.
Su m m a r y
F IGURE 6. Comparison of spectroscopic properties of syn-
thetic and natural anhydrolandomycinone (31).
The originally proposed structure (2) of landomycinone,
the aglycone of landomycin A, has been synthesized and
shown to be nonidentical to the naturally derived lando-
mycin A aglycone. Our synthesis of 2 features the Do¨tz
benzannulation reaction of chromium carbene 5 and
alkyne 6 and the intramolecular Michael-type cyclization
reaction of the phenolic naphthoquinone 19. It is pro-
posed that natural landomycinone possesses the alterna-
tive structure 3, but attempts to access this structure via
the Michael-type cyclization of the isomeric phenolic
naphthoquinone 38 have been unsuccessful. Therefore,
an alternative strategy for synthesis of 3 must be
developed in order to verify the structural reassignment
of landomycinone and therefore of landomycin A itself.
F IGURE 7. Revised structures of landomycin A (32) and
landomycinone (3).
(38) Cabezas, J . A.; Pereira, A. R.; Amey, A. Tetrahedron Lett. 2001,
42, 6819.
J . Org. Chem, Vol. 69, No. 15, 2004 4911

Roush and Neitz
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
tocols and characterization data for selected new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Ack n ow led gm en t. We gratefully acknowledge the
National Institutes of Health (GM38907) for gener-
ous support of this research and thank Prof. Rohr
for providing spectroscopic data for natural landomyci-
none.
J O049426C
4912 J . Org. Chem., Vol. 69, No. 15, 2004