Abstract
The objective of the research described in the first part of this thesis involves the
application of carbon monoxide and transition metals in key steps of a synthetic route
to lavendamycin, an antic cancer compound, and its analogues. Lavendamycin is a
pentacyclic compound that possesses a quinoline-5,8-quinone AB ring linked to a b-
carboline CED ring. The development of general routes to the synthetic equivalents
of the lavendamycin AB quinoline system together with a linker atom, quinoline -2-
carboxaldehydes, as well as to the lavendamycin DE indole ring system, namely
tryptophan derivatives, was addressed.
The Pictet-Spengler cyclisation approach towards lavendamycin involves the reaction
between quinoline-2-carboxaldehyde and tryptophan methyl ester to furnish the
pentacyclic precursor of the methyl ester of lavendamycin. This synthetic approach
requires the availability of quinoline-2-carboxaldehydes, previously prepared by the
oxidation of 2-methylquinolines with toxic selenium dioxide. A general strategy
towards the synthesis of the AB ring moiety utilising a pre-formed ring system such
as commercially available 8-hydroxyquinoline has been successfully developed. It
involved the high pressure palladium catalysed formylation of 2-bromo or other
suitable 2-substituted quinoline derivatives under syngas (1:1 CO:H2). The
preparation of the required 2-substituted quinoline derivative involved the
methylation of the 8-hydroxylgroup followed by N-oxidation and then a
rearrangement step.
In both the Pictet-Spengler and Bischler-Napieralski synthetic approaches to
lavendamycin, the CDE ring moiety is introduced using tryptophan methyl ester as
building block. The application of this approach to the synthesis of lavendamycin
analogues with a substituted D-ring required the availability of substituted tryptophan
methyl esters. A general strategy towards the tryptophan derivatives starting with a
Wittig reaction between a suitable 2-nitrobenzaldehyde precursor and 1,3-dioxolan-2-
yl-methyltriphenylphosphonium bromide, followed by a two-stage, one -pot rhodium
catalysed hydroformylation/reduction reaction, has been successfully developed. This
methodology yielded ten different possible tryptophan precursors in moderate to good
yields.
The second part of the research described in this thesis included the identification of
factors effecting the rate and regioselectivity of palladium catalysed
methoxycarbonylation of a-olefins. The results showed that fast reactions under
polar conditions give mainly linear esters. However, reactions under less polar
conditions are slower, yielding mainly branched esters. Detailed analysis of the
results suggest the operation of a so-called “cationic” mechanism (involving cationic
palladium intermediates) in the formation of mainly linear esters, but the operation of
a so-called “neutral” mechanism (involving neutral palladium intermediates) in the
formation of mainly branched esters. The nature of the phosphine ligands was found
to play a significant, but secondary role in determining regioselectivity of
methoxycarbonylation.
Another objective was the optimisation of the palladium catalysed hydroformylation
of a-olefins. An evaluation of the efficiency of the palladium catalysed
hydroformylation process required a comparison with the hydroformylation processes
based on cobalt and rhodium. Variation of ligands (diphosphines of the type
R2P(CH2)nPR2), solvents, acids, etc. had a dramatic effect on the products and the rate
of the reaction. In the presence of trifluoroacetic acid 1-pentene is converted to C-6 aldehydes, while in the presence of trifluoromethanesulfonic acid 1-pentene is
converted to C-11 ketones. Corresponding results were obtained with 1-octene as
substrate. The palladium catalysts were found to also effect isomerisation of the a-
olefin into internal olefins, but isomerisation was not a rate limiting process with
respect to the hydroformylation reaction. Palladium catalysed isomerisation reactions
occurred at a slower rate than the corresponding cobalt catalysed isomerisation
process. However, with rhodium no isomerisation occurred.
The comparison between cobalt, rhodium and palladium showed that rhodium is the
best catalyst for the hydroformylation of a-olefins. The pressures and temperatures
required for this process are much lower than that required for palladium and cobalt.
The ligand used is triphenylphosphine, which is relatively inexpensive and non-toxic,in contrast with the more expensive ligands required for the cobalt and palladium
hydroformylation processes.
The use of palladium opens up the unique possibility of converting a-olefins into
“dimeric” ketones, which show promise as precursors for the new class of geminidetergents.
Prof. C.W. Holzapfel