in this case does not substitute for the OH group and if indeed the unusually high activity of 83c is due to hydrogen bonding interactions within the colchicine binding site o f tubulin, then the fluorine atom o f 83g is unable to undergo the same interactions.
In the same way the 3-bromo-4-methoxy chalcone, 83h, was much less cytotoxic - 10 times less active than 83g - as would be expected.
The 3,4-dimethoxy compound 83d had around the same cytotoxicity as 83b and 83g which suggests that the methoxy group on the 3 position was also unable to interact in the same way as the hydroxy group in that position but at the same time it was not enough o f a steric hindrance to reduce the activity. The same can be said for the 2-fluoro-4-methoxy and 3,4- methylenedioxy B rings, 83f and 83a. In the case o f the 2,3,4-trimethoxy chalcone 83e, however, there must be a steric hindrance to tubulin binding because this was the only compound in this series to have a cytotoxicity greater than 10 pM. Likewise the inactivity of 83h is probably due to steric factors.
3.2.2 a-Trifluoromethyl Chalcones
The next series o f chalcones to be examined were those with a CF3 group in the alpha position 88. Synthesis o f this series proved less straightforward than that o f previous chalcones. In order to make them by the Knoevenagel reaction that had been used before, the following ketone 89 would have to be made {Scheme 3.8).
[CF3] A r ^ O
89 88
Scheme 3.8 - Knoevenagel synthesis o f a-trifluoromethyl chalcones.
This would require a CF3+ synthon, all sources o f which are unstable and expensive (as discussed in chapter one).
Another option was to add the trifluoromethyl group to the a-bromochalcones 90 via a Pd catalysed coupling reaction {Scheme 3.9). There is some precedent in the literature for this146
but the reagent used (methyl difluorofluorosulfonylacetate acid) is very expensive and has to be used in excess. O c 0 F .s 0 P d(pph3)4 90 88
Scheme 3.9 Synthesis o f a-trifluoromethyl chalcones via P d cross-coupling.
This highlights the problems associated with incorporation o f a trifluoromethyl group into an organic molecule, as seen in chapter one.
The method eventually employed in the synthesis o f 88 was via the diketone 94. This was made from the dimethylhydrazone 92 o f 3,4,5-trimethoxybenzaldehyde, by reaction with trifluoroacetic anhydride followed by hydrolysis, according to the method developed by Kamitori et al. {Scheme 3.10).U1
/ h2n- n \ 92, 98% N. TFAA 93, 80% H2S 0 4 OK CF OMe 94a MeOH 94, 77%
Scheme 3.10 - Synthesis o f diketone 94.
Chapter 3 Fluorinated Chalcones
form o f 94. The *H NMR spectrum o f 94 showed two methoxy peaks at 3.84 and 3.92 ppm and the 2 and 6 aromatic protons as a singlet at 7.57 ppm. The 19F NMR spectrum showed the CF3 group as a singlet at -81.0 ppm.
Aldehydes were converted to phosphonium salts 97a-h via alcohols 95a-h and chlorides 96a-h (with the exceptions o f the 4-methoxy and 4-trifluoromethyl compounds where the chlorides were purchased directly) according to Scheme 3.11. All these reactions gave quantitative yields and the known alcohols and chlorides were identified by NMR spectroscopy and used in subsequent reactions without further purification.
Ketone 94 was used in Wittig reactions with the appropriate phosphonium salt 97 to make the chalcones 88. It was found that the conventional Wittig reaction using NaOMe as base in MeOH gave very low yields (5%-10 %) in this case so the reaction was carried out in THF using lithium hexamethyldisilazane base according to the method developed by Katzenellenbogen et a /,148 this resulted in much more satisfactory yields ( Table 3.3).
p p h 3 + Cl A T X I --- ► A r " " p p h3 96 97, quantitative yield. LiHMDS CF CF: 88
Scheme 3.11 - Wittig reaction between 94 and phosphonium salts 97.
Because there were two carbonyl groups in the diketone 94 it was possible that the Wittig reaction had taken place on the other carbonyl group to that which was intended and formed compound 98 (Scheme 3.12), although this was thought to be unlikely because o f the activating, electron-withdrawing effects that the trifluoromethyl group would have on the alpha carbonyl.
Ar/ ^ 0 NaBH< - A r ^ O H —
H[k . Ar LHMDS Ar PPh3 94 A r \ ^ P P h 3 LHMDS O Hp 88
Scheme 3.12 Alternative Wittig products.
The Wittig reactions all resulted in pure compounds, as seen in the ]H and 13C NMR spectra so there is no doubt that only one o f these isomers was made in the reactions. The proton NMR data o f the chalcones 88 were compared with that o f known compound 99.149
CL H2' OMe H6' ,0 99
Figure 3.13 - Known unsaturated ester.
In the 'h NMR spectrum o f 99 protons H 2’ and H 6’ are reported to form a singlet peak at 6.44 ppm, whereas in the chalcones described here the ortho proton peak comes at roughly 7.0-7.2 ppm (see chapter 8). This is consistent with these protons being in conjugation with the electron withdrawing carbonyl group in the chalcones 88 and hence shifted downfield. In addition, the nOe data discussed below revealed no enhancement at HP when irradiated at H2’/H6’, suggesting that these protons were not spatially close as they would be in Z-98. Since if 98 were formed it is unlikely, from both a steric and electronic point o f view, that it would form E isomer, this is further evidence that the structure 88 is correct.
Chapter 3 Fluorinated Chalcones
The Wittig reaction was successful with all the aryl groups with the exception o f the 3-fluoro- 4-methoxy and 3-bromo-4-methoxy ylids. Those chalcones which were isolated were tested for cytotoxicity and tubulin binding and the results summerised in Table 3.3, once again cytotoxicity and tubulin inhibition are expressed as I C 5 0 values.
Ar group substitution Compound number Yield (%) Cytotoxicity IC50 Tubulin inhibition IC50 3,4-Methylenedioxy 88a 58 0.44 pM 3.2 pM 4-Methoxy 88b 64 84.6 nM 3.7 pM 3-OTBDMS-4- methoxy 88c 49 >10 pM N/A 3-Hydroxy-4-methoxy 88d 86 79.4 nM 3.9 pM 3,4-Dimethoxy 88e 71 0.61 pM >10 pM 2,3,4-Trimethoxy 8 8f 36 1.99 pM N/A 4-T rifluoromethyl 88g 83 0.21 pM N/A 2-Fluoro-4-methoxy 88h 22 0.18 pM 0.7 pM
Table 3.3 - Yields, cytotoxicities and tubulin inhibition assays for a-trifluoromethyl chalcones.
As before, in the case o f 88d the chalcone was initially made with a TBDMS protecting group on the hydroxy substituent which was later removed with TBAF {Scheme 3.13).
OTBDMS