Using chirality to influence supramolecular gelation

Different self-assembled structures can be formed by varying the chirality of a functionalised dipeptide allowing gels with different properties to be prepared.

. Summary of fits to the SANS data for the solutions of the different 2NapFF. Each data set was fitted using a hollow cylinder model. No fit was attempted for the (rac)-2NapFF (see main text). For (mix)-2NapFF, a fixed polydispersity in radius of 0.5 was chosen on the basis of quality of fit to the data.
For the (L,D)-and (D,L)-2NapFF, the fits to a hollow cylinder gave the lowest  2 values.
However, the lengths are surprisingly short and do not agree with the observations by cryo-S7 TEM of long structures (see below). A satisfactory fit could also be achieved to a core shell model (where the core SLD is set to the SLD of the solvent). In these cases, the  2 values were less good, but the fits imply that the structures have a reasonable length. In both cases however the values for the radius and thickness were very similar. Hence, we interpret the length from the fit to the hollow cylinder as being the Kuhn length.        Table S2. Summary of fits to the SANS data for the gels formed for the different 2NapFF

Further analysis of SANS data of gels
Log-log plots of I(q) v q for the gels formed from (L,L)-2NapFF, (D,D)-2NapFF, (mix)-2NapFF and (rac)-2NapFF show significant differences from the SANS data taken from the corresponding solutions (Fig. S10). By considering the gradients of these graphs, we can qualitatively describe the samples. These plots are given in Figure S12. The SANS data for exponent has a value of -4 is reliant on there being an infinitely sharp transition between the two phases within the material which may not be seen in real samples. It is also of note that in (rac)-2NapFF, at higher q, there is a third gradient of -2.6, commonly observed in scattering for self-similar rough surfaces such as those seen in gels. 4 A global fit of each of the SANS patterns shows that the data sets can be best described by a flexible cylinder with an elliptical cross-section. These are shown in Figure S11, with a summary of the fits shown in Table S2. This agrees with the observation in the log-log plots that the data are non-rigid rods. Minor radii, axis ratios and Kuhn lengths are given in Table   S2. Due to the q range of the experiment, it is not possible to fit an overall length to the data.
The radii of the (L,L)-, (D,D)-and (mix)-2NapFF are similar, at approximately 2.7-3.0Å, whereas the (rac)-2NapFF has a larger radius of approximately 5.1nm. It is clear from the cryo-TEM of the solution of the (rac)-2NapFF that the sample contains a mixture of fibres and tubes and is very polydisperse. It is likely that this polydispersity persists in the gel phase.
Examination of the Kuhn lengths obtained from these global fits shows that the (L,L)-2NapFF is significantly more flexible than the (D,D)-2NapFF; interestingly the rheological data shows that the (D,D)-2NapFF is a much stronger gel than that of the (L,L)-2NapFF. There is no obvious reason as to why changing the chirality should have such a dramatic effect on the gel S14 properties and this will form the basis of further work. The (mix)-2NapFF sample can be seen to be less flexible than the (L,L)-2NapFF but more than the (D,D)-2NapFF. This can be rationalised on the assumption that there is no sorting of the molecules into individual (L,L)-2NapFF and (D,D)-2NapFF fibres and that all fibres will contain some (L,L)-2NapFF and some (D,D)-2NapFF molecules, although each fibre contains a different (and random) amount of each. Thus the overall flexibility of the sample will lie between that of the (L,L)-2NapFF and the (D,D)-2NapFF. The Kuhn length of the (rac)-2NapFF sample shows that it is stiffer than the (L,L)-2NapFF or the (mix)-2NapFF, but less stiff than the (D,D)-2NapFF. As now the sample contains all of the possible combinations of enantiomers, the resulting sample is seen, as highlighted above, to be a mix of fibres and tubes and to have contributions to the stiffness from the packing of each combination of chiral centres.
As an additional confirmation of the size and shape of the fibres within the gel, a modified Guinier plot was produced, by plotting ln(Q α I) vs Q 2 where α = 1 for a rigid rod, and 2 for a ribbon-like fibre with an elliptical (or rectangular) cross-section. 4 It was found that the best linear fit of the data was to a plot where α = 2. These modified Guinier plots are shown in Figure S13. A value for the cross-sectional thickness of the scattering object (t) can be extracted from this data 3 and showed that for the RR, t = 5.3 nm. Based on the radius of the global fit, the thickness here was found to be 5.4 nm, which is in good agreement. Similarly, for the SS, t = 6.3 nm, with a diameter from the global fit found to be 6.0nm. The same analysis was performed on the racemic gel and mixed gel. For the racemic gel, t = 10.6 nm, corresponding well to the global fit which gave a diameter of 10.2 nm. In the case of the mixed gel, t = 6.8 nm, which compares to a value of the diameter from the global fit of 6.1nm. In all cases, prior to the linear region in the ln(Q 2 I) v Q 2 plot there is a bump as Q tends to 0 which is suggestive of thicker bundles of fibres which form the gel network. 4 S15 Figure S15. Plots of ln(q 2 I(q)) v q 2 to determine the cross section of the fibre in (a) (L,L)-   and 10 mg/mL at a shear rate of 10 s -1 . Red circle symbols show data for (L,L)-; mustard diamond data are for (D,L)-; purple square data are for (rac)-1ThNapFF.

Synthetic Details
All reagents and solvents were purchased from the usual commercial suppliers and used as there was concern that precipitation from diethyl ether might skew the isomeric distribution of the mixture due to differential solubilities of its components. It was indeed found that initial trituration of the evaporated reaction mixture with diethyl ether provided a white solid which was highly enriched in the (S,S) and (R,R) isomers, while subsequent crystallisation of the mother liquor provided a white solid highly enriched in the (R,S) and (S,R) isomers. The yields for both crops of isomers were roughly equal and represented a 73 % total yield. The deprotection of the (rac,rac) compound was instead carried out with hydrogen chloride in diethyl ether or dioxane. The workup for this reaction is simple evaporation of the excess HCl and solvent, however it provides no means of washing out impurities. The precursor to this step (ED-006) must therefore be pure, or any impurities will remain in the product.

Methyl (2R)-2-[(2R)-2-amino-3-phenylpropanamido]-3-phenylpropanoate trifluoroacetic acid salt (DF-004)
To a solution of DF-003 (1.65 g, 3.87 mmol) in chloroform (10 mL) was added trifluoroacetic acid (ca. 10 eq, 3 mL) and the mixture was stirred at ambient temperature overnight. After this time, TLC indicated the absence of starting material. The reaction mixture was poured into diethyl ether (ca. 200 mL) and stirred for 1 hour. The precipitate was filtered off, washed in the filter with additional diethyl ether, and dried under vacuum at 50 °C overnight. The title compound DF-004 was thus obtained as a white solid (1.45 g, 85%) containing < 0.5% (NMR) residual diethyl ether, which was observed in the proton and carbon NMR spectra.
After this time, it was diluted with chloroform, washed in turn with 1M hydrochloric acid and brine, dried (MgSO 4 ), and evaporated to dryness under reduced pressure. Crude EE-007 was thus obtained as a grey foam (1.84 g, 94%) and used as is in the next step. A small amount was purified via column chromatography (1:9 ethyl acetate/dichloromethane) to afford a sample for characterisation (sticky white solid). The NMR spectra of the material are complex S68 due to the presence of diastereoisomers but identical to an overlay of the individual spectra of the enantiopure (S,R) (or (R,S)) and (S,S) (or (R,R)) isomers.     12 g, 69%) as a white solid and in close (44:56) to the expected (50:50) ratio of (S,S)/(R,R) to (S,R)/(R,S) isomers. NMR data is complex due to the presence of diastereoisomers but is identical to an overlay of the NMR spectra of optically pure (S,S) (or (R,R)) and (S,R) (or (R,S)) isomers.      To a solution of EG-002 (597 mg, 1.13 mmol) in tetrahydrofuran (10 mL) was added a solution of lithium hydroxide (4 eq, 108 mg) in water (10 mL) and the mixture was stirred overnight.
After this time, it was poured into 1M hydrochloric acid (ca. 400 mL) and stirred for one hour.
Filtration, washing with water in the filter then drying by repeated azeotropic distillation with acetonitrile afforded the title compound as a white solid (1.41 g). The aqueous washings developed a precipitate on standing. This was filtered off, washed with water and dried as above to afford another crop of title compound (135 mg). Combined yield 1.54 g (97 %). Purity        In a similar fashion to the above, 6-bromo-2-naphtoxyacetic acid was coupled to alanine ethyl ester (either the L-, the D-, or rac-), and deprotected using lithium hydroxide to give 6-Br-2NapAOH. This was then coupled to glycine ethyl ester, followed by deprotection using lithium hydroxide as described above.