2.1 Synthesis

The bile acid amides 1–6 were synthesized through the route presented in Scheme 1, frequently utilized by our research group.17, 18, 30-36 Compounds 1–6 were purified either by column chromatography or by recrystallization methods. The isolated yields varied from 11 % to 54 %. The syntheses and characterisation details for compounds 1–6 can be found in the Supporting Information.

The synthetic route to bile acid derivatives 1–6.

2.2 Solid-State Structures

Single crystals of the amides 1–5 suitable for X-ray diffraction were obtained by slow evaporation of either their acetonitrile or 1,4-dioxane solutions. Unfortunately, not all of the crystallization attempts were successful. Despite numerous attempts, compound 6 was only obtained as a gel or amorphous powder. The solid-state structures of compounds 1–5 are depicted in Figure 1.

The X-ray structures of 1–5: (a) N-hexyllithocholamide, 1, (b) N-hexyldeoxycholamide, 2, (c) N-hexylcholamide, 3, (d) N-cyclohexyllithocholamide, 4, and (e) N-cyclohexyldeoxychol-amide, 5. Displacement ellipsoids are drawn at the 50 % probability level. Atom sites with minor occupancies have been omitted for clarity. Color code: grey (C), white (H), red (O), and blue (N).

Compound 2 crystallized in the highly symmetrical tetragonal space group P43, whereas all the other hexyl and cyclohexyl derivatives, similar to the previously studied lithocholyl, deoxycholyl, and cholyl amides bearing ethyl,34 propyl, butyl, and isopentyl side chains,33 crystallize in the monoclinic space group P21.

As the previously studied ethyl, propyl, butyl, and isopentyl derivatives,33, 34 also the current series of hexyl- and cyclohexyl amides 1–5 crystallized without solvent molecules. Within the series of bile acid alkyl amides, bile acid-solvent interactions have hitherto only been observed for the N-ethyldeoxycholamide and N-ethylcholamide.34 In addition to stabilizing the crystal lattice, the co-crystallized acetonitrile molecules had a significant influence on the structure of the side chain orientation of the respective bile acid, as can been seen in Figure 2 (red capped stick models).

Overlay of the X-ray crystal structures of all (a) lithocholamide, (b) deoxycholamide, and (c) cholamide derivatives shown in capped stick model. Color code: red (N-ethyl),34 gold (N-propyl),33 green (N-butyl),33 blue (N-isopentyl),33 grey (N-cyclohexyl), and purple (N-hexyl). Co-crystallized solvent molecules as well as atom sites with minor occupancies have been omitted for clarity.

In amides 1, 2, 4, and 5, intermolecular bile acid-bile acid interactions, i.e. O−H⋅⋅⋅O and N−H⋅⋅⋅O hydrogen bonds between hydroxyl and amide groups, are responsible for forming the typical ordered 1D bilayered structures (Figure 3(a) and Figures S8–S12).33, 39-41 Differing from the other derivatives, compound 3 forms a 2D hydrogen-bonded polymer where neighbouring 1D bilayered, tubular assemblies are interconnected through N−H⋅⋅⋅O hydrogen bonds as illustrated in Figure 3(c) – black broken lines. Since compound 3 crystallizes not with one molecule but two crystallographically independent molecules in the asymmetric unit, the N-alkyl side chains are oriented in different conformations (Figure S5) and are engaged in hydrogen bonding different to what had hitherto been observed.

The 1D hydrogen-bonded polymer viewed along the a-axis (a) and the tubular assembly viewed along the b-axis of 1 (b), 2D hydrogen-bonded polymer viewed along the a-axis (c) and along the b-axis of 3 (d) in capped stick model. Atom sites with minor occupancies have been omitted for clarity. Broken lines represent N−H⋅O (turquoise and black) and O−H⋅O (green) hydrogen bonding. Color code: grey (C), white (H), red (O), and blue (N).

The infinite 1D bilayer structure observed in virtually all bile acid amide structures33, 34 is stabilized by several intermolecular O−H⋅⋅⋅O and N−H⋅⋅⋅O interactions between the hydroxyl groups at the 3α(/7α/12α)-position(s) as well as by interactions between the carbonyl and N−H groups in the side chains of the bile acid amides. This known motif was observed for one of the molecules of 3. The other molecule in the X-ray structure of 3 forms the bilayered structures only with O−H⋅⋅⋅O hydrogen bonds. The “free” N−H groups act as linkers between the bilayers and are stabilized by N−H⋅⋅⋅O(carbonyl) interactions with d(N⋅⋅⋅O)=3.008 Å. The polar and apolar layers alternate along the b-axis (Figure 3(d)).

Compared to the previously studied ethyl, propyl, butyl, and isopentyl derivatives,33, 34 the n-hexyl and cyclohexyl groups in compounds 1–5 impart flexibility with high degrees of conformational freedom (Figures S4 and S5). In addition, the flexibility of the alkyl chains increases with the number of hydroxyl groups attached to the steroidal backbone (Figure 2). Hirshfeld surface analyses42-45 for the crystal structures indicate that the bilayers form a compact 3D crystal lattice through a high percentage of H⋅⋅⋅H contacts (Table S4 and Figures S13–17, Supporting Information). Surprisingly, the highest percentage of H⋅⋅⋅H contacts was observed for 2, N-hexyldeoxycholamide, which contrasts the previously observed behavior where the H⋅⋅⋅H interactions decrease with an increasing number of hydroxyl groups attached to the steroidal backbone (Table S4). The X-ray structures of N-hexyldeoxycholamide, 2, and N-cyclohexyldeoxycholamide, 5, do not offer any apparent molecular or intermolecular-interaction-based explanation (see Hirshfeld surface analysis, Table S4) on not forming gels similarly to the OH-group-containing lithocholamides and the three OH-group-containing hexylcholamides (Table 1). Despite very similar molecular conformations and packing in the crystal lattice (Figures 1 and 2), the likely cause for the non-gelation is either too low or too high solubility in the target solvent.

Table 1. Solvents in which compounds 1–6 formed gel systems.

Solvent

1

2

3

4

5

6

Benzene

PG

PGs

Gf

–

–

Gf

Toluene

Gf

–

Gf

PGs

–

PGs

Ethylbenzene

PG

–

–

PG

–

Gs

o-Xylene

PG

–

Gf

PG

–

Ges

m-Xylene

Gf

–

–

PG

–

PGs

p-Xylene

Gf

–

–

PG

–

PGs

Mesitylene

PG

–

–

–

–

–

tert-Butylbenzene

Gf

–

Gf

PG

–

–

Cumene

PG

–

–

PG

–

–

Chlorobenzene

–

–

Gs

–

–

Ges

Anisole

PGs

–

Gs

–

–

Ges

ACN

Gf

–

–

–

–

–

Ethyl acetate

Gf

–

–

PG

–

–

n-Hexane

Gf

–

–

PG

–

–

Diethyl ether

PGf

–

–

–

–

–

DMSO

–

–

–

PG

–

–

Formamide

–

–

–

PG

–

–

Ethylene glycol

–

Gs

–

PG

–

–

Cyclohexene

PG

–

–

G

–

–

Abbreviations: Gf (gel formed under 30 min), G (gel formed within an hour), Gs (gel formed after 1 day), Ges (gel formed extremely slowly, after months), PGf (partial gel formed in 30 min), PG (partial gels formed within an hour), PGs (partial gel formed after 1 day), and “–“ (no gel). 2.3 Gelation Studies

The results of the gelation test of compounds 1–6 conducted in 36 solvents were in good congruence with the results obtained previously in our group.17, 18, 30-36 Most of the gel systems were formed in aromatic solvents and are presented in Table 1. More detailed information on the gelation experiments can be found in the Supporting Information. A total of 44 gel systems were formed, 28 of which by lithocholic acid derivatives (compounds 1 and 4), two by a deoxycholic acid derivative (compound 2), and 14 by cholic acid derivatives (compounds 3 and 6). No gelation was observed in alcohols or chlorinated solvents with the exception of chlorobenzene.

The deoxycholic acid derivative 5, containing a cyclohexyl side chain, did not form any gels. The other deoxycholic acid derivative with the hexyl side chain (compound 2), however, formed two gel systems. This was unexpected, since deoxycholic acid derivatives do not usually self-assemble into gels as previously observed.17, 18, 30-36 The majority of the gel systems were formed by lithocholic acid derivatives 1 and 4. Compound 1 was observed to form gel systems in the largest number of solvents. The cholic acid derivative 3 was the only compound exclusively forming flawless gels, half of which started to degrade after four hours. According to our previous studies,33, 34 lithocholic acid derivatives have been more efficient gelator molecules, whereas in this study, mostly partial gels were obtained. For example, in the case of compound 1 which formed the largest number of gel systems, eight of the fifteen gels formed were partial.

The time of the gel formation was exceptionally short for compound 3 in most of the solvents: four out of the six gels formed in only three minutes. On the other hand, in chlorobenzene and anisole, the gel formation took four days. There appears to be a correlation between the temporal stability of the gel and the formation time: slowly formed gels were stable, whereas extremely quickly formed gels started to collapse within four hours (except the gel in tert-butylbenzene). A similar phenomenon was not observed with compound 6, where both the quickly and the slowly formed gels were equally stable. Compound 1 self-assembled into gels fast as well, but the gelation time was considerably slower (over fifteen minutes) when compared with compound 3. When comparing the melting points of the benzene gels of compounds 1, 2, 3, and 6, differences were observed. The benzene gels of compounds 1–3 had similar melting points ranging from 64 °C to 69 °C. Hence, the compound or the formation time of the gel did not have a pronounced effect on the melting point. However, the benzene gel of compound 6 deviated from this trend as it had a significantly higher melting point (80 °C). In the case of compound 6, melting points of the gels were also determined in differing solvents. The gels in chlorobenzene, o-xylene and anisole exhibited lower melting points ranging from 70 °C to 76 °C than those observed for the corresponding benzene gel. The highest melting point was observed for the ethylbenzene gel of compound 6 (94 °C).

When considering the melting points of the gels in different solvents and the formation time of the gels, the least stable ones corresponded to the most slowly formed gels.

An interesting congruence with respect to the gelator molecule and the solvent was observed, showing that the lithocholic acid derivatives (1 and 4) formed gels when having a complimentary structure of the side chain and the solvent. Compound 1, with its linear hexyl side chain, formed a gel in n-hexane, whereas compound 4 only formed a partial gel. In cyclohexene, compound 4, with its cyclohexyl side chain, formed a gel and compound 1 self-assembled into a partial gel.

Differences in the appearance and consistence of the gel systems in different solvents were also observed. Compound 1 produced cleavable, bright gels in benzene and cyclohexene, whereas the gels in acetonitrile and diethyl ether were clearly fibrous. The gel of compound 4 in toluene was also fibrous. Gels in cumene were constructed of small spheres in the cases of compounds 1 and 4. In addition, the gel of compound 4 in cyclohexene also consisted of spheres. The gels of compound 2 in benzene and ethylene glycol were ductile, as was the gel of compound 3 in benzene. Especially in the case of compound 3 in benzene, the gel appeared extremely stretchy. The gel formed by compound 6 in benzene, for one, resembled a jelly. All the gel systems were qualitatively stable; each either maintaining its shape or returning to a smooth-surfaced gel after being disturbed with a spatula. This was surprising, since in our previous studies the gels collapsed immediately when disturbed.33, 34

Similar to our previous studies,33, 34 all the gels in the current study were thixotropic by nature.

Kamlet-Taft solvent parameters can be used in determining the solvents’ properties.46-48 They include three parameters that can be utilized in the analysis of solvent interactions in gel systems. The α- and β-parameters describe the hydrogen-bond-donating and accepting ability of the solvent, respectively. The third parameter, the π*-parameter, refers to the polarizability of the solvent. In the current study, all solvents in which gels or partial gels were formed were analyzed with respect of the three Kamlet-Taft parameters (see Figure 4). The gelating and non-gelating solvents are divided by a vertical line in Figure 4.

Kamlet-Taft parameters (α, β, and π*) for gel-forming (1–19) and not gel-forming (20–36) solvents (Table 1 and Supporting Information). The vertical line divides the gel-forming and the non-gel-forming solvents.

In our previous studies,36 the α-parameter (hydrogen bond donor ability) was observed to be an important Kamlet-Taft parameter effecting the immobilization of the solvent by a gelating compound. The value was zero for most of the gel-forming solvents. In this study, the α-parameter exhibited similar properties (Figure 1) as it was zero for all the gel-forming solvents, with the exception of acetonitrile (0.19), formamide (0.71), and ethylene glycol (0.90). The majority of the β-parameter values (hydrogen bond acceptor ability) were in accordance with previous studies of bile acid derivatives having moderate values from 0.10 to 0.31.34 The gelating solvents diethyl ether, ethyl acetate, dimethyl sulphoxide, and ethylene glycol have higher β-parameter values (0.47, 0,45, 0.76, and 0.52, respectively). This differs from our previous results,36 where a higher β-value seemed to completely prevent gelation. The lack of the functional group at the end of the aliphatic side chain probably enables the gel formation in a larger versatility of solvents. On the other hand, the steroidal backbone may also play an important role: in the current study, lithocholic acid, deoxycholic acid, and cholic acid backbones were used, whereas in the previous one,36 instead of cholic acid, chenodeoxycholic acid was used. When comparing the π*-parameter (polarizability) values, acetonitrile, formamide, and ethylene glycol again stand out with significantly higher values, as do chlorobenzene and anisole. For all the other gelating solvents, the value of the π*-parameter varies between 0 and 0.56.

When looking into the Kamlet-Taft parameters regarding the chlorinated solvents, an interesting difference is observed. Chlorobenzene, in which gels were formed, is clearly divergent from chloroform and dichloromethane, which did not induce gel formation. Chloroform and dichloromethane both possess a slightly higher hydrogen bond donor ability (α-value) than chlorobenzene. The hydrogen bond acceptor ability (β-value), however, seems to be a more significant factor for gelation within chlorinated solvents used in this study, since chlorobenzene has the highest value of the three. The same observation applies to the polarizability of the solvent (π*-value), which means that the solvation of peripheral groups of the gelator molecule and the fiber-fiber interactions play a major role in gel formation in chlorinated solvents.

When comparing the aromatic solvents in which gels were formed (excluding chlorobenzene), anisole seems different from the other aromatic solvents at first glance. Anisole has markedly higher β- and π*-values than other the aromatic solvents, but the α value is zero, as is the case with all the other aromatic solvents. When the ratio of β/π* is calculated (Table S2, Supporting Information), all aromatic solvents, however, show similar ratios around 0.30.

Compounds 1–6 were highly soluble in alcohols, meaning that gelation was not observed. All Kamlet-Taft parameters are, in general, in the same range for the alcohols, which leads to the conclusion that solvents which possess relatively equal hydrogen bond donor and acceptor abilities are not good solvents for gelation with bile acid amides.

2.4 Morphology

In the SEM images of the selected dried gel systems (xerogels), mostly fibrous structures were seen. In the fibrous gel networks, the larger fibres were observed to have formed from thinner fibres (Figure S2).

The gels of compounds 3 and 6 formed in benzene were divergent: smooth ball-shaped structures were observed (Figure 5(a) and (b), respectively). For compound 3, the diameter of the spheres varied from 1.3 μm to 3.9 μm. In the case of compound 6, the spheres were fused https://www.merriam-webster.com/dictionary/elided and the gel appeared as dough-like in the SEM images. The spherical structures were uniform in size, the average diameter being 665 nm. The gel of compound 3 in benzene was extraordinary by its physical appearance also, because the gel was very elastic and stretchy.

SEM images of xerogels formed from compound 3 (a), and compound 6 (b) in benzene.

Compounds 1 and 2 in benzene and compound 4 in toluene formed a fibrous gel network consisting of uniform-looking fibres (Figure 6). The width and length of the fibres varied within each system. The longest fibres were formed by compound 4 (from 61 μm to 775 μm) and the shortest fibres by compound 2 (from 12 μm to 242 μm). In the case of compound 1, the length of the fibres varied from 19 μm to 376 μm. When comparing the width of the fibres, compound 4 possessed the thickest fibres (from 1.5 μm to 7.8 μm) and compounds 1 (from 654 nm to 6.6 μm) and 2 (from 732 nm to 4.7 μm) thinner ones. The texture of the gels formed in benzene and toluene was different. The gel formed by compound 4 in benzene was clearly fibrous (already to the naked eye), whereas the gel of 1 in benzene was cleavable. The gel formed by compound 2, on the other hand, was clearly more ductile. It seems that the smaller and more even-sized fibres give rise to a more elastic gel as illustrated by compound 2 in benzene, and fibres with more versatility in size and shape produce less elastic gels as exemplified by compound 1 in benzene.

SEM images of compound 1 in benzene (a), compound 2 in benzene (b), and compound 4 in toluene (c).

When inspecting the gel in acetonitrile and the partial gel in diethyl ether formed by compound 1 at a macroscopic level, both were fibrous. In the SEM images (Figures S1(a) and (b), respectively), both consisted of fibres in various sizes. Particularly, in acetonitrile, the gel network resembled a mixture of pieces of fettuccine and tagliatelle pasta. The length of the fibres varied from 24 μm to 244 μm, and the width from 1.5 μm to 6 μm. The average size of the plate-like structures amounted to approximately 37×66 μm2. In the partial gel formed in diethyl ether, the length of the fibres varied between 9.3 μm and 220 μm and the width from 863 nm to 13.5 μm, respectively.

The gels formed by compounds 1 and 4 in cyclohexene (Figures S1(c) and (d), respectively) and cumene (Figures S1(e) and (f), respectively) resembled each other. In the case of compound 4, a correlation between the appearance of the gel and the structures seen in SEM images was observed. In both solvents, compound 4 formed a gel system consisting of spherical shapes. In the SEM images, the gel network fibres were observed to form larger spherical structures. In cyclohexene, the length of the fibres was 27–314 μm and the width was measured to 309 nm–2.6 μm, whereas in cumene, the length of the fibres varied from 3.6 μm to 274 μm and the width ranged from 860 nm to 9.8 μm, respectively. The partial gel in cyclohexene formed by compound 1 was cleavable, whereas the partial gel in cumene consisted of spherical structures. Both of the gel systems were observed to consist of evenly distributed fibres with various sizes. In cyclohexene, the length of the fibres was 55–378 μm and the width was found at 1.4–13.6 μm. The partial gel formed in cumene by compound 1 resembled the gel formed in acetonitrile by the same compound, although in cumene, the fibres were smaller. The length of the fibres in cumene was 570 nm–14.8 μm and the width was measured to 7.9–239 μm, respectively.

The gel formed by compound 2 in ethylene glycol (Figure S1(g)), which formed extremely slowly, consisted of beam-like short fibres of various sizes. Their width varied from 1.7 μm to 6.3 μm, and their length ranged from 3.5 μm to 43 μm. The texture of the gel was notably more viscous and it stretched more than the other gel systems studied, an exception being compound 3 in benzene.