Surfactant-facilitated crystallisation of water-soluble foldamers† †Electronic supplementary information (ESI) available: Detailed experimental procedures, additional figures and tables describing structures as well as crystallography details. CCDC 1050867–1050870. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c6sc00090h

Common surfactants promote the crystallisation of a series of water-soluble oligourea foldamers which had previously proven resistant to crystallisation efforts.


Introduction
There is considerable interest in the creation of articial folded molecules able to mimic certain desirable qualities of natural biomolecules. Such moleculestermed foldamers 1,2have been developed within the contexts of a broad range of applications, including biopolymer surface recognition, [3][4][5][6][7][8] hostguest chemistry, [9][10][11][12][13][14][15] catalysis, 11,[16][17][18][19] aqueous self-assembly, [19][20][21][22][23][24][25] and nano-technology. [26][27][28] X-ray crystallography continues to contribute considerably to the growth and progress of the foldamer eld, permitting the structures and functions of novel foldamer architectures to be understood at the atomic level. The growth of single, well-ordered crystals suitable for structure determination by X-ray diffraction methods is consequently an important procedure in the foldamer eld, however, this process can oen be challenging, particularly with respect to the crystallisation of water-soluble foldamers using aqueous biocrystallographic methods (such as vapour diffusion). One factor which almost certainly contributes to the difficulty in obtaining well-ordered aqueous crystals is the typically high solvent contents of such crystals (up to 70%), which hinders the formation of strong intermolecular contacts within a crystal lattice. As one potential method to overcome this obstacle to aqueous crystal growth, we report here the use of cationic and anion surfactants as a means to crystallise a series of fully watersoluble foldamers which had previously proven to be highly resistant to crystallisation efforts. A series of short water-soluble aliphatic oligoureas bearing proteinogenic side-chains were crystallised in the presence of cetrimonium bromide{ (CTAB) or sodium dodecylsulfate (SDS), permitting high resolution X-ray structures to be determined with resolutions ranging from 1.19Å to 1.84Å. Crystals of the oligoureas in the absence of surfactant could not be obtainedindeed, analysis of the crystal structures reveals the surfactant molecules to play a crucial role in crystal packing, forming key intermolecular packing contacts and thereby acting as 'molecular glue' in the crystal lattice. Currently, by far the most commonly reported use of surfactants in biocrystallography is as a means to aid the solubilisation of membrane-associated proteins [29][30][31] (which naturally have limited solubility in water), with very few reports of alternative uses of such molecules. 32,33 Although several innovative techniques for obtaining well-ordered single crystals of short oligomers including nucleic acids, peptides and foldamers have been reportedsuch as racemic [33][34][35][36][37][38][39][40] and quasi-racemic 41,42 crystallographic methodsto our knowledge, the use of surfactants as a means to facilitate the crystallisation of watersoluble yet otherwise difficult-to-crystallise foldamers or peptides has not been reported. We believe that the ndings reported herein may be of interest to those engaged in recalcitrant aqueous foldamer (or peptide) crystallogenesis studies, as a possible means to generate well-formed crystalsand consequently atomic-scale detailsof synthetic biomimetic molecules.
succinimidyl carbamate building blocks following previously reported methods 43 (for a full description of the chemistry methods see ESI †), yielding a pure and highly water-soluble molecule.k In order to understand the folding of this oligomer in an aqueous environment, we sought to crystallise oligourea 1 using standard protein crystallisation methods.
Crystallisation trials were performed at 20 C in standard (aqueous) hanging drops composed of 0.5 mL of a solution of oligourea 1 at a concentration of 20 mg ml À1 (in pure water) plus an equal volume of crystallisation reagent. Several hundred unique crystallisation conditions from standard commercial sparse-matrix screens were used in an attempt to crystallise oligourea 1, however, no crystals were obtained. We then employed focussed, specialised crystallisation screens (involving the systematic screening of salts and pH, for example), these too, however, proved unsuccessful in producing crystals. It was not until we serendipitously investigated surfactants as co-crystallising agents that we were able to grow single, well-formed crystals suitable for X-ray diffraction analysis. Oligourea 1 was nally crystallised using a crystallisation reagent composed of 0.5 M sodium chloride, 10 mM magnesium chloride, 100 mM sodium HEPES buffer (pH 7.0) and 10 mM of the cationic detergent cetrimonium bromide (CTAB) (Fig. 1c).
X-ray diffraction analysis of these crystals (using synchrotron radiation**) revealed good quality, high-resolution diffraction, indicative of well-ordered crystals. A full dataset was collected and processed to 1.44Å, with the data belonging to space group C222 1 , with cell dimensions of a ¼ 40.05Å, b ¼ 40.72Å, c ¼ 18.49Å. The structure was solved by molecular replacement using a previously reported crystal structure of an oligourea obtained from crystals grown from an organic solvent crystallisation system 44 (i.e. non-aqueous crystallisation). Matthews analysis indicated the asymmetric unit to be composed of a single copy of oligourea 1, which was easily modelled into the initial electron density maps provided by the molecular replacement solution. Following this, however, a signicant region of positive (i.e. unaccounted-for) electron density was evident in both 2F o À F c and F o À F c maps. A single molecule of CTAB was comfortably modelled into this electron density, resulting in a total of eight surfactant molecules and eight oligourea molecules per unit cell. The nal model was rened using data to a resolution of 1.46Å, with nal R work and R free factors of 17.76 and 24.36%, respectively (Table 1). Full data collection and renement details can be found in the ESI (Table S1 †).
The crystal structure of oligourea 1 reveals the foldamer to be fully helical, forming the expected canonical 2.5-helix typical of aliphatic oligoureas, 44,49 with all possible intra-helical hydrogen bonds present ( Fig. 2a and b, Table S2 †). The 1 : 1 oligourea : surfactant ratio of the crystal structure results in the surfactant occupying a signicant volume (8.54%, see Table 1) of the unit cell, with crystal packing contacts resulting in a total  of four distinct surfactant molecules interacting with a single oligourea molecule ( Fig. 2c and d).
Analysis of the helix geometry of the crystal structure of oligourea 1 reveals helical parameters almost identical to those of an exemplative canonical oligourea crystal structure derived from surfactant-free crystallisation conditions (Table 2), indicating that the surfactant does not negatively impact the secondary structure of this foldamer. Structural alignment of the CTAB-bound oligourea 1 structure reported here with this non-surfactant bound oligourea helix also reveals a high level of structural agreement (Ca r.m.s. deviation, 0.368Å) ( Table 3). In addition to apparently exerting no inuence on foldamer secondary structure, the surfactant molecules do not appear to locally disrupt any regions of the foldamer eitherall oligourea side-chains, backbone atoms and rst hydration sphere are well-resolved in electron density (Fig. 2e), permitting details of the folding of this molecule in aqueous conditions to be understood at the atomic level. Of particular note is the observation of a chloride ion bound to the free N-terminal urea NH groups of the oligourea. A single chloride ion is hydrogen bonded to the NH groups of two oligourea helices arranged in head-to-head fashion, an arrangement analogous to that of natural chloride ion channels 50 (ESI Fig. S1 †). Oligoureas have shown promise as articial anion receptors in non-aqueous conditions 51,52the observation of a chloride ion bound to oligourea 1 in the crystal structure reported here suggests that the anion-binding ability of oligoureas could feasibly be extended to aqueous conditions, which would provide further possibilities for future application of these foldamers.
CTAB molecules interact with hydrophobic regions of the oligourea, such as the leucine-type side-chains of the urealeucine residues (Leu u [superscript 'u' denotes urea-residue]), as well as with charged regions, such as the glutamate-type sidechains of the urea-glutamate residues (Glu u ), effectively 'sticking' foldamer molecules together in the crystal lattice ( Fig. 2c and d). For example, the charged trimethylammonium head of 'CTAB 1' (as labelled in Fig. 2c and d) is situated within electrostatic-bonding distance of the charged glutamate-type side-chains of residues Glu u 2 and Glu u 7. The alkyl chain of this CTAB molecule then associates through hydrophobic interactions with the leucine-type side-chains of residues Leu u 4 and Leu u 9 of the same oligourea molecule, with the terminus of the CTAB alkyl chain interacting with two additional molecules of oligourea 1 in the crystal lattice (ESI Fig. S2 †). In addition, it appears that the CTAB molecules occupy regions of the lattice otherwise lled with disordered bulk solvent. Indeed, based on  Matthews estimates, a signicant percentage of solvent (almost 10%) is replaced with the more structurally ordered surfactant molecules (Table 1), which would be expected to contribute signicantly to crystal packing interactions and improve the overall ordering of the crystal. Thus it seems that the CTAB surfactant promotes crystal growth through two routes: (1) by bridging inter-foldamer contacts in the crystal lattice, thereby acting as 'molecular glue' and (2) by displacing disordered bulk solvent with better-ordered lattice components (i.e. the surfactant).
In order to investigate whether alternative surfactants could be employed in a similar manneri.e. to promote crystal growth by acting as 'molecular glue'we attempted to crystallise oligourea 1 in the presence of the common anionic surfactant sodium dodecylsulfate (SDS). Crystallisation experiments similar to those described abovebut with SDS in place of CTAB yielded good-quality crystals suitable for X-ray diffraction analysis. A 1.84Å resolution dataset was collected for a crystal of oligourea 1 grown in the presence of SDS. The data were processed as above, and were indexed and integrated in space group P2 1 2 1 2 1 , with cell dimensions (inÅ) of: a ¼ 18.49, b ¼ 40.26, c ¼ 41.02. The structure was solved by molecular replacement, with two copies of oligourea 1 in the asymmetric unit. Towards the end of the renement process, two molecules of SDS were modelled into appropriate residual electron density, resulting in a 1 : 1 surfactant : oligourea ratio, analogous to the equivalent CTAB structure above. The nal model was rened to a resolution of 1.84Å, with R work and R free factors of 23.61 and 26.89%, respectively (see Tables 1 and S1 †).
As with the CTAB-oligourea 1 co-crystal structure described above (structure 1), the oligourea molecules of the SDS complex (structure 2) are fully helical, with no deviations from the expected helical geometry and with all expected intra-helical hydrogen bonds present ( Fig. 3a and b, Tables 2 and S2 †).  Structural alignment of the two crystallographically unique oligourea chains of the SDS complex reveals a high level of correlation, with an r.m.s. deviation of 0.703Å (for 115 vs. 115 atoms) (Fig. 3c and Table 3). More signicantly, structural alignment of the oligourea 1 helices from the CTAB and SDS cocrystal structures also reveals a high level of similaritywith r.m.s. deviation values of 0.393 and 0.805Å for alignments of structure 1 vs. chain A and chain B of structure 2, respectivelyindicating that the nature of the co-crystallising surfactant does not appear to impact the oligourea folding or geometry (Fig. 3c and Table 3). Interestingly, the SDS molecules do not occupy the same positions in the crystal lattice (relative to oligourea 1) as the CTAB molecules in the equivalent structure reported above, but instead are re-orientated by 90 (Fig. 3d and e) seemingly in order to maximise electrostatic contacts between the surfactant sulfate groups and the positively charged ornithinetype (Orn u ) side-chains of the oligourea. Despite the SDS molecules occupying different positions in the crystal lattice (compared to CTAB in structure 1), the anionic SDS molecules play a similar role to the CTAB in crystal packing, involving: (1) the intermolecular bridging of oligourea molecules and (2) the replacement of disordered bulk solvent (Table 1). This suggests that certain surfactants possess a degree of intrinsic versatility as opportunistic components of aqueous crystal lattices, making these molecules potentially useful crystallogenesispromoting tools.
In addition, the re-orientation of the SDS molecules in the crystal lattice generates a curious arrangement comparable, to some degree, to a phospholipid bilayer of a cell membrane (Fig. 3f). Although the foldamers reported here have no known anti-bacterial activity, there is considerable interest in the development of peptides and peptidomimetic molecules (including oligoureas 56,57 ) as anti-bacterial agents. 58 One mode of action of some such peptides is thought to involve peptidemembrane interactions, however, high-resolution structural details of anti-bacterial peptide-membrane interactions are understandably challenging to obtain. Thus, our results suggest that it may be worthwhile for those engaged in such research to consider the use of anionic surfactants as co-crystallising reagents (in conjunction with the peptide-of-interest) as a means to generate a 'model-membrane' in a crystal lattice, and thereby permit potentially valuable structural information to be obtained.
In order to test whether the method of using surfactants to promote crystal growth could be extended to additional difficult-to-crystallise foldamers, we performed crystallisation trials in the presence and absence of surfactants for two additional oligourea moleculesoligoureas 2 and 3 (Fig. 4). These foldamers are analogues of oligourea 1oligourea 2 contains lysine-type urea residues (Lys u ) in place of the ornithine-type urea residues (Orn u ), with oligourea 3 bearing this same Orn u to Lys u replacement, in addition to all leucine-type (Leu u ) urea residues being replaced with isoleucine-type (Ile u ) urea residues (Fig. 4). As with oligourea 1, sparse-matrix crystallisation screening of oligoureas 2 and 3 failed to yield crystals (full details of crystallisation experiments can be found in the ESI †), however, use of crystallisation reagents containing CTAB yielded well-ordered good quality single crystals for both foldamers. Diffraction data collected for crystals of oligoureas 2 and 3 resulted in successful structure determination for both foldamers. Both structures belong to space group C222 1 , with resolutions of the nal rened models of 1.49Å and 1.19Å for oligoureas 2 and 3, respectively (further crystallographic details can be found in Tables 1 and S1 †). As expected, the crystal structures reveal oligoureas 2 and 3 to form well-folded 2.5helices, with average helical metrics and folding characteristics almost identical to those of oligourea 1 (Tables 2 and 3 and S2 †). Importantly, the crystal structures of oligoureas 2 and 3 both reveal the presence of well-ordered CTAB molecules playing key roles in forging crystal packing contacts (Fig. 4). These additional high-resolution crystal structures thus provide further evidence in support of the proposition that certain surfactant molecules can exert a dramatic and positive inuence on the outcome of aqueous foldamer crystallisation endeavours.

Conclusions
The amphiphilic nature of surfactants makes these molecules well-suited for use as 'molecular glue' in crystal lattices, as they have the ability to interact with both hydrophilic and hydrophobic moieties. In the examples reported here, the surfactants appear to aid crystal growth in two ways: (1) by linking foldamer molecules in the crystal lattice, and (2) by replacing disordered bulk solvent with ordered crystalline components. We have shown here that both anionic and cationic surfactants can Helical-wheel diagrams and co-crystal structures of oligoureas 2 (a) and 3 (b) bound by cetrimonium bromide (CTAB). Grey labels on helical-wheels correspond to residues modified relative to oligourea 1. The crystal structures are in the same orientation as oligourea 1 in Fig. 2c and were refined to final resolutions of 1.49Å and 1.19Å for oligoureas 2 and 3, respectively.
promote the crystallogenesis of fully water-soluble oligourea foldamers which had previously proven to be resistant to crystallisation, permitting high resolution crystal structures to be determined with resolutions of up to 1.19Å. The considerable level of structural information provided by these highresolution crystal structuresmade possible only through the inclusion of surfactants as co-crystallising reagentssuggests that the use of surfactants as promoters of crystallogenesis may be well worth considering for those engaged in (or embarking upon) challenging aqueous crystallographic studies of watersoluble foldamers (or peptides). Although it should of course be noted that, as the ndings described herein are conned to a single class of foldamer, further investigations will be required in order to determine whether surfactants are indeed able to aid the crystallisation of additional classes of foldamers (such as b-peptides).