Organization of Peptides at Membrane Interfaces

Even the simplest protocell must have had the capability to catalyze the chemical reactions needed for its survival and growth. One group of potential early catalysts were peptides -- possible precursors of contemporary protein enzymes. In modern enzymes, catalytic activity almost invariable depends upon the structure into which the protein folds which, in turn, depends upon the specific sequence of the amino acid residues along the protein backbone. This poses two problems for peptides to act as protocellular catalysts: First, in the absence of information molecules, high sequence specificity of peptides could not have been required for their catalytic activity. Second, short peptides typically do not exhibit secondary structure in aqueous solution and, therefore, do not appear to be suitable candidates for protoenzymes. There is, however, a growing body of evidence that peptides, which are disordered in water, acquire secondary structure at water-air or water-membrane interfaces if they have a proper sequence of polar and nonpolar residues. Structures that are stable at the interface are amphiphilic; polar residues are immersed in water and nonpolar residues are exposed to air or the membrane interior. The specific identity of the residues is less important, a desirable property in the protocellular environment. All main elements of secondary structure -- -helix [34, 35] -sheet [34] and -turn [] -- have been observed at aqueous interfaces, sometimes for peptides less than 10 residues long.

To examine the effect of sequence amphiphilicity on the secondary structure of simple peptides at aqueous interfaces, we studied two heptamers placed at the water-air interface. A similar behavior is expected at more complex water-oil and water-membrane interfaces [36, 37]. The peptides were composed of two residues, nonpolar leucine (L), and polar glutamine (Q). Their specific sequences were (LQQLLQL) and (LQLQLQL). These sequences were designed to maximize the amphiphilicity of an -helix and a -strand, respectively, by exposing their polar side chains to the aqueous phase and their nonpolar residues to the air (see Fig. 4).

 
Figure 4: Axial projection of the (LQQLLQL) -helix peptide at the water-air interface. Hydrophobic leucine residues are exposed toward the air, whereas hydrophilic glutamine residues are buried in the aqueous phase

In one set of molecular dynamics simulations, (LQQLLQL) and (LQLQL QL) were initially arranged at the water-air interface in amphiphilic secondary structures (-helix and -strand, respectively.) After approximately 3 ns. of molecular dynamics trajectory, both peptides remained at the interface. The -helix remained stable, showing only small deviations from the initial set of angles (; ). The only exceptions were the tail-end leucine residues, for which an equilibrium between () and () was observed.

In contrast, fluctuations within the backbone of the -strand were much larger, with several excursions of the angles from ca. 150 to -40. Examples of these fluctuations are shown in Fig. 5. This clearly reveals the instability of the -strand for a single peptide molecule at the water-air interface. However, at higher concentrations of the peptide this conformation might still be favored due to the association of several molecules into -sheets stabilized by a network on intermolecular hydrogen bonds. Since our system contained only one peptide molecule this possibility was not explored.

 
Figure 5: Time-history of angles (left), and the corresponding free energy profiles (right), for residues 1 and 4 of the (LQLQLQL) -strand

To investigate further the relationship between the sequence of peptides and their secondary structure at the interface, a second set of molecular dynamics simulations was performed. In these simulations, (LQQLLQL) was placed at the interface in the -strand conformation and (LQLQLQL) was arranged as the -helix. Neither of these initial structures was amphiphilic. A trajectory 15 ns. long was generated for this system.

During the course of simulations both peptides remained interfacially active. For (LQQLLQL), the first two angles of the -strand rapidly shifted to -40. The resulting structure, although different from an -helix, was rigorously amphiphilic. The initial, non-amphiphilic -helix of (LQLQLQL) was also found to be unstable at the interface and never refolded to the -strand.

It is clear that the current molecular dynamics simulations are not sufficient to describe all possible folding pathways of the two peptides. To do so, the free energy as a function of backbone angles () has to be fully explored using the approach outlined at the end of the method section (see Eq. (1)). This work is currently in progress.

The results of our simulations illuminate three important properties of small peptides at aqueous interfaces. First, peptides that contain both polar and nonpolar amino acids tend to accumulate at the interface. Second, amphiphilicity provides a strong force driving the peptides towards specific, organized structures. This force is absent in bulk media, such as water or the membrane interior. This tendency to organize at the interface, driven by the amphiphilicity of the structure rather than a specific sequence, is consistent with the concept of an active interface and might have been conducive to primitive catalysis under protobiological conditions. Finally, the degree of structural organization of the peptide backbone changes with the position in the sequence. The backbone is considerably more disordered at the ends of the peptide than in the middle.

The existence of secondary structure in membrane-bound peptides does not necessarily imply their catalytic activity, however. Only a few examples of such activity are known to to date. Peptides containing an alternating Leu-Lys sequence, which folded at the interface to the -sheet geometry, have been shown to hydrolyze polyribonucleotides by the general acid-base mechanism [38]. A decapeptide exhibited almost the same activity as the polypeptide. In another example, a 14-residue peptide, which formed an amphiphilic -helix, catalyzed decarboxylation of oxaloacetate [39]. In both cases, a binding pocket was not necessary to achieve catalytic activity. A simple active center was created by placing a small ligand (e.g. iron) between a bundle of four amphiphillic -helices [40]. However, in general, the link between the interfacial structure of peptides and their catalytic activity remains largely unexplored.