INTRODUCTION

The ability of organic molecules to self-organize into self-sustaining, reproducing and evolving structures governed the transformation of matter from inanimate to animate on the early earth. Probably the earliest such structures were protocells -- membrane-enclosed, cell-like structures capable of supporting essential life functions, such as the capture and utilization of energy and synthesis of proteins. [5] In modern organisms, most of these life functions are performed by proteins which are, in turn, synthesized on an RNA template. It is, however, unlikely that both proteins and RNA arose simultaneously and immediately became interconnected. The discovery of catalytic properties of RNA led to a suggetion that the present world of nucleic acids and proteins was preceded by the ``RNA World,'' wherein RNA molecules alone acted as both catalysts and information storage systems. [2, 1] This concept, however, encounters considerable difficulties. RNA is fragile and no efficient prebiotic syntheses of its building blocks have been found. Furthermore, RNA cannot be readily incorporated into membranes to perform functions which, in modern cells, include energy transduction and transport. Finally, since there is no relationship between the function of a catalytic RNA and the function, if any, of the protein for which it can code, there is no clear path from the RNA World to today's world of protein catalysis and nucleic acid information storage. We therefore hypothesize that initially protocells evolved in the absence of a nucleic acid-based genome and only later did coded information storage emerge. While peptides do not suffer from similar problems as RNA, amino acids cannot base-pair like nucleic acids, so it is not clear how peptides, alone, could transfer information between generations. Thus a new conception of ``evolution'' is necessary that does not require a nucleic acid-based, or similar, genome.

Central to this new concept of non-genomic evolution is the emergence of peptide-bond forming protoenzymes (ligases). In all likelihood, they were initially very weak, non-specific catalysts, joining amino acids to form peptides of various lengths and sequences. A few of the peptides so generated could have been better catalysts of peptide bond formation than the protoenzymes which formed them. These better protoenzymes would, in turn, generate even more peptides, increasing the rate at which a protocell ``searched'' the space of all peptides for functional ones. Some of the peptides generated in this search would undoubtedly function as proteases, cutting peptide bonds. Since proteases cleave unstructured peptides more rapidly than structured ones, and since functional peptides have to have some degree of ordered structure, the proteases would preferentially destroy non-functional peptides. Occasionally, the newly produced peptides would be capable of performing novel functions. If they integrated into the protocellular metabolism, they could increase its capabilities. This process would eventually lead to the emergence (or utilization) of nucleic acids and their coupling with peptides to yield a genomic system.

For this process to be effective, it is required that protocells grow and divide either by acquiring amphiphilic material from the environment or by producing it internally. The contents of the two ``offspring'' protocells would not be identical and some would not contain the proper suite of components for self-maintenance. Nevertheless, over time, the catalytic efficiency of a community of protocells might increase. This increase in overall efficiency is non-genomic evolution.

Recent breakthroughs in experimental protein chemistry open the gates for systematic experimental and theoretical tests of the ideas undelying non-genomic evolution. Szostak and Roberts [6] have modified the methods of in vitro evolution, previously only applicable to nucleic acids, to select peptides with specific properties. This work will provide needed information on the distribution of catalytic abilities among small peptides. In a series of elegant papers, Ghadiri and co-workers [4, 3, 7] have produced a self-replicating peptide system with an inherent error-correction mechanism and have demonstrated the evolution of populations of peptides. Most recently, Chmielewski, et al. [8] have constructed another peptide system capable of auto- and cross-catalysis and generating self-replicating peptides that were not present in the original mixture.