Amino acids show up every where in metabolism, both D- and L-forms. A quick perusal of a few biosynthetic pathways will reveal that there are many more amino acids that are intermediates in the synthesis of the canonical 20. Why were any one of those not selected to be coded into protein? Perhaps some of them were, e.g., ornithine is an intermediate in arginine synthesis and could have filled a role for a basic amino acid in primitive proteins.
Actually, I ran across a paper that is really stunning. Michael Yarus's lab found a 5 nucleotide RNA that could take an activated amino acid (Phe-AMP, Phe-UMP, or Met-AMP) and catalyze the transfer of the amino acid onto the 2'-OH of a partially complementary RNA. The sequence of the RNAs are GUGGC (the ribozyme) and GCCU (the substrate). What really blows me away is that further amino acids could be added to form short polypeptides on the end of the GCCU substrate.... and they've managed to sythesise ribosomes from scratch ! (Well, technically speaking its called 'bio-engineering' which capitalises on the flow-on benefits of DNA sequencing research, and a whole bunch of clever equipment).
Of course, this is only a small step and is somewhat artificial: the activated Phe-AMP was present at 1-2 mM, but just demonstrating this really, really simple reaction is important in discovering other interesting catalytic possibilities. Many of the ribozymes that have been discovered are much larger, though this may just be a byproduct of the methods used to make and select the molecules. Most methods I've seen are derived from the SELEX method of generating RNA aptamers.
Certainly, classifying the functions of different components is a powerful way of organizing the processes of a cell, but at their core heritability and replication are a series of reactions in which one component meets another and something happens. One pitfall is to inappropriately assign a limiting definition of a function to a molecule or class of molecules. A prime example is the thought that only proteins were catalysts. Once RNA was show to perform catalysis, people started noticing that it was not an inert carrier of genetic information or a structural component of the ribosome.Hmm .. well, clearly metabolism is essential. Heritability and replication are also essential. Whilst these latter two, clearly require metabolism, they are separable functions which also make use of the same cell componentry.
I guess the point that I was trying to make was in thinking about prebiotic and early evolution, we have to be careful not to limit ourselves by the admittedly successful conceptual framework built around our knowledge of how cells function. In a self-sustaining system, every component is required for its own synthesis, be they intermediates in a metabolic pathway or a complex structure such as RNA polymerase. Coming to grips with the structure of biological networks is a challenge.
Finally, here is a fabulous lecture about the ribosome.