Variety truly is the spice of life. Without variation, evolution is impossible. Variation is most extreme between the kingdoms, but persists down to the level of individuals within the same species. Variation is easily observable in the phenotypic traits of multi-cellular organisms, but also occurs on a microscopic level, in cell walls, organelle inventory, metabolic pathways, protein sequence and modification, signaling molecules and pathways, all the way down to the genetic blueprint. The genetic blueprint, the nucleotide sequence of DNA, is the root of all variability throughout biology. Changes in DNA sequence infer new selectable phenotypes and thus fuel evolution, leading to the vast diversity in life.

But variation seems to have limits. All known living systems are built with the same basic genetic code, and with the same biopolymers; DNA, RNA and protein. The universality of biopolymer backbones is implicit in Francis Crick’s Central Dogma: DNA to RNA to protein (Crick, 1970). The backbones of DNA (the deoxyribophosphodiester backbone), RNA (the ribophosphodiester backbone) and protein (the peptide backbone) are everywhere the same (Voet & Voet, 2003). Or are they? This fundamental paradigm of biology, that biopolymers are universal, has been challenged. Wolfe-Simon and coworkers recently proposed (Wolfe-Simon et al., 2010) that in a bacterium that lives in an unusual chemical environment in Mono Lake, arsenic substitutes for phosphorus in the nucleic acid backbone.

It has been said that a living system with an alternative biopolymer backbone is an extraordinary premise that should require extraordinary proof (Zimmer, 2010). But what specifically is so extraordinary about a bacterium with a reshaped biopolymer backbone?

Invariance with time. The biopolymer backbones are among the oldest chemical entities in biology (Woese, 2000). The backbones are older than any particular sequence of sidechains, of DNA, RNA or protein.

Invariance with speciation. The biopolymer backbones are invariant in all extant living systems (with the possible exception in Mono Lake). The 16S ribosomal RNA genes of many thousands of organisms have been sequenced, without any evidence of alternative backbone chemistries.

Dependencies. There are vast catalogs of dependencies on the exact (invariant) chemical composition of the biopolymer backbones. Reshaping a biopolymer backbone would break these dependencies, and cause comprehensive non-incremental changes, which are not allowed in evolutionary processes. Reshaping a biopolymer backbone equals catastrophe. This is not to say that polymer backbones did not change and evolve, only that the evolution of backbones was completed long ago, and reached stasis well before the last universal common ancestor of life, over 3.5 billion years ago.

The genetic code. The role of the genetic code in the translation of mRNA to protein perfectly illustrates relationships between biochemical dependencies and invariance over time and over speciation. The observation of a novel enzyme might attract attention, much less so than a serious proposal of an alternative genetic code with, for example, a quartet codon instead of the conventional triplet codon. The difference between an enzyme and the genetic code is in age, in universality and in the catalog of dependencies. The genetic code is ancient, far older than the last universal common ancestor of life, far older than the protein sequence of any enzyme.

The genetic code is universal over all extant species, with only minor variations even from the superheated cauldron of mitochondrial evolution (Watanabe, 2010). The genetic code is involved in innumerable dependencies. Changing the genetic code from a triplet to a quartet would change the sequence of every coded protein, breaking countless dependencies, resulting in comprehensive catastrophe. Therefore, changing the genetic code from triplet to quartet cannot be accomplished by any known evolutionary process, and is considered impossible.

The protein backbone. The structures and stabilities of α-helices and β-sheets are exactly dependent on every atom in the peptide backbone. The function of essentially every protein enzyme in biology is built on specific assemblies of α-helices and β-sheets. Complex metabolic networks are dependent on precise structures, functions and regulation of enzymes. Similarly the structural properties of proteins such keratin, actin, myosin, fibroin and collagen are comprehensively dependent on peptide backbone interactions. A systemic change in the chemical composition of the polypeptide backbone, like a change in the genetic code, would cause enormous changes in structure, function and stability of all proteins, and would entirely disrupt metabolic networks and biological structure. In addition, a change in the chemical composition of the polypeptide backbone would require a massive rewiring of protein biosynthesis.

And finally, since amino acids are used as fodder for the bio-synthesis of many other critical cellular components such as nucleosides, the change would necessarily cascade into essentially all biochemical systems. As in changing the genetic code, reshaping the polypeptide backbone would alter the structures and stabilities of all proteins, breaking countless dependencies, resulting in comprehensive non-incremental catastrophe. Therefore changing the polypeptide backbone cannot be accomplished by any known evolutionary process, and is considered impossible.

The nucleic acid backbone. The same type of dependencies have developed around the nucleic acid backbones (Figure 1), which are generally considered to be even older and more intricately interwoven into the chemistry of life than protein. Ignoring DNA, tRNA, mRNA and the nucleotide metabolites, consider the ribosomal RNA. The rRNA-based core of the ribosome is thought to be older than the genetic code, older than coded protein (Fox & Ashinikumar, 2004; Fox, 2010). The extant ribosome is a large assembly of rRNA and protein, which synthesizes coded protein in all living organisms. We know that the core of the ribosome has been invariant since well before the last universal common ancestor (Koonin, 2003; Woese, 2001; Bokov & Steinberg, 2009; Fox, 2010; Hsiao et al., 2009). As illustrated in Figure 2, the rRNA sequence, the positions of the rRNA atoms in three dimensions, even the positions of magnesium ions and water molecules are conserved in the core of the bacterial and archaean ribosomes and so have been impervious to billions of years of evolutionary pressures (Hsiao & Williams, 2009). The forces maintaining the sequence, structure and molecular interactions of rRNA have yielded a structure of astounding resilience.

Could ribosomal structure and function tolerate a systemic chemical change in the backbone such as the conversion of one atom type to another: phosphorus to arsenic, carbon to silicon, oxygen to sulfur or hydrogen to fluorine? A conversion of one atom type to another would change positions and the electronic structure of every atom of the rRNA, altering every intra- and inter-molecular interaction. The net folding stability, chemical stability and structure of the ribosome would be dramatically altered. Reshaping the rRNA backbone would break innumerable dependencies within the translation system. Therefore, systemic conversion of the RNA backbone cannot be accomplished by any known evolutionary process, and is considered impossible.

Summary. In biology, some things, like the genetic code and the backbones of biopolymers, are universal and have not changed for billions of years (Koonin, 2003; Woese, 2001; Bokov & Steinberg, 2009; Fox, 2010; Hsiao et al., 2009). The basis for this invariance over time and speciation is obvious; a vast catalog of dependencies would break upon a reshaping of a biopolymer backbone, leading to catastrophe. The recent paper by Wolfe-Simon suggests this paradigm is violated by a bacterium grown in an unusual chemical environment. The scientific community has greeted the proposed discovery of a reshaped nucleic acid backbone with a reasonable demand for very strong supporting evidence (Danchin, 2010; Pratt 2010; Zimmer, 2010). To gain acceptance, the evidence supporting this paradigm shift must be significantly stronger than that presented so far.