An old paper and the MGD

In recent email discussions with one of the pioneers of molecular evolution, I became aware of a 1986 paper that suggested the idea of a limit on protein sequence divergence, similar to the MGD. But these authors only recognized a limit for bacteria but failed to see similar limit also exists for nearly all organisms regardless of time of divergence, so long evolution is long enough or more than a few million years. They failed to see the inverse relationship between 'limit' and organismal complexity. They correctly recognized that the limit is determined by function but failed to see the key role of epigenetic complexity in adding more limit in addition to that determined by barebone function. They also did not explicitly point out that the idea of limit is contradictory to the molecular clock paradigm that does not recognize the concept of limit and assumes that everything is more or less on a linear range of divergence during the past 3 billion years of life.

Meyer, T. E., M. A. Cusanovich, and M. D. Kamen. 1986. Evidence against use of bacterial amino acid sequence data for construction of all-inclusive phylogenetic trees. Proc. Natl. Acad. Sci. USA 83:217–220.

abstract
It has been proposed that phylogenetic trees, intended to show divergence of eukaryotic protein and nucleic acid sequences, be extended to include those from bacteria. However, we have compared the amino acid sequences of 18 of the most divergent mitochondrial cytochromes c with those of 18 bacterial cytochromes c2 and have found that the average percentage difference between these mitochondrial cytochromes c and cytochromes c2 was not significantly greater than that among the cytochromes c2 alone. The large discontinuities in physical-chemical properties recognized between the prokaryote and eukaryote cytochromes render it highly improbable that members of the two classes should be no more different from one another than members of either class alone, assuming that sequence differences can accurately reveal evolutionary divergence. Instead, we propose that divergent amino acid sequences approach a limit of change considerably less than for comparison of random sequences. This limit of change presumably is determined by the structure/function relationship. When two homologous protein sequences have reached such a limit, convergence or back-mutations and parallel mutations become as frequent as divergent mutations. As two diverging proteins approach this steady-state condition, sequence differences no longer reflect the numbers of mutations resulting in amino acid substitution and therefore species cannot be positioned on a phylogenetic tree. Insertions and deletions are less reversible than are amino acid substitutions and, provided they are well-documented, might be more reliable indicators of bacterial relationships. Nevertheless, we suggest that data available on bacterial protein sequences do not permit construction of all-inclusive phylogenetic trees. Comparisons of protein and rRNA trees suggest that similar restrictions apply to use of rRNA sequence data.