Lunzer, Miller, Felsheim & Dean. 2005. Science 310:499-501. |
All IMDHs
use NAD as a coenzyme. We
replaced the wildtype gene of E. coli with an engineered gene that encodes a kinetically
equivalent NADP-specific "RKYVYR" enzyme to break a 4 billion year
old constraint.
The GP map.
Engineered
genotypic intermediates reveal each mutation contributes additively to enzyme
performance (kcat/Km) with NAD and with NADP. A savage trade-off in performance
ensures all genotypes hug the axes.
The GF map.
Engineered
genes were recombined into the leu operon and fitnesses determined in competition with
wildtype. The same mutations
often produce marked epistatic contributions to fitness.
The PF map.
A
metabolic model describing fitness as a function of enzyme performance
explains 97% of the data. Epistasis
is produced by the non-linear dependence of fitness on performance. NADP-specific mutants are inferior
because intense product inhibition by intracellular NADPH lowers in vivo performance.
The
landscape is smooth like Fisher's - not rugged like WrightÕs.
The
landscape has a single adaptive "peak" where the wildtype is..
The.landscape would have two "peaks" (wildtype and "RKYVYR")
were Fisher's assumption, of many mutations each of small effect, correct
(since the trade-off would now force all evolutionary paths into the
maladaptive valley at the origin).
|
