Through ancestral sequence reconstruction (ASR) techniques, ancient enzymes can be recreated and biochemically tested, giving insight into the enzymes’ evolutionary history. A previous study by Hobbs et al. (2012) has shown that some ancestral 3-isopropylmalate dehydrogenase (IPMDH) enzymes of the Bacillus lineage are more catalytically efficient and kinetically stable than extant counterparts. Given these characteristics, this trend raises questions as to why ancestral Bacillus IPMDH enzymes have been superseded by catalytically slower and less kinetically stable counterparts. The homology between IPMDH and the dehydrogenases of tartrate, malate and isocitrate makes IPMDH an interesting model enzyme in terms of the evolution of substrate specificity. Here, the reconstruction of a 2.7 billion year old enzyme has been attempted to extend the reconstruction of IPMDH back to the last common ancestor of the Firmicutes. This reconstruction tested the limits of ASR techniques in terms of time and levels of sequence divergence, especially for such a structurally complex enzyme. However, upon expression and purification, the enzyme was found to form an inactive, soluble aggregate. This suggests that current ASR techniques are too simplistic to reconstruct the complexity and divergence of IPMDH back as far as the last common ancestor of the Firmicutes. Enzyme evolution was investigated with ancestors from the Bacillus genus. Substrate promiscuity of ancestral enzymes was compared to a contemporary counterpart. It was concluded that the ancestral IPMDH enzymes tested do not show additional substrate promiscuity when compared to contemporary counterparts. The fitness of organisms carrying the IPMDH ancestors was assessed to establish what effects the high turnover rates and kinetic stability possessed by some ancestral IPMDH enzymes had on cells when functioning within the normal catalytic pathway for leucine. In vivo, the fastest and most kinetically stable ancestral IPMDH resulted in slower growth rates. This detrimental effect in vivo clarifies why this enzyme has been lost over evolutionary time. The X-ray crystal structure of the most recent IPMDH ancestor was also determined at 2.6 Å resolution. The structure of this ancestral IPMDH was found to be similar to other IPMDH structures, including the previously solved IPMDH from the last common ancestor of the Bacillus.