A novel finding was a nearly six degree increase in melting temperature by single point mutant N404C. The BglB crystal structure reveals a weak hydrogen bond between N404 and the backbone of a L402. Molecular modeling of N404C predicts the loss of this hydrogen bond to the protein's alpha helix, in which the protein is allowed to repack into more energetically favorable states.
Similarly, the point mutation W120F resulted in a delta Tm of +1.6 C. The BglB crystal structure indicates a weak hydrogen bond between W120 and the backbone of N163, which could be construed as an unsatisfied polar interaction. The mutation to Phe maintains the structural integrity at the mutation site as well as removes the unsatisfied hydrogen bond donor to the neighboring alpha helix. The increased stability is then due to destabilization of the unfolded state, which exposes a hydrophobic Phe to bulk solvent. It is also worth noting that the mutant W120A results in no soluble protein production after 3 independent attempts, indicating that W120 plays a key role in stabilizing the protein. Previous studies have shown a similar increase in stability upon Trp to Phe point mutations \cite{12600203}. Analysis of a multiple sequence alignemnt of 5,000 proteins from the Pfam database reveals approximately equal probablity of finding Trp, Phe, or Tyr here, and less than 1% representation of all of the other amino acids combined. Thus, the experimental measurements and the evolutionary record agree that W120 plays a key role in stabilizing BglB. No major structural changes are predicted via Rosetta modeling.
Point mutation E222H had a melting temperature of 34.7 C, a nearly five degree decrease than that of native BglB. Previous studies show strong hydrogen bond interaction, 2.6 and 3.1 Å, between E222 and its neighboring R240 residue (Carlin 2016). The introduction of histidine at the mutation site causes the loss of these strong hydrogen bonds as well as the creation of electrostatic repulsion between the partially positively charged and positively charged amino acids. The cumulative effect of this mutation results in the protein's decreased stability at lower temperatures.
Mutants that did not express followed rules broadly consistant with previous work and sequence conservation, such as the large destabilizing effect of any substitation W407X and XXX. Other mutants that did not express mostly followed well-established observations of destabilizing effects, such as the introduction of proline into an alpha helix (Y166P, Q19P), the mutation of topology-defining/helix-ending proline residues (P329N), mutations to and from glycine (), the introduction of charged residues into the hydrophobic core (A236E, F72H), and extreme small-to-large mutations (A227W).
It is interesting to note that all the mutations we made in our study to the catalyic nucelophile (E353) and acid-base (E164) polar residues resulted in soluble protein, consistent with the idea that enzymes must trade structural stability for function in placing these two negatively-charged groups less than 5 Å apart (the sole exception, E164G, could be the result of an altered folding trajectory due to the conformational freedom of glycine).