Biosynthetic Approaches to Protein Engineering Using Fluorinated Amino Acids

Citation

Son, Soojin (2007) Biosynthetic Approaches to Protein Engineering Using Fluorinated Amino Acids. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/4dsq-m054. https://resolver.caltech.edu/CaltechETD:etd-09172008-144626

Abstract

Non-canonical amino acids provide a new set of building blocks and a potential route to new chemistries that extend beyond those achieved through the 20 common amino acids. Among various methods developed for this purpose, residue-specific incorporation in vivo has the potential to economically engineer proteins of various sizes with altered physical and chemical behavior. The work presented in this thesis explored the incorporation of fluorinated aliphatic residues and their effect on protein structure, stability, and function.

In Chapter 2, two fluorinated amino acids, 5,5,5-trifluoroisoleucine (5TFI) and 4,4,4-trifluorovaline (4TFV), were incorporated in vivo into mutant GCN4 proteins. Both 5TFI and 4TFV showed replacement levels higher than 88%, as indicated by quantitative amino acid and MALDI-MS analysis. The incorporation of 5TFI into a-positions of the coiled-coil protein raised the thermal denaturation temperature (T_m) by 27°C from that of its isoleucine counterpart. However, when valines were replaced by 4TFV in the same positions, T_m only increased by 4°C. Similar trends were observed in response to chemical denaturation by guanidine hydrochloride; ΔΔG_folding upon incorporation of 5TFI and 4TFV was -2.1 and -0.3 kcal/mol, respectively. Secondary and higher order structures as well as biological activity were retained in the presence of both 5TFI and 4TFV. These results indicate that, even when introduced into the same positions within the protein, the effect of fluorination differs depending on which amino acid is fluorinated.

In Chapter 3, the stereochemical effects of 5,5,5-trifluoroleucine (5TFL) were studied using (2S, 4R)-5',5',5'-trifluoroleucine and (2S, 4S)-5',5',5'-trifluoroleucine. The results from in vitro activation assays correlated well with efficiency of their incorporation in vivo. The (2S, 4S) isomer, whose k_cat/K_m was 100-fold lower than that of leucine, was incorporated at high levels, with 91% replacement of the encoded leucine residues in a de novo engineered coiled-coil protein A1. The (2S, 4R) isomer exhibited 9- fold lower k_cat/K_m than the (2S, 45) isomer, resulting in a slightly lower level of incorporation, 80%. The secondary structure of A1 was undisturbed upon the incorporation of either isomer and their impact on thermostability was similar, with an increase of 11°C in T_m as compared to that of A1 containing leucine. However, the equimolar mixture of A1 containing (2S, 45)-TFL and A1 containing (2S, 4R)-TFL displayed a further increase in T_m of 3°C. Although this further enhancement in thermostability was modest, it may be attributed to the ability of the coils to pack more compactly into dimers due to the stereochemical differences.

In Chapter 4, laboratory evolution was utilized to recover the catalytic activity of chloramphenicol acetyltransferase (CAT) after replacement of isoleucine by 5TFI. Upon global incorporation of 5TFI into CAT, the catalytic efficiency, k _cat/K_m, was reduced by more than 2-fold, from 10.2 ± 0.8 μM^-1 to 3.9 ± 0.5 μM^-1 min ^-1. Four rounds of random mutagenesis, enrichment, and screening were performed, yielding a 7-fold fluorinated mutant, tfi-G4, whose activity in fluorinated form was 2.8-fold higher than that of the fluorinated parent enzyme, tfi-WT. The total number of isoleucines decreased only by one in the 7-fold mutant, and the gap in activity between the hydrogenated (ile-G4) and fluorinated (tfi-G4) forms narrowed. Despite similar secondary structure, the incorporation of fluorinated amino acids decreased the stability of CAT for both the wild- type and G4 pairs based on both functional and structural analysis. Fluorinated forms were more sensitive towards thermal and chemical denaturation. However, both forms of G4 enzymes had increased stabilities as compared to their wild-type counterparts. This resulted in tfi-G4 exhibiting similar thermostability as that of ile-WT. Although structural changes were noted at both high and low pH, the pK_a of the catalytically essential histidine (His-193) was not affected by the incorporation of 5TFI. Based on these results, the incorporation of 5TFI appears to be adversely affecting protein folding, which resulted in decreased activity and stability. However, laboratory evolution effectively recovered these losses and yielded a fluorinated enzyme that performed similarly to the wild-type.

Finally, in Chapter 5, the effect of fluorination on sensitivity to proteolytic degradation was explored. GCN4 proteins containing either 5TFI (tfi-INL) or 4TFV (tfv-VNL) as well as CAT proteins containing either 5TFI (tfi-G2) or 5TFL (tfl-L2A1) were treated with two proteases, trypsin and elastase. As evidenced by gel electrophoresis and densitometry analysis, the half life of tfi-INL in the presence of elastase was 4 times that of its hydrogenated counterpart INL. The increased resistance was less significant upon incorporation of 4TFV as well as in response to trypsin. The opposite trend was observed in the analysis of CAT mutants. Both of the fluorinated CAT mutants were more susceptible to elastase and trypsin as compared to their hydrogenated counterparts. Upon incorporation of fluorinated amino acids, two factors that affect proteolytic degradation are altered, the stability of the protein as well as substrate recognition and hydrolysis by the protease. It appears that changes in both of these factors contributed to the observed rates of proteolysis.

The work explored in this thesis has expanded our understanding of fluorinated amino acids and their effects on proteins. Based on these insights, we are continuing to expand the use of fluorinated amino acids (and other non-canonical amino acids) to control protein structure, function, and stability.

Thesis Files