Cryo-EM and directed evolution reveal how Arabidopsis nitrilase specificity is influenced by its quaternary structure

Abstract

Nitrilases are helical enzymes that convert nitriles to acids and/or amides. All plants have a nitrilase 4 homolog specific for ß-cyanoalanine, while in some plants neofunctionalization has produced nitrilases with altered specificity. Plant nitrilase substrate size and specificity correlate with helical twist, but molecular details of this relationship are lacking. Here we determine, to our knowledge, the first close-to-atomic resolution (3.4 Å) cryo-EM structure of an active helical nitrilase, the nitrilase 4 from Arabidopsis thaliana. We apply site-saturation mutagenesis directed evolution to three residues (R95, S224, and L169) and generate a mutant with an altered helical twist that accepts substrates not catalyzed by known plant nitrilases. We reveal that a loop between α2 and α3 limits the length of the binding pocket and propose that it shifts position as a function of helical twist. These insights will allow us to start designing nitrilases for chemoenzymatic synthesis.

Keywords: Biocatalysis; Electron microscopy; Hydrolases; Protein design; Secondary metabolism.

Fig. 1

Data quality. a In our hands purified recombinant AtNIT4 forms relatively short heterogeneous filaments (scale bar = 40 nm). b The sample after incubating and blotting in a low-humidity environment: the concentrating effect of dehydration resulted in the formation of a monolayer of long, straight filaments stacked up against one another (scale bar = 40 nm). c The filaments showed helical order to a limit of ~9 Å. d Clear secondary structure can be visualized in the classified average, but this deteriorates with increasing distance from the center because of the curved nature of the filaments (scale bar = 5 nm). e By masking out a single helical turn and treating this as a single particle we were able to improve the resolution substantially

Fig. 2

NIT helical structure. a In vitro, AtNIT4 assembles into an extended left-handed helical tube with 4.9 subunits per turn (helical twist of −73.0°) and outer and inner diameters of ~13 nm and ~2 nm, respectively. Arrows indicate the positions of the two diad-axes (#) and (*). b, c The NIT4 filament is comprised of a left-handed helix with a pitch of 8.62 nm. Monomers associate across previously described interfaces^,; while no interaction was observed across the D- or F-interfaces, they are shown for the sake of completeness. The positions of the two diad-axes (#) and (*) relative to individual monomers are shown (EMD-0320). d About 80% of the model was built with high certainty, 34 N- and 8 C-terminus residues and 24 residues on the outer surface of the filament were disordered and could not be visualized in the map. ß-sheets stabilizing the helical structure can be seen in the core of the enzyme (pdb id: 6i00). e The C-interface is predominantly an interaction across monomers 2 and 3. f The C-terminal tail of each sequential monomer contributes two ß-strands, which form one ß-sheet for every dimer

Fig. 3

The substrate-binding pocket. a AtNIT4 structure showing two dimers interacting across the C-interface, the position of the catalytic site is indicated by a square. b The active-site pocket, the catalytic residues (E76, K163, E170, and C197) are shown (**), residues conserved in all plant NITs (*). A loop (tan) arising from the adjacent monomer (between α2 and α3) extends over the entrance to the active-site pocket. c As expected, the amino acid residues surrounding the binding pocket correlate with substrate specificity. Three amino acid positions in particular are responsible for the majority of the variation between known plant NITs: 169, 224, and 95 (numbered according to AtNIT4). d A lid (tan), formed by the loop, limits the length of the binding pocket (#). R95 is in the lid loop; S224 is on the outer border of the binding pocket and L169 points toward the light green subunit across the C-interface

Fig. 4

Generating the library. Mutations were introduced sequentially: R95 (1 mutation), R95 + L169 (2 mutations) and R95 + L169 + S224 (3 mutations). a Graphical representation of the bottlenecks in the experiment. As the library was assembled (one mutation to three mutations), more colony forming units (CFUs) are required to satisfactorily sample the full library diversity. Triangles represent the number of unique mutants present required for 95% probability of full-coverage (all possible protein mutants present), top-mutant (the best amino acid sequence present) and top-ten (one of the 10 best amino acid sequences present) in magenta, yellow, and green, respectively. The upper (dark gray) and lower (light grey) 95% confidence interval values for number of CFUs show that by the end of library creation at least one of the top ten mutants was present (stars). This diversity was maintained through all cloning and selection steps. b The final library mix was sequenced and showed successful overlapping NNS peaks at amino acid positions 95, 169, and 224 (sample size: 1% of the total CFUs in each sample)

Fig. 5

NIT4 mutant selection, screening and characterization. a, b, c, d A variable number of colonies were visible after 2 days of incubation on the selection plates. The plates, substrate structures, and sequencing results are shown. e Screening results indicate measured absorbance after incubating with AtNIT4 wt, AtNIT4 R95T L169A S224Q and an inactivated enzyme control, and assaying with Nessler’s reagent. Black indicates that AtNIT4 R95T L169A S224Q produced more ammonia than AtNIT4 wt. Gray indicates that the substrate reacts to form a high-absorbance product, but after incubating with the mutant, the absorbance was reduced. White indicates a slightly higher AtNIT4 wt activity compared to the mutant. f Enzyme activity results, asterisk indicates no measurable ammonia formation after incubating for 30 min with 5× excess enzyme. AtNIT4 wt showed no nitrilase activity against any of the listed substrates except for β-cyano-l-alanine. The mutant showed high activity against potassium cyanide, 2-cyanopyridine and 2-furonitrile and moderate activity against fluoroacetonitrile and 4-cyanopyridine (1st quartile, median and 2nd quartile ± sd) n = 3 independent activity measurements

Fig. 6

Comparison of plant NIT helical twists. a A comparison of different plant NITs; each monomer in the filament has been colored differently to illustrate the number of protein subunits in a helical turn (360°), the brown subunit (dimer 6) is more visible with decreasing helical twist. b AtNIT4 negative stain map vs cryo-EM model, AtNIT4 has ~5 dimers per turn. c Nit6803 ∆291 negative stain map vs crystal structure (**) indicates the missing C-terminal residues. d AtNIT4 R95T L169A S224Q negative stain reconstruction, the low helical twist and large diameter show similarities to Nit6803 ∆291, density accounted for by the interacting C-terminal tails can be seen in the enzyme core. e Comparison between the structure of AtNIT4 (labeled −73°) and AtNIT4 with the helical operators of the Nit6803 ∆291 crystal structure applied (labeled −60°). The lid loop shifts in the direction shown by arrows (in the plane of the page), resulting in steric clashes. Geometrically, this movement is greater when the diameter of the fiber remains constant, but the helical twist decreases. Scale bar = 10 nm (applies to all end-on views)