Bioprospecting Reveals Class III ω-Transaminases Converting Bulky Ketones and Environmentally Relevant Polyamines

Amine transaminases of the class III ω-TAs are key enzymes for modification of chemical building blocks, but finding those capable of converting bulky ketones and (R) amines is still challenging. Here, by an extensive analysis of the substrate spectra of 10 class III ω-TAs, we identified a number of residues playing a role in determining the access and positioning of bulky ketones, bulky amines, and (R)- and (S) amines, as well as of environmentally relevant polyamines, particularly putrescine. The results presented can significantly expand future opportunities for designing (R)-specific class III ω-TAs to convert valuable bulky ketones and amines, as well as for deepening the knowledge into the polyamine catabolic pathways.

. TR3 to TR8 were recovered from 6 positive clones originated from chronically polluted seawater samples from Milazzo harbor (Sicily, Italy) (13,15). TR2 derived from a positive clone from beach acidic pool in Vulcano Island (13). TR9 and TR10 were recovered from a positive clone from P. oleovorans genomic DNA. TR1 and TR3 to TR8 was recovered using benzaldehyde as amine acceptor and 2-(4-nitrophenyl)ethan-1-amine as amine donor, whereas TR2 , TR9 and TR10 were recovered using o-xylylenediamine hydrochloride as amine donor. Table S2. (A) The pairwise sequence similarities for all 10 class III ω-TAs investigated in this study and the 3 similar ATAs previously identified by metagenomics, as calculated using Needleman-Wunsch alignments performed against all the other candidates ("all-vs.-all"). (B) The specific activity given as U mg -1 protein tested against a set of 22 structurally different keto acids, aldehydes and ketones as amine acceptors and 2-(4-nitrophenyl)ethan-1-amine as donor. Reactions performed at 40ºC and pH 7.5. (C) The specific activity given as U mg -1 protein tested against a set of structurally different amine donors and benzaldehyde as acceptor. Reactions performed at 40ºC and pH 7.5. (D) Raw data corresponding to the data presented in Figure 4; the data represent the relative percentages (%) of specific activity at pH 7.5 at each of different temperatures, expressed as U g -1 , compared with the maximum activity when benzaldehyde and 2-(4-nitrophenyl)ethan-1-amine were used as acceptor and donor, respectively. (E) Raw data corresponding to the data presented in Figure 5; the data represent the relative percentages (%) of specific activity at pH 7.5 at each of different solvent concentrations, expressed as U mg -1 , compared with the maximum activity when benzaldehyde and 2-(4-nitrophenyl)ethan-1-amine were used as acceptor and donor, respectively. In all cases, the assays were performed as replicates, with the average value given, and the standard deviations were less than 5% in all cases. S12 Panel (A)   TR1  TR2  TR3  TR4  TR5  TR6  TR7  TR8  TR9  TR10 3A8U KX505389 KX505387 KX505388  TR1 35  TR1  TR2  TR3  TR4  TR5  TR6  TR7  TR8 TR9  TR10 Specific activity (U g -1 protein) A1  TR1  TR2  TR3  TR4  TR5  TR6  TR7  TR8  TR9   1 Activity for (R)-(-+)-2-aminohexane, (R)-(+)-α-methylbenzylamine, and (R)-(−)-1-(2-naphthyl)-ethylamine was below detection limit under our assay conditions.

Panel (D)
Relative activity (%) at each temperature (in °C)  25°C  30°C  35°C  40°C  45°C  50°C  55°C  60°C  65°C  70°C  TR-1  Scheme S1 Enzyme assays for determinations of amino acceptor substrates. Assay reactions were conducted as described in Experimental Section. Specific activity (U/g) was initially evaluated at 600 nm by quantifying the formation of the product in red color. Additionally, the formation of amine products was confirmed by ESI-MS. Acceptor abbreviations correspond to amines in Table S1B, and the correspondence between ID, name and structure is given.
Scheme S2 Enzyme assays for determinations of amino donor substrates. Assay reactions were conducted as described in Experimental Section. Specific activity (U/g) was initially evaluated at 600 nm by quantifying the formation of the product in red color. Additionally, the formation of amine products was confirmed by ESI-MS. Amine abbreviations correspond to amines in Table S1C, and the correspondence between ID, name and structure is given.
2-(4-Nitrophenyl)ethan-1-amine D14 1,2,3,4-Tetrahydro-1-naphthylamine Scheme S3 Enzyme assays for determinations of enantioselectivity based on kinetic resolution. Prior to the assay a stock solution of each racemix amine ((R/S)-2-aminohexane or (R/S)-2-aminonane) was first prepared in acetonitrile at a concentration of 200 mM. A 200 mM benzaldehyde stock solution, used as acceptor, was secondly prepared in acetonitrile. Finally, a pyridoxal-5'-phosphate (PLP) solution at a concentration of 0.2 mM was prepared in 100 mM K2HPO4 buffer pH 7.5. Reactions were performed as described in Scheme 2. Specific activity (U/g) was initially evaluated by quantifying the remaining amount of benzaldehyde by adding at the end of the assay 2-(4nitrophenyl)ethan-1-amine which allows the formation of the product in red color which can be followed colorimetrically at 600 nm. Additionally, the remaining amount of (R) and (S) amines was confirmed by GC-MS and the formation of reaction products by ESI-MS/MS. S21 Figure S1 Agar plate-based ATA screening assay. Class III ω-TA from Pseudomonas putida (PDB 3A8U) is also included as a reference for a well characterized enzyme of this class.
S23 Figure S3 A Coomassie-stained SDS-PAGE gel showing the purity level of proteins after expression in E. coli at 16 °C (TR2, TR9 and TR10) or 37ºC (TR1, TR3-TR8) and further purification by metal affinity chromatography. As shown, purity was assessed as >98% using SDS-PAGE in a Bio-Rad Mini Protein system after a single His6-tag purification step. The molecular weight marker is shown on the left. Note that TR1, TR3 and TR6 were expressed in pBXCH vectors, TR4, TR5, TR7 and TR8 in pBXNH3 vector, and TR1, TR9 andTR12 in pRhokHi-2 vector. They were found as the vectors producing higher level of soluble protein.    Figure S5 GC chromatograms representing the abundance of (R)-aminononane and (S)-aminononane in control (upper panel in black color) and reaction in the presence of TR2 (panel in red), which was used as example of enzyme converting both (R) and (S) amines. Reactions were performed as detailed in Experimental section at 40ºC for 60 min, and reaction products analyzed by GC. Retention time for (R)-aminononane and (S)-aminononane were 6.6 and 6.7 min, respectively. Note that the separation of chiral compounds is really challenging and it is not always possible a full resolution between the peak of the two enantiomers. Being aware of that, we have been conscious that in our case it was not possible to fully separated (R)-aminononane and (S)-aminononane and, in order to avoid any possible mistake in the integration of their respective areas we have based our calculation in the deconvoluted peak area. The deconvoluted peak area is calculated using a mathematical model, which considers the individual Gaussian peak shape that the software (MassHunter Qualitative Analysis, version B.08.00) algorithm is able to construct based on the chromatographic profile.
S27 Figure S6 Conservation of key active site residues examined by multiple sequence alignment of the class III -transaminases herein reported, that of Pseudomonas putida (PDB 3A8U) and those included in Figure   S2. Sequences encoding (R)-specific class IV ATAs in Figure S2 are not included in the alignment.
Residues suggested to be implicated in substrate specificity (6) and PLP binding are shown in bold underlined red colored letters in 3A8U. The hairpin region located in the proximity of the residue R414 (following PDB 3A8U numeration) is highlighted in grey color background with underlined letters. -    IANLA       - - TLKEKGLI   - - - -  Figure S7 Structural investigation of class III ω-TAs herein reported to elucidate key active site residues and active site pocket architecture and size around the Lys288 (following PDB 3A8U numeration) which is essential to mediate the transaminase reaction as a catalytic base. The PDB code of templates used to create the models, and the sequence identity, is shown. The figure illustrates the key active site residues determining the substrate specificity in previously reported ω-TAs (6). Figure S8 Structural investigation of class III ω-TAs herein reported to elucidate key active located in the hairpin where the conversed Arg414 (following PDB 3A8U numeration) is located and the residues constituting the helix in its proximity. The PDB code of templates used to create the models, and the sequence identity, is shown.

S51
The Lys288 (following PDB 3A8U numeration) which is essential to mediate the transaminase reaction as a catalytic base, and the positioning of the conserved Arg414 are shown together with 3 residues whose presence and nature define the access of bulky substrates to active site. S52 Figure S9 Structural investigation of TR1 class III ω-TA to elucidate the positioning of Tyr408 in the access tunnel to the active site pocket (left panel). The presumptive movement, after amine interaction, leading to a higher access channel is shown on the right panel. Figure S10 Structural investigation of TR1 class III ω-TA to elucidate the role and positioning of Ser141 in S-pocket. The surface generated by Ser241 is shown in yellow color, whereas the modelled structure when Ser241 is mutated in silico by Ala in highlighted in blue color. As shown, the presence of an Ala in TR1 does open any space window, which may be enough to allow positioning of amines with larger alkyl substituents. S54 Figure S11 Multiple-sequence alignment of (R)-selective PLP-dependent fold class IV proteins and TR2 class III ω-TA. The key residues found to play a role in recognition of (R)-amines in class IV ω-TA are highlighted in red background in the model enzyme ID 89899273 (17,18). TR2 was selected as example of the class III ω-TA capable of converting (R)-and (S)-amines in this study.