Old yellow enzymes (OYEs) are a group of enzymes that catalyze the reduction of C=C bonds in α,β-unsaturated carbonyl compounds and/or C=C bonds in α,β-unsaturated nitro compounds by using flavin mononucleotide (FMN) as a cofactor. OYE was first isolated from the brewer bottom yeast, Saccharomyces pastorianus, in the 1930s, and since that time OYEs have been found in bacteria, yeasts and plants. A strictly conserved histidine/asparagine or histidine/histidine pair and a tyrosine residue are involved in the catalysis of the asymmetric reduction; a hydride derived from FMNH2 is stereoselectively transferred to Cβ of the bound α,β-unsaturated carbonyl compound, and a tyrosine residue adds a proton to Cα of the α,β-unsaturated carbonyl compound from the opposite side.

Some OYEs display excellent enantio-selectivity in the asymmetric reduction of the C=C bonds of α,β-unsaturated carbonyl compounds. For example, two yeast OYEs, CYE and TYE, isolated from the yeasts Candida macedoniensis AKU4588 and Pichia sp. AKU4542 (formerly, Torulopsis sp.), respectively, can be used to obtain an industrially important chiral compound, (4R,6R)-4-hydroxy-2,2,6-trimethylcyclohexanone [(4R,6R)-actinol], which is applicable for the synthesis of xanthoxin, zeaxanthin and related compounds. These OYEs can reduce 3,5,5-trimethyl-2-cyclohexene-1,4-dione [ketoisophorone] to yield (6R)-2,2,6-trimethylcyclohexane-1,4-dione [(6R)-levodione] in an enantio-selective manner (Figure 1). In addition, TYE can reduce 4-hydroxy-2,6,6-trimethyl-2-cyclohexanone [(4S)-phorenol] to yield (4R,6R)-actinol, although CYE hardly reduces this substrate.

To reveal the structural basis for the different substrate preferences between CYE and TYE, we have solved the crystal structures of these OYEs in the presence or absence of a substrate analog, p-hydroxybenzaldehyde (p-HBA). We have revealed that Loop 5 (located between β5 and α5) and Loop 6 (located between β6 and α6) of CYE and TYE play pivotal roles in determining their substrate preferences and demonstrated that some artificial OYEs with mutation(s) in these loops can be applicable for the asymmetric reduction of various enone compounds.

Binding Model of (4S)-phorenol to CYE based on the crystal structures of CYE-p-HBA

To gain insights into the different substrate preferences between CYE, the binding model of (4S)-phorenol to CYE was built based on the crystal structures of the CYE-p-HBA complex (Figure 2). The binding model of (4S)-phorenol to CYE implies that the bulky dimethyl group at C6 of (4S)-phorenol would not be favored by CYE because the dimethyl group in (4S)-phorenol, and Phe250 (Loop 5) and Pro295 (Loop 6) in CYE would collide.

The Pro295 in Loop 6 Plays Pivotal Roles in Substrate Preferences

. The putative binding model of (4S)-phorenol to CYE suggests that the dimethyl group at C6 may collide with Pro295 in Loop 6 and Phe250 on Loop 5 (Figure 2). To examine the effect of Pro295 on the substrate preference of CYE, CYE (P295G) was constructed to avoid the possible collision between the dimethyl group in (4S)-phorenol and Pro295 in CYE. Furthermore, to investigate the effect of Phe296 on Loop 6 on the substrate preference of CYE, mutants of CYE (F296X) and CYE (P295G/F296X), where X represents one of 20 amino acid residues, were also constructed, and a series of mutational analyses were performed. However, no remarkable tendency was toward ketoisophorone and (4S)-phorenol. Each point mutant of CYE (P295G/F296X) exhibits an enhanced catalytic activity toward both ketoisophorone and (4S)-phorenol compared to the corresponding point mutant of CYE (F296X), demonstrating that Pro295 on Loop 6 of CYE does affect the catalytic activity of CYE toward ketoisophorone and (4S)-phorenol, probably via a collision with the dimethyl group of ketoisophorone and (4S)-phorenol. Thus, Pro295 on Loop 6 of CYE acts as a substrate filter. Phe296 does not have a significant effect on the catalytic activity when Pro295 is replaced with Gly295.

To investigate further, a series of chimeric mutants of CYE were constructed and mutational analyses were performed. The mutants used were named as follows: CYE (CYE Loop 6→Glyn) means the CYE mutant where the Loop 6 of CYE (289-EPRVTDPFLPEFEKWFKEGT-308) is replaced with a poly-glycine linker Glyn, where n = 2-5. When Loop 6 of CYE is replaced with a poly-glycine linker Glyn, where n = 2-5, CYE (CYE Loop 6→Gly-Gly), CYE (CYE Loop 6→Gly-Gly-Gly), CYE (CYE Loop 6→Gly-Gly-Gly-Gly) and CYE (CYE Loop 6→Gly-Gly-Gly-Gly-Gly) exhibit higher catalytic activity toward (4S)-phorenol (Figure 3). These results support the hypothesis that the Loop 6 plays a significant role in substrate preferences.

The Gly250 in Loop 5 Plays Pivotal Roles in Substrate Preferences

The binding model of (4S)-phorenol to CYE implies that the bulky dimethyl group at C6 of (4S)-phorenol would not be favored by CYE because the dimethyl group in (4S)-phorenol, and Phe250 in CYE would collide in addition to Pro295 in Loop 6. To examine the effect of Phe250 on the substrate preferences of CYE, CYE (F250G), CYE (F250A) mutants were constructed to avoid the possible collision between dimethyl group in (4S)-phorenol and Phe250 in CYE. CYE (F250G) and CYE (F250A), which would have larger substrate binding pockets than CYE (WT), show increased catalytic activity toward (4S)-phorenol (Figure 3), indicating that the substrate binding pocket of CYE (WT) would hardly accommodate (4S)-phorenol. Mutations to Loop 5 and Loop 6, CYE (F250A/P295G) and CYE (F250G/P295G/F296G), have also been constructed, and catalytic activities were assayed toward 2-cyclohexene-1-one, ketoisophorone and (4S)-phorenol, and the catalytic data were shown in Figure 3. The data indicates that the CYE (P295G) mutant shows the most increased catalytic activities among them.

The Grafting Mutant of Loop 6 Showed Highest Catalytic Activity.

No electron density of Loop 6 in TYE was observed (Figure 4). To investigate the role of Loop 6 of TYE, grafting mutant, CYE (CYE Loop 6→TYE Loop 6) mutant where the Loop 6 of CYE (289-EPRVTDPFLPEFEKWFKEGT-308) is replaced with the Loop 6 of TYE (286-EPRVNGIADAPENSED-301), was constructed. CYE (CYE Loop 6→TYE Loop 6) mutant showed the highest catalytic activity compared to other CYE mutants (Figure 5), and this indicates that Loop 6 plays pivotal roles in substrate preferences and catalytic efficiencies.

Conclusions

In this study, we have designed and constructed mutants in Loops 5 and 6 of CYE and TYE based on their crystal structures, and succeeded in obtaining a CYE mutant, CYE (CYE Loop6 → TYE Loop 6), which shows remarkably higher catalytic activities than the wild-type CYE (3- and 300-fold toward ketoisophorone and (4S)-phorenol, respectively). Although the relative activities of this CYE mutant are still lower than those of the wild-type TYE (0.6- and 0.5-fold toward ketoisophorone and (4S)-phorenol, respectively), the present mutational data demonstrate the importance of Loops 5 and 6 in determining the substrate preferences of OYEs for the first time and suggest that mutations in these loops could improve the catalytic efficiencies of OYEs and/or confer a new substrate preference to OYEs.

Figure 1. Synthesis of (4R,6R)-actinol from ketoisophorone

Figure 2. Left: crystal structure of CYE-p-HBA, Right: binding model of CYE-(4S)-phorenol.

Figure 3. Specific activity of CYE mutants toward 2-cyclo-hexene-1-one, ketoisophorone and (4S)-phorenol.

Figure 4. The electron density of Loop 6 was not observed.

Figure 5. Specific activity of CYE and TYE mutants toward 2-cyclohexene-1-one, ketoisophorone and (4S)-phorenol.

本論文は、基質特異性の異なる2種類の旧黄色酵素CYEとTYEのX線結晶構造解析、ならびに得られた結晶構造に基づいて作製された変異体の活性測定の結果について述べている。CYEとTYEはそれぞれ、Candida macedoniensis AKU4588とPichia sp. AKU4542に由来するNADPH依存性の脱水素酵素であり、エノン化合物を立体選択的に不斉還元する。本研究では、2種類の旧黄色酵素CYEとTYEの構造生物学的解析と生化学的解析を行った結果、基質選択性と触媒効率に関与するループ領域を明らかにし、両酵素の工業的応用に向けた展望を示している。本論文は7章からなる。

第1章は序論で、まずキラル化合物の有用性について説明し、酵素法を用いたキラル化合物の不斉合成技術を紹介している。次に、CYEとTYEに関する現在までの知見を説明している。CYEとTYEはそれぞれ403および405アミノ酸残基から成り、KIPから(4R,6R)-actinolを合成する過程で単離された酵素である。KIPから(4R,6R)-actinolのワンポット合成において、CYEを用いた場合、反応中間産物である(4S)-phorenol がある一定の割合で蓄積するが、TYEを用いた場合、(4S)-phorenol が蓄積しないことから、CYEとTYEの基質特異性の違いについて述べている。CYEとTYEの基質特異性の違いが、KIPから(4R,6R)-actinolの合成にどのような影響を与えるのかを説明した上で、基質特異性の異なる2種類の旧黄色酵素の構造基盤解明の重要性が説明されている。

第2章から第4章では、CYEとCYE-p-HBAのX線結晶構造解析および構造比較について述べられている。構造比較の結果、Loop 6とよばれるループ領域において大きな構造変化があることが述べられており、Loop 6が運動性を有した活性部位の蓋として働くことを記述している。次に、論文中では、CYE-p-HBA複合体構造からCYE-(4S)-phorenolの構造モデルを構築している。その結果、(4S)-phorenolのジメチル基がPro295、Phe296残基と近接することを示し、Pro295残基をGly295残基に置換することによって、KIPへの活性を2.4倍、(4S)-phorenolに対する活性を190倍上昇させることに成功したことを述べている。Pro295残基をGly295残基に置換することにより活性が上昇した要因として、Loop 6が活性部位の蓋として働き、還元反応が進行するための適した閉構造をとることで酵素活性が上昇することを示唆している。また活性部位の蓋が失われた場合、酵素活性に対する影響を検証するため、Loop 6を欠損させた変異体の作製を試みている。その結果、Loop 6を欠損させ閉構造を取れなくしたような変異体であってもある程度の活性は保持されることを述べている。第2章から第4章の解析を総括すると、Loop 6が閉構造をとれなくても反応は進行するが、Loop 6が閉構造をとることによって、より効率的に反応が進むことを示唆している。

第5章から第6章では、TYEのX線結晶構造解析およびCYEとTYEの構造比較について述べられている。CYEとTYEの異なる基質特異性の構造基盤を解明するため、CYEとTYEの結晶構造からLoop 6を置換して活性測定を行いTYEのLoop 6の役割について考察するという戦略をとっている。その結果、KIPへの活性を3.0倍、(4S)-phorenolに対する活性を300倍上昇させることに成功しており、TYEにおいてもLoop 6が酵素活性に非常に重要な役割を果たしていることを示している。また、CYEとTYEの活性部位の比較からLoop 5に存在するPhe残基の位置が異なっており、この位置の違いが酵素活性に影響を与えていることを示している。

第7章では、CYEとTYEのX線結晶構造解析と変異体の活性測定の結果について総合考察が述べられている。まず、CYEとTYEの異なる基質特異性を有する構造基盤の解明によって、Loop 5とLoop 6が重要であること述べている。その結果を踏まえ、Loop 5やLoop 6に注目し改変体の創製を試みている。現在までにTYEを超える活性を有する変異酵素は得られていないものの、今後さらなる高活性化に向けた展望を示している。

以上、本研究はCYEとTYEの異なる基質特異性を有する構造基盤を解明しただけでなく、両酵素を用いた工業応用も見据えて研究を行っており学術上、応用上貢献するところが少なくない。よって、審査委員一同は、本論文が博士(農学)の学位論文として価値あるものと判断した。