Bioluminescence of fire y is a photon emission process accompanying oxidation of the substrate luciferin in the enzyme luciferase (Luc) with cofactors ATP and magnesium ion. Its color can be shifted from the natural yellow-green to red in in vitro reaction by decrease in the pH value of the buffer solutions, addition of transition metal ions, or mutagenesis to the luciferase.

The color-change mechanism of fire y bioluminescence was explained as equilibrium or competition between two species or states of the light emitter oxyluciferin (OL). However, Ando et al. (Ando et al. Nat. photonics 2(1), 44-47, 2007) found that the mere intensity change in a green peak determined the spectral shape and peak position by studying the pH-sensitive quantitative spectra recently. The first objective of this dissertation is to investigate the in uence of the micro-environmental change on the green, orange, and red emission components, and thus to further verify the feasibility of Ando and coworkers' interpretation. I measured the quantitative spectra of fire y bioluminescence under circumstances modified by addition of transition metal ions or mutagenic Lucs. Then I carried out curve fitting with Gaussian-type functions to quantitatively analyze how the reaction environment affect each emission component. Meanwhile, the spectroscopic properties of free OL in various solvents with different acidity have been clarified recently, but those of OL in Luc environment (OL-Luc complex) still remain to be studied. The second objective is to understand the in uence of Luc environment on the emission of OL. I studied the spectroscopic properties of OL in the active pocket of pH-sensitive and pH-insensitive Lucs and revealed the real form it takes in Luc.

I used a homemade setup (Figure 1) for total-photon- ux spectrum measurement to quantitatively study the color change of firefly bioluminescence following the method of Ando and coworkers' (Ando et al. Photochem. Photobiol. 83(5), 1205-10, 2007). I calibrated the light-collection efficiency of the sample cell, the transmissivity of the slit, and the absolute sensitivity of the system to convert relative spectrum to absolute one scaled by photon number. I constructed setups for in situ absorption monitoring and uorescence spectrum measurement to characterize the stability of OL in Luc formed in bioluminescent reaction and to study its spectroscopic properties. The setups and materials used in the measurement are stated in Chapter 2.

First I studied the quantitative spectra of Photinus pyralis (North-American fire y) bioluminescence in the presence of various transition metal ions at pH 8.0. Figure 2 shows the bioluminescence spectra in the presence of different concentrations of zinc ions. The intensity of the green peak ( 560 nm) was sensitive to the concentration of zinc ions, while that of the red peak (620 nm) seemed to be unchanged. I analyzed the spectra by decomposing them into three Gaussian-type functions (shown in the inset of Figure 2): green (~ 2:2 eV), orange (~ 2:0 eV), and red (~ 1:9 eV) Gaussian components. The peak energy and full-width-at-half-maximum (FWHM) value of each component are independent of the concentration of zinc ions. Only the intensity of the green component is sensitive to the concentration of zinc ion, while those of the other two are not. Similar color change was observed by adding Cd(2+), Ni(2+), Co(2+), Fe(2+), or Hg(2+) ions, while concentrated Mg(2+), Ca(2+), or Mn(2+) ions caused no shift in spectral peak or quantum yield. According to the concentration of the metal ion required to produce identical spectral change, the sensitivity of quantum yield and spectrum to these metal ions were ordered as Hg(2+) > Zn(2+); Cd(2+) > Ni(2+); Co(2+); Fe(2+) >> Mg(2+); Mn(2+); Ca(2+). The order had no correlation with the electron-withdrawing strength of these metal ions. It is likely that the metal ions bind to Luc and the modified Luc affects bioluminescent emission of OL. The strong difference and the peculiar order in spectral sensitivities for the various metal ions may stem from their special binding affinity to different amino acid residues in luciferase. The experimental results mentioned above and discussions on possible bindings between the metal ion and protein are presented in Chapter 3.

To understand the color change caused by microenvironmental change in the Luc, I studied the spectra of bioluminescence catalyzed by mutagenic Luciola cruciata (Genji-botaru) Luc in which the amino acid residue tyrosine (Y) 257 was substituted by phenylalanine (F), alanine (A), glutamate (E),or arginine (R) in Chapter 4. The representative spectra of Y257F and Y257E are shown in Figure 3. Curve fitting with Gaussian-type functions were employed to quantitatively analyze their bioluminescence spectra that exhibited sensitivity or insensitivity to change in the pH value of solutions. For the pHsensitive Luc (wild type and Y257F), intensity change in the green region was observed accompanying the spectral peak shift. Their spectra were reproduced by three Gaussian components, similar to the results in Figure 2. Change in the intensities of the green Gaussian components at various pH values determined the peak positions and shapes of the total bioluminescence spectra. In addition, the maximum intensity of the green component in the spectra of Y257F was weaker than that in wild-type Luc, directly correlating with the less pH-sensitivity of the bioluminescence spectra of Y257F. In contrast, the spectra of the pH-insensitive bioluminescence were well reproduced by only two components, the orange and red ones. As an example, the spectra catalyzed by Y257E are shown in Figure 3 (b).The green component was not required in fitting the bioluminescence spectra catalyzed by Y257A/E/R, and hence resulted in no/minor change in the spectral intensities or peak positions. The FWHM and peak energies of the three Gaussian components were close for all the Lucs, although the strong intensity and blue shift of the orange component of Y257R were exceptions. The sensitivity of each component to the pH value was unchanged: the green component was pH-sensitive, and the orange and red ones were still pH-insensitive. The change in the intensity of the green component mainly determined the spectral shape and peak position. Therefore, modification to the Luc by mutagenesis in uences the intensity of the green emission component, and thus regulates the pH-sensitivity of the bioluminescence. The correlation between the color change and the properties of the substituted amino acid residues are also discussed in this chapter.

Chapter 5 aimed at characterizing the properties of the light emitter OL in Luc environment and thus revealing the real forms of OL in the active pocket of the enzyme. I demonstrated the stability of OL in Luc formed in bioluminescent reaction via monitoring the in situ absorption spectra. OL in complex with Luc was stable for about an hour under aerobic basic conditions, and the stable period was longer in acid solution. This enabled the studies on the in situ spectroscopic properties of OL.

For OL in complex with pH-sensitive Photinus pyralis Luc (Figure 4), a peak with λ(abs) = 380 nm is dominant in the absorption spectra. The trade-off in its intensity with that of a weak shoulder λ(abs) ~ 430 nm reveals equilibrium between two species of OL. Excitation at 380 nm generated a blue (λ(fl)= 450 nm) and green (λ(fl) = 560 nm) uorescence bands with comparable intensities (Figure 4 (a)). Excitation at 430 nm generated a single green uorescence band with peaks at 560 nm (Figure 4 (b)). Very weak red uorescence ~ 610 nm was observed with excitation at 510 nm, although no corresponding absorption peaks was detected.

According to the spectroscopic properties of free OL and NMR data, the three emitters were assigned to neutral enol-OL (blue), monoanionic phenolate-enol-OL- (green), and monoanionic phenolate-keto-OL- (red), respectively. The strong intensity of the green uorescence with excitation at 380 nm (Figure 4 (a)) is not proportional to the weak absorbance at 430 nm, which ruled out the possible direct excitation of the emitter corresponding to λ(abs) ~ 430 nm. A possible reason is that partial excited-state proton transfer occurred to the neutral OL generated the green-emitting phenolate-enol-OL-. The absorbance at 430 nm was not proportional to the intensity of corresponding green uorescence shown in Figure 4 (b), which indicated that its uorescence efficiency depended on pH value of the solution. The absorption and uorescence spectra reveal that OL mainly exists as neutral enol form in the active pocket of Luc after bioluminescent emission. A minority of greenemitting phenolate-enol-OL- displays equilibrium with enol-OL, which is moderately regulated by change in pH value. The red emitter phenolateketo- OL- is absent in the ground state, reason of which may be its transition to enol-OL that has lower free energy.

The spectroscopic properties of OL in a redemitting H433Y mutant of Luciola cruciata Luc were studied and compared with those in the wild-type one in Chapter 6. Both the intensities and peak positions of the absorption and uorescence spectra of OL in complex with the pH-insensitive red H433Y mutant were very similar to those of pH-sensitive wildtype Luc. The similarities indicate that the difference in the environment of wild-type and H433Y mutant Lucs does not significantly change the spectroscopic properties of OL in the ground state. They also indicate that the total concentrations of all species of final products of the green- and red-bioluminescent reaction are similar.

The quantitative spectra and analyses show that the intensity of the green bioluminescent emission component is sensitive to the changes in reaction environment and thus determines the spectral shape and peak energy, verifying the assumption of Ando and coworkers'. This suggests the necessity of explaining the intensity change in thegreen peak to understand the color-change mechanism of fire y bioluminescence. S tudies on the in situ spectroscopic properties of OL clarified the real forms of OL in the Luc and enabled assignment of bioluminescent emitters. Similarities between its properties in pHsensitive and -insensitive Luc reveal that different Luc environment causes little change in the properties of OL in the ground state, but affects the excited state generated via chemical reaction, indicating the importance of studying the intermediate state of the reaction to explain the color-change mechanism.

Figure 1: Scheme of the setup for quantitative spectrum measurement.

Figure 2: Quantitative spectra of bioluminescencein the presence of 5 mM of Mg(2+) ions and Zn(2+) ions with different final concentrations at pH 8.0. The insets show the results of Gaussian fitting and each component.

Figure 3: Quantitative spectra of bioluminescence catalyzed by (a) Y257F and (b) Y257E mutants of Genji-botaru luciferase at various pH values. The insets show the results of Gaussian fitting and each component.

Figure 4: In situ absorption (left axis) and uorescence (right axis) spectra of oxyluciferin in Genjibotaru luciferase formed in bioluminescent reaction at various pH values with excitation at (a) 380 nm and (b) 430 nm.

本論文は,ホタルにおける生物発光色の変化を定量的な「その場分光」の手法によって研究したものであり,7章から構成されている.

第1章ではホタルが示す生物発光に関わる化学反応,発光効率,環境(pHや溶媒)による発光色の変化などに関する過去の実験研究および発光色変化を説明するモデルについて紹介している.この研究は,低いpH(酸性環境)で発光色が赤になる原因は緑発光の強度の減少によるという最近の発見に基づいて,様々の条件での発光色変化が同様に理解できるかという問題提起から始まっており,発光色の変化を通じて,発光のメカニズムを解明することを最終目標にすることが述べられている.

第2章では実験手法と試料の調製方法が述べられている.試料は,生物発光に必要なLiciferin,2価金属イオン,ATPおよび発光を起こさせる酵素Luciferaseを緩衝液に混合したものである,発光色は,酵素の種類によって異なるが,ここでは北米ホタル,ゲンジボタル由来の酵素,および後者の変異体であるY257F/E/R,H433Yを用いている,生物発光はATPをトリガーとして注入することにより開始され,発光スペクトルは,反応終了までの総発光量を定量することにより,反応分子あたりの発光絶対量として定量的に求められることを述べている.

第3章では各種2価イオンの添加による発光色の変化が調べられている,その結果,Zn(2+),Cd(2+)の添加によって変化するのは主に緑発光であり,赤発光の変化は僅かであることを見出した.さらに緑発光の減少への影響は,大きい方からHg,(Zn,Cd),(Ni,Co,Fe),(Mg,Mn,Ca)の順であることを見出した.

第4章ではゲンジボタルの野生種と,257番目のアミノ酸残基を置換した4種類の変異体(Y257F/A/E/R)について発光スペクトルを測定し,やはり緑発光のみが減少または消滅することを見出している.

第5章では生物発光の反応が進行している途中の光吸収スペクトルを逐次測定し,総発光量と吸収強度の間によい相関があること,反応開始から約1時間後まで反応生成物は安定に残存することを確認し,反応終了後の溶液について吸収と光励起発光の詳細な測定を行った.光吸収スペクトルのpH依存性は裸の酸化Luciferin(OL)より小さいことを見出し,これは発光分子が収まっている活性ポケットの疎水性のためであると解釈された.光励起発光によっても緑発光が観測され,生物発光の場合と同様に酸性環境で減少することから,発光分子は生物発光時とほぼ同じ状態で活性ポケット内にとどまっていることがわかった.赤発光は非常に微弱で,赤発光を示す分子はほとんど残留していないことも示された.

第6章では赤発光を示す変異体(H433Y)について,吸収,生物発光,光励起発光スペクトルについて論じている.発光スペクトルのpH依存性は小さかったが,この試料は元々緑発光を示さないので,自然な結果といえる.吸収スペクトルもpH依存性が小さいが,その一方で,光励起発光では北米ホタルと同様にpH依存性を持つ緑発光が見られた,これらの結果および溶液中分子の吸収スペクトルとの比較などから,緑発光の起源は活性ポケット内のpheonolate-enol-OL-であると同定された,これは,赤発光変異体の場合でも緑発光分子そのものは存在していることを示しており,したがって緑発光を示さないのは,その始状態が化学的に生成されないか,もしくは励起状態が無輻射的に緩和してしまうためと結論された.

第7章には以上のまとめと今後の課題が書かれている.緑発光は,pHの低下,Zn(2+)などの2価イオンの添加,アミノ酸残基置換などによって減少し,それによって生物発光色が変化することが示された.これらの変化はいずれも活性ポケットの形状や大きさの変化をとおして,緑発光の始状態に作用するためと理解された.

この研究は,所属研究室で開発された発光量子効率を定量できる測定装置と技術を最大限に利用し,2価イオンの添加をはじめいくつかの条件による発光色の変化を極めて精密に定量的に研究したものである.国内外に同種の研究は知られておらず,独創性の高いものである,すべてのケースにおいて緑発光の増減が発光色の変化の起源であることを示した意義は大きい,また,光励起発光測定から最終生成物に大きな違いがないことがわかり,緑発光の強度を支配している要因は化学過程による励起状態の生成またはその緩和にあることも明らかになった,これは生物発光の研究における非常に重要な知見であり,さらなる生物発光の解明に結びっく大きな貢献をなしたものと認められる.

本研究のテーマには,複数の共同研究者が関与しているが,実験,解析および解釈にいたるまで,本人がほぼ独力で遂行したものであり,全体として申請者が主導的に研究を進めたものと認められる.

以上の理由により,提出された論文は,博士(理学)の学位を授与するにふさわしいものであると,審査委員全員の一致によって判断した.