发光生物在自然界中的分布十分广泛，它们可以通过体内的化学反应产生生物发光。由于生物发光具有成本低廉、高灵敏性、高信噪比和非侵入性等优点，目前已被广泛应用于生物医学、生物成像以及食品安全等领域。大多数发光生物通过一套荧光素-荧光素酶体系或光蛋白体系发光，然而有些发光生物具有两套发光体系，第一套体系产生的能量并未转化为可见光，而是转移到第二套体系并导致第二套体系发光。这种能量转移过程涉及蛋白-蛋白相互作用，其相互作用时间很短，实验上很难得到相互作用位点的信息以及蛋白-蛋白复合物的晶体结构，所以其微观机制从未被解释清楚。在分子和电子态水平上清晰理解这种微观机制的意义不仅仅有助于对生物发光现象的理解，更重要的是能够找到调节发光波长和强度的手段，从而得到理想的具有实用价值的生物发光成像系统。本论文选择了在能量转移发生后，发光波长发生蓝移（明亮发光杆菌）和红移（水母和海肾）的两种生物发光体系来研究其微观机制。

阐明这种能量转移过程的机制，我们必须清晰回答以下三个科学问题：能量的来源是什么？能量转移的详细过程是什么？最终发光的机制是什么？其中，对能量转移过程的研究是最复杂也是最重要的。在明亮发光杆菌生物发光中，能量来源于第一套发光体系细菌荧光素酶中直链型过氧化物解离生成的激发态4a-羟基黄素（HFOH^*）。但是该过氧化物的解离过程非常复杂，目前文献提出了七种可能的解离机制。因此，我们首先要详细计算该过氧化物的解离，结合文献提出的各种机制，真正搞清明亮发光杆菌发光的能量来源。这是本文的研究内容之一。明亮发光杆菌比其他发光细菌的发光波长蓝移约20 nm且发光强度增强。实验证实明亮发光杆菌是通过其体内存在第二套发光体系——鲁马嗪蛋白和6,7-二甲基-8-核糖鲁马嗪（DLZ）体系发光的，发光的能量来源是HFOH^*。研究这种发光波长蓝移现象所涉及的能量转移机制以及解释最后发光强度增强的根本原因，是本文的研究内容之二。水母和海肾生物发光中的能量来源在我们课题组之前发表的文章中已经被清晰阐述。其能量来源于第一套发光体系水母光蛋白和海肾荧光素酶中腔肠素二氧环丁酮的解离生成的S₁态的氧化腔肠素（S₁-CTD）。对这两个体系发光的能量来源问题，本论文不再做重复研究。水母和海肾的最大发光波长均为509 nm，实验研究发现是由于在它们体内存在第二套发光体系——绿色荧光蛋白体系，能够分别与水母光蛋白和海肾荧光素酶形成蛋白-蛋白复合物，并发生从CTD向绿色荧光蛋白发色团的能量转移过程，导致其最大发光波长分别从470 nm和480 nm红移至509 nm。实验还发现对于海肾来说，不仅发光波长会红移，生物发光的量子产率也会增加，研究这种发光波长红移现象所涉及的能量转移机制以及解释最后生物发光量子产率增加的根本原因，是本文的研究内容之三。综上所述，本论文综合使用了量子力学（QM）、分子动力学（MD）模拟以及组合量子力学和分子力学（QM/MM）等多种手段和方法，分别对以上三个研究内容进行了理论研究和计算。论文解决的主要问题和主要创新点如下：

（1）首次全面、详细地考虑了直链型过氧化物解离生成生物发光体S₁-HFOH的七种机制，采用（含时）开壳层密度泛函理论、自旋翻转密度泛函理论以及完全活化空间二阶微扰理论，基于合理的简化模型，详细描述了解离过程中涉及到的电荷转移、势能面交叉等信息，最后发现该解离过程遵循电荷转移诱导发光（CTIL）机制，还将直链型过氧化物的解离机制与另外两种过氧化物（1,2-二氧环丁酮和环内桥接型过氧化物）的解离机制进行了比较，我们得出电荷转移和电荷回迁的发生与化学键的断裂相关，且电荷转移过程可以使势能面之间的非绝热跃迁更有可能发生。

（2）首次在真实蛋白环境中研究了在明亮发光杆菌中从HFOH^*到DLZ的能量转移过程，采用蛋白-蛋白对接、MD模拟、高精度的QM/MM等方法，确认了在荧光素酶-鲁马嗪蛋白复合物中的能量转移过程遵循Förster共振能量转移（FRET）机制，计算得到的能量转移速率显著快于能量供体的辐射速率和非辐射速率，意味着能量转移过程能够高效发生。S₁-DLZ的荧光量子产率高于S₁-HFOH的荧光量子产率，解释了实验中观察到的加入鲁马嗪蛋白后生物发光强度变强这一现象的原因。最后，我们将发光波长的蓝移的根本原因归结于：能量供体强的电子-振动耦合和能量受体小的斯托克斯位移。

（3）首次在真实蛋白环境中研究了水母和海肾中从氧化腔肠素到绿色荧光蛋白的能量转移过程。采用蛋白-蛋白对接、MD模拟、高精度的QM/MM等方法，确认了在水母光蛋白-水母绿色荧光蛋白和海肾荧光素酶-海肾绿色荧光蛋白复合物中的能量转移过程遵循FRET机制。计算得到的能量转移速率显著快于能量供体的辐射速率和非辐射速率，意味着能量转移过程能够高效发生。荧光量子产率的计算印证了加入海肾绿色荧光蛋白后生物发光量子产率增加这一现象的原因。

Luminescent organisms are widely distributed in nature. They can produce bioluminescence (BL) through chemical reactions in their body. Due to its low cost, high sensitivity, signal-noise-ratio and non-invasive properties, BL has a wide range of applications in biomedical, biological imaging and food safety fields. Most luminescent organisms emit light through luciferin-luciferase or photoprotein systems. Some luminescent organisms have two sets of luminescent systems, and the energy generated by the first system is not converted into visible light, but transferred to the second system, thus resulting in luminescence of the second system. This energy transfer (ET) process involves protein-protein interactions, which occur over very short timescales, making it difficult to obtain the interaction residues and crystal structures of protein-protein complexes in experiment. Therefore, the microscopic mechanism of this ET process has never been fully understood. Clearly understanding the ET mechanism at the molecular and electronic levels is not only beneficial for understanding the phenomenon of BL but also crucial for finding ways to modulate the wavelength and intensity of BL, thereby obtaining ideal and practical BL imaging systems. This dissertation focuses on studying the mechanisms of two bioluminescent systems where the BL wavelength shifts towards blue (Photobacterium phosphoreum, PP) and red [jellyfish Aequorea victoria (AV) and sea pansy Renilla reniformis (RR)] after the ET process.

To elucidate the mechanism of the ET process, we must clearly answer the following three scientific questions: What is the source of energy? What is the detail of ET process? What is the mechanism of final luminescence? Among them, the study of ET process is the most complex and important. In PP BL, the energy originates from the excited-4a-hydroxyflavin (HFOH^*) generated by the dissociation of linear peroxide in luciferase (PPLuc) of the first luminescent system. However, the dissociation process of this linear peroxide is very complex. Currently, seven possible mechanisms have been proposed in the literature. Therefore, combined with the various mechanisms proposed in the literature, we need to calculate the dissociation mechanism of the linear peroxide in detail at first, thus leading to the fully understanding of the energy source of PP BL. This is first research content in this dissertation. Compared to other luminescent bacteria, PP emits a blue shift of about 20 nm in BL wavelength and an enhanced BL intensity. Experimental studies have shown that PP emits light through the second luminescent system, the lumazine protein (PPLumP) and 6,7-dimethyl-8-ribityllumazine (DLZ) system. The energy source of PP BL is HFOH^*. The second research content in this dissertation is to investigate the ET mechanism involved in the blue-shifted phenomenon of the emission wavelength and explain the fundamental reason for the enhancement of the final emission intensity. The energy sources of AV and RR BL have been clearly explained in previous articles published by our research group. The energy originates from S₁-state-coelenteramide (S₁-CTD) generated by the dissociation of coelenterazine dioxetane (CDO) in the first luminescent system, aequorin (AVaequorin) and luciferase (RRLuc). This dissertation will not repeat the research on the energy sources of AV and RR BL. The maximum BL wavelengths of AV and RR are both 509 nm. Experimental studies have shown that this is due to the presence of a second luminescent system, the green fluorescent protein (GFP) system, which can form a protein-protein complex with AVaequorin (AVaequorin-AVGFP) and RRLuc (RRLuc-RRGFP), respectively. In AVaequorin-AVGFP and RRLuc-RRGFP, an ET process from CTD to the chromophore of GFP takes place, resulting in a red-shifted BL wavelength from 470 nm and 480 nm to 509 nm. Furthermore, experiments have shown that for RR BL, not only is the maximum BL wavelength red-shifted, but the quantum yield of BL also increases. We need to investigate the ET mechanism involved in this red-shifted BL and explain the fundamental reasons for the increase in quantum yield of BL. This is third research content in this dissertation. Based on the discussion above, this dissertation comprehensively employs quantum mechanics (QM), molecular dynamics (MD) simulations, as well as combined quantum mechanics and molecular mechanics (QM/MM) methods to theoretically investigate and compute the three aforementioned research content. The main problems and innovative points in this dissertation are as follows:

(1) For the first time, the seven proposed mechanisms for the dissociation process of linear peroxide to generate S₁-HFOH were comprehensive and detailed studied. The open shell (time-dependent) density functional theory [(TD) DFT], spin-flip density functional theory (SF-DFT), and complete active space second-order perturbation theory (CASPT2) are used. Based on a reasonable simplified model, we made detailed descriptions of charge transfer process, potential energy surface crossings and so on in the dissociation process. It was found that the dissociation process of linear peroxide follows the charge transfer induced luminescence (CTIL) mechanism. Additionally, a comparison was made between the decomposition mechanism of linear peroxide and the other two peroxides (1,2-dioxetanone and endoperoxide). By comparison, the occurrence of charge transfer (CT) and back charge transfer (BCT) is related to the breaking of chemical bonds and the CT process can make nonadiabatic transitions likely to occur.

(2) For the first time, the ET process from HFOH to DLZ in PP BL was studied in real protein environment. Protein-protein docking, MD simulation, and high-precision QM/MM methods were used to confirm that the ET process in PPLuc-PPLumP complex follows the Förster resonance energy transfer (FRET) mechanism. The calculated FRET rate is much larger than the radiative and non-radiative rates of the donor, indicating that the FRET process can occur efficiently. The fluorescence quantum yield of S₁-DLZ is higher than that of S₁-HFOH, which clearly explains the experimentally observed enhancement of the emission intensity. Finally, we attribute the fundamental reason for the blue shift of the emission wavelength to the strong vibronic coupling of the donor and the small Stokes shift of the acceptor

(2) For the first time, the ET process from CTD to GFP in AV and RR BL was studied in real protein environment. Protein-protein docking, MD simulation, and high-precision QM/MM methods were used to confirm that the ET process in AVaequrorin-AVGFP and RRLuc-RRGFP complexes follow the FRET mechanism. The calculated FRET rate is much larger than the radiative and non-radiative rates of the donor, indicating that the FRET process can occur efficiently. The calculation of fluorescence quantum yield confirms the reason for the increase in bioluminescence quantum yield after adding RRGFP.