【作者】 闫秀芬；

【导师】 肖鹤鸣；

【作者基本信息】 南京理工大学 ， 材料物理与化学， 2007， 博士

【摘要】 硝基芳烃是一类重要的污染物，应用广泛且在环境中大量存在，对生态环境造成了严重污染，威胁到人们的生产、生活和身体健康，因此对硝基芳烃的毒性进行系统研究具有重要意义。定量结构-活性关系（Quantitative Structure-Activity Relationship，QSAR）是一种非常重要的毒性研究方法，已广泛用于硝基芳烃类化合物的毒性研究。目前，在建立硝基芳烃毒性QSAR模型过程中，普遍存在如何获得分子描述符的问题，先前的研究主要通过实验或采用半经验分子轨道计算获得，前者费时费力且误差较大，后者属于近似研究，结果欠准确。随着计算机和量子化学理论的飞速发展，考虑了电子相关的第一性原理方法密度泛函理论（Density Functional Theory，DFT）得以广泛应用，其计算结果较为精确，又比较节省机时，但迄今为止，尚未见有将DFT方法应用于硝基芳烃毒性研究的文献报道。此外，文献研究表明，2，4，6-三硝基甲苯（TNT）及其代谢产物可与血红蛋白（Hemoglobin，Hb）发生共价结合，且Hb加合物水平与动物的染毒剂量相关。其它硝基芳烃也有类似报道。但是至今亦未见有与此相关的动态理论计算工作报道。对硝基芳烃衍生物的致毒机理有待深入研究。在此背景下，本论文围绕硝基芳烃衍生物的毒性进行了系统的理论计算研究，全部工作主要分为三章。1．以硝基芳烃对黑呆头鱼的急性毒性研究为例，运用不同理论水平的量子化学计算方法，包括半经验AM1和PM3方法，从头算HF／6-31G^*和DFT-B3LYP／6-311G^**方法等计算28种硝基芳烃化合物的分子几何和电子结构，获得它们的分子结构描述符，结合硝基芳烃对黑呆头鱼的急性毒性实验值，建立了相应的QSAR模型。结果表明，基于DFT-B3LYP方法计算所得QSAR模型稳定性和显著性明显优于AM1和PM3方法；相比于HF方法，B3LYP模型能对硝基芳烃的致毒机理作出更为合理的解释。故就本例而言，B3LYP方法更适合于硝基芳烃急性毒性的QSAR研究。2．运用DFT-B3LYP方法研究硝基芳烃的分子结构与其毒性之间的定量关系（QSAR），寻找影响其毒性大小的关键结构因素，对同系物的已有毒性进行阐明、对未知毒性进行预测，并为硝基芳烃的毒性降解处理提供有意义的建议。采用DFT-B3LYP方法，在6-311G^**基组水平下，对单（多）硝基苯、单（多）硝基甲苯、硝基苯胺、卤代硝基苯等多系列的硝基芳烃化合物进行分子几何全优化计算，获得相关量子化学参数作为分子描述符，结合硝基芳烃对梨形四膜虫、圆腹雅罗鱼和斜生栅列藻等不同物种的急性毒性实验数据，分类建立了毒性QSAR模型，讨论了硝基、氨基和卤素等取代基对化合物毒性的影响。结果表明，硝基是硝基芳烃衍生物的主要致毒基团，硝基数越多，化合物的毒性越大。单硝基取代芳烃和多硝基取代芳烃具有不同的致毒机理，前者为极性麻醉型毒性化合物，其毒性大小由疏水性和电子反应性共同决定，毒性作用主要为对细胞膜的穿透刺激作用；后者为反应性毒性化合物，其毒性大小主要由电子反应性决定，分子可能发生单电子还原或亲核取代等反应而致毒。苯环上的其它取代基视其电子特性不同对硝基芳烃的毒性影响也不同，给电子基团如氨基会削弱硝基芳烃的毒性；吸电子基团如卤素则会增强化合物的毒性。故采用合适的包覆剂将硝基包覆或用微生物对其进行降解应该是治理硝基芳烃类化合物污染的有效途径。文中给出的稳健QSAR模型可用于相应化合物对不同物种急性毒性的预测。3．在前两章研究的基础上，以乙硫醇（C₂H₅SH）作为含还原性巯基的生物蛋白大分子（R-SH）的简单模型物，对TNT及其代谢产物、硝基苯和二硝基苯以及卤代硝基芳烃等不同类型的硝基芳烃的还原代谢中间体亚硝基芳烃与乙硫醇的反应，进行了量子化学动态计算，寻找反应过渡态，求得反应活化能，从理论上探讨了硝基芳烃的致毒机理。以DFT-B3LYP／6-31G^*方法全优化计算了TNT、2-氨基-4，6-二硝基甲苯和4-氨基-2，6-二硝基甲苯在生物体内的代谢还原中间体亚硝基芳烃及其与乙硫醇反应各驻点的结构，求得质子化亚硝基芳烃中亚硝基氮与巯基硫间距离分别为2.005（?）、1.935（?）和1.955（?）时的反应过渡态（TS1，TS2和TS3）以及相应的活化能15.443 kJ／mol、36.286kJ／mol和45.994 kJ／mol。表明硝基数越多，反应越易进行；氨基具致钝作用；当亚硝基处于甲基邻位时，因受到位阻作用，故反应活化能增高。以类似方法和步骤计算研究了硝基苯、1，3-二硝基苯和1，4-二硝基苯的亚硝基化合物与乙硫醇的反应机理，获得类似结果和结论。求得氮-硫原子间距分别等于1.966（?）、2.045（?）和2.054（?）的三个反应过渡态和相应的活化能35.707kJ／mol、17.783kJ／mol和12.702kJ／mol。表明反应速度受硝基数目和位置的影响：硝基数越多反应活化能越小；对二硝基苯快于间二硝基苯。三个反应活化能的大小顺序与硝基芳烃毒性大小排序恰好相反，预示硝基芳烃与生物蛋白大分子中还原性巯基间的反应可能是其致毒的关键步骤。这与第二部分QSAR研究结果相吻合。最后，以类似方法和步骤对卤代硝基芳烃的致毒机理进行了计算研究。求得4-氟硝基苯、4-氯硝基苯和4-溴硝基苯的亚硝基代谢产物与乙硫醇的反应过渡态和相应的活化能。过渡态的氮-硫原子间距分别为1.940（?）、1.947（?）和1.941（?），活化能依次为43.397kJ／mol、41.529kJ／mol和44.332kJ／mol。不同卤素在同一位置取代对反应速度几无影响，表明卤代硝基芳烃至少应存在两种致毒反应：一是其被还原代谢前，与卤素相连的苯环碳原子先受到亲核试剂进攻，使卤素原子被取代；二是其亚硝基代谢中间体与含巯基的蛋白质大分子发生了共价加合，形成了加合物。总之，本文用第一性原理DFT-B3LYP方法，对硝基芳烃的分子结构与毒性的定量关系进行了系统研究，找到影响硝基芳烃毒性的关键因素，给出了较好的毒性预测模型（QSAR），并提出了硝基芳烃污染物毒性降解的初步建议。首次对硝基芳烃的致毒机理进行了动态理论计算，建议了硝基芳烃在生物体内与蛋白质作用的简单模型和致毒的具体过程，找到关键基元反应的过渡态和活化能，探讨了取代基效应和位阻效应。如上关于硝基芳烃衍生物结构-毒性定量关系和致毒机理的理论研究，为理论化学、材料化学和环境化学的交叉研究和发展，提供了新的示例和结果，提供了丰富的原始数据和相应规律。更多还原

【Abstract】 Nitroaromatics are widely used either as materials or as intermediates in explosives,dyestuffs, pesticides, and organic synthesis. They exist as industrial wastes and directpollutants in the environment, and are relatively soluble in water and detectable in rivers,ponds, and soil. Nitroaromatics are representative of electrophilic toxicants. Nitroaromaticcompounds have attracted considerable attention because of their varied toxic effects, suchas narcosis, mutagenicity, and carcinogenicity. Furthermore, some of them can be degradedinto more toxic molecules. Thus, it is necessary to study the toxicities of nitroaromatics.Quantitative structure-activity relationship （QSAR） is a mathematical modeldescribing the relationship between toxic potency and one or more descriptors of achemical. QSAR has been widely used for studying nitroaromatics’ toxicity. However,there is a very serious problem, that is, how to obtain the molecular descriptors for QSAR.In earlier QSAR works, the descriptors were gotten by experiments or semi-empiricalmolecular orbital （MO） calculations. Both of the two means have limitations. The formerneeds more time, money, and manpower, while the latter uses the empirical orexperimental parameters to deal with the Schrodinger equation and omit some molecularintegral calculations, so its result is not accurate. As the development of computer andquantum chemical theory, first principle method density functional theory （DFT） has beenused to study nitroaromatics’ structures and activities now, and it can get the compounds’molecular descriptors easily and accurately. But there is still no reference about the DFTapplication for QSAR study of nitroaromatics’ toxicity.Another evidence obtained in studies on human and model laboratory animals havedemonstrated the formation of hemoglobin （Hb） adducts upon inhalation, ingestion, or skincontact with 2,4,6-trinitrotoluene （TNT）, and there is a correlation between total TNTexposure level and its Hb adduct content. The Hb adducts of other nitroaromaties havebeen documented too. The mechanism of covalent binding of TNT or other nitroaromatiesto critical cellular proteins has been of interest. However there is still no dynamictheoretical calculation study on it.In this paper, the hybrid density functional scheme B3LYP is employed to give thesystemic theoretical studies on the QSAR of nitroaromatics and an explicit illustration onnitroaromaties’ toxic mechanism have been done. The whole work consists of threechapters.1. QSAR models of nitroaromatics toxicity to fathead minnow are established based on different theoretical levels, which are compared with each other to select the best one.Semiempirical MO AM1 and PM3 methods, ab initio MO HF method with 6-31G^*,and DFT-B3LYP method with 6-311G^** were used to calculate the electronic andstructural properties of 28 nitroaromatics. QSARs were established based on theseproperties and the toxicity of nitroaromatics to the fathead minnow. The results show thatthe models established based on the first principle methods （HF and DFT-B3LYP） arebetter than those based on semiempirical methods （AM1 and PM3）. HF model is a littlebetter than DFT-B3LYP model on correlation and significance. But the B3LYP model givesmore reasonable interpretation of nitroaromatics toxic mechanism. As far as this exampleis concerned, the B3LYP method is the best choice for nitroaromatics toxicity study.2. DFT-B3LYP method was used to study the QSAR of nitroaromatics toxicity. Thekey factors affecting their toxicity are found, the toxic mechanism of nitroaromatics isdiscussed based on the QSAR studies, and the toxic values of some nitroaromatics arepredicted by QSAR models. This work gives significant informations for nitroaromaticstoxic mechanism study and provides usable suggestions for detoxifying the river or soilpolluted by nitroaromatics.The DFT-B3LYP method, with the basis set 6-311G^**, was employed to optimize themolecular geometries and electronic structures of nitroaromatics such as mono- （multi-）nitrobenzene, mono- （multi-） nitrotoluene, nitroaniline, and halogenated nitrobenzenederivatives. The quantum chemical parameters were selected as molecular structuraldescriptors. According to the type and number of substituents, the acute toxicity of suchnitroaromatics to Tetrahymena pyriformis, golden orfe fish, and the algae （Scenedesmusobliguus） along with the structural descriptors, was used to establish the QSARs. The nitro,amido, and halogen substituents’ influences on nitroaromatics toxicity were discussed. Theresults indicate that the toxicity of nitroaromatics increases with the increase of the numberof the nitro substituents. The toxic mechanisms of mono-nitrobenzene andmulti-nitrobenzene are different. Mono-nitrobenzenes belong to polar narcosis toxicants,whose toxicity is decided by both its hydrophobicity and electronic activity.Multi-nitroaromatics are reactive toxicants. The electrophilic or nucleophilic reactions willoccur between these chemicals and the proteins in vivo. The other substituents on benzenering have important influences on nitroaromatics toxicity too. The amidoes decrease thetoxicity of nitroaromtics while the halogen substituents increase it. In one word, the nitrogroup is the primary toxic group of nitroaromatics. Therefore, wrapping or reducing thenitro groups on benzene ring is an available approach to degrade the nitroaromatics toxicity. In addition, the stable and remarkable QSAR models established in the part can be used topredict the nitroaromatics toxicity.3. The theoretical calculations have been done on the reactions of the nitrosoaromaticswith the ethanethiol instead of protein that has the thiol （-SH） group. The nitrosoaromaticsare the nitrosoarene intrermediates of nitroaromatics yielded by 2-electron reduction ofnitro groups in vivo. The nitroaromatics studied here are TNT and its two metabolites2-amino-4,6-dinitrotoluene （2A） and 4-amino-2,6-dinitrotoluene （4A）, nitrobenzene （NB）,1,3-dinitrobenzene （13-DNB） and 1,4-dintrobenzene （14-DNB）, three halogenatednitrobenzenes including 4-fluoronitrobenzene（4-FNB）, 4-chloronitrobenzene （4-CNB）, and4-bromonitrobenzene （4-BrNB）. Reaction profiles between the nitrosoaromatics and theethanethiol have been studied. Stationary points including their transition states weresuccessfully located and characterized for the first time at the B3LYP/6-31G^* level withoutany restriction on the internal coordinates. Studies on the geometry, charge, and energy ofthe stationary points were carried out to illustrate the adduct process.All the nitrosoaromatics studied here could bind covalently with the ethanethiol. Theorder of the activation barrier height is TNT＜2A＜4A, 14-DNB＜13-DNB＜NB, 4-FNB≈4-CNB≈4-BrNB. This binding is found to be largely dependent upon the number ofnitro substituents. The type and position of the substituents on the benzene ring is relatedwith the compound’s reactivity directly too. The more nitro groups the benzene ring has,the more the reactivity of nitroaromatics are. The compounds with nitro groups positionedortho to each other are more reactive than the isomers having the nitro groups in the metaorientation. The amido on the benzene ring will decrease the activity of nitroaromacties.These relative reactivity orders are consistent with their toxic value orders. These resultsindicate that the covalent binding of nitrosoaromatics with proteins, for example,hemoglobin, may be the key step of nitroaromatics toxic reaction in vivo.For 4-FNB, 4-CNB, and 4-BrNB, their near activation barrier heights indicate thedifference of halogen substituents on the same position of the benzene ring have littleinfluence on their reaction with ethanethiol. Further analysis suggests that the halogens onthe benzene ring may be replaced by nuloephilic compounds such as DNA in vivo, whichmay be another toxic action of halogenated nitroaromatics.In summary, the first principle DFT-B3LYP method has been employed in this thesisto study the QSAR of nitroaromatics toxicity and their toxic mechanism for the first time.The stable QSAR models of nitroaromtics toxicity to different species have beenestablished, which provide the primary structural factors affecting the compounds’ toxicity and can be used to predict some unknown congeneric compounds’ toxicity. The dynamictheoretical calculations have been done on nitroaromatics toxic mechanism for the firsttime. A simple model of nitroaromatics’ reaction with proteins and the concrete toxicprocedure has been suggested. The transition states and activation energy of the keyelementary reaction have been successfully located here. The substituent effects and sterichindrance in the toxic reaction were discussed. The above theoretical study onnitroaromatics QSAR and toxic mechanism presents new example and results, providesabundant original data and correlating rules to the cross research and the developmentbetween theoretical chemistry, material chemistry and environment chemistry.更多还原