<b>Systematic Analysis of Chemical Constituents, Pharmacokinetic Properties, and Toxicity Risks of </b><i>Lithocarpus litseifolius (Hance) Chun</i><b>: A Computational Study</b>

Abstract Objective: Lithocarpus litseifolius (Hance) Chun is a valuable medicinal and edible plant with multiple pharmacological activities. This study systematically reviews its chemical constituents and evaluates their drug-likeness, pharmacokinetic properties, and toxicity risks to provide evidence for its application and development. Methods: A comprehensive literature search was performed using the keyword "Lithocarpus litseifolius (Hance) Chun" to identify flavonoids, polyphenols, and triterpenoids. The Simplified Molecular Input Line Entry System (SMILES) structures were retrieved from PubChem. Pharmacokinetic parameters were predicted using SwissADME, including Lipinski's rule of five, ESOL solubility, oral bioavailability (BA), gastrointestinal (GI) absorption, log skin permeability coefficient (log Kp), and P-glycoprotein substrate (P-gp S) status. PAINS and Brenk structural alerts were assessed for toxicity screening. Target prediction was conducted using SwissTargetPrediction (Homo sapiens, probability > 0). Results: A total of 40 flavonoids, 6 polyphenols, and 14 triterpenoids were identified. After excluding compounds without explicit SMILES structures, 34 flavonoids, 6 polyphenols, and 8 triterpenoids were retained for analysis. Among these 48 compounds, 23 satisfied the criteria of violating ≤ 1 Lipinski rule and having BA ≥ 30%. Additionally, 22 compounds (45.8%) had neither PAINS nor Brenk alerts, while 12 compounds (25.0%) exhibited both alerts, with 3 compounds containing ≥ 2 Brenk fragments. Conclusion: L. litseifolius demonstrates favorable drug-likeness and low toxicity risk. However, compounds with dual PAINS/Brenk alerts require structural optimization or exclusion. This computational analysis provides a foundation for further pharmacological development and clinical application of L. litseifolius constituents. Keywords: Lithocarpus litseifolius; flavonoids; polyphenols; triterpenoids; drug-likeness; SwissADME; toxicity assessment; bioavailability Background Lithocarpus litseifolius (Hance) Chun, also known as Lithocarpus litseifolius or sweet tea, is a plant belonging to the genus Lithocarpus in the family Fagaceae of the order Fagales [1]. It is a "medicinal and edible" resource in southern China, with provenances from Guangxi exhibiting superior quality compared to those from other regions [2]. As a precious plant integrating three functions—tea, sweetener, and medicine—its medicinal value has gradually attracted attention [3]. In 2017, it was approved as a "new food raw material" at the national level, and in 2024, it was included in the Guangxi characteristic food and medicine catalog. L. litseifolius contains various bioactive constituents, such as flavonoids [4], phenolic glycosides [5], and triterpenoids [6]. Studies have shown [6, 7] that these components possess multiple biological activities, including antioxidant, anti-inflammatory, hypoglycemic, hypolipidemic, uric acid-lowering, and antitumor effects. This study aims to systematically review its main chemical constituents to provide evidential support for the application of L. litseifolius. 1 Methods 1.1 Screening of Active Constituents from L. litseifolius A literature search was conducted using the keyword "Lithocarpus litseifolius (Hance) Chun" to identify flavonoids, phenolic compounds, and triterpenoids. The Simplified Molecular Input Line Entry System (SMILES) of these compounds were then retrieved from the PubChem database. The SwissADME database (http://www.swissadme.ch/) was used to predict whether the compounds complied with Lipinski's rule of five (molecular weight [MW] ≤ 500, lipophilicity [logP] ≤ 5, ≤ 10 hydrogen bond acceptors, ≤ 5 hydrogen bond donors). Additionally, key pharmacokinetic parameters were summarized and evaluated, including ESOL solubility (mg/ml) ≥ 0.01, oral bioavailability (BA) ≥ 30%, gastrointestinal (GI) absorption: HIGH, log skin permeability coefficient (log Kp [cm/s]), and P-glycoprotein substrate (P-gp S) status [8, 9]. "PAINS" and "Brenk" alerts [10] were also assessed to identify potential nonspecific bioactivity interference, toxicity risks, or chemical instability, providing structural alerts for early toxicological screening. 1.2 Target Prediction of Active Constituents from L. litseifolius Compound target prediction was performed using the SwissTargetPrediction database (http://www.swisstargetprediction.ch/), with the species restricted to Homo sapiens and probability > 0 [11]. After data organization and deduplication, the pharmacological characteristics and targets of L. litseifolius were obtained. 2 Results A systematic literature review was conducted to comprehensively compile compounds from L. litseifolius. The results revealed a total of 40 flavonoid components; however, explicit SMILES structures could not be retrieved from the PubChem database for 6 of them. Additionally, 6 polyphenolic compounds and 14 triterpenoid compounds were identified, of which 5 triterpenoids yielded undefined results and 1 triterpenoid lacked a clear SMILES structure in PubChem. Based on these findings, constituents without definite SMILES structures were excluded, finally retaining 34 flavonoids, 6 polyphenols, and 8 triterpenoids. Regarding the pharmacological properties of L. litseifolius, 33 compounds had a molecular weight ≤ 500; 22 compounds had ≤ 10 hydrogen bond acceptors; 25 compounds had ≤ 5 hydrogen bond donors; 39 compounds had a lipophilicity (logP) ≤ 5; 23 compounds violated ≤ 1 Lipinski rule; 37 compounds had ESOL solubility > 0.01 mg/ml; 23 compounds had BA ≥ 30%; 11 compounds exhibited HIGH gastrointestinal absorption; 0 compounds had positive log skin permeability coefficient (log Kp [cm/s]); and 27 compounds were identified as P-gp substrates (detailed data are available in the table). In terms of potential activity interference, toxicity risks, or chemical instability of L. litseifolius constituents, 45.8% of compounds (22/48) had neither PAINS nor Brenk alerts, while 75.0% of compounds (36/48) had no PAINS or Brenk alerts. Conversely, 25% of compounds (12/48) exhibited both PAINS and Brenk alerts, and 3 compounds contained ≥ 2 Brenk fragments. 3 Discussion Analysis of Pharmacological Properties, Toxicity Risks, and Administration Routes Pharmacological property analysis demonstrated that L. litseifolius exhibits favorable drug-likeness and high bioavailability. ESOL solubility indicated that some compounds are suitable for oral administration. However, only 11 compounds showed HIGH gastrointestinal absorption, which may limit the bioavailability of certain constituents. Nevertheless, solubility can be enhanced using techniques such as micronization and nanocrystallization, solid dispersions, surfactants, and cosolvents [12]. The log skin permeability coefficient (log Kp) values were all negative, indicating weak transdermal penetration capacity. If transdermal administration is considered, penetration enhancers such as borneol, menthol, and camphor [13] should be added to improve percutaneous absorption and thereby enhance local or systemic therapeutic effects. Toxicity risk analysis revealed that L. litseifolius possesses high structural reliability and low toxicity risk. However, 25% of compounds exhibited both PAINS and Brenk alerts, and 3 compounds contained ≥ 2 Brenk fragments, classifying them as high-risk candidates that should be prioritized for exclusion or structural modification. Future research should further explore the pharmacological and toxicological properties of these compounds and optimize their structures to meet clinical application requirements. References [1] He WW, Huang A, Tang XY, et al. Research progress on chemical constituents, pharmacological effects, and quality evaluation of Lithocarpus litseifolius [J]. Central South Pharmacy, 2022, 20(11): 2648-2657. [2] Wang K, Huang XL, Li BC, et al. Analysis and evaluation of main economic traits and active components in 30 provenances of Lithocarpus polystachyus [J]. Southwest China Journal of Agricultural Sciences, 2019, 32(05): 1051-1056. [3] Lin XQ, Wang QS, Qin XH, et al. Textual research on sweet tea "Lithocarpus litseifolius" in materia medica [J]. Journal of Fujian Agriculture and Forestry University (Natural Science Edition), 2023, 52(06): 785-792. [4] Xu J, Zhao Y, Wen J, et al. Composition and toxicity of total water-soluble flavonoids from Lithocarpus litseifolius (Hance) Chun (sweet tea) [J]. Food and Agricultural Immunology, 2024, 35(1). [5] Wang M, Liu X, Zhang Z, et al. Phytochemicals and bioactive analysis of different sweet tea (Lithocarpus litseifolius [Hance] Chun) varieties [J]. Journal of Food Biochemistry, 2021, 45(3): e13183. [6] Cheng Y, Liu F, Wang C, et al. Bioactive Triterpenoids from the Leaves and Twigs of Lithocarpus litseifolius and L. corneus [J]. Planta Medica, 2018, 84(1): 49-58. [7] Wei YQ, Cai YL, Yang Y, et al. Research progress on dihydrochalcones from Lithocarpus litseifolius extracts in treatment of type 2 diabetes mellitus and its complications [J]. Chinese Journal of Chinese Materia Medica, 2025, 50(3): 658-671. [8] Avdeef A. Prediction of aqueous intrinsic solubility of druglike molecules using Random Forest regression trained with Wiki-pS0 database [J]. ADMET DMPK, 2020, 8(1): 29-77. [9] Pillai O, Dhanikula AB, Panchagnula R. Drug delivery: an odyssey of 100 years [J]. Current Opinion in Chemical Biology, 2001, 5(4): 439-446. [10] Rani N, Kumar P. Exploring Natural Compounds as Potential CDK4 Inhibitors for Therapeutic Intervention in Neurodegenerative Diseases through Computational Analysis [J]. Molecular Biotechnology, 2025, 67(8): 3310-3329. [11] Zhong T, Sun J, Miao M, et al. The mechanism of probiotics in pregnancy outcomes in overweight or obese pregnant women based on meta-analysis, network pharmacology and molecular docking [J]. BMC Pregnancy and Childbirth, 2025, 25(1): 886. [12] Bhalani DV, Nutan B, Kumar A, et al. Bioavailability Enhancement Techniques for Poorly Aqueous Soluble Drugs and Therapeutics [J]. Biomedicines, 2022, 10(9). [13] Jiang X, Ma X, Liu WH, et al. Research progress on permeability of transdermal patches [J]. Chinese Pharmaceutical Affairs, 2023, 37(3): 312-320.摘要 目的：硬叶柯（Lithocarpus litseifolius (Hance) Chun）是一种兼具药用与食用价值的珍贵植物，具有多种药理活性。本研究对其化学成分进行系统综述，并评估其成药性、药代动力学特性及毒性风险，为其应用与开发提供证据支持。 方法：以关键词“Lithocarpus litseifolius (Hance) Chun”进行全面文献检索，筛选出黄酮类、多酚类及三萜类化合物。从PubChem数据库获取简化分子线性输入规范（Simplified Molecular Input Line Entry System, SMILES）结构。采用SwissADME预测药代动力学参数，包括Lipinski五规则（Lipinski's rule of five）、ESOL溶解度、口服生物利用度（BA）、胃肠道（GI）吸收、皮肤渗透系数对数（log Kp）及P-糖蛋白底物（P-glycoprotein substrate, P-gp S）状态。通过PAINS与Brenk结构警示进行毒性筛选。采用SwissTargetPrediction进行靶点预测（物种限定为智人，概率>0）。 结果：共鉴定出40种黄酮类化合物、6种多酚类化合物及14种三萜类化合物。排除无明确SMILES结构的化合物后，最终保留34种黄酮类、6种多酚类及8种三萜类化合物用于分析。在上述48种化合物中，23种符合违反≤1条Lipinski规则且口服生物利用度≥30%的标准。此外，22种化合物（45.8%）无PAINS及Brenk警示，12种化合物（25.0%）同时存在两类警示，其中3种化合物包含≥2个Brenk片段。 结论：L. litseifolius展现出良好的成药性与较低的毒性风险，但同时存在PAINS与Brenk警示的化合物需进行结构优化或排除。本计算分析为L. litseifolius活性成分的进一步药理开发与临床应用奠定了基础。 关键词：硬叶柯（Lithocarpus litseifolius）；黄酮类；多酚类；三萜类；成药性；SwissADME；毒性评估；生物利用度 背景：硬叶柯（Lithocarpus litseifolius (Hance) Chun）又名甜茶，隶属于壳斗目（Fagales）壳斗科（Fagaceae）柯属（Lithocarpus），是中国南方的药食同源资源，广西产区的种质品质优于其他区域[2]。作为集茶用、甜味剂及药用三种功能于一体的珍贵植物，其药用价值逐渐受到关注[3]。2017年，该植物被获批为国家级“新食品原料”，2024年被纳入广西特色药食同源目录。硬叶柯含有多种生物活性成分，如黄酮类[4]、酚苷类[5]及三萜类[6]。已有研究[6,7]表明，此类成分具备抗氧化、抗炎、降糖、调脂、降尿酸及抗肿瘤等多种生物活性。本研究旨在系统综述其主要化学成分，为硬叶柯的应用提供证据支撑。 1 方法 1.1 硬叶柯活性成分筛选 以关键词“Lithocarpus litseifolius (Hance) Chun”开展文献检索，鉴定黄酮类、酚类化合物及三萜类化合物，随后从PubChem数据库获取这些化合物的简化分子线性输入规范（SMILES）。采用SwissADME数据库（http://www.swissadme.ch/）预测化合物是否符合Lipinski五规则（分子量[MW]≤500、脂溶性[logP]≤5、氢键受体数≤10、氢键供体数≤5）。同时汇总并评估关键药代动力学参数，包括ESOL溶解度（mg/ml）≥0.01、口服生物利用度（BA）≥30%、胃肠道（GI）吸收水平为高、皮肤渗透系数对数（log Kp [cm/s]）及P-糖蛋白底物（P-gp S）状态[8,9]。此外，评估“PAINS”与“Brenk”警示[10]，以识别潜在的非特异性生物活性干扰、毒性风险或化学不稳定性，为早期毒理学筛选提供结构警示依据。 1.2 硬叶柯活性成分靶点预测 采用SwissTargetPrediction数据库（http://www.swisstargetprediction.ch/）进行化合物靶点预测，物种限定为智人（Homo sapiens），概率阈值设为>0[11]。经数据整理与去重后，获取硬叶柯的药理特性及作用靶点。 2 结果 通过系统文献综述全面整理硬叶柯中的化合物，结果显示共鉴定出40种黄酮类成分，但其中6种无法从PubChem数据库获取明确的SMILES结构。此外还鉴定出6种多酚类化合物与14种三萜类化合物，其中5种三萜类化合物结果未明确，1种三萜类化合物在PubChem中无清晰SMILES结构。据此排除无明确SMILES结构的成分，最终保留34种黄酮类、6种多酚类及8种三萜类化合物。 关于硬叶柯的药理特性：33种化合物分子量≤500；22种化合物氢键受体数≤10；25种化合物氢键供体数≤5；39种化合物脂溶性（logP）≤5；23种化合物违反≤1条Lipinski规则；37种化合物ESOL溶解度>0.01 mg/ml；23种化合物口服生物利用度≥30%；11种化合物胃肠道吸收水平为高；0种化合物皮肤渗透系数对数（log Kp [cm/s]）为正值；27种化合物被鉴定为P-糖蛋白底物（详细数据见附表）。 关于硬叶柯成分的潜在活性干扰、毒性风险或化学不稳定性：45.8%的化合物（22/48）无PAINS及Brenk警示，75.0%的化合物（36/48）无PAINS或Brenk警示。与之相反，25%的化合物（12/48）同时存在PAINS与Brenk警示，其中3种化合物包含≥2个Brenk片段。 3 讨论 3.1 药理特性、毒性风险与给药途径分析 药理特性分析表明，硬叶柯具备良好的成药性与较高的生物利用度。ESOL溶解度结果显示部分化合物适合口服给药，但仅11种化合物胃肠道吸收水平为高，这可能限制部分成分的生物利用度。不过可通过微粉化、纳米结晶、固体分散体、表面活性剂及助溶剂等技术提升溶解度[12]。皮肤渗透系数对数（log Kp）均为负值，表明透皮渗透能力较弱。若考虑经皮给药途径，需添加冰片、薄荷醇、樟脑等透皮促进剂[13]以改善经皮吸收，从而增强局部或全身治疗效果。 毒性风险分析显示，硬叶柯具有较高的结构可靠性与低毒性风险，但25%的化合物同时存在PAINS与Brenk警示，3种化合物包含≥2个Brenk片段，属于高风险候选物，应优先排除或进行结构修饰。未来研究应进一步探索此类化合物的药理与毒理特性，优化其结构以满足临床应用需求。 参考文献 [1] He WW, Huang A, Tang XY, et al. Research progress on chemical constituents, pharmacological effects, and quality evaluation of Lithocarpus litseifolius [J]. Central South Pharmacy, 2022, 20(11): 2648-2657. [2] Wang K, Huang XL, Li BC, et al. Analysis and evaluation of main economic traits and active components in 30 provenances of Lithocarpus polystachyus [J]. Southwest China Journal of Agricultural Sciences, 2019, 32(05): 1051-1056. [3] Lin XQ, Wang QS, Qin XH, et al. Textual research on sweet tea "Lithocarpus litseifolius" in materia medica [J]. Journal of Fujian Agriculture and Forestry University (Natural Science Edition), 2023, 52(06): 785-792. [4] Xu J, Zhao Y, Wen J, et al. Composition and toxicity of total water-soluble flavonoids from Lithocarpus litseifolius (Hance) Chun (sweet tea) [J]. Food and Agricultural Immunology, 2024, 35(1). [5] Wang M, Liu X, Zhang Z, et al. Phytochemicals and bioactive analysis of different sweet tea (Lithocarpus litseifolius [Hance] Chun) varieties [J]. Journal of Food Biochemistry, 2021, 45(3): e13183. [6] Cheng Y, Liu F, Wang C, et al. Bioactive Triterpenoids from the Leaves and Twigs of Lithocarpus litseifolius and L. corneus [J]. Planta Medica, 2018, 84(1): 49-58. [7] Wei YQ, Cai YL, Yang Y, et al. Research progress on dihydrochalcones from Lithocarpus litseifolius extracts in treatment of type 2 diabetes mellitus and its complications [J]. Chinese Journal of Chinese Materia Medica, 2025, 50(3): 658-671. [8] Avdeef A. Prediction of aqueous intrinsic solubility of druglike molecules using Random Forest regression trained with Wiki-pS0 database [J]. ADMET DMPK, 2020, 8(1): 29-77. [9] Pillai O, Dhanikula AB, Panchagnula R. Drug delivery: an odyssey of 100 years [J]. Current Opinion in Chemical Biology, 2001, 5(4): 439-446. [10] Rani N, Kumar P. Exploring Natural Compounds as Potential CDK4 Inhibitors for Therapeutic Intervention in Neurodegenerative Diseases through Computational Analysis [J]. Molecular Biotechnology, 2025, 67(8): 3310-3329. [11] Zhong T, Sun J, Miao M, et al. The mechanism of probiotics in pregnancy outcomes in overweight or obese pregnant women based on meta-analysis, network pharmacology and molecular docking [J]. BMC Pregnancy and Childbirth, 2025, 25(1): 886. [12] Bhalani DV, Nutan B, Kumar A, et al. Bioavailability Enhancement Techniques for Poorly Aqueous Soluble Drugs and Therapeutics [J]. Biomedicines, 2022, 10(9). [13] Jiang X, Ma X, Liu WH, et al. Research progress on permeability of transdermal patches [J]. Chinese Pharmaceutical Affairs, 2023, 37(3): 312-320.