兽研简要总结
作为具有重要经济价值的毛皮动物，貉在全球范围内被广泛养殖。然而，制定标准化养殖方案以提高生长期貉的成活率仍是该行业面临的主要挑战。肠道微生物群对幼龄动物的健康至关重要，而出生后早期是微生物定植的关键时期。然而，关于貉肠道微生物群组成和功能的研究仍然有限。本研究系统采集了貉在五个关键发育时间点（2、3、4、6和8月龄）的粪便样本，并利用16S rRNA测序对其肠道微生物组成进行了表征。本研究结果为优化貉养殖管理提供了理论支持。
关键词：貉，肠道微生物群，生长期，代谢功能
1. 引言
貉（Nyctereutes procyonides）广泛分布于东亚和俄罗斯远东地区，以其珍贵的毛皮而闻名[1]。貉浓密的被毛和独特的毛色使其特别适合生产优质毛皮产品，包括大衣、帽子和其它奢侈服装[2]。近几十年来，受全球时尚趋势和奢侈品市场扩张的推动，对貉毛皮的商业需求不断增长[3]。然而，由于该物种的肉食性饮食、攻击性气质和高应激易感性，圈养繁殖仍然困难。这些行为特征要求制定标准化的饲养方案以确保动物福利和生产效率。生长期是幼龄貉的关键发育窗口期，在此期间适当的营养和管理直接影响毛皮质量和产量[4]。因此，维持貉在此期间的健康对于商业毛皮生产至关重要。
肠道微生物群包括细菌、古菌、真菌、病毒、噬菌体和真核寄生虫[5]。该微生物生态系统通过代谢营养物质[6]、维持免疫稳态[7]和调节生长[8]对宿主生理产生深远影响。平衡的肠道微生物组成和稳定的生态系统动态对于动物正常的消化功能、代谢调节和生殖健康至关重要。共生微生物的保护作用通过多种途径介导：刺激宿主免疫反应、产生生物活性代谢物以及通过定植抵抗竞争性排斥病原微生物[9]。作为宿主的“第二基因组”，肠道微生物群在应对饮食、生理应激、宿主遗传和健康状况等因素时表现出动态可塑性[10,11]。肠道微生物群落失衡与多种病理状态相关，从自身免疫性疾病[12]和代谢失调[10,13]到传染病易感性增加[14]。这些发现强调了维持微生物稳态对优化貉养殖的重要性。值得注意的是，肠道微生物群在宿主发育过程中经历显著的组成和功能转变[15,16]。已有报道表明，貉毛皮质量与肠道微生物群之间存在相关性[17]。尽管先前的研究已调查了貉的肠道微生物生态学，但主要集中于特定生长阶段。这些发现是否能完全反映整个生长期肠道微生物群的动态变化仍不清楚。这一知识空白阻碍了为这一具有重要商业价值的物种制定循证饲养策略。
本研究采集了貉在五个关键发育时间点（2、3、4、6和8月龄）的粪便样本，并通过16S rRNA测序对肠道微生物组成进行了表征，旨在鉴定不同生长阶段的核心微生物群落并阐明其主要代谢功能。
2. 材料与方法
2.1. 样本采集
实验在吉林省左家狐狸和貉养殖基地进行。养殖场位于缓坡地带，光照充足，空气流通良好。圈舍采用钢板屋顶，檐高1.5–2.0 m，每个单元宽2–4 m。笼舍用钢架加固并用铁丝网围封。所有貉均单只饲养于1.0 m × 1.0 m的笼中，养殖设施内未配备人工温湿度控制系统。每天07：00和16：00各喂食一次。基础日粮组成见表1。选取10只健康雄性貉（4.12 ± 0.47 kg），生长速度一致，无疾病史。在出生后2（M2组，n = 8）、3（M3组，n = 8）、4（M4组，n = 8）、6（M6组，n = 8）和8（M8组，n = 10）月龄对同一批10只个体重复采样，以追踪其肠道微生物群的周期性变化。在每个采样时间点，于早晨饲喂后约2–3 h采集1–2 g新鲜粪便。采集的粪便样品放入无菌离心管中，采集后立即置于液氮罐中保存。每个阶段的所有样品采集完毕后，转移至实验室−80 °C冰箱中长期保存。所有样品采集完成后，进行统一批次处理和测序。
Table 1. Composition of basal diets (DM basis, g/kg).
| Ingredients | Content |
|---|---|
| Extruded corn | 405.00 |
| Soybean meal | 210.00 |
| Meat meal | 50.00 |
| Fish meal | 25.00 |
| Blood meal | 50.00 |
| Distillers dried grains with solubles | 100.00 |
| Corn germ meal | 100.00 |
| Soybean oil | 35.00 |
| CaHPO 4 | 8.00 |
| Limestone powder | 7.00 |
| Premix | 10.00 |
2.2. DNA提取与PCR扩增
使用DNeasy PowerSoil试剂盒（QIAGEN，德国希尔登）从粪便样品中提取总DNA。随后，为获得高质量的总DNA，对样品进行离心以去除杂质。使用针对16S rRNA基因保守区设计的特异性引物进行PCR扩增，产物用磁珠纯化以去除小分子副产物。PCR产物纯度通过1%琼脂糖凝胶电泳评估。随后对V3/V4区进行PCR扩增，在2%琼脂糖凝胶上验证[18]，并通过添加平台特异性接头和样品特异性条形码制备测序文库。然后将文库混合并在Illumina平台上测序。
2.3. 16S rRNA测序数据预处理
使用Fastp（版本0.23.1）处理原始标签以获得高质量标签数据。应用Cutadapt（版本1.9.1）去除引物序列并降低噪音[19]。随后将序列与物种注释数据库（https://www.arb-silva.de/；2025年3月20日）比对以去除嵌合序列，获得最终有效数据。将序列相似性为97%的ASV聚类在一起。使用QIIME2软件包1.8.0版本平台对微生物进行鉴定和分类[20，21]。生成的ASV表导出至R-studio，进行ASV聚类、分类和多样性评估。
2.4. Alpha多样性分析和Beta多样性分析
使用QIIME2软件估计微生物丰富度和均匀度。进行PCA、主坐标分析（PCoA）、非度量多维标度（NMDS）分析、相似性分析（ANOSIM）和置换多元方差分析（PERMANOVA）以探究组间差异[22]。使用R-studio中Vegan包的speccaccum函数生成物种累积曲线。使用QIIME2中的UPGMA.tre函数绘制UPGMA聚类图。
2.5. 组间差异显著性分析
应用Kruskal–Wallis秩和检验筛选组分丰度存在差异的物种，并使用线性判别分析（LDA）计算每个差异物种的效应量。为探究不同年龄组肠道微生物群之间的潜在相互作用，使用Spearman相关分析筛选属水平总丰度前25的物种。保留绝对相关系数> 0.6且p值显著≤ 0.05的物种用于进一步研究。
2.6. 功能预测
应用PICRUSt（https://github.com/picrust/picrust2；2025年3月20日）预测微生物群落的潜在分子功能[23]。为分析Tax4Fun，使用runRefBlast和make Functional Prediction功能，采用默认参考数据库和99%的聚类阈值分析16S rRNA基因[24，25]。使用BugBase预测测序结果中的细菌组成[26]。使用过滤后的序列，以Silva 138 99% ASV全长序列数据库对序列进行分类[27]。
3. 结果
3.1. 肠道微生物群的丰富度和多样性
如图1A、C所示，不同年龄貉的肠道微生物群在多样性和丰富度方面存在显著差异，其中M4组的Chao1指数和Observed features最低，表明物种丰度随年龄增长先下降后趋于稳定。同样，M4组的Shannon和Simpson指数最低，表明群落多样性随年龄增长先降低后升高（图1E、F）。Pielou指数用于评估物种均匀度（图1D）。尽管M4组的Pielou值最低，但肠道微生物群的均匀度随年龄增长显著增加。此外，Dominance值的变化趋势与Pielou指数相反，进一步印证了这一结果（图1B）。物种累积箱线图可描述样本的物种组成和丰度。随着样本量增加，箱线图趋于平缓，表明样本量足以进行数据分析（图1G）。图1H显示，曲线随年龄增长变得更宽更平，表明物种丰度和均匀度随年龄增长而增加。Venn图分析（图1I、J）显示，五个组共有8个微生物门。此外，M2组有6个独有的微生物物种，M6组有1个独有的细菌。在属水平上，五个组共有的物种总数为106，M2、M3、M4、M6和M8组分别有28、2、12、8和19个独有物种。
[图：肠道微生物群丰富度和多样性（M2、3、4、6：n = 8；M8：n = 10）。（A–F）：各组间Alpha多样性指数。]
3.2. 肠道微生物群的差异
如图2A所示，M2组与M4、M6和M8组存在显著差异，而M3组与M4和M6组存在显著差异。Beta指数组间差异检验是一种用于比较组间差异的统计方法。可以推断，随着年龄增长，肠道微生物群发生了变化。图2C、D显示PCoA和NMDS图中存在明显的组间分离。ANOSIM分析证实各组间肠道微生物群落结构存在显著差异（R = 0.907，p = 0.001），组间差异大于组内差异（图2E）。图2F显示，PERMANOVA表明各组间群落同质性存在明显差异（R2 = 0.553，F = 11.43，p = 0.001）。ZH12的组内变异最大，而ZH6的肠道微生物群落最为同质。组间肠道微生物群和细菌群落结构的相似性可通过算术平均非加权成对法（UPGMA）检测。层次聚类树见图2G、H。M2组（ZH601-ZH608）和M3组（ZH901-ZH908）聚在一起，而M4组（ZH1201-ZH1208）和M6组（ZH1803-ZH1807、ZH1809-ZH18011）分散在两端，M8组（ZH2401-ZH24010）更接近M4和M6组。
[图：不同组间肠道微生物群的差异（M2、3、4、6：n = 8；M8：n = 10）。（A）：Beta指数组间差异]
3.3. 肠道微生物群落组成
如图3A所示，厚壁菌门、拟杆菌门、梭杆菌门、放线菌门和变形菌门在各组中占主导地位。在属水平上，Prevotella 9、链球菌属、乳杆菌属、消化链球菌属和布劳特菌属是各组中的优势菌属（图3B）。物种丰度聚类热图用于发现各组中物种聚集含量的高低。在门水平上，蓝细菌门、厚壁菌门和放线菌门在M2组中富集；梭杆菌门在M4组中富集；广古菌门、螺旋体门和弯曲菌门在M6组中富集；变形菌门在M3组中富集；弯曲菌门、脱硫杆菌门、拟杆菌门和变形菌门在M8组中富集（图3C）。在属水平上，M4组中丰度较高的物种为鲸杆菌属和链球菌属；M2组中丰度较高的物种为Ligilactobacillus和巨球藻属；M6组中丰度较高的物种为乳杆菌属和Prevotella 9；M3组中丰度较高的物种为布劳特菌属、巨单胞菌属和消化链球菌属；M8组中丰度较高的物种为Prevotella 9和Turicibacter（图3D）。
[图：肠道微生物群落组成（M2、3、4、6：n = 8；M8：n = 10）。（A）：门水平的相对丰度。（]
3.4. 肠道微生物群的差异丰度
如Kruskal–Wallis H检验条形图（图4A、B）所示，差异丰度门为厚壁菌门、拟杆菌门、梭杆菌门、放线菌门和变形菌门（图4A），差异丰度属为Prevotella 9、鲸杆菌属、链球菌属、Blautia和巨球藻属（图4B）。LEfSe分析（LDA评分>4，p<0.05）见图4C：M2组中富集最多的分类群为厚壁菌门、毛螺菌目和毛螺菌科，而Prevotella 9在M3组中富集。乳杆菌目、芽孢杆菌纲、乳杆菌科、梭杆菌科、梭杆菌纲、梭杆菌门、梭杆菌目和鲸杆菌属在M4组中富集。Prevotellaceae是M6组中富集最多的分类群，而M8组中富集的分类群为拟杆菌目、拟杆菌门、拟杆菌纲和梭菌纲。图4D显示了不同物种的进化分支。辐射状圆圈代表从门到属的分类水平，每个小圆圈表示该水平上的一个分类群，其直径与相对丰度成正比。无显著差异的分类群统一为黄色，而有显著差异的分类群则根据分组着色。
[图：不同组间肠道微生物群的差异丰度（M2、3、4、6：n=8；M8：n=10）。（A）：Kruskal–Wallis]
3.5. 微生物相互作用的共现网络分析
共现网络显示，M2组有73条边，包括57条正边和16条负边，而Prevotellaceae UCG 001、Prevotella 7和Monoglobus包含大量相关性（图5A）。M3组有157条边，包括116条正边和41条负边，而Lactiplantibacillus、双歧杆菌属、Turicibacter、Slackia和Lachnoclostridium包含大量相关性（图5B）。M4组有50条边，包括43条正边和7条负边，而Dorea、葡萄球菌属和Prevotella 7包含大量相关性（图5C）。M6组有62条边，包括49条正边和13条负边，而Slackia、链球菌属、甲烷短杆菌属和Romboutsia包含大量相关性（图5D）。M8组有42条边，包括39条正边和3条负边，而Prevotellaceae UCG 001和Lachnospiraceae NK3A20群包含大量相关性（图5E）。所有组共有117条边，包括103条正边和14条负边，而Prevotellaceae UCG 001、Negativibacillus、鲸杆菌属、Mailhella和消化球菌属包含大量相关性（图5F）。
[图：属水平微生物群落的共现网络（M2、3、4、6：n=8；M8：n=10）。（A）：M2组]
3.6. 肠道微生物群的功能预测
使用PICRUSt对不同注释水平下肠道微生物群丰度最高的前10个功能进行统计分析（图6A–C）。水平1中的优势功能为代谢、遗传信息处理、环境信息处理、细胞过程、人类疾病和生物体系统（图6A）。水平2中的主要功能为膜转运、碳水化合物代谢、复制与修复、氨基酸代谢、翻译、能量代谢、特征不明、核苷酸代谢、辅因子和维生素代谢以及细胞过程与信号传导（图6B）。水平3中的主要功能为转运蛋白、仅一般功能预测、ABC转运蛋白、DNA修复与重组蛋白、核糖体、嘌呤代谢、嘧啶代谢、肽酶、转录因子和染色体（图6C）。此外，我们通过Tax4Fun分析肠道微生物群功能获得了类似结果。各组之间的功能差异已通过热图展示（图6D–F）。BugBase是一种基于16S rRNA预测微生物表型并进行组间差异分析的工具。它可以预测七个方面的信息，包括革兰阳性/阴性、生物膜形成、致病潜力、需氧/厌氧、耐应激、兼性厌氧以及含可移动元件。通过BugBase分析，我们的结果表明各组功能相似（图6G）。原核分类群功能注释（FAPROTAX）是由美国俄勒冈大学生物系Louca实验室开发的数据库。该数据库利用当前关于培养菌株的文献来预测属或种等原核生物的功能注释。我们的结果表明，各微生物群的核心功能主要为化能异养、发酵、动物寄生虫或共生体、哺乳动物肠道和人类肠道（图6H）。
[图：肠道微生物群的功能预测（M2、3、4、6：n=8；M8：n=10）。（A–C）：不同水平的PICRUSt分析。]
4. 讨论
目前，生长期貉肠道微生物群的动态变化仍不清楚。本研究系统研究了不同生长阶段肠道微生物群的组成动态和功能潜力，最终目标是为商业貉养殖场改进饲养管理实践建立循证框架。α多样性指标，包括物种丰富度、多样性和优势度，共同表征了肠道生态系统中微生物的复杂性[28]。我们的分析揭示了貉肠道微生物群明显的时间模式：在生长期内，多样性和丰富度先下降，然后上升，最终趋于稳定。这一发育轨迹与其他哺乳动物中报道的年龄依赖性微生物演替一致[29]。β多样性分析量化了发育阶段之间的组成差异[30]。在生长早期和晚期之间检测到显著的微生物群落变化，PCoA、NMDS、ANOSIM和PERMANOVA均显示出明显的时间分离[31]。层次聚类分析进一步证实了这些发现，揭示了微生物群落的时间依赖性重构，与已记录的肠道微生物群成熟模式相平行[32]。
我们的结果表明，肠道微生物群主要由厚壁菌门、拟杆菌门、放线菌门和梭杆菌门组成，这与先前在貉中的发现一致[17]。值得注意的是，肠道微生物群主要由厚壁菌门和拟杆菌门组成，这是在哺乳动物肠道生态系统中观察到的特征性分布模式[33]。厚壁菌门/拟杆菌门（F/B）比值与能量吸收呈正相关[34,35,36]。在本研究中，F/B比值在生长早期相对较高，这可能与从日粮中快速吸收能量以满足生长的高能量需求有关。随着生长随年龄减慢，F/B比值逐渐下降。当貉在第8个月停止生长时，能量需求进一步下降，这也可能导致F/B比值降低。尽管放线菌在肠道微生物群中丰度相对较低，但其丰度下降在维持微生物平衡、宿主免疫调节和代谢健康方面的作用可作为多种慢性疾病的微生物标志物[37]。梭杆菌门呈现出显著的时间模式，在4个月时丰度显著增加，这可能有利于短链脂肪酸的合成，从而维持肠上皮细胞的能量供应并维持屏障完整性[38]。属水平分析确定Prevotella 9是所有年龄组中的优势分类群。作为拟杆菌门的有益成员，Prevotella能够高效发酵抗性淀粉和果胶等复杂多糖[39]。此外，源自Prevotella的短链脂肪酸（SCFAs）通过增加紧密连接蛋白表达来改善肠道屏障功能[40]。值得注意的是，我们观察到鲸杆菌属丰度在第四个月出现短暂暴发。研究发现，鲸杆菌属能显著提高胰蛋白酶、脂肪酶和刷状缘酶的活性，并通过产生短链脂肪酸为肠上皮提供能量。此外，鲸杆菌属可上调ZO-1和Occludin的表达，维持肠道健康[41]。共现网络分析进一步揭示，鲸杆菌属与多种微生物相关，可能是因为它通过产生维生素B12增强与肠道微生物群的相互作用[42]。然而，到6月龄时，鲸杆菌属丰度显著下降，同时Prevotella 9重新建立优势地位。
尽管肠道微生物群的宿主调节效应已得到充分确立，但其潜在机制仍不完全清楚[43]。此外，虽然16S rRNA测序能有效分析微生物组成，但其提供的功能信息有限。为解决这一问题，使用Tax4Fun和PICRUSt对发育各阶段进行了功能预测。值得注意的是，PICRUSt的实用性受限于其对过时参考数据库的依赖[44]，而Tax4Fun通过其改进的算法和更新的基因组数据库展现出更优的预测准确性[45]。我们的双重分析策略一致鉴定出代谢和遗传信息处理为主要功能类别。
尽管这项16S rRNA研究为貉肠道微生物群动态提供了基础性见解，但若干局限性值得考虑。在测序过程中，DNA提取、PCR扩增和引物选择所产生的偏倚，加之物种/菌株水平分辨率有限以及相对丰度数据未必能反映绝对细胞数量或活性，均制约了结果的准确性[46]。探究生长貉肠道微生物群的周期性变化有助于确定最佳饲喂阶段，并为调整日粮和管理措施以改善健康和生长性能提供科学依据。然而，本研究未测定生长性能或毛皮品质；因此，这些参数与生长貉肠道微生物群之间的关系仍有待未来研究探讨。未来研究还应纳入个体差异分析、饮食习惯相关性以及畜牧管理元数据的整合。这种多维度方法将有助于更深入地理解商业毛皮动物生产中微生物组与宿主的相互作用。
5. 结论
我们的结果揭示了生长貉肠道微生物群组成存在显著的阶段特异性变化。在门水平上，不同时期生长貉的优势菌门为厚壁菌门、拟杆菌门、放线菌门和变形菌门。属水平分析表明，包括Prevotella 9、巨球藻属、Holdemanella、Romboutsia和Fusionnibacter在内的若干关键分类群在不同生长阶段存在显著差异丰度。功能预测分析显示，这些微生物群落主要参与碳水化合物、氨基酸和能量代谢。
作者贡献
概念化，Z.L.；方法学，S.Y.；软件，W.N.；验证，C.Y.；形式分析，C.X.；调查，L.S.和C.X.；资源，C.X.；数据管理，S.Y.；写作——初稿准备，C.Y.；写作——审阅与编辑，W.N.；可视化，C.Y.；监督，Z.L.；项目管理，Z.L.；资金获取，Z.L.。所有作者均已阅读并同意稿件的出版版本。
机构审查委员会声明
所有动物实验均经吉林农业大学伦理委员会批准（编号20230606001），批准日期：2024年6月6日。
知情同意声明
已获得本研究涉及动物所有者的书面知情同意。
数据可用性声明
所有数据和材料可应通讯作者要求提供。
利益冲突
作者声明无利益冲突。
资金声明
作者声明本研究、 authorship 和/或本文的发表获得了财政支持。本工作由国家重点研发计划（2023YFD1302000）资助。
脚注
[内容请查看原文]
参考文献
[内容请查看原文]
关联数据
数据可用性声明
所有数据和材料可应通讯作者要求提供。
查看英文原文▾
Simple Summary
As economically important fur-bearing animals, raccoon dogs are widely farmed around the world. Nevertheless, developing standardized breeding protocols to improve the survival of growing raccoon dogs remains a major challenge for the industry. The gut microbiota is essential for the health of young animals, and the early postnatal stage is a critical period for microbial colonization. However, research on the composition and function of the gut microbiota in raccoon dogs remains limited. In this study, we systematically collected fecal samples from raccoon dogs at five key developmental time points (2, 3, 4, 6, and 8 months of age) and characterized their gut microbial composition using 16S rRNA sequencing. The present findings provide theoretical support for the optimization of raccoon dog breeding management.
Keywords: raccoon dog, gut microbiota, growth period, metabolic function
1. Introduction
Raccoon dogs (Nyctereutes procyonides) are widely distributed in East Asia and the Russian Far East and are known for their valuable pelage [1]. The dense fur and distinctive coloration of raccoon dogs make them particularly suitable for the production of premium fur products, including coats, hats, and other luxury apparel [2]. Recent decades have seen growing commercial demand for raccoon dog fur, driven by global fashion trends and the expansion of luxury markets [3]. Nevertheless, captive breeding remains difficult due to the carnivorous diet, aggressive temperament, and high stress susceptibility of this species. These behavioral characteristics necessitate the development of standardized husbandry protocols to ensure animal welfare and production efficiency. The growth phase represents a critical developmental window for juvenile raccoon dogs, during which proper nutrition and management directly affect fur quality and yield [4]. Maintaining the health of raccoon dogs during this period is therefore essential for commercial fur production.
The gut microbiota includes bacteria, archaea, fungi, viruses, phages, and eukaryotic parasites [5]. This microbial ecosystem exerts profound influences on host physiology by metabolizing nutrients [6], sustaining immune homeostasis [7], and modulating growth [8]. Balanced gut microbial composition and stable ecosystem dynamics are essential for proper digestive function, metabolic regulation, and reproductive health in animals. The protective roles of commensal microbiota are mediated through multiple pathways: stimulation of host immune responses, production of bioactive metabolites, and competitive exclusion of pathogenic microorganisms through colonization resistance [9]. As the host’s “second genome,” the gut microbiota exhibits dynamic plasticity in response to factors such as diet, physiological stress, host genetics, and health status [10,11]. Imbalances in gut microbial communities have been associated with various pathological states, from autoimmune disorders [12] and metabolic dysregulation [10,13] to increased susceptibility to infectious diseases [14]. These findings underscore the importance of maintaining microbial homeostasis for optimal raccoon dog husbandry. Notably, the gut microbiota undergoes significant compositional and functional transformations throughout host development [15,16]. It has been reported that there is a correlation between the quality of raccoon dog fur and gut microbiota [17]. Although previous studies have investigated the gut microbial ecology of raccoon dogs, they have mainly focused on specific growth stages. It remains unclear whether these findings can fully reflect the dynamic changes in the gut microbiota across the entire growth period. This knowledge gap hinders the development of evidence-based feeding strategies for this commercially important species.
In this study, fecal samples were collected from raccoon dogs at five key developmental time points (2, 3, 4, 6, and 8 months of age), and gut microbial composition was characterized by 16S rRNA sequencing, aiming to identify the core microbial communities across different growth stages and elucidate their predominant metabolic functions.
2. Materials and Methods
2.1. Sample Collection
The experiment was conducted at the Zuojia Fox and Raccoon dog Breeding Base in the Jilin Province. The breeding farm lies on gently sloped land and includes sufficient sunlight and good air circulation. The housing units adopt steel sheet roofs; the eaves are 1.5–2.0 m high, and each single unit covers a width of 2–4 m. The cages are reinforced with steel frameworks and enclosed by wire mesh. All raccoon dogs were raised individually in 1.0 m × 1.0 m cages, and no artificial temperature and humidity control system was equipped in the breeding facility. Feed was provided at 07:00 and 16:00 every day. The composition of the basal diet is presented in Table 1. A total of 10 healthy male raccoon dogs (4.12 ± 0.47 kg) with consistent growth rates and no history of disease were selected. The same 10 individuals were sampled repeatedly at 2 (M2 group, n = 8), 3 (M3 group, n = 8), 4 (M4 group, n = 8), 6 (M6 group, n = 8), and 8 (M8 group, n = 10) months after birth to track the periodic changes in their gut microbiota. At each sampling time point, 1–2 g of fresh feces was collected in the morning, approximately 2–3 h after the morning feeding. The collected fecal samples were placed in sterile centrifuge tubes and stored in a liquid nitrogen tank immediately after collection. After all samples at each stage were collected, they were transferred to the laboratory’s −80 °C freezer for long-term storage. After all the samples had been collected, they were processed and sequenced in a single batch.
Table 1. Composition of basal diets (DM basis, g/kg).
| Ingredients | Content |
|---|---|
| Extruded corn | 405.00 |
| Soybean meal | 210.00 |
| Meat meal | 50.00 |
| Fish meal | 25.00 |
| Blood meal | 50.00 |
| Distillers dried grains with solubles | 100.00 |
| Corn germ meal | 100.00 |
| Soybean oil | 35.00 |
| CaHPO 4 | 8.00 |
| Limestone powder | 7.00 |
| Premix | 10.00 |
2.2. DNA Extraction and PCR Amplification
Total DNA was extracted from fecal samples with the DNeasy PowerSoil kit (QIAGEN, Hilden, Germany). Subsequently, in order to obtain high-quality total DNA, the samples were centrifuged to remove impurities. Primers specific to conserved regions of the 16S rRNA gene were used for PCR amplification, and the products were purified with magnetic beads to remove small molecular byproducts. PCR product purity was assessed by 1% agarose gel electrophoresis. The V3/V4 regions were then amplified by PCR, verified on a 2% agarose gel [18], and prepared for sequencing by adding platform-specific adapters and sample-specific barcodes. The libraries were then pooled and sequenced on the Illumina platform.
2.3. 16S rRNA Sequencing Data Preprocessing
Raw tags were processed with Fastp (version 0.23.1) to obtain high-quality tag data. Cutadapt (version 1.9.1) was applied to remove primer sequences and reduce noise [19]. The sequences were then compared with the species annotation database (https://www.arb-silva.de/; 20 March 2025) to remove chimeric sequences, yielding the final valid data. ASVs with 97% sequence similarity were clustered together. Microorganisms were identified and classified using the QIIME2 software package version 1.8.0 platform [20,21]. The generated ASV table was exported to R-studio and subjected to ASV clustering, classification, and diversity assessment.
2.4. Alpha Diversity Analysis and Beta Diversity Analysis
QIIME2 software was used to estimate microbial richness and evenness. PCA, principal coordinates analysis (PCoA), non-metric multidimensional scaling (NMDS) analysis of similarities (ANOSIM), and permutational multivariate analysis of variance (PERMANOVA) were conducted to explore group differences [22]. The speccaccum function in the Vegan package of R-studio was used to generate species accumulation curves. UPGMA clustering maps were drawn using the UPGMA.tre function in QIIME2.
2.5. Significance Analysis of Intergroup Differences
The Kruskal–Wallis rank sum test was applied to screen species with differences in component abundance, and linear discriminant analysis (LDA) was used to calculate the effect size of each differential species. To investigate the potential interactions between gut microbiota in different age groups, Spearman correlation analysis was used to screen the top 25 species at the genus level for total abundance. Species with absolute correlation coefficients > 0.6 and significant p-values ≤ 0.05 were retained for further research.
2.6. Functional Prediction
PICRUSt (https://github.com/picrust/picrust2; 20 March 2025) was applied to predict the potential molecular functions of the microbial communities [23]. In order to analyze Tax4Fun, runRefBlast and make Functional Prediction functions were used, with a default reference database and a clustering threshold of 99% for analyzing 16S rRNA genes [24,25]. BugBase was used to predict bacterial composition in sequencing results [26]. Filtered sequences were used to classify sequences with the Silva 138 99% ASV full-length sequence database [27].
3. Results
3.1. The Richness and Diversity of Gut Microbiota
As shown in Figure 1A,C, notable differences were observed in the diversity and richness of the gut microbiota among raccoon dogs of different ages, with the lowest Chao1 and Observed features in the M4 group, indicating that species abundance first decreases and then stabilizes with age. Similarly, the Shannon and Simpson were the lowest in the M4 group, indicating that community diversity first decreased and then increased with age (Figure 1E,F). The Pielou is used to assess species evenness (Figure 1D). Although the M4 group had the lowest Pielou value, the evenness of the gut microbiota increased significantly with age. In addition, the trend of the Dominance value is opposite that of the Pielou index, which further corroborates this result (Figure 1B). A species accumulation boxplot can describe the species composition and abundance of the sample. As the sample size increased, the boxplots tended to flatten, suggesting that the sample size was sufficient for data analysis (Figure 1G). Figure 1H shows that the curve became wider and flatter with age, suggesting that species abundance and evenness increased with age. Venn diagram analysis (Figure 1I,J) showed that the five groups shared eight microbial phyla. In addition, there were 6 microbial species unique to group M2, and 1 bacteria unique to group M6. At the genus level, the total number of species shared by the five groups was 106, with 28, 2, 12, 8, and 19 species unique to groups M2, M3, M4, M6, and M8, respectively.
[图：Gut microbiota richness and diversity (M2, 3, 4, 6: n = 8; M8: n = 10). ( A – F ): Alpha diversity indices among groups.]
3.2. Differences in Gut Microbiota
As shown in Figure 2A, the M2 group differed significantly from the M4, M6, and M8 groups, while the M3 group differed significantly from the M4 and M6 groups. A beta index intergroup difference test is a statistical method used to compare differences between groups. It can be inferred that as age increased, the gut microbiota changed. Figure 2C,D shows clear group separation in the PCoA and NMDS plots. ANOSIM analysis confirmed significant divergence in gut microbial community structure across the groups (R = 0.907, p = 0.001), with greater dissimilarity between groups than within groups (Figure 2E). Figure 2F shows that PERMANOVA illustrated distinct differences in community homogeneity among groups (R2 = 0.553, F = 11.43, p = 0.001). ZH12 had the highest within-group variation, whereas ZH6 displayed the most homogenous gut microbial community. The similarity of gut microbiota and bacterial community structure between groups can be detected through the arithmetic mean unweighted pairwise method (UPGMA). The hierarchical clustering tree was shown in Figure 2G,H. The M2 (ZH601-ZH608) and M3 (ZH901-ZH908) groups clustered together, while the M4 (ZH1201-ZH1208) and M6 (ZH1803-ZH1807, ZH1809-ZH18011) groups were dispersed at both ends, and the M8 (ZH2401-ZH24010) group was closer to the M4 and M6 groups.
[图：Differences in gut microbiota among different groups (M2, 3, 4, 6: n = 8; M8: n = 10). ( A ): Beta index intergroup diff]
3.3. Gut Microbial Community Composition
As shown in Figure 3A, Firmicutes, Bacteroidota, Fusobacteriota, Actinobacteriota, and Proteobacteria dominated the phyla across the groups. At the genus level, Prevotella 9, Streptococcus, Lactobacillus, Peptoclostridium, and Blautia were the dominant genera across the groups (Figure 3B). The species abundance clustering heatmap is used to discover the high or low aggregation content of species in each group. At the phylum level, Cyanobacteria, Firmicutes, and Actinobacteriota were enriched in M2; Fusobacteriota in M4; Euryarchaeota, Spirochaetota, and Campylobacterota in M6; Proteobacteria in M3; and Campylobacterota, Desulfobacterota, Bacteroidota, and Proteobacteria in M8 (Figure 3C). At the genus level, the species with higher abundance in the M4 group were Cetobacterium and Streptococcus; the species with higher abundance in the M2 group were Ligilactobacillus and Megasphaera; the species with higher abundance in the M6 group were Lactobacillus and Prevotella 9; the species with higher abundance in the M3 group were Blautia, Megamonas, and Peptoclostridium; and the species with higher abundance in the M8 group were Prevotella 9 and Turicibacter (Figure 3D).
[图：Gut microbiota community composition (M2, 3, 4, 6: n = 8; M8: n = 10). ( A ): Relative abundance at the phylum level. ( ]
3.4. Differential Abundance of Gut Microbiota
As demonstrated by the Kruskal–Wallis H-test bar plots (Figure 4A,B), the differentially abundant phyla were Firmicutes, Bacteroidota, Fusobacteriota, Actinobacteriota, and Proteobacteria (Figure 4A), and the differentially abundant genera were Prevotella 9, Cetobacterium, Streptococcus, Blautia, and Megasphaera (Figure 4B). LEfSe analysis (LDA score > 4, p < 0.05) is shown in Figure 4C: the most enriched taxa in the M2 group were Firmicutes, Lachnospirales, and Lachnospiraceae, while Prevotella 9 was enriched in the M3 group. The Lactobacillates Bacilli, Lactobacillaceae, Fusobacteriaceae, Fusobacteriia, Fusobacteriota, Fusobacteriales, and Cetobacterium were enriched in the M4 group. Prevotellaceae was the most enriched taxon in the M6 group, whereas the enriched taxa in the M8 group were Bacteroidales, Bacteroidota, Bacteroidia, and Clostridia. Figure 4D shows the evolutionary branching of different species. The radiating circles represent taxonomic levels from phylum to genus, and each small circle indicates a taxon at that level, with its diameter proportional to relative abundance. Taxa without significant differences are uniformly yellow, whereas those with significant differences are colored according to group.
[图：Differential abundance of gut microbiota among different groups (M2, 3, 4, 6: n = 8; M8: n = 10). ( A ): Kruskal–Wallis ]
3.5. Co-Occurrence Network Analysis of Microbial Interactions
The co-occurrence network showed that 73 edges including 57 positive edges and 16 negative edges were in the M2 group, while Prevotellaceae UCG 001, Prevotella 7, and Monoglobus contained numerous correlations (Figure 5A). Group M3 has 157 edges, including 116 positive edges and 41 negative edges, while Lactiplantibacillus, Bifidobacterium, Turicibacter, Slackia, and Lachnoclostridium contained numerous correlations (Figure 5B). Group M4 has 50 edges, including 43 positive edges and 7 negative edges, while Dorea, Staphylococcus, and Prevotella 7 contained numerous correlations (Figure 5C). Group M6 has 62 edges, including 49 positive edges and 13 negative edges, while Slackia, Streptococcus, Methanobrevibacter, and Romboutsia contained numerous correlations (Figure 5D). Group M8 has 42 edges, including 39 positive edges and 3 negative edges, while the Prevotellaceae UCG 001 and Lachnospiraceae NK3A20 groups contained numerous correlations (Figure 5E). There are 117 edges, including 103 positive edges and 14 negative edges in all groups, while Prevotellaceae UCG 001, Negativibacillus, Cetobacterium, Mailhella, and Peptococcus contained numerous correlations (Figure 5F).
[图：The co-occurrence network of the microbial community at the genus level (M2, 3, 4, 6: n = 8; M8: n = 10). ( A ): M2 grou]
3.6. Function Prediction of Gut Microbiota
The top 10 most abundant functions of the gut microbiota at different annotation levels were statistically analyzed using PICRUSt (Figure 6A–C). The dominant functions in level 1 were metabolism, genetic information processing, environmental information processing, cellular processes, human diseases, and organismal systems (Figure 6A). The main functions in level 2 were membrane transport, carbohydrate metabolism, replication and repair, amino acid metabolism, translation, energy metabolism, poorly characterized, nucleotide metabolism, metabolism of cofactors and vitamins, and cellular processes and signaling (Figure 6B). The main functions in level 3 were transporters, general function prediction only, ABC transporters, DNA repair and recombination proteins, ribosome, purine metabolism, pyrimidine metabolism, peptidases, transcription factors, and chromosome (Figure 6C). In addition, we obtained similar results by analyzing the function of gut microbiota through Tax4Fun. The functional differences among the groups have been demonstrated using heatmaps (Figure 6D–F). BugBase is a tool based on 16S rRNA for predicting microbial phenotype and performing intergroup differential analysis. It can predict seven aspects of information, including Gram positive/negative, biofilm formation, pathogenic potential, aerobic/anaerobic, stress tolerant, facultatively anaerobic, and contains mobile elements. Through BugBase analysis, our results indicated that the functions of each group were similar (Figure 6G). Function Annotation of Prokaryotic Taxa (FAPROTAX) is a database developed by the Louca Laboratory of the Department of Biology at the University of Oregon in the United States. The database utilizes current literature on cultured strains to predict functional annotations of prokaryotes such as genera or species. Our results showed that the core functions of each microbiota were mainly chemoheterotrophy, fermentation, animal parasites or symbionts, mammal gut, and human gut (Figure 6H).
[图：Function prediction of gut microbiota (M2, 3, 4, 6: n = 8; M8: n = 10). ( A – C ): PICRUSt analysis in different levels.]
4. Discussion
At present, the dynamic changes in the gut microbiota of raccoon dogs during their growth period are still unclear. This study systematically investigates the compositional dynamics and functional potential of gut microbiota across growth stages, with the ultimate goal of establishing an evidence-based framework for improved husbandry practices in commercial raccoon dog breeding operations. Alpha diversity metrics, including species richness, diversity, and dominance, collectively characterize the microbial complexity within gut ecosystems [28]. Our analysis revealed a distinct temporal pattern in the gut microbiota of raccoon dogs: diversity and richness initially decreased, then increased, and ultimately stabilized during the growth period. This developmental trajectory is consistent with the age-dependent microbial succession reported in other mammalian species [29]. Beta diversity analysis quantified compositional differences between developmental stages [30]. Significant microbial community shifts were detected between early and late growth phases, with PCoA, NMDS, ANOSIM, and PERMANOVA demonstrating clear temporal separation [31]. Hierarchical clustering analysis reinforced these findings, revealing time-dependent restructuring of microbial communities that parallels documented patterns of gut microbiota maturation [32].
Our results indicate that the gut microbiota was predominantly composed of Firmicutes, Bacteroidota, Actinobacteriota, and Fusobacteriota, consistent with previous findings in raccoon dogs [17]. It is worth noting that the gut microbiota mainly consists of Firmicutes and Bacteroidetes, a characteristic distribution pattern observed across mammalian gut ecosystems [33]. The Firmicutes/Bacteroidetes (F/B) ratio is positively correlated with energy absorption [34,35,36]. In this study, the F/B ratio was relatively high during the early growth stage, which may be related to the rapid absorption of energy from the diet to meet the high energy demand for growth. As growth slowed down with age, the F/B ratio gradually decreased. When the raccoon dogs stopped growing at the 8th month, the energy demand further declined, which may also contribute to the decrease in the F/B ratio. Although Actinobacteria have relatively low abundance in the gut microbiota, their decreased abundance in maintaining microbial balance, host immune regulation, and metabolic health can serve as a microbial marker for various chronic diseases [37]. A striking temporal pattern emerged for Fusobacteriota, with a significant increase in abundance at 4 months, which may be beneficial for the synthesis of short chain fatty acids, thereby maintaining the energy supply of intestinal epithelial cells and maintaining barrier integrity [38]. Genus-level analysis identified Prevotella 9 as the predominant taxon across all age groups. As a beneficial member of the Bacteroidetes phylum, Prevotella can efficiently ferment complex polysaccharides such as resistant starch and pectin [39]. Furthermore, short-chain fatty acids (SCFAs) derived from Prevotella improve intestinal barrier function by increasing tight junction protein expression [40]. Notably, we observed a transient bloom of Cetobacterium abundance in the fourth month. Studies have found that Cetobacterium can significantly increase the activity of trypsin, lipase, and brush border enzymes, and supply energy to the intestinal epithelium by producing short-chain fatty acids. In addition, Cetobacterium can upregulate the expression of ZO-1 and Occludin, maintaining intestinal health [41]. Co-occurrence network analysis further revealed that Cetobacterium is correlated with multiple microorganisms, possibly because it enhances its interactions with the gut microbiota by producing vitamin B12 [42]. However, by 6 months of age, Cetobacterium abundance declined significantly, coinciding with the re-establishment of Prevotella 9 dominance.
Although the host–modulatory effects of the gut microbiota are well established, the underlying mechanisms remain incompletely understood [43]. Moreover, while 16S rRNA sequencing effectively profiles microbial composition, it provides limited functional insights. To address this, functional prediction across developmental stages was performed using both Tax4Fun and PICRUSt. Notably, PICRUSt’s utility is constrained by its reliance on outdated reference databases [44], whereas Tax4Fun demonstrates superior predictive accuracy through its refined algorithm and updated genomic repositories [45]. Our dual-analysis approach consistently identified metabolism and genetic information processing as the predominant functional categories.
While this 16S rRNA study provides foundational insights into the dynamics of the raccoon dog gut microbiota, several limitations warrant consideration. During sequencing, biases arising from DNA extraction, PCR amplification, and primer selection, together with limited resolution at the species/strain level and relative abundance data that may not necessarily reflect absolute cell count or activity, constrain the accuracy of the results [46]. Exploring the periodic changes in the gut microbiota of growing raccoon dogs can help identify optimal feeding stages and provide a scientific basis for adjusting diet and management practices to improve health and growth performance. However, this study did not measure growth performance or fur quality; therefore, the relationship between these parameters and the gut microbiota of growing raccoon dogs remains to be explored in future research. Future investigations should also incorporate individual variation analysis, dietary habit correlations, and the integration of livestock management metadata. Such multidimensional approaches will enable a more robust understanding of microbiome–host interactions in commercial fur animal production.
5. Conclusions
Our results revealed significant stage-specific variations in the gut microbiota composition of growing raccoon dogs. At the phylum level, the dominant species of growth raccoon dogs at different times were Firmicutes, Bacteroidota, Actinobacteriota, and Proteobacteria. Genus-level analysis demonstrated significant differential abundance of several key taxa including Prevotella 9, Megasphaera, Holdemanella, Romboutsia, and Fusionnibacter across growth stages. Functional prediction analysis showed that these microbial communities were mainly involved in carbohydrate, amino acid, and energy metabolism.
Author Contributions
Conceptualization, Z.L.; methodology, S.Y.; software, W.N.; validation, C.Y.; formal analysis, C.X.; investigation, L.S. and C.X.; resources, C.X.; data curation, S.Y.; writing—original draft preparation, C.Y.; writing—review and editing, W.N.; visualization, C.Y.; supervision, Z.L.; project administration, Z.L.; funding acquisition, Z.L. All authors have read and agreed to the published version of the manuscript.
Institutional Review Board Statement
All animal experiments were approved by the Ethics Committee of Jilin Agricultural University (No. 20230606001), approval date: 6 June 2024.
Informed Consent Statement
Written informed consent was obtained from the owner of the animals involved in this study.
Data Availability Statement
All data and materials are available from the corresponding authors upon request.
Conflicts of Interest
The authors declare no conflict of interest.
Funding Statement
The authors declare financial support was received for the research, authorship, and/or publication of this article. This work was supported by the National Key Research and Development Program of China (2023YFD1302000).
Footnotes
[内容请查看原文]
References
[内容请查看原文]
Associated Data
Data Availability Statement
All data and materials are available from the corresponding authors upon request.