兽研引言
胆汁酸（BAs）是由肝脏合成的胆固醇衍生物。它们作为强效的代谢和免疫信号分子，参与脂质吸收、胆固醇稳态和肠肝轴信号调控（1, 2）。宿主肝细胞利用胆固醇合成初级胆汁酸，随后经肠道微生物群化学修饰为次级胆汁酸（3）。这种化学修饰通过四种主要转化途径发生：去结合、脱羟基、氧化和差向异构化（4）。参与这些途径的关键酶及相应基因已被充分表征：去结合由胆盐水解酶（BSH）介导；7α-脱羟基由bai操纵子介导；氧化和差向异构化由羟基类固醇脱氢酶（HSDHs）介导（5）。胆汁酸池的组成和功能受多种因素影响，包括宿主遗传、饮食和肠道微生物群，其中肠道微生物群发挥关键作用（6–8）。不同微生物群落的酶活性决定了胆汁酸池的独特组成和功能（9, 10），而不同肠道区域的理化环境通过选择性塑造微生物群落组成进一步调控胆汁酸转化（11）。因此，肠道微生物群与胆汁酸之间存在密切的相互作用（12）。尽管关于胆汁酸的研究日益增多，但关于肠道微生物群在猫科动物胆汁酸转化中作用的研究仍然有限。
值得注意的是，大量研究表明寄生虫感染深刻影响宿主免疫和代谢稳态，并显著改变肠道微生物群的组成和功能（13–16）。在弓形虫感染期间，肠道微生物多样性的变化和特定功能微生物群丰度的改变可能间接影响肠道微生物群转化次级胆汁酸的能力（16）。响应弓形虫感染，宿主经历代谢重编程，其中胆固醇代谢和胆汁酸信号通路在免疫调节和细胞防御中发挥关键作用（17–19）。因此，弓形虫感染诱导的宿主胆固醇代谢改变以及肠道微生物群的结构和功能变化可能影响胆汁酸转化，从而影响宿主的免疫代谢稳态。然而，在弓形虫感染期间宿主如何调节胆固醇水平以及肠道微生物群的胆汁酸转化能力如何变化仍不清楚。
为解决上述问题，我们分析了可用的宏基因组数据集，并整合了先前发表的血清代谢组学数据，以评估弓形虫感染对猫肠道微生物群和次级胆汁酸转化的影响。我们对猫肠道微生物组进行了表征，重点关注参与胆汁酸转化的关键基因，并分析了血清胆固醇水平。本研究阐明了猫肠道微生物群的胆汁酸转化能力以及弓形虫感染对该过程的影响，为弓形虫病期间肠道微生物群相关代谢失调提供了新见解，并为未来宿主-寄生虫-肠道微生物群相互作用的研究奠定了理论基础。
结果
猫肠道微生物组的分析
本研究共分析了2,474个非冗余代表性基因组。这些基因组大小范围为0.66至6.20 Mbp（平均2.35 Mbp），GC含量范围为23.49%至72.53%（平均47.76%）（图1A）。基因组完整度范围为80.01%至100%（平均92.63%），污染度范围为0%至4.99%（平均1.17%）（图1B）。其中，1,613个基因组（65.20%）符合高质量标准（完整度≥90%且污染度<5%）。每个基因组的开放阅读框（ORFs）数量从653至5,248不等（平均2,203）。在分类注释分析中，2,474个基因组被归类于细菌谱系，涵盖17个门、21个纲、48个目、90个科和279个属。绝大多数基因组归属于Bacillota_A（n = 988，39.94%），其次是Actinomycetota（n = 598，24.17%）和Bacteroidota（n = 298，12.05%）。在纲水平上，最丰富的是Clostridia（n = 988，39.94%），其次是Coriobacteriia（n = 454，18.35%）和Bacteroidia（n = 298，12.05%）。在属水平上，基因组涵盖广泛的分类多样性；最丰富的属为Collinsella（n = 303，12.25%）、Blautia_A（n = 139，5.62%）和Bifidobacterium（n = 106，4.28%）（图1C和表S1）。此外，通过基因预测构建了包含5,449,269个基因的微生物基因目录。这些基因组和基因目录为研究猫肠道微生物组的分类和功能提供了重要资源。
[图：收集的猫肠道基因组汇总。（A）基因组的基因组大小（以兆碱基对计，Mbp）和GC含量]
胆汁酸转化相关基因组目录
基于KEGG功能注释，2,474个基因组中有2,045个被预测含有至少一个与次级胆汁酸转化相关的KO。对这2,045个基因组的分析显示，950个基因组携带一个KO，636个基因组携带两个，455个基因组携带三至五个，4个基因组携带六个（表S2）。与各酶或bai基因相关的基因组数量如下：baiN（1,446个基因组）、BSH（766个基因组）、3α-HSDH（631个基因组）、baiH（288个基因组）、7α-HSDH（258个基因组）、baiCD（220个基因组）、7β-HSDH（83个基因组）、3β-HSDH（64个基因组）和baiB（37个基因组）（图2A）。值得注意的是，没有任何基因组与关键酶baiE相关，表明猫肠道微生物组中可能不存在次级胆汁酸转化所需的完整7α-脱羟基途径。
[图：与胆汁酸（BA）转化相关的基因组汇总。（A）以下各项的存在（着色）或缺失（未着色）]
分类学注释显示，这些基因组涵盖12个门、13个纲、34个目、64个科和226个属（表S2）。猫肠道中BA相关KO的存在主要归因于Bacillota_A和Actinomycetota。具体而言，3α-HSDH、7β-HSDH、baiN和baiCD主要与这两个门的基因组相关：Bacillota_A分别贡献了53.09%、18.07%、60.03%和31.36%，而Actinomycetota分别贡献了43.42%、81.93%、20.95%和39.09%。baiB仅存在于Bacillota_A的基因组中（100%）。3β-HSDH和baiH主要存在于Actinomycetota和Bacillota的基因组中，Actinomycetota贡献了50.00%和42.01%，Bacillota贡献了25.00%和31.25%。7α-HSDH主要与Bacillota_A（37.21%）和Campylobacterota（29.07%）的基因组相关，而BSH主要与Actinomycetota（42.82%）和Bacteroidota（20.76%）的基因组相关（图2B）。这些结果表明，次级BA转化是通过肠道微生物群落的共同协作完成的。
猫与其他宿主物种之间参与BA转化的肠道微生物基因组比较
为阐明猫肠道微生物基因组参与BA转化的特异性，我们将来自猫的2,045个BA相关基因组与来自其他宿主物种的基因组进行了比较：人类（2,294个宏基因组组装基因组[MAGs]）、猪（1,411个MAGs）和鸡（2,113个MAGs）。注释分析显示，1,741个人类MAGs、1,162个猪MAGs和1,621个鸡MAGs也参与BA转化。在四个宿主物种中，baiN（K07007）和BSH（K01442）的比例最高（图3A）。在门水平上，Bacillota_A是所有四个宿主物种中BA相关MAGs的主要贡献者。Actinomycetota是猫肠道微生物群中的次要贡献者，占26.61%，而Bacteroidota是其他三个宿主物种的次要贡献者，在人类、猪和鸡肠道微生物群中分别占13.96%、12.39%和15.15%（图3B和表S3）。这些发现揭示了BA转化主要基因组贡献者方面存在宿主物种特异性差异。
[图：不同宿主物种中与胆汁酸（BA）转化相关的肠道基因组汇总。（A）比例]
BA相关基因的分布及区域差异
我们分析了猫肠道微生物群中BA相关基因在不同肠道区域和不同分类群中的分布。随着样本数量的增加，KO丰富度的稀释曲线趋于饱和，表明我们的分析完全覆盖了猫肠道微生物组中与次级BA转化相关的KO基因（图4A）。采用α多样性和β多样性指数评估这些功能基因在不同肠道区域间的分布差异。通过Richness和Shannon指数测定的α多样性表明，大肠（LI）中BA相关基因的多样性和丰富度显著高于小肠（SI）（P < 0.05）（图4B至C）。基于Bray-Curtis距离的主坐标分析（PCoA）显示，大肠和小肠之间BA相关基因的组成存在显著差异（R² = 0.1145，P < 0.05）（图4D）。
[图：不同肠道中与胆汁酸（BA）转化相关的微生物基因丰度、多样性和群落结构]
对与BA转化相关基因相对丰度的分析显示，与baiN和BSH相关的基因在小肠和大肠中均丰度较高，而与3β-HSDH、7β-HSDH和baiB相关的基因丰度较低。同时，与7α-HSDH相关基因的相对丰度在小肠中高于大肠，而与3α-HSDH、baiCD和baiH相关基因的相对丰度在小肠中低于大肠（图4E）。我们还对不同肠道区域中BA相关基因进行了分类学分析。在门水平上，小肠和大肠之间的分布模式存在差异。例如，Fusobacteriota是小肠中编码baiN基因的最主要贡献者，而Bacillota_C在大肠中占主导地位。对于编码BSH的基因，小肠和大肠中丰度最高的门均为Bacillota，其次是Bacillota_A（图4F）。
弓形虫感染扰乱BA转化
为探究弓形虫感染对猫肠道微生物群中BA转化的影响，我们重新分析了感染前后收集的宏基因组数据。Alpha多样性分析显示，仅在感染后第3天的小肠中，BA相关基因的丰富度显著增加，其他部位未观察到显著变化（图5A至B）。基于Bray-Curtis距离的PCoA进一步表明，感染后BA相关基因的组成主要在小肠中发生改变（图5C至D）。这些发现提示，小肠对感染诱导的BA转化变化反应更为敏感，而大肠则保持相对稳定。
[图：跨肠道区域和感染阶段的胆汁酸（BA）转化相关基因分析。（A和B）丰富度]
我们还分析了两个肠道区域中与各种酶或bai基因相关基因相对丰度的变化。结果显示，仅在感染后第3天的小肠中，编码3α-HSDH和BSH的基因相对丰度显著增加，而其他BA相关酶或bai基因未观察到显著变化（图5E至F）。值得注意的是，弓形虫感染后，编码3α-HSDH和BSH的基因相对丰度呈现先升高后降低的趋势，提示弓形虫感染对小肠微生物群BA转化潜力的影响可能具有时间依赖性。为确定导致这些变化的微生物物种，对BA相关基因进行了分类学分析。在门水平上，Bacillota_C在感染后第3天的小肠中显著增加（图5G）。在相对丰度最高的前10个属中，以下属表现出显著变化：Blautia_A、Enterococcus_E和Ligilactobacillus在感染后第3天的小肠中显著增加（图5H至J）。这些结果表明，特定微生物群可能参与了弓形虫感染早期BA转化潜力的改变。
为进一步探究弓形虫感染对BA合成和转化的影响，我们整合了同一批猫的血清代谢组学分析结果。结果显示，与感染前水平相比，感染后胆固醇水平显著升高（P < 0.05，表1），表明弓形虫感染可能影响宿主胆固醇代谢，并与肠道微生物群介导的BA转化改变相关。
TABLE 1. Analysis of differential serum metabolites after T. gondii infection a
| Comparison | Metabolite | VIP | FC | log2FC | P -value | Type |
|---|---|---|---|---|---|---|
| CON vs D3 | Cholesterol | 1.61523 | 92.707 | 6.53461 | 0.01759 | Up |
| CON vs D8 | Cholesterol | 1.9843 | 67.6656 | 6.08035 | 0.00087 | Up |
讨论
为探究猫肠道微生物群在次级胆汁酸转化中的功能作用，我们构建了一个包含2，474个非冗余代表性基因组的目录，以阐明其关键功能特征。分类学注释显示，这些基因组涵盖了广泛的分类学多样性，反映了猫肠道微生物群落的多样性和复杂性。本研究中猫肠道微生物群的优势纲为梭菌纲、红蝽菌纲和拟杆菌纲，这与Ricardo David Avellán-Llaguno等人(20)的研究结果一致，他们报道猫肠道微生物群中拟杆菌纲和梭菌纲的相对丰度较高。功能注释在2，045个基因组中鉴定出4，787个参与胆汁酸转化的基因，表明猫肠道微生物组具有广泛的胆汁酸转化遗传潜力。值得注意的是，与不同胆汁酸相关酶或bai基因簇相关的基因组数量差异显著。在本研究中，baiN（1，446个基因组）和BSH（766个基因组）是最为普遍的两个。BSH催化初级胆汁酸的去结合反应，广泛分布于主要细菌和古菌门中，表明去结合是猫肠道微生物群中的常见功能(21)。相比之下，baiN作为bai基因簇的成员，在大量基因组中被检测到；然而，其他关键bai基因的普遍性则低得多。例如，仅有37个基因组与baiB相关，该基因编码胆汁酸-CoA连接酶，参与7α-脱羟基过程中的底物活化(22)。重要的是，催化7α-脱羟基途径中关键脱水步骤的关键酶baiE在任何基因组中均未被检测到(23)。这些结果表明，尽管bai基因簇的某些组分存在，但猫肠道微生物组中可能缺乏完整的7α-脱羟基途径。这些发现表明，猫体内的胆汁酸转化主要由广泛分布的功能介导，如去结合和羟类固醇脱氢酶介导的修饰，而通过7α-脱羟基完整形成次级胆汁酸的能力可能仅限于极少数微生物群或不存在。
综合分析显示，Bacillota_A和放线菌门是与大多数关键酶和bai基因簇相关的主要门，表明这两个主要门可能在猫肠道微生物群的次级胆汁酸转化中发挥主导作用。Bacillota_A是Bacillota门的一个分支；先前研究表明，Bacillota（特别是梭菌类群）是7α-脱羟基途径和bai基因簇的主要宿主(22，24，25)，这与本研究的发现基本一致。特别值得注意的是，编码7α-HSDH和BSH的基因来源于多种微生物类群，而7β-HSDH主要由放线菌门编码，baiB则仅来源于Bacillota_A。这种分布模式表明，猫肠道微生物群内的胆汁酸转化依赖于由多种细菌组成的相互作用网络，这对于维持肠道微生态的代谢稳态具有重要意义(18，26)。这也表明猫肠道微生物群在次级胆汁酸转化方面存在功能分化，绝大多数微生物群仅参与初步转化步骤。
对不同宿主物种的比较分析显示，各宿主物种肠道微生物基因组中胆汁酸相关MAG的比例高度相似，表明胆汁酸转化是不同宿主物种肠道微生物群中的一项基本且必要的功能(27，28)。BSH在所有四种被检测宿主物种的肠道微生物群中均占主导地位，进一步证实了其在微生物胆汁酸转化中的核心作用。分类学注释结果显示，拟杆菌门在人类、猪和鸡中是一个次要但稳定的贡献者，但在猫科动物中的贡献显著较低。这一差异可能与宿主基因型、饮食和肠道环境有关(29–31)。
通过分析猫科动物不同肠道区域的微生物群，我们揭示了小肠和大肠之间与胆汁酸转化相关的功能基因组成和多样性存在显著差异。这种功能基因的区域特异性分布模式为理解不同肠道区域胆汁酸相关酶的功能差异提供了重要背景。进一步分析胆汁酸相关基因的丰度揭示了不同肠道区域具有不同的功能特征。编码BSH的基因在小肠和大肠中均高度丰富，表明其在胆汁酸转化中的核心作用。值得注意的是，编码7α-HSDH的基因在小肠中富集，提示小肠微生物群可能在胆汁酸7α-羟基的氧化还原反应中发挥关键作用，生成参与后续次级胆汁酸转化的关键中间代谢产物(32)。相反，编码3α-HSDH、baiCD和baiH的基因主要在大肠中富集，表明大肠微生物群可能更倾向于承担胆汁酸转化的中间步骤，包括部分7α-脱羟基反应和胆汁酸中间体的进一步结构修饰(33, 34)。这种功能分布可能是由不同肠道区域的物理化学环境（如pH、氧含量和胆汁酸浓度）对微生物群的选择性塑造所形成的(35)。
在本研究中，α多样性和PCoA分析共同表明，弓形虫感染诱导的胆汁酸相关基因变化主要集中在小肠。感染后第三天小肠中基因多样性和群落结构发生显著变化，而大肠保持相对稳定。这一发现与不同肠道区域的生理特征和免疫应答模式高度吻合。小肠是弓形虫入侵和复制的主要部位，经历炎症和黏膜损伤等剧烈的局部变化。这种破坏可能直接且迅速地影响常驻微生物群，从而驱动功能转变(36, 37)。相比之下，大肠微生物群可能因其更高的复杂性和稳定性而对急性感染表现出更强的抵抗力(38)。
分析了不同肠道区域中与各种酶或bai基因相关的基因丰度，发现感染后小肠微生物群中编码3α-HSDH和BSH的基因相对丰度显著增加。这提示编码这些酶的微生物群丰度增加，可能增强小肠微生物群介导的次级胆汁酸转化潜力，从而影响初级胆汁酸向次级胆汁酸的转化过程。尽管在稳态条件下次级胆汁酸可通过法尼醇X受体信号通路抑制胆汁酸合成，但在感染背景下，这种负反馈可能受到炎症反应和代谢重编程的调节，从而维持活跃的肝脏胆汁酸合成。由于胆固醇是胆汁酸生物合成的唯一前体，宿主可能代偿性上调胆固醇合成以满足增加的需求，导致血清胆固醇水平升高(8, 17, 39, 40)。这些变化表明弓形虫感染可能诱导涉及宿主代谢调控和肠道微生物群功能重塑的协同反应(41)。进一步的分类学分析将这些变化与特定微生物分类群联系起来。Bacillota_C门以及Blautia_A属、Enterococcus_E属和Ligilactobacillus属丰度的显著增加表明，这些微生物分类群可能是感染期间参与或影响胆汁酸转化的关键参与者。相关研究报道Ligilactobacillus可编码BSH酶并参与次级胆汁酸转化(42)，支持其潜在的功能重要性。本研究结果表明，弓形虫感染优先破坏小肠微生物群的胆汁酸转化能力，并伴随关键功能基因和特定微生物分类群的动态变化。
材料与方法
实验设计
所有实验数据均来源于我们先前发表的研究（43）。原始实验设计和样本信息描述如下。简言之，18只4至5月龄的无特定病原体猫（Felis catus）根据感染弓形虫阶段的不同被随机分为三组（每组n=6）。CON组为对照组（感染前）。D3组感染无性繁殖阶段（感染后第3天），D8组感染有性繁殖阶段（感染后第8天）。感染组猫经口接种800个弓形虫PRU株（基因II型）包囊，而对照组猫接受等体积的磷酸盐缓冲液（PBS）。收集小肠和大肠肠道内容物样本进行宏基因组测序，并对相同动物的血清样本在先前研究中进行了非靶向代谢组学分析。
公共数据收集
我们从Recherche Data Gouv（法国国家研究数据存储库）检索了六个公共宏基因组项目：PRJNA758898、PRJEB9357、PRJEB4391、PRJNA944553、PRJNA908260和PRJNA923753。此外，从NCBI数据库获得了两个项目（PRJNA961122和PRJNA1237286）。由于项目PRJNA1237286与第一个数据库中的五个公共项目存在重叠，我们进行了整合和去重以确保每个样本的独特性。最终，我们保留了来自三个独立项目（PRJNA961122、PRJNA1237286和PRJNA923753）的肠道基因组和444个宏基因组样本。
预处理和生物信息学分析
原始测序数据使用FASTP（v0.23.0）（44）进行质量控制和过滤，参数为“-q 20 -u 30 -n 5 -y -y 30 -l 90 -trim_poly_g”。高质量reads使用Bowtie2（v2.5.0）（45）比对到宿主参考基因组以去除宿主来源序列。每个样本的宏基因组组装使用MEGAHIT（v1.2.9）（46）进行以生成contigs，参数为“-k-list 21,41,61,81,101,121,141”。Reads使用BWA-MEM（v0.7.17）（47）比对到组装的contigs，并使用SAMtools（v0.1.19）（48）转换为BAM格式。Contigs的测序深度使用MetaBAT2（v2.12.1）（49）中的jgi_summarize_BAM_contig_depth脚本计算。Contigs的分箱使用MetaBAT2（v2.12.1）（49）进行，参数为“-M 2000 -s 200,000 -saveCls -unbinned”。根据上述步骤，本研究共获得22,197个基因组。
使用CheckM2（v1.0.1）（50）评估基因组的完整度和污染度，并应用以下筛选标准以消除偏倚：完整度≥80%、污染度<5%、质量评分（完整度－5×污染度）≥50%，最终保留10,684个基因组。保留的基因组使用dRep（v3.4.5）（51）进行聚类和去冗余，参数如下：-pa 0.90 -sa 0.99 -nc 0.30。平均核苷酸一致性（ANI）>99%的基因组归为一组，每组中质量评分最高的基因组被视为代表性基因组。总共获得2,474个非冗余代表性基因组用于本研究。此外，使用Prodigal（v2.6.3）（52）预测这些代表性基因组中的开放阅读框（ORFs）。使用GTDB-tk（v2.3.2）（53）进行分类学注释，并使用PhyloPhlAn（v3.0.51）（54）构建系统发育树，使用iTOL进行可视化。
BA相关微生物基因的功能分析
通过使用DIAMOND（v2.1.8.162）（56）和BLASTP（57）将蛋白编码基因序列比对到京都基因与基因组百科全书（KEGG）数据库（55），获得功能注释。基于先前研究（4, 58–60），选择了15个KEGG Orthology（KO）标识符（例如K00076和K01442），这些标识符对应于当前注释的编码参与次级BA转化关键酶的基因，用于基因筛选。这些KO对应于引言中概述的四条主要BA转化途径中涉及的关键酶：去结合（BSH）、氧化/差向异构化（7α-、7β-、3α-和3β-HSDHs）以及7α-脱羟基化（bai基因簇成员，包括baiA、baiB、baiCD、baiE、baiF、baiH、baiI和baiN）。对于样本间比较，使用Bowtie2（v2.5.0）（45）将36个样本的高质量测序reads比对到所选基因。基因的比对read计数归一化为每百万转录本（TPM）。
为了实现跨宿主物种的功能比较，我们纳入了来自人类（61）、猪（62）和鸡（63）的宏基因组组装基因组（MAGs）。这些数据集由我们的研究团队从主流和公开来源收集和整理，均具有代表性。所有MAGs均使用与上述猫肠道微生物基因组相同的分析流程处理。从这些MAGs预测的基因经过与猫肠道微生物基因相同的KEGG注释流程，以确保跨宿主物种的可比性。
血清代谢组学分析
本文所采用的血清代谢组学原始数据来源于我们此前已发表的研究（41）。样品前处理和仪器检测的实验流程简要概述如下。血清样品使用甲醇/乙腈（1：1，体积比）进行蛋白沉淀。经涡旋、低温超声和离心后，收集上清液并在真空浓缩器中蒸干。干燥后的提取物重新溶解于100 μL含5 ppm 2-氯苯丙氨酸的50%甲醇中，再次离心。澄清的上清液随后用于LC–MS/MS分析。从每个样品中取等体积混合制备质控（QC）样品，用于监测系统稳定性。代谢物检测在Thermo Vanquish UHPLC–Orbitrap Exploris 120平台上进行，数据依赖型采集（DDA）在正离子和负离子模式下分别进行。原始数据在Compound Discoverer 3.3中处理，进行峰提取、对齐、过滤和归一化。代谢物鉴定使用mzCloud、LIPID MAPS、人类代谢组数据库（HMDB）、MoNA和NIST_2020_MSMS等数据库进行。
统计分析与可视化
分类学和功能基因丰度数据用于计算丰富度和Shannon多样性指数。β多样性基于Bray-Curtis距离采用主坐标分析（PCoA）进行评估，组间差异采用置换多元方差分析（PERMANOVA）进行评价（64）。采用Wilcoxon秩和检验评估各组间多样性指数、分类单元和功能基因特征丰度差异的显著性。代谢组学差异采用偏最小二乘判别分析（PLS-DA）进行可视化。所有统计分析均在R（v4.4.0）中完成，所有可视化图形均使用“ggplot2”包（v3.5.1）生成。
结论
本研究构建了包含2，474个猫肠道微生物基因组的目录，并整合血清代谢组学数据，以表征胆汁酸（BA）转化及其在弓形虫感染后的变化。结果揭示，猫肠道微生物组具有广泛的BA转化遗传潜力，但缺乏完整的7α-脱羟基途径。此外，参与BA转化的特定功能在不同肠道区域存在差异。小肠微生物群可能在BA 7α-羟基的氧化还原反应中发挥关键作用，而大肠微生物群似乎更倾向于进行BA转化的中间步骤。此外，弓形虫感染破坏了小肠中的BA转化能力，这得到了BA转化相关基因丰度增加和血清胆固醇水平升高的支持。本研究结果为猫弓形虫病期间肠道微生物群相关的代谢紊乱提供了新的见解。
致谢
我们感谢中国上海派森诺生物科技有限公司在宏基因组测序和初步分析中提供的技术协助。
本研究由国家自然科学基金（批准号32172887）、国家重点研发计划（2023YFF0724603）、国家自然科学基金–云南联合基金（批准号U2202201）、山西农业大学高层次人才专项研究基金（批准号2021XG001）和山东省自然科学基金（编号ZR2025MS501）资助。
L.Y.T.：撰写—初稿、形式分析、数据管理、调查研究、方法学、可视化。H.Z.：经费获取、形式分析。J.X.Z.：撰写—审阅与编辑、调查研究、形式分析、数据管理。Y.Q.G.：形式分析、调查研究、撰写—审阅与编辑。X.Z.：形式分析、数据管理、可视化。S.C.X.：调查研究、方法学、资源。Z.Z.：调查研究、方法学、资源。X.Q.Z.：概念化、经费获取、项目管理、资源、监督、撰写—审阅与编辑。X.X.Z.：监督、撰写—审阅与编辑。C.X.：概念化、经费获取、项目管理。R.L.：形式分析、项目管理、监督、撰写—审阅与编辑。H.M.：概念化、形式分析、撰写—审阅与编辑。所有作者均已阅读并批准最终稿件。
贡献者信息
夏长友，电子邮箱：xiachangyou@caas.cn。
刘睿，电子邮箱：201901106@qau.edu.cn。
马贺，电子邮箱：mahe@qau.edu.cn。
Christopher A. Elkins，美国佐治亚州亚特兰大市，疾病控制与预防中心
数据可用性
本研究的宏基因组测序数据已存入美国国家生物技术信息中心（NCBI），BioProject编号为PRJNA1203944。支持本研究结论的所有其他数据均可在论文和补充材料中获取。
伦理批准
所有动物实验程序均经山西农业大学实验动物护理与使用委员会批准（批准号SXAU-EAW-2021XM121001）。所有动物实验程序均符合《中华人民共和国生物安全法》。
补充材料
以下材料可在线获取：https://doi.org/10.1128/aem.00674-26。
[内容请查看原文]
[内容请查看原文]
[内容请查看原文]
ASM 不拥有可能链接至文章或通过文章访问的补充材料的版权。作者已授予 ASM 非独占、全球范围的许可，以发布补充材料文件。如需重复使用，请直接联系通讯作者。
参考文献
[内容请查看原文]
关联数据
补充材料
[内容请查看原文]
[内容请查看原文]
[内容请查看原文]
数据可用性声明
本研究的宏基因组测序数据已存入美国国家生物技术信息中心（NCBI），BioProject 编号为 PRJNA1203944。支持本研究结论的所有其他数据均可在论文及补充材料中获取。
查看英文原文▾
INTRODUCTION
Bile acids (BAs) are cholesterol derivatives synthesized by the liver. They serve as potent metabolic and immune signaling molecules and are involved in lipid absorption, cholesterol homeostasis, and enterohepatic axis signaling regulation (1, 2). Host hepatocytes utilize cholesterol to synthesize primary BAs, which are subsequently chemically modified by the intestinal microbiota into secondary BAs (3). This chemical modification occurs through four main transformation pathways: deconjugation, dehydroxylation, oxidation, and epimerization (4). The key enzymes and corresponding genes involved in these pathways have been well characterized: deconjugation is mediated by bile salt hydrolase (BSH); 7α-dehydroxylation is mediated by the bai operon; and oxidation and epimerization are mediated by hydroxysteroid dehydrogenases (HSDHs) (5). The composition and function of the BA pool are influenced by multiple factors, including host genetics, diet, and the gut microbiota, with the gut microbiota playing a pivotal role (6–8). The enzymatic activities of different microbial communities determine the unique composition and function of the BA pool (9, 10), and the physicochemical environment of different intestinal regions further regulates BA transformation by selectively shaping the composition of the microbial community (11). Therefore, there is a close interaction between the gut microbiota and BAs (12). Although there have been increasing studies on BAs, research on the role of the gut microbiota in BA transformation in felines remains limited.
Notably, numerous studies have demonstrated that parasite infection profoundly impacts host immunity and metabolic homeostasis and significantly alters the composition and function of the gut microbiota (13–16). During Toxoplasma gondii infection, changes in gut microbial diversity and the abundance of specific functional microbial groups may indirectly affect the gut microbiota’s capacity for secondary BA transformation (16). In response to T. gondii infection, the host undergoes metabolic reprogramming, in which cholesterol metabolism and BA signaling pathways play key roles in immune modulation and cellular defense (17–19). Therefore, the alterations in host cholesterol metabolism and the structural and functional changes in the gut microbiota induced by T. gondii infection may influence BA transformation, thereby affecting the host’s immunometabolic homeostasis. However, it remains unclear how the host regulates cholesterol levels and how the BA transformation capacity of the gut microbiota changes during T. gondii infection.
To address the above questions, we analyzed available metagenomic data sets and integrated previously published serum metabolomic data to evaluate the impact of T. gondii infection on the feline gut microbiota and secondary BA transformation. We characterized the feline gut microbiome, focusing on key genes involved in BA transformation, and analyzed serum cholesterol levels. This study elucidated the BA transformation capacity of feline gut microbiota and the impact of T. gondii infection on this process, providing new insights into gut microbiota-related metabolic dysregulation during toxoplasmosis and laying a theoretical foundation for future research on host-parasite-gut microbiota interactions.
RESULTS
Profiling of feline gut microbiome
A total of 2,474 nonredundant representative genomes were analyzed in this study. These genome sizes ranged from 0.66 to 6.20 Mbp (average 2.35 Mbp), with the GC content spanning 23.49% to 72.53% (average 47.76%) (Fig. 1A). The genomes exhibited completeness ranging from 80.01% to 100% (average 92.63%) and contamination ranging from 0% to 4.99% (average 1.17%) (Fig. 1B). Among them, 1,613 genomes (65.20%) met high-quality standards (completeness ≥90% and contamination <5%). The number of open reading frames (ORFs) per genome varied from 653 to 5,248 (average 2,203). For taxonomic annotation analysis, 2,474 genomes were classified within bacterial lineages, covering 17 phyla, 21 classes, 48 orders, 90 families, and 279 genera. The vast majority of genomes were assigned to Bacillota_A (n = 988, 39.94%), followed by Actinomycetota (n = 598, 24.17%) and Bacteroidota (n = 298, 12.05%). At the class level, the most abundant was Clostridia (n = 988, 39.94%), followed by Coriobacteriia (n = 454, 18.35%) and Bacteroidia (n = 298, 12.05%). At the genus level, the genomes spanned a wide taxonomic diversity; the most abundant genera were Collinsella (n = 303, 12.25%), Blautia_A (n = 139, 5.62%), and Bifidobacterium (n = 106, 4.28%) (Fig. 1C and Table S1). Additionally, a microbial gene catalog comprising 5,449,269 genes was constructed by gene prediction. These genome and gene catalogs served as important resources for studying the classification and function of the feline gut microbiome.
[图：Summary of collected feline intestinal genomes. ( A ) The genome size (in megabase pairs, Mbp) and GC content of genomes]
Catalog of genomes related to BA transformation
Based on KEGG functional annotation, 2,045 out of 2,474 genomes were predicted to contain at least one KO associated with secondary BA transformation. Analysis of these 2,045 genomes revealed that 950 genomes carried one KO, 636 genomes carried two, 455 genomes carried three to five, and four genomes carried six (Table S2). The number of genomes related to each enzyme or bai gene was as follows: baiN (1,446 genomes), BSH (766 genomes), 3α-HSDH (631 genomes), baiH (288 genomes), 7α-HSDH (258 genomes), baiCD (220 genomes), 7β-HSDH (83 genomes), 3β-HSDH (64 genomes), and baiB (37 genomes) (Fig. 2A). Notably, none of the genomes were related to the key enzyme baiE, indicating that a complete 7α-dehydroxylation pathway required for secondary BA transformation may not be present in the feline gut microbiome.
[图：Summary of genomes related to bile acid (BA) transformation. ( A ) The presence (colored) or absence (uncolored) of the ]
Taxonomic annotation revealed that these genomes encompassed 12 phyla, 13 classes, 34 orders, 64 families, and 226 genera (Table S2) The presence of BA-related KOs in the feline gut was mainly attributed to Bacillota_A and Actinomycetota. Specifically, 3α-HSDH, 7β-HSDH, baiN, and baiCD were predominantly associated with genomes from these two phyla: Bacillota_A contributed 53.09%, 18.07%, 60.03%, and 31.36%, respectively, while Actinomycetota contributed 43.42%, 81.93%, 20.95%, and 39.09%, respectively. baiB was exclusively found in genomes from Bacillota_A (100%). 3β-HSDH and baiH were mainly present in genomes from Actinomycetota and Bacillota, with Actinomycetota contributing 50.00% and 42.01% and Bacillota contributing 25.00% and 31.25%, respectively. 7α-HSDH was primarily associated with genomes from Bacillota_A (37.21%) and Campylobacterota (29.07%), while BSH was mainly related to genomes from Actinomycetota (42.82%) and Bacteroidota (20.76%) (Fig. 2B). These results indicated that secondary BA transformation was accomplished through the collective collaboration of the gut microbial community.
Comparison of gut microbial genomes involved in BA transformation between felines and other host species
To elucidate the specificity of the feline gut microbial genomes involved in BA transformation, we compared the 2,045 BA-related genomes from cats with those from other host species: humans (2,294 metagenome-assembled genomes [MAGs]), pigs (1,411 MAGs), and chickens (2,113 MAGs). Annotation analysis showed that 1,741 human MAGs, 1,162 pig MAGs, and 1,621 chicken MAGs were also involved in BA transformation. Among the four host species, the proportions of baiN (K07007) and BSH (K01442) were the highest (Fig. 3A). At the phylum level, Bacillota_A was the primary contributor to BA-related MAGs across all four host species. Actinomycetota served as a secondary contributor in the feline gut microbiota, accounting for 26.61%, whereas Bacteroidota was the secondary contributor in the other three host species, accounting for 13.96%, 12.39%, and 15.15% in the human, pig, and chicken gut microbiota, respectively (Fig. 3B and Table S3) These findings revealed host species-specific differences in the primary genomic contributors to BA transformation.
[图：Summary of the intestinal genomes related to bile acid (BA) transformation in different host species. ( A ) Proportion o]
Distribution and regional differences of BA-related genes
We analyzed the distribution of BA-related genes in the gut microbiota of felines across different intestinal regions and among different taxa. As the number of samples increased, the rarefaction curve for KO richness approached saturation, indicating that our analysis fully covered the KO genes related to secondary BA transformation in the feline gut microbiome (Fig. 4A). The alpha and beta diversity indices were used to evaluate the distribution differences of these functional genes among different intestinal regions. The alpha diversity measured by Richness and Shannon indices indicated that the diversity and richness of BA-related genes were significantly higher in the large intestine (LI) than in the small intestine (SI) (P < 0.05) (Fig. 4B through C). Principal coordinates analysis (PCoA) based on Bray-Curtis distance showed a significant difference in the composition of BA-related genes between the large and small intestines (R² = 0.1145, P < 0.05) (Fig. 4D).
[图：Microbial gene abundance, diversity, and community structure related to bile acid (BA) transformation in different intes]
Analysis of the relative abundance of genes associated with BA transformation showed that genes related to baiN and BSH were abundant in both the small and large intestines, whereas genes associated with 3β-HSDH, 7β-HSDH, and baiB were present at low abundance. Meanwhile, the relative abundance of genes related to 7α-HSDH was higher in the small intestine than in the large intestine, whereas the relative abundances of genes associated with 3α-HSDH, baiCD, and baiH were lower in the small intestine than in the large intestine (Fig. 4E). We also performed taxonomic analysis of BA-related genes across different intestinal regions. At the phylum level, the distribution patterns differed between the small and large intestines. For instance, Fusobacteriota was the most abundant contributor to genes encoding baiN in the small intestine, whereas Bacillota_C dominated in the large intestine. For genes encoding BSH, the most abundant phylum in both the small and large intestines was Bacillota, followed by Bacillota_A (Fig. 4F).
T. gondii infection disrupts BA transformation
To investigate the impact of T. gondii infection on BA transformation in the feline gut microbiota, we reanalyzed metagenomic data collected before and after infection. Alpha diversity analysis revealed a significant increase in the richness of BA-related genes only in the small intestine on day 3 post-infection, with no significant changes observed elsewhere (Fig. 5A through B). PCoA based on Bray-Curtis distance further indicated that the composition of BA-related genes was primarily altered in the small intestine following infection (Fig. 5C through D). These findings suggested that the small intestine responded more sensitively to infection-induced changes in BA transformation, whereas the large intestine remained relatively stable.
[图：Analysis of bile acid (BA) transformation-related genes across intestinal regions and infection stages. ( A and B ) Rich]
We also analyzed changes in the relative abundance of genes related to various enzymes or bai genes in both intestinal regions. Results showed that only the relative abundance of genes encoding 3α-HSDH and BSH increased significantly in the small intestine on day 3 post-infection, while no significant changes were observed in other BA-related enzymes or bai genes (Fig. 5E through F). Notably, following T. gondii infection, the relative abundance of genes encoding 3α-HSDH and BSH exhibited a trend of initially increasing and then decreasing, suggesting that the impact of T. gondii infection on the BA transformation potential of the small intestinal microbiota may be time-dependent. To identify the microbial species responsible for these alterations, taxonomic analysis was performed on BA-related genes. At the phylum level, Bacillota_C showed a significant increase in the small intestine on day 3 post-infection (Fig. 5G). Among the top 10 genera with the highest relative abundance, the following exhibited significant changes: Blautia_A, Enterococcus_E, and Ligilactobacillus significantly increased in the small intestine on day 3 post-infection (Fig. 5H through J). These results indicated that specific microbial groups may be involved in the alteration in BA transformation potential during the early stages of T. gondii infection.
To further explore the effects of T. gondii infection on BA synthesis and transformation, we integrated the findings from serum metabolomic analysis of the same cats. The results revealed that cholesterol levels significantly increased after infection compared to pre-infection levels (P < 0.05, Table 1), indicating that T. gondii infection may affect host cholesterol metabolism and be associated with alterations in gut microbiota-mediated BA transformation.
TABLE 1. Analysis of differential serum metabolites after T. gondii infection a
| Comparison | Metabolite | VIP | FC | log2FC | P -value | Type |
|---|---|---|---|---|---|---|
| CON vs D3 | Cholesterol | 1.61523 | 92.707 | 6.53461 | 0.01759 | Up |
| CON vs D8 | Cholesterol | 1.9843 | 67.6656 | 6.08035 | 0.00087 | Up |
DISCUSSION
To investigate the functional role of the feline gut microbiota in secondary BA transformation, we constructed a catalog comprising 2,474 non-redundant representative genomes to illustrate its key functional characteristics. Taxonomic annotation revealed that the genomes encompassed broad taxonomic diversity, reflecting the diversity and complexity of the feline gut microbial community. The dominant classes within the feline gut microbiota in this study were Clostridia, Coriobacteriia, and Bacteroidia, which aligned with findings from Ricardo David Avellán-Llaguno et al. (20), who reported high relative abundances of Bacteroidia and Clostridia in the feline gut microbiota. Functional annotation identified 4,787 genes involved in BA transformation across 2,045 genomes, indicating that the feline gut microbiome harbors a widespread genetic potential for BA transformation. Notably, the number of genomes associated with different BA-related enzymes or bai gene clusters varied significantly. In this study, baiN (1,446 genomes) and BSH (766 genomes) were the two most prevalent. BSH, which catalyzes the deconjugation of primary BAs, is widely distributed across major bacterial and archaeal phyla, indicating that deconjugation was a common function in the feline gut microbiota (21). In contrast, baiN, a member of the bai gene cluster, was detected in a large number of genomes; however, other key bai genes were much less prevalent. For example, only 37 genomes were associated with baiB, which encodes BA-CoA ligase involved in substrate activation during the 7α-dehydroxylation process (22). Importantly, the key enzyme baiE, which catalyzes the essential dehydration step in the 7α-dehydroxylation pathway, was not detected in any genomes (23). These suggested that, although certain components of the bai gene cluster were present, a complete 7α-dehydroxylation pathway is likely absent in the feline gut microbiome. These findings indicated that BA transformation in cats was primarily mediated by widely distributed functions such as deconjugation and hydroxysteroid dehydrogenase-mediated modifications, whereas the capacity for complete secondary BA formation via 7α-dehydroxylation may be limited to a very small fraction of the microbiota or absent.
Comprehensive analysis revealed that Bacillota_A and Actinomycetota were the primary phyla associated with most key enzymes and bai gene clusters, suggesting that these two major phyla may play a dominant role in secondary BA transformation in the feline gut microbiota. Bacillota_A is a branch of the Bacillota phylum; previous studies have indicated that Bacillota (particularly Clostridia taxa) are the primary hosts for the 7α-dehydroxylation pathway and bai gene clusters (22, 24, 25), which is largely consistent with the findings of this study. Of particular note, genes encoding 7α-HSDH and BSH were derived from various microbial taxa, whereas 7β-HSDH was predominantly encoded by Actinomycetota, and baiB was exclusively derived from Bacillota_A. This distribution pattern indicated that BA transformation within the feline gut microbiota relied on an interactive network composed of diverse bacteria, which is significant for maintaining the metabolic homeostasis of the gut microecology (18, 26). It also suggested functional differentiation within the feline gut microbiota regarding secondary BA transformation, with the vast majority of microbiota participating only in preliminary transformation steps.
Comparative analysis of different host species showed that the proportions of BA-related MAGs in the gut microbial genomes of each host species were highly similar, indicating that BA transformation is a fundamental and essential function in the gut microbiota of different host species (27, 28). BSH was predominant in the gut microbiota of all four host species examined, further confirming its central role in microbial BA transformation. Taxonomic annotation results showed that the phylum Bacteroidota was a minor yet stable contributor in humans, pigs, and chickens, but its contribution was significantly lower in felines. This difference may be related to host genotype, diet, and intestinal environment (29–31).
By analyzing the microbiota from different intestinal regions of felines, we revealed significant differences in the composition and diversity of functional genes related to BA transformation between the small and large intestines. This region-specific distribution pattern of functional genes provided an important background for understanding the functional differences in BA-related enzymes across different intestinal regions. Further analysis of the abundance of BA-related genes uncovered distinct functional characteristics in different intestinal regions. Genes encoding BSH were highly abundant in both the small and large intestines, demonstrating its core role in BA transformation. Notably, genes encoding 7α-HSDH were enriched in the small intestine, suggesting that small intestinal microbiota may play a key role in the oxidation-reduction reactions of the 7α-hydroxy group of BAs, generating critical intermediate metabolites involved in subsequent secondary BA transformation (32). Conversely, genes encoding 3α-HSDH, baiCD, and baiH were primarily enriched in the large intestine, indicating that the large intestinal microbiota may be more inclined to undertake the intermediate steps of BA transformation, including partial 7α-dehydroxylation reaction and further structural modification of bile acid intermediates (33, 34). This functional distribution may have been formed by the selective shaping of the microbiota by the physical and chemical environments of different intestinal regions (e.g., pH, oxygen content, and BA concentration) (35).
In this study, alpha diversity and PCoA analyses collectively indicated that changes in BA-related genes induced by T. gondii infection were primarily concentrated in the small intestine. Significant changes in the diversity of genes and community structure in the small intestine on the third day post-infection occurred, while the large intestine remained relatively stable. This finding aligned well with the physiological characteristics and immune response patterns of different gut regions. The small intestine is the primary site for T. gondii invasion and replication, experiencing drastic local changes such as inflammation and mucosal damage. This disruption may directly and rapidly impact the resident microbiota, thereby driving functional shifts (36, 37). In contrast, the microbiota of the large intestine may exhibit stronger resistance to acute infections due to its higher complexity and stability (38).
The abundance of genes related to various enzymes or bai genes in different intestinal regions was analyzed, and it was found that the relative abundance of genes encoding 3α-HSDH and BSH in the small intestinal microbiota significantly increased after infection. This suggested an increase in the abundance of microbiota encoding these enzymes, which may enhance the potential for secondary BA transformation mediated by the small intestinal microbiota, thereby affecting the transformation process from primary BAs to secondary BAs. Although secondary BAs can inhibit BA synthesis via the farnesoid X receptor signaling pathway under homeostatic conditions, in the context of infection, this negative feedback may be modulated by inflammatory responses and metabolic reprogramming, thereby sustaining active hepatic BA synthesis. As cholesterol is the sole precursor for BA biosynthesis, the host may compensatorily upregulate cholesterol synthesis to meet the increased demand, resulting in elevated serum cholesterol levels (8, 17, 39, 40). These changes indicated that T. gondii infection may induce a coordinated response involving host metabolic regulation and functional remodeling of the gut microbiota (41). Further taxonomic analysis linked these changes to specific microbial taxa. The significant increase in the abundance of the phylum Bacillota_C and the genera Blautia_A, Enterococcus_E, and Ligilactobacillus indicated that these microbial taxa may be key participants involved in or influencing BA transformation during infection. Relevant studies have reported that Ligilactobacillus can encode BSH enzymes and participate in secondary BA transformation (42), supporting its potential functional importance. The results of this study demonstrated that T. gondii infection preferentially disrupted the BA transformation capacity of the small intestinal microbiota, accompanied by dynamic changes in key functional genes and specific microbial taxa.
MATERIALS AND METHODS
Experimental design
All the experimental data were derived from the study we published previously (43). The original experimental design and sample information are described below. Briefly, 18 4-month- to 5-month-old specific-pathogen-free cats (Felis catus) were randomly divided into three groups (n = 6 per group) based on the different Toxoplasma gondii stages they were infected with. Group CON served as the control group (pre-infection). Group D3 was infected with the asexual reproduction stage (day 3 post-infection), and group D8 was infected with the sexual reproduction stage (day 8 post-infection). Cats in the infected groups were orally inoculated with 800 cysts of the Toxoplasma gondii PRU strain (Genotype II), while cats in the control group received an equivalent volume of phosphate-buffered saline (PBS). Intestinal content samples from the small and large intestines were collected for metagenomic sequencing, and serum samples from the same animals were subjected to untargeted metabolomics analysis in the previous study.
Public data collection
We retrieved six public metagenomic projects from Recherche Data Gouv (the French national research data repository): PRJNA758898, PRJEB9357, PRJEB4391, PRJNA944553, PRJNA908260, and PRJNA923753. Additionally, two projects (PRJNA961122 and PRJNA1237286) were obtained from the NCBI database. Since project PRJNA1237286 overlapped with the five public projects from the first database, we performed integration and deduplication to ensure the uniqueness of each sample. Ultimately, we retained intestinal genomes from three independent projects (PRJNA961122, PRJNA1237286, and PRJNA923753) and 444 metagenomic samples.
Preprocessing and bioinformatic analysis
Raw sequencing data were subjected to quality control and filtering using FASTP (v0.23.0) (44) with the parameters “-q 20 -u 30 -n 5 -y -y 30 -l 90 -trim_poly_g.” High-quality reads were aligned to the host reference genome using Bowtie2 (v2.5.0) (45) to remove host-derived sequences. Metagenomic assembly of each sample was performed with MEGAHIT (v1.2.9) (46) to generate contigs, using the parameter “-k-list 21,41,61,81,101,121,141.” Reads were mapped to the assembled contigs using BWA-MEM (v0.7.17) (47) and converted to BAM format with SAMtools (v0.1.19) (48). Sequencing depth for contigs was calculated using the jgi_summarize_BAM_contig_depth script from MetaBAT2 (v2.12.1) (49). Binning of contigs was conducted with MetaBAT2 (v2.12.1) (49) using the parameters “-M 2000 -s 200,000 -saveCls -unbinned.” According to the above steps, a total of 22,197 genomes were retrieved in this study.
CheckM2 (v1.0.1) (50) was used to estimate the completeness and contamination of the genomes, and the following selection criteria were applied to eliminate bias: completeness ≥80%, contamination <5%, and a quality score (completeness – 5 × contamination) ≥50%, resulting in the retention of 10,684 genomes. The retained genomes were clustered and dereplicated using dRep (v3.4.5) (51) with the following parameters: -pa 0.90 -sa 0.99 -nc 0.30. Genomes > 99% average nucleotide identity (ANI) were grouped together, and the genome with the highest quality score in each group was considered as the representative genome. In total, 2,474 nonredundant representative genomes were obtained for use in this study. Additionally, open reading frames (ORFs) in these representative genomes were predicted using Prodigal (v2.6.3) (52). Taxonomic annotation was assigned with GTDB-tk (v2.3.2) (53), and a phylogenetic tree was constructed using PhyloPhlAn (v3.0.51) (54) and visualized with iTOL.
Functional analysis of BA-related microbial genes
Functional annotations for protein-coding genes were obtained by aligning their sequences against the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (55) using DIAMOND (v2.1.8.162) (56) and BLASTP (57). Based on previous studies (4, 58–60), 15 KEGG Orthology (KO) identifiers (e.g., K00076 and K01442) that correspond to the currently annotated genes encoding key enzymes involved in secondary BA transformation were selected for gene screening. These KOs correspond to key enzymes involved in the four main BA transformation pathways outlined in the introduction: deconjugation (BSH), oxidation/epimerization (7α-, 7β-, 3α-, and 3β-HSDHs), and 7α-dehydroxylation (members of the bai gene cluster, including baiA, baiB, baiCD, baiE, baiF, baiH, baiI, and baiN). For inter-sample comparison, Bowtie2 (v2.5.0) (45) was used to align high-quality sequencing reads from the 36 samples to the selected genes. The mapped read count of the gene was normalized to transcripts per million (TPM).
To enable functional comparison across host species, we included metagenome-assembled genomes (MAGs) from humans (61), pigs (62), and chickens (63). These data sets were collected and compiled by our research team from mainstream and publicly available sources and are all representative. All MAGs were processed using the same analytical pipeline as the feline gut microbial genomes described above. Genes predicted from these MAGs were subjected to the same KEGG annotation pipeline as the feline gut microbial genes to ensure comparability across host species.
Serum metabolomics analysis
The serum metabolomic raw data adopted herein were generated in our previously published study (41). The experimental procedures for sample pretreatment and instrumental detection are briefly summarized below. Serum samples were subjected to protein precipitation using methanol/acetonitrile (1:1, vol/vol). After vortexing, cryo-sonication, and centrifugation, the supernatant was collected and evaporated in a vacuum concentrator. The dried extract was re-dissolved in 100 μL of 50% methanol supplemented with 5 ppm 2-chlorophenylalanine and centrifuged again. The clarified supernatant was subsequently used for LC–MS/MS analysis. An equal volume from each sample was pooled to prepare quality control (QC) samples for monitoring system stability. Metabolite detection was performed on a Thermo Vanquish UHPLC–Orbitrap Exploris 120 platform, with data-dependent acquisition (DDA) conducted in both positive and negative ion modes. Raw data were processed in Compound Discoverer 3.3 for peak extraction, alignment, filtering, and normalization. Metabolite identification was carried out using databases including mzCloud, LIPID MAPS, the Human Metabolome Database (HMDB), MoNA, and NIST_2020_MSMS.
Statistical analysis and visualization
Taxonomic and functional gene abundance data were used to calculate richness and the Shannon diversity index. β-diversity was assessed using Principal Coordinates Analysis (PCoA) based on Bray-Curtis distance, and differences between groups were evaluated using permutational multivariate analysis of variance (PERMANOVA) (64). The Wilcoxon rank-sum test was utilized to assess the significance of differences in diversity indices, taxonomic units, and functional gene feature abundances among groups. The metabolomics differences were visualized using partial least squares discriminant analysis (PLS-DA). All statistical analyses were conducted in R (v4.4.0), and all visualizations were produced using the “ggplot2” package (v3.5.1).
Conclusion
In this study, we constructed a catalog of 2,474 feline gut microbial genomes and integrated serum metabolomic data to characterize bile acid (BA) transformation and its alteration upon T. gondii infection. The results revealed that the feline gut microbiome harbored widespread genetic potential for BA transformation but lacked the complete 7α-dehydroxylation pathway. Furthermore, the specific functions involved in BA transformation differed across intestinal regions. The small intestinal microbiota may play a key role in the oxidation-reduction reactions of the BA 7α-hydroxy group, whereas the large intestinal microbiota appeared to be more inclined to undergo the intermediate steps of BA transformation. Additionally, T. gondii infection disrupted BA transformation capacity in the small intestine, supported by increased abundance of BA transformation-related genes and elevated serum cholesterol levels. The findings of this study provided new insights into gut microbiota-associated metabolic perturbations during feline toxoplasmosis.
ACKNOWLEDGMENTS
We thank Shanghai Personal Biotechnology Co., Ltd., China, for technical assistance in metagenomic sequencing and preliminary analyses.
The study was supported by the National Natural Science Foundation of China (Grant No. 32172887), the National Key R&D Program of China (2023YFF0724603), the NSFC–Yunnan Joint Fund (Grant No. U2202201), the Special Research Fund of Shanxi Agricultural University for Highhigh-level Talentstalents (Grant No. 2021XG001), and the Shandong Provincial Natural Science Foundation (No. ZR2025MS501).
L.Y.T.: Writing – original draft, Formal analysis, Data curation, Investigation, Methodology, and Visualization. H.Z.: Funding acquisition and Formal analysis. J.X.Z.: Writing – review & editing, Investigation, Formal analysis, and Data curation. Y.Q.G.: Formal analysis, Investigation, and Writing – review & editing. X.Z.: Formal analysis, Data curation, and Visualization. S.C.X.: Investigation, Methodology, and Resources. Z.Z.: Investigation, Methodology, and Resources. X.Q.Z.: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, and Writing – review & editing. X.X.Z.: Supervision and Writing – review & editing. C.X.: Conceptualization, Funding acquisition, and Project administration. R.L.: Formal analysis, Project administration, Supervision, and Writing – review & editing. H.M.: Conceptualization, Formal analysis, and Writing – review & editing. All authors have read and approved the final manuscript.
Contributor Information
Changyou Xia, Email: xiachangyou@caas.cn.
Rui Liu, Email: 201901106@qau.edu.cn.
He Ma, Email: mahe@qau.edu.cn.
Christopher A. Elkins, Centers for Disease Control and Prevention, Atlanta, Georgia, USA
DATA AVAILABILITY
Metagenomic sequencing data for this study have been deposited in the National Center for Biotechnology Information (NCBI) under BioProject no. PRJNA1203944. All other data supporting the findings of this study are available in the paper and supplemental material.
ETHICS APPROVAL
All animal experimental procedures were approved by the Institutional Animal Care and Use Committee of Shanxi Agricultural University (Approval No. SXAU-EAW-2021XM121001). All animal experimental procedures were in compliance with the Biosafety Law of the People's Republic of China.
SUPPLEMENTAL MATERIAL
The following material is available online at https://doi.org/10.1128/aem.00674-26.
[内容请查看原文]
[内容请查看原文]
[内容请查看原文]
ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.
REFERENCES
[内容请查看原文]
Associated Data
Supplementary Materials
[内容请查看原文]
[内容请查看原文]
[内容请查看原文]
Data Availability Statement
Metagenomic sequencing data for this study have been deposited in the National Center for Biotechnology Information (NCBI) under BioProject no. PRJNA1203944. All other data supporting the findings of this study are available in the paper and supplemental material.