Seventy-two specific pathogen free male BALB/c mice weighing 16g were purchased from the Animal Center of the Nanjing General Hospital of Chinese People’s Liberation Army. The mice were housed in JV222 IVC cages under a 12–12-h dark–light cycle with food pellets and water ad libitum. The room temperature maintained 20-22 °C. All animal experiments complied with the ARRIVE guidelines and were conducted in accordance with the U.K. Animals (Scientific Procedures) Act, 1986 and the National Research Council’s Guide for the Care and Use of Laboratory Animals. All experimental protocols used in this study, including animal husbandry and slaughter, were approved by the Institution of Animal Science and Welfare of Jiangsu Province (No. JSIASWAP2022070393).

The treatment of the mice was showed in Supplementary Table 1 . After adaptive feeding for 7 d, 10 average body weight (24.76 ± 0.43g) in total of 72 mice was selected and sacrificed by cervical dislocation in a bio-safety cabinet. The organs of heart, lung, and liver were sampled. Organs were cut into small pieces and allocated in 2mL sterilized tubes to prepare the flushing liquid. One copy of the flushing liquid from the heart was disposed of to perform culturomic analysis.

A Samples of the lung, heart and liver (0.2 g) were immersed in 20mL sterilized ice cold double distilled water in a 9 cm bacteria-free plate for 20 min, and the tissues were flashed with the immersed liquid. Then flushing soak was harvested and precipitated in 5000 rpm/min for 10 min to collect the bacterial cells. To increase the harvesting production of cells, the supernatant was re-used to flush the tissues, and employed with a repeated precipitation (5000 rpm/min for 10 min). The pellet was resuspended in Tris-HCl (50mM, pH 6.8) to extract total DNA. A volume of 1mL of the cardiac flushing liquid was centrifuged at the speed of 5000rpm/min to get the sediment. Sediments were dissolved in Tris-HCl for DNA extraction.

One total of 1 μg DNA per sample was used as the input material for the DNA sample preparation. Sequencing libraries were generated and enriched using a reaction of polymerase chain reaction (PCR). TruSeq™ DNA Sample Prep Kit (Abcam, USA) was used to extract the total samples DNA following the manufacturer’s protocol, which were fragmented by sonication to a size of 300 base pairs. The DNA fragments were end-polished, A-tailed, and ligated with a full-length adaptor as a bridge for PCR amplification (HiSeq 3000/4000 PE Cluster Kit). Finally, the PCR products were sequenced using polymerase and four fluorescently labeled base pairs (HiSeq 3000/4000 SBS Kits). After cluster generation, the library preparations were sequenced on an Illumina HiSeq platform (16). The low-quality raw reads containing N reads in the primary sequencing data were filtered to obtain a clean read. MEGAHIT (v1.1.2 https://github.com/voutcn/megahit) and CD-HIT software (v4.6.1 http://weizhongli-lab.org/cd-hit/) were used. To remove interference originating from host genetic reads, BWA software (v0.7.17 http://bio-bwa.sourceforge.net/) was used to compare with the Ensembl Release 104 database (https://ensembl.org/index.html). The Software SOA Paligner (soap2.21release https://github.com/ShujiaHuang/SOAPaligner) was used to align the nucleotide sequences. Beta diversity distance measurements were performed with weighted UniFrac to investigate the structural variation in microbial communities across samples, and then visualized via principal coordinate analysis (PCoA) and principal component analysis (PCA). In addition, Clusters of Orthologous Groups (COG, 2020; http://eggnog5.embl.de/#/app/downloads) was used to predict the function of the microbiota in the organs.

The raw Illumina sequencing data of the heart, liver, and lungs were deposited in the Sequence Read Archive database (SRP) of NCBI (SRR29213176, 29234301, 29240348). The BioProject accession numbers were PRJNA1117782, 1118213, and 1118345.

The classic culture method combined with sequencing was used to verify the presence of lactic acid bacteria in the heart (17). Sterilized de Man, Rogosa, and Sharpe (MRS) broth and Yeast Extract Peptone Dextrose (YEPD) medium were prepared (18). A volume of 0.1 mL of flushing liquid of heart were cultured in MRS broth and YEPD medium. After cultured for 18 h, 0.1 mL of the fermented liquids was soaked to test its inhibitory action on pathogenic E. coli 25922 (purchased from Central Microorganism Storage, Beijing, China). E. coli 25922 cells were cultured in Luria-Bertani (LB) broth. Then, 0.1mL of the bacterial fluid was spread evenly onto plates containing MacConkey solid culture medium. The inhibitory action of the fermented liquids was measured using the cylinder-plate method (19). Then, the plates were laid in facultative anaerobic cultivation for 20 h at 37 °C to observe the flora of bacteria. The composition of bacteria in the heart and their fermentation in the MRS were measured using metagenomic sequencing. The pH of fermentation in the MRS was also tested.

Raw Illumina sequencing data of the heart were deposited in the Sequence Read Archive database (SRP) of NCBI (SRR29254823). BioProject accession number: PRJNA1119040.

Through the assays described in the first section, the intrinsic bacteria existing in the heart and liver and their potential function in the inhibition of pathogenic E. coli 25922 (purchased from Central Microorganism Storage, Beijing, China) were detected. To further evaluate the role of Ligilactobacillus salivarius in modulating the translocation of gastrointestinal core bacteria to the liver and lungs under heat stress, another animal experiment was conducted.

Ligilactobacillus salivarius was isolated by our research group at Jiangsu Vocational College of Agriculture and Forestry. It was stored at the China General Microbiological Culture Collection Center (CGMCC), Beijing, China (storage number CGMCC17718). The bacteria was inoculated in MRS medium following 1% inoculum size, cultured in 37 °C for 16 h. The live number of bacteria reached 1.6×10¹⁰ colony forming units per milliliter (CFU/mL).

Forty mice with an average body weight of 20 g were used in this study. Twenty mice were allocated to Ligilactobacillus salivarius supplemented group, and another 20 mice were fed as the control indicated in Supplementary Table 2 . There were four replicates of each group, with five mice per group. Feed and management were in agreement with those described in a previous trial. Mice were housed in an air conditioned room at 37 °C to create heat stress. In the supplemented group, mice were infused with 0.4 mL fermentation liquid of Ligilactobacillus salivarius. In the gavage administration of 20 g weight of mice, the volume of infusion cannot exceed 0.5mL. Hence, a volume of 0.4mL is suitable. After Ligilactobacillus salivarius cultured, the fermentation solution was diluted with saline until the live number reached 1×10⁸ CFU/mL according to the suggested dosage (20).

After feeding for 7 d, two mice with average body weights of 27.93 ± 0.39 g were selected from each replicate. The heart, liver, and segment of the ileum of eight mice from each group were harvested in an aseptic environment.

Samples (0.2 g) of the liver and heart were cut into small pieces and washed with a flushing liquid, as previously described. The flushing liquid was then harvested. The ileal samples were cut into 2-cm pieces, which were spread out, and the mucosa was scraped from the intestine with sterilized slides after the contents had been washed with Tris-buffered saline containing 0.1% Tween 20 to extract DNA. The total DNA was extracted using the E.Z.N.A.^® tissue DNA Kit (Omega Bio-tek, Norcross, GA, USA) according to the manufacturer’s protocol. DNA concentration and purity were determined using a NanoDrop 2000 UV-vis spectrophotometer (Thermo Scientific, Waltham, MA, USA), and DNA quality was determined by 1% agarose gel electrophoresis. The V3-V4 hypervariable regions of the bacterial 16S rRNA gene were amplified with primers 338F (5-ACTCCTACGGGAGGCAGCAG-3′ and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) using a thermocycler PCR system (GeneAmp 9700, Applied biosystems, Foster City, CA, USA). Polymerase Chain Reaction (PCR) was conducted as follows: 3 min of denaturation at 95 °C, 27 cycles: 30 s at 95 °C, 30 s of annealing at 55 °C, 45 s of elongation at 72 °C, and a final extension at 72 °C for 10 min. PCR was performed in triplicate in 20-μL mixtures containing 4 μL of 5×FastPfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of FastPfu polymerase, and 10 ng of template DNA. The resulting PCR products were analyzed by electrophoresis on a 2% agarose gel to detect the copies of bacteria in different samples (21).

The PCR products were extracted from a 2% agarose gel and further purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) and quantified using QuantiFluor™-ST (Promega, Madison, WI, USA) according to the manufacturer’s protocol. Purified amplicons were pooled at equimolar concentrations and paired-end sequencing was performed (2 × 300) on an Illumina MiSeq platform (Illumina, San Diego, CA, USA) according to standard protocols. Raw Illumina sequencing data were deposited in the Sequence Read Archive database (SRP) of NCBI SRR18190482. The BioProject accession number is PRJNA811797. Diversity metrics were calculated using the core diversity plugin in QIIME2 (21). The alpha diversity was conducted through pan/core analysis. The index of operational taxonomic units (OTUs) were used to estimate the microbial diversity within an individual sample. In addition, potential Kyoto Encyclopedia of Genes and Genomes (KEGG) (22) ortholog functional profiles of the microbial communities were predicted using PICRUSt. Raw Illumina sequencing data were deposited in the Sequence Read Archive database (SRP) of NCBI (SRR18190482). The BioProject accession number is PRJNA811797.

For mice supplemented with Ligilactobacillus salivarius supplementation, the heart, liver, and ileum were harvested. Translocation of Lactobacillus reuteri to the heart and liver was investigated using FISH (23). The probe was designed based on the 16S ribosomal sequence of Lactobacillus reuteri (24), which was isolated and stored at the China Center for Type Culture Collection (CCTCC), Wuhan University, Wuhan, China (strain number CCTCC M2015660). Probes with carboxytetramethylrhodamine were designed and conjugated to the DNA of Lactobacillus retuteri with sufficient length to ensure specific binding ( Supplementary Table S1 ). Organs (0.2 g) such as the liver, heart, and ileal mucous membrane were harvested and homogenized with 1.8mL ice cold saline, and 50μL of homogenate was fixed by immersion in 2mL 10% formaldehyde for 24 h, which was transferred to poly-L-lysine-coated slides and air-dried on a sterile benchtop for 3 h. Samples were then incubated with lysozyme at 32 °C for 10 min, washed with distilled water, immersed in 70% ethanol for 2 min, and air-dried. The probe was diluted to 60 nM before use. 12 μL of probe were then added to the tissue, followed by incubation at 46 °C for 12 h, and washed with phosphate buffer solution (pH 7.4). The tissue was stained with 4′,6-diamidino-2-phenylindole for 5 min and washed thrice with distilled water for 5 min each. After drying, the slides were mounted using Fluoromount-GTM (Abcam, Cambridge, UK) and observed under a fluorescence microscope (BX53; Olympus, Tokyo, Japan).

Strain of Lactobacillus reuteri was cultured in beef extract peptone medium and a tenfold dilution series of Lactobacillus reuteri was performed. The CFU of Lactobacillus reuteri were counted using the plate method under a microscope to obtain samples of 1 × 10⁴, 10⁵, 10⁶, and 10⁷. Total RNA in each dilution was extracted using the RNA Extraction Kit (Invitrogen, Carlsbad, CA, USA). Reverse transcription was performed using the GoScript Reverse System (Invitrogen). Primers were designed according to the 16S ribosomal RNA of Lactobacillus reuteri, and are listed in Supplementary Table 3 . Quantitative PCRs was performed using SuperReal PreMix Plus (SYBR Green, FP205) on an ABI 7900HT Fast Real-Time PCR System. Relative expression levels of target genes were quantitatively normalized against the expression of β-actin using the 2^-ΔΔCT method. First-strand cDNA was synthesized by incubating a reaction mixture containing 11 μL RNA and 1 μL RNase-free dH2O at 70 °C for 3 min, followed by 0 °C for 5 min. The cDNAs was also used for PCR. A dNTP mixture (1 μL; 10 mmol/l), 4 μL GoScript 5X reaction buffer, 1 μL GoScript reverse transcriptase, 1.5 μL Mg²⁺ (25 mM), and 0.5 μL RNase inhibitor were combined in a total volume of 20 μL and incubated at 37 °C in a water bath. A standard curve was established based on the number of Lactobacillus reuteri and CT values. Samples of the ileum, heart, and liver from mice with and without Ligilactobacillus salivarius supplementation were harvested. 0.15 g sample was used to extract total RNA, and qPCR was performed as described above. The levels of Lactobacillus reuteri in the samples were converted to real numbers using a standard curve.

Samples of the heart, liver, and ileal mucosa (0.15 g) were weighed and placed in 5 mL sterilized tubes. Tissues were incubated and digested with 100 U/mL collagenase (Sigma, USA) buffer (20 μg/mL neutral protease II, 32 mM HEPES, 127.5 mM NaCl, 3.15 mM KCl, pH 7.60, 37 °C) for 60 min. The cell suspension was filtered through mesh (grid size about 150 μm) and further ten diluted with1.35 mL ice cold saline to collect the cellular samples for reactive oxygen species (ROS) and superoxide dismutase (SOD) assays.

For ROS test (25), a volume of 30 μL cellular sample was sucked and added into 270 μL dimethyl sulfoxide dilution (10mM) contained 5 μmol/L 2,7-dichlorofuorescin diacetate as a fluorescence primer can be hydrolyzed into dichlorofluorescin in cells and were oxidized in strong green fluorescence. Mix the samples and reagent, and incubate the tubes in 37 °C for 30 min. The tubes were centrifuged at 5, 000 rpm/min for 5 min, and the supernatant was removed. The precipitate was resuspended in ice-cold, double-distilled water. 200 μL liquid was added into 96 well plate for fluorescence detection for excitation and emission wavelength was 488 nm and 525 nm respectively. The results are shown as fluorescence values.

For SOD measurement (20), a volume of 30 μL cellular sample was sucked and added into 270 μL 0.01 mol/L phosphate buffer solution. The manufacturing protocol followed the instructions of the testing kit (Nanjing Jiancheng Bioengineering Institute, China).

Body weight, qRT-PCR, and sequencing data were subjected to one-way ANOVA using the GLM procedure in SPSS, with significance reported at P < 0.05. The correlation of bacterial composition with anti-oxidation was measured with Spearman’s analysis (R,version 3.3.1, pheatmap package).

The intrinsic microbiota can be detected in the heart and liver of mice. After removing the genetic interface from the host, 842, 179, and 176 valid clean reads were harvested comprising 99.39% of the total genome. The heart had the lowest richness of bacteria; only 157 OTUs were detected, and 159 OTUs were found in the cultivation of cardiac flushing liquid, which was both lower than those in the liver and lungs, with 218 and 431 OTUs respectively manifested in Figure 1A . The overlapping parts of the OTUs between the lungs and liver were greater than those between the lungs and cardiac flushing liquid. At the genus level, Burkholderia was the primary bacterial species found in the heart, liver, and lungs. However, the bacterial composition of the liver differed. Some strains of Klebsiella pneumoniae, Shigella sonnei, and unclassified Enterobacteriaceae were also detected in the liver. In addition to more strains of Streptomyces samsunensis, unclassified _g_ Rhodococcus was found in the lungs treated with streptomycin and ampicillin ( Figure 1B ). The sequencing data further validated strains of Lactobacillus sp., namely Lactobacillus reuteri, Lactobacillus crispatus, and Lactobacillus plantarum only residing in the lungs. And Lactobacillus murinus, which exists only in the liver. B._longum and Bifidobacterium sp. only resided in the heart showed in Figure 1C . The species of Burkholderia and Ralstonia in organs function in organic metabolism, environmental information processing, cellular processes, and genetic information processing are shown in Figure 1D .

UniFrac-based principal coordinate analysis (PC_OA) on similarities with function revealed significant differences in the bacterial community structure among the groups ( Figure 1E ). The bacterial composition and use of antibiotics were the two primary factors, accounting for 56.50% and 26.04%, respectively, reflecting the functions of the heart, liver, and lungs.

Culturomics techniques were used to unveil the dominate species of bacteria contained in heart ( Figure 2A ) for scarce bacteria. The most cultivable microorganism were genus of Burkholderia manifested on an MRS plate. The pH value of the culture in the MRS medium reached 3.5. Intrinsic bacteria reside in the heart because of their inhibitory action on pathogenic E. coli, as shown in Figure 2B . Burkholderia and Ralstonia were the two primary genera in the cultivation of cardiac flushing liquid. Strains of Ralstonia_pickettii, Ralstonia_sp._3PA37C10, R._insidiosa, Burkholderia_lata, unclassified _g_ Ralstonia, and unclassified _p_ Pseudomonadota in heart was accounted for most of the population in the heart. After cultivation, additional strains of Burkholderia stabilis, Trichinella britovi, and Trichinellanativa were detected ( Figures 1B , 2C ). Cultivated nutrition strictly required bacteria; B. longum and Bifidobacterium sp. were unidentified ( Figure 1C ). The bacterial function in the culturomics of heart-flushing liquid indicated fewer roles in environmental information processing, cellular processes, and metabolism compared with the flushing liquid of the heart ( Figure 1D ). This function strongly correlated with the bacterial composition of the heart and its cultivation, as shown in the PCA plot ( Figure 2D ).

Mice with Ligilactobacillus salivarius supplemented for 7 d under heat conditions, then the intrinsic bacterial composition of the heart, liver, and ileal mucosa were measured using 16S rDNA sequencing. A total of 1, 003, 673 raw reads were obtained in total. After removing the low-quality sequences, 368, 213 clean tags were identified. Mice with Ligilactobacillus salivarius supplementation the bacterial composition of the heart, liver, and ileum improved at the phylum level indicated in Figure 3A . Principal coordinates analysis (PCoA) indicated that two primary factors accounted for 82.69% of all attributes of the differences in bacterial composition at the phylum level ( Figure 3B ). Five phyla: Bacillota, Bacteroidota, Pseudomonadota, Thermodesulfobacteriota, and Actinomycetota, accounted for nearly 98.2% of the total bacterial composition ( Figure 3C ). The richness of the phyla Bacillota and Bacteroidota was reduced, and more phylum of Pseudomonadota was replaced with a reduced proportion. The bacterial composition at the genus level was also rich upon supplementation with Ligilactobacillus salivarius ( Figure 3D ). The dispersed area indicated that the bacterial composition was more uniform. The main factors contributed to 72.39% of the differences in composition ( Figure 3E ). The richness of the genera Lactobacillus (P<0.001), Paenibacillus (P<0.001), and norank_f_Mitochondria (P<0.01) in heart, liver, and ileal mucosa were all increased ( Figure 3F ). The richness of unclassified_k:norank_d:Bacteria in the heart also improved, whereas it decreased in the liver and ileum (P<0.001). The proportion of the genera Alistipes, norank_f_ Muribaculaceae, and Bacteroides were reduced in all three organs of Ligilactobacillus salivarius supplemented mice (P<0.001).

The anti-oxidative capacity in heart, liver and ileum was measured shown in Figure 4A , both in tissue of heart and liver, the level of ROS produced (in per second) in Ligilactobacillus salivarius supplemented mice was lower than control mice (P < 0.01). The levels of produced ROS were also reduced in the ileum following Ligilactobacillus salivarius supplementation (P < 0.01). SOD activity in the heart, liver, and ileum was significantly higher than that in control (P < 0.01). The differences in microbial composition at the species level showed that more strains of Lactobacillus reuteri (P < 0.001), B. stabilis (P < 0.001), and Ralstoniapickettii (P < 0.01) became richer than in the control in all three organs ( Figure 4B ). Lactobacillus reuteri, Lactobacillus murinus, Lactobacillus johnsonii owning strong correlation with anti-oxidative parameters. Some bacterial species, Lachnospiraceae_bacterium, uncultured_bacterium_g:norank_f:Muribaculaceae, uncultured_Muribaculaceae_bacterium, unclassified_g:Muribaculum, and norank_f_ Mitochondria also play significant roles in the defense against heat stress ( Figure 4C ).

Under heat stress, oral supplementation with Ligilactobacillus salivarius significantly increased the richness of Lactobacillus in the heart, liver, and ileum ( Figure 3F ) and increased the proportion of Lactobacillus reuteri ( Figure 4B ). To further validate the translocalization of core bacteria in the heart and liver, FISH and qPCR assays were performed, and the results are shown in Figure 5A . More Lactobacillus reuteri 16S fluoresce-labeled green spots were found in the homogenates of the heart, liver, and ileum. qPCR results indicated that the number of Lactobacillus reuteri Reached 10⁵ CFU/g in both the heart and liver of mice supplemented with Ligilactobacillus salivarius supplementation, showed in Figures 5B, C . However, a low number of Lactobacillus reuteri was found in the control mice.