The study subjects were healthy volunteers who met the following inclusion criteria: an age between 19 and 45 years, a body mass index (BMI) between 18.5 and 25 kg/m², voluntary acceptance of and adherence to the experimental protocol, and ability to participate in a timely review and follow-up. The exclusion criterion was any autoimmune or other chronic disease. The study was conducted at Henan University of Technology from April 21 to 31 June 2022, and the flow of the trial is shown in Figure 1. Of the 110 initial volunteers, 100 met the above-listed criteria and were included in the study. To exclude the influence of volunteers’ use of other probiotic products on the trial, the probiotic products of all kinds were stopped by communicating with volunteers 2 weeks before enrollment. Prior to formal enrollment, we conducted further inquiry, confirmed that all volunteers meet the conditions before enrollment. The study followed the guidelines of the Declaration of Helsinki, and all procedures involving the human body were approved by the Ethics Committee of Henan University of Technology (No. HautEC202277). Written informed consent was obtained from the study subjects after they had received a detailed explanation of the nature of the study.

Following a 14 day washout period, the subjects were divided into placebo (CTL) and LRa05 groups. The CTL group received a 2-g supplement of maltodextrin per day, and the LRa05 group received 1.9 g maltodextrin plus 0.1 g LRa05 bacterial powder (1 × 10¹⁰ CFU) per day. The placebo and probiotic products used in this study were obtained from Wecare Probiotics Co., Ltd. The experimental period was 4 weeks, and the volunteers did not consume products containing probiotics during this period and 2-week washout period. No additional dietary restrictions were imposed, and the subjects maintained their original lifestyle habits during the experimental period.

This was a single-blind placebo-controlled trial, meaning that the subjects were unaware of their assigned group. The trial design and process of the study are shown in Figure 2. The study consisted of two visits. During the first visit (T0), each subject’s height, weight, and blood pressure were measured. In the end (T1), each subject’s body fat percentage was measured. We used a questionnaire to evaluate the effect of the placebo or probiotic intervention on the subjects’ defecation. As a questionnaire, Bristol Stool Form Scale (BSFS) is the most widely used evaluation of stool consistency (19), in which types 1 and 2 are considered hard stools (associated with symptoms of constipation); types 3, 4, and 5 are generally considered normal stool forms; and types 6 and 7 are considered abnormally loose or liquid stools (associated with symptoms of diarrhea). To eliminate individual differences, a longitudinal study approach was applied, using samples collected on Day 0 as the control. Fecal samples were collected at T0 and T1 and used to assess the subjects’ gut microbiota (Figure 2).

The subjects were asked to self-collect stool samples, freeze them immediately at −20°C, and bring them to the visit site for storage at −80°C until analysis. DNA was isolated from the samples for testing using the QIAamp DNA kit according to the manufacturer’s instructions. Polymerase chain reaction was used to amplify the V3–V4 variable region of the prokaryotic 16S rRNA gene as previously described (20) using the primers F1 and R2 (5′- CCTACGGGNGGCWGCAG-3′ and 5′-GACTACHVGGGTATCTAATCC-3′). Sequencing was performed on the Miseq platform (Illumina, San Diego, CA, USA) using a 2 × 300 bp paired-end protocol. In preparation for sequencing experiments, a sample that has been sequenced was prepared as a positive control to complete DNA extraction and PCR simultaneously with the sample to be tested.

Methods for bioinformatics analysis of amplicons were adopted from our previous publications (20, 21). Trimmomatic was used to filter low quality sequences (22). Denoising was performed using the UNOISE algorithm, reads were clustered into amplicon sequence variants (ASVs), and classification assignments and construction of ASV tables were performed by using USEARCH software. The sequencing results were analyzed using USEARCH software (version 11.0667),¹ which produced amplicon sequence variants (ASVs). The 16S rRNA database from the RDP reference training set (version 18) was used as the reference database for sequence annotation.² Community diversity (Shannon and Simpson indices) and richness [Chao1 and abundance-based coverage estimators (ACE)] in the gut microbiota were analyzed at the level of ASVs using the vegan 2.5-7 package (23) on the R platform. The PICRUSt v2.5.0 (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) pipeline (24) was applied to the 16S rRNA sequencing data, and the imputed relative abundances of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways in each sample were used to predict alterations in fecal microbiome function using the picrust2_pipeline.py command.

Continuous variable that followed a normal distribution were analyzed using the t-test. Continuous variable that did not follow a normal distribution were analyzed using the non-parametric test to identify differences between the two groups. Kruskal–Wallis test followed by dunn.test function (in the dunn.test package) were used for multiple comparison testing. Linear discriminant analysis combined with effect size (LEfSe) algorithm measurements was used to identify biomarkers unique to each group based on the abundance values (25). The cutoff of LDA score in the LEfSe analysis is 2.0. Principal coordinate analysis (PCoA) of data on the gut microbiota was conducted according to the Bray–Curtis distance, and significant differences between the groups were determined using the adonis2 function of vegan 2.5-7. PICRUSt data were analyzed using Statistical Analysis of Metagenomic Profiles (STAMP, version 2.1.3) (26). The Mantel test (23) (mantel function in vegan 2.5-7) was used to test the correlation between the gut microbiota at T0 and T1 according to the Bray–Curtis distance, and the reported mantel r and p-values are based on 999 permutations. All graphs were generated using the ggplot2 package in R (27). Statistical analyses were performed using R version 4.2, and p-values < 0.05 were considered statistically significant.

The sequence data used in this article have been deposited in the NCBI database (ac-cession number, SRA: PRJNA899929).

One hundred subjects were included in this study, and four and two subjects in the CTL and LRa05 groups, respectively, dropped out due to restrictions pertaining to COVID-19 (Figure 1). Table 1 illustrates that there were no significant differences in age, sex, blood pressure, and BMI between the two groups at baseline. The metadata of all volunteers were provided in Supplementary Table 1.

As shown in Table 2, no significant changes in BMI, body fat percentage, and BSFS results were observed in either the CTL or the LRa05 group. This outcome demonstrates the safety of LRa05 supplementation.

The species accumulation curves (Figure 3A) show successive increases in species in each group, indicating that saturation had been reached in species accumulation, and additional samples could not provide new records. In addition, the Venn diagram in Figure 3A indicates 1,155 species that were common between the CTL and LRa05 groups, accounting for 93.8% of the 1,232 found species. This result suggests that both the placebo and LRa05 intervention resulted in detectable changes in species diversity in the subjects’ gut microbiota. Further diversity analysis (Figure 3B) revealed that both the placebo and LRa05 intervention led to increased species richness (Chao1 and ACE) in the gut microbiota but did not cause significant changes in community diversity (Shannon and Simpson indices). These findings are consistent with the findings from our species accumulation analysis and Venn diagram.

Although alpha diversity analysis did not reveal significant changes in the gut microbiota with placebo and LRa05 intervention, beta diversity analysis revealed (Figure 3C) that both the placebo and LRa05 intervention had a significant effect on the gut microbiota composition. No significant differences in the gut microbiota were observed between the CTL and LRa05 groups at T0 and T1 (Figure 3D, p = 0.317 and p = 0.857). This result suggests that changes in the gut microbiota after intervention were mainly caused by maltodextrin. The effect of probiotic supplementation on the gut microbiota of healthy individuals was not significantly different from the effect of the placebo.

At the phylum level, Firmicutes, Bacteroidetes, Proteobacteria, and Actinobacteria were the four most predominant bacterial phyla in the subjects’ gut microbiota (Figure 4A). Our analysis of the gut microbial data at T0 and T1 showed no significant phylum-level difference in the gut microbiota between the CTL and LRa05 groups at either time point (Figure 4B), except for Proteobacteria. However, both the placebo and LRa05 intervention resulted in significant changes in the gut microbiota composition at T1 relative to T0; these changes mainly involved significant reductions in the relative abundances of Firmicutes and Actinobacteria and a significant increase in the relative abundance of Bacteroidetes.

We further investigated changes in the gut microbiota at the genus level using LEfSe analysis. Compared with T0, placebo supplementation resulted in significant increases in the relative abundances of Prevotella, Phocaeicola, Bacteroides, and Parabacteroides and significant decreases in the relative abundances of Escherichia/Shigella, Gemmiger, and Bifidobacterium at T1 (Figure 4C). LRa05 intervention led to changes in the gut microbiota similar to those observed with the placebo, including significant increases in the relative abundances of Phocaeicola, Bacteroides, and Parabacteroides and significant decreases in the relative abundances of Escherichia/Shigella, Gemmiger, and Bifidobacterium at T1 relative to T0 (Figure 4D). Further investigation of the differences between the LRa05 and CTL groups revealed that probiotic intervention led to significant increases in the relative abundances of Weissella, Lacticaseibacillus, Enterococcus, and Mitsuokella and significant decreases in the relative abundances of Dialister, Negativibacillus, and Anaeromassilibacillus at T1 (Figure 4E). However, these differences were not observed at T0 (Figure 4F). These results indicate that although the observed changes in the gut microbiota were mainly caused by maltodextrin, the addition of LRa05 led to changes in the relative abundances of some specific microorganisms, such as an increase in Lacticaseibacillus. As LRa05 belongs to the genus Lacticaseibacillus, this result implies that LRa05 colonization of the gut occurred. Gut microbiota composition information at the family and genus level was also analyzed (Supplementary Figures 1, 2).

PICRUST analysis showed that the placebo and LRa05 intervention caused similar changes in the function of the microbiota, consistent with the LEfSe analysis results. Compared with T0, both the placebo and LRa05 intervention resulted in the increased abundance of pathways such as chaperones and folding catalysts, alanine, lysosome, and other glycan degradation, as well as significantly decreased abundance of pathways such as ABC transporters, transporters, secretion system, and bacterial motility proteins, at T1 (Figures 5A, B). Only small differences in gut microbiota function were observed between the CTL and LRa05 groups at T0 (Figure 5C), as indicated by smaller differences in mean proportion. At the end of the trial, the abundances of the Salmonella infection and NOD-like receptor signaling pathway were reduced and the abundance of the nitrogen metabolism pathway was increased in the LRa05 group compared with the CTL group (Figure 5D). The reduced abundance of the Salmonella infection pathway implies the inhibitory potential of LRa05 against harmful bacteria. Therefore, we further compared the abundance of the genus Sellimonas at T1 between the CTL and LRa05 groups (Figure 5E). We found that the relative abundance of Sellimonas tended to decrease in the probiotic group compared with the placebo group, but this difference was not significant (p = 0.6872).

We also assessed the association between the gut microbiota before and after the intervention in both groups using Mantel correlation analysis. As shown in Figure 6, there was a significant correlation (p = 0.001) between the gut microbiota at T0 and T1 in all subjects (Figure 5A), the CTL group (Figure 5B), and the LRa05 group (Figure 5C) after the intervention. However, the correlation coefficient was higher in the CTL group (r = 0.417) than in the LRa05 group (r = 0.182), indicating the effect of LRa05 on the gut microbiota.