Abstract
Breast milk contains human milk oligosaccharides (HMOs) that cannot be digested by infants, yet nourish their developing gut microbiome. While Bifidobacterium are the best-known utilizers of individual HMOs, a longitudinal study examining the evolving microbial community at high-resolution coupled with mothers’ milk HMO composition is lacking. Here, we developed a high-throughput method to quantify Bifidobacterium longum subsp. infantis (BL. infantis), a proficient HMO-utilizer, and applied it to a longitudinal cohort consisting of 21 mother-infant dyads. We observed substantial changes in the infant gut microbiome over the course of several months, while the HMO composition in mothers’ milk remained relatively stable. Although Bifidobacterium species significantly influenced sample variation, no specific HMOs correlated with Bifidobacterium species abundance. Surprisingly, we found that BL. infantis colonization began late in the breastfeeding period both in our cohort and in other geographic locations, highlighting the importance of focusing on BL. infantis dynamics in the infant gut.
Similar content being viewed by others
Introduction
Breast milk is considered the ideal nutrition for infants during their first 6 months of life1. Millions of years of evolution have shaped breast milk composition such that its third most abundant component, human milk oligosaccharides (HMOs), cannot be digested by the infant, but serves as substrate for the infant’s gut bacteria2,3. There are many different types of HMOs, which can be largely classified into three groups: fucosylated, sialylated, or neutral. Each HMO is composed of 3 to 32 monomers, and a single milk sample typically contains 50 to 200 distinct types of HMOs4. Among other factors, maternal genetics plays a role in the production of specific HMOs in breast milk5. For example, mothers with an inactive fucosyltransferase 2 (FUT2) gene, termed non-secretors, fail to form alpha-1,2 bonds between fucose and lactose or other HMO backbone structures, resulting in the lack of 2′FL and other alpha-1,2-fucosylated glycans6. Additionally, environmental factors coupled with infant age can affect HMO composition7.
On average, infants have a higher relative abundance of Bifidobacterium species while they are breastfed8,9,10. Bifidobacterium species were previously shown capable of utilizing multiple HMOs11,12,13, however this ability varies between species and even within a single species14,15,16,17. Among all Bifidobacterium species and subspecies, the best-known HMO utilizer is Bifidobacterium longum subsp. infantis (BL. infantis) which grows efficiently on most types of HMOs18, and possesses a large variability of HMO utilizing genes19. In contrast, other Bifidobacterium species have a lower capability of HMO utilization, for example B. breve strains cannot utilize 3′SL and 6′SL at all, and most of them cannot utilize fucosylated HMOs15. Since HMOs serve as food for the gut microbiome, one may hypothesize that different HMO compositions in mothers’ milk affects the developing gut microbial community.
To date, most research addressing the HMO-bacteria relationship in the infant gut focused on a single time point20. A longitudinal cohort study is needed in order to examine how changes in HMO composition impact the infant gut microbiome over time.
To quantify the abundance of various Bifidobacterium species in microbiome communities two approaches are commonly used: 16S-rRNA sequencing and shotgun metagenomics. While metagenomic sequencing allows classification at the species level, 16S-rRNA sequencing provides only genus-level classification of microbiome communities. The basic annotated unit in 16S-rRNA sequencing is referred to as operational taxonomic unit (OTU), which can be assigned to genus-level classification and may represent multiple species of the assigned genus. Multiple species can be annotated as the same OTU, hence using 16S-rRNA sequencing so far provided mostly weak or no associations with abundances of specific HMOs21,22,23,24. A single OTU can include multiple species (or subspecies) with various HMO utilization capabilities, thus, a higher-resolution taxonomic definition is needed.
The largest variability in HMO-utilization capability can be found within the Bifidobacterium longum species. Overall, this species can be divided into two subspecies found in humans: B. longum subsp. longum (BL. longum) and B. longum subsp. infantis (BL. infantis). BL. longum is found both in infants and adults, while BL. infantis is unique to the infant gut. Studies have shown that BL. infantis can utilize almost all HMOs25, while BL. longum has a limited repertoire. To study the HMO-microbe relationship, taking into account these differences in HMO-utilization within B. longum subspecies, a high-throughput, higher-resolution method is needed. Past studies have used different methods to differentiate between BL. infantis and BL. longum, such as qPCR26, PCR27 or the Bifidobacterium Longum-Infantis Ratio (BLIR) method28,29, yet these methods require the original DNA and are not high-throughput. Others have searched for BL. infantis specific genes such as the H1 cluster20 or other BL. infantis clusters30,31, however these methods do not give an exact ratio between the subspecies, rather they indicate their presence or absence. The new method we propose here could be applied also to the massive amounts of data available in public repositories.
Here, we establish a new matched cohort of breast milk and infant stool samples collected longitudinally throughout the first year of life. We develop a method to allow BL. infantis quantification from existing metagenomic data, and apply it to samples from our cohort to study the relationship between the abundance of Bifidobacterium species in the infant gut and HMO composition in mothers’ milk over time. Finally, we apply our B. longum subspecies quantification method to existing infant gut datasets to examine the timing of BL. infantis colonization across geographic locations.
Results
Cohort design
We have established a new and unique longitudinal cohort to test the relationship between HMOs in mothers’ milk and the developing infant gut microbiome. Our cohort consists of 21 mother-infant dyads with matched infant stool samples and breast milk samples collected on the same day. Altogether, we collected 80 stool samples and 50 breast milk samples together with the infant nutritional information and antibiotic treatments (Supplementary Fig. 1, Supplementary Data 1). We collected these samples between the age of 2 weeks and 41 weeks, and each dyad contributed between one to eight paired samples.
Specific marker genes allow better quantification of B. longum subspecies
Bifidobacterium longum subsp. infantis (BL. infantis) is the best known utilizer of HMOs19,32, however current methods for taxonomic classification from metagenomes are unable to separate the Bifidobacterium longum (B. longum) species into its main subspecies; Bifidobacterium longum subsp. longum (BL. longum) and BL. infantis20,29. MetaPhlAn is one of the most common tools for profiling the composition of microbial population from metagenomic data, by using specific marker genes for each taxonomic group33. However MetaPhlAn has no specific marker genes for BL. infantis and therefore classifies B. longum at the species taxonomic level. Due to the differences between B. longum subspecies in the context of HMO utilization, there is a rising need for a high-throughput method that will allow specific identification and quantification of BL. infantis from metagenomics data.
Here we define B. longum subspecies specific markers and use them in a tailored MetaPhlAn33 database which allows abundance quantification of two B. longum subspecies: BL. infantis and BL. longum (Supplementary Fig. 2A). To construct our new dataset, we searched for marker genes that are unique to each subspecies. A marker gene was selected if it was present in at least 90% of reference genomes of one subspecies and not in a single genome of the other subspecies (Fig. 1A, Methods). We chose to discard two subspecies of B. longum: Bifidobacterium longum subsp. suis and Bifidobacterium longum subsp. suillum since they are rarely found in humans34, and a limited amount of reference genomes exist for these subspecies (Methods).
A Identification of unique genes in BL. infantis (blue) and BL. longum (green) reference genomes that can serve as marker-genes in the quantification of B. longum subspecies. HMO utilizing genes (HUGs) are marked on the bottom right and constitute only a small fraction of BL. infantis marker genes. B, C Validation of the computational approach, comparing the tailored MetaPhlAn marker-gene quantification results to the experimental qPCR results of B BL. infantis and C BL. longum. P-value for regression was calculated using a two-sided t-test.
In order to validate our results, we applied MetaPhlAn with our new set of marker genes coupled with subspecies-specific qPCR to metagenomic sequencing data from 68 infant stool samples. When comparing the relative abundance of BL. infantis and BL. longum in each method, we observed a strong correlation between our computational approach and qPCR (R2 = 0.999 for BL. infantis and 0.997 for BL. longum; Fig. 1B, C). This finding confirms our method’s specificity and sensitivity for both BL. infantis (Fig. 1B) and BL. longum (Fig. 1C). In some samples, MetaPhlAn failed to assign a classification to a small percentage of BL. longum and therefore it was designated as unclassified B. longum (Supplementary Fig. 2B, C).
The infant gut microbiome shows excessive changes while HMO composition in mothers’ milk is fairly stable
To examine the dynamics of Bifidobacterium species in the infant gut, we conducted metagenomic sequencing and analyzed the data using our novel MetaPhlAn database. We observed a significant prevalence of Bifidobacterium (at least one sample with >25%) in all the infants (Fig. 2A), in line with our expectation as most infants in our study were breastfed8,14,35. Bifidobacterium remained highly abundant even after solid foods were introduced to infants (Fig. 2A, arrows). The abundant species included Bifidobacterium breve, Bifidobacterium bifidum, Bifidobacterium pseudocatenulatum and B. longum subspecies (Fig. 2A), along with Bacteroides species such as Bacteroides dorei and Bacteroides vulgatus (Supplementary Fig. 3). Interestingly, the presence of BL. longum and BL. infantis was mutually exclusive, reflecting potential intra-species competition, as previously suggested36.
A Relative abundance of Bifidobacterium species in the infant gut (colorful) with all other species classified as “other” (gray). Samples from the same infant are grouped together, sorted by age. Arrows indicate samples taken after the introduction of solid food. B Temporal changes in the relative abundance of the microbial community in three infants (inf03, inf16, and inf18) in the first 30–40 weeks of life. The most prevalent bacteria are colored, and the remaining are indicated as “Other” (gray). Colors as in (A), with additional colors for the non-Bifidobacterium species. C Relative abundance of 16 HMOs measured in mothers’ milk. Samples are categorized into three groups based on their HMO profiles, and arches connect samples obtained from the same mother (Methods).
Overall, we found that the infant gut microbiome underwent significant changes over the course of several months. While the bacterial composition tended to be stable over a few weeks, there were certain time points when a switch in composition occurred (Fig. 2A, B). In infants that we had samples from many time points we found this switch to occur around 10 weeks of life. For example, in infant03 (inf03) over the course of the first 15 weeks the dominant bacteria in the gut constantly changed, including B. vulgatus, Klebsiella pneumoniae, and finally B. breve; the gut of infant18 was initially dominated by B. breve, followed by a complete switch to BL. infantis; and the gut of infant16 constantly changed its microbial composition (dominated by Escherichia coli, Veillonella seminalis and B. pseudocatenulatum; Fig. 2B). It was not always clear what triggered these microbial shifts, however since Bifidobacterium and other gut microbes utilize HMOs, we hypothesized that changes in the HMO composition in mothers’ milk might be causing bacterial changes in the infant gut.
To examine the impact of HMO composition in mothers’ milk on the infant gut microbiome, we quantified 16 common HMOs in 50 milk samples from 20 mothers using high performance liquid chromatography with fluorescence detection (HPLC-FLD; Methods). In contrast to the dynamic infant gut microbiome, the composition of HMOs in mothers’ milk remained relatively stable over the course of months, in terms of both their concentration in milk (Supplementary Fig. 4A) and the relative abundance of specific HMOs (Fig. 2C). We divided the milk samples into three main groups, based on their HMO composition: those with low or no 2′FL abundance (Group 1, non-secretors37); samples from secretor mothers with very high abundance of 2′FL (>40%) and LNFP1 (>10%) and smaller amounts of other HMOs (Group 2); and samples from secretor mothers with lower abundance of 2′FL (<30%), and no clear dominant HMO (Group 3; Fig. 2C). We found no major changes over time in the abundance of specific HMOs, other than LSTc which was reduced to almost 0 over the course of ~40 weeks (Supplementary Fig. 4B, C), in line with previous findings7. Overall, the changes found in the HMO composition in consecutive milk samples were significantly less pronounced than those found in the microbial population from consecutive infant gut samples (t-test, p = 2.88e-5, Supplementary Fig. 5A).
The dominant Bifidobacterium species shape the infant gut microbiome
General comparison of the infant gut microbiome composition determined that Bifidobacterium species play a significant role in the breastfed infant gut (Fig. 2A). We next searched for differences across samples in an attempt to characterize the various microbial profiles of the infant gut. We examined the diversity of the infant gut samples in our cohort, using a dimension reduction approach (PCoA with Bray-Curtis dissimilarity, Methods), and found that samples cluster into distinct clusters (using K-means, k = 4). The first three groups had samples with mostly one main Bifidobacterium dominant species (with relative abundance of >30%): B. breve, BL. longum and BL. infantis, and the fourth group contained samples dominated by either a different Bifidobacterium species (Bifidobacterium adolescentis, B. pseudocatenulatum) or other species (named “Mixed”, Fig. 3A). While usually consecutive samples from the same infant were assigned to the same cluster, occasional cluster switches were observed, strengthening our finding that the microbiome changes in this time frame (Fig. 3B). Overall, these analyses highlight the importance of Bifidobacterium species in our samples, as these are major factors that impact the variation in the infant gut microbial composition.
A Principal Coordinate Analysis (PCoA) of infant gut microbiome samples using Bray-Curtis dissimilarity. Points are color-coded based on the dominant Bifidobacterium species in each sample, and non-Bifidobacterium dominated samples are colored in gray. Samples without a dominant species (>30% relative abundance) are labeled as “No dominant” (yellow). The size of each point represents the relative abundance of the dominant bacteria in that sample. Ellipses encompass four groups identified using k-means clustering. Each group represents a primary dominant species (indicated by the ellipse color): BL. infantis (blue), BL. longum (green), B. breve (red), and “Mixed” (gray). B Changes in group assignment observed in consecutive samples from the same infant (colors as in A). C Microbial diversity of samples within each group (measured by the Shannon index; colors as in A; two sided t-test, n = 80 samples from 21 infants, ***p ≤ 0.001, ****p ≤ 0.0001). Box boundaries are the 25th and 75th percentiles, and the median is highlighted. Whiskers represent 1.5 * IQR and points past them are outliers. D Relative abundance of BL. infantis (x-axis) versus the relative abundance of B. breve (y-axis) in each sample, indicating the mutual exclusiveness of the two species.
To further investigate these four groups of infant gut samples, we examined the alpha diversity of the microbiome population in samples found in each group using Shannon index. We found that the alpha diversity of samples within the BL. infantis group was lower compared to samples from other groups (t-test, p ≤ 0.001, Fig. 3C), indicating that when BL. infantis is found, it dominates the community at such high levels of relative abundance leaving a smaller ecological niche for other bacteria in the infant gut. We next focused on the “Mixed” group, and asked whether additional variables may play a role in these microbial profiles. We examined breastfeeding, maternal or infant antibiotic use, delivery mode, breastfeeding type (pumped or direct), and introduction of solid foods, yet we did not find any specific variable that was associated with the microbial profile of the “Mixed” group. As expected, the BL. infantis group consisted solely of infants who received none-to-low amounts of infant formula38.
To characterize the relationships among dominant Bifidobacterium species in the infant gut, we examined their occurrence within groups where they are not dominant. We found that BL. infantis and B. breve are mutually exclusive, consistent with a previous study in Hazda infants39, implying competition for the same niche in the infant gut (Fig. 3D). However, it remains unclear what specific niche B. breve and BL. infantis are competing for, given B. breve’s limited ability to utilize a variety of HMOs15. Finally, we observed BL. longum in some B. breve-dominant samples (Fig. 2A), suggesting that B. breve may rely on derivatives from BL. longum through cross-feeding in these samples40.
Single HMOs are not associated with specific Bifidobacterium species
It is well established that different Bifidobacterium species have different HMO-utilization capabilities14,40,41. Therefore, specific HMOs may benefit specific Bifidobacterium species in the infant gut based on their HMO utilization profiles. However, we found no significant correlation between the abundance of Bifidobacterium in general and the main Bifidobacterium species and subspecies with specific HMOs (Fig. 4A) or HMO groups (fucosylated, sialylated or neutral; Supplementary Fig. 5B). In addition, linear association models accounting for individual infants showed no significant association between specific HMOs and Bifidobacterium species and subspecies in the infant gut (when requiring FDR q < 0.2). Nevertheless, we found that BL. infantis exhibited a high abundance (>25%) exclusively in infants to secretor mothers (Fig. 4B, t-test p = 0.021). In addition, we observed a modest and non-significant negative correlation (r = −0.27) between BL. infantis and LSTc (Fig. 4C). It is worth noting the importance of considering the timing factor in interpreting these findings. The delayed presence of BL. infantis in the gut (which will be discussed in more detail later) and the gradual decrease of LSTc over time (Supplementary Fig. 4B,C) could contribute to the observed correlation. The lack of variation in the HMO composition together with the lack of HMO-microbes associations indicate that the microbial shifts, specifically within Bifidobacterium species, can not be explained by a change in mothers’ milk HMO composition.
A Pearson correlations between all measured HMOs and the Bifidobacterium genus, as well as the main individual species and subspecies (BL. infantis, BL. longum, B. breve and B. bifidum). None of these correlations were found to be statistically significant (t-test). B Comparison of the relative abundance of LSTc in mothers’ milk (x-axis) with the relative abundance of BL. infant is in the infant gut microbiome (y-axis). C Maximum relative abundance of BL. infantis in each infant divided into two groups of infants with secretor and non-secretor mothers, highlighting the high abundance of BL. infantis only in infants to secretor mothers, but not in all of them.
Metagenomes with BL. infantis contain more HMO utilizing genes
Metagenomes obtained from various time points of multiple infants contain distinct strains and species, resulting in variable gene abundance profiles which can enable various patterns of HMO utilization. To assess the HMO utilization potential of specific Bifidobacterium species in our dataset, we investigated the presence of HMO-utilizing genes (HUGs)42 organized into five distinct clusters (H1-H532; Fig. 5). We observed that the dominant species in each sample significantly influenced the metagenome’s theoretical capacity for HMO utilization. As expected, BL. infantis-dominated samples exhibited the highest abundance of HUGs, confirming its exceptional capability in utilizing HMOs19. Notably, some of these samples displayed high variation in gene abundance from clusters H1 and H5, indicating a potential lower capacity to transport some HMOs19, and utilize lacto-N-tetraose (LNT) and lacto-N-neotetraose (LNnT)16.
Normalized genes abundance of HMO utilization genes (HUGs) in all infant samples. Genes (rows) are categorized and labeled according to their respective gene cluster (H1-H5). Samples (columns) are annotated based on their assigned dominant-species group.
Interestingly, metagenomes that were dominated by B. breve or BL. longum also contained genes from the H5 cluster, emphasizing their ability to utilize HMOs based on lacto-N-biose (LNB)16. However, samples from the “Mixed” groups exhibited minimal or no HUGs, suggesting either alternative genes for HMO utilization or a lack of capacity to utilize HMOs altogether.
BL. infantis does not colonize the infant gut in early breastfeeding weeks
BL. infantis is the most proficient HMO-utilizer in the infant gut11,19, thus we expected that BL. infantis will have a fitness advantage in the breastfed infant gut from the initial days of breastfeeding. However, despite the majority of infants in our cohorts that were breastfed since birth, BL. infantis was primarily detected starting only at 10 weeks of age (Fig. 6A). Linear association models showed a clear positive association of the relative abundance of BL. infantis with age (coefficient = 3.16, q = 6.63e-5), which was not found for any other Bifidobacterium species or subspecies. Overall, BL. infantis exhibited the highest abundance at 10–25 weeks, followed by a gradual decrease in abundance (Fig. 6A).
A Relative abundance of BL. infantis in our Israeli cohort, among infants where BL. infantis was detected at any time point. B–D Relative abundance of BL. infantis at different time points in samples from B Sweden (n = 300 samples from 100 infants) C UK (n = 376 samples from 169 infants) and D Finland (n = 228 samples from 108 infants). For Sweden and the United Kingdom where samples were collected at defined time points box plots are shown, with box boundaries at the 25th and 75th percentiles, and the median highlighted. Whiskers represent 1.5 * IQR and points past them are outliers. E The percentage of infants in multiple geographical locations harboring BL. infantis in the first 10 weeks of life and at 10 weeks and later. F Phylogenetic tree displaying BL. infantis strains found across all cohorts. The dominant strain from each sample is represented, and strains are color-coded based on their corresponding cohort, highlighting that some countries (like Italy) have very similar strains in all samples.
Our innovative computational methodology enabled the exploration of the delayed colonization of BL. infantis across various geographical regions, providing a gateway for further in-depth investigation into this phenomenon. Thus, to corroborate our findings, we examined additional infant cohorts from Sweden9, United Kingdom43, Estonia44, Italy45, Russia44, United States46 and two cohorts from Finland44,47 comprising samples from a total of 1,017 infants throughout the first year of life. Across all cohorts, a similar pattern of late-colonization of BL. infantis was observed: In Sweden, BL. infantis was not observed at birth almost at all, reaching its peak prevalence at 17 weeks, followed by a gradual decline in both prevalence and relative abundance by 52 weeks (Fig. 6B). In the UK, BL. infantis was observed in only four out of 178 infants in the first 3 weeks of life, and during the later infancy period (17–52 weeks) BL. infantis was found in 28 infants (Fig. 6C). In Finland, one cohort with 126 infants had only a single infant with detectable levels of BL. infantis in the first 10 weeks of life, and additional 11 infants gained it later on (Fig. 6D). In a second Finnish cohort, BL. infantis was not observed in any infant samples (ages ≤ 13 weeks; Finalnd2, Supplementary Fig. 6A). In the Italian and Estonian cohorts, a similar pattern was observed however the relative abundance of BL. infantis was lower in most infants (Supplementary Fig. 6B, C). The cohort from the the United states contained samples only from the first 2 weeks of life, revealing no infants with BL. infantis in the first week of life, and only one out of 77 infants with detectable levels in the second week of life (Supplementary Fig. 6D). Finally, in the Russian cohort which consisted of samples collected from infants aged 12 weeks and beyond, BL. infantis was detectable in 14/69 infants,with its presence becoming noticeable only at 20 weeks (Supplementary Fig. 6E). Overall, all cohorts exhibited a late-colonization of BL. infantis, commonly starting at 10 weeks of age, or later (paired t-test, p = 0.023; Fig. 6E, Supplementary Fig. 6F).
To examine variations between BL. infantis strains across the different countries we next focused on BL. infantis strain-level composition, using the SNP profiles on the newly-identified BL. infantis marker genes (Methods; Fig. 6F). We found that BL. infantis strains within the same infant in our cohort were more similar to each other than strains between unrelated infants (t-test, p ≤ 2.2e-16, Supplementary Fig. 6G). Furthermore, in some countries (i.e., Italy & Russia), most of the strains were very similar, while in other countries (i.e., Sweden & UK) there was a much larger variation between strains (Fig. 6F, Supplementary Fig. 6H). Strains from our Israeli cohorts were clustered in two distinct groups, one more similar to strains found in the US (light purple, Fig. 6F) and the other closer to strains found in the UK (brown, Fig. 6F). Specifically, the strains from the Italian samples were very distinct from all other strains, with the exception of a single UK strain. Finally, BL. infantis reference genomes were clustered into three groups, some identical to the Italian strains, while others clustered closely with Russian and Estonian strains (gray, Fig. 6F). These findings allow us to explore the variation found within BL. infantis and highlight interesting variability across geographic locations.
Discussion
In this study we introduced an innovative approach to quantify BL. infantis and distinguish it from BL. longum in metagenomic data. Our method enables researchers to concentrate on studying this distinct subspecies and its associations with HMOs from existing metagenomic data. This approach can be adapted to differentiation of subspecies in additional microbial species, and specifically further within B. longum (such as BL. suis & BL. suillum subspecies) once sufficient numbers of reference genomes become available. We employed our approach to explore the diversity within the infant gut microbiome and discovered the lack of associations with individual HMOs present in mothers’ milk. Our analysis revealed that the variability between samples was greatly influenced by the dominant Bifidobacterium species in each sample.
Previous research has suggested that colonization of Bifidobacterium species in the infant gut may be influenced by priority effects40. However, our study revealed substantial changes in the dominant Bifidobacterium species within the same infant over the course of several weeks (Fig. 3A, B). This indicates that over time there are additional factors responsible for Bifidobacterium species prosperity, such as species competition and cross-feeding. For example, it was reported that B. breve, despite having limited ability to utilize HMOs, can outcompete stronger competitors if introduced early into a microbial community40. In addition, B. breve has the capacity to cross-feed on monosaccharides derived from HMOs by other Bifidobacterium species48,49. This implies that B. breve may initially dominate the population when carbohydrates are available, but subsequently loses the competition to other Bifidobacterium species once these carbohydrates are depleted. In our data we found that the B. breve dominated group had a diverse microbial population (Fig. 3C), perhaps since it cross feeds on HMOs derivatives from other species. More research is needed to understand the microbial shifts in the infant gut and the effect cross-feeding has on the microbial dynamics in the infant gut.
While the HMOs present in a mother’s milk remained relatively stable over numerous weeks, we observed notable changes in the infant gut microbiome during this time period. This suggests that even subtle variations in the composition of breast milk may have an impact on the development of the gut microbiome. Alternatively, it is possible that other components present in breast milk, such as cytokines, microRNAs and antibodies, play a role in influencing the infant’s microbiome50,51. In addition, we can not rule out environmental effects that may contribute to these changes such as starting daycare which in Israel typically occurs at 15–26 weeks of age. Our findings indicate that there were no significant correlations between HMOs and Bifidobacterium, further supporting the idea that additional factors beyond HMOs are involved in shaping the infant’s microbiome. Previous studies22,23,24,26 have examined the composition of the microbiome and its relationship with HMOs using 16S-rRNA amplicon sequencing. These studies have reported varying results, with some finding no significant correlations, while others observed modest correlations. Interestingly, some of the studies identified a negative correlation between Bifidobacterium OTUs and multiple HMOs22,24. This could be attributed to a decrease in the overall Bifidobacterium abundance within the infant gut over time, coupled with an increase in specific HMOs, in line with our findings regarding LSTc and BL. infantis (Fig. 4B). Additionally, it is possible that such specific correlations may be observable only using a larger cohort.
Finally, we found that BL. infantis does not commonly colonize the infant gut in the early weeks of breastfeeding and that not all breastfed infants have detectable levels of BL. infantis. Importantly, when BL. infantis was found in the infant gut, it commonly dominated the gut microbiome community. Analyzing additional infant cohorts from multiple geographical locations strengthened our findings regarding the late colonization of BL. infantis in infants’ gut. Previous studies have reported a low prevalence of BL. infantis during early time points29,36, however these studies did not incorporate frequent sampling in the first months of life, thus lacking the ability to precisely determine the timing of BL. infantis arrival. A previous study proposed that the arrival of BL. infantis is influenced by the history of breastfeeding practices of a given country29, which impacts the overall exposure to BL. infantis strains. Countries with historically lower breastfeeding rates are likely to have a lower prevalence of BL. infantis, resulting in infants acquiring BL. infantis at a later stage through horizontal transfer29.
Although BL. infantis is one of the more-studied infant gut commensals, its publicly-available reference genomes do not span the entire genomic variation of this subspecies (gray, Fig. 6F). To expand the BL. infantis-related research, additional reference genomes should be characterized in full, from natural isolates representing diverse geographic locations and lifestyles. Further research, especially longitudinal sampling of infants and their surroundings, is required to elucidate the timing and sources from which infants acquire BL. infantis and to comprehend the differences observed between countries.
Methods
Sample collection
Breast milk and stool samples were collected as part of the Breast Milk Baby (BMB) cohort from mothers and infants from birth till 1 year old. Stool samples were collected using eSwab® with 1 ml of liquid Amies medium + 1 regular FLOQSwabs® (Copan) in order to preserve bacterial population. Breast milk samples were collected by pump or manually and stored in sterile tubes. Both sample types were collected by mothers in their homes and stored at 4 °C for up to 24 h and then shipped to the lab and stored long term at −80°C.
Metagenomic library construction and sequencing
DNA was extracted from stool samples using DNeasy PowerSoil Pro Kit (#47014, QIAGEN). Illumina sequencing libraries were prepared using Nextera XT DNA Library Preparation kit (FC-131-1096, Illumina) according to the manufacturer’s recommended protocol with half of the volume and the DNA. Samples were sequenced using Illumina single-end 150 bp sequencing on a NextSeq 500 device.
B. longum subspecies quantification
To identify B. longum subspecies-specific markers we started with 116 B. longum reference genomes downloaded from the NCBI with completeness of >90% and contamination of <5% (Supplementary Data 2). Reference genomes were classified to B. longum subsp. longum (BL. longum), B. longum subsp. infantis (BL. infantis) and unknown based on NCBI annotation, leaving 30 BL. longum and 16 BL. infantis references. As some of the unknown references could belong to either BL. infantis or BL. longum, we removed them all for specificity. We decided not to include subspecies BL. suis and BL. suillum in our method due to the limited availability of reference genomes for these subspecies leading to difficulty to generate reliable marker genes. In addition, these subspecies were rarely reported in humans34, thus are less relevant in our settings. Therefore it is important to note that using our method, BL. suis and BL. suillum can be misidentified as a different subspecies or alternatively identified as B. longum species with unclassified subspecies.
PanPhlan352 was used to analyze the pangenome of all 46 reference genomes. Clustering all presence/absence profiles in the pangenome revealed two clear clusters, the first of 14 BL. infantis strains and the second included all BL. longum strains and two BL. infantis strains (BL. infantis 157 F, BL. infantis CCUG52486; Supplementary Fig. 7). Additional analysis showed that these two BL. infantis reference genomes that were clustered with BL. longum reference genomes did not contain the H1 HMO utilization cluster which defines BL. infantis19. In addition, a previous study showed that these two BL. infantis reference genomes are most probably BL. longum, based on the phylogeny of the core pangenome of 158 B. longum strains53. Taken together, BL. infantis 157 F and BL. infantis CCUG52486157 were suspected as mis-annotated and were excluded from further analysis.
Subspecies-specific marker genes were chosen in two steps. First, using the pangenome we found 331 genes that were present in 90% of one subspecies and not present at all in the other subspecies. For example, a gene that was present in 13 out of the 14 BL. infantis references and not in any BL. longum references was selected to be a marker gene for BL. infantis. Next, all selected genes were filtered to be specific at the species level to B. longum and to confirm they do not exist in other Bifidobacterium species. To do so, we used Blastn 2.12.0 to map all marker genes to the nr/nt nucleotide database. Of the 331 putative marker genes, 84 matched other species (such as B. breve) with >90% alignment and over 50% coverage and therefore were filtered out. Our final set of maker genes included 119 BL. infantis and 128 BL. longum markers (Supplementary Data 3). The MetaPhlAn database was customized to include the newly defined marker genes using described MetaPhlAn instructions (https://github.com/biobakery/MetaPhlAn/wiki/MetaPhlAn-4) and then MetaPhlAn4 was used with the --index and --bowtie2db parameters and our customized marker-gene database.
To verify the results, subspecies of B. longum were determined using qPCR with subspecies-specific primers54 for BL. longum (F: GTGTGGATTACCTGCCTACC, R: GTCGCCAACCTTGACCACTT) and BL. infantis (F: ATGATGCGCTGCCACTGTTA, R: CGGTGAGCGTCAATGTATCT). The efficiency of the primers was assessed by testing them in five dilutions. qPCR was performed at 95 °C for 10 s, followed by 40 cycles of 95 °C for 10 s and 60 °C for 30 s. The ratio between BL. infantis and BL. longum was calculated using the delta-delta Ct method.
To use our tailored MetaPhlAn database see our GitHub page (https://github.com/yassourlab/MetaPhlAn-B.infantis/).
Metagenomic analysis
Host reads were removed using an in house pipeline by aligning reads to the human genome by Bowtie255 (2.4.5-1). Samples were filtered and trimmed for Nextera adapters using fastq-mcf, ea-utils56 (1.05). Taxonomic profiling was done using MetaPhlAn433 with our unique database as described above. Functional profiling was done using HUMAnN352. HUGs42 were selected from previously described HMO clusters32. Strain analysis was performed using StrainPhlAn 433 with default parameters and --sample_with_n_markers 50. SNPs for BL. infantis marker-genes were calculated using the multiple-sequence alignment (MSA) produced by StrainPhlAn 4 (--mutation rates parameter). BL. infantis was considered detectable when its relative abundance was ≥0.5%. Further analysis was done using an in house R (4.2.2) script utilizing dplyr57 (1.1.2), tidyr58 (1.3.0) and tidyverse59 (2.0.0). Plots were created using ggplot260 (3.4.2) and ggforce61 (0.4.1), colors were used from RColorBrewer62 (1.1–3) and pals63 (1.7). Heatmaps were created using pheatmap64. Alpha and beta diversity were calculated using “diversity” (Shannon index) and “vegdist” (Bray-Curtis dissimilarity) from the vegan65 (2.6–4) package and the PCoA was created using the ape66 (5.7-1) package. Phylogenetic tree was produced using ggtree67 (3.6.2) and the sankey plots were created using ggsankey (0.0.99999). Additional cohorts were downloaded from NCBI Sequence Read Archive as following: Sweden9 (PRJEB6456), United states46 (PRJNA591079), Italy45 (PRJNA352475), United Kingdom43 (PRJEB32631), Finland, Estonia and Russia44 (PRJNA497734) and an additional cohort from finland47 (Finland2, PRJNA475246).
HMO quantification
HMO standards used in this study were purchased from Dextra Laboratories, United Kingdom. These included 2′-fucosyllactose (2′FL), 3-fucosyllactose (3FL), 3′-sialyllactose (3′SL), 6′-sialyllactose (6′SL), lacto-N-tetraose (LNT), disialyllacto-N-tetraose (DSLNT), Lactodifucotetraose (LDFT), lacto-N-difucohexaose 1 (LNDFH), lacto-N-fucopentaose (LNFP) 1, 2, and 3, lacto-N-hexaose (LNH), lacto-N-neotetraose (LNnT) and sialyl-lacto-N-tetraose (LST) a, b and c. Linear B6-Trisaccharide was used as an internal standard.
HMO quantification was performed as previously described68,69. Briefly 5 µl of human milk was combined with Linear B-6 Trisaccharide (Dextra Laboratories, UK) and HPLC grade water, then subjected to C18 columns (Thermo Scientific #60108-390) and carbograph columns (Thermo Scientific #60302-606) to remove proteins and salts respectively. Samples were labeled using 2-aminobenzamide (2-AB, Sigma) for 2 h at 65 °C. Excess 2-AB was removed using Silica columns (Thermo Scientific, #60300-482). Samples were separated by HPLC with fluorescence detection on a TSKgel Amide-80 column (Tosoh Bioscience, Tokyo, Japan) with a linear gradient of a 50 mM ammonium formate/acetonitrile solvent system. Retention times of purchased standard HMOs were used to annotate HPLC peaks. 6′SL and LNT peaks could not be separated and therefore, were calculated together. The amount of each individual HMO was calculated based on normalization to the internal standard (Supplementary Data 4). The relative abundance of each of the individual HMOs was determined by setting the sum of the 16 identified oligosaccharides as 100% total HMOs.
Statistical analysis
No statistical method was used to predetermine the sample size. The investigators were not blinded to allocation during experiments and outcome assessment. Independent t-test was performed to test between groups when mentioned using the R function “t-test”. Paired t-test was done between the percentages of infants in each cohort that had BL. infantis prior to 10 weeks and the percentage of infants that had detectable levels of BL. infantis after 10 weeks. Distances between consecutive infant gut microbiome samples and between breast milk HMO compositions were calculated using Bray-Curtis dissimilarity using the “vegdist” function from the vegan65 (2.6–4) package. Correlation between the microbial population and HMO composition was performed using Pearson correlation with the “cor” R function. Adjusted p-values were calculated using “corr.test” from the psych package70.
Linear association models
The “Maaslin2”71 R package was used to perform linear models in order to find associations between 16 HMOs and Bifidobacterium species and in the infant gut bacteria. The individual was set as a random factor to account for the effect of each mother-infant pair. In addition, “Maaslin2” was used to perform linear association models of age compared to Bifidobacterium species, adjusted to individuals.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
The Human-filtered metagenomic sequencing data generated in this study has been deposited in the SRA database under BioProject PRJNA994433. Metadata of the cohort is provided in Supplementary Data 1. HMO quantification results are provided in Supplementary Data 4 and MetaPhlAn results in Supplementary Data 5.
Code availability
Our tailored MetaPhlAn database is available on our GitHub page (https://github.com/yassourlab/MetaPhlAn-B.infantis/).
References
Westerfield, K. L., Koenig, K. & Oh, R. Breastfeeding: Common questions and answers. Am. Fam. Physician 98, 368–373 (2018).
Sela, D. A. & Mills, D. A. Nursing our microbiota: molecular linkages between bifidobacteria and milk oligosaccharides. Trends Microbiol. 18, 298–307 (2010).
Marcobal, A. et al. Consumption of human milk oligosaccharides by gut-related microbes. J. Agric. Food Chem. 58, 5334–5340 (2010).
Zivkovic, A. M., German, J. B., Lebrilla, C. B. & Mills, D. A. Human milk glycobiome and its impact on the infant gastrointestinal microbiota. Proc. Natl Acad. Sci. USA 108, 4653–4658 (2011).
Han, S. M. et al. Maternal and infant factors influencing human milk oligosaccharide composition: beyond maternal genetics. J. Nutr. https://doi.org/10.1093/jn/nxab028. (2021)
Lewis, Z. T. et al. Maternal fucosyltransferase 2 status affects the gut bifidobacterial communities of breastfed infants. Microbiome 3, 13 (2015).
Plows, J. F. et al. Longitudinal changes in human milk oligosaccharides (HMOs) over the course of 24 months of lactation. J. Nutr. 876-882 https://doi.org/10.1093/jn/nxaa427. (2021)
Yassour, M. et al. Natural history of the infant gut microbiome and impact of antibiotic treatment on bacterial strain diversity and stability. Sci. Transl. Med. 8, 343ra81 (2016).
Bäckhed, F. et al. Dynamics and stabilization of the human gut microbiome during the first year of life. Cell Host Microbe 17, 690–703 (2015).
Le Doare, K., Holder, B., Bassett, A. & Pannaraj, P. S. Mother’s Milk: a purposeful contribution to the development of the infant microbiota and immunity. Front. Immunol. 9, 361 (2018).
Sela, D. A. et al. Bifidobacterium longum subsp. infantis ATCC 15697 α-fucosidases are active on fucosylated human milk oligosaccharides. Appl. Environ. Microbiol. 78, 795–803 (2012).
Underwood, M. A. et al. A comparison of two probiotic strains of bifidobacteria in premature infants. J. Pediatr. 163, 1585–1591.e9 (2013).
Garrido, D., Dallas, D. C. & Mills, D. A. Consumption of human milk glycoconjugates by infant-associated bifidobacteria: mechanisms and implications. Microbiology 159, 649–664 (2013).
Garrido, D. et al. Comparative transcriptomics reveals key differences in the response to milk oligosaccharides of infant gut-associated bifidobacteria. Sci. Rep. 5, 13517 (2015).
Ruiz-Moyano, S. et al. Variation in consumption of human milk oligosaccharides by infant gut-associated strains of bifidobacterium breve. Appl. Environ. Microbiol. 79, 6040–6049 (2013).
Duar, R. M. et al. Comparative genome analysis of bifidobacterium longum subsp. infantis strains reveals variation in human milk oligosaccharide utilization genes among commercial probiotics. Nutrients 12, 3247 (2020).
LoCascio, R. G. et al. Glycoprofiling of bifidobacterial consumption of human milk oligosaccharides demonstrates strain specific, preferential consumption of small chain glycans secreted in early human lactation. J. Agric. Food Chem. 55, 8914–8919 (2007).
Duboux, S., Ngom-Bru, C., De Bruyn, F. & Bogicevic, B. Phylogenetic, functional and safety features of 1950s B. infantis strains. Microorganisms 10, 203 (2022).
Sela, D. A. et al. The genome sequence of bifidobacterium longum subsp. infantis reveals adaptations for milk utilization within the infant microbiome. Proc. Natl Acad. Sci. USA. 105, 18964–18969 (2008).
Derrien, M. et al. Gut microbiome function and composition in infants from rural kenya and association with human milk oligosaccharides. Gut Microbes 15, 2178793 (2023).
Pace, R. M. et al. Variation in human milk composition Is related to differences in milk and infant fecal microbial communities. Microorganisms 9, 1153 (2021).
Borewicz, K. et al. The association between breastmilk oligosaccharides and faecal microbiota in healthy breastfed infants at two, six, and twelve weeks of age. Sci. Rep. 10, 4270 (2020).
Barnett, D. J. M. et al. Human milk oligosaccharides, antimicrobial drugs, and the gut microbiota of term neonates: observations from the KOALA birth cohort study. Gut Microbes 15, 2164152 (2023).
Borewicz, K. et al. Correlating infant faecal microbiota composition and human milk oligosaccharide consumption by microbiota of one-month old breastfed infants. Mol. Nutr. Food Res. 63, e1801214 (2019).
Locascio, R. G. et al. A versatile and scalable strategy for glycoprofiling bifidobacterial consumption of human milk oligosaccharides. Microb. Biotechnol. 2, 333–342 (2009).
Seppo, A. E. et al. Infant gut microbiome is enriched with bifidobacterium longum ssp. infantis in old order mennonites with traditional farming lifestyle. Allergy 76, 3489–3503 (2021).
Young, S. L. et al. Bifidobacterial species differentially affect expression of cell surface markers and cytokines of dendritic cells harvested from cord blood. Clin. Diagn. Lab. Immunol. 11, 686–690 (2004).
Davis, J. C. C. et al. Growth and morbidity of gambian infants are influenced by maternal milk oligosaccharides and infant gut microbiota. Sci. Rep. 7, 40466 (2017).
Taft, D. H. et al. Bifidobacterium species Colonization in Infancy: a global cross-sectional comparison by population history of breastfeeding. Nutrients 14, 1423 (2022).
Tso, L., Bonham, K. S., Fishbein, A., Rowland, S. & Klepac-Ceraj, V. Targeted high-resolution taxonomic ldentification of bifidobacterium longum subsp. infantis using human milk oligosaccharide metabolizing genes. Nutrients 13, 2833 (2021).
Casaburi, G. et al. Metagenomic insights of the infant microbiome community structure and function across multiple sites in the United States. Sci. Rep. 11, 1472 (2021).
LoCascio, R. G., Desai, P., Sela, D. A., Weimer, B. & Mills, D. A. Broad conservation of milk utilization genes in bifidobacterium longum subsp. infantis as revealed by comparative genomic hybridization. Appl. Environ. Microbiol. 76, 7373–7381 (2010).
Blanco-Míguez, A. et al. Extending and improving metagenomic taxonomic profiling with uncharacterized species using MetaPhlAn 4. Nat. Biotechnol. https://doi.org/10.1038/s41587-023-01688-w (2023)
Yanokura, E. et al. Subspeciation of bifidobacterium longum by multilocus approaches and amplified fragment length polymorphism: description of B longum subsp. suillum subsp. nov., isolated from the faeces of piglets. Syst. Appl. Microbiol. 38, 305–314 (2015).
Yatsunenko, T. et al. Human gut microbiome viewed across age and geography. Nature 486, 222–227 (2012).
Avershina, E. et al. Bifidobacterial succession and correlation networks in a large unselected cohort of mothers and their children. Appl. Environ. Microbiol. 79, 497–507 (2013).
Bode, L. Human milk oligosaccharides: every baby needs a sugar mama. Glycobiology 22, 1147–1162 (2012).
Dai, D. L. Y. et al. Breastfeeding enrichment of B. longum subsp. infantis mitigates the effect of antibiotics on the microbiota and childhood asthma risk. Med (N Y) https://doi.org/10.1016/j.medj.2022.12.002 (2022).
Olm, M. R. et al. Robust variation in infant gut microbiome assembly across a spectrum of lifestyles. Science 376, 1220–1223 (2022).
Ojima, M. N. et al. Priority effects shape the structure of infant-type bifidobacterium communities on human milk oligosaccharides. ISME J. https://doi.org/10.1038/s41396-022-01270-3 (2022)
Asakuma, S. et al. Physiology of consumption of human milk oligosaccharides by infant gut-associated bifidobacteria. J. Biol. Chem. https://doi.org/10.1074/jbc.M111.248138 (2011).
Henrick, B. M. et al. Bifidobacteria-mediated immune system imprinting early in life. Cell https://doi.org/10.1016/j.cell.2021.05.030 (2021).
Shao, Y. et al. Stunted microbiota and opportunistic pathogen colonization in caesarean-section birth. Nature 574, 117–121 (2019).
Vatanen, T. et al. Genomic variation and strain-specific functional adaptation in the human gut microbiome during early life. Nat. Microbiol 4, 470–479 (2019).
Ferretti, P. et al. Mother-to-infant microbial transmission from different body sites shapes the developing infant gut microbiome. Cell Host Microbe 24, 133–145.e5 (2018).
Mitchell, C. M. et al. Delivery mode affects stability of early infant gut microbiota. Cell Rep. Med. 1, 100156 (2020).
Yassour, M. et al. Strain-level analysis of mother-to-child bacterial transmission during the first few months of Life. Cell Host Microbe 24, 146–154.e4 (2018).
Egan, M., O’Connell Motherway, M., Ventura, M. & van Sinderen, D. Metabolism of sialic acid by bifidobacterium breve UCC2003. Appl. Environ. Microbiol. 80, 4414–4426 (2014).
Lawson, M. A. E. et al. Breast milk-derived human milk oligosaccharides promote Bifidobacterium interactions within a single ecosystem. ISME J. 14, 635–648 (2020).
Carr, L. E. et al. Role of human milk bioactives on infants’ gut and immune health. Front. Immunol. 12, 604080 (2021).
Collado, M. C. et al. Longitudinal study of cytokine expression, lipid profile and neuronal growth factors in human breast milk from term and preterm deliveries. Nutrients 7, 8577–8591 (2015).
Beghini, F. et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife 10, e65088 (2021).
Díaz, R., Torres-Miranda, A., Orellana, G. & Garrido, D. Comparative genomic analysis of Novel bifidobacterium longum subsp. longum strains reveals functional divergence in the human gut microbiota. Microorganisms 9, 1906 (2021).
Kim, H.-B. et al. Development of real-tme PCR assay to specifically detect 22 bifidobacterium Species and subspecies using comparative genomics. Front. Microbiol. 11, 2087 (2020).
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Aronesty, E. ea-utils: Command-line Tools for Processing Biological Sequencing Data. https://www.researchgate.net/publication/319159822_ea-utils_Command-line_tools_for_processing_biological_sequencing_data (2011).
Wickham, H., François, R., Henry, L., Müller, K. & Vaughan, D. dplyr: A Grammar of Data Manipulation. https://dplyr.tidyverse.org/ (2023).
Wickham, H., Vaughan, D. & Girlich, M. tidyr: Tidy Messy Data. https://tidyr.tidyverse.org/ (2023).
Wickham, H. et al. Welcome to the tidyverse. J. Open Source Softw. 4, 1686 (2019).
Wickham, H. in ggplot2: Elegant Graphics for Data Analysis (ed. Wickham, H.) 241–253 (Springer International Publishing, 2016).
Pedersen, T. L. ggforce: Accelerating ‘ggplot2’. https://ggforce.data-imaginist.com/ (2022).
Neuwirth, E. RColorBrewer: ColorBrewer palettes. Nat. Genet. https://doi.org/10.1038/s41588-023-01586-6 (2022).
Wright, K. pals: Color Palettes, Colormaps, and Tools to Evaluate Them. https://kwstat.github.io/pals/ (2021).
Kolde, R. pheatmap: Pretty Heatmaps. https://rdrr.io/cran/pheatmap/ (2019).
Oksanen, J. et al. vegan: Community Ecology Package. https://cran.r-project.org/web/packages/vegan/vegan.pdf (2022).
Paradis, E. & Schliep, K. ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics 35, 526–528 (2019).
Xu, S. et al. Ggtree: A serialized data object for visualization of a phylogenetic tree and annotation data. iMeta 1, e56 (2022).
Jantscher-Krenn, E. et al. The human milk oligosaccharide disialyllacto-N-tetraose prevents necrotising enterocolitis in neonatal rats. Gut 61, 1417–1425 (2012).
Jantscher-Krenn, E. et al. Human milk oligosaccharides are present in amniotic fluid and how specific patterns dependent on gestational age. Nutrients 14, 2065 (2022).
Revelle, W. psych: Procedures for Psychological, Psychometric, and Personality Research. https://www.researchgate.net/publication/281345624_psych_Procedures_for_Psychological_Psychometric_and_Personality_Research_R_Package_Version_10-95 (2023).
Mallick, H. et al. Multivariable association discovery in population-scale meta-omics studies. PLoS Comput. Biol. 17, e1009442 (2021).
Acknowledgements
This work was funded in part by the Azrieli Foundation grant for faculty fellows (M.Y.), by the Israel Science Foundation grant 2660/18 (D.E. & M.Y.) by the Waterloo foundation (D.E. & M.Y.), and by the Austrian Science Fund (FWF), under project number KLI 784 (E.J.K). M.Y. is the Rosalind, Paul and Robin Berlin Faculty Development Chair in Perinatal Research.
Author information
Authors and Affiliations
Contributions
D.E. established the cohort, generated the sequencing data, quantified the HMO abundance and performed all analyses. S.S. assisted with the experimental setup. E.J.K. guided and taught the HMO-quantification method. M.Y. guided the work. D.E. and M.Y. wrote the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Ethics declarations All mothers have agreed to participate in our study, which was approved by the Hebrew University’s Institutional Review Board (IRB, approval number 20042021), and signed our consent forms for themselves and their infants. No compensation was provided to participants.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Ennis, D., Shmorak, S., Jantscher-Krenn, E. et al. Longitudinal quantification of Bifidobacterium longum subsp. infantis reveals late colonization in the infant gut independent of maternal milk HMO composition. Nat Commun 15, 894 (2024). https://doi.org/10.1038/s41467-024-45209-y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-024-45209-y
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.
