Geographical location specific composition of cultured microbiota and Lactobacillus occurrence in human breast milk in China

Mengfan Ding ab, Ce Qi ad, Zhengyu Yang c, Shan Jiang c, Ye Bi c, Jianqiang Lai *c and Jin Sun *ab
aState Key Laboratory of Food Science and Technology, Jiangnan University, Wuxi, China. E-mail:
bSchool of Food Science and Technology, Jiangnan University, Wuxi, China
cNational Institute for Nutrition and Health, Chinese Center For Disease Control And Prevention, Beijing, China. E-mail:
dInstitute of Nutrition and Health, Qingdao University, Qingdao, China

Received 6th November 2018 , Accepted 5th January 2019

First published on 7th January 2019

Breast milk bacteria play an important role in the early development of the gut microbiota and the immune system. Dominant living bacteria of 89 healthy Chinese women from 11 cities in five regions were analysed by broad-range yeast extract, casitone, and fatty acid and de Man, Rogosa, and Sharpe-based culturing coupled with 16S rRNA sequence and quantitative polymerase chain reaction. Principal coordinate analysis showed that human breast milk samples were classified into three groups, driven by Enterococcus (abundance in group 1, 63.13%), Streptococcus (abundance in group 2, 68.16%) and Staphylococcus (abundance in group 3, 55.17%). The microbiota profile was highly region-specific. Samples from the Northwest and North of China showed higher alpha diversity compared to other regions (p < 0.05). Staphylococcus, Streptococcus, and Enterococcus were the dominant genera in all samples. Lactobacillus had a high occurrence in samples from the Northwest and North, dominated by Lactobacillus reuteri and Lactobacillus gasseri. Samples of mothers with a high postpartum body mass index showed more Staphylococcus and less Lactobacillus and Streptococcus. Staphylococcus was negatively correlated with Lactobacillus and Streptococcus. The mode of delivery also affected the composition of microbiota, even after culture. These findings indicate differences between the North and South, provide effective information for collection of samples in which Lactobacillus is the predominant genus, and lower the detection limit for small amounts of bacteria.

1. Introduction

Human breast milk (HBM)-associated microbes are among the first to colonise the infant gut and they protect the newborn from pathogenic bacteria and contribute to the development and maturation of the infant immune system.1 Recently, a study based on next generation sequencing (NGS) revealed that the presence of specific HBM bacteria was dependent on the geographical location.2 Several studies have focused on the microbiota of HBM in China based on their geography, and China's vast size and its long history contribute to dietary and cultural differences between geographical regions,3 which is a good independent variable for correlation with variations in the composition of HBM. A NGS-based study in Taiwan and mainland China identified the predominant bacterial families, including Streptococcaceae, Staphylococcaceae, and Pseudomonadaceae.2 Samples from the most developed cities, Beijing, Guangzhou and Suzhou, were dominated by streptococci and staphylococci.4 HBM microbiota composition of China also differed considerably from that of countries of Europe (Spain and Finland) and Africa (South Africa).5

Several research studies suggested that HBM is a rich source of many mutualistic, beneficial, and potentially probiotic bacteria.6 However, recent studies conducted in China showed that Bifidobacterium and Lactobacillus were present in few samples and in low abundance.4 In addition, Staphylococcus and Streptococcus and low levels of Bifidobacterium, Enterobacteriaceae/Enterococcus and Clostridium were detected, whereas no Bacteroides or Lactobacillus were identified because of medium selectivity/electivity bias related to culture.3 When isolated, Bifidobacterium and Lactobacillus only represented approximately 1.7% and 0.4% of all the isolated bacteria, respectively.3 The low detection rate of HBM probiotics may be related to the low diversity of the samples studied and the weak sensitivity of the research methods.

The Chinese Centre for Disease Control and Prevention (CDC) recently developed a regional HBM composition databank based on the analysis of samples from 11 provinces, municipalities, or autonomous regions.7 HBM samples in this database have high geographical diversity that can help to discover the possible source of HBM with high probiotic. Recent studies characterizing the HBM microbiome relied on culture-independent sequencing methods.8 However, there are limitations to the information that can be obtained from molecular approaches alone, and the isolation of organisms is required to define the roles of specific bacteria in maintaining the health and development of the newborn. Culturing determines the viable population in a community, while most molecular methods do not distinguish between DNA obtained from live or dead cells.9 Furthermore, culturing using selective media allows the growth and detection of less abundant bacteria that can be missed because of insufficient sequencing depth in culture-independent studies.10 Recently, it has been shown that yeast extract-casitone-fatty acid (YCFA) medium allows the growth of many species of microbes of complex systems such gut samples.11 We observed that the highest target HBM bacterial counts were generated by enrichment cultivation by a combination of YCFA and de Man, Rogosa, and Sharpe (MRS) media. Therefore, in this study, we aimed to (i) find the cultured microbiota in samples from 11 cities belonging to five regions, combining two culture media, MRS and YCFA; and (ii) discover possible geographical sources of potentially probiotic-rich HBM.

2. Materials and methods

2.1 Collection of milk samples

The HBM samples used in this study were collected from voluntary donors (n = 89) throughout China from January 2010 to December 2012,7 belonging to five regions, namely Northeast (n = 28): Heilongjiang Mudan (HM), n = 9; Heilongjiang Qiqihaer (HQ), n = 9; Heilongjiang Fuyu (HF), n = 10; South (n = 31): Yunnan Dali (YD), n = 9; Guangxi Zhuangzu (GZ), n = 6; Guangzhou Huangpu (GH), n = 9; Guangzhou Zengcheng (GZC), n = 7; Northwest (n = 9): Gansu Lintan (GL), n = 9; East (n = 14): Shanghai (S), n = 8; Zhejiang Zhoushan (ZZ), n = 6; and North (n = 7): Shandong Weihai (SW) n = 7. All participants received detailed information about the study and provided written informed consent. The study protocol was approved by the Medical Ethics Committees of the National Institute for Nutrition and Health, China CDC (NA20110322). The clinical trial registration number was ChiCTR-EOC-16009016. Participants were lactating women 20–35 years of age, healthy self-reported, who did not smoke or drink alcohol, and who had delivered and were breastfeeding a single healthy baby. Exclusion criteria included mastitis, any infectious disease (especially tuberculosis, viral hepatitis, and human immunodeficiency virus infection), cardiovascular disease, metabolic diseases (such as diabetes), mental health disorders, cancer or other malignant or degenerative diseases, inability to answer questions, and current participation in any other study related to nutrition or drug intervention. To obtain a sample, one full breast was emptied using a portable automatic breast pump (HNR/X-2108Z, Shantou, Guangdong, China) between 9 am and 11 am. Breast-milk samples were obtained in sterile tubes by manual expression using sterile gloves after cleaning the nipples and areola by wiping with a swab soaked in sterile water and discarding the first drops to minimize contamination according to the protocol of Albesharat et al.12 Immediately after collection, samples were placed in an ice-box and transported to a −80 °C freezer in the nearest participating laboratory. In order to eliminate the effects of the lactation stage, we only obtained samples on the 42nd day postpartum for analysis.

2.2 Enrichment cultivation, DNA extraction, and microbial diversity analysis

Milk samples were thawed on ice and centrifuged at 10[thin space (1/6-em)]000 rpm for 10 min,13 and the supernatant was discarded. The sediment was dissolved in phosphate-buffered saline (PBS) buffer. Aliquots (100 μL) underwent enrichment cultivation in MRS and YCFA media respectively at 37 °C for 24 h. MRS is always used to culture Lactobacillus, while YCFA contains a more complex microbiota, broad range bacteriological medium.11 Enriched microbiota was harvested by centrifugation at 10[thin space (1/6-em)]000 rpm for 10 min, respectively. Total DNA was extracted by repeated bead-beating and using a previously described column method14 with some modifications. Briefly, cell lysis was achieved by bead beating with zirconium beads (0.1 g, 0.7 mm: 0.3 g, 0.15 mm) on an oscillator (Precellys 24, Bertin Technologies, Montigny-le-Bretonneux, France) in the presence of 4% (w/v) sodium dodecyl sulfate (SDS), 500 mM NaCl, and 50 mM ethylenediaminetetraacetic acid (EDTA). The rotating speed of the oscillator was 6300 rpm with two circulations (30 s per circulation). Ammonium acetate was used to precipitate and remove the impurities, and SDS in addition to isopropanol precipitation was used for the recovery of the nucleic acids. RNA and proteins were removed or degraded using RNase and proteinase K, respectively, followed by the use of an Ezup Column Bacteria Genomic DNA Purification kit (Sangon Biotech Co., Ltd, Shanghai, China). The genomic DNA concentration was determined using a Nanodrop 1000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA), and the integrity of the DNA was determined by electrophoresis on agarose gels. The genomic DNA concentration was standardised for all samples to 20 ng μL−1 and mixed DNA extracted from two media in equal volume for further analysis.

2.3 16S rRNA gene sequencing

Microbial profiles were analysed by 16S rRNA sequencing at GENEWIZ, Inc. (Suzhou, China). For the library preparation, a library sequence of the V3 and V4 regions of the 16S rRNA was constructed using a 10 ng DNA aliquot. The V3 and V4 regions were amplified by polymerase chain reaction (PCR) using the following primer pair: forward 5′-CCT ACG GRR BGC ASC AGK VRV GAA T-3′ and reverse 5′-GGA CTA CNY VGG GTW TCT AAT CC-3′. The first-round PCR products were used as templates for a second round of amplicon enrichment by PCR (94 °C for 3 min, followed by 24 cycles at 94 °C for 5 s, 57 °C for 90 s, and 72 °C for 10 s, and a final extension at 72 °C for 5 min). At the same time, indexed adapters were added to the ends of the 16S rRNA amplicons to generate indexed libraries that were ready for downstream sequencing on the MiSeq platform (Illumina, San Diego, CA, USA). DNA libraries were validated using an Agilent 2100 bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) and quantified using a Qubit 2.0 fluorometer. DNA libraries were multiplexed and loaded on an Illumina MiSeq instrument according to the manufacturer's instructions. Sequencing was performed using a 2 × 300 paired-end configuration, and image analysis and base calling were conducted using the MiSeq Control Software embedded in the MiSeq instrument.

2.4 Quantitative PCR (qPCR)

The qPCR system consisted of 10 μL SYBR Green, 0.4 μL forward primer, 0.4 μL reverse primer, 1 μL DNA, and 8.2 μL double-distilled H2O (ddH2O). The amplification conditions were 95 °C, 5 min; 95 °C, 20 s; Tm shown in Table 1, 30 s; 72 °C, 1 min. A universal bacterial primer set was included as a reference gene. All qPCRs were performed in duplicate in transparent 96-well optical reaction plates sealed with optical adhesive film on an ABI Prism 7900HT system (Thermo Fisher Scientific). Following the thermocycling program, the raw fluorescence data recorded by the software-defined storage (SDS) software were exported to the LinRegPCR program.15,16 This software was used to perform baseline correction and calculate the mean PCR efficiency per amplicon group. This was used to calculate the initial quantities N0 (arbitrary fluorescence units) for each amplicon using the formula: N0 threshold/(EffCTmean), where Effmean denotes the mean PCR efficiency per amplicon, threshold is the optimal cut-off in the exponential region, and CT is the cycle number where each sample exceeds this threshold. The relative abundances of the eight specific amplicon species were obtained by normalisation to the N0 value obtained for the universal bacterial amplicon group determined in the same array.
Table 1 Primers used in quantitative real-time PCR
Bacteria Sequence (5′–3′) T m (°C) Ref.
Total bacteria EUB338: ACTCCTACGGGAGGCAGCAG 56 17
Lactobacillus fermentation F: AATACCGCATTACAACTTTG 60 18
Lactobacillus rhamnosus F: CAACCGGTCCRTGAACCGT 60 19
Lactobacillus salivarius F: GTCGTAACAAGGTAGCCGTAGGA 60 20
Lactobacillus casei F: ACCGCATGGTTCTTGGC 60 21
Lactobacillus reuteri F: GAACGCAYTGGCCCAA 60 21
Lactobacillus plantarum F: CTCTGGTATTGATTGGTGCTTGCAT 60 21
Lactobacillus delbrueckii F: GGRTGATTTGTTGGACGCTAG 60 22
Streptococcus salivarius F: AACGTT GAC CTT ACG CTA GC 55 23

2.5 Statistical analysis

The Quantitative Insights Into Microbial Ecology (QIIME; ver. 1.9.1) data analysis package was used for 16S rRNA data analysis. The forward and reverse reads were joined and assigned to samples based on barcodes and truncated by cutting off the barcode and primer sequences. Quality filtering of the joined sequences was performed, and the sequences that did not fulfill the following criteria were discarded: sequence length <200 bp, no ambiguous bases, and the mean quality score ≥20. The effective sequences were used in the final analysis. Sequences were grouped into operational taxonomic units (OTUs) using the clustering program VSEARCH (ver. 1.9.6) against the SILVA 119 database that was pre-clustered at a 97% sequence identity. The Ribosomal Database Project (RDP) classifier was used to assign a taxonomic category to all the OTUs at a confidence threshold of 0.8. Taxonomic categories at the species level were predicted with the RDP classifier and the SILVA 119 database.

Sequences were rarefied prior to calculation of alpha and beta diversity statistics. Alpha diversity indices were calculated in QIIME from rarefied samples using the Shannon and Simpson indexes for diversity, and the Chao1 and ACE indexes for richness. Beta diversity was calculated using weighted and unweighted UniFrac distances, and principal coordinate analysis (PCoA) was performed. An arithmetic mean phylogenetic tree was constructed from the beta diversity distance matrix with an unweighted pair group method.

The SPSS 20.0 software package and Origin 8.5 were used to compare alpha diversity indexes, relative abundance at the phylum and genus levels, and qPCR results. Results about specific Lactobacillus strains were also analysed using Cluster 3.0 available on the Internet and the TreeView tool ( Results were expressed as the mean ± standard error of the mean (SEM). Data were tested for normality (by the Shapiro–Wilk test) and homoscedasticity (by the Levene test). When the data met the normal distribution criteria, one-way analysis of variance was performed. If the differences were statistically significant, the data were analysed using the Kruskal–Wallis test comparing medians across groups. LDA Effect Size (LEfSe) (http: //hutten hower.sph.harvard. edu/galaxy/) was used to identify taxa that differed consistently between sample types.

3. Results

3.1 Participant characteristics

In this study, 6848 mothers who met the inclusion criteria were recruited (Fig. 1). A total of 6481 HBM samples were collected, including colostrum, transitional milk, and mature milk. Samples (n = 660) from 42 days were collected. Excluding samples that were used for other analyses, no more volumes for analysing (n = 102) and samples reserved for other analyses (n = 469), only 89 samples from five regions were used for HBM microbiota analysis (Fig. 2). No significant difference was found for all characteristics among the five regions; however, there was a significant difference among 11 cities. The average body mass index (BMI) of mothers before pregnancy in the HM group was significantly higher than that of the HQ, HF, and GH groups (p < 0.01, p < 0.05, and p < 0.01, respectively). A higher body weight before pregnancy was found in the HM group compared to the HQ, YD, and GH groups (p < 0.05). There was also a significantly different weight gain of mothers before delivery in the HM and YD groups (p < 0.05), HM and ZZ groups (p < 0.05), HQ and YD groups (p < 0.05), HQ and ZZ groups (p < 0.05), S and YD groups (p < 0.05), and GH and YD groups (p < 0.05), and mothers after delivery in GZC had a relatively lower BMI than that in HM (Table 2).
image file: c8fo02182a-f1.tif
Fig. 1 Flow diagram of the enrollment and collection of breast-milk samples and choice samples for BM microbiota analysis.

image file: c8fo02182a-f2.tif
Fig. 2 Alpha diversity of microbiota in human breast milk and that cultured using a combination of different media.
Table 2 Basic characteristics of study participants
Region and city n Preterm Full term Vaginal delivery Caesarean delivery MWb (kg) MWGa (kg) BMIb BMIa
Abbreviations: BMI, Body Mass Index; MW, maternal weight; MWG, maternal weight gain; GH, Guangzhou Huangpu; GL, Gansu Lintan; GZ, Guangxi Zhuangzu; GZC, Guangzhou Zengcheng; HF, Heilongjian, Fuyu; HM, Heilongjian Mudanjiang; HQ, Heilongjian Qiqihaer; S, Shanghai; SW, Shandong Weihai; YD, Yunnan Dali; ZZ, Zhejiang Zhoushan. Data are shown as the median and interquartile.a During pregnancy until 48 h before delivery.b Before pregnancy. —No information. *p < 0.05 compared with HM. **p < 0.01 compared with HM. #p < 0.05 compared with YD. &p < 0.05 compared with ZZ.
Northeast 28 2 26 12 16 53.7 (40–70) 17.5 (10–31) 20.8 (17.2–25.7) 26.4 (19.5–36)
HM 9 2 7 3 6 57.3 (53–65) 18.2 (11–26)# 22.6 (19.5–25.7) 30 (25.3–36)
HQ 9 0 9 4 5 50.3 (43–70) 18.1 (11.5–31)#& 20.1 (17.2–24.5)** 26.5 (22–31.6)
HF 10 0 10 5 5 53.4 (40–65) 16.1 (10–20.5) 19.9 (17.3–23.4)* 26.3 (19.5–32)
South 31 1 30 17 14 51.3 (39.5–80) 15.4 (3–26) 20.8 (15.8–35.1) 27.3 (19.7–42
YD 9 0 9 3 6 50.4 (39.5–65) 13.8 (9–15) 20.8 (16.9–26.3) 26.8 (23.6–31)
GZ 6 1 5 5 1 50.5 (47–55) 15 (12–18) 21.3 (18.4–21.3) 27.6 (25.7–30)
GH 9 0 9 6 3 50.2 (40–61) 17.2 (3–26)# 18.8 (15.8–23.8)** 26.2 (19.7–31)
GZC 7 4 3 3 4 53.9 (40–80) 15.6 (12–20) 22.4 (18.4–35.1) 29 (24.6–42)*
Northwest 9 0 9 8 1
GL 9 0 9 8 1
East 14 1 13 5 9 54.8 (45–65) 14.5 (4–23) 21.5 (18.9–25.4) 27 (21–35.2)
-S 8 1 7 2 6 55.9 (45–65) 19.1 (15–23)# 21.9 (19.1–25.4) 29 (24.6–35.2)
ZZ 6 0 6 3 3 53.6 (47.5–62) 9.8 (4–19)# 21.1 (18.9–24.7) 25(21–28.1)
North 7 1 6 4 3 55 (45–57) 16.9 (8–23) 20.8 (16.7–25.7) 27.5 (23–30.7)
SW 7 1 6 4 3 55 (45–57) 16.9 (8–23) 20.8 (16.7–25.7) 27.5 (23–30.7)

3.2 Microbiota characteristics of all samples

We first used samples to examine the microbial diversity of HBM after pre-enrichment using five media, namely MRS, YCFA, M17, glucose peptone extract medium (GYP), and gut microbiota medium (GMM). Combined with the Simpson and Shannon index, the highest diversity of HBM was obtained by a combination of MRS and YCFA pre-culturing (Fig. 2).

We obtained 63[thin space (1/6-em)]599.4 ± 13[thin space (1/6-em)]163.73 reads and 93 OTUs of all participants. The OTUs were classified into three phyla, six classes, 14 orders, 23 families, 28 genera, six species, and an unclassified group. Fig. 3A shows the top 30 bacterial genera in all samples. Relative abundance was detected in more than 0.01% and 1%. Staphylococcus was detected in each sample when the limit was more than 0.01%, followed by Bacillus (78), Enterococcus (68), Streptococcus (68), and Lactobacillus (36); the other genera only appear in part of individuals less than 20. Bacterial occurrence is quite different when the content limit was increased to 1%. BM containing the relative abundance of Staphylococcus more than 1% decreased from 89 to 79. Bacillus, Enterococcus, Streptococcus, and Lactobacillus occurrence was reduced from 78 to 30, 68 to 16, 68 to 36, and 36 to 15, respectively.

image file: c8fo02182a-f3.tif
Fig. 3 Dominant bacteria of all samples and region independent cluster analysis. (A) The top 30 genera discovered in all samples with their occurrence. (B) OUT cluster heatmap. Six specific Clusters I, II, III, IV, V and VI were enriched with OTU4 (g_Bacillus), OUT2 (g_Enterococcus), OUT1 (g_Staphylococcus), OUT7 (g_Lactobacillus), OUT8 (f_Enterobacteriaceae) and OUT3 (g_Streptococcus), respectively. The corresponding value of color is the value of the relative abundance of each OTU after normalization. (C) Principal coordinate analysis (PCoA) of unweighted UniFrac distances of the breast milk microbiota at each sample point. (D) Bacterial composition of three groups at the genus level. (E) Alpha diversity index (Chao1, Shannon, and Simpson). (F) Relative abundance of four main bacteria in three groups; bars with different letters mean significant discrepancy between groups (P < 0.05).

Cluster analysis (Fig. 3B) divided HBM into six specific Clusters, I, II, III, IV, V, and VI, which were enriched with OTU4 (g_Bacillus), OUT2 (g_Enterococcus), OUT1 (g_Staphylococcus), OUT7 (g_Lactobacillus), OUT8 (f_Enterobacteriaceae), and OUT3 (g_Streptococcus), and PCoA (Fig. 3C) showed that all samples could be divided into three groups. Group 1 was enriched with Enterococcus, group 2 was enriched with Streptococcus, and group 3 was enriched with Staphylococcus, accounting for 63.13%, 68.16%, and 55.17%, respectively (Fig. 3D). Bacillus was a group 3-specific genus accounting for 15.17% of all genera. Lactobacillus was a dominant genus in group 3 (10.49%), and its relative abundance was less than 1% in groups 1 and 2 (Fig. 3D). MRS medium is preferred for the growth of Lactobacillus. MRS enrichment is helpful to decrease the baseline of the detection rate, but it can cause potential bias in evaluating the abundance of Lactobacillus. Group 3 had a significantly higher relative abundance of Staphylococcus compared to the other two groups (p < 0.05). Group 2 showed a significantly higher Streptococcus abundance compared to the other two groups (p < 0.05). Group 2 showed a significantly higher alpha diversity based on the Chao1 and Shannon indices compared to the other two groups (Fig. 3E).

3.3 Correlations between maternal BMI and the four main genera

There was no correlation between pre-pregnancy BMI and the four dominant genera (Lactobacillus, Streptococcus, Staphylococcus and Enterococcus). Postpartum BMI showed a significantly negative correlation with Lactobacillus (R = −0.204, p = 0.085) and a positive correlation with Staphylococcus (R = 0.325, p = 0.005) (Fig. 4A). Staphylococcus had a negative correlation with Lactobacillus (R = −0.408, p < 0.01) and Streptococcus (R = −0.502, p < 0.01) (Fig. 4B). HBM samples from mothers with vaginal delivery had a significantly lower relative abundance of Staphylococcus (p < 0.05) and Enterococcus (p < 0.05) but a higher abundance of Streptococcus (p < 0.071) and Lactobacillus (p < 0.05) compared to those with a caesarean section (Fig. 4C).
image file: c8fo02182a-f4.tif
Fig. 4 (A) The correlation between Lactobacillus, Streptococcus, Staphylococcus and BMI of mothers during pregnancy. (B) The correlation between Lactobacillus and Staphylococcus, Streptococcus and Staphylococcus. (C) Effect of delivery mode on a major genus. (*P < 0.05 versus vaginal delivery.)

3.4 Region-independent microbiota pattern and alpha diversity

Fig. 5A shows the region-specific profile of bacterial genera in HBM. There were five dominant genera in all regions: Staphylococcus, Streptococcus, Bacillus, Enterococcus, and Lactobacillus. Staphylococcus (30–60%) and Streptococcus (3–30%) were the most prevalent genera in HBM samples of all regions except Northwest, followed by Enterococcus (0–22%) and Bacillus (1–18%). Samples from Northwest had a significant higher mean relative abundance (44.6%) of Lactobacillus compared to Northeast, South, and East (p < 0.01). Bifidobacterium was not detected in any samples. Fig. 5B revealed that HBM samples of both Northwest and North had the highest alpha diversity. The Shannon and Simpson indices of samples from both Northwest and North were significantly higher and lower than those of East, South, and Northeast, respectively (p < 0.05). The Chao1 index of samples from both Northwest and North regions was significantly higher than that of East and South (p < 0.05).
image file: c8fo02182a-f5.tif
Fig. 5 (A) Sampling location and genus more than 1%. (B) Alpha diversity of BM microbiota of 5 parts including the Shannon Index, Simpson Index and Chao1 Index. Data represent the means ± SEM (#P < 0.05 versus the North group, *P < 0.05 versus Northwest, **P < 0.01 versus Northwest). (C) Predominant bacteria of each city described by linear discriminant analysis. (D) Four main bacteria at the genus level in 11 cities; different letters above the line mean significant difference between cities (P < 0.05).

3.5 City-specific microbiota composition of HBM

LEfSe analysis identified the most differentially abundant taxa among HBM samples from 10 cities. Staphylococcus was the genus specific for samples from HM and GZC, while Streptococcus was the typical genus in the samples from YD, GH, and HF. Lactobacillus and Enterococcus were the marker genera in the samples from GL and S (Fig. 5C). No marker bacteria were found in samples from HQ.

Statistical comparison was carried out to investigate the relative abundance of the four most dominant genera (Fig. 5D).

Lactobacillus was enriched in samples from GL, and its abundance was significantly higher than that of samples from other cities, except SW and YD. Streptococcus was enriched in samples from HF, and its abundance was significantly higher than that of samples from GL, S, ZZ, GZC, and HM. Staphylococcus was the most dominant genus in samples from most cities, especially GH and GZC, and its abundance was significantly higher than that of GL. The abundance of Enterococcus in samples from ZZ and HM was significantly higher than that of samples from all other cities, except SW and S.

3.6 The relative abundance of specific Lactobacillus strains and Streptococcus salivarius determined by qPCR

According to a previous report, eight Lactobacillus strains in HBM that had potential for application in food production in China, namely L. plantarum, L. salivarius, L. reuteri, L. gasseri, L. fermentation, L. casei, L. delbrueckii, L. rhamnosus, and S. salivarius, have potential for application in yogurt production. Table 3 shows that S. salivarius was found in 26% HBM, and L. reuteri was detected in 16% HBM, followed by L. gasseri (12%), L. plantarum, L. fermentation (8%), and L. casei (2%), whose relative abundance was more than 0.001%. L. reuteri was mainly found in samples from GL and YD, L. gasseri was mainly detected in samples from S, GH, HQ, and HF, and L. casei and L. fermentation appeared in samples from SW, GH, HQ, and HF. The remaining Lactobacillus had almost no regional specificity (Fig. 6).
image file: c8fo02182a-f6.tif
Fig. 6 The occurrence of specific Lactobacillus and S. salivarius in HBM from 11 cities by cluster analysis. The colour key represents the adjusted value of relative abundance for specific Lactobacillus and S. salivarius species. LR, Lactobacillus reuri; LP, Lactobacillus plantarum; LC, Lactobacillus casei; LG, Lactobacillus gasseri; LF, Lactobacillus fermentation; LS, Lactobacillus salivarius; SS, Streptococcus salivarius; GL, Gansu Lintan; SW, Shandong Weihai; S, Shanghai; ZZ, Zhejiang Zhoushan; YD, Yunnan Dali; GZ, Guangxi Zhuangzu; GH, Guangzhou Huangpu; GZC, Guangzhou Zengcheng; HM, Heilongjian Mudanjiang; HQ, Heilongjian Qiqihaer; HF, Heilongjian, Fuyu.
Table 3 The occurrence (>0.001%) of specific Lactobacillus and Streptococcus salivarius in 11 cities
Abbreviations: LR, Lactobacillus reuri; LP, Lactobacillus plantarum; LC, Lactobacillus casei; LG, Lactobacillus gasseri; LF, Lactobacillus fermentation; LS, Lactobacillus salivarius; SS, Streptococcus salivarius; GL, Gansu Lintan; SW, Shandong Weihai; S, Shanghai; ZZ, Zhejiang Zhoushan; YD, Yunnan Dali; GZ, Guangxi Zhuangzu; GH, Guangzhou Huangpu; GZC, Guangzhou Zengcheng; HM, Heilongjian Mudanjiang; HQ, Heilongjian Qiqihaer; HF, Heilongjian, Fuyu.
GL 5 1 0 4 1 0 1
SW 1 0 1 0 1 0 4
S 1 1 0 1 0 0 2
ZZ 0 0 0 0 0 0 0
YD 4 2 0 1 0 0 2
GZ 0 0 0 0 0 0 0
GH 0 1 1 2 0 0 3
GZC 1 0 0 0 1 0 0
HM 0 1 0 0 0 0 3
HQ 2 1 0 1 2 0 4
HF 1 0 0 2 2 0 4
Total 16% 8% 2% 12% 8% 0 26%

4. Discussion

In this study, HBM from samples from 11 cities belonging to five regions was cultured in MRS and YCFA and analysed by 16S rRNA sequencing and qPCR. HBM bacterial composition showed regional specificity and Northwest and North had higher alpha diversity. Lactobacillus had a higher occurrence in Northwest and North compared with L. reuteri and L. gasseri. Postpartum BMI was negatively correlated with Lactobacillus and positively correlated with Staphylococcus. Lactobacillus and Streptococcus showed negative correlation with Staphylococcus. We also showed that the relative abundance of the four main genera was affected by delivery mode.

The four dominant genera were Staphylococcus, Streptococcus, Enterococcus, and Lactobacillus, depending on specific individuals. Finegoldia, Bifidobacterium, Propionibacterium, Yersinia, and other bacteria could also be detected (less than 1% abundance). These results are in agreement with previous studies24–28 that focused on how geographical location directly affected the microbiota in BM samples from mothers. However, there are also studies that have identified an higher abundance of Bacillus and Enterobacteriaceae in samples from Canadian and Irish populations compared to other taxonomic groups, which were inconsistent with our study because of differences in sample collection.29

Using NGS, Shiao-Wen found that HBM samples from Taiwan and mainland China could be divided into three clusters, dominated by Staphylococcaceae, Streptococcaceae, and Pseudomonadaceae at the family level.2 In our study, HBM could be divided into three groups according to PCoA analysis, driven by the genera Enterococcus, Streptococcus, and Staphylococcus. A wide variety of skin- and enteric-associated bacteria observed in BM samples are potentially opportunistic pathogens; however, our results suggested that the growth of these pathogens was suppressed by bacteria such as Lactobacillus and Streptococcus.

Our study gives insights into the host–microbe cross-talk in HBM associated with the mother's BMI before pregnancy and after delivery. High pre-pregnancy BMI and excessive weight gain during pregnancy have previously been associated with aberrations in the composition of microbiota in the maternal gut.30 Thus, high BMI may also have effects on the quantity of HBM bacteria. Here, we found high BMI in mothers with high relative abundance of Staphylococcus and low Streptococcus and Lactobacillus. This may indicate an imbalance in the microbiota of HBM, as previously documented for gut microbiota in inflammatory and obese states.31

Studies about HBM Lactobacillus, potentially probiotic bacteria, of healthy Iranian mothers suggested that the HBM microbiome was significantly influenced by several factors, the mode of delivery, rural or urban location, and lactation time.13,32 We also proved that the mode of delivery was a key factor in influencing HBM bacteria even after culturing with two media. It is known that caesarean delivery is associated with more Staphylococcus and Enterococcus present in HBM, while vaginal delivery is associated with more Lactobacillus and Streptococcus. Different lactation stages can affect the composition of the mother's milk flora, resulting in different cultured active microbiota in vitro. Mature milk shows more Pseudomonas and Bifidobacterium compared to colostrum and transition milk. Colostrum milk contains more Pseudomonas and Rothia compared to transition and mature milk.3 Our study only focuses on the cultured bacteria in mature milk due to sample collection.

Different climates, eating habits, and diets may result in a unique composition of microorganisms in participants. Increasing evidence shows that bacterial populations in the large intestine responded to changes in the diet, in particular to the type and quantity of dietary carbohydrate.16,33 A study also described gluten diet-associated changes in microorganisms in the small intestine of adults and children.34 In the Northwest and North of China, pasta is the main food, while in the East, South, and Northeast, it is mainly rice, which contains more carbohydrate. Northwestern people live in prairies with more dairy food,35 and we suppose that this is the reason why their BM has more Lactobacillus and high diversity.

Several studies extracted DNA from milk samples directly; thus, only a small amount of Lactobacillus (3.2%) could be tested.4,36 In this study, we observed a high quantity of Lactobacillus in the YD and GL groups (more than 90% after culturing). Therefore, the method that we used resulted in milk samples with an abundance of Lactobacillus, indicating the effectiveness of this method in contrast with studies that used selective agar.37,38 We identified eight specific Lactobacillus. L. reuteri and L. gasseri are two popular species that have been identified using qPCR. However, surprisingly, one study aimed to see the composition of Lactobacillus and found L. gasseri to be dominant, while just a few L. fermentation strains were found.39 Besides using a different culturing method, dietary, host, and sampling time factors, different outcomes in pyrosequencing analyses and qPCR may be due to geographical differences.

Whole genome sequencing showed that L. rhamnosus, L. ruminis, L. delbrueckii, L. plantarum, L. casei, and L. acidophilus were the most represented strains, with estimated amounts ranging between 6 and 8[thin space (1/6-em)]log, while L. reuteri, L. gasseri, L. fermentation, and L. salivarius were detected at 4 or 5[thin space (1/6-em)]log in human faeces at three time points (0, 670, and 700 days).40 Nevertheless, L. reuteri and L. gasseri seem to be the most represented strains that have regional specificity.

Although L. reuteri and L. gasseri are the predominant lactobacilli species, they are only found in individual HBM. Overall, 16% of mothers had detectable L. reuteri and 12% L. gasseri in their milk, similarly to a previous study.28 Reuter and Mitsuoka intensively studied the Lactobacillus biota of the human digestive tract and reported that L. reuteri was one of the dominant Lactobacillus species, was regularly detected, and was similar to other intestinal tract bacteria. However, their low prevalence in humans observed in more recent studies suggests the reduction of the L. reuteri population size over the past 50 years, and we need more data to identify whether it is because of the development of diets.41L. gasseri was recovered only from the milk of women with normal weight. An attractive association of L. gasseri with prevention of obesity in animals and humans has been described42. L. gasseri commonly colonises the oral cavity, vagina, and gastrointestinal (GI) tract, and it has the ability to tolerate acidic gastric conditions.43 Thus, the isolates of L. gasseri are potential candidates for application as probiotics.43

Streptococcus is regarded as the second most abundant bacterium in HBM. S. salivarius is abundant in the human mouth and always colonises the jejunum and ileum.44 It can inhibit the production of interleukin 8 (IL-8) to influence the immune response.15S. salivarius VM18 can inhibit R5 HIV-1 Bal, and heat-killed S. salivarius has the ability to reduce the infectious capacity of X4 HIV-1 Hc4 and R5/X4 HIV-1c7/86;45 thus, S. salivarius may play an important role in structuring the immune system of infants. S. salivarius is a more common bacterium compared to Lactobacillus in our study. S. salivarius also can release acid, thus reducing the pH of the environment46 and inhibiting the growth of Gram-positive bacteria.

We stored the collected breast milk samples in a −80 °C freezer for further analysis. An emerging study suggests that cold storage of milk at a −20 °C freezer for 6 weeks does not significantly affect either the quantitative or the qualitative bacterial composition of breast milk.47 Excessive storage time may have effects on the composition of HBM.

The limitations of our study include: (1) the strict anaerobic preservation of samples cannot be satisfied when sampling, which may lead to potential bias in the detection of anaerobic bacteria. (2) The difference in proliferation efficiency of different bacteria in the enrichment medium may also cause potential bias.

5. Conclusions

In conclusion, HBM samples can be divided into three groups, driven by Enterococcus, Streptococcus, and Staphylococcus. The bacterial composition of HBM of mothers in North and Northwest regions is different from that of mothers from East, Northeast and South because of the eating habits and degree of industrialization. In our study, the culture method can efficiently selectively increase the relative abundance of Lactobacillus and other microbiota. In addition, culturing HBM before DNA extraction decreases the detection limit. L. reuteri and L. gasseri are the two lactobacilli that most individuals have, and S. salivarius is more common than Lactobacillus. BMI influences bacterial content that negatively correlates with Streptococcus and Lactobacillus and positively correlates with Staphylococcus. Streptococcus and Lactobacillus are two bacteria that may limit the growth of Staphylococcus. Pregnancy time also affects the composition of the flora, even after culturing. These findings provide information for researchers to collect samples that are rich in Lactobacillus as well as a method to detect Lactobacillus in low abundance.

Conflicts of interest

There are no conflicts to declare.


This study was funded by the National Key R&D Program of China (2017YFD0400601) and the National High Technology Research and Development Program of China (863 Program) (2010AA023004) and the Medical Ethics Committees of the China CDC (NA20110322) and the Wuxi Municipal Science and Education Strengthening Health Engineering Medical Key Discipline Construction Program (ZDXK003).

Notes and references

  1. M. Toscano, R. D. Grandi, E. Grossi and L. Drago, Front. Microbiol., 2017, 8, 2100–2105 CrossRef PubMed.
  2. S. W. Li, K. Watanabe, C. C. Hsu, S. H. Chao, Z. H. Yang, Y. J. Lin, C. C. Chen, Y. M. Cao, H. C. Huang and C. H. Chang, Front. Microbiol., 2017, 8, 965–980 CrossRef PubMed.
  3. T. Jost, C. Lacroix, C. Braegger and C. Chassard, Br. J. Nutr., 2013, 110, 1253–1262 CrossRef CAS PubMed.
  4. O. Sakwinska, D. Moine, M. Delley, S. Combremont, E. Rezzonico, P. Descombes, G. Vinyes-Pares, Y. Zhang, P. Wang and S. K. Thakkar, PLoS One, 2016, 11, e0160856 CrossRef PubMed.
  5. K. Himanshu, T. E. Du, K. Amruta, A. Juhani, K. M. Linderborg, Y. Zhang, M. P. Nicol, I. Erika, B. Yang and M. C. Collado, Front. Microbiol., 2016, 7, 1619 Search PubMed.
  6. M. K. Mcguire and M. A. Mcguire, Adv. Nutr., 2015, 6, 112–123 CrossRef CAS PubMed.
  7. Y. W. Hsu, T. H. Liou, Y. M. Liou, H. J. Chen and L. Y. Chien, Asia Pac. J. Clin. Nutr., 2016, 25, 108–117 CAS.
  8. J. L. Fitzstevens, K. C. Smith, J. I. Hagadorn, M. J. Caimano, A. P. Matson and E. A. Brownell, Nutr. Clin. Pract., 2017, 32, 354–364 CrossRef PubMed.
  9. J. M. Auchtung, C. D. Robinson and R. A. Britton, Microbiome, 2015, 3, 42–57 CrossRef PubMed.
  10. J. T. Lau, F. J. Whelan, I. Herath, C. H. Lee, S. M. Collins, P. Bercik and M. G. Surette, Genome Med., 2016, 8, 1–10 CrossRef PubMed.
  11. H. P. Browne, S. C. Forster, B. O. Anonye, N. Kumar, B. A. Neville, M. D. Stares, D. Goulding and T. D. Lawley, Nature, 2016, 533, 543–546 CrossRef CAS PubMed.
  12. R. Albesharat, M. A. Ehrmann, M. Korakli, S. Yazaji and R. F. Vogel, Appl. Microbiol., 2011, 34, 148–155 CrossRef CAS PubMed.
  13. P. Khodayarpardo, L. Mirapascual, M. C. Collado and C. Martínezcosta, J. Perinatol., 2014, 34, 599–605 CrossRef CAS PubMed.
  14. Z. Yu and M. Morison, Biotechniques, 2004, 36, 808–812 CrossRef CAS PubMed.
  15. G. Kaci, O. Lakhdari, J. Doré, S. D. Ehrlich, P. Renault, H. M. Blottière and C. Delorme, Appl. Environ. Microbiol., 2011, 77, 4681–4684 CrossRef CAS PubMed.
  16. H. J. Flint, S. H. Duncan, K. P. Scott and P. Louis, Environ. Microbiol., 2007, 9, 1101–1111 CrossRef CAS PubMed.
  17. M. J. Prol-García, M. Planas and J. Pintado, Aquaculture, 2010, 302, 26–35 CrossRef.
  18. E. M. Dickson, M. P. Riggio and L. Macpherson, J. Med. Microbiol., 2005, 54, 299 CrossRef CAS.
  19. C. H. Huang, M. T. Chang, L. Huang and W. S. Chu, Mol. Cell. Probes, 2015, 29, 531–533 CrossRef CAS.
  20. S. Harrow, V. Ravindran, R. Butler, J. Marshall and G. Tannock, Appl. Environ. Microbiol., 2007, 73, 7123–7127 CrossRef CAS PubMed.
  21. H. Tsuji, R. Oozeer, K. Matsuda, T. Matsuki, T. Ohta, K. Nomoto, R. Tanaka, M. Kawashima, K. Kawashima and S. Nagata, Benefic. Microbes, 2012, 3, 113–125 CrossRef CAS PubMed.
  22. R. Zhang, K. Daroczy, B. Xiao, L. Yu, R. Chen and Q. Liao, J. Med. Microbiol., 2005, 61, 729 CrossRef PubMed.
  23. V. Srinivasan, R. E. G. Jr, P. L. Shewmaker, S. Patrick, A. S. Chitnis, H. O'Connell, I. Benowitz, P. Patel, A. Y. Guh and J. Noblewang, PLoS One, 2012, 7, e32169 CrossRef CAS.
  24. K. M. Hunt, J. A. Foster, L. J. Forney, U. M. E. Schütte, D. L. Beck, Z. Abdo, L. K. Fox, J. E. Williams, M. K. Mcguire and M. A. Mcguire, PLoS One, 2011, 6, e21313 CrossRef CAS PubMed.
  25. E. Jimenez, J. de Andres, M. Manrique, P. Pareja-Tobes, R. Tobes, J. F. Martínez-Blanch, F. M. Codonar, D. Ramon, L. Fernandez and J. M. Rodriguez, J. Hum. Lact., 2015, 31, 406–415 CrossRef PubMed.
  26. C. Urbaniak, M. Angelini, G. B. Gloor and G. Reid, Microbiome, 2016, 4, 1 CrossRef.
  27. T. Jost, C. Lacroix, C. P. Braegger, F. Rochat and C. Chassard, Environ. Microbiol., 2014, 16, 2891–2904 CrossRef CAS PubMed.
  28. G. Sinkiewicz and L. Ljunggren, Microb. Ecol. Health Dis., 2008, 20, 122–126 CrossRef CAS.
  29. C. Urbaniak, J. Cummins, M. Brackstone, J. M. Macklaim, G. B. Gloor, C. K. Baban, L. Scott, D. M. O'Hanlon, J. P. Burton and K. P. Francis, Appl. Environ. Microbiol., 2014, 80, 3007 CrossRef PubMed.
  30. M. C. Collado, K. Laitinen, S. Salminen and E. Isolauri, Pediatr. Res., 2012, 72, 77–85 CrossRef CAS PubMed.
  31. N. M. Delzenne and P. D. Cani, Annu. Rev. Nutr., 2011, 31, 15–31 CrossRef CAS PubMed.
  32. M. Taghizadeh, M. Mirlohi, F. Poursina, G. Madani, M. Khoshhali, N. Bahreini and H. G. Safaei, Adv. Biomed. Res., 2015, 4, 141 CrossRef PubMed.
  33. P. Louis, K. P. Scott, S. H. Duncan and H. J. Flint, J. Appl. Microbiol., 2007, 102, 1197–1208 CrossRef CAS PubMed.
  34. E. Nistal, A. Caminero, A. R. Herrán, L. Arias, S. Vivas, J. M. de Morales, S. Calleja, L. E. de Miera, P. Arroyo and J. Casqueiro, Inflamm. Bowel. Dis., 2012, 18, 649–656 CrossRef PubMed.
  35. J. C. Ogier, O. Son, A. Gruss, P. Tailliez and A. Delacroixbuchet, Appl. Environ. Microbiol., 2002, 68, 3691–3701 CrossRef CAS PubMed.
  36. L. Qian, H. Song and W. Cai, Benefic. Microbes, 2016, 7, 559 CrossRef CAS PubMed.
  37. E. Jiménez and J. M. Rodríguez, Cultivation-Independent Assessment of the Bacterial Diversity of Breast Milk Among Healthy Women, Springer US, 2015 Search PubMed.
  38. G. Solís, C. G. de Los Reyes-Gavilan, N. Fernández, A. Margolles and M. Gueimonde, Anaerobe, 2010, 16, 307–310 CrossRef PubMed.
  39. R. Martín, S. Langa, C. Reviriego, E. Jimínez, M. L. Marín, J. Xaus, L. Fernández and J. M. Rodríguez, J. Pediatr., 2003, 143, 754–758 CrossRef PubMed.
  40. M. Rossi, D. Martínez-Martínez, A. Amaretti, A. Ulrici, S. Raimondi and A. Moya, Environ. Microbiol., 2016, 8, 399–406 CrossRef CAS PubMed.
  41. J. Walter, R. A. Britton and S. Roos, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 4645–4652 CrossRef CAS PubMed.
  42. Q. S. Damaceno, J. P. Souza, J. R. Nicoli, R. L. Paula, G. B. Assis, H. C. Figueiredo, V. Azevedo and F. S. Martins, Probiotics Antimicrob. Proteins, 2017, 9, 1–9 CrossRef PubMed.
  43. K. Selle and T. R. Klaenhammer, FEMS Microbiol. Rev., 2013, 37, 915–935 CrossRef CAS PubMed.
  44. C. Delorme, A. L. Abraham, P. Renault and E. Guédon, Infect., Genet. Evol., 2015, 33, 381–392 CrossRef CAS PubMed.
  45. V. Martín, A. Maldonado, L. Fernández, J. M. Rodríguez and R. I. Connor, Breastfeed. Med., 2010, 5, 153–158 CrossRef.
  46. S. Y. Yang, F. X. Lü, Z. X. Lu, X. M. Bie, Y. Jiao, L. J. Sun and B. Yu, Amino Acids, 2008, 34, 473–478 CrossRef CAS PubMed.
  47. M. L. Marín, A. Rebeca, J. Esther, G. Adolfo, F. Leonides and J. M. Rodríguez, J. Pediatr. Gastroenterol. Nutr., 2009, 49, 343–348 CrossRef.


These authors contributed equally to this work.

This journal is © The Royal Society of Chemistry 2019