It is now well documented that the best way for an infant’s gut (largely sterile at birth) to get populated by the bacterial components of a healthy microbiota is through its mother’s milk. Over 700 species of bacteria have been found in breast milk including the prominent probiotics Lactococcus, Leuconstoc and Bifidobacterium. In “Body by Darwin” I described an extraordinary evolved symbiosis between Bifidobacterium longum and humans whereby human mothers express large quantities of complex long-chain sugars called oligosaccharides in breast milk even though their babies cannot digest them – they completely lack the enzymes to do the job. It turns out that the oligosaccharides are never meant for the baby – but to allow Bifidobacteria a competitive edge in the large intestine. B. longum, for instance, has a suite of 700 unique genes that allow it to digest oligosaccharides such that breast-fed babies become the perfect ecological niche for this bacterium. Now a group of scientists from Japan have added the enzymatic specifics to this story. They have shown that the two main species of Bifidobacterium – longum and bifidum – have evolved different ways to crack open specific oligosaccharides. Bifidobacterium bifidum, they say, possesses a glycoside hydrolase family 20 lacto-N-biosidase for liberating lacto-N-biose I from lacto-N-tetraose, an abundant oligosaccharide unique to human milk, while Bifidobacterium longum subsp. longum has a non-classified enzyme (LnbX), thanks to molecular evolution from the former enzyme, to exploit the same sugar. By determining the crystal structure of the catalytic domain of LnbX they provide evidence for the creation of a novel glycoside hydrolase family, GH136, and show that it is indispensable for B. longum growth on lacto-N-tetraose and is therefore a key genetic factor for persistence in the gut of breast-fed infants. Their results, they say, suggest that human milk oligosaccharides have been the main selective pressure for the evolution of the gene lnbX in B. longum, because a stronger correlation between the gene and bacterial persistence was detected in the stools of breast-fed infants than in those of mixed-fed infants. It seems likely that a whole range of bifidobacterial species and subspecies employ different strategies, involving the evolution of a family of enzymatic complexes, to exploit lacto-N-tetraose, so that this oligosaccharide can be shared among members of the genus and even among strains belonging to the same species/subspecies, but withheld from other bacterial species. it is a wonderful example of the evolution of a complex gut ecosystem of various bifidobacteria which avoids competition for a strategic sugar, to which they all have exclusive access, all of which has been driven by the complexity of mother’s milk.