Nobody buys Lactobacillus plantarum. The species name covers thousands of strains held in culture collections around the world, and two of them can behave so differently that sharing the name is a courtesy rather than a specification. What a buyer of probiotic cultures actually purchases is a strain — a deposit carrying a designation such as CCFM8724 or BA05, a fermentation history and, increasingly, a sequenced genome. Everything else on the quotation, the CFU count and the shelf life included, hangs off that designation. Four questions separate a usable specification from a marketing claim: which strain, counted how, counted when, and what happens to the cells between the fermenter and the filling line.

The strain suffix is the product

The codes appended to species names are not decoration. CCFM designations come from the food-microorganism collection at Jiangnan University, GDMCC from the Guangdong provincial collection, and house codes such as HH28 or LA-G80 are the manufacturer's own deposit numbers. Each identifies one isolate, kept and propagated under one set of conditions, with one dossier behind it. Bifidobacterium lactis BL03 and every other B. lactis on the market are different products that happen to share a species, and a quotation that stops at the species name has not described anything a formulator can defend on a label. Insist that the strain designation appears on the certificate of analysis, not only in the brochure.

CFU is a method, not a property

A CFU figure is the output of a plate count: dilution, plating on a selective agar — MRS for the lactobacilli, TPY or MRS with cysteine for the bifidobacteria — and anaerobic incubation for a day or two. Change the agar, the atmosphere or the counting rules and the number moves, which is why a count quoted without its method is not comparable to any other count. Two complications deserve attention. Cells clump, so one colony may come from a pair or a chain rather than a single cell, and some counting conventions divide by the clump size while others do not. Where flow cytometry is used instead of plating the figure usually reads higher, because it counts cells that are metabolically active but no longer able to form a colony. For a blend there is a further question of allocation: a ten-strain mix quoted at "100 billion CFU total" can hide a composition in which the hardiest organism holds most of the count. Ask for per-strain figures.

Counted when

The number that matters commercially is the one at the end of shelf life, and the number most suppliers quote first is the one at release. Both are legitimate as long as each is labelled for what it is, because release counts carry an overage — commonly one and a half to two times the label claim — to absorb losses in storage. The variable that governs those losses more than temperature alone is water activity: dried cultures are specified with an a_w typically at or below 0.25, and the moisture figure on the certificate is a proxy for it. Ask for stability data at the storage temperature you can actually offer rather than an accelerated condition, and check the sampling date printed against every figure supplied.

Bifidobacterium is the harder half of the catalogue

The bifidobacteria are strict anaerobes and noticeably less forgiving than the lactobacilli, in the fermenter and after drying alike. Bifidobacterium breve CCFM1025, Bifidobacterium animalis BA05, Bifidobacterium bifidum BB67 and Bifidobacterium infantis CCFM687 all demand oxygen-barrier packaging — nitrogen flush, aluminium foil laminates, an oxygen absorber in the drum — and they lose count faster than the lactobacilli when exposed. Their gastric and bile tolerance is lower too, which is why encapsulation and enteric-coating decisions belong in the discussion at quotation stage rather than after the first trial. The trade-off is real: what the genus gives up in robustness it returns in claims the lactobacilli cannot make, particularly in the infant-nutrition and gut-transit areas.

The genus changed under everyone

In 2020 the Lactobacillus genus was split. Lactobacillus plantarum CCFM8724 and Lactobacillus plantarum CN2018 now sit formally in Lactiplantibacillus; Lactobacillus casei LC-G11 in Lacticaseibacillus; Lactobacillus reuteri CCFM1040 in Limosilactobacillus — while Lactobacillus helveticus HH28, Lactobacillus acidophilus LA-G80, Lactobacillus crispatus HLC28 and Lactobacillus gasseri HLG13 stayed in the shrunken genus. Suppliers differ in whether their paperwork has caught up, so a label name and a certificate name may disagree without anyone being wrong. The fix is procedural: make the strain designation the contract identifier and let the taxonomy follow whichever current name the supplier uses consistently.

One of them is not a bacterium

Saccharomyces boulardii HS5 is a yeast, and the difference is not cosmetic. Antibiotics aimed at bacteria — cell-wall synthesis, the bacterial ribosome — leave it untouched, which is why it is specified alongside antibiotic courses, while antifungals are what kill it. Quality control runs on different agar, Sabouraud or YPD rather than MRS; Gram staining is irrelevant; and the cells are an order of magnitude larger than bacteria, which matters at the encapsulation step. It survives gastric passage well, one of the reasons it keeps its place in travellers' diarrhoea formulations.

Specialists, starters and workhorses

The remaining profiles show how far the rest of the range differs from itself. Streptococcus thermophilus ST67 is a yoghurt starter — thermophilic, happiest near 42 °C, and growing in chains that complicate the CFU arithmetic above. Pediococcus pentosaceus HPP27 earns its keep in fermentation and food protection, where bacteriocin production does the work rather than colonisation. Lactobacillus crispatus HLC28 is the dominant organism of a healthy vaginal microbiota, producing L-lactic acid to hold pH near 4, and it is specified for a delivery route — vaginal or oral — that changes the excipients around the cells more than the cells themselves.

What goes on the enquiry

Seven lines make probiotic quotations comparable: the strain designation and its culture-collection deposit; per-strain CFU with the counting method; the sampling point of each figure, release or end of shelf life, and the overage behind it; water activity and moisture; the packaging — barrier material, nitrogen flush, absorber — together with the storage temperature the claim assumes; the safety file, including screening for acquired antimicrobial resistance genes, which is what a QPS-type assessment will ask for; and a retention sample held against the lot. Formats across the range are freeze-dried powder in 1 kg nitrogen-flushed foil bags inside 25 kg PE-lined drums, with frozen DVS pellets available for dairy work, and shelf life of 18 to 24 months refrigerated. None of this makes probiotics hard to buy; it makes them hard to buy on a species name and a large number.