Blog

A starter culture does more than lower pH. This review sets out how lactic acid bacteria turn milk sugars, proteins and fats into acid, texture and flavour, and how the underlying genetics affect those functions. For cheesemakers, its value is in explaining why cultures with similar species lists can behave differently in the vat and during ripening.

The question and the scope

Thomas Bintsis brings together research on the metabolism and genetics of the main groups of lactic acid bacteria used in fermented foods. This is a 2018 review, not a new cheese trial. The useful question is how a culture’s metabolic pathways and genetic traits translate into acidification, flavour development, texture and consistency. The discussion extends beyond dairy; the points below concentrate on cheese.

1. Acidification is the first job, but not the only one

Homolactic fermentation directs most of the sugar carbon towards lactic acid. Heterolactic fermentation also produces carbon dioxide and ethanol or acetate. That distinction matters when selecting for a close texture, gas formation or a particular flavour profile. The pathway used depends on the organism, the substrate and the conditions; a culture’s behaviour cannot be reduced to a single species label.

Pyruvate sits at a major branch point. Carbon passing through it can end up as lactate or be diverted towards compounds including diacetyl, acetoin, acetaldehyde and acetate. The balance changes with the strain and its environment. An acceptable final pH therefore tells us relatively little on its own about the aroma the culture will produce.

2. Citrate metabolism links buttery flavour and gas

Citrate-positive lactococci and Leuconostoc can use citrate and contribute diacetyl and related aroma compounds, as well as carbon dioxide. Citrate transport is a critical part of that ability; in citrate-positive Lactococcus lactis, it is associated with a plasmid. The review also describes the influence of pH and the availability of sugars on the products formed.

For cheesemakers, this helps explain why an aromatic culture can affect both flavour and openness. A culture selected for buttery notes may also bring gas production. The result still depends on its partners and the make conditions; adding a citrate user does not guarantee a fixed diacetyl level or a particular eye pattern.

3. Protein breakdown is central to ripening

Milk contains relatively little free amino acid. Many LAB cannot make all the amino acids they need, so their ability to break down casein and use peptides is important to growth as well as ripening. Proteinases and peptidases produce progressively smaller peptides and amino acids, which then provide substrates for further flavour reactions.

The practical distinction is between producing acid promptly and developing the desired flavour over time. A useful adjunct may contribute to the latter without being the main vat acidifier. Non-starter LAB, including organisms described in the paper as L. paracasei, L. plantarum, L. rhamnosus and L. casei, can become important during maturation. Their contribution is strain dependent; their presence alone does not establish a beneficial result.

4. LAB have a limited direct role in fat breakdown

The review describes LAB lipolysis as limited, while noting esterase activities and the importance of fatty acids as flavour compounds and precursors. Fat breakdown can supply compounds that contribute directly to flavour or feed subsequent reactions producing esters and other volatiles. A claim that a starter “develops flavour” should therefore be read more specifically: which pathway, which strain and under what ripening conditions?

5. Important dairy traits can sit on plasmids

In lactococci, plasmids can carry functions involved in lactose use, casein breakdown, bacteriophage resistance, bacteriocin production and exopolysaccharide production. Plasmids are DNA elements separate from the main chromosome. Their exchange and rearrangement can support adaptation, but can also contribute to instability in useful functions.

This is one reason strain identity and culture maintenance matter. Two cultures carrying the same species name need not have the same acidification performance, proteolytic activity or phage resistance. The review gives bacteriophage resistance and metabolic stability as explicit selection criteria for lactococcal starters. A species list is a starting point, not a complete performance specification.

6. The main culture groups do different work

Group	Relevance to cheese
Lactococci	Mesophilic acidification and flavour development; several important dairy functions may be plasmid encoded.
Streptococcus thermophilus	A dairy-adapted thermophilic starter. The review discusses sugar and protein metabolism, acetaldehyde and polysaccharide production. Its plasmids generally play a smaller functional role than those of lactococci.
Lactobacilli	A diverse group containing starter and non-starter organisms. The paper covers thermophilic dairy species such as L. helveticus and L. delbrueckii, as well as organisms associated with maturation. Do not treat all lactobacilli as thermophilic.
Pediococci	P. pentosaceus can form part of the non-starter flora of ripening cheese. Sugar use and bacteriocin production can be linked to plasmids.
Leuconostocs	Heterolactic metabolism and citrate use can contribute carbon dioxide, diacetyl and acetoin.

7. Texture and protective functions need separate consideration

Some LAB produce exopolysaccharides, which can affect the physical properties of fermented products. Others produce bacteriocins or other inhibitory metabolites. These are distinct culture functions, not automatic properties of every strain within a species. The review’s discussion of enterococci also makes clear that occurrence in a traditional fermentation is not sufficient grounds for starter selection.

8. Genetics explains possibilities, not a ready-made make schedule

The review describes metabolic engineering that redirects pyruvate away from lactate and towards compounds such as diacetyl. These examples show how changing a particular genetic function can alter the fermentation output. They do not show that ordinary commercial cultures have been engineered in this way, or supply a dosing instruction for a home make.

Genome information helps identify potential functions, but actual performance still needs testing in milk and cheese. For my purposes, I would compare cultures against four separate requirements: the acidification curve through the make, the intended flavour contribution, the effect on texture and gas, and reliability across repeated batches. A promising genetic trait is useful evidence; it is not a substitute for those measurements.

Limits and reading notes

This is a broad review of the literature available in 2018. It does not provide a controlled comparison of current retail starter blends, universal pH targets or culture dose conversions. Organism names are retained as printed in the paper; several names and classifications have changed since publication. Its strongest use is as a framework for asking better questions about a culture, rather than as a catalogue of guaranteed effects.

Source and free full text

Bintsis, T. (2018). Lactic acid bacteria as starter cultures: An update in their metabolism and genetics. AIMS Microbiology, 4(4), 665–684. DOI and publisher record.

Open access — free to read. Read the complete article on PubMed Central. The article is distributed under CC BY 4.0. No payment or ResearchGate account is needed for this copy.

The useful question here is whether native cultures can give a pasteurised goat’s milk cheese more character without losing control of acidification. Taboada and colleagues compared a native starter, the same starter organisms with native adjuncts, and a commercial Streptococcus thermophilus culture. The native combinations acidified effectively and produced cheeses rated more favourably by a trained panel after 30 days.

Cultures and the comparison

The native starter (S) contained Lactobacillus rhamnosus UNSE308 and L. delbrueckii subsp. bulgaricus UNSE309 at 30% each, with S. thermophilus UNSE314 and UNSE321 at 20% each. These are the names used in the paper.

The adjunct combination (A) retained the two lactobacilli at 30% each, reduced the two streptococci to 18% and 17%, and included four adjuncts at 1.25% each: L. plantarum UNSE316 and UNSE317, and Pediococcus pentosaceus UNSE22 and UNSE253. The comparison culture (C) was commercial S. thermophilus.

All cultures were activated and propagated before use. The native combinations were added to the cheese milk at 2% by volume; the commercial culture at 1%. These are active liquid culture additions, not equivalent doses of freeze-dried DVI powder.

Acidification curves

Figures 1A and 1B below compare 1%, 1.5% and 2% active inocula. The tests used pasteurised goat’s milk held at 38°C, with calcium chloride and chymosin, and pH readings every 30 minutes. Both native combinations reached pH 5.2 in four hours at 2%. The plotted 1.5% and 1% additions reached that point at five and six hours respectively. The authors report 4.5 hours for the commercial culture at 2%; that comparison was not plotted.

For cheesemakers: the inoculation rate changes the time available before the same pH is reached. The adjunct blend did not prevent rapid acidification at the higher addition. These are controlled milk acidification tests, not a measured pH curve through cutting, draining, pressing and brining. In particular, pH 5.2 is the test endpoint, not a reported cutting target for the make.

Make conditions

The cheeses were made in 35 L batches from milk pasteurised at 65°C for 30 minutes and cooled to 38°C. Coagulation lasted 40 minutes, followed by cutting to corn-sized curd, moulding and two hours of pressing. The cheeses were held at 5°C for 12 hours, brined in 16% w/v brine at 5°C for two hours, then ripened at 12°C and 85% relative humidity for 30 days. These conditions matter when comparing the results with a different make.

What happened during ripening

Cheese	pH, day 1	pH, day 30	Salt, day 30
Native starter (CS)	5.22	4.92	1.29%
Starter and adjuncts (CA)	5.20	4.88	1.27%
Commercial culture (CC)	5.24	5.05	1.25%

These are the reported mean pH measurements at two stages, not a continuous ripening curve. The native cheeses finished at a lower pH than the commercial comparison. Moisture was close to 49.7% in all three at day 30.

Proteolysis was slight. Non-protein nitrogen reached 4.54% and 4.49% of total nitrogen in CS and CA, compared with 4.18% in CC. That measure describes the soluble breakdown products of protein; it is not a direct percentage of free amino acids. Total free fatty acids were approximately 1.51, 1.53 and 1.47 g/kg respectively. The differences point to somewhat greater biochemical development with the native cultures, rather than a large change in overall cheese composition.

The trained panel rated the native cheeses “good” and the commercial comparison “regular”. The adjunct cheese had the highest creaminess and spiciness scores and the lowest bitterness. Its aroma also showed stronger fruity, propionic and goat’s cheese notes. For a cheesemaker, this is the more useful result: adjunct selection can change the character of the cheese even where the basic composition remains similar.

How far to take it

There were three independent makes per treatment and nine trained panellists. This supports a useful culture comparison under the stated conditions, not a general consumer preference claim or a transferable recipe for any commercial culture carrying the same species names. Strain identity, culture activity, milk and temperature still matter. I would use the curves to frame a trial and record pH through my own make, rather than lift the four-hour endpoint as a universal schedule.

Source and access

Taboada, N., Medina, R., Núñez, M. and López Alzogaray, S. (2017). Biochemical and microbiological changes throughout the ripening of Argentinean fresh goat’s milk cheeses made with native cultures. Journal of Microbiology, Biotechnology and Food Sciences, 6(5), 1174–1180. DOI and publisher record.

Open access: the article is licensed under CC BY 4.0. Open the full paper (PDF). Figures 1A and 1B are reproduced above from Taboada et al. (2017); the image is cropped from page 1177, with no changes to the plotted data.

This is the latest one I was able to find on the web. Hopefully not too much has changed with DVI cultures since.

The book contains pretty rich information on the nature of the bacteria, flavour impact and typical styles of use. There is considerable overlap with the Culture Characteristics table which draws on this document as one of its key sources.