This article discusses the major functions of starters in dairy fermentations. Recent research on the relative importance of the antimicrobial agents produced by starters is included. The importance of undissociated lactic acid (HLac) is discussed with regard to the inhibition of the growth of Listeria monocytogenes, Escherichia coli and Staphylococcus aureus.

The author recommends that regulators should require manufacturers of raw milk cheeses to meet a minimum value for HLac that must be achieved prior to product release for retail sale.

The inclusion of limits for HLac in HACCP plans would provide additional safety for consumers of cheese and in particular raw milk cheeses. An interactive calculator is also provided for the calculation of HLac in cheese. The calculator provides advice on whether the value calculated is sufficient to inhibit the growth of Listeria monocytogenes  in cheese and uses an amended value for the pKa of lactic acid that is relevant to the ionic environment in cheese. Further information on starter cultures is available in the section on starters.

It is concluded that a high concentration of HLac in a cheese does not guarantee freedom from pathogens, however, it will normally prevent their further growth and under ideal conditions will contribute to their decline in cheeses that have not been mould-ripened.

Major starter functions

The major functions of starters in dairy fermentations are shown in Table 1.

 Table 1.  Major functions of starters in milk fermentations 

Acid production

        Gel formation

        Whey expulsion (syneresis)


        Flavour development

Flavour compound production

        Formation of diacetyl and           acetaldehyde


        Lowering of pH and redox potential

        Production antibiotic substances e.g.         bacteriocins such as nisin

        Production of hydrogen peroxide that is inhibitory per se and through activation of the LP system

        Formation of inhibitory compounds e.g.  D-leucine, diacetyl and reuterin

        Production of lactic acid / lactate

  Production of acetic acid / acetate  

Out competition of other organisms for nutrients

Gas formation

        Eyehole formation

        Production of openness to facilitate           'blue veining'

Stabiliser formation

        Body and viscosity improvement

        Increase cheese yield?

        Result in reduced use of milk powder           in yoghurt manufacture

Lactose utilisation

        Reduce potential for gas and off-flavour         development

        Make products more acceptable to the         lactose intolerant

Lowering of redox potential


         Aids flavour development

Although lactic acid production is the major task required of starters in many milk fermentations, starters have other - sometimes equally important - roles.

The preservative properties of starters have been used to extend the shelf life of products e.g. Cottage cheese and to inhibit the growth of psychrotrophic micro-organisms in silo milk in cheese factories.  The mechanism for these effects involves competition for nutrients, decrease in pH, the formation of inhibitory compounds, lowering of redox potential and possibly activation of the lactoperoxidase system and is discussed further in the section on the relative importance and effectiveness of the antimicrobial mechanisms of starters. 

The production of openness e.g. slits or seams, is important in some varieties of blue-veined cheese such as Cambazola or Blue Brie and can be achieved by the use of starters containing strains of Lc. lactis subsp. lactis biovar. diacetylactis and/or Leuconstoc species.

Some starters produce complex polysaccharide-type materials which can be utilised by technologists to improve the body or viscosity of products such as yoghurt, increase cheese yield or to produce the essential product characteristics (ropiness) of Finnish long milk - Viili.

LAB starters, particularly strong acid producers, have the potential to significantly lower the oxidation–reduction (Eh) potential in dairy products, particularly in cheese or in liquid products in sealed systems. The Eh of raw milk is about +150 mv, whereas Cheddar cheese has an Eh of around -250 mv. Reduction of Eh to such low values creates an environment in which only facultative or obligate anaerobic microorganisms can grow. Low redoxpotential is also an essential requirement for the formation of the typical flavour of Gouda and Cheddar cheeses.

Cheddar cheese at the start of pressing contains about 0.5% (w/w) lactose.  Failure of starters to utilise residual lactose (due to phage-induced cell lysis) during pressing and in the first few weeks of maturation may result in sweet cheese and/or gas production by heterofermentive non-starter bacteria.

Because acid production by the starter influences syneresis, variations in starter activity may result in variations in the moisture content of the final cheese. Obviously these effects have the potential to affect the profitability of dairy plants.

In cheesemaking, lysis of starter cells can result in problems which range from slow acid production to completely lost vats.  Lysis of starters during the late stages of cheesemaking can sometimes be missed by production staff and may result in cheese of high pH, high lactose content, high redox potential and low lactic acid and lactate content.

Disruption of starter functions can have a range of consequences. These range from major product quality problems, including the growth of pathogens e.g. Staphylococcus aureus (especially if raw milk has been used), to lost product.

The relative importance and effectiveness of the antimicrobial mechanisms of starters

It has been known for decades that starter bacteria can inhibit the growth of many spoilage and pathogenic bacteria. In some instances it has been demonstrated that lactic acid bacteria may also kill pathogens.

Reuterin, 3-hydroxypropionaldehyde, produced by Lactobacillus reuteri has potent bactericidal activity against E. coli 01575:H7, Listeria monocytogenes and many Gram-negative bacteria. The bactericidal activity of reuterin is enhanced in the presence of the LP system (Arqués et al., 2008). While L. reuteri is not a normal component of defined starters it is present in some artisanal starters. Reuterin is a byproduct of glycerol metabolism and to obtain significant concentrations of this metabolite supplementation of milk with glycerol is required.

While particular strains may demonstrate potent bactericidal activity against particular pathogens there is general acceptance that starter cultures can inhibit the growth, rather than kill, many pathogens and spoilage bacteria. 

An interesting report by Charlie Daly at Oregon State University (Daly et al., 1972) provided an early indication of the antimicrobial spectrum of a lactic culture, the effectiveness of its antimicrobial effects and the mechanisms involved.

In milk the inhibitory effects on growth ranged from 70 to 99.9% after 24 hours incubation (Table 2). Note the inhibitory effects were often less in milk compared with broth media.

Table 2. Organisms inhibited by Streptococcus diacetilactis
  % Inhibition at 24 hours.
Organism Milk Broth
Pseudomonas fluorescens 99.9 99.99
P. fragi 99.9 99.99
P. viscosa 99.9 99.99
P. aeruginosa 90 99.99
Alcaligenes metalcaligenes 99.9 99.99
A. viscosa 90 99.99
Escherichia coli 80.0. 99.99
Serratia marcescens 90 99.9
Salmonella seftenberg 99.99
S. tennesse 70 99.99
Staphylococcus aureus 99.9 99.99 

Modified from Daly et al. (1972)

Daly et al. (1972) investigated the nature of the antimicrobial activity and found evidence for several inhibitory factors. Convincing evidence that HLac was responsible for the larger part of the inhibitory effect was presented.

The greater antimicrobial activity of undissociated organic acids, including lactic, acetic and propionic acids, compared with the fully ionised form has been known for many years. Typically the undissociated acid can have several to hundreds of times the antimicrobial effect of the fully dissociated (ionised) form.  

Values for the concentration of HLac required to inhibit the growth of some non pathogenic Listeria species and the non-pathogenic E.coli M23 have been known for many years (e.g. Presser et al., 1997) e.g. Table 3.

Table 3. Experimentally determined values for MICs of undissociated lactic acid
Organism MIC*
Escherichia coli M23  8.32
Yersinia enterocolitica  5-10
Listeria innocua **  4.9
*   Minimum inhibitory concentration of undissociated lactic acid (mM)
** Determined using sodium lactate
Modified from Presser et al. (1997)

However, this data is more indicative than definitive for cheese since the experimental conditions used were not representative of the cheese environment. Salt, sodium chloride, increases the resistance of most bacteria to acid conditions and this was not included in this series of experiments. Because of the protective effects of salt, it is likely that the MIC values for HLac are higher than those stated in Table 3 for E. coli and listeria in cheese.

Later work by Presser et al. (1998) using broth media containing sodium chloride concentrations relevant to cheese established that a concentration of 11 mM undissociated lactic acid was required to inhibit the growth of E. coli M23.

Most pathogens of relevance to cheese and milk products cease growth when the HLac concentration is somewhere in the range 5 to 15 mM e.g. Charlier et al. (2008) showed the total growth inhibition of lactic acid adapted S. aureus in skim milk medium corresponded to 6–12 mM HLac. This range applies when other growth factors are optimal. Note that growth may not occur at lower concentrations of acid if other preservative factors are operating.

 Wemmenhove et al. (2017) evaluated factors relevant to Gouda for their potential to inhibit growth of Listeria monocytogenes in Gouda cheese. Factors included water activity, pH, undissociated acetic and lactic acid, diacetyl, free fatty acids, lactoferrin, nitrate, nitrite and nisin. Of these factors pH, undissociated acetic and lactic acid, diacetyl and nisin are directly associated with starter growth.

This work revealed that while other inhibitors were important, HLac is the most important factor for growth inhibition of L. monocytogenes in Gouda cheese.  Growth of L. monocytogenes was inhibited when the HLac concentration was >6.35 mM.  Note in this research the starter culture used did not produce significant concentrations of nisin.

This information has direct relevance in the quality assurance of cheese production. These workers also calculated a more accurate value for the dissociation constant (pK a) of lactic acid and this is discussed below.

 Determination of the concentration of undissociated lactic acid in cheese

In the cheese moisture, weak organic acids such as lactic or acetic exist as a mixture of the dissociated (ionized) and the undissociated molecular species. The proportions of the various molecular species are dependent on the pH of the cheese.

Once the total lactic acid concentration has been determined, there are several methods for measuring total lactic acid, the Henderson–Hasselbalch equation (Equation 1, the brackets indicate the molar concentration of the acid.) can be used to calculate the relative proportions of undissociated and dissociated forms of lactic acids at particular pH values using the pKa of the acid.

The pKa of an acid is the negative log of its dissociation constant. In practice this is the pH at which the acid is 50% dissociated into its constituents. 

Equation 1.   [dissociated acid]  = 10 pH–pKa
                   [undissociated acid]

Solving for undissociated acid gives equation 2.

Equation 2. [undissociated acid] = [dissociated acid]
                                                 (1 + 10 pH–pKa)                 

Until recently the pKa of lactic acid was accepted as 3.86 (Dawson,1986).  Wemmenhove et al. (2017) have determined that the actual value in Gouda cheese is 3.71.  Note it is likely that this lower value probably reflects the true value in most normal fat and medium moisture cheeses. The lower value will also give lower values for HLac in calculations compared with those using the previously accepted value.

Assuming a cheese with a total lactic acid concentration of 1.1%, a pH of 5.25 and a cheese moisture of 37% calculate the concentration of HLac.

This can be done as follows:

Calculate the concentration of lactic acid (M) in the cheese moisture.

1000/37 x 1.1 
molecular weight of lactic acid                      


=0.33 M

The concentration of HLac (M) is calculated using equation 2

1+10 5.25-3.71

= 0.33          

= 0.00925   M

Converting to millimolar concentration

= 0.00925  x 1000

= 9.25 mM HLac

This calculation would suggest that L. monocytogenes  would not be expected to grow in this cheese.

A free calculator for undertaking or checking calculations has been provided. 

Note that this value assumes that all the lactic acid (dissociated and undissociated) is present only in the water phase. The situation in reality is more complicated. Undissociated lactic acid would also be expected to be soluble in milk fat. However, since milk fat is protected by the milk globule membrane and a significant proportion of fat will be solid at normal temperatures (around 20% at 20 °C) partition in fat is likely to be low. Nevertheless the calculations described here may slightly over estimate the HLac available for antimicrobial activity.

Access calculator for determining undissociated lactic acid

Are cheeses with high concentrations of undissociated lactic acid free from pathogens?

Pathogens such as E. coli and L. monocytogenes, if present in cheese milk, will normally grow during cheese manufacture and their initial numbers will typically be multiplied by 10 to >10,000 depending on various factors including the strain, the starter used, rate of acidification, and temperatures used in manufacture. There is no evidence that this growth can easily be prevented in cheese, containing normal concentrations of HLac, if these organisms are present in the cheese milk or have been introduced as contaminants during manufacture.

Providing the cheese has been manufactured properly, late phage-induced starter lysis has not occurred and an appropriate concentration of HLac is present further pathogen growth would not be expected. While it may be expected that the numbers of pathogens will decline significantly with storage time, particularly at elevated storage temperatures,  it may take many weeks or months for their numbers to decline to their initial concentration in cheese milk. Some strains of E. coli  can survive in cheese e.g. Cheddar for many months.

To ensure safety it is paramount that milk for cheese manufacture contains no, or only low concentrations, of pathogens and that pathogens are not introduced during manufacture.

The situation with mould-ripened cheeses, manufactured using milk contaminated with L. monocytogenes is particularly noteworthy. 

The pH in mould ripened cheeses rises during storage and can reach pH 6.5 in some weeks. As the pH increases, the concentration of HLac decreases (Table 4).


Table 4. Effect of ripening time on the pH and concentration of  undissociated lactic acid in a mould ripened cheese that initially contained a total lactic acid concentration of 1% (w/w).
Time, days pH Concentration of undissociated lactic acid (mM)*
0 5.0  12
7 5.1  9.7
20 5.5  3.9
30 5.8  2
50 5.9  1.6
60 6.1  1
70 6.8  0.2
Cheese had a total lactic acid concentration of 1% (w/w) at manufacture and a moisture content of 45%.
*The concentration of undissociated lactic acid was determined using the calculator on this website.


In the case illustrated in Table 4, the concentration of HLac has dropped to 3.9 mM after 20 days ripening. In the absence of other antimicrobial agents it is likely that any listeria present will commence growth. Since listeria can grow well at the temperatures used in cheese ripening, distribution and storage this high pH, low undissociated acid concentration environment, that is developing with storage may create an ideal environment for the growth of listeria to high cell densities with the high probability of food poisoning and its consequent risks including deaths of consumers.

It should be clear from the above that a high concentration of HLac in a cheese does not guarantee freedom from pathogens, however, it will normally prevent their further growth and under ideal conditions will contribute to their decline in cheeses that have not been mould-ripened.


That regulators should require manufacturers of raw milk cheeses to meet a minimum value for HLac that must be achieved prior to product release for retail sale.

Literature cited

Arqués, J.L., Rodríguez, E., Nuñez, M. et al. (2008).Inactivation of Gram-negative pathogens in refrigerated milk by reuterin in combination with nisin or the lactoperoxidase system. Eur Food Res Technol. 227, 77.

Charlier, C., Even, S., Gautier, M., Le Loir, Y. (2008). Acidification is not involved in the early inhibition of Staphylococcus aureus growth by Lactococcus lactis in milk. Intern. Dairy J. 18, 197–203.

 Daly, C., Sandine, W.E. and Elliker, P. R. (1972). Interactions of food starter cultures and food-borne pathogens: Streptococcus diacetilactis versus food pathogens. Milk Food Technol. 35, 349-357.

Dawson, R. M. C. (1986). Data for biochemical research. Oxford: Clarendon Press. 

Presser, K. A., Ratkowsky, D. A. and Ross, T. (1997). Modelling the growth rate of Escherichia coli as a function of pH and lactic acid concentration. Appl. Environ. Microbiol. 63, 2355–2360.

Presser, K.A., Ross, T., and Ratkowsky, D.A.(1998). Modelling the growth limits (growth/no growth interface) of E. coli as a function of pH, lactic acid and temperature. Appl. Environ. Microbiol. 64, 1773–1779.

Wemmenhove, E., van Valenberg, H.J.F., van Hooijdonk, A.C.M.,Wells-Bennik, M.H.J., Zwietering, M.H. (2017). Factors that inhibit growth of Listeria monocytogenes in nature-ripened Gouda cheese: A major role for undissociated lactic acid. Food Control 84, 413-418.  

How to cite this article

Mullan, W.M.A. (2017) . [On-line]. Available from: . Accessed: 15 April, 2024. Based on an technical article authored in 2005. Updated September 2018; February 2022.