Head, School Of Science and Technology, CSU-Riverina
The School of Science and Technology at Charles Sturt University-Riverina is currently engaged on two research projects in collaboration with the cheese industry. Whilst I would like to report on the work in progress, one aspect of work with industry is that it is more often than not subject to confidentiality agreements. For this reason I have chosen to talk about some work we did a few years ago under an ARC grant on a topic which is at least related to our current work. The primary aim of the project was to investigate some of the structural features of lactoperoxidase. The enzyme is a haemoprotein which is present not only in milk but also in saliva and tears. Its origin in these fluids almost certainly arises from eosinophils or similar cell types which also contain the enzyme in high concentration. Its function is to act as the basis for a powerful antibacterial system. This is especially clear in relation to its role in saliva where it has a protective role with respect to the development of dental caries.
There are three components to the lactoperoxidase antibacterial system. Naturally the enzyme itself is needed; it also requires hydrogen peroxide as the primary oxidant but at quite low micromolar concentrations. It is now clear that hydrogen peroxide itself is not the antibacterial agent but an oxidation product of the thiocyanate union. Hypothiocyanate (OSCN) appears to act catalytically to oxidise nonsterically hindered sulphhydryl groups on bacterial cell walls. Hydrogen peroxide producing catalase negative bacteria such as many lactobacilli and streptococci are especially sensitive to the LPO antibacterial system.
Preparation of the enzyme meant gaining access to unpasteurised whey on a massive scale and we were especially grateful to Haberfields Dairy for giving us the whey as well as equipment to carry out the work. The procedure we used was based upon the high isolectric point of lactoperoxidase which allows it to be absorbed onto the cation exchange resin Amberlite CG50. Lactoferrin, which is another antibacterial protein in milk, is also absorbed and is a by-product of the separation procedure. A good preparation can result in as much as 5 grams of freeze dried enzyme of reasonably high purity as judged by polyacrylamide gel electrophoresis (Nichol, et al., 1987). However, we have noticed that the yield of lactoperoxidase is extremely variable, with some batches of whey yielding virtually no LPO. Our subjective impressions were that spring whey gave the best yields. In parallel to this, we were also made aware by the Dairy Industry Authority that a number of cheese producers in the area were experiencing difficulties which appeared to be seasonal in character. These problems revolved around poor development of acidity following inoculation of milk and during the cheddaring processes itself. Naturally this aroused our suspicions that variations in lactoperoxidase activity could be associated with inhibition of growth of the starter cultures.
The net result of all this was that we began a study aimed at systematically following the activity of lactoperoxidase in whole milk on a daily basis. The study extended over a period of 18 months and was carried out on milk from two farms, one located at Ladysmith, about 20 kms east of Wagga, and one at Tumut, about 100 kms east.
We used the oxidation of pyrogallol to purpurogallin as the basis of the assay. Parallel protein determinations were also carried out and results expressed as specific activity to remove any variabilty due to changing total protein concentration.
It was immediately apparent that there were major variations in the concentration of lactoperoxidase in milk relative to total protein. There are obvious peaks during spring as we suspected but there were also peaks in late autumn/early winter as well.
In an attempt to detect the determinants associated with this pattern, we analysed the data using a spectrum analyser computer software package called Prospect. This allows the spectrum to be analysed for statistically significant periodicity. It also allows cross checking of the data for correlation with other seasonal factors. We were particularly interested in rainfall and the likelihood that changes might be associated with changing feed quality. Calving patterns could also be a factor.
Variance preserving plots provide most information on statistically significant periodicity in data such as these. Such a plot is shown in Figure 1.
Major significant peaks occur in the 22-15 week period with the 15 week period being especially well defined in relation to the Ladysmith data. Interestingly, a secondary peak occurs in the three week region. This perhaps gives a lead to causes of variation. The oestrus cycle of the cow is approximately 21 days. It is well known that herds tend to cycle together and come into heat at around about the same time. This suggests that the longer period peaks, although well separated from the three week peak, may also be associated with the reproductive cycle and is probably associated with the calving pattern established on the two farms.
The other possibility is that variation is due to a fluctuating seasonal influence.
Figure 2 tests for coherence of the data at the two different locations and clearly shows that long period cycles are highly coherent at both Tumut and Ladysmith. This means that the periodicity patterns are very similar at the two sites. A phase difference test (Figure 3) shows that the Tumut cycles lead the Ladysmith cycles by about 1-2 weeks. This could suggest a rainfall or soil moisture associated pattern as the autumn break in particular tends to occur earlier as one moves east.
Although there is a weak peak with a period 15 weeks in the rainfall autospectrum, significant peaks are only seen in the high frequency region. These probably correspond to the relatively regular fronts passing through this region in the autumn/winter period and their associated rainfall.
Clearly this project is by no means complete. However, we have demonstrated long period cyclic patterns in lactoperoxidase specific activity in milk. We have also shown that very marked variations occur. These variations are much more marked than variations in total protein level. Where milk for cheese manufacture is taken from a single herd, it is possible that high LPO levels could have an inhibitory effect on peroxide producing lactic acid bacteria. This is certainly a factor of which the cheese maker needs to be aware.