Detection of residual enzyme activity for assessment of whole milk powder shelf-life stability : a thesis presented in partial fulfilment of the requirements for the degree of Master of Food Technology at Massey University, Manawatū, New Zealand

Abstract

Whole milk powder (WMP) is one of the major exported commodities of New Zealand. It is significantly important to maintain quality stability during exportation and storage. The attention to the enzymatic degradation in WMP is limited. This study aims to directly detect enzyme activity in WMP samples and evaluate its relationship with bacterial counts and nutrient content stability, thereby find out the value of direct enzyme activity in assessing WMP shelf-life stability. Sixteen WMP samples were divided into two groups classified by raw milk bacterial loading, ten high-quality samples and six poor-quality samples. Mesophilic and thermophilic bacterial counting were conducted using pouring plate method, incubating at 30 °C and 55 °C for 48 h. Protease activity was detected using azocasein (incubated at 37 ± 1 °C for 20 h) and commercial EnzChek® kit assay (incubated at 22 ± 1 °C for 20 h). Lipase activity was determined using p-nitrophenyl palmitate (p-NPP) assay (incubated at 37 ± 1 °C for 2 h). Fat and protein contents were monitored using FOSS Milkoscan for 6 7 months. Linear mixed-effects models and linear regression models were applied for statistical analysis. Mesophilic bacterial counts in high quality samples were significantly lower than those in poor-quality samples (estimated mean: 2.88 ± 0.04 log cfu/g vs. 3.16 ± 0.05 log cfu/g, p < 0.001), which was consistent with the difference of raw milk bacterial counts between these two quality groups. There was no significant difference of thermophilic bacterial counts between the high- and poor-quality groups (estimated mean: high quality 2.94 ± 0.11 log cfu/g, poor-quality 3.04 ± 0.14 log cfu/g, p = 0.576). Protease activity was not detected in all samples using azocasein assay (LOD = 11.53mU/g WMP) and was detected only in one sample that came from high quality group using EnzChek® kit assay (LOD = 3.69 mU/g WMP, LOQ = 9.86mU/g WMP). The protease activity of that sample was between the LOD and LOQ of the EnzChek® assay. Lipase activity was detected in all samples using p-NPP assay (LOD = 1356 U/g WMP, LOQ = 3002 U/g WMP). Only one sample that came from high-quality group had lipase activity between the LOD and LOQ of the assay, all other samples had quantitative lipase activities. There was no significant difference of lipase activity between high- and poor-quality groups (estimated mean: high-quality 5045 U/g WMP, poor-quality 5213 U/g WMP, p = 0.521). However, the sample that had the highest lipase activity (7894 ± 490 U/g WMP) was from high-quality group, indicating that bacterial counts might not provide reliable enzymatic deterioration in WMP during storage. The fat content maintained stable in high-quality samples over 7-month storage, poor-quality samples over 6-month storage (poor-quality samples were manufactured one month later) (p = 0.197). And there was no significant difference between those two groups (p = 0.114). Protein content also exhibited stable in high- and poor-quality groups over their respective storage time (p = 0.352) and did not significantly vary between the two groups (p = 0.236). The FOSS Milkoscan results provided evidence that the fat and protein content conformed regulatory limits over the storage period but failed to provide information about the deteriorative reactions in WMP during storage. The results in the present study suggest that direct enzyme activity may offer complementary information to conventional analysis in a comprehensive evaluation of WMP shelf-life.