739

Journal of Protection, Vol. 49, No. 9, Pages 739-753 (September 1986) Copyright® International Association of , Food and Environmental Sanitarians

Assessment of Product Quality and Potential Shelf-Life - A Review

J. R. BISHOP1 and C. H. WHITE*

Dairy Science Department, Mississippi Agricultural and Experiment Station, Mississippi State University, Mississippi State, Mississippi 39762 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 (Received for publication May 28, 1985)

ABSTRACT organisms and their metabolites, especially those that occur from post- contamination (134). Over the years, many tests and assays have been developed In determining the presence of post-pasteurization con­ to estimate the quality and potential shelf-life of dairy products. taminants, two major problems are encountered. First, These have ranged from simple, standard bacterial enumerations time is critical. Rapid results are needed so that the to more complex metabolite detections. This paper is a review sources of contamination can be found and eliminated. of the parameters that have been used to estimate, or indicate Second, the number of contaminants in a freshly pas­ the inherent quality of dairy products. teurized sample may not be sufficient to detect contami­ nation. The problem of detecting post-pasteurization con­ Milk has been shown to be ideal for the growth of tamination at a very early stage is serious. There may due to its water, protein, carbohydrate, never be a perfect test for a specific plant, but any im­ and content, all of which are important provement would be helpful, especially if results can be to many forms of (62). In the days before ade­ obtained sooner than with some of the presently used quate refrigeration, bacterial spoilage ran rampant, and methods (19). economic losses were staggering (5). Today, farm bulk The major goal of any test or assay used to assess the tanks, every-other-day pick-up at farms, 5-d and 4-d-a- quality of a perishable dairy product must be to provide week plant operations, discontinuing of home delivery, reliable and accurate results within a test period (less than and purchasing of milk only on shopping days have in­ 48 h) that would allow for effective corrective measures. creased the age of milk before consumption (149). These measures may mean removal of product, selling These methods and a more critical consuming public quicker or adjusting the next batch. As a result of this have caused the food industries to re-evaluate their qual­ need, the objective of this review is to compare methods ity standards. Quality means different things to different that have been used, and are currently being used to as­ people and standards are ever changing. It has been sess dairy product quality. pointed out that concepts of milk quality depend on many factors: (a) educational background of the individual, (b) economic and social status of the individual, (c) scientific KEEPING QUALITY information available, and (d) the ordinances, laws, regu­ The period between processing/packaging and when lations, etc., under which the dairy operates milk becomes unacceptable to consumers is referred to (62). The milk industry has been a leader in establishing as the "shelf-life" or "consumable life" of milk. Al­ programs to give the consumer a product that is pure, though it reflects the "keeping quality" of milk, Baker of good flavor, of attractive appearance and of desirable (7) states that there is no adequate working definition of keeping quality. These programs emphasize rigorous lab­ shelf-life. Consumers of dairy products must have some oratory examination of milk and dairy products to ensure indication of the expected or potential shelf-life of the that this quality is maintained (62). products they buy. Therefore, to allow consumers to as­ It is a frequently heard axiom that production of such sess the age of products at the time of purchase, a date superior quality dairy products possessing extended shelf- is placed on the container that indicates either the date life requires high quality milk. Milk and other ingredients of packaging or the last date that the product may be used should be free of off-odors and -flavors, abnormal sold or offered for sale. The consumer then expects that chemical and physical properties, and undesirable micro- a product purchased at any time up to that date is of acceptable quality. Also, if properly treated, it should re­ 'Present address: Department of Food Science and Technology, Vir­ ginia Polytechnic Institute and State University, Blacksburg, VA. main acceptable beyond the last date of sale (78).

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 740 BISHOP AND WHITE

On a daily basis, the consumer uses taste and smell Gillis and Custer (66), who found that repasteurization to judge how long milk, kept under refrigeration main­ did not significantly extend the shelf-life beyond that of tains an acceptable flavor. This, in effect, is the con­ the same milk that was not repasteurized. In addition, sumer's own test of keeping quality (75); therefore, Mistry and Kosikowski (138) reported that sorbate levels flavor and keeping quality are of paramount importance of .15 and .20% in milk at 7°C effectively prolonged to maintaining consumption of fluid milk (8). The most quality and reduced the rate of psychrotrophic bacteria common descriptions given to unacceptable milk flavors growth, 19-20 d vs. 12-14 d, respectively. by consumers are staleness, putrid, and/or curdled, bitter There exist opinions that the consumer may exercise and fruity (7). the most efficient indicator of keeping quality - flavor When dairy products leave the plant and are displayed (7,8,75,76). These researchers report that the flavor score in supermarket showcases and, finally, reach the home of freshly pasteurized milk, although subjective, is much refrigerator, low temperature refrigeration is the exception better than standard microbiological and chemical tests as rather than the rule for each of these (149). Temperatures a predictor of keeping quality. Also, several other re­ below 7°C are one of the key factors to extending shelf- searchers (13,27,68,75,208) have concluded that the Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 life (10). It has been generally accepted that an increase keeping quality of milk is generally unrelated to standard in storage temperature of 5°F will decrease the shelf-life microbiological and chemical tests. by approximately one-half (94). Hankin et al. (79) con­ Differences of opinion concerning efficient methods of cluded that keeping quality at any storage temperature shelf-life estimation are evident. Patel and Blankenagel was unrelated to the manufacturer's code date (last day (151) state that there is no doubt that in commercially product is to be sold). There was a significant correlation pasteurized and packaged milk, post-pasteurization con­ between keeping quality at 10°C and the other two stor­ tamination is by far the most common cause of flavor age temperatures of 1.7 and 5°C, suggesting a practical defects of microbial origin. However, they found that test to measure keeping quality at the lower temperatures. even in the absence of post-pasteurization contaminants, Commercially processed milk stored at 0°C averaged off-flavors may be encountered if the contained 31 d of shelf-life but only 6 d at 7°C, with summer milk large populations of psychrotrophs. Genera of bacteria possessing twice the shelf-life of winter milk, as reported most frequently involved in keeping quality problems in by Finely et al. (55). Grosskopf and Harper (69) found pasteurized milk include Pseudomonas, Achromobacter, that aseptically packaged milk stored at 4°C had a shelf Chromobacterium, Alcaligenes, Proteus, Escherichia and -life of 28 d. Hankin et al. (79) reported that keeping Enterobacter (49). The gelatinous and flavor defects quality of commercially stored at 1.7, of cottage produced by Pseudomonas viscosa, 5.6 and 10°C remained organoleptically acceptable, on Pseudomonas fragi and Alcaligenes metalcaligenes are the average 17.5, 12.1 and 6.9 d, respectively. Harmon well known. In milk as well as cottage cheese, P. viscosa et al. (82) found that the shelf-life of commercial samples produces bitter, rancid and unclean flavors while that pro­ of milk was 21 to 63 d when stored at .5°C, 7 to 17 d duced by P. fragi is a fairly common fruity defect (49). at 6°C, 3 to 7 d at 12°C, and only 1 d at 20°C. They Elliker (50) found the major source of post-pasteurization found keeping quality to be directly related to rate of de­ contamination by bacteria to be the filler. velopment of bacterial populations. They added that the The occurrence of off-flavors in raw milk, market maximum number of microorganisms ultimately attained milk, laboratory pasteurized milk and milk inoculated at spoilage was the same regardless of storage tempera­ with Pseudomonas fluorescens coincided with psychrot­ ture. Janzen et al. (94) noted a whole milk shelf-life of rophic counts of 107/ml. It was apparent that deterioration 9 d at 7°C. Langlois et al. (110) indicated an average of pasteurized milk during storage is of microbiological shelf-life of skim milk at 4.5°C was 13.3 d. Randolph origin, but may not be due to the growth of bacteria per et al. (160) found that the average keeping quality of se (7). Matoba et al. (126) reported that the occurrence market milk purchased at outlets was 7.7 d at 7°C, of a bitter flavor in milk appears to be related to the and 6.2 d at 10°C. The samples were from 0 to 14 d presence of a heat-stable protease. Hydrolysis of old when purchased. Most of the samples possessed psyc- and lactalbumin yields bitter peptides. Therefore, keeping hrophilic (now known as psychrotrophic) type of defects quality depends both on numbers of bacteria and their at the time the flavor quality was judged unacceptable. biochemical activities (121). This fact is illustrated by the Sherman et al. (181) reported that pasteurized milk had research of Patel and Blankenagel (151) where milk with a shelf-life of 8 to 12 weeks at 0°C and spoilage marked counts of 106/ml before heating frequently developed ob­ by extensive proteolysis, was of bacterial origin, predo­ jectionable flavors after pasteurization and subsequent minantly gram-negative pseudomonads. Shipe et al. (182) storage. The most common defect was a bitter flavor reported that the average keeping quality of commercial which developed in spite of small numbers of organisms milk at 7°C was 13 d. in the pasteurized products and in the absence of post- Shelf-life was affected by days held raw (95), pasteuri­ pasteurization contamination. zation temperature, storage temperature and season as de­ BACTERIAL ENUMERATION termined by taste panel and bacteriological tests (55). Other considerations regarding shelf-life were reported by Microbiological tests done on milk have two main pur-

JOURNAL OF FOOD PROTECTION. VOL. 49, SEPTEMBER 1986 ASSESSING SHELF-LIFE AND QUALITY OF DAIRY 741 poses: (a) to determine whether sanitary practices are ad­ tor which is measured spectrophotometrically and related equate, and (b) to gain information that will aid in pre­ to pyruvate concentration. This method relies essentially dicting keeping quality (727). The Standard Plate Count on two factors. First, that pyruvate production by all bac­ (724) is a widely used method for determining the bacte­ teria in milk is in some way equal. This is not true be­ rial count of raw and pasteurized milk for regulatory pur­ cause some organisms do not produce pyruvate while poses. The method was originally described in the first others produce various amounts depending on their phys­ edition of Standard Methods, but has undergone several iological state. Second, the method uses the colony count modifications since then (766). The basic underlying as­ procedure as its reference point (148). Ledford et al. sumption of the Standard Plate Count - as well as other (113) reported a correlation between pyruvate and Stan­ plate count procedures - is that a single viable cell, when dard Plate Count of .72, while Marshall and Harmon placed at an appropriate temperature, will multiply to the (727) found no association between Standard Plate Count point where a visible colony is produced (790). However, and pyruvate concentration. in some instances colonies arise from cells which were Adenosine-5'-triphosphate (ATP) detection has been originally aggregated in pairs, small chains, or small greatly simplified by the discovery of the luciferin- Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 clusters and cannot be differentiated from colonies which luciferase bioluminescent reaction, which contains two arise from single cells (98,124). Therefore, the basic as­ key properties that render it applicable to the quantitative sumption of plate count procedures is to some degree in­ measure of ATP. First, the amount of light produced dur­ valid (772). ing the reaction is directly proportional to the concentra­ Even though the Standard Plate Count has historically tion of luciferin and ATP in the reaction mixture (130). been the main cultural procedure used to determine viable Second, ATP is the primary high-energy nucleotide that bacterial populations in dairy products (724), literature participates in the bioluminescent reaction (772). suggests that the scientific community is not entirely The luciferin-luciferase assay has previously been used satisfied with this method of microbial enumeration. to measure viable organisms in many diverse environ­ Aside from being inaccurate, the Standard Plate Count ments including water (777), soils (47,97), rumen fluids has been criticized as being time consuming and expen­ (51,52,60,220) and foods (24,180). Several dairy applica­ sive (12,64,65,98,154), not indicative of raw milk quality tions of the luciferin-luciferase assay have been reported, (12,43,84,85,101,102,111,162,170,215,216), and slow including estimation of somatic cell and bacterial contents as a means of generating bacterial population estimates of raw milk (27). Correlation of the ATP content of 48 (65,67). Regardless of the drawbacks or problems as­ tank truck samples and Standard Plate Count was .93 sociated with the Standard Plate Count, it is officially (27). Correlation between the ATP content of 209 farm considered as a suitable method for measuring bacterial bulk tank samples and Standard Plate Count was also .93 populations in most types of dairy products (724). It is (772). Measurement of milk sample ATP content was not a suitable method for detection of post-pasteurization considered to be an acceptable method of detecting high contamination or indication of keeping quality as it does bacterial count milk (21,172). not differentiate between those organisms that survived Another measure of bacterial content was suggested by the heat treatment and those that gained entry into the O'Toole (148) who stated that cell numbers are no more product after pasteurization. (79). than a crude measure of cell mass. The dry mass of The coliform count (724) is also used as an indication bacterial cells in a sample is probably the best measure of contamination of dairy products. If coliform bacteria of the bacterial content, if its determination could be are present in adequately pasteurized products, one can be achieved. sure post-pasteurization contamination has occurred. PSYCHROTROPHIC BACTERIA However, the absence of coliforms is no guarantee that When considering the shelf-life of refrigerated milk the product is free of contaminants. This method is lim­ and milk products, the concern is almost exclusively with ited because only a relatively small group of all possible those microorganisms which grow rapidly at storage tem­ contaminants are coliforms (79). peratures. These bacteria were referred to as "psyc- Although all contaminants are important, the main con­ hrophiles", a term which means "cold loving". This is cern of the dairy processor is the group of organisms that a misnomer as they are not "cold-loving" but rather can grow at refrigeration temperatures - the psychrot- "cold-enduring"; while they are able to grow below rophs. Zall et al. (223) found significant correlation 10°C, they grow much better at ambient or higher tem­ (r=.878) between Standard Plate Count and psychrot- peratures. A more appropriate name is "psychrotrophs" rophic bacteria count values. - "cold enduring" (100). These psychrotrophs are de­ Determination of the pyruvate content of milk is a fined in essentially four major ways based on (a) op­ means of measuring bacteriological quality. One advan­ timum growth temperature, (b) ability to grow at low tage of this approach is that the analysis can be auto­ temperatures, (c) method of enumeration and (d) criteria mated to test 80-120 samples/h. The analysis involves the which are independent of the incubation temperature breakdown of pyruvate catalyzed by lactic dehydrogenase (279). in the presence of reduced nicotinamide-adenine dinuc- Psychrotrophic microflora consist of microorganisms leotide (NADH), and it is the concentration of this cofac- that can grow relatively rapidly at commercial refrigera-

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 742 BISHOP AND WHITE tion temperatures. This does not imply that the optimum negative bacteria (43,187). Oehlrich and McKellar (144) temperature of growth of this group of organisms lies in evaluated an 18°C/45 h plate count technique for enumer­ the temperature range encountered in normal commercial ation of psychrotrophs with correlation values to the 7°C, refrigeration of foods. Optimum growth temperature of 10 d method of .866 in raw milk and .936 in pasteurized many psychrotrophs is in the mesophilic range (20-45°C). milk. The term psychrotrophic was coined to accommodate the Many psychrotrophs produce lipases which can degrade heterogeneous group that did not fall within the strict milkfat and cause flavor problems in dairy products (59). physiological definition of the psychrophilic group. Cur­ Common thermal processing may leave heat-stable lipases rently, the term psychrophilic is reserved for microor­ almost intact (3). The major producer of heat-stable ganisms that optimally grow at temperatures below 10°C lipases in milk appears to be P. fluorescens (59). Frieden (obligate psychrophiles) (203). (63) stated the lipase from P. fluorescens is activated Psychrotrophs are mostly gram-negative, non-spore after a short heat treatment. Frieden also stated that en­ forming rods; they are usually aerobic; nearly all of them zymes and enzymatic activities are connected with associ­ produce heat-resistant metabolites; most of them are fairly ation to other molecules, e.g. proteins, and this activation Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 resistant to penicillin and some other compounds. This is probably from dissociation of an enzyme-inhibitor- gram-negative group includes bacteria belonging to the complex as a result of the heat treatment. The lipase from genera Pseudomonas, Achromobacter, Flavobacterium P. fluorescens was thermostable enough to be active in and Alcaligenes (134,219). Among the gram-positive bac­ foods which were subjected to high-temperature-short- teria, enterococci, micrococci and certain species of ther- time (HTST) sterilizing processes. As a consequence, moduric Bacillus and Clostridium are important lipases can cause quality changes in foods and reduce (134,207). Additionally, certain and yeasts are their shelf-life (3,168). capable of rapid growth in refrigerated foods (203). During psychrotrophic growth in milk, certain gross Among psychrotrophs producing heat-resistant proteases, changes in the protein occur which include: (a) a de­ pseudomonads are the most common contaminants of crease in the total protein content, (b) changes in the rela­ milk (752). tive amounts of the protein fractions and (c) the appear­ Psychrotrophs are generally found in water and soil ance of two different atypical protein fractions (149). and are introduced into milk through these sources and Psychrotrophic microorganisms may have an indirect as become established on milk contact surfaces, equipment, well as a direct effect on the quality of dairy products. flooring and drains in the processing plant. They may be Indirectly, psychrotrophs produce off-flavors and -odors introduced via wash water (203). Milk produced under during growth in stored refrigerated raw milk which may sanitary conditions usually contains less than 10% of the carry over into the finished product even though the or­ total microbial flora as psychrotrophs, but milk produced ganisms fail to survive pasteurization. Directly, or­ under unsanitary conditions can contain more than 75% ganisms surviving pasteurization or resulting from post- psychrotrophs (96). pasteurization contamination may multiply in sufficient While there have been a few reports of psychrotrophs numbers during manufacture and storage of dairy prod­ surviving laboratory pasteurization in very small num­ ucts so as to reduce the shelf-life, the quality and the bers, there is abundant evidence that in commercial oper­ yield of the finished product (134). These bacteria cause ations this rarely happens. Consequently, there is a high bitter, fruity, rancid or yeasty flavors (36). Psychrot­ 6 7 certainty that their presence in a pasteurized product rep­ rophic levels in excess of 10 -10 /ml are usually required resents recontamination at or beyond the pasteurizer before organoleptic changes are detected in milk through leaking gaskets, hairline or pinhole cracks in (114,134,151,159). plates, etc. (100). With conventionally pasteurized milk, Cottage cheese manufactured from milk in which psyc­ 8 psychrotrophs increased from 1/ml to 10 /ml after 10 d hrotrophs had multiplied to levels of 107/ml and then pas­ at 7.2°C (134). El-Farekh (48) calculated the generation teurized at 62.8°C for 30 min was found to have a firmer time of P. fluorescens to be 7.2 h at 7.2°C. curd and to require less time than control The standard enumeration method for psychrotrophic milk. However, the cottage cheese was organoleptically bacteria involves pour-plating a sample in tryptone yeast unacceptable (88). Mohamed and Bassette (140) attri­ extract agar (TGE), and incubation for 10 d at 7°C (124). buted cottage cheese vat failure to milk heavily contami­ A more rapid technique was developed by Oliveria and nated with psychrotrophs whether the cheese was man­ Parmalee (146), who reported that enumeration of psyc­ ufactured by the direct-acid-set method or by the conven­ hrotrophic bacteria at 21°C for 25 h (mPBC) produced tional starter culture procedure. The formed curd disinteg­ counts in very good agreement with those obtained by rated and shattered during the manufacturing procedure. standard psychrotrophic count. Correlation coefficients in It was found that average cottage cheese yields decrease raw milk were .992 and in pasteurized milk .996. More (6,86). Principal psychrotrophic defects in cottage cheese specific detection and enumeration of gram-negative bac­ are slime formation, surface discoloration, off-odors and teria is accomplished by the use of inhibitors. Crystal off-flavors (123). Stone and Naff (191) found increases violet and neotetrazolium chloride inhibits growth of in soluble nitrogen and bitter flavor to be associated with gram-positive bacteria without causing inhibition of gram- increases in psychrotrophic bacteria in cottage cheese.

JOURNAL OF FOOD PROTECTION. VOL. 49, SEPTEMBER 1986 ASSESSING SHELF-LIFE AND QUALITY OF DAIRY FOODS 743

Contaminated water has been shown to be a source of trated and (d) the proteolysis proceeded in sterile solu­ cottage cheese spoilage organisms (142). tions containing enzymes but no living organisms. Peter­ Survival of psychrotrophic coliforms in a cultured son and Gunderson (155), in studying certain characteris­ product was found to be limited, and it was concluded tics of proteolytic enzymes from P. fluorescens, reported that they have little effect on shelf-life (174). that extracellular, proteolytic enzyme elaboration was in­ versely proportional to the temperature, at least from 0 PROTEOLYSIS to 30°C. It would appear temperature plays an important role in the effect of proteolytic activity on dairy quality. The first indication of naturally occurring proteolytic Sandvik (176) concluded that the interpretation of the enzymes in milk was published as early as 1879 (164). food spoilage potential of proteolytic bacteria should in­ Storrs and Hull (192) also contributed substantially to the clude the possible residual effect of the enzyme after pas­ belief that the protease enzyme is a natural constituent teurization or sterilization. Extracellular enzymes, produc­ of milk. Harper (83) stated that raw milk usually (but ed by microbes in refrigerated foods before heat treat­ not always) contains a small and variable amount of pro­

ment, may not be completely inactivated by the heating Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 tease enzyme, and that proteolysis is slight. The most fre­ and may be active in the stored product (129). Protease quently represented protease-producing psychrotrophic enzymes survived 70°C for 30 min; however, when heat­ gram-negative bacterial genus is Pseudomonas, with P. ing was at 40°C for 30 min, proteolytic activity decreased fluorescens as the most common species (112). significantly (33). Several researchers have shown pro­ Shahani (197) reported that at least 19 enzymes have tease inactivation at 63°C for 15 h, 71°C for 8 h and been found in normal cow's milk. Enzymes occurring 121°C for 9 min (1,122,129, respectively). Adams (i) naturally in milk could be classified as follows: (a) en­ compared inactivation of MC 60 protease at 149°C to in­ zymes acting as a hydrolytic group, (b) enzymes having activation of bacterial spores to establish the ultra-high- a physiological role, (c) enzymes associated with the temperature (UHT) sterilization parameters. The protease microsomal particles of milk and (d) enzymes having an was more than 4,000 times more resistant than Bacillus unknown role. Proteinases and peptidases constitute the stearothermophilus spores. The resistance of psychrot­ primary enzyme forms in bacteria responsible for pro­ rophic proteases suggests that their destruction by ex­ teolysis of milk proteins (209). treme heat is impractical. Most of the work on identification of native milk pro­ teinases has been with trypsin-like enzymes The destruction of heat-resistant bacterial proteases at (30,46,109,199). These enzymes belong to the group of sub-sterilization temperature may be feasible. Maximum proteinases and predominantly hydrolyze beta-ca­ low temperature inactivation occurred at 55°C. For low sein (70). Results indicate that the trypsin-like enzyme temperature inactivation to be beneficial, it is essential system consists of two serine proteinases and, in addi­ that inactivation occur at all protease concentrations. The tion, a chymotrypsin-like enzyme. Gamma- and extent of protease inactivation appeared to be independent related phosphoproteins are fragments of beta-casein. of protease concentration and, therefore, could occur at However, gamma-casein formation during storage of milk the low protease levels which might be in raw milk (9). could be expected to be slow as the enzyme reaction The practicality of this procedure in a dairy plant is ques­ would be limited by association of both substrate (beta- tionable but it is possible. casein) and enzyme with casein micelles. Optimum temperature and pH ranges for protease ac­ An entirely different situation exists during cold-stor­ tivity are 5 to 50°C and pH 6 to 8 (1,32,122). Half age of bulk milk at 2 to 6°C. Large quantities of beta-ca­ (46%) of the variability of protease activity was explained sein dissociate from micelles into milk serum primarily by milk pH, by psychrotrophic bacteria and by the stage because of its hydrophobicity and changes in the salt of . The second half (54%) of this variability in equilibrium (163). Behavior has been similar for pro­ activity could be only due to variations of the native milk teinases associated with micelles. Proteolytic activity proteinase system (91). brings about a modification of the electrophoretic pattern The degree of measurable proteolysis is directly corre­ of dairy products (90). lated with the incidence of the naturally occurring "bit­ Pennington et al. (153) made the observation in 1913 ter" flavor (16,17,39,87,108,210,212,213,214). Janzen that when milk was held at 0°C, the following et al. (96) reported a significant relationship between pro­ phenomena were noted: (a) proteolysis of casein was tease activity and flavor score of whole and skim milk primarily of bacterial origin, (b) proteolysis of lactalbu- (r= ..936 and -.917, respectively) at 7°C. Also con­ min was primarily due to native enzymes of the milk and cluded in the study was that, because the shelf-life of (c) both enzyme systems combined gave rise to more skim milk was significantly less than that of whole milk, rapid proteolytic changes than either system alone. and protease activity was significantly higher in skim Warner and Polis (206) stated that proteolysis which milk as compared to whole milk, the increased protease occurred in casein solutions was attributed to the presence activity in the skim may be partially responsible for its of an enzyme on the basis of the following evidence: (a) decreased shelf-life. the activity was affected by heat, (b) the activity had a The successful use of UHT treatment of milk can be definite optimum pH, (c) the activity could be concen­ hindered by heat-stable enzymes in milk (141). White

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 744 BISHOP AND WHITE

and Marshall (211), and Thomas and Mills (198) showed counts (before PI) under 10,000/ml, which did not reflect a reduction in shelf-life of Cheddar and cottage cheese the probable true hygienic conditions of milk production. due to the addition of a heat-stable protease enzyme. Other Scottish workers (184), following Chalmer's (29) Speck and Adams (189) introduced the following lead, repeated the analysis on similar samples after hold­ methods for controlling heat-stable proteases in milk: (a) ing for 24 h at 12.8-15.6°C. Because certain samples prevent contamination by psychrotrophs, (b) prevent with initial counts under 10,000/ml showed a 100-fold growth and metabolism by psychrotrophs by lower tem­ increase following PI, they concluded that there would perature and decreased aeration, addition of antibiotics seem to be little correlation between production methods and other inhibitors, and addition of starter organisms, and the relative increase in count. Meany (132) applied (c) UHT-inactivation, and (d) inactivation at sub-pasteuri­ PI at 15.6°C for 24 h to a series of samples from selected zation temperatures. farms in the Chicago area. Differences in the degree of Many methods exist for detection of proteolysis. The increase were related to conditions of production; when trichloroacetic acid (TCA) method of Hull (89) is the equipment was thoroughly cleaned, much smaller most widely used to date, but not necessarily the best. increases in counts were noted. Davis and Killmeier (44) Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 Samples et al. (175) and McKellar (131) found the trinit- reported excellent results using PI conditions of 12.8°C robenzene sulfonic acid (TNSB) method to be more suita­ for 18 h. ble for detecting proteolysis than the Hull procedure. A Several workers (84,99,102,165) have reported that the sensitive assay for protease activity based on the reaction PI count is a better indicator of raw milk microbiological of primary amino groups of trichloroacetic acid-soluble quality than the SPC. A large magnitude of change in peptides and amino acids with fluorescamine was found the total bacterial count following PI indicates psychrot­ to be suitable for determining protease activity in general rophic contamination (102,161,215). Using the PI count (31). Halambeck et al. (72) stated that because the Hull (18 h at 12.8°C), 100,000-200,000/ml is suggested as a procedure is specific for the release of aromatic amino desirable standard. This count would serve to emphasize acids, it lacks sufficient sensitivity to detect small differ­ thorough cleaning along with efficient cooling in the pro­ ences in the release of other amino acids and peptides. duction of quality milk (11,99,124,173,200). More sensitive methods are based on amine reaction with Additionally, the PI count is considered to provide a TNBS or fluorescamine. Church et al. (34) introduced a good indication of potential keeping quality of milk and spectrophotometric assay using o-phthaldialdehyde for de­ potential shelf-life of dairy foods processed from raw termination of proteolysis in milk and isolated milk pro­ milk (11,215,216,217,218). Phillips et al. (157) in­ teins. Because all hydrolytic products are assayed, the troduced a test involving preincubation of samples at method is more accurate than procedures that depend 21°C for 25 h in the presence of a mixture of ni- upon properties of aromatic residues. Furthermore, the o- sin:penicillin:crystal violet to prevent growth of gram- phthaldialdehyde assay is more rapid and convenient than positive organisms to identify post-pasteurization con­ methods using ninhydrin, TNBS or fluorescamine. An tamination of milk and single (16-18%). The test immunological detection procedure (92) and a dialysis (P-INC test) could successfully predict the level of con­ method (139) for measuring protease activity have not had apparent widespread use. tamination after storage at 6°C for 7 d in 85% of the milk samples. Bishop and White (16,17) found that PI at 21°C PRELIMINARY INCUBATION for 14 h followed by plating for mPBC produced the highest correlation of those tested (none, 2TC-7 h, 13°C- Preliminary incubation (PI) is based upon the theory 18 h, and 18°C-18 h) to potential shelf-life of pasteurized that as the holding temperature of milk is lowered, a fluid milk and cottage cheese, r= -.814 and -.586, re­ point is reached where the udder flora no longer multi­ spectively. plies while many of the psychrotrophic contaminants Ledford (115,116) examined the practice of PI before grow actively. Preliminary incubation encourages growth coliform enumeration in violet red bile agar. He found of bacterial contaminants, and aids in detecting milk in that PI for 14 d at 6.7°C indicated coliforms in 75% of which cooling masks unsanitary production and handling the samples. Unsatisfactory flavor scores were observed practices. Experimental evidence has indicated that the in 91% of the samples positive for coliforms. Shorter PI ratio of bacteria detected before and after PI is closely periods resulted in much lower percentages of related related to the cleanliness of milking equipment, especially flavor scores. Preliminary incubation of milk samples at milking machines (99). 25°C for 20 h resulted in no significant increase in The PI count is simply a Standard Plate Count (SPC) values, as an indication of proteolysis, com­ following incubation of raw milk for 18 h at 12.8°C pared to unincubated samples (107). (124). The test is designed to detect sanitation problems Johns (99) stated that it is not contended that PI will (172). Chalmers (29) in Scotland had suspected that effi­ detect every type of unsanitary condition or practice. cient cooling in farm bulk tanks might cover up insanit­ Nevertheless, Johns believed that PI furnished more in­ ary practices. To check this, he held samples at 15°C formation regarding the care taken in production and han­ for 12 h, then ran a second set of plate counts at 37°C. dling than can be obtained from methods that excluded He concluded that 24% of the 103 samples gave initial PI.

JOURNAL OF FOOD PROTECTION. VOL. 49, SEPTEMBER 1986 ASSESSING SHELF-LIFE AND QUALITY OF DAIRY FOODS 745

MOSELEY TEST of fluid dairy products are geared to measure these pseudomonads (276). Over 25 years ago, personnel at the W. K. Moseley The chief objection to Mosley's storage quality test Laboratory in Indianapolis began conducting a second (724) has been that a week or more elapses before the SPC on processed samples after holding them at 7.2°C results are available. Various workers have suggested dif­ for 5 d (700). The basis of the Moseley Keeping Quality ferent ways of shortening the waiting period. Gyllenberg Test is that practically all of the bacteria that grow well (71) used ammonium lactate agar and ammonium lactate in milk in 5-7 d at 7°C are destroyed by pasteurization. crystal violet agar, and incubated plates 2 to 4 d at 28°C, By conducting this test, two important types of informa­ for determining the incidence of organisms responsible tion can be obtained. First, if the bacterial numbers in­ for spoilage of refrigerated milk. Freeman et al. (62) in­ crease appreciably during storage, the evidence is strong vestigated the inhibitory action of various chemicals on that post-pasteurization contamination has occurred. Sec­ gram-positive organisms. They reported that .5% sodium ond, if the plate count on the fresh sample is high, ther- desoxycholate (SDC) was the most effective of all those moduric bacteria are probably in the raw product, a situa­

tested. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 tion that now occurs less frequently (49) than in earlier Olson (147) recommended the CVT test, in which 2 times. ppm of crystal violet plus 50 ppm of 2,3,5-triphenyltet- Application of the 5 d at 7.2°C keeping quality test fol­ razolium chloride (TTC) are incorporated into the plate lowed by careful study of contamination sources has count agar to repress gram-positive bacteria and also greatly improved a shelf-life of pasteurized fluid milk and make gram-negative colonies more distinctive. Sing et al. has represented a real economic advantage to plants (784) reported poor correlation between CVT counts and adopting the program (49). Several researchers SPC values after 7 d of storage at 7.2°C, with only 15% (54,110,160,186) have reported highly significant correla­ of 90 samples agreeing. Lightbody (720) used 10 IU of tions between the keeping quality of pasteurized fluid penicillin plus TTC to suppress gram-positive organisms milk and the Moseley test. Bishop et al. (75) obtained and found a highly significant correlation between count a correlation of -.770 between log Moseley count vs. po­ and samples after holding 24 h at 20°C, 4 d at 5°C, or tential shelf-life of pasteurized whole milk. 4 d at 10°C with plates incubated 3 d at 30-32°C. Blan- kenagel and Humbert (78) described a surface disc method using crystal violet as an inhibitor. SHELF-LIFE TESTS Taking advantage of the fact that gram-negative conta­ If there would be such a thing as an ideal test for de­ minants are catalase-positive, strongly aerobic, and not termining post-pasteurization contamination, and there­ repressed by surface-active agents, Maxcy (727) de­ fore, determining shelf-life, it would have the following veloped the following procedure: (a) plates are poured characteristics: (a) The ideal test would be accurate, i.e. with nutrient agar containing .5% alkyl aryl sulfonate, (b) it would indicate exactly the number of organisms which plates are dried 48 h at 32°C, then .5 ml milk is carefully contaminated the product after pasteurization. It would "spotted" onto the surface in separate droplets, (c) plates also differentiate between thermoduric bacteria that sur­ are allowed to stand undisturbed for 30 min while the vived the heat treatment and post-pasteurization conta­ milk is absorbed before being incubated in an upright minants, regardless of how few bacteria there may be. position for 16-20 h at 32°C, plus 1 h at 7°C with covers (b) The ideal test would provide results within the short­ removed, and (d) plates are flooded with 5% solution of est time possible. A little accuracy could be sacrificed hydrogen peroxide. Catalase-positive colonies decompose to get results the morning after processing, (c) The test the hydrogen peroxide and are recognized by the gas bub­ would be simple to do and it would be economical (79). bles released; only these colonies are counted. This count Examples of attempts to develop less time-consuming is related to post-pasteurization contamination. Hankin tests are: (a) use of preincubation to build up relatively and Dillman (73) used the fact that pseudomonads are large populations that are easier to enumerate, (b) use of strongly oxidase-positive to correlate the oxidase test with selective media that permit growth of contaminants but keeping quality. A gram-negative psychrotrophic bacteria inhibit growth of other bacteria, (c) testing for certain count has been highly correlated (77,217,218) and not end-products of microbial metabolism or changes in the correlated at all (133) to keeping quality. Hankin and substrate, (d) applying the sample on the surface of agar others (74,79,80) found an inverse correlation between media to accelerate growth of aerobes, (e) flavor tests, flavor score and SPC and the oxidase count. and (f) use of a combination of two or more of these Janzen (93) suggested that shelf-life of pasteurized (79,276). milk is influenced by somatic cell concentration in raw The entire concept of shelf-life is based on the well- milk. High somatic cell concentrations are indicative of documented supposition that the primary cause of shelf- abnormalities that frequently involve increased microbial life deterioration is gram-negative psychrotrophic bac­ flora. teria. Most of these bacteria are members of the genus Boyko and Blankenagel (23) utilized the findings of Pseudomonas. Consequently, most of the tests which Freeman et al. (62) that .5% SDC was the best inhibitor have been designed to predict or estimate the shelf-life of gram-positive organisms. They found that all 45 test

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 746 BISHOP AND WHITE organisms which grew in milk containing SDC were mals (4), and as a method of detecting pyrogen in paren­ killed by laboratory pasteurization and, therefore, as­ teral pharmaceuticals (37). This method has been shown sumed that organisms present in processed milk which to be the most sensitive available for detection of endoto­ grew in the presence of SDC were contaminants. Catch- xin (38,104). ick and Gibon (28) sought a way to shorten the SDC According to the reported mechanism of the Limulus test to 16 h so that results would be available the next Amoebocyte Lysate (LAL) assay, endotoxin activates a day. They modified the SDC test by adding resazurin. proenzyme-enzyme cascading reaction resulting in activa­ The advantages of this modified resazurin reduction test tion of coagulase (145,195). Coagulase hydrolyzes coagu- are: (a) use of a large (9 ml) sample, (b) use of PI to len, a clottable heat-stable protein (194,196), at - build up small numbers, (c) detection of only gram-nega­ and arginine- linkages, resulting in re­ tive organisms, and (d) providing results the next morn­ moval of the peptide-C chain to form a gel (coagulin) ing (79). The methylene blue test after pre-incubation at (207). 18°C for 18 h was a good index of keeping quality, as This particular assay included the following "grading determined by Jooste and Groeneveld (103). Custer and of lysate gelation": 4+ = firm gel with considerable Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 Knight (41) found that Early Detection (SDC + resazu­ opacity; 3+ = soft gel with moderate to considerate rin) test to be as accurate as the Moseley test in measur­ opacity; 2 + = weak gel with slight to moderate opacity ing the shelf-life and the results can be obtained 6 d and adhesion of starch-like floccules to sides of the tube sooner. Waes and Bossuyt (204) reported that the benzal- when slanted; 1+ = very weak gel-with slight opacity kon-crystal violet-ATP method was useful for predicting and with some starch-like floccules adhering to sides of keeping quality. the tube; negative = no visible increase in viscosity or Marshall and Harmon (121) found pyruvate concentra­ opacity. As little as 100 pg of certain endotoxins/ml may tion to be of limited value as a quality test for milk due be reliably detected (103). • . to great variability of results. Senyk et al. (178) and DiLuzio and Friedman (45) suggested the Limulus Shipe et al. (182) obtained correlation values of flavor assay might be used for detection of bacterial endotoxin score to pyruvate content of -.81 and -.78, respectively. in drinking water and other surface waters. Jorgensen et Asher and Sargent (4) produced the best fitted regression al. (705) reported that the Limulus assay procedure was 2 equation (r = .615) with data on changes of acidity, pH, easily adapted to the testing of water samples for endoto­ acid degree value and free fatty acids as related to keep­ xin. Measured endotoxin concentrations varied from .78 ing quality. Shipe and Hsu (183) measured fluorescent ng/ml to 1,250 ng/ml. The Limulus assay is currently compounds in milk to provide an index of milk quality. permitted by the Food and Drug Administration to deter­ Rodrigues and Pettipher (769) reported that the keeping mine possible endotoxin contamination of ingredients quality of pasteurized milk stored at 5 and 11 °C was pre­ used to prepare parenteral and biological products in the dicted within 24 h by pre-incubating samples and count­ . The great advantage of the assay ing bacteria by the Direct Epifluorescent Filter Technique is its rapidity - a total test time of less than 2 h. (DEFT). The Limulus endotoxin assay was compared to the For any of the aforementioned shelf-life tests to be ef­ SPC and coliform count for assessment of the bac­ fective, there must exist a control of heterogeneity of teriological quality of reclaimed waste water by Jorgensen samples. Maxcy and Wallen (728) stated that a single et al. (106). LAL assays were technically simpler to per­ package provides low probability for predicting behavior form and provided results much sooner than conventional of the entire production lot. The extreme differences in culture methods. However, the endotoxin values did not spoilage rates of individual units with sample sets indi­ correlate extremely well with determinations of viable cated sensory evaluation of multiple samples to be the bacterial numbers. This lack of correlation may have most logical, simple criterion for evaluating shelf-life. been due to alterations in the normal ratio of viable gram- The aim of shelf-life tests should be to detect and elim­ negative cells to endotoxin caused by water reclamation inate all sources of recontamination. When this has been procedures. Seiter and Jay (777) determined endotoxins accomplished, shelf-life determination and detection in ground beef by the LAL assay. should no longer be a problem (100). Clark (35) evaluated various methods to enumerate psychrotrophic bacteria in fluid milk and concluded that ENDOTOXIN (LIPOPOLYSACCHARIDE) the LAL test could be used as a simple shelf-life projec­ DETECTION tion test. Mikolajcik and Brucker (135) studied the LAL assay's potential for detection of psychrotrophic spoilage Levin and Bang (778) originally described the ability organisms in foods. They stated that the LAL assay is of amoebocyte lysate from Limulus polyphemus, the a rapid and highly sensitive method to detect endotoxins horseshoe crab, to form a gel in the presence of minute in biological systems. Endotoxin activity is associated amounts of endotoxin. Since that time, the Limulus assay with the Lipid A segment of lipopolysaccharides con­ has been used as a method of detecting endotoxin in pa­ stituting the outer cell membrane of gram-negative bac­ tients with gram-negative septicemia (26,119,167), for teria (GNB). Both viable and non-viable GNB are de­ the study of experimental endotoxemia and shock in ani­ tected which makes this test particularly useful in tracing

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 ASSESSING SHELF-LIFE AND QUALITY OF DAIRY FOODS 747 the history of the milk supply. The LAL assay is finding increased turbidity by strongly bridging gel micelles, or increased use as a rapid indirect indicator of gram-nega­ all three. tive spoilage organisms in refrigerated meat and dairy Bodyfeldt et al. (20) reported the LAL test included products. The LAL assay will detect 103-104 GNB/ml of the following limitations: expensive reagents, lack of suf­ milk. Mikolajcik and Brucker (137) also found that for ficient sensitivity to microquantities of endotoxins, and each log increment in GNB population, the lipopolysac- complexity of the procedure. This lack of sensitivity was charide (LPS) titers also increased one log cycle with a alleviated by Bishop et al. (14) by removing the problem correlation coefficient of .88. Average LPS values in­ of varied background turbidity by "blanking" each sam­ creased from .86 ng/ml to 4,350 ng/ml as log increments ple with an unincubated duplicate. They also detected of GNB/ml increased from 1.0 to 8.0 with correlation significant linear relationships between LPS content of coefficient of .95. Currently used GNB enumeration pro­ milk and psychrotrophic bacteria count (r=.715), flavor cedures can detect viable GNB in the product resulting score (r = -.918) and days of storage at 4.5°C (r=.939). from post-pasteurization contamination but not GNB Bishop and White (16,17) evaluated the LAL assay as

which were present before pasteurization. Pasteurization related to potential shelf-life of pasteurized fluid milk and Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 did not affect the LPS titer as it is heat-stable. When cottage cheese. The spectrophotometric LAL method was the mean GNB counts of commercial pasteurized milk used for milk and a correlation value of -.913 resulted. were plotted for each LPS value, a correlation coeffi­ With cottage cheese shelf-life, the LAL gelatin method cient of .99 was obtained. For 186 individual trials, a gave a correlation value of -.811. correlation coefficient of .98 was obtained. Sudi (193) pointed out that the Limulus test may easily Mikolajcik (136) revealed that the LAL assay will de­ develop into an entirely new double-purpose method of tect within 1 h as few as 100 GNB/ml of milk. The accu­ food analysis. The test may not only prove to be suitable racy of the test improves where counts exceed 10,000 for detection of manipulations of raw materials, but it GNB/ml of milk. The outstanding feature of the proce­ may also provide means for demonstrating that good dure is its ability to detect the quality of the raw milk manufacturing practices have been followed as closely as supply even after pasteurization. Thus it furnishes infor­ possible. Mikolajcik (136) feels the Limulus test, because mation about the number of GNB which were present in of its sensitivity, accuracy and speed, will be of tremend­ the raw milk supply and the possibilities that proteases ous value to the dairy industry as a rapid screening test and lipases might be present in the milk which would for raw milk and as a guide for predicting the shelf-life affect the keeping quality of the finished product. of pasteurized and UHT . Sudi (193) discovered that, because of the heat stability of gram-negative lipopolysaccharides, the evidence of IMPEDANCE DETECTION bacterial growth in raw milk cannot be destroyed by UHT treatment, but can be detected by the LAL assay. In fact, Several rapid automated methods have been proposed there was an apparent 3.6-fold increase of GN-LPS con­ to overcome the disadvantages inherent in plate counts. centration following UHT treatment. The reason for this These include: ATP photometry (222), fluorescence increase was not explained. It is proposed that the main microscopy (156), radiometric measurements (158) and advantage of the Limulus assay lies in the assessment of impedance detection (25). The impedance method has the bacteriological quality of heat-treated dairy products, been considered to be the most promising of the instru­ like UHT milk, where cultural methods are negative. mental methods (221). Evans et al. (53) modified the Limulus lysate assay Impedance is the resistance to the flow of an alternat­ so that the reaction of lysate and endotoxin formed a tur­ ing electrical current through a conducting material, and bid suspension instead of a firm clot. Absorbency at 360 has been shown to be a complex entity composed of a nm is measured. The firm-clot method was found to be resistive (or conductive) component and a reactive com­ less sensitive and reproducible for detection of endotoxin ponent (capacitance). The impedance technique relies on than the spectrophotometric modification of the Limulus the fact that metabolizing microorganisms alter the chem­ lysate assay. Tsuji and Steindler (201) found the LAL ical composition of the growth medium and that these sensitivity to endotoxin increased approximately 20- to chemical changes cause a change in the impedance of the 30-fold when lyophilized LAL was reconstituted with de- medium (2,150). Total impedance change represents the pyrogenated natural seawater. Magnesium was identified vector sum of changes in the resistance and capacitance as the component in natural seawater and commercial sea components of the growth medium. Impedance changes salts responsible for the increased sensitivity. The greatest are detectable when the concentration of microorgansims 6 7 endotoxin sensitivity for the LAL reagent was achieved exceeds a threshold level of 10 to 10 cells per milliliter. with solutions of magnesium ranging from 50 to 60 nM. The time required for the initial inoculum to reach the Magnesium may increase the LAL sensitivity by: (a) act­ threshold level is designated as the detection time and ing at the initial endotoxin recognition phase by modify­ is a function of both the initial concentration and the spe­ ing the size of endotoxin for easier recognition by LAL, cific growth kinetics of the organism in the given (b) accelerating the rate of proenzyme-enzyme cascading medium. By comparing the detection time obtained to the reaction, (c) contributing to formation of a firmer gel or results of a standard calibration curve, an estimate of the

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 748 BISHOP AND WHITE initial concentration of microorganisms can be made. The psychrotrophic organisms, have two severe limitations. impedance detection time (IDT) is defined as the ac­ First, spoilage is not always directly related to the celerating change of impedance associated with microbial number of organisms present (159). Second, it appears growth and metabolism. The computer uses a complex that psychrotrophs are only part of the milk spoilage algorithm that determines the onset of acceleration in the problem. Poor flavor and keeping quality can also be at­ impedance curve. The calculation of the IDT from impe­ tributed to the presence of microbial enzymes and dance curves is a critical part of the impedimetric metabolic products (149,151) from organisms present be­ method. To obtain consistent interpretations of IDT by fore pasteurization. Thus, present methods, although of­ the computer algorithm, it is important to generate impe­ fering some useful information, are too slow and often dance curves with smooth baselines and sharp accelera­ too inaccurate to meet the needs of milk producers and tions. It is important to realize that the basis of the im­ processors. An optimal microbiological test would pro­ pedimetric estimation is metabolic change, while that of vide counting and keeping quality estimates within a time a colony count estimation is biomass production. The allowing for effective corrective measures. Cady et al.

(25) also found that detection times correlate better with Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 consequences of this difference must be carefully consid­ the shelf-life than do the SPC and psychrotrophic count. ered in the preparation of samples for impedimetric esti­ Cady's early data show promise as a 9-14 h impedance- mation of the number of microorganisms (56). based keeping quality prediction. Bossuyt and Waes (22) Impedance measurements are particularly useful for determined that impedance measurements are useful to rapidly screening various foods to determine whether they trace post-pasteurization contamination in pasteurized meet the desired microbiological criterion. Rowly et al. milk. Martins (125) evaluated a rapid impedimetric (171) found the impedance method to be promising as method for determining the keeping quality of milk and a rapid screening technique to determine if cooked meats concluded that positive correlation coefficients indicate have less than one coliform per gram. Hardy et al. (18) that samples with short shelf-life tended to produce early used the impedance method to rapidly assess whether a impedance response detection times and samples with sample of frozen vegetable contains greater or less than long shelf-life tended to produce late detection times. 105 organisms per gram. Firstenberg-Eden and Klein (57) Bishop et al. (75) and Bishop and White (16,17) obtained a correlation coefficient of .90 between the im­ evaluated the value of impedance detection as related to pedimetric test and violet red bile agar (VRBA) plate potential shelf-life of pasteurized fluid milk and cottage counts for coliform using a total of 91 meat samples. cheese. It proved the most valuable on a combined time O'Connor (143) used an impedance method for deter­ and relationship basis as compared to standard micro­ mination of bacteriological quality of raw milk, and biological procedures, and proteolysis and endotoxin de­ found that an IDT of 8.5 h differentiated samples with tection. Resultant correlation values for milk and cottage 5 less than or greater than 10 bacteria per milliliter. Gnan cheese shelf-life and impedance detection were .930 and and Luedecke (67) stated that impedance detection offers .897, respectively. a definite time advantage, approximately 7 h vs. 48 h for the SPC, when examining raw milk samples. A classifi­ CONCLUSION cation system that selects a cut-off detection time of ap­ proximately 7 h is helpful in screening out samples likely Each method used for shelf-life estimation has its ad­ to have a SPC exceeding 105/ml. vantages and its drawbacks. Standard microbiological Waes and Bossuyt (205) described a simple impedimet­ methods, including the associated Moseley keeping qual­ ric method to detect, within 2 h, complete failure of a ity test, are familiar to laboratory personnel in the dairy starter due to bacteriophages in the manufacture of Ched­ industry, and offer some useful information, though their dar cheese. This method is based on the observation that estimative value is limited. Bacterial metabolite detection about 105 disturbing bacteriophages per milliliter will is of increasing interest as emphasis is decreasing with cause complete failure of the starter and prevent the nor­ bacterial numbers. The most commonly investigated mal change of impedance due to lactic starter bacteria. metabolites include protease and lipase enzymes, and en­ Firstenberg-Eden and Tricarico (58) revealed that a dotoxins produced by gram-negative bacteria. rapid impedimetric determination for total, mesophilic Automated has made its way into the and psychrotrophic counts in raw milk showed correla­ dairy industry with the introduction of ATP measurement tions between IDT and bacteria counts of -.96, -.95 and and impedance detection. ATP concentration has not -.96, respectively. were most often seen as proven very useful for shelf-life estimation. Impedance the predominant population. The impedimetric method al­ detection has been shown to offer relatively rapid results lowed for these samples containing above 105 cfu/ml to with high correlation values to shelf-life, which result in be screened out within 4 h. Psychrotrophic levels of 105/ fairly accurate estimation. ml and above were screened within 21 h, while total con­ The method ultimately used for shelf-life estimation centration of samples containing above 105 cfu/ml were will depend on many factors. Each situation evaluated screened within 16 h. will require an analysis of need based on product type Cadey et al. (25) stated that present keeping quality and volume, information needed, time available, cost of tests, which try to predict spoilage based on presence of test method/instrument, laboratory set-up and personnel,

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 ASSESSING SHELF-LIFE AND QUALITY OF DAIRY FOODS 749 and top management support of the quality assurance pro­ biological counts in milk. S. Afr. J. Dairy Technol. 12:89-91. gram. Thus there is no clear-cut method for shelf-life es­ 25. Cady, P., D. Hardy, S. Martins, S. W. Dufour, and S. J. timation for each and every situation. Kraeger. 1978. Automated impedance measurements for rapid sc­ reening of milk microbial count. J. Food Prot. 41:277-283. REFERENCES 26. Cardis, D. T., R. B. Reinhold, P. W. H. Woodruff, and J. Find. 1972. Endotoxemia in man. Lancet 2:1381-1385. 1. Adams, D. M., J. T. Barach, and M. L. Speck. 1975. Heat 27. Carey, K. D., and R. E. Anderson. 1982. Low-temperature resistant proteases produced in milk by psychrotrophic bacteria microflora and keeping quality of pasteurized dairy products. J. of dairy origin. J. Dairy Sci. 58:828-834. Dairy Sci. 65:76 (Abstract). 2. Allison, J. B., J. A. Anderson, and W. H. Cole. 1938. The 28. Catchick, A., and D. L. Gibon. 1967. Sixteen hour method for method of electrical conductivity in studies on bacterial detecting post-pasteurization contamination in market milk. Can. metabolism. J. Bacterid. 36:571-586. Dairy J. 46:25-27. 3. Anderson, R. E., C. B. Hedlund, and U. Jonsson. 1979. Ther­ 29. Chalmers, C. H. 1956. The bulk cooling, storage, and transporta­ mal inactivation of a heat-resistant lipase produced by the psyc­ tion of milk. Dept. Health (Scotland) Health Bull. 14:1. hrotrophic bacterium Pseudomonas fluorescens. J. Dairy Sci. 30. Chen, J. H., and R. A. Ledford. 1971. Purification and character­ 62:361-367. ization of milk protease. J. Dairy Sci. 54:763.

4. Asher, Y. J., and A. G. Sargent. 1983. Relation between the 31. Chism, G. W., A. E. Hung, and J. A. Marshall. 1979. Sensitive Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 biochemical changes in milk under varying incubation conditions assay for proteases in sterile milk. J. Dairy Sci. 62:1798-1800. and the keeping quality of pasteurized milk stored at 7.2°C. J. 32. Christen, G. L., and R. T. Marshall. 1984. Selected properties Dairy Sci. 66:64 (Abstract). of lipase and protease of Pseudomonas fluorescens 27 produced 5. Atherton, H. V., E. J. Doan, and G. W. Watrous, Jr. 1953. in four media. J. Dairy Sci. 67:1680-1687. Observations on bacterial population and characteristics of bottled 33. Christen, G. L., and R. T. Marshall. 1984. Thermostability of milk under refrigerated conditions. J. Dairy Sci. 36:570. lipase and protease of Pseudomonal fluorescens 27 produced in 6. Aylward, E. B., J. O'Leary, and B. E. Langlois. 1980. Effect various broths. J. Dairy Sci. 67:1688-1693. of milk storage on cottage cheese yield. J. Dairy Sci. 63:1819- 34. Church, F. C, H. E. Swaisgood, D. H. Porter, and G. L. Catig- 1825. nani. 1983. Spectrophotometric assay using o-phthaldialdehyde 7. Baker, S. K. 1983. The keeping quality of refrigerated pas­ for determination of proteolysis in milk and isolated milk pro­ teurized milk. Austral. J. Dairy Technol. 38:124-127. teins. J. Dairy Sci. 66:1219-1227. 8. Bandler, D. K. 1971. Milk flavor: The true test of quality. J. 35. Clark, M. G. 1980. Evaluation of methods to enumerate psychrot­ Milk Food Technol. 34:385-387. rophic bacteria in fluid milk. M. S. Thesis, Oregon State Univ. 9. Barach, J. T., D. M. Adams, and M. L. Speck. 1976. Low 36. Collins, E. B. 1981. Heat resistant psychrotrophic microor­ temperature inactivation of heat-resistant proteases from psychrot­ ganisms. J. Dairy Sci. 64:157-160. rophic bacteria. J. Dairy Sci. 59:391-395. 37. Cooper, J. F., J. Levin, and H. N. Wagner. 1971. Quantitative 10. Barnard, S. E. 1974. Flavor and shelf-life of fluid milk. J. Milk comparison of in-vitro and in-vivo methods for the detection of Food Technol. 37:346-349. endotoxin. J. Lab. Clin. Med. 78:138-147. 11. Barnard, S. E. 1981. Getting good preliminary incubation counts. 38. Cooper, J. F., H. D. Hochstein, and E. B. Seligmann, Jr. 1972. Dairy Food Sanitation 1:248-249. The Limulus test for endotoxin (pyrogen) in radiopharmaceuticals 12. Barnum, H. J. 1969. The evils of the standard plate count. J. and biologicals. Bull. Parenter. Drug Assoc. 26:153-162. Milk Food Technol. 32:219. 39. Cousin, M. A. 1982. Presence and activity of psychrotrophic 13. Bassette, R., D. Y. C. Fung, H. Roberts, and G. Ward. 1982. microorganisms in milk and dairy products. A review. J. Food A survey of milk flavor and quality. J. Food Prot. 45:135-138. Prot. 45:172-207. 14. Bishop, J. R., A. B. Bodine, and J. J. Janzen. 1983. Relation­ 40. Cuevas, P., and J. Fine. 1972. Role of intraintestinal endotoxin ship of endotoxin concentration to shelf-life of pasteurized fluid in death from peritonitis. Surg. Gynecol. Obstet. 134:953-957. milk. J. Dairy Sci. 66:268 (Abstract). 41. Custer, E. W., and H. Knight. 1980. An evaluation of the relia­ 15. Bishop, J. R., C. H. White, and R. Firstenberg-Eden. 1984. bility of the C. V. T., Early Detection, and Moseley tests as Rapid impedimetric method for determining the potential shelf-life predictors of shelf-life of pasteurized fluid milk. J. Dairy Sci. of pasteurized whole milk. J. Food Prot. 47:471-475. 63:169 (Abstract). 16. Bishop, J. R., and C. H. White. 1985. Estimation of potential 42. Cyzeska, F. J., J. A. Seiter, S. N. Marks, and J. M. Jay. 1981. shelf-life of pasteurized fluid milk utilizing bacterial numbers and Culture medium for selective isolation and enumeration of gram- metabolites. J. Food Prot. 48:663-667. negative bacteria from ground meats. Appl. Environ. Microbiol. 17. Bishop, J. R., and C. H. White. 1985. Estimation of potential 42:303-307. shelf-life of cottage cheese utilizing bacterial numbers and 43. Dabbah, R., W. A. Moats, and J. C. Olson, Jr. 1970. Correlation metabolites. J. Food Prot. 48:1054-1057. between fieldman ratings of sanitary conditions on dairy farms 18. Blankenagel, G., E. S. Humbert. 1965. A rapid method to deter­ and six commonly used bacterial tests. J. Dairy Sci. 53:642. mine post-pasteurization milk contamination. Can. Dairy Ice 44. Davis, R., and J. T. Killmeier. 1959. A new method for deter­ Cream J. 44:28-29. the quality of bulk tank milk. Am. Milk Rev. and Milk 19. Blankenagel, G. 1976. An examination of methods to assess post- Plant Mo. 21:37-38. pasteurization contamination. J. Milk Food Technol, 39:301-304. 45. DiLuzio, N. R., and T. J. Friedman. 1973. Bacterial endotoxins 20. Bodyfelt, F. W., M. G. Clark, and W. E. Sandine. 1981. Evalu­ in the environment. Nature (London) 244:49-51. ation of three methods to enumerate psychrotrophic bacteria in 46. Dulley, J. R. 1971. Bovine milk protease. J. Dairy Res. 39:1-9. fluid milk. J. Dairy Sci. 64:50 (Abstract). 47. Eiland, F. 1979. An improved method for determination of 21. Bossuyt, R. G. 1981. Bacteriological quality determination of raw adenosine triphosphate (ATP) in soil. Soil Biol. Biochem. 11:31- milk by an ATP assay technique. J. Dairy Sci. 64:59 (Abstract). 35. 22. Bossuyt, R. G., and G. M. Waes. 1983. Impedance measure­ 48. El-Farekh, T. A. 1962. Quantitative study of growth of some ments to detect post-pasteurization contamination of pasteurized dairy psychrophiles. M. S. Thesis, Univ. Tenn. milk. J. Food Prot. 46:622-624. 49. Elliker, P. R., E. L. Sing, L. J. Christen, and W. E. Sandine. 23. Boyko, A., and G. Blankenagel. 1966. Detection of contamina­ 1964. Psychrophilic bacteria and keeping quality of pasteurized tion in pasteurized milk. Can. Dairy Ice Cream J. 45:20-21. dairy products. J. Milk Food Technol. 27:69-75. 24. Britz, T. J., J. J. Bezuidenhour, J. M. Dreyer, and P. L. Steyn. 50. Elliker, P. R. 1968. Shelf-life of food products. Am. Dairy Rev. 1980. Use of adenosine triphosphage as an indicator of the micro­ 30:80-81.

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 750 BISHOP AND WHITE

51. Erfle, J. D., S. Mahadevan, and F. D. Sauer. 1979. Effect on 74. Hankin, L., and E. O. Anderson. 1969. Correlations between quality of adenosine-5'-triphosphate concentration and adeny­ flavor score, flavor critism, standard plate count, and oxidase late energy charge of rumen microbes from fistualted cows. J. count on pasteurized milks. J. Milk Food Technol. 32:49-51. Dairy Sci. 62:284-291. 75. Hankin, L., and G. R. Stephens. 1972. What tests usefully pre­ 52. Erfle, J. D., S. Mahadevan, and F. D. Sauer. 1981. Relationship dict keeping quality of perishable foods? J. Milk Food Technol. between adenylate energy charge, rumen volatile con­ 35:574-576. centrations, and rates of production and dry matter digestibility 76. Hankin, L., and W. F. Dillman. 1972. Further studies in the in the cow. J. Dairy Sci. 64:634-642. flavor quality of retail milk in Connecticut. J. Milk Food Tech­ 53. Evans, T. M., J. E. Schillinger, and D. G. Stuart. 1978. Rapid nol. 35:710-712. determination of bacteriological water quality by using Limulus 77. Hankin, L., G. R. Stephens, and W. F. Dillman. 1975. Quality lysate. Appl. Environ. Microbiol. 35:376-382. control significance of special media for enumeration of microbial 54. Felingham, D., and H. S. Juffs. 1977. Comparison and evalua­ groups in cottage-type cheese. J. Milk Food Technol. 38:738-744. tion of keeping quality tests for pasteurized milk and cream prod­ 78. Hankin, L., W. F. Dillman, and G. R. Stephens. 1977. Relation ucts. Austral. J. Dairy Technol. 32:158-612. of code dates to quality of milk sold in retail markets. J. Food 55. Finley, R. D., H. B. Warren, and R. E. Hargrove. 1969. Storage Prot. 40:116-119. stability of commercial milk. J. Milk Food Technol. 32:382-387. 79. Hankin, L., W. F. Dillman, and G. R. Stephens. 1977. Keeping 56. Firstenberg-Eden, R. 1983. Rapid estimation of the number of quality of pasteurized milk for retail sale related to code date, Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 microorgansims in raw meat by impedance measurement. J. Food storage temperature, and microbial counts. J. Food Prot. 40:848- Technol. 37:64-70. 853. 57. Firstenberg-Eden, R., and C. S. Klein. 1983. Evaluation of a 80. Hankin, L., and D. Shields. 1983. Keeping quality and flavor, good rapid impedimetric procedure for the quantitative estimation and microorganisms, proteases, and lipases in raw cow and of coliforms. J. Food Sci. 48:1307-1311. milk at collection and after storage. J. Food Prot. 46:873-877. 58. Firstenberg-Eden, R., and M. K. Tricarico. 19S3. Impedimetric 81. Hardy, D., S. J. Kraeger, S. W. Dufour, and P. Cady. 1977. determination of total, mesophilic, and psychrotrophic counts in Rapid determination of microbial contamination in frozen vegeta­ raw milk. J. Food Sci. 48:1750-1754. bles using automatic impedance measurements. Appl. Environ. 59. Fitz-Gerald, C. H., and H. C. Deeth. 1983. Factors influencing Microbiol. 34:14-17. lipolysis by skim milk cultures of some psychrotrophic microor­ 82. Harmon, L., G., G. M. Trout, and T. I. Hedrick. 1959. Shelf- ganisms. Austral. J. Dairy Technol. 38:97-103. life of homogenized milk. Milk Dealer 48:40-42. 60. Forsberg, C. W., and K. Lam. 1977. Use of adensine-5'-triphos- 83. Harper, W. J. 1960. Observations on milk protease. J. Dairy Sci. phate as an indicator of the microbiota biomass in rumen con­ 43:1850-1851. tents. Appl. Environ. Microbiol. 33:528-537. 84. Hartley, J. C, G. W. Reinhold, and E. R. Vedamuthu. 1968. 61. Foster, M. J., and O. Rahn. 1936. Growth and fermentation of Bacteriological methods for evaluation of raw milk quality. A re­ bacteria near their minimum temperature. J. Bacteriol. 32:485- view. 1. Use of bacterial tests to evaluate production conditions. 497. J. Milk Food Technol. 31:315-322. 62. Freeman, T. R., N. V. Nanavati, and W. E. Glenn. 1964. Faster 85. Hartley, J. C, G. W. Reinhold, E. R. Vedamuthu, and W. S. enumeration of psychrophilic bacteria in pasteurized milk. J. Milk Clark, Jr. 1968. Bacterial tests of grade A raw milk samples as Food Technol. 27:304-307. a measure of farm production conditions. J. Milk Food Technol. 63. Frieden, C. 1971. Protein-protein interaction and enzymatic activ­ 31:388-392. ity. Ann. Rev. Biochem. 40:653-660. 86. Hicks, C. L., M. Allauddin, B. E. Langlois, and J. O'Leary. 64. Fung, D. Y. C, and W. S. LaGrange. 1969. Microtiter method 1982. Psychrotrophic bacteria reduce cheese yield. J. Food Prot. for enumerating viable bacteria in milk. J. Milk Food Technol. 45:331-334. 32:144-146. 87. Hockney, L. J., and M. A. Cousin. 1981. Detection of heat-resis­ 65. Fung, D. Y. C, and A. A. Kraft. 1969. Rapid evaluation of tant proteases produced by psychrotrophs in refrigerated pas­ viable cell counts using microtiter system and MPN technique. teurized milk. J. Dairy Sci. 64:51 (Abstract). J. Milk Food Technol. 32:408-409. 88. Honer, C. 1981. Raw milk psychrotrophs and the importance of 66. Gillis, W. T., and E. W. Custer. 1982. Effect of repasteurization their control. Dairy Rec. 82:104-108. on shelf-life of fluid milk. J. Dairy Sci. 65:220 (Abstract). 89. Hull, M. E. 1947. Studies on milk proteins. II. Colormetric deter­ 67. Gnan, S., and L. O. Luedecke. 1982. Impedance measurements mination of the partial hydrolysis of the proteins in milk. J. Dairy in raw milk as an alternative to the standard plate count. J. Food Sci. 30:881-884. Prot. 45:4-7. 90. Humbert, G., and C. Alais. 1979. The milk proteinase system. 68. Gomez-de-Lasse, C. 1970. Bacteriological and organoleptic J. Dairy Res. 46:5590-571. analyses of milk samples exposed to different time-temperature 91. Humbert, G., J. Brun-Bellut, T. Mora, and C. Alais. 1983. Pro­ relationships. M. S. Thesis, Univ. Vermont. teinase activity in bovine raw milk. J. Dairy Sci. 66:2271-2277. 69. Grosskopf, J. C, and W. J. Harper. 1969. Role of psychrot­ 92. Hutkins, R. W., and R. T. Marshall. 1981. Immunological detec­ rophic sporeformers in long life milk. J. Dairy Sci. 52:897 tion of a heat-stable protease of Pseudomonas fluorescens P26 in (Abstract). milk. J. Dairy Sci. 64:47 (Abstract). 70. Groves, M. L., W. G. Gordon, E. B. Kalan, and S. B. Jones. 93. Janzen, J. J. 1972. The effect of somatic cell concentration in 1972. Composition of bovine y-casein A1 and A3, and further the raw milk on the shelf-life of the processed product. J. Milk evidence for a relationship in the biosynthesis of Y- and (3-casein. Food Technol. 35:112-114. J. Dairy Sci. 55:1041-1049. 94. Janzen, J. J., A. B. Bodine, and J. R. Bishop. 1981. Effects 71. Gyllenberg, H. 1960. Contamination and deterioration of market of package temperature and days of storage on the flavor score milk. III. A selective plating test for the demonstration of signifi­ of processed milk. J. Food Prot. 44:455-458. cant numbers of pseudomonads. Acta Agric. Scand. 10:65-68. 95. Janzen, J. M., J. R. Bishop, A. B. Bodine, and C. A. Caldwell. 72. Halambeck, S. M., R. H. Schmidt, and H. A. Morris. 1983. 1982. Shelf-life of pasteurized fluid milk as affected by age of Evaluation of spectrophotometric and fluorometric methods of as­ raw milk. J. Dairy Sci. 65:2233-2236. sessing proteolysis by lactic acid bacteria in milk. J. Dairy Sci. 96. Janzen, J. M., J. R. Bishop, and A. B. Bodine. 1982. Relation­ 66:56 (Abstract). ship of protease activity to shelf-life of skim and whole milk. 73. Hankin, L., and W. F. Dillman. 1968. A rapid test to find "po­ J. Dairy Sci. 65:2237-2240. tentially" psychrophilic organisms in pasteurized dairy products. 97. Jenkinson, D. S., S. A. Davidson, and D. S. Powlson. 1979. J. Milk Food Technol. 31:141-145. Adenosine triphosphate and microbial biomass in soil. Soil. Biol.

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 ASSESSING SHELF-LIFE AND QUALITY OF DAIRY FOODS 751

Biochem. 11:521-527. J. Dairy Technol. 21:94-99. 98. Johns, C. K. 1959. Applications and limitations of quality tests 121. Marshall, R. T., and C. C. Harmon. 1978. The automated pyru­ for milk and milk products. J. Dairy Sci. 42:1625-1650. vate method as a quality test for grade A milk. J. Food Prot. 99. Johns, C. K. 1960. Preliminary incubation for raw milk samples. 41:168-177. J. Milk Food Technol. 23:137-141. 122. Marshall, R. T., and J. K. Marstiller. 1980. Low temperature la­ 100. Johns, C. K. 1969. Tests for estimating shelf-life of milk and bility vs. high temperature stability of protease from Pseudomonas milk products. J. Milk Food Technol. 32:301-303. fluorescens Mt. J. Dairy Sci. 63:57 (Abstract). 101. Johns, C. K. 1971. Bacteriological testing of milk for regulatory 123. Marth, E. H. 1970. Spoilage of cottage cheese by psychrophilic purposes - usefulness of current procedures and recommendations microorganisms. Cult. Dairy Prod. J. 5:14-16. for change. 2. Bacteriological testing of raw milk for regulatory 124. Marth, E. H. (ed.). 1978. Standard methods for the examination purposes. J. Milk Food Technol. 34:172-180. of dairy products. 14th ed. Am. Public Health Assoc, 102. Johns, C. K. 1975. Use of counts after preliminary incubation Washington, DC. to improve raw milk quality for a Denver plant. J. Milk Food 125. Martins, S. B. 1982. Evaluation of a rapid impedimetric method Technol. 38:481-482. for determining the keeping quality of milk. J. Food Prot. 103. Jooste, P. J., and H. T. Groeneveld. 1971. Bacterial counts and 45:1221-1226. dye reduction in relation to the keeping quality of South African 126. Matoba, T., R. Hayashi, and T. Hata. 1970. Isolation of bitter pasteurized milk. S. Afr. Dairy Technol. 3:193-201. peptides from tryptic hydrolysate of casein and their chemical Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 104. Jorgensen, J. H., and R. F. Smith. 1973. Preparation, sensitivity, structure. Agric. Biol. Chem. 34:1235-1243. and specificity of Limulus lysate for endotoxin assay. Appl. 127. Maxcy, R. B. 1967. A rapid test for determining post-pasteuriza­ Microbiol. 26:43-48. tion contamination of market milk. J. Dairy Sci. 50:938. 105. Jorgensen, J. H., J. C. Lee, and H. R. Pahren. 1976. Rapid de­ 128. Maxcy, R. B., and S. E. Wallen. 1983. Heterogeneity of samples tection of bacterial endotoxins in drinking water and renovated as a problem in shelf-life prediction. J. Food Prot. 46:542-544. wastewater. Appl. Environ. Microbiol. 32:347-351. 129. Mayerhofer, H. J., R. T. Marshall, C. H. White, and M. Lu. 106. Jorgensen, J. H., J. C. Lee, G. A. Alexander, and H. W. Wolf. 1973. Characterization of a heat-stable protease of Pseudomonas 1979. Comparison of Limulus assay, standard plate count, and fluorescens P26. Appl. Microbiol. 25:44-48. total coliform count for microbiological assessment of renovated 130. McElroy, W. D., and A. Green. 1956. Function of adenosine wastewater. Appl. Environ. Microbiol. 37:928-921. triphosphate in the activation of luciferin. Arch. Biochem. 107. Juffs, H. S. 1973. Proteolysis detection in milk. J. Dairy Res. Biophy. 64:257-271. 40:371-392. 131. McKellar, R. C. 1981. Development of off-flavors in UHT and 108. Juffs, H. S. 1975. Proteolysis detection in milk. J. Dairy Res. pasteurized milk as a function of proteolysis. J. Dairy Sci. 42:31-41. 64:2138-2145. 109. Kaminogawa, S., F. Sato, and K. Yamauchi. 1971. Purification 132. Meany, J. A. 1959. Changing needs in the bacteriological exami­ and some properties of milk protease. Agric. Biol. Chem. nation or raw milk. Presented at the 46th Ann. Meeting, Interna­ 35:1465-1467. tional Assoc. Milk and Food Sanitarians, Glenwood Springs, 110. Langlois, B. E., H. E. Randolph, and N. M. Curme. 1966. Keep­ Colo. ing quality of dairy products obtained at retail outlets. I. Low- 133. Mikolajcik, E. M., and N. T. Simon. 1978. Heat resistant psyc­ milk, skim milk, and chocolate . J. Milk Food Tech­ hrotrophic bacteria in raw milk and their growth at 7°C. J. Food nol. 29:214-217. Prot. 41:93-95. 111. Langlois, B. E., and C. Sanhirum. 1977. Effect of various combi­ 134. Mikolajcik, E. M. 1979. Psychrotrophic bacteria and dairy prod­ nations of medium, diluent and incubation conditions on recovery uct quality. 1. Major organisms involved and defects produced. of bacteria from manufacturing grade and grade A raw milk. J. Cult. Dairy Prod. J. 14:6-10. Food Prot. 40:222-227. 135. Mikolajcik, E. M., and R. B. Brucker. 1981. Limulus amebocyte 112. Law, B. A. 1979. Enzymes of psychrotrophic bacteria and their lysate assay - its potential for detection of psychrotrophic spoilage effects on milk and milk products. J. Dairy Res. 46:573-588. organisms in food systems. J. Food Prot. 44:802 (Abstract). 113. Ledford, R. A., G. F. Senyk, W. F. Shipe, D. K. Bandler, and 136. Mikolajcik, E. M. 1983. GNB organisms test yields reliable re­ E. T. Wolff. 1981. Rates of change in bacterial counts of com­ sults. Dairy Field 166:38. mercial milks during storage. J. Dairy Sci. 64:46 (Abstract). 137. Mikolajcik, E. M., and R. B. Brucker. 1983. Limulus amebocyte 114. Ledford, R. A., G. F. Senyk, and E. Kotsides. 1983. Growth lysate assay - a rapid test for the assessment of raw and pas­ relationships of psychrotrophs Pseudomonas and Enterobacter iso­ teurized milk quality. Dairy Food Sanit. 3:129-131. lates in milk. J. Dairy Sci. 66:63 (Abstract). 138. Mistry, V. V., and F. V. Kosikowski. 1982. Effect of potassium 115. Ledford, R. A., G. F. Senyk, E. Kotsides, and E. T. Wolff. sorbate on the keeping quality and psychrotrophic bacteria of 1983. Preliminary incubation method for detection of coliforms milk. J. Dairy Sci. 65:63 (Abstract). in commercially processed milk samples. J. Dairy Sci. 66:1607- 139. Mistry, V. V., and F. V. Kosikowski. 1983. Application of 1610. dialysis to measure protease activity in milk. J. Dairy Sci. 66:687- 116. Ledford, R. A., G. F. Senyk, E. Kotsides, and E. T. Wolff. 693. 1983. Influence of growth of colifonns on flavor acceptability of 140. Mohamed, F. O., and R. Bassette. 1979. Quality and yield of commercially processed milk samples. J. Dairy Sci. 66:1611- cottage cheese influenced by psychrotrophic microorganisms in 1615. milk. J. Dairy Sci. 48:22-226. 117. Levin, G. V., J. R. Schrot, and W. C. Hess. 1975. Methodology 141. Naikai, S., H. K. Wilson, and E. O. Herreid. 1965. Assaying for application of adenosine triphosphate determination in waste­ sterile concentrated milk from native proteolytic enzymes. J. Dairy water treatment. Environ. Sci. Technol. 9:961-965. Sci. 48:754-757. 118. Levin, J., and F. B. Bang. 1968. Clottable protein in Limulus: 142. Nielsen, V. H. 1969. Prolonging cottage cheese life. Am. Dairy its localization and kinetics of its coagulation by endotoxin. Rev. 31:54-56. Thromb. Diath, Haemorrh. 19:186-197. 143. O'Connor, F. 1979. An impedance method for the determination 119. Levin, J., T. E. Poore, N. S. Young, S. Margolis, N. P. Zeuber, of bacteriological quality of raw milk. Ir. J. Food Sci. Technol. A. S. Townes, and W. R. Bell. 1972. Gram-negative sepsis: de­ 3:93-100. tection of endotoxemia with the Limulus test. Ann. Int. Med. 144. Oehlrich, H. K., and R. C. McKellar. 1983. Evaluation of an 76:1-7. 18°C/45-hour plate count technique for the enumeration of psyc­ 120. Lightbody, L. 1966. The effect of low temperature storage of pas­ hrotrophic bacteria in raw and pasteurized milk. J. Food Prot. teurized milk on bacterial counts using penicillin agar. Austral. 46:528-529.

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 752 BISHOP AND WHITE

145. Ohki, M. S., S. Nakamura, T. Morita, and S. Iwanaga. 1980. ence of time and temperature of incubation on the plate count A new endotoxin sensitive factor associated wtih hemolymph of milk. J. Milk Food Technol. 37:209-212. coagulation system or horseshoe crab (Limulidae) FEBS Lett. 171. Rowley, D. B., P. Vandemark, D. Johnson, and E. Shattuck. 120:217-220. 1979. Resuscitation of stressed fecal coliforms and their sub­ 146. Oliveria, J. S., and C. E. Parmalee. 1976. Rapid enumeration sequent detection by radiometric and impedance techniques. J. of psychrotrophic bacteria in raw milk and pasteurized milk. J. Food Prot. 42:335-341. Milk Food Technol. 39:269-272. 172. Ryan, J. J. 1982. Assessment of raw milk microbiological quality 147. Olson, H. C. 1964. How to improve shelf-life. Proc. 57th Ann. by adenosine-5'-triphosphate content and preliminary incubation Conv. Milk Ind. Foundation, Chicago. bacteria counts. Ph.D. dissertation, Louisiana State Univ. 148. O'Toole, D. K. 1983. Methods for the direct and indirect assess­ 173. Ryan, J. J., R. H. Gough, and C. H. White. 1984. Preliminary ment of the bacterial content of milk. J. Appl. Bacteriol. 55:187- incubation counts as an index of raw milk microbiological quality 201. during storage. J. Food Prot. 47:206-208. 149. Overcast, W. W. 1968. Psychrophilic microorganisms and keep­ 174. Sadovski, A. Y., S. Gordin, and I. Foreman. 1980. Psychrot­ ing quality of milk and its products. J. Dairy Sci. 51:1336-1338. rophic growth of microorganisms in a cultured milk product. J. 150. Parsons, L. B., and W. S. Sturges. 1926. Conductivity as applied Food Prot. 43:765-768. to studies of bacterial metabolism. J. Bacteriol. 12:267-272. 175. Samples, D. R., R. L. Richter, and C. W. Dill. 1983. Measuring Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 151. Patel, G. B., and G. Blankenagel. 1972. Bacterial counts of raw proteolysis in slurries: A comparison of the Hull milk and flavor of the milk after pasteurization and storage. J. and TNBS procedures. J. Dairy Sci. 67:60-63. Milk Food Technol. 35:203-206. 176. Sandvik, O. 1962. Studies on casein precipitating enzymes of 152. Pate, T. R., F. M. Bartlett, and J. Hamid. 1983. Extracellular aerobic and facultatively anaerobic bacteria. From the Dept. of heat-resistant proteases of psychrotrophic pseudomonads. J. Food Food Hygiene and Microbiol., Vet. College of Norway, Oslo. Prot. 46:90-94. 177. Seiter, J. A., and J. M. Jay. 1980. Comparison of direct serial 153. Pennington, M. E., J. S. Hepburn, E. Q. St. John, E. Witmer, dilution and most-probably-number methods for determining en­ M. O. Stafford, and J. I. Bunell. 1913. Bacterial and enzymatic dotoxins in meats by the Limulus amoebocyte lysate test. Appl. changes in milk and cream at 0°C. J. Biol. Chem. 16:331-368. Environ. Microbiol. 40:177-178. 154. Peterkin, P. I., and A. N. Sharpe. 1980. Membrane filtration of 178. Senyk, G. F., W. F. Shipe, R. A. Ledford, and P. Wienberg. dairy products for microbiological analysis. Appl. Environ. Micro­ 1984. Measurement of pyruvate content in milk ultrafiltrates for biol. 39:138-143. evaluation of bacterial quality. J. Dairy Sci. 67:1660-1665. 155. Peterson, A. C, and M. F. Gunderson. 1960. Some characteris­ 179. Shahani, K. M. 1966. Milk enzymes: Their role and significance. tics of proteolytic enzymes from Pseudomonas fluorescens. Appl. J. Dairy Sci. 49:907-920. Microbiol. 8:98-104. 180. Sharpe, A. N., M. N. Woodrow, and A. K. Jackson. 1970. 156. Pettipher, G. L., R. Mansell, C. H. McKinnon, and C. M. Adenosine triphosphate (ATP) levels in foods contaminated by Cousins. 1980. Rapid membrane filtration-epifluorescent micro­ bacteria. J. Appl. Bacteriol. 33:758-767. scopy technique for direct enumeration of bacteria in raw milk. 181. Sherman, J. M., G. M. Cameron, and J. C. White. 1941. The Appl. Environ. Microbiol. 39:423-429. bacteriological spoilage of milk held near the freezing point. J. 157. Phillips, J. D., M. W. Griffiths, and D. D. Muir. 1984. Preincu­ Dairy Sci. 24:526-527. bation test to rapidly identify post-pasteurization contamination in 182. Shipe, W. F., G. F. Senyk, R. A. Ledford, D. K. Bandler, and milk and single cream. J. Food Prot. 47:391-393. E. T. Wolff. 1981. Changes in flavor, acid degree value, tyrosine 158. Previte, J. J. 1972. Radiometric detection of some food-borne value, and pyruvate content of commercial milks during storage. bacteria. Appl. Microbiol. 24:535-539. J. Dairy Sci. 64:46 (Abstract). 159. Punch, J. D., J. C. Olson, and E. L. Thomas. 1966. Psyc­ 183. Shipe, W. F., and H. Hsu. 1982. Measurement of fluorescent hrophilic bacteria. III. Population levels associated with flavor or compound production during storage of market milk samples. J. physical changes in milk. J. Dairy Sci. 48:1179-1183. Dairy Sci. 65:46 (Abstract). 160. Randolph, H. E., T. R. Freeman, and R. W. Peterson. 1965. 184. Sing, E. L., P. R. Elliker, L. J. Christensen, and W. E. Sandine. Keeping quality of market milk obtained at retail outlets and at 1967. Effective testing procedures for evaluating plant sanitation. processing plants. J. Milk Food Technol. 28:92-96. J. Milk Food Technol. 30:103-111. 161. Randolph, H. E., B. K. Chakroborty, O. Hampton, and D. L. 185. Smillie, D. M., M. J. Orr, and R. M. McLarty. 1958. Bulk milk Bogard. 1973. Microbial counts of individual producer and com­ collection in Wigtownshire. Bacterial counts at 6 and 30 hours. mingled grade A raw milk. J. Milk Food Technol. 36:146-151. Dairy Ind. 23:560-562. 162. Randolph, H. E., B. K. Chakraborty, O. Hampton, and D. L. 186. Smith, K. L., J. J. Jezeski, and D. W. Wilkins. 1982. The Bogart. 1973. Effect of plate incubation temperature on bacterial Moseley test as a predictor of the shelf-life of pasteurized milk. counts of grade A raw milk. J. Milk Food Technol. 36:152-154. J. Dairy Sci. 65:221 (Abstract). 163. Reimerdes, E. H., and E. Herlitz. 1979. The formation of y-ca- 187. Smith, T. L., and L. D. Witter. 1979. Evaluation of inhibitors seins during cooling of raw milk. J. Dairy Res. 46:219-221. for rapid enumeration of psychrotrophic bacteria. J. Food Prot. 164. Reimerdes, E. H- 1983. New aspects of naturally occurring pro­ 42:158-160. teases in bovine milk. J. Dairy Sci. 66:1591-1600. 188. Speck, M. L. (ed.). 1976. Compendium of methods for the micro­ 165. Reinbold, G. W., C. K. Johns, and W. S. Clark. 1969. Modifica­ biological examination of foods. Am. Public Health Assoc., tion of the preliminary incubation treatment for raw milk samples. Washington, DC. J. Milk Food Technol. 32:42-43. 189. Speck, J. L., and D. M. Adams. 1976. Heat resistant proteolytic 166. Reinbold, G. W. 1971. Bacteriological testing of raw milk and enzymes from bacterial sources. J. Dairy Sci. 59:786-789. dairy products. J. Milk Food Technol. 34:613-619. 190. Stanier, R. Y., E. A. Adelberg, and J. Ingraham. 1976. The 167. Reinhold, R. B., and J. Fine. 1971. A technique for quantitative microbial world, 4th ed. Englewood Cliffs: Prentice Hall. measurement of endotoxin in human plasma. Proc. Soc. Exp. 191. Stone, W. K., and D. M. Naff. 1967. Increases in soluble nitro­ Biol. Med. 137:334-340. gen and bitter flavor development in cottage cheese. J. Dairy Sci. 168. Richter, R. 1983. Determining raw milk quality. Dairy Field 50:1497-1500. 166:70-71. 192. Storrs, A. B., and M. E. Hull. 1956. Proteolytic enzymes in 169. Rodriques, U. M., and G. L. Pettipher. 1984. Use of the direct human and cow's milk. J. Dairy Sci. 39:1097-1103. epifluorescent filter technique for predicting the keeping quality 193. Sudi, J. 1982. Limulus assay of bacterial lipopolysaccharides in of pasteurized milk within 24 hours. J. Appl. Bact. 57:125-130. UHT milk. Kieler Milchwirtschaftliche Forschungsberichte 170. Roughley, F. R., C. K. Johns, and K. L. Smith. 1974. The influ­ 34:141-146.

JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986 ASSESSING SHELF-LIFE AND QUALITY OF DAIRY FOODS 753

194. Sullivan, J. D., and S. W. Watson. 1975. Purification and proper- 209. Webb, B. H., and A. J. Johnson. 1965. Fundamentals of dairy ties of the clotting enzyme from Limulus lysate. Biochem. chemistry. AVI Publishing Company, Inc., Westport, CT. Biophys. Res. Commun. 66:848-855. 210. White, C. H., and R. T. Marshall. 1973. Heat stable protease 195. Sweadner, K. J., M. Forte, and L. L. Nelson. 1977. Filtration from Pseudomonas fluorescens P26 degrades ultra-high tempera­ removal of endotoxin (pyrogens) in solution in different states of ture pasteurized milk. J. Dairy Sci. 56:624 (Abstract). aggregation. Appl. Environ. Microbiol. 34:382-385. 211. White, C. H., and R. T. Marshall. 1973. Reduction of shelf-life 196. Tai, J. Y., and T. Liu. 1977. Studies on Limulus amebocyte ly­ of dairy products by a heat-stable protease from Pseudomonas sate - isolation of pro-clotting enzyme. J. Biol. Chem. 252:2178- fluorescens P26. J. Dairy Sci. 56:849-853. 2181. 212. White, C. H., W. T. Gillis, D. L. Simmler, M. K. Galal, J. 197. Thomas, S. B., and B. F. Thomas. 1973. Psychrotrophic bacteria R. Walsh, and J. T. Adams. 1978. Evaluation of raw milk quality in refrigerated bulk-collected raw milk. Dairy Ind. 38:11-12. tests. J. Food Prot. 41:356-360. 198. Thomas, T. D., and O. E. Mills. 1981. Proteolytic enzymes of 213. White, C. H., and M. Bulthaus. 1979. Defects caused by adding starter bacteria. Neth. Milk Dairy J. 35:255-258. protease-producing psychrotrophs to raw milk. J. Dairy Sci. 62:46 199. Tomich, J. M., and W. N. Eigel. 1976. Evidence for the presence (Abstract). of plasminogen and plasmin in bovine milk. Fed. Proc. 35:1487- 214. White, C. H., M. Bulthaus, and R. T. Marshall. 1980. Changes 1489. in milk caused by addition of protease-producing psychrotrophs 200. Trobaugh, W. S. 1982. Quality assurance through the preliminary to raw milk. J. Dairy Sci. 63:56 (Abstract). Downloaded from http://meridian.allenpress.com/jfp/article-pdf/49/9/739/1657494/0362-028x-49_9_739.pdf by guest on 25 September 2021 incubation count. Dairy Food Sanit. 2:184-189. 215. White, C. H. 1982. A bacteriological approach to testing the raw 201. Tsuji, K., and K. A. Steindler. 1983. Use of magnesium to in­ milk supply. Dairy Rec. 83:53. crease sensitivity of Limulus amoebocyte lysate for detection of 216. White, C. H. 1982. Bacteriological testing a useful tool in predict­ endotoxin. Appl. Environ. Microbiol. 45:1342-1350. ing fluid milk shelf-life. Dairy Rec. 83:112-113. 202. United States Department of Health, , and Welfare. 217. White, C. H., and R. D. Little. 1983. Comparison of tests to 1978. Grade A pasteurized milk ordinance. U.S. Dept. HEW, rapidly predict shelf-life of fluid milk products. J. Dairy Sci. Washington, DC. 66:87 (Abstract). 203. Vedamuthu, E. R. 1984. Testing for psychrotrophic bacteria. 218. White, C. H., and R. D. Little. 1983. Evaluation of tests to pre­ Technics/Topics 14:2. dict the shelf-life of fluid milk products. J. Dairy Sci. 66:268 204. Waes, G. M., and R. G. Bossuyt. 1982. Usefulness of the benzal- (Abstract). kon - crystal violet - ATP method for predicting the keeping qual­ 219. Witter, L. D. 1961. Psychrophilic bacteria - A review. J. Dairy ity of pasteurized milk. J. Food Prot. 45:928-931. Sci. 44:983-1001. 205. Waes, G. M., and R. G. Bossuyt. 1984. Impedance measure­ 220. Wolstrup, J., and K. Jensen. 1976. Adenosine triphosphate in the ments to detect bacteriophage problems in cheddar cheesemaking. bovine rumen during maximum nutrient supply and starvation. J. J. Food Prot. 47:349-351. Appl. Bacteriol. 41:243-259. 206. Warner, R. C, and E. Polis. 1945. On the presence of a proteoly­ 221. Wood, J. M., J. H. Lach, and B. Jarvis. 1978. The evaluation tic enzyme in casein. J. Am. Chem. Soc. 67:529-532. of bacterial populations in foods. Leatherhead Food R. A. Re­ 207. Washam, C. J., H. C. Olson, and E. R. Vedamuthu. 1977. Heat- search Report No. 289. resistant psychrotrophic bacteria isolated from pasteurized milk. J. 222. Wood, J. M., and C. J. Stannard. 1981. The estimation of viable Food Prot. 40:101-108. microorganisms in raw meat by measurement of microbial ATP. 208. Watrous, G. H., S. E. Barnard, and W. W. Coleman. 1971. A Presented at 3rd Intern. Symp. Rapid Methods and Automation survey of the actual and potential bacterial keeping quality of pas­ in Microbiol., Washington, DC. teurized milk from 50 Pennsylvania dairy plants. J. Milk Food 223. Zall, R. R., J. H. Chen, and S. C. Murphy. 1982. Estimating Technol. 34:145-149. the number of psychrotrophs in milk using the direct microscopic method. Cult. Dairy Prod. J. 17:24-38.

SANI-MATIC "Your Total Sanitation Equipment Headquarters • REDUCE MANUAL LABOR • SHORTEN WASH TIME • SAVE CHEMICALS/WATER/STEAM • IMPROVE CLEANLINESS Jet Recirculation Parts & Fittings Washers Model 119-1SR Please circle No. 119 on your Exclusive "Rifle-Like" Jets Reader Service Card provide high volume recir­ 3 Tank CIP System culating cross-tank jet scrub­ Please circle No. 121 on your Push Pull Washers bing of parts and fittings. Reader Service Card Cleans pipes, hoses & cylin­ CIP Systems drical shapes. Save energy and detergent with the Ecology 3 Tank CIP System. Shell and tube Model RWJ-150 "¥ heater reduces BTU require­ Push-Pull Cleaning Please circle No. 120 on your ments. Reader Service Card Unique "Push-Pull" cleaning Detergent re-use feature system effectively cleans DEC SANI-MATIC SYSTEMS reduces both water con­ pipe and hose on the inside sumption and BOD sewer­ VZ^ P.O. Box 8662, Madison, Wl 53708 and outside simultaneously. age requirements. Ph.: (800) 356-3300 or in Wl: (608) 222-2399 Please circle No. 119, 120, 121 on your Reader Service Card JOURNAL OF FOOD PROTECTION, VOL. 49, SEPTEMBER 1986