Listeria monocytogenes - Threat to a Safe Food Supply: A Review1 LAURA J. PEARSON and ELMER H. MARTH Department 01 Food Science and The Food Research Institute University 01 Wisconsin-Madison Madison 53706 ABSTRACT
Listeria monocytogenes can cause circling disease, encephalitis, meningitis, septicemia, and mastitis in dairy cattle. Shedding of the pathogen from the udder or contamination from the environment can lead to presence of L. m01wcytogenes in raw milk. Surveys indicate the pathogen is in about 4% of US raw milks. Although HTST pasteurization commonly inactivates L. monocylogenes, evidence suggests that under unusual circumstances minimal survival is possible. The pathogen grows well in liquid dairy products at 4 to 35·C and achieves higher populations in chocolate than in unflavored milks. When present in cheese milk, growth of L. monocylOgenes may be retarded but not stopped by lactic starter cultures. The pathogen is concenlrated in the curd with only a small fraction of ceUs in milk appearing in whey. Once in curd, the behavior of the pathogen ranges from growth (feta cheese making) to death of most but not all cells (cottage cheese making). During ripening of cheese, the numbers of L. monocylogenes decrease gradually (as in Cheddar or Colby cheese), decrease precipitously early during ripening, and then stabilize (as in blue cheese) or increase markedly (as in Camembert cheese). Consumption of foods containing L. monocylOgenes can lead to listeriosis in susceptible humans (adults with a compromised immune system), pregnant women, and infants). In large outbreaks of human listeriosis, mortality rates of ca. 30% are common.
(Key words: listeriosis, foodbome illness, dairy foods) INTRODUCTION
Listeria monocytogenes has been called an "emerging foodbome pathogen" because only recently have we recognized that it can be transmitted in food. After several foodbome outbreaks and associated deaths caused by this pathogen in the United States and other countries, many believe L. monocytogenes has indeed emerged! The goals of this paper are to: 1) describe the organism and the disease it causes, 2) highlight some of the characteristics that enable L. monocytogenes to survive and grow in dairy products and some other foods, 3) provide proof that this pathogen is a threat to a safe food supply, and 4) describe some procedures to detect the pathogen in food, avoid contamination of food, and inactivate the pathogen in food and in food factories. GENERAL CHARACTERISTICS Morphology
Listeria monocytogenes is a gram-positive, acapsular, asporogenous bacterium that was first described in 1926. The diphtheroid-like rod measures 1.0 to 2.0 I!ffi by .5 I!ffi. Listeriae are motile, predominantly by means of perittichous flagella, and exhibit characteristic "turnbling motility" when grown at 20 to 25·C (51). Colonies have a characteristic blue-green appearance when viewed by the Henry method of oblique lighting (56). On blood agar, colonies produce a narrow zone of beta-hemolysis caused by a soluble hemotoxin (98). Growth Requirements
Received Seplember II. 1989. Accepted November 2. 1989. lInvited paper. 1990 J Dairy Sci 73:912-928
Listeriae are not fastidious. They can survive in feces, milk, soil, water, silage, and on plants. 912
USTERJA AND DAIRY FOODS
913
Furthennore, bacteria in the genus Listeria can multiply in high salt (up to 10% sodium chloride) or bile concentrations (98). Listeria monocytogenes grows best at neutral to slightly alkaline pH but will grow from pH 5.0 to 9.6 (51, 98). Listeria monocytogenes grows from 3 to 45°C (optimum for growth, 30 to 37°C). Listeriae are aerobic to microaerophilic; growth is enhanced in an atmosphere with reduced oxygen and 5 to 10% C~. Listeria monocytogenes requires biotin, riboflavin, thiamine, thioctic acid, and some amino acids for growth.
merous vertebrates can be infected, including ruminants (cattle, sheep, and goats), pigs, poultry (61), foxes, gerbils, rabbits, and guinea pigs (81). The pathogen can cause circling disease (61, 106), encephalitis, meningitis, and septicemia (61). Listeria monocytogenes also can cause bovine mastitis and then be excreted in milk (48), and it is generally accepted that listeriosis can be transmitted to humans by contaminated milk (98). Vizcaino and Garcia (107) found that 13 of 36 sero-positive otherwise healthy cows excreted Listeria in the milk.
Biochemical Properties
Transmission
Listeria monocytogenes possesses a split noncyclic citric acid pathway with an oxidative and reductive portion that is important for biosynthesis of cell constituents. Energy is supplied by glycolysis (66). Listeria monocytogenes has a heme-containing catalase that is inhibited by KCN and NaN3. Glucose is fermented and oxidized by the bacterium with production of acid, but no gas. The end product of fermentation is mainly lactic acid. Gelatin, casein, and milk are not hydrolyzed (98).
Listeria monocytogenes can infect humans and other animals by oral, ocular, cutaneous, respiratory, or urogenital routes (87). Modes of transmission from animals to humans include ingestion of contaminated food and contact with infected animals. Milk can be contaminated by environmental sources, such as feces, or the bacterium can be shed in the milk itself (58).
Serotyping
Antigenic types are distinguished by somatic (0) and flagellar (H) antigens. Based on these
antigens, strains of L. monocytogenes are subdivided into serotypes: 1l2a, 1/2b, 1/2c, 3a, 3b, 3c, 4a, 4ab, 4b, 4c, 4d, 4e, 7 (79), 4f, 4g, 5, and 6 (87). Of 722 human cases of listeriosis in the US, 59% were caused by serotype 4b and 18% by serotype l/2a. A bacteriophage typing system, based on lytic properties of a set of phages, also has been developed. It is reproducible and discriminatory; however, only 64% of the serotypes can be typed by the bacteriophages (79). USTERIOSIS Zoonosis
Listeriosis, the disease caused by L. monocytogenes, is generally accepted as a zoonosis-a disease or infection naturally transmitted between vertebrate animals and humans (87). Nu-
Virulence
Virulence of L. monocytogenes may be related to two toxins: hemolysin and cytotonic toxin. The hemolysin is a cytolysin able to lyse tissue and red blood cells. The cytotonic toxin stimulates cyclic AMP production similar to cholera toxin (77). Additionally, almost all strains have phosphatase activity, and this may playa role in virulence (82). Upon infection, L. monocytogenes is taken up by white blood cells, where it survives and multiplies (10). All serotypes of L. monocytogenes have the ability to provoke monocytosis (98). Listeria strains have low virulence in comparison to other pathogenic microorganisms, based on animal studies (87). Most healthy humans can fight off infections with only minor symptoms (stomach upset or "flu") or no symptoms (13). The incidence of listeriosis is highest in newborns, pregnant women, elderly persons, and immunosuppressed individuals (10). Immunosuppression can occur in patients receiving corticosteroids or cancer chemotherapy (13), or in individuals with preexisting diseases such as cirrhosis, AIDS, or malignancies (10). Journal of Dairy Science Vol. 73. No.4, 1990
914
PEARSON AND MARTH
Disease Manifestations In Humans
Disorders associated with L. monocytogenes include meningoencephalitis, influenza-like, low grade septicemia in the prenatal period, infectious mononucleosis-like syndrome, septicemia in adults, pneumonia, endocarditis. localized abscesses, papular or pustular cutaneous lesions, conjunctivitis. urethritis, and recurring abortion. Listeria monocytogenes also may cause cerebral damage. mental retardation, and mononucleosis in some cases (51). Listeriosis usually occurs sporadically in the human population, but epidemics develop occasionally (61). Earlier literature stated the mortality rate to be 90% (61); currently it is estimated at 30% (87). The high mortality rate results from underlying immune deficiencies in susceptible individuals. Treatment of listeriosis involves drug therapy; ampicillin, erythromycin. and rifampicin are the antibiotics of choice (87). A tetracycline and trimethoprim mixture showed in vivo synergy (67). Estradiol-17~ at concentrations >5 J.Lg/ml completely inhibits the hemolysin produced by L. monocytogenes (54). Chloramphenicol and streptomycin have relatively little effect on the pathogen (20). ENVIRONIENTAL RESISTANCE Psychrotrophlc PropertiM
Listeria monocytogenes is generally considered to be a mesophile (optimum growth at 30 to 37"C); however, it also has psychrotrophic properties that enable it to grow at refrigeration temperatures. Using an Arrhenius plot, the minimum temperature for growth of the pathogen was extrapolated to be 3·C (109). The duration of the lag phase increases with decreasing temperature (109); at 4·C the lag phase lasts 5 to 10 d (7).
Mechanical refrigeration maintains the temperature at 4 to 5"C to preserve many foods including dairy products and hopefully prevent growth of pathogenic organisms. Since the 196Os, several foodborne pathogens have been recognized as able to grow at ca. 5"C, including Clostridium botulinum type E. Yersinia enlerocolitica. Aeromonas hydrophila. and L. monocytogenes (83). Additionally. slight deviation upward in temperature from that desired can greatly increase the growth rate of L. monocylOJoumal of Dairy Science Vol. 13. No.4. 1990
genes (at 13·C the generation time was 5.0 h for strain V7 in whole milk. whereas at 4·C it was ca. 35 h) (89). The ability to grow at refrigeration temperatures. while other competing microflora cannot, selects for L. monocytogenes; this may play a role in foodbome outbreaks (83). Listeria monocytogenes is reportedly more virulent when grown at low rather than at high temperatures (51). and virulent strains (serotype 4b) exhibited enhanced psychrotrophic growth in whole milk (29). Composition of liquid milk products affects the psychrotrophic growth of L. monocytogenes. Whipping cream supported development of the lowest population when compared with whole. skim. and chocolate milk, perhaps because of the higher fat water ratio (89). Two strains of serotype 4b grew better at 4 and lO·C in whole milk than in skim milk and reconstituted nonfat milk solids (11 %) (29). Growth of strain Scott A (also serotype 4b) in whole versus skim milk was insignificantly different at 4, 8, and l3·C (89). Conversely, Pseudomonas spp. grew slightly faster in skim milk (8.7% total solids) that was fortified to contain 10 or 12% total solids than in unfortified skim milk (63). Competition with other psycmotrophic bacteria (especially Pseudomonas spp.) has varied effects on growth of L. monocytogenes in milk. Studies with Pseudomonas fragi and L. monocytogenes strain Scott A co-inoculations resulted in neither inhibition nor stimulation of either organism (73). Other studies found that Pseudomonas jluorescens P26 and B52 moderately inhibited L. monocytogenes strains Scott A, California (CA) and V7 at 7 and 13·C (41). When milks (whole, skim. and reconstituted nonfat milk solids) were preincubated for 3 d with P. jluorescens P26, TI5. and 852 and P. fragi, the generation times of L. monocytogenes strain Scott A at 10·C were shorter (from 10.0 h without Pseudomonas spp. to 7.0 to 8.5 h with various Pseudomonas spp.). The proteolytic activity of the Pseudomonas spp. and resultant availability of nutrients to the pathogen may have been responsible for this stimulation (73). The ability of L. monocytogenes to grow at refrigeration temperatures has created unique sanitation problems. In a study of simulated milk cooling systems by Petran and Zottola
USTERIA AND DAIRY FOODS
(85), it was found that sweet water coolant systems (refrigerated potable water) could be conductive to growth of L. monocytogenes if small holes or cracks in equipment allowed leakage of some milk into the coolant. Usteria monocylogenes was inhibited in concentrations of propylene glycol greater than 12%, so most dairies are not threatened. since 30% glycol is used. Another mode of transmission of the pathogen could be on the exterior of waxed cardboard or plastic milk containers, according to Stanfield et al. (101). After 14 d, nonheatshocked L. monocylogenes serotypes la, Ib, and 4b could be isolated from the exterior of previously inoculated plastic containers held at -8 to 6.6·C. Heat-shocked organisms could not be isolated after 4 d. Thermal ResIstance
It is questionable whether L. monocylogenes can survive the pasteurization process as defined by the Grade A pasteurized milk ordinance (43). However, evidence on this issue is conflicting, partially because of 1) the methods used to determine heat resistance of L. rnonocytogenes and 2) the physiological state of the bacterium during heating. Strains of L. monocytogenes differ in their heat resistance. Strains most commonly used for testing include those isolated from the Massachusetts listeriosis outbreak (described later in this paper)-Scon A (serotype 4b), V7 (serotype la), and V37 (serotype 4b) (l4)-and those most commonly isolated from raw milk-ATCC 19115 (serotype 4b) and F5069 (serotype 4b) (29). Of these, strain Scott A is the most heat resistant (14). The holding technique for pasteurization (61.7·C for 35 min) used by Beams and Girard (11) failed to free milk from viable cells of L. monocylOgenes when the initial inoculum of the pathogen was 5 x 1()4/ml or greater. Using sealed borosilicate glass tubes, Bradshaw et al. (16) showed that 15 10g1O cells of L. monocytogenes would be inactivated in 15 s at 71.7·C. Donnelly et al. (30) proposed that differences between the results in these two repons were caused by condensate and bacteria collecting in the cap of the test tube or by cells coating the walls of the test tube above the level of the
915
heating menstruum in the method used by Beams and Girard. Following heating of samples containing L. monocytogenes, researchers used immediate direct plating techniques (11, 15, 14, 19, 29, 30) or direct plating and enrichment techniques (32, 46) to recover survivors. Using only direct plating, all researchers except Beams and Girard (11) determined present pasteurization guidelines to be adequate to eliminate L. monocytogenes. Garayzabal et al. (46) isolated L. monocylogenes from pasteurized samples with cold enrichment but never with plating immediately following thermal treatment. Enrichment procedures allow use of a larger sample than does direct plating and may allow repair of thermally injured cells (32). As a facultative intracellular parasite, L. monocytogenes can exist within polymorphonucleocyte leukocytes (pMNL) and be shed in milk (32). Freely suspended cells of L. monocytogenes strain Scott A in whole milk were destroyed at 71.7·C for 15 s (14). Intracellular L. monocytogenes strain Scott A survived HTST pasteurization at 72.2·C for 16.4 s; however, this study used enrichment techniques to recover survivors (32). The D71.7·C value (time at 71.7·C needed for population to decrease by one order of magnitude) for L. monocytogenes strain Scott A inside phagocytes was 1.9 s with a ZD value (change in temperature needed to change D-value by one order of magnitude) of 6·C, using only direct plating techniques (18). Doyle et al. (32) also noted that PMNL in milk degraded after milk was held for 3 to 4 d at 4·C, and this may increase the susceptibility of L. monocytogenes to inactivation by pasteurization because the cells become freely suspended. Bradshaw et al. (15) confirmed that pasteurization guidelines for whole and skim milk, heavy cream, and ice cream mix are adequate to ensure inactivation of L. monocytogenes (D71.7·C values ranged from .9 to 2.7 s). The study also showed that the thermal resistance of L. monocytogenes strain Scott A increased when the products had been previously sterilized. This may have implications when already processed products are pasteurized. The heat resistance of five strains of L. monocytogenes was not affected by prior growth in skim, whole, or reconstituted nonfat milk (29). Journal of Dairy Science Vol. 73. No.4, 1990
916
PEARSON AND MARTH
EVIDENCE FOR A FOODBORNE THREAT
Outbreaks of Foodbome liS18riosis
Four major recent outbreaks of listeriosis, two in the US and one each in Canada and Switzerland, caused by ingestion of contaminated food confmn that L. monocytogenes can be present in food and can cause illness when such food is consumed by susceptible persons. In the Maritime Provinces of Canada, 34 cases of perinatal listeriosis and seven of adult listeriosis occurred between March 1 and September I, 1981. A food preference survey identified coleslaw as the vehicle of transmission. Additionally, L. monocytogenes serotype 4b was isolated from coleslaw in a patient's refrigerator and from patients. The contamination probably resulted from sheep manure (from sheep with documented cases of circling disease) used to fertilize the cabbage field. Cabbage is often held in prolonged cold storage before distribution, and this may have allowed proliferation of the psychrotrophic L. monocytogenes while the number of competing microorganisms decreased (97). An epidemiological investigation by the Centers for Disease Control (CDC) showed pasteurized whole or 2% milk of a specific brand to be the vehicle of infection in a listeriosis outbreak in Massachusetts. Between June 30 and August 30, 1983, 49 patients (7 fetuses or infants and 42 adults) acquired listeriosis. All of the adults were immunosuppressed because of preexisting illness. Mortality was 29%. In addition to the two case-control studies, investigators also used the dose-response effect, the protective effect of skim milk, comparison with cases of listeriosis in Connecticut, and phage typing to assess the validity of their findings. The implicated milk was pasteurized at a single factory, and proper pasteurization guidelines were followed. Twelve percent of raw milk samples collected after the outbreak contained the organism. Veterinarians that year had diagnosed four cases of bovine listeriosis (encephalitis) on farms supplying the dairy. Postpasteurization contamination was thought unlikely for two reasons: skim milk was not contaminated with the pathogen, even though it was pasteurized with the same equipment on the same day, and no source of the organism could be found in the factory (42). During the last several years Journal of Dairy Science Vol. 73. No.4. 1990
questions have been raised about the adequacy of the epidemiologic investigation, and the real cause of the outbreak may have been something other than milk. In another outbreak, Mexican-style fresh cheese was determined to be the cause of at least 86 cases of listeriosis from January 1 to June 14, 1985 in California. Fifty-eight of the cases involved mother-infant pairs. Twentynine deaths occurred (33.7% mortality). Using case-control studies, consumption of cheese made in a single factory in the Los Angeles area was determined to be the cause of listeriosis. Cheese samples with different ex.piration dates contained L. monocytogenes serotype 4b, suggesting a continuing contamination problem in the factory (62). Some cheese was phosphatase-positive, indicative of improper pasteurization or no pasteurization of cheese milk. However, application of the phosphatase test to cheese to determine whether or not it was made from pasteurized milk has been questioned. Producers of the cheese may have mixed raw and pasteurized milk or postpasteurization contamination may have been the problem (99). The food-related outbreak of listeriosis in Switzerland was associated with consumption of Vacherin Mont d'Or cheese. This is a soft surface-ripened (bacterial) cheese. Continued testing of products for L. monocylogenes by regulatory agencies has resulted in recalls of contaminated food. Several brands of Brie cheese, produced in Sorcy, France were contaminated with L. monocytogenes, according to a Food and Drug Administration (FDA) warning on February 12, 1986. The factory making the cheese operated under the current imported soft-ripened cheese agreement that requires pasteurization of the milk before cheese manufacture (4). A month later. more brands of Brie cheese were recalled because of contamination (2). In the US, a major producer of ice cream bars recalled 25,000 cases of product because of contamination with L. monocytogenes. There were no reports of illness, probably because the frozen state of the product did not allow proliferation of the bacterium to numbers sufficient to cause infection of consumers (3). Another major manufacturer of ice cream recalled a million gallons of product from 11 Midwestern states when L. monocytogenes was found in the product (106). Additional recalls included those of Liederkranz cheese in 1985
USTERIA AND DAIRY FOODS
and the soft Mexican-style cheese in Arizona in 1986 (99). Recalls of products are costly and obviously detrimental to the reputation of the company. Occurrence In Dairy Products
The Massachusetts and California outbreaks pointed to milkborne L. monocytogenes as a serious public health problem. Several studies have been done to determine the incidence of L. monocytogenes in raw milk. Lovett et al. (72) tested raw milk from farm bulk tanks in three areas of the United States-California, Massachusetts, and a Tri-State area (Ohio, Kentucky, and Indiana). The overall incidence was 4.2% (milks from California had 0% and from Massachusetts had 7%). Seasonal variation occurred in Massachusetts and in the Tri-State area; incidence was lowest in hot weather months and peaked in the cold months. Most of the TriState isolates were serotype 4b. and the Massachusetts isolates were mainly serotype I (72). Other studies support these results. Liewen and Plautz (69) detected L. monocytogenes is 4% of raw milk samples in Nebraska The seasonal variation also was confirmed in this study. Raw milk samples collected throughout the year from a factory in Madrid, Spain contained L. monocytogenes at percentages ranging from 14.3 to 75 (44). Although L. monocytogenes was not detected, other species of Listeria were isolated from raw milk in New Zealand (103). In a survey in the Netherlands by Beckers et al. (12), L. monocytogenes was detected in 7 of 69 samples of imported soft cheese made from raw milk. Another study of blue-veined cheeses found the pathogens in 2 of 28 samples (23). Garayazabal et al. (45) found L. monocytogenes in 21.4% of pasteurized milk samples (3.2% fat milk treated at 78T for IS s) marketed by a Madrid processing plant.
917
zen bean, peas, pea pods, and spinach were examined and were found to be Listeria-free using a 7-d enrichment procedure. Another study did isolate L. monocytogenes from a head lettuce and also showed that the population of the pathogen increased during storage (102). Gitter (47) isolated L. monocytogenes serotypes 4 and 112 from 14.7% of "oven-ready" poultry from supermarkets. Various surveys have shown that L. monocytogenes may be present in S to SO% of fresh retail red meat and poultry (31). A study by the FDA found 26% of seafood samples to be positive for L. monocytogenes. Positive products included shrimp, crabmeat, lobster tail, langostinos, scallops, sunmi-based imitation seafoods (108), and cold smoked fish (S). Beh8vlor During Food Proceulng
To explore behavior of L. monocytogenes during food preparation or processing. researchers have monitored survival and growth of the pathogen in artificially inoculated food products. These studies show what foods support growth of L. monocytogenes and what food processing practices increase or decrease changes in high populations of the pathogen in the food. As discussed previously, milk and milk products have been implicated as sources of listeriosis. Rosenow and Marth (89) documented growth of four strains of L. monocytogenes (Scon A, V7. V37CE. and CA) in skim, whole, and chocolate milk and in whipping cream at temperatures ranging from 4 to 3SoC. Generation times ranged from 29.S h (strain V37CE) in chocolate milk) to 45.5 h (strain V7 in whipping cream) at 4°C. When the temperature was increased to 13T-a temperature conceivably encountered during storage and transport of food by consumers-generation times were shortened to 4.5 h (strain V7 in chocolate milk) or 6.8 h (strain CA in whole milk). All strains eventually attained populations of at Occurrence In Other Foods least 107 cfu/ml. Chocolate milk always supIn similar studies, various vegetables. beef, ported development of the highest populations poultry, and seafood were tested for presence of of L. monocytogenes strains at all temperatures L. monocytogenes. Petran et al. (86) tested fresh tested. Further investigation showed that both lettuce. potato peels. com husks, broccoli cane sugar and cocoa powder, present in the stems, cabbage outer leaves, carrot peels, cauli- chocolate milk, served to increase the maxiflower stems, mushroom stems, spinach, and mum populations attained by L. monocytogenes beet peels from a local supermarket. They did (88). The enhanced growth of the pathogen in not isolate L. monocytogenes. Additionally, fro- chocolate milk and the ability to grow during Joumal of Dairy Science Vol. 73. No.4, 1990
918
PEARSON AND MARTH
refrigeration emphasize the need to avoid postpasteurization contamination of this product during the handling it receives. Concentrated (30% solids) and unconcentrated skim milks, inoculated with lOS to 1()6 L. monocytogenes strain Scott A or V7/ml, were spray-dried to a moisture content of 3.6 to 6.4%. The spray drying process itself resulted in a 1- to 1.5-loglO drop in the population of the pathogen initially/ml of milk when that milk was converted to the dry form. Storage of nonfat dry milk at 25"C resulted in a continued decrease in the population of the pathogen with strain V7 surviving the longest (12 wk) in nonfat dry milk made from concentrated skim milk (33). Cheeseborne outbreaks of listeriosis show cheese to be susceptible to contamination by L. monocytogenes. Two surface-ripened cheeses (Camembert and brick) were made from milk containing L. monocytogenes, and behavior of the pathogen was monitored through manufacture and storage. Slice, surface, and interior brick cheese (prepared by the washed curd method with 5 x 102 L. monocytogenes/ml in the milk) samples reached populations of 5.11 to 7.01 log10 L. monocytogenes strain Scott A or Ohio (OH)/g after 28 d of ripening. The surface samples had highest populations of the pathogen, perhaps because of their nearly neutral pH (6.5 to 6.9). After 154 d of ripening. numbers of the pathogen decreased by lo-fold. Strains CA and V7 failed to grow significantly, and cold enrichment was required to detect the low populations (74). Camembert cheese (prepared from milk containing 500 L. monocytogenes strain Scott A. V7, CA, or OH/ml) supported development of populations of 1 x 1()6 to 5 x 107 cfu of the pathogen/g after 65 d of ripening. When the pH of some samples of cheese was adjusted, growth of the pathogen was greatest at pH 6.1 rather than pH 4.6 or 7.4 (91). During manufacture of blue cheese by the Modified Iowa Method, artificially added L. monocytogenes mainly went with the curd rather than the whey. This also was true for other varieties of cheese. The population of L. monocytogenes (all strains) decreased in blue cheese during the first 50 d of ripening (at 9 to 12 C); however, the population of strain Scott A leveled off after the pH increased with mold growth during 50 to 120 d of ripening (84). D
Journal of Dairy Science Vol. 73, No.4, 1990
Studies show hard cheeses also can pose a health hazard if made from contaminated milk. Cheddar cheese (with normal values for moisture, fat-in-dry matter, salt, and pH) made from milk containing 500 L. monocytogenes strain CA, V7, or Scott Nml did not support growth of the pathogen during manufacture. Storage at 6 or 13 C resulted in a .1 to .8 log10 cfu/g increase in population initially, and then populations decreased. All samples contained a detectable number of L. monocytogenes after 60 d of ripening (90). In fact, viable cells of the pathogen were found in some 1-year-old cheese. Cheddar cheese can be used to make cold pack cheese food. Ryser and Marth (92) found that if inoculated cheese food was made without preservatives of acidifying agents, samples contained detectable populations (2.05 to 2.73 10g1O cfu/g) of strains Scott A, V7, CA, and OH after 182 d of storage at 4 C. Listeria monocytogenes was unable to grow in the coldpack cheese food with pH of 5.0 to 5.45. The pH of this product can be lowered by adding lactic or acetic acid. Sorbic acid and sodium propionate, common mold inhibitors that may be added to cheese food, are also somewhat antibacterial toward the pathogen (92). As with other varieties of cheese. L. monocytogenes is mainly entrapped in the curd during manufacture of Colby cheese. During storage at 4 C, populations of the pathogen remained constant for a few weeks and then decreased. Two factors increased the size of the surviving population of the pathogen: a higher rather than lower moisture content in cheese and a higher rather than lower initial inoculum level (111). Preparation of semi-hard cheese (Manchegotype) with milk containing 5.3 or 3.6 10glO L. monocytogenes 4b cfulml resulted in populations after 60 d of storage of 5.3 and 3.5 10glO cfulg respectively. Brining of the cheese was accompanied by a drop in population (25). Cottage cheese was manufactured according to the short-set procedure from pasteurized skim milk containing 104 to lOS L. monocytogenes strain Scott A or V7/ml. Cooking the curd decreased the population to -hemolysis. Recently the method has been modified so the amount of sample actually tested is the same as that used by the FDA. Beside enrichment procedures just described, there are several newly developed methods. Doyle and Schoeni (34) used an enrichment broth with tryptose, defibrinated sheep blood. glucose. K2HP04, polymyxin B. acriflavine-HCI, and nalidixic acid. The sample is homogenized in 100 ml of the broth and incubated with agitation at 37·C under microaerobic conditions for 24 h and then plated on MLA. Slade and Collins-Thompson (100) developed a two-stage enrichment technique. The primary enrichment is with nonselective tryptose broth for 3, 7. or 14 d at 4·C. The secondary enrichment broth consists of thiocyanate-nalidixic acid broth with acriflavine. Mter primary and secondary enrichment, samples are streaked on MLA. Several studies have compared the various enrichment procedures. Doyle and Schoeni (34) found that the FDA procedure. cold enrichment. and their selective enrichment procedure isolated the pathogen from 46% of soft, surface-ripened cheese. but no single method detected the organisms in all Listeria-positive samples. The USDA procedure was significantly worse at recovering L. monocytogenes from Brie cheese than the FDA procedure, even when the USDA broth was supplemented with D-glucose and phytone (105). However. Farber et al. (40) found that the USDA procedure was more efficient in isolating Listeria from ground meat than the FDA procedure. Monoclonal Antibody-Enzyme Immunoassay
Enzyme immunoassay (EIA), also known as enzyme-linked immunosorbent assay (ELISA), can be used to isolate and detect Listeria in food. Farber et al. (40) tested an EIA method with microtiter plates and on membrane filters using a monoclonal antibody directed against the B flagellar antigen of Listeria spp. Using the membrane system, naturally contaminated ground meats gave positive results as evidenced by purple spots on the filter. Two procedures developed for the microtiter plate system (coating plates with a flagella-containing supernatant fluid and a capture sandwich assay) were sucJournal of Dairy Science Vol. 73. No.4, 1990
922
PEARSON AND MARTH
cessful. Efforts to produce a monoclonal antibody to the listerial hemolysin continue in the hope that this will be a more specific immunological probe than the flagellar antigen. Mattingly et al. (75) produced genus-specific monoclonal antibodies against Listeria and used two of the antibodies to create an ELISA technique to detect the pathogen. Samples were grown in an enrichment broth (containing nalidixic acid. acriflavine. and cycloheximide) before the ELISA. All Listeria spp. were positive (as determined by optical density readings) except Listeria denitrificans (this species may be erroneously placed in the genus Listeria). The test was very sensitive (a 24-h-old pure culture from the enrichment broth was detected as positive through a 1:40.000 dilution). Over 20 common contaminating bacterial species yielded negative results. Deoxyribonucleic Acid Hybridization
Datta et al. (21) developed a DNA colony hybridization technique to detect L. monocytogenes. The extracellular ~-hemolysin from the pathogen is thought to be important in intracellular survival and virulence. and therefore. the hemolysin gene could be used to detect virulent L. monocytogenes. Crude DNA could not be used for hybridization because of the debris present in homogenized foods. The ~-hemoly sin gene fragment (500 base pairs) was cloned using recombinant DNA techniques. Lysing of the cells by microwaves was necessary to amplify the radiolabeled DNA probe. Only ~ hemolytic (cAMP test-positive) strains of L. monocytogenes were positive. This DNA probe was then used to detect hemolytic L. monocytogenes in naturally contaminated dairy products (22). The dairy products-raw milk. ricotta cheese. and semisoft cheeses-were plated and the resulting colonies were tested by DNA colony hybridization. Use of the gene probe saves time; however. there are some limitations. Only .1 ml can be plated. so for detection the contaminated product must have>10 cells/g. Additionally. not all L. monocytogenes strains react with the probe. Immunofluorescence and Flow Cytometry
Early fluorescent antibody techniques for identification of L. monocytogenes resulted in Journal of Dairy Science Vol. 73. No.4. 1990
many false-positive results because of crossreactivity with Streptococcus. Micrococcus. and Staphylococcus spp. Visual differentiation of these bacteria is difficult, so fluorescence microscopy is not very reliable. To increase the specificity of this technique. Donnelly and Baigent (27) employed an automated fluorescentantibody procedure through use of flow cytometry. Flow cytometry characterizes a bacterial cell population by morphology. nucleic acid content. and surface antigens as the cells in suspension pass through a laser beam. The species was labelled using polyvalent antibodies for cell surface antigens and propidium iodide for cellular DNA. Donnelly and Baigent (27) found that light scatter was not sufficient to distinguish between Listeria and non-Listeria contaminants. Sonicating the cells improved the fluorescent labelling procedure and consequently improved the resolution. In testing raw milk. use of a Listeria enrichment broth (containing nalidixic acid and acriflavine) before flow cytometry, decreased the number of competing microorganisms. Additionally. the Listeria enrichment broth allowed growth of the pathogen (from 8.5 x 1(jJ to 1.7 x 1()6 cfu/ml within 24 h at 3TC); this is necessary since resolution is not possible at
Listeria monocytogenes can cause circling disease, encephalitis, meningitis, septicemia, and mastitis in dairy cattle. Shedding of the pathogen from the udder or contamination from the environment can lead to presence of L. m01wcytogenes in raw milk. Surveys indicate the pathogen is in about 4% of US raw milks. Although HTST pasteurization commonly inactivates L. monocylogenes, evidence suggests that under unusual circumstances minimal survival is possible. The pathogen grows well in liquid dairy products at 4 to 35·C and achieves higher populations in chocolate than in unflavored milks. When present in cheese milk, growth of L. monocylOgenes may be retarded but not stopped by lactic starter cultures. The pathogen is concenlrated in the curd with only a small fraction of ceUs in milk appearing in whey. Once in curd, the behavior of the pathogen ranges from growth (feta cheese making) to death of most but not all cells (cottage cheese making). During ripening of cheese, the numbers of L. monocylogenes decrease gradually (as in Cheddar or Colby cheese), decrease precipitously early during ripening, and then stabilize (as in blue cheese) or increase markedly (as in Camembert cheese). Consumption of foods containing L. monocylOgenes can lead to listeriosis in susceptible humans (adults with a compromised immune system), pregnant women, and infants). In large outbreaks of human listeriosis, mortality rates of ca. 30% are common.
(Key words: listeriosis, foodbome illness, dairy foods) INTRODUCTION
Listeria monocytogenes has been called an "emerging foodbome pathogen" because only recently have we recognized that it can be transmitted in food. After several foodbome outbreaks and associated deaths caused by this pathogen in the United States and other countries, many believe L. monocytogenes has indeed emerged! The goals of this paper are to: 1) describe the organism and the disease it causes, 2) highlight some of the characteristics that enable L. monocytogenes to survive and grow in dairy products and some other foods, 3) provide proof that this pathogen is a threat to a safe food supply, and 4) describe some procedures to detect the pathogen in food, avoid contamination of food, and inactivate the pathogen in food and in food factories. GENERAL CHARACTERISTICS Morphology
Listeria monocytogenes is a gram-positive, acapsular, asporogenous bacterium that was first described in 1926. The diphtheroid-like rod measures 1.0 to 2.0 I!ffi by .5 I!ffi. Listeriae are motile, predominantly by means of perittichous flagella, and exhibit characteristic "turnbling motility" when grown at 20 to 25·C (51). Colonies have a characteristic blue-green appearance when viewed by the Henry method of oblique lighting (56). On blood agar, colonies produce a narrow zone of beta-hemolysis caused by a soluble hemotoxin (98). Growth Requirements
Received Seplember II. 1989. Accepted November 2. 1989. lInvited paper. 1990 J Dairy Sci 73:912-928
Listeriae are not fastidious. They can survive in feces, milk, soil, water, silage, and on plants. 912
USTERJA AND DAIRY FOODS
913
Furthennore, bacteria in the genus Listeria can multiply in high salt (up to 10% sodium chloride) or bile concentrations (98). Listeria monocytogenes grows best at neutral to slightly alkaline pH but will grow from pH 5.0 to 9.6 (51, 98). Listeria monocytogenes grows from 3 to 45°C (optimum for growth, 30 to 37°C). Listeriae are aerobic to microaerophilic; growth is enhanced in an atmosphere with reduced oxygen and 5 to 10% C~. Listeria monocytogenes requires biotin, riboflavin, thiamine, thioctic acid, and some amino acids for growth.
merous vertebrates can be infected, including ruminants (cattle, sheep, and goats), pigs, poultry (61), foxes, gerbils, rabbits, and guinea pigs (81). The pathogen can cause circling disease (61, 106), encephalitis, meningitis, and septicemia (61). Listeria monocytogenes also can cause bovine mastitis and then be excreted in milk (48), and it is generally accepted that listeriosis can be transmitted to humans by contaminated milk (98). Vizcaino and Garcia (107) found that 13 of 36 sero-positive otherwise healthy cows excreted Listeria in the milk.
Biochemical Properties
Transmission
Listeria monocytogenes possesses a split noncyclic citric acid pathway with an oxidative and reductive portion that is important for biosynthesis of cell constituents. Energy is supplied by glycolysis (66). Listeria monocytogenes has a heme-containing catalase that is inhibited by KCN and NaN3. Glucose is fermented and oxidized by the bacterium with production of acid, but no gas. The end product of fermentation is mainly lactic acid. Gelatin, casein, and milk are not hydrolyzed (98).
Listeria monocytogenes can infect humans and other animals by oral, ocular, cutaneous, respiratory, or urogenital routes (87). Modes of transmission from animals to humans include ingestion of contaminated food and contact with infected animals. Milk can be contaminated by environmental sources, such as feces, or the bacterium can be shed in the milk itself (58).
Serotyping
Antigenic types are distinguished by somatic (0) and flagellar (H) antigens. Based on these
antigens, strains of L. monocytogenes are subdivided into serotypes: 1l2a, 1/2b, 1/2c, 3a, 3b, 3c, 4a, 4ab, 4b, 4c, 4d, 4e, 7 (79), 4f, 4g, 5, and 6 (87). Of 722 human cases of listeriosis in the US, 59% were caused by serotype 4b and 18% by serotype l/2a. A bacteriophage typing system, based on lytic properties of a set of phages, also has been developed. It is reproducible and discriminatory; however, only 64% of the serotypes can be typed by the bacteriophages (79). USTERIOSIS Zoonosis
Listeriosis, the disease caused by L. monocytogenes, is generally accepted as a zoonosis-a disease or infection naturally transmitted between vertebrate animals and humans (87). Nu-
Virulence
Virulence of L. monocytogenes may be related to two toxins: hemolysin and cytotonic toxin. The hemolysin is a cytolysin able to lyse tissue and red blood cells. The cytotonic toxin stimulates cyclic AMP production similar to cholera toxin (77). Additionally, almost all strains have phosphatase activity, and this may playa role in virulence (82). Upon infection, L. monocytogenes is taken up by white blood cells, where it survives and multiplies (10). All serotypes of L. monocytogenes have the ability to provoke monocytosis (98). Listeria strains have low virulence in comparison to other pathogenic microorganisms, based on animal studies (87). Most healthy humans can fight off infections with only minor symptoms (stomach upset or "flu") or no symptoms (13). The incidence of listeriosis is highest in newborns, pregnant women, elderly persons, and immunosuppressed individuals (10). Immunosuppression can occur in patients receiving corticosteroids or cancer chemotherapy (13), or in individuals with preexisting diseases such as cirrhosis, AIDS, or malignancies (10). Journal of Dairy Science Vol. 73. No.4, 1990
914
PEARSON AND MARTH
Disease Manifestations In Humans
Disorders associated with L. monocytogenes include meningoencephalitis, influenza-like, low grade septicemia in the prenatal period, infectious mononucleosis-like syndrome, septicemia in adults, pneumonia, endocarditis. localized abscesses, papular or pustular cutaneous lesions, conjunctivitis. urethritis, and recurring abortion. Listeria monocytogenes also may cause cerebral damage. mental retardation, and mononucleosis in some cases (51). Listeriosis usually occurs sporadically in the human population, but epidemics develop occasionally (61). Earlier literature stated the mortality rate to be 90% (61); currently it is estimated at 30% (87). The high mortality rate results from underlying immune deficiencies in susceptible individuals. Treatment of listeriosis involves drug therapy; ampicillin, erythromycin. and rifampicin are the antibiotics of choice (87). A tetracycline and trimethoprim mixture showed in vivo synergy (67). Estradiol-17~ at concentrations >5 J.Lg/ml completely inhibits the hemolysin produced by L. monocytogenes (54). Chloramphenicol and streptomycin have relatively little effect on the pathogen (20). ENVIRONIENTAL RESISTANCE Psychrotrophlc PropertiM
Listeria monocytogenes is generally considered to be a mesophile (optimum growth at 30 to 37"C); however, it also has psychrotrophic properties that enable it to grow at refrigeration temperatures. Using an Arrhenius plot, the minimum temperature for growth of the pathogen was extrapolated to be 3·C (109). The duration of the lag phase increases with decreasing temperature (109); at 4·C the lag phase lasts 5 to 10 d (7).
Mechanical refrigeration maintains the temperature at 4 to 5"C to preserve many foods including dairy products and hopefully prevent growth of pathogenic organisms. Since the 196Os, several foodborne pathogens have been recognized as able to grow at ca. 5"C, including Clostridium botulinum type E. Yersinia enlerocolitica. Aeromonas hydrophila. and L. monocytogenes (83). Additionally. slight deviation upward in temperature from that desired can greatly increase the growth rate of L. monocylOJoumal of Dairy Science Vol. 13. No.4. 1990
genes (at 13·C the generation time was 5.0 h for strain V7 in whole milk. whereas at 4·C it was ca. 35 h) (89). The ability to grow at refrigeration temperatures. while other competing microflora cannot, selects for L. monocytogenes; this may play a role in foodbome outbreaks (83). Listeria monocytogenes is reportedly more virulent when grown at low rather than at high temperatures (51). and virulent strains (serotype 4b) exhibited enhanced psychrotrophic growth in whole milk (29). Composition of liquid milk products affects the psychrotrophic growth of L. monocytogenes. Whipping cream supported development of the lowest population when compared with whole. skim. and chocolate milk, perhaps because of the higher fat water ratio (89). Two strains of serotype 4b grew better at 4 and lO·C in whole milk than in skim milk and reconstituted nonfat milk solids (11 %) (29). Growth of strain Scott A (also serotype 4b) in whole versus skim milk was insignificantly different at 4, 8, and l3·C (89). Conversely, Pseudomonas spp. grew slightly faster in skim milk (8.7% total solids) that was fortified to contain 10 or 12% total solids than in unfortified skim milk (63). Competition with other psycmotrophic bacteria (especially Pseudomonas spp.) has varied effects on growth of L. monocytogenes in milk. Studies with Pseudomonas fragi and L. monocytogenes strain Scott A co-inoculations resulted in neither inhibition nor stimulation of either organism (73). Other studies found that Pseudomonas jluorescens P26 and B52 moderately inhibited L. monocytogenes strains Scott A, California (CA) and V7 at 7 and 13·C (41). When milks (whole, skim. and reconstituted nonfat milk solids) were preincubated for 3 d with P. jluorescens P26, TI5. and 852 and P. fragi, the generation times of L. monocytogenes strain Scott A at 10·C were shorter (from 10.0 h without Pseudomonas spp. to 7.0 to 8.5 h with various Pseudomonas spp.). The proteolytic activity of the Pseudomonas spp. and resultant availability of nutrients to the pathogen may have been responsible for this stimulation (73). The ability of L. monocytogenes to grow at refrigeration temperatures has created unique sanitation problems. In a study of simulated milk cooling systems by Petran and Zottola
USTERIA AND DAIRY FOODS
(85), it was found that sweet water coolant systems (refrigerated potable water) could be conductive to growth of L. monocytogenes if small holes or cracks in equipment allowed leakage of some milk into the coolant. Usteria monocylogenes was inhibited in concentrations of propylene glycol greater than 12%, so most dairies are not threatened. since 30% glycol is used. Another mode of transmission of the pathogen could be on the exterior of waxed cardboard or plastic milk containers, according to Stanfield et al. (101). After 14 d, nonheatshocked L. monocylogenes serotypes la, Ib, and 4b could be isolated from the exterior of previously inoculated plastic containers held at -8 to 6.6·C. Heat-shocked organisms could not be isolated after 4 d. Thermal ResIstance
It is questionable whether L. monocylogenes can survive the pasteurization process as defined by the Grade A pasteurized milk ordinance (43). However, evidence on this issue is conflicting, partially because of 1) the methods used to determine heat resistance of L. rnonocytogenes and 2) the physiological state of the bacterium during heating. Strains of L. monocytogenes differ in their heat resistance. Strains most commonly used for testing include those isolated from the Massachusetts listeriosis outbreak (described later in this paper)-Scon A (serotype 4b), V7 (serotype la), and V37 (serotype 4b) (l4)-and those most commonly isolated from raw milk-ATCC 19115 (serotype 4b) and F5069 (serotype 4b) (29). Of these, strain Scott A is the most heat resistant (14). The holding technique for pasteurization (61.7·C for 35 min) used by Beams and Girard (11) failed to free milk from viable cells of L. monocylOgenes when the initial inoculum of the pathogen was 5 x 1()4/ml or greater. Using sealed borosilicate glass tubes, Bradshaw et al. (16) showed that 15 10g1O cells of L. monocytogenes would be inactivated in 15 s at 71.7·C. Donnelly et al. (30) proposed that differences between the results in these two repons were caused by condensate and bacteria collecting in the cap of the test tube or by cells coating the walls of the test tube above the level of the
915
heating menstruum in the method used by Beams and Girard. Following heating of samples containing L. monocytogenes, researchers used immediate direct plating techniques (11, 15, 14, 19, 29, 30) or direct plating and enrichment techniques (32, 46) to recover survivors. Using only direct plating, all researchers except Beams and Girard (11) determined present pasteurization guidelines to be adequate to eliminate L. monocytogenes. Garayzabal et al. (46) isolated L. monocylogenes from pasteurized samples with cold enrichment but never with plating immediately following thermal treatment. Enrichment procedures allow use of a larger sample than does direct plating and may allow repair of thermally injured cells (32). As a facultative intracellular parasite, L. monocytogenes can exist within polymorphonucleocyte leukocytes (pMNL) and be shed in milk (32). Freely suspended cells of L. monocytogenes strain Scott A in whole milk were destroyed at 71.7·C for 15 s (14). Intracellular L. monocytogenes strain Scott A survived HTST pasteurization at 72.2·C for 16.4 s; however, this study used enrichment techniques to recover survivors (32). The D71.7·C value (time at 71.7·C needed for population to decrease by one order of magnitude) for L. monocytogenes strain Scott A inside phagocytes was 1.9 s with a ZD value (change in temperature needed to change D-value by one order of magnitude) of 6·C, using only direct plating techniques (18). Doyle et al. (32) also noted that PMNL in milk degraded after milk was held for 3 to 4 d at 4·C, and this may increase the susceptibility of L. monocytogenes to inactivation by pasteurization because the cells become freely suspended. Bradshaw et al. (15) confirmed that pasteurization guidelines for whole and skim milk, heavy cream, and ice cream mix are adequate to ensure inactivation of L. monocytogenes (D71.7·C values ranged from .9 to 2.7 s). The study also showed that the thermal resistance of L. monocytogenes strain Scott A increased when the products had been previously sterilized. This may have implications when already processed products are pasteurized. The heat resistance of five strains of L. monocytogenes was not affected by prior growth in skim, whole, or reconstituted nonfat milk (29). Journal of Dairy Science Vol. 73. No.4, 1990
916
PEARSON AND MARTH
EVIDENCE FOR A FOODBORNE THREAT
Outbreaks of Foodbome liS18riosis
Four major recent outbreaks of listeriosis, two in the US and one each in Canada and Switzerland, caused by ingestion of contaminated food confmn that L. monocytogenes can be present in food and can cause illness when such food is consumed by susceptible persons. In the Maritime Provinces of Canada, 34 cases of perinatal listeriosis and seven of adult listeriosis occurred between March 1 and September I, 1981. A food preference survey identified coleslaw as the vehicle of transmission. Additionally, L. monocytogenes serotype 4b was isolated from coleslaw in a patient's refrigerator and from patients. The contamination probably resulted from sheep manure (from sheep with documented cases of circling disease) used to fertilize the cabbage field. Cabbage is often held in prolonged cold storage before distribution, and this may have allowed proliferation of the psychrotrophic L. monocytogenes while the number of competing microorganisms decreased (97). An epidemiological investigation by the Centers for Disease Control (CDC) showed pasteurized whole or 2% milk of a specific brand to be the vehicle of infection in a listeriosis outbreak in Massachusetts. Between June 30 and August 30, 1983, 49 patients (7 fetuses or infants and 42 adults) acquired listeriosis. All of the adults were immunosuppressed because of preexisting illness. Mortality was 29%. In addition to the two case-control studies, investigators also used the dose-response effect, the protective effect of skim milk, comparison with cases of listeriosis in Connecticut, and phage typing to assess the validity of their findings. The implicated milk was pasteurized at a single factory, and proper pasteurization guidelines were followed. Twelve percent of raw milk samples collected after the outbreak contained the organism. Veterinarians that year had diagnosed four cases of bovine listeriosis (encephalitis) on farms supplying the dairy. Postpasteurization contamination was thought unlikely for two reasons: skim milk was not contaminated with the pathogen, even though it was pasteurized with the same equipment on the same day, and no source of the organism could be found in the factory (42). During the last several years Journal of Dairy Science Vol. 73. No.4. 1990
questions have been raised about the adequacy of the epidemiologic investigation, and the real cause of the outbreak may have been something other than milk. In another outbreak, Mexican-style fresh cheese was determined to be the cause of at least 86 cases of listeriosis from January 1 to June 14, 1985 in California. Fifty-eight of the cases involved mother-infant pairs. Twentynine deaths occurred (33.7% mortality). Using case-control studies, consumption of cheese made in a single factory in the Los Angeles area was determined to be the cause of listeriosis. Cheese samples with different ex.piration dates contained L. monocytogenes serotype 4b, suggesting a continuing contamination problem in the factory (62). Some cheese was phosphatase-positive, indicative of improper pasteurization or no pasteurization of cheese milk. However, application of the phosphatase test to cheese to determine whether or not it was made from pasteurized milk has been questioned. Producers of the cheese may have mixed raw and pasteurized milk or postpasteurization contamination may have been the problem (99). The food-related outbreak of listeriosis in Switzerland was associated with consumption of Vacherin Mont d'Or cheese. This is a soft surface-ripened (bacterial) cheese. Continued testing of products for L. monocylogenes by regulatory agencies has resulted in recalls of contaminated food. Several brands of Brie cheese, produced in Sorcy, France were contaminated with L. monocytogenes, according to a Food and Drug Administration (FDA) warning on February 12, 1986. The factory making the cheese operated under the current imported soft-ripened cheese agreement that requires pasteurization of the milk before cheese manufacture (4). A month later. more brands of Brie cheese were recalled because of contamination (2). In the US, a major producer of ice cream bars recalled 25,000 cases of product because of contamination with L. monocytogenes. There were no reports of illness, probably because the frozen state of the product did not allow proliferation of the bacterium to numbers sufficient to cause infection of consumers (3). Another major manufacturer of ice cream recalled a million gallons of product from 11 Midwestern states when L. monocytogenes was found in the product (106). Additional recalls included those of Liederkranz cheese in 1985
USTERIA AND DAIRY FOODS
and the soft Mexican-style cheese in Arizona in 1986 (99). Recalls of products are costly and obviously detrimental to the reputation of the company. Occurrence In Dairy Products
The Massachusetts and California outbreaks pointed to milkborne L. monocytogenes as a serious public health problem. Several studies have been done to determine the incidence of L. monocytogenes in raw milk. Lovett et al. (72) tested raw milk from farm bulk tanks in three areas of the United States-California, Massachusetts, and a Tri-State area (Ohio, Kentucky, and Indiana). The overall incidence was 4.2% (milks from California had 0% and from Massachusetts had 7%). Seasonal variation occurred in Massachusetts and in the Tri-State area; incidence was lowest in hot weather months and peaked in the cold months. Most of the TriState isolates were serotype 4b. and the Massachusetts isolates were mainly serotype I (72). Other studies support these results. Liewen and Plautz (69) detected L. monocytogenes is 4% of raw milk samples in Nebraska The seasonal variation also was confirmed in this study. Raw milk samples collected throughout the year from a factory in Madrid, Spain contained L. monocytogenes at percentages ranging from 14.3 to 75 (44). Although L. monocytogenes was not detected, other species of Listeria were isolated from raw milk in New Zealand (103). In a survey in the Netherlands by Beckers et al. (12), L. monocytogenes was detected in 7 of 69 samples of imported soft cheese made from raw milk. Another study of blue-veined cheeses found the pathogens in 2 of 28 samples (23). Garayazabal et al. (45) found L. monocytogenes in 21.4% of pasteurized milk samples (3.2% fat milk treated at 78T for IS s) marketed by a Madrid processing plant.
917
zen bean, peas, pea pods, and spinach were examined and were found to be Listeria-free using a 7-d enrichment procedure. Another study did isolate L. monocytogenes from a head lettuce and also showed that the population of the pathogen increased during storage (102). Gitter (47) isolated L. monocytogenes serotypes 4 and 112 from 14.7% of "oven-ready" poultry from supermarkets. Various surveys have shown that L. monocytogenes may be present in S to SO% of fresh retail red meat and poultry (31). A study by the FDA found 26% of seafood samples to be positive for L. monocytogenes. Positive products included shrimp, crabmeat, lobster tail, langostinos, scallops, sunmi-based imitation seafoods (108), and cold smoked fish (S). Beh8vlor During Food Proceulng
To explore behavior of L. monocytogenes during food preparation or processing. researchers have monitored survival and growth of the pathogen in artificially inoculated food products. These studies show what foods support growth of L. monocytogenes and what food processing practices increase or decrease changes in high populations of the pathogen in the food. As discussed previously, milk and milk products have been implicated as sources of listeriosis. Rosenow and Marth (89) documented growth of four strains of L. monocytogenes (Scon A, V7. V37CE. and CA) in skim, whole, and chocolate milk and in whipping cream at temperatures ranging from 4 to 3SoC. Generation times ranged from 29.S h (strain V37CE) in chocolate milk) to 45.5 h (strain V7 in whipping cream) at 4°C. When the temperature was increased to 13T-a temperature conceivably encountered during storage and transport of food by consumers-generation times were shortened to 4.5 h (strain V7 in chocolate milk) or 6.8 h (strain CA in whole milk). All strains eventually attained populations of at Occurrence In Other Foods least 107 cfu/ml. Chocolate milk always supIn similar studies, various vegetables. beef, ported development of the highest populations poultry, and seafood were tested for presence of of L. monocytogenes strains at all temperatures L. monocytogenes. Petran et al. (86) tested fresh tested. Further investigation showed that both lettuce. potato peels. com husks, broccoli cane sugar and cocoa powder, present in the stems, cabbage outer leaves, carrot peels, cauli- chocolate milk, served to increase the maxiflower stems, mushroom stems, spinach, and mum populations attained by L. monocytogenes beet peels from a local supermarket. They did (88). The enhanced growth of the pathogen in not isolate L. monocytogenes. Additionally, fro- chocolate milk and the ability to grow during Joumal of Dairy Science Vol. 73. No.4, 1990
918
PEARSON AND MARTH
refrigeration emphasize the need to avoid postpasteurization contamination of this product during the handling it receives. Concentrated (30% solids) and unconcentrated skim milks, inoculated with lOS to 1()6 L. monocytogenes strain Scott A or V7/ml, were spray-dried to a moisture content of 3.6 to 6.4%. The spray drying process itself resulted in a 1- to 1.5-loglO drop in the population of the pathogen initially/ml of milk when that milk was converted to the dry form. Storage of nonfat dry milk at 25"C resulted in a continued decrease in the population of the pathogen with strain V7 surviving the longest (12 wk) in nonfat dry milk made from concentrated skim milk (33). Cheeseborne outbreaks of listeriosis show cheese to be susceptible to contamination by L. monocytogenes. Two surface-ripened cheeses (Camembert and brick) were made from milk containing L. monocytogenes, and behavior of the pathogen was monitored through manufacture and storage. Slice, surface, and interior brick cheese (prepared by the washed curd method with 5 x 102 L. monocytogenes/ml in the milk) samples reached populations of 5.11 to 7.01 log10 L. monocytogenes strain Scott A or Ohio (OH)/g after 28 d of ripening. The surface samples had highest populations of the pathogen, perhaps because of their nearly neutral pH (6.5 to 6.9). After 154 d of ripening. numbers of the pathogen decreased by lo-fold. Strains CA and V7 failed to grow significantly, and cold enrichment was required to detect the low populations (74). Camembert cheese (prepared from milk containing 500 L. monocytogenes strain Scott A. V7, CA, or OH/ml) supported development of populations of 1 x 1()6 to 5 x 107 cfu of the pathogen/g after 65 d of ripening. When the pH of some samples of cheese was adjusted, growth of the pathogen was greatest at pH 6.1 rather than pH 4.6 or 7.4 (91). During manufacture of blue cheese by the Modified Iowa Method, artificially added L. monocytogenes mainly went with the curd rather than the whey. This also was true for other varieties of cheese. The population of L. monocytogenes (all strains) decreased in blue cheese during the first 50 d of ripening (at 9 to 12 C); however, the population of strain Scott A leveled off after the pH increased with mold growth during 50 to 120 d of ripening (84). D
Journal of Dairy Science Vol. 73, No.4, 1990
Studies show hard cheeses also can pose a health hazard if made from contaminated milk. Cheddar cheese (with normal values for moisture, fat-in-dry matter, salt, and pH) made from milk containing 500 L. monocytogenes strain CA, V7, or Scott Nml did not support growth of the pathogen during manufacture. Storage at 6 or 13 C resulted in a .1 to .8 log10 cfu/g increase in population initially, and then populations decreased. All samples contained a detectable number of L. monocytogenes after 60 d of ripening (90). In fact, viable cells of the pathogen were found in some 1-year-old cheese. Cheddar cheese can be used to make cold pack cheese food. Ryser and Marth (92) found that if inoculated cheese food was made without preservatives of acidifying agents, samples contained detectable populations (2.05 to 2.73 10g1O cfu/g) of strains Scott A, V7, CA, and OH after 182 d of storage at 4 C. Listeria monocytogenes was unable to grow in the coldpack cheese food with pH of 5.0 to 5.45. The pH of this product can be lowered by adding lactic or acetic acid. Sorbic acid and sodium propionate, common mold inhibitors that may be added to cheese food, are also somewhat antibacterial toward the pathogen (92). As with other varieties of cheese. L. monocytogenes is mainly entrapped in the curd during manufacture of Colby cheese. During storage at 4 C, populations of the pathogen remained constant for a few weeks and then decreased. Two factors increased the size of the surviving population of the pathogen: a higher rather than lower moisture content in cheese and a higher rather than lower initial inoculum level (111). Preparation of semi-hard cheese (Manchegotype) with milk containing 5.3 or 3.6 10glO L. monocytogenes 4b cfulml resulted in populations after 60 d of storage of 5.3 and 3.5 10glO cfulg respectively. Brining of the cheese was accompanied by a drop in population (25). Cottage cheese was manufactured according to the short-set procedure from pasteurized skim milk containing 104 to lOS L. monocytogenes strain Scott A or V7/ml. Cooking the curd decreased the population to -hemolysis. Recently the method has been modified so the amount of sample actually tested is the same as that used by the FDA. Beside enrichment procedures just described, there are several newly developed methods. Doyle and Schoeni (34) used an enrichment broth with tryptose, defibrinated sheep blood. glucose. K2HP04, polymyxin B. acriflavine-HCI, and nalidixic acid. The sample is homogenized in 100 ml of the broth and incubated with agitation at 37·C under microaerobic conditions for 24 h and then plated on MLA. Slade and Collins-Thompson (100) developed a two-stage enrichment technique. The primary enrichment is with nonselective tryptose broth for 3, 7. or 14 d at 4·C. The secondary enrichment broth consists of thiocyanate-nalidixic acid broth with acriflavine. Mter primary and secondary enrichment, samples are streaked on MLA. Several studies have compared the various enrichment procedures. Doyle and Schoeni (34) found that the FDA procedure. cold enrichment. and their selective enrichment procedure isolated the pathogen from 46% of soft, surface-ripened cheese. but no single method detected the organisms in all Listeria-positive samples. The USDA procedure was significantly worse at recovering L. monocytogenes from Brie cheese than the FDA procedure, even when the USDA broth was supplemented with D-glucose and phytone (105). However. Farber et al. (40) found that the USDA procedure was more efficient in isolating Listeria from ground meat than the FDA procedure. Monoclonal Antibody-Enzyme Immunoassay
Enzyme immunoassay (EIA), also known as enzyme-linked immunosorbent assay (ELISA), can be used to isolate and detect Listeria in food. Farber et al. (40) tested an EIA method with microtiter plates and on membrane filters using a monoclonal antibody directed against the B flagellar antigen of Listeria spp. Using the membrane system, naturally contaminated ground meats gave positive results as evidenced by purple spots on the filter. Two procedures developed for the microtiter plate system (coating plates with a flagella-containing supernatant fluid and a capture sandwich assay) were sucJournal of Dairy Science Vol. 73. No.4, 1990
922
PEARSON AND MARTH
cessful. Efforts to produce a monoclonal antibody to the listerial hemolysin continue in the hope that this will be a more specific immunological probe than the flagellar antigen. Mattingly et al. (75) produced genus-specific monoclonal antibodies against Listeria and used two of the antibodies to create an ELISA technique to detect the pathogen. Samples were grown in an enrichment broth (containing nalidixic acid. acriflavine. and cycloheximide) before the ELISA. All Listeria spp. were positive (as determined by optical density readings) except Listeria denitrificans (this species may be erroneously placed in the genus Listeria). The test was very sensitive (a 24-h-old pure culture from the enrichment broth was detected as positive through a 1:40.000 dilution). Over 20 common contaminating bacterial species yielded negative results. Deoxyribonucleic Acid Hybridization
Datta et al. (21) developed a DNA colony hybridization technique to detect L. monocytogenes. The extracellular ~-hemolysin from the pathogen is thought to be important in intracellular survival and virulence. and therefore. the hemolysin gene could be used to detect virulent L. monocytogenes. Crude DNA could not be used for hybridization because of the debris present in homogenized foods. The ~-hemoly sin gene fragment (500 base pairs) was cloned using recombinant DNA techniques. Lysing of the cells by microwaves was necessary to amplify the radiolabeled DNA probe. Only ~ hemolytic (cAMP test-positive) strains of L. monocytogenes were positive. This DNA probe was then used to detect hemolytic L. monocytogenes in naturally contaminated dairy products (22). The dairy products-raw milk. ricotta cheese. and semisoft cheeses-were plated and the resulting colonies were tested by DNA colony hybridization. Use of the gene probe saves time; however. there are some limitations. Only .1 ml can be plated. so for detection the contaminated product must have>10 cells/g. Additionally. not all L. monocytogenes strains react with the probe. Immunofluorescence and Flow Cytometry
Early fluorescent antibody techniques for identification of L. monocytogenes resulted in Journal of Dairy Science Vol. 73. No.4. 1990
many false-positive results because of crossreactivity with Streptococcus. Micrococcus. and Staphylococcus spp. Visual differentiation of these bacteria is difficult, so fluorescence microscopy is not very reliable. To increase the specificity of this technique. Donnelly and Baigent (27) employed an automated fluorescentantibody procedure through use of flow cytometry. Flow cytometry characterizes a bacterial cell population by morphology. nucleic acid content. and surface antigens as the cells in suspension pass through a laser beam. The species was labelled using polyvalent antibodies for cell surface antigens and propidium iodide for cellular DNA. Donnelly and Baigent (27) found that light scatter was not sufficient to distinguish between Listeria and non-Listeria contaminants. Sonicating the cells improved the fluorescent labelling procedure and consequently improved the resolution. In testing raw milk. use of a Listeria enrichment broth (containing nalidixic acid and acriflavine) before flow cytometry, decreased the number of competing microorganisms. Additionally. the Listeria enrichment broth allowed growth of the pathogen (from 8.5 x 1(jJ to 1.7 x 1()6 cfu/ml within 24 h at 3TC); this is necessary since resolution is not possible at
