Detecting and Controlling Microorganisms in Foods

NON-TECHNICAL SUMMARY: Mycotoxin-producing and spoilage molds are common contaminants of raw agricultural commodities and foods. Hence, it is important to detect these molds before they can produce mycotoxins and before the food is spoiled to prevent safety and quality concerns with the food supply. Traditional enumeration methods take five to seven days to detect molds and an additional one to two weeks to identify them to the genus and species level. In recent years, fresh fruits (cantaloupe, tomatoes) and vegetables (green onions, lettuce, spinach) that are consumed raw have been involved in several well-publicized multistate outbreaks of foodborne illness in the USA, which have been linked to strains of Escherichia coli O157:H7 and several Salmonella strains. Normally, E. coli O157:H7 and Salmonella strains are associated with animal reservoirs. These outbreaks have led to research on how these pathogens become associated with fresh produce and whether they attach to surfaces, form biofilms or enter the plant tissues.
<P>

APPROACH: Objective 1: The major molds involved with mycotoxin production (Aspergilluis, Fusarium and Penicillium) will be studied. Since it has been shown by current research that enzymes and physical disruption are effective in extracting DNA fom Fusarium spores, similar methods will be tried for mycotoxigenic species of Aspergilluis and Penicillium. The extracted DNA will be analyzed by DNA-based methods. Objective 2: Escherichia coli O157:H7, which were genetically modified in Dr. B. M. Applegate's lab to contain a fluorescent gene, will be used in the research. This strain of E. coli O157:H7 will be grown on leafy greens in a growth chamber to determine if it attaches or goes into the tissue of the greens. This will be observed by fluorescence and microscopic examination. Also, normal microflora will be isolated from leafy greens and will also be put into the growth chamber to determine if they compete with new strain of E. coli O157:H7 to attach or go inside the leaves. This also will be observed by fluorescence and microscopic examination. The results of all the research will be published in peer-reviewed journals, presented at scientific meetings and to food companies.

<P>PROGRESS: 2006/10 TO 2007/09 <BR>
OUTPUTS: Experiments were done to determine which methods would be most effective at releasing the DNA from spores of Fusarium species for use in mold detection by PCR. Digestion with lyticase provided detection of Fusarium verticillioides to 1000 spores per reaction tube and Fusarium graminearum to 100 spores per reaction tube after real-time PCR analysis. When the lyticase protocol was modified to include incubation times from two to six hours at 37 C and further purification and concentration steps, Fusarium graminearum could be detected at 10 spores per reaction tube when analyzed by real-time PCR. Similar experiments are being done with Fusarium verticillioides to improve detection levels. Experiments were done with immunomagnetic capture of Fusarium verticillioides in buffer and cornmeal by using beads pre-coated with anti-rabbit goat antibodies; however, the capture rates varied from 15 to 66% making the method inefficient for routine mold detection. Experiments were done with lactic acid bacteria and Bacillus species isolated from ready-to-eat ham to determine if they could inhibit Listeria monocytogenes on the surface of ham. The strains of lactic acid bacteria were Gram-positive rods that were identified as either Carnobacterium or Lactobacilus based on biochemical tests. The Bacillus strains were identified as Bacillus thuringiensis by the Vitek identification system. Three strains of these lactic acid bacteria and one strain of Bacillus inhibited all eleven strains of Listeria monocytogenes used in the experiments. When experiments were done by co-inoculating 10 to 100 cells of either lactic acid bacteria or Bacillus species and Listeria monocytogenes onto the surface of ham followed by incubation at 4 and 10 C for 14 days, growth of Listeria monocytogenes was inhibited. The pH was usually decreased by one unit or less; therefore, the inhibition was possibly due to the production of antimicrobial compounds. <BR> PARTICIPANTS: Charles P. Woloshuk (co-PI) Karen A. Chong (graduate student) Willette M. Crawford (graduate student) Janaka S. Morandage (graduate student) <BR> TARGET AUDIENCES: Food Scientists and food industry safety and quality control managers.
<BR> <BR>
IMPACT: 2006/10 TO 2007/09<BR>
The ability to break open the spore of Fusarium species to release the DNA means that a method can be developed to rapidly detect these mycotoxin-producing molds in grains and foods. The ability to detect as few as 10 to 100 spores of these Fusarium species also means that a sensitive method can be developed. This is the first report of low levels (10 to 100 cells) of spoilage microflora, which includes lactic acid bacteria and Bacillus species, inhibiting the growth of Listeria monocytogenes on the surface of ready-to-eat meat. These results could be used in future risk assessments for prevention and control of Listeria monocytogenes in ready-to-eat foods.
<BR> <BR>

PROGRESS: 2005/10/01 TO 2006/09/30<BR>
Three areas were studied for the detection and control of microorganisms in foods. Objective one was to develop a method to release DNA from Fusarium conidia for use in polymerase chain reaction (PCR) for mold detection in foods. Objective two was to complete a project on the predictive model for post-process pasteurization of frankfurters that contain Listeria monocytogenes that have been grown under different conditions because L. monocytogenes continues to be a concern in processed meats and more information is needed to determine post-process lethality. Objective three was to study the growth of spoilage bacteria on ham and to determine how it then affects Listeria monocytogenes because there have been several recalls of ham due to contamination by L. monocytogenes. Objective 1: The FastPrep extraction of DNA using up to 6 treatments cycles of 40 sec produced a 40 to 52% lysis rate for Fusarium verticillioides conidia. The use of lyticase gave a 3% lysis rate to release the DNA from 100,000 conidia of Fusarium verticillioides. Research is continuing to improve the rate of DNA lysis from the conidia by these two methods. Objective 2: L. monocytogenes cells grown in tryptic soy broth containing 0.6% yeast extract (TSBYE) and in 8.5% fat frankfurter slurry had D-values at 60 C of 2.2 min and cells grown in 23% and 11% fat frankfurter slurries had D-values at 60 C of 1.7 min. L. monocytogenes cells grown in tryptic soy broth with 25 ppm quaternary ammonium compounds had D-values at 60 C of 1.0 min. The death of L. monocytogenes cells on frankfurter surfaces took longer than in frankfurter slurries. When a hot water bath at 60 C was used to post-process pasteurize frankfurters inoculated with L. monocytogenes, cells grown in both TSBYE and 23% fat frankfurter slurries took 19.6 min for a 3-log10 reduction of cells compared to 10.8 min for cells grown in 8.5% fat frankfurter slurries. This research begins to show what happens in an actual food versus a model system. Objective 3: Twenty-eight strains each of lactic acid bacteria and Bacillus species were screened for their ability to inhibit the growth of Listeria monocytogenes on agar spot assays. Those that showed the greater inhibition were further screened against six Listeria monocytogenes strains at various concentrations and for sensitivity to proteinase K. Two lactic acid bacteria had the same level of inhibition of L. monocytogenes whether the levels were 100 or 1 million. One lactic acid bacterium had the same degree of inhibition when acid was produced and when it was suppressed. Bacillus species were not as effective as lactic acid bacteria at low levels and only one species inhibited all six strains of L. monocytogenes. Three lactic acid bacteria and two Bacillus species were inhibited by proteinase K suggesting that bacteriocins were produced. Research is continuing with inoculation of ham with the L. monocytogenes, lactic acid bacteria and Bacillus species to determine their actions and whether they produce bacteriocins that can inhibit L. monocytogenes.
<BR> <BR>
IMPACT: 2005/10/01 TO 2006/09/30<BR>
The research on the development of a protocol to release DNA from mold spores suggests that if this method can be further optimized, then a rapid method for mold detect can be developed. The predictive model for post-process killing of Listeria monocytogenes in hot dogs was submitted to the USDA to add to the Pathogen Modeling Program and to ComBase, which are computer programs maintained at the USDA website that can be used by the meat industry to develop thermal processes for their products. The research on the growth of spoilage bacteria in the presence of L. monocytogenes may result in the discovery of a natural antimicrobial agent that may be effective in preventing L. monocytogenes from growing in ham and other processed meats.

<BR> <BR>
PROGRESS: 2004/10/01 TO 2005/09/30<BR>
Research was done on the three objectives. For objective 1, which is to develop and characterize rapid methods to detect mycotoxin-producing molds in foods, research continued on the development of a protocl to release DNA from spores (conidia) of Fusarium species. It was determined that additional time was needed to rotate the tubes in the FastPrep lysing instrument to release DNA from Fusarium micro- and macroconidia. Also, if any mycelial fragments were left in the samples, they were easily released and detected by the PCR and agarose gel. Currently, about 100,000 spores of Fusarium graminearum and 1 million spores of Fusarium verticillioides can be detected. The methods are being further revised and optimized to be able to detect lower levels of spores. For objective 2, which is to work in teams to develop predictive models for death of pathogens in foods, a model was developed to predict how L. monocytogenes reacted to temperature using cells that were grown in trypticase soy broth (TSB) and heated in hot dog broths with different levels of two preservatives currently used in processed meats (sodium lactate and sodium diacetate). This research answers questions about how L. monocytogenes reacted to 55,60 and 65 C in processed meats with different levels of sodium lactate (0, 2.4 or 4.8% wt/vol) and sodium diacetate (0, 0.125, or 0.25% wt/vol). In addition, L. monocytogenes cells were grown in TSBYE, sanitizer solutions and hot dog broths and the resulting data were analyzed with mathematical equations to determine the death of L. monocytogenes in hot dog broth and on the surfaces of packaged hot dogs at 60 C. The death of L. monocytogenes was affected by the temperature where an increase from 55 to 65 C increased death, by the type of preservative where sodium diacetate produced most death; however, the presence of sanitizer (quaternary ammonium compounds), fat level and strain of L. monocytogenes used were not significant when packaged hot dogs were heated but they were significant when heated in hot dog broth systems; therefore, meat processing companies could use a predictive model to determine what heat or preservative treatment to give packaged ready-to-eat hot dogs and processed meats to kill L. monocytogenes if it were to contaminate the meat after processing.A post process treatment for packaged hot dogs at 60 C for 10.8 to 19.6 min reduced L. monocytogenes by 3 logs with the maximum decrease in 15 min at 60 C being from 2.3 to 5.1 log reductions depending on the treatment, strain and growth medium. As the amount of fat in the hot dogs decreased from 23 to 8.5%, L. monocytogenes became more resistant to the heat treatments. This result was opposite what has been reported in the literature. For objective 3, which is to work in multidisciplinary teams to develop rapid method s to detect bacterial pathogens in foods, microbial advice was given to the group working on the development of FTIR to detect Salmonella strains and Escherichia coli O157:47 in foods.
<BR> <BR>
IMPACT: 2004/10/01 TO 2005/09/30<BR>
The research on the development of a protocol to release DNA from mold spores suggests that if this method can be further optimized, then a rapid method for mold detect can be developed. The predictive model for post-process killing of Listeria monocytogenes in hot dogs was submitted to the USDA to add to the Pathogen Modeling Program and to ComBase, which are computer programs maintained at the USDA website that can be used by the meat industry to develop thermal processes for their products.

<BR> <BR>
PROGRESS: 2003/10/01 TO 2004/09/29<BR>
A procedure for capturing antibodies produced against two molds, Fusarium graminearum and Fusarium verticillioides, in both peptone and cornmeal was developed to use as an immunocapture before doing quantitative real-time polymerase chain reaction (qPCR). The use of the 96-well microtiter plates was the best method to capture the antibodies of the Fusarium species; however, there was nonspecific binding. To overcome this binding, blocking agents (bovine serum albumin, beta-casein, nonfat dry milk and Triton-X) were evaluated. None of these agents gave significantly different results from the controls without blocking agents. In order to access the DNA before qPCR, sonication for 45 second for F. verticillioides and 120 seconds for F. graminearum were used. The qPCR targeted the amplification of two genes involved in mycotoxin production, the Tri6 gene for F. graminearum and the Fum1 gene for F. verticillioides, which were confirmed on agarose gel electrophoresis. The concentration of F. verticillioides affected its detection by qPCR but that of F. graminearum did not. Detection of F. graminearum was 100 CFU/ml and 100,000 CFU/ml for the buffer and cornmeal, respectively; however, for F. verticillioides, it was 100,000 CFU/ml and 10,000,000 CFU/ml for the buffer and cornmeal, respectively. The immunocapture method did not increase the sensitivity of detection of F. graminearum and F. verticillioides when compared to the PCR results without capture. The q-PCR did not detect other mold when samples containing species of Byssochlamys, Penicillium, Trichoderma, and Ulocladium were included in the assay. Research is continuing on the improvement for the immunocapture step to improve the sensitivity. Thermal inactivation studies with five strains of Listeria monocytogenes were done using trypticase soy broth with yeast extract (TSBYE) and hot dog slurries containing different levels of fat. After several formulation trials, two hot dog slurry broths were developed with varying fat levels. Emulsifier levels, dilution effect, and hot dog ingredients were carefully used in the development of these formulations. Thermal inactivation studies were done at three temperatures (55, 60 and 65 C) in both the TSBYE and the hot dog slurries for the five strains individually and in a mixed cocktail. Since the most heat resistant strain did not alter the thermal inactivation curves, all five strains will be used in future research. These preliminary experiments have generated thermal inactivation models that will be statistically analyzed to develop the mathematical models and will be used to calculate D-values for each strain and log-reduction for the mixtures. In addition, data from the preliminary experiments will be used as the basis for doing thermal inactivation of L. monocytogenes inoculated onto the surface of hot dogs.
<BR> <BR>
IMPACT: 2003/10/01 TO 2004/09/29<BR>
Rapid methods are needed to detect molds and pathogenic bacteria in foods. This project will develop rapid methods to detect molds and pathogenic bacteria in foods and develop predictive microbiological models for bacterial pathogens in foods to control their growth or death during and after processing. There is a need to develop rapid methods to detect molds in grains and foods before they can produce mycotoxins that pose a health hazard when used in human food. This research on the immunocapture of Fusarium molds shows that the mycelia and spores can be linked to antibodies and then these mycelia and spores can be used in the polymerase chain reaction assay (PCR). The method has to be further refined because once these mycelia or spores are captured, the DNA needs to be released. Methods to do this are still being evaluated. Once this hurdle is overcome, then this method has potential for a rapid method for mold detection. Listeria monocytogenes has been implicated in foodborne disease and death from consumption of hot dogs and other processed ready-to-eat (RTE) meats. The current research on the thermal inactivation of L. monocytogenes in hot dog slurries can be used to develop predictive models for the processing of RTE meats that can be used in the USDA Pathogen Modeling Program as well as the USDA maintained ComBase.

<BR> <BR>
PROGRESS: 2002/10/01 TO 2003/09/30<BR>
Research has been done on the development of immunocapture methods for Fusarium graminearum and Fusarium verticillioides. Three different methods of immunocapture were studied, namely magnetic and nonmagnetic beads, microcentrifuge tubes and microtiter plates. The magnetic beads were not successful in binding the antibody needed for immunocapturing F. graminearum and F. verticillioides. The nonmagnetic beads based on protein A were able to bind up to 15.0 milligrams of antibody per milliliter of beads and successful immunocaptured both F. graminearum and F. verticillioides. This was confirmed by digital photos taken using 100X magnification and by positive polymerase chain reaction (PCR) results for the Tri6 and Fum5 genes, involved with trichothecene biosynthesis and fumonisin biosynthesis, respectively. The second method used a microcentrifuge tube to which the antibody was conjugated. Although the antibody was conjugated to the surface of the tube, the amount was too small to capture mycelial fragments, microconidia, or macroconidia from either F. graminearum or F. verticillioides. The third method used a 96 well microtiter plate to immobilize the antibodies for the immunocapture of F. graminearum and F. verticillioides. F. graminearum was immunocaptured and confirmed by positive PCR results for the Tri6 gene. The 96 well microtiter plate worked best for the immunocapture because a large number of samples could be prepared easily on one plate, the method was simple to do, and the sample set-up was similar to that used with the real-time PCR assay. Research is beginning on the immunocapture of F. verticillioides. This method will be optimized and used to detect Fusarium species in both inoculated and naturally contaminated grain and food samples. A new project was started on the use of predictive microbiology to study the thermal inactivation of Listeria monocytogenes in hot dogs. The cultures currently are being grown for use in the intial experiments.
<BR> <BR>
IMPACT: 2002/10/01 TO 2003/09/30<BR>
The ability to rapidly get molds out of grains and foods will be important for rapid method developmet. This research on the immunocapture will be one way to easily and rapidly pull Fusarium species away from the food for easy identification and enumeration. A polynmerase chain reaction (PCR) was developed by a colleague and will be combined with the immunocapture for the rapid detection of Fusarium species that produce mycotoxins. This should help the grain and food industries quickly determine if these molds are present and, thereby, allow them to prevent their growth and production of mycotoxins. Listeria monocytogenes has caused foodborne illness from the consumption of hot dogs and luncheon meats. The meat processing industries need ways to process these ready-to-eat foods after packaging. Most reseach has been done with growth of the cells in laboratory media. This research will focus on growth of the cells in the food and the use of predictive microbiology to determine a reliable thermal process for the meat.
<BR> <BR>

PROGRESS: 2001/10/01 TO 2002/09/30<BR>
The first phase of the project on the detection of Fusarium spoecies in grains and foods has been completed with the production of antibodies to Fusarium graminearum and Fusarium moniliforme. Since these two antibodies were able to detect many other Fusarium species with only minor cross-reactivities to two uncommon molds, the next phase of the project is to use these antibodies to capture the molds that can then be identifiied by PCR primers developed to these two mycotoxin-producing Fusarium species by a colleague. This research on the development of the immunocapture procedure is in the intial stages. Research is progressing on the evaluation of immunomagnetic beads, direct immunocapture on solid supports and the use of protein A-antibody columns.
<BR> <BR>
IMPACT: 2001/10/01 TO 2002/09/30<BR>
The development of a rapid way to detect molds such as Fusarium in grains and foods could help grain storage operators and food processors to screen for potential problems with mold growth and mycotoxin production. The combination of this method to capture molds in foods with molecular methods to identify the molds to the species level could benefit the grain and food industries.
<BR> <BR>

PROGRESS: 2000/10/01 TO 2001/09/30<BR>
The antibodies that were produced to Fusarium graminearum and Fusarium moniliforme were used to develop an indirect ELISA because the double sandwich ELISA had a high degree of nonspecific binding that was attributed to the antibody binding with the conjugate. In the indirect ELISA, the minimal detectable mold antigen in the carbonate buffer was 1 ng to 0.01 microgram/ml and 0.01 to 0.1 microgram/ml for F. graminearum and F. moniliforme, respectively. When 70 fungi from 22 genera were tested for cross-reactivity to the Fusarium antibodies in a competitive assay,only Monascus species and Phoma exigua showed reactions that could interfere with the assay. These two molds are not commonly found in foods; hence, they should not cause much false positive binding. When freeze-dried mycelium was added to cornmeal, F. graminearum could be detected at a minimum of 0.1 microgram/g and F. moniliforme, at a minimum of 1 microgram/g. Using the competitive ELISA, antigens to F. graminearum and F. moniliforme could be detected in cornmeal within 12 to 24 h after inoculation with 10 to 20 spores/g and incubation at 25 C. These ELISA readings corresponded to between 100 and 1000 CFU/g of each mold as determined by plate counts. The F. graminearum antibody was partially characterized using a Western blot, which showed that the antibody recognized the protein antigen plus the mycelium from F. graminearum and other Fusarium species. This suggests that the antibody could be used for immuno-capture of Fusarium species from foods or grains.
<BR> <BR>
IMPACT: 2000/10/01 TO 2001/09/30<BR>
The research on the detection of Fusarium species using the antibodies and antigens suggest that this research could be further developed to produce a method to detect Fusarium molds in foods and grains. This method could be used to collect Fusarium species from foods and grains and these molds could then be identified by other methods.
<BR> <BR>

PROGRESS: 1999/10/01 TO 2000/09/30<BR>
A predictive microbiological model for the inactivation of Listeria monocytogenes in different levels of milkfat (0%, 2.5%, 5.0%) and pH (5.0, 6.0, 7.0) at temperatures of 55,60 and 65 C was studied. The cells were grown and inactivated in media at the same pH and fat level. When the modifieid Gompertz equation was used to model death, temperature, milkfat and the interaction of temperature and milkfat significantly affected the shouldering and maximum death rate of the survival curves. For the cells grown in different levels of milkfat and at pH 6.0, the 4D-values were 30 to 100% higher than those obtained when cells were grown at pH 7.0 and 0% milkfat. These results indicate that models that have been developed from ideal laboratory media may underestimate the heat resistance of L. monocytogenes. Antibodies were raised to Fusarium graminearum and F. moniliforme in rabbits. These antibodies have been purified , conjugates have been made and enzyme-linked immunosorbent (ELISA)are being developed to be able to detect these molds rapidly in foods and grains. From preliminary research on the sensitivity and specificity, it appears that the minimium detection is 0.1 microgram and no other genera of molds or yeasts are detected.
<BR> <BR>
IMPACT: 1999/10/01 TO 2000/09/30<BR>
This microbiological predictive model shows that conditions used for the growth of bacteria, especially those that cause foodborne illness, are important in determining the heat resistance of microorganisms. Many of the current predictive models have been developed using ideal laboratory media that would not take into account different growth conditions in foods. The research with the production of a test to detect the Fusariun species suggests that a rapid method can be developed for early screening of foods and grains for these molds that produce many toxic substances.
<BR> <BR>

PROGRESS: 1998/10/01 TO 1999/09/30<BR>
To compare the first predictive microbiological model for the relationship of milkfat, pH and processing temperature on the thermal inactivation of Listeria monocytogenes in broth supplemented with milkfat, a second model was developed using the growth conditions of L. monocytogenes to determine heat resitance. Cells of L. monocytogenes were grown in broths with pH adjusted to 5.0, 6.0 and 7.0 and with 0, 2.5 and 5% milkfat. The same 3x3x3 factorial design was used in the inactivation studies with these pH and milkfat values and with heating at 55, 60, and 65 C. All three regions of the survival curve are being evaluated and these will be compared to the first model to determine if stress conditions affect thermal inactivation, especially the acid tolerance response. Green Colombian coffee beans were examined for the amount of internal mold contamination. There were 651 molds isolated from internal contamination of 26 samples of green Colombian coffee beans. A total of 41 different species in 15 genera were isolated and identified with Aspergillus niger var. niger, Aspergillus tamarii, Aspergillus flavus, Eurotium amstelodami, Penicillium crustosum, Penicillium citrinum and Syncephalastrum racemosum being predominant. None of the Aspergillus flavus or Aspergillus parasiticus produced aflatoxin as measured by thin layer chromatography.
<BR> <BR>
IMPACT: 1998/10/01 TO 1999/09/30<BR>
Microbial predicitve computer models are being used to determine the processing conditions to destroy pathogenic bacteria in foods. The information generated by this model could be used when developing pasteurization conditions for foods that contain fat. Colombia exports coffee to the United States and there is currently little known about the molds that are associated with green coffee beans. This research shows that although molds that produce aflatoxins are present, these strains do not produce toxin in coffee beans.
<BR> <BR>

PROGRESS: 1997/10/01 TO 1998/09/30<BR>
A predictive microbiological model was done using a 3x3x3 factorial design to determine the relationship of 0, 2.5, 5% milkfat, pH 5, 6, 7 and 55, 60, 65 C processing temperatures on the thermal resistance of Listeria monocytogenes in broth supplemented with milkfat. The data was evaluated using a modified Gompertz equation that plotted the survival curve including the shoulder, tail and maximum slope. The pH affected the shoulder; however, the temperature and milkfat affected the slope. When the model was validated, predictions were acceptable for temperatures above 62 c. This predictive model could be used to estimate death of L. monocytogenes in foods where fat and pH are important. Shelf-life studies were done with pourable low-fat ranch dressing with and without added nisin (an antimicrobial bacteriocin produced by Lactococcus lactis) at 5 ppm. At 25 c, the samples with nisin showed a slower increase in spoilage lactic acid bacteria than control samples for l to 3 weeks; whereas at 5 C there was no change in spoilage numbers over 65 days in nisin treated dressings compared to controls. The pH of the dressings decreased by 0.5 pH units or less during these shelf-life studies. Lactobacillus plantarum and Lactobacillus brevis were identified as the major spoilage microorganisms in these dressings. Strains of L. brevis were inhibited by 150 to 250 ppm nisin, but L. plantarum was not inhibited by nisin. Green coffee beans were obtained directly from Colombia, South America and from a coffee roaster in the United States after receiving shipments from Colombia. The beans were surface disinfected with 0.4% chlorine and directly plated onto four different agars to recover molds. Samples had from 10 to 100% infection rates with molds. Most of the mold isolates have been identified as species of Aspergillus, Penicillium, Fusarium, and Mucor.
<BR> <BR>

PROGRESS: 1996/10/01 TO 1997/09/30<BR>
Antibodies that were produced to Aspergillus parasiticus and Fusarium moniliforme were used in enzyme-linked immunosorbent assays (ELISA) to detect the presence of mycotoxin-producing molds in foods. A. parasiticus (100 spores/g) was inoculated into peanuts and incubated at 15 deg.C for 14 days and 21 deg.C for 7 days. At 21 deg.C, mold antigens were detected by day 2; however, no aflatoxin was present at this time. Likewise, at 15 deg.C, mold antigens were detected by day 4, but no aflatoxin was present. Aflatoxin was detected starting on day 3 and 6 at 21 deg. and 15 deg.C, repectively. Therefore, in peanuts A. parasiticus antigens could be detected before aflatoxin was present. Mycelial antigens of F. moniliforme were used to produce antibodies for use in ELISA. This antibody reacted with all Fusarium species and was not specific for F. moniliforme. The antigen to F. moniliforme was used to detect Fusarium species in naturally contaminated corn samples. All 25 samples showed the presence of Fusarium antigens; however, only three samples has fumonisin B1 detected. There was no apparent correlation between the ELISA reading and the presence of fumonisin B1.

<BR> <BR>
PROGRESS: 1995/10 TO 1996/09<BR>
Antibodies were produced to Fusarium poae for use in developing an enzyme-linkedimmunosorbent assay (ELISA). This assay detected all 17 Fusarium species tested. Four non-Fusarium molds showed cross reactions; however, these molds, which were species of Aspergillus and Penicillium, could not compete with the Fusarium species for binding sites in the competitive ELISA. Partially purified extracellular and mycelial antigens contained 16% protein, but the carbohydrate contents were 37 and 51%, respectively. These mycelial and extracellular antigens had 11 and 7% galactose, 40 and 24% glucose and 32 and 73% mannose, respectively. Enzymatic digestion of the F. poae mycelium suggested that protein, glucose and mannose were part of the immunodominant site. Commercial food products were surveyed using ELISAs for three different antibodies: Aspergillus parasiticus, Fusarium species, and a mixed antibody that detected several molds. Mold antigens were detected in 45 of 49 retail food products; low levels of A. parasiticus antigens were detected in 25 of the 49 foods; and 32 of the 49 foods reacted positively with the Fusarium moniliforme antibody. The mold counts of these foods ranged from <10 (est.) to 103 cfu/g of food and only six foods (nuts or grains) contained A. parasiticus or Fusarium species.

<BR> <BR>
PROGRESS: 1994/10 TO 1995/09<BR>
An antibody produced to Aspergillus parasiticus was used in an enzyme-linked immunosorbent assay (ELISA) to detect the presence of this mold in inoculated corn. The results from ELISA were compared to those from the ergosterol assay and plate count methods with ELISA giving the earliest detection. Also, the ELISA detected the A. parasiticus antigens before aflatoxin was produced to any significant level as determined by chromatographic and ELISA methods. As the level of aflatoxin continued to increase, the level of A. parasiticus antigen also incresaed. This suggested that the assay developed could be used to screen corn for early detection of aflatoxin-producing molds. Similar work is beginning done with inoculated peanuts to determine if the assay has early detection potential in different commodities associated with aflatoxin production. Antigens to fumonisin-producing molds, F. moniliforme and F. napiforme, and to F. poae are being purified using affinity chromatography followed by purity checks on gel electrophoresis. The F. poae antigens have produced two peaks that show activity by ELISA. The mycelial antigens from F. poae showed cross-reactivity with other Fusarium species and with Aspergillus wentii. In competitive inhibition assays, only reaction with Fusarium species has been shown so far. Additional competitive assays are being done with Aspergillus species. The antigens produced to F. moniliforme and F. napiforme only react with some other Fusarium species. Research is being done t.

<BR> <BR>
PROGRESS: 1993/10 TO 1994/09<BR>
Fusarium species were screened for fumonisin production on corn and two strains,F. moniliforme and F. napiforme, were chosen to produce antibodies in rabbits. The antibodies were produced after masking the antigens from these strains with common Fusarium antibodies. In addition, another antibody was produced to Fusarium poae to develop a general assay for Fusarium species because the first antibody cross-reacted with Aspergillus species. Research also continued with the development of an enzyme-linked immunosorbent assay (ELISA) for the detection of aflatoxin-producing Aspergillus species. Both the competitive inhibition and cross-reactivity assays showed reactions with A. parasiticus and A. flavus plus two closely related fermentation strains, A. soyae and A. oryzae. Several corn samples with aflatoxin ranging from 0 to 500 ppm were analyzed to determine if the mold antigen correlated with the aflatoxin level. When there was a small amount of antigen, there was a low aflatoxin level. Also, when a high antigen level was detected there was a high aflatoxin level. The increase in antigen level paralleled the increase in aflatoxin. Addition research was done with the mixed antibody that was produced to six genera of molds and it was shown that this antibody could detect molds of many other genera that were not used to produce the antibody. This research showed that a general antibody could be produced to detect nonspecific mold contamination of foods.

<BR> <BR>
PROGRESS: 1992/10 TO 1993/09<BR>
Enzyme-linked immunosorbent assays (ELISA) were developed to detect general and specific mold spoilage in foods. A general antibody was produced to detect species of Aspergillus, Cladosporium, Fusarium, Geotrichum, Mucor, and Penicillium. These molds could be detected in dairy products whether they were spiked into or growing in them. Plate counts compared with the ELISA readings when mixtures of four molds (Aspergillus versicolor, Cladosporium herbarum, Fusarium poae, and Penicillium chrysogenum) were inoculated into yogurt and cottage cheese and stored at 7(degree)C or 22(degree)C for 10 to 22 days. The ELISA readings could detect 10(superscript 3) cfu/g of mold or more. A specific antibody was produced to Fusarium poae. This antibody detected several different Fusarium species and only cross-reacted with Aspergillus versicolor and A. wentii. However, when a competition inhibition assay was done, both antigens from A. versicolor and A. wentii were inhibited by the antibody producing mold. F. poae could be detected in fruits, dairy products and grains to a level of 0.1 to 1 (mu)g/ml. When plate counts and ELISA readings were compared for detection of F. poae in cottage cheese, 10(superscript 3) cfu/g or more could be detected by the ELISA. Research is continuing on making the assay more specific and sensitive. An antibody was produced to Aspergillus parasiticus that was highly sensitive for the detection of the aflatoxin producing molds (A. flavus and A. parasiticus).
.
<BR> <BR>
PROGRESS: 1991/10 TO 1992/09<BR>
An enzyme-linked immunosorbent assay (ELISA) was developed to detect mold spoilage in foods. One antibody was produced that detected species of Aspergillus, Cladosporium, Fusarium, Geotrichum, Mucor, and Penicillium. When cottage cheese, yogurt, and Cheddar cheese were spiked with these molds, the ELISA readings increased as the amount of mold contamination increased. Yogurt was inoculated with mold spore suspensions of these six molds and incubated at 7deg.C for 3 weeks. The ELISA readings showed an increase after 6 days. Research is being done on the detection of these molds in foods and their recovery. The antibody to Fusarium poae was further characterized. It was specific for only antigens from other Fusarium species and had a sensitivity of 0.1 micro gram/ml in both buffer and food. Further purification of Fusarium antigens is being done to be able to analyze the chemical composition and determine the immunodominant site. An antibody was produced to Aspergillus parasiticus that was highly sensitive for the detection of the aflatoxin producing molds (A. flavus and A. parasiticus). Competitive assays were done with these mold antigens in the presence of other mold antigens and the assays were very specific at detecting aflatoxin-producing antigens. Grains and nuts were spiked with 1 mg/ml of the A. parasiticus and recovery using the ELISA ranged from 70 to 100%. Research is continuing on the recovery of the aflatoxin-producing molds from foods by ELISA.

<BR> <BR>
PROGRESS: 1990/10 TO 1991/09<BR>
An antibody was produced to a mixture of Aspergillus versicolor, Cladosporium herbarum, Fusarium poae, Geotrichum candidum, Mucor circinelloides, and Penicillium chrysogenum. The four molds with the highest antibody titers are being evaluated, namely, A. versicolor, C. herbarum, F. poae, and P. chrysogenum. The mold antigens from these four molds were used to determine if the mixed antibody could detect them. With an antibody-conjugate ratio of 10/20 all four molds could be individually detected by this antibody. The sensitivity for all enzyme-linked immunosorbent assays (ELISA) was 1 (mu)g/ml or less. Procedures were evaluated for the purification of the antibodies and protein A affinity chromatography gave the best results. Antibodies were produced to spores and extracellular extracts of Aspergillus flavus, Aspergillus ochraceus, and Aspergillus parasiticus. Antigens from these three molds cross-reacted with the antibodies from all three species. This suggests that these three mycotoxin producing molds have very similar spores and cellular components. The antibodies produced to these molds were purified using protein A affinity chromatography. These antibodies were conjugated to horseradish peroxidase and ELISA evaluations are being done.