|
CFSAN/Office of Food Safety, Defense, and Outreach
June 2006
Following the conduct of vulnerability assessments with industry on a variety of foods regulated by the Food and Drug Administration a number of research questions were generated. The commodities evaluated were dairy products, fruit juices, bottled water, water used for food processing, and infant formula. The research questions fell into the following general categories:
A summary of the main research results is provided below. Specific details on each project follows in subsequent sections.
If you require additional information on any of the research projects discussed in this report, please contact LeeAnne Jackson at LeeAnne.Jackson@fda.hhs.gov or call 301-436-1593.
Abrin is a class II, ribosome inactivating protein (RIP-2) isolated from the seeds of the tropical plant Abrus precatorius also known as rosary peas, jequirity beans, crab's eye, and precatory bean. The effects of pasteurization on the ability to detect abrin in a crude extract of A. precatorius in milk were examined and compared to the effects of pasteurization on toxicity (oral and intraperitoneal) in female BALB/c mice. Pasteurization parameters of 30 minutes at 145 °F (LTLT) and 15 seconds at 161 °F (HTST) were studied.
Both forms of pasteurization examined had no effect on the ability to detect abrin in milk. The oral and intraperitoneal toxicity of rosary pea extract in raw milk and following HTST and LTLT were determined for BALB/c female mice. The low number of mice per group (n = 3) along with the high apparent oral LD50 (i.e. low oral toxicity), estimated at > 1 mg per kg body weight, precluded an extensive titration and detailed estimation of the oral LD50. In contrast, the higher intraperitoneal toxicity of rosary pea extract enabled a more precise estimation of an LD50 and comparison between the effects of pasteurization on toxicity. The i.p. toxicity of the rosary pea extracts were not significantly affected by pasteurization and were between 3 and 9 ug/kg body weight. The pasteurization parameters evaluated cannot be relied upon to inactivate abrin if present in milk.
Ricin is a water soluble protein toxin produced from Ricinus communis (castor beans). Alpha-amanitin is a water soluble protein toxin from a poisonous mushroom, Amanita phalloides. T-2 toxin is a lipid soluble polypeptide toxin produced by a group of mycotoxin-producing molds, Fusarium acuminatum, F. poae, and F. sporotrichioides. The goal of this study was to determine the partitioning of ricin, α-Amanitin and T-2 toxin into the aqueous and lipid phases of cream, skim milk, cheese and whey during the processing of whole milk. Occurrence and quantities of the toxins in the phases were determined by immunochemical methods (BUHLMANN EK-AM1-Amanitin; Tetracore ELISA-Ricin; Veratox® and Ridascreen®fast-T-2 toxin). Partition (distribution) coefficients were expressed as the common logarithm of the quotient of the total quantities of the toxin in the two major phases of each process or food type.
The distribution coefficients for α-Amanitin, Ricin and T-2 toxin in cream and skim milk were -1.53 (95% CI: -1.62, -1.45); -1.27 (95% CI: -1.62, -1.45); and -0.17 (95% CI: -0.23, -0.12), respectively. Between cheese and whey, the distribution coefficients were determined to be -0.80 (95% CI:-0.84, -0.75); -0.68 (95%CI: -0.89, -0.46); and 0.08 (95%CI: 0.00, 0.17) for α-Amanitin, Ricin and T-2 toxin respectively. The distribution coefficients were not adjusted for differences in volumes or mass to reflect practical similarities with industrial processing.
The distribution of each of the toxins in the major phases of each food commodity could be attributed to solubility properties. The water soluble octapeptide, α-Amanitin largely [94%; (95%CI: 81,107%)] isolated into the skim (aqueous) portion of milk with approximately 2-4% going into the cream. The quantity of α-Amanitin that isolated into the cream could be attributed to solubility in the skim component (40-65%) of the cream. Ricin, a 66kD protein, is also water soluble and largely (86%) isolated in the skim (aqueous) phase of the cream separation process with approximately 4-5% partitioning into the cream. On the other hand, a substantial amount (50 ± 8%) of the relatively more lipophilic T-2 toxin partitioned into the cream portion of milk yielding a distribution coefficient of -0.17 (95%CI:-0.23, -0.12).
The objective of this study was to determine the partition coefficients for aconitine, nicotine, and strychnine in raw milk. Since milk consists of skim milk and cream phases, it is of interest to determine into which phase the toxin will partition. An HPLC method for analysis of aconitine, nicotine, strychnine from milk products was developed. Sample clean up techniques consisted of liquid-liquid partitioning (hexane/water-acetonitrile), solid phase extraction (OASIS HLB), and manipulation of pH of sample to avoid volatility and hydrolysis losses. Analysis was conducted with an HPLC with dual band UV detector. Nicotine and strychnine levels were measured at 260nm and aconitine at 232nm. Centrifugation of whole milk was used to simulate commercial separation. Whole milk was placed into 50 ml centrifuge tubes, spun for 30 minutes at 2000 x g and 5°C. Skim milk from tubes were decanted and mixed. Cream layer adhering to the side of bottles was dissolved and mixed together. The mixed samples were used for fat testing or if spiked, for testing of toxins. Centrifugations were conducted at 30 minutes and 5 days after spiking to simulate contamination in the plant and on the farm, respectively.
Mean recoveries for the three analytes in skim milk, whole milk and cream ranged from 72.1% to 89.2%. Centrifugation of 3.25% whole milk resulted in a fat content of 39.5% and 0.07% for cream and skim milk, respectively. Whole milk was spiked with 1 ppm of aconitine, nicotine and strychnine. Each of the three toxins was found in both cream and skim milk streams. Initial partition coefficient (toxin conc. in cream/toxin conc. in whole milk) for aconitine was 0.769 in cream, and increased to 1.121 by day 5. Initial partition coefficient for nicotine was 0.49 in cream, and increased to 0.761 by day 5. Initial partition coefficient for strychnine was 1.064 in cream, and increased to 1.135 by day 5. Between day 0 and day 5, partition coefficients for the toxic compounds in skim milk decreased.
A HPLC-UV Method was developed for analysis of 3 nitrogen alkaloids from milk products. Centrifugation of milk was used to simulate commercial separation. Aconitine and nicotine were recovered in both the cream and skim milk streams. Initial recovery was low in cream; however, by day 5 the level had increased in cream for both compounds. Strychnine was also recovered in the cream and skim milk streams. A slight increase concentration was seen in cream, however, by day 5 the level further increased in cream. This research shows that separation cannot be used to remove any of these three nitrogen alkaloids from contaminated whole milk.
The acute toxicity of sodium arsenite (As), when administered in half and half unpasteurized cream (HH) by gavage, was assessed in Charles River Sprague-Dawley rats. Male and non-pregnant female rats received a single oral dose of As in HH at doses of 0.41, 4.1, 41.0 and 410.0 mg/kg body weight. Pregnant rats received a single oral dose of As in HH at doses of 0.41, 4.1 and 41.0 mg/kg body weight on gestation day 10 (GD- 10). Control rats received deionized water alone, HH alone or 41.0 mg/kg As in deionized water (41 mg/kg As-water). Male and non-pregnant female rats (n=8 rats/group) were observed for 14 consecutive days after dosing. Pregnant rats (n=8 rats/group) dosed on GD 10 were observed until GD 20 when fetuses were collected.
All male and non-pregnant female rats placed in the 410.0 mg/kg As-HH treatment group died within 8 hours of exposure. Two of eight non-pregnant female rats placed in the 41 mg/kg As-water treatment group died within 24 hours of exposure. Animal deaths were not observed in any other control or treatment group. A statistically significant increase in the number of male and non-pregnant female animals exhibiting diarrheal soft stool was observed in both the 41 mg/kg As-HH and 41mg/kg As-water groups in comparison to their respective controls (HH alone or deionized water alone). Statistically significant differences were not observed between the 41.0 mg/kg As-water control group and the 41.0 mg/kg As-HH treatment group. Adverse effects were not observed in pregnant animals from either the control or As treatment groups.
In summary, long term effects of As exposure were not observed in male, non-pregnant female and pregnant rats after the animals received a single oral dose of sodium arsenite in half and half. Female rats and female fetuses appeared to be more sensitive to the effects of As when As was administered in water. However similar effects were not observed when As was administered in HH. Finally, all animals receiving 410 mg/kg As HH died shortly after compound administration suggesting that the oral toxicity of As was not reduced when administered in HH. Although a 410 mg/kg As water control was not run concurrently, this dose is 10 times the LD 50 for rats and it was assumed that animals receiving this concentration in water would also die. Since no animals survived these findings suggest that HH did not reduce the oral toxicity of As when administered at a dose of 410 mg/kg. However, at lower doses (41 mg/kg body weight) HH did appear to reduce the oral toxicity of AS on the female and developing fetus.
Little information exists about the survival and/or growth of B. anthracis in foods. This study was undertaken to determine the thermal resistance of B. anthracis spores in three juices. Apple, orange without pulp, and calcium-fortified orange juices were inoculated with spores (ca. 108 spores/ml), subjected to heat treatments (79.4, 85, 90 and 93.3°C) and plated on tryptic soy agar supplemented with sheep blood.
In general, the inactivation of B. anthracis spores was similar in all three juices. The spores declined slowly when heated at 79.4°C. There was approximately a 2 log reduction after heating for 60 min. A greater than a 4-log reduction was observed after heating at 85°C for 30 min. A nearly 5-log reduction was noted after heating at 90°C for 8 min and at 93.3°C, a 5-log reduction was achieved in 4 min.
B. anthracis spores had the highest D-values in calcium-fortified orange juice as compared to the other juices tested. D-values in all three juices were 25.56-36.15 min at 79.4°C, 5.27-6.20 min at 85°C, 1.61-1.82 min at 90°C, and 0.72- 0.79 min at 93.3°C. The z-values ranged from 8.3°C in orange juice to 8.9°C in apple juice.
| Temperature | Apple juice (avg ± s.d.) |
Orange juice (avg ± s.d.) |
Orange juice-Calcium (avg ± s.d.) |
|---|---|---|---|
| 79.4°C (175°F) | 25.56 ± 5.4 | 33.16 ± 6.8 | 36.15 ± 7.7 |
| 85°C (185°F) | 5.88 ± 0.9 | 5.27 ± 0.7 | 6.20 ± 1.2 |
| 90°C (194°F) | 1.70 ± 0.3 | 1.61 ± 0.2 | 1.82 ± 0.1 |
| 93.3°C (200°F) | 0.72 ± 0.0 | 0.77 ± 0.1 | 0.79 ± 0.1 |
| z-value | 8.87 | 8.42 | 8.31 |
The times and temperatures required to destroy B. anthracis spores in juices clearly exceeds commercial pasteurization processes. Based on these results, to achieve a 5-log reduction in B. anthracis Sterne spores in any of the juices tested, heating for a minimum of 127-180 min at 79.4°C (175°F), 26-31 min at 85°C (185°F), or 8-9 min at 90°C (194°F) would be required. Application of other alternatives in juice processing will be necessary to inactivate B. anthracis spores if present.
Shiga toxins and Shiga-like toxins (Stx) are a relatively large group of cytotoxins produced by certain serotypes of Shigella and shiga-toxin producing Escherichia coli (STEC). They are responsible for the induction of hemolytic uremic syndrome, a sometimes fatal condition with especially severe consequences in young children. This study examined the inactivation of Shiga Toxins 1 and 2 (Stx1 and Stx2) by acidic or alkaline pH as well as temperature.
The induction of apoptosis in the human monocyte cell line, THP-1, was used as a biological endpoint for inactivation of Stx1 and Stx2 by acidic or alkaline pH. The flow cytometry method using PE-annexin V and 7-AAD fluorescent staining is sensitive and highly quantitative, in addition to being widely accepted as a marker for early apoptotic events.
Aliquots of spiked fruit punch were heat treated at 95°C for 10 minutes and tested in parallel with unheated aliquots from the same spiked sample. For both Stx 1 and 2, complete inactivation was seen at acid, neutral and alkaline pH.
Fruit punch was adjusted to pH ranging from 2-9 with either 1N HCl or 1N NaOH and the toxins were added at 1:100 dilution of the original stock. These punch samples were incubated at 4 or 20°C for 0-90 days. Aliquots (0.1 ml) were taken at different time points to test in THP-1 cell apoptosis assay. A standard curve was generated at each time point using the same toxin stored at -80°C for comparison and for the assessment of toxin inactivation.
Samples of fruit punch were spiked with Stx 1 and 2 at a 1:100 dilution from the original toxin stocks. This corresponded to the highest concentration in the dilution curve. For the degree of inactivation to be equivalent to a 1000 fold dilution, the level of apoptosis/necrosis induced form spiked samples would have to be less than of equal to the apoptotic effects seen with a 1:100,000 dilution on the curve. With Stx-1 at 4°C no inactivation at the 1000 or 100 fold level was seen at any point out to 90 days of storage. There was a detectable decrease in activity that was apparent at pH 2-3. With Stx-1 stored at 20°C, inactivation equivalent to a 100-fold dilution was seen by 12 days of storage at pH 2. At neutral and alkaline pH, less than 100-fold decrease in activity was seen. Results from Stx-2 at 4°C and 20°C, while similar to those seen with Stx 1, indicated that there was somewhat less inactivation. At no point was the level of inactivation at the 100 or 1000-fold level. However, pH, temperature and storage time effects on apoptosis were seen. The most inactivation was seen with the combination of acid pH and higher storage temperature at the 30 Day time point.
Stx 1 and 2 are relatively immunologically stable across a pH range from 2-9 when spiked into fruit punch. There was greater loss of biological activity at acid pH than at neutral pH. The combination of acid pH and higher storage temperature resulted in the greatest degree of toxin inactivation. This loss of activity was estimated to be between 10 and 100 fold at low pH.
Francisella tularensis is a gram negative bacterium that can cause gastrointestinal or oropharyngeal tularemia from ingestion of contaminated food or water. Despite the potential for accidental or intentional contamination of foods with this organism, little information exists on the thermal stability in specific foods. In the current study, the effects of three food products (apple juice, mango juice and orange juice) on the thermal stability of the Live Vaccine Strain (LVS) of Francisella tularensis at four different temperatures were investigated.
Survivor curves were calculated for F. tularensis LVS heated at 57.5, 55, 52.5 and 50°C in apple juice. D-values ranged from 9 (57.5°C) to 53 (50°C) sec in apple juice. LVS failed to survive at temperatures above 55°C in mango juice and 53°C in orange juice. Survival curves were calculated at temperatures of 47.5, 50, 52.5 and 55°C for mango juice and 45, 47.5, 50 and 52.5°C for orange juice. D-values at 55°C to 47.5°C were 10 to 62 sec in mango juice and 12 to 143 sec at 52.5°C to 45°C in orange juice. Z-values for LVS at temperature ranges of 47.5°C to 55°C were 9.51 and 6.67 in mango juice and orange juice, respectively.
This study is the first to determine thermal inactivation of F. tularensis in specific foods and will allow comparisons with the thermal inactivation data of other more traditional food pathogens to determine whether similar pasteurization parameters are required.
In conclusion, F. tularensis LVS is susceptible to relatively mild heat treatment in and thus should not survive after standard pasteurization processes employed for other more traditional enteric pathogens.
A literature review indicates virulent Yersinia pestis will survive and grow at low pH. However, initial studies showed non-virulent Y. pestis strains did not survive in juice. Consequently, a virulent Y. pseudotuberculosis strain was chosen as a surrogate for virulent Y. pestis. The thermal resistance of Y. pseudotuberculosis was measured in orange and apple juice and in juice concentrates.
The thermal resistance in single strength apple juice was similar to that measured in buffer at similar pH values. The thermal resistance in concentrated apple juice was significantly lower (approximately 10X) from that of single strength apple juice and buffer at similar pH. Populations of Y. pseudotuberculosis in concentrated orange juice at 10°C showed greater than a 3-log10 reduction after one hour; consequently further testing at higher temperatures was not done.
| Temperature (°C) | Apple Juice | Orange Juice | Apple Juice Concentrate* | Buffer pH 7.0 |
|---|---|---|---|---|
| 49.9 | 4.94 ± 0.93 | 2.82 ± 0.76 | 0.37 | N.D. |
| 51.9 | 2.15 ± 0.71 | 1.18 ± 0.23 | 0.17 | N.D. |
| 53.9 | 0.99 ± 0.15 | 0.71 ± 0.20 | N.D. | 22.99 ± 4.54 |
| 55.8 | 0.39 ± 0.05 | 0.24 ± 0.10 | N.D. | 9.22 ± 2.54 |
| 57.8 | 0.18 ± 0.04 | 0.12 ± 0.03 | N.D. | 2.57 ± 1.37 |
| z-value | 5.45 ± 0.50 | 5.52 ± 0.29 | 5.29 | 3.75 ± 0.25 |
*Two trials only.
The thermal resistance of Y. pseudotuberculosis was significantly less in apple and orange juice than in buffer at neutral pH. This change is clearly related to pH as the D- and z-values obtained in both single strength juice products were similar to that of buffer with a similar pH. The pH, however, does not explain the further loss in heat resistance in concentrated juices. In concentrated juices, the pH was not different from single strength juices; however, the thermal resistance of the microorganism was dramatically lower. This likely reflects the difference in the concentration of the acids and the accompanying anion effect.
Currently, juice manufacturers are required to treat juice to provide a 5-log10 reduction in the most resistant pertinent pathogen. In general, the pertinent pathogens would be either E. coli O157:H7 (apple juice) or Salmonella (orange juice). Both of these microorganisms have greater thermal resistance then Y. pseudotuberculosis in juice. A published D-value at 52°C for E. coli O157:H7 in apple juice/cider is 18 min. The corresponding z-value published for E. coli O157:H7 is 4.8. This value is similar to that found for Y. pseudotuberculosis in this study (5.45 ± 0.50), consequently, a thermal process designed to provide a 5-log10 reduction of E. coli O157:H7 in juice would also inactivate any Y. pseudotuberculosis present. Y. pseudotuberculosis would not survive a normal pasteurization process for a juice product.
The stability of six compounds (aconitine, colchicine, strychnine, monofluoroacetic acid, nicotine sulfate and sodium nitrite) in fruit punch to changes in pH (pH 2 to 9) was determined. The stability protocol included storage at 5 and 20 °C and time point sampling from 0 hours to 90 days. LC-UV and LC-MS were used to assess analyte stability versus positive control samples (spiked water or mobile phase).
Of the six analytes tested, four (colchicine, strychnine, monofluoroacetic acid, and nicotine sulfate) were stable in fruit punch at all conditions tested. Of the two remaining analytes, both aconitine and sodium nitrite showed some level of degradation at a number of the test conditions. However, neither showed complete degradation under all of the conditions. Aconitine was shown to be stable at pH 6 and below. However at pH 7 and above the extent and rate of degradation was seen to be directly proportional to storage temperature, but inversely proportional to the hydrogen ion concentration. For pH 7 and above, storage at 5 °C showed only partial degradation over the 90 day period, while at 20°C aconitine completely degraded at or before the 90 day stability testing point. At both temperatures, the rate of degradation increased with pH. The effect of pH was most obvious with the 20 °C storage conditions. The pH 9 samples show complete degradation by day 10, but the pH 7 samples are not fully degraded until day 90.
Sodium nitrite recovery showed a similar dependence on storage temperature, with the samples stored at 20 °C showing a more rapid degradation than those stored at 5 °C. However, the pH dependence noted for sodium nitrite was opposite that seen for aconitine. For sodium nitrite, the extent and rate of degradation increased with hydrogen ion concentration.
Based on the results of this stability study, the degradation of the tested chemical agents occurs only in very limited conditions and is usually a slow (days) process. Sodium nitrite at pH 2, is the only exception, with degradation occurring too fast (minutes) to collect accurate recoveries for the 0 hour time point. When degradation does occur, the results showed that increased storage temperature increases the rate of degradation. Therefore, in the case of aconitine, when degradation was observed it is unknown if the resulting products were less toxic than the starting material.
The purpose of the study was to determine the effect of various toxic chemicals on the conductivity of liquid foods and the ability to detect a lethal or half-lethal dose of an agent. Sodium monofluoroacetate, sodium cyanide, sodium nitrite, sodium arsenite and strychnine sulfate were evaluated. The following liquid foods were evaluated: bottled water, orange juice, apple juice, lemonade, passion fruit juice, cranberry juice cocktail and apricot nectar. Aliquots of foods were spiked with chemical agents at several concentrations ranging from 0.1 to 5 mg/mL followed by a conductivity measurement after 3 hours.
A lethal dose of sodium arsenite and strychnine sulfate was unable to be detected by conductivity in cranberry juice and lemonade, while strychnine sulfate was unable to be detected in apricot nectar. An unexpected observation was that many of the agents, except sodium monofluoroacetate and strychnine sulfate, yielded a change in the color of the beverage. None of the agents yielded a color change for bottled water. A change in conductivity was observed in several fruit juices and bottled water after various toxic chemicals were added.
In many cases, color changes could be observed after agent addition which would also support the fact that reactions are taking place. The color change varied by sample and also by brand for a juice. It is interesting to note that in some cases, such as nitrite addition to apple juice, the greatest color change was observed for the lowest spike level. The color changes can be summarized as follows:
A change in color can be an indication of the inclusion of an agent, however, strychnine and sodium fluoroacetate did not usually produce a color change. Samples with highest spiked concentrations resulted in a noticeable color change. The belief is that in-line sensors at food plants might be useful for monitoring.
Conductivity does appear to be a means of indication of toxic chemical agent addition to fruit juice and bottled water. The method was able to measure a change in conductivity in several fruit juices and bottled water after various toxic chemical agents were added. Agent addition at concentrations equal to a lethal dose per serving were able to be detected in most cases and in many cases agent addition at one half the lethal dose per serving concentration could be detected. Each brand of a sample should be evaluated on its own because of the wide variation in conductivity between brands. However, the pooled results and conclusion from this study still provides valuable insight to industry on a simple, low cost method to monitor their food process stream for possible toxic agent addition.
Terrorist threats have precipitated the need for information on the resistance of uncommon pathogens to food processing. No studies exist measuring the effect of UV treatment on Yersinia pestis. The objective of the study was to characterize resistance of Y. pestis to UV treatment in apple juice and water using a single lamp annular UV reactor under turbulent and laminar flow regimes. Three commercial brands of apple juice were purchased from a local distributor. The UV reactor consists of four annular sections with lamp lengths of 20, 40, 87.5, and 87.5 inches. Treatment samples were taken between each section and at the end. Flow rate was controlled with a peristaltic pump and measured with a flow meter. Measured viscosity and density were used to determine flow conditions needed for turbulence. To determine the inactivation coefficient the finite element program was used to model the treatment dose and flow profile.
The suitability of surrogates for Yersinia pestis was assessed with respect to the ability to survive in apple juice at pH of 3.67 for extended periods of time. Strains of Yersinia pestis 1122 and Yersinia pseudotuberculosis were tested. Cultures were suspended in apple juice and sampled every two hours during 24 hours. An attenuated strain of Y. pestis performed poorly in apple juice at pH 3.57 showing a 4.4 log reduction in only 6 hours. Yersinia pseudotuberculosis yielded a 1.34 log reduction after 6 hours. The latter was chosen as the surrogate. The inactivation tests of Y. pseudotuberculosis were conducted in model caramel solutions and three brands of apple juice in laminar and turbulent regimes respectively. The inactivation rate of Escherichia coli K12 (ATCC 25253) was measured in the same reactors by the same method. A value for the rate constant k=0.325 with the coefficient of determination R2= 0.907 was obtained. It shows that Y. pseudotuberculosis is less resistant to UV than Escherichia coli K12.
Y. pestis 1122 (attenuated strain) did not survive in juice at pH 3.57 (4.4 log reduction in 6 hours). Yersinia pseudotuberculosis proved more pH resistant (1.34 log reduction in 6 hours), which was used as the surrogate. Under laminar flow conditions water with 0.1% caramel exhibited reduction of 1.5 logs after a 127 sec treatment. Inactivation in water with 0.05% caramel reached 5-logs after approximately 60 seconds of residence time. When UV fluence and residence time were accounted for the inactivation rate was determined to follow a first-order reaction model. For turbulent flow conditions (Reynolds No.~2000) the reduction in Y. pseudotuberculosis was severely limited when juice with added vitamin C was treated. There was over a 50% loss in vitamin C after one complete pass. For one pass the vitamin C free juice yielded ~0.6-log reduction. At turbulent flow conditions (Reynolds No.~4000) the reduction in clear water was complete (>5-log inactivation).
For clear juice, a properly designed UV treatment system appears to be able to eliminate Yersinia type species. Juices enriched with vitamin C require significantly higher doses of UV. If vitamin C is added to juice post-processing, reprocessing of the juice needs to include vitamin C effect in the dose requirements.
Picrotoxin is an equimolar mixture of toxic picrotoxinin and non-toxic picrotin and is found in the dried fruits of Anamirta cocculus (fish berries) distributed in south-eastern Asia (particularly the Malabar coast of India) and the East Indies. The lowest published lethal dose for picrotoxin (LDLo) in humans by oral administration is 357 µg/kg. There is a lack of information about the thermal stability of picrotin and picrotoxinin. Thermal stability of picrotin and picrotoxinin (components of picrotoxin) in 100% apple juice was tested at various processing temperatures and times. Juice samples were spiked at 5 µg/ml or 10 µg/ml. Three replicates of spiked apple juice samples (1 ml) and blank apple juice (1 ml) were preheated at 63, 72, 80, 90, or 100 °C for 10 min and then incubated further for 15, 30 sec, 1, 15, or 30 min.
The results suggest these compounds are very stable at typical food processing temperatures in apple juice that is acidic (pH 3.7). Picrotin and picrotoxinin were decomposed quickly in alkaline solution. At pH 10, both compounds almost disappeared within four hours at room temperature.
The results confirmed that picrotin and picrotoxinin were very stable in 100% apple juice. There were no significant differences about thermal stability of either analyte. Both toxins were very stable at typical food processing temperatures and times. These results suggest that picrotoxinin and non-toxic picrotin are very stable in acidic solution and unstable in alkaline solution. The basic/alkaline cleaners and disinfectants could be very useful not only to neutralize contaminated food products but also to clean contaminated product lines.
Francisella tularensis is a gram negative bacterium that can cause gastrointestinal or oropharyngeal tularemia from ingestion of contaminated food or water. Despite the potential for accidental or intentional contamination of foods with this organism, little information exists on the thermal stability in specific foods. In the current study, the effect of liquid infant formula on the thermal stability of the Live Vaccine Strain (LVS) of Francisella tularensis at four different temperatures were investigated. LVS was heated in a submerged coil heating apparatus in infant formula at temperatures ranging from 45°C to 65°C to obtain preliminary thermal inactivation data. Linear regressions were then conducted for the average log10 cfu/ml survivors of LVS from duplicate platings from experiments performed in duplicate. D- and z-values were calculated from these results.
Survivor curves of LVS heated in infant formula failed to display a log-linear decrease in surviving bacteria. That is, an initial lag phase was seen in which the numbers of surviving LVS in the original inoculum remained constant for a specific amount of time prior to the subsequent exponential decrease in survivors. LVS did not survive at temperatures above 58°C in any food tested (data not shown) and hence linear regressions could not be calculated.
Survivor curves were calculated for LVS heated at 57.5, 55, 52.5 and 50°C in infant formula. D-values ranged from 12 (57.5°C) to 580 (50°C) sec in liquid infant formula with a z-value of 4.37. The greater heat tolerance exhibited in infant formula may be due to the high fat content found in that product. This is consistent with previous reports of food borne bacteria displaying a greater resistance to heat in high fat foods. The higher heat resistance of some bacterial pathogens of food has been attributed to a lower water activity in high fat foods. A greater heat resistance of Salmonella, Listeria monocytogenes, E. coli O157:H7, Staphylococcus aureus and Streptococcus has been reported in environments with lower water activity levels.
In conclusion, F. tularensis LVS is susceptible to heat and thus should not survive after standard pasteurization processes employed for other more traditional enteric pathogens.
Ricin is a potent protein toxin found in the seeds of the castor bean plant, Ricinus communis. Several reports indicate that ricin can be detoxified by thermal treatment, however the conditions required for inactivation are not well characterized. In addition, little information exists on the thermal stability of ricin added to foods. The objective of this work was to determine the effects of heat treatments on the detection and toxicity of ricin added to milk- and soy-based infant formulas. Reconstituted infant formula powders containing 100 µg ricin/mL were heated at 60-90°C for up to 5 h. The heat-treated formulas were analyzed by ELISA to determine levels of detectable ricin. The residual cytotoxicity of ricin-containing infant formula after heat treatments was determined in an anchorage-dependent transformed macrophage cell line (RAW264.7 macrophage cells).
The effects of processing time and temperature on the residual cytotoxicity of ricin in an intact milk protein-based infant formula and a soy protein-based infant formula was determined. The rate of loss in ricin cytotoxicity or ELISA detection is highly dependent on temperature and in general, the extent of loss of ricin increases with processing temperature and time. Minor losses of ricin cytotoxicity and detection were found in the infant formulas processed at 60°C for < 2 h. At 90°C, over 90% loss in ricin cytotoxicity and ELISA detection was found only after > 4 min processing.
The ELISA and the cytotoxicity assay indicated that ricin detection and toxicity decreased with increasing heating times and temperatures. Minimal losses in detection and toxicity were found for ricin heated at 60°C for 2 h. The half-lives of ricin cytoxic activity in a milk-based infant formula at 60, 70, 75, 80, 85 and 90°C were >100 min, 7.2 ± 0.3 min, 5.4 ± 0.5 min, 4.5 ± .5 min, 3.9 ± 0.4 min and 3.1 ± 0.1 min, respectively. The half-lives of ricin cytotoxic activity in a soy-based infant formula at 60, 70, 75, 80, 85 and 90°C were >100 min, 14.9 ± 1.4 min, 7.5 ± 0. 7 min, 5.5 ± 0.5 min, 3.9 ± 0.3 min and 1.9 ± 0.2 min, respectively. The results indicate that ricin is a relatively heat stable protein and may remain toxic under some food processing conditions.
Infant formulas mixes that are used in the production of powdered infant formulas are typically pasteurized under more extreme thermal conditions (higher temperatures and longer hold times) than fluid milk due to their higher solids and fat content. However, it is unlikely that even these conditions would fully inactivate ricin. After pasteurization, the infant formula mix undergoes a spray drying step. Although high air temperatures (up to 250°C) are used during spray drying, the product typically does not exceed 60°C during the procedure. Consequently, spray drying is unlikely to cause inactivation of ricin. In conclusion, infant formula spiked with ricin at any stage of manufacture, would likely have active toxin in the final, spray-dried product.
The objective of the study was to determine the length of time for the test medium to remove toxin, the amount of toxin removed, and to determine what happens when contaminated water flows through the medium. Agents for analysis were selected based on their water solubility and availability. Filters selected were Braun and Pur Ultimate filter. Each filter had a carbon and resin-based form. Compounds selected were strychnine, colchicine, and nicotine sulfate (40%). Solutions of each agent were prepared to approximately 10mg/ml.
Agents removed (µg/mg) by filter comparison is as follows:
| Agent | Braun | Pur | ||
|---|---|---|---|---|
| Carbon | Resin | Carbon | Resin | |
| Strychnine | 44 | 44 | 46 | 46 |
| Colchicine | 36 | 35 | 29 | 40 |
| Nicotine Sulfate | 44 | 39 | 44 | 33 |
The Pur carbon filter was able to remove approximately 60%, 90%, and 95% of colchicine, nicotine sulfate, and strychnine, respectively, after approximately 200 minutes of exposure. The Pur resin filter was able to remove approximately 80%, 75%, and 95% of colchicine, nicotine sulfate, and strychnine, respectively, after approximately 200 minutes of exposure. The Braun resin filter was able to remove approximately 75%, 85%, and 95% of colchicine, nicotine sulfate, and strychnine, respectively, after approximately 200 minutes of exposure. Carbon and resin-based filters removed all toxic agents to some extent. Carbon-based filters removed slightly more agent and resin-based filters worked faster than granular carbon.
Micro-filtration processes using membrane filters with pore sizes of 1-5 mm have been commonly used as part of treatment procedures for bottled water manufacturing. This research evaluated the efficacy of such micro-filtration in removing Escherichia coli O157:H7, Salmonella, and Yersinia pestis from bottled water and from produce rinse water. A 4-strain cocktail of E. coli O157:H7 (F4637, F4546, 960212, 960218), 4 serotypes of Salmonella (S. Tennessee, S. Muenchen, S. Cubana, and S. St. Paul) and an avirulent strain (A1122) of Y. pestis were used for this study. The filters evaluated in this study were the Polycap 36 HD and the Polycap 36 AS filter capsules obtained from Fisher Scientific, IL. The Polycap 36HD filters are made of a monofilament anisotropic polypropylene material and have pore sizes of 1 and 5 mm nominal and an effective filtration area of 400 cm2. The Polycap AS filters are made of nylon and have pore sizes of 0.2 mm and a filtration area of 400 cm2. The filtration system consists of a Polycap HD or a Polycap AS filter capsule (Fisher Scientific, IL) and a peristaltic pump.
The extent of microbial removal increased with decreasing filter pore size regardless of the level of inoculation. The filters tested were most effective in the removal of E. coli O157:H7 and least effective in the removal of Salmonella from spiked bottled water regardless of the level of inoculation. Microbial removal was less efficient during filtration of spiked lettuce rinse water. No detectable release of biological agents was observed in filters that were used over a period of 24 hours. Filters with 1µm pore size were able to remove 1.7, 3.0, and 5.6 logs of Salmonella, Y. pestis and E .coli O157:H7 respectively, from spiked bottled water. Filters with 0.2 µm pore in general can remove the target organisms to levels below 1 cfu/100 ml.
The effects of disinfectants on staphylococcal enterotoxin B (SEB) and α-amanitin in water and three selected surfaced (0.03" thick Teflon, 0.003" thick stainless steel and 0.01" latex rubber) were studied. Disinfectants based on alkyl ammonium disinfectants, chlorine bleach, chlorine dioxide, and ozone were evaluated. A method capable of verifying the molecular structure of the compound, liquid chromatography/mass spectrometry, was used to evaluate degradation processes, first in water, and then on selected surfaces. The information obtained from these studies will help establish the manner in which disinfection procedures could be used if protein/peptide toxins (i.e. staphylococcal enterotoxin B (SEB) and α-amanitin) are used as non-traditional adulterants in a food processing environment.
Disinfectants analyzed were 0.1% CTACl (quaternary ammonium surfactant); 1, 5, 10, and 100 ppm NaClO; 10, 100, and 500 ppm NaClO2; 0.1 ppm and continuous O3 for exposure times of 5 and 60 minutes. The least effective disinfectant was CTACl, which is a quaternary ammonium surfactant similar to those used in common household disinfectants. Of the remaining disinfectants, NaOCl is the most aggressive, and can completely degrade the protein in a short period of time at concentrations of 10 ppm. Ozone represents a case where extensive degradation can occur, if the concentration can be raised high enough. At 0.1 ppm, only partial degradation occurs. If, however, the ozone is continuously bubbled through the solution (creating a much higher short term concentration), there is rapid and complete degradation of the protein. Finally, chlorine dioxide, as produced here (using the in-situ acid chlorite reaction) is an oxidizing agent which is less aggressive than bleach, but nevertheless can completely oxidize SEB at 10 ppm.
Disinfectants analyzed were 1 and 10 ppm NaClO; 1, 10, and 100 ppm NaClO2; 0.1 ppm and continuous O3. The surfaces analyzed were stainless steel, Teflon and latex. The effect of disinfectants on SEB attached to surfaces largely parallels the results for SEB in water.
Disinfectants analyzed were 10 and 100 ppm NaClO; 10, 100, and 500 ppm NaClO2; 0.1 ppm and continuous O3 for exposure times of 5 and 60 minutes. Concentrations of 10 ppm NaClO, 100 ppm NaClO2, and 0.1 ppm O3 were able to completely degrade α-amanitin in 5 minutes.
Disinfectants analyzed were 1, 10 and 100 ppm NaClO; 1, 10, and 100 ppm NaClO2; 0.1 ppm and continuous O3. The surfaces analyzed were stainless steel, Teflon and latex. Concentrations of 10 and 100 ppm NaClO; 1, 10 and 100 ppm NaClO2; were able to eliminate α-amanitin after 5 minutes of exposure when attached to stainless steel or Teflon. Latex offered a protective effect. Only 100 ppm NaClO and 100 ppm NaClO2 was able to eliminate α-amanitin after 5 minutes of exposure when attached to latex.
Three of four disinfectants evaluated can be used to substantially degrade SEB. The results are less definitive for α-amanitin. However, as with SEB, higher concentrations of the disinfectants evaluated can produce complete degradation (100 ppm and greater for bleach and ClO2, continuous ozone). α-amanitin does appear to be more difficult to degrade than SEB. The type of surface was found to be of potential importance only at low concentration levels for disinfectants. Under such conditions latex rubber surfaces can reduce the effects of a disinfectant. Teflon and stainless steel interfere to a smaller degree.
Food defense concerns about accidental or intentional contamination of food contact surfaces underscore the need to assess the efficacy of promising disinfectants against threat agents or their surrogates using in situ studies. Antibacterial activity of acidified sodium chlorite (ASC, SanovaTM) was evaluated against Pseudomonas aeruginosa (Pa) and Burkholderia cepacia (Bc) on smooth ("new") and sanded ("worn") conveyor belt materials (polyethylene, polypropylene, acetal, polyester). Belt coupons were inoculated with 6 to 9 log 10 CFU Pa or Bc in 1% fish slurry or PBS. Inoculum was spotted or spread over the coupon surface and allowed to dry for 10 min to 1.5 h at room temperature or overnight at 4°C. Inoculated coupons were sprayed with PBS or with freshly prepared 50 to 2000 ppm ASC and allowed to stand for 2 min; they were then dipped in 0.1 % sodium thiosulfate for 5 to 10 s. Treated coupons were placed in whirlpakTM bags containing 25 ml wash buffer (PBS, 0.5% Tween 20.0.1 % glycine) and sonicated for 10 min. Surviving cells were enumerated on TSA-YE after overnight incubation at 37°C.
The efficacy of ASC was similar on all belt materials. Treatment with 400 ppm ASC resulted in a 2-log decrease in numbers of Pa cells spread on new belt material. Spot-inoculated surfaces were difficult to disinfect; a 1-log decrease in numbers of Pa recovered was seen only after treatment with 1000 ppm ASC. Two-thousand ppm ASC was required to disinfect worn coupons. Results showed that Bc was more sensitive than Pa to ASC, suggesting that Pa can be used as a conservative surrogate for Bc. ASC effectively removed/inactivated Pa from new belt surfaces.
There is limited data on the on the effectiveness of disinfection/sterilization techniques against Burkholderia pseudomallei. Pa and Bc were used as surrogates to assess the efficacy of ASC for disinfection of food processing conveyor belts. Excellent recovery of Pa was achieved from all belt materials. The efficacy of ASC was similar on all belt materials. Treatment with 100 ppm of ASC resulted in a 2-log decrease in numbers of Pa cells spread on new belt material, while 200 ppm was required to disinfect worn coupons. Bc was more sensitive than Pa to ASC. Treatment with 50 ppm ASC resulted in a 3-log and 1-log reduction respectively, of Bc cells attached to new and worn belts; treatment of belts with 100 ppm ASC reduced numbers of Bc cells attached to worn surfaces to non-detectable levels.
ASC effectively removed/inactivated Pa from new belt surfaces provided the contamination is diffuse. Reduced efficacy on worn conveyor belts suggest that a public health hazard could arise if belts are not adequately disinfected.
