GB2220066A - Detecting biological contamination - Google Patents

Abstract

A sample suspected of harbouring biological contamination is exposed to hydrogen peroxide and any resulting evolution of oxygen is measured with an oxygen electrode. The method utilises the presence of catalase in many food contaminants. The electrode may be a Clark-type electrode.

Description

DETECTING BIOLOGICAL CONTAMINATION The present invention relates to methods and apparatus for detecting and measuring biologicial contamination of materials, especially foodstuffs, for example contamination by microbes.

For perishable items like foods, rapid and reliable methods of assessing microbial contamination are needed; the retrospective results provided by plate counts are of little practical value. Ideally, such new methods should not only be simple and quick to perform and interpret but they should also be cheap in terms of both capital and running costs. Unfortunately, many of the currently available methods do not meet these criteria.

Assay of the enzyme catalase (H202:H20 oxidoreductase; EC1.11.1.6) has been suggested as just such a possible method. This enzyme is present in bacteria that employ a respiratory cytochrome electron transport system for energy generation, i.e. all aerobes and most facultative anaerobes, and these organisms are common spoilage organisms, particularly of aerobically coldstored foods (Wang & Fung 1986). Catalase is not present in strict anaerobes or the lactic acid bacteria although some strains contain true or pseudo-catalase activity.

Catalase activity can be determined by many methods, which can be categorised as titrimetric, spectrophotometric or manometric methods (Wang & Fung 1986) but the most notable method developed for estimating catalase activity was that devised by Gagnon et al. (1959). This method relies on the measurement of the rate of flotation of paper discs due to the evolution of gaseous oxygen after the addition of hydrogen peroxide (H202) and has been shown to give a reliable estimate of the cell densities of several different bacteria: the catalasepositive spoilage organisms in vacuum-packed turkey could be rapidly estimated down to 104g-1 meat. Wang and Fung (1986) used this and a pasteur pipette method devised by them to examine the surface microbiology of chicken.

They found that the psychrotrophic population could be reliably detected giving typical disc flotation times of < 0.5, < 3, < 10, < 22, < 42 and < 71 minutes corresponding to cell densities of < 109, 1O11Os, 106 -107 , 105-196, 104-105 and < 103 bacteria cm- 2 respectively.

More recently Phillips and Griffiths (1987) examined this method for use with raw milk but, due to the presence of endogenous milk catalase, they could find no correlation with bacterial densities. Indeed, assay of catalase activity has been shown to correlate with the leucocyte content of raw milk. However) Phillips and Griffiths (1987) did conclude that the test had potential for determining the quality of pasteurised milk following preincubation of samples with inhibitors of Gram-positive organisms at 21oC for 25 hours.

The present invention provides an alternative approach in which the oxygen generated by the catalaseinduced degradation of He or is measured by use of an oxygen electrode, preferably following concentration of the microbes on an electropositive filter. That this should be successful is surprising, since oxygen electrodes are generally thought to measure dissolved oxygen, whereas the catalase produces actual bubbles of oxygen. Nadler et al (1986) did propose the use of an oxygen electrode to measure catalase activity, but not in the context of detecting microorganisms. It would not previously have been thought that such a method would be sufficently sensitive for this use. Notwithstanding this, the method has been found to give surprisingly good results and provides the basis of a simple, reliable and quick test.A particular advantage is that the direct electrical output of the oxygen electrode allows easter data analysis and automation. Furthermore, the use of disposable materials may be kept to a minimum.

Thus, one aspect of the present invention provides 9 method of detecting biological contamination of a sample by exposing hydrogen peroxide to the sample and detecting the resulting evolution of oxygen with an oxygen electrode.

By "biological contamination" we mean contamination with those living respiring organisms which contaln measurable catalase activity, principally bacteria but also fungi (including yeasts), protozoa and animals such as insects (including insect larvae).

The electrode is preferably a Clark-type electrode.

This has a platinum working electrode and a silver/silver chloride reference electrode (poised at 600mV) separated from the sample by an oxygen permeable membrane, for example one made of Teflon (Regd. T.M. for polytetrafluoroethylene).

The sample may be the original material in which the contamination has occurred or it may be an extract thereof. If the original material is food, especially raw food such as raw meat, endogenous catalase will often be present irrespective of any biological contamination. To separate this endogenous catalase from microbial catalase, it is advantageous to concentrate the organisms on an electropositive filter, and then to measure the catalase activity on the filter. This approach was described in principle in Hirsch & Martin, 1984 and Kroll, 1985, although the filter was used only to remove the bacteria from the food and the bacteria were not repidly assayed in any way. The filter preferably has a pore size range of 2-9 um, although other pore sizes may be suitable, depending upon the material used.

Alternatively, an inhibitor which is selective for eucaryotic catalase may be employed or, because catalase is predominantly an intra-cellular enzyme, a suitable cell-disrupting agent or process may be employed to selectively disrupt the microbial cells. Indeed nonselective cell disruption with a detergent such as Triton X100, preferably following concentration of the microbes, can release greater quantities of catalase for assay.

Only some organisms, predominantly Gram negative bacteria, have a measurable catalase activity. However, it is often these organisms which one is most concerned to detect in food, for example in pasteurised milk and non-vacuum packed meat. The suitability of the method of the invention in respect of any given material to be analysed, and the extent to which measuring the catalasegenerated oxygen provides a quantitative test, can be determined by the man skilled in the art by routine methods.

The method of the invention may be used for detecting contamination of raw and pasteurised milk and cream, raw and cooked meat (such as beef steak, pork, lamb and chicken), fish, vegetables (such as carrots and cauliflower), processed foods (especially non-sterile chilled products such as processed meats and quiches) and frozen foods.

In addition, contamination of cosmetic and pharmaceutical preparations such as shampoos and creams may be detected.

An embodiment of the invention will now be described by way of example and with reference to the accompanying drawings, in which: Figure la illustrates the relationship between catalase activity and cell densities of B.licheniformis (O; m = 0.A7, c = -0.53, r = 0.82), P.vulgaris (O; m = 0.30, c = -0.71, r = 0.56) and E.coli (A; m = 0.33, c = -1.23, r = 0.86) as determined by the initial rates of increase in 02 concentration using an 02 electrode after the addition of lOul 10 vol He02; Figure ib illustrates the relationship between catalase activity and cell density of M.roseus , rn = 0.28, c = -0,32, r = 0.38), ?s.s,utzeri (# ; m = U.-?, c = -2.39, # 39 , r = 0.83) and Pseudomonas sp. P4 ( ; m = c = -4.9s, r = 0.92) as determined by the initial r--.ts of increase in 02 concentration using an 02 electrode after the addition of 10ul 10 vol Hz02; Figure 2 illustrates the relationship between catalase activity and plate counts of cell suspensions made from minced beef (O ; m = 0.14, c = 0.75, r = 0.52) or braising steak ( ; m = 0.14, c = 0.75, r = 0.52); Figure 3 illustrates the relationship between catalase activity and plate counts of pasteurised milk (#) or cell suspensions made from samples of cauliflower (o); Figure 4 illustrates the relationship between plate counts of raw milk and catalase activity associated with electropositive filters after filtering the raw milk through 05S grade 'Zeta plus' filters; Figure 5 illustrates the relationship between plate counts of suspensions from braising steak and the catalase activity associated with electropositive filters after filtering the meat suspension through 05S grade 'Zeta plus' filters; Figure 6 illustrates apparatus for filtering bacteria from a sample; and Figure 7 illustrates apparatus in accordance with the invention for determining microbial contamination.

Materials and Methods PREPARATION OF PURE CULTURES Pure cultures of Bacillus licheniformis KRM007 (Hannah Research Institute), Micrococcus roseus NCDO 767, Pseudomonas stutzeri HRI B4/4 (H.R.I.), Escherichia coli NCDO 2328, Proteus vulgaris NCTC 10020 and Pseudomonas sp. P4 were subcultured at monthly intervals on Yeastrel Milk agar (YMA;Oxoid Lid,) and stored at 4 C. nrotn cultures for experimentation were obtained by inoculatIng 0.lml of a Nutrient Broth (Oxoid Ltd.) culture, which na been inoculated the previous day, into 50ml of Nutriont Broth in sterile 250ml conical flasks at 30or. After 15h shaking incubation, the cell suspensions were harvested by centrifugation at 5000 xd for ten minutes, washed once with 0.1M potassium phosphate buffer pH 7.0, and finally resuspended in fresh buffer and stored on ice for use that day.Serial dilutions of each culture were made in sterile buffer as required to give a range of cell densities between approximately 104 -1OS cfu ml-1. Further serial dilutions were made of each sample in 1/4 strength Ringers solution and lml samples of the appropriate dilutions were plated in YMA and incubated at 300C for 3d so that the original cfu ml-l in each sample could be determined.

FOOD SAMPLES Fresh cauliflower, braising steak or minced beef were obtained from a local commercial outlet. To approximately 10g of each food sample, 20ml of 1/4 strength Ringers solution was added and the bacteria dislodged from the samples by 'stomaching' for 2 minutes.

The resulting suspensions, together with samples of raw or pasteurised milk which were obtained from a local dairy, were stored on ice for immediate use. In some experiments, food samples were stored at 50C for up to 7d before analysis to increase the density of spoilage organisms.

In some experiments samples of raw milk (10ml) or suspensions from solid food samples (15ml) were filtered through 25mm discs of 05S grade Zetaplus filters (Gelman Sciences, Northampton) which had been previously sterilised in situ by autoclaving in 'Swinex' in-line filter holders (Millipore Ltd.). Filtration was aided by the manual application of a slight positive pressure using a 2Oml sterile syringe. After filtration ten ml of air was passed through the filter to remove excess liquid. For the solid food suspensions, particulate matter was removed by the inclusion of 25um nylon mesh discs (25mm diameter) which were placed in the filter holder before sterilisation, directly above the Zetaplus filters.Serial dilutions of the filtrate and the original samples were made in 1/4 strength Ringers solution, plated in YMA and incubated at 3OoC for 3d to determine the original cfu g-1/ml-1 in each sample and the efficiency of absorption to the Zetaplus filters.

ASSAY OF CATALASE ACTIVITY Catalase activity was measured using a magnetically stirred Clark-type oxygen electrode (Rank Bros.

Cambridge) connected via a polarising circuit (0.6V) to a chart recorder with back off facilities (Type CR600, J.J.

Lloyd Instruments, Southampton). The temperature of the reaction was controlled by circulating water at 250C through the jacket of the electrode. The electrode was calibrated before use and several times throughout the day by adding air-saturated distilled water to the reaction chamber of the vessel (magnetic stirring on) and adjusting the sensitivity of control of the polarising circuit to give a full scale deflection of about 90% on the chart recorder. The zero oxygen value was obtained by adding a few crystals of sodium dithionite, which in practice is close to the zero value obtained by short circuiting the electrode.From the amount of oxygen dissolved in air-saturated distilled water at 250C (259 nmoles 02 ml-l) the number of nmoles of oxygen ml-1 corresponding to lem of the chart recorder could be calculated.

After calibration, the distilled water was aspirs Weå out of the reaction chamber and replaced bp 3.9ml of 0.1x potassium phosphate buffer pH 7.0 to which 100ul alicts of the range of concentrations or suspensions of sc-e cultures were added. With the food samples, 4ml of milk or the stomached suspensions were added directly. When the catalase activity associated with an electropositIve filter was assayed, the filter was cut in half using ethanol flamed scissors and forceps. Half the filter was lodged in the reaction chamber to which 4ml of 1/4 strength Ringers solution or buffer had been added.The other half of the filter was treated similarly in a subsequent experiment and the average value of rates of oxygen release obtained. The electrode compartment top was replaced and the reading was allowed to stabilise ( < 1 min). An aliquot of hydrogen peroxide solution was then quickly injected into the reaction chamber using a glass microsyringe through the central hole in the electrode top. Preliminary experiments showed that the addition of 10ul of 10 vol H202 was the minimum amount of H2 02 which produced a maximum response and this was used in all subsequent experiments. The oxygen evolution was followed on the chart recorder (3cm min-1 chart speed) for 2-3 minutes, and the initial rate of increase in the concentration of oxygen was calculated directly from the chart paper.

Results Initial control experiments were designed to show that this method of measuring catalase activity was valid. The addition of Hz02 to 4ml of sterile phosphate buffer or Ringers solution produced a negligible deflection on the O2 electrode reading and no measurable rate of increase in O2 concentration was seen, demonstrating that no direct interaction between the H202 and the buffer solutions or electrode was occurring.

Simirar'ty, the addition of H2O2 to ceil suspnsion- (appro..~mately 107 cfu ml-1) of the catalase nega-ive bacteria Lactobacillus plantarum NCDO 1752 or Lactoccocus lactis HRI A2 produced a similar response.

On addition of H202 to suspensions of pure cultures of aerobic or facultative anaeorbic bacteria a rapid increase in the concentration of 02 was observed (results not shown). The initial rates of increase in 02 concentration were directly proportional to the densities of bacteria (Figure 1) and reasonable correlation coefficients of 0.82, 0.56, 0.86, 0.88, 0.83 and 0.92 between the initial rate of increase in 02 concentration and plate counts were observed (Figure 1). The method was also reasonably sensitive and could detect minimum ceil densities of between approximately 103-105 cfu ml-1.

However, these relationships strongly depended on the identity of the bacterial culture (Figure 1). Indeed, it is evident that the specific catalase activity of the different organisms li.e. given by the value of the intercept) we have examined here and the relationship between catalase activity and the cell densities of the different organisms (i.e. the slope of the line) are markedly genus/species dependent such that an initial rate of increase in oxygen concentration of 10 nmoles ml-l min-1 corresponded to cell densities of 1.5 x 103 ml-1 (B.licheniformis), 3.1 x 105 ml-l (P.vulgaris), 2.5 x 106 ml-l (E.coli), 2.5 x 104 ml-l (M.roseus), 8.9 x 105 ml-1 (Pseudomonas sp. P4) or 1.7 x 106 ml-1 (Ps.stutzeri).

With minced beef and braising steak there was no relationship between the initial rate of increase in O2 concentration and bacterial numbers (Figure 2). Not only was there much variation in the former but high rates of increase in 2 concentration were also observed at the lower cell densities. This suggests that significant amounts of endogenous catalase must be present in the samples and that the amount of this non microbial catalase is variable, masking , in most cases, any bacterial catalase present. Similar results w:-reobtained with raw milk (results not shown) as found by Phillips and Griffiths (1987) using the disc flotation technique.

With cauliflower and pasteurised mil more encouraging results were obtained (Figure 3). There did indeed appear to be a relationship between catalase activity and cell numbers but only above 105 cfu g-1/ ml-l. Again it was not that lower catalase activities/ cell densities were not detectable by the method but that endogenous catalase activity of approximately @ (pasteurised milk) and 7 (cauliflower) nmoles 2 ml-1 min-1 appeared to be present.

When samples of raw milk or suspensions from braising steak were passed through electropositive filters (Kroll 1985) the bacteria were reasonably efficiently entrapped by the filters with efficiencies of 56.1 + 20.1% and 84.8 + 9.0% respectively (as determined by plate counts) although the efficiency of retention from raw milk was lower and showed more variation than that reported previously (Kroll 1985). When the catalase activity associated with the filters was assayed a distinctly improved relationship with bacterial numbers was observed (Figures 4, 5). With the steak an extremely good relationship of catalase activity and cell numbers was observed (Figure 5) although the relationship is apparently curvilinear.This is probably due to a decrease in the efficiency of absorption of the bacteria by the filters at higher cell densities (which reduced to 65.9 + 17.1% above 1 x 108 bacteria g-l) perhaps due to saturation of available attachment sites. This good relationship was probably produced by the effective reduction of the non-microbial catalase in the sample for analysis; however the method could detect only over 105 cfu g-l, Similar improved results were obtained with caullfiower and carrots (not shown). With raw :nick (Figure 4) considerable scatter was observed at the lover cell densities and it was assumed that this was due to some endogenous catalase attaching to the filters and the low and variable adsorption of bacteria to the filters.

Even if, with some foodstuffs, detection of contamination below 105 c.f.u. ml-t is not reliable, the detection of bacterial contamination above 105 c.f.u.

ml-1 and its simple, rapid and low cost quantification are potentially very useful.

References Hirsch, D.C. & Martin, L.D. 1984 Rapid detection of Salmonella in certified raw milk by using charde-modified filters and Felix-Ol bacteriophage. Journal of Food Production 47, 388-390.

Kroll, R.G. 1985 Electropositively charged filters for the concentration of bacteria from foods. Food Microbiology 2, 183-186.

Nadler et al, 1986, Biochim. et Biophys. Acta 882, 234241.

Phillips, J.D. & Griffith, M.W. 1987 A note on the use of the Catalasemeter in assessing the quality of milk.

Journal of ApPlied Bacteriology 62, 223-226.

Wang, G.I.J. & Fung, D.Y.C. 1986 Significance of bacterial catalase in food microbiology: A review.

Journal of Food Safety 8, 47-67.

Claims (9)

1. C',rnMS

   i. A method of detecting biological contamination of a sample by exposing hydrogen peroxide to the sample an detecting the resulting evolution of oxygen with an oxygen electrode.
2. 2. A method according to Claim 1 wherein the oxygen electrode is a Clark-type electrode.
3. 3. A method according to Claim 1 or 2 wherein the sample is an extract from material suspected to be contaminated.
4. 4. A method according to Claim 3 wherein the extract is obtained by a process which concentrates microbes.
5. 5. A method according to Claim 4 wherein the process involves passing the material or part thereof across or through an electropositive filter.
6. 6. An apparatus for detecting biological contamination of a sample, comprising means for supplying hydrogen peroxide to the sample, an oxygen electrode to detect resulting evolution of oxygen, and means for indicating the electrical output of the electrode.
7. 7. An apparatus according to Claim 6 additionally comprising an electropositive filter through which or across which material to be analysed is passed in order to yield the said sample.
8. 8. A method substantially as described herein.
9. 9. An apparatus substantially as described herein.