The direct epifluorescent filter technique (DEFT): increased selectivity, sensitivity and rapidity

With the direct epifluorescent filter technique (DEFT), differentiation of bacteria was achieved by a modified Gram-staining procedure using acridine orange as the counterstain. The method enumerated viable Gram-negative and all Gram-positive bacteria. Counts of clumps of orange fluorescent cells (Gram-negative DEFT count) correlated well with colony counts of Gram-negative bacteria in samples of raw milk (r = 0.94). The use of stainless steel membrane filter supports and the addition of citrate-NaOH buffer (0.1 M, pH 3.0) during filtration enabled 10 ml samples of milk to be filtered, thereby increasing the sensitivity of the DEFT five-fold. The relationship between colony and DEFT counts with 10 ml samples was better (r = 0.90) than that using standard 2 ml samples (r = 0.88). Alternatively, these modifications in procedure allowed the preincubation time for 2 ml milk samples to be reduced from 10 to 2 min. Sonication was successful in dispersing bacterial clumps in both pure cultures and in raw milk samples to yield a bacterial count by DEFT which should give a better indication of the hygienic status and keeping quality of a product, than counts of colony forming units.     

Semi-automated counting of bacteria and somatic cells in milk using epifluorescence microscopy and television image analysis

An epifluorescent microscope fitted with a 'Chalnicon' closed circuit television camera linked to an Optomax System III image analyser was used to count bacteria and somatic cells on membrane filters prepared from milk by the direct epifluorescent filter technique (DEFT). For both bacteria and somatic cells, the semi-automated DEFT count of low count milk exceeded the visual DEFT count but this difference became proportionately less as the count increased. There was close agreement between the semi-automated and visual DEFT counts over the range 10⁵-5 times 10⁶/ml for bacteria and 3 times 10⁵-5 times 10⁶/ml for somatic cells. The semi-automated DEFT count of bacteria and somatic cells correlated well with the plate count and Coulter count respectively, with correlation coefficients of 0.83 and 0.81.     

Use of the direct epifluorescent filter technique for the enumeration of yeasts

The ability of the direct epifluorescent filter technique (DEFT) to enumerate the viable numbers of various species of yeasts was evaluated. A DEFT count could be made in less than 10 min and the DEFT counts of non-heat-treated samples agreed well with plate counts. The DEFT was unsuitable for the enumeration of yeasts in heat-treated samples because non-viable cells fluoresced orange. A double staining technique using Janus Green B and acridine orange was developed to overcome this problem. The modified DEFT enabled viable and non-viable yeasts to be differentiated in heat-treated samples of pure cultures and improved the relationship between the DEFT count and plate count. The method proved to be of little value, however, for use with beverage products because of unreliable staining patterns.     

Use of the direct epifluorescent filter technique for the enumeration of yeasts

The ability of the direct epifluorescent filter technique (DEFT) to enumerate the viable numbers of various species of yeasts was evaluated. A DEFT count could be made in less than 10 min and the DEFT counts of non-heat-treated samples agreed well with plate counts. The DEFT was unsuitable for the enumeration of yeasts in heat-treated samples because non-viable cells fluoresced orange. A double staining technique using Janus Green B and acridine orange was developed to overcome this problem. The modified DEFT enabled viable and non-viable yeasts to be differentiated in heat-treated samples of pure cultures and improved the relationship between the DEFT count and plate count. The method proved to be of little value, however, for use with beverage products because of unreliable staining patterns.     

Preliminary evaluation of COBRA, an automated DEFT instrument, for the rapid enumeration of micro-organisms in cultures, raw milk, meat and fish

A preliminary evaluation of an automated direct epifluorescent filter technique instrument, COBRA, showed that it enumerated rapidly bacteria in cultures, raw milk, meat and fish. The correlation coefficients of regression lines with corresponding pour plate counts were 0·99 (pure cultures), 0·81 (raw milk) and 0·91 (raw meat and fish), respectively. The system was simple to use, one operator could process > 100 samples/h and results were available in < 1 h.     

Evaluation of a commercial fluorochromic system for the rapid detection and estimation of wine lactic acid bacteria by DEFT

A commercial fluorochromic system was evaluated for the rapid detection of lactic acid bacteria in fortified wines by the epifluorescent filter technique (DEFT). The viability test used, employing the fluorescence dyes SYTO 9 and propidium iodide, was able to detect and clearly differentiate viable from non-viable cells (killed with a 50% v/v ethanol solution). A good overall agreement ( r ₌ 0·92) was obtained between the DEFT count and the plate count in the range studied (5 × 10²–4 × 10⁹ cells ml⁻¹). Wine components which might otherwise interfere with the method could be removed by including simple wash steps in the protocol. This measure proved critical to the success of the procedure. For practical purposes, the rapid method studied seems to be a good alternative to the traditional cultural methods as part of quality control programmes in wine making. It may also be useful when studying the efficacy of certain treatments in the elimination of wine bacterial contaminants.     

Collaborative trial of the direct epifluorescent filter technique (DEFT), a rapid method for counting bacteria in milk

The direct epifluorescent filter technique (DEFT) is a new rapid method which uses membrane filtration and epifluorescent microscopy for counting bacteria in milk. A collaborative trial of the DEFT was conducted between six laboratories. Each laboratory obtained a highly significant relationship between the DEFT count and plate count with a correlation coefficient generally > 0.9 but there were significant differences between these relationships. The repeatability of the DEFT, although ca 1·5 times worse than that of the plate count, was of a level acceptable in practice. Reproducibility of the DEFT was ca 3 times that of the plate count. This poor reproducibility was probably mainly due to counting errors. Possible reasons for this and ways of reducing counting errors are discussed.     

Rapid Enumeration of Bacteria in Heat-treated Milk and Milk Products Using a Membrane Filtration-Epifluorescent Microscopy Technique

A rapid direct epifluorescent filter technique (DEFT), taking ca. 20 min to complete, was used to enumerate bacteria in heat-treated milk and milk-products. The DEFT count could detect as few as 5750 bacteria/ml in pasteurized and skim milk, pasteurized cream, whey and sweet cream butter and was in agreement with the standard plate count. The technique was also used to determine the sterility of cartoned ultra heat-treated milk after incubation at 37°C. The agreement between the DEFT and plate count was poor for evaporated and condensed milks, some reconstituted skim milk powders, pasteurized whey and ripened cream butter. Possible reasons for this are discussed.     

A bacteriostatic mixture for milk samples and its effect on bacteriological, cytological and chemical compositional analysis

Of the bacteriostats tested, that comprising boric acid, glycerol, potassium sorbate and nystatin was the most suitable in preventing the multiplication of bacteria, yeasts and moulds in milk whilst causing least change in the Direct Epifluorescent Filter Technique (DEFT) count of bacteria and somatic cells. The DEFT and plate count of bacteria and the DEFT count of somatic cells were reduced after preservation and storage. The relationships between the plate count of fresh milk samples and the DEFT count and plate count of the same samples after preservation and storage for 3 d at room temperature (20–22°C) had correlation coefficients ( r ) of 0.80 and 0.85, respectively. The relationship between the Coulter count of somatic cells in fresh milk and the DEFT count of somatic cells in stored preserved milk had a correlation coefficient of 0.90. The addition of the bacteriostatic mixture to milk affected the determination of protein by the Milkoscan and lactose by the Infra Red Milk Analyser. It did not affect the estimation of the other milk components (fat, protein or lactose) by either instrument.     

Microbiological characteristics of anevato: a traditional greek cheese

Nine batches of Anevato, raw goat milk cheese, were examined throughout a 60 day storage time at three different periods within the lactation season of the goat. High mean log counts per gram of cheese for aerobic bacteria (7·92–9·56), lactic acid bacteria (7·78–9·32), Gram-negative organisms 5·64–9·67), psychrotrophs (7·90–11·79) and proteolytic bacteria (7·57–9·36) were found. Enterobacteriaceae, coliforms and yeasts were considerably lower. Enterobacteriaceae and coliforms in the curd of cheese made in May were lower by approximately 3·0 log₁₀ cfu g⁻¹ than counts in curd made in January, and were lower by about 2·5 log₁₀ cfu g⁻¹ than those in cheese made in March. This coincided with lower pH and higher counts of lactic acid bacteria in cheese made in March and May. Yeast populations were affected by the season and were higher in May than March and/or January. Lactococci dominated in the cheese until 15 days, but lactobacilli became predominant after 30 days. Lactococcus lactis was the most abundant species of lactic acid bacteria found in Anevato cheese. Results suggest the need for improving milk quality and/or using heat-treated milk to produce Anevato cheese; the use of L. lactis as a starter would possibly eliminate or suppress the growth of undesirable organisms.     

Rapid Membrane Filtration-Epifluorescent Microscopy Technique for Direct Enumeration of Bacteria in Raw Milk

Membrane filtration and epifluorescent microscopy were used for the direct enumeration of bacteria in raw milk. Somatic cells were lysed by treatment with trypsin and Triton X-100 so that 2 ml of milk containing up to 5 × 10⁶ somatic cells/ml could be filtered. The majority of the bacteria (ca. 80%) remained intact and were concentrated on the membrane. After being stained with acridine organe, the bacteria fluoresced under ultraviolet light and could easily be counted. The clump count of orange fluorescing cells on the membrane correlated well (r = 0.91) with the corresponding plate count for farm, tanker, and silo milks. Differences between counts obtained by different operators and between the membrane clump count and plate count were not significant. The technique is rapid, taking less than 25 min, inexpensive, costing less than 50 cents per sample, and is suitable for milks containing 5 × 10³ to 5 × 10⁸ bacteria per ml.     

The use of Alcalase 2·5L in the acridine orange direct count technique for the rapid enumeration of bacteria in beef mince

The acridine orange direct count (AODC) technique was used to count bacterial numbers in beef mince. The ability of Alcalase 2·5L to degrade beef mince was compared with the previously used Alcalase 0·6L. Both methods were evaluated against the standard plate count. The results showed that Alcalase 2·5L can be used in the AODC technique for the rapid counting of bacterial numbers in beef mince. Prediction equations obtained for the relationship between the AODC and standard plate counts were validated with commercial beef mince samples.     

A note on the use of the Catalasemetre in assessing the quality of milk

The Catalasemetre, for assessing the quality of raw and pasteurized milk, has been studied. No correlation was found between catalase activity and bacterial counts for farm bulk tank milks within the range 5.2 ± 10²-5.4 ± 10⁵ cfu/ml. Similarly, no relation was observed between catalase activity and somatic cell counts of milk (range of counts from 0.08 to 3.5). However, the catalase activity and bacterial count of pasteurized milks which had been pre-incubated at 21.C for 25 h in the presence of crystal violet-penicillin-nisin to inhibit Gram-positive bacterial growth were significantly related. Thus, the use of this pre-incubation procedure coupled with the Catalasemetre to estimate bacterial growth, has potential in assessing the keeping quality of pasteurized milk samples within 25.5 h of production. Results on the thermostability of native milk catalase are also presented.     

Evaluation of the stain Viablue for the rapid estimation of viable yeast cells

Viablue can be used as a differential viability stain for yeast cells. Differential staining using fluorescence microscopy and viability showed a good correlation ( r = 0·99). For practical purposes the staining procedure takes only a few minutes and appears ideally suited to such rapid analytical procedures as the direct epifluorescent filter technique (DEFT) and automated quality control testing of foods and beverages.     

THE DETERMINATION OF ESCHERICHIA COLI I IN SEA WATER

SUMMARY: For the determination of Escherichia coli I in sea water lactose broth frequently gave higher presumptive and confirmed counts than MacConkey's broth. In the presumptive count there were 53 cases where lactose broth gave larger numbers than MacConkey's broth, with 11 equal counts, and only 25 cases with smaller counts ( P =0·00137). After confirmation the corresponding numbers of cases were 39, 10 and 27 ( P =0·134).
In the samples giving most probable numbers (MPN) of less than 100 E. coli /100 ml lactose broth was superior to MacConkey's broth ( P =0·021). At higher MPN values both media were satisfactory, but with highly polluted water MacConkey's broth might give better recoveries due to the suppression of high concentrations of non-coli-aerogenes bacteria.
Samples stored for 24 or 48 hr before testing gave higher presumptive recoveries when examined with lactose broth than with MacConkey's broth, the values for P being 0·028 and 0·0027 respectively.
It appears that lactose broth without inhibitory ingredients could be used with advantage in the examination of sea water.     

THE ASSESSMENT OF THE BACTERIOLOGICAL CONDITION OF MILK BOTTLES

SUMMARY: A study of the relative values of a number of bacteriological tests for assessing the condition of milk bottles indicated that the colony count of the bottle rinse solution on yeastrel milk agar incubated for 4 days at 30°, combined with a clot-on-boiling test applied to 1 ml. of rinse in 9 ml. of sterile milk after incubation for 72 hr. at 19–20°, gave the most useful results.
The mean of the ratios of colony counts at 30° to those at 37° was 15·1, while it was as high as 22·9 for rinses with 37° of over 600 for an unsatisfactory bottle should be retained when the test is done at 30°. The thermoduric colony count of rinses of milk bottles, even when laboratory pasteurized in milk, did not provide any additional information to that given by the colony count at 30° made without pasteurization. A high proportion of the organisms in bottle rinses survived laboratory pasteurization in milk, the survival rate being highest in efficiently treated bottles.
The clot-on-boiling test gave results in general agreement with colony counts and served to indicate the potential influence of badly contaminated bottles on the keeping quality of milk placed in them. A substantial proportion of rinses with satisfactory colony counts reduced methylene blue within 48 hr. at 19–20°.
Colony counts at 37° were on the average much lower for bottles treated with steam than for bottles submitted to detergent treatment in various types of bottle washing machines. Treatment of bottles by steam or hypochlorite was more efficiently done on the farms than at the dairies.     

Real-time PCR detection of Paenibacillus spp. in raw milk to predict shelf life performance of pasteurized fluid milk products

Psychrotolerant sporeformers, specifically Paenibacillus spp., are important spoilage bacteria for pasteurized, refrigerated foods such as fluid milk. While Paenibacillus spp. have been isolated from farm environments, raw milk, processing plant environments, and pasteurized fluid milk, no information on the number of Paenibacillus spp. that need to be present in raw milk to cause pasteurized milk spoilage was available. A real-time PCR assay targeting the 16S rRNA gene was designed to detect Paenibacillus spp. in fluid milk and to discriminate between Paenibacillus and other closely related spore-forming bacteria. Specificity was confirmed using 16 Paenibacillus and 17 Bacillus isolates. All 16 Paenibacillus isolates were detected with a mean cycle threshold (C(T)) of 19.14 ± 0.54. While 14/17 Bacillus isolates showed no signal (C(T) > 40), 3 Bacillus isolates showed very weak positive signals (C(T) = 38.66 ± 0.65). The assay provided a detection limit of approximately 3.25 × 10(1) CFU/ml using total genomic DNA extracted from raw milk samples inoculated with Paenibacillus. Application of the TaqMan PCR to colony lysates obtained from heat-treated and enriched raw milk provided fast and accurate detection of Paenibacillus. Heat-treated milk samples where Paenibacillus (≥1 CFU/ml) was detected by this colony TaqMan PCR showed high bacterial counts (>4.30 log CFU/ml) after refrigerated storage (6°C) for 21 days. We thus developed a tool for rapid detection of Paenibacillus that has the potential to identify raw milk with microbial spoilage potential as a pasteurized product.     

Rapid Counting of Bacteria in Rinses of Milking Equipment by a Membrane Filtration - Epifluorescent Microscopy Technique

Membrane filtration and epifluorescent microscopy were used for the rapid enumeration of bacteria in rinses of milking equipment. Rinse solution (10 ml) was filtered through a 0.6 µm pore size Nuclepore membrane filter and the bacteria retained stained with acridine orange. The clump count of orange fluorescing bacteria on the membrane correlated well with the corresponding plate count ( r =0.83). The technique is rapid, taking approximately 10 min, and is sufficiently sensitive to detect 1000 bacteria ml rinse (equivalent to 5 × 10³ bacteria/m² of milking equipment surfaces by the method currently in use).     

Variations in concentrations of bacterial metabolites, enzyme activities, moisture, pH and bacterial composition between and within individuals in faeces of seven healthy adults

Variations between and within individuals, and correlations between concentrations of bacterial metabolites, including putrefactive products, ammonia and short chain fatty acids (SCFAs), enzyme activities, moisture and pH, as well as bacterial composition, were studied in faecal samples from seven healthy adults over a period of 7 months. Large variations, both between and within individuals, were observed in concentrations of putrefactive products. Although values for ammonia, SCFAs, enzyme activities, moisture and pH were generally variable, significant person-to-person differences were observed.
While ranges of log viable counts of the predominant bacteria such as eubacteria, bifidobacteria and bacteroides in each subject remained between 0·2 and 1·3, those of enterobacteria, streptococci (including enterococci) and lecithinase-negative clostridia varied between 0·4 and 3·0. Levels of bifidobacteria, enterobacteria, streptococci and total aerobic bacteria showed inter-individual variations. Correlations were found among certain of the parameters: moisture correlated negatively with p -cresol ( r = -0·707), pH ( r = -0–671) and β-glucosidase activity (GS) ( r = -0·608), and positively with acetic acid ( r = 0·621), while negative correlations were observed in pH with acetic and butyric acids ( r = -0·690 and -0·623, respectively).
No significant correlations were found between bacterial compositions, and other faecal factors such as pH, moisture, metabolic enzyme activities and concentrations of putrefactive products.     

Use of the Direct Epifluorescent Filter Technique for predicting the keeping quality of pasteurized milk within 24 hours

The keeping quality (KQ) of pasteurized milk stored at 5°C and 11°C was predicted within 24 h by pre-incubating samples and counting bacteria by the Direct Epifluorescent Filter Technique (DEFT). For samples from 5°C storage, 0.03% (w/v) benzalkonium chloride and 0.002% (w/v) crystal violet (final concentration) were added to inhibit the growth of Gram positive bacteria during pre-incubation. The samples from milk stored at 11°C were pre-incubated without the addition of inhibitors. After pre-incubation there was a satisfactory relationship between the DEFT count and the KQ of milks at both 5°C and 11°C. The DEFT count following pre-incubation correctly classified > 80% of pasteurized milks on the basis of KQ.