in Beverages - PubMed Central Canada

5 downloads 0 Views 713KB Size Report
Jun 6, 1986 - grid were counted in each field of the filter not covered by the gasket. Budding yeasts were counted as one cell as long as the daughter cell was ...
APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Sept. 1986, p. 599-601

0099-2240/86/090599-03$02.00/0 Copyright © 1986, American Society for Microbiology

Vol. 52, No. 3

Fluorescence Microscopy Procedure for Quantitation of Yeasts in Beverages HERBERT A. KOCH, RUTH BANDLER,* AND REGINA R. GIBSON Division of Microbiology, Food and Drug Administration, Washington, D.C. 20204 Received 25 February 1986/Accepted 6 June 1986

Existing methods for quantitating yeasts in beverages include time-consuming plate counts that detect only living cells and hemacytometer counts that are reliable only at very high concentrations (e.g., 106 to 20 x 106 cells per ml). The new method described here involves the use of fluorescence microscopy with the fluorescent stain aniline blue to differentiate yeasts (and other fungi) from backgrounds for easy counting and also may be used in conjunction with membrane filtration to concentrate yeasts from liquids before cell enumeration. Recoveries averaged 91.5% for beverages spiked with levels of 500 to 600,000 organisms per ml. The correlation coefficient of count to spike level was 0.996.

The determination of yeast concentration in liquids is important in many fields of microbiology, both for stock solutions and for analytical samples. Spoilage yeasts can cause severe problems in nearly all kinds of beverages, and even low numbers may alter the product esthetics irrespective of whether the yeasts are still alive. The microbiological evaluation of a liquid deals with the identification and quantitation of the microorganisms found in it. Plate counts are time-consuming and only valid for viable cells. The use of a hemacytometer is limited to liquids with concentrations of 1 x 106 to 20 x 106 cells per ml and thus cannot extend to thick or pulpy liquids. We have developed a simple and accurate method for counting yeasts (and other fungal particles) in all types of liquids. The method uses filtration to concentrate large volumes and the fluorescent stain aniline blue to facilitate microscopic counting. Aniline blue, which has been used to localize cell wall polysaccharides in plants (1) and molds (2), is extremely useful in inducing fluorescence in a broad spectrum of fungi. Preparation of spiked liquids. Cultures of several genera of yeasts (Saccharomyces, Candida, Pichia, Trichosporon, and Cryptococcus) were autoclaved to halt growth, added to beverages (water, grape juice, orange juice, and cola), and examined by the methods described below. Autoclaved cultures of the yeast Saccharomyces cerevisiae were counted with a hemacytometer and then used to spike water and grape juice at levels ranging from 500 to 600,000 cells per ml and orange juice at levels ranging from 125,000 to 5,000,000 cells per ml. Test portions were analyzed by several analysts, using the methods described below. Background yeast levels were determined for unspiked beverages. Replicate counts were performed with the hemacytometer on test portions at levels ranging from 55,000 to 12.5 x 106 yeast cells per ml. Method for filterable liquids (e.g., water and grape juice). Portions of 10 ml were filtered through a membrane filter (AABG, pore size, 0.8 ,um; Millipore Corp.) with a Millipore disk filter holder which attaches to a standard syringe. (Portion size can be increased or decreased, depending on the level of contamination.) Because the gradations on the syringes were inaccurate, the plunger was removed, the *

syringe was attached to the filter holder, and 10 ml of liquid was pipetted into the syringe. An air cushion of about 3 ml was inserted between the plunger and the test liquid so that all of the liquid could be pressed through the filter. The filter holder was kept vertical to ensure even distribution of liquid over the filter. After filtration, the filter was placed on a microscope slide; the grid was kept parallel to the edges of the slide to facilitate counting. Method for nonfilterable liquids (e.g., orange juice) which clog the filter. Portions of 4 ml were mixed with 1 ml NaOH (25 g in 100 ml water) and shaken to suppress background interference in the fluorescence microscope (Fig. 1). After approximately 10 min, a 0.1- or 0.01-ml portion (depending on the level of contamination) was spread over a Millipore filter (AABG, 0.8 ,um) placed on a piece of bibulous paper. When the filter was dry at the surface, it was placed on a microscope slide, as described above. Microscopic counting procedure. The filter was covered with 1 drop of aniline blue (1% in 15 M K2HPO4, adjusted to pH 8.9 with K3PO4; a stock solution can be made, and age improves the fluorescence). The drop was spread over the whole filter. After approximately 5 min, a cover slip was placed over the filter. Yeasts were counted under a fluorescence microscope with excitation from blue light, lOx eyepieces with a Howard mold count or other eyepiece grid, and a 20x (or 40x) objective. Three squares of the eyepiece grid were counted in each field of the filter not covered by the gasket. Budding yeasts were counted as one cell as long as the daughter cell was obviously smaller than the mother cell. If they were approximately equal in size, they were counted as two cells. We counted only those yeasts located completely within an eyepiece square or touching the left and lower borders of the eyepiece square. We omitted yeasts touching the right and upper borders. Calculations. The area of the filter covered by one square of the eyepiece grid was determined by an objective (stage) micrometer. For filtered samples, the working area of the Millipore filter (the portion not covered by the gasket during filtration) was 380 mm2. For nonfilterable liquids, the working area of the Millipore filter was the entire filter, or 491 mm2, since no gasket was used. The number of yeasts per milliliter is expressed as (number of yeasts counted/number of eyepiece squares examined) x (working area of filter/area

Corresponding author. 599

600

APPL. ENVIRON. MICROBIOL.

NOTES

TABLE 1. Comparison of the new fluorescence procedure with the hemacytometer for recovery of yeasts from spiked beverages Procedure

Fluorescence Filtered

level Spike (yeasts/ml)

Avg count

% Coefficient No. of % Avg of variation liquids recovery

3 1 5

97.3 89.9 92.7

5.9

2 1 2

85.5 83.6 102.5

10.7 22.6 76.6

4 6 4

96.8 114.7 227.3

583,965 22,468 463

10.1

5,000,000 4,273,518 417,872 500,000 128,092 125,000

35.5

Hemacytometer 12,500,000 12,100,000 1,250,000 1,433,333 125,000 55,000

Nonfiltered

600,000 25,000 500

5.4

of one eyepiece square) x (volume of liquid)-'. Note that for nonfilterable liquids, the volume included only the net amount used and not the volume of NaOH added (i.e., 80% of the total volume applied to the filter). All genera of yeasts tested showed strong fluorescence with aniline blue. Data from the spiked beverages showed a linear relationship between count and spike level over the range of spike used. Recoveries are presented for high, medium, and low spike levels (Table 1). Results obtained by the various analysts were not significantly different. Background yeast levels in the water and juices were zero. Data from the

FIG. 1. Orange juice, not filtered, with addition of NaOH to suppress background fluorescence. Magnification, x 750.

hemacytometer were used as the reference, even though they were only an approximation. For the new fluorescence method, percent recoveries ranged from 80.6 to 108.6% with a mean of 91.5% for filtered liquids and from 54.6 to 106.9% with a mean of 86.9% for orange juice. The correlation coefficients for the relationship of count to spike level were 0.996 for filtered liquids and 0.961 for orange juice, both of which indicate highly significant relationships (P < 0.01). For cases in which replicate analyses were performed at a spike level, coefficients of variation ranged from 1.0 to 10.1% for filtered samples and from 5.9 to 35.5% for orange juice. These summary statistics show that the percent recovery is high and the coefficients of variation are low, i.e., the method is reliable. Replicate counts with the hemacytometer yielded average recoveries of >100% for all but the 12.5 x 106 yeast cells per ml level, indicating a lack of precision at low levels (Table 1). The results were especially poor at the lowest spike level, 55,000 yeast cells per ml, which had an average recovery of 227% and a coefficient of variation of 76.6% The precision of this new fluorescent staining method is good. It is more reliable than the hemacytometer at low levels of yeast contamination because it permits analysis of larger samples. The method is also easy to perform, and can be interrupted without difficulty. Once the filters are stained, they can be stored without fading before being read. This method also offers the possibility of discriminating between dead (heat or formaldehyde killed) and living yeast cells. The dead cells show fairly uniform fluorescence, and the plasma may be granular. In the living cells, the cell wall stains brighter and is more defined than the plasma, which is less prominent and very uniformly stained (Fig. 2). Aniline

FIG. 2. Discrimination of (A) dead and (B) living yeast cells. Magnification, x 1,000.

VOL. 52, 1986

blue-induced fluorescence is also useful for investigating fungal morphology and has been tested with several genera of molds (Penicillium, Aspergillus, Sporothrix, and Geotrichum), all of which showed strong fluorescence. We appreciate the assistance of Stanley M. Cichowicz and Robert S. Ferrera in testing the method.

NOTES

601

LITERATURE CITED 1. Smith, M. M., and M. E. McCully. 1978. A critical evaluation of the specificity of the aniline blue-induced fluorescence. Protoplasma 95:229-254. 2. von Sengbusch, P., J. Hechler, and U. Mueller. 1983. Molecular architecture of fungal cell walls. An approach by use of fluorescent markers. Eur. J. Cell Biol. 30:305-312.