Rapid Enumeration of Microorganisms in Foods by the Direct Epifluorescent Filter Technique
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Oct. 1982, p. 809-813 Vol. 44, No. 4
0099-2240/82/100809-05$02.00/0
Copyright © 1982, American Society for Microbiology
Rapid Enumeration of Microorganisms in Foods by the Direct
Epifluorescent Filter Technique
GRAHAM L. PETTIPHER* AND UBALDINA M. RODRIGUES
National Institute for Research in Dairying, Shinfield, Reading RG2 9AT, England
Received 12 April 1982/Accepted 7 June 1982
Downloaded from http://aem.asm.org/ on March 11, 2021 by guest
Filtration of "stomachered" food suspensions through nylon filters (pore size, 5
,um) removed most of the food debris without affecting the recovery of microorga-
nisms. Two to ten milliliters of these prefiltered suspensions could be filtered in
the direct epifluorescent filter technique (DEFT). The technique takes less than 30
min to complete and has a lower sensitivity of810 PETTIPHER AND RODRIGUES APPL. ENVIRON. MICROBIOL.
la removal of food debris, microorganisms can be
concentrated by a number of methods, e.g.,
9 _ *i membrane filtration, centrifugation, or use of
cm
* ion-exchange resins. This may lead to an im-
8 provement in the sensitivity of the techniques
and possibly, in the case of impedance, a more
rapid result. To be of practical use, any method
~0 7 for removing debris from food suspensions
G)
should not affect the recovery of microorga-
40. 6 nisms.
For suspensions of most foods tested, enu-
Downloaded from http://aem.asm.org/ on March 11, 2021 by guest
C
01) 5 meration of microorganisms in the DEFT
L-
proved difficult because of the presence of food
debris. Prefiltration of the suspensions through
4 nylon filters (pore size, 5 ,um) removed most of
the food debris and only slightly reduced the
recovery of microorganisms, as determined by
3 4 5 6 7 8 9 10 the plate count (Fig. 1). The recovery rates fell
Log10 unfiltered plate count/g as the microbial content of the food increased,
FIG [.1. Relationship between plate counts of pre- e.g.,
about 100% at 104/g, 87% at 106/g, 72% at
18/g ad6%t101/g
filterec1 (y) and unfiltered (x) stomachered suspensions 10n/g, and 60% /g. Th oso mcora
for all foods tested. Line represents fitted regression
10pi
at The loss of microorga-
nisms during prefiltration is unlikely to substan-
line (y = 0.18 + 0.96x; r = 0.99). tially affect the count of microorganisms in the
food. Similar results were obtained with reus-
able stainless steel filters (pore size, 4 ,um; Don
Whitley Scientific Ltd., Baildon, Shipley, West
into a sterile disposable plastic syringe and expressed Yorkshire, England), which gave a relationship
throug,h the Swinnex filter unit into a sterile glass of y = 0.04 + 0.99x (r = 0.99), where y repre-
contaiiner. A new nylon filter was used for each sents loglo prefiltered plate count per gram, and
samplet. x represents log1o unfiltered plate count per
DEFrT count. The apparatus and reagents used, the gram. Nylon filters were used in preference to
treatmlent and filtration of samples, and the staining stainless steel filters because they were suffi-
and co)unting of bacteria were as described previously
by Petttipher et al. (6), except where indicated. The ciently inexpensive to be disposable. This was
freeze--dried enzyme trypsin (Difco Laboratories, De- more convenient; there was no need to wash
troit, Mich.), reconstituted according to the manufac- filters before assembly of the Swinnex units.
turer's instructions, was used throughout for the treat- Prefiltration of food suspensions has been
ment of food suspensions before filtration in the used to facilitate the counting of colonies on
DEFT '. In a preliminary study, the effect of various hydrophobic grid membrane filters (5) and to
combi nations of Bactotrypsin, a-amylase (Sigma clarify food suspensions without reducing the
Chemiical Co., St. Louis, Mo.; A 6880), pectinase recovery of bacteria (3). Our results support
(Sigmai, P 5146), and cellulase (Sigma, C 7502) on the these observations. After filtration through ei-
quality of the final microscopic preparations was ex- ther a nylon (pore size, 5 ,um) or a stainless steel
amine(
Platt dcount.
Decimal dilutions offood suspensions in
sterile '/4-strength Ringer solution were plated (1) with
(pore size, 4 iim) filter, most of the debris was
removed from suspensions without a large re-
plate c-ount agar (Oxoid Ltd., Basingstoke, England). duction in the plate count of the liquid. This
Coloniies were counted after 3 days of incubation at suggests that the majority of microorganisms in
300C. Unless stated otherwise, plate counts given in stomachered food suspensions are not firmly
the Re sults section are for unfiltered food suspensions. attached to large particles of food. The action of
RESULTS AND DISCUSSION the stomacher may be to wash organisms free of
the food (8). The type of material used for
Foc id debris insuspensions of samples pre- prefiltration is important, since microorganisms
pared for microbiological analysis can be trou- can adhere to membrane filter materials with
blesorne, blocking pipettes and obscuring colo- pore size diameters much larger than the parti-
nies o:n plates prepared from low dilutions (4). In cles themselves (9).
micro: scopic preparations, debris may make the Treatment with enzyme and surfactant was
countiing of microorganisms difficult or impossi- necessary to achieve efficient filtration of some,
ble. 1rhus, the removal of debris from food but not all, prefiltered food suspensions in the
suspeinsions is preferable for microscopic tech- DEFT. For unfiltered suspensions, the combina-
niquess and may also be a useful preparative tion of amylase and cellulase produced good
separr ition step for other rapid methods. After microscopic preparations from peas and corn,VOL. 44, 1982 MICROORGANISM ENUMERATION BY EPIFLUORESCENCE 811
1C 0.87) (Fig. 3). The relationship between the two
0 « counting techniques for frozen fish thawed and
4-a
c 9 _0°
/ stored at 70C for 1 to 4 days was y = 0.47 + 0.88x
0 / (r = 0.94), where y represents log1o prefiltered
U *<
I-
DEFT counts per gram, and x represents log1o
uw 8 plate counts per gram. Counts were in the range
LLI
of 104 to 1010/g.
a, 7 _S*
/ For initial counts on frozen vegetables (range,
5 x 102 to 5 x 105/g), there was poorer agree-
L.
/ ment between the prefiltered DEFT count and
a)
6
Downloaded from http://aem.asm.org/ on March 11, 2021 by guest
oo the plate count. The relationship had a correla-
0-,
0
tion coefficient of 0.66 (Fig. 4). This was mainly
o)0T- i
_/ due to counts obtained for Brussels sprouts;
V0 wO other vegetables, e.g., peas, corn, and broad
beans, had individual correlation coefficients of
4 0.8 or greater. This poor agreement is unlikely to
4 5 6 7 8 9 10 be due solely to the effect offreezing, since there
Log10 plate count/g was reasonable agreement between the two
counting techniques for frozen meat and frozen
FIG. 2. Relationship between prefiltered DEFT fish. Unlike meat and fish, vegetables are
count anid plate count for fresh meat (0) and fish (0). blanched at a temperature of about 98 to 99°C
Line rep)resents fitted regression line (y = 1 Ox - 0.41; before freezing to inactivate enzymes which
r = 0.91 would otherwise lead to deterioration of the
product during storage. The time of blanching is
based on the heat penetrability of the product;
but for prefiltered suspensions this treatment small vegetables such as peas and corn are
was no better than that obtained with trypsin. blanched for about 1 min, whereas larger vegeta-
Since tireatment with trypsin and Triton X-100 bles such as Brussels sprouts are blanched for 2
improv4ed the quality of the microscopic prepa- to 3 min. Heat treatment is known to affect the
rations for all prefiltered food suspensions test- agreement between the DEFT count and plate
ed, the treatment was used throughout this count, especially for products containing a large
study. Since this work was completed, it has number of streptococci or micrococci (7). The
been foound that the use of isopropanol in place DEFT count of peas, which comprised mainly
of etharnol as the final rinse in the DEFT consid- streptococci, exceeded the plate count by about
erably improves the intensity of fluorescence of
the microorganisms.
Depending on the food, 4 to 15 ml of stoma- 8r
chered food suspensions could be filtered
through a single nylon prefilter, and 3 to 10 ml of C"
these prefiltered suspensions could be filtered in 7F
the DEFT. A sample size of 2 ml was routinely 0
used in the DEFT, together with a microscope
factor of 57,500 (i.e., 1 microorganism per field LL 61-
= 57,500 microorganisms per g if 2 ml of food
suspension is filtered). Microorganisms were a,4) 5
morphologically distinguishable as vegetative a,
bacterial cells, spores, fungal hyphae, or yeasts. 0
Unless stated otherwise, these organisms were 4
included in the DEFT count if they fluoresced
orange. The coefficient of variation between 0
triplicate log DEFT counts on food suspensions -Jl
3'-
was812 PETTIPHER AND RODRIGUES APPL. ENVIRON. MICROBIOL.
8 plate count of the cooked chicken roll was
comprised mainly of psychrotrophic bacteria.
4J- For cream doughnut, the numerous orange-red-
7 * fluorescing yeasts seen in the DEFT were not
U- included in the count, since it was assumed they
*-
6 i _ / Z / had originated from the dough and would be
0C.)
a
0
FO nonviable after baking. Colonies on plates pre-
V
-v pared from this product were of bacterial origin.
LU 5 The DEFT counts of whole white and black
0 *3A/ o^
peppers were comprised mainly of pieces of
Afungal hyphae and bacterial spores, respective-
Downloaded from http://aem.asm.org/ on March 11, 2021 by guest
4 _
a A^Aly. This was reflected in the type of colonies
J)
0) 3 -
A/ ^ seen on the corresponding plates. For most of
the dry products tested, the DEFT count ex-
-i
ceeded the plate count by 3- to 20-fold (Table 1).
2 _________________________________ It is possible that certain procedures in the
2 3 4 5 6 7 8 manufacture of these products, e.g., heating or
drying, may have reduced the viable count but
Log10 plate count/g that these nonviable organisms are detected and
FIG. 4. Relationship between prefiltered DEFT included in the DEFT. In these cases the DEFT
count alnd plate count for frozen vegetables. Symbols: count probably reflects the microbiological qual-
0, peaLS; 0, com; A, Brussels sprouts; A, broad ity of the food before heating.
beans; 'V, carrots; O, runner beans; *, mixed vegeta- With prefiltration, the DEFT can be used to
bles. Liine represents fitted regression line (y = 1.22 + obtain a count of microorganisms in a variety of
0.88x; r = 0.66). foods in less than 30 min. As an added advantage
of the technique, the organisms can be tentative-
ly identified as bacteria, spores, yeasts, or fungi.
1.5 log units. This is probably due to nonviable For products for which agreement between the
organissms fluorescing orange and therefore be- DEFT count and plate count is good, e.g., fresh
ing in(cluded in the DEFT count. Following and frozen meat and fish, the DEFT could be
microb ial growth after thawing and refrigerated used to speed quality control during production
storage., the DEFT count and plate count of and public health investigations. For products
frozen vegetables were in closer agreement. The for which agreement between the two counting
relatioinship between the two counting tech- methods is reasonable, such as frozen vegeta-
niques for frozen vegetables thawed and stored bles, the DEFT may be useful for monitoring
at 7°C for 1 to 4 days was y = 0.89 + 0.86x (r = growth during storage and transport resulting
0.97), where y represents loglO prefiltered DEFT from inadequate cooling or mishandling. For
count per gram, and x represents loglo plate products for which agreement between the
count Iper g. Counts were in the range of 104 to DEFT count and plate count is poor, the DEFT,
1010/g. because of its rapidity, may still prove useful in
Of t;he other foods tested, there was good detecting contamination by monitoring the con-
agreemient between the prefiltered DEFT count dition of the product at various stages during its
and thiutrJnnte, cmrint for
iand- whVUoIL LUe crnkepl mentc
(VTablte 1);at. Tecrensm
h1iag manufacture.
doughrrnut, and whole pepper (Table 1). The high ACKNOWLEDGMENTS
We thank R. J. Fulford (National Institute for Research in
Dairying) for statistics and S. Phillips (Tropical Products
TABLE 1. DEFT counts and plate counts of cooked Institute, London, England) for arranging the supply of whole
pepper samples.
meats, cream cakes, spices, and dry products
LITERATURE CITED
Food
Logl0 count per g
1. American Public Health Association. 1978. Standard meth-
DEFT Plate ods for the examination of dairy products, 14th ed. Ameri-
Cooked chicken roll 7.63 7.80 can Public Health Association, Inc., New York.
Cooked ham 4.26 4.59 2. Breed, R. S., and J. D. Brew. 1916. Counting bacteria by
means of the microscope. Technical Bulletin no. 49. New
Cream doughnut 6.10 6.41 York Agricultural Experimental Station, Albany, N.Y.
Whole black pepper 7.14 6.32 3. Entis, P. 1982. Practical considerations in the filtration of
Whole white pepper 5.96 5.11 foods, p. 58-61. In R. C. Tilton (ed.), Rapid methods and
Curry powder 5.87 5.08 automation in microbiology. Proceedings of the 3rd Inter-
Flour 4.06 3.52 national Symposium on Rapid Methods and Automation in
Desiccated coconut 5.50 4.02 Microbiology. American Society for Microbiology, Wash-
Custard powder 3.24 1.90 ington, D.C.
4. Peterklin, P. I., and A. N. Sharpe. 1981. Filtering out foodVOL. 44, 1982 MICROORGANISM ENUMERATION BY EPIFLUORESCENCE 813
debris before microbiological analysis. Appl. Environ. Mi- meration of bacteria in heat-treated milk and milk products
crobiol. 42:63-65. using a membrane filtration-epifluorescence microscopy
5. Pettipher, G. L. 1981. Rapid methods for assessing bacteri- technique. J. Appl. Bacteriol. 50:157-166.
al numbers in milk. Dairy Ind. Int. 46:15-23. 8. Sharpe, A. N., and A. K. Jackson. 1972. Stomaching: a new
6. Pettipher, G. L., R. Mansell, C. H. McKinnon, and C. M. concept in bacteriological sample preparation. AppI. Mi-
Cousins. 1980. Rapid membrane filtration-epifluorescent crobiol. 24:175-178.
microscopy technique for direct enumeration of bacteria in 9. Zierdt, C. H. 1979. Adherence of bacteria, yeast, blood
raw milk. Appl. Environ. Microbiol. 39:423-429. cells and latex spheres to large-porosity membrane filters.
7. Pettipher, G. L., and U. M. Rodrigues. 1982. Rapid enu- Appl. Environ. Microbiol. 38:1166-1172.
Downloaded from http://aem.asm.org/ on March 11, 2021 by guestYou can also read
