1372

Journal of Food Protection, Vol. 62, No. 12, 1999, Pages 1372–1375

Biological Control of Postharvest Decays of Apple Can Prevent

Growth of Escherichia coli O157:H7 in Apple Wounds
W. J. JANISIEWICZ,1* W. S. CONWAY,2 AND B. LEVERENTZ2

1U.S. Department of Agriculture, Agricultural Research Service, Appalachian Fruit Research Station, Kearneysville, West Virginia 25430; and
2U.S. Department of Agriculture, Agriculture Research Service, HCQL, Beltsville, Maryland 20705, USA

MS 99-157: Received 27 May 1999/Accepted 13 August 1999

ABSTRACT
Fresh cells of the antagonist Pseudomonas syringae at 2.4 3 108 CFU/ml inoculated into wounds of ‘Golden Delicious’
apple prevented Escherichia coli O157:H7 (concentrations ranging from 2.4 3 105 to 2.4 3 107 CFU/ml) from growing in
the wounds. This occurred when the two microorganisms were co-inoculated or inoculation with E. coli O157:H7 was con-
ducted 1 or 2 days after inoculation with the antagonist. In similar tests, application of the commercial formulation of this
antagonist prevented the growth of E. coli O157:H7 in wounds when inoculated 1 or 2 days after application of the antagonist.
Populations of E. coli O157:H7 in wounds treated with water (control) before inoculation with this pathogen increased ap-
proximately 2 log units during the first 48 h after inoculation. These results indicate that biocontrol agents developed for
controlling storage decays of fruits may have the additional benefit of preventing the growth of foodborne pathogens in freshly
wounded tissue of intact and fresh-cut fruits.

Microbial contamination of fresh fruits and vegetables O157:H7 on apple tissue will reduce the risk of potential
has become of increasing concern in recent years as new illnesses due to consumption of the contaminated fruits and
outbreaks of foodborne illnesses traced to the consumption their products. It will also reduce the risk of cross-contam-
of fruits and vegetables or their products have been reported ination of fruit during postharvest handling.
(3, 5, 7, 23, 26). Outbreaks of illnesses caused by the con- Microbial control of foodborne pathogens mainly, Lis-
sumption of unpasteurized apple cider were caused by con- teria monocytogenes, Salmonella Typhimurium, and Staph-
tamination with Escherichia coli O157:H7 (3, 5). The ylococcus aureus, on meat, cheese, and vegetables has been
sources of contamination of the cider were not determined, suggested following successful laboratory experiments with
but various potential sources and contributing events before antimicrobial-producing lactic acid bacteria such as Lacto-
and after harvest have been suggested (6). Although those bacillus acidovorus, L. casei, L. plantarum, Lactococcus
outbreaks are fewer in comparison to outbreaks from the lactis, and Pediococcus spp. (19, 21, 25). The mechanism
consumption of contaminated meat and meat products, of action of these antagonists is mainly due to the produc-
there is a growing concern about such contaminations (2– tion of bacteriocins, organic acids, and hydrogen peroxide.
4). This is a concern to the fruit and vegetable industry, This has positive and negative implications as the com-
particularly to the rapidly expanding use of fresh-cut fruits pounds produced by the antagonists may rapidly kill food-
and vegetables where large areas of freshly cut uncolonized borne pathogens, but the pathogens may eventually develop
tissue are exposed. New Food and Drug Administration resistance to these compounds (11, 20). The possibility of
food safety regulations are expected for producers and han-
developing resistance to a biocontrol agent whose mecha-
dlers of fruits, vegetables, and their products to reduce the
nism of resistance is based on competitive exclusion is less
risk of future outbreaks (27).
likely to occur.
Knowledge of microbial contamination of fruits and
Microbial control of plant pathogens causing fruit de-
vegetables in general, and foodborne pathogens such as E.
cay after harvest also has been successful (14, 17). Two
coli O157:H7 in particular, is in its infancy (1, 12, 22, 29).
commercial pioneering products, one based on an antago-
There is a need to develop basic knowledge regarding
nistic bacterium and the other on a yeast, were registered
sources of contamination and the fate of these organisms
on fruits and vegetables (8). In our earlier work, we dem- by the U.S. Environmental Protection Agency in 1995 for
onstrated that E. coli O157:H7 can grow exponentially on the control of postharvest diseases of pome fruits and citrus,
freshly cut apple tissue and concluded that this should and they have been produced under the names BioSave 110
be considered by the industry during the handling and pro- and Aspire, respectively. For the past 3 years BioSave 110
cessing of the apples (16). It is axiomatic that any measures has been used successfully on a large scale for protecting
taken to prevent or reduce the establishment of E. coli pome fruits against fruit decay after harvest (24). The
mechanism of biological control of these and other antag-
* Author for correspondence. Tel: 304-725-3451. Fax: 304-728-2340; onists in this system has not been fully elucidated, but com-
E-mail: wjanisie@afrs.ars.usda.gov. petition for nutrients and space appears to play a major role
J. Food Prot., Vol. 62, No. 12 BIOLOGICAL CONTROL OF E. COLI IN APPLES 1373

(9, 13, 18, 28). There is no information as to how these

and other antagonists, developed for the control of plant
pathogens, affect foodborne pathogens.
The objective of this study was to determine the effect
of an antagonist, a saprophytic strain of Pseudomonas sy-
ringae developed for controlling postharvest decays of ap-
ple and pear and used in BioSave 110 product, on the col-
onization of fresh-cut apple tissue by E. coli O157:H7.
MATERIALS AND METHODS
Fruit. ‘Golden Delicious’ apples were harvested from a com-
mercial orchard at harvest maturity and stored at 18C. The exper-
iments were conducted on apples stored from 2 to 6 months. Be-
fore use, the apples were washed with water on a brush washer,
allowed to dry, surface sterilized with 70% ethanol, and left in
the laboratory overnight to dry and equilibrate to room tempera-
ture before inoculation the next day.

Pathogen and antagonist. The antagonist, strain L-59-66 of

the saprophytic P. syringae, was discovered in our laboratory and
was commercialized by EcoScience Corp. (Orlando, Fla.) under
the name BioSave 11 (dry formulation) and BioSave 110 (frozen
liquid formulation) for controlling postharvest decays (blue mold,
gray mold, and Mucor rot) on apples and pears. The fresh cells
of P. syringae and E. coli O157:H7 strain F510 isolated from
human patient (obtain from Dr. M. Jianhong, Department of Nu- FIGURE 1. Populations of E. coli O157:H7 in wounds of ‘Golden
trition and Food Science, University of Maryland, College Park, Delicious’ apple protected with a fresh cell preparation of P. sy-
Md.) were prepared by growing cultures in Luria Bertani broth ringae strain L-59-66. Fruit were wounded and co-inoculated with
overnight at 268C. The cultures were harvested by centrifugation P. syringae/E. coli or water/E. coli (control) mixtures, or inocu-
(Centrifuge 5416, Eppendorf, Hamburg, Germany) at 6,500 3 g lated with P. syringae or water (control) and then inoculated with
for 10 min, washed with 0.85% saline, resuspended in sterile dis- E. coli 24 h or 48 h later (24 h and 48 h precolonization). The
tilled water, and adjusted to the desired stock concentration that concentration of P. syringae was constant (;2.4 3 108 CFU/ml),
was approximately 2.4 3 108 CFU/ml. The BioSave 110 suspen- but concentrations of E. coli varied and were 2 3 105 CFU/ml
sion was prepared by suspending 0.88 g of frozen pellets of the (row A), 2 3 106 CFU/ml (row B), and 2 3 107 CFU/ml (row
product in 1 liter of water, resulting in a concentration of approx- C). Populations of E. coli were recovered within 1 h of inocula-
imately 2.4 3 108 CFU/ml. This corresponds to the minimum tion with E. coli (0 h) and after 24 and 48 h incubation at 24 C.
recommended commercial application rate of this product. Three Bars indicate standard error of the means.
10-fold dilutions of the E. coli stock suspension and a full-strength
suspension of P. syringae, both fresh cells and the commercial taining the entire wound was removed with a sterile cork borer,
formulation, were used to inoculate fruits. put in a Stomacher bag with 4.5 ml of 0.05 M phosphate buffer
(pH 6.5), and blended in a Stomacher 80 blender (Seward Medi-
Biocontrol of E. coli O157:H7 by P. syringae. ‘Golden De- cal, London) for 2 min at normal speed. The resulting slurry was
licious’ apples were wounded (one wound per fruit) by aseptically filtered through glass wool, plated in duplicate on Luria Bertani
removing a cylinder of tissue 3 mm in diameter and 3 mm deep medium using a spiral plater (Spiral Biotech, Bethesda, Md.), and
midway between the calyx and stem end. The apples were placed incubated at 378C for 24 h. These incubation conditions allowed
in plastic boxes on fruit pack trays, and 25 ml of the P. syringae/ E. coli cells to grow to visible colonies while P. syringae colonies
E. coli suspension mixture or P. syringae/water were deposited in were still not visible. The colonies were counted with a laser bac-
each wound. There were three mixture combinations containing terial colony counter (model 500 A, Spiral Biotech) and the counts
equal volumes of a full-strength P. syringae (fresh cells or the in CFU per wound were determined with the BEN 4.0 software
commercial formulation) and E. coli stock suspension diluted 10-, program (Spiral Biotech). There were three replications per treat-
100-, or 1,000-fold. Fruit inoculated with P. syringae/water at the ment and the treatments were arranged in a completely random-
beginning of the experiment were inoculated with the three con- ized block design. The experiment was conducted twice. In ad-
centrations of E. coli after 24 or 48 h of incubation. Controls dition, one set of biocontrol experiments was conducted with non-
consisted of wounded apples inoculated with water/E. coli at the pathogenic E. coli F-11775 (16) and a fresh cell preparation of P.
three concentrations at the onset of the experiment and after syringae L-59-66.
wounded fruits were incubated for 24 or 48 h. The wounds on the
fruit that were inoculated with E. coli after 24 or 48 h incubation RESULTS
were treated with 25 ml of water at the onset of the experiment.
The concentration of E. coli in the apple wounds was enumerated
Populations of E. coli O157:H7 co-inoculated with
within 1 h after application (0 h) and after incubation for 24 and fresh cells of P. syringae L-59–66 in wounds of ‘Golden
48 h at 24 C. This time span was chosen because in our earlier Delicious’ apple did not change significantly during 48 h
work (16) we had shown that there was no additional growth of of incubation. This occurred at all three initial inoculum
E. coli in apple wounds 48 h after inoculation. To recover E. coli, levels of E. coli O157:H7 (Fig. 1). During the same period,
a cylinder of apple tissue (1 cm diameter and 1 cm deep) con- populations of E. coli O157:H7 inoculated alone (control)
1374 JANISIEWICZ ET AL. J. Food Prot., Vol. 62, No. 12

formulated product 48 h prior to inoculation with E. coli

O157:H7 resulted in a slight but significant decline in E.
coli O157:H7 populations at all inoculum levels after 48 h
incubation (Fig. 2).
DISCUSSION
Our results clearly demonstrated that the antagonist P.
syringae L-59-66, used for controlling postharvest decay of
pome fruits caused by fungi such as Penicillium expansum
and Botrytis cinerea (17, 18), can also prevent the growth
of the foodborne pathogen E. coli O157:H7 on wounded
apple tissue. Co-inoculation of apple wounds with fresh
cells of the antagonist and E. coli O157:H7 prevented the
growth of the pathogen as effectively as preinoculation of
the wounds with fresh cells of the antagonist 24 or 48 h
prior to inoculation with the pathogen. The growth of E.
coli O157:H7 by as much as 2 log units in control treat-
ments was not unexpected. Previously, we demonstrated in-
creases in E. coli O157:H7 populations on freshly cut apple
tissue by more than 3 log units during 48 h at 248C (16).
In those studies, the extent of the increase depended on the
initial inoculum level and was greatest for lower and only
slight or not significant for the higher inoculum level. Al-
FIGURE 2. Populations of E. coli O157:H7 in wounds of ‘Golden though the range of the inoculum levels of E. coli O157:
Delicious’ apple protected with BioSave 110. Fruit were wounded
H7 used in this study was not as broad, it fell within the
and co-inoculated with BioSave/E. coli or water/E. coli (control)
range of the earlier studies and the population increases
mixtures, or inoculated with BioSave or water (control) and then
inoculated with E. coli 24 h or 48 h later (24 h and 48 h pre- reported here were similar to the earlier work.
colonized). The concentration of P. syringae from BioSave was P. syringae strain L-59-66 has been shown to be a very
constant (;2.4 3 108 CFU/ml), but concentrations of E. coli var- effective colonizer of apple and pear tissue (17, 18). The
ied and were 2 3 105 CFU/ml (row A), 2 3 106 CFU/ml (row mechanism of antagonism of this bacterium against E. coli
B), and 2 3 107 CFU/ml (row C). Populations of E. coli were O157:H7 may be competition for nutrients and space. This
recovered within 1 h of inoculation with E. coli (0 h) and after may be advantageous because it is very unlikely that the
24 and 48 h incubation at 248C. Bars indicate standard error of pathogens will develop resistance to this mechanism. Our
the means. experience (unpublished results) suggests that secondary
metabolites produced by this antagonist have no inhibitory
effect on bacteria. The lack of change or only a small de-
increased approximately 2 log units, 1 log unit, and re- cline in populations of E. coli O157:H7 in wounds treated
mained about the same at the lowest, intermediate, and the with the antagonist further indicate the lack of involvement
highest inoculum levels, respectively (Fig. 1). In apple of any bactericidal compounds.
wounds precolonized by P. syringae for 24 h prior to in- The various antagonist/E. coli O157:H7 ratios used in
oculation with E. coli O157:H7, populations of E. coli our studies were selected to determine interactions over a
O157:H7 declined slightly during 48 h incubation while in range of high pathogen pressure. However, under natural
the control treatment, the populations increased at a rate conditions, concentrations of E. coli most likely will be
similar to those in wounds without precolonization. The much lower than the lowest concentration used in our study.
populations in apple wounds precolonized with P. syringae Thus, the effectiveness of the antagonist in preventing col-
for 48 h prior to inoculation with E. coli O157:H7 were onization of the apple tissue by E. coli O157:H7 is expected
similar to those with 24 h precolonization. Results from to be even greater.
experiments with nonpathogenic strain of E. coli F-11775 The commercial formulation of P. syringae reduced the
and fresh cell preparation of P. syringae L-59-66 were sim- growth of E. coli O157:H7 on fruit inoculated with the
ilar to those with the pathogenic strain of E. coli O157:H7 antagonist at least 24 h before inoculation with the patho-
(data not shown). gen. This delay in comparison with the fresh cell applica-
Populations of E. coli O157:H7 co-inoculated with the tion may have resulted from inert ingredients in the for-
formulated product in wounds of ’Golden Delicious’ apple mulation of the product that may have delayed the onset of
increased to the same extent as did populations of E. coli competition or from an extended lag phase of the cells in
O157:H7 in the control treatments at all inoculum levels the formulated product. The second possibility appears to
after 48 h incubation (Fig. 2). However, inoculation of E. be less likely because, in earlier work, this antagonist was
coli O157:H7 into wounds where the formulated product equally effective in colonizing apple tissue and controlling
was applied 24 h earlier resulted in significantly lower pop- apple decay in the dry formulation as it was in the fresh
ulations than in the control treatments. Application of the cell preparation (17). Thus, it is important, when develop-
J. Food Prot., Vol. 62, No. 12 BIOLOGICAL CONTROL OF E. COLI IN APPLES 1375

ing a commercial antagonist formulation for controlling 9. Bull, C. T., M. L. Wadsworth, K. N. Sorensen, J. Y. Takemoto, R.
K. Austin, and J. L. Smilanick. 1998. Syringomycin E produced by
postharvest fruit decay, that the effect of the formulation
biological control agents controls green mold on lemons. Biol. Con-
on foodborne pathogens should be considered as well. trol 12:89–95.
Our study was restricted to the effect of the P. syringae 10. Conway, W. S., W. J. Janisiewicz, A. E. Watada, and C. E. Sams.
antagonist on E. coli O157:H7, but other foodborne path- 1998. Survival and growth of Listeria monocytogenes on fresh cut
ogens, such as L. monocytogenes, that have been shown to apple slices. Phytopathology 88(Suppl.):S18.
11. Crandall, A. D., and T. J. Monville. 1998. Nisin resistance in Listeria
grow on freshly cut apple tissue may also be affected by
monocytogenes ATCC 700302 is a complex phenotype. Appl. En-
this antagonist (10). Both of these pathogens may cause viron. Microbiol. 64:231–237.
contamination problems in the rapidly expanding fresh-cut 12. del Rosario, B. A., and L. R. Beuchat. 1995. Survival and growth
industry, which, in the near future, most likely will include of enterohemorrhagic Escherichia coli O157:H7 in cantaloupe and
apple slices. Biocontrol agents could be selected from an watermelon. J. Food Prot. 58:105–107.
13. Droby, S., E. Chalutz, C. L. Wilson, and M. Wisniewski. 1989. Char-
array of antagonists such as those isolated from exposed
acterization of the biocontrol activity of Debaromyces hansenii in
apple tissue in the orchard in our earlier work (15). Good the control of Penicillium digitatum on grapefruit. Can. J. Microbiol.
candidates for this competitive exclusion can also be yeasts 35:794–800.
that naturally occur in apple cider that may be presumed to 14. Droby, S., L. Cohen, A. Daus, B. Waiss, B. Horev, E. Chalutz, H.
be safe for human consumption and may not adversely af- Katz, M. Keren-Tzur, and A. Shachnai. 1998. Commercial testing of
Aspire: a yeast preparation for the biological control of postharvest
fect the organoleptic or aesthetic qualities of the slices. The decay of citrus. Biol. Control 12:97–101.
potential usefulness of these yeasts is further supported by 15. Janisiewicz, W. J. 1996. Ecological diversity, niche overlap, and co-
our earlier observations that in cider, artificially contami- existence of antagonists used in developing mixtures for biological
nated with E. coli O157:H7, increases in yeast populations control of postharvest diseases of apples. Phytopathology 86:473–
over time coincided with the decline of E. coli O157:H7 479.
16. Janisiewicz, W. J., W. S. Conway, M. W. Brown, G. M. Sapers, P.
populations (16). Fratamico, and R. L. Buchanan. 1999. Fate of Escherichia coli
The successful use of P. syringae or its commercial O157:H7 on fresh-cut apple tissue and its potential for transmission
formulation for control of postharvest decays of pome fruits by fruit flies. Appl. Environ. Microbiol. 65:1–5.
(24) has the additional, previously unforseen, benefit of re- 17. Janisiewicz, W. J., and S. N. Jeffers. 1997. Efficacy of commercial
formulation of two biofungicides for control of blue mold and gray
ducing the risk of potential contamination by the foodborne
mold of apples in storage. Crop Protect. 16:629–633.
pathogen E. coli O157:H7 and possibly other foodborne 18. Janisiewicz, W. J., and A. Marchi. 1992. Control of storage rots of
pathogens as well. This adds new dimension to biological various pear cultivars with a saprophytic strain of Pseudomonas sy-
control research on harvested fruit and other commodities. ringae. Plant Dis. 76:555–560.
19. Lewus, C. B., A. Kaiser, and T. J. Montville. 1991. Inhibition of
ACKNOWLEDGMENT food-borne bacterial pathogens by bacteriocins from lactic acid bac-
teria isolated from meat. Appl. Environ. Microbiol. 57:1683–1688.
We thank C. Sharer for technical assistance with the experiments 20. Ming, X., and A. Daeschel. 1993. Nisin resistance of foodborne bac-
and for preparing figures. teria and the specific resistance response of Listeria monocytogenes.
Scott A. J. Food Prot. 56:944–948.
REFERENCES 21. Muriana, P. M. 1996. Bacteriocins for control of Listeria spp. in
food. J. Food Prot. 1996(Suppl.):54–63.
1. Abdul-Raouf, U. M., L. R. Beuchat, and M. S. Ammar. 1993. Sur- 22. Nguyen, T.-C., and F. Carlin. 1994. The microbiology of minimally
vival and growth of Escherichia coli O157:H7 on salad vegetables. processed fresh fruit and vegetables. Crit. Rev. Food. Sci. Nutr. 34:
Appl. Environ. Microbiol. 59:1999–2006. 371–401.
2. Anonymous. 1994. Foodborne pathogens: risk and consequences. 23. Slutsker. L., A. A. Reis, K. D. Greene, J. G. Wells, L. Hutwagner,
Task force report, ISNN 01944088; Council for Agricultural Science and P. M. Griffin. 1997. Escherichia coli O157:H7 diarrhea in the
and Technology no. 122, Sept. 1994. Ames, Iowa. United States: clinical and epidemiological features. Ann. Intern.
3. Anonymous. 1997. Outbreaks of Escherichia coli O157:H7 infection Med. 126:505–513.
associated with drinking unpasteurized apple cider—October 1996. 24. Stack, J. P. 1998. Postharvest biological control: commercial suc-
Update on emerging infections from the Center for Disease Control cesses and a model for public and private sector cooperation. Invited
and Prevention. Ann. Emerg. Med. 29:645–646. Papers, Abstracts—1, 7th International Congress of Plant Pathology,
4. Bean, N. H., and P. M. Griffin. 1990. Foodborne disease outbreaks Edinburgh, Scotland, BSPP, London.
in the United States, 1973–1987: pathogens, vehicles, and trends. J. 25. Stevens, K. A., B. W. Sheldon, N. A. Klapes, and T. R. Klaenham-
Food Prot. 53:804–817. mer. 1991. Nisin treatment for inactivation of Salmonella species and
5. Besser, R. E., S. M. Lett, J. T. Weber, M. P. Doyle, T. J. Barrett, J. other gram-negative bacteria. Appl. Environ. Microbiol. 57:3613–
G. Wells, and P. M. Griffin. 1993. An outbreak of diarrhea and he- 3615.
molytic uremic syndrome from Escherichia coli O157:H7 in fresh- 26. Tauxe, R. V. 1997. Emerging foodborne diseases: an evolving public
pressed apple cider. JAMA 269:2217–2220. health challenge. Emerg. Infect. Dis. (CDC) 3:425–434.
6. Beuchat, R. L., and J. H. Ryu. 1997. Produce handling and process- 27. Warner, G. 1997. Fresh packers await new food safety guidelines.
ing practices. Emerg. Infect. Dis. (CDC) 3:459–465. Goodfruit Grower 48:7–8.
7. Brackett, R. E. 1993. Microbiological spoilage and pathogens in 28. Wisniewski, M., C. Biles, S. Droby, R. McLaughlin, C. L. Wilson,
minimally processed refrigerated fruits and vegetables. In R. Wiley and E. Chaluz. 1991. Mode of action of postharvest biocontrol yeast,
(ed.), Minimally processed refrigerated fruits and vegetables. Van Pichia guilliermondi. I. Characterization of attachment to Botrytis
Nostrand Reinhold, New York. cinerea. Physiol. Mol. Plant. Pathol. 39:245–258.
8. Buchanan, R. L. 1997. Identifying and controlling emerging food- 29. Zhao, T., M. P. Doyle, and R. E. Besser. 1993. Fate of enterohem-
borne pathogens: research needs. Emerg. Infect. Dis. (CDC) 3:517– orrhagic Escherichia coli O157:H7 in apple cider with and without
521. preservatives. Appl. Environ. Microbiol. 59:2526–2530.