Serratia marcescens as a bacterial pathogen of Rhagoletispomonella flies (Diptera: Tephritidae)

A nonpigmenting strain of Serratia marcescens Bizio isolated from dead and apparently diseased wild apple maggot flies, Rhagoletis pomonella (Walsh), was shown to he pathogenic to healthy apple maggot flies upon ingestion. The microorganism was detected in live adult alimentary canal organs four days post ingestion hut produced death in some flies within 24 h when flies fed on a cell concentration of 4.7 x 104 cfu/ml and within 8 h when flies fed on filter-sterilized culture medium that previously contained a 21 h culture of S. marcescens. Increasing the cell concentration 10,000 fold did not lead to an increased rate of kill. Young flies (7-10 days old) were more susceptihle to infection leading to death than were older flies (21-28 days old). The potential use of S. marcescens cells as control agents against apple maggot flies is negated hy their pathogenicity to vertehrates; however, the potential use of toxic compounds produced hy this strain of S. marcescens is discussed.


INTRODUCTION
Control of Rhagoletis pomonella (Walsh), a major apple pest in the United States, depends almost exclu sively on the application of organophosphate insecticides azinphosmethyl or phosmet. Growing concerns about public health, environmental pollution, insecticide resis tance, and rising costs associated with pest control press growers to consider alternative methods of pest control that are less invasive or more cost effective. The use of natural pathogens to cause epizootics is an alluring possi bility. Specificity, persistence, and transovarial infection are a few characteristics of insect pathogens that make them attractive for insect pest management programs. Ample information exists regarding the success of certain natural pathogens for insect control, notably Bacillus thuringiensis used to control several genera of numerous insect orders, and Bacillus popilliae used to control Japa nese Beetle, Popillia japonica Newman. Little is known about diseases in the apple maggot fly, and the informa tion that does exist (e.g. Jacques et al., 1969;Fay, 1989) does not report specific causative agents and associated pathogenicity.
To begin to understand and evaluate any potential use of a pathogen for control of apple maggot fly, i.e. use of a toxic metabolite from a pathogen, we chose to examine the pathogenicity of a nonpigmenting strain of Serratia marcescens Bizio that was isolated earlier from dead and apparently diseased wild apple maggot flies and identified using standard methods (Brenner, 1992). Strains of S. marcescens have been reported to be recovered from other tephritids such as diseased Ceratitis capitata Wei demann and Dacus (Bactrocera) dorsalis Hendel flies (Grimont & Grimont, 1978) and these hacteria may pos sess some utility as insect control agents. With this idea in mind, we examined this microorganism for its ahility to produce disease; that is, to hecome estahlished within the fly, to resist the host's immune defenses, and to exploit the host to the point of disahility and/or death. We sought to identify conditions for pathogenicity of S. marcescens to R. pomonella hy feeding S. marcescens to adult R. pomonella and examining (1) the effects of two different ingested doses, (2) age-mortality relationships, (3) persis tence and/or estahlishment of S. marcescens in alimentary canal organs of R. pomonella, and (4) the effect of the presence of S. marcescens on other hacterial populations present within alimentary canal organs of R. pomonella. We also demonstrated Koch's postulates and sought to determine if this strain produced any toxic metaholite(s) against R.pomonella.

MATERIALS AND METHODS
Bacterial cell preparation and fly feeding procedure. An 18 h culture of S. marcescens, ohtained from dead R. pomonella eclosed from pupae collected in the field, was grown in 7 ml trypticase soy hroth (Difco Lahoratories, Detroit, MI) (TSB) at 28°C, not shaken, was placed in a Dynac™ centrifuge (Clay-Adams Inc., Parsinnany, NJ) and spun at 5000 rpm for 20 min at room temperature. The supernatant fluid was discarded, and the remaining pellet of hacterial cells was resuspended in 500 pl of sterile distilled water (pH 7), which was just enough liquid to assist in the removal of the cells hy micropipette. The cell-water suspension was enumerated hy serial dilution using the spreadplate technique on prepoured trypticase soy agar (TSA) plates in triplicate. Colony-forming units were enumerated after 18 h of incuhation at 30°C. We determined that after 18 h of incuhation in TSB, this strain of Serratia in culture had reached a density * Current address. Department of Biological Sciences, California State University, Hayward, CA 94542, USA. E-mail: clauzon@csuhayward.edu; fax: (510)-885-4747. of 4.7 x 104 colony-forming units (cfu)/ml. Fifty pl of the cell suspension was then dispensed aseptically and equally onto 8 sterile paper disks (7 mm dia) made from Whatman No. 1 filter paper (Whatman Internat'l. Ltd, Maidstone, England) contained in a sterile Petri dish. Ten pl of a 10% yeast hydrolysate/sugar solution (fly food) was also added to each disk and mixed well with the bacterial cells. Apple maggot flies, eclosed from wild origin pupae, were separated into groups based on age and were starved of protein and sugar for 18 h. Flies were then placed individually into Petri dishes containing the bacterial cells on paper disks and were allowed to feed. Flies that fed only on the yeast hydrolysate/sugar solution (plus 50 pl sterile dH20) on paper disks served as controls. Flies were watched carefully and, after all flies had fed and displayed distended abdomens, the flies were removed and placed in plexiglass-wire cages con taining fresh water, sugar, and protein. Percent mortality was recorded after 24 and 120h.
High and low dose tests. A concentration of 4.7 * 108 cfu/ml, obtained by dilution technique, was chosen arbitrarily and considered to be a "high" dose of S. marcescens, while a concentration of 4.7 * 104 cfu/ml was considered to be a "low" dose of the bacterial cells. Recall that this low dose did cause mortality within 24 h (Table 1.). From this point onward, high and low doses refer to these cell concentrations respectively. Four groups of 15 adult R. pomonella were fed the high dose and 4 groups of 15 adult R. pomonella were fed the low dose of S. marcescens following the disk method described above. Four groups of 15 adult R. pomonella were fed yeast hydro lysate/sugar and served as control groups. Percent mortality was recorded at 24 and 120 h.
Age-mortality relationship tests. High and low doses of S. marcescens were fed to R. pomonella considered to be young (7-10 days old) and old (21-28 days old). Six cages for each age and dose (15 flies per cage) were tested using the paper disk method described above. "Young" and "old" fly groups serving as control groups (N = 90) were given only yeast hydrolysate/ sugar. Percent mortality was recorded at 24 and 120 h.
Detection, establishment and persistence of Serratia marc escens within the host. R. pomonella (7-10 days old) were fed a dose of 4.7 * 104 cfu/ml and were sacrificed 1-2 min, 15 min, 30 min, 1 h, 24 h, and 4 d post-feeding while at the same time parts of their alimentary canal tract (esophageal bulb, crop, and a section of midgut) were removed aseptically from each fly. The organs were placed individually into 7 ml of TSB and incu bated for 18-24 h at 28°C. The cultures were streaked individu ally onto TSA to obtain pure colonies, incubated at 28°C, and bacterial colonies were subcultured onto biochemical media for identification. The API 20E (Analytab Products, Inc., Plainview, NY) microbiological identification system was also used as part of the identification process. R. pomonella that fed on yeast hydrolysate/sugar alone were likewise dissected at the same time intervals and served as control groups. The presence, not quantity, of S. marcescens in the removed parts of the alimen tary canal was recorded. Recovered S. marcescens were fed to small groups (10) of flies (6) ranging in age between 10 to 28 days old to confirm Koch's postulates.
Toxicity tests using S. marcescens supernatant. A 21 h cul ture of S. marcescens grown in TSB and centrifuged at 5000 rpm for 15 min at 4°C. The supernatant was removed and filtersterilized using a 0.22 pm filter (Acrodisc, Gelman Sciences, Ann Arbor, MI). A sample of the supernatant was placed into TSB which was incubated overnight and observed for any bac terial growth. Lack of growth confirmed successful filter sterili zation. Fifty pl of sterile supernatant was applied to sterile paper disks as described earlier. Adult R. pomonella "young" and "old" flies, as described earlier, were allowed to feed on the soaked disks contained in a sterile Petri dish. Once the flies were observed to have fed and their abdomens were distended, the flies were individually removed and placed by age into sepa rate cages containing fresh food and water. Preliminary feeding tests revealed that some flies that fed on the supernatant died just hours after feeding. Therefore, percent mortality was recorded at 8 and 24 h. Statistical analysis. Data acquired from all studies were ana lyzed using the Statistix© Analytical Software for analysis of variancetests (SAS Institute, 1985).

RESULTS
Preliminary tests showed that a cell concentration of 4.7 * 104 cfu/ml produced death in approximately 48% of flies that were approximately 14 days old. Tests were then designed to determine a concentration of this S. marcescens needed to produce approximately 50% death of flies within a predetermined time period for a broad age range of flies. These tests showed that a cell concen tration of 4.7 * 104 cfu/ml produced death in 43% of the flies after 24 h with no appreciable additional death after 120 h (data not shown). All control flies (not exposed to the pathogen) were alive at 120 h.
Flies that fed on the two different concentrations of S. marcescens, 4.7 * 104 cfu/ml and 4.7 * 108 cfu/ml respec tively, produced nearly identical results (Table 1). There was no significant difference in fly mortality between concentrations. However, significant differences in mor tality did occur between fly ages; younger flies were sig nificantly more susceptible to infection and death than older flies at each concentration (Table 2).
To determine the presence and persistence of S. marc escens within infected R. pomonella, flies fed on bacterial suspensions and alimentary canal organs were sampled for S. marcescens starting at 1-2 min and ending at 4 days post-ingestion. S. marcescens were recovered 1-2 min after feeding and until the end of the study (Table 3 ). By day's end of the fourth day (post-ingestion), only 30 percent of the infected flies remained alive and of these 30%, 66% harbored S. marcescens. The control flies did not harbor any Serratia sp. in samples of their alimentary canals (data not shown).
Microorganisms present within the alimentary canal of R. pomonella maintained in the laboratory before experi- mentation were primarily Pseudomonas sp., Alcaligenes sp., and Acinetobacter sp. The control flies were found to contain large numbers of Pseudomonas sp., and lesser numbers of Alcaligenes sp. and Acinetobacter sp. (non fermenters), with very few Enterobacter spp. and Kleb siella spp. (butylene glycol fermenters) throughout the study. After feeding on S. marcescens, however, infected flies showed an increase in isolation of the number of bacterial species belonging to the family Enterobacteriaceae (Table 4). In these test flies, Klebsiella sp. domi nated, followed closely by Enterobacter sp., with only 1 Pseudomonas sp. and no Alcaligenes or Acinetobacter sp. recovered. S. marcescens re-isolated from test flies was subsequently used to test Koch's Postulates. We deter mined that, indeed, this strain of S. marcescens was responsible for apple maggot fly death (data not shown) in mortality studies. The supernatant from a 21 h old culture of S. marces cens was highly toxic to flies, killing on average 59% of the young flies within 8 h and 72% within 24 h. Eighty percent of old flies were dead within 24 h (Table 5). There was no difference in mortality in relation to age. DISCUSSION Our findings indicate that a nonpigmenting strain of S.  marcescens isolated from diseased and/or dead apple maggot flies kills, on average, 40% of test flies (10-28 days old) within 24 h post-ingestion at a concentration of 4.7 x 104 cfu/ml. Increasing the number of cells ingested by the flies 10,000 fold did not lead to a substantial increase in the rate or amount of kill. There may be at least five possible explanations for this occurrence: (1) the rapidity of kill eliminates or disguises any obvious dosage effect, (2) the doses tested are actually two closely-aligned points on a dose-response curve, (3) viru lence is related to an exotoxin produced by the bacteria, whose lethal concentration is not based solely on cellular number, (4) enough S. marcescens were digested by some flies to release toxic or lethal amounts of Lipid A (endo toxin) (Atlas, 1988) or (5) cells entered the insects' hemolymph via small wounds or other entrance sites during feeding. Further investigation is needed to evaluate these possibilities and/or any other subtle effects of dosage. Young R. pomonella appeared to be more sensitive to the presence of this bacterial pathogen than older more mature flies (Table 2). This is contrary to studies that sug gest decreased enzyme action and immune response are characteristic of older mature flies (Christensen et al., 1986, Jianyong et al., 1992, making them more sensitive to the effects of bacterial pathogens. In addition, during the course of our study a few of the test flies (young and old) did not appear to be adversely affected upon expo sure to the pathogen. This suggests that those flies were insensitive to the pathogen's presence or that some pro tective mechanism was present in such flies (i.e. induction of bactericidal peptides) but not in the majority of the fly population.
Several studies exist that describe acquired humoral immunity against bacteria and toxins in insects (e.g. Krieg, 1987), although Chadwick (1971) found that S. marcescens survived treatment with immunized hemolymph. In our experiments, such a protective mechanism is not known but merits further investigation. In addition, tighter age-dependency experiments should be conducted with companion microbiological and immunological analyses. It is also possible that these flies did not ingest or feed as much as the other flies, however, this is unlikely as each fly was observed to feed and only flies with extended abdomens post-feeding were used for the study.
We have also shown that S. marcescens was respon sible for the death of R. pomonella flies by fulfilling the requirements of Koch's postulates. We isolated continu ously this strain of S. marcescens from dead flies and used these cultures to reinfect additional flies whom also died. The gut environment of the fly appears to favor growth of S. marcescens and possibly other bacteria once introduced. Several different tephritid pest species pos sess few bacteria on eclosion (Lauzon et al., 2000, unpub lished). It is likely that the first bacteria that arrive to or are already present in the gut that can tolerate or thrive under gut conditions are those that become established. Enteric bacteria, Enterobacter spp. and Klebsiella spp. are isolated repeatedly from the alimentary canal organs of R. pomonella despite the numerous different types of bacteria and fungi that these flies consume while feeding on natural food sources, such as fecal material (e.g. Lauzon et al., 1998Lauzon et al., , 2000. Serratia spp., Enterobacter and Klebsiella spp. are all members of the family Enterobacteriaceae. Perhaps the gut environment is exception ally supportive of these types of bacteria. In addition, a dramatic shift in the kinds and numbers of bacteria within the alimentary canal of flies infected with S. marcescens suggests that fundamental changes are occurring within the flies when they are exposed to this pathogen. The exact mechanism for this shift is not known but could be related to pH effects on bacterial growth. In preliminary experiments, we demonstrated that S. marcescens grows well in an acidic (pH 5.0) environment (data not shown). Pseudomonas spp. and related bacteria, which were found to be dominant inhabitants of alimentary canals of laboratory-maintained apple maggot flies before testing, favor utilization of amino acids as an energy source and are sensitive to acidic pH. Conversely, Serratia spp. are rapid fermenters of carbohydrates. The rapid utilization of carbohydrates results in the release of small amounts of acids from the breakdown of glucose, and appear to tol erate an acidic environment very well. A pH shift down ward would tend to favor bacteria tolerant of an acidic environment; however, further work is necessary before a more definitive statement can be made as to why a popu lation shift occurred. Kodama & Nakasuji (1971) described a situation where the presence of Streptococ cus faecalis and S. faecium in the midgut of silkworms enhances the pathogenicity of Serratia piscatorum by lowering the midgut pH to one that favors the growth of the Serratia sp. Several investigators have reported numerous factors that affect susceptibility of insects to bacterial infection, including age (Beegle et al., 1981), plant allelochemical-induced stress (Felton & Dahlman, 1984), diet (James & Lighthart, 1992), and general changes of microbiotia within insects housed in laborato ries (Lighthart, 1988). Whatever the explanation, the strain of S. marcescens tested here does kill R.pomonella, which indicates that further study is merited to understand the nature of pathogenicity associated with this strain of S. marcescens.
S. marcescens has been reported as a pathogen of several economically important insect pests such as the boll weevil, Anthonomus grandis Boheman (Ourth & Smalley, 1980), the tobacco hornworm, Manduca sexta (Linnaeus) (Dunn & Drake, 1983), and the house fly, Musca domestica Linnaeus (Benoit et al., 1990), as well as beneficial arthropods such as the predatory mite Metaseiulus occidentalis (Nesbitt) , and the honey bee, Apis mellifera Linnaeus (El Sanousi et al., 1987). S. marcescens also has been reported as an insect pathogen of moderate virulence that causes a fatal septicemia after penetration through the insect's gut wall and subsequent invasion ofthe hemocoel (Lysenko, 1985;Krieg, 1987). Though pathogenicity has been reported for several insect species, most methodologies used include only injection of Serratia sp. into the insect (Steinhaus, 1963) rather than introduction of the bacteria into the insect through ingestion. Ingestion of S. marcescens is generally seen as of little consequence to insects (e.g. Podgwaite & Cosenza, 1976), however, Bracken & Buchner (1967) did show that the ichneumonid parasitoid Exeristes comstockii (Cresson) acquires enough S. marcescens from its larval form to cause a fatal septicemia in its own and other adults.
S. marcescens is ubiquitous in nature (Brenner, 1992), but its presence and persistence is not well characterized and it is not often recovered from wild insects (Steinhaus, 1959;Krieg, 1987). Studies aimed at examining further the ecology of S. marcescens are discouraged because the pathogenicity of S. marcescens to humans, other verte brates, and beneficial insects prevents widespread use of this microorganism per se to control insect pests. Our data indicate, however, that this strain of S. marcescens pro duces a metabolite(s) toxic to apple maggot flies, sug gesting the possible use of the toxic metabolite(s) pro duced by S. marcescens in apple maggot fly control strategies.
Toxins or proteases secreted by Serratia sp. may facili tate entrance of the bacteria into the hemocoel. At present, some 11 proteases have been described as being produced by several different strains of S. marcescens (e.g. Brenner, 1992), including chitinases (Lysenko, 1976). Poinar et al. (1979) found that S. marcescens was capable of entering the hemocoel in tsetse flies, which died soon after from fatal septicemia. Also, Asano et al. (1999) determined that synergistic effects occurred when supernatant from a culture of Serratia marcescens was used in conjunction with Bacillus thuringiensis Cry1C toxin. It remains to be determined if our strain of S. marc escens enters the hemocoel of apple maggot flies and if this entrance is facilitated by one or more toxin(s); how ever, samples of several midguts of R. pomonella that died in our study looked partially or completely disinte grated. Perhaps the condition of the midgut was due to autolysis and/or proteinases produced by Serratia growing post mortem. Serratia spp. typically produce compounds other than proteases and possibly, these com pounds were produced by our strain that feasibly poi soned the flies. These compounds would include peptides, acids and alcohols (e.g. Brenner, 1992). Current studies are underway to characterize the candidate toxin(s) produced by this strain of S. marcescens. Thus far, we have conducted experimentation aimed to charac terize the nature of the supernatant (CRL et al., unpubl.) and isolate any toxic components. Our efforts show some promising preliminary findings including indication that flies find toxic fractions palatible.