JOURNAL OF BACTERIOLOGY, Jan. 1976, p. 119-124 Copyright O 1976 American Society for Microbiology
Vol. 125, No. 1
Printed in U.S.A.
Starvation Survival of Salmonella enteritidis ROBERT E. DRUILHETI* AND JOSEPH M. SOBEK Department of Microbiology, University of Southwestern Louisiana, Lafayette, Louisiana 70501 Received for publication 18 August 1975
Washed cells of Salmonella enteritidis harvested from a defined medium during logarithmic growth were subjected to starvation in pH 7 phosphate buffer at 37 C. Viability was measured by slide cultures and plate counts. The survival of cell suspensions equivalent to 1 to 10 mg (dry wt)/ml was influenced by cryptic growth. The rate of cryptic growth, assessed by plate counts, increased with cell density and could not be alleviated by starvation with dialysis. Dialysis of the starving culture did retard the onset of cryptic growth but did not eliminate it, indicating that the major substrates for regrowth were relatively large cellular components. In phosphate buffer, 6.7 homologous heat-killed cells allowed for the doubling of one S. enteritidis cell. Cryptic growth was not observed when cells were starved on the surface of membrane filters or in suspensions equivalent to 20 jig (dry wt)/ml (10' cells/ml). Similar half-life survival times were calculated for both these populations, but the shape of their survival curves differed significantly. These differences were attributed to stress factors encountered during cell preparation and during starvation. The half-life survival time of S. enteritidis starved at 20 jig (dry wt)/ml was 140 h in phosphate buffer, 82 h in 3,6-endomethylene-1,2,3,6-tetrahydrophthalic acid buffer, and 77 h in tris(hydroxymethyl)aminomethane buffer. The starvation survival of a variety of sapro- 0.5; (NH4)2SO4, 2.0; NaCl, 0.5; glucose, 5.0; final pH, phytic bacteria has been discussed in compre- 7.0. Liquid cultures in 300-ml nephelo culture flasks ml of hensive articles (3, 4, 16, 22, 23, 25), but no (Bellco Glass Inc., Vineland, N.J.) containing a50model medium were incubated at 37 C on starvation survival studies of Salmonella defined rotary shaker (New Brunswick Scientific Co., Inc., enteritidis have been reported. Tannock and VS N.J.). Growth was assessed by moniNew Smith (32) related bovine salmonellosis to the toringBrunswick, absorbancy at 540 nm. survival of S. typhimurium and S. bovismorStarvation procedure. Survival studies were bificans on pasture and in water. Tichacek (33) begun with middle-logarithmic-phase cells collected studied the survival of a dried suspension of S. at an optical density at 540 nm (OD.4o) of 0.38 to 0.45 enteritidis in relation to environmental in- by centrifugation at 5,000 x g for 5 min at room fluences. Other reports relating Salmonella temperature (23 C). The cell pellet was washed three survival to the natural environment have been times in buffer (37 C) and suspended in buffer to the cell density. The period between collection concerned with recovery from unpolluted water desired initiation of starvation experiments averaged 35 (7), growth in dialysis sacs placed in river water and (12, 13), and recovery from stream sediments min. Three starvation buffer systems were used: potas(11). sium phosphate, tris(hydroxymethyl)aminomethane The present study was undertaken to deter- (Tris), and 3,6-endomethylene-1,2,3,6-tetrahydrophmine the survival characteristics of S. thalic acid (EMTA). Each buffer was prepared at enteritidis in a non-nutrient environment. The 0.067 M, pH 7.0, and sterilized by membrane filtraeffects of cell density, buffers, and the method tion. Viability studies with 20-ml cell suspensions at of starvation were considered. densities equivalent to 1.0 to 10 mg (dry wt)/ml were
made in glass tubes (22 by 200 mm), and the effect of buffers on the survival of dilute cell suspensions (20 500-ml round-bottom ;sg/ml) was determined inwere aerated with sterile flasks. The cell suspensions water-saturated air to compensate for evaporation I Present address: Department of Internal Medicine, Uni- during starvation at 37 C. Cellulose dialysis memversity of Texas Medical School at Houston, Houston, Tex. branes (Union Carbide Corp., Chicago, Ill.) were used to study the effect of removal of extracellular material 77025. 119
MATERIALS AND METHODS Growth conditions. S. enteritidis (ATCC 13076) was grown in a defined medium composed of (grams per liter): K,HPO4, 9.2; KH2PO4, 5.2; MgSO4-7H2O,
120
DRUILHET AND SOBEK
on the survival pattern of S. enteritidis. The membranes were washed repeatedly in hot distilled water before use. A washed-cell suspension of known density was placed in the sterile dialysis chamber (21.5 by 260 mm) and aerated at a rate of 20 ml of air per min. The cells were dialyzed against 1,400 ml of aerated phosphate buffer replaced at a rate of 1 or 4 ml/min. Samples were removed at regular intervals for viability determinations. In one series of experiments the dialysis apparatus was modified to correct for any volume changes due to sampling. The chamber used for this purpose was a graduated glass cylinder (35 by 250 mm) closed at one end with cellulose membrane. The cylinder, supplied with internal aeration and sample tubes, was suspended in a 2-liter flask by a rubber cap cut to receive the 35-mm tube. After sampling, the menisci of the cell suspension and the dialysis buffer were made to coincide by adjusting the cylinder depth in the reservoir. Starvation on membrane filters. Cells were grown at 37 C in the defined medium under shake conditions, and 10 ml (OD,40 = 0.4) was removed, sedimented at 5,000 x g, washed twice, and suspended in 10 ml of phosphate buffer. A 10- 8 dilution of cells was prepared in phosphate buffer (37 C) and 10 ml was filtered through each of 24 filter membranes (Millipore type HA, 47 mm, 0.45 lAm). The filter was placed on a washed, sterile, membrane adsorbent pad (Millipore Corp.) saturated with phosphate buffer. Each petri dish reservoir contained three adsorbent pads, a filter membrane with 120 to 150 dispersed cells, and 10 ml of buffer. The petri dish lid was lined with filter paper to absorb water condensate formed during starvation. The entire procedure required approximately 45 min. Viability determinations. The one-step slide culture method of Postgate et al. (24) was used to assess cell viability. This method was reported to determine the ratio of viable to total cells with Poissonian precision and to be accurate within the range of 5 to 100% viability (23). The method was modified by the use of an annular glass ring (18 by 8 mm) supported on a cover glass (22 by 40 mm) and the use of 0.2 ml of Trypticase soy agar (TSA). The TSA (Baltimore Biological Laboratories) was filtered hot before sterilization. Slide culture inocula used with concentrated cell suspensions were made by adding 0.1 ml of a 10-2 dilution to the agar surface; 0.02 ml of undiluted solution was used when cells were starved in dilute suspension. An inoculum size of 5 x 108 cells provided 20 to 40 cells per high-power microscopic field. Slide cultures were incubated at 37 C for 6 to 8 h in a moisture chamber and observed with a phase contrast microscope. A microcolony resulting from one or more cell division was scored as one viable unit; a single cell, often coccoid in appearance, was scored as one dead unit. Sufficient microscopic fields were counted to reach a total of 300 objects; coincidences were disregarded, and the viability was determined as the percentage of viable units in the total number of objects counted. Slide cultures were prepared in duplicate, and viabilities were calculated from the average number of viable and dead cells. Variabilities
J. BACTERIOL.
of 1% or less were routinely observed with replicate preparations. In some experiments plate counts on TSA were used to estimate viability. Plate count viabilities were determined in triplicate and reported as change in colony-forming units with time of starvation. The survival of cells starved on membrane filters was determined as follows. At each sample time, three filters were selected at random, transferred to TSA (MacConkey agar in one experiment), and incubated at 37 C for 24 h. The filters were removed from the agar surface, placed in ethanol (95%, vol/vol) for 1 min, and transferred to methylene blue (0.6 g/100 of ml ethanol) for 30 s and then to a 1:750 aqueous solution of Zephiran chloride (Sterling Drugs, Inc., New York, N.Y.). Colonies were removed from the filter by shaking in the Zephiran solution, and the filter was placed in 5% (wt/vol) phenol for 2 min. The colonies developed from viable cells were seen as clear areas in a pale blue background. Mean colony counts of three filters were used to calculate viability at each sample interval. Cryptic growth. A modification of the method described by Nioh and Furusaka (20, 21) was designed to determine the minimal concentration of dead cells that would support growth of S. enteritidis. Cells were grown in defined medium to an OD.40 of 0.60, collected by centrifugation, and washed with physiological saline. The suspension was diluted in saline to 2.35 x 109 cells/ml (determined by plate count), and four aliquots were sedimented at 10,000 x g. The cells were suspended in the defined medium minus the carbon source or minus the carbon and nitrogen source, or 0.067 M phosphate buffer. Each suspension was heated for 15 min in a boiling water bath and diluted 1:10 and 1:100 with the appropriate solution. The heat-killed suspension (20 ml) was transferred to 100-ml flasks, inoculated with 106 cells, and incubated at 37 C. Viable counts on TSA were made at 0, 8, and 24 h.
RESULTS Starvation of concentrated cell suspensions and cryptic growth. Figure 1 is representative of the survival pattern of dense cell suspensions of S. enteritidis starved at 37 C in 0.067 M phosphate buffer. The viability of a 10-mg/ml cell suspension decreased 3% in the first 6 h of starvation. Later samples indicated an increase in the percentage of viability. The number of viable units in the population starved for 36 h was 40% greater than the number originally added to the starvation system. Plate counts were used to follow the viability of these cells. Similar results were obtained with cell densities from 1 to 10 mg (dry wt)/ml. With cell suspensions of 1.0 mg/ml, regrowth was observed after 18 h and the population reattained the initial viability (100%) after 40 h.
STARVATION SURVIVAL OF S. ENTERITIDIS
VOL. 125, 1976
121
growth. The less dense cell suspension, 2.0 mg/ml, also showed this characteristic increase in viability after 48 h. A fourfold increase in the buffer exchange rate (4.2 ml/min) failed to /4< significantly alter the survival pattern (Fig. / -~ 2B). Complete renewal of dialyzing buffer was achieved every 30 min. The faster exchange rate retarded the onset of cryptic growth but did not eliminate it. To minimize the influence of volume changes on the viability of dialysis suspensions, the experiment was repeated in a different dialysis system designed to correct for volume changes due to sampling. The results of this study are shown in Fig. 2C. The viability pattern in this 2 ,4 ,* ,system was not significantly different from the 3 io survival curves obtained in other dialysis sys9 24 12
,-_~ 4~ ~
5, 100 a
c0
0:
0
Time
(hours)
F IG. 1. Survival curve of S. enteritidis in 0.067 M phowsphate buffer (pH 7) showing cryptic growth. Cell demsities equivalent to 10 mg (dry wt)/ml (0) and 1 mg (dry wt)/ml (0) were starved under forced aeration at 37 C. Points were determined by plate counts. Dasi!hes represent plate count values greater than thos ;eobtained at 0 h (100% viability).
(27, 30) was [dhe cryptlc growth phenomenon of S. enteritidis diedbyfollowingcthengrowth in varying concentrations of its homologous stu
heeEt-killed cell suspension (Table 1). When hefat-killed cells were substituted for glucose in tne caetinect meciium, z.;3 x iuw ceiis increasea the 24-h plate count by 2,200-fold and 2.35 x 108 dead cells supplied sufficient carbon to increase the colony-forming units by 178-fold. Cryptic growth was less when dead cells supplied both carbon and nitrogen in the otherwise defined medium. Cryptic growth diminished when the dead-cell concentration was decreased. When the concentration of heat-killed cells was 2.35 x 107 cells/ml of phosphate buffer, one new cell was produced at the expense of approximately 6.7 heat-killed cells. Starvation in dialysis. The survival of S. enteritidis was studied in several dialysis systems in an attempt to reduce cryptic growth in dense cell suspensions. Dialysis was used to remove small-molecular-weight components released by dead cells. Figure 2A represents the viability of two 100-ml cell suspensions (2 and 8 mg [dry wtyml) dialyzed against 1,400 ml of phosphate buffer at an exchange rate of 1.0 ml/min. A complete buffer turnover was accomplished every 24 h. The percentage of viability of the 8.0-mg/ml cell suspension decreased by 25% for the first 24 h of starvation and then increased in a manner characteristic of cryptic
tems.
Starvation in dilute suspensions and the effects of buffers on survival. The effect of
phosphate, Tris, and EMTA buffers on the survival pattern of cell suspensions equivalent to 20 ,gg (dry wt)/ml is shown in Fig. 3. A 50% loss of viability was observed at 4 days in EMTA (Fig. 3A) compared with only 30% loss of viability in phosphate buffer (Fig. 3C). Since possible contaminants in EMTA could have caused a decrease in viability through substrate-accelerated death (26, 31) or growth inhibition, the experiment was repeated with buffer prepared from recrystallized material. An apTABLE 1. Growth of S. enteritidis in varying concentrations of homologous heat-killed cells Heat-
Medium
Definedd minus
carbon
killed cells' (x1O-)
235 23.5 2.35
Defined minus 235 23.5 carbon and
CFU°/ml (x 10-') 8h
0.50 750 0.46 35 0.49 53
nitrogen
2.35
0.54 320 0.53 21 0.48 16
0.067 M
235 23.5 2.35
0.57 350 0.50 28 0.51 27
phosphate buffer, pH 7.0
in
__
0h
Overall increase
24 h
CFUc
(x10-')
1,100 22.0 82 30
1.78 0.61
820 15.2 42 0.80 20 0.41 360 54 35
6.32 1.08 0.70
a Concentration determined by plate counts on TSA before heating. Values represent total number of dead cells in the medium. b CFU, Colony-forming units. c Increase in CFU is equal to the 24-h growth yield divided by the initial CFU. d The defined medium is described in the text.
122
J. BACTERIOL.
DRUILHET AND SOBEK
creased more slowly to 55%, and thereafter the death rate was increased and approached that observed in the initial 2 days of starvation. At the end of 9 days, 11% of the population remained viable. Starvation on membrane filters. This method of starvation was used to completely rule out the effects of cryptic growth on survival. The viability of dispersed cells starved on membrane filters saturated with phosphate buffer is shown in Fig. 4. The survival pattern of these cells is characterized by an initial rapid loss of viability during the first 2 days, a 3- to 4-day period of stasis, and a terminal slow decrease in viability beginning at day 6. The medium used to assess viability, TSA (Fig. 4A) or MacConkey agar (Fig. 4B), had little effect on the recovery of cells after the 2nd day of starvation. The possibility that increased longevity of cells starved on membrane filters resulted from nutrients remaining in the adsorbent pads was examined. Cells were starved on filters supported on the surface of agar prepared with 0.67 M phosphate buffer and 1.5% (wt/vol) washed agar. The viability pattern of cells starved in this manner did not differ from the previous results.
100-
7550A
100a
750.1C)0C
50-
B
100 -
7550-
Ic 0
1
I
I
12
24
36
48
60
60
100-
96
Timo (hours)
FIG. 2. Cryptic growth of S. enteritidis starved in dialysis at densities equivalent to 8 mg (dry wt)/ml (0) and 2 mg/ml (0). Cells (100 ml) were enclosed in cellulose membrane and continuously dialyzed against 1,400 ml of aerated phosphate buffer (0.067 M, pH 7) at 37 C. Buffer exchange rates of I ml/min (A) and 4 ml/min (B and C) were used and volume corrections due to sample removal were made (C). Points were determined by slide culture (B and C) and by both slide culture and plate count (A).
=
0-
>
60-
1.
*
20.
U0 a
A
0
2
3
2
1
Ti m-
3
4
(days)
100-
parent pKa' of 6.30 was calculated from the titration curve of recrystallized EMTA. The viability of cells starved in recrystallized EMTA was approximately 10% higher than in buffer prepared from commerical material. However, the cells survived better in phosphate buffer than in EMTA. After 4 days of starvation, 67% of the population in phosphate buffer was viable whereas the viability in recrystallized EMTA was only 58%. The survival of S. enteritidis in Tris buffer (Fig. 3B) was no better than in EMTA. The percentage of viability of S. enteritidis in phosphate buffer (Fig. 3C) decreased to 80% in 2 days, between day 2 and day 5 viability de-
20so
2
4
Time
6
0
10
(days)
FIG. 3. Survival pattern of S. enteritidis in commercial (0) and recrystallized (0) EMTA (A), Tris (B), and phosphate buffers (C). Washed cell suspensions equivalent to 20 gug (dry wt)/ml were starved in 0.067 M buffer, pH 7, at 37 C. Viabilities were determined by slide culture.
STARVATION SURVIVAL OF S. ENTERITIDIS
VOL. 125, 1976 0.-
7A
a
A
°
2
4
os * Days
2 of
6 *
Starvtion
FIG. 4. Survival characteristics of S. epteritidis starved on the surface of membrane filters and subsequently transferred to TSA (A) or MacConkey agar (B). Points were determined as described in text.
DISCUSSION Penicillin and chloramphenicol will prevent cryptic growth (25), but these inhibitors may alter intracellular changes taking place during starvation. Although dense cell suspensions would make measurements of starvationinduced intracellular changes more feasible, the cryptic growth response of S. enteritidis requires that endogenous metabolism studies be performed with dilute suspensions. Even the estimate of 6.7 heat-killed cells accounting for the doubling of a single S. enteritidis cell may be too high since the nutrient supplied by unstarved heat-killed cells may differ greatly from those liberated by the death of starved cells. One would expect that starvation before death would increase the number of cells required per cell division because, since readily utilizable endogenous components would be the first to go during starvation, they would not be available as nutrients for cryptic growth. The ability of this organism to grow on the products of dead cells was greater than that reported for Arthrobacter (20, 21), a soil organism of marked ability to survive starvation. It would be interesting to see whether this ability also extends to heterologous heat-killed cells. Postgate and Hunter (25) reported that death of 50 members of a starved suspension of Aerobacter aerogenes allowed the doubling of one survivor. Ambiguity that might arise from cryptic growth during starvation of A. aerogenes was eliminated by using a 20-,g/ml cell suspension (25). With S. enteritidis, this cell concentration was equivalent to 106 cell/ml, far below the levels supporting cryptic growth on homologous heat-killed cells. However, even at these low cell concentrations, materials released by dead cells could affect survival. Dialysis of heavy cell suspensions during starvation did not eliminate the cryptic growth response but confirmed the essentially nondi-
123
alyzable nature of cryptic growth substrates. The utilization of dialyzable cell products probably accounted for the rapid onset of cryptic growth in survival studies with concentrated cell suspensions since dialysis extended the death phase before the onset of cryptic growth. Dispersed cells starved on membrane filters showed differences in the survival pattern when compared with cells starved in dilute suspensions. These differences may be the result of stress of the population. The stress of filtration could have caused the rapid initial loss of 40 to 50% of the membrane-starved population. This rapid rate of death ceased after the 2nd day and no significant change in viability was observed until the 5th or 6th day of starvation. Cells that survived during this 2- to 6-day period were either not affected or had recovered from the filtration stress. Cells starved in dilute suspensions were not subjected to this stress and showed a much slower loss in viability during the first 2 days of starvation. The 50% survival time of S. enteritidis starved in dilute suspension with phosphate buffer is 140 h, and is 132 h when starved on membrane filters. Between the 2nd and 6th days of starvation, the death rate of cells starved on filters was only 2% per day, whereas cells starved in suspension lost viability at an average rate of 7.8% per day. Cells in suspension were subjected to the stress of forced aeration and agitation, whereas cells starved on filters were not forcibly aerated. In EMTA and Tris buffers the half-life survival time was 82 and 77 h, respectively. EMTA was reported as a non-metabolizable buffer suitable for bacterial studies at pH 7.0 in media free from nitrogen, phosphate, and sulfur (18). Tris-saline buffer was used by Postgate and Hunter (25) in survival studies of A. aerogenes. Decreased viability in the absence of phosphate was possibly due to trace contaminants in EMTA, since survival was enhanced with recrystallized EMTA, or to the poor buffering capacity of Tris at pH 7. The presence of phosphate in the starvation environment does not seem to be critical to the starvation survival of S. enteritidis. If phosphate were an important factor in maintaining cell longevity, a much more drastic change in the viability pattern would probably have been observed in EMTA and Tris buffers. This was not the case. Several workers have suggested that polyphosphates may serve as potential maintenance energy sources (1, 14, 15, 19). This view was dispelled, at least for A. aerogenes, by Harold and Harold (10). Although the polyphosphate content of S. enteritidis was not measured, according to cur-
124
J. BACTERIOL.
DRUILHET AND SOBEK
rent theories of polyphosphate accumulation (8, 9, 28) the growth environment was not conductive to the accumulation of this potential re-
water. Appl.
Microbiol. 21:379-380.
12. Hendricks, C. W. 1972. Enteric bacterial growth rates in
river water. Appl. Microbiol. 24:168-174. 13. Hendricks, C. W., and S. M. Morrison. 1967. Multiplica-
tion and growth of selected enteric bacteria in clear mountain stream water. Water Res. 1:567-576. 14. Hou, C. I., A. F. Gronlund, and J. J. R. Campbell. 1966. Influence of phosphate starvation on cultures of Pseudomonas aeruginosa. J. Bacteriol. 92:851-855. 15. Kornberg, S. R. 1957. Adenosine triphosphate synthesis from metaphosphate by an enzyme from Escherichia coli. Biochim. Biophys. Acta 26:294-300. 16. Lamanna, C. 1963. Endogenous metabolism with special referance to bacteria. Ann. N.Y. Acad. Sci. 102:515-793. 17. McFeters, G. A., and D. G. Stuart. 1972. Survival of coliform bacteria in natural water: field and laboratory studies with membrane-filter chambers. Appl. Microbiol. 24:805-819. 18. Mallette, M. F. 1967. A pH 7 buffer devoid of nitrogen, sulfur, and phosphorus for use in bacteriological systems. J. Bacteriol. 94:283-290. 19. Mallette, M. F., C. I. Cowan, and J. J. R. Campbell. 1964. Growth and survival of Escherichia coli in medium limited in phosphate J. Bacteriol. 87:779-785. 20. Nioh, I., and C. Furusaka. 1968. Growth of bacteria in the heat-killed cell suspensions of the same bacteria. J. Gen. Appl. Microbiol. 14:373-385. 21. Nioh, I., and C. Furusaka. 1970. Factors affecting growth yield of Arthrobacter simplex in its heat-killed cell suspension. J. Gen. Appl. Microbiol. 16:115-126. 22. Postgate, J. R.1967. Viability measurements and the survival of microbes under minimal stress. Microb. Physiol. 1:1-23. 23. Postgate, J. R. 1969. Viable counts and viability, p. LITERATURE CITED 611-628. In J. R. Norris and D. W. Ribbons (ed.), 1. Bohinski, R. C., and M. F. Mallette. 1967. Response of Methods in microbiology, vol. 1. Academic Press Inc., log-phase cells of Escherichia coli to medium limited in New York. both sulfate and phosphate. J. Bacteriol. 93:1316-1326. 24. Postgate, J. R., J. E. Crumpton, and J. R. Hunter. 1961. 2. Clifton, C. E. 1966. Aging of Escherichia coli. J. Bacteriol. Measurement of bacterial viability by slide culture. J. 92:905-912. Gen. Microbiol. 24:15-24. 3. Dawes, E. A., and D. W. Ribbons. 1962. The endogenous 25. Postgate, J. R., and J. R. Hunter. 1962. The survival of metabolism of microorganisms. Annu. Rev. Microbiol. starved bacteria. J. Gen. Microbiol. 29:233-263. 16:241-264. 26. Postgate, J. R., and J. R. Hunter. 1964. Acceleration of 4. Dawes, E. A., and D. W. Ribbons. 1964. Some aspects of death of Aerobacter aerogenes starved in the presence the endogenous metabolism of bacteria. Bacteriol. Rev. of growth-limiting substrates. J. Gen. Microbiol. 28:126-149. 34:459-473. 5. Dawes, E. A., and D. W. Ribbons. 1965. Studies on the 27. Ryan, F. J. 1955. Spontaneous mutation of non-dividing endogenous metabolism of Escherichia coli. Biochem. bacteria. Genetics 40:726-738. J. 95:332-343. 28. Smith, I. W., J. F.Wilkinson, and J. P. Duguid. 1954. 6. Ensign, J. C. 1970. Long-term starvation survival of rod Volutin production in Aerobacter aerogenes due to and spherical cells of Arthrobacter crystallopoietes. J. nutrient imbalance. J. Bacteriol. 68:450-463. Bacteriol. 103:569-577. 29. Strange, R. E. 1968. Bacterial "glycogen" and survival. 7. Fair, J. F., and S. M. Morrison. 1967. Recovery of Nature (London) 220:606-607. bacterial pathogens from high quality surface water. 30. Strange, R. E., F. A. Dark, and A. G. Ness. 1961. The Water Resour. Res. 3:799-803. survival of stationary phase Aerobacter aerogenes 8. Harold, F. M. 1963. Accumulation of inorganic polyphosstored in aqueous suspension. J. Gen. Microbiol. phate in Aerobacter aerogenes. I. Relationship to 25:61-76. growth and nucleic acid synthesis. J. Bacteriol. 31. Strange, R. E., and J. R. Hunter. 1966. 'Substrate86:216-221. accelerated death' of nitrogen-limited bacteria. J. Gen. 9. Harold, F. M. 1964. Enzymatic and genetic control of Microbiol. 44:255-262. polyphosphate accumulation in Aerobacter aerogenes. 32. Tannock, G. W., and J. M. B. Smith. 1971. Studies on the J. Gen. Microbiol. 35:81-90. survival of Salmonella typhimurium and Salmonella 10. Harold F. M., and R. L. Harold. 1965. Degradation of bovismorbificans on pasture and in water. Aust. Vet. J. inorganic polyphosphate in mutants of Aerobacter 47:557-559. aerogenes. J. Bacteriol. 89:1262-1270. 33. Tichacek, B. 1972. Survival of nonsporulating microorga11. Hendricks, C. W. 1971. Increased recovery rate of salnisms in laboratory and natural conditions. J. Hyg. monellae from stream bottom sediments versus surface Epidemiol. Microbiol. Immunol. 16:179-185. serve.
The longevity of S. enteritidis cells starved in dilute suspension was not anticipated since most Enterobacteriaceae do not survive beyond a few days in a nutrient-free environment (2, 5, 17, 25, 29, 30). Ensign (6) compared the half-life survival times for starving vegetative cells of several bacteria. Calculating the half-lives from the results of other investigators, Ensign reported the following 50%'survival times: Streptococcus mitis, 22 h; S. lactis, 30 h; Escherichia coli, 36 h; A. aerogenes, 45 h; Azotobacter agilis, 50 h; and Pseudomonas aeruginosa, 84 h. These values are significantly lower than the half-life of S. enteritidis in phosphate buffer. The marked ability of this organism to grow on products released by dead cells, the striking patterns of cryptic growth observed in heavy suspensions, and the relatively long half-life when compared with coliform organisms may result in significantly longer survival times in natural habitats, such as soil or water, than coliforms or other indicators of fecal pollution.
Vol. 125, No. 1
Printed in U.S.A.
Starvation Survival of Salmonella enteritidis ROBERT E. DRUILHETI* AND JOSEPH M. SOBEK Department of Microbiology, University of Southwestern Louisiana, Lafayette, Louisiana 70501 Received for publication 18 August 1975
Washed cells of Salmonella enteritidis harvested from a defined medium during logarithmic growth were subjected to starvation in pH 7 phosphate buffer at 37 C. Viability was measured by slide cultures and plate counts. The survival of cell suspensions equivalent to 1 to 10 mg (dry wt)/ml was influenced by cryptic growth. The rate of cryptic growth, assessed by plate counts, increased with cell density and could not be alleviated by starvation with dialysis. Dialysis of the starving culture did retard the onset of cryptic growth but did not eliminate it, indicating that the major substrates for regrowth were relatively large cellular components. In phosphate buffer, 6.7 homologous heat-killed cells allowed for the doubling of one S. enteritidis cell. Cryptic growth was not observed when cells were starved on the surface of membrane filters or in suspensions equivalent to 20 jig (dry wt)/ml (10' cells/ml). Similar half-life survival times were calculated for both these populations, but the shape of their survival curves differed significantly. These differences were attributed to stress factors encountered during cell preparation and during starvation. The half-life survival time of S. enteritidis starved at 20 jig (dry wt)/ml was 140 h in phosphate buffer, 82 h in 3,6-endomethylene-1,2,3,6-tetrahydrophthalic acid buffer, and 77 h in tris(hydroxymethyl)aminomethane buffer. The starvation survival of a variety of sapro- 0.5; (NH4)2SO4, 2.0; NaCl, 0.5; glucose, 5.0; final pH, phytic bacteria has been discussed in compre- 7.0. Liquid cultures in 300-ml nephelo culture flasks ml of hensive articles (3, 4, 16, 22, 23, 25), but no (Bellco Glass Inc., Vineland, N.J.) containing a50model medium were incubated at 37 C on starvation survival studies of Salmonella defined rotary shaker (New Brunswick Scientific Co., Inc., enteritidis have been reported. Tannock and VS N.J.). Growth was assessed by moniNew Smith (32) related bovine salmonellosis to the toringBrunswick, absorbancy at 540 nm. survival of S. typhimurium and S. bovismorStarvation procedure. Survival studies were bificans on pasture and in water. Tichacek (33) begun with middle-logarithmic-phase cells collected studied the survival of a dried suspension of S. at an optical density at 540 nm (OD.4o) of 0.38 to 0.45 enteritidis in relation to environmental in- by centrifugation at 5,000 x g for 5 min at room fluences. Other reports relating Salmonella temperature (23 C). The cell pellet was washed three survival to the natural environment have been times in buffer (37 C) and suspended in buffer to the cell density. The period between collection concerned with recovery from unpolluted water desired initiation of starvation experiments averaged 35 (7), growth in dialysis sacs placed in river water and (12, 13), and recovery from stream sediments min. Three starvation buffer systems were used: potas(11). sium phosphate, tris(hydroxymethyl)aminomethane The present study was undertaken to deter- (Tris), and 3,6-endomethylene-1,2,3,6-tetrahydrophmine the survival characteristics of S. thalic acid (EMTA). Each buffer was prepared at enteritidis in a non-nutrient environment. The 0.067 M, pH 7.0, and sterilized by membrane filtraeffects of cell density, buffers, and the method tion. Viability studies with 20-ml cell suspensions at of starvation were considered. densities equivalent to 1.0 to 10 mg (dry wt)/ml were
made in glass tubes (22 by 200 mm), and the effect of buffers on the survival of dilute cell suspensions (20 500-ml round-bottom ;sg/ml) was determined inwere aerated with sterile flasks. The cell suspensions water-saturated air to compensate for evaporation I Present address: Department of Internal Medicine, Uni- during starvation at 37 C. Cellulose dialysis memversity of Texas Medical School at Houston, Houston, Tex. branes (Union Carbide Corp., Chicago, Ill.) were used to study the effect of removal of extracellular material 77025. 119
MATERIALS AND METHODS Growth conditions. S. enteritidis (ATCC 13076) was grown in a defined medium composed of (grams per liter): K,HPO4, 9.2; KH2PO4, 5.2; MgSO4-7H2O,
120
DRUILHET AND SOBEK
on the survival pattern of S. enteritidis. The membranes were washed repeatedly in hot distilled water before use. A washed-cell suspension of known density was placed in the sterile dialysis chamber (21.5 by 260 mm) and aerated at a rate of 20 ml of air per min. The cells were dialyzed against 1,400 ml of aerated phosphate buffer replaced at a rate of 1 or 4 ml/min. Samples were removed at regular intervals for viability determinations. In one series of experiments the dialysis apparatus was modified to correct for any volume changes due to sampling. The chamber used for this purpose was a graduated glass cylinder (35 by 250 mm) closed at one end with cellulose membrane. The cylinder, supplied with internal aeration and sample tubes, was suspended in a 2-liter flask by a rubber cap cut to receive the 35-mm tube. After sampling, the menisci of the cell suspension and the dialysis buffer were made to coincide by adjusting the cylinder depth in the reservoir. Starvation on membrane filters. Cells were grown at 37 C in the defined medium under shake conditions, and 10 ml (OD,40 = 0.4) was removed, sedimented at 5,000 x g, washed twice, and suspended in 10 ml of phosphate buffer. A 10- 8 dilution of cells was prepared in phosphate buffer (37 C) and 10 ml was filtered through each of 24 filter membranes (Millipore type HA, 47 mm, 0.45 lAm). The filter was placed on a washed, sterile, membrane adsorbent pad (Millipore Corp.) saturated with phosphate buffer. Each petri dish reservoir contained three adsorbent pads, a filter membrane with 120 to 150 dispersed cells, and 10 ml of buffer. The petri dish lid was lined with filter paper to absorb water condensate formed during starvation. The entire procedure required approximately 45 min. Viability determinations. The one-step slide culture method of Postgate et al. (24) was used to assess cell viability. This method was reported to determine the ratio of viable to total cells with Poissonian precision and to be accurate within the range of 5 to 100% viability (23). The method was modified by the use of an annular glass ring (18 by 8 mm) supported on a cover glass (22 by 40 mm) and the use of 0.2 ml of Trypticase soy agar (TSA). The TSA (Baltimore Biological Laboratories) was filtered hot before sterilization. Slide culture inocula used with concentrated cell suspensions were made by adding 0.1 ml of a 10-2 dilution to the agar surface; 0.02 ml of undiluted solution was used when cells were starved in dilute suspension. An inoculum size of 5 x 108 cells provided 20 to 40 cells per high-power microscopic field. Slide cultures were incubated at 37 C for 6 to 8 h in a moisture chamber and observed with a phase contrast microscope. A microcolony resulting from one or more cell division was scored as one viable unit; a single cell, often coccoid in appearance, was scored as one dead unit. Sufficient microscopic fields were counted to reach a total of 300 objects; coincidences were disregarded, and the viability was determined as the percentage of viable units in the total number of objects counted. Slide cultures were prepared in duplicate, and viabilities were calculated from the average number of viable and dead cells. Variabilities
J. BACTERIOL.
of 1% or less were routinely observed with replicate preparations. In some experiments plate counts on TSA were used to estimate viability. Plate count viabilities were determined in triplicate and reported as change in colony-forming units with time of starvation. The survival of cells starved on membrane filters was determined as follows. At each sample time, three filters were selected at random, transferred to TSA (MacConkey agar in one experiment), and incubated at 37 C for 24 h. The filters were removed from the agar surface, placed in ethanol (95%, vol/vol) for 1 min, and transferred to methylene blue (0.6 g/100 of ml ethanol) for 30 s and then to a 1:750 aqueous solution of Zephiran chloride (Sterling Drugs, Inc., New York, N.Y.). Colonies were removed from the filter by shaking in the Zephiran solution, and the filter was placed in 5% (wt/vol) phenol for 2 min. The colonies developed from viable cells were seen as clear areas in a pale blue background. Mean colony counts of three filters were used to calculate viability at each sample interval. Cryptic growth. A modification of the method described by Nioh and Furusaka (20, 21) was designed to determine the minimal concentration of dead cells that would support growth of S. enteritidis. Cells were grown in defined medium to an OD.40 of 0.60, collected by centrifugation, and washed with physiological saline. The suspension was diluted in saline to 2.35 x 109 cells/ml (determined by plate count), and four aliquots were sedimented at 10,000 x g. The cells were suspended in the defined medium minus the carbon source or minus the carbon and nitrogen source, or 0.067 M phosphate buffer. Each suspension was heated for 15 min in a boiling water bath and diluted 1:10 and 1:100 with the appropriate solution. The heat-killed suspension (20 ml) was transferred to 100-ml flasks, inoculated with 106 cells, and incubated at 37 C. Viable counts on TSA were made at 0, 8, and 24 h.
RESULTS Starvation of concentrated cell suspensions and cryptic growth. Figure 1 is representative of the survival pattern of dense cell suspensions of S. enteritidis starved at 37 C in 0.067 M phosphate buffer. The viability of a 10-mg/ml cell suspension decreased 3% in the first 6 h of starvation. Later samples indicated an increase in the percentage of viability. The number of viable units in the population starved for 36 h was 40% greater than the number originally added to the starvation system. Plate counts were used to follow the viability of these cells. Similar results were obtained with cell densities from 1 to 10 mg (dry wt)/ml. With cell suspensions of 1.0 mg/ml, regrowth was observed after 18 h and the population reattained the initial viability (100%) after 40 h.
STARVATION SURVIVAL OF S. ENTERITIDIS
VOL. 125, 1976
121
growth. The less dense cell suspension, 2.0 mg/ml, also showed this characteristic increase in viability after 48 h. A fourfold increase in the buffer exchange rate (4.2 ml/min) failed to /4< significantly alter the survival pattern (Fig. / -~ 2B). Complete renewal of dialyzing buffer was achieved every 30 min. The faster exchange rate retarded the onset of cryptic growth but did not eliminate it. To minimize the influence of volume changes on the viability of dialysis suspensions, the experiment was repeated in a different dialysis system designed to correct for volume changes due to sampling. The results of this study are shown in Fig. 2C. The viability pattern in this 2 ,4 ,* ,system was not significantly different from the 3 io survival curves obtained in other dialysis sys9 24 12
,-_~ 4~ ~
5, 100 a
c0
0:
0
Time
(hours)
F IG. 1. Survival curve of S. enteritidis in 0.067 M phowsphate buffer (pH 7) showing cryptic growth. Cell demsities equivalent to 10 mg (dry wt)/ml (0) and 1 mg (dry wt)/ml (0) were starved under forced aeration at 37 C. Points were determined by plate counts. Dasi!hes represent plate count values greater than thos ;eobtained at 0 h (100% viability).
(27, 30) was [dhe cryptlc growth phenomenon of S. enteritidis diedbyfollowingcthengrowth in varying concentrations of its homologous stu
heeEt-killed cell suspension (Table 1). When hefat-killed cells were substituted for glucose in tne caetinect meciium, z.;3 x iuw ceiis increasea the 24-h plate count by 2,200-fold and 2.35 x 108 dead cells supplied sufficient carbon to increase the colony-forming units by 178-fold. Cryptic growth was less when dead cells supplied both carbon and nitrogen in the otherwise defined medium. Cryptic growth diminished when the dead-cell concentration was decreased. When the concentration of heat-killed cells was 2.35 x 107 cells/ml of phosphate buffer, one new cell was produced at the expense of approximately 6.7 heat-killed cells. Starvation in dialysis. The survival of S. enteritidis was studied in several dialysis systems in an attempt to reduce cryptic growth in dense cell suspensions. Dialysis was used to remove small-molecular-weight components released by dead cells. Figure 2A represents the viability of two 100-ml cell suspensions (2 and 8 mg [dry wtyml) dialyzed against 1,400 ml of phosphate buffer at an exchange rate of 1.0 ml/min. A complete buffer turnover was accomplished every 24 h. The percentage of viability of the 8.0-mg/ml cell suspension decreased by 25% for the first 24 h of starvation and then increased in a manner characteristic of cryptic
tems.
Starvation in dilute suspensions and the effects of buffers on survival. The effect of
phosphate, Tris, and EMTA buffers on the survival pattern of cell suspensions equivalent to 20 ,gg (dry wt)/ml is shown in Fig. 3. A 50% loss of viability was observed at 4 days in EMTA (Fig. 3A) compared with only 30% loss of viability in phosphate buffer (Fig. 3C). Since possible contaminants in EMTA could have caused a decrease in viability through substrate-accelerated death (26, 31) or growth inhibition, the experiment was repeated with buffer prepared from recrystallized material. An apTABLE 1. Growth of S. enteritidis in varying concentrations of homologous heat-killed cells Heat-
Medium
Definedd minus
carbon
killed cells' (x1O-)
235 23.5 2.35
Defined minus 235 23.5 carbon and
CFU°/ml (x 10-') 8h
0.50 750 0.46 35 0.49 53
nitrogen
2.35
0.54 320 0.53 21 0.48 16
0.067 M
235 23.5 2.35
0.57 350 0.50 28 0.51 27
phosphate buffer, pH 7.0
in
__
0h
Overall increase
24 h
CFUc
(x10-')
1,100 22.0 82 30
1.78 0.61
820 15.2 42 0.80 20 0.41 360 54 35
6.32 1.08 0.70
a Concentration determined by plate counts on TSA before heating. Values represent total number of dead cells in the medium. b CFU, Colony-forming units. c Increase in CFU is equal to the 24-h growth yield divided by the initial CFU. d The defined medium is described in the text.
122
J. BACTERIOL.
DRUILHET AND SOBEK
creased more slowly to 55%, and thereafter the death rate was increased and approached that observed in the initial 2 days of starvation. At the end of 9 days, 11% of the population remained viable. Starvation on membrane filters. This method of starvation was used to completely rule out the effects of cryptic growth on survival. The viability of dispersed cells starved on membrane filters saturated with phosphate buffer is shown in Fig. 4. The survival pattern of these cells is characterized by an initial rapid loss of viability during the first 2 days, a 3- to 4-day period of stasis, and a terminal slow decrease in viability beginning at day 6. The medium used to assess viability, TSA (Fig. 4A) or MacConkey agar (Fig. 4B), had little effect on the recovery of cells after the 2nd day of starvation. The possibility that increased longevity of cells starved on membrane filters resulted from nutrients remaining in the adsorbent pads was examined. Cells were starved on filters supported on the surface of agar prepared with 0.67 M phosphate buffer and 1.5% (wt/vol) washed agar. The viability pattern of cells starved in this manner did not differ from the previous results.
100-
7550A
100a
750.1C)0C
50-
B
100 -
7550-
Ic 0
1
I
I
12
24
36
48
60
60
100-
96
Timo (hours)
FIG. 2. Cryptic growth of S. enteritidis starved in dialysis at densities equivalent to 8 mg (dry wt)/ml (0) and 2 mg/ml (0). Cells (100 ml) were enclosed in cellulose membrane and continuously dialyzed against 1,400 ml of aerated phosphate buffer (0.067 M, pH 7) at 37 C. Buffer exchange rates of I ml/min (A) and 4 ml/min (B and C) were used and volume corrections due to sample removal were made (C). Points were determined by slide culture (B and C) and by both slide culture and plate count (A).
=
0-
>
60-
1.
*
20.
U0 a
A
0
2
3
2
1
Ti m-
3
4
(days)
100-
parent pKa' of 6.30 was calculated from the titration curve of recrystallized EMTA. The viability of cells starved in recrystallized EMTA was approximately 10% higher than in buffer prepared from commerical material. However, the cells survived better in phosphate buffer than in EMTA. After 4 days of starvation, 67% of the population in phosphate buffer was viable whereas the viability in recrystallized EMTA was only 58%. The survival of S. enteritidis in Tris buffer (Fig. 3B) was no better than in EMTA. The percentage of viability of S. enteritidis in phosphate buffer (Fig. 3C) decreased to 80% in 2 days, between day 2 and day 5 viability de-
20so
2
4
Time
6
0
10
(days)
FIG. 3. Survival pattern of S. enteritidis in commercial (0) and recrystallized (0) EMTA (A), Tris (B), and phosphate buffers (C). Washed cell suspensions equivalent to 20 gug (dry wt)/ml were starved in 0.067 M buffer, pH 7, at 37 C. Viabilities were determined by slide culture.
STARVATION SURVIVAL OF S. ENTERITIDIS
VOL. 125, 1976 0.-
7A
a
A
°
2
4
os * Days
2 of
6 *
Starvtion
FIG. 4. Survival characteristics of S. epteritidis starved on the surface of membrane filters and subsequently transferred to TSA (A) or MacConkey agar (B). Points were determined as described in text.
DISCUSSION Penicillin and chloramphenicol will prevent cryptic growth (25), but these inhibitors may alter intracellular changes taking place during starvation. Although dense cell suspensions would make measurements of starvationinduced intracellular changes more feasible, the cryptic growth response of S. enteritidis requires that endogenous metabolism studies be performed with dilute suspensions. Even the estimate of 6.7 heat-killed cells accounting for the doubling of a single S. enteritidis cell may be too high since the nutrient supplied by unstarved heat-killed cells may differ greatly from those liberated by the death of starved cells. One would expect that starvation before death would increase the number of cells required per cell division because, since readily utilizable endogenous components would be the first to go during starvation, they would not be available as nutrients for cryptic growth. The ability of this organism to grow on the products of dead cells was greater than that reported for Arthrobacter (20, 21), a soil organism of marked ability to survive starvation. It would be interesting to see whether this ability also extends to heterologous heat-killed cells. Postgate and Hunter (25) reported that death of 50 members of a starved suspension of Aerobacter aerogenes allowed the doubling of one survivor. Ambiguity that might arise from cryptic growth during starvation of A. aerogenes was eliminated by using a 20-,g/ml cell suspension (25). With S. enteritidis, this cell concentration was equivalent to 106 cell/ml, far below the levels supporting cryptic growth on homologous heat-killed cells. However, even at these low cell concentrations, materials released by dead cells could affect survival. Dialysis of heavy cell suspensions during starvation did not eliminate the cryptic growth response but confirmed the essentially nondi-
123
alyzable nature of cryptic growth substrates. The utilization of dialyzable cell products probably accounted for the rapid onset of cryptic growth in survival studies with concentrated cell suspensions since dialysis extended the death phase before the onset of cryptic growth. Dispersed cells starved on membrane filters showed differences in the survival pattern when compared with cells starved in dilute suspensions. These differences may be the result of stress of the population. The stress of filtration could have caused the rapid initial loss of 40 to 50% of the membrane-starved population. This rapid rate of death ceased after the 2nd day and no significant change in viability was observed until the 5th or 6th day of starvation. Cells that survived during this 2- to 6-day period were either not affected or had recovered from the filtration stress. Cells starved in dilute suspensions were not subjected to this stress and showed a much slower loss in viability during the first 2 days of starvation. The 50% survival time of S. enteritidis starved in dilute suspension with phosphate buffer is 140 h, and is 132 h when starved on membrane filters. Between the 2nd and 6th days of starvation, the death rate of cells starved on filters was only 2% per day, whereas cells starved in suspension lost viability at an average rate of 7.8% per day. Cells in suspension were subjected to the stress of forced aeration and agitation, whereas cells starved on filters were not forcibly aerated. In EMTA and Tris buffers the half-life survival time was 82 and 77 h, respectively. EMTA was reported as a non-metabolizable buffer suitable for bacterial studies at pH 7.0 in media free from nitrogen, phosphate, and sulfur (18). Tris-saline buffer was used by Postgate and Hunter (25) in survival studies of A. aerogenes. Decreased viability in the absence of phosphate was possibly due to trace contaminants in EMTA, since survival was enhanced with recrystallized EMTA, or to the poor buffering capacity of Tris at pH 7. The presence of phosphate in the starvation environment does not seem to be critical to the starvation survival of S. enteritidis. If phosphate were an important factor in maintaining cell longevity, a much more drastic change in the viability pattern would probably have been observed in EMTA and Tris buffers. This was not the case. Several workers have suggested that polyphosphates may serve as potential maintenance energy sources (1, 14, 15, 19). This view was dispelled, at least for A. aerogenes, by Harold and Harold (10). Although the polyphosphate content of S. enteritidis was not measured, according to cur-
124
J. BACTERIOL.
DRUILHET AND SOBEK
rent theories of polyphosphate accumulation (8, 9, 28) the growth environment was not conductive to the accumulation of this potential re-
water. Appl.
Microbiol. 21:379-380.
12. Hendricks, C. W. 1972. Enteric bacterial growth rates in
river water. Appl. Microbiol. 24:168-174. 13. Hendricks, C. W., and S. M. Morrison. 1967. Multiplica-
tion and growth of selected enteric bacteria in clear mountain stream water. Water Res. 1:567-576. 14. Hou, C. I., A. F. Gronlund, and J. J. R. Campbell. 1966. Influence of phosphate starvation on cultures of Pseudomonas aeruginosa. J. Bacteriol. 92:851-855. 15. Kornberg, S. R. 1957. Adenosine triphosphate synthesis from metaphosphate by an enzyme from Escherichia coli. Biochim. Biophys. Acta 26:294-300. 16. Lamanna, C. 1963. Endogenous metabolism with special referance to bacteria. Ann. N.Y. Acad. Sci. 102:515-793. 17. McFeters, G. A., and D. G. Stuart. 1972. Survival of coliform bacteria in natural water: field and laboratory studies with membrane-filter chambers. Appl. Microbiol. 24:805-819. 18. Mallette, M. F. 1967. A pH 7 buffer devoid of nitrogen, sulfur, and phosphorus for use in bacteriological systems. J. Bacteriol. 94:283-290. 19. Mallette, M. F., C. I. Cowan, and J. J. R. Campbell. 1964. Growth and survival of Escherichia coli in medium limited in phosphate J. Bacteriol. 87:779-785. 20. Nioh, I., and C. Furusaka. 1968. Growth of bacteria in the heat-killed cell suspensions of the same bacteria. J. Gen. Appl. Microbiol. 14:373-385. 21. Nioh, I., and C. Furusaka. 1970. Factors affecting growth yield of Arthrobacter simplex in its heat-killed cell suspension. J. Gen. Appl. Microbiol. 16:115-126. 22. Postgate, J. R.1967. Viability measurements and the survival of microbes under minimal stress. Microb. Physiol. 1:1-23. 23. Postgate, J. R. 1969. Viable counts and viability, p. LITERATURE CITED 611-628. In J. R. Norris and D. W. Ribbons (ed.), 1. Bohinski, R. C., and M. F. Mallette. 1967. Response of Methods in microbiology, vol. 1. Academic Press Inc., log-phase cells of Escherichia coli to medium limited in New York. both sulfate and phosphate. J. Bacteriol. 93:1316-1326. 24. Postgate, J. R., J. E. Crumpton, and J. R. Hunter. 1961. 2. Clifton, C. E. 1966. Aging of Escherichia coli. J. Bacteriol. Measurement of bacterial viability by slide culture. J. 92:905-912. Gen. Microbiol. 24:15-24. 3. Dawes, E. A., and D. W. Ribbons. 1962. The endogenous 25. Postgate, J. R., and J. R. Hunter. 1962. The survival of metabolism of microorganisms. Annu. Rev. Microbiol. starved bacteria. J. Gen. Microbiol. 29:233-263. 16:241-264. 26. Postgate, J. R., and J. R. Hunter. 1964. Acceleration of 4. Dawes, E. A., and D. W. Ribbons. 1964. Some aspects of death of Aerobacter aerogenes starved in the presence the endogenous metabolism of bacteria. Bacteriol. Rev. of growth-limiting substrates. J. Gen. Microbiol. 28:126-149. 34:459-473. 5. Dawes, E. A., and D. W. Ribbons. 1965. Studies on the 27. Ryan, F. J. 1955. Spontaneous mutation of non-dividing endogenous metabolism of Escherichia coli. Biochem. bacteria. Genetics 40:726-738. J. 95:332-343. 28. Smith, I. W., J. F.Wilkinson, and J. P. Duguid. 1954. 6. Ensign, J. C. 1970. Long-term starvation survival of rod Volutin production in Aerobacter aerogenes due to and spherical cells of Arthrobacter crystallopoietes. J. nutrient imbalance. J. Bacteriol. 68:450-463. Bacteriol. 103:569-577. 29. Strange, R. E. 1968. Bacterial "glycogen" and survival. 7. Fair, J. F., and S. M. Morrison. 1967. Recovery of Nature (London) 220:606-607. bacterial pathogens from high quality surface water. 30. Strange, R. E., F. A. Dark, and A. G. Ness. 1961. The Water Resour. Res. 3:799-803. survival of stationary phase Aerobacter aerogenes 8. Harold, F. M. 1963. Accumulation of inorganic polyphosstored in aqueous suspension. J. Gen. Microbiol. phate in Aerobacter aerogenes. I. Relationship to 25:61-76. growth and nucleic acid synthesis. J. Bacteriol. 31. Strange, R. E., and J. R. Hunter. 1966. 'Substrate86:216-221. accelerated death' of nitrogen-limited bacteria. J. Gen. 9. Harold, F. M. 1964. Enzymatic and genetic control of Microbiol. 44:255-262. polyphosphate accumulation in Aerobacter aerogenes. 32. Tannock, G. W., and J. M. B. Smith. 1971. Studies on the J. Gen. Microbiol. 35:81-90. survival of Salmonella typhimurium and Salmonella 10. Harold F. M., and R. L. Harold. 1965. Degradation of bovismorbificans on pasture and in water. Aust. Vet. J. inorganic polyphosphate in mutants of Aerobacter 47:557-559. aerogenes. J. Bacteriol. 89:1262-1270. 33. Tichacek, B. 1972. Survival of nonsporulating microorga11. Hendricks, C. W. 1971. Increased recovery rate of salnisms in laboratory and natural conditions. J. Hyg. monellae from stream bottom sediments versus surface Epidemiol. Microbiol. Immunol. 16:179-185. serve.
The longevity of S. enteritidis cells starved in dilute suspension was not anticipated since most Enterobacteriaceae do not survive beyond a few days in a nutrient-free environment (2, 5, 17, 25, 29, 30). Ensign (6) compared the half-life survival times for starving vegetative cells of several bacteria. Calculating the half-lives from the results of other investigators, Ensign reported the following 50%'survival times: Streptococcus mitis, 22 h; S. lactis, 30 h; Escherichia coli, 36 h; A. aerogenes, 45 h; Azotobacter agilis, 50 h; and Pseudomonas aeruginosa, 84 h. These values are significantly lower than the half-life of S. enteritidis in phosphate buffer. The marked ability of this organism to grow on products released by dead cells, the striking patterns of cryptic growth observed in heavy suspensions, and the relatively long half-life when compared with coliform organisms may result in significantly longer survival times in natural habitats, such as soil or water, than coliforms or other indicators of fecal pollution.
