Historical and Contemporary NaCl Concentrations Affect the Duration and Distribution of Lag Times from Individual Spores of Nonproteolytic Clostridium botulinum

Abstract

In this study we determined the effect of NaCl concentration during sporulation (0 or 3.0% [wt/vol] added NaCl) and subsequent growth (0 or 2.0% [wt/vol] added NaCl) on the distributions of times associated with various stages of the lag phase of individual spores of nonproteolytic Clostridium botulinum strain Eklund 17B. The effects of NaCl on the probability of germination and the probability of subsequent growth were also determined. Spore populations exhibited considerable heterogeneity at all stages of lag phase for each condition tested. Germination time did not correlate strongly with the times for later stages in the lag phase, such as outgrowth and doubling time. Addition of NaCl to either the sporulation or growth media increased the mean times for, and variability of, all the measured stages of the lag phase (germination, emergence, time to one mature cell, and time to first doubling). There was a synergistic interaction between the inhibitory effects of NaCl in the sporulation medium and the inhibitory effects of NaCl in the subsequent growth medium on the total lag time and each of its stages. Addition of NaCl to either the sporulation medium or the growth medium reduced both the probability of germination and the probability of a germinated spore developing into a mature cell, but the interaction was not synergistic. Spores formed in medium with added NaCl were not better adapted to subsequent growth in suboptimal osmotic conditions than spores formed in medium with no added NaCl were. Knowledge of the distribution of lag times for individual spores and quantification of the biovariability within lag time distributions may provide insight into the underlying mechanisms and can be used to improve predictions of growth in food and to refine risk assessments.

The species Clostridium botulinum is comprised of four physiologically and phylogenetically distinct groups of anaerobic spore-forming bacteria that produce the potentially lethal botulinum neurotoxin (13). Although botulism is rare, the severity of the intoxication ensures that considerable effort is directed toward preventing growth of this pathogen in food. Nonproteolytic C. botulinum is one of the two groups most frequently associated with food-borne botulism. This group produces spores which can survive pasteurization heat treatments and are able to germinate, grow, and produce toxin at 3°C (9). These characteristics make nonproteolytic C. botulinum a particular concern in the production of mild-heat-treated refrigerated foods (17).

The ability to model and predict the development of a bacterial population underpins many areas of food safety, such as hazard analysis of critical control points and quantitative risk assessment. Understanding variability in times to growth is an important part of these processes. Although widespread in the environment, spores of nonproteolytic C. botulinum are usually present at low concentrations (13); thus, if growth occurs in food packs, it is likely to initiate from just a few spores. This has consequences when researchers try to predict growth as the uncertainty around estimates of growth times of nonproteolytic C. botulinum increases as the initial spore concentration decreases (29). Improved prediction of growth from low initial concentrations requires knowledge of the variation in individual lag times (11). It has also been shown that variability in lag times of individual spores is an important component in risk assessment for products contaminated at low levels (2). Such variability cannot be derived from observations made at the population level; instead, it must be determined from studies of individual spores (1).

When the initial number of cells is very small, the variability in the times to detection is close to the variability in the lag times of individual cells (15). Similarly, the variability in times to growth events, such as time to toxin production, is expected to closely follow the distribution of individual spores' lag times when spore concentrations are very low. While lag time distributions have been reported for vegetative cells of several species, data for distributions from spore inocula are less readily available. The lag time from spores includes the time for germination and outgrowth in addition to the time for the first replication cycle. As there are more stages in the lag phase of spores and each stage has an associated variability, in most cases distributions of lag time are wider when growth initiates from spores than when growth starts from the equivalent number of vegetative cells in the same conditions. There have been relatively few studies of variability in the lag time of individual spores, and they have tended to concentrate only on germination (4, 5, 6, 12, 27). More recently, we (25) examined germination, outgrowth, and doubling time from spores of nonproteolytic C. botulinum in good growth conditions. We found that the population was heterogeneous at each stage of the lag phase, and the time spent in each stage was independent of the duration of the other stages. While germination is an important part of lag phase variability and thus the accuracy of growth models, other stages of the lag phase are also highly variable and contribute to the total variability during the lag phase of spores.

The lag phase of bacterial growth from vegetative cells depends not only on the environmental conditions prevalent during growth but also on the initial physiological state of the organism, which is influenced by previous culture conditions or treatments (1). Usually a lag is considered to be related to the ability of an individual cell to adapt to a new environment. It is conceivable that the environment in which a spore is formed not only alters the ability of that spore to germinate but also influences its subsequent ability to grow and start to multiply in different conditions. For example, spores formed in environments containing NaCl, such as marine sediments, may be better adapted to subsequently grow in the presence of NaCl than spores formed in higher-water-activity environments, such as plant or animal tissue. Any alteration in the ability of spores to grow in the presence of reduced water activity is important as decreasing water activity, whether as a result of NaCl or another mechanism, is one of the most frequently used methods of food preservation (3). In this paper we describe the effect of historic exposure to NaCl during spore formation on the duration of the lag phase of C. botulinum spores in media with and without added NaCl.

The aims of this study were to determine the effect of NaCl addition on the variability of stages in the lag phase of individual spores of nonproteolytic C. botulinum strain Eklund 17B and to assess the extent that the conditions under which the spores were formed affected the length and variability of the subsequent lag phase. Greater understanding of the variability of the stages in the lag phase of individual spores and the relationships between them should allow more precise calculation of the risk of toxin production in food.

MATERIALS AND METHODS

Spore preparation.

Duplicate crops of nonproteolytic C. botulinum type B strain Eklund 17B spores were produced for each sporulation condition using a two-phase medium. Sporulation medium (SM) without added NaCl (SM+0%) consisted of a solid phase of 250 ml of Robertson's cooked-meat medium (Southern Group Laboratories, London, United Kingdom) with 4.5 g of agar and a liquid phase of 100 ml of deoxygenated distilled water. Unsupplemented cooked-meat medium contains 0.3% (wt/vol) NaCl. For sporulation medium with 3% NaCl (SM+3%) both the solid and liquid phases were supplemented with 3.0% (wt/vol) NaCl. Both sporulation media were inoculated with 500 μl of culture (grown overnight in PYGS broth [23] at 30°C) and incubated for 7 days at 30°C. The resulting spores were harvested, washed, and separated from cell debris by discontinuous density gradient centrifugation, as described previously (25) except that the wash solutions contained 0.1 and 3.0% (wt/vol) NaCl for spores formed in the presence of 0 and 3% added NaCl, respectively. Similarly, cleaned spore suspensions were stored at 1°C in 0.1 or 3.0% saline solution. The concentrations of spores in the prepared suspensions were determined on PYGS agar incubated for 48 h at 30°C under a headspace containing 10% CO₂ and 90% H₂. Cultural purity and the absence of high levels of proteolytic activity were checked on VL blood agar and reinforced clostridial medium containing 5% (wt/vol) skim milk incubated in the same conditions (19).

Microscopy and image analysis of growth from spores.

Slides inoculated with between 4,000 and 6,400 spores mm⁻² in a 2-mm² area were prepared and used as described previously (25). Briefly, spores were overlaid with 8 μl of molten anaerobic PYGS medium containing 0.5% (wt/vol) agar, sealed under a coverslip (18 by 18 mm), and then observed by phase-contrast microscopy using a ×40 objective. Samples were maintained at 22°C using a stage-mounted Peltier device, and a motorized XYZ stage controller was used to allow images of multiple set positions to be captured at regular times throughout the experiment.

Growth was observed in two different growth media (GM); PYGS medium with no added NaCl (GM+0%) and PYGS medium supplemented with 2% (wt/vol) NaCl (GM+2%). Unsupplemented PYGS medium contains 0.1% (wt/vol) NaCl. Images of spores were acquired every 5 min for 4 h and then at 10-min intervals for 20 h in GM+0% and at 5-min intervals for 4 h and then at 10-min intervals for 44 h in GM+2%. Three replicate experiments were performed with each of the two independent spore suspensions prepared in SM+0% or SM+3% (a total of six replicates for each growth medium NaCl concentration). The total numbers of spores observed were as follows: for cultures grown in SM+0% and then in GM+0% (SM+0%/GM+0%), 1,009 spores; for SM+3%/GM+0%, 2,522 spores; for SM+0%/GM+2%, 1,659 spores; and for SM+3%/GM+2%, 4,493 spores. The numbers of spores examined at each stage of the lag phase are shown in Table 1. Not all spores germinated or resulted in two cells. The smaller the probability of germination and growth, the greater the number of spores observed.

TABLE 1.

Averages and standard deviations for events occurring during germination and subsequent outgrowth from spores of nonproteolytic strain C. botulinum Eklund 17B as affected by added NaCl in the SM and the GM

Sporulation conditions	Lag phase event	Mean ± SD in the following growth conditions (h):
		GM+0%	GM+2%
SM+0%	tgerm	1.05 ± 1.59 (940)a	1.96 ± 2.31 (1,202)
	tem	3.74 ± 2.03 (864)	5.10 ± 2.95 (1,040)
	tC1	6.36 ± 2.27 (800)	9.27 ± 3.18 (935)
	tC2	7.47 ± 2.07 (724)	10.20 ± 2.82 (715)
	tdet	22.94 ± 1.73 (466)	28.65 ± 6.52 (421)
	tem − tgerm	2.71 ± 1.33 (864)	3.15 ± 2.05 (1,040)
	tC1 − tgerm	5.45 ± 1.86 (800)	7.45 ± 2.56 (935)
	tC1 − tem	2.87 ± 1.15 (800)	4.59 ± 1.69 (935)
	tC2 − tC1	1.36 ± 0.55 (724)	1.75 ± 0.69 (715)
SM+3%	tgerm	1.27 ± 1.91 (1,917)	4.81 ± 4.26 (2,060)
	tem	4.25 ± 2.05 (1,677)	7.43 ± 4.78 (1,499)
	tC1	6.72 ± 2.00 (1,490)	10.80 ± 5.23 (1,155)
	tC2	8.10 ± 2.10 (1,213)	11.23 ± 4.50 (800)
	tdet	23.23 ± 2.67 (296)	32.13 ± 12.87 (270)
	tem − tgerm	3.25 ± 1.54 (1,677)	3.77 ± 2.80 (1,499)
	tC1 − tgerm	5.87 ± 1.63 (1,490)	8.27 ± 4.02 (1,155)
	tC1 − tem	2.81 ± 1.03 (1,490)	4.96 ± 3.10 (1,155)
	tC2 − tC1	1.62 ± 0.80 (1,213)	2.10 ± 1.62 (800)

^aThe numbers in parentheses are the numbers of spores examined for each treatment.

Analysis of results.

Individual images were compiled into a sequence of frames for each field of view so that the same spore could be followed throughout dormancy, germination, emergence, elongation, and cell multiplication. Maximum pixel intensity and object length measurements were determined for each spore/cell in each frame of each sequence of images and used to calculate the times to germination (t_germ) and emergence (t_em), the time to a length equivalent to one mature cell (t_C1), and the time to a length equivalent to two mature cells (t_C2), as described previously (25). The length of a mature cell was defined as 5.5 μm, which was calculated by determining the average length of newly divided mid-exponential-phase cells at 22°C in PYGS broth with no added NaCl.

Measurement of time to turbidity.

The time to turbidity in GM+0% or GM+2% was determined using a Bioscreen C automated turbidity reader (Lab Systems, Finland) installed in an anaerobic cabinet with an atmosphere containing 5% CO₂, 10% H₂, and 85% N₂. A 50-μl aliquot of spore suspension containing 20 spores ml⁻¹ diluted in the appropriate growth medium and 300 μl of GM+0% or GM+2% were dispensed into each of the 200 wells of the Bioscreen plates. The plates were incubated at 22°C, and the optical density at 600 nm (OD₆₀₀) was measured every 10 min for 5 days. The detection time (t_det) was defined as the time that it took the measured OD₆₀₀ to reach 0.14 U. Uninoculated medium had an OD₆₀₀ of 0.10 U.

The number of cells present at t_det and the specific growth rates of nonproteolytic C. botulinum 17B in GM+0% and GM+2% at 22°C were determined from plate count measurements. Counts were determined using appropriate dilutions of cells plated on PYGS agar and incubated under a headspace containing 90% H₂ and 10% CO₂ for 24 h at 30°C.

Statistical analysis.

Data from replicate experiments were pooled for each spore crop and then analyzed. Mean times were compared using the two-way analysis of variance of SAS. An in-house program (Varifit) written in Visual Basic was used to compare the shapes and variabilities of the distributions. A chi-square test was carried out to determine whether the distributions were homogeneous by comparing binned data between distributions recentered on the same mean. A Bartlett test was used to compare the variances of the distributions. Differences were considered to be significant when the P value was less than 0.05. Pearson coefficients were calculated to determine the correlation between growth event times and intervals during germination and outgrowth. Conditions that had a Pearson coefficient of >0.6 were considered to be strongly correlated.

Logistic regression was used to model the effects of historic and contemporary exposure to NaCl on the probability of spore germination and the probability of growth from germinated spores of C. botulinum Eklund 17B. The initial model was: log [P/(1 − P)] = intercept + NaCl_SM + NaCl_GM + D(P) + NaCl_SM × D(P) + NaCl_GM × D(P), where P is either the probability of germination or the probability of growth to one mature cell length from a germinated spore, NaCl_SM is the added NaCl concentration in the SM, NaCl_GM is the added NaCl concentration in the GM, and D(P) is a dummy variable with only two possible values depending on P (0 for the probability of germination and 1 for the probability of growth to one mature cell length from a germinated spore). The D(P) variable allowed us to study the relationship between the probability of a spore germinating and the probability of subsequent growth. Terms for the interaction between sporulation medium NaCl concentration and growth medium NaCl concentration were not included in the model as preliminary plots of the logit function against the growth medium NaCl concentration for spore crops produced in SM+0% and SM+3% showed a lack of interaction between these parameters. A stepwise selection was used to select the significant terms of the model.

RESULTS

If lag time is defined as the time between a dormant spore being exposed to conditions suitable for growth and the formation of a mature vegetative cell, then the distribution of t_C1 can be considered to represent the distribution of lag times for individual spores (Fig. 1). The time to detection (t_det) (Fig. 2) includes both the lag time and the time required for cells to multiply from the lag time to the detection level. The shape of the t_det distribution should mimic that of the lag time distribution (displaced by the time required to multiply from t_C1 to t_det); as long as the growth in each well originates from a single spore, the growth rates are identical for all measured wells (i.e., the growth conditions are identical in all wells), and all cells that leave the lag phase subsequently lead to turbidity. It should be noted that the time to multiply from t_C1 to t_det is a function of the growth conditions; therefore, only the shapes, and not the durations, of the t_det distribution curves for different growth media can be compared.

FIG. 1.

Frequency distributions of t_C1 for individual spores of C. botulinum Eklund 17B formed in SM with NaCl (SM+0% or SM+3%) and grown at 22°C in PYGS medium (GM+0% or GM+2%). The following treatments were used: SM+0%/GM+0% (solid bars), SM+3%/GM+0% (shaded bars), SM+0%/GM+2% (bars with dark stripes), and SM+3%/GM+2% (bars with light stripes).

FIG. 2.

Frequency distributions of t_det for individual spores of C. botulinum Eklund 17B formed in SM with NaCl (SM+0% or SM+3%) and grown at 22°C in PYGS medium (GM+0% or GM+2%). The following treatments were used: SM+0%/GM+0% (solid bars), SM+3%/GM+0% (shaded bars), SM+0%/GM+2% (bars with dark stripes), and SM+3%/GM+2% (bars with light stripes).

The turbidity experiments were based on observation of 1,999 inoculated wells, and growth was detected in 1,453 of the wells. The fact that growth occurred in only 72% of the wells suggests that growth started from a single spore in a large number of the wells. The actual distribution of spores in the wells could be calculated by assuming that it followed a Poisson distribution, and this calculation suggested that the average number was 1.3 cells well⁻¹. The specific growth rates of C. botulinum Eklund 17B in GM+0% and GM+2% at 22°C, measured using a standard agar plate count technique, were 0.63 ± 0.07 and 0.61 ± 0.07 h⁻¹, respectively, equivalent to population doubling times of 1.10 and 1.13 h, respectively. The detection limit (OD₆₀₀, 0.14) was equivalent to 3 × 10⁶ CFU ml⁻¹ in both GM+0% and GM+2%. If growth following a lag occurs at a constant rate, the detection time distribution can be determined from the t_C1 distribution shifted in time. The measured mean detection times were between 23 h (no salt in the GM) and 29 to 32 h (2% salt) (Table 1), compared to ca. 29 to 30 and 34 to 35 h, respectively, estimated from the mean time to one cell plus the time for 21 doublings calculated using the population doubling times. Although there is a discrepancy of some hours, it is small considering the error accumulation and the fact that the first division time is expected to be longer than subsequent division times (18).

An advantage of using microscopy to monitor the lag phase of spores is that it is possible to determine which stages are most affected when conditions are altered. The numbers of observations, mean times, and standard deviations for times to germination (t_germ), emergence (t_em), length equivalent to one mature cell (t_C1), length equivalent to two mature cells (t_C2), and detectable turbidity (t_det) for each of the test conditions are shown in Table 1. The effect of exposure to NaCl during sporulation or growth on the distributions of times for germination, for outgrowth (t_C1 − t_germ), and for a mature cell to double in length (t_C2 − t_C1) are shown in Fig. 3, 4, and 5, respectively. In all conditions tested, the mean time for outgrowth was much longer than the mean time for germination or cell doubling. On average, outgrowth (t_C1 − t_germ) accounted for 71, 71, 69, and 65% of the time to doubling (t_C2) for spores grown in SM+0%/GM+0%, SM+3%/GM+0%, SM+0%/GM+2%, and SM+3%/GM+2%, respectively.

FIG. 3.

Frequency distributions of t_germ for spores of C. botulinum Eklund 17B formed in SM with NaCl (SM+0% or SM+3%) and grown at 22°C in PYGS medium (GM+0% or GM+2%). The following treatments were used: SM+0%/GM+0% (solid bars), SM+3%/GM+0% (shaded bars), SM+0%/GM+2% (bars with dark stripes), and SM+3%/GM+2% (bars with light stripes).

FIG. 4.

Frequency distributions of times in outgrowth (t_C1 − t_germ) for spores of C. botulinum Eklund 17B formed in SM with NaCl (SM+0% or SM+3%) and grown at 22°C in PYGS medium (GM+0% or GM+2%). The following treatments were used: SM+0%/GM+0% (solid bars), SM+3%/GM+0% (shaded bars), SM+0%/GM+2% (bars with dark stripes), and SM+3%/GM+2% (bars with light stripes).

FIG. 5.

Frequency distributions of times for first doubling (t_C2 − t_C1) for spores of C. botulinum Eklund 17B formed in SM with NaCl (SM+0% or SM+3%) and grown at 22°C in PYGS medium (GM+0% or GM+2%). The following treatments were used: SM+0%/GM+0% (solid bars), SM+3%/GM+0% (shaded bars), SM+0%/GM+2% (bars with dark stripes), and SM+3%/GM+2% (bars with light stripes).

Effect of growth NaCl concentration on the lag phase.

Adding 2% NaCl to the growth medium significantly increased the mean lag time as measured by t_C1, increased lag time variability, and altered the shape of the t_C1 distribution curves (Fig. 1). The addition of NaCl to the GM affected all stages of the lag phase, increasing both the mean and the variance of the times for germination, outgrowth, and doubling for spores produced in either SM+0% or SM+3%. The shape and variability of the distribution curves obtained with GM+2% were significantly different from the shape and variability of the equivalent curves obtained with GM+0%.

Effect of sporulation NaCl concentration on the lag phase.

The effect of adding 3% NaCl to the sporulation medium was more subtle than the effect of adding 2% NaCl to the growth medium. For example, adding 3% NaCl to the sporulation medium increased the mean lag time (t_C1) in GM+0% by only 6%, while adding 2% NaCl to the growth medium increased the mean lag time (t_C1) of spores produced in SM+0% by 46% (Table 1). Although small, the observed increase in t_C1 was significant. There was no significant difference in the mean t_det between spores produced in SM+0% and spores produced in SM+3% when incubated in GM+0%. The means and variabilities of times for germination, outgrowth, and doubling were all significantly greater for spores formed in SM+3% than for spores formed in SM+0%, and statistical testing did not reveal homogeneity between the distributions of the two treatments for any of the parameters measured. Although the shapes of the t_germ distribution curves were not homogeneous statistically, they were similar visually (Fig. 3). Many spores formed in SM+3% could germinate as rapidly as spores formed in SM+0%, but a subpopulation had extended germination times. An increase in the tail of the t_germ distribution curve was particularly noticeable for germination in GM+2%.

Combined effect of sporulation and growth NaCl concentrations on the lag phase.

Spores formed in medium containing added NaCl were not better able to germinate or subsequently grow on medium containing added NaCl than spores formed in the absence of added NaCl. The mean times for all stages of germination and growth were longer for spores formed in SM+3% and grown on GM+2% than for spores subjected to any of the other treatments (Table 1). The combined effect of NaCl added to the sporulation medium and NaCl added to the growth medium was proportionally greatest on the t_germ. The mean time for germination of spores formed in SM+3% on GM+2% was 3.76 h longer than that of spores formed in SM+0% on GM+0%, an increase of 358%. The equivalent increases in times for outgrowth and first doubling were 2.82 and 0.74 h (increases of 52 and 55%, respectively).

Statistical analysis of the mean times for t_C1 showed there was a significant interaction between addition of NaCl to the sporulation medium and addition of NaCl to the growth medium. This interaction appeared to be synergistic. Adding 3% NaCl to the sporulation medium increased the mean t_C1 by 0.36 h, and adding 2% NaCl to the growth medium increased t_C1 by 2.91 h, but the mean time for spores produced in SM+3% growing in GM+2% was 4.44 h longer than the mean time for spores formed in SM+0% growing in GM+0%. A similar synergistic effect was observed for the mean times of all the measured stages in the lag phase and the times to specific events in the lag phase.

Correlation between stages of growth.

The times for germination, outgrowth, and doubling for individual spores/cells did not correlate strongly (P > 0.6) with each other for any of the spore types or growth conditions tested, except that outgrowth correlated with doubling time for spores formed in SM+3% and grown in GM+2%. The correlations between all time events are shown in Table 2. The first spore to germinate was not the first to reach the length of one mature cell for any of the combinations of sporulation and growth NaCl concentrations tested. For example, the fastest times to germination were 5, 5, 8, and 8 min for SM+0%/GM+0%, SM+3%/GM+0%, SM+0%/GM+2%, and SM+3%/GM+2%, respectively. The spores that resulted in the shortest times to the length of two cells (4.30, 5.11, 6.35, and 6.21 h) took 13, 25, 49, and 65 min to germinate, respectively. In GM+0%, 59% of the spores formed in SM+0% that reached t_C2 did so quicker than the first spore to germinate did.

TABLE 2.

Correlation matrix (Pearson coefficients) between the times for events during germination and outgrowth of spores of C. botulinum Eklund 17B^a

Event	NaCl concn (%)		Pearson product moment coefficientsb
	SM	GM	tem	tC1	tC1	tem − tgerm	tC1 − tgerm	tC2 − tgerm	tC1 − tem	tC2 − tem	tC2 − tC1
tgerm	0	0	0.754c	0.578	0.468	0.032	0.048	−0.009	0.098	0.075	−0.039
	3	0	0.660c	0.600c	0.478	0.096	0.195	0.145	0.402	0.297	0.273
	0	2	0.718c	0.593	0.542	0.017	0.026	0.093	0.101	0.173	0.428
	3	2	0.818c	0.698c	0.709c	0.188	0.362	0.414	0.493	0.524	0.527
tem	0	0		0.883c	0.839c	0.681c	0.579	0.570	0.371	0.398	−0.016
	3	0		0.881c	0.819c	0.811c	0.768c	0.721c	0.388	0.399	0.486
	0	2		0.861c	0.820c	0.708c	0.524	0.572	0.297	0.359	0.653
	3	2		0.873c	0.837c	0.718c	0.723c	0.706c	0.569	0.529	0.658
tC1	0	0			0.979c	0.670c	0.843c	0.842c	0.764c	0.779c	0.637c
	3	0			0.952c	0.601c	0.902c	0.870c	0.778c	0.727c	0.592
	0	2			0.987c	0.587	0.821c	0.871c	0.742c	0.807c	0.740c
	3	2			0.978c	0.498	0.920c	0.882c	0.897c	0.842c	0.814c
tC1	0	0				0.655c	0.805c	0.880c	0.674c	0.833c	0.781c
	3	0				0.610c	0.842c	0.939c	0.624c	0.853c	0.811c
	0	2				0.582	0.803c	0.887c	0.652c	0.829c	0.840c
	3	2				0.426	0.848c	0.936c	0.735c	0.907c	0.918c
tem − tgerm	0	0					0.834c	0.806c	0.455	0.474	0.018
	3	0					0.783c	0.732c	0.156	0.251	0.369
	0	2					0.781c	0.727c	0.333	0.327	0.468
	3	2					0.671c	0.575	0.278	0.179	0.371
tC1 − tgerm	0	0						0.969c	0.871c	0.841c	0.555
	3	0						0.933c	0.737c	0.681c	0.538
	0	2						0.972c	0.849c	0.842c	0.588
	3	2						0.953c	0.898c	0.803c	0.759c
tC2 − tgerm	0	0							0.793c	0.903c	0.745c
	3	0							0.614c	0.844c	0.806c
	0	2							0.770c	0.887c	0.761c
	3	2							0.733c	0.908c	0.921c
tC1 − tem	0	0								0.916c	0.440
	3	0								0.835c	0.426
	0	2								0.934c	0.441
	3	2								0.909c	0.667c
tC2 − tem	0	0									0.763c
	3	0									0.854c
	0	2									0.732c
	3	2									0.917c

^aSpores were formed in SM containing 0 or 3.0% (wt/vol) NaCl and were grown in PYGS medium (GM) supplemented with 0 or 2.0% (wt/vol) NaCl.

^bBoldface type indicates independent variables. Normal type indicates nested variables, where one outgrowth event was a component of the other.

^cPairs of events that are strongly correlated (Pearson coefficient, ≥0.6).

Effects of sporulation and growth NaCl concentrations on the probability of growth.

Addition of 3% NaCl to the sporulation medium significantly decreased both the probability of spore germination and the probability that a germinated spore subsequently developed into a cell 5.5 μm long (Table 3). After stepwise selection, the significant terms of the logistic regression model were as follows: log (P/1 − P) = 1.9059 − 0.277NaCl_SM − 0.5315NaCl_GM + 0.274D(P). The positive value associated with the dummy variable, D(P), in the model described above indicates that the probability of growth from a germinated spore was greater than the probability of germination. Minus signs in front of the NaCl_SM and NaCl_GM terms show that the addition of NaCl to either the sporulation medium or the growth medium decreased both the probability of germination and the probability of subsequent growth. The constant associated with NaCl_GM is larger than the constant associated with NaCl_SM, indicating that NaCl in the growth medium decreased the probability of germination and subsequent growth more than the same concentration of NaCl in the sporulation medium did. No interaction was found between the addition of NaCl to the sporulation medium and the addition of NaCl to the growth medium. The terms for NaCl concentrations multiplied by the dummy variable were not significant, indicating that adding NaCl to the sporulation medium or the growth medium decreased the probability of germination and the probability of subsequent growth to similar extents.

TABLE 3.

Probability of germination and probability of subsequent outgrowth to the length of a mature cell for spores of nonproteolytic C. botulinum Eklund 17B as affected by added NaCl in the SM and the GM

Lag phase event	Sporulation conditions	Probability of lag phase event in:
		GM+0%	GM+2%
Germination	SM+0%	0.869	0.715
	SM+3%	0.767	0.467
Development of mature cell	SM+0%	0.865	0.785
from germinated spore	SM+3%	0.778	0.585

DISCUSSION

In this study we examined the effect of NaCl concentration during sporulation and subsequent growth on the lag time of individual spores of nonproteolytic C. botulinum and the stages within the lag phase. Cells must maintain a suitable level of cytoplasmic water and must maintain turgor pressure for effective functioning and growth (28, 20). Exposing microorganisms to osmotic environments outside the optimum range results in inhibition of a variety of physiological processes (3). The concentration of NaCl necessary to prevent growth of nonproteolytic C. botulinum under otherwise optimal conditions is considered to be 5.0% (wt/vol) (13). Two percent NaCl (equivalent to a water activity of 0.988) is a concentration found in many foods and has been reported to have relatively little effect on the times to turbidity during growth from spores of nonproteolytic C. botulinum (29). In the present study, the specific growth rates of populations in broth containing 0% or 2% added NaCl were both approximately 0.6 h⁻¹.

Although vegetative cells of nonproteolytic C. botulinum Eklund 17B could multiply and sporulate in medium containing 3% NaCl, preliminary studies showed that emerged cells had difficulty elongating into normal rod-shaped cells in growth medium containing 3% added NaCl. Emerged cells often swelled into large irregular spherical objects and then lysed or divided, producing a mixture of rods and spheres. These observations suggest that newly emerged cells are more sensitive to NaCl in the environment than mature vegetative cells. The highly variable morphology associated with emerging cells in medium containing 3% NaCl made it difficult to measure growth. Instead, growth experiments were conducted using growth medium with 0 or 2% added NaCl, in which the diameter of elongating cells was consistent.

Despite very similar population growth rates, cells in GM+0% and cells in GM+2% were microscopically distinct, suggesting that there were physiological differences. Cells in GM+2% were shorter and were frequently observed to divide when they reached a length of 7 to 9 μm, whereas their counterparts in GM+0% were frequently much longer than 11 μm without visible septation. As it was often difficult to determine when cells had divided, lag phases in different conditions were compared by measuring the time to a defined length. The length chosen was the length of one newly divided mid-exponential-phase cell (t_C1) in standard PYGS broth culture (5.5 μm).

In the present study, addition of 2% NaCl to the growth medium significantly reduced the probability of growth from nonproteolytic C. botulinum spores and significantly increased the lag time and the variability of the distribution of lag times from individual spores. This extended lag time was related to multiple stages, with the mean times and standard deviations increased for lag, germination, outgrowth, first doubling, and time to turbidity. This inhibitory effect could not have been due solely to increased energy required for growth in the presence of NaCl as NaCl inhibited germination as well as later stages of the lag phase. Germination and outgrowth involve different cellular processes (16). Germination is initiated by small molecules binding to receptors located in the spore inner membrane. This is followed by a cascade reaction using preformed enzymes which transforms the dormant spore to a metabolically active germinated spore. Subsequent outgrowth requires macromolecular synthesis. As the spore expands, the spore coats are shed and a young cell emerges. Synthesis continues with the cells growing to obtain the full complement of molecules found in mature vegetative cells and then continuing to increase in size until cell division occurs. Environmental NaCl increased the duration of many or all of these stages.

In a previous study we reported no correlation between germination time and total lag time for individual nonproteolytic C. botulinum spores incubated in good conditions (25). The results in the present paper show the lack of correlation between germination and later stages in the lag phase for spores in suboptimal conditions. Although spores must germinate before they can grow and thus germination limits the shortest time to growth, the time subsequently spent in outgrowth is highly variable and also longer than the germination time. Consequently, the first spore to germinate is not necessarily the critical individual that is the first cell to start multiplication or toxin production. It is thus impossible to predict total lag time from studies of the kinetics of germination alone, although knowledge of the rate and extent of germination, combined with data on outgrowth, is critical for improving predictions of growth from low numbers of spores.

The lag time of vegetative cells is more difficult to predict than their growth rate as the lag time is affected by cell history as well as the current environment (1). Historical conditions are known to affect lag times for spore inocula. For example, it is well documented that germination of many species is more rapid and extensive if the spores have been heat activated. It has previously been shown that the temperature during spore formation, the storage temperature, and a pregrowth heat shock can alter the ability of spores of nonproteolytic C. botulinum to germinate (7). For Bacillus subtilis it has also been shown that increasing the sporulation temperature decreased the proportion of spores able to grow on medium supplemented with NaCl (26). Nonproteolytic C. botulinum can be found in a range of environments, and it is possible that spores formed in the presence of increased NaCl concentrations, such as marine sediments, could be better adapted to germination and growth in foods with reduced water activity. When subjected to osmotic stress, bacterial cells respond by taking up potassium ions or accumulating or synthesizing compatible solutes, such as glycine betaine, proline, glutamine, ectoine, n-acetylornithine, and trehalose (8). These compounds can accumulate to high concentrations in the cell without impeding core metabolism and can balance the osmotic potential of the surrounding medium, enabling the cell to maintain turgor pressure (28). Other adaptations to growth in high-NaCl environments include changes to the composition of lipid membranes and alterations in global gene expression (8). It is possible that some of these adaptations that help vegetative cells grow in osmotic stress conditions could be incorporated into spores formed in such conditions.

If spores formed by cells growing in environments with increased NaCl concentrations gave rise to new cells that were better adapted to growth in the presence of elevated NaCl concentrations, then the time required for outgrowth in GM+2% would have been shorter than the time observed for spores formed in the absence of added NaCl. Conversely, the results of the present study showed that the mean time for outgrowth in GM+2% was extended when spores were formed in SM+3% rather than in SM+0% (Table 1 and Fig. 4). The spores formed in SM+3% had longer mean times for, and greater variability in, the distribution of times to germination, outgrowth, first doubling, lag time, and time to detection when they were subsequently grown in medium either with or without added NaCl. The inclusion of 3% NaCl in the sporulation medium also significantly reduced the probability of a spore germinating and the probability of outgrowth after germination. This suggests that spores formed in the presence of NaCl were damaged in some way so that they were less able to grow, irrespective of the growth conditions.

The reason that sporulation in the presence of NaCl adversely affects subsequent growth remains to be elucidated. C. botulinum Eklund 17B grew and sporulated in medium containing 3% NaCl, but it has previously been observed that the final turbidity is reduced at this concentration, and 4% NaCl adversely affects cell separation (24). For B. subtilis it has been shown that increased NaCl concentrations in the growth medium can alter the pattern of expression of early, phase II sporulation genes and can affect asymmetric septation (21). Similar alterations in C. botulinum sporulation or global gene expression brought on by stress conditions could affect fundamental properties of the spores formed. The decreased probability of germination observed for spores formed in SM+3% and the increase in germination time indicate that some part of the germination system was adversely affected, but the visual observation method does not allow further targeting to a particular stage of germination. It is possible that spores formed in SM+3% have fewer or damaged germinant receptors than spores produced in optimal conditions. A reduction in the number of germinant binding sites could adversely affect the rate or probability of germination but does not explain the observed extension of times for outgrowth or doubling. The adverse effect on these later stages suggests that whatever changes were induced by the presence of NaCl were major and affected several different functions within the cell.

It is possible that the stress associated with growth in the presence of an elevated NaCl concentration results in the formation of spores containing levels of components required for subsequent growth that are lower than those in spores produced in optimal conditions. Additional time required to make good any reduced macromolecular content could explain the increased time required for germination and outgrowth but may not explain increases in doubling time as doubling time was determined only after all cells had reached the same defined size. It has been suggested that osmoresistance in B. subtilis spores might be related to energy generation by the outgrowing spore (26), with energy coming from either endogenous generation from amino acids released by degradation of small acid-soluble proteins or from a readily taken up and metabolized energy source, such as glucose. Although the mechanisms may be different in clostridia, similar principles may apply. Spores contain a number of energy reserves (22), and any of these reserves could have been affected such that osmotically stressed mother cells produced spores with lower energy reserves than those produced by unstressed mother cells. The ability to take up glucose and amino acids from the medium may also have been adversely affected by alterations to transporters or membranes. It has previously been shown that the sporulation temperature significantly alters the spore coat, cortex composition, and core water content of B. subtilis spores (14) and that the osmoresistance of B. subtilis spores is influenced by the sporulation temperature (26). Ruzal et al. showed that spore formation by B. subtilis was adversely affected and then prevented by increasing NaCl concentrations, and they postulated that the absence of sporulation during an osmotic upshock could be related to modifications in membrane composition (21). It has been shown that NaCl-stressed B. subtilis cells show signs of iron limitation, possibly because iron transporters are partially inhibited (10). It has also been suggested that critical uptake systems for amino acids are compromised in high-salt medium (26). To fully determine the cause of the adverse effect of NaCl on subsequent germination and growth requires more knowledge concerning the mechanisms of sporulation and germination in nonproteolytic C. botulinum.

The interactions between the effects of suboptimal NaCl concentrations during sporulation or growth on the probability of germination and outgrowth by spores of nonproteolytic C. botulinum and the time required for germination and outgrowth are complicated. Exposure to NaCl during sporulation and subsequent growth appeared to have a synergistic effect on the mean time required for stages in the lag phase, particularly germination, but it did not have a synergistic effect on the probability of germination. This complexity suggests that there were multiple independent mechanisms for NaCl inhibition during the lag phase.

Predictive microbiology aims to map the relationship between the environments encountered by microorganisms and their responses. Predicting growth rate is easier than predicting lag time, as growth rate depends only on the current environment, whereas the lag time additionally depends on the historic environment. The observed independence of the growth rate and the times required for different stages in lag phase makes accurate prediction of lag times difficult in situations where the spore history is unknown. Fortunately, spores formed by NaCl-stressed cells were less able to germinate and subsequently grow in the presence of NaCl than spores formed under optimal conditions; thus, not including any assumption about spore history in predictive models may give a failsafe result. The sporulation NaCl concentration also had less effect on the shortest times to germination and growth and on the variability of the spore population than did the concentration of NaCl in the growth medium. However, historical NaCl exposure during sporulation did have a significant impact on the probability of growth, so that a failure to account for historic conditions could lead to an overestimate of growth and thus the risk associated with a product. Models should include separate data for probability of growth and time to growth.