Ralstonia solanacearum Facing Spread-Determining Climatic Temperatures, Sustained Starvation, and Naturally Induced Resuscitation of Viable but Non-Culturable Cells in Environmental Water

Ralstonia solanacearum is a bacterial phytopathogen affecting staple crops, originally from tropical and subtropical areas, whose ability to survive in temperate environments is of concern under global warming. In this study, two R. solanacearum strains from either cold or warm habitats were stressed by simultaneous exposure to natural oligotrophy at low (4 °C), temperate (14 °C), or warm (24 °C) temperatures in environmental water. At 4 °C, the effect of temperature was higher than that of oligotrophy, since R. solanacearum went into a viable but non-culturable (VBNC) state, which proved to be dependent on water nutrient contents. Resuscitation was demonstrated in vitro and in planta. At 14 °C and 24 °C, the effect of oligotrophy was higher than that of temperature on R. solanacearum populations, displaying starvation-survival responses and morphological changes which were stronger at 24 °C. In tomato plants, starved, cold-induced VBNC, and/or resuscitated cells maintained virulence. The strains behaved similarly regardless of their cold or warm areas of origin. This work firstly describes the natural nutrient availability of environmental water favoring R. solanacearum survival, adaptations, and resuscitation in conditions that can be found in natural settings. These findings will contribute to anticipate the ability of R. solanacearum to spread, establish, and induce disease in new geographical and climatic areas.


Introduction
The plant pathogen Ralstonia solanacearum is a relevant species and former constituent of the R. solanacearum species complex [1][2][3]. It causes severe wilt disease and economic losses in solanaceous and other basic crops for human consumption worldwide as well as in important ornamentals [4][5][6][7][8]. The pathogen frequently cannot be effectively controlled due to its high pathogenic potential and persistence in natural settings. Bacterial wilt control in the field has frequently been addressed by conventional methods, mainly agrochemicals and/or cultural practices, with variable results, and often with environmental impact [9,10]. Alternatively, biological control methods are being explored, such as bacteriophage-based treatments. The use of lytic bacteriophages may be an eco-sustainable strategy because of their specificity and bactericidal activity, although until now no bacteriophage-based product is commercially available against R. solanacearum [10]. Therefore, this pathogen poses a threat to the maintenance of global food security. In fact, the species has a quarantine status in the European Union (EU), the USA and Canada [5,11,12], and is considered a priority pathogen in agriculture for control and containment [13]. A major concern is that R. solanacearum seems to hold great potential for geographical expansion even under environmentally unsuitable conditions, as it appears to infect plants and persist during This work addressed, for the first time, the simultaneous effect of different temperatures and starvation on stress induction of two R. solanacearum strains (former R. solanacearum phylotype II) from either cold or warm habitats, and their responses in environmental water microcosms. Further, the capability of these R. solanacearum strains to resuscitate and keep pathogenic under conditions that can be found in natural settings was also firstly demonstrated. Knowledge derived from this work will help to foresee tendencies in R. solanacearum persistence and dissemination in aquatic systems within the frame of global warming, as well as their capability for establishment and disease induction in new geographical and climatic areas.

Bacterial Strains and Culture Conditions
Two bacterial strains of the present species R. solanacearum [2,3] isolated from either warm or cold habitats were used: strain IVIA-1602.1, from a diseased potato tuber from Canary Islands (Spain), and strain IPO-1609, from a diseased potato plant from The Netherlands [29], both are race 3, biovar 2 of the former R. solanacearum phylotype II. They were kept at −80 • C in a 30% (v/v) glycerol medium and routinely grown on the non-selective Yeast Peptone Glucose Agar (YPGA) [39] for 72 h at 29 • C. In stress induction assays, bacterial culturability was tested on YPGA and the Semiselective Medium South Africa (SMSA) agar developed for R. solanacearum isolation [40] after incubation at 29 • C for 72 h. SMSA medium was also used to re-isolate the pathogen from the host (tomato plants). Both media are frequently used for R. solanacearum isolation, since the colonies of the pathogen can be easily recognized as typically smooth. YPGA contains filtered-sterilized glucose, from which R. solanacearum produces a large amount of extracellular polysaccharide. Colonies are fluidal with pearly cream-white whorls. With respect to SMSA, semiselectivity is mainly based on the action of four antibiotics (penicillin, polymyxin, chloramphenicol and bacitracin), triphenyl-tetrazolium chloride, and crystal violet. Colonies are fluidal with reddish whorls.

Characteristics of Environmental and Distilled Water Samples
River water samples were collected according to [16] from four different locations in Spain, and nutrient contents were separately determined for each of them. In the different water samples, organic matter levels were from 2 to 3.73% (w/v), and the main ion concentrations ranged as follows (values per liter): Na + , 9.7-9.9 mg; K + , 2.1-2.9 mg; Ca 2+ , 10.1-13.0 mg; Mg 2+ , 3.9-5.0 mg; dissolved Fe, 0.24-0.27 mg; Mn, 0.06-0.11 mg; Cu, <0.024 mg; dissolved Zn 2+ , <0.018 mg; CO 3 2− , <1.8 mg; NO 3− , 4.37-5.93 mg; P 2 O 5 , 0.374-0.583 mg, and Cl − , 9.4-11.4 mg. Salt contents in the samples were correspondent with conductivity values from 151 to 168 µSiemens/cm at 20 • C, and pH values were from 7.48 to 7.83. Distilled water, used for comparative purposes in some assays, had no organic matter and only trace mineral ions: at 20 • C, conductivity was ≤20 µSiemens/cm, and pH value was 7. All water samples were stored in the refrigerator.

Preparation and Monitoring of Stressed R. solanacearum in Water Microcosms
All water samples were autoclaved and filtered through 0.22-µm-pore-size membranes and used for microcosm preparation and inoculation with either of the strains IVIA-1602.1 or IPO-1609 at a range of 5 × 10 6 -1 × 10 7 CFU (colony-forming units)/mL similarly to [6]. To induce stressed R. solanacearum populations, cells in microcosms were incubated at 4 • C, 14 • C, and 24 • C without shaking for 40 days or until loss of culturability. The temperature of 4 • C was selected because it had induced the VBNC state in R. solanacearum (former R. solanacearum phylotype II) in non-environmental pure water [26,28,29]. The temperatures of 14 • C and 24 • C were within the range in which R. solanacearum had been detected in environmental water [16,17]. Initially, microcosms were prepared with each of the four river water samples from different locations, and the survival of R. solanacearum monitored at 4 • C and 24 • C. Based on the results obtained, one river water sample was Microorganisms 2022, 10, 2503 4 of 17 selected for a comparative study on the survival of R. solanacearum in microcosms of river water versus distilled water at 4 • C, 14 • C, and 24 • C. Microcosms from river water samples and distilled water were prepared in triplicate.
Sampling from each microcosm was performed at inoculation time (day 0) and at 1, 2, 4, 8, 14, 28, and 40 days post-inoculation (dpi) to monitor: 2.3.1. Total, Viable, and Culturable Bacterial Populations Microscopic counts of total and viable R. solanacearum cells were done by a direct viable count (DVC) method [41], extended to 16 h [26] and subsequent staining with either the polyclonal antiserum 1546-H IVIA against R. solanacearum or acridine orange [16,42]. Plate counts of culturable cells were done on two media, the general YPGA and the semiselective SMSA, both recommended by EU legislation to isolate the pathogen from environmental samples [43]. R. solanacearum colonies for culturable cell counts were confirmed by PCR as described [43] with primers Ps-1 and Ps-2 based on the sequence of the 16S rRNA gene.

Cell Morphology
Bacterial cell shape was observed by specific immunofluorescence staining with the polyclonal antiserum 1546-H IVIA against R. solanacearum [16]. Cell morphology was observed with a Nikon Eclipse E800 microscope at a magnification of ×1000. Pictures were taken with an adapted digital camera DXM1200 using ACT-1 version 2.62 software, and no processing of the images was performed. At each sampling time and for each temperature and environmental water microcosm, the number of bacilli and/or cocci from at least 20 random fields was counted (approximately 300 cells).

Pathogenicity
The ability of starved R. solanacearum cells incubated in the environmental water microcosms at 4 • C, 14 • C and 24 • C to induce disease was tested from each triplicate microcosm at each sampling time on groups of 72 tomato plants cv. 'Roma' aged three weeks (two plants per microcosm, six plants per water sample at each temperature). Inoculations were performed by injecting into the stem volumes of 10 µL directly taken from the microcosms. Plant inoculations were carried out according to EU Legislation [43]. Positive and negative controls were performed on groups of 12 tomato plants cv. 'Roma' (six plants per positive/negative control) at each sampling time. In the case of positive controls, inoculations were performed by injecting 10 µL of a freshly growing cell suspension from either of the two R. solanacearum strains. Each suspension was previously washed and adjusted in sterile 10 mM phosphate buffered saline solution (PBS), pH 7.2 (NaCl, 8 g/L; PO 4 H 2 Na·2H 2 O, 0.4 g/L; PO 4 HNa 2 ·12H 2 O, 2.7 g/L) to OD 600nm = 0.1 (approximately 10 8 CFU/mL), and diluted to a final concentration of about 10 7 CFU/mL. In the case of negative controls, inoculations were done by injecting 10 µL of sterile 10 mM PBS. Incubation of the plants and monitoring of disease symptoms were performed in a growth chamber (16 h light, 8 h dark; 26 • C) under quarantine conditions. The pathogen was re-isolated from the wilting plants by cutting 2-3 cm of the stems above the inoculation point, and plating the obtained extracts onto SMSA. The colonies were PCR-identified as described [43] with primers Ps-1 and Ps-2 based on the sequence of the 16S rRNA gene. Stems from inoculated non-wilted plants were processed in the same way.

Resuscitation of R. solanacearum Populations from the VBNC State Induced in Environmental Water Microcosms
Assays for resuscitation were performed in three different conditions with VBNC R. solanacearum populations from the environmental water microcosms at 4 • C in triplicate.
To determine if the appearance of culturable cells was due to true resuscitation instead of regrowth of a few remaining culturable but undetected cells, serial ten-fold dilutions were carried out with the VBNC cells, as described [44,45]. The first series of ten-fold dilutions was performed when the microcosms were containing initially approximately 10 6 viable Microorganisms 2022, 10, 2503 5 of 17 cells/mL and <10 culturable cells (CFU)/mL, until reaching concentrations of 10 −2 viable cells/mL and <10 −7 culturable cells (CFU)/mL. From then and prior to the resuscitation assays, non-culturability was tested at each sampling time by plating volumes of 1 mL directly taken from the microcosms. Resuscitation assays were performed with aliquots (1-10 mL) taken from each environmental water microcosm and their ten-fold dilutions, both in vitro and in planta, according to three different procedures, as follows: 2.4.1. By Enrichment in a Modified Wilbrink (WB) Broth [16] Direct aliquots and their ten-fold dilutions in WB broth were incubated at 29 • C with shaking (200 r.p.m.) until appearance of turbidity or for at least one week. Moreover, additional aliquots were taken, transferred to WB broth, and maintained at 4 • C without shaking. Sampling was at time 0 and once a week during a month. To check culturability from turbid tubes, streaks were plated onto YPGA. To test non-culturability from nonturbid tubes, 100-µL volumes were plated onto YPGA. Colonies appeared on plates were PCR-identified. To prove pathogenicity of the colonies on plates, bacterial suspensions were prepared from them and 10-µL volumes were stem-inoculated onto tomato plants (two plants per suspension). Moreover, 10-µL volumes were directly taken from the turbid tubes and inoculated onto stems (two plants per tube) for pathogenicity tests of the cells. Plants were processed for re-isolation and identification of the bacterial pathogen as abovementioned.

By Temperature Upshift in Environmental Water
Briefly, 10-mL aliquots and their ten-fold dilutions in sterile environmental water were incubated at 24 • C without shaking. Sampling was at time zero and each two weeks during a month. To check culturability, 100-µL volumes from the dilutions were daily plated onto YPGA until appearance of colonies or for at least one week. The colonies were PCR-identified and their pathogenicity tested as abovementioned. Moreover, pathogenicity of the cells in the ten-fold dilutions was proved by inoculating 10-µL volumes directly from the dilutions onto the stems (two plants per dilution), which were processed as described.

In Planta
From direct aliquots and their ten-fold dilutions in sterile environmental water, 10-µL volumes were stem-inoculated onto tomato plants aged three weeks (two plants per dilution). Sampling was at time 0 and over one month. Plants were processed for re-isolation and identification of the pathogen as abovementioned.

Statistical Analysis
Survival assays were performed at three incubation temperatures with environmental water sample and distilled water in triplicate R. solanacearum-inoculated microcosms. Total, viable, and culturable data of R. solanacearum cell counts were normalized by logtransformation, and mean values analyzed by a linear regression model considering the following factors: incubation temperature, type of water (environmental or distilled), period of incubation, media, and bacterial strain. Differences among means of coccoid percentages at the three temperatures were estimated by variance analysis (ANOVA). A p value < 0.05 was defined as significant.

R. solanacearum Goes into a Nutrient-Dependent Cold-Induced VBNC State in Environmental Water
At low (4 • C) temperature, in environmental or distilled water (Figure 1), total populations of the strain IVIA-1602.1 of R. solanacearum remained above their initial inoculation numbers throughout the 40-day experiments, while viability was slightly lower, with declines approximately from 25-30 dpi in both types of water. In contrast, culturable bacterial populations significantly decreased (p < 0.05) about one log unit up until eight and four dpi for river and distilled water respectively, pointing out a proportion of cells sensitive to low-temperature conditions. Thereafter, progressively and significant stronger losses in culturability occurred, with values below detection level (10 1 CFU/mL) by 40 ± 7 and 20 ± 3 days, depending on the water sample, in river and distilled water, respectively ( Figure 1). These drops in culturable counts with high numbers of cells still viable indicated a majority of the populations becoming VBNC. The strain IVIA-1602.1 displayed similar trends in the microcosms of the other environmental water samples at 4 • C, only with differences in non-culturability between environmental and distilled water (p < 0.05). The non-selective medium YPGA and the semiselective SMSA medium yielded similar results for each of the water samples (p > 0.05).
sitive to low-temperature conditions. Thereafter, progressively and significant stronger losses in culturability occurred, with values below detection level (10 1 CFU/mL) by 40 ± 7 and 20 ± 3 days, depending on the water sample, in river and distilled water, respectively ( Figure 1). These drops in culturable counts with high numbers of cells still viable indicated a majority of the populations becoming VBNC. The strain IVIA-1602.1 displayed similar trends in the microcosms of the other environmental water samples at 4 °C, only with differences in non-culturability between environmental and distilled water (p < 0.05). The non-selective medium YPGA and the semiselective SMSA medium yielded similar results for each of the water samples (p > 0.05).
At temperate (14 °C) and warm (24 °C) temperatures, trends in total, viable and culturable populations were similar (p > 0.05) (Figures 1 and S1), and so only those at 24 °C have been plotted in Figure 1. At both temperatures, total populations of the strain IVIA-1602.1 remained above 10 7 cells/mL in environmental water and around this value in distilled water, and viability was slightly lower in both types of water for the 40-day experiments. During the period, culturability remained roughly at 10 7 CFU/mL in environmental water while in distilled water culturable cells stabilized below this value ( Figure 1). Assays with the other water samples yielded analogous results, also on both media (p > 0.05). At temperate (14 • C) and warm (24 • C) temperatures, trends in total, viable and culturable populations were similar (p > 0.05) ( Figure 1 and Figure S1), and so only those at 24 • C have been plotted in Figure 1. At both temperatures, total populations of the strain IVIA-1602.1 remained above 10 7 cells/mL in environmental water and around this value in distilled water, and viability was slightly lower in both types of water for the 40-day experiments. During the period, culturability remained roughly at 10 7 CFU/mL in environmental water while in distilled water culturable cells stabilized below this value ( Figure 1). Assays with the other water samples yielded analogous results, also on both media (p > 0.05).
Similarity in trends of culturable data from the microcosms of the four water samples inoculated with the R. solanacearum strains could be observed by the statistical analyses. For comparative purposes, increments of culturable data at 4 • C and 24 • C were jointly calculated with respect to the initial value and plotted with time to assess the effect of water sample ( Similarity in trends of culturable data from the microcosms of the four water samples inoculated with the R. solanacearum strains could be observed by the statistical analyses. For comparative purposes, increments of culturable data at 4 °C and 24 °C were jointly calculated with respect to the initial value and plotted with time to assess the effect of water sample (Figure 2, left) and the effect of media (Figure 2, right). Population dynamics of total, viable, and culturable cells of the strain IPO-1609 of R. solanacearum were similar to those of the strain IVIA-1602.1 (p > 0.05) in triplicate microcosms from environmental water samples ( Figure S2).

R. solanacearum Changes Their Shape in Environmental Water with Increased Temperatures
Cells of the strain IVIA-1602.1 were examined in triplicate environmental water sample microcosms at 4 °C, 14 °C, and 24 °C. Data are plotted in Figure 3. Population dynamics of total, viable, and culturable cells of the strain IPO-1609 of R. solanacearum were similar to those of the strain IVIA-1602.1 (p > 0.05) in triplicate microcosms from environmental water samples ( Figure S2).

R. solanacearum Changes Their Shape in Environmental Water with Increased Temperatures
Cells of the strain IVIA-1602.1 were examined in triplicate environmental water sample microcosms at 4 • C, 14 • C, and 24 • C. Data are plotted in Figure 3. At 4 °C and throughout the 40-day experiments, bacterial cells showed the typical R. solanacearum bacillar morphology and coccoid cells were seldom detected, with a frequency <1% depending on the water sample (Figure 3).
At 14 °C, a great majority of R. solanacearum cells kept bacillar shape. Coccoids were observed in a low proportion, with constant percentages around a value between 1-3% At 4 • C and throughout the 40-day experiments, bacterial cells showed the typical R. solanacearum bacillar morphology and coccoid cells were seldom detected, with a frequency <1% depending on the water sample (Figure 3).
At 14 • C, a great majority of R. solanacearum cells kept bacillar shape. Coccoids were observed in a low proportion, with constant percentages around a value between 1-3% throughout the first 28 days, and then a slight increase up to values ranging 2-6% by 40 dpi, depending on the water sample (Figure 3).
At 24 • C R. solanacearum bacilli remained a majority but, coccoid cells were more frequent: depending on the water sample, percentages by the first week were around 8-10%, progressively increasing to 13-16% by the second week, and then up to 22-32% by 28 dpi which stabilized to the end of the 40 days (Figure 3).
Among the low, temperate, and warm temperatures, the average percentage of coccoids significantly increased with temperature (p < 0.05). Cell shape of the strain IPO-1609 showed the same trends in one-off trials at each of the three temperatures.

Starved and/or Cold-Induced VBNC R. solanacearum in Environmental Water Keeps Virulent in Planta
During the 40-day periods of incubation at 4 • C, 14 • C, and 24 • C in the water microcosms under starvation conditions, aliquots were taken at different times to be inoculated in planta. R. solanacearum cells of strain IVIA-1602.1 incubated at 4 • C and inoculated in tomato stems induced disease in 98-100% of the plants (Figure 4). Similar wilting percentages were obtained with cells from microcosms at 14 • C and at 24 • C ( Figure 4) and were comparable to those of the strain IPO-1609 in one-off trials. At the three temperatures and depending on the water sample, viable R. solanacearum cells inoculated per plant were about 10 5 throughout the 40-day sampling periods, and only from approximately 28 dpi at 4 • C there was a slight decline to values around 10 4 viable cells per plant (Figure 4). At 4 • C, culturable cells inoculated per plant ranged from 10 5 to 10 4 in the initial dpi depending on the water sample. Then, they were decreasing until 10 4 -10 3 CFU per plant by the first week, and progressively to <10 CFU per plant by 28 dpi and to undetectable levels by 40 dpi (Figure 4). At 14 • C and 24 • C, culturable cells were 10 5 -10 4 per plant throughout the sampling periods ( Figure 4). Plants started to show symptoms within 8-11 dpi and completely wilted within four weeks. The pathogens were re-isolated on SMSA agar from the diseased plants and PCR-identified. Positive control plants yielded 100% wilting. Negative control plants did not show any symptoms.

R. solanacearum Resuscitates from the Cold-Induced VBNC State in Environmental Water and Is Fully Pathogenic in the Host
Assays carried out to assess the resuscitation capability of the VBNC R. solanacearum cells of strain IVIA-1602.1 yielded similar results in triplicate microcosms from environmental water samples. Data are summarized in Table 1, in the three different conditions. The viability of the VBNC cells when the microcosms were containing approximately 10 6 viable cells/mL and <10 culturable cells (CFU)/mL is illustrated in Figure 5.

By Enrichment in WB Broth
From cold-induced VBNC cells of R. solanacearum strain IVIA-1602.1, and after the temperature upshift with shaking and nutrients, monitoring was of: (i) turbidity by R. solanacearum growth in the direct aliquots and their serial ten-fold dilutions, (ii) culturability on YPGA, and (iii) pathogenicity in the host. These were observed in all the direct aliquots and their serial ten-fold dilutions up to 10 −6 , corresponding to 1 VBNC cell/mL, at time 0 of the VBNC induction of the R. solanacearum populations (Table 1). Thereafter, the resuscitation capability of these VBNC cells in the microcosms was decreasing with time until reaching about two orders of magnitude by one month from the VBNC induction. Time for resuscitation (estimated as time for observation of turbidity by R. solanacearum growth) was 24 h for direct aliquots, 36 h for dilutions 10 −1 , 10 −2 , and 10 −3 , 48 h for dilutions 10 −4 and 10 −5 , and four days for dilutions 10 −6 . These rates of growth were maintained throughout the experimental period. Culturability of the cells in the turbid dilutions was positive in all cases, and colonies were PCR-identified as R. solanacearum. However, cells from aliquots in WB broth maintained at 4 • C without shaking remained non-culturable. Pathogenicity assays were positive in all plants, either when inoculated directly from the turbid dilutions or from the colonies on the plates. The pathogen was re-isolated from the wilted plants and PCR-identified. Sensitivity of the detection of resuscitated R. solanacearum cells from the VBNC state was 1 VBNC cell/mL.
Microorganisms 2022, 10, x FOR PEER REVIEW 9 o 4 °C, culturable cells inoculated per plant ranged from 10 5 to 10 4 in the initial dpi depe ing on the water sample. Then, they were decreasing until 10 4 -10 3 CFU per plant by first week, and progressively to <10 CFU per plant by 28 dpi and to undetectable level 40 dpi (Figure 4). At 14 °C and 24 °C, culturable cells were 10 5 -10 4 per plant through the sampling periods ( Figure 4). Plants started to show symptoms within 8-11 dpi completely wilted within four weeks. The pathogens were re-isolated on SMSA agar fr the diseased plants and PCR-identified. Positive control plants yielded 100% wilting. N ative control plants did not show any symptoms.    Assays carried out to assess the resuscitation capability of the VBNC R. solanacearum cells of strain IVIA-1602.1 yielded similar results in triplicate microcosms from environmental water samples. Data are summarized in Table 1, in the three different conditions. The viability of the VBNC cells when the microcosms were containing approximately 10 6 viable cells/mL and <10 culturable cells (CFU)/mL is illustrated in Figure 5.  [41,42], and estimated in 10 6 VBNC cells/mL. Scale bars: 5 µm.

By Enrichment in WB Broth
From cold-induced VBNC cells of R. solanacearum strain IVIA-1602.1, and after the temperature upshift with shaking and nutrients, monitoring was of: (i) turbidity by R. solanacearum growth in the direct aliquots and their serial ten-fold dilutions, (ii) culturability on YPGA, and (iii) pathogenicity in the host. These were observed in all the direct aliquots and their serial ten-fold dilutions up to 10 −6 , corresponding to 1 VBNC cell/mL, at time 0 of the VBNC induction of the R. solanacearum populations (Table 1). Thereafter,  [41,42], and estimated in 10 6 VBNC cells/mL. Scale bars: 5 µm.

By Temperature Upshift in Environmental Water
From cold-induced VBNC cells of R. solanacearum strain IVIA-1602.1 and after the temperature upshift, monitoring was performed regarding: (i) culturability on YPGA and (ii) pathogenicity in the host. These were observed in cells from the direct aliquots and their serial ten-fold dilutions up to 10 −4 , corresponding to 10 2 VBNC cells/mL, and up to 10 −3 , corresponding to 10 3 VBNC cells/mL, respectively, at time 0 of the VBNC induction (Table 1). Then, the resuscitation capability of the cells in the microcosms was decreasing with time until about three orders of magnitude by one month from the induction (Table 1). Time for resuscitation (estimated as time for observation of culturability after plating) was 48 h from the temperature upshift. This was observed after plating 100-µL volumes from the same direct aliquots incubated during 24 h, 48 h and 72 h from the temperature upshift, and then on plates for 3 days at 29 • C. Sampling at 24 h yielded no growth on the plates, sampling at 48 h turned out in countable colonies, and continuous bacterial growth was observed after sampling at 72 h from the temperature upshift, reaching in all cases 10 6 CFU/mL. Colonies were PCR-identified as R. solanacearum. Pathogenicity assays to test cells from the colonies were positive in all plants. Pathogenicity assays to test cells from the direct aliquots and their serial ten-fold dilutions were positive in at least one of the two inoculated plants per microcosm. The pathogen was re-isolated and PCR-identified from the wilted plants. Sensitivity of the detection of resuscitated R. solanacearum cells from the VBNC state was 10 2 VBNC cells/mL.

In Planta
From cold-induced VBNC cells of R. solanacearum strain IVIA-1602.1 and after the temperature upshift in the host, pathogenicity was monitored. This was observed in cells from the direct aliquots and their serial ten-fold dilutions up to 10 −3 , corresponding to 10 3 VBNC cells/mL (10 VBNC cells/plant), at time 0 of the VBNC induction, and then the resuscitation capability was decreasing until about one order of magnitude by one month from the induction ( Table 1). The pathogen was re-isolated and PCR-identified from the wilted plants. Sensitivity of the detection of resuscitated R. solanacearum cells from the VBNC state was 10 3 VBNC cells/mL.

Discussion
To anticipate the spread of the disease, current prevention and control strategies against bacterial wilt should take into account knowledge on the potential behavior of R. solanacearum in response to global warming. In this work, adaptations by strains of the pathogen from different climatic regions were observed under exposure to environmental temperatures in oligotrophic freshwater, which allowed survival without losing wilting capacity.
With respect to low temperatures, this is the first report of viable R. solanacearum populations induced to the VBNC state in environmental water, in conditions more approaching those of natural settings. Previous work described either: (i) a loss in R. solanacearum culturability under low temperature in natural or distilled, ultrapure water, but without determining the viability of the bacterial populations, thus without confirming the presence of VBNC cells [26,27,32,46]; or (ii) the VBNC induction in distilled, ultrapure, non-environmental water [28,29]. The fact that other cold-adapted water bacteria are not likely to be cold-induced VBNC [47], contrarily to what was observed in this work with R. solanacearum from different climates, suggests that this pathogen is not naturally coldadapted, even when introduced to cold habitats. Thus, this work demonstrated that, in R. solanacearum, low temperature plays a major role than starvation in inducing the VBNC state ( Figure 6), contrarily to what has been reported for other bacteria [47]. However, starvation-induced stress proteins could have protected R. solanacearum from temperature damage, since it was less vulnerable to cyclic cold stress in pure water than in host tissue [27]. Moreover, the VBNC R. solanacearum cold-induction period occurred more slowly in environmental water, pointing to an effect of water nutrient contents, namely trace organic matter and some dissolved salts available for the cells but absent in distilled water, and so nutrient concentrations not supporting R. solanacearum growth would act as an additional stress contributing to the cold-induced VBNC state. In the field, latent VBNC cells maintain structure, biology, and significant gene expression, and global climate change might be resuscitating them when low temperature is the inducing factor, leading to increased outbreaks [48]. Similar to R. solanacearum, a lower mineral salt concentration markedly shortened the VBNC Vibrio parahaemolyticus induction period [49]. Bacterial species, such as V. vulnificus and Aeromonas hydrophila, also behaved similarly to R. solanacearum under low-temperature and nutrient-limiting conditions [50,51], whilst others, such as Campylobacter jejuni and Erwinia amylovora, displayed different responses [52][53][54][55].
At temperate and warm temperatures in environmental water, R. solanacearum populations displayed starvation-survival responses as described at 24 • C [6] and similar in terms of population levels. Lack of unculturability was in agreement with previous work reporting the isolation and persistence of the pathogen in environmental water at temperatures allowing R. solanacearum multiplication [15][16][17]26]. The presence of organic matter and salts in environmental water contributed to stimulate R. solanacearum survival, similarly to Aerobacter aerogenes [56], E. amylovora [53,54], and Leuconostoc mesenteroides [57], where trace minerals facilitated culturability, since mineral salts can affect not only cell growth, but also cell survival during nutrient limitation conditions [31].
Morphological changes are a visible indicator of adaptation to the environment [31,54,58]. Starved R. solanacearum cells transformed from the typical bacilli into coccoids, since shape rounding off and size reduction allow nutrients to be sequestered more efficiently [31]. This was observed in different proportions according to temperature. Although cells entering the VBNC state often exhibit dwarfing [37,59], R. solanacearum coccoids were seldom observed during this process, probably because low temperature rapidly causes decrease in R. solanacearum metabolism and uptake of water nutrients, with constitutive expression of genes associated with survival and stress response for a stable maintenance of their transcript level [60]. Likewise, copper-induced VBNC R. solanacearum cells were unchanged in size [61]. Therefore, at both starvation-and survival-inducing temperatures, the transition to coccoids would be mostly influenced by nutrient limitation and to a lesser extent by low temperature, as reported elsewhere [62]. At these two temperatures, the proportions of coccoids differed, with significantly higher numbers at warm temperature, probably to improve the speed for exchange of material with the surrounding environment to hold a faster energy-consuming metabolism, which becomes a requirement at elevated temperatures [58,63]. Thus, in natural nutrient-deprived environments, the stress of oligotrophy would be less intense for the pathogen at temperatures around 14 • C than at values nearer to the optimum as 24 • C, and so temperature would be modulating this adaptation to oligotrophy (Figure 6), acting on cell metabolism rate and nutrient requirement frequency. Moreover, in the presence of indigenous microbiota, R. solanacearum survived longer at 14 • C than at 24 • C in oligotrophic environmental water [35], and the culturability of R. solanacearum strain IPO-1609 was favored at 12 • C and 20 • C rather than at 28 • C in agricultural water in both the presence and absence of other aquatic microorganisms [26]. Similar to R. solanacearum, a number of bacterial species decreased their sizes with increasing environmental temperatures [51,52,63]. Notwithstanding, this cannot be considered a general bacterial behavior [47,50,58,63].
Although R. solanacearum has frequently been described as cold tolerant [14,28], the strains introduced to either cold or warm areas were apparently better temperate-adapted than cold-adapted as considered [12], and similarly to [28], where data indicated that R. solanacearum had no special adaptation to survive cold temperatures in water under controlled conditions. Likewise, the cold-water-adapted Vibrio tasmaniensis did not enter the VBNC state at 4 • C while the warm-water-adapted V. shiloi did [47].
R. solanacearum resuscitation from the cold-induced VBNC state was observed after stress removal by placement of the VBNC cells in three different favorable conditions, including the host plant, and all of them implying, at least, an upshift in temperature. In enrichment conditions, culturability on solid medium and pathogenicity of the resuscitated R. solanacearum cells from the dilutions were both confirmed. Restoration of culturability was more dependent on the temperature upshift and shaking than the presence of nutrients, since VBNC cells in enrichment liquid medium at 4 • C were not able to form colonies, and so they maintained their VBNC status. In environmental water and after the temperature upshift, culturability and pathogenicity of the resuscitated cells were similarly confirmed, the only resuscitation-inducing factor here being the temperature upshift, which has not been reported for R. solanacearum up until now. That would explain the seasonal variation of R. solanacearum populations in environmental water [16,17]. If temperature is so critical, increases in water temperature corresponding to rising global surface temperatures will likely lead to a wider geographic distribution of R. solanacearum and a higher incidence of infections in planta resulting from resuscitated cells of the pathogen, as it is being observed in Vibrio species [48]. Likewise, a simple reversal of temperature was sufficient to allow the resuscitation of other bacterial species [37,38,59]. In contrast, it was not effective to resuscitate the close R. pseudosolanacearum (former R. solanacearum phylotype I) in soil and water [60,64], since the addition of hydrogen peroxide-degrading compounds, such as catalase or sodium pyruvate, was necessary. The number of bacterial cells resuscitated by temperature upshift in environmental water was equal to the initial inoculum, similarly to [37,44,60]. Resuscitation in planta of the VBNC R. solanacearum cells was evidenced by the occurrence of wilting symptoms, and progressively declined over time, accordingly to [29]. Virulence in tomato plants was also observed in revived cells of R. pseudosolanacearum after exiting a cold-induced VBNC state in pure water [64]. In all the three different resuscitation conditions, the resuscitated R. solanacearum cells displayed similar phenotypes to the original culturable cells, including virulence in tomato plants, as described [60]. Moreover, also in the three different conditions, a decrease in the proportion of VBNC cells capable of resuscitation occurred over time, this process being dependent on the age of the VBNC cells, as stated [38]. In that respect, several authors agree to consider the existence of gradual stages within the VBNC state, namely a reversible non-culturable stage where cells can be resuscitated and an irreversible non-culturable stage, where cells cannot be resuscitated, although they keep respiratory activity [28,60,64] (Figure 6). Among these conditions, in this work, the enrichment was the most effective for R. solanacearum resuscitation and the most sensitive for their detection. This is probably because it combines a temperature upshift with nutrients and shaking, which supplies with oxygen and disperses oxidative compounds (peroxides, other free radicals) accumulated extra-or intracellularly either produced in the cells in response to low temperature stress or commonly present in rich culture media [37,65]. Reversal of adverse VBNC-state-inducing factors can be efficiently applied to the detection of R. solanacearum resuscitated cells.
On the basis of all these results, starved and/or cold-stressed and/or cold-induced VBNC R. solanacearum cells could be present in environmental water, being a threat to secure crop production as they are not easily detected [61], can survive cold temperature fluctuations [27], and can revert to a fully pathogenic state just by a temperature upshift, which can be favored within the frame of global climate change conditions. All of these R. solanacearum survival forms maintained their capacity for in planta multiplication and colonization, causing disease symptoms in the host, as observed elsewhere for similar time periods [6,16,29]. Not only can global warming contribute to the pathogen spread and virulence, but crops resistant to R. solanacearum at moderate temperatures can also become more susceptible at high ambient temperatures [14,24], increasing the probability of infections. The bacterial wilt disease is most severe on plants at temperature values ranging from 25 • C to 35 • C [14,28,66].
On the other hand, taking into account the temperature interval of 4-10 • C applied by EU legislation [43] to transport suspected water samples, temperatures above 4 • C up to around 10 • C would be more advisable than 4 • C, since cultivation-based methods are required to confirm pathogen detection. For the inspection of environmental samples, it should be determined whether to also test for these VBNC cells [64] to improve the sensi-tivity of the detection. This is a relevant point since the early detection of R. solanacearum in irrigation water and its eradication would contribute to improve any integrated management program of the bacterial wilt disease [61].
Overall, R. solanacearum strains from either cold or warm origin were able to adapt to a combined effect of temperature and oligotrophy. At low temperature, the delay in the induction of the VBNC state in environmental water suggested a protective effect of water nutrient contents on bacterial cells and pointed out the relevance of performing survival studies in conditions better approaching those in the environment. At temperate and warm temperatures, adaptations to oligotrophy were starvation-survival responses and morphological changes influenced by temperature. It appeared that, when temperature was the main stress (cold conditions), nutrient deprivation acted as an additional stress, contributing to accelerate the effect of temperature, and conversely, when oligotrophy was the main stress (temperate and warm conditions), temperature increased the effect of oligotrophy ( Figure 6). In all conditions, R. solanacearum cells remained pathogenic and capable of resuscitation by a simple reversal of temperature.
Microorganisms 2022, 10, x FOR PEER REVIEW 15 of 18 Figure 6. Influence of temperature and oligotrophy on Ralstonia solanacearum stress induction in environmental water and stress responses by the bacterium. Adapted to R. solanacearum from a proposed model by [67]. See text for details.
Under circumstances of global warming, understanding R. solanacearum adaptations to environmental abiotic stresses can help to design strategies to prevent and control their spread and dissemination in waterways and other natural settings. This is particularly important in the case of R. solanacearum-contaminated water, since it cannot be used for irrigation, contributing to the global problem of the increased water scarcity in the environment due to climate change, which has serious implications, among others, for food production and health.  Influence of temperature and oligotrophy on Ralstonia solanacearum stress induction in environmental water and stress responses by the bacterium. Adapted to R. solanacearum from a proposed model by [67]. See text for details.
Under circumstances of global warming, understanding R. solanacearum adaptations to environmental abiotic stresses can help to design strategies to prevent and control their spread and dissemination in waterways and other natural settings. This is particularly important in the case of R. solanacearum-contaminated water, since it cannot be used for irrigation, contributing to the global problem of the increased water scarcity in the environment due to climate change, which has serious implications, among others, for food production and health.