Gram-Negative Bacteria from Organic and Conventional Agriculture in the Hydrographic Basin of Loja: Quality or Pathogen Reservoir?

: Organic and conventional agriculture are vital for the development of human society; however, the use of contaminated water and the inappropriate use of organic chemical fertilizers can lead to an increase in the microbial load (potentially pathogenic) of the normal microbiota of the agricultural soil. In this context, the aim of our study was to isolate Gram-negative bacteria from the superﬁcial soil layer and irrigation water of agricultural areas (11 organic farms and nine conventional farms) and consider their potential ecological and health risk importance. Through culture isolation using three bacterial media (TSA) trypticase soy agar (general nutritive media); MacConkey Gram-negative bacteria and (EMB) eosin methylene blue agar (selective for Enterobacteriaceae) and classical biochemical tests, we recorded a total of 12 bacterial species, most belonging to the Enterobacteriaceae family, such as Enterobacter , Escherichia , Klebsiella , Salmonella and Shigella , which can be pathogenic for humans and animals. In contrast, bacteria such as Pantoea agglomerans , Pseudomonas aeruginosa , P. ﬂuorescens and Burkholderia mallei could facultatively work as diazotrophic or plant growth-promoting rhizobacteria. Soil bacteria richness detected with the media applied was signiﬁcantly higher than water bacteria, but we found no signiﬁcant differences between organic and conventional agriculture. We conclude that the isolated bacteria in water and soil mostly be-longs to enteropathogenic bacteria which could be pathogenic to animals and humans. While other bacteria like Pseudomonas aeruginosa could be viewed as useful by improving nutrient availability in agricultural soil.


Introduction
Agriculture is of vital importance for the development of human societies and has an enormous impact in the soil functional system [1], since almost 40% of the total land area is used for food production through agriculture [2]. Currently, organic agriculture has been revived and represents a growing proportion in the world economy [3], since the population seeks to obtain products which are "healthier" and free of chemicals harmful for humans and the environment [4]. Conventional agriculture is characterized by the usage of chemicals during its production [5]; on the contrary, organic agriculture is based on the effective natural management of plant-soil nutrition cycling [6] and makes use of natural organic compounds such as manure, compost and homemade fertilizers from animal waste such as chicken manure to compensate for nutrient deficiencies [5].
Agronomy 2021, 11, 2362 3 of 13 lands and urbanization [31] with a population of over 200,000 inhabitants (INEC = Instituto Nacional de Estadísticas y Censos). Twenty rural farms (11 organic farms and nine conventional farms) were evaluated, where vegetables such as lettuce, cabbage, carrots, potatoes and legumes are grown combined, mainly under mixed soils (Vertisol or Mollisols) [7]. In all cases, the land had been used as a vegetable farm for at least three years.
The selection criteria for soil and water samples for each type of farm were defined as follows: I.
Organic: farms that apply organic fertilizers such as poultry manure and/or handprocessed compost to the agricultural soil and avoid or restrict the use of chemical compounds for soil fertilization and plague control. II. Conventional: farms that apply commercial chemical fertilizers (e.g., N10-P30-K10, urea 46%, ammonium nitrate 36%) and plague control (e.g., metaldehyde; chlorpyrifos) for agricultural management. All conventional farms also use animal manure as a base fertilizer before applying chemical treatments according to the particular needs of each farm.

Experimental Design and Sampling
A total of 40 samples were taken corresponding to 20 rural farms (replicates = 11 organic farms and nine conventional farms). For water analysis, 20 water samples (500 mL of piped, non-chlorinated water) were collected at the intakes at the entrance of the farms, which come directly from a stream, using sterile glass bottles. The water spring used for crop irrigation may be shared by more than one farm, as indicated in the map (Figure 1). The irrigation caption of water to each farm is protected by wood fencing against cattle.
The studied sites were located around Loja city (canton) between 2000 to 2400 m a.s.l., with an environmental annual mean temperature of 20 °C, 80% relative humidity and 900 mm 3 of precipitation [30]. Loja basin is mainly covered by native vegetation, grasslands and urbanization [31] with a population of over 200,000 inhabitants (INEC = Instituto Nacional de Estadísticas y Censos). Twenty rural farms (11 organic farms and nine conventional farms) were evaluated, where vegetables such as lettuce, cabbage, carrots, potatoes and legumes are grown combined, mainly under mixed soils (Vertisol or Mollisols) [7]. In all cases, the land had been used as a vegetable farm for at least three years.
The selection criteria for soil and water samples for each type of farm were defined as follows: I.
Organic: farms that apply organic fertilizers such as poultry manure and/or handprocessed compost to the agricultural soil and avoid or restrict the use of chemical compounds for soil fertilization and plague control. II.
Conventional: farms that apply commercial chemical fertilizers (e.g., N10-P30-K10, urea 46%, ammonium nitrate 36%) and plague control (e.g., metaldehyde; chlorpyrifos) for agricultural management. All conventional farms also use animal manure as a base fertilizer before applying chemical treatments according to the particular needs of each farm.

Experimental Design and Sampling
A total of 40 samples were taken corresponding to 20 rural farms (replicates = 11 organic farms and nine conventional farms). For water analysis, 20 water samples (500 mL of piped, non-chlorinated water) were collected at the intakes at the entrance of the farms, which come directly from a stream, using sterile glass bottles. The water spring used for crop irrigation may be shared by more than one farm, as indicated in the map ( Figure 1). The irrigation caption of water to each farm is protected by wood fencing against cattle.
For soil microbiological analysis, 20 soil samples (1 kg) were obtained, one for each farm (11 organic farms and nine conventional farms), resulting from five randomly placed sub-samples. These sub-samples correspond to 200 g of sieved material (sieve 2 mm 2 ) from 500 g of soil taken at approximately 10 cm deep using a bore tool around the cultivated For soil microbiological analysis, 20 soil samples (1 kg) were obtained, one for each farm (11 organic farms and nine conventional farms), resulting from five randomly placed sub-samples. These sub-samples correspond to 200 g of sieved material (sieve 2 mm 2 ) from Agronomy 2021, 11, 2362 4 of 13 500 g of soil taken at approximately 10 cm deep using a bore tool around the cultivated plants. All the samples were kept in a cooler (cooling boxes filled with ice blocks) and after sampling, these were analyzed immediately (1 h)

Isolation and Biochemical Identification of Bacterial Strains
For bacterial isolation, 1 g of each sample was suspended in 9 mL of 0.1% peptonized water and homogenized for 30 sec. Decimal dilutions of 10 −1 to 10 −2 were prepared from the water and soil samples, based on the probable microbial density (APHA, 1998), to obtain isolated colonies. The viable bacterial cells for both dilutions were counted as colony-forming units (CFUs). Duplicate aliquots of 100 µL of the dilutions were taken and spread on three types of solid media: trypticase soy agar (TSA) general nutritive media, MacConkey agar selective for Gram-negative bacteria and eosin methylene blue agar (EMB) selective for Enterobacteriaceae. Additional revision of colonies was done at 72 h to detect any color change or new growth of new colonies.
After the incubation period, 1 or 2 colonies per plate were selected for subculturing based on: differences in morphology, coloration or brightness and colony size. All pure cultures were analyzed via differential Gram staining. Then, another subculturing was performed on solid and semi-solid media to assure classical species identification with positive/negative biochemical reaction tests [32][33][34]. The latter test was carried out applying: Simmons citrate agar, iron-tripe sugar agar (TSI), urea agar, sulfide indole motility (SIM) and lysine iron agar (LIA). All subcultures and differential biochemical tests were incubated in aerobiosis for 24 h at 37 • C.

Data Analysis
The sampling effort of bacterial species in each sample type was estimated using species accumulation curves and a Chao 2 nonparametric richness estimator based on the incidence of species in a determined sample. The richness of microorganisms at the level of farm type (organic vs. conventional) and sample type (soil vs. water) was evaluated using an unpaired two-samples Wilcoxon test, because the data did not present a normal distribution (p < 0.05).
For community composition, a nonmetric multidimensional scaling analysis (NMDS) was performed to visualize the similarity between bacterial species composition according to farm type and sample type, using Euclidean distance and 999 Monte Carlo permutations. To analyze the effect of the farm type and sample type variables on the composition of the microorganism communities, a one-way analysis of similarities (ANOSIM) was performed. All analyses were done using the statistical "vegan" 2.5-7 package in the R environment [35].
From pure isolates (40 in total), 12 bacterial species were identified according to their biochemical reaction, out of which two were identified to genus level (Table 1). No new strains were obtained at 72 h.   Most bacteria can be correlated by urea, motility and lysine reactions and a few by strong generation of gas. On the other hand, Shigella spp., was not reactive for the biochemical test.
The highest isolated Gram-negative bacteria were found in soil samples taken from organic farms ( Table 2). Five species present ubiquitous habitat, four species are frequently found in animals, humans or plants and two species are exclusively intracellular in human intestine. According to the ecology, eight bacteria species have a functional role related to plant-soil nutrition cycling or growth promotion. On the other hand, ecological information for three bacteria species was not found. Based on the bibliography, all bacteria species could present some pathogenic activity (Table 2).  The species accumulation curves indicated an adequate sampling for the two types of substrates ( Figure 2). The Chao 2 estimator (https://search.r-project.org/CRAN/refmans/ fossil/html/chao1.html, accessed on 4 May 2021) for the water samples indicated an adequate sampling, with six estimated and six observed species. As for the soil samples, the Chao 2 estimator indicated fifteen estimated species, compared to the 10 observed species, suggesting that the sampling is close to 70%.  Soil samples had a higher bacterial richness (Mann-Whitney test, p < 0.0001) compared to water samples ( Figure 3). The type of farm did not have an effect on bacterial species richness (p = 0.1433). Soil samples had a higher bacterial richness (Mann-Whitney test, p < 0.0001) compared to water samples (Figure 3). The type of farm did not have an effect on bacterial species richness (p = 0.1433).
The NMDS analysis indicated an ordering between bacterial communities related to farm and sample type (Figure 4). This was corroborated via ANOSIM analysis, which pointed out significant differences between the two sample types (r = 0.1131, p = 0.0027).
Conversely, type of farm (organic and conventional sampling points) does not show significant differences (r = 0.0427, p = 0.1004). Soil samples had a higher bacterial richness (Mann-Whitney test, p < 0.0001) compared to water samples (Figure 3). The type of farm did not have an effect on bacterial species richness (p = 0.1433). The NMDS analysis indicated an ordering between bacterial communities related to farm and sample type (Figure 4). This was corroborated via ANOSIM analysis, which pointed out significant differences between the two sample types (r = 0.1131, p = 0.0027).
Conversely, type of farm (organic and conventional sampling points) does not show significant differences (r = 0.0427, p = 0.1004).

Discussion
Generally, microbiota are considered excellent catalysts, facilitating nitrogen fixation, mineralization and solubilization [10,11], transforming organic compounds, amino acids and proteins and improving their uptake for plants [66,67]. However, microbial diversity does not always have a positive impact on the functioning of ecosystems, including soil

Discussion
Generally, microbiota are considered excellent catalysts, facilitating nitrogen fixation, mineralization and solubilization [10,11], transforming organic compounds, amino acids and proteins and improving their uptake for plants [66,67]. However, microbial diversity does not always have a positive impact on the functioning of ecosystems, including soil [68].
The results obtained in this study show a higher diversity of isolated Gram-negative bacteria in agricultural soil, compared to the water used for irrigation, which is consistent with previous research [21,69]. Although water can be a carrier of microorganisms, and thus a source of contamination, the fact that we have found a relatively low microbial load suggests that the water could be qualified as suitable for agricultural irrigation. However, it would be necessary to promote microbiological control of irrigation water as an additional element of food safety in the study area [70].
At the soil level, the dominant microbial pathogen group was Salmonella spp. This indicates a high contamination rate, with potential microbiological risk for humans, as demonstrated by Johannessen [71] and De Quadros et al. [72]. This bacterial load may be due to the high environmental survival rate of this type of bacteria; this can be for weeks in some types of cultivated plants such as lettuce or on substrates like compost [73]. Balkhair [74] indicates that several bacterial genera, amongst which are those found in our study, can have a high survival in soil, which can represent a pathogenic reservoir. This effect is frequent in soils considered organic, i.e., treated with animal manure (e.g., poultry manure) and compost [21,69,75,76], which generally change the composition of the agricultural soil microbiota [77]. On the other hand, the bacterium Proteus vulgaris, besides being an opportunist that causes urinary tract infections [78], possesses urease activity, which can increase the microbially-induced calcium carbonate precipitation (MIPC), thus affecting calcium availability and increasing soil hardness [79,80]. As mentioned before, all the farms sampled in the current study use animal manure as base fertilizer, applied 3-4 weeks before the planting of vegetables. This type of management may explain the overlap of bacterial species between organic and conventional farms, which are slightly different but non-statistically significant. High conservation of agrobiodiversity has been shown experimentally in small-scale organic farming in Ecuador's highlands [8]. Other fertilizers can be applied afterwards (e.g., N10-P30-K10; urea 46%; ammonium nitrate 36%) and plague control (e.g., metaldehyde; chlorpyrifos), usually between 10 to 15 days, or according to the particular needs of each of the conventional farms. This practice would explain the presence of pathogenic bacteria referenced as more resistant to extreme environmental conditions or chemicals [74,81].
Regarding bacteria isolated from water samples, the genus Shigella with the indicated species is the most common. Their presence is considered as an indicator of fecal pollution, exclusively from human excreta [82]. This genus easily accumulates at the interface formed between water and soil, in channels commonly called "MO deposits" [83]. Species of the genera (Escherichia, Salmonella, Shigella) are considered the main enteropathogens, causing diarrheal diseases in humans through food ingestion, many times coming directly from agriculture [26,29,84,85]. Similarly, Burkholderia mallei of the Burkholderiaceae family is considered pathogenic, causing "glanders" disease (contagious and re-emerging) in horses, donkeys and mules, but with low probability of transmission to humans, except for direct contacts with sick animals and other media [86]. Several genera of the Enterobacteriaceae family, such as Enterobacter, Escherichia, Klebsiella, Salmonella and Shigella, can be found in both soil and water; they include bacteria recognized as part of the microbiota of the digestive tract of animals such as cattle, horses and even humans, which can be natural dispersers to these ecosystems [87]. However, we cannot discard some human contamination due to the work of many farmers there who do not have knowledge about microbiological pollution.
On the other hand, our samples revealed the presence of the bacteria Pseudomonas fluorescens, a natural inhabitant of the phyllosphere, but which can survive and multiply in microhabitats associated with the rhizosphere and rhizoplane and be passively transported, forming biofilms in water flowing through saturated soils [88]. This species is a rootcolonizing bacterium (rhizobacterium) [89], with plant growth-promoting capacity through auxin-like phytohormones such as indole acetic acid [61,90]. In addition, P. fluorescens is a pathogen controller via secondary metabolites, such as siderophores [91,92] or antibioticlike substances (i.e., phenazine, pyrrolnitrin and pyocyanin) [93].
The present study indicates that sampled farmlands are a potential source of contamination and can affect human health, since the encountered type of bacteria has previously been reported by several authors as pathogenic [63,[93][94][95][96][97]. The found frequency and the CFU measurements (≤300) of isolation suggests that the risk of contamination of cultivated plants is a random process mostly associated to the specific manure in each farm, regardless of whether there are complementary organic or conventional management activities. Consequently, it is advisable to perform a microbiological quality analysis on these vegetables to establish acceptability for human consumption [72,98].
Probably some tests for solubilization and fixation of phosphorus, silicate, zinc and nitrogen or evaluation of the production of auxin (indole-3-acetic acid) or volatile organic compounds could be developed to evaluate our strains because some that are considered potential pathogens for humans are also important diazotrophic endophytes or PGPR, alleviating salinity stress and promoting growth in plants [39,41,[45][46][47]54,59].
External environmental factors, chemicals and, most importantly, the introduction of foreign bacteria to the natural microbiota can upset the balance in the functionality of the natural microbiome by means of antagonistic reactions, parasitism, action of phytopathogens or natural competition for space [99]. Therefore, it is essential to maintain stability through the diversity ratio between "beneficial" and "harmful" microorganisms, especially in ecosystems such as those analyzed in this study.

Conclusions
The isolated Gram-negative bacteria determined in agricultural surface water and soil shows mostly the presence of enteropathogenic bacteria of animals and humans, along with other bacteria that, functionally, could be beneficial for improving nutrient availability in agricultural soil. The culture media as applied here does not allow isolate the whole Gram-negative bacteria diversity present in both types of samples, but allowed isolate the rare bacterium Burkholderia mallei.
Further studies are needed to improve the microbial culturability and understand the influence of organic and conventional management on the diversity of the natural microbiota of agricultural soil. Meanwhile, it is recommended that processes of maturation or previous fermentation of manure are implemented, which reduce the presence of pathogenic microorganisms.  Acknowledgments: The authors would like to thank Diana Szekely for their critical language revision and some advice on writing. The authors are very grateful to the student Diego Piedra and many other students belonging to the Biological School and Biochemistry and Farmacy School who helped during sampling and laboratory work.