MICROBIAL ARSENIC TRANSFORMATIONS IN CONTAMINATED SOILS

The work presented in this thesis explores the potential role of bacteria in the arsenic cycle an agricultural soil from Scarlino (Tuscany), and a soil from Torviscosa (Fiuli), pyrite cinders contaminated. Both soils contained high levels of total arsenic. The soil of Scarlino contained 250 mg kg-1, mainly associated to iron oxides (70% of the total), while the labile forms were 10% of the total. Total As in the soil of Torviscosa was 446 mg kg-1, mainly present in the residual phase (50% of the total), while the labile forms represented 15% of the total. This thesis started to investigate the presence of rhizobacteria associated to Cirsium arvense (L.), a wild plant present in the soil of Scarlino, with the aim to isolate and characterize As-resistant bacteria with potential plant growth promoting characteristics and potentially involved in As transformations. FISH analysis of the rhizobacterial community of C. arvense revealed the presence mainly of Alphaproteobacteria, Betaproteobacteria and Gammaproteobacteria. Several isolates mainly belonged to Bacillus, Achromobacter, Brevundimonas, Microbacterium, and Ochrobactrum genera and resulted highly resistant to As(III) and As(V). In most of these strains ArsC, ArsB and ACR3 genes, peculiar to the ars operon of the arsenic detoxification system have been detected. The ArsB genotype predominated over the ACR3, highlighting that the ArsB gene family is extensive; and the ArsB-type efflux pumps seemed to be predominant in Firmicutes and in Betaproteobacteria, and the ACR3-type in Alphaproteobacteria. Several isolates reduced As(V) and oxidized As(III). In particular, Ancylobacter dichloromethanicum strain As3-1b was able to transform both forms of As. Moreover, this strain was able to oxidize As(III) also under chemoautotrophic conditions. Numerous As-resistant isolates possessed plant growth-promoting characteristics, being able to produce siderophores, IAA, and ACC deaminase. For this reason, these isolates could potentially support plant growth in As-polluted soil and reduce stress symptoms. From a greenhouse experiment conducted on the soil from Scarlino with sunflower plants bacterized with an As-resistant PGPR strain, it was evidenced a synergistic role of the PGPR strain and the sunflower plants in As uptake. Moreover, the As-resistant strain colonization of the rhizozphere of sunflower was monitored by quantifying ACR3(2) gene, coding for the As(III) efflux pump. The results of the quantification of this gene evidenced that the strain colonized the sunflower rhizosphere. As second part of this thesis we have investigated the effect of the addition of a C source on the fate of As and of the microbial population in the soil from Torviscosa, when is flooded. Moreover as the soil contained Fe and Mn oxides, we have investigated also their solubilization. Our results have confirmed that the presence of a readily utilizable carbon source for microorganisms can enhance microbial As solubilization. The two C sources (glucose and citrate) used in the experimental microcosms, differently influenced the solubilization of As and Fe, as well as the bacterial community. Solubilization of As and Fe resulted differently affected either in term of amount released or in term of kinetics of solubilization. In fact, citrate promoted two-fold higher solubilization of As and Fe, compared to glucose. This can be due to the effect of citrate on microbial community and to the wide and complex interactions among citrate and the dissolved ions occurring in a soil. In fact, citric acid, being a chelating agents, can modify the solubility of elements, such as iron and manganese. In addition, we found that citric acid influenced copper solubilization, thus increasing the ecotoxicological risk deriving from the flooding of multipolluted soils. Different kinetics of solubilization of As and Fe under glucose and citrate were observed, confirming the complexity of the relationship between As and Fe release. By adding glucose to the soil, Fe solubilization was not concomitant to that of As, thus suggesting that desorption rather than dissolution was the main mechanism controlling release of arsenic from pyrite cinders. In glucose-amended soil, the As(III) formation was governed by biological processes. The ability to reduce As(V) and the presence in soil DNA of genes belonging to the ars operon and of arrA genes confirmed that bacteria reduced As(V) by means of a detoxification systems and/or by a metabolic processes. In the presence of citrate, As(III) was the main species present in solution at oxic conditions, possibly due to the detoxifying activity of As(V) reducing bacteria; while As(V) was predominant at reducing condition. This was possibly due to the concomitant iron oxides dissolution that could have liberate As(V) occluded into Fe oxides. Diverse bacterial populations were enriched by glucose or citrate. In fact, glucose stimulated populations of Flavobacterium and Paenibacillus, while citrate incremented those of Bacillus, Pseudomonas, Clostridium, and Geobacter. As-resistant bacteria were isolated either from glucose- or from citrate-amended microcosms, and with some of them we demonstrated their reducing capacities. Most strains possessed the ars genes, but not arrA genes. Quantification of arsC and arrA genes performed on soil microcosms revealed the constant presence of arsC in all microcosms, while arrA genes became evident only in glucose-amended microcosms, suggesting that the type of C differently stimulated the As-resistant microbial communities. The two functional genes arsC and arrA can be used a reliable biomarkers for detection of As-resistant bacteria responsible of As(V) reduction in contaminated soils. In particular, regarding arrA the method proved to be sensitive to detect a very low copy number (10 copy number g-1 soil). In the final part of the thesis we have investigate the effect of milled alfalfa, a more complex C source, on As solubilization in the soil, under flooding condition. Beside microcosms experiment, a greenhouse experiment was also set up using the soil amended with milled alfalfa, and Salix purpurea (L.), in order to investigate As uptake and translocation in the plants. The solubilization of arsenic observed in submersion was mainly related to biological induced redox modification. However, the reductive conditions observed both in the microcosms and in the greenhouse experiment were not sufficient to allow Fe oxides dissolution and only a negligible release of Fe occurred. For this reason, it can be hypothesized that As mobilization was due to reductive desorption from solid surfaces rather than to dissolution of Fe oxides, as in the case of glucose. Interestingly, solubilization of As in flooded soil during the greenhouse experiment was lower than that in microcosms soils, suggesting a direct or indirect effect of S. purpurea in the control of As level in the aqueous phase. From the results of As content in the leaves of willow, it possible to note that As accumulation capability of willows was low. Nevertheless, the plants demonstrated a good attitude to grow in a high contaminated soil subjected to flooding. For this reason, phytostabilization with willow could be a possible strategy in soils subjected to flooding.