**4. Conclusions**

*Ecosystem and Biodiversity of Amazonia*

are resistant to ampicillin, vancomycin, gentamicin, kanamycin, streptomycin, erythromycin, clindamycin, tetracyclines, and chloramphenicol [4, 7]. Another study has reported that Lactobacilli isolated from fermented olives were resistant to cephalosporins, streptomycin, and kanamycin [41]. Overall, the observed pattern is for high variability or diversity in the response to antibiotics as there is considerable dispersion in the response within and among bacterial species or plant hosts (**Figure 6**). However, there were also isolates with unique resistance patterns, which will require a further inquiry into their molecular and physiological properties. Undoubtedly, a full safety assessment with a robust identification of the strains and an in vitro evaluation of the potential risks is needed; particularly if these are

*Non-metric multidimensional scaling (NMDS) that represents, in two dimensions, the distances among isolates and host plants according to the observed pattern of antibiotic resistance. a) Tile for the spatial relationships in NMDS space between antibiotics and the studied bacterial isolates. b) Tile for spatial relationships in NMDS space between antibiotics and the host plants of the studied bacterial isolates. Short distances among these objects in the canonical space represent strong associations of either resistance or susceptibility. The antimicrobial spectrum is shown here as a cofactor showing a directional trend. AMX25 was removed from this analysis as it was not informative (i.e. susceptibility as a zero-response constant). The colored areas facilitate the interpretation of the antibiotics position.* L. fermentum *and* L. plantarum *were used as a reference in the tests.*

It is known that some LAB strains produce a wide variety of anti-pathogenic compounds, like bacteriocins, ethanol, organic acids, diacetyl, acetaldehydes, hydrogen peroxide (H2O2), and peptides [42, 43]. When we analyzed the antimicrobial spectrum against ten Gram-negative and Gram-positive bacteria, including closely related species and pathogens such as *Salmonella enterica, Shigella sonnei, Escherichia coli, Enterobacter, Staphylococcus aureus,* we observed that the isolates showed high inhibitory potential, as none had values below six and were defined as broad-spectrum (**Figures 3** and **4**). The inhibitory effect of LAB strains may result from a combination of competition for metabolic substrates, growth suppression by organic acids, and bacteriocin secretion. Recently, we showed that some of the Lactobacilli strains inhibited *Salmonella enterica subsp. enterica* ATCC51741 and *E. coli* ATCC25922 at both the early and logarithmic stages of bacterial growth *in vitro* and *ex vitro* [44, 45, 46]. Also, we showed that one selected LAB strain from the

intended to be used as additives in food products.

**26**

**Figure 6.**

The Amazon rainforest is a sizeable reservoir of plants, animals, and bacterial diversity. For Ecuador, the Amazon region could be a significant source of new bioproducts, based on the transformation of biodiversity [47]. Subtropical wild fruits have a relevant ethnobotanical significance, as they are mostly consumed by indigenous people as food or natural medicine; however, the bacterial microbiota of those fruits has not been assessed. In this research, we investigated the lactic acid bacteria diversity associated with several wild fruits collected from the Amazon region of Ecuador. Their remarkable inhibitory potential towards Gram-negative bacteria might be related to their capacity to produce various antimicrobial substances, that when applied to food products might prevent the growth of undesirable microorganisms. A better understanding of the metabolic capacity of these microorganisms will further complement our knowledge about the development of a novel starter or preservative culture for fruit- and vegetable-based foods. The prospective comparative exploration of the genomes of LAB strains from various plant or fruit origins would be of particular interest to provide information on their adaptations to different food-matrices and to further explore biotechnological applications.

Genotype-functional correlation studies contribute to the discovery of new biotechnological properties for several species. The results from the present study supported our hypothesis that LAB strains from wild fruits of the Amazon Region of Ecuador carry noteworthy characteristics that could be inherent to their ecological niches or environmental origin and that could be developed for biotechnological applications. Several strains were found capable of producing antimicrobials with high inhibitory potential against commensal and spoilage bacteria and are promissory natural food preservatives.
