Fruits of Guatemalan avocado were purchased from a retail market of Ibarra city (capital of Imbabura province, northern Ecuador) in two ripening stages: an immature firm (AF) recognized as light green fruit (ready to eat in 4 days) and mature intense green (AM) with soft skin fruit (ready to eat). The fruits with no visible damage (5 fruits × 2 stages × 3 repetitions: total 30 fruits) were taken and transported to the laboratory for further analyses. The fruits were washed twice with tap and distillate water before further manipulation. In addition, a bulk of peel and pulp (CPE) from visibly spoiled fruits (5 fruits × 3 repetitions: total 15 fruits) were used for comparison. The cutting utensils (knife, spatula) and board were surface-sterilized with 70% ethanol to avoid cross-contamination. Figure 1 shows an overview of the workflow used in the analysis process.

The bacteriological analysis was performed as previously described (Tenea et al., 2023). In brief, the exocarp (AFPE and AMPE samples) was gently removed, chopped into small pieces, and mixed, and 25 g/sample was inoculated in pre-enrichment buffered peptone water (0.1%), homogenized, and incubated for 4 h at 37°C. Similarly, 25 g pulp/sample (AFPU, AMPU) was used. In addition, following the same procedure, CPE sample (visibly spoiled) was analyzed. After incubation, decimal dilutions made with sterile water were inoculated on 3M Petrifilm Aerobic (3M Science Applied to Life, Detroit, MI, USA), to determine the total aerobic microbial population (37°C, 48 h). To detect and differentiate the presumptive presence of SALM/SHIGA, aliquots (100 μl) were plated on Shigella–Salmonella (SS) agar (Difco, Detroit, MI, USA) and incubated for 48 h at 37–40°C. The presence of SALM was confirmed as previously described (ISO, 2002). Independent experiment aliquots (100 μl) were placed on Chromocult Coliform agar (Merck Millipore, Kenilworth, NJ, USA), to determine the total coliforms and EC, and eosin methylene blue (Difco, Detroit, MI, USA), to detect the presence of ENT and EC. In addition, 3M Petrifilm Yeast and Mold Count Plates (3M Science Applied to Life, Detroit, MI, USA) were used for the enumeration of YM (incubation at 25–28 °C for 7 days). The presence of STAPHY was determined in Brilliance Staph 24 agar medium (Oxoid Limited, Wade Road, Basingstoke, Hampshire, UK) (ISO, 2018), whereas the presence of LIST was evaluated on GranuCult FRASER broth (Merck, Millipore, Kenilworth, NJ, USA) (ISO, 2017). The experiments were run in triplicate, and the microbial counts were expressed as logCFU/g.

The pH was determined using a pH meter (SevenCompact S210, Mettler Toledo LCC, Columbus, OH, USA). Total acidity, expressed as a percent of citric acid, was determined by titrating with 0.01 M NaOH to pH 8.2 as described (Tenea et al., 2023). The total soluble solids (°Brix) was evaluated using a digital refractometer to determine the total sugar in fruits.

A total of 52 random colonies from both peel and pulp (8–10 colonies/stage/indicator) were screened for antibiotic susceptibility according to the Kirby–Bauer disk diffusion procedure (CLSI, 2021) against the following panel of antibiotics: amoxicillin (AMX: 25 μg), ampicillin (AM: 10 μg), gentamicin (CN: 10 μg), kanamycin (K: 30 μg), tetracycline (TE: 30 μg), and cefuroxime (CXM: 30 μg). E. coli ATCC25922, S. enterica subsp. enterica ATCC51741, L. monocytogenes ATCC19115, Enterobacter spp. UTNEnt1 (a laboratory AR strain isolated from strawberry), S. aureus ATCC43300 (methicillin sensible), and S. aureus ATCC1026 (methicillin-resistant) strains were used as references. The microbiological breakpoints reported by the FEEDAP standards were used to categorize the clones as susceptible, intermediary, or resistant (EFSA, 2012). The Scan500 was used to determine the inhibitory halos automatically (Interscience, Fr). The percentage of resistance was determined as the number of total bacteria resistant/number of total isolates tested. The MAR index was calculated as the ratio between the number of antibiotics that an isolate is resistant to and the total number of antibiotics the organism is exposed to.

The 16S rRNA gene sequencing was used for the taxonomical assignment of the selected MR clones (9) and ENT (2) clones with a MAR index above 0.50 (Macrogen Inc., Korea, custom service). The PCR was performed using 27F 5' (AGA GTT TGA TCM TGG CTC AG) 3' and 1492R 5' (TAC GGY TAC CTTGTT ACG ACT T) 3' primers (Weisburg et al., 1991). The PCR reaction was carried out with EF-Taq (SolGent, Korea), with the following protocol: activation of Taq polymerase at 95°C for 2 min, 35 cycles of 95°C for 1 min, at 55°C and 72°C for 1 min, and extension of 10 min at 72°C. The amplification products were purified on a multiscreen filter plate (Millipore Corp., Bedford, MA, USA). The sequencing reaction was performed using a PRISM BigDye Terminator v3.1 cycle sequencing kit. The DNA samples containing the extension products were added to Hi-Di formamide (Applied Biosystems, Foster City, CA). The sequencing was conducted using 785F 5' (GGA TTAGAT ACC CTG GTA) 3' and 907R 5' (CCG CAA TTC MTT TRA GTT T) 3' primers, which are the inter-primers of 16S RNA V3 region (Muyzer et al., 1993). The mixture was incubated at 95°C for 5 min, followed by 5 min on ice, and then analyzed by an ABI Prism 3730XL DNA analyzer (Applied Biosystems, Foster City, CA). A fast homology search of the sequences was conducted using the megablast algorithm against the 16S ribosomal RNA database at NCBI (http://www.ncbi.nlm.nih.gov/BLASTN), as implemented in Geneious Prime 2020.2.3 (Kearse et al., 2012). This first search was used to obtain a maximum of 100 hits and associated search quality parameters that provided an initial reference for taxonomic classification. A final taxonomic assignment was made by the RDP Bayesian classifier algorithm (Wang et al., 2007) with 100 bootstrap replicates and a K-mer of size 8, as implemented in the function “accurate, high-resolution sample inference from the amplicon sequencing data” (assigned taxonomy) of the DADA2 package (Callahan et al., 2020).

The results revealed ~56,228 and 58,111 clean reads for AFPE and AFPU, while 53,104 and 45,236 reads were found in AMPE and AMPU, respectively (Table 1). In total, 99.85% of reads for AFPE/AMPE and 99.94% for AFPU/AMPU samples were classified as bacteria. Alike, 60,296 reads (99.78%) were classified for CPE sample. At the phylum level, Cyanobacteria, Acidobacteria, and Proteobacteria were the most abundant among the groups. The total phylum-level taxonomic categories identified were 33 for AFPE and AFPU, 35 for AMPE, 28 for AMPU, and 29 for CPE. The top 10 abundant bacterial community structures at the phylum level are shown in Figure 2A and Supplementary Table 1. A major abundant class of “chloroplast bacterial genome” was observed in AMPU and CPE samples. The plastid genome contains a bacterial genome signature explaining the high abundance of the Cyanobacteria phylum (Robinson et al., 2022). In addition, Firmicutes were the most abundant in AFPE (2.26%), while Bacteroidetes were less abundant in the AMPU (0.15%) samples (Supplementary Table 1). Early metagenomic studies indicated that Proteobacteria, Acidobacteria, Bacteroidetes, and Firmicutes were the most abundant phyla in grapes (Zarraonaindia et al., 2015), while Firmicutes, Actinobacteria, and Proteobacteria were more abundant in melon pulp Cucumis melo L. (Glassner et al., 2018). In addition, the abundance of Proteobacteria in watermelon was linked to the ability of the fruit to utilize a wide variety of carbohydrates, amino acids, and lipids (Xia et al., 2015). The relative abundance and the top 10 categories at the genus level are shown in Figure 2B and Supplementary Table 2. A total of 770 and 404 reads were categorized for AFPE and AFPU samples, whereas 401 and 381 reads were categorized for AMPE and AMPU samples. A total of 422 reads were identified in CPE samples. To detect changes in bacterial community composition and visually compare the overall absolute abundance among the groups at the phylum and genus levels, heatmaps and hierarchical clustering were developed (Figures 3A, B). The AMPE and AMPU were grouped into the same clade with CPE samples, suggesting that the mature and firm fruits differ in bacterial composition. This may correlate with the differences in the physicochemical parameters with the high value of total solids detected in the mature stage (Supplementary Table 3). Most likely, this microenvironment allows bacteria and other germs to invade, adapt, and survive during storage. According to the findings, Gp15, a clade of unidentified reads, and Streptophyta, a clade of plants, were the most abundant categories. Recent studies investigating the hospital airborne microbiome indicated that Gp15, Pseudomonas, Staphylococcus, Corynebacterium, and Acinetobacter genera can be responsible for different types of nosocomial infections (Perrone et al., 2022). All sample groups contained pleomorphic bacteria, such as those generally associated with Staphylococcus; however, AMPU and AFPE samples showed 0.60% relative abundance of these species. Streptomyces, the most widespread and perhaps most important genus of Actinomycetes, was found in AFPE samples. The species of this genus are a good source of bioactive substances, antibiotics, and extracellular enzymes (Olanrewaju and Babalola, 2019). Although low abundant (0.80%), E. coli and Shigella were detected in AMPE samples (Supplementary Table 2). Additionally, using the Venn diagram, 20 (19.8%) and 18 (31%) families were shared between peel and pulp, respectively (Supplementary Figure 1). While Enterobacteriaceae were found in all samples (firm, mature, and spoilage fruits), bacteria belonging to the Marinilabiliaceae, Vibrionaceae, and Helicobacteraceae families were detected in AMPE. A larger sample size and additional research are required to confirm these findings and provide a biological context; hence, no precise role could be assigned. At the species level, the AFPE group showed the highest Shannon index diversity of 2.22 (1,327 species), while the lowest species diversity (1.21) (363 species) was detected within the AMPE group (Table 1). AFPE had the highest number of designated species, and at this point, no potential human pathogens had been found. The relative abundance of reads assigned at the species level is shown in Figure 4. Staphylococcus strepanovicii was prevalent in AMPE and CPE samples (relative abundance of 0.12%), whereas S. gallinarum, S. warneri, and S. pasteuri were prevalent in AMPE samples (relative abundance of 0.07, 0.13, and 0.12%, respectively) (Supplementary Figure 2). S. strepanovicii was mainly associated with mammals. S. gallinarum spp. are opportunistic human pathogens that have primarily been identified in poultry (Shi et al., 2015). S. warneri was detected as an endophyte in the peel of apples from major commercial markets in Tamil Nadu, India (Phukon et al., 2013). In addition, in less abundance (0.12%), Weissella oryzae was detected in the AFPE and AFPU samples, while Lactobacillus fermentum was detected in AFPE samples. Thus, the 16S metagenome study is the first to provide a diagnosis of bacterial diversity in ripe and unripe Guatemalan avocados; nonetheless, the composition of fungal communities associated with postharvest avocados is recommended. These results may contribute to a better understanding of the microbial composition of harmful bacteria, which will aid in the development of safety measures to prevent their spread before selling. The microbial 16S rRNA sequences were deposited in the Sequence Read Archive (SRA, https://www.ncbi.nlm.nih.gov/sra/) under accession code PRJNA972543 (15 May 2023).

In this study, bacteriological analysis revealed a high content of indicator bacteria in the mature ripe stage than in the immature (firm) stage (Table 2). The highest number of total aerobes was detected in AMPE (3.62 ± 0.18 logCFU/g), AMPU (3.68 ± 0.37 logCFU/g), and CPE (3.85 ± 0.38 logCFU/g) samples, suggesting that during shifting from immature to mature stage, these fruits become susceptible to contamination. Nonetheless, the increase in microbial population could be part of fruit natural flora. Interestingly, EC was detected in AMPU only, while SALM and SHIGA were detected in AMPE and AMPU samples. Early research indicated the presence of Salmonella and Shigella species in avocado mature fruits collected from the market of Northern Nigeria (Shiferaw and Kibret, 2018). In addition, Coetzee et al. (2017) reported the presence of E. coli and Salmonella spp. in avocado fruits collected from Southern Nigeria. Moreover, the highest amount of STAPHY and LIST was detected in mature (AMPE and AMPU) and spoiled (CPE) avocado fruits (Table 2). Recent research indicated a high frequency of occurrence (29.2%) of S. aureus in spoiled avocado fruits obtained from three market locations in Sokoto city, Nigeria (Aliero et al., 2022). In addition, avocado fruits purchased from retail markets in Guadalajara (Mexico) contained E. coli, Salmonella spp., Listeria spp., and L. monocytogenes, according to previous studies (Shiferaw and Kibret, 2018; García-Frutos et al., 2020). Avocados are regularly sold in bulk in retail markets in Ecuador, where they are exposed to potential sources of pathogens while being kept at outdoor temperature for several days. More likely, these pathogens may adhere, survive, and form biofilms under these circumstances. Although there is no proof of how these bacteria can reach the pulp, we suspect that during handling, germs may be transported to the pulp (comestible part) from the spoon used to remove the peel. While yeasts were not detected, high content of molds was found in the AMPE, AMPU, and CPE samples (Table 2). The differences in physicochemical properties may be responsible for the divergence in the prevalence of some microbe species in these fruits. Altogether, our analysis supported earlier research (García-Frutos et al., 2020; Aliero et al., 2022; Cabrera-Díaz et al., 2022), linking the fruit contamination with storage conditions, the use of microbiological hazardous containers, poor handling techniques, and unsanitary market conditions.

Antibiotic resistance is a global issue (Ayandele et al., 2020). Currently, there is substantial proof that the improper management of contamination has resulted in a severe problem with MAR (Catalano et al., 2022). The percentage of AR resistance among STAPHY, LIST, ENT, and SALM clones isolated from avocado fruits is shown in Figure 5. In addition, the prevalence of AR among isolates from various classes of antibiotics, such as tetracycline (TE30), aminoglycosides (K30, CN10), cephalosporins (VAN30), beta-lactamases (AM10), and penicillin-like antibiotics (AX25, MET5), is shown in Supplementary Tables 4–7. This research revealed that out of 21 selected STAPHY clones, only one was methicillin sensitive but resistant to four different antibiotic classes (Supplementary Table 4). Additionally, 13 clones were resistant to vancomycin. Similarly, among LIST, 11 clones were resistant to vancomycin and 6 to tetracycline (Supplementary Table 5). Among ENT clones, 9 were resistant to kanamycin (Supplementary Table 6). Although only 4 SALM clones were detected in mature fruits (AMPE and AMPU), they were resistant to amoxicillin (Supplementary Table 7). Moreover, the MAR index was calculated for each clone (Supplementary Tables 4–7). The MAR index is an efficient, reliable, and effective method for locating the sources of antibiotic-resistant bacteria (Davis and Brown, 2016). According to previous studies, a MAR >0.2 denotes a source of contamination with significant risk (Davis and Brown, 2016). However, this study found that 100% of STAPHY clones, 66.67% of LIST clones, and 75% of ENT clones had MAR indices >0.25 (Figure 6). In addition, complementary hemolysis and gelatinase assays indicated that 100% of the selected STAPHY clones showed beta-hemolysis and were positive for the presence of enzyme gelatinase (data not shown). Previous research indicates that hemolysin and gelatinase from Staphylococcus are significant virulence factors with cytotoxic actions (Bertelloni et al., 2021).

The 16S rRNA sequencing of the nine selected STAPHY clones showing a MAR index above 0.63 was identified as S. xylosus (clones A1 and A2), S. saprophyticus (clones A3, A4, and A8), Mammaliicoccus sciuri (clone A10), and S. epidermis (clones A5 and A11). Clone A6 presented no match in the GenBank database. S. xylosus is a ubiquitous species that lives on the mucous membranes and epithelium of animals, especially mammals (Nagase et al., 2002). Although this species is defined as non-pathogenic, a few strains were related to animal opportunistic infections (Siqueira and Lima, 2002; Dordet-Frisoni et al., 2007). BLASTN analysis against the 16S ribosomal RNA NCBI database indicated a 98.83% sequence identity of the A1 clone with S. caeli strain 82B (NR_180106.1) isolated from air sampling of an industrial rabbit held in Italy (MacFadyen et al., 2019). A2, A3, A4, and A8 clones showed 99.42% to 99.85% sequence identity with S. saprophyticus subsp. saprophyticus ATCC 15305 (NR_074999.2) isolated from urine, S. edaphicus strain CCM8730 (NR_156818.1) isolated in Antarctica (Pantuček et al., 2018), and S. casei strain SB72 (NR_037053.1) isolated from ripened cheese (Place et al., 2002). In addition, S. saprophyticus was found as a food contaminant with a high prevalence of 34% in beef and pork meat (Hedman et al., 1990). Clone A10, found in mature pulp, showed 100% sequence identity with M. sciuri strain DSM 20345 (NR_025520.1), a mesophilic human pathogen that was isolated from the skin of eastern gray squirrels (Jayne et al., 2015). Former known as S. sciuri, M. sciuri belongs to the novel order Mammaliicoccus and is a commensal animal-associated bacterium but is also found in food (Marino et al., 2011; Madhaiyan et al., 2020). These strains showed 97% sequence similarity with M. sciuri strain B9-58B isolated from retail pork meat products (Neyaz et al., 2020). Interestingly, clone A5 showed 100% sequence identity with S. epidermis clone D02 (GU003840.1) isolated from active sludge (Parsley et al., 2010), whereas clone A11 showed 98.11% identity with S. epidermis clone 2322 (MT604781.1). Additionally, two ENT clones (AMPE clone 14 and AMPU clone 15) showing the highest MAR index (0.83) were identified as Enterobacter sp. with 96.44% and 99.37% sequence identity with Enterobacter bacterium strain 35 (KY681875.1) found in forest mushrooms (Pent et al., 2017). We stipulate that the presence of these microorganisms may be related to storage facilities as the fruits are deposited directly on soil or wet cellars where small animals can cross through, as well as human manipulation, but this statement needs to be supported by further analysis. These isolates were deposited at the NCBI Gene Bank database with the following accession numbers: OQ372998 (S. xylosus FMCShyA1, https://www.ncbi.nlm.nih.gov/nuccore/2439391154), OQ372999 (S. xylosus FFCShyA2, https://www.ncbi.nlm.nih.gov/search/all/?term=OQ372999), OQ373000 (S. saprophyticus FFCShyA3, https://www.ncbi.nlm.nih.gov/search/all/?term=OQ373000.1), OQ373001 (S. saprophyticus FFCShyA4, https://www.ncbi.nlm.nih.gov/search/all/?term=OQ373001), OQ876755 (S. epidermis dFMCShy5, https://www.ncbi.nlm.nih.gov/search/all/?term=OQ876755), OQ373002 (S. saprophyticus FMCShyA8, https://www.ncbi.nlm.nih.gov/search/all/?term=OQ373002), OQ373003 (M. sciuri FMCShyA10, https://www.ncbi.nlm.nih.gov/search/all/?term=OQ373003), OQ876756 (S. epidermis FMCShy11, https://www.ncbi.nlm.nih.gov/search/all/?term=OQ876756), OQ876757 (Enterobacter sp. dFMCEag14, https://www.ncbi.nlm.nih.gov/search/all/?term=OQ876757), and OQ876758 (Enterobacter sp. dFMCEag15, https://www.ncbi.nlm.nih.gov/search/all/?term=OQ876758).