The new B. thuringiensis isolates were obtain from uncultivated and cultivated soil samples from the Papaloapan region. The cultivated soil samples were from sugarcane, coffee and banana crops. The geographic location and coordinates of the soil sampling is show in the Electronic Supplementary Material, Table S1. Papaloapan region is a humid tropical region with annual average conditions of 70% of humidity, 33˚C and 3000 mm of precipitation. Five samples of 500 g at 15 cm deep of soil per crop (four corners and the center) were collected. The samples were stored at 4˚C for 24 h.

The isolation method was design considering the principal characteristics of B. thuringiensis which is forming an insoluble and crystalline protein (Cry protein). The method consist them in two steps: 1) colony and microscopically morphology and 2) production of an insoluble protein at the end of the submerged culture. In order to compare the isolates B. thuringiensis var. kurstaki HD-73 (ATCC-35866), which produces a 133.3 kDa Cry1A(c) insecticidal crystal protein, was used as a reference. The isolates were phylogenetically identified comparing the molecular chronometer 16S rRNA from the isolates with 16S rRNA from B. thuringiensis strains reported in databases.

Steps 1: Colony and Microscopically Morphology. From the collected sample serial dilutions were done with sterile distilled water. The 1 × 10⁻³ dilutions were inoculated by triplicate in nutrient agar and were incubated at 30˚C for 24 hours. Colonies were selected by their morphology looking for white to off-white color, opaque, slightly raised elevation, and regular outlined colonies [10] . The selected colonies were inoculated in serial subcultures of nutrient agar until pure cultures were obtained. After that, the selected isolates were inoculated in 250 ml of nutrient broth by duplicate at 30˚C and 180 rpm during 48 h. At eight hours of culture (during the exponential growth) samples were collected for Gram staining, looking for gram positive bacteria. At 24 hours of culture (during the sporulation phase) samples were collected for spore formation staining. The stains were analyzed with a microscope under dark field illumination, with a 40× objective and 10× ocular lenses. It is important for a proper Gram staining of B. thuringiensis to take the sample during the exponential growth; otherwise, according to our experiments, the bacteria could be stained as a gram negative. At the end of the cultures the broth were collected for Cry protein detection.

Step 2: Detection of Insoluble Protein. The Cry proteins are insoluble at neutral pH, they are soluble just under alkaline conditions [9] [11] . The detection of Cry protein was done by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Bovine serum albumin (BSA) was used as protein standard. Considering the insolubility of Cry protein, the culture broth were centrifuged in 50 ml corning tubes at 5580 g during 30 min. The pellets were resuspended in 2 ml of distilled water, concentrated in a 15 ml corning tube and centrifuged at 5580 g during 30 min. The samples preparation and the solubilization process were done using the methodology described by Navarro et al. [12] . Four replicates were done for each sample.

The selected isolates were inoculated in Gerry Rowe culture medium, and incubated in orbital shaker (Thermo Scientific) at 30˚C and 200 rpm during 48 h. At eight hours of culture, 5 ml samples were centrifuged at 11,180 g during 15 min at 4˚C. Genomic DNA was extracted using the Ultra Clean microbial DNA isolation kit (MOBIO) following the manufacturer’s recommendations. DNA was quantified using a nano-spectrophotometer (Nanodrop 2000, Thermo Scientific).

The PCR amplification of 16S DNA gene was done using 100 ng of genomic DNA, 10 pM fD1 primer (CCGAATTCGTCGACAACAGAGTTTGATCCTGGCTCAG) and rD1 primer (CCCGGGATCCAAGCTTAA GGAGGTGATCCAGCC) [13] . 16S DNA PCR was performed in a MaxyGene (Axygen Scientific^®) thermal cycler under the following conditions: 95˚C 5 min; 35 cycles of 95˚C for 1 min, 55˚C for 40 s and 72˚C for 30 s; finally, 72˚C for 10 min. The PCR amplification of nprX gene was done using the primers NprR-1 (GGGCAT TTGTTCTGTCTC) and NprR-2 (GCTAACACTAACGCTAAAC) [4] for NprR-NprX in 50 μL reaction (DreamTaq kit ™ 2X Green PCR Master Mix, Fermentas). nprX gene PCR was performed in a MaxyGene (Axygen Scientific^®) thermal cycler under the following conditions: 95˚C 5 min; 35 cycles of 95˚C for 1 min, Ta 1 min and 72˚C for 1 min; finally, 72˚C for 10 min. The amplifications were observed with an imaging and analyzing system INGENIUS SYNGENNE^®. The PCR products were purified using Thermo Scientific GeneJET PCR Purification Kit. The sequencing service was performed in Macrogen Inc. Korea. The nucleotide sequences obtained in this study were compared with sequences retrieved from databases in a pairwise mode, with the BLAST2 sequences tool (http://www.ncbi.nlm.nih.gov/BLAST/bl2seq/wblast2.cgi). The identification criteria proposed by Pothathil and Lazzazzera [6] , and Perchat et al. [3] were used to identify the isolates putative signal pentapeptide sequence.

Likelihood trees of 16S rRNA gene and nprX gene nucleotide sequences were built for isolates bioinformatic analysis. For phylogenetic tree based on 16S rRNA, the sequences from 29 isolates and 42 sequences from B. cereus group strains were included. For phylogenetic tree based on nprX gene, 24 isolates and 33 sequences from B. cereus group strains were included (Electronic Supplementary Material Table S2 and Table S3). The B. cereus group strains sequences were obtained from the GenBank database of the NCBI (http://www.ncbi.nlm.nih.gov/genbank) and IMG database (Integrated Microbial genomes, http://img.jgi.doe.gov). The phylogenetic tree based on 16S rRNA gene was built under the Statistical Method Maximum Likelihood. The Substitution Model used was Tamura Nei [14] . The phylogenetic tree based on 16S rRNA gene topology was optimized with the nearest neighbor interchanges (NNI). The phylogenetic tree based on nprX gene was built under the general time reversible (GTR) substitution model using Phyml [15] , while the topology was optimized with the nearest neighbor interchanges and sub-tree pruning and regrafting approaches. Statistical support was determined by 10,000 bootstrap replicates. The phylogenetic trees based on 16S rRNA and nprX gene were built using MEGA 6.0 software and CLUSTALW analysis.

The results of the isolation method are shown in Table 1. From colony and microscopically morphology 45 isolates were selected as B. thuringiensis; from those, 41 isolates produce an insoluble protein between 60 and 130 kDa (91%). Considering just colony morphology, approximately 63% of the isolates produce an insoluble protein. Therefore, the combination of colony and microscopically morphology increase the possibility to isolate correctly B. thuringiensis strains.

For isolates identification three consecutive phylogenetic trees were built. The first one was built using 25 reported sequences of 16S rRNA gene: 19 of B. cereus group strains (10 for B. thuringiensis, 6 for B. cereus, 3 for B. anthracis), 4 of Bacillus subtilis (B. subtilis), 1 of Bacillus licheniformis (B. licheniformis) and 1 of Escherichia coli (E. coli), Figure 1. This phylogenetic tree shows three groups, one with all the strains of B. cereus

group, the second one with B. subtilis and B. licheniformis and the last one with E. coli.

A ClustalW analysis [16] using the isolates 16S rRNA gene sequences and the sequences used for the first tree with the MEGA 6.0 was done. With this alignment the second phylogenetic tree was built using the Tamura Nei model [14] (Figure 2). The second phylogenetic tree shows two groups, one with the B. cereus strains and

the other one with B. subtilis. All the isolates sequences are closer to the B. cereus group (Figure 2). The 16S rRNA gene sequences of 16 isolates (A-8, A-6, A-37, A-36, A-33, A-32, A-31, A-30, A-3, A-28, A-27, A-26, A-22, A-15, A-14 and A-11) have a comb-like configuration. Through the alignment of these sequences, it was observed that all the sequences are identical. Therefore, they were considered as one group (A-X group).

Considering the results of the first and the second phylogenetic tree, the third one was built (Figure 3). For the third phylogenetic tree 22 sequences of B. thuringiensis strains were added (Electronic Supplementary Material Table S2). In this third phylogenetic tree (Figure 3) it can be observed that isolated AC1 presents similarity with B. thuringiensis alesti; as well as AC2 with B. thuringiensis HD-771, AC3 with B. thuringiensis higo and A-13 with B. thuringiensis IAM 12077. Besides, A-9 and A-24 show similarity not only with B. thuringiensis IAM 12077 but also with different varieties of B. thuringiensis such as chinensis, oswaldocruzi, kenyae and thompsoni. Isolates AC5, AC6, AC7 and AC8 show similarity with B. thuringiensis and B. cereus. The A-X group share information with B. thuringiensis (ostriniae, seoulensis and BMB) and with B. cereus (ATCC 14579). The isolated A-34 shows similarity with B. thuringiensis tolworthi. The isolated A-35 and A-5 show similarity with the group of B. thuringiensis, B. cereus and B. anthracis species. Table 2 shows the results from

the phylogenetic identification of the isolates from the Papaloapan region. For AC-5, AC-6, AC-7, AC-8, A-5 and A-35 isolates the phylogenetic analysis is not conclusive, therefore, they were identified as Bacillus spp. However, all of these isolates produce an insoluble protein at the end of the culture.

The phylogenetic tree based on nprR gene nucleotide sequence built on 29 strains isolated and 32 strains of B. cereus group from the literature (Figure 4, Electronic Supplementary Material Table S3) show a perfect correlation of seven putative mature NprX heptapeptide sequences [3] . Analysis of the NprX peptide was performed using the software SignalP 3 [15] , which incorporates a prediction of cleavage sites and a signal peptide/non- signal peptide prediction based on a combination of several artificial neural networks and hidden Markov models. Alignment of the secreted region was done using ClustalX to identify putative mature signaling peptides. The putative mature NprX for isolated strains are show in Table 2.

In phylogenetic tree based on nprR gene (Figure 4) different putative mature NprX can be observed: SKPDIVG for 15 B. thuringiensis strains, 3 B. anthracis strains and one B. cereus strain; WTSDIVG for 5 B. thuringiensis strains and 4 Bacillus spp. strains; SRPDVLT^** for 2 strains of B. thuringiensis, one of B. cereus and two of Bacillus spp.; WKPDVLG^* for 2 B. thuringiensis strains; WKPDTLG^* for 2 B. thuringiensis strains and one B. cereus strain; SNPDIY^** for one B. thuringiensis strain and one B. cereus strain; SKPDTYG for 2 strains of B. cereus.

The phylogenetic tree based on nprR gene also shows the type of Cry protein produced for the reported strains (Figure 4, Electronic Supplementary Material Table S3). No relationship was found between the Cry protein type produced and the putative mature peptides of all strain studied.