Characterization of Genetically Modified Microorganisms Using Short- and Long-Read Whole-Genome Sequencing Reveals Contaminations of Related Origin in Multiple Commercial Food Enzyme Products

Abstract

Despite their presence being unauthorized on the European market, contaminations with genetically modified (GM) microorganisms have repeatedly been reported in diverse commercial microbial fermentation produce types. Several of these contaminations are related to a GM Bacillus velezensis used to synthesize a food enzyme protease, for which genomic characterization remains currently incomplete, and it is unknown whether these contaminations have a common origin. In this study, GM B. velezensis isolates from multiple food enzyme products were characterized by short- and long-read whole-genome sequencing (WGS), demonstrating that they harbor a free recombinant pUB110-derived plasmid carrying antimicrobial resistance genes. Additionally, single-nucleotide polymorphism (SNP) and whole-genome based comparative analyses showed that the isolates likely originate from the same parental GM strain. This study highlights the added value of a hybrid WGS approach for accurate genomic characterization of GMM (e.g., genomic location of the transgenic construct), and of SNP-based phylogenomic analysis for source-tracking of GMM.

Keywords: Bacillus velezensis; SNP phylogenomic analysis; food enzyme; genetically modified microorganism (GMM); hybrid genome assembly; whole-genome sequencing.

Figure 1

Assembly graphs of Unicycler, hybridSPAdes, and Canu assemblies of GM isolate Pilsner1-2. (A): Unicycler assembly of Pilsner1-2, with the pUB110 sequence on the plasmid contig in the assembly highlighted in blue, and the contig representing the putative plasmidic prophage highlighted in green. (B): HybridSPAdes assembly, with the pUB110 sequence inside the chromosome scaffold highlighted in blue, and the contig representing the putative plasmidic prophage highlighted in green. (C): Canu assembly with the three contigs that display recombinant plasmid concatemers framed in blue. (D): Kablammo visualization of the largest of the three contigs carrying concatemers of the Canu assembly. The top ruler represents the plasmid pUB110, and the bottom ruler represents the contig, containing 8 complete copies of the recombinant plasmid.

Figure 2

Map of the recombinant 6756 bp plasmid carried by the GM isolates. aadD: kanamycin and neomycin resistance gene, ble: bleomycin resistance gene, rep: replicase, mob: mobilization protein. A part of the mob gene is absent compared to the original pUB110 vector, and is replaced with a recombinant insert of 2385 bp in length, highlighted in blue, encompassing a complete protease encoding gene, highlighted in orange, which is also present in the chromosome of the isolates, as well as in the chromosome of B. velezensis 10075. Figure created with Circos 0.69-6 [38].

Figure 3

Alignment to reference read of ~44 kb, originating from isolate Crystal-1, displaying a two-copy integration of the recombinant plasmid. The alignment shows all raw long reads of all isolates (as labeled in the first column ‘Annotations’) that match pUB110 as well as chromosomal sequence upstream or downstream of the sequence used as insert in the transgenic construct, which carries the wild-type protease (hereafter referred to as ‘insert’). These reads are thus expected to cover the junction of chromosomal integration of the recombinant plasmid. If a supplementary alignment was reported, its connection to the primary alignment is shown via a thin grey line. The primary, or representative, alignment is the highest scoring alignment for a particular read. If the read does not align in its entirety, parts of the read sequence may be clipped, i.e., removed from the alignment. In some cases, (mostly) non-overlapping parts of the read may align at different positions to the reference. In that case, the non-primary alignment(s) is/are referred to as supplementary alignment(s). Primary and supplementary alignments that are inverted compared to each other are marked in red (forward alignment compared to the reference) and blue (reverse complement alignment). Clipped regions are hidden, and, if they are >30 bp, flagged with a red mark. For clarity, secondary alignments (same part of the read from the primary alignment, but aligning at another location to the reference), mismatched bases (variants) and short indels are hidden. Green vertical lines mark the edges of the first and third copy of the insert, which is present three times in the reference, as a consequence of the double plasmid integration event that is represented by the reference. The blue vertical lines mark the end of the first, and beginning of the second, integrated pUB110 copy in the reference read, with in between another (third) copy of the insert. If a continuous alignment extends beyond two adjacent green lines, it supports a chromosomal integration of at least one pUB110 copy. These reads are marked with a green arrow. If the alignment extends beyond both blue lines, in addition to two adjacent green lines, it supports the presence of (at least) two contiguous integrated pUB110 copies. These reads are flagged with a blue arrow. Figure created with IGV.

Figure 4

SNP-based phylogenetic tree, visualized with FigTree (github.com/rambaut/figtree, 29 October 2021), with associated SNP addresses. Both for SNP phylogeny and SNP typing, the Unicycler assembly of isolate Pilsner1-2 was used as reference. The scale is expressed as average substitutions per site in the SNP matrix. Node values represent bootstrap support values.