Novel purification reagents for the study of the human microbiome

Nucleic-acid-based techniques such as hybridization, PCR, quantitative PCR and next-generation sequencing offer a rapid and highly sensitive alternative to culture-based techniques for detecting pathogenic bacteria directly in specimens. However, in addition to the inherent limitations of amplification and identification of biological samples, the human genome itself may interfere with pathogen detection and diagnosis, owing in part to the higher proportion of human genomic DNA that is present relative to the target microbiome. As a consequence, analyses of a metagenome or microbiome from clinical samples using next-generation sequencing or PCR are inefficient, difficult and time-consuming.

We have developed a unique method for the separation of large pieces of human DNA (about 20 kb) from similar sizes of Escherichia coli DNA (about 20 kb) using methyl-binding protein 2a (MBD2a) fused to the Fc portion of a human antibody heavy chain (MBD2a-Fc). This MBD2a-Fc protein is expected to bind to a hydrophobic protein-A-coated magnetic bead that has been engineered to show almost no or no non-specific DNA binding. The bacterial/human DNA is added to the beads and incubated for 15 minutes, and then a magnetic field is applied to capture MBD2a-Fc matrix bound to methylated DNA (for example, human DNA). The majority (about 95%) of the human DNA, which is CpG methylated, is expected to remain bound to the magnetic bead matrix, whereas the bacterial DNA, which is generally CpG methylation free, remains in the supernatant. The recovery rate of the input microbial DNA is >95%.

The supernatant, which is enriched for microbial DNA, is fractionated by sonication or restriction enzyme digestion. A second reagent, CXXC-Fc, an unmethylated CpG moiety binding protein, further concentrates the bacterial DNA fraction. We have established methodologies whereby these reagents can be used to decode entire microbiomes in a cost-effective manner using existing next-generation sequencing platforms and newer single-molecule sequencing technologies. Sequence analysis from Ion Torrent data will be presented. The availability of large microbiome datasets will allow us to identify unique markers specific to bacterial species, as well as single nucleotide polymorphism targets that are associated with normal and disease states. Additionally, we plan to develop a rapid in situ solid phase platform using the above reagents to lyse, enzymatically fractionate, purify and concentrate minuscule quantities of bacterial DNA from blood or other bodily fluids from mammalian species. We believe these new purification reagents will have a major impact on the broader biomedical community interested in studying the human microbiome and host-pathogen interactions.