High-throughput genome scanning in constant tension fluidic funnels.

Genome Sequence Scanning (GSS) is a bacterial identification technology that detects sparse sequence-specific fluorescent tags on long DNA molecules linearized in a continuous-flow microfunnel. The efficiency and sensitivity of GSS depends on the detection throughput of well-stretched molecules. Previous studies have investigated the fundamental roles of elongational and shear flow on DNA stretching in continuous flow devices. Here, we present a novel compound funnel design that significantly improves effective molecule throughput over previously described devices. First, exploring the relationship between fluid strain rate and molecule tension allows for design of funnel geometries that accommodate high fluid velocities without compromising molecules due to overstretching under high tension. Second, a constant-strain detection channel is utilized to significantly reduce the number of molecules lost to further analysis due to shear-induced molecular tumbling and relaxation. Finally, application of the constant-strain detection channel allows for a priori prediction of spatial resolution bias due to accelerating flow. In all, the refined funnel geometries presented here yield over thirty-fold increase in effective molecule throughput due to increased fluid flow and improved retention of stretched molecules, compared to previously described devices.

Fast high-resolution mapping of long fragments of genomic DNA based on single-molecule detection.

Here we describe bacterial genotyping by direct linear analysis (DLA) single-molecule mapping. DLA involves preparation of restriction digest of genomic DNA labeled with a sequence-specific fluorescent probe and stained nonspecifically with intercalator. These restriction fragments are stretched one by one in a microfluidic device, and the distribution of probes on the fragments is determined by single-molecule measurement of probe fluorescence. Fluorescence of the DNA-bound intercalator provides information on the molecule length. Because the probes recognize short sequences, they encounter multiple cognate sites on 100- to 300-kb-long DNA fragments. The DLA maps are based on underlying DNA sequences of microorganisms; therefore, the maps are unique for each fragment. This allows fragments of similar lengths that cannot be resolved by standard DNA sizing techniques to be readily distinguished. DNA preparation, data collection, and analysis can be carried out in as little as 5h when working with monocultures. We demonstrate the ability to discriminate between two pathogenic Escherichia coli strains, O157:H7 Sakai and uropathogenic 536, and we use DLA mapping to identify microorganisms in mixtures. We also introduce a second color probe to double the information used to distinguish molecules and increase the length range of mapped fragments.

Staphylococcus aureus strain typing by single-molecule DNA mapping in fluidic microchips with fluorescent tags.

BACKGROUND: Epidemiologic studies require identification or typing of microbial strains. Macrorestriction DNA mapping analyzed by pulsed-field gel electrophoresis (PFGE) is considered the current gold standard of genomic typing. This technique, however, is difficult to implement because it is labor-intensive and difficult to automate, it requires a long time to obtain results, and results often vary between laboratories. METHODS: We used direct linear analysis (DLA), which uses a single reagent set and long fragments of microbial genomic DNA to identify various microbes. In this technique, an automated system extracts fragments exceeding 100 kb from restriction enzyme digests of genomic DNA from microbial isolates and hybridizes them with a sequence-dependent fluorescent tag. These fragments are then stretched in a microfluidics chip, and the patterns of the distribution of the tags are discerned with fluorescence confocal microscopy. The tag pattern on each DNA fragment is compared with a database of known microbial DNA sequences or with measured patterns of other microbial DNAs. RESULTS: We used DLA to type 71 Staphylococcus aureus strains. Of these, 9 had been sequenced, 10 were representative of the major pulsed-field types present in the US, and 52 were isolated recently in a hospital in Cambridge, MA. Matching DNA fragments were identified in different samples by a clustering algorithm and were used to quantify the similarities of the strains. CONCLUSIONS: DLA-based strain typing is a powerful technique with a resolution comparable to macrorestriction mapping with PFGE, but DLA is faster, more automated, and more reproducible.