X-ray crystallography is the gold standard for determining the structure of proteins with atomic resolution. Current methods are labor intensive and tend to falter with large protein complexes and membrane proteins. Automation of the structure determination process is in its infancy, and a central roadblock is the difficulty of creating the highly ordered protein crystals that are needed for x-ray crystallography. Part of my research has been focused on developing a new paradigm for crystal growth and analysis, and using this paradigm to solve the structures of a set of evolutionarily selected glycerol kinase mutants.
We used microfluidic large-scale integration to develop a set of tools that enable a rational approach to crystal growth. A "formulations" chip allows one to measure the phase behavior of a protein over thousands of conditions and to create phase diagrams for the most promising conditions. With these phase diagrams in hand, it is possible to create a "rational" set of crystal growth reagents. We use these reagents to screen for crystal growth in a microfluidic chip; the chip takes advantage of the novel fluid physics available at the nanoliter scale to favorably manipulate the crystal nucleation and growth kinetics. The most promising conditions are exported to a "growth chip," which scales up the reaction and allows one to take diffraction data in situ, so the crystals never suffer mechanical damage from handling.
Transcriptional regulatory networks are vital to cellular function. Cells react to intra- and extracellular perturbations by transducing signals to transcription factors that specifically bind to cis-regulatory elements, thereby inducing or repressing target genes. Identifying cis-regulatory elements and the emergent transcriptional regulatory networks has become one of the outstanding challenges of systems biology. A number of powerful approaches, both informatic and experimental, have been applied to this problem with varying degrees of success. We combined systems biology with high-throughput microfluidic assays to measure the complete topography of transcription factor–DNA-binding energy for several transcription factors.
To do so, we developed a method to detect transient binding events based on the mechanically induced trapping of molecular interactions (MITOMI), which eliminates the off-rate problem facing current array platforms and allows for measurements of absolute affinity. MITOMI was used to map the binding energy landscapes of four eukaryotic transcription factors belonging to the basic helix-loop-helix (bHLH) family by collecting more than 41,000 individual data points covering titrations over 464 target DNA sequences. These binding energy topographies allowed us to (1) predict in vivo function for two yeast transcription factors, (2) make a comprehensive test of the base additivity assumption, and (3) test the hypothesis that the basic region alone determines binding specificity of bHLH transcription factors.
A major challenge facing environmental science is the identification of microbial species that catalyze or may be capable of catalyzing important activities in situ. Nucleic acid–based studies have proved quite effective in assessing species and metabolic diversity and potential in microbial communities. PCR-based techniques that use single genes as proxies for organisms or key microbial activities continue to provide valuable insights into the diversity of a microbial community. It has been difficult, however, to interrelate PCR-derived gene inventories to derive correspondences between any two or more specific genes of interest.
Metagenomics, in which the genomes present in a microbial community are pooled, processed into clone libraries, and sequenced, allows for broad and comprehensive examinations of the genetic potential of organisms present in the environment. Unless the microbial community is dominated by one or a few species, however, complete sequence coverage, much less individual genome reconstruction of the resident genome(s) via computation, remains a challenging task. A common characteristic of both PCR-based and metagenomic studies is the use of homogenized, whole-community genomic DNA extracts that result in irreversible mixing of genes and genome "shrapnel." The cell as a distinct informational entity is lost, as is, by extension, any concept of the organism.
We are approaching this problem via two routes. First, we have employed microfluidic digital PCR to amplify and analyze multiple genes obtained from single bacterial cells harvested from nature. A gene encoding an enzyme involved in the mutualistic symbiosis occurring between termites and their gut microbiota was used as an experimental hook to discover the previously unknown ribosomal RNA–based species identity of several key symbionts. Second, we are using microfluidic devices to isolate and amplify the genomes of individual microbes from complex environmental samples. These genomes are then analyzed by high-throughput sequencing. An important part of this work has been the development of new sequencing technologies.
Nanoliter-Scale Synthetic Chemistry
Molecular imaging probes are an important and growing class of chemical compounds for biology and medicine. In conjunction with PET (positron emission tomography) imaging, the identification of molecular imaging targets and the development of new radiolabeled molecular probes for those targets are crucial for expanding the capability of in vivo molecular imaging for biological research, molecular diagnostics, and drug discovery. The United States already has a vast network of cyclotron production sites in place as convenient sources for radionuclides (e.g., [18F]fluoride, [11C]CO2, and [11C]MeI). The ability to provide a broad range of radiolabeled molecular probes needed in various applications is therefore only limited by the ability to generate probes rapidly, in high radiochemical yields, on demand, with flexible structural diversity, and at low cost. The synthesis of the [18F]-labeled molecular imaging probe 2-deoxy-2-[18F]fluoro-D-glucose ([18F]FDG) in an integrated microfluidic chip was chosen as a proof of principle since it is the most widely used radiolabeled molecular probe, with more than a million patient doses produced in 2004. The short half-life of [18F]fluorine makes rapid synthesis of doses essential, and the synthetic process includes common steps required in many chemical syntheses. The nanogram mass of PET molecular imaging probes administered to subjects is ideal for miniaturized architecture of integrated microfluidics.
We developed a microfluidic device and used it to implement the multistep synthesis of [18F]FDG in doses large enough for animal imaging. We are currently developing the ability for nanoliter-scale synthesis of a variety of biological molecules.