Advances in biomarker discovery, synthetic biology and next-generation genomics promise to make designer cells a reality in the near future. Such cells can be engineered to report on diverse aspects of their physiology, potentially opening up new avenues of biological inquiry and enhancing our understanding of cell function.

My first exposure to experimental science required microinjecting antibodies into fibroblasts and then asking if the delivered antibodies had any effect. I was looking specifically for defects or delays in mitotic progression, which was the focus of my advisor's work. Several months and several thousand injections later, in the basement of Dick McIntosh's lab at the University of Colorado at Boulder, I came across an interesting antibody that (after an additional 3 years of experiments) we found recognized a human mitotic kinesin (1). The best way to master microinjection was (and still is) practice, practice, practice. During one of these practice sessions, it was not uncommon for me to clog a needle; accidently aspirating some nucleoplasm led to the same frustration every time. One day, when yanking back on the spoiled microneedle, something remarkable happened and I dragged the entire genome out of the nucleus at one fell swoop. Staring through the microscope, it was inspiring to think here was a single genome, and if one could simply figure out a way to decode this gummy mess, individual genomics would be a reality. Fast-forward 17 years to the post-genomic age, and we now know the sequence of several human beings, a few cancers, and a growing menagerie of creatures: some living, and some long dead, like our Neanderthal cousins and the woolly mammoths that inhabit many museums.

These are indeed heady times for biologists, and it is not hyperbole to say that if you can imagine an experiment, the technologies exist to pull it off. We now have an opportunity to extract not simply some information, but every last bit/base of data from a single nucleus. By combining many recent innovations, we can design a cell that faithfully reports its complete genomic sequence. These self-reporting cells are the logical extension of biomarker biology; great strides have been made to develop cell lines that express reporters that inform, for example, their cell cycle state (2), their degree of stress (3), or their state of differentiation (4). Such self-reporting cells will make genomic interrogation as straightforward as going to your computer's control panel icon and asking for the details of its operating system.

The motivation behind this innovation derives, in part, from the efforts of Ronald W. Davis (Stanford, University), Mark Johnston (Washington University at St. Louis), and Guri Giaever (University of Toronto) (as well as other members of the yeast deletion consortium) who decided in the late 1990s to delete each gene in the model cell, Saccharomyces cerevisiae (5,6). This was an innovative idea, but early on, a question arose. How could one keep all of the 24,000 strains straight? This was shaping up to be a book keeping nightmare. Swap a plate, introduce an off-by-one error, and you'll sink the experiment and perhaps sacrifice a graduate student's career. Drs. Davis and Johnston had the inspiration, based on earlier work using DNA to mark bacterial strains (7), to include a DNA barcode that would be unique to each of the strains. This idea was ahead of the technology at the time. The scale of DNA synthesis throughput required for the effort had yet to be developed, but this is what Dr. Davis had always excelled at—building technology to enable previously unimaginable biology. So they cobbled together several 96-well oligonucleotide synthesizers (a technology that, at the time, was thought to be too ridiculously high-throughput to be of any use), and synthesized the barcodes (oligonucleotides) required to construct the strains. These collections have been used by hundreds of labs both in clonal and competitive assays over the past decade.

The barcode story serves to illustrate the power of knowing, with certainty, the identity of an individual cell within a mixed population. We are not going to barcode every cell in the human body; even if one could there are biological complications of differentiation, for example. So a reverse engineering approach is required: rather than add a barcode, can we directly sequence each cell, so that its genome becomes the ultimate barcode? I suggest that self-reporting cells that can sequence themselves and faithfully report their sequences to the investigator cheaply and quickly would be a fantastic development for basic research and an aid for bettering human health. We actually have all the tools available now. So the innovation is not really an invention but rather a new application of existing tools.

These tools include: (i) microfluidics devices to isolate individual cells (8); (ii) biomarkers (e.g., antibodies to cell-surface antigens to select particular cell types); (iii) whole-genome amplification protocols of single cells (8,9); and (iv) next generation sequencing technologies (for a review see Reference (10), a technology that proudly wore the crown of method of the year in 2007 (11) and continues to advance at a dizzying rate with new applications seeming to arrive weekly.

So imagine, if you will, that a sample of cells arrives in the lab (or clinic), and it needs to be sequenced now, whether it be to determine tissue compatibility of a donor organ, or drug responsiveness of a dying patient. We load the cells in a flow cell coated with the relevant biomarkers, route that cell to a reaction chamber, and flow in the reagents to amplify its genome. That DNA is then exported to my next-generation sequencer. If my sequencing is of the single molecule variety, I'll get the data in a few hours and the surgeon can make the call.

The technology in practice

As you arrive at the doctor's office and finish checking in (via retinal scan?) you hand her your genome on a memory stick. Next, she takes a milliliter of blood and injects it into a microfluidic card that has channels coated with a variety of cancer biomarkers and collects a few samples from exit ports of the device. To these samples, she adds a “sequencing virus,” which will retrofit your cells into self-sequencing machines. Inside the virus is a chimeric protein that contains a portion of the nuclear lamina α-helical domain and the catalytic site of the phage 29 isothermal polymerase. This chimera, once expressed via a doxycycline-induced promoter, will take up residence within your nucleus, tethered to the nuclear periphery to aid imaging. In addition to the chimera, the virus also brings in a polymerase silencer, a doxycyline-repressed 30-mer stretch of DNA that folds over the active site of the polymerase. Upon adding doxycyline, the polymerase is expressed and a short repressing DNA aptamer is shut off. Co-delivered with the virus is a liposome that carries the fluorescent phosphate-labeled deoxynucleotide fuel that will provide the energy for the sequencing reaction (12). These converted cells are then deposited on a micropatterned slide, a miniature confocal microscope (13) is attached, and the sequencing reaction commences. Given current sequencing rates of 5 nucleotides per second, your data and diagnosis will be ready for download the next day.

The potential of this technology is not difficult to imagine. Knowing the genetic identity of a patient's sample or cells derived from experiment will allow the most direct linking of phenotype to genotype and from hypothesis to diagnosis.