That single-cell biology is one of the biggest new trends in the field comes as no surprise to Michelle Khine, an assistant professor in the school of engineering at the University of California, Merced. “We are seeing more and more the need to do single-cell measurements because traditional bulk-population methods output only the mean response of the cells. This population average could be very different from the actual response of each cell,” she says. “In biomedical research, a lot of times, we’re interested in the outliers’ behavior.”
That interest has finally been met with technology sensitive and precise enough to manipulate and interrogate individual cells, drastically changing the types of questions scientists can ask. “People have been, for hundreds of years, grinding up organisms … and measuring average concentrations across that population,” says Norm Dovichi at the University of Washington. “Now that technology is maturing, we are able to characterize single cells.”
In this field, the chance to look at something new might as well be a double-dog-dare. Scientists with expertise in a range of technology areas from RNAi to sequencing are racing to apply their tools and wrest new secrets from lone cells. At the Sidney Kimmel Cancer Center, Robert Margolis and his team have figured out how to inject siRNA into a cell to silence a gene while simultaneously inserting a second plasmid bearing a rescue protein to restore proper function to the cell. At UC Merced, Khine is using technology she helped develop to perform electroporation on individual cells. Meanwhile, Kun Zhang in George Church’s lab has figured out how to amplify DNA from one cell and is currently working to couple that with next-gen sequencing technology to enable true single-cell genomic analysis.
On the pages that follow, you’ll meet some of the scientists who are getting this field up and running. Today, their technologies tend to be homemade, but from the response in the field, it’s a safe bet that in the not-too-distant future these single-cell tools will be coming to a lab near you.
Single-Cell Pioneer Builds New Tools, Tracks Proteins
Norm Dovichi is one of the leaders and pioneers of the single-cell movement. A faculty member at the University of Washington for the past six years, Dovichi leads a 15-member lab that’s contributing technology and protocols to enable the community at large to get a peek at what’s going on inside individual cells.
Dovichi says his interest in the field stems from a postdoc he completed at Los Alamos National Laboratory, where he “managed to spend half my time with the flow cytometry group.” Trained as a laser spectroscopist, Dovichi was impressed with the flow cytometry technology and says that “ever since then I have been looking for ways to parlay that experience” into his own research.
“The focus of analytical chemistry, certainly since 1980, has been to develop techniques to study smaller and smaller [components],” he says. Now that scientists can follow the activities of a single cell, they can track what happens to a particular protein in a particular cell. Talk about small.
In Dovichi’s lab, where several technological advances have enabled scientists to watch cells on this up-front-and-personal level, there are now a few projects underway to use these tools to probe biological questions. One revolves around characterizing the protein content of individual cells. The goal is to “essentially generate the equivalent of an SDS PAGE” and map a single cell, he says. In another project, scientists are looking to monitor a cell’s metabolism, one area where Dovichi expects to see a good deal of difference between cells. “These are difficult experiments and you don’t want to do them if they’re not going to be informative,” he says. “We do them where we expect there to be significant cell-to-cell variation.” That includes neurons, whose heterogeneity is well known and a good target for Dovichi’s variety-seeking team. His lab is working to characterize neurons with this technology, and he says immune system cells — likely the most complex set of cells in the body second only to brain cells — will also be a focal point for the group.
A relatively new project in the Dovichi lab involves embryology. Using C. elegans, an organism whose development and cellular structure is very well known, scientists aim to “catalog the evolution of each and every cell” as the nematode grows from embryo phase to adult.
All of the work utilizes technology homegrown in Dovichi’s lab, relying heavily on ultrasensitive, laser-based methods. Currently, the laser-sensitive detector is coupled to an electrophoretic capillary system, and Dovichi says the main challenge is separation issues for complex protein samples. He and the team are hoping to get the single-cell technology already functioning to work in tandem with a mass spectrometer to help iron out those problems.
At this point, the technology has come a long way, so Dovichi is now pushing to open it up to new applications, allowing a whole host of biological questions that scientists haven’t been able to ask.
Yang to Molecules: I’ve Got My Eye on You
No matter how persuasive you may be, there’s simply “no way [to] ask a molecule, ‘Hey, don’t move, we’re trying to watch you,’” says Haw Yang, a scientist at the Lawrence Berkeley National Laboratory. So as an expert in imaging, Yang realized that he was going to need some sort of technological innovation if he was going to be able to perform long-term imaging to see what was going on at the single-cell level.
Yang has long been interested in better understanding complex macromolecules and how they function within cells. “That’s a general area that is relatively unknown,” he says. Questions about molecular activity, conformation changes, and localization within the cell have long puzzled scientists in this field.
Unlike techniques where scientists immobilize a molecule on a substrate, what Yang wanted to be able to do was to choose a particular molecule and then follow it around with a camera, watching its every interaction over time. The problem from an imaging perspective was that these molecules move in three dimensions — how could an onlooking camera keep focus?
The breakthrough was in tricking the imaging technology into keeping an eye on a nanoprobe, which is in turn attached to the molecule of interest. “That’s a completely new type of imaging — a three-dimensional type of imaging with nanoscale precision,” he says, noting that much of the work was performed by Hu Cang and Shan Xu, two postdocs in his lab. Add to that real-time spectroscopy, and you’ve got yourself a molecule that wants to be noticed.
Yang expects the implications of the technology to be fairly broad in range. “It’s not limited to a certain type of problem,” he says. “We think we now have more possibilities than we can handle.”
And Yang’s already thinking of ways to improve the imaging tool. “I have new ideas for the next generation, how to do it better,” he says.
Gene Knockdown Goes One Cell at a Time, Finds Success Every Time
Robert Margolis started off the single-cell work that got him noticed by the Faculty of 1,000 as “kind of a dream project,” he says. Margolis, it would seem, is a man of complicated dreams.
When he first heard about RNA interference, Margolis says, he hit on the idea that if you could go in and suppress a gene, surely you should be able to replace it at the same time. “We began to think about how to engineer a rescue product,” says Margolis, a scientist with the Sidney Kimmel Cancer Center. “If you could introduce the RNAi and be sure that in the very same cell you were introducing the effective rescue plasmid — you’re not replacing the gene, but you’re replacing the gene product.” While RNAi itself has an iffy record, knocking down anywhere from 5 percent to 90 percent of its target product, Margolis says it was widely known that if you could get a cell to accept a plasmid, it would also accept a second. “Of those that accept a plasmid, we found that all will accept a second plasmid at the same time.” This cellular twofer made him realize that on a single-cell basis, he could get RNAi to be effective all the time. It took years and major advances in single-cell technology, but his brainstorm finally paid off.
Margolis had no shortage of challenges in getting his concept into practice. First, he had to demonstrate effective suppression of the unwanted gene, so he needed a reliable phenotype he could follow. Next up: the second plasmid with the rescue product. But that plasmid had to be exquisitely engineered so “the RNAi doesn’t recognize it, but the same protein is made,” he says. Then, to document the success of the experiment, Margolis had to tag the rescue protein to ensure that it can be recognized by “an antibody that does not recognize anything [else] in the cell.”
Successful gene knockdown, carefully engineered undetected second plasmid, uniquely tagged rescue protein — is it any wonder this was years in the making? But when Margolis tested it out, first on an Aurora kinase and later on two additional proteins, “it always works like a charm,” he says. With the Aurora kinase, the N-terminus can be truncated with no functional change, so his team was able to change a few amino acids in that region to create a protein that an antibody wouldn’t confuse with the wild type protein.
The secret lies in engineering the replacement protein. Margolis has demonstrated it in mitotic proteins including survivin and PRC1, and he says the major breakthrough his team found was in using global alignments of proteins from different species to locate the right epitope. In comparing orthologous proteins across species, “what you’re looking for is regions that can be changed at will — they’re not conserved,” he says. “That’s the epitope you want to target in the native protein. Then you can in fact swap in the region from an ortholog protein from another species” to be sure the rescue protein is distinguishable in the cell from the original. The actual detection work is accomplished using highly sensitive microscopy.
The method enables the identification of key mutations in a protein, and allows for far greater flexibility and repetition of experiment than its nearest relative, homologous recombination. “In homologous recombination, for each and every mutation you have to redo [the process],” Margolis says. With the siRNA technique, plasmids are introduced as a cassette and the process can be repeated for “as many mutations as you want.” That’s key for research looking at potential drug binding sites, he notes, adding that the method can be used to prime cells with various mutations and then perform high-throughput screening against the wild type control cell.
“It has simply never been possible before to rapidly do a genetic screen of a spectrum of mutants in a single mammalian cell,” Margolis says. “This is the tip of the iceberg for an amazing amount of things that one can do.”
Ploning Cultivates DNA from Single Cells in the Church Lab
For the past two and a half years, Kun Zhang, a scientist in George Church’s Harvard lab, has been listening to people say that clean amplification of DNA from a single cell “was not possible.”
Zhang may be too polite to say “I told you so,” but the papers he’s published recently are sure to silence the pessimists.
Zhang says the major challenge facing amplification from single cells is, of course, dealing with contamination. In most experiments, trace contamination is the norm — and doesn’t pose any real threat to the collection of high-quality data. But on the individual cell level, even minute amounts of contamination — including environmental DNA or trace contaminants from enzymes, gloves, water, and the like — can be enough to shut down an experiment.
But Zhang came up with a way around that. While some papers trying to solve this problem were reporting that about 30 percent of their amplified product came from the target DNA, his breakthrough was “a very, very clean amplification system” that enabled him to cut background to about 3,000-fold less than a single copy of the E. coli genome.
“Basically what we’re doing is we use a real-time amplification method so we can monitor the amplification kinetic” as it’s taking place, Zhang says. That way, researchers can continually monitor the product for contamination and pull the plug if things get dicey. The process, polymerase cloning or “ploning,” is based on making “clonal copies of the DNA molecule using DNA polymerase instead of traditional cloning,” Zhang says. That avoids the step of inserting the DNA into E. coli and introducing another possible source of contamination.
Putting his method to the test, Zhang amplified a particular lab strain of E. coli and then sent the results off to the Joint Genome Institute for sequencing. He could then compare that to JGI’s sequence of the original strain to see what toll the amplification had taken. He says he was glad he did: as it turned out, the ploning process introduced chimerism, and Zhang has since worked out a Perl code to deal with that.
Currently, Zhang is trying to use the protocol to amplify DNA from highly heterogeneous samples such as ocean water. Generally in such an experiment, about 60 percent of the DNA amplified from the sample would be that of the target; Zhang says that with the ploning process, “we got about 96.5 percent of good reads that mapped to our target genome.”
He’s also working with his lab colleagues to try to connect ploning with next-generation sequencing to “try to do super-deep sequencing on a single cell,” he says.
Total Cell Control from a Bioengineering Background
Michelle Khine was a graduate student at Berkeley when she stumbled onto a great project and wound up helping develop a “polymer single-cell electroporation array” to enable intracellular delivery of anything from drug compounds to DNA. It was a microfluidic platform based on a 96-well plate that allowed scientists to individually control cells, apply electric fields, and electroporate the cells repeatedly, if so desired.
That technology was in fact so promising that the team filed patents on it and started a company called Fluxion Biosciences. Khine completed her PhD, worked with the company for a time, and last year decided to head back to academia, landing an assistant professorship in bioengineering at the newest school in the University of California system, Merced.
Khine says that her lab will continue working in the single-cell arena. “I’ve come to realize that this technology is not just important for electroporation,” she says. One application underway in her lab is studying stem cell differentiation with similar kinds of microfluidic chips. “We’re applying electrical and chemical cues to the cells to try to differentiate them into different cell types,” she says. Also of interest: better understanding cell fate decisions and looking into “how receptors are expressed in response to certain chemokines.”
Though her background is in engineering, biomedical applications of that expertise lured Khine, who felt that such a combination could lead to work that “has the potential impact … to save lives and improve the quality of lives,” she says. That’s not just lip service. In a class Khine will be teaching next year, her students will be designing real chips that can be used as diagnostic tools in third world countries, she says.
From Cytometry, the Western Blot for Single Cells
Studying protein abundance certainly isn’t new. But as John Newman, a postdoc in Jonathan Weissman’s lab at the University of California, San Francisco, points out, you need a solid 1 million to 2 million cells to perform a decent western blot to track protein quantity.
“There’s always some question about how those bulk measurements would play out if you got down to the single-cell level,” Newman says. Now there are some answers.
Newman was part of the team that figured out a way to use fluorescence combined with flow cytometry to measure protein abundance. The goal was to evaluate the variation in that abundance on a cell-to-cell basis. “The relative change between cells can be large,” he says. “The secondary motivation was just to try to understand a little bit about how cells can actively partition proteins.”
Using fluorescent detection turned out to be an especially good avenue for looking at abundance, Newman says, because unlike other technologies, it meant “you don’t have to break open the cell” to see what’s going on inside. The downside was that, due to the single-cell factor, the team had to use a lot of cells to see different proteins. “If you want to look at 1,000 proteins, you have to look at 1,000 samples,” he says.
All in all, the automation and other development work to get the project going took about a year and a half. The original part of the work proved that there was indeed considerable variation in protein abundance between cells; now, Newman and his colleagues are trying to figure out which genes are responsible for those fluctuations. Flow cytometry for single cells is playing a role in that effort as well.