By Meredith W. Salisbury
If you’ve been in this field long enough, you’ve probably heard — or perhaps contributed to — the grumbling by now. Researchers who traffic in plant and animal genomics say they often feel like second-hand citizens in the broader genomics community, arguing that human genome researchers get all the glory.
That’s obviously a matter of judgment. But what is clearly true no matter which side of the line you’re on is that scientists working on plant and animal genomes have been responsible for some of the biggest breakthroughs — technological or otherwise — in the community. RNAi, as you’ll recall, was first recognized in petunias. And of course, the many technological insights coming out of comparative genomics, such as CGH and ROMA, wouldn’t be possible without model organism genomes to compare.
In the following pages, Genome Technology pays tribute to the plant and animal genomics folks with a look at scientists involved in some of today’s most compelling research and innovations. These are mostly in the early stages now, but will one day likely be adopted — as have so many advances from plant and animal researchers — by the human genome field as well. And in an attempt to let you in on what’s next, we also included, where possible, research topics that these scientists consider particularly intriguing.
With more than 2,000 people attending the community’s biggest meeting, Plant & Animal Genome XIII in San Diego this January, there’s obviously far too much fascinating work going on to be able to cover it all here. We narrowed our universe by looking only at research in model organisms. The science presented here is meant to be a sampling, rather than a comprehensive view, of the field.
Life Sciences Institute, University of Michigan
Colocalization to track down protein-protein interactions
Somewhere between the false positives of the yeast two- hybrid method and the false negatives of FRET lie the true answers about protein-protein interaction — or so goes the theory behind Anuj Kumar’s yeast research.
The germs of Kumar’s current work came up while he was a postdoc studying protein localization in Mike Snyder’s Yale lab, he says. Now a research assistant professor at the Life Sciences Institute in Michigan, Kumar has spent the past year getting started on his work. He’s still very much “in the pilot stages,” he says, adding that for now his group has “cloned 60 genes … obviously we’d like to do a whole lot more than that.”
Kumar’s approach to understanding multiple proteins’ localizations — a good indicator of function and potential to interact — begins with building a collection of yeast genes that have been fused to a fluorescent protein, he says. That’s accomplished using “a recombination-based cloning strategy that’s commercially available” that avoids the use of restriction enzymes or ligases. “We’re making these plasmids so that the yeast genes are under the control of their own promoters,” he says. Working with protein pairs, the fluorescent protein is fused to the C terminus of the yeast gene. Using pairs of proteins enables Kumar to co-localize the proteins — “we’re trying to determine if two proteins have the potential to interact,” he says.
Existing strategies for looking at protein-protein interactions have known problems, Kumar says. The yeast two- hybrid approach leads to “a lot of false positives” — generating what may be a comprehensive list, but includes too much garbage to be of real use. “You can also use FRET,” he adds, but “this method is prone to false negatives.” The results from a FRET experiment tend to be high-confidence, but far from comprehensive.
Currently, Kumar’s progress has extended to designing the vectors for the cloning, used for some 60 genes. Studies are still in the earliest phases — to some extent, Kumar is still settling into his post at Michigan’s LSI and gearing up his lab. He’ll be collaborating with Michigan colleagues Bob Fuller, Joel Swanson, and Adam Hoppe on the FRET-based part of his research.
Ultimately, he hopes his work will generate “a high-confidence protein-protein interaction data set,” he says. “The idea there is to complement the other protein-protein interaction data sets which exist in yeast which may be prone to more false positives.” After that, Kumar anticipates applying the same concept to studying pathways in yeast.
Yeast is the chosen vehicle for Kumar’s work partly because of how easy it is to manipulate and how well its genetics are understood, he says. “It’s quite similar to a single human cell,” he adds. “Two-thirds of the genes in humans that have been implicated in diseases have an ortholog in yeast.” Yeast is also stable in both haploid and diploid form, and when DNA is introduced into it, “it goes where it should,” Kumar says. “That’s a pretty powerful tool for a lot of applications.” Still, he notes, there’s nothing specific about colocalizing proteins in yeast, so that type of research could certainly be done on other organisms.
On his radar:
• Protein arrays
• Advances in mass spectrometry to measure protein abundance
• Large-scale applications of RNAi
• Methods of studying post-translational modification
Cold Spring Harbor Laboratory
Methylation, other epigenetic breakthroughs advance plant genomics
Rob Martienssen started working with transposable elements during his Berkeley postdoc in the ’80s and hasn’t stopped since. That dedication has led to new technologies and a fuller understanding throughout the genomics community of plant biology and epigenetics.
By the time Martienssen headed to Cold Spring Harbor Laboratory in 1989, he had already “found an example of a transposable element that, when it became methylated, regulated the gene into which it was inserted,” he remembers. At the time transposable elements were just starting to be explored, and biologists were looking for explanations for their activity. “As a geneticist I wanted to take a molecular approach to it,” he says.
That interest led to a collaboration with Eric Richards, a CSHL member who is now at Washington University, that relied on Arabidopsis for “isolating mutants in DNA methylation,” Martienssen says. From his point of view, a key advantage to using a model organism like Arabidopsis is that ability to isolate mutants, which tend to show up in human genes as well. In the decade or so since then, “we’ve been using those mutants but also many others that have come out of all sorts of screens.”
That work has paid off in spades. Thanks to a joint effort with CSHL’s Dick McCombie that turned up methylation patterns in maize, Martienssen says, “we found … that DNA methylation seemed to be restricted almost entirely to transposable elements. … The mutants that we isolated in Arabidopsis turned out to be master regulators of transposons.” Boiled down, that epiphany meant that seemingly intractable plant genomes could be tackled simply by removing all of the methylated DNA — effectively stripping away all the nongenic regions — and sequencing what remained. That technology was eventually licensed out to Orion Genomics, a St. Louis company he cofounded.
The same concept led to a second technology, based on work with Vincent Colot in Paris: using microarrays to determine DNA methylation patterns. Chips are spotted with “PCR fragments that are tiled across the whole of an Arabidopsis chromosome in 1 KB pieces,” Martienssen says. It was necessary to build their own arrays because commercial versions focus on genes, and the methylation patterns would be lost on those. (That technology has also been licensed to Orion.)
Recently, Martienssen has thrown another new technology at Arabidopsis — chromatin IP, better known as ChIP-on-chip — which has helped him understand histone modification, another epigenetic mark like methylation. “DNA methylation and histone modification … can be rewritten and erased and so on without changing the DNA sequence itself,” he says. “We believe that the way that that happens is through RNA-mediated guiding.”
Now Martienssen’s lab is “looking very intently at the mechanisms by which these regions are targeted [as well as] the long-term effects of epigenetic change.” There are potential effects in chromosome imbalance, he says, and his team is studying the role of small RNAs in that and other regions.
On his radar:
No fowl play, just a new approach to WGS
Even at a production-scale sequencing facility as established as the Genome Sequencing Center at Washington University, there’s still room for innovations around model organisms. Take the chicken, for example: the genome itself was “pretty straightforward,” says Director Rick Wilson — but it proved an ideal testing ground for the first map-assisted whole-genome shotgun sequence done at the genome center.
Like the other organisms sequenced at the major genome centers, chicken went through a lengthy selection process before NHGRI gave the green light for its genome project. The reason chicken is such a good model organism, says Wilson, is that “chicken embryos and mammalian embryos share a lot of similarity in how they develop in terms of their skeletal system and their muscular system.” Additionally, of course, the chicken is essential to the agricultural community, and, as Wilson points out, “if you can better understand chicken health, you can make healthier food.” It also provides for studies on how diseases like avian flu jump species, wreaking havoc the way SARS did two years ago. And from an “evolutionary point of view, chickens are the closest living relatives to dinosaurs,” Wilson adds.
Of course, model organism work is nothing new to Wilson, who points out that the genome center at WashU worked on C. elegans and yeast, the first models to be sequenced. Recently, Wilson’s team added two more fly genomes to its list of accomplishments, and is currently improving the 4x chimp draft it did in collaboration with the Broad Institute. The platypus and the macaque are also underway, and WashU is gearing up to start sequencing the Sumatran orangutan and possibly the lamprey. “The primary reason for sequencing them all is to better understand and annotate the human genome,” Wilson says. “Every gene has evolved at a different rate — that’s why there’s no one perfect model for all of the human genes.”
Chicken, which by measures of evolutionary distance is “in a bit of a sweet spot” for its ability to explain certain human genes, provides a unique comparison point. Sequencing took place on Gallus gallus, the jungle fowl widely agreed to be the best chicken model because it seems to be the progenitor of all modern farm chicken breeds. Because the whole-genome shotgun method does so badly with repeats, WashU researchers decided to build a BAC-based map at the same time sequencing was going on, and “that worked extremely well for chicken,” Wilson says. Sequence production took about nine months, and the draft was released a year ago, with a paper published in December.
On his radar:
• Studies led by Eddy Rubin of the Joint Genome Institute looking at micro-organisms in different environments
University of Notre Dame
Second mosquito genome to aid dengue, immune research
The mosquito, which has garnered attention as a model organism for immune system response, was first sequenced a few years ago — completed in October 2003 — by a team led by Celera Genomics with other participating researchers at TIGR, NIAID, and Sanger. That mosquito, Anopheles gambiae, is the primary vector of malaria.
Step aside, Anopheles. The genome sequence of a cousin, Aedes aegypti, is just about finished, and Notre Dame’s David Severson, academic coordinator of the project, says it’s going to be quite a boon for the field. In addition to the obvious comparative genomics opportunity, aegypti is the main carrier of dengue fever and is expected to provide a wealth of medical information.
Severson, an entomologist and devoted mosquito researcher since the late ’80s, says he finds the implications for human health of his research very rewarding. Much of his work focuses on “understanding at the molecular level, at the gene level, what makes a given mosquito [able] to take up, harbor, and subsequently transmit a pathogen from one vertebrate host to another.”
He remembers a seminal moment in the progress of the mosquito community: it was around 1990, and a group of “fairly distinguished mosquito researchers … got together and made a decision that genomics was an important issue in terms of where the field should go,” he says. Existing remedies for handling mosquito-related health risks all had problems: vaccine development wasn’t as promising as some had hoped, and antimalarial therapeutics were facing challenges with drug resistance. “Everything that’s been put out there has some problem,” Severson says. In what seems like a prescient move, the mosquito community hung its hopes on genomics — the idea being to use “genetic manipulation of the mosquito vector itself to prevent it from transmitting [pathogens].”
The major push Severson has headed up, getting aegypti’s genome sequenced — his lab developed the mosquito stock and produced the RNA to make ESTs for the project — is almost complete. “We’re in the process of trying to deal with all that information,” he says. The mosquito community is hard at work figuring out what genes are present, annotating the genome, and making sense of the biology, he adds. Having the genome sequence will likely have a perk: from what he’s seen happen in other organism communities, the sequence availability will likely “bring a lot of expertise into our field from people working on other systems,” Severson says. Comparative work with Anopheles is an obvious early step for the community, and he looks forward to being “able to take advantage of microarrays” once the sequence is in place.
And now genetic manipulation is well within reach. Making transgenic mosquitoes has become a fairly routine process, Severson says, which “does create the opportunity to start testing the effect of either knocking out particular genes or adding foreign genes.” Ideally, such work could prevent mosquitoes from passing on pathogens like malaria or dengue, possibly by upregulating the insect’s immune system and training it to kill those pathogens upon entry.
Model legume provides test bed for integrated functional genomics
His organization jump-started the project to sequence the genome of model legume Medicago truncatula, and now biochemist Rick Dixon is hip-deep in functional genomics studies to try to make sense of it all.
It was in 2001 that the Samuel Roberts Noble Foundation, an agricultural research and funding group, issued a $5 million grant to start work with the University of Oklahoma’s Advanced Center for Genome Technology for the sequence of Medicago, a model for major crops like soybean and alfalfa, and a smaller-scale crop of its own. Since then, Medicago has gone on to have its full genome sequenced by a consortium that includes Oklahoma, the University of Minnesota, and TIGR, among others. Researchers have produced microarrays for this model plant, and almost 95 percent of the spots have proven to correspond with alfalfa, says Dixon, director of the plant biology division at the Noble Foundation. In orthologous genes, that rate can be bumped up to 100 percent.
Information from the genome sequence is bound to be used by all sorts of groups. “If you look at the Medicago community in general,” Dixon says, “different people are plugged into it for different reasons.” One popular scientific reason is Medicago’s model for nitrogen fixation, a critical function that has been hard to study because not all plants do it.
Dixon and his group are using Medicago to look at the chemicals and compounds it produces — including a variety implicated in human health needs, such as antioxidants, flavonoid compounds, and isoflavones. “We’ve been utilizing the genomic information to get at the genes that make these complex natural products,” he says. Those products have ramifications for human diseases that range from cardiovascular disease to cancer, he adds.
Genomics has radically altered how scientists look at the chemical products of Medicago. “In the old days we used to disappear into the cold rooms for six months and try to purify all these things out,” Dixon says. With a sequence on his side, these days he instead builds a bare-bones model of the molecule in which he’s interested, and systematically adds to it, noting the enzymes that play a role. The main goal of the work is to eventually use this information to better understand “the pathways for health-promoting phytochemicals and then manipulating those in plants” to make healthier food for animals and people, he says.
Dixon and his lab are currently finishing up an NSF grant project focused on an integrated functional genomics approach to Medicago. Starting with a single cell and adding various chemical stimuli to coerce the cell into making different products, Dixon’s team has been working on measuring changes at the transcriptome and metabolome level throughout this. Adding proteomic measurements is underway. “I don’t think many people are doing that right now,” he says.
On his radar:
• The need for high-throughput structural biology. According to Dixon, “If you have a gene sequence, you still can’t predict what the function of the gene will be from [it]. … We might be able to predict these things if we had the structure.”
Fred Hutchinson Cancer Research Center
Point mutations, SNP discovery with TILLING tool
A plant genome breakthrough from Steve Henikoff’s lab has proven so popular that, in addition to more than four papers since its earliest version in 1999, demand has been so high that Henikoff has taken to running workshops and pilot projects for other researchers interested in bringing TILLING — or Targeted Induced Local Lesions IN Genomes — to their labs.
TILLING got its start in 1999 with grad student Claire McCallum, who was studying DNA methylation in Arabidopsis when she realized she could put mutations into the plant’s genome. That led to a method for routinely inducing mutations in plants to study point mutations. After four years of development, the process has become so robust that Henikoff has started a TILLING service for the Arabidopsis and maize communities and is starting one for Drosophila.
Henikoff starts with the gene sequence of interest and then mutagenizes the organism — he began with Arabidopsis — with chemical mutagens. Then he adds a cleaving enzyme and follows that by making “primers that will amplify up a region of interest, [usually] 1.5KB,” he says. When he runs an electrophoretic gel, he can find which strands have bulged and come apart from the enzyme activity, indicating the sequence mismatches that were found. The process runs on a Li-Cor system, and Henikoff estimates that his team screens about 1 million base pairs with each run (he generally runs it twice daily).
“There’s major demand for it,” Henikoff says of TILLING. “It’s not very easy to get mutations in your genes. There’s not a high-throughput method for doing that.” And because the mutagenesis can be used to cause partial loss of function instead of full loss — which can result in the organism’s death — it has proved a more fruitful source of study for scientists using the technology.
In fact, demand is so high that scientists routinely trek to Henikoff’s Seattle lab, where he hosts workshops about once a month to show people how the process works. For those researchers seriously interested, he’ll even run a pilot study on their population — crops can be done for free, thanks to an NSF grant — to ascertain whether TILLING has potential for their particular organism. “We’ve had failures and we’ve had successes,” he says. “The failures save people a lot of trouble because the mutation rate might not be high enough.”
Now, Henikoff is starting to adapt TILLING for use as a discovery tool for SNPs. “What we’ve been finding is that the methodology we’ve been developing for TILLING is really good for rare polymorphisms, something that you might find in a gene region once in a thousand [bases],” he says. This approach, called Ecotilling, has been tested out in Arabidopsis and Henikoff’s team is working on developing tools to enhance the process. “We’re part of the way there,” he says.
On his radar:
• Use of microarray technology for more than gene expression studies
Johns Hopkins University
Transposable elements: clue to V(D)J recombination
Nancy Craig has been interested in site-specific recombination since her days in grad school, so in some ways you might say her latest brainstorm is the culmination of her odyssey in the field so far. In a Nature paper published in late December, a team led by Craig, a professor at the Institute for Basic Biomedical Sciences at Johns Hopkins, announced the finding that a transposable element causes DNA-level changes on par with those seen in V(D)J recombination, the complex process through which cells recognize countless different antigens.
Using a transposon known as Hermes, found in the house fly and part of a superfamily that also shows up in organisms such as maize and petunia, Craig and her team went to work expressing the Hermes transposase in E. coli. After purifying it, Craig says, “we found out that we could make it do transposition in the test tube.” That seemingly simple step was a breakthrough in the hunt for which enzyme handles the bulk of the transposition process. “This really makes it seem that this is the only protein that’s necessary to do it,” Craig continues. “This is the heart of the reaction — it does the breakage and joining.”
Based on the work with E. coli, Craig and her team found that the transposase frees a transposon from its surrounding DNA strand by making nicks in the ends of specific recognition sequences. A chemical reaction causes the ends of the donor DNA sequence to loop into a hairpin, effectively releasing the transposon. That unusual formation has been a long-noted characteristic of V(D)J recombination and not seen in other transposase activity so far. Up until the work with the Hermes transposase, Craig says, “There had not been an element characterized which did it in exactly this way. … This had been hinted at by in vivo experiments, but this was the first ‘OK, here’s the biochemistry.’”
The upshot of this finding, she adds, is “evidence that this very fancy system for doing V(D)J recombination … very likely [evolved] from a transposable element.” In what will likely also prove a boon to this field of research, Craig’s team has managed a feat that has always been elusive in the study of transposases and integrases: getting a crystal structure of the Hermes transposase, which she hopes to publish in the near future.
Most recently, Craig is applying this knowledge to studies using yeast as a proving ground for using transposons in vivo to mutagenize DNA. In a collaboration with Mike Snyder at Yale, she says, “We took our transposon, modified it, made this huge insertion library, and analyzed all these proteins that were disrupted by this element.” Craig’s transposon research has also shown use in sequencing plasmids where primer walking hasn’t succeeded: “If you inserted a transposon into [the plasmid], you just use primers off the end of the transposon” to find your sequence’s location in the region, she says.
All of this work stems from Craig’s early days in graduate school, where site-specific recombination drew her interest. Transposable elements had been discovered and scientists were just finding how prevalent they were, she remembers. “Understanding how discrete pieces of DNA moved from place to place — that just captured my fancy,” she says.