Clear Evidence for Two Rounds of Vertebrate Genome Duplication

October 2005 | Volume 3 | Issue 10 | e347 | e353 Viewed under a microscope, the coronavirus appears almost beautiful, thanks to the halo-like crown formed by its surface proteins. (“Corona” means “crown” in Latin.) Aesthetics aside, this genus of viruses is responsible for a wide range of animal and human diseases, from the common cold to the deadly severe acute respiratory syndrome, familiarly known as SARS. Research efforts to design antiviral agents to combat coronaviruses intensifi ed after SARS killed at least 800 people in 2003 and have focused mostly on just this virus. But Haitao Yang, Dawei Ma, Zihe Rao, and colleagues reasoned that it might prove more effi cient to develop wide-spectrum drugs and vaccines that could work against all coronaviruses— signifi cantly reducing the health and economic burden associated with the 25 species of coronavirus. Scientists fear that vaccines may prove ineffectual against coronaviruses because the viruses, like HIV, change their protein sequences and structures so often that a vaccine targeting one strain would likely be ineffective against another. The success of such a vaccine strategy depends on fi nding a protein target that is present, or well conserved, among all the different coronaviruses. By combining structural and biochemical analyses, Yang et al. not only identifi ed such a target in a conserved region of a viral enzyme but also designed compounds with antiviral activity against multiple coronaviruses. Casting a Wide Net to Fight Coronaviruses

Viewed under a microscope, the coronavirus appears almost beautiful, thanks to the halo-like crown formed by its surface proteins. ("Corona" means "crown" in Latin.) Aesthetics aside, this genus of viruses is responsible for a wide range of animal and human diseases, from the common cold to the deadly severe acute respiratory syndrome, familiarly known as SARS. Research efforts to design antiviral agents to combat coronaviruses intensifi ed after SARS killed at least 800 people in 2003 and have focused mostly on just this virus. But Haitao Yang, Dawei Ma, Zihe Rao, and colleagues reasoned that it might prove more effi cient to develop wide-spectrum drugs and vaccines that could work against all coronaviruses-signifi cantly reducing the health and economic burden associated with the 25 species of coronavirus.
Scientists fear that vaccines may prove ineffectual against coronaviruses because the viruses, like HIV, change their protein sequences and structures so often that a vaccine targeting one strain would likely be ineffective against another. The success of such a vaccine strategy depends on fi nding a protein target that is present, or well conserved, among all the different coronaviruses. By combining structural and biochemical analyses, Yang et al. not only identifi ed such a target in a conserved region of a viral enzyme but also designed compounds with antiviral activity against multiple coronaviruses. DOI: 10.1371/journal.pbio.0030353 The tree of life has long served as a useful tool for describing the history and relationships of organisms over evolutionary time. One species is represented as a branching point, or node, on the tree, and the branches represent paths of descent from a parental node. The tree diagram carries an implicit assumption that genes are transferred vertically, from parent to child, and that all the genes in a new species come from the ancestral species. In theory, one should be able to trace the origin of each gene in a species back to its ancestor. In practice, however, the ancestral gene is rarely available, so researchers look for the gene in a closely related species. (These similar genes, which diverge slightly after a speciation event, are called orthologs.)

Casting a Wide Net to Fight Coronaviruses
But as the tools of genome analysis became more refi ned, searches for similar genes sometimes turned up sequences that belonged to a species on a different branch of the evolutionary tree. Clearly, vertical gene transfer was not the only mechanism of genetic transmission. Organisms, it turns out, can acquire genes from non-ancestral species through a mechanism called horizontal gene transfer (HGT)-think of it as acquiring genes from your neighbor instead of your parents. Such genetic exchanges, most common among bacteria and other microbes, are not represented in the tree of life-no single branch connects the two unrelated species. Initial studies suggested that HGT events were extremely common, prompting some to say it was time to replace the tree with a netlike diagram. Other studies have since suggested that methods used to calculate HGT overestimated its frequency: researchers detect HGT events by fi nding inconsistencies between gene trees and organism, or whole-genome, trees, but statistical errors can artifi cially increase the number of HGT events.
In a new study, Fan Ge, Li-San Wang, and Junhyong Kim estimate the frequency of HGT events by using a novel statistical method to compare the gene trees and whole-genome trees of microbes. Their method solves the statistical problem by directly testing for discrepancies between trees that arise from statistical error versus true HGT events. Analyzing over 40 microbial genomes, Kim and colleagues estimate that HGT infi ltrates just 2% of the average microbial genome. Even when relatively common, the authors conclude, HGT events do not disrupt the integrity of the tree of life, contributing just small bits of genetic material, "much like cobwebs on tree branches." To construct both gene trees and a whole-genome tree for the microbes, the authors selected core sets of orthologous gene groups from the NIH database of clusters of orthologous genes. (Clusters are derived by comparing protein sequences encoded in complete genomes, which represent major lineages on the evolutionary tree. Each cluster corresponds to an ancient, conserved protein domain.) Kim and colleagues created gene trees for each cluster of orthologous genes they selected, then created whole-genome trees from the gene trees and compared each gene tree to the whole-genome tree, using their new method. HGT events were detected when two species appeared close together on a gene tree but far apart on the wholegenome tree. Overall, just over 11% of the orthologous gene clusters showed statistically signifi cant HGT events, with HGTs accounting for about 2% on average of each of the 40 microbial genomes.
Altogether, these results suggest that HGT is not as common as once thought. And even when large-scale HGT events do occur-which Kim simulated in a previous study-they do not obscure the evolutionary path of most genes and lineages. If you imagine a tree with 10,000 taxa, the authors explain, and 1,000 HGTs per genome across all the taxa, the HGTs would form "extremely thin connections like cobwebs," leaving the backbone of the tree intact. Infrequent though it may be, HGT likely has some impact on the evolutionary history of life-impacts that advances in genome analysis technology may help uncover.

-Liza Gross
Anyone who winces remembering the challenges of childhood might consider the fate of the developing worm. Four times in its life, a growing nematode worm fl ips on its side and writhes around to shed its exoskeleton, or cuticle. During each molt, a worm casts aside its cuticle and synthesizes a new protective shell, its primary defense against a harsh environment. Though it's clear that a complex array of signaling proteins and enzymes are required to engineer these rites of passage, only a few genes have been implicated in the process.
Scientists have found critical molting genes in the fruitfl y, but most of these are not present in the worm. It may be that the worm has unique molting genes, because its cuticle is more elastic than the hardened casing of an insect. If scientists fi nd molting genes that exist only in the worm, they can begin to unravel the mechanisms that govern these critical phases of a worm's life. They can also develop treatments that target the worm's parasitic cousinswhich wreak havoc on humans, livestock, and plants-without producing harmful side effects. In a new study, Alison Frand, Sascha Russel, and Gary Ruvkun searched the entire genome of the worm Caenorhabditis elegans for molting genes and identifi ed 159 candidates, using a technique called RNA interference.
In RNA interference, researchers use double-stranded RNA (dsRNA) to block the expression of, or silence, a specifi c gene by destroying the gene's messenger RNA transcript before it can be translated into protein. These dsRNAs can be expressed in bacteria-a staple food for lab worms-which then multiply into large colonies expressing the same dsRNA. Using a pre-existing "library" of bacterial clones that each express a particular dsRNA, the authors fed groups of larvae one bacterial clone at a time-until thousands of larvae had eaten bacteria with dsRNA designed to The Molting Worm Sheds Its Genetic Secrets DOI: 10.1371/journal.pbio.0030345 Because coronavirus species show great diversity among their structural proteins-which include the glyocoproteins that form the halo-the authors turned to three enzymes as potential targets. But since structural data were available for only one of the enzymes, called the main protease (M pro ), the authors focused on M pro . Having structural data in hand greatly accelerates drug development, and since humans and other animals have no proteins similar to M pro , the likelihood of deleterious side effects is low.
Initial computer analysis showed that the M pro primary protein sequences (the linear amino acid sequences) have only 38% sequence identity between coronavirus species in some cases. But because three-dimensional structures tend to be more conserved than amino-acid sequences, the authors chose representative viruses from each group of coronavirus to study and compare the structure of their M pro . This protease normally binds to its target protein (called the substrate) via a specifi c region, called the substrate-binding site. Structural analysis determined that this site is well conserved among coronaviruses, and biochemical tests confi rmed that it would make a promising target for antiviral agents.
To test this hypothesis, the researchers created a synthetic version of the substrate that normally binds to the protease's substrate-binding site-reasoning that if they could inhibit the substrate's access to the binding site by the mimic (known as suicide inhibitors), they should be able to block the protease's activity and maybe halt viral replication. By studying the structure of the protease-substrate/inhibitor complex, Yang et al. continually improved their synthetic inhibitor until it bound strongly to the protease. Using this initial inhibitor as a base, the authors designed a panel of inhibitors and identifi ed compounds that rapidly blocked proteases from multiple coronaviruses and kept the coronaviruses from reproducing. The compounds caused no obvious damage in human cells in the experiments.
The substrate-binding site identifi ed by the researchers is an especially attractive target for drug development because evolutionarily conserved regions do not undergo high mutation rates like the rest of the viral genome, allowing antiviral drugs to maintain their effectiveness. Support for this hypothesis comes from the fi nding that a compound developed in this study also inhibits M pro from new coronavirus strains that cause conjunctivitis, bronchiolitis, and pneumonia. By identifying promising candidates for drugs capable of targeting the entire silence nearly every one of the worm's 19,427 genes.
After the larvae ate the bacterial clones, the authors screened them for molting defects-which is how they identified the 159 genes. Molting defects mostly left larvae trapped in their old cuticle; those that managed to escape often failed again during the next molt. The majority of candidate molting genes appear to play a role in all four molts, the authors argue, since their inactivation foils molting at several stages. And, significantly, the majority of genes-many of which are found only in worms-exist in worm parasites that infect humans, animals, and plants.
Among the genes identifi ed, the authors found several transcription factors (proteins that activate genes), indicating that molting requires "extensive changes in gene expression." Other genes are associated with signaling proteins that likely coordinate the activity of different cell types during molting, and many genes code for proteins that are required for protein synthesis-likely to build the new cuticle. Still other genes may help remodel the cuticle.
To monitor the expression of some of these genes and infer their function, the authors tagged a subset of genesrepresenting many of the functional categories found in the screen-with green fl uorescent protein. Since green fl uorescent protein glows when a gene is activated, the researchers can see where and when genes are expressed. Fluorescence levels were high just before each molt and dropped off soon after. All of these genes were expressed in the epithelial cells that secrete new cuticle. These results provide strong evidence for the genes' role in molting, since they were expressed both at the right time and the right place. These experiments also allowed Frand et al. to propose a model describing the timing and order of gene expression during molting.
With all the genes identifi ed in this screen, researchers can now start to piece together the overlapping pathways that guide the worm through its formative years. And with the discovery of wormspecifi c genes, it's likely that more effective treatments await patients with elephantiasis, African river blindness, and other diseases caused by pathogenic nematodes. -Liza Gross As one of the most important sources of novel gene functions, gene duplications play a major role in evolutionary change. Though a gene copy will generally become inactive after duplication, it can be saved-either by acquiring a new function or dividing aspects of the original gene's function-on its way to becoming ubiquitous, or "fi xed," within the population.
The notion of "evolution by gene duplication" was proposed in 1970 by Susumu Ohno, who argued that gene and whole genome duplication provided the raw material for evolutionary innovations such as subcellular compartments, fi ns, and jaws. Having "extra" copies of genes provides the opportunity for duplicate genes to escape the constraints of purifying selection, and allows the genes to diverge and acquire novel functions. Ohno also proposed that two rounds of whole genome duplication occurred at some point in early vertebrate evolution-a possibility that could explain the relatively large size and complexity of the vertebrate genome.
Investigators equipped with far more powerful genomemining tools than were available to Ohno have long sought evidence of this hypothesis (known as the 2R hypothesis, for "two rounds" of whole genome duplication), but with confl icting results. The observation that some gene families have four members in vertebrates but just one in invertebrates (the 4:1 rule) appeared to support the 2R hypothesis, until it was discovered that less than 5% of homologous gene families (similar genes with shared ancestry) followed the rule. But even when gene families do follow the rule, their confi guration could Clear Evidence for Two Rounds of Vertebrate Genome Duplication Understanding the recombination patterns across a chromosomedetermining the positions and frequency of genetic exchanges between homologous chromosomes-is crucial for understanding and tracking inheritance of traits. Mapping genes that affect parasites' traits, such as responses to various antimalarial agents, is possible because, during meiosis, homologous chromosomes line up and may exchange segments. Genes-or any polymorphic bits of DNA-that are close together tend to remain linked during this process, while those far apart tend to become separated. Identifying and following polymorphic markers through multiple generations is a key technique for genetic mapping.
For Plasmodium falciparum, the microbe that causes malaria, chromosomal mapping is necessary for understanding the evolution of the parasite and development of drug resistance, but multiple factors make this a complex task. In this issue, Jianbing Mu and colleagues use single nucleotide polymorphisms (SNPs) to evaluate some of these factors, and set the stage for further mapping of this important parasite's genome.
The authors began by locating 183 SNPs spaced across Chromosome 3 in 99 P. falciparum populations from throughout the world. Not all SNPs were found in all populations, indicating a more recent evolutionary origin for some SNPs; these differences were then used to track evolution and migration in parasites. Statistical analysis of the SNPs allowed the populations to be parsed into fi ve groups, largely corresponding to continents. More refi ned analysis of the SNPs revealed possible migratory history, including a recent migration of an African variety to coastal South America.
Mu and colleagues also showed for the fi rst time that the historical rate of recombination varies widely-over 20-fold-among different populations. A large part of the variation is due to a combination of the frequency of infections with multiple parasite strains (because sexual recombination occurs only within an infected mosquito) and the degree of inbreeding within a parasite population. Inbreeding tends to lower the extent of detectable recombination events, while multiple infections by different strains increase it.
Clues to the Evolution of the Malarial Chromosome Plasmodium falciparum, the microbe that causes malaria, infects red blood cells.
By analyzing different populations of the pathogen from around the world, researchers found clues to its genome structure that will be important for identifying genes that contribute to drug resistance and virulence. just as likely arise from two rounds of single gene duplication as from whole genome duplication. And because duplicate genes are far more likely to degrade than to assume new or shared functions, the signal of whole genome duplication disappears.
Recent studies have shown that the global pattern of the physical location of homologous genes provides evidence of ancient whole genome duplications in yeast and plants, even when most of the duplicates have degraded. Now Paramvir Dehal and Jeffrey Boore have taken this approach to test the 2R hypothesis, by comparing the recently completed genome sequence of the invertebrate sea squirt with the genomes of three vertebrates-pufferfi sh, mouse, and human. (Because the sea squirt is a close relative of vertebrates, its genome can help reconstruct a more accurate tree of the organisms' evolutionary relationships than a more distant relative like the fruitfl y could.) After generating gene clusters that each contained "all, and only, those genes that descended from a singe gene in their common ancestor," the authors used a method to infer the evolutionary relationships of the genes in each cluster. They could then compare these gene trees to the known evolutionary relationships of the organisms to determine when each gene duplicated in relation to when the lineages diverged. From this analysis, Dehal and Boore identifi ed over 3,500 gene duplications present in multiple vertebrates, indicating they had occurred at the base of the vertebrate tree, dating back some 450 million years. But did these early duplication events arise from some large-scale duplication event, or were they simply the result of a great number of smaller scale duplications?
To explore this question, the authors analyzed the relative positions of the resulting paralogs in the vertebrate genome with the highest-quality data-the human genome. When considering only this subset of 3,500-plus early vertebrate duplications, they found a global pattern of human genome segments with similar arrangements of paralogous genes and multiple chromosomes with long linear stretches of interdigitated sets of paralogous genes-evidence that the duplications occurred in large segments. Even stronger support for the 2R hypothesis comes from the observation that the colinear arrangement of these genes is predominantly in a 4fold pattern; this repetitive pattern is seen across almost all the human chromosomes. It's unlikely, the authors argue, that any combination of smaller, independent duplication events could have generated the same pattern. Now that strong evidence for Ohno's hypothesis exists, researchers can investigate both the mechanism of genome duplication events and their possible effects on vertebrate evolution. It seems likely that a whole genome duplication would provide combinatorial possibilities that could permit a greater leap in evolution than could single gene duplications, even if the single gene duplications affected the complete set of genes. Studies that examine the function of these paralogous genes can explore whether these large-scale genomic events helped drive organismal complexity and diversifi cation within the vertebrate lineage. Despite the wide differences in recombination rates, all populations had a similar clustering of recombination "hot spots" at the middle and ends of the chromosome. Recombination is most likely to occur at these spots, and the similar localization refl ects either the common evolutionary history of all the populations or localization of crossover events to particular genomic regions.
The authors compared their results from population structure analysis with those using SNPs from genes that might be infl uenced by drug pressure. Their results showed that misleading inferences about the parasite population structures could be derived using information from genes that are potentially under drug selection.
These results are important because they provide information on the multiple complex factors that must be considered in understanding the genomic structure of P. falciparum, which is critical for identifying genes that contribute to phenotypes such as drug resistance and virulence. Reseachers conducting future mapping studies will be able to draw on the important fi ndings and caveats revealed by this work to refi ne their own methods and interpret their results. In one of biology's most impressive engineering feats, specialized proteins package some six-and-a-half feet of human DNA into a nucleus that averages just 5 microns (0.0001969 inches) in diameter. In the fi rst of a series of supercondensing steps, DNA winds around proteins called histones, which together form a complex called the nucleosome. Histones package DNA into repetitive coils, which not only provide genomic structure but also help regulate gene expression. These tasks are mediated in part by chemical modifi cations to histone proteins-most commonly to histone "tails," long, unstructured chains of amino acids that protrude from nucleosomes. Different chemical modifi cations are associated with different functional effects. Acetylation, which adds an acetyl group to an amino acid on the histone tail, has been linked to both gene activation and silencing, depending on which amino acid is modifi ed. Methylation (addition of a methyl group to the histone tail) has also been linked to gene activation and repression, although the chemical effects of methylation differ dramatically from those of acetylation.
Even in yeast, amino acid modifi cations in the histone tails can number in the tens and twenties. Given the number of possible permutations of modifi cation types and amino acids, the question arose, might different combinations of histone modifi cations produce discrete outcomes? The notion that a sequence or combination of specifi c modifi cations on histone tails acts as a signal to other proteins and produces distinct biological effects was advanced as the "histone code" hypothesis in 2000.
Progress in deciphering the vocabulary, mechanics, and function of the histone code has been hindered by the coarse resolution of available tools. Nucleosomes typically cover about 146 base pairs, but existing technology could only average over 500 to 1,000 base pairs at a time-confounding the effects of single nucleosomes. In a new study, Oliver Rando and colleagues take advantage of the high resolution afforded by their custom-made microarray, which has a resolution of 20 base pairs. Working with the budding yeast Saccharomyces cerevisiae, the scientists examined 12 different histone modifi cations in individual nucleosomes and found only a small number of distinct combinations with "few discrete histone modifi cation patterns." The concurrent modifi cations fall into two categories: one set targets a transcriptional start site but is the same no matter what the level of transcription, while the other occurs throughout gene coding regions and is linked to transcription. Importantly, the only modifi cations that appear to correlate with transcription occur over transcribed regions, as though they were the consequence, rather than the cause, of transcription.
Why might histone tails exhibit so many modifi cations if they form only two independent categories? It's possible that histone-modifying enzymes may work best in groups and so the marks that recruit them-acetyl and methyl groups-also come in groups. Another possible explanation relates to how histone modifi cations signal transcription enzymes that a particular gene requires more or less transcription. When the positively charged amino acid lysine acquires an acetyl group, it loses its charge, and charge-charge interactions play a major role in many interactions between proteins and other molecules. Multiple lysine acetylations on the histone tail may thereby aid certain chemical reactions necessary for transcription in a continuous way; having multiple levels of acetylation, for example, may allow the cell to "tune" Histones can undergo many potential modifi cations, and it has been hypothesized that these can occur in many different combinatorial histone modifi cation patterns (A). In this study, researchers found that only a few modifi cation patterns occur in yeast, with many of the modifi cations co-occurring in groups (B).
visual stimuli, including natural, random, and synthesized images; the synthesized images helped them distinguish between the power spectrum and phase effects. Feature sensitivity depends on a neuron's preferred features, which the authors estimated from a neuron's response to a set of natural images, including a man's face, a building, a lion, and a hand. The authors created a set of random images with global and feature contrasts that matched the natural images, based on the preferred features, and recorded neuronal response to both the natural and random image sets. Overall, the majority of complex cells showed higher feature sensitivity for the natural stimuli, indicated by the contrast-response function, which plots neuronal response against the contrast of the feature. (A steep contrastresponse function shows high neuronal sensitivity to a preferred feature, while a fl at function indicates insensitivity.) Simple cells showed no difference in their response to natural and random stimuli.
Since the contrasts of natural and random image sets were matched, the complex cells' sensitivity to natural images could not be explained by differences in overall contrast. More likely, the authors reasoned, the cells were responding to the power or phase spectra of the light. Felsen et al. manipulated each property in synthesized image sets to distinguish their effects on neuronal response. In one image set, each image had a natural power and random phase spectrum, while a second image set had the reverse. Comparing the feature sensitivity of complex cells to each of these image sets with a random image set, the contrast-response function, and thus feature sensitivity, was highest for the synthesized natural phase image set.
By experimentally linking visual statistics with neuronal responses, this study not only reveals a novel coding response property of complex cells but also provides evidence for the theory of effi cient coding. The fi nding that complex cells selectively respond to properties of natural stimuli that simple cells don't shows how brain circuits divide tasks to make the most of available resources. protein-protein interactions, and thus gene expression, up and down, rather than simply turn it on or off.
Rando and colleagues propose that the histone modifi cations associated with transcription may facilitate rather than trigger gene expression, perhaps by clearing a path for the transcription machinery or attracting proteins needed for the job. The authors are careful to point out, however, that histone modifi cations may also play some role in initiating gene expression, but that any transcription pattern would likely be obscured, or "erased," as transcription occurs. While future studies will help determine which role proves more common, these results suggest that histone modifi cations are facilitators rather than activators and that the histone code is more a transcription footprint than a starting signal. An Amur tiger roaming the snowcovered forests of the Russian Far East sees life differently than an arboreal primate raised in the thick canopy of a tropical jungle. Adaptations in the structure of the eye and the visual centers of the brain facilitate these different worldviews, fi ne-tuning each animal's vision to the light levels and visual properties of its environment.
The notion that brain circuits are adapted to represent natural stimuli is known as the effi cient coding hypothesis. In this framework, the visual system responds most effectively to the features found in the natural environment, like the specifi c arrangement of trees, another animal's face, or the contrast of land abutting a river. Natural images possess different statistical features than random noise (such as static on a TV monitor), so responses to these two classes of stimuli could be distinct. According to the prevailing hypothesis, the retina and the lateral geniculate nucleus, the brain area that relays neural signals to the cortex, are tuned to the power spectrum of light signals found in natural scenes. In other words, these neurons are sensitive to the dominant energy at particular spatial or temporal frequencies in natural scenes and less sensitive to random noise, which has an overall fl at power spectrum.
In a new study, Gidon Felsen, Yang Dan, and colleagues investigate how neurons in the cat primary visual cortex (V1) respond to images with natural statistics, and discover something truly novel about a V1 neuron's sensitivity to features in natural scenes: a specifi c class of V1 neurons, called complex cells, are preferentially tuned to the phase regularities of light signals in natural scenes, not to the power spectrum.
(Phase relates to how the signals in different spatial frequencies align, which gives rise to edges in the image.) The V1 contains two types of neurons, known as simple and complex cells, that were originally distinguished based on how they respond to light, determined by shining a fl ashlight on a wall and mapping the space that triggered neuronal activation. This activation space defi nes the cell's receptive fi eld. The receptive fi eld properties of simple and complex cells vary signifi cantly: simple cells behave linearly (two spots of light in the receptive fi eld doubles the response), and complex cells behave nonlinearly.
To characterize the feature sensitivity of these neurons, the authors measured their response to different classes of Complex Cells in the Brain's Vision Center Tune in to Natural Scenes Single-celled bacteria may appear simple by some standards, but these tiny cells employ sophisticated systems for processing stimuli. One especially important class of signaling molecules that help bacteria coordinate the activities of daily life is called the two-component signal transduction system. This system-comprised of enzymes called histidine kinases and their target molecules, the response regulators-allows bacteria to sense and respond to their surroundings by transforming various environmental cues, such as sugars, peptides, and antibiotics, into physiological responses. These environmental signals trigger a chemical reaction in histidine kinases called autophosphorylation, in which the kinase transfers a phosphoryl group from a molecule of ATP (which powers many cellular processes) to one of its own amino acids. The enzyme then transfers the phosphoryl group to its target response regulator, producing changes in gene expression, motility, protein breakdown, and various other cellular processes.
Though it's possible to identify histidine kinases and response regulators in bacteria by analyzing their genome sequences, it's far more diffi cult to determine how the two components interact: do they form monogamous pairs or behave more promiscuously? To characterize the range of possible interactions and the intracellular changes they bring about, Michael Laub and colleagues developed a novel method of mapping the connections between histidine kinases and response regulators in the freshwater bacterium Caulobacter crescentus. By combining genetic and biochemical analyses on a system-wide scale, the authors rapidly identifi ed two-component signaling pathways in C. crescentus, including pathways required for core cell processes.
Laub and colleagues fi rst identifi ed 106 two-component signal transduction genes (62 histidine kinases and 44 response regulators) from the genome sequence. Then they created mutant strains of bacteria that each had one of the 106 genes Mapping Core Communication Networks in Bacteria DOI: 10.1371/journal.pbio.0030359 Unlike virtually every other type of cell, muscle cells contain dozens or even hundreds of nuclei. These multinucleate cells, called myofi bers, form by fusion of precursor cells, called myoblasts, with "founder cells." In the fruit fl y, embryonic founder cells are formed by a well-known signaling pathway, but the same mechanism is not used to form adult founder cells. In this issue, K. VijayRaghavan and colleagues identify several key molecules involved in adult founder cell formation, and show that the process occurs through a novel mechanism.
In the fl y embryo, founder cells differentiate from myoblasts through the actions of a membrane-bound receptor called Notch, an important player in several "signaling cascades" that use environmental signals to trigger changes in gene expression. However, in previous work, the authors have shown that Notch does not play a role in establishing adult founder cells. Instead, several clues pointed to the Fibroblast growth factor (FGF) family of receptors, one of which, in the fl y, is called Heartless. Among other locations, Heartless is found on the surface of adult myoblasts in the abdomen. Reducing its expression in these cells, the authors showed, reduced the number of founder cells, while elevating it increased them.
But since Heartless is found in all adult myoblasts, it could not be responsible by itself for converting a myoblast into a founder cell. Another protein, called Heartbroken, seemed like a good candidate, since it functions exclusively within the FGF pathway. The authors showed that while the gene for Heartbroken is initially expressed in all myoblasts, over time its expression becomes restricted to those cells that develop into founder cells. Furthermore, by artifi cially maintaining Heartbroken expression, the authors dramatically elevated the number of founders in developing muscle, strengthening the case that Heartbroken is a key promoter of founder cell development.
But since both Heartless and Heartbroken are initially present in early myoblasts, what prevents wholesale Heartless signaling and premature, widespread founder cell formation? The authors show that a third factor, called Sprouty, declines in expression as founders are specifi ed, and is not detected after founders are established. Sprouty is known to be a negative regulator of FGF signaling. VijayRaghavan and colleagues suggest that Sprouty interferes with Heartless signaling in early myoblasts, preventing founder cell formation even in the presence of Heartbroken. The gradual decline in the level of Sprouty may then "release the brakes" on Heartless signaling. Not every myoblast becomes a founder cell at that point, though, because the level of Heartbroken has also declined. Exactly which cells will maintain suffi cient Heartbroken to become founders, and how those cells are specifi ed, remains to be worked out.
Still more remains to be discovered about the development of muscle in the adult fl y. The signifi cance of this work lies in identifying the FGF pathway as a critical component in muscle cell development, which provides leads that can be used to fi ll in the missing members of the pathway. While the details of vertebrate muscle development differ, the FGF pathway is known to be involved there too, and this work may shed light on aspects of that process as well. deleted (called deletion strains) to learn how the genes function. The phenotypes, or physical characteristics, of the mutant strains allowed the scientists to identify 39 genes required for cell cycle progression, growth, and morphogenesis, including nine genes essential to survival. To address the promiscuity question and fi gure out the likely phosphotransfer pairings among the components, the authors developed a biochemical method, called phosphotransfer profi ling, that quickly identifi es a histidine kinase target by tracking the transfer of radioactive (called radio-labeled) phosphates from the kinase to the target.
The authors validated their in vitro technique on twocomponent proteins from E. coli (a system in which many of the living bacteria's kinase-target pairings are known) by showing that the kinases preferentially phosphorylated their known targets in the test tube as well, forming promiscuous unions only after prolonged periods. Confi dent that their method would also work for other bacteria, the researchers applied it to C. crescentus and determined the likely pairings for previously identifi ed histidine kinases. Since the authors again observed a high preference for known targets, they were confi dent that these pairings were real, and the method could be used to identify targets of other, uncharacterized kinases. Based on the deletion analysis results, Laub and colleagues focused on a histidine kinase that appears essential for growth or survival and identifi ed a single target response regulator. Eliminating the expression of each gene produced nearly identical phenotypes, revealing the pair's role in a signaling pathway that maintains the integrity of the bacteria's cellular envelope. Because twocomponent systems aren't found in humans, this cellularenvelope pathway may prove an effective antibiotic target against pathogenic bacteria, based on the effectiveness of other antibacterial therapies aimed at the cell membrane.
The library of deletion strains the authors developed will be a valuable resource, not only for identifying other two-component pathways in C. crescentus but also for studying aspects of the cell cycle and development, thanks to the bacterium's unusual life cycle-it divides asymmetrically, producing a motile daughter cell and a stationary one that can then differentiate into the motile version. And since the techniques presented here should work in any organism with two-component signaling (most bacteria, fungi, and many plants), researchers can apply them to the tall task of decoding the labyrinthine communication systems that sustain cellular life. Serious gardeners know a healthy harvest depends on the right mix of ingredients at the right time. Nitrogen, which spurs leafy growth at the expense of fl owers, works best soon after plants emerge. Phosphate, which promotes blossoms, should come later. Indoor growers often gas their plants with CO 2 to boost growth, development, and budding. The really serious indoor gardener can pay $599 for a microclimate controller that automatically regulates temperature, humidity, CO 2 , lighting, and irrigation.
These components are just as important to natural ecosystems as they are to the well-tended garden. But their composition in the atmosphere and soil is changing at an unprecedented rate. Rising concentrations of CO 2 and other greenhouse gases-which have increased apace with global agricultural and industrial development-are linked to warmer temperatures and changes in precipitation patterns. The by-products of industrial and agricultural operations have also increased nitrogen concentrations, saturating systems in highly polluted areas.
Plants use CO 2 (and water and light) during photosynthesis to make sugar and grow. A plant absorbs CO 2 through its leaves and takes up nitrogen (often in the form of nitrate) and other soil nutrients through its roots. Increasing CO 2 concentrations can facilitate plant growth by increasing the rate of photosynthesis, though this effect varies depending on the photosynthetic pathway the plant uses. Trees, for example, use a different pathway than most tropical grasses. Higher carbon inputs can also trigger more effi cient nitrogen use. Because plants are primary producers, at the base of the food chain, major shifts in the global cycles that support them will have wide-ranging consequences.
To observe how a natural system might respond to these changes over the long term, a team led by researchers from the Carnegie Institution of Washington and Stanford University created an experimental system in their Northern California backyard. Jeffrey Dukes, Christopher Field, and their colleagues treated grassland plots to every possible combination of current or increased levels of four environmental factors-CO 2 , temperature, precipitation, and nitrogen infl ux-to simulate likely regional changes over the next 100 years. Previous studies have tested natural systems' responses to one or two of these changes, but none has tested the longterm, simultaneous impacts of each.
The strongest effects on grassland production came from elevated levels of nitrogen (which typically reaches a fertilization limit). Elevated temperature, precipitation, and, surprisingly, CO 2 , had minimal impacts. These results suggest, the authors argue, that California grasslands, and ecosystems that respond Field Tested: Grasslands Won't Help Buffer Climate Change as CO 2 Levels Rise DOI: 10.1371/journal.pbio.0030329 similarly, are not likely to help buffer the rate of climate change by acting as a carbon "sink"-slowing the rise of CO 2 levels by storing more carbon in new growth. It's thought that ocean and terrestrial ecosystems have stored nearly half the carbon emissions produced by humans since the industrial revolution.
If it turns out that other natural systems also fail to sequester as much carbon as scientists once thought, atmospheric CO 2 concentrations will rise even faster than expected-with serious implications for future climate change.
The experiments were part of the Jasper Ridge Global Change Experiment (JRGCE), which started on Stanford's 1,200-acre biological preserve in 1997. Since 1998, this grassland ecosystem has been outfi tted with an ecologist's version of a microclimate controller (complete with CO 2 pumps, heaters, and irrigation tubing) and subjected to experimentally controlled atmospheric, climatic, and nutrient conditions. (This study examines the experiment's fi rst fi ve years.) To quantify the grassland response to these treatments, the authors estimated net primary production, or NPP (the amount of carbon left over after cellular respiration) by measuring shoot and root growth in 36 circular plots scattered across roughly two acres. Four control plots experienced the natural variations of California's Mediterranean climate.
Overall, increased rainfall, warming, and elevated CO 2 had little effect on NPP. (More rain triggered shoot growth but stunted root growth, so NPP wasn't affected.) In some experimental treatments and years, elevated CO 2 actually reduced grassland production. Increased nitrate, on the other hand, led to shoot and root growth imbalance, with shoots growing faster than roots. And this added nitrogen "strongly increased" NPP in every year but one. These results suggest that increasing concentrations of atmospheric CO 2 are not likely to increase growth of the roots and leaves of plants in this grassland. Why not?
One possibility involves phosphorus. High levels of CO 2 and nitrogen can reduce phosphorus concentrations or limit its uptake in these plants. Ongoing JRGCE experiments are exploring how this and other factors-such as grazing or shifts in seasonal events-might limit the growth effects of CO 2 .
Because grasslands and forests operate in complex feedback loops with both the atmosphere and soil, understanding how ecosystems respond to global changes in climate and element cycling is critical to predicting the range of global environmental changes-and attendant ecosystem responses-likely to occur. Ecosystem responses may well vary according to their composition and location. By studying natural systems over multiple years on an ecosystem scale, experiments like JRGCE offer a valuable tool for simulating predicted global changes and assessing their likely impacts, region by region.

Researchers work on some of the 36 plots that lie below four infrared heaters in the Jasper
Ridge Global Change Experiment. Total plant growth in grassland plots like these rarely responded to changes in climate or atmospheric CO 2 concentration.
Embedded in the brain's hypothalamus on top of the intersection of the two optic nerves lies a group of some 10,000 cells known as the suprachiasmatic nucleus. This neural region determines the biological equivalent of Greenwich Mean Time: the circadian cycle. Circadian rhythms infl uence phenomena as diverse as leaf position in plants, fatigue patterns in mice, and latenight hunger in human adolescents. By measuring exactly how long an organism takes to complete one cycle, scientists hope to gain insight into the mechanisms underlying these broad effects. "Circadian" translates from Latin to "approximately a day," but scientists search for more precise answers about how genes and environment account for fl uctuations in circadian cycle length.
Theoretically, scientists could measure the circadian signal from its source in the suprachiasmatic nucleus-but no method exists to do this in a living subject. Instead, scientists rely on prolonged observation of live mice or humans. They document when mice use their mouse wheel, for example, and when they sleep. But to screen for and identify circadian rhythm variations in humans, the required period of lengthy observation is prohibitively costly and labor intensive.
Circumventing these technical limitations, Ueli Schibler and colleagues have recently designed a method to measure circadian cycles in mammalian cells cultured from tissues other than the suprachiasmatic nucleus. The researchers infected Fibroblast Culture Cells Keep Track of Circadian Rhythms DOI: 10.1371/journal.pbio.0030348 human and mouse fi broblast tissue cultures with a virus engineered to report when a certain host circadian rhythm gene was expressed. They found that their data jibed with the previously accepted length of the human circadian cycle: 24.5 hours. Because of the sensitivity of their method, Schibler and colleagues also confi rmed that, for both humans and mice, circadian rhythms vary substantially between individuals. This suggests that the genetics of the circadian clock likewise varies between individuals.
The so-called reporter gene, lucerifase, produces a protein that illuminates the cell. Steve Brown, a postdoc in Schibler's laboratory, created a virus system that specifi cally inserts lucerifase into the host genome near a gene important for establishing the circadian rhythm. With this system, the same gene that promotes expression of the circadian rhythm gene also promotes expression of lucerifase-consequently, the fi broblast cultures light up in time with the circadian clock.
Schibler and colleagues obtained human fi broblast skin cells from the abdomen, buttocks, male foreskin, and other areas to measure the circadian rhythm indirectly. Unlike an individual's true rhythm, a fi broblast cell culture's rhythm does not vary with changes in light exposure or sleep habits. The researchers point out that their method can expose differences in circadian rhythms; it does not, however, directly measure the signal from the suprachiasmatic nucleus that coordinates each individual cell's clock.
The authors also found that circadian rhythm variations between cultures from the same individuals differed far less than differences between individuals. This discovery indicates that fi broblast cell cultures are reliable tools for approximating an individual's genetically determined circadian rhythm. The scientists also found that, for mice, circadian rhythm time obtained from prolonged observation and that obtained from the new viral method were showing the same tendency, although the differences observed in behavior were exacerbated in molecular fi broblast rhythms. This result suggests that the fi broblast circadian clocks might give relevant information about the brain's circadian clock in humans as well as in mice.
In the future, scientists may use the new method to screen large populations for genetically linked sleep disturbances such as advanced and delayed sleep phase syndromes. They may also use this test to home in on the genetic mechanism responsible for such conditions. Outside the realm of medicine, future genetic studies of circadian rhythm may exploit the method developed by Schibler and colleagues to explore questions about the exact relationship between the suprachiasmatic nucleus and the circadian rhythms of cultured fi broblast cells. Just how does this brain structure coordinate genetic imperatives with environmental input including light fl uctuations? Finally, frequent fl iers, bleary-eyed in foreign time zones, may get an answer to why waking up is so hard to do. Studying circadian rhythms in humans isn't as straightforward as it is with lab rodents. As a proxy for humans, researchers compared circadian gene expression in primary fi broblasts from different human individuals and found unexpectedly large differences between the circadian clocks of different human subjects.
Organelles in cells communicate with each other via small membrane-bound vesicles that carry molecular cargo from one part of the cell to another. Just as a cargo ship must know where to dock in order to empty goods at the correct port, vesicles in a cell must dock and empty their contents in the appropriate part of the cell. Studying the regulation of vesicle transport and membrane fusion is a major area of research in cell biology. Though the rules governing vesicle transport and fusion in the sea of cellular organelles have been deciphered in bits and pieces from various cell systems, all the components required for vesicle fusion have not been characterized for any single cell type. In a new study, Gerardo De Blas, Carlos Roggero, Claudia Tomes, and Luis Mayorga have elucidated the molecular mechanics of membrane fusion by studying the single vesicle of the sperm, the acrosome.
Enzymes released from the acrosome facilitate contact between the sperm and egg membranes during fertilization by dissolving the sheath surrounding an egg. For this, the acrosome membrane must fuse with the outer sperm membrane-a process called the acrosome reaction. This reaction happens only once in the lifetime of a sperm. In other systems, proteins EnSNAREd by the Sperm Acrosome DOI: 10.1371/journal.pbio.0030352 involved in membrane fusion must be recycled so they can be reused in the fusion of a subsequent vesicle. But the acrosome reaction is unidirectional and requires no recycling, making it easier to decipher the steps involved.
In previous work, Mayorga's group had shown that an increase in the concentration of intracellular calcium activates a molecule called Rab3A, thus initiating the acrosome reaction. The reaction proceeds with the help of various proteins, including NSF, SNAREs, and synaptotagmin VI, as well as calcium release from within the acrosome. Synaptotagmin VI is a calcium-sensitive protein. SNAREs are highly specialized proteins that exist in complexes of three molecules wound together in a helix and are present on both the acrosome and plasma membranes. NSF unwinds these helices so that molecules on opposite membranes can interact. To determine whether the SNARE proteins are in a single or triplet confi guration at any given time, the authors used bacterial neurotoxins that can degrade single SNAREs but have no effect against the triplets.
In this study, De Blas and colleagues combine fl uorescent techniques, a light-sensitive calcium chelator (which depletes all the calcium in the acrosome), and chemicals that inhibit specifi c steps in the cascade, to decipher whether each reaction occurs before or after the release of calcium. The researchers show that Rab3A activates NSF, which goes on to untwine helical, neurotoxin-resistant SNARE complexes on the acrosomal and sperm membranes, allowing opposing SNAREs to interact. Once SNAREs on opposite membranes form neurotoxin-sensitive loose complexes, calcium is released from within the acrosome. One protein that is required after the release of calcium is synaptotagmin VI. This protein, the authors suggest, could be responsible for the tight zippering of SNAREs on opposite membranes, converting them into toxin-resistant complexes. As the helix between molecules on opposite membranes becomes tighter, the membranes get pulled closer to each other, enabling membrane fusion.
How calcium activates Rab3A or synaptotagmin VI and how these proteins carry out their roles at the molecular level remain to be elucidated. However, this study elegantly demonstrates the cascade of players involved in the acrosomal membrane fusion reaction, from start to fi nish, in a single cell-something that had not been shown before.