Multi-state Modeling of Biomolecules

Multi-state modeling of biomolecules refers to a series of techniques used to represent and compute the behavior of biological molecules or complexes that can adopt a large number of possible functional states. Biological signaling systems often rely on complexes of biological macromolecules that can undergo several functionally significant modifications that are mutually compatible. Thus, they can exist in a very large number of functionally different states. Modeling such multi-state systems poses two problems: the problem of how to describe and specify a multi-state system (the “specification problem”) and the problem of how to use a computer to simulate the progress of the system over time (the “computation problem”). To address the specification problem, modelers have in recent years moved away from explicit specification of all possible states and towards rule-based formalisms that allow for implicit model specification, including the κ-calculus [1], BioNetGen [2]–[5], the Allosteric Network Compiler [6], and others [7], [8]. To tackle the computation problem, they have turned to particle-based methods that have in many cases proved more computationally efficient than population-based methods based on ordinary differential equations, partial differential equations, or the Gillespie stochastic simulation algorithm [9], [10]. Given current computing technology, particle-based methods are sometimes the only possible option. Particle-based simulators fall into two further categories: nonspatial simulators, such as StochSim [11], DYNSTOC [12], RuleMonkey [9], [13], and the Network-Free Stochastic Simulator (NFSim) [14], and spatial simulators, including Meredys [15], SRSim [16], [17], and MCell [18]–[20]. Modelers can thus choose from a variety of tools, the best choice depending on the particular problem. Development of faster and more powerful methods is ongoing, promising the ability to simulate ever more complex signaling processes in the future.

This is a “Topic Page” article for PLOS Computational Biology.

     

Inside the Mucosal Immune System

An intricate network of innate and immune cells and their derived mediators function in unison to protect us from toxic elements and infectious microbial diseases that are encountered in our environment. This vast network operates efficiently by use of a single cell epithelium in, for example, the gastrointestinal (GI) and upper respiratory (UR) tracts, fortified by adjoining cells and lymphoid tissues that protect its integrity. Perturbations certainly occur, sometimes resulting in inflammatory diseases or infections that can be debilitating and life threatening. For example, allergies in the eyes, skin, nose, and the UR or digestive tracts are common. Likewise, genetic background and environmental microbial encounters can lead to inflammatory bowel diseases (IBDs). This mucosal immune system (MIS) in both health and disease is currently under intense investigation worldwide by scientists with diverse expertise and interests. Despite this activity, there are numerous questions remaining that will require detailed answers in order to use the MIS to our advantage. In this issue of PLOS Biology, a research article describes a multi-scale in vivo systems approach to determine precisely how the gut epithelium responds to an inflammatory cytokine, tumor necrosis factor-alpha (TNF-α), given by the intravenous route. This article reveals a previously unknown pathway in which several cell types and their secreted mediators work in unison to prevent epithelial cell death in the mouse small intestine. The results of this interesting study illustrate how in vivo systems biology approaches can be used to unravel the complex mechanisms used to protect the host from its environment.Higher mammals have evolved a unique mucosal immune system (MIS) in order to protect the vast surfaces bathed by external secretions (which may exceed 300 m² in humans) that are exposed to a rather harsh environment. The first view of the MIS is a single-layer epithelium covered by mucus and antimicrobial products and fortified by both innate and adaptive components of host defense (Figure 1). To this, we can add a natural microbiota that lives in different niches, i.e., the distal small intestine and colon, the skin, the nasal and oral cavities, and the female reproductive tract. The largest microbial population can reach ∼10¹² bacteria/cm³ and occurs in the human large intestine [1]–[3]. This large intestinal microbiota includes over 1,000 bacterial species and the individual composition varies from person-to-person. Other epithelial sites harbor a separate type of microbiota, including the mouth, nose, skin, and other wet mucosal surfaces, that contributes to the host; in turn, the host benefits its microbial co-inhabitants. Gut bacteria grow by digesting complex carbohydrates, proteins, vitamins, and other components for absorption by the host, which in return rewards the microbiota by developing a natural immunity and tolerance (reviewed in [4]–[7]). Finally, the host microbiota influences the development and maturation of cells within lymphoid tissues of the MIS [8],[9].Open in a separate windowFigure 1The gut, nasal, upper respiratory and salivary, mammary, lacrimal, and other glands consist of a single layered epithelium.Projections of villi in the GI tract consist mainly of columnar epithelial cells (ECs), with other types including goblet and Paneth cells. Goblet cells exhibit several functions including secretion of mucins, which form a thick mucus covering. Paneth cells secrete chemokines, cytokines, and anti-microbial peptides (AMPs) termed α-defensins.Mucosal epithelial cells (ECs) are of central importance in host defense by providing both a physical barrier and innate immunity. For example, goblet cells secrete mucus, which forms a dense, protective covering for the entire epithelium (Figure 1). Peristalsis initiated by the brush border of gastrointestinal (GI) tract ECs allows food contents to be continuously digested and absorbed as it passes through the gut. In the upper respiratory (UR) tract, ciliated ECs capture inhaled, potentially toxic particles, and their beating moves them upward to expel them, thereby protecting the lungs. Damaged, infected, or apoptotic ECs in the GI tract move to the tips of villi and are excreted; newly formed ECs arise in the crypt region and continuously migrate upward. Paneth cells in crypt regions of the GI tract produce anti-microbial peptides (AMPs), or α-defensins, while ECs produce β-defensins [10],[11] for host protection (Figure 1). A major resident cell component of the mucosal epithelium are intraepithelial lymphocytes (IELs). The IELs consist of various T cell subsets that interact with ECs in order to maintain normal homeostasis [12]. Regulation is bi-directional, since ECs can also influence IEL T cell development and function [12]–[14].The MIS, simply speaking, can be separated into inductive and effector sites based upon their anatomical and functional properties. The migration of immune cells from mucosal inductive to effector tissues via the lymphatic system is the cellular basis for the immune response in the GI, the UR, and female reproductive tracts (Figure 2). Mucosal inductive sites include the gut-associated lymphoid tissues (GALT) and nasopharyngeal-associated lymphoid tissues (NALT), as well as less well characterized lymphoid sites (Box 1). Collectively, these comprise a mucosa-associated lymphoid tissue (MALT) network for the provision of a continuous source of memory B and T cells that then move to mucosal effector sites 13,14. The MALT contains T cell regions, B cell–enriched areas harboring a high frequency of surface IgA-positive (sIgA⁺) B cells, and a subepithelial area with antigen-presenting cells (APCs), including dendritic cells (DCs) for the initiation of specific immune responses (Figure 2). The MALT is covered by a subset of differentiated microfold (M) cells, ECs, but not goblet cells, and underlying lymphoid cells that play central roles in the initiation of mucosal immune responses. M cells take up antigens (Ags) from the lumen of the intestinal and nasal mucosa and transport them to the underlying DCs (Figure 2). The DCs carry Ags into the inductive sites of the Peyer''s patch or via draining lymphatics into the mesenteric lymph nodes (MLNs) for initiation of mucosal T and B cell responses (Figure 2). Retinoic acid (RA) producing DCs enhance the expression of mucosal homing receptors (α₄β₇ and CCR9) on activated T cells for subsequent migration through the lymphatics, the bloodstream, and into the GI tract lamina propria [15],[16]. Regulation within the MIS is critical; several T cell subsets including Th1, Th2, Th17, and Tregs serve this purpose [13],[14],[17] (Figure 2).Open in a separate windowFigure 2The mucosal immune system (MIS) is interconnected, enabling it to protect vast surface areas.This is accomplished by inductive sites of organized lymphoid tissues, e.g., in the gut the Peyer''s patches (PPs) and mesenteric lymph nodes (MLNs) comprise the GALT. Lumenal Ags can be easily sampled via M cells or by epithelial DCs since this surface is not covered by mucus due to an absence of goblet cells. Engested Ags in DCs trigger specific T and B cell responses in Peyer''s patches and MLNs. Homing of lymphocytes expressing specific receptors helps guide their eventual entry into major effector tissues, e.g., the lamina propria of the gut, the upper respiratory (UR) tract, the female reproductive tract, or acinar regions of exocrine glands. Terminal differentiation of plasma cells producing polymeric (mainly dimeric) IgA is then transported across ECs via the pIgR for subsequent release as S-IgA Abs.

Box 1. Major Inductive Sites for Mucosal Immune Responses

1. GALT (gut-associated lymphoid tissues)
   • Peyer''s patches (PPs)
   • Mesenteric lymph nodes (MLNs)
   • Isolated lymphoid follicles (ILFs)
2. NALT (nasopharyngeal-associated lymphoid tissues)
   • Tonsils/adenoids
   • Inducible bronchus-associated lymphoid tissue (iBALT)
   • Cervical lymph nodes (CLNs)
   • Hilar lymph nodes (HLNs)

Mucosal effector sites, including the lamina propria regions of the GI, the UR and female reproductive tracts as well as secretory glandular tissues (i.e., mammary, lacrimal, salivary, etc.) contain Ag-specific mucosal effector cells such as IgA-producing plasma cells, and memory B and T cells [18]. Adaptive mucosal immune responses result from CD4⁺ T cell help (provided by both CD4⁺ Th2 or CD4⁺ Th1 cells), which supports the development of IgA-producing plasma cells (Figure 2). Again, the ECs become a central player in the MIS by producing the polymeric Ig receptor (pIgR) (which binds both polymeric IgA and IgM) [19]. Lamina propria pIgA binds the pIgR on the basal surface of ECs, the bound pIgA is internalized, and then transported apically across the ECs (Figure 2). Release of pIgA bound to a portion of pIgR gives rise to secretory IgA (S-IgA) antibodies (Abs) with specificities for various Ags encountered in mucosal inductive sites [13],[14],[19]. In addition, commensal bacteria are ingested by epithelial DCs, which subsequently migrate to MLNs for induction of T cell–independent, IgA B cell responses [20]. In summary, two broad types of S-IgA Abs reach our external secretions by transport across ECs and protect the epithelial surfaces from environmental insults, including infectious diseases.It should be emphasized that several unique vaccine strategies are being developed to induce protective mucosal immunity. In this regard, delivery of mucosal vaccines by oral, nasal, or other mucosal routes requires specific adjuvants or delivery systems to initiate an immune response in MALT [21],[22]. However, a major benefit of mucosal vaccine delivery is the simultaneous induction of systemic immunity, including CD4⁺ Th1 and Th2, CD8⁺ cytotoxic T lymphocytes (CTLs), and Ab responses in the bloodstream, which are predominantly of the IgG isotype [21]. This, of course, provides a double layer of immunity in order to protect the host from microbial pathogens encountered by mucosal routes. This is especially promising for development of vaccines for developing countries, as well as those to protect our aging population [23],[24].In this issue of PLOS Biology, Lau et al. used a multi-scale in vivo systems approach to assess how cells of the intestinal MIS communicate with intestinal ECs in response to an inflammatory signal [25]. The present study centered on the use of the proinflammatory cytokine tumor necrosis factor-alpha (TNF-α) given intravenously (i.v.) to assess its effects on the gut epithelium in the presence (wild-type [WT] mice) or absence (Rag1 knockout mice) of adaptive T and B lymphocytes. It is well known that TNF-α regulates many EC effects, including programmed cell death (apoptosis), survival, proliferation, cell cycle arrest, and terminal differentiation [26]. The authors had previously shown that TNF-α given i.v. to WT mice resulted in two different response patterns in the small intestine [27]. In the duodenum, which adjoins the stomach, TNF-α enhanced EC apoptosis, while in the ileum, the part next to the colon, an enhancement of EC division was seen [27]. In the present study, i.v. injection of TNF-α induced apoptosis in the duodenum (but not ileum) of WT, with heightened cell death in Rag1 mice [25]. Loss of either T or B lymphocytes also led to increased EC apoptosis, suggesting that both cell types are required to protect the epithelium from cell death. Also intriguing was the finding that eliminating the gut microbiota by antibiotic treatment did not affect the degree of EC apoptosis seen. Mathematical modeling allowed the group to show that TNF-α-induced apoptosis involved several steps in mice lacking functional T and B cells. Analysis of potential cytokines involved revealed that only a single chemokine, monocyte chemotactic protein-1 (MCP-1, C-C motif ligand 2 [CCL2]), protected ECs from apoptosis [25]. This new finding complements recent studies showing that IL-22, which is produced by several immune cells in the gut, plays a major role in protecting ECs from inflammation, infection, and tissue damage (Figure 3) [28].Open in a separate windowFigure 3The gut epithelium exhibits several pathways that protect the integrity of this organ.Intestinal epithelial cells (ECs) produce stem cell factor (SCF), which induces proliferation and resistance to bacterial invasion. In addition, neighboring γδ intraepithelial lymphocytes (IELs) produce keratinocyte growth factor (KGF), which also stabilizes ECs. IL-22 produced by Th17, Th22, and γδ T cells as well as natural killer (NK) and lymphoid tissue inducer (LTi) cells plays a key role in both early and late phases of innate immunity in order to maintain the EC barrier. In addition, monocyte chemotactic protein (MCP-1) produced by Paneth cells and goblet cells down-regulates migration of plasmacytoid DCs (pDCs) into the intestinal lamina propria in order to decrease TNF-α-induced EC apoptosis.Several unexpected discoveries followed. First, both goblet and Paneth cells were the major sources of MCP-1, and not the lymphoid cell populations that normally produce this chemokine (Figure 3). Second, the MCP-1 produced did not directly protect ECs, but instead acted via downregulation of plasmacytoid DCs (pDCs), a lymphocyte-like DC that produces various cytokines [29]. Finally, the study established that loss of adaptive (immune) lymphocytes resulted in decreased MCP-1 production, leading to increased pDC numbers and enhanced EC apoptosis. In the final experiment, the authors again showed that pDCs in the duodenum of Rag-1 mice produced increased levels of interferon-gamma that directly induced EC apoptosis. The message given by this intricate study is that systems biology approaches are quite useful in unraveling the complexities posed by the MIS in both health and disease.The model developed by Lau et al. [25] could be useful to study several major problem areas. For example, a paucity of murine models exist to study food or milk allergies that usually affect the duodenum of the small intestine [30]. It is known that chemokine receptors control trafficking of Th2-type cells to the small intestine for IgE-dependent allergic diarrhea [31],[32]. The multi-scale systems approach could be used to assess much earlier responses to food or milk allergies in TNF-α-treated mice. A second avenue could well include the cell and molecular interactions that lead to intestinal EC damage resulting in IBD [33]. Clearly, progress is being made to study genetic aspects, regulatory T cells, and the contributions of the host microbiota to IBD development [17],[34]. Nevertheless, current mouse models have their “readout” as weight loss and chronic inflammation of the colon [17],[33],[34]. The Lau et al. approach could reveal cell-to-cell linkages that ultimately resulted in EC damage [25]. Further, this approach could reveal the earliest stages of pathogenesis of IBD before the influx of inflammatory cells causes the macroscopic changes characteristic of these diseases. Since the duodenum is normally sterile, one could have predicted their finding that antibiotic treatment to remove the gut microbiota would indeed be without effect. However, one wonders what effects would be seen in the stomach or in the colon, both of which can harbor a natural microbiota. Does TNF-α and antibiotic treatment alter the EC program in these mucosal tissue sites?Finally, the intriguing question arising from the Lau et al. study [25] involves the finding that a full-blown adaptive immune system was required to maintain homeostasis and thus reduce EC apoptosis in the GI tract. Note that the response to in vivo TNF-α was assessed after only 4 hours, well before T and B cell responses could be manifested. How are early T and B cell signals transmitted to ECs? What are the mediators involved between the innate and adaptive components in the MIS for communication with the epithelium? As always, insightful studies raise many more questions than are answered. Nevertheless, the multi-scale in vivo systems analysis identified effects on the epithelium in a manner not appreciated up to now. The advantages of using an in vivo perturbation system is far superior to cell culture studies where only a few cell types are present.     

What Can Article-Level Metrics Do for You?

Article-level metrics (ALMs) provide a wide range of metrics about the uptake of an individual journal article by the scientific community after publication. They include citations, usage statistics, discussions in online comments and social media, social bookmarking, and recommendations. In this essay, we describe why article-level metrics are an important extension of traditional citation-based journal metrics and provide a number of example from ALM data collected for PLOS Biology.The scientific impact of a particular piece of research is reflected in how this work is taken up by the scientific community. The first systematic approach that was used to assess impact, based on the technology available at the time, was to track citations and aggregate them by journal. This strategy is not only no longer necessary—since now we can easily track citations for individual articles—but also, and more importantly, journal-based metrics are now considered a poor performance measure for individual articles [1],[2]. One major problem with journal-based metrics is the variation in citations per article, which means that a small percentage of articles can skew, and are responsible for, the majority of the journal-based citation impact factor, as shown by Campbell [1] for the 2004 Nature Journal Impact Factor. Figure 1 further illustrates this point, showing the wide distribution of citation counts between PLOS Biology research articles published in 2010. PLOS Biology research articles published in 2010 have been cited a median 19 times to date in Scopus, but 10% of them have been cited 50 or more times, and two articles [3],[4] more than 300 times. PLOS Biology metrics are used as examples throughout this essay, and the dataset is available in the supporting information (Data S1). Similar data are available for an increasing number of other publications and organizations.Open in a separate windowFigure 1Citation counts for PLOS Biology articles published in 2010.Scopus citation counts plotted as a probability distribution for all 197 PLOS Biology research articles published in 2010. Data collected May 20, 2013. Median 19 citations; 10% of papers have at least 50 citations.Scientific impact is a multi-dimensional construct that can not be adequately measured by any single indicator [2],[5],[6]. To this end, PLOS has collected and displayed a variety of metrics for all its articles since 2009. The array of different categorised article-level metrics (ALMs) used and provided by PLOS as of August 2013 are shown in Figure 2. In addition to citations and usage statistics, i.e., how often an article has been viewed and downloaded, PLOS also collects metrics about: how often an article has been saved in online reference managers, such as Mendeley; how often an article has been discussed in its comments section online, and also in science blogs or in social media; and how often an article has been recommended by other scientists. These additional metrics provide valuable information that we would miss if we only consider citations. Two important shortcomings of citation-based metrics are that (1) they take years to accumulate and (2) citation analysis is not always the best indicator of impact in more practical fields, such as clinical medicine [7]. Usage statistics often better reflect the impact of work in more practical fields, and they also sometimes better highlight articles of general interest (for example, the 2006 PLOS Biology article on the citation advantage of Open Access articles [8], one of the 10 most-viewed articles published in PLOS Biology).Open in a separate windowFigure 2Article-level metrics used by PLOS in August 2013 and their categories.Taken from [10] with permission by the authors.A bubble chart showing all 2010 PLOS Biology articles (Figure 3) gives a good overview of the year''s views and citations, plus it shows the influence that the article type (as indicated by dot color) has on an article''s performance as measured by these metrics. The weekly PLOS Biology publication schedule is reflected in this figure, with articles published on the same day present in a vertical line. Figure 3 also shows that the two most highly cited 2010 PLOS Biology research articles are also among the most viewed (indicated by the red arrows), but overall there isn''t a strong correlation between citations and views. The most-viewed article published in 2010 in PLOS Biology is an essay on Darwinian selection in robots [9]. Detailed usage statistics also allow speculatulation about the different ways that readers access and make use of published literature; some articles are browsed or read online due to general interest while others that are downloaded (and perhaps also printed) may reflect the reader''s intention to look at the data and results in detail and to return to the article more than once.Open in a separate windowFigure 3Views vs. citations for PLOS Biology articles published in 2010.All 304 PLOS Biology articles published in 2010. Bubble size correlates with number of Scopus citations. Research articles are labeled green; all other articles are grey. Red arrows indicate the two most highly cited papers. Data collected May 20, 2013.When readers first see an interesting article, their response is often to view or download it. By contrast, a citation may be one of the last outcomes of their interest, occuring only about 1 in 300 times a PLOS paper is viewed online. A lot of things happen in between these potential responses, ranging from discussions in comments, social media, and blogs, to bookmarking, to linking from websites. These activities are usually subsumed under the term “altmetrics,” and their variety can be overwhelming. Therefore, it helps to group them together into categories, and several organizations, including PLOS, are using the category labels of Viewed, Cited, Saved, Discussed, and Recommended (Figures 2 and ​and4,4, see also [10]).Open in a separate windowFigure 4Article-level metrics for PLOS Biology.Proportion of all 1,706 PLOS Biology research articles published up to May 20, 2013 mentioned by particular article-level metrics source. Colors indicate categories (Viewed, Cited, Saved, Discussed, Recommended), as used on the PLOS website.All PLOS Biology articles are viewed and downloaded, and almost all of them (all research articles and nearly all front matter) will be cited sooner or later. Almost all of them will also be bookmarked in online reference managers, such as Mendeley, but the percentage of articles that are discussed online is much smaller. Some of these percentages are time dependent; the use of social media discussion platforms, such as Twitter and Facebook for example, has increased in recent years (93% of PLOS Biology research articles published since June 2012 have been discussed on Twitter, and 63% mentioned on Facebook). These are the locations where most of the online discussion around published articles currently seems to take place; the percentage of papers with comments on the PLOS website or that have science blog posts written about them is much smaller. Not all of this online discussion is about research articles, and perhaps, not surprisingly, the most-tweeted PLOS article overall (with more than 1,100 tweets) is a PLOS Biology perspective on the use of social media for scientists [11].Some metrics are not so much indicators of a broad online discussion, but rather focus on highlighting articles of particular interest. For example, science blogs allow a more detailed discussion of an article as compared to comments or tweets, and journals themselves sometimes choose to highlight a paper on their own blogs, allowing for a more digestible explanation of the science for the non-expert reader [12]. Coverage by other bloggers also serves the same purpose; a good example of this is one recent post on the OpenHelix Blog [13] that contains video footage of the second author of a 2010 PLOS Biology article [14] discussing the turkey genome.F1000Prime, a commercial service of recommendations by expert scientists, was added to the PLOS Article-Level Metrics in August 2013. We now highlight on the PLOS website when any articles have received at least one recommendation within F1000Prime. We also monitor when an article has been cited within the widely used modern-day online encyclopedia, Wikipedia. A good example of the latter is the Tasmanian devil Wikipedia page [15] that links to a PLOS Biology research article published in 2010 [16]. While a F1000Prime recommendation is a strong endorsement from peer(s) in the scientific community, being included in a Wikipedia page is akin to making it into a textbook about the subject area and being read by a much wider audience that goes beyond the scientific community. PLOS Biology is the PLOS journal with the highest percentage of articles recommended in F1000Prime and mentioned in Wikipedia, but there is only partial overlap between the two groups of articles because they focus on different audiences (Figure 5). These recommendations and mentions in turn show correlations with other metrics, but not simple ones; you can''t assume, for example, that highly cited articles are more likely to be recommended by F1000Prime, so it will be interesting to monitor these trends now that we include this information.Open in a separate windowFigure 5 PLOS Biology articles: sites of recommendation and discussion.Number of PLOS Biology research articles published up to May 20, 2013 that have been recommended by F1000Prime (red) or mentioned in Wikipedia (blue).With the increasing availability of ALM data, there comes a growing need to provide tools that will allow the community to interrogate them. A good first step for researchers, research administrators, and others interested in looking at the metrics of a larger set of PLOS articles is the recently launched ALM Reports tool [17]. There are also a growing number of service providers, including Altmetric.com [18], ImpactStory [19], and Plum Analytics [20] that provide similar services for articles from other publishers.As article-level metrics become increasingly used by publishers, funders, universities, and researchers, one of the major challenges to overcome is ensuring that standards and best practices are widely adopted and understood. The National Information Standards Organization (NISO) was recently awarded a grant by the Alfred P. Sloan Foundation to work on this [21], and PLOS is actively involved in this project. We look forward to further developing our article-level metrics and to having them adopted by other publishers, which hopefully will pave the way to their wide incorporation into research and researcher assessments.     

Status and the Brain

Social hierarchy is a fact of life for many animals. Navigating social hierarchy requires understanding one''s own status relative to others and behaving accordingly, while achieving higher status may call upon cunning and strategic thinking. The neural mechanisms mediating social status have become increasingly well understood in invertebrates and model organisms like fish and mice but until recently have remained more opaque in humans and other primates. In a new study in this issue, Noonan and colleagues explore the neural correlates of social rank in macaques. Using both structural and functional brain imaging, they found neural changes associated with individual monkeys'' social status, including alterations in the amygdala, hypothalamus, and brainstem—areas previously implicated in dominance-related behavior in other vertebrates. A separate but related network in the temporal and prefrontal cortex appears to mediate more cognitive aspects of strategic social behavior. These findings begin to delineate the neural circuits that enable us to navigate our own social worlds. A major remaining challenge is identifying how these networks contribute functionally to our social lives, which may open new avenues for developing innovative treatments for social disorders.

“Observing the habitual and almost sacred ‘pecking order’ which prevails among the hens in his poultry yard—hen A pecking hen B, but not being pecked by it, hen B pecking hen C and so forth—the politician will meditate on the Catholic hierarchy and Fascism.” —Aldous Huxley, Point Counter Point (1929)

From the schoolyard to the boardroom, we are all, sometimes painfully, familiar with the pecking order. First documented by the Norwegian zoologist Thorleif Schjelderup-Ebbe in his PhD thesis on social status in chickens in the 1920s, a pecking order is a hierarchical social system in which each individual is ranked in order of dominance [1]. In chickens, the top hen can peck all lower birds, the second-ranking bird can peck all birds ranked below her, and so on. Since it was first coined, the term has become widely applied to any such hierarchical system, from business, to government, to the playground, to the military.Social hierarchy is a fact of life not only for humans and chickens but also for most highly social, group-living animals. Navigating social hierarchies and achieving dominance often appear to require cunning, intelligence, and strategic social planning. Indeed, the Renaissance Italian politician and writer Niccolo Machiavelli argued in his best-known book “The Prince” that the traits most useful for attaining and holding on to power include manipulation and deception [2]. Since then, the term “Machiavellian” has come to signify a person who deceives and manipulates others for personal advantage and power. 350 years later, Frans de Waal applied the term Machiavellian to social maneuvering by chimpanzees in his book Chimpanzee Politics [3]. De Waal argued that chimpanzees, like Renaissance Italian politicians, apply guile, manipulation, strategic alliance formation, and deception to enhance their social status—in this case, not to win fortune and influence but to increase their reproductive success (which is presumably the evolutionary origin of status-seeking in Renaissance Italian politicians as well).The observation that navigating large, complex social groups in chimpanzees and many other primates seems to require sophisticated cognitive abilities spurred the development of the social brain hypothesis, originally proposed to explain why primates have larger brains for their body size than do other animals [4],[5]. Since its first proposal, the social brain hypothesis has accrued ample evidence endorsing the connections between increased social network complexity, enhanced social cognition, and larger brains. For example, among primates, neorcortex size, adjusted for the size of the brain or body, varies with group size [6],[7], frequency of social play [8], and social learning [9].Of course, all neuroscientists know that when it comes to brains, size isn''t everything [10]. Presumably social cognitive functions required for strategic social behavior are mediated by specific neural circuits. Here, we summarize and discuss several recent discoveries, focusing on an article by Noonan and colleagues in the current issue, which together begin to delineate the specific neural circuits that mediate our ability to navigate our social worlds.Using structural magnetic resonance imaging (MRI), Bickart and colleagues showed that the size of the amygdala—a brain nucleus important for emotion, vigilance, and rapid behavioral responses—is correlated with social network size in humans [11]. Subsequent studies showed similar relationships for other brain regions implicated in social function, including the orbitofrontal cortex [12] and ventromedial prefrontal cortex [13]. Indeed, one study even found an association between grey matter density in the superior temporal sulcus (STS) and temporal gyrus and an individual''s number of Facebook friends [14]. Collectively, these studies suggest that the number and possibly the complexity of relationships one maintains varies with the structural organization of a specific network of brain regions, which are recruited when people perform tests of social cognition such as recognizing faces or inferring others'' mental states [15],[16]. These studies, however, do not reveal whether social complexity actively changes these brain areas through plasticity or whether individual differences in the structure of these networks ultimately determines social abilities.To address this question, Sallet and colleagues experimentally assigned rhesus macaques to social groups of different sizes and then scanned their brains with MRI [17]. The authors found significant positive associations between social network size and morphology in mid-STS, rostral STS, inferior temporal (IT) gyrus, rostral prefrontal cortex (rPFC), temporal pole, and amygdala. The authors also found a different region in rPFC that scaled positively with social rank; as grey matter in this region increased, so did the monkey''s rank in the hierarchy. As in the human studies described previously, many of these regions are implicated in various aspects of social cognition and perception [18]. These findings endorse the idea that neural plasticity is engaged in specifically social brain areas in response to the demands of the social environment, changing these areas structurally according to an individual''s experiences with others.Sallet and colleagues also examined spontaneous coactivation among these regions using functional MRI (fMRI). Measures of coactivation are thought to reflect coupling between regions [19],[20]; these measures are observable in many species [21],[22] and vary according to behavior [23],[24], genetics [25], and sex [26], suggesting that coactivation may underlie basic neural function and interaction between brain regions. The authors found that coactivation between the STS and rPFC increased with social network size and that coactivation between IT and rPFC increased with social rank. These findings show that not only do structural changes occur in these regions to meet the demands of the social environment but these structural changes mediate changes in function as well.One important question raised by the study by Sallet and colleagues is whether changes in the structure and function of social brain areas are specific outcomes of social network size or of dealing with social hierarchy. After all, larger groups offer more opportunity for a larger, more despotic pecking order. In the current volume, Noonan and colleagues address this question directly by examining the structural and functional correlates of social status in macaques independently of social group size [27]. The authors collected MRI scans from rhesus macaques and measured changes in grey matter associated with social dominance. By scanning monkeys of different ranks living in groups of different sizes, the authors were able to cleave the effects of social rank from those of social network size (Figure 1).Open in a separate windowFigure 1Brain regions in rhesus macaques related to social environment.Primary colors indicate brain regions in which morphometry tracks social network size. Pastel colors indicate brain regions in which morphometry tracks social status in the hierarchy. Regions of interest adapted from [48], overlaid on Montreal Neurological Institute (MNI) macaque template [49].The authors found a network of regions in which grey matter measures varied with social rank; these regions included the bilateral central amygdala, bilateral brainstem (between the medulla and midbrain, including parts of the raphe nuclei), and hypothalamus, which varied positively with dominance, and regions in the basal ganglia, which varied negatively with social rank. These regions have been implicated in social rank functions across a number of species [28]–[32]. Importantly, these relationships were unique to social status. There was no relationship between grey matter in these subcortical areas and social network size, endorsing a specific role in social dominance-related behavior. Nevertheless, grey matter in bilateral mid-STS and rPFC varied with both social rank and social network size, as reported previously. These findings demonstrate that specific brain areas uniquely mediate functions related to social hierarchy, whereas others may subserve more general social cognitive processes.Noonan and colleagues next probed spontaneous coactivation using fMRI to examine whether functional coupling between any of these regions varied with social status. They found that the more subordinate an animal, the stronger the functional coupling between multiple regions related to dominance. These results suggest that individual differences in social status are functionally observable in the brain even while the animal is at rest and not engaged in social behavior. These findings suggest that structural changes associated with individual differences in social status alter baseline brain function, consistent with the idea that the default mode of the brain is social [33] and that the sense of self and perhaps even awareness emerge from inwardly directed social reasoning [34].These findings resonate with previous work on the neural basis of social dominance in other vertebrates. In humans, for example, activity in the amygdala tracks knowledge of social hierarchy [28],[35] and, further, shows activity patterns that uniquely encode social rank and predict relevant behaviors [28]. Moreover, recent research has identified a specific region in the mouse hypothalamus, aptly named the “hypothalamic attack area” [36],[37]. Stimulating neurons in this area immediately triggers attacks on other mice and even an inflated rubber glove, while inactivating these neurons suppresses aggression [38]. In the African cichlid fish Haplochromis burtoni, a change in the social status of an individual male induces a reversible change in the abundance of specialized neurons in the hypothalamus that communicate hormonally with the pituitary and gonads [39]. Injections of this hormone in male birds after an aggressive territorial encounter amplifies the normal subsequent rise in testosterone [40]. Serotonin neurons in the raphe area of the brainstem also contribute to dominance-related behaviors in fish [29],[31] and aggression in monkeys [41].Despite these advances, there are still gaps in our understanding of how these circuits mediate status-related behaviors. Though regions in the amygdala, brainstem, and hypothalamus vary structurally and functionally with social rank, it remains unknown precisely how they contribute to or respond to social status. For example, though amygdala function and structure correlates with social status in both humans and nonhuman primates [27],[28],[35],[42], it remains unknown which aspects of dominance this region contributes to or underlies. One model suggests that the amygdala contributes to learning or representing one''s own status within a social hierarchy [28],[35]. Alternatively, the amygdala could contribute to behaviors that support social hierarchy, including gaze following [43] and theory of mind [44]. Lastly, the amygdala could contribute to social rank via interpersonal behaviors or personality traits, such as aggression [45], grooming [45], or fear responses [46],[47]. Future work will be critical to determine how signals in these regions relate to social status; direct manipulation of these regions, possibly via microstimulation, larger-scale brain stimulation (e.g., transcranial magnetic stimulation and transcranial direct current stimulation), or temporary lesions, will be critical to better understand these relationships.The work by Noonan and colleagues suggests new avenues for exploring how the brain both responds to and makes possible social hierarchy in nonhuman primates and humans. The fact that the neural circuits mediating dominance and social networking behavior can be identified and measured from structural and functional brain scans even at rest suggests the possibility that similar measures can be made in humans. Although social status is much more complex in people than it is in monkeys or fish, it is just as critical for us and most likely depends on shared neural circuits. Understanding how these circuits work, how they develop, and how they respond to the local social environment may help us to understand and ultimately treat disorders, like autism, social anxiety, or psychopathy, that are characterized by impaired social behavior and cognition.     

An inordinate fondness for eukaryotic diversity

Why do some groups of organisms, like beetles, have so many species, and others, like the tuataras, so few? This classic question in evolutionary biology has a deep history and has been studied using both fossils and phylogenetic trees. Phylogeny-based studies have focused on tree balance, which compares the number of species across clades of the same age in the tree. These studies have suggested that rates of speciation and extinction vary tremendously across the tree of life. In this issue, Rabosky et al. report the most ambitious study to date on the differences in species diversity across clades in the tree of life. The authors bring together a tremendously large dataset of multicellular eukaryotes, including all living species of plants, animals, and fungi; they divide these organisms into 1,397 clades, accounting for more than 1.2 million species in total. Rabosky et al. find tremendous variation in diversity across the tree of life. There are old clades with few species, young clades with many species, and everything in between. They also note a peculiar aspect of their data: it is difficult or impossible to predict how many species will be found in a particular clade knowing how long a clade has been diversifying from a common ancestor. This pattern suggests complex dynamics of speciation and extinction in the history of eukaryotes. Rabosky et al.''s paper represents the latest development in our efforts to understand the Earth''s biodiversity at the broadest scales.J.B.S. Haldane—one of the most quotable of all evolutionary biologists—had a favorite saying about what patterns of species richness tell us about the nature of the Universe: “God has an inordinate fondness for beetles” (see [1] for more details). With this quip, Haldane is referring to the overwhelming number of beetle species on Earth. We still don''t know exactly how many species of beetles there are on the Earth—perhaps around 400,000—but certainly, there are a lot.One of the primary mysteries in macroevolution is the tremendous difference in numbers of species among different taxonomic groups. Modern systematists classify species into clades (groups of species that represent all of the descendents of a common ancestor, like turtles or arthropods). Different clades in the tree of life have dramatically different diversities. This might not be surprising—after all, species in one clade can be distinct from other species in size, energy use, and a thousand other ways. Also, some clades are much older than others. However, even when we control for differences in age by comparing sister clades—that is, pairs of clades that are each others'' closest relative—we still see profound differences in number of species. For example, there are currently two living species of tuatara, a clade of lizard-like reptiles that currently inhabit small islands around New Zealand (Figure 1, left). These tuatara are the sister clade to the squamates, a clade of 7,000 species that includes all living snakes and lizards (Figure 1, right) [2]. Since these two groups are sister clades, they diverged from a common ancestor at exactly the same time (∼250 million years ago [3]). Tuataras used to be far more diverse in the past (though almost certainly not as diverse as squamates [4]), but their current diversity is dwarfed by the tremendous number and variety of snakes and lizards around the globe. Similar patterns occur across the whole tree of life. In fact, old, low diversity clades contain some of the most enigmatic species on Earth: ginkgo trees, coelacanths, tailed frogs, horseshoe crabs, and monotremes, among others. These species are sometimes called “living fossils,” although only some of them are actually thought to resemble their ancient ancestors [5].Open in a separate windowFigure 1The tuatara (Sphenodon punctatus, left) is one of only two surviving species in the clade Sphenodontia.The sister clade to the tuatara is Squamata, which includes the ∼7,000 living species of snakes and lizards, including the ornate day gecko (Phelsuma ornata, right). (Left) from Wikimedia commons, taken by user KeresH, http://commons.wikimedia.org/wiki/File:Henry_at_Invercargill.jpg; (Right) by the author.We can learn a lot about the differences in diversity across clades from phylogenetic trees. In particular, phylogenetic tree balance summarizes the pattern of differences in the number of species between sister clades across a whole phylogeny [6]. A phylogenetic tree can be completely balanced, such that each pair of sister taxa in the tree have exactly the same number of species (Figure 2A; this is only possible if the number of species in the tree is a power of 2: 2, 4, 8, 16, 32, etc.). A phylogenetic tree can also be completely imbalanced, so that every comparison of sister clades has a single species in one clade and the remainder in the other (such a tree is also called pectinate; Figure 2B). There are a few different ways to quantify tree balance, but they all work in basically the same way: compare the number of species between sister clades in the tree, and summarize those differences across a whole phylogeny [7].Open in a separate windowFigure 2Balanced (A), imbalanced (B), and random birth-death (C) phylogenetic trees of eight species (a–h).Imagine for a moment that you have the tree of life (a phylogenetic tree of all species on Earth). We don''t have such a tree yet, but scientists are moving in that direction and trees are getting bigger and bigger (e.g., [8]). The tree of life is a huge and complicated structure, and—as one might imagine at a scale that encompasses all living things, bacteria to beetles to beagles—resists generalizations. But even as the tree of life takes shape, we already know that it is highly imbalanced. This statement applies broadly across living things, and applies equally well to plants as it does to animals, and everything else (as far as we know). Dramatic differences in diversity among clades is a characteristic feature of life on Earth.To better understand patterns of balance in the tree of life, we can start with the birth-death model, a simple model of how phylogenetic trees might grow through time (reviewed in [9]). Under a birth-death model, phylogenetic trees “grow” through time following two processes: speciation, where one species splits into two, and extinction, when one species dies out. For simple birth-death models we assume that both of these processes happen at a constant rate through time for each species alive at that moment, and use the parameters λ (speciation rate) and μ (extinction rate). If we simulate this birth-death model using a computer, we will obtain a phylogenetic tree (Figure 2C). If the extinction rate is greater than zero, such a tree will include surviving species as well as species that have gone extinct. Since we will be comparing this tree to phylogenies of living species, we can assume that any species that went extinct before the present has been pruned. We can then measure the balance of the resulting birth-death tree (our simulated tree will also have branch lengths, but we will ignore those for now). If we repeat this process a large number of times, we obtain a statistical distribution of our tree balance measure, which represents the expectation of that distribution under the birth-death model. It turns out, perhaps surprisingly, that this distribution of balance depends only on the fact that the trees are simulated under a birth-death model; in terms of tree balance, the actual rates of speciation and extinction do not matter, as long as they are constant across clades [10].There is another counterintuitive feature of the balance of birth-death trees: these trees are surprisingly imbalanced compared to what, perhaps, your intuition might suggest. For example, imagine that a certain phylogenetic tree contains 100 species. If you look at the deepest split in that tree and compare the diversity of the two sister clades, what do you expect to find? Are you more likely to find an even number of species in each of these two clades—say, 50/50 or 49/51—or a very uneven number, like 2/98 or 1/99? The surprising answer is that all four of these listed possibilities are equally likely. In fact, all possible combinations of diversity for each of the two clades are equally probable [11].This property of birth-death models means that birth-death trees are quite imbalanced: it is not uncommon, for example, to find sister clades that differ in diversity by a factor of 20 just by random chance. However, the tree of life is imbalanced even compared to birth-death trees! This general observation was first discovered in an influential paper by Arne Mooers and colleagues [7]. In that paper, the authors measured the balance of phylogenetic trees that had been reconstructed using trait data, DNA sequences, or both. They then compared their balance to what we might expect under a birth-death model. They found that phylogenetic trees from a wide range of taxa are extraordinarily imbalanced. This paper showed that the general “shape” of the tree of life is highly imbalanced.The classical interpretation of the imbalanced tree of life is that clades vary in their rates of speciation and/or extinction. There are many reasons to suspect that species in some clades might speciate more frequently, or go extinct less frequently, than their relatives. For example, perhaps a species'' range affects its probability of speciating or going extinct—as suggested by a recent paper in PLOS Biology by Pigot et al. [12]—so that clades of species with different distributions of range sizes will experience different rates of diversification. Many studies have attempted to measure speciation and extinction rates in groups with a good fossil record, and compare these rates across different types of organisms and time periods (e.g., [13],[14]). These studies have generally found wide variation in both rates and in their difference (speciation−extinction = net diversification rate).Recent studies go beyond measures of tree balance by using the tree''s branch lengths to gain information about speciation rates. One simple way to do this is to compare the diversity of a clade to its age; one can then estimate the speciation rate as λ = ln(n)/t, where n is the number of living species in the clade and t is the clade''s stem age (the time since divergence from the clade''s sister group) [15]. There are also modifications of this equation that incorporate extinction and that can use the clade''s crown age (the time since all living species in the clade shared a common ancestor; see [15]).Perhaps surprisingly, one can even estimate extinction from a phylogenetic tree based only on living species [16]. This is because old and young lineages are hit by extinction with different probabilities: since young lineages have not been around very long, they are less likely to have gone extinct than older lineages. The phenomenon is called the “pull of the present” [16] and means that extinction leads to an overabundance of very young lineages in a tree. We can look for this pull in patterns of lineage accumulation through time, which can thus be used to estimate both speciation and extinction rates and to compare these rates across clades (e.g., [17]; but see [18], which points out that this method does not work well when its assumptions of rate constancy are strongly violated). More recently, new methods have been developed to search for variation in speciation and/or extinction rates across large phylogenetic trees, and to try to correlate these rates with the traits of lineages [19],[20].In this issue, Rabosky et al. [21] attempt the most ambitious study to date investigating the differences in species diversity across clades in the tree of life. The authors bring together a tremendously large dataset that spans the multicellular eukaryotes, including all living species of plants, animals, and fungi. For each of 1,397 eukaryotic clades, the authors gathered estimates of age and diversity from the literature – accounting for more than 1.2 million species in total. The authors also summarize the evolutionary relationships among these clades using a “backbone” phylogenetic tree with branch lengths in millions of years. This provides a remarkably complete view of what we currently know about the species diversity of clades across a huge section of the tree of life.The diversity data in Rabosky et al. [21] are broadly consistent with the historical background above: there are major differences in diversification rates across the tree of life. There are old clades with few species, young clades with many species, and everything in between. But Rabosky et al. also note a peculiar aspect of their data: there is typically either a very weak or no relationship between the number of species in a clade and its age. That is, in the data they analyze, it is difficult to guess how many species are in a clade on the basis of how long it has been diversifying from a common ancestor. The traditional explanation for this pattern would be differences in diversification rates across clades—although the authors use simulations to show that, at least under one scenario about how rates might vary across trees, one rarely finds such weak or absent relationships between age and diversity. The authors speculate about other possible explanations for this peculiar (lack of) pattern, from bias in the way clades are named to ecological processes that limit the number of coexisting, competing species.Rabosky et al.''s [21] analysis is not the final chapter; the tree of life is still under construction, and the total number of species in some clades is best viewed as an educated guess. Specifically, I suspect that we have very poor estimates of the extant diversity of many eukaryotic groups, particularly small, understudied organisms. Indeed, new techniques that use genomic sequencing to identify undiscovered species from DNA sequences from environmental samples often reveal that the species we know are only a small component of natural ecosystems [22]. One might also note that the total diversity of multicellular eukaryotes counted in this study might be a vast underestimate compared to recent estimates that use statistical analyses to correct for incomplete sampling [23]. Still, the results in Rabosky et al. [21] are intriguing and will certainly inspire further study, which I expect will be focused on testing more sophisticated mathematical models, beyond the constant-rate birth-death models prevalent today, that might be able to explain patterns in the data.I first learned of the Huxley quote that opens this article in a classic paper, “Homage to Santa Rosalia, or why are there so many kinds of animals?” [24]. This paper, written by the great ecologist G.E. Hutchinson, speculates about the mechanisms that allow so many different species to coexist in natural communities. Hutchinson describes collecting Italian water boatmen from a small pond in the shadow of Santa Rosalia''s shrine, and wondering why the pond contained two species of water beetle—no more, no less. Hutchinson says “Nothing in her history being known to the contrary, perhaps for the moment we may take Santa Rosalia as the patroness of evolutionary studies…” [24]. Rabosky et al.''s [21] paper represents the latest development in our efforts to understand why the Earth has the particular number of species that it has – no more, no less. Santa Rosalia would be proud.     

Cooperation and the Fate of Microbial Societies

Microorganisms have been cooperating with each other for billions of years: by sharing resources, communicating with each other, and joining together to form biofilms and other large structures. These cooperative behaviors benefit the colony as a whole; however, they may be costly to the individuals performing them. This raises the question of how such cooperation can arise from natural selection. Mathematical modeling is one important avenue for exploring this question. Evolutionary experiments are another, providing us with an opportunity to see evolutionary dynamics in action and allowing us to test predictions arising from mathematical models. A new study in this issue of PLOS Biology investigates the evolution of a cooperative resource-sharing behavior in yeast. Examining the competition between cooperating and “cheating” strains of yeast, the authors find that, depending on the initial mix of strains, this yeast society either evolves toward a stable coexistence or collapses for lack of cooperation. Using a simple mathematical model, they show how these dynamics arise from eco-evolutionary feedback, where changes in the frequencies of strains are coupled with changes in population size. This study and others illustrate the combined power of modeling and experiment to elucidate the origins of cooperation and other fundamental questions in evolutionary biology.How much cooperation does it take to maintain a society? Many biological populations, from microbes to insects to humans, depend on the cooperation of their members in order to access resources, raise offspring, and avoid danger. Yet in any cooperative activity, there is the risk of “cheaters,” who benefit from the generosity of others while making no contribution of their own. Consider, for example, the layabout in a communal household who refuses to cook or clean dishes. If this cheating behavior spreads through the population, the society as a whole may collapse.Evolutionary biologists since Darwin have been fascinated by how populations can overcome this dilemma. Studying this question can be challenging. While the products of evolution are evident in the natural world, the process that produced them is mostly hidden from view. As a consequence, direct observation of the evolution of cooperation in action is often limited.Much of our current understanding of this conundrum arises from mathematical modeling. Ever since the birth of population genetics about a century ago, it has been recognized that the theory of evolution can be set upon exact mathematical foundations. This approach has flourished ever since, and especially in the last few decades. The theory of choice to study social phenomena is evolutionary game theory [1]–[5], in which behaviors that affect others are represented as strategies. Simple mathematical models describe the dynamics of these strategies under mutation and selection, depending on the population structure [6]–[12]. Applied to the problem of cooperation, these models show that if a cooperating individual receives some of the benefit of his or her own labors—as in Snowdrift games or some nonlinear public goods games—then evolutionary dynamics may lead to an equilibrium in which cooperators and cheaters coexist [1],[13]. On the other hand, if benefits accrue only to others—as in Prisoners'' Dilemma games—then cooperation is expected to disappear unless some mechanism is present to support it [14].Recently, experiments with microbes have afforded us an unprecedented opportunity to observe evolution in action [15]–[20]. Bacteria, yeast, and other single-celled organisms divide rapidly enough that evolutionary change—the arrival and fixation of beneficial mutations—can be observed in the laboratory. Moreover, the experimenter is able to control the population size, environmental conditions, and other variables, and can therefore test hypotheses regarding how the course of evolution depends on these variables. Experimenters can also preserve specimens of the population from all phases of its evolution as a “living record” of genotypic and phenotypic change. In short, experiments with microbes are a powerful tool for testing evolutionary hypotheses.Microorganism experiments hold particular promise for shedding light on how cooperative behaviors emerge from evolution [21]–[26]. Microbial species cooperate in a variety of ways: They form biofilms, produce iron-scavenging agents, produce chemicals to resist antibiotics, and form fruiting bodies when local resources are depleted. By mixing wild-type strains that display a particular cooperative behavior with “cheater” mutants that do not, researchers can test hypotheses about what conditions favor wild-type “cooperators” over cheaters.In one such experiment, Gore et al. [26] studied a social dilemma in the yeast Saccharomyces cerevisiae. The preferred nutrient sources for this yeast are the simple sugars glucose and fructose; however, it can subsist on the compound sugar sucrose by producing the enzyme invertase, which breaks down sucrose into glucose and fructose. A crucial point is that, since this reaction occurs near the cell wall, only about 1% of these simple sugars are captured by the cell in which they are produced. The remaining 99% diffuse away and are available to other cells. Thus producing invertase is a cooperative behavior, with the bulk of the benefit going to cells other than the producer. Moreover, this cooperation is costly, in that the production of invertase carries a metabolic cost to the producer. To study the evolution of this behavior, Gore et al. created cheater strains that do not produce invertase, and thereby avoid the associated cost. Letting these strains compete with each other, they found that, in most cases, cooperator and cheater strains converged to an equilibrium in which both strains coexisted—a result consistent with theoretical predictions regarding Snowdrift games and nonlinear public goods games [1],[13].Much theoretical work on the evolution of cooperation and other traits has assumed, for the sake of simplicity, that the population size remains roughly constant while the strains in question are competing. However, it is entirely possible that population dynamics—changes in population size—may occur on the same timescale as evolutionary dynamics—changes in the frequencies of competing types. In this case, these two dynamical processes may affect one another, a phenomenon known as eco-evolutionary feedback [27]–[30]. Mathematical modeling has shown that eco-evolutionary feedback may lead to a variety of complex dynamical behaviors, including multiple equilibria, cycling, chaos, and Turing patterns [28],[30]–[33].In this issue of PLOS Biology, Sanchez and Gore [34] have—for the first time, to our knowledge—empirically demonstrated eco-evolutionary feedback in the evolution of cooperation. Using the yeast system described above, the authors studied the coupled dynamics of the population density and the proportion of cooperator types within the population. The mechanism for eco-evolutionary feedback in this system is intuitive: the growth of the population as a whole depends on the concentration of simple sugars, which in turn depends on the density of cooperators. If there are insufficient cooperators, the overall population density declines. With low population density, cooperators have an advantage due to the simple sugars they manage to retain for themselves. At this point, cooperators increase in frequency, and the concentration of simple sugars increases, leading to overall population growth. But once this happens, cheaters proliferate faster than cooperators due to their lower metabolic costs. This in turn depresses the frequency of cooperators, and the cycle repeats itself. We would therefore expect to see cycling or spiraling behavior in the eco-evolutionary dynamics of these types, consistent with theoretical predictions [32],[33].In their experiment, Sanchez and Gore observed not only spiraling, but also bistability—the presence of two equilibria to which the system might converge, depending on the initial conditions [35]. If the initial population density and/or the initial proportion of cooperators is too low, not enough simple sugars are produced and the population collapses. On the other hand, if there are sufficiently many cooperators in the initial population, the population converges in spiraling fashion to an equilibrium in which population density is high and cooperators and cheaters coexist (Figure 1). To complement their experiment, the authors developed a simple Lotka-Volterra–type model describing the interdependent growth of the competing strains. This model reproduces the observed eco-evolutionary dynamics with remarkable fidelity, given its simplicity.Open in a separate windowFigure 1Dynamics of eco-evolutionary feedback in cooperator and cheater strains of the yeast S. cerevisiae , as observed in the experiment of Sanchez and Gore.There are two basins of attraction, with a different outcome expected from each. If there are too few cooperators to start, not enough simple sugars are produced and the population collapses. On the other hand, if the initial number of cooperators is sufficient, the system converges in spiraling fashion to an equilibrium in which cooperators and cheaters coexist.Interestingly, the proportion of cooperators in the coexistence equilibrium is low—less than 10%—but is nonetheless sufficient to maintain the viability of the population. Does the predominance of cheaters in this equilibrium hurt the population as a whole? The authors found that the overall density and productivity of the population in the coexistence equilibrium is not much less than what cooperators would achieve in the absence of cheaters. However, the predominance of cheaters does impact the population''s resilience to an ecological shock—in this case, rapid and significant dilution of the population. Cooperators in monomorphic equilibrium survive this shock, but populations in mixed equilibrium between cooperators and cheaters do not. In short, mixed populations are comparably productive to, but significantly less resilient than, cooperator-only populations.The study of Sanchez and Gore illustrates the synergistic power of theory and experiment when carefully combined. The opportunities for further such combinations are immense. Population genetics and evolutionary game theory have provided us with a wealth of testable hypotheses about evolution, and we now have the experimental technology to test them. Some of the most interesting hypotheses regard the effect of spatial structure on the evolution of cooperation. Well-known results in evolutionary game theory show that spatial structure can promote cooperation [6],[36]–[39], though this effect depends strongly on the details of spatial reproduction and replacement [40]. Thus far, experimental studies have addressed this question only indirectly, with reduced pathogen virulence representing an indirect form of cooperation [41], or with group subdivision standing in for spatial structure [22],[23]. The effects of spatial structure on the evolution of cooperation in microbial colonies remains an important open question.At the same time, we must also allow experimental results to inform the development of new mathematical models. The field of social bacterial evolution requires well-defined, simple models that describe how populations of bacteria change over time, taking into account the reproductive events, social interactions, and population structures particular to these populations. This approach ultimately brings together the methods of population genetics, evolutionary game theory, ecology, and experimental microbiology.     

The Origins of Specialization: Insights from Bacteria Held 25 Years in Captivity

Examples of ecological specialization abound in nature but the evolutionary and genetic causes of tradeoffs across environments are typically unknown. Natural selection itself may favor traits that improve fitness in one environment but reduce fitness elsewhere. Furthermore, an absence of selection on unused traits renders them susceptible to mutational erosion by genetic drift. Experimental evolution of microbial populations allows these potentially concurrent dynamics to be evaluated directly, rather than by historical inference. The 50,000 generation (and counting) Lenski Long-Term Evolution Experiment (LTEE), in which replicate E. coli populations have been passaged in a simple environment with only glucose for carbon and energy, has inspired multiple studies of their potential specialization. Earlier in this experiment, most changes were the side effects of selection, both broadening growth potential in some conditions and narrowing it in others, particularly in assays of diet breadth and thermotolerance. The fact that replicate populations experienced similar losses suggested they were becoming specialists because of tradeoffs imposed by selection. However a new study in this issue of PLOS Biology by Nicholas Leiby and Christopher Marx revisits these lines with powerful new growth assays and finds a surprising number of functional gains as well as losses, the latter of which were enriched in populations that had evolved higher mutation rates. Thus, these populations are steadily becoming glucose specialists by the relentless pressure of mutation accumulation, which has taken 25 years to detect. More surprising, the unpredictability of functional changes suggests that we still have much to learn about how the best-studied bacterium adapts to grow on the best-studied sugar.The wonder of biological diversity belies a puzzling subtext. Species are defined as much by their limits as their capabilities. Very few species in our common vernacular tolerate life in a wide range of environments, and those that do—the Norway rat, say—are not generally appealing. More often, we celebrate specialization to a particular condition: for example, orchid epiphytes growing tenuously in the cloud forest, only a subtle climate shift from extinction. Even grade school natural history teaches us that species are often unfit when living beyond their natural range.So it comes as a surprise that the causes of this rampant ecological specialization are poorly understood. “Use it or lose it,” but why? One common explanation is that natural selection tends to favor traits that simultaneously enhance fitness in one environment but compromise fitness elsewhere. This selective process is known as “antagonistic pleiotropy.” Another explanation is that a selective shadow falls upon unused traits, rendering them susceptible to mutational erosion by random genetic drift. This neutral process is known as “mutation accumulation” (Figure 1). These processes inevitably co-occur, and can be enhanced by the hybrid dynamic of genetic hitchhiking, in which neutral mutations affecting unused functions become linked to different mutations under positive selection. In most cases, the functional decay of a species can only be studied retrospectively, and distinguishing the roles of antagonistic pleiotropy and mutation accumulation is hampered by weak historical inference. Did selection, or an absence of selection, produce the blind cavefish [1]? There is little controversy that the sum of these dynamics can produce specialists, but their timing and relative importance is an open question.Open in a separate windowFigure 1Hypothetical dynamics of fitness in foreign environments by pleiotropy or mutation accumulation during long-term adaptation.Prolonged adaptation to one environment leads to decelerating fitness gains in the selective environment (solid black line), as beneficial mutations become limiting. Consequences of this adaptation for fitness in other environments may take different forms. No net change may occur if beneficial mutations generate no or inconsistent side effects (neutrality). However, the same mutations responsible for adaptation may also increase fitness in other environments (synergistic pleiotropy, dotted line), may decrease fitness in foreign environments at an equivalent rate if antagonistic effects correlate with selected effects (antagonistic pleiotropy, dotted line), or may decrease fitness at an increasing rate if subsequent mutations generate greater tradeoffs (antagonistic pleiotropy, dashed and dotted line). The uncertainty of the form of pleiotropic effects reflects a general lack of understanding of how mutations interact to affect fitness, particularly over the long term. Mutation accumulation (MA) in traits hidden from selection is expected to reduce fitness randomly but linearly on average, more slowly during evolution at a low mutation rate (MA, low U) or more rapidly at a high mutation rate (MA, high U). Evidence of all processes is now evident in this latest study of the evolution of diet breadth in the LTEE [20].The study of “evolution in action” using model experimental populations of rapidly reproducing organisms allows researchers to quantify both adaptation and any functional declines simultaneously. This approach is especially powerful when samples of evolving populations can be stored inanimate and studied at a later time under various conditions. Perhaps the best example of this approach is Richard Lenski''s Long-Term Evolution Experiment (LTEE), in which 12 populations of E. coli have been grown under simple conditions for more than 25 years and 50,000 generations [2],[3].When as a graduate student I wondered aloud whether the LTEE lines had become specialists, a colleague remarked: “Of course! You''ve selected for streamlined E. coli that have scuttled unused functions.” But with only a small amount of glucose as the sole carbon source available to the ancestor (the innovation by one population of using citrate for growth more than 30,000 generations in the future notwithstanding [4]), all anabolic pathways to construct new cells remain under strong selection to preserve their function. Moreover, because some catabolic reactions use the same intermediates as anabolic pathways (a form of pleiotropy) [5], growth on alternative carbon sources may be nonetheless preserved. Thus, we wondered whether the physiology of E. coli might actually prove to be robust during long-term evolution on glucose alone.Over the first 2,000 generations, the LTEE lines gained more often than lost fitness across a range of different environments [6]. In addition, a high-throughput screen of cellular respiration (Biolog) for the best-studied clone from these lines showed 171 relative gains and only 32 losses [7]. Even these losses in substrate respiration did not translate to reduced fitness versus the ancestor; rather, the evolved clone was simply relatively worse in the foreign resources than in glucose [7]. Evidently, each of the five beneficial mutations found in this early clone was broadly beneficial and imparted few tradeoffs [8]. Generalists rather than specialists were the rule.Between 2,000 and 20,000 generations, fitness losses in foreign conditions became more obvious but not always consistent. Some lines became less fit than the ancestor in a dilute complex medium (LB) [9], all lines grew worse at high (>40°C) and low (<20°C) temperature [10], and all lines became sensitive to the resource concentration in their environment, even for glucose [9]. Did subsequent beneficial mutations cause these tradeoffs (antagonistic pleiotropy), or did other, neutral or slightly harmful mutations accumulate by drift (Figure 1)? We must consider the population genetic dynamics of these LTEE populations. The hallmark of neutral theory [11] is that mutations with no effect in the selective environment should become fixed in the population at the rate of mutation. For the ancestor of this experiment, the mutation rate is ∼10⁻³ per genome per generation [12],[13], so only a handful of neutral mutations would have fixed by the time tradeoffs became evident, and would not likely explain the early specialization.However, an important extension of neutral theory is that slightly harmful mutations—those whose effects are roughly the inverse of the population size or below, 1/N—can also be fixed by drift [14]. Millions of slightly deleterious mutations were produced in these populations, which cycled between 5×10⁶ and 5×10⁸ cells each day. Might these mutations account for tradeoffs over the first 10–20,000 generations? In small populations, the effect of these mutations can be substantial, which explains why bottlenecked populations may experience fitness declines or even the genome erosion frequently seen in bacterial endosymbionts [15]. But in the large LTEE populations, most deleterious mutations are weeded out by selection and only those with the slightest effects may accumulate over very long time scales. Thus, because these early losses tended to occur when adaptation in the selective environment was most rapid, and because the randomness and rarity of mutation accumulation should not produce parallel changes over these time scales, early specialization is best explained by antagonistic pleiotropy [9],[10].Later in the LTEE, elevated mutation rates began to evolve in certain lines, resulting in a fundamental change in the population genetic environment [16],[17] that should increase the rate of functional decay in unused, essentially neutral functions. These mutator populations tended to perform worse in multiple environments, and in theory should continue to specialize more rapidly by accelerated mutation accumulation. As a first test, we used Biolog plates to assay respiration on 95 different carbon sources over the first 20,000 generations [18]. Although mutators tended to exhibit a reduced breadth of function in this assay, the difference was not statistically significant [18]. Rather, a surprising number of losses of function were shared among replicate lines, and we took this parallelism as further support of antagonistic pleiotropy driven by selection for common sets of adaptive mutations.Here the LTEE offers its greatest advantage: more time, both for evolution and innovative research. Over subsequent generations, mutator lines should continue to accumulate greater mutational load by drift and hence become more specialized than lines retaining the low ancestral rate. Genomic sequences of the evolved lines now have confirmed this increased mutational load [3],[19] in the six of 12 lines that are now mutators [16]. In this issue, Leiby and Marx [20] have readdressed these questions by retracing old steps, applying the prior Biolog assays to lines spanning 50,000 generations of evolution, and by pioneering new high-throughput assays of fitness in many resources. Somewhat surprisingly, these methods disagree and challenge the reliability of Biolog data as a fitness proxy. As a proprietary measure of cellular respiration, it can demonstrate major functional shifts but is less reliable than growth rate as a fitness parameter.More importantly, Leiby and Marx provide clear evidence that niche breadth in the LTEE was shaped by both mutation accumulation and pleiotropy. Growth rates actually increased on several resources, and hence the pleiotropic effects of adaptation to glucose were synergistic, broadening functionality particularly over the first 20,000 generations, as well as antagonistic, producing fewer tradeoffs than previously thought [20]. Pleiotropic effects were also somewhat unpredictable: a sophisticated flux-balance analysis [21] of foreign substrates did not reveal more gains for resources similar to glucose or losses for dissimilar resources. Some early losses linked to selection (maltose, galactose, serine) [6] became complete, but also subtle gains of function for dicarboxylic acid metabolism, perhaps related to growth on metabolic byproducts, became amplified. The most striking pattern was that mutator populations became specialists, diminished for many functions owing to their greater mutational burden, and this only became evident after 50,000 generations in a single resource. These convergent functional losses were not caused by selection, as is often argued, but rather by an absence of selection in the face of mutational pressure. Mutational decay by genetic drift takes a long time, and it will take much longer for the non-mutator lines, it seems.Although Leiby and Marx [20] correctly emphasize the importance of truly long-term selection combined with deficient DNA repair to reveal effects of mutation accumulation, decay has been witnessed in other systems undergoing regular population bottlenecks over shorter time scales [22],[23]. Antagonistic pleiotropy can also reveal its effects much more rapidly than was seen in the LTEE, especially when selection discriminates among discrete fitness features in a heterogeneous environment, such as in the colonization of a new landscape [24],[25]. What this study uniquely illustrates is the unpredictability of pleiotropic effects of adaptation to a simple environment, which in turn shows how chance draws from a distribution of contending beneficial mutations may produce divergent outcomes, ranging from generalists to specialists. A sample of the first mutants competing to prevail in the LTEE system showed variable niche breadth [26] so perhaps we should not be surprised that the footprints of these large-effect mutations endure. Further study of the precise mechanisms by which different mutations produce more fit offspring will teach us more about the origins of diversity that beguile us. We can also gain a broader perspective on the longstanding tension between chance and necessity [27]—a motivator of the LTEE—by focusing more on what is unnecessary, such as how organisms grow in foreign environments. Often insight comes from studying at the margins of a problem, and here, the limits to the growth of these bacteria have allowed us to focus more on how exactly they have accomplished their most essential tasks.     

Caring for Offspring in a World of Cheats

Parents providing care to offspring face the same problem that exists in every biological system in which some individuals offer resources to others: cheaters, who exploit these benefits. In almost all species in which males contribute to parental care, females mate with multiple males. As a result, males frequently provide efforts for unrelated offspring at a cost to their own reproductive fitness. In a new study, Griffin et al. find that across a wide range of animal species, males flexibly adjust their contribution to parental care in relation to extra-pair paternity. However, adjustment is not perfect, because males are limited by the potential costs of withholding help to their own offspring, which is only outweighed if cheating occurs frequently and if providing care reduces a male''s future reproductive success. These findings illustrate how in biological systems cheater and cheated can adapt to changes in each other, preventing either one from gaining control.Whatever your personal feelings, evolutionary biologists will tell you that caring for offspring is not an easy affair. Pick up any current textbook on behavioural ecology, and you will find that the word “family” is invariably followed by the word “conflict” (e.g., [1]). Conflicts between family members arise because selection favours individuals aiming to maximize reproductive fitness, and these aims frequently collide because selection pressures differ even among related individuals [2]–[4]. Offspring can improve their reproductive fitness by obtaining the maximal amount of investment from both of their parents. However, parents frequently provide less than the maximum because any increased investment into current offspring impacts their ability to produce additional offspring in the future. Caring for offspring in all its forms is energetically expensive and may impair a parent''s ability to have additional offspring in a variety of ways. For example, when a female of the golden egg bug (Phyllomorpha laciniata) lays her eggs on a male rather than on a plant, her offspring will have increased survival, but the father carrying the eggs has a higher risk of being eaten by a bird [5]. In bighorn sheep (Ovis canadiensis) [6], mothers are less likely to have a surviving offspring in the year after rearing a son, as males are generally heavier at birth and suckle more frequently because being larger provides an advantage when competing against other males. In European starlings (Sturnus vulgaris), males who participate less in the incubation of the offspring have a higher chance of gaining a secondary female [7].Given the costs of providing parental care, we would expect that individuals should not expend energy if they do not gain any fitness at all, as is the case when they care for offspring that are not their own [2],[8]–[10]. Individuals that are potential victims of cheating are predicted to have evolved a range of counteradaptations to reduce the risks and costs of raising unrelated offspring [11],[12] ([13].

Table 1

Strategies to minimize the risks and costs of being exploited by cheaters.

Hitchhiking to Speciation

The modern evolutionary synthesis codified the idea that species exist as distinct entities because intrinsic reproductive barriers prevent them from merging together. Understanding the origin of species therefore requires understanding the evolution and genetics of reproductive barriers between species. In most cases, speciation is an accident that happens as different populations adapt to different environments and, incidentally, come to differ in ways that render them reproductively incompatible. As with other reproductive barriers, the evolution and genetics of interspecific hybrid sterility and lethality were once also thought to evolve as pleiotripic side effects of adaptation. Recent work on the molecular genetics of speciation has raised an altogether different possibility—the genes that cause hybrid sterility and lethality often come to differ between species not because of adaptation to the external ecological environment but because of internal evolutionary arms races between selfish genetic elements and the genes of the host genome. Arguably one of the best examples supporting a role of ecological adaptation comes from a population of yellow monkey flowers, Mimulus guttatus, in Copperopolis, California, which recently evolved tolerance to soil contaminants from copper mines and simultaneously, as an incidental by-product, hybrid lethality in crosses with some off-mine populations. However, in new work, Wright and colleagues show that hybrid lethality is not a pleiotropic consequence of copper tolerance. Rather, the genetic factor causing hybrid lethality is tightly linked to copper tolerance and spread to fixation in Copperopolis by genetic hitchhiking.New species arise when populations gradually evolve intrinsic reproductive barriers to interbreeding with other populations [1]–[3]. Two species can be reproductively isolated from one another in ways that prevent the formation of interspecific hybrids—the species may, for instance, have incompatible courtship signals or occupy different ecological habitats. Two species can also be reproductively isolated from one another if interspecific hybrids are formed but are somehow unfit—the hybrids may be sterile, inviable, or may simply fall between parental ecological niches. All forms of reproductive isolation limit the genetic exchange between species, preventing their fusion and facilitating their further divergence. Understanding the genetic and evolutionary basis of speciation—a major cause of biodiversity—therefore involves understanding the genetics and evolutionary basis of the traits that mediate reproductive isolation.Most reproductive barriers arise as incidental by-products of selection—either ecological adaptation or sexual selection. For these cases, the genetic basis of speciation is, effectively, the genetics of adaptation. But hybrid sterility and lethality have historically posed two special problems. Darwin [4] devoted an entire chapter of his Origin of Species to the first problem: as the sterility or lethality of hybrids provides no advantage to parents, how could the genetic factors involved possibly evolve by natural selection? The second problem was recognized much later [5], after the rediscovery of Mendelian genetics: if two species (with genotypes AA and aa) produce, say, sterile hybrids (Aa) due to an incompatibility between the A and a alleles, then how could, e.g., the AA genotype have evolved from an aa ancestor in the first place without passing through a sterile intermediate genotype (Aa)? Not only does natural selection not directly favor the evolution of hybrid sterility or lethality, but there is reason to believe natural selection positively prevents its evolution.Together these problems stymied evolutionists and geneticists for decades. T.H. Huxley [6] and William Bateson [5], writing decades apart, each branded the evolution of hybrid sterility one of the most serious challenges for a then-young evolutionary theory. Darwin had, in fact, offered a simple solution to the first problem. Namely, hybrid sterility and lethality are not advantageous per se but rather “incidental on other acquired differences" [4]. Then Bateson [5], in a few short, forgotten lines solved the second problem (see [7]). Later, Dobzhansky [2] and Muller [8] would arrive at the same solution, showing that hybrid sterility or lethality could evolve readily, unopposed by natural selection, under a two-locus model with epistasis. In particular, they imagined that separate populations diverge from a common ancestor (genotype aabb), with the A allele becoming established in one population (AAbb) and the B allele in the other (aaBB); while A and B alleles must function on their respective genetic backgrounds, there is no guarantee that the A and B alleles will be functionally compatible with one another. Hybrid sterility and lethality most likely result from incompatible complementary genetic factors that disrupt development when brought together in a common hybrid genome. Dobzhansky [2] and Muller [8] could point to a few supporting data in fish, flies, and plants. Notably, like Darwin, neither speculated on the forces responsible for the evolution of the genetic factors involved.Today, there is no doubt that the Dobzhansky-Muller model is correct, as the data for incompatible complementary genetic factors is now overwhelming [1],[9]. In the last decade, a fast-growing number of speciation genes involved in these genetic incompatibilities have been identified in mice, fish, flies, yeast, and plants [9]–[11]. Perhaps not surprisingly, these speciation genes often have histories of recurrent, adaptive protein-coding sequence evolution [10],[11]. The signature of selection at speciation genes has been taken by some as tacit evidence for the pervasive role of ecological adaptation in speciation, including the evolution of hybrid sterility and lethality [12]. What is surprising, however, from the modern molecular analysis of speciation genes is how often their rapid sequence evolution and functional divergence seems to have little to do with adaptation to external ecological circumstances. Instead, speciation genes often (but not always [9]–[11]) seem to evolve as by-products of evolutionary arms races between selfish genetic elements—e.g., satellite DNAs [13],[14], meiotic drive elements [15], cytoplasmic male sterility factors [16]—and the host genes that regulate or suppress them [9]–[11],[17]. The notion that selfish genes are exotic curiosities is now giving way to a realization that selfish genes are common and diverse, each generation probing for transmission advantages at the expense of their bearers, fueling evolutionary arms races and, not infrequently, contributing to the genetic divergence that drives speciation. Indeed, the case has become so strong that examples of hybrid sterility and lethality genes that have evolved in response to ecological challenges (other than pathogens) appear to be the exception [9],[11],[17].Perhaps the most clear-cut case in which a genetic incompatibility seems to have evolved as a by-product of ecological adaptation comes from populations of the yellow monkey flower, Mimulus guttatus, from Copperopolis (California, U.S.A.). In the last ∼150 years, the Copperopolis population has evolved tolerance to the tailings of local copper mines (Figure 1). These copper-tolerant M. guttatus plants also happen to be partially reproductively isolated from many off-mine M. guttatus plants, producing hybrids that suffer tissue necrosis and death. In classic work, Macnair and Christie showed that copper tolerance is controlled by a single major factor [18] and hybrid lethality, as expected under the Dobzhansky-Muller model, by complementary factors [19]. Surprisingly, in crosses between tolerant and nontolerant plants, hybrid lethality perfectly cosegregates with tolerance [19],[20]. The simplest explanation is that the copper tolerance allele that spread to fixation in the Copperopolis population also happens to cause hybrid lethality as a pleiotropic by-product. The alternative explanation is that the copper tolerance and hybrid lethality loci happen to be genetically linked; when the copper tolerance allele spread to fixation in Copperopolis, hybrid lethality hitchhiked to high frequency along with it [20]. But with 2n = 28 chromosomes, the odds that copper tolerance and hybrid lethality alleles happen to be linked would seem vanishingly small [20].Open in a separate windowFigure 1Yellow monkey flowers (Mimulus guttatus) growing in the heavy-metal contaminated soils of copper-mine tailings.In this issue, Wright and colleagues [21] revisit this classic case of genetic incompatibility as a by-product of ecological adaptation. They make two discoveries, one genetic and the other evolutionary. By conducting extensive crossing experiments and leveraging the M. guttatus genome sequence (www.mimulusevolution.org), Wright et al. [21] map copper tolerance and hybrid necrosis to tightly linked but genetically separable loci, Tol1 and Nec1, respectively. Hybrid lethality is not a pleiotropic consequence of copper tolerance. Instead, the tolerant Tol1 allele spread to fixation in Copperopolis, and the tightly linked incompatible Nec1 allele spread with it by genetic hitchhiking. In a turn of bad luck, the loci happen to fall in a heterochromatic pericentric region, where genome assemblies are often problematic, putting identification of the Tol1 and Nec1 genes out of immediate reach. Wright et al. [21] were, however, able to identify linked markers within ∼0.3 cM of Tol1 and place Nec1 within a 10-kb genomic interval that contains a Gypsy3 retrotransposon, raising two possibilities. First, the Gypsy3 element is unlikely to cause hybrid lethality directly; instead, as transposable elements are often epigenetically silenced in plants, it seems possible that the Nec1-associated Gypsy3 is silenced with incidental consequences for gene expression on a gene (or genes) in the vicinity [22]. Second, although the Nec1 interval is 10-kb in the reference genome of M. guttatus, it could be larger in the (not-yet-sequenced) Copperopolis population, perhaps harboring additional genes.With Tol1 and Nec1 mapped near and to particular genomic scaffolds, respectively, Wright et al. were able to investigate the evolutionary history of the genomic region. Given the clear adaptive significance of copper tolerance in Copperopolis plants, we might expect to see the signatures of a strong selective sweep in the Tol1 region—a single Tol1 haplotype may have spread to fixation so quickly that all Copperopolis descendant plants bear the identical haplotype and thus show strongly reduced population genetic variability in the Tol1-Nec1 region relative to the rest of the genome [23],[24]. After the selective sweep is complete, variability in the region ought to recover gradually as new mutations arise and begin to fill out the mutation-drift equilibrium frequency spectrum expected for neutral variation in the Copperopolis population [25],[26]. Given that Tol1 reflects an adaptation to mine tailings established just ∼150 generations ago, there would have been little time for such a recovery. And yet, while Wright et al. find evidence of moderately reduced genetic variability in the Tol1-Nec1 genomic region, the magnitude of the reduction is hardly dramatic relative to the genome average.How, then, is it possible that the Tol1-Nec1 region swept to fixation in Copperopolis in fewer than ∼150 generations and yet left no strong footprint of a hitchhiking event? One possibility is that rather than a single, unique Tol1-Nec1 haplotype contributing to fixation, causing a “hard sweep," multiple Tol1-Nec1 haplotypes sampled from previously standing genetic variation contributed to fixation, causing a “soft sweep" [27]. A soft sweep would be plausible if Tol1 and Nec1 both segregate in the local off-mine ancestral population and if the two were, coincidentally, found on the same chromosome more often than expected by chance (i.e., in linkage disequilibrium). Then, after the copper mines were established, multiple plants with multiple Tol1 haplotypes (and, by association, Nec1) could have colonized the newly contaminated soils of the mine tailings. Tol1 segregates at ∼9% in surrounding populations, suggesting that standing genetic variation for copper tolerance may well have been present in the ancestral populations.Two big questions remain for the Tol1-Nec1 story, and both would be readily advanced by identification of Tol1 and Nec1. The first question concerns the history of Tol1 haplotypes in Copperopolis and surrounding off-mine populations. As Nec1-mediated hybrid lethality is incomplete, the ∼9% Tol1 frequency in surrounding populations could reflect its export via gene flow from the Copperopolis populations. Conversely, if there was a soft sweep from standing Tol1 variation in surrounding off-mine populations, then Tol1 and Nec1 may still be in linkage disequilibrium in those populations (assuming ∼150 years of recombination has not broken up the association). Resolving these alternative possibilities is a matter of establishing the history of movement of Tol1 haplotypes into or out of the Copperopolis population. The soft sweep scenario, if correct, presents a population genetics puzzle: during the historical time that mutations accumulated among the multiple tolerant but incompatible Tol1-Nec1 haplotypes in the ancestral off-mine populations, why did recombination fail to degrade the association, giving rise to tolerant but compatible haplotypes?The second question concerns the identity of Nec1 (or if it really is a Gypsy3 element, the identity of the nearby gene whose expression is disrupted as a consequence). The answer bears on one of the new emerging generalizations about genetic incompatibilities in plants [9]. Recently, Bomblies and Weigel [28] synthesized a century''s worth of observations on the commonly seen necrosis phenotype in plant hybrids and, based on their own genetic analyses in Arabidopsis [29], suggested that many of these cases may have a common underlying basis: incompatibilities between plant pathogen resistance genes can cause autoimmune responses that result in tissue necrosis and hybrid lethality. Hybrid necrosis, indeed, appears to involve pathogen resistance genes across multiple plants groups [9],[28]. It remains to be seen if the Nec1-mediated lethality provides yet another instance.     

Viral OTU Deubiquitinases: A Structural and Functional Comparison

Recent studies have revealed that proteases encoded by three very diverse RNA virus groups share structural similarity with enzymes of the Ovarian Tumor (OTU) superfamily of deubiquitinases (DUBs). The publication of the latest of these reports in quick succession prevented proper recognition and discussion of the shared features of these viral enzymes. Here we provide a brief structural and functional comparison of these virus-encoded OTU DUBs. Interestingly, although their shared structural features and substrate specificity tentatively place them within the same protease superfamily, they also show interesting differences that trigger speculation as to their origins.The covalent attachment of ubiquitin (Ub) to protein substrates, i.e., ubiquitination, plays a pivotal regulatory role in numerous cellular processes [1]–[5]. Ubiquitination can be reversed by deubiquitinases (DUBs) [6] and, not surprisingly, various virus groups encode such DUBs to influence ubiquitin-mediated host cell processes [7]–[21]. Some of these viral DUBs resemble proteases belonging to the Ovarian Tumor (OTU) superfamily [22]–[28]. Makarova et al. previously identified OTU proteases as a novel superfamily of cysteine proteases from different organisms [29], and their bioinformatics-based analysis included several of the viral enzymes discussed here. Recently reported structures of these viral DUBs include the OTU domains of the nairoviruses Crimean-Congo hemorrhagic fever virus (CCHFV) [22]–[24] and Dugbe virus (DUGV) [25], the papain-like protease (PLP2) domain of the arterivirus equine arteritis virus (EAV) [26], and the protease (PRO) domain of the tymovirus turnip yellow mosaic virus (TYMV) (Figure 1A–1D) [27], [28]. These viruses are strikingly diverse, considering that nairoviruses are mammalian negative-strand RNA viruses, while the mammalian arteriviruses and plant tymoviruses belong to separate orders of positive-strand RNA viruses.Open in a separate windowFigure 1Viral and eukaryotic OTU domain structures and viral protein context.Crystal structures of (A) CCHFV OTU (3PT2) [23], (B) DUGV OTU (4HXD) [25], (C) EAV PLP2 (4IUM) [26], (D) TYMV PRO (4A5U) [27], [28], (E) yeast OTU1 (3BY4) [57], and (F) human OTUD3 (4BOU) [46]. The β-hairpin motifs of CCHFV OTU and DUGV OTU are indicated in boxes in panels A and B, respectively, and the zinc-finger motif of EAV PLP2 is boxed in panel C. Active sites are indicated with arrows. The CCHFV OTU, DUGV OTU, EAV PLP2, and yeast OTU1 domains were crystallized in complex with Ub, which has been removed for clarity. Structure images were generated using PyMol [60]. (G) Schematic representation of the CCHFV large (L) protein [61], [62]. A similar organization is found in the DUGV L protein, but is not depicted. The OTU domain resides in the N-terminal region of this protein and is not involved in autoproteolytic cleavage events [48]. (H) Schematic representation of the EAV polyprotein 1ab [63]. PLP2 resides in nonstructural protein 2 (nsp2) and is responsible for the cleavage between nsp2 and nsp3 [51]. (I) Schematic representation of the TYMV ORF1 polyprotein [50]. PRO resides in the N-terminal product of this polyprotein and is responsible for two internal cleavages [49], [50]. Key replicative enzymes are indicated in G, H, and I. Colored arrowheads denote cleavage sites for the indicated protease domains. HEL, helicase; PLP, papain-like protease; RdRp, RNA-dependent RNA polymerase; SP, serine protease.Ubiquitination often involves the formation of polyubiquitin chains [1], which can target the ubiquitinated substrate to the proteasome for degradation [2] or modulate its protein–protein interactions, as in the activation of innate immune signaling pathways [3], [4]. Interestingly, several cellular OTU DUBs were found to negatively regulate innate immunity [30]–[33]. Likewise, both nairovirus OTU and arterivirus PLP2 were recently shown to inhibit innate immune responses by targeting ubiquitinated signaling factors [7]–[9], [26], [34], [35]. In contrast to eukaryotic OTU DUBs, both of these viral proteases were found to also deconjugate the Ub-like protein interferon-stimulated gene 15 (ISG15) [7], [36], which inhibits viral replication via a mechanism that is currently poorly understood [37]. Interestingly, coronaviruses (which, together with the arteriviruses, belong to the nidovirus order) also encode papain-like proteases targeting both Ub and ISG15 that were shown to inhibit innate immunity [11]–[13], [38]–[42] but belong to the ubiquitin-specific protease (USP) class of DUBs [6], [43], [44]. The presence of functionally similar, yet structurally different proteases in distantly related virus families highlights the potential benefits to the virus of harboring such enzymes.The proteasomal degradation pathway is an important cellular route to dispose of viral proteins, as exemplified by the turnover of the TYMV polymerase [45]. Moreover, the degradation of this protein is specifically counteracted by the deubiquitinase activity of TYMV PRO, which thus promotes virus replication [10]. The functional characterization of viral OTU DUBs remains incomplete and future studies will likely reveal additional roles in replication and virus–host interplay.Polyubiquitin chains can adopt a number of different configurations, depending on the type of covalent linkage present within the polymer [1]. A distal Ub molecule can be linked via its C-terminus to one of seven internal lysine residues present in a proximal Ub molecule via an isopeptide bond. Alternatively, in the case of linear chains, the C-terminus of the distal Ub is covalently linked to the N-terminal methionine residue of the proximal Ub via a peptide bond. While human OTU proteases often show a distinct preference for one or two isopeptide linkage types [46], nairovirus OTUs and TYMV PRO appear to be more promiscuous in their substrate preference [22], [25]. However, like most human OTU proteases, they seem unable to cleave linear polyubiquitin chains in vitro [22], [25], [46]. Arterivirus PLP2 has not been extensively studied in this respect.It is important to note that many positive-strand RNA viruses, including arteriviruses and tymoviruses, encode polyproteins that are post-translationally cleaved by internal protease domains [47]. Thus, while CCHFV OTU is not involved in viral protein cleavage and its activity seems dispensable for replication (Figure 1G) [48], both arterivirus PLP2 and tymovirus PRO are critically required for viral replication due to their primary role in polyprotein maturation (Figure 1H, 1I) [49]–[53]. Interestingly, while both EAV PLP2 and TYMV PRO can process peptide bonds in cis and in trans [50], [51], PRO does not cleave peptide bonds in linear polyubiquitin chains in vitro [25]. To date, activity of EAV PLP2 towards linear polyubiquitin chains has not been reported.Based on mutagenesis of putative catalytic residues, arterivirus PLP2 and tymovirus PRO were initially generally classified as papain-like cysteine proteases [51], [54], [55]. Now that crystal structures of these proteases are available, it is possible to search the DALI server [56] in order to identify structurally similar domains. Using the 3-dimensional coordinates of TYMV PRO, the most recently solved structure of a viral OTU protease, such a query identifies structural similarity with eukaryotic OTU DUBs as well as the nairovirus OTU domains and EAV PLP2 ([57] further highlights their similarities (Figure 2A–2C), and these comparisons together clearly position them within the OTU DUB superfamily. Sequence comparisons alone were insufficient to demonstrate this conclusively, as the similarity of viral OTU domains to each other and to eukaryotic OTU proteases is very limited and mostly restricted to the areas surrounding the active site residues [29].Open in a separate windowFigure 2Superpositions of the viral OTU proteases with yeast OTU1 and one another.Superpositions of yeast OTU1 (3BY4) [57] with (A) CCHFV OTU (3PT2) [23], RMSD: 1.8 Å over 112 residues, (B) EAV PLP2 (4IUM) [26], RMSD: 2.8 Å over 69 residues, and (C) TYMV PRO (4A5U) [27], [28], RMSD: 1.4 Å over 76 residues. Superpositions of the yeast OTU1-Ub complex with (D) the CCHFV OTU-Ub complex and (E) the EAV PLP2-Ub complex, highlighting the difference in the orientation of Ub between the two viral OTU domains versus the eukaryotic yeast OTU1 domain. The Ub that is complexed with yeast OTU1 is depicted in yellow, while the Ub complexed with CCHFV OTU or EAV PLP2 is depicted in orange. (F) Superposition of EAV PLP2 and TYMV PRO, RMSD: 2.5 Å over 53 residues. (G) Close-up of the active site region (boxed) of the superposition depicted in F. Side chains of the catalytic cysteine (Cys270 and Cys783 for EAV PLP2 and TYMV PRO, respectively) and histidine (His332 and His869 for EAV PLP2 and TYMV PRO, respectively) residues are shown as sticks, as well as the active site Asn263 for EAV PLP2. The backbone amide group of Asp267 likely contributes to the formation of the oxyanion hole in the active site of EAV PLP2, yet a functionally equivalent residue is absent in TYMV PRO. The Gly266 and Gly268 residues flanking Asp267 in EAV PLP2 are depicted as sticks as well, for clarity. Note the alternative orientation of the active site cysteine residue of TYMV PRO which, unlike EAV PLP2, was not determined in covalent complex with an Ub suicide substrate. All alignments were generated using the PDBeFOLD server [64], and thus the reported RMSD values differ from those reported in [60]. RMSD, root-mean-square deviation.

Table 1

Three-dimensional structural alignment of viral OTU domains against selected structures in the Protein Data Bank using the DALI server [56].

Focus Issue on Calcium Signaling: Analyses of a Gravistimulation-Specific Ca2+ Signature in Arabidopsis using Parabolic Flights

Gravity is a critical environmental factor affecting the morphology and functions of organisms on the Earth. Plants sense changes in the gravity vector (gravistimulation) and regulate their growth direction accordingly. In Arabidopsis (Arabidopsis thaliana) seedlings, gravistimulation, achieved by rotating the specimens under the ambient 1g of the Earth, is known to induce a biphasic (transient and sustained) increase in cytoplasmic calcium concentration ([Ca2+]_c). However, the [Ca2+]_c increase genuinely caused by gravistimulation has not been identified because gravistimulation is generally accompanied by rotation of specimens on the ground (1g), adding an additional mechanical signal to the treatment. Here, we demonstrate a gravistimulation-specific Ca²⁺ response in Arabidopsis seedlings by separating rotation from gravistimulation by using the microgravity (less than 10⁻⁴g) conditions provided by parabolic flights. Gravistimulation without rotating the specimen caused a sustained [Ca2+]_c increase, which corresponds closely to the second sustained [Ca2+]_c increase observed in ground experiments. The [Ca2+]_c increases were analyzed under a variety of gravity intensities (e.g. 0.5g, 1.5g, or 2g) combined with rapid switching between hypergravity and microgravity, demonstrating that Arabidopsis seedlings possess a very rapid gravity-sensing mechanism linearly transducing a wide range of gravitational changes (0.5g–2g) into Ca²⁺ signals on a subsecond time scale.Calcium ion (Ca²⁺) functions as an intracellular second messenger in many signaling pathways in plants (White and Broadley, 2003; Hetherington and Brownlee, 2004; McAinsh and Pittman, 2009; Spalding and Harper, 2011). Endogenous and exogenous signals are spatiotemporally encoded by changing the free cytoplasmic concentration of Ca²⁺ ([Ca2+]_c), which in turn triggers [Ca2+]_c-dependent downstream signaling (Sanders et al., 2002; Dodd et al., 2010). A variety of [Ca2+]_c increases induced by diverse environmental and developmental stimuli are reported, such as phytohormones (Allen et al., 2000), temperature (Plieth et al., 1999; Dodd et al., 2006), and touch (Knight et al., 1991; Monshausen et al., 2009). The [Ca2+]_c increase couples each stimulus and appropriate physiological responses. In the Ca²⁺ signaling pathways, the stimulus-specific [Ca2+]_c pattern (e.g. amplitude and oscillation) provide the critical information for cellular signaling (Scrase-Field and Knight, 2003; Dodd et al., 2010). Therefore, identification of the stimulus-specific [Ca2+]_c signature is crucial for an understanding of the intracellular signaling pathways and physiological responses triggered by each stimulus, as shown in the case of cold acclimation (Knight et al., 1996; Knight and Knight, 2000).Plants often exhibit biphasic [Ca2+]_c increases in response to environmental stimuli. Thus, slow cooling causes a fast [Ca2+]_c transient followed by a second, extended [Ca2+]_c increase in Arabidopsis (Arabidopsis thaliana; Plieth et al., 1999; Knight and Knight, 2000). The Ca²⁺ channel blocker lanthanum (La³⁺) attenuated the fast transient but not the following increase (Knight and Knight, 2000), suggesting that these two [Ca2+]_c peaks have different origins. Similarly, hypoosmotic shock caused a biphasic [Ca2+]_c increase in tobacco (Nicotiana tabacum) suspension-culture cells (Takahashi et al., 1997; Cessna et al., 1998). The first [Ca2+]_c peak was inhibited by gadolinium (Gd³⁺), La³⁺, and the Ca²⁺ chelator EGTA (Takahashi et al., 1997; Cessna et al., 1998), whereas the second [Ca2+]_c increase was inhibited by the intracellular Ca²⁺ store-depleting agent caffeine but not by EGTA (Cessna et al., 1998). The amplitude of the first [Ca2+]_c peak affected the amplitude of the second increase and vice versa (Cessna et al., 1998). These results suggest that even though the two [Ca2+]_c peaks originate from different Ca²⁺ fluxes (e.g. Ca²⁺ influx through the plasma membrane and Ca²⁺ release from subcellular stores, respectively), they are closely interrelated, showing the importance of the kinetic and pharmacological analyses of these [Ca2+]_c increases.Changes in the gravity vector (gravistimulation) could work as crucial environmental stimuli in plants and are generally achieved by rotating the specimens (e.g. +180°) in ground experiments. Use of Arabidopsis seedlings expressing apoaequorin, a Ca²⁺-reporting photoprotein (Plieth and Trewavas, 2002; Toyota et al., 2008a), has revealed that gravistimulation induces a biphasic [Ca2+]_c increase that may be involved in the sensory pathway for gravity perception/response (Pickard, 2007; Toyota and Gilroy, 2013) and the intracellular distribution of auxin transporters (Benjamins et al., 2003; Zhang et al., 2011). These two Ca²⁺ changes have different characteristics. The first transient [Ca2+]_c increase depends on the rotational velocity but not angle, whereas the second sustained [Ca2+]_c increase depends on the rotational angle but not velocity. The first [Ca2+]_c transient was inhibited by Gd³⁺, La³⁺, and the Ca²⁺ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid but not by ruthenium red (RR), whereas the second sustained [Ca2+]_c increase was inhibited by all these chemicals. These results suggest that the first transient and second sustained [Ca2+]_c increases are related to the rotational stimulation and the gravistimulation, respectively, and are mediated by distinct molecular mechanisms (Toyota et al., 2008a). However, it has not been demonstrated directly that the second sustained [Ca2+]_c increase is induced solely by gravistimulation; it could be influenced by other factors, such as an interaction with the first transient [Ca2+]_c increase (Cessna et al., 1998), vibration, and/or deformation of plants during the rotation.To elucidate the genuine Ca²⁺ signature in response to gravistimulation in plants, we separated rotation and gravistimulation under microgravity (μg; less than 10⁻⁴g) conditions provided by parabolic flight (PF). Using this approach, we were able to apply rotation and gravistimulation to plants separately (Fig. 1). When Arabidopsis seedlings were rotated +180° under μg conditions, the [Ca2+]_c response to the rotation was transient and almost totally attenuated in a few seconds. Gravistimulation (transition from μg to 1.5g) was then applied to these prerotated specimens at the terminating phase of the PF. This gravistimulation without simultaneous rotation induced a sustained [Ca2+]_c increase. The kinetic properties of this sustained [Ca2+]_c increase were examined under different gravity intensities (0.5g–2g) and sequences of gravity intensity changes (Fig. 2A). This analysis revealed that gravistimulation-specific Ca²⁺ response has an almost linear dependency on gravitational acceleration (0.5g–2g) and an extremely rapid responsiveness of less than 1 s.Open in a separate windowFigure 1.Diagram of the experimental procedures for applying separately rotation and gravistimulation to Arabidopsis seedlings. Rotatory stimulation (green arrow) was applied by rotating the seedlings 180° under μg conditions, and 1.5g 180° rotation gravistimulation (blue arrow) was applied to the prerotated seedlings after μg.Open in a separate windowFigure 2.Acceleration, temperature, humidity, and pressure in an aircraft during flight experiments. A, Accelerations along x, y, and z axes in the aircraft during PF. The direction of flight (FWD) and coordinates (x, y, and z) are indicated in the bottom graph. The inset shows an enlargement of the acceleration along the z axis (gravitational acceleration) during μg conditions lasting for approximately 20 s. B, Temperature, humidity, and pressure in the aircraft during PF. Shaded areas in graphs denote the μg condition.     

Archaeological Support for the Three-Stage Expansion of Modern Humans across Northeastern Eurasia and into the Americas

Background

Understanding the dynamics of the human range expansion across northeastern Eurasia during the late Pleistocene is central to establishing empirical temporal constraints on the colonization of the Americas [1]. Opinions vary widely on how and when the Americas were colonized, with advocates supporting either a pre-[2] or post-[1], [3], [4], [5], [6] last glacial maximum (LGM) colonization, via either a land bridge across Beringia [3], [4], [5], a sea-faring Pacific Rim coastal route [1], [3], a trans-Arctic route [4], or a trans-Atlantic oceanic route [5]. Here we analyze a large sample of radiocarbon dates from the northeast Eurasian Upper Paleolithic to identify the origin of this expansion, and estimate the velocity of colonization wave as it moved across northern Eurasia and into the Americas.

Methodology/Principal Findings

We use diffusion models [6], [7] to quantify these dynamics. Our results show the expansion originated in the Altai region of southern Siberia ∼46kBP , and from there expanded across northern Eurasia at an average velocity of 0.16 km per year. However, the movement of the colonizing wave was not continuous but underwent three distinct phases: 1) an initial expansion from 47-32k calBP; 2) a hiatus from ∼32-16k calBP, and 3) a second expansion after the LGM ∼16k calBP. These results provide archaeological support for the recently proposed three-stage model of the colonization of the Americas [8], [9]. Our results falsify the hypothesis of a pre-LGM terrestrial colonization of the Americas and we discuss the importance of these empirical results in the light of alternative models.

Conclusions/Significance

Our results demonstrate that the radiocarbon record of Upper Paleolithic northeastern Eurasia supports a post-LGM terrestrial colonization of the Americas falsifying the proposed pre-LGM terrestrial colonization of the Americas. We show that this expansion was not a simple process, but proceeded in three phases, consistent with genetic data, largely in response to the variable climatic conditions of late Pleistocene northeast Eurasia. Further, the constraints imposed by the spatiotemporal gradient in the empirical radiocarbon record across this entire region suggests that North America cannot have been colonized much before the existing Clovis radiocarbon record suggests.     

Sequencing the Connectome

Connectivity determines the function of neural circuits. Historically, circuit mapping has usually been viewed as a problem of microscopy, but no current method can achieve high-throughput mapping of entire circuits with single neuron precision. Here we describe a novel approach to determining connectivity. We propose BOINC (“barcoding of individual neuronal connections”), a method for converting the problem of connectivity into a form that can be read out by high-throughput DNA sequencing. The appeal of using sequencing is that its scale—sequencing billions of nucleotides per day is now routine—is a natural match to the complexity of neural circuits. An inexpensive high-throughput technique for establishing circuit connectivity at single neuron resolution could transform neuroscience research.Neuroscientists seek neural explanations of perception, thought, and behavior. What does such an explanation look like? One of the earliest examples is Descartes'' account [1] of the reflex withdrawal of a foot from a fire (Figure 1A). Descartes hypothesized that small particles of the fire displace the skin of the foot, which pulls on a tiny thread and thereby opens a pore in the pineal gland, releasing animal spirits, which flow back via a hollow tube into the foot to cause retraction. Although more modern accounts of the spinal reflex arc differ in important mechanistic and anatomical details, the kernel of Descartes'' explanation is both correct and intellectually satisfying: the neural circuit he describes immediately implies the causal relationship between the stimulus and the resultant action. Circuit-level explanations of computation and behavior represent the gold standard.Open in a separate windowFigure 1The wiring of neural circuits is highly structured.(A) Descartes'' model of the foot withdrawal reflex. (B) Two similar circuits in which the computation is readily deduced from the wiring. The circuit on the top is directionally selective, whereas the one on the bottom performs a center-surround computation. (C) The costs of DNA sequencing are falling exponentially. From 2001 to 2007, the costs of sequencing dropped exponentially, in pace with Moore''s law [18] for computation. Since the introduction of “next generation” sequencing technologies in 2008, the cost of sequencing has fallen more than 10-fold every year, compared with the steady 1.4-fold yearly drop for computing power. Data from http://www.genome.gov/sequencingcosts/.     

From Dinosaurs to Modern Bird Diversity: Extending the Time Scale of Adaptive Radiation

What explains why some groups of organisms, like birds, are so species rich? And what explains their extraordinary ecological diversity, ranging from large, flightless birds to small migratory species that fly thousand of kilometers every year? These and similar questions have spurred great interest in adaptive radiation, the diversification of ecological traits in a rapidly speciating group of organisms. Although the initial formulation of modern concepts of adaptive radiation arose from consideration of the fossil record, rigorous attempts to identify adaptive radiation in the fossil record are still uncommon. Moreover, most studies of adaptive radiation concern groups that are less than 50 million years old. Thus, it is unclear how important adaptive radiation is over temporal scales that span much larger portions of the history of life. In this issue, Benson et al. test the idea of a “deep-time” adaptive radiation in dinosaurs, compiling and using one of the most comprehensive phylogenetic and body-size datasets for fossils. Using recent phylogenetic statistical methods, they find that in most clades of dinosaurs there is a strong signal of an “early burst” in body-size evolution, a predicted pattern of adaptive radiation in which rapid trait evolution happens early in a group''s history and then slows down. They also find that body-size evolution did not slow down in the lineage leading to birds, hinting at why birds survived to the present day and diversified. This paper represents one of the most convincing attempts at understanding deep-time adaptive radiations.

“It is strikingly noticeable from the fossil record and from its results in the world around us that some time after a rather distinctive new adaptive type has developed it often becomes highly diversified.” – G. G. Simpson ([1], pp. 222–223)

George Gaylord Simpson was the father of modern concepts of adaptive radiation—the diversification of ecological traits in a rapidly speciating group of organisms (Figure 1; [2]). He considered adaptive radiation to be the source of much of the diversity of living organisms on planet earth, in terms of species number, ecology, and body form [1]–[3]. Yet more than 60 years after Simpson''s seminal work, the exact role of adaptive radiation in generating life''s extraordinary diversity is still an open and fundamental question in evolutionary biology [3],[4].Open in a separate windowFigure 1An example of adaptive radiation and early bursts in rates of speciation and phenotypic evolution.(a) The adaptive radiation of the modern bird clade Vanginae, which shows early rapid speciation, morphological diversity, and diversity in foraging behavior and diet [15],[32]. (b) Hypothetical curve of speciation rates through time that would be expected in adaptive radiation. The exponential decline in speciation rates shows that there was an “early burst” of speciation at the beginning of the clade''s history. (c) Hypothetical curve of rates of phenotypic evolution through time that would be expected in adaptive radiation, also showing an early burst of evolution with high initial rates. Part (a) is reproduced from [32] with permission (under CC-BY) from the Royal Society and the original authors.To address this question, researchers have looked for signatures of past adaptive radiation in the patterns of diversity in nature. In particular, it has been suggested that groups that have undergone adaptive radiation should show an “early-burst” signal in both rates of lineage diversification and phenotypic evolution through time—a pattern in which rates of speciation and phenotypic evolution are fast early in the history of groups and then decelerate over time (Figure 1; [3]–[5]). These predictions arise from the idea that clades should multiply and diversify rapidly in species number, ecology, and phenotype in an adaptive radiation and that rates of this diversification should decrease later as niches are successively occupied [2].Early bursts have been sought in both fossils and phylogenies. Few fossil studies have discussed their results in the context of adaptive radiation (but see [6]), but they often have found rapid rises in both taxonomic and morphological diversity early in the history of various groups [7], ranging from marine invertebrates [8] to terrestrial mammals [9]. However, fossils often lack the phylogeny needed to model how evolution has proceeded [7]. On the other hand, studies that test for early bursts in currently existing (extant) species typically use phylogenies, which allow us to model past evolution in groups with few or no fossils [5]. Phylogenies have most often been used to test early bursts in speciation (see, e.g., [10]). However, such tests may be misled by past extinction, which will decay the statistical signal of rapid, early diversification [11]. Furthermore, diverse evolutionary scenarios beyond adaptive radiation can give rise to early bursts in speciation [12]. By contrast, studies of phenotypic diversification may be more robust to extinction [13] and they test the distinguishing feature that separates adaptive from nonadaptive radiation [2],[12].Thus, studies of adaptive radiation in extant organisms increasingly have focused on phylogenetic tests of the early-burst model of phenotypic evolution. Some studies show strong support for this prediction in both birds [14],[15] and lizards [5],[16]. However, the most extensive study to date showed almost no support for the early-burst model. In this study, Harmon et al. [17] examined body size in 49 (and shape in 39) diverse groups of animals, including invertebrates, fishes, amphibians, reptiles, birds, and mammals. They found strong support for the early-burst model in only two of these 88 total datasets.This result raises an important question: if adaptive radiation explains most of life''s diversity [1], how is it possible that there is so little phylogenetic evidence for early bursts of phenotypic evolution? One possibility is that early bursts are hard to detect. This can be due to low statistical power in the most commonly employed tests [18]. It may also be due to a lack of precision in the way “early burst” is defined (and thus tested), as the ecological theory of adaptive radiation suggests that the rate of phenotypic evolution will decrease as species diversity increases in a group, not just over time [14],[16]. Indeed, recent studies [14],[16] detected a decline in rates with species diversity in clades that were also in the Harmon et al. [17] study, yet for which no decline over time was detected.A second possible reason for why early-burst patterns are uncommon is more fundamental: the patterns of phenotypic diversity that result from adaptive radiation may be different at large time scales. Many of the best examples of adaptive radiation are in groups that are relatively young, including Darwin''s finches (2.3 million years old [myr]; [19]) and Lake Malawi and Victoria cichlids (2.3 myr; [20]), whereas most groups that are examined for early bursts in phenotypic evolution are much older (e.g., 47 of 49 in Harmon et al. [17]; mean ± sd = 23.8±29.2 myr). So there may be an inherent difference between what unfolds over the relatively short time scales emphasized by Schluter [2] and what one sees at macroevolutionary time scales (see [21] for an in-depth discussion of this idea as it relates to speciation).The time scale over which adaptive radiations unfold has been little explored. As a result, the link between extant diversity and major extinct radiations remains unclear. Simpson [1] believed that adaptive radiation played out at the population level, but that it should manifest itself at larger scales as well—up to phyla (e.g., chordates, arthropods). He suggested that we should see signals of adaptive radiations in large, old clades because they are effectively small-scale adaptive radiation writ large [1]. Under this view, we should see the signal of adaptive radiation even in groups that diversified over vast time scales, particularly if adaptive radiation is as important for explaining life''s diversity as Simpson [1] thought it was.Part of the reason why potential adaptive radiations at deep time scales remain poorly understood is that studies either focus on fossils or phylogenies, but rarely both. In this issue, Benson et al. [22] combine these two types of data to address whether dinosaurs show signs that they adaptively radiated. Unlike most other studies, the temporal scale of the current study is very large—in this case, over 170 million years throughout the Mesozoic era, starting at 240 million years ago in the Triassic period. This characteristic allowed Benson et al. to shed light on deep-time adaptive radiation.The authors estimated body mass from fossils by using measurements of the circumference of the stylopodium shaft (the largest bone of the arm or leg, such as the femur), which shows a consistent scaling relationship with body mass in extant reptiles and mammals [23]. They then combined published phylogenies to obtain a composite phylogeny for the species in their body-size dataset. The authors finally conducted two types of tests of the rate of body-size evolution—tests of early bursts in phenotypic evolution that are the same as those of Harmon et al. [17], as well as an additional less commonly used test that estimates whether differences between estimated body size at adjacent phylogenetic nodes decreases over time.Benson et al. [22] found two striking results. First, in both of their analyses, the early-burst model was strongly supported for most clades of dinosaurs. This early burst began in the Triassic period, indicating that diversification in body size in dinosaurs began before the Triassic-Jurassic mass extinction event would have opened competition-free ecological space (as commonly hypothesized; [24],[25]). Rather, the authors [22] suggest that a key innovation led to this rise in dinosaurs, though it is not clear what this innovation was [26]. In general, though, the finding of an early burst in body-size evolution in most dinosaurs—if a consequence of adaptive evolution—suggests that adaptive radiation may play out over large evolutionary time scales, not just on the short time scales typical of the most well-studied cases of extant groups.Second, one clade—Maniraptora, which is the clade in which modern-day birds are nested—was the only part of the dinosaur phylogeny that did not show such a strong early burst in body-size evolution. Instead, this clade fit a model to a single adaptive peak—an optimum body size, if you will—but also maintained high rates of undirected body-size evolution throughout their history. Benson et al. [22] suggest that this last result connects deep-time adaptive radiation in the dinosaurs, which quickly exhausted the possibility of phenotypic space, with the current radiation in extant birds, which survived to the present day because their constant, high rate of evolution meant that they were constantly undergoing ecological innovation. This gives a glimpse into why modern birds have so many species (an order of magnitude higher than the nonavian dinosaurs) and so much ecological diversity.The use of fossils allowed Benson et al. [22] to address deep-time radiation in dinosaurs and its consequence on present-day bird diversity. Nevertheless, the promise of using fossils to understand adaptive radiation has its limits. The paleontological dataset presented here is exceptional, yet still insufficient to explore major components of adaptive radiations like actual ecological diversification. As in many paleontological studies, Benson et al. used body-size data to represent ecology because body size is one of the few variables that is available for most species. But it is unclear how important body size really is for ecological diversification and niche filling, because body size is important for nearly every aspect of organismal function. Consequently, evolutionary change in body size can result not only from the competition that drives adaptive radiation, but also from predation pressure, reproductive character displacement, and physiological advantages of particular body sizes in a given environment, among other reasons [27].Despite the broad coverage of extinct species presented in Benson et al. [22], the data were insufficient to study another major part of adaptive radiation: early bursts of lineage diversification. While new approaches are becoming available to study diversification with phylogenies containing extinct species [28],[29] or with incomplete fossil data [30], these approaches are limited when many taxa are known from only single occurrences. This is the case in the Benson et al. dataset, and more generally in most fossil datasets.Given that few fossils exist for many extant groups, a major goal for future studies will be the incorporation of incomplete fossil information into analyses primarily focused on traits and clades for which mostly neontological data are available. For example, Slater et al. [31] developed an approach to include fossil information in analyses of phenotypic evolution. They showed that adding just a few fossils (12 fossils in a study of a 135-species clade) drastically increased the power and accuracy of their analyses of extant taxa. Thus, the combination of fossil data and those based on currently living species is important for future studies, as are new approaches that allow analyzing early bursts of lineage diversification along with phenotypic evolution in fossils.So what answers do Benson et al. [22] bring to Simpson''s original question of the importance of adaptive radiation for explaining diversity on earth? The authors present an intriguing and unconventional link between adaptive radiation and the diversity of modern-day birds. They argue that bird diversification was possible because the dinosaur lineage leading to birds did not exhaust niche space, potentially thanks to small body sizes; in contrast, other dinosaur groups adaptively radiated, filled niche space, and thus could not produce the ecological innovation that may have been necessary to survive the Cretaceous-Paleogene mass extinction. This intriguing hypothesis suggests an important role for the relative starting points of successive adaptive radiations in explaining current diversity, giving a new spin to the pivotal question raised by Simpson more than 60 years ago.     

Rapid Assessment Procedure for Loiasis and Mapping Lymphatic Filariasis: Two Perfect Illustrations of “To Be in English or Not to Be”

Interest in filariasis has found a new impetus now that neglected tropical diseases have their own journal. However, some of the advances published in renowned international journals have completely ignored previous publications on the subject, particularly those in languages other than English. The rapid assessment procedure for loiasis and the mapping of lymphatic filariasis provide two perfect illustrations of this. This problem may seem a bit outdated, given that all “good authors” now publish exclusively in English. It certainly is outdated for most areas of medicine. But, surely, this should not be the case for neglected tropical diseases, for which certain long-standing findings are every bit as important as what may be presented as new discoveries. One possibility would be for certain journals, such as PLOS Neglected Tropical Diseases, to include a specific heading permitting the publication in English of older studies that initially appeared in a language other than English. The texts would be English versions respecting the entirety of the original text. Submission should be accompanied by a presentation of the problem, with details and explanatory comments, with submission at the initiative of the authors of the former article in question or their students or sympathizers.Interest in filariasis has found a new impetus now that neglected tropical diseases (NTDs) have their own journal. However, some of the advances published in renowned international journals have completely ignored previous publications on the subject, particularly those in languages other than English. This Viewpoint article is intended to make us ponder the issue of a language gap or discrimination existing in publishing outcomes and reference citations. This is also the question of deleterious effects of the obligation “to be in English or not to be”.The rapid assessment procedure for loiasis (RAPLOA) and the geographical distribution of lymphatic filariasis provide two perfect illustrations of this.The RAPLOA has recently been widely used to determine the regional endemicity of loiasis and to update existing endemicity data for this disease over its global distribution range. This important work has been recently published in PLOS NTDs [1]. The determination, within a population of the prevalence or, preferably, the annual incidence of episodes, of conjunctival migration by adult worms is a simple, non-invasive, relatively sensitive and specific method for evaluating the endemicity of Loa loa. This approach has proved particularly useful in areas in which both loiasis and onchocerciasis are observed: the mass treatment program to control onchocerciasis is based on the use of ivermectin and there is a risk of adverse treatment outcomes in patients carrying large numbers of L. loa worms [2]. In regions of high endemicity, the correlation between the conjunctival migration index and the microfilarial index is strong overall, both for villages and for age groups. Its use as an epidemiological index was clearly proposed in a publication in 1994 [3]. However, as this article was published in French, in Médecine Tropicale (Marseille), it has never been cited, despite being listed in international databases, including PubMed. A poster communication concerning the same issue had no real impact either, despite being presented at an international congress [4]. The origin of this new epidemiological index (RAPLOA) is systematically attributed to two World Health Organization (WHO) publications in 2001 [5] and 2002 [6]. It is true that the studies reported in these publications validated the concept at a large scale and in different endemic foci.The usefulness of specific clinical manifestations (eye worm and Calabar swelling) to assess L. loa had been recommended by different authors as early as 1950 [7]. But the correlation between the microfilarial index and the frequency of ocular migration has not been studied, and even less attention was paid to the notion of an epidemiological index until the epidemiological studies carried out in Congo Republic (former People''s Republic of the Congo) during the 1980s [8]. However, the index as such was clearly defined in 1994 [3]. Here is a direct translation of an excerpt of the French text published in 1994: “For loiasis, the usual parasitological indices (microfilarial index and mean microfilarial density) are the only measures recognized as providing information about the level of endemicity in humans. In addition to requiring blood samples standardized in terms of both volume and sampling time, these indices do not reflect the real level of parasitism, given the high frequency of infected subjects without microfilaria in the blood. Subjects infested with mature, fertile adult worms, as demonstrated by the removal of a subconjunctival female containing microfilaria from a patient with no detectable microfilaria in the blood, are frequently observed. Two symptoms are both specific and frequent in infected subjects both with and without microfilaria in the blood: subconjunctival migration of an adult worm and elusive, migrating edemas of the hands, wrists and lower part of the forearm” [9]. “The index of subconjunctival filarial migration over the preceding year is particularly useful, because it correlates well with the microfilarial index but is more sensitive” (Figure 1). “Its determination involves precise questioning of the patient, which can be facilitated by the use of a demonstration chart, with diagrams and photographs” (see Figure 2).Open in a separate windowFigure 1Index of the subconjunctival migration of Loa loa adult worms and microfilarial index.Reproduced from Medicine Tropicale [3], released under CC BY 2.0 by Medicine Tropicale. IMSC = Indice de Migration Sous-Conjonctivale in French and Index of the SubConjunctival Migration in English. IM = Indice Microfilarien in French and Microfilarial Index in English.Open in a separate windowFigure 2Illustration of the passage of an adult worm (Loa loa) across the eye.This illustration (diagram and photograph) was made for presentation to patients questioned in endemic regions.The conclusion of this article was formulated as follows: “Screening for foci of filarial endemicity could be improved by the use of a simplified method and the validation of simple, inexpensive indices. Once these foci have been identified, a more precise evaluation can be carried out.”What is most astounding about the two WHO publications cited as the origin of this “new epidemiological index” [5], [6] is that the principal authors come from French-speaking African countries and/or work in or with this institution. They would therefore have been able to understand articles written in French. Furthermore, the WHO has a long-standing culture of multilingualism, particularly in English and French.Against this background, the rejection by the Bulletin of the World Health Organization and by other international journals published in English of an opinion article dealing with this issue and using the example of lymphatic filariasis does not seem to be justified, and is another illustration of “to be in English or not to be.”Indeed, filariasis due to Wuchereria bancrofti is systematically described as endemic in Congo and Gabon, two French-speaking countries, in non-specialist works on tropical medicine and in more specialist publications (WHO) despite a total absence of epidemiologic studies and/or confirmed case report over the last 30 years to prove it. What is certain is that no case was found when the last studies were conducted in these countries at the end of the 1970s and during the 1980s but unfortunately published in French. The studies that we carried out in the Congo as part of the National Project on Onchocerciasis and Other Filariases (between 1982 and 1987) confirmed the presence of four types of human filariasis: onchocerciasis, loaiasis, and the filariases caused by Mansonella perstans and M. streptocerca. There was a total absence of confirmed cases of lymphatic filariasis (bancroftosis). In this case, it is not a question of the attribution of merit for a particular “discovery”, but of basic knowledge of the geographic distribution of a scarcely studied disease, lymphatic filiariasis, in French Central Africa. Taking into account only publications in English, even older and poorly structured data, have been, in our opinion, a source of confusion and has led to false conclusions being drawn about the distribution range of this disease. This undoubtedly highlights the need to update knowledge by carrying out prospective studies (which seem to be underway), but these studies do not seem to be considered a matter of priority given the low levels of resources available and current health priorities.This has drawn us to publish this article in a French-language journal, but together with an entire translation into English [10]. Despite the bilingual nature of this publication, the international PubMed database identifies this article as being published in French, effectively ensuring that it will never be consulted, a classic “catch 22” situation! Indeed, this reference has never yet been cited by another author in a journal published in English. It may be that publication of an article in another language than English makes it more likely that it will not be cited, even if the authors of a subsequent article have access to the journal in which it was published and can understand the language used. Here, we begin to encroach on ethical problems and it is probably best not to delve too deeply. However, suffice it to say that the limited dissemination of publications in a language other than English may account for such equivocal attitudes.The problem is not a rivalry between French and English, but the confrontation between English and all other languages of the world. Moreover, the problem is undoubtedly worse for works published in non–Western European languages such as Chinese, Russian, and Japanese, which are arguably even less accessible.All things considered, this problem may seem a bit outdated, given that all “good authors” now publish exclusively in English. It certainly is outdated for most areas of medicine, where everything that is old is assigned to being nothing more than the history of medicine. But, surely, this should not be the case for NTDs, for which certain long-standing findings are every bit as important as what may be presented as new discoveries.One possibility would be for certain journals, such as PLOS NTDs, to include a specific heading permitting the publication of older studies that initially appeared in a language other than English (and are therefore currently little known). The texts included in this heading would essentially be English versions of these articles previously published in other languages, respecting the entirety of the original text.This would concern studies considered of importance because they highlight a point that remains unclear or describe an aspect considered innovative in a review but for which the originality of the article is due more to an incomplete reference list than to a true advance in knowledge. These articles should be judged in light of the knowledge and technical and methodological means available at the time at which they were initially published. Submission should be accompanied by a presentation of the problem, with details and explanatory comments, with submission at the initiative of the authors of the article in question or their students or sympathizers.