Phylogenetic inference: how much evolutionary history is knowable?

In order to reconstruct phylogenetic trees from extremely dissimilar sequences it is necessary to estimate accurately the extent of sequence divergence. In this paper a new method of sequence analysis, Markov triple analysis, is developed for determining the relative frequencies of nucleotide substitutions within the three branches of a three-taxon dendrogram. Assuming that nucleotide sites are independently and identically distributed and assuming a Markov model for nucleotide (or protein) evolution, it is shown that the unique Markov matrices can be reconstructed given only the joint probability distribution relating three taxa. (In the much simpler case involving only two taxa and two character states, Markov matrices can also be reconstructed, provided symmetry assumptions are placed on the elements of the matrices.) The method is illustrated using sequence data from the combined first and second codon positions derived from complete human, mouse, and cow mitochondrial sequences.      

DNA hybridization evidence for the principal lineages of hummingbirds (Aves:Trochilidae)

The spectacular evolutionary radiation of hummingbirds (Trochilidae) has served as a model system for many biological studies. To begin to provide a historical context for these investigations, we generated a complete matrix of DNA hybridization distances among 26 hummingbirds and an outgroup swift (Chaetura pelagica) to determine the principal hummingbird lineages. FITCH topologies estimated from symmetrized delta TmH-C values and subjected to various validation methods (bootstrapping, weighted jackknifing, branch length significance) indicated a fundamental split between hermit (Eutoxeres aquila, Threnetes ruckeri; Phaethornithinae) and nonhermit (Trochilinae) hummingbirds, and provided strong support for six principal nonhermit clades with the following branching order: (1) a predominantly lowland group comprising caribs (Eulampis holosericeus) and relatives (Androdon aequatorialis and Heliothryx barroti) with violet-ears (Colibri coruscans) and relatives (Doryfera ludovicae); (2) an Andean-associated clade of highly polytypic taxa (Eriocnemis, Heliodoxa, and Coeligena); (3) a second endemic Andean clade (Oreotrochilus chimborazo, Aglaiocercus coelestis, and Lesbia victoriae) paired with thorntails (Popelairia conversii); (4) emeralds and relatives (Chlorostilbon mellisugus, Amazilia tzacatl, Thalurania colombica, Orthorhyncus cristatus and Campylopterus villaviscensio); (5) mountain-gems (Lampornis clemenciae and Eugenes fulgens); and (6) tiny bee-like forms (Archilochus colubris, Myrtis fanny, Acestrura mulsant, and Philodice mitchellii). Corresponding analyses on a matrix of unsymmetrized delta values gave similar support for these relationships except that the branching order of the two Andean clades (2, 3 above) was unresolved. In general, subsidiary relationships were consistent and well supported by both matrices, sometimes revealing surprising associations between forms that differ dramatically in plumage and bill morphology. Our results also reveal some basic aspects of hummingbird ecologic and morphologic evolution. For example, most of the diverse endemic Andean assemblage apparently comprises two genetically divergent clades, whereas the majority of North American hummingbirds belong a single third clade. Genetic distances separating some morphologically distinct genera (Oreotrochilus, Aglaiocercus, Lesbia; Myrtis, Acestrura, Philodice) were no greater than among congeneric (Coeligena) species, indicating that, in hummingbirds, morphological divergence does not necessarily reflect level of genetic divergence.      

The sensitivity of Afromontane tarns in the Maloti-Drakensberg region of South Africa and Lesotho to acidic deposition

Despite their remoteness from sources of atmospheric pollutant emissions, the Afromontane tarns in the Maloti-Drakensberg region are perfect candidates to study the negative effects of acidifying atmospheric pollution, because mountain lakes are widely recognised as sentinel ecosystems, unimpacted by direct human disturbance within their catchments. Thirty-four tarns were sampled in the Maloti-Drakensberg region and most were found to be extremely sensitive to acidic deposition, as indicated by their low acid neutralising capacity. There are very few studies of freshwater critical loads for any region within South Africa. The steady-state water chemistry model (SSWC) was adapted and used to determine critical loads, whereas exceedance was estimated relative to modelled regional deposition data, in order to understand the risk of harmful effects to aquatic ecosystems. Seventy-six percent of sampled sites across the Maloti-Drakensberg would exceed critical loads even at the lowest modelled deposition levels, but there are no current measured deposition data for the region. The sensitivity of the Maloti-Drakensberg tarns needs to be considered in future policy formulation regarding acceptable levels of acidifying atmospheric pollution from South Africa’s energy sector and indicates the need for assessing aquatic ecosystem impacts in other regions of South Africa.     

Differences in the carbon tetrachloride-induced damage to components of the smooth and rough endoplasmic reticulum from rat liver

There is a higher activity of ethyl morphine N-demethylase (EM-ase) and cytochrome P-450 (P-450) reductase as well as higher P-450 content in the smooth endoplasmic reticulum (SER) than in the rough endoplasmic reticulum (RER). The extent of the irreversible binding of the¹⁴C from¹⁴CCl₄ to lipids and proteins, as well as the CCl₄-induced destruction of P-450 is more intense in SER than in RER while the opposite was found for glucose 6-phosphatase (G6P-ase) destruction. CCl₄-induced lipid peroxidation is as intense in SER as is in RER.¹⁴C from¹⁴CCl₄ gets irreversibly bound to ribosomal proteins.     

Genomic characterisation of the effector complement of the potato cyst nematode Globodera pallida

Background

The potato cyst nematode Globodera pallida has biotrophic interactions with its host. The nematode induces a feeding structure – the syncytium – which it keeps alive for the duration of the life cycle and on which it depends for all nutrients required to develop to the adult stage. Interactions of G. pallida with the host are mediated by effectors, which are produced in two sets of gland cells. These effectors suppress host defences, facilitate migration and induce the formation of the syncytium.

Results

The recent completion of the G. pallida genome sequence has allowed us to identify the effector complement from this species. We identify 128 orthologues of effectors from other nematodes as well as 117 novel effector candidates. We have used in situ hybridisation to confirm gland cell expression of a subset of these effectors, demonstrating the validity of our effector identification approach. We have examined the expression profiles of all effector candidates using RNAseq; this analysis shows that the majority of effectors fall into one of three clusters of sequences showing conserved expression characteristics (invasive stage nematode only, parasitic stage only or invasive stage and adult male only). We demonstrate that further diversity in the effector pool is generated by alternative splicing. In addition, we show that effectors target a diverse range of structures in plant cells, including the peroxisome. This is the first identification of effectors from any plant pathogen that target this structure.

Conclusion

This is the first genome scale search for effectors, combined to a life-cycle expression analysis, for any plant-parasitic nematode. We show that, like other phylogenetically unrelated plant pathogens, plant parasitic nematodes deploy hundreds of effectors in order to parasitise plants, with different effectors required for different phases of the infection process.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-15-923) contains supplementary material, which is available to authorized users.     

Anecdotal,Historical and Critical Commentaries on Genetics: The Impact of Whole Genome Sequencing on Model System Genetics: Get Ready for the Ride

Much of our understanding of how organisms develop and function is derived from the extraordinarily powerful, classic approach of screening for mutant organisms in which a specific biological process is disrupted. Reaping the fruits of such forward genetic screens in metazoan model systems like Drosophila, Caenorhabditis elegans, or zebrafish traditionally involves time-consuming positional cloning strategies that result in the identification of the mutant locus. Whole genome sequencing (WGS) has begun to provide an effective alternative to this approach through direct pinpointing of the molecular lesion in a mutated strain isolated from a genetic screen. Apart from significantly altering the pace and costs of genetic analysis, WGS also provides new perspectives on solving genetic problems that are difficult to tackle with conventional approaches, such as identifying the molecular basis of multigenic and complex traits.GENETIC model systems, from bacteria, yeast, plants, worms, flies, and fish to mice allow the dissection of the genetic basis of virtually any biological process by isolating mutants obtained through random mutagenesis, in which the biological process under investigation is defective. Such forward genetic analysis is unbiased and free of assumptions. The rigor and conceptual simplicity of forward genetic analysis is striking, some may say, beautiful; and the unpredictability of what one finds—be that an unexpected phenotype popping out of a screen or the eventual molecular nature of the gene (take the discovery of miRNAs as an example; Lee et al. 1993)—appeals to the adventurous. Even though mutant phenotypic analysis alone can reveal the logic of underlying biological processes (take Ed Lewis'' analysis of homeotic mutants as an example; Lewis 1978)—it is the identification of the molecular lesions in mutant animals that provides the key mechanistic and molecular details that propel our understanding of biological processes.The identification of the molecular lesion in mutant organisms depends on how the mutation was introduced. Classically, two types of mutagens have been used in most model systems: biological agents such as plasmids, viruses, or transposons whose insertions disrupt functional DNA elements (either coding or regulatory elements) or chemical mutagens, such as ethyl methane sulfonate (EMS) or N-ethyl N-nitroso urea (ENU), that introduce point mutations or deletions. Point mutation-inducing chemical mutagens are in many ways a superior mutagenic agent because their mutational frequency is high and because the spectrum of their effects on a given locus—producing hypomorphs, hypermorphs, amorphs, neomorphs, etc.—is hard to match by biological mutagens. Moreover, chemical mutagens do not display the positional bias of many biological agents. In addition, point mutations in a gene are often crucial in dissecting the functionally relevant domains of the gene product. In spite of the advantages of chemical mutagens, model system geneticists often prefer biological mutagens simply because the molecular lesions induced by those agents are characterized by the easily locatable DNA footprint that these agents generate. In contrast, the location of a point mutation (or deletion) has to be identified through conventional mapping strategies, which tend to be tedious and time consuming. Even in model systems in which positional cloning is quite fast and straightforward (e.g., Caenorhabditis elegans, which has a short generation time and a multitude of mapping tools available), it nevertheless is a significant effort that can occasionally present hurdles that are difficult to surmount (e.g., if the gene maps into a region with few genetic markers that allow for mapping). These difficulties explain why RNAi-based “genetic screens” have gained significant popularity in C. elegans; they circumvent mapping and reveal molecular identities of genes involved in a given process straight away (Kamath and Ahringer 2003). However, genes and cells show differential susceptibility to RNAi; off-target effects and lack of reproducibility can be a problem, and the range of effects that RNAi has on gene activity is generally more limited compared to chemically induced gene mutations.The recent application of next generation, deep sequencing technology (see Bentley 2006; Morozova and Marra 2008 for technology reviews) is beginning to significantly alter the landscape of genetic analysis as it allows the use of chemical mutagens without having to deal with its disadvantages. Deep sequencing technology incorporated into platforms such as Illumina''s Genome Analyzer or ABI''s SOLiD, allows one-shot sequencing of the entire model system''s genome, resulting in the detection of mutagen-induced sequence alterations compared to a nonmutagenized reference genome. Proof-of-concept studies have so far been conducted in bacteria, yeast, plant, worms, and flies, all published within the last year (Sarin et al. 2008; Smith et al. 2008; Srivatsan et al. 2008; Blumenstiel et al. 2009; Irvine et al. 2009; Rigola et al. 2009). Many more studies are under way; for example, since our first proof-of-principle study (Sarin et al. 2008), my laboratory has identified the molecular basis of >10 C. elegans strains defective in neuronal development and homeostasis (V. Bertrand, unpublished data; M. Doitsidou, unpublished data; E. Flowers, unpublished data; S. Sarin, unpublished data).The advantages of whole genome sequencing (WGS) are obvious. The process is extraordinarily fast with the sequencing taking only ∼5 days and the subsequent sequence data analysis only a few hours, particularly if the end user employs bioinformatic tools customized for mutant detection (Bigelow et al. 2009). The process is also remarkably cost effective. For example, a C. elegans genome can be sequenced with a required sequence coverage of ∼10 times for <$2,000 in reagent and machine operating costs. The capacity of deep sequencing machines—and hence the costs associated with sequencing a genome—apparently follow Moore''s law of doubling its capacity about every 2 years, like many technological innovations do (Pettersson et al. 2009). That is, the <$1,000 genome for C. elegans (∼100-Mb genome) and Drosophila (∼123-Mb genome) is just around the corner and other models will sooner or later follow suit. The cost effectiveness becomes particularly apparent if one compares the cost of WGS to the personnel and reagent costs associated with multiple-month to multiple-year mapping-based cloning efforts.WGS identifies sequence variants between a mutated genome and a premutagenesis reference genome. Chemical mutagens randomly introduce many mutations in the genome and, therefore, the phenotype-causing sequence variant needs to be identified as such out of a large pool of sequence variants. Sequence variants that have no impact on the phenotype can be outcrossed before sequencing or eliminated through some rough mapping of the mutation, which allows the experimenter to focus only on those variants contained in a specific sequence interval. Ensuing functional tests such as transformation rescue or phenocopy by RNAi and the availability of other alleles of the same locus are critical means to validate a phenotype-causing sequence variant (Sarin et al. 2008). The latter approach—the availability of multiple alleles of the same locus—is in many ways the most powerful one to sift through a number of candidate variants revealed by WGS. In this approach, candidate loci revealed by WGS are resequenced by conventional Sanger sequencing in allelic strains and only those that are indeed phenotype causing will show up mutated in all allelic variants of the locus (Sarin et al. 2008). The availability of multiple alleles of a locus is highly desirable for many aspects of genetic analysis anyway and therefore does not represent an additional and specific burden for undertaking a WGS project.The importance of WGS on model system genetics will be substantial and wide ranging. Speed and cost effectiveness means that the wastelands of genetic mapping can be trespassed fast enough to allow an experimenter to multitask a whole mutant collection in parallel, thereby closing in on the “holy grail” of genetic analysis—the as-complete-as-possible mutational saturation of a biological process and the resulting deciphering of complete genetic pathways and networks. What will become limiting steps are not any more the tediousness of mapping, but rather the effectiveness with which mutant collections can be built. Novel technologies that involve machine-based, semiautomated selection of mutant animals have been developed over the past few years to study a variety of distinct biological processes in several metazoan model systems, e.g., gfp-based morphology or cell fate screens in worms (Crane et al. 2009; Doitsidou et al. 2008) or behavioral screens in flies (Dankert et al. 2009) and are important steps in this direction. Such an “industrial revolution” of genetic screening (i.e., the mutant selection part, followed by WGS) moves us geneticists away from, not into the trenches of factory life and frees us up to do what we should like to enjoy most—thinking of designing interesting screens, seeing how genes interact, and interpreting it all.Another important impact of WGS is that it will allow tackling problems that were previously hard to deal with. For example, the tediousness of following subtle phenotypes, low penetrance phenotypes, or phenotypes that are cumbersome to score often hampers positional cloning approaches that rely on identifying rare recombinants in a large sibling pool. Moreover, many genetic traits such as behavioral genetic traits are very sensitive to genetic background and are therefore also often hard to map in the conventional way. WGS hones in on candidate genes straight away. Taking this notion a step further, WGS will also be able to get at the molecular basis of multigenic traits and quantitative trait loci, which again are hard to molecularly identify through conventional mapping strategies; a proof-of-principle study has made this point already in bacteria (Srivatsan et al. 2008). In principle, such multigenic traits may have been mutationally induced or could be present in natural variants of a species, which provides intriguing perspective for the population geneticist.Model organisms of biological interest that were previously relatively intractable for classic genetic mutant analysis due to the absence of genetic markers or other practical problems such as prohibitive generation times, may also now be movable into the arena of genetic model systems, through the WGS-mediated molecular analysis of mutagen-induced variants or through the study of natural variants.The sequencing of human cancer genomes has already begun to illustrate the impact of WGS on human genetics (Campbell et al. 2008; Ley et al. 2008). However, those human WGS studies illustrate why model systems will continue to be extremely important—their experimental accessibility allows us to address which of the many variants detected by WGS is indeed the phenotype-causing one.The message to model system geneticists is clear: Get access to a deep sequencer, buckle up, and get ready for the ride.     

High yield expression of leptospirosis vaccine candidates LigA and LipL32 in the methylotrophic yeast <Emphasis Type="Italic">Pichia pastoris</Emphasis>

Background  

Leptospirosis, a zoonosis caused by Leptospira spp., is recognized as an emergent infectious disease. Due to the lack of adequate diagnostic tools, vaccines are an attractive intervention strategy. Recombinant proteins produced in Escherichia coli have demonstrated promising results, albeit with variable efficacy. Pichia pastoris is an alternative host with several advantages for the production of recombinant proteins.