Using Ancestral Sequence Reconstruction to Understand Protein Function
30-31 March 2005
Kristineberg, Sweden

Organisers:

David Liberles, University of Bergen, Norway
Giorgio Matassi, University of Paris VI, France
David Ardell, Uppsala University, Sweden

Report

Summary

"Using Ancestral Sequence Reconstruction to Understand Protein Function" took place in Kristineberg, Sweden March 30-31, 2005. A total of 38 participants from 12 countries, including 11 from Norway, 6 from Denmark, 4 from USA, and 3 each from Sweden, France, and Ireland took part. The international composition of the meeting coupled with the high level of scientific discourse resulted in a very successful meeting. A total of 18 scientific presentations over the two day period covered topics ranging from the use of ancestral sequence reconstruction in the pharmaceutical industry to computational methodology, including assessments of accuracy and systematic biases, to genome and proteome reconstruction to experimental synthesis for evolutionary hypothesis testing to purely computational applications of ancestral sequences. Each day closed with a lively discussion session. An edited volume is currently being prepared based upon the meeting, which is under consideration for publication at Oxford University Press.

Scientific Content and Future Directions

Following a meeting introduction written by Emile Zuckerkandl (Stanford University, USA), one of the grandfathers of molecular evolution, read by Giorgio Matassi (University of Paris VI, France), the scientific content of the meeting opened with a talk by Marie Skovgaard (Novo Nordisk, Denmark and University of Bergen, Norway). She presented ongoing work examining the phylogeny and evolution of the proglucagon gene family. A novel member of this family isolated from gila monster has interesting pharmacoactive properties and computational ancestral sequence reconstruction coupled to experimental peptide synthesis is being used to test the link between sequence and function in the evolution of these novel properties. This case study was presented as an example of how ancestral sequence reconstruction can be used to link sequence to function with potential pharmaceutic applications.

The second talk by David Liberles (University of Bergen, Norway) presented computational approaches for detecting significant changes in gene expression driven by promoter substitution and in protein coding sequence potentially linked to functional change using a phylogenetic approach. These approaches have been applied systematically to large scale datasets, yielding a set of candidates for experimental analysis of functional shifts along specific branches of gene family trees using ancestral sequence reconstruction.
The next set of talks presented computational methodologies for ancestral sequence reconstruction. Tal Pupko (Tel Aviv University, Israel) presented an algorithm for maximum likelihood joint ancestral sequence reconstruction. Julien Duthiel (University of Montpellier II, France) and Jonathan Bollback (University of Copenhagen, Denmark) presented statistical methods for mapping substitutions onto branches of phylogenetic trees, enabling analysis of coevolution of sites along branches. Richard Edwards (Royal College of Surgeons, Ireland) presented an ancestral sequence reconstruction method that explicitly deals with gaps as an additional character state.

David Pollock (Louisiana State University, USA) gave a cautionary presentation on the potential perils and biases of ancestral sequence reconstruction. He found that Bayesian methods were less biased than likelihood or parsimony in his examination of DNA compositional bias. Further, he warned that a telltale sign of a problem of bias occurred when the activity of the ancestral state was more extreme than any of the extant sequences. A second caution was that ancestral sequences should be generated by sampling from the posterior distribution rather than taking the maximum likelihood estimate at every position. This led to a spirited discussion, where not everyone agreed with that point. Some discussion of criticality and the type of function under selection followed, with the view that traits effected by a small number of residues (for example binding) may be different from traits influenced by criticality over a large number of sites (like protein stability or DNA melting temperature). This talk ultimately served as one of the main focal points of the meeting.

Toni Gabaldon (University of Nijmegen, Netherlands) and Pierre Pontarotti (University of Provence, France) presented genomic and proteomic computational ancestral sequence reconstructions. Gabaldon analyzed the metabolism of the ancestral mitochondrium while Pontarotti presented a detailed reconstruction of gene content and order of the major histocompatibility region of the last common ancestor of all species with bilateral symmetry (urbilateria).

On the second day, attention turned to the growing field of those experimentally synthesizing reconstructed ancestral states to test specific evolutionary hypotheses about the evolution of biological function. Slim Sassi (University of Florida, USA) presented an analysis of the evolution of seminal ribonuclease in bovids and their emergent immunosuppressive role.

David Ardell (Uppsala University, Sweden) presented a comparison of sequence alignment and structural alignment from an evolutionary perspective, calling into question the view of structural alignment as the gold standard for alignment. He analyzed cases where structure may be sliding through sequence, generating positions that are aligned structurally, but not through sequence homology.

Returning to experimental ancestral sequence reconstruction, Belinda Chang (University of Toronto, Canada) presented a reconstruction of the ancestral visual pigments of the last common ancestor of birds and crocodiles, which is also the last common ancestor of these species with dinosaurs. The ancestral protein was found to be slightly red-shifted in comparison with extant birds and alligator.

Joe Thornton (University of Oregon, USA) analyzed the evolution of hormones and binding specificity. Ancestral sequence reconstruction suggested that the original hormone receptor bound estrogen, which was consistent with a segmental evolution of hormone metabolism. Denis Shields (Royal College of Surgeons, Ireland) presented a systematic analysis of binding specificity, using ancestral sequence reconstruction.

Eric Gaucher (The Foundation for Applied Molecular Evolution, USA) attempted to go back much further in evolution, addressing the temperature that the last common ancestor of eubacteria lived in. Elongation Factor-Tu is one of the few molecules that aligns well enough to go back that far and this was used to address this question. Another case study involving alcohol dehydrogenase was also presented.

Victor Albert (University of Oslo, Norway) presented a hypothesized role for a C-C motif effecting activity of COX 1 in bladderworts and Welwitschia as a key link between environment and energy.

In the last purely computational session, Richard Goldstein (National Institute for Medical Research, UK) presented a phylogeny-based HMM with the potential utility of generating ancestral sequences. In another variation, Gina Cannarozzi (Swiss Federal Institute of Technology-Zurich, Switzerland) presented empirical codon matrices as an alternative to parameterized matrices from maximum likelihood for use in reconstructing ancestral sequences. Tal Pupko (Tel Aviv University, Israel) also formulated a similar matrix.

When all was summed up, the three themes of the meeting were:
1. 1. Embrace uncertainty in ancestral sequence reconstruction
2. 2. Think about criticality and the types of selective pressures
3. 3. Watch your alignments, as good alignments are the key to accurate ancestral sequence reconstruction

The meeting brought together many of the important researchers in the field, especially on the European side, together with a few trans-Atlantic participants. A follow-up meeting is envisioned in North America. Further, an edited volume is being prepared based upon this meeting with ongoing discussions with Oxford University Press towards publication of the volume. This is a technique of growing importance that is in the process of transforming from a niche of molecular evolution to a part of mainstream molecular biology. This meeting was important in discussing the problems and prospects as the field matures.

Summary Programme

Detailed Programme with Abstracts

Wednesday, March 30, 2005
8:45-9:00 Organizers' Welcome (Ardell, Liberles, Matassi)
9:00-9:45 Marie Skovgaard (Novo Nordisk, Denmark and University of Bergen, Norway)
A search for future drugs in evolutionary history
A number of native human peptides have potential as therapeutic drugs, as they often are involved in signaling pathways in the human body. Nature provides a wide variety of orthologous and paralogous peptides which often maintain the ability to activate human receptors. Phylogeny serves as an important method to establish the evolutionary relationship between two such peptides, and can restrict the area of organisms in which a new drug candidate should be expected to be found. Moreover, ancestral sequence reconstruction can be used not only to see how the functionality evolved but also potentially to discover new drugs. As an example, the glucagon like peptide 1 as a potential drug is discussed and it is shown how ancestral sequence reconstruction can be used to widen the number of potential drug candidates.

9:45-10:15 David Liberles (University of Bergen, Norway)
Genes, genomes and evolution: Insights from ancestral character states
Ancestral sequence reconstruction is a powerful method for understanding the evolution of genes and their function in genomes. To understand the evolution of gene expression, promoter sequences were reconstructed at the last common ancestor of human and chimpanzee with mouse as an outgroup. Gene expression evolution was analyzed over the tree using a minimum evolution algorithm and substitutions in transcription factor binding sites along lineages with particularly rapid expression level evolution were identified.
Protein-coding sequences have also been systematically analyzed using ancestral sequence reconstruction. Explicit reconstruction of ancestral sequences across chordate and embryophyte gene families has been coupled with calculation of Ka/Ks ratios to detect lineage-specific positive selection in The Adaptive Evolution Database (TAED). This database contains examples of protein-encoding sequences where function is hypothesized to have changed under selective pressure in specific lineages. Several examples, including myostatin in Artiodactyls, are currently being studied experimentally in more detail using various methods, including experimental ancestral sequence reconstruction.
10:15-10:45 BREAK

10:45-11:30 Tal Pupko (Tel Aviv University, Israel)
Algorithmic challenges in ancestral sequence reconstruction: Taking variation among sites into account
More realistic models of sequence evolution are constantly being developed. Probabilistic models are considered the state-of-the-art approach for phylogenetic reconstruction. Applying these models for ancestral sequence reconstruction can be computationally trivial in some cases. For example, marginal reconstruction of ancestral sequences, with and without assuming rate variation among sites can be done in linear time. However, there is no efficient algorithm for joint reconstruction of ancestral sequence taking into account among site rate variation. I will present a branch-and-bound algorithm that guarantees finding the maximum likelihood joint reconstruction when the rate distribution is Gamma (I will also discuss alternative, more complex rate distributions). Additionally, I will also discuss a novel codon-based model for phylogenetic inference that can also be used for better ancestral reconstruction of coding sequences.

11:30-12:15 Julien Dutheil (University of Montpellier II, France)
Statistical approaches to the substitution mapping problem and detection of coevolution between sites
We introduce a new, general probabilistic substitution mapping procedure taking into account uncertainty over ancestral states, multiple substitutions and among site rate variations.
This methods provides an unbiased estimate of the number of changes having occurred at a specific site in a specific branch of the tree. We then define a new statistic for the detection of co-evolving sites in a molecule, based on this mapping.
This method was first applied to a benchmark rRNA dataset, and succeeded in retrieving significantly correlated pairs already known as co-evolving (stem-pairs and documented tertiary interactions). Some extensions of the method for the protein case are then presented.
12:15-13:30 LUNCH

13:30-14:15 Jonathan Bollback (University of Copenhagen, Denmark)
Ancestral character reconstruction: Bayesian ancestral state reconstruction and stochastic character histories
Reconstructing ancestral character states at a particular focal point in history (node or interval) and summarizing states across the known history (phylogeny) are important to our understanding of morphological and molecular evolution. Ancestral state reconstruction (ASR) remains important because it is the basis for studies concerned with correlated evolution, reconstructing ancient molecules or behaviors, and understanding historical patterns and trends of character (morphological or molecular) evolution. Stochastic character histories (SCH) moves beyond ASR by providing information on not only reconstructions at the internal nodes of a tree, but also detailed information about the timing and types of changes along the branches. The inferences made in these kinds of studies depend, often critically, on the reliability of the ancestral estimates at the internal nodes, which in turn may depend on the particulars of the analysis such as the topology, branch lengths, and substitution parameters. Phylogenetic uncertainty has usually been ignored in ancestral state reconstruction. A hierarchical Bayesian approach is presented that accommodates uncertainty in the topology, branch lengths, and model parameters using MCMC. The full hierarchical approach is compared to an empirical Bayesian approach, in which these values are fixed. A discussion of more recent work which has extended our understanding of character evolution by stepping beyond ASR to describing a more complete character history that includes not only reconstructions at the internal nodes but the timing (placement) and types of character changes along the branch between the nodes. An introduction to stochastic character histories will be given with emphasis on the differences between typical ancestral state reconstruction, possible applications, and the program SIMMAP will be introduced as a tool for sampling character histories and performing hierarchical Bayesian analyses of ancestral states.

14:15-15:00 Richard Edwards (Royal College of Surgeons, Ireland)
GASP (Gapped Ancestral Sequence Prediction)
Ancestral sequence reconstructions can aid prediction of residues conferring functional specificity within families of related proteins. We have made use of the Burst After Duplication (BAD) method for such predictions. BAD identifies potentially specific residues as those that experience a radical amino acid substitution directly following duplication, which is subsequently conserved. For accurate predictions, it is helpful to maximise the number of sequences in the protein family alignment. This can often be at the cost of incorporating sequences with insertions/deletions (indels) events or sequence fragments. Most implementations of existing ancestral prediction algorithms are unable to process residues with gaps, which may result in a loss of vital data. We implemented a new algorithm, GASP (Gapped Ancestral Sequence Prediction), for predicting ancestral sequences from phylogenetic trees and the corresponding multiple sequence alignments. Alignments may be of any size and contain gaps. GASP first assigns the positions of gaps in the phylogeny before using a likelihood-based approach centred on amino acid substitution matrices to assign ancestral amino acids. Important outgroup information is used by first working down from the tips of the tree to the root, using descendant data only to assign probabilities, and then working back up from the root to the tips using descendant and outgroup data to make predictions. GASP was tested on a number of simulated datasets based on real phylogenies. Prediction accuracy for ungapped data was similar to three alternative algorithms tested, with GASP performing better in some cases and worse in others. Adding simple insertions and deletions to the simulated data did not have a detrimental effect on GASP accuracy.

15:00-15:45 David Pollock (Louisiana State University, USA)
Ancestral reconstruction bias: the monster under the bed
Reconstruction of ancestral protein sequences has provided many exciting inferences about ancestral protein function and how modern protein functions have come about. It has recently become clear, however, that bias in ancestral reconstruction may affect results regardless of the method, particularly for more ancient proteins. The ?monster under the bed? is that this bias can easily lead to incorrect functional interpretation for reconstructed proteins. It no longer seems acceptable to infer function or stability based on a single protein, nor is it enough to check for functional alteration by substituting obviously ambiguous sites. The only means to eliminate this bias is to sample ancestral proteins from the posterior probability space; if a residue has only 5% or 10% probability at an ancestral site, it still should be sampled 5% or 10% of the time. The question of how many reconstructed ancestral samples are sufficient to estimate probable ancestral function is an open one, and it may be specific to the variability in inferred function among likely ancestors.
15:45-16:15 BREAK

16:15-17:00 Toni Gabaldon (University of Nijmegen, Netherlands)
Reconstructing ancestral proteomes and tracing their evolution: the mitochondrial case
Mitochondria are eukaryotic organelles derived from the endosymbiosis of a single alpha-proteobacterium. After the endosymbiosis, the mitochondrial genome underwent an extensive reduction process involving the loss of many genes and the transfer of others to the nuclear genome. We used large-scale phylogenetic analyses and the availability of completely sequenced alpha-proteobacterial and eukaryotic genomes to search for eukaryotic genes likely derived from the proto-mitochondrion. These results were used to reconstruct the proto-mitochondrial proteome and its corresponding metabolism, which suggests a (facultatively) aerobic endosymbiont catabolizing lipids, glycerol and amino acids provided by the host. Recent data from proteomics of human and yeast mitochondria allow us to compare this ancestral endosymbiotic metabolism with that of modern mitochondria and therefore trace its evolution. Here I present our last data from the reconstruction of the ancestral mitochondrial proteome and its comparison with modern mitochondrial proteomes. The comparison reveals that the transition from endosymbiont to modern organelle was accompanied by major changes not only in the distribution of functional groups represented in the proteome but also in their composition. Finally I will zoom in on the evolution through the eukaryotic lineages of the mitochondrial NADH:ubiquinone oxidoreductase (complex I). References: 1-Gabaldon T. and Huynen MA. Reconstruction of the Proto-mitochondrial Metabolism. Science. (2003). 301(5633):609 2-Gabaldon T. and Huynen MA. Shaping the mitochondrial proteome. (2004) BBA-Bioenergetics. 1659(2-3):212-220

17:00-17:45 Pierre Pontarotti (University of Provence, France)
Towards the Reconstruction of the Urbilateria Genome: The Major Histocompatibility Complex Region Example
Our long term project is to trace the evolution of urbilateria (common ancestor of all the bilaterian species). The comparison of genomes from different species as well as the evidence for an ancestral organisation is investigated at several levels: genomic and post genomic, proteome, interactome, transcriptome and at long term, at the anatomical and physiological levels. I will focus my talk on one of the aspect of the project: the analysis of the ancestral genome reconstruction of the Major Histocompatibility Complex region. Up to now, the majority of the studies aiming to reconstruct ancestral genomic organisation consisted in the deciphering of conserved synteny between two species. Most of the comparative maps corresponded therefore to the comparison of genomic organisation between two species. However, these last years, parsimonious reconstructions as well as the corresponding algorithms have been described (for example, "MGR-multiple genome rearrangement published by Bourque and Pevzner, 2002). Therefore, we are confident that based on such approaches, the use of complete mammalian genome sequences (Human, Mouse, Rat, Dog, and very soon cow) (Bourque et al. 2004) will make possible the reconstruction of the ancestral mammalian genome. Blanchette et al (2004) were able to reconstruct a region of 1.1 Megabase of the mammalian cenoancestor by comparing the actual genomic CFTR sequences from 19 mammals. The ancestral state can be then compared with the other available vertebrate genomes such as chicken (Bourque et al 2005) and the teleosts such as Tetraodon nigrovidis (Jaillon et al 2004). The vertebrate ancestor state can be in turn deduced and compared to other phyla's ancestor organisations in order to decipher the Urbilateria genome organization. However, as the data are not yet available, we are performing the analysis in a different way. Indeed, we have reported (Abi Rached et al 2002, Vienne et al 2003, Danchin et al. 2003, Danchin and Pontarotti 2004 a, Danchin and Pontarotti 2004 b) that an ancestral genomic organization was at the origin of two sets of conserved paralogous regions in vertebrates: the MHC and its three paralogous regions, and the 8-10-4-5 paralogy group. We have first shown that these two sets of paralogous regions originated from an en bloc duplication that occurred at the beginning of the vertebrate history (Abi Rached et al 2002, Vienne et al 2003). In both cases, we have compared the genomic distribution of a set of conserved orthologues from informative species, cloned and mapped at the laboratory, and those of bilaterian species for which the genomic information were available, and deduced the minimal gene content for the ancestral region from which they originated. The choice of informative species, the design of the experiments and the algorithm for the genome reconstruction will be presented at the meeting. Abi-Rached L, Gilles A, Shiina T, Pontarotti P, Inoko H Nat Genet. 2002;31(1):100-5 Bourque G, Pevzner PA.. Genome Res. 2002 ;12(1):26-36. Bourque G, Pevzner PA, Tesler G.. Genome Res. 2004;14(4):507-16. Bourque G, Zdobnov EM, Bork P, Pevzner PA, Tesler G. Genome Res. 2005 Jan;15(1):98-
110.
Blanchette M, Green ED, Miller W, Haussler D..Genome Res. 2004 ;14(12):2412-23.
Danchin EG, Abi-Rached L, Gilles A, Pontarotti P.Immunogenetics. 2003 ;55(3):141-8.
Danchin EG, Pontarotti P.. Journal of Molecular Evolution. 2004 59 (5): 587-597
Danchin EG, Pontarotti P Trends Genet. 2004 ;20(12):587-91
Jaillon O, et. Nature. 2004;431(7011):946-57
Vienne A, Rasmussen J, Abi-Rached L, Pontarotti P, Gilles A. Mol Biol Evol. 2003;20(8):
1290-8.

17:45-18:45 DISCUSSION (led by Giorgio Matassi (University of Paris VI, France))

18:45 DINNER

20:00 Informal discussion on an edited volume based upon this meeting, on ancestral sequence reconstruction

Thursday, March 31, 2005
8:45-9:30 Slim Sassi (University of Florida, USA)
The past as a key to unlock the present: The resurrection of ancestral proteins to elucidate the function of Seminal Ribonuclease
Seminal Ribonuclease is an unusual member of the RNase A superfamily in that it is endowed with several unexpected in vitro biological activities (e.g., immune suppression, tumor cytotoxicity). These activities are absent from its closest homologs, pancreatic RNase A and brain ribonuclease. Seminal ribonuclease is found in a subgroup of artiodactyls, the bovids, and is only present in the plasma of seminal fluid. This protein and its gene presented an interesting biological problem rooted in evolution. What function of seminal ribonuclease has been selected during its evolution? What is its mechanism of action? Why is its gene a pseudogene in most lineages except for the bovids? What is its evolutionary history? We have used a combination of evolutionary computational tools and wet lab biochemistry to answer these questions. The phylogeny of the relevant group of artiodactyls was reconstructed based on the gene and protein sequence of seminal ribonucleases using a combination of maximum likelihood, parsimony and paleontology methods. The sequences of the relevant ancestors were reconstructed using a maximum likelihood approach based on a nucleotide, codon and amino acid models. The resulting ancestral protein sequences were resurrected using site directed mutagenesis, expression in bacteria, and purification of the recombinant proteins. We hypothesize that the selected biological function of seminal ribonuclease is immune suppression. Suppression of the female immune response against male sperm is an important factor contributing to reproductive success. Therefore, this function was assayed in the different ancestral proteins with a well-known lymphocyte proliferation test. Immune suppression results correlate with the ribonuclease enzymatic activity and dimerization states of the different ancestors. We observe high immune suppression activity in the early ancestor after gene duplication, followed by a decrease in this activity in intermediate ancestors. Interestingly, immune suppression activity is regained in the contemporary enzyme. These results are consistent with a decrease of function after gene duplication, followed by reacquisition of a new function grounded in a slightly different mechanism. The new function, immune suppression, has apparently been fixed in bovids due to a selective reproductive advantage. These findings support our hypothesis concerning the immunosuppressive function of seminal ribonuclease, and demonstrate the utility of the paleobiochemical approach in elucidating mechanisms by which evolutionary forces drive the invention of new biological functions in proteins.

9:30-10:00 David Ardell (Uppsala University, Sweden)
Evolution in the sequence-to-structure mapping of proteins
Structural alignments are used both as gold standards for evaluting sequence alignment methods and sometimes, in evolutionary studies, as substitutes for sequence alignments when the latter are too dificult to make because of great sequence divergence. The use of structural alignments for evolutionary reconstructions and inferences begs the question of whether analogous parts of homologous proteins are always occupied by homologous residues. We describe the execution of one approach we used to screen Pfam release 15 for events of "sequences sliding through structures." We found very few interesting candidates of such events in the data, although we believe this may reflect limitations of the data and approach. We describe briefly our attempt to follow-up an interesting candidate event and our plans for future work in this area.
10:00-10:30 BREAK

10:30-11:15 Belinda Chang (University of Toronto, Canada)
Reconstructing ancestral visual pigments

11:15-12:00 Joe Thornton (University of Oregon, USA)
Reconstructing the evolution of steroid hormone/receptor relationships
12:00-13:15 LUNCH

13:15-14:00 Denis Shields (Royal College of Surgeons, Ireland)
Ancestral reconstruction and the evolution of specificity and diversity
Broad survey analysis of protein evolution suggests that after paralogue duplication, there is an excess of change at conserved sites, which may reflect the evolution of specificity. Ancestral sequence reconstructions help in the prediction of residues that confer specificity. Application to single protein families enables the prediction of specificity conferring residues. Subsequent experimental evaluations include the construction of protein chimeras, and the design of synthetic peptides delivering protein-specific motifs involved in protein interactions. Use of information from ancestral sequences is likely to increase the power to detect amino acid residues that have evolved functional specificity. There is no good training set of known residues that confer specificity, and this presents a challenge for the evaluation of the optimal algorithms for predicting such sites.

14:00-14:45 Eric Gaucher (The Foundation for Applied Molecular Evolution, USA)
Reconstructing and resurrecting ancient sequences: Applications from evolutionary and synthetic biology to astrobiology
Recent developments in computational algorithms and experimental techniques have laid the foundation for the use of reconstructed and resurrected sequences as a valid and broadly used 'tool' in biology and chemistry. However, validity and applicability require our discipline to generate more sophisticated models, and move beyond anecdotal stories describing sequence evolution. This talk will summarize previous work with resurrected proteins and their implications for Astrobiology. Further, I will discuss current research with resurrected sequences as they relate to 'experimental evolution' and 'directed evolution'. The goal, as it relates to current applications, is to have ancient sequence space provide a significant component to the burgeoning field of synthetic biology. These developments will influence a variety of disciplines ranging from developmental and evolutionary biology, to pharmaceuticals and biomedicine, and thus hopefully satisfy the visions originally proposed by Linus Pauling and Emile Zuckerkandl.
14:45-15:15 BREAK

15:15-16:00 Richard Goldstein (National Institute of Medical Research, UK)
Ancestral reconstruction and homologue identification
The reconstruction of ancestral sequences can provide many insights into how proteins have changed through evolution. It can also yield practical benefits. We describe a method for recreating ancestral sequences in the form of hidden Markov models, and then demonstrate how these models can be used to identify distant homologies with improved accuracy.

16:00-16:45 Gina Cannarozzi (ETH-Zürich, Switzerland)
Empirical codon mutation matrices and their uses in molecular evolution, including ancestral sequence reconstruction
Codon mutation matrices constructed from observed sequence alignments provide empirical codon transition probabilities which eliminate the need for using parameterized models or heuristics. Their applications in molecular evolution, particularly in molecular dating and ancestral reconstruction, will be discussed.

16:45-17:15 Victor Albert (University of Oslo, Norway)
Convergent adaptive evolution of cytochrome c oxidase between seed plant lineages sharing a 300 million-year-old common ancestor: protein structural constraints vs. rare labilities, and their engineering implications for enhancing cellular energetics
The highly conserved respiratory machinery of eukaryotic cells might seem an unlikely target for selection supporting novel morphologies. We demonstrate that a dramatic molecular evolutionary rate increase in subunit I of cytochrome c oxidase (COX) from an active-trapping lineage of carnivorous plants is caused by positive Darwinian selection. Bladderworts (Utricularia) trap plankton when water-immersed, negatively pressured suction bladders are triggered. The resetting of traps involves active ion transport, requiring considerable energy expenditure. As judged from the quaternary structure of bovine COX, the most profound adaptive substitutions are two contiguous cysteines absent in {approx} 99.9% of databased COX I sequences from Eukaryota, Archaea, and Bacteria. This motif lies directly at the docking point of COX I helix 3 and cytochrome c, and modeling of bovine COX I suggests the possibility of an unprecedented helix-terminating disulfide bridge that could alter COX/cytochrome c dissociation kinetics. The helix3-4 loop makes crucial contacts with the active site of COX, and we postulate that the C-C motif could act as a redox sensor for conformational change that decouples (or partly decouples) electron transfer from proton pumping. Such decoupling would permit bladderworts to optimize power output (which equals energy times rate) during times of need, with only 20% reduction in overall energy efficiency. A desert-dwelling gymnosperm (Welwitschia) that intakes water only during seasonal, torrential rains is the only other organism (of which we are aware) with the C-C motif, and we consider the possibility of engineering arid-land crops for fine tuning of power (and thus growth) when water is most plentiful.
17:15-18:15 DISCUSSION (led by Giogio Matassi (University of Paris VI, France))
18:15 DINNER
20:00 GOODBYE PARTY

List of participants

1 Victor Albert, University of Oslo Norway
2 David Ardell, Uppsala University Sweden
3 Matthew Betts, University of Bergen Norway
4 Jonathan Bollback, University of Copenhagen Denmark
5 Øivind Braaten, University of Oslo Norway
6 Nick Braun, University of Tromsø Norway
7 Gina Cannarozzi, ETH-Zürich Switzerland
8 Belinda Chang, University of Toronto Canada
9 Anna Chernova, National Institute of Medical Research UK
10 Julien Dutheil, University of Montpellier II France
11 Richard Edwards, Royal College of Surgeons Ireland
12 Esben Friis, Novozymes Denmark
13 Toni Gabaldon, University of Nijmegen Netherlands
14 Eric Gaucher, The Foundation for Applied Molecular Evolution USA
15 Richard Goldstein, National Institute of Medical Research UK
16 Rodrigo Gouveia-Oliveira, Danish Technical University Denmark
17 Kristoffer Illergård, Uppsala University Sweden
18 Janos Kodra, Novo Nordisk Denmark
19 David Liberles, University of Bergen Norway
20 Dennis Madsen, Novo Nordisk Denmark
21 Giorgio Matassi, University of Paris VI France
22 Morten Mattindsdal, University of Oslo Norway
23 Mary O'Connell, Dublin City University Ireland
24 Anders Gorm Pedersen, Danish Technical University Denmark
25 David Pollock, Lousiana State University USA
26 Pierre Pontarotti, University of Provence France
27 Tal Pupko, Tel Aviv University Israel
28 Christian Roth, Stockholm University Sweden and University of Bergen Norway
29 Gisle Sælensminde, University of Bergen Norwa
30 Abhiman Saraswathi, Karolinska Institute Sweden
31 Slim Sassi, University of Florida USA
32 Adrian Schneider, ETH-Zürich Switzerland
33 Denis Shields, Royal College of Surgeons Ireland
34 Marie Skovgaard, Novo Nordisk Denmark and University of Bergen Norway
35 Eva Sykorova, Czech Academy of Sciences Czech Republic
36 Josef Thingnes, University of Oslo Norway
37 Gård Thomassen, University of Oslo Norway
38 Joe Thornton, University of Oregon USA