What is Ancestral Sequence Reconstruction?
Ancestral Sequence Reconstruction is a method used to test molecular evolutionary hypotheses by deriving ancestral gene sequences from extant species. The gene sequences (usually in the form of amino acids) provide the template to synthesize biomolecules present in extinct species. This enables scientists to compare and contrast the extinct species from their diverged descendants and trace the origins and evolution of different proteins.
1963: L. Pauling and E. Zuckerkandl introduced the idea of reconstructing genes from extinct species from the inferred ancestral sequences 
1971: W. Fitch proposed an algorithm to resurrect genes from extinct ancestors using the parsimony principle (Thornton)
1980s: advancement in computer programs increase confidence in generated sequences, which are used by biologists to study morphological and behavioral evolution
1990s: advancement in DNA synthesis methods leads to more practical discoveries about genes from ancestors that existed around 10 million years ago
2000s: greater knowledge of structure-function relationships in proteins, advancement in technology, and improved techniques allow reconstruction of genes from more than 1 billion years ago
Pipeline for Generating Ancestral Genes
- Sequences from extant species of the desired common ancestral gene and outgroup genes are aligned.
- The ancestral gene is inferred based on evolutionary models (typically maximum likelihood).
- Ancestral genes are synthesized via overlapping oligonucleotides and PCR assembly or through site-directed mutagenesis to extant gene.
- Ancestral genes are cloned and expressed in vector.
- Ancestral protein tested for functionality and specific characteristics.
Methods of Inferring Ancient Sequences
- Consensus Sequence - the most frequently occurring residue of a extant organisms is assumed to be the ancestral state.
- Limitations: accuracy of generated sequence relies heavily on the chosen sequences phylogenetic relationships among extant species are ignored.
- Method not commonly used in modern-day reconstruction studies.
- Maximum Parsimony - minimization of the total number of changes required to account for the terminal sequences.
- Useful approach when using sequences from closely related extant species.
- Limitations: produce ambiguous results for unconserved sequences and does not take into account the length of branches on phylogenetic tree which provides information on substitution frequencies.
- Maximum Likelihood (ML) - ancestral states are evaluated at each internal node in the tree based on the likelihood of each mutation. This process uses a statistical framework of molecular evolution which takes into account differences in certain types of mutations. The generated ancestral sequence gives the probability that each residue is correctly predicted.
- More accurate than both consensus sequence and maximum parsimony method.
- Algorithms require the sequences from extant species, branched tree that shows phylogenetic relationships, and evolutionary model.
- Posterior probability takes into account the branch lengths on trees, which are scaled to show expected number of substitutions per site.
- Bayesian Markov Chain Monte Carlo (BMCMC) - technique that incorporates the phylogenetic uncertainty of the tree
- Comparisons to ML sequences do not show a statistically significant improvement to accuracy or greater degree of confidence in produced ASRs.
Modern advances in oligonucleotide synthesis allow for the cheap construction of synthetic ancestral genes. Using overlapping PCR assembly most synthetic genes can be constructed for several hundred dollars. If the ancestral gene only differs by a few amino acids from a certain extant sequence, then many biologists resort to using site-directed mutagenesis.
Expression of archaic proteins in extant organisms such as E. coli may require codon optimization to allow for proper expression. Synthesized genes are typically cloned into a vector and used to transform a host organism.
Testing Ancestral Variants
After synthesis is complete, ancestral genes can be tested for their function. For example, the ancestral genes can be cloned into bacterial expression vectors and transformed into E.coli. The proteins can be purified and tested for their activity, whether that be fluorescence spectra, binding affinity, thermoactivity, reporter-gene expression, ligand or substrate binding, etc.
Examples of Ancestral Sequence Reconstructions
As the technology for ancestral sequence reconstruction has advanced the technique has popularized. It offers a unique way to peer into the past and revisit the biomolecules of our forefathers. The sensitivity of genetic information to degrade limits our understanding to extant organisms or circumstantially preserved ones, such as specimens preserved by ice. The ability to resurrect sequences from the dead and test them relieves us from these temporal limitations. Also, predicted mutations and their effects on proteins can be tested out and backed up with data by ASRs, which cannot be done by just structural observations.
Evolution of Coral Pigments
One example of ancestral sequence reconstruction was done by the Matz group (currently residing at the University of Texas at Austin). Fluorescent proteins from related coral species had wavelengths corresponding to Cyan, Green, and Red. The details of the evolution of fluorescent color in the GFP superfamily was not fully understand. That is, what fluorescent spectra did the common ancestors of the modern corals have?
Different models for reconstruction based on amino acids, codons and nucleotides resulted in reconstructed proteins differing in 4-8 amino acids out of 217. Gene synthesis utilized codons designed "to be degenerate in order to incorporate alternative predictions."  The reconstructed sequences included red, pre-red, Red/Green and ALL. The ancestral sequences revealed an interesting evolutionary history. The ancestor to all the coral fluorescent proteins had a single green peak at 505nm. Gradually, a shift towards greater expression of lower wavelengths developed, culminating in the pre-red protein. One branch of pre-red went on to lose the ancestral green peak and develop a strong peak around 580nm. Modern red features significantly less emission at outside of the primary peak, thereby increasing its specificity compared to recent ancestors such as pre-red and Red/Green. According to Ugalde et al., incremental gains in protein function "[have] been anticipated since Darwin, but has only recently been demonstrated in computer simulation experiments"
Inferring the Paleoenvironment of ancient Earth
Ancestral sequence reconstruction has been used to infer the environmental conditions on the early Earth. The study sought to explore the evolutionary history of an elongation factor thermo unstable (EFtu). This elongation factor functions optimally at the temperature which the organism lives, for example thermophilic organisms have an EFtu optimized for high temperatures. By resurrecting the EFtu protein in the common ancestors of bacteria, the temperature profile might elucidate what the environmental conditions were. Interestingly the EFtu common ancestor to all mesophilic bacterium (~1BYA) has an optimal temperature of a thermophile. This suggests that the hypothesis of a hot early Earth is true.
Precambrian β lactamases
The structure of the last common ancestor to extant β lactamase has been reconstructed and tested using Ancestral Sequence Reconstruction. Sequences for β lactamase from 75 bacterial strains, whose last CA lived 2-3 Gya, were compared. To exclude the effects of recent antibiotic driven evolution, only non-clinical varieties were chosen. Using Bayesian statistics, Risso et al. calculated the most probabilistic ancestral amino acid for each point in the sequence. Resistance to various β lactam based antibiotics was conferred to modern microbes engineered to produce the Precambrian enzyme. The extremely high promiscuity of the Precambrian enzyme compared to extant sequences indicates that over the past few billion years, β lactamases have evolved greater substrate specificity. The Precambrian enzyme featured a denaturation temperature of ~90°C, which is 35°C above that of the highest extant sequence.
Perez-Jiminez et. al resurrected various Precambrian Thioredoxins. These belonged to the last common ancestors of the last bacterial common ancestor (LBCA), the last archaeal common ancestor (LACA) and the archaeal-eukaryotic common ancestor (AECA)". All of these sequences are believed to have been present over 3.5 Gya. The amino-acid sequence in each case was determined by maximum likelihood inference from 200 extant sequences. The genes to code for these proteins were synthesized and cloned into E. coli for expression and purification.
All three reconstructed proteins showed a Tm around 113°C, with a ΔTm between the highest extant Thioredoxin and the ancestors of around 25°C. The three paleo Thioredoxins also showed substantially greater activity at pH5 than representative extant enzymes from the each domain. The lower substrate specificity exhibited by the reconstructed enzymes indicates the abundance of sulfur rich compounds in the early oceans of Earth, and also hints at the generalist nature of archaic enzymes.
Ancestral Alcohol Dehydrogenase 
Present-day yeasts use two homologs of alcohol dehydrogenases (Adhs): Adh1 converts acetaldehyde to ethanol and Adh2 oxidizes ethanol to acetaldehyde. Thomson "et. al" reconstructed the Adh1A sequence from Adh1 and Adh2 sequences from S. cerevisiae and their relatives through the maximum likelihood approach. Studies on Adh1A show that its functions and kinetic properties more closely resembled Adh1 and conclusions are made that ancestral yeasts only make and do not attempt to accumulate ethanol. Two hypothesis was made about the development of the make-accumulate-consume strategy: selection due to humans domestication of yeast or selection due to competition for resources in fruits. Since the Adh1-Adh2 duplication was predicted to occur around 80 million years ago, it is reasonable to deduce that the emergence of fleshy fruits was the force that drove the evolution of alcohol dehydrogenases. Although the pathways to accumulate and consume ethanol consume ATP, the outcome of being able to outcompete other microorganisms for carbohydrates in fruits seems to outweigh the cost in energy.
Steroid Hormone Receptors 
Mineralocorticoid receptor (MR) is mostly activated by aldoesterone and sometimes cortisol, while the glucocorticoid recepot (GR) is specifically activated by cortisol. Brigham "et. al" computed the ancestral gene sequence AncCR from 59 steroid receptor sequences using ML, BMCMC, and maximum parsimony to study the ancestral corticoid receptor. The existence of MR and GR is caused by the gene duplication of AncCR more than 450 million years ago, and AncCR is found to be more functionally similar to MR because it is has high affinity for aldosterone. Two crucial mutations in AncCR sequence, S106P followed by L111Q, converts the receptor to have GR phenotype.
Interestingly, aldosterone molecules and AncCR did not co-exist because aldosterone molecules only emerged during the evolution of tetrapods. Therefore, AncCR and its descendant genes were structurally preadapted to a hormone that wasn't produced at the time.
HCH degrading genotype 
Hexachlorocyclohexane (HCH) is a pesticide and pollutant that have only be found to be degraded by two Sphingobium species, Sphingobium japonicum UT26 found in Japan and Sphingobium indicum B90A found in India. Both species have the lin genes that enable them to degrade HCH; particularly significant are the linA, linB and linC genes. ASR results produced an ancestral sequence that only have the genes lin D, E, R, F, G, and H which suggest that the lin gene components that are responsible for degrading HCH appeared through lateral gene transfer. The relatively small genome size of the last common ancestor compared to its descendants and the observation that most genes responsible for degradation of xenobiotics are located in plasmids support the hypothesis of a successful horizontal transfer event.
Ancestral sequence reconstruction holds various benefits. The most direct benefit is a better understanding of archaic organisms and the world they inhabited. Numerous ancient protein reconstructions have yielded enzymes that are much more thermally stable than extant derivatives, even those found in thermophiles, leading to stronger evidence for extremely warm global conditions prior to 1 Gya. Archaic proteins also allow for greater study of evolvability, as many reconstructed proteins feature high promiscuity. The patterns in the evolution of high promiscuity Precambrian β lactamase, for instance, could help further the fight against the development of antibiotic resistance. High promiscuity is also key to directed protein evolution. In addition, the reconstruction of archaic enzymes serves as a sort of paleo-bioprospecting, and is currently the best approach to expand thermal stability and pH tolerance in enzymes. The enzymes discovered through this form of prospecting and those developed through directed evolution of promiscuous enzymes could serve functional roles in human engineered systems.
- Pauling, L. and E. Zuckerkandl. 1963. Chemical paleogenetics: molecular restoration studies of extinct forms of life. Acta Chem. Scand 17:89. CrossRef doi=10.3891
- pmid = 20368266
- Perez-Jimenez, R., Inglés-Prieto, A., Zhao, Z.-M., Sanchez-Romero, I., Alegre-Cebollada, J., Kosuri, P., Garcia-Manyes, S., et al. (2011). Single-molecule paleoenzymology probes the chemistry of resurrected enzymes. Nature structural & molecular biology, 18(5), 592–6. doi:10.1038/nsmb.2020
- Ugalde, J. a, Chang, B. S. W., & Matz, M. V. (2004). Evolution of coral pigments recreated. Science (New York, N.Y.), 305(5689), 1433. doi:10.1126/science.1099597
- [http://www.nature.com/nature/journal/v451/n7179/full/nature06510.html Gaucher, E. a, Govindarajan, S., & Ganesh, O. K. (2008).
Palaeotemperature trend for Precambrian life inferred from resurrected proteins. Nature, 451(7179), 704–7. doi:10.1038/nature06510]
- Risso, V. a, Gavira, J. a, Mejia-Carmona, D. F., Gaucher, E. a, & Sanchez-Ruiz, J. M. (2013). Hyperstability and Substrate Promiscuity in Laboratory Resurrections of Precambrian β-Lactamases. Journal of the American Chemical Society. doi:10.1021/ja311630a
- pmid = 15864308
Research on the yeasts' alcohol dehydrogenase and linking the appearance of Adh2 to the emergence of fleshy fruits.
- pmid = 24030592
Research on bacterial species that are able to degrade HCH and using ASRs to determine that the genes with the degrading function was acquired through lateral gene transfer.
- Gaucher, E. a, Thomson, J. M., Burgan, M. F., & Benner, S. a. (2003). Inferring the palaeoenvironment of ancient bacteria on the basis of resurrected proteins. Nature, 425(6955), 285–8. doi:10.1038/nature01977
- Kaganman, I. (2011). Resurrected enzymes. Nature Methods, 8(6), 452–452. doi:10.1038/nmeth0611-452
- Thornton, J. W. (2004). Resurrecting ancient genes: experimental analysis of extinct molecules. Nature reviews. Genetics, 5(5), 366–75. doi:10.1038/nrg1324
Example ASR process and methods and gave examples of research using ASRs.
- pmid = 16601189
Research on original steroid hormone receptor and identified the LCA of MR and GR which happen to have a great affinity of aldosterone even though the molecule did not exist during the time that AncCR existed.