Genetics and Evolutionary Aspects

Hypothesis: SARS-associated Coronavirus is a separate genus from other Coronavirus

• It would seem that none of the publications on SARS thus far included an outgroup in the phylogenetic trees presented.

• Lack of an outgroup means that the tree cannot be rooted

• An unrooted tree does not indicate the direction of evolution, or the polarity of characters²

• Thus, goals of this analysis is to:
  a) discern the evolutionary relationship between SARS and coronavirus by rooting the tree with torovirus as an outgroup;
  b) the relative rate of mutation of SARS in relation to other coronavirus, using Tajima's relative-rates test;
  c) determine whether SARS is under positive selection pressure by comparing the rates of synonymous versus nonsynomymous mutations

• Results from this analysis must be interpreted in light of the host-virus relationship and interaction of susceptible host with the environment in order to provide a more coherent picture. Also, several assumptions and constraints have been made in drawing these conclusions.

Methods

The nucleotide sequence of Urbani-SARS coronavirus ORF1b RNA polymerase was obtained from the WHO Collaborative Networks³.  The nucleotide sequence was translated to amino acid sequence (Figure 1).  The translation was verified with BLASTX (translated BLAST search - nucleotide query to protein db).  Nucleotide sequences of RNA-directed RNA polymerase from Canine coronavirus UWSMN-1 (gi|22774013); Porcine transmissible gastroenteritis virus (gi|4927036); Turkey coronavirus (gi|4927034); Rat sialodacryoadenitis coronavirus (gi|4927032); Human coronavirus (strain OC43) (gi|4927030); Porcine hemagglutinating encephalomyelitis virus (gi|4927028); Feline infectious peritonitis virus (gi|4927026); Canine coronavirus (gi|4927024); Bovine coronavirus (gi|4927022); and Breda torovirus (gi|4204756) were obtained from NCBI nucleotide database. The sequences were aligned using the program ClustalW 1.81 (ftp://ftp.ebi.ac.uk/pub/software/dos/clustalw) with default parameters (gap opening penalty=10, gap extension penalty=0.20). The alignments were then formatted to a MEGA format. The alignments were visualized in MEGA version 2.1 (Kumar et al. 2002). A preliminary phylogenetic tree was constructed with the following method in place: Unweighted Pair Group Method with Arithmetic mean (UPGMA), Kimura 2-parameter, and bootstrap with 100 replications. Tajima relative-rate tests are then performed between SARS-associated coronavirus, Bredavirus, and coronavirus from all three antigenic groups.

Results and Discussions

Figure 2a: Radiation Tree

Figure 2b: Traditional Rectangular View

Figure 2c: Bootstrap Consensus Tree

Table 1: Pairwise Distance between all viruses examined, using the Kimura 2-Parameter method

Average Pairwise Distance between the Three Antigenic Groups and SARS-associated Coronavirus:

Viruses within the coronavirus family separate into three antigenic groups as expected. However, the genetic distance between SARS and other coronaviruses is similar to the distance between SARS and Bredavirus but different from the distance between coronaviruses from all three antigenic groups (see the above table). This is something not found in other publications on SARS thus far. This, in conjunction with the UPGMA tree obtained, seems to suggest that the SARS-associated coronavirus diverged very early in the evolution from the ancestors of coronavirus, soon after the divergence of Bredavirus (a torovirus subfamily) from its common ancestor. Preliminary results from Tajima relative-rates test between SARS-associated coronavirus, Bredavirus, and coronavirus from all three antigenic groups suggest that the SARS-associated coronavirus mutated at the same rate as other coronaviruses. This decreases the likelihood that the SARS-associated coronavirus is mutating faster than expected. These results suggest that perhaps the new virus is a separate genus distinct from other known coronavirus.

Next Step

• complete the Tajima's relative-rate test

• construct phylogenetic trees using other methods, like maximum parsimony, minimum evolution

• use more than one outgroup in the construction of the trees, such as including Arterivirus

• repeat the experiment with other genes in the SARS-associated coronavirus