Principles of Microbial Diversity. James W. Brown
as a dendrogram.
Example analysis
In about 1990, an organism designated ES-2 was isolated from a deep-sea hydrothermal vent sample. ES-2 grows heterotrophically at 65°C. A lipid analysis of the isolate was unusual, showing that it contained a number of apparently branched lipids as well as fatty acids. Electron microscopy and standard microbiological tests were not helpful in identifying the organism.
Phylogenetic analysis of the organism was performed essentially as above. DNA was isolated from cells, and the SSU rRNA was amplified by PCR using primers near both the 5′ and 3′ ends of the gene. The amplified DNA was cloned and sequenced (Fig. 4.7), and the secondary structure of the encoded RNA was determined for use in the alignment process (Fig. 4.8).
Figure 4.7 Sequence of the ES-2 SSU rRNA, in GenBank format. doi:10.1128/9781555818517.ch4.f4.7
With the sequence in hand, the next step in the process was to align the sequence with the SSU rRNA database. This was originally done by hand in a local database, but these days it can be done automatically at the Ribosomal Database Project (RDP), using an algorithm called “infeRNAl” (from INFER RNA ALignment). After this, the next step was to calculate phylogenetic trees; the RDP uses a weighted neighbor-joining method. First, a “representative” tree was generated using representative sequences from each major phylogenetic group of bacteria to see which group ES-2 seems to belong to (Fig. 4.9).
Figure 4.8 Secondary structure of the ES-2 SSU rRNA sequence. This is a hand-drawn image, because that is the way these structures are usually sorted out. doi:10.1128/9781555818517.ch4.f4.8
Figure 4.9 Phylum-scale tree including ES-2, generated using the RDP II website. doi:10.1128/9781555818517.ch4.f4.9
ES-2 is clearly most closely related to the representative of the phylum Firmicutes (represented by Bacillus mycoides). So, the next step was to generate another tree by using representative sequences from the Firmicutes (Fig. 4.10). At the time this analysis was performed, only a few such sequences were available. Now there are thousands of Firmicute sequences in the database.
This tree shows that ES-2 is a member of the Clostridium/Eubacterium group of the Firmicutes. So a final tree with representatives of this group, and some especially close relatives identified using BLAST was generated (Fig. 4.11).
Figure 4.10 Tree of representative Firmicutes, including ES-2, generated using the RDP II website. doi:10.1128/9781555818517.ch4.f4.10
Figure 4.11 Fine-scale phylogenetic tree of ES-2, generated using the RDP II website. doi:10.1128/9781555818517.ch4.f4.11
Interpretation
ES-2 is a member of the Clostridium/Eubacterium group of the Firmicutes, and is particularly close (probably deserves to be in the same genus or even species) to Caloranaerobacter azorensis, a more recent deep-sea vent isolate. At the time of this analysis (but no longer), organisms from this group that produce spores were considered to be in the genus Clostridium, while those that did not produce spores were classified as Eubacterium. Because of the environment from which it was isolated, the new species was named Eubacterium thermomarinus.
What good was this information?
This information was very useful; we were interested in identifying organisms that were very distinct from those already in hand for analysis of RNase P RNA structure. Given that this was basically a member of the genus Clostridium, from which we already had several sequences, this organism became a low priority and was not pursued further. Knowing the phylotype of ES-2 prevented us from spending a fair amount of effort for what would have been trivial gain.
Questions for thought
1 1. Which organism would you choose for an outgroup for an rRNA tree of mammals? Does it matter which nonmammal you choose? Why? What might you choose as an outgroup for an rRNA tree of Bacteria? What about for a “universal tree” containing sequences of all kinds of organisms?
2 2. What would a tree of some animals look like if constructed from globin genes where some of the sequences were alpha globins and others were beta globins? What if some of these were adult alpha or beta globins and others were juvenile or fetal globins?
3 3. What would a tree (no pun intended) of plants look like if some of the sequences (rRNAs) were accidentally taken from the chloroplast instead of the nucleus? What if all of the sequences were from the chloroplasts?
4 4. On the initial “phylum-scale” representative tree of Bacteria shown above, can you show where we might have hoped ES-2 would be?
5 5. Which properties can you predict for ES-2 based on its phylotype? Which properties can you not predict?
5
Tree Construction Complexities
The process of generating phylogenetic trees as described in chapter 4 is straightforward. This is a gross simplification. Phylogenetic analysis is an entire scientific area of study, and the material that has been presented is very highly simplified. In this chapter we touch (just touch) on some of the complexities.
However, before thinking about more refined substitution models, treeing algorithms, or alternative sequences, keep one thing in mind: the most important thing by far that is needed to get a good, robust tree is to start with a good alignment. The simple and fast neighbor-joining method, using the Jukes and Cantor substitution model, usually gives perfectly usable trees if given a good small-subunit ribosomal RNA alignment to work with. Combined with bootstrapping (see below), this method is probably used more than any other for the creation of published trees.
Substitution models
In chapter 4 we talked about the Jukes and Cantor method to estimate evolutionary distance from sequence similarity. This is a simple method, but there are several other more sophisticated methods. The Jukes and Cantor method and other methods for estimating evolutionary distance amount to an attempt to describe how sequences change. In other words, they are mathematical models of the process of evolution of these sequences, and they are therefore usually called “substitution models.” The choice of an appropriate substitution model is critical and often underappreciated.
Figure 5.1 Two-parameter substitution models distinguish between transitions and transversions and score them differently. Each parameter is represented by an arrow. The values of these parameters can be predetermined (typically