Audio recording of this lecture

Previous or Next lecture

Readings for this lecture:

• Reysenbach AL, Wickham GS & Pace NR 1994 Phylogenetic analysis of the hyperthermophilic pink filament community in Octopus Spring, Yellowstone National Park. Appl. Env. Microbiol. 60:2133-2199
• Huber R, et al. 1998 Thermocrinus ruber, gen. nov., sp. nov., a pink-filament-forming hyperthemophilic bacterium isolated from Yellowstone National Park. Appl. Env. Microbiol. 64:3576-3583
• Suau A, Bonnet R, Sutren M, Godon J-J, Gibson GR, Collins MD and Doré J. 1999 Direct analysis of genes encoding 16S rRNA from complex communities reveals many novel molecular species within the human gut. Appl. Env. Microbiol. 65:4799-4807.

Molecular Ecology: Basic Sequence-based Approaches

Most of what we know about the microbial world is taken from the properties of cultivated organisms. But most bacterial species are not readily cultivatable (as we will see in lab this week), so you inevitably end up examining unrepresentative organisms if you rely on cultivation - all you can study are the "weeds". Inferences from these species are unlikely to be realistic.

Molecular phylogenetic analysis solves this problem (or at least improves the result), because cultivation is not required. Molecular phylogenetic analysis can be performed on isolated (cultivated or not) material, or even of microbial populations. Molecular phylogenetic analysis of uncultivated organisms is similar to analysis of cultivated species, but starts with DNA extracted directly from a purified sample or the environment, rather than from a pure culture:

Some uses of molecular phylogenetic analysis in microbial ecology:

• Identification of unculturable organisms. e.g. the identification of uncultivatable pathogens in tissue samples.
• Identify the predominant microbial groups in a population
• Count sequences from various phylogenetic group to assess their abundance in the ecosystem
• Assess enrichments aimed at cultivating organisms previously identified by rRNA sequence.
• Use fluorescently-labeled rRNA sequences (oligonucleotides) as probes for the identification or enumeration of specific organisms or groups in environmental samples.

Detection & identification of unculturable organisms

We'll start with the simplest case, in which molecular phylogenetic analysis is applied to organisms that can be obtained in at least a fairly pure (or at least visually distinguishable) form but cannot be cultivated.

Identification of the Octopus Spring Pink Filaments organism

Reysenbach, A.L., Wickham, G.S. and Pace, N.R. 1994 Phylogenetic analysis of the hyperthermophilic pink filament community in Octopus Spring, Yellowstone National Park. Appl. Env. Microbiol. 60:2133-2199.

Question being asked : What is the phylogenetic identity (phylotype) of the pink filamentous organism?

Described in the 60's by Thomas Brock (the namesake of the General Microbiology textbook), there is an abundance of pink filaments in the direct outflow from Octopus Spring. Similar pink filament mats are common, but have been best studied in Octopus Spring because of their abundance and the relative ease of access to the spring. Early attempts to label cells by feeding them radioactive organic compounds failed, as did early attempts to extract nucleic acids, causing some to suggest that they weren't alive, but just the fried remains of mesophilic groundwater organisms that had been regurgitated by Octopus spring. Thermus aquaticus (which is often also pink) was isolated as a by-product of attempts to cultivate these filaments, but it turns out Thermus is a minor constituent of this microbial community.

Ultimately, the pink filaments were characterized without cultivation by molecular phylogenetic analysis, and this information was used in the successful enrichment, isolation, and pure cultivation of the organism.

Pink filament biomass was collected following growth for several weeks on cotton filter pads placed in situ, to make sure the organisms actually grew in that environment rather than being toasted groundwater organisms. These "furnace filters" quickly became colonized by the pink filamentous growth.

DNA was isolated from a washed pink filament sample, and ssu-rDNA was amplified by PCR using universal ssu-rRNA primers. Interestingly, only universal or bacterial-specific primers yielded PCR products - archaeal-specific primers did not amplify anything. The PCR products were cloned and analyzed phylogenetically. Cloning is required when analyzing PCR products from natural samples because they will be a mixture of sequences - cloning is used to separate this population into individuals.

The sequences were sorted using "T-tracts". The purpose of this is to only sequence one clone from each set of highly-similar seuences; there's little reason to sequence a slew of things you know in advance are identical or nearly so. These are standard sequencing reactions, but instead of performing the reactions to identify all 4 bases (A,G,C, and T), only the "T" reactions are performed. These are runs side-by-side on a gel, and the patterns are sorted to prevent full sequence analysis of identical or nearly-identical clones. Representative clones (there were only 3 types) were fully sequenced. T-tracts are not used any more, but were run in the "old days" when you had to run each of the 4 sequencing reactions separately and run them on separate lanes of a gel. In todays automated sequencing systems, all 4 reactions are run in the same tube and run on a single lane of a gel.

Most of the sequences (26 of the 35 clones obtained) were a novel sequence (EM17) related to Aquifex. There were two other sequences that showed up less often among the clones:

EM19 (2/35) is a less-close relative of Aquifex, and EM3 (7/35) is a relative of Thermotoga.

A fluorescently-labeled oligonucleotide probe complementary to a species-specific (highly variable) region of the EM17 sequence was used to probe environmental samples for the EM17 organism; this is called fluorescent in situ hybridization, or FISH. The EM17 probe did fluorescently-label the pink filaments, and not other inhabitants of the sample, showing that the EM17 ssu-rRNA sequence really is that of the pink filament microbe. The EM19 and EM3 specific probes failed to hybridize to anything observable in the samples.

By the way, "EM" stands for Electric Monk, the haywire labor-saving device described in "Dirk Gently's Holistic Detective Agency" by Douglas Adams. When the Electric Monk is introduced in the story, he is stuck because he believes (believing is, after all, the function of an Electric Monk) that everything around him is a uniform shade of pink, making distinguishing any one thing from any other thing impossible. If you spend too much time with your head in the fumes of a hydrothermal vent, this sort of thing begins to seem profound.

Cultivation and isolation of the Pink Filaments organism

Huber, R., et al. 1998 Thermocrinus ruber, gen. nov., sp. nov., a pink-filament-forming hyperthemophilic bacterium isolated from Yellowstone National Park. Appl. Env. Microbiol. 64:3576-3583

The task: Cultivate EM17 on the basis of its phylotype, and isolate it as a pure culture from a single cell.

Since the pink filamentous organism is apparently closely related to Aquifex, and all of the cultivated members of this phylogenetic group (i.e. both of them, at the time) are physiologically similar, a series of enrichment media were based on the optimal growth conditions for Aquifex. In particular, all enrichments were provided with both hydrogen and oxygen, in low concentrations of 3% each. In addition, the chemical composition of Octopus Spring was mimicked. These enrichments were innoculated with pink filaments from Octopus Spring. The progress of the enrichments was assessed by microsccopic examination and by testing samples for hybridization with the EM17 probe (by FISH). Some enrichments yielded good growth of a rod-shaped pink bacterium that hybridized strongly to the EM17-specific probe.

In order to isolate this organism in pure culture, a 1064nm (about 1 micron) wavelength focused infrared laser was used as 'optical tweezers' to capture a single cell, which was used to start a pure culture. This cell was injected into a fresh culture bottle for growth. Of course, many cells were collected and cultured in this way, and the resulting growth, if any, was tested using the EM17 probe.

With a pure culture in hand, the ssu-rRNA sequence of this organism was determined and found to be nearly identical to the EM17 sequence.

...although not exactly the same. This is typical - most microbial populations are collections of closely-related organisms, specialized for different micro-environments. Enrichment cultures typically capture only the fastest-growing, robust types (i.e. the weeds), whether they are the predominant strain or not (usually not). The organism was named Thermocrinus ruber ("hot red hair").

Thermocrinus ruber grows as long rods in suspension, but when cultivated in media prepared to match the composition of Octopus Spring water in an artificial 'creek', grows as nice pink filaments that are just like those seen in the wild.

This paper is a monogram - a formal description and naming of a newly isolated organism, and at the end of the paper, a series of more-or-less standard tests are performed to describe the general properties of the organism, the sort of data you'd find in Bergy's Manual, for example.

Some other cases of uncultivated species characterized by molecular phylogenetic analysis

• The sulfur-oxiding symbionts of the giant tube worm, the vent clam, and the scaly snail. These organisms are found only at deep-sea hydrothermal vents, and have no digestive tracts - they are feed entirely by their sulfur-oxidizing proteobacterial endosymbionts.
  
  
• Magnetotactic Bacteria - can be isolated by taking advantage of their magnetotaxis, but cannot be grown in culture.
  
  
• Bacillary angiomatosis - a disease that turns out to be caused by a close relative of the causative agent of cat scratch disease, Rochalimaea quintana. This knowledge, obtained by molecular phylogenetic analysis, has allowed the cultivation of the organism, now known as Bartonella bacilliformis, using culture methods used for Rochalimaea.
  
  
• Cenarchaeum symbiosum is a psychrophilic (grows optimally at 10C) crenarchaeon symbiont of an unnamed marine sponge of the genus Axinella. They (the symbionts) are related in ssu-rRNA sequence to abundant crenarchaeal species found in marine and soil environments, although none have been cultivated.
  
  
• Epulopiscium fishelsonii, for a long time the largest known prokaryote, is a gut symbiont of some marine herbivorous fish (surgeonfish, doctorfish, & tangs), was originally thought to be a protist, perhaps a ciliate, based on size (big enough to see clearly with the naked eye), but didn't seem to have any organelles. It turns out to be a Firmicute!
  
  
• Sap-sucking and wood-eating insects contain an unusual organ, called the bacteriome, consisting of about 75 cells called bacteriocytes. These cells contain endosymbiotic oval bacteria, that provide the insect with the vitamins and essential amino aids that are absent in their nutritionally poor diets. This is an obligate relationship - the insect cannot survive without the bacteria (antibiotic treatment is lethal), and the bacteria have never been cultivated outside of the insect. However, they have been characterized using ssu-rRNA analysis - most are gamma purple Bacteria, but some are beta-purples and some are Cytophagales. One such insect endosymbiont, Buchnera aphidicola, has had it's entire genome sequenced even though it can't be cultivated except in aphids.

And many, many, more....

PCR-based rDNA analysis of microbial populations

In a typical molecular phylogenetic analysis, you start wih genomic DNA isolated from a pure culture an organism. The ssu-rRNA sequence obtained by PCR amplification is used to determined the place of that organism in the 'big tree'.

It is also possible, however, to start a molecular phylogenetic analysis with DNA extracted directly from an environmental sample instead of a pure culture. The PCR amplification products in this case are a collection of ssu-rRNA sequences representing the population of rRNA genes in the DNA, in turn representing the population of organisms in the original sample. The ssu-rRNA sequences are separated by cloning, and then ssu-rRNA sequences from each clone are determined, in order to survey the microbial inhabitants of an environment.

This paper is based on this molecular phylogenetic approach to surveying microbial populations without resorting to cultivation. This approach is far superior to the older cultivation-based approaches, and has been used a lot. However, it has become clear that several aspects of PCR/cloning/sequencing of microbial populations limits their interpretation. In fact, bias exists in every step of the process, and so estimation of the relative abundance of organisms from molecular phylogenetic "surveys" is not quantitative. Some cells will not be opened using standard methods (or will not be opened as efficiently), some will be selectively lost to some extent in the DNA purification (perhaps tightly bound by protein), some are easier than others to amplify rDNAs from (depending on primer sequences, PCR conditions, etc), different organisms have different rDNA copy numbers, some will clone more readily than others, &c. Nevertheless, even a qualitative molecular phylogenetic survey of the microbial population is very useful, and far more informative than examination of whatever happens to grow on plates.

One aspect of molecular phylogenetic surveys that is often overlooked is that human samples, either normal or pathogenic, are microbial populations, and can be examined using this method. In this paper, the authors examine the microbial population of a single human fecal sample using the more-or-less standard PCR/rDNA based method.

Suau A, Bonnet R, Sutren M, Godon J-J, Gibson GR, Collins MD and Doré J. 1999 Direct analysis of genes encoding 16S rRNA from complex communities reveals many novel molecular species within the human gut. Appl. Env. Microbiol. 65:4799-4807.

Purpose: To perform a census of the normal gut flora of humans.

In this paper, the authors have done a PCR-based rDNA analysis of a microbial population of direct human relevance, i.e. feces. Don't laugh - the gut flora are without a doubt the single most important human:microbe interface. The gut flora are critical for good human nutrition - digestion is largely a controlled fermentation and absorbtion of the products. A lot of traditional cultivation work has been done on the gut environment, given it's importance to humans, and you might imagine that everything about the gut flora was known. Of course this is incorrect, as this paper demonstrates, but on the other hand it is true that human environments are probably the best characterized environments by cultivation methods, since that's what we've have the incentive to pursue hardest. In fact, in this study, they could cultivate 20-30% of the organisms in the sample - far, far better than most environmental samples where a single plating method will give far less than 1%.

In this paper, they isolated DNA from a 125g fecal sample from an unnamed 40 year old healthy man (probably one of the authors, who wisely chose to remain anonymous) and used along with a bacterial-specific forward primer (against nucleotides 8-20, right at the 5' end of the RNA) and the same 1492 reverse primer you used in lab. Note that they used a low annealing temp - 48C (we used 55C) and as few cycles as possible (10) in attempt to be as non-selective as possible in the amplification. Although they amplified nearly the entire 16S rRNA gene (ca. 1500bp), they only sequenced the first 500 or so bp of each clone for this analysis - this is pretty reasonable, although for further work they should (and apparently did) get the entire sequence.

A note about chimeras Before phylogenetic analysis, a final screening was performed to eliminate any "chimeric" sequences. These types of sequences show up in most rDNA PCR experiments from natural mixed populations. They seem to arise when a molecule of Taq polymerase stalls partway through a PCR round. If the resulting truncated DNA strand ends in a highly-conserved part of the rDNA, it can anneal to another DNA molecule, whether or not it is from the same organism, and serve in the next PCR round as a primer for the synthesis of a DNA molecule that is from one organism at one end and another organism at the other end. Such chimeras can be identified using three standard methods: Making phylogenetic trees based independently on the 5' half and 3' half of each sequence. if the two trees disagree significantly, it's probably a chimeric sequence. drawing the secondary structure of the RNA - if the basepairs involving the 5' and 3' sequences don't work, it's probably a chimera. Using the CHECK_CHIMERA function and the Ribosomal Database Project. This program compares the similarity of a sequence along it's length to other sequences in the database - a "break" in this similarity, where the sequence begins to look less and less like one sequence and more like another, indicates that the sequence is probably a chimera. Anything that seems to be a chimeric sequence is, of course, discarded from the analysis.

They obtained 295 clones, 11 of which seeemed to be chimeras and so were discarded, leaving 284 clones for analysis. These fell into 82 "operational taxonomic units" (OTUs), or "molecular species", collections of sequences that are >=98% identical. Their statistical analysis of the distribution of these sequences estimates that these sequences represent about 85% of the organisms in the sample. Almost all of their sequences (95%) fell into 3 major phylogenetic groups already known to be major players in the human gut flora: Bacteroides & relatives, Clostridium coccoides & relatives, and Clostridium leptum & relatives.

The Bacteroides
88 of their clones (31%) fell into 20 OTUs scattered about within this group.

The Clostridium coccoides group
125 (44%) of their clones fell into 31 OTUs in this group.

The Clostridium leptum group
57 (20%) of their clones fell into 20 OTUs in this other Clostridium group.

Other groups
Only 13 (5%) of their sequences fell outside of these major groups:

• Firmicutes:
  • 2 Streptococcus clones (1 OTU)
  • 4 clones (3 OTUs) very distantly related to Mycoplasma (which in turn is part of the Clostridium innocuum group)
  • 2 clones (1 OTU) in the Sporomusa group (a deep branch with a typical Gram-negative envelope)
  • 5 clones (5 OTUs) fell out into various other Clostridium/Eubacterium groups
• Verrucomicrobium group:
  • 1 clone

Notice that of the 284 sequences (representing organisms) they examined, 69% (195) were members of the Firmicutes, a.k.a. the low G+C Gram-positive Bacteria, and 31% (88) were Bacteroids, leaving only a single leftover, a Verrucomicrobium relative.

Questions for thought

• More and more diseases that were previously thought to be non-infectious are being discovered to be actually caused by bacterial or viral infection. What kind of "non-infectious" diseases can you think of that might eventually be found to have an infectious cause?
  
  
• Why do you suppose T. ruber grows as rods when cultuvated in staandard culture conditions, but as filaments in the artificial (or natural) spring?
  
  
• The products of the PCR from DNA isolated from the pink filaments were cloned before analysis. What would have happened if the PCR product DNA was sequenced directtly, as we did in lab?
  
  
• Why do you suppose microbiologists had such a hard time showing (or believing!) that the T. ruber mats were actually living?
  
  
• The authors of the first paper seem to have been surprized by their lack of detection of Archaea in the pink filaments sample. Why do you suppose this is? What could they have done to investigate this further?
  
  
• Can you think of any bacterial:bacterial symbioses other than that of Chlorobium? What about bacterial:eukaryotic symbioses?
  
  
• So, where was the E. coli or Lactobacillus in the gut flora survey?
  
  
• Where do you think the greatest bias might occur in a molecular phylogenetic analysis of a microbial population? How would you find out?
  
  
• If you did a "survey" of a population, and thought the results might be off somehow, perhaps biased by the PCR primers, how would you test it?
  
  
• How do you think "OTUs" compare to "species"?

Previous or Next lecture