In recent years, while sequencing technologies have improved dramatically, DNA extraction methods have seen little advance. It’s a required step for metagenomic sequencing of microbiomes, which allows scientists to gain a complete picture of the diversity of microbes, many of which cannot be successfully cultured in the lab. As a result of this lack of progress, DNA extraction remains a major bottleneck in the process of analyzing large numbers of microbiome samples, even when robots are used. This problem has been especially relevant to scientists working on the National Institutes of Health's Human Microbiome Project—a study of microbial diversity in humans whose results were reported in Nature and PLoS ONE this June—as well as the Earth Microbiome Project, a proposed multidisciplinary effort to analyze microbial communities across the globe.
For example, in a study published in PLoS ONE, Gilberto Flores, a microbiologist at the University of Colorado, Boulder, and colleagues tested a direct PCR approach provided by the Extract-N-Amp Plant PCR Kit from Sigma-Aldrich (St. Louis, MO, USA), performing 16S rRNA gene-based analyses of bacterial communities found in various parts of the human body (1). The researchers directly compared this method to the standard approach in which DNA is first extracted and purified from samples using a series of steps prior to PCR amplification. They found that the two approaches provided similar estimates of taxon richness, phylogenetic diversity, relative abundances of individual taxa, and differences between body sites and between individuals in bacterial community composition.
In the end, the direct PCR approach was at least six to eight times faster than the DNA extraction and purification approach, making it especially suitable for high-throughput studies aimed at cheaply and efficiently analyzing the structure and diversity of a large number of human microbial communities. Currently, Flores is using this approach to study how microbial communities in various body habitats change over time. However, this direct PCR approach may not be suitable for shotgun metagenomic analyses, which usually require larger amounts of highly purified DNA. He hopes that "people will use it for studies where they need to deal with large numbers of samples relatively quickly."
Beyond improving the efficiency of DNA extraction protocols, another challenge is to isolate DNA that accurately represents the diversity of microbes in the community sampled. Microbes differ in how susceptible they are to chemical lysis procedures. For instance, Gram-positive bacteria have cell walls with thicker layers of peptidoglycan compared with Gram-negative bacteria. So, DNA extraction procedures may be biased toward isolating DNA from certain types of microbes. But because researchers don't know ahead of time what microbes—and how many of each kind—are in the community under most circumstances, it's impossible to know whether or not the method being used accurately represents the microbial diversity in a sample.
Traditionally, microbiologists studying samples from humans have chosen DNA extraction methods without an obvious rationale, and the accurate representation of microbial diversity is often ignored as a criterion for evaluating DNA extraction methods. But when Larry Forney, a microbiologist at the University of Idaho, and collaborators studied the temporal dynamics of the composition of vaginal bacterial communities, as reported in Science Translational Medicine (2), they took special care to ensure their analysis accurately represented the diversity within the microbial sample.
"We started out optimizing the procedures that we used for characterizing the vaginal microbiome a very long time ago—17 years ago—and one of the things that's always sort of bothered me was that there are different criteria that people have used to assess protocols, and mostly what you hear about is increasing the yield of DNA," he says. "People will choose a method that they've been sold on for whatever reason—the vendor or papers they’ve read, because somebody else used the method, and so on. And the criterion that's missing, that I thought is probably more important than any of the others, is representativeness."
To address this problem, Forney and his team compared six commonly used DNA extraction procedures using a mock community that contained equal numbers of cells of 11 human-associated bacterial species. As reported in a paper published in PLoS ONE, the researchers analyzed 16S rRNA gene sequences from the mock community and found that all six DNA extraction methods had significant differences between the observed species abundances and the expected species abundances (3). In other words, none of the methods provided an accurate representation of the bacterial diversity of the mock community.
However, protocols that used bead beating—a physical disruption technique that improves the lysis of bacterial cells—and the enzyme mutanolysin were significantly better at representing the bacterial community structure than methods without both of these steps. In general, a cocktail of enzymes lysed cells of different species more effectively than a single enzyme alone or no enzyme at all.
The researchers also found no significant relationship between DNA yields and the representation of microbial diversity within or across methods, so kits that tout high DNA yields may not accurately reflect the relative abundances of various bacteria in the community.
"If you have two species of bacteria in a mixture, and you have 100% lysis of one and 0% lysis of the other, you might have a fair amount of DNA, but in no way does it resemble what was in the sample," says Forney. "One of the things I wanted to accomplish by publishing the article was to heighten awareness of the difference between methods so that people would think about the method they use and not simply use whatever was available on the shelf or whatever had the lowest price."
Given the observed differences in DNA yield and representativeness across methods, Forney's study highlights the importance of identifying methods and strategies that are suitable for making comparisons across samples. For this reason, the Earth Microbiome Project has selected a standard DNA extraction protocol, says project participant Janet Jansson, a microbial ecologist at Lawrence Berkeley National Laboratory. "I truly believe it's important to rely on a standardized approach as much as possible—except for difficult samples—for the purpose of comparison between studies."
As a Ph.D. student, Jansson had to develop protocols for DNA extraction from soil because there were no kits available. Since discovering that a commercially available kit for extracting DNA from soils provides similar yields of DNA compared to a method she developed in her own lab, she has used the proprietary kit for human fecal and biopsy samples. "Soil is the most difficult type of sample to work with," she says. "If you can get DNA out of soil, you can pretty much get it out of anything."
Jansson asserts that bead beating is the best type of mechanical disruption technique for increasing the DNA yield and the representativeness of bacterial communities. But it can also shear genomic DNA into small fragments, which introduces biases into subsequent PCR reactions. Moreover, a study published in the Journal of Microbiological Methods found that this type of mechanical disruption can decrease the quality of microbial DNA extracted from intestinal tissue (4).
In this study, bead beating increased the quantity of microbial DNA, but it also appeared to increase the presence of host DNA. Surprisingly, bead beating decreased the amount of DNA extracted from Gram-positive bacteria. This suggests that relatively gentle mechanical disruption techniques, such as vortexing, in combination with chemical procedures may be sufficient for the lysis of microbial cells from intestinal tissues. "The take-home message of the paper is that bead beating is probably too harsh for these biopsies," says senior study author H. Rex Gaskins, an expert in host-intestinal microbiota interactions at the University of Illinois at Urbana-Champaign.
In the end, despite all of these efforts to compare the effectiveness of different DNA extraction techniques, microbiologists are still faced with vast and daunting unknowns. "You never know what you have with these kinds of samples. It's kind of a black box situation," says Jansson. "The problem is when you have an unknown sample, you know that you must have a bias, but unless you sequence every single organism in that sample, you're not going to know what your bias is. So, you never know if you extract DNA from every cell or not. That's still a major hurdle, and I'm not really sure how to get around that one."
1. Flores, G. E., J. B. Henley, and N. Fierer N. 2012. A direct PCR approach to accelerate analyses of human-associated microbial communities. PLoS ONE 7(9): e44563. doi:10.1371/journal.pone.0044563.
2. Gajer, P., R. M. Brotman , G. Bai, J. Sakamoto, U. M. Schütte , X. Zhong, S. S. Koenig, L. Fu, Z. S. Ma, X. Zhou, Z. Abdo, L. J. Forney, and J. Ravel. 2012. Temporal dynamics of the human vaginal microbiota. Sci Transl Med. 4(132):132ra52.
3. Yuan, S., D. B. Cohen, J. Ravel, Z. Abdo, and L. J. Forney. 2012. Evaluation of methods for the extraction and purification of DNA from the human microbiome. PLoS ONE 7(3): e33865. doi:10.1371/journal.pone.0033865.
4. Carbonero, F., G. M. Nava, A. C. Benefiel, E. Greenberg, and H. R. Gaskins. 2011. Microbial DNA extraction from intestinal biopsies is improved by avoiding mechanical cell disruption. J Microbiol Methods 87(1):125-7.