In the early 1980s, Steve Benner had just earned a Ph.D. in chemistry from Harvard University, working under Nobel Prize winner Robert Burns Woodward. At the time, chemists like Woodward and Elias James Corey were synthesizing a wide variety of molecules such as steroids, alkaloids, and proteins, but spent little time looking at nucleic acids. So, Benner decided to synthesize an unnatural base-pair that would extend the genetic alphabet, a feat that would have valuable applications in diagnostics, therapeutics, and biotechnology.
One of the reasons is that natural polymerases, the enzymes that replicate DNA by matching bases to their counterparts, have evolved over four billion years to faithfully recognize and match four substrates: adenine (A), thymine (T), guanine (G), and cytosine (C). So, when unnatural base-pairs are added, they often gets lost because of mismatching, eventually dropping out completely from the sequence.
“When we started this, it wasn’t clear that any of this would work,” says Benner. “But we are now are going much further; we want these “funny” bases copied, and we want these these copies copied and those copies copied as well.”
And now, the three labs at the forefront of the field have each reported that DNA containing their candidate unnatural base-pairs have been amplified by PCR, with varying degrees of success. For each team, their individual milestone has been years in the making, overcoming the challenges inherent in trying to coax an unnatural substrate to behave like its natural counterpart. And now, the work on the applications that were impossible without the ability to PCR amplify DNA containing these unnatural base-pairs is beginning.
One Small Step
At the Foundation for Applied Molecular Evolution in Gainesville, FL, Benner has been focused on the work that he began almost 30 years ago. Being a pioneer in the field, Benner has already tasted success in the form of several applications of his unnatural base-pairs. For example, in 2003, the U.S. Food and Drug Administration approved clinical assays that incorporated his unnatural base-pairs as tags to detect HIV, hepatitis B, and hepatitis C. While qPCR could detect these viruses, it is more labor-intensive than his clinical assays, and the amplification involved could potentially lead to contamination as well. “You’re amplifying the thing you’re trying to detect,” says Benner. “That’s not necessarily something that you want to do.”
But what excites him now is moving beyond using his synthetic nucleotides just as tags and actually incorporating them into the sequence of living cells. Of course, incorporating those nucleotides into living cells requires convincing polymerases to add those unnatural base-pairs in a specific way.
“We were going to understand what a polymerase looks for in the four different natural substrates and try to accommodate as much of the desires of natural polymerases without abandoning our goal of increasing the information density of DNA,” says Benner.
To ensure that his unnatural base-pairs would be compatible with natural polymerases, Benner’s group took the smallest step away from natural substrates as they could. To develop synthetic base-pairs, the group rearranged the hydrogen-bond geometries that linked the four natural base-pairs together. During the early 1990s, the group reported one notable pair called isoC and isoG, isomers of C and G. But because these bases were derived from natural bases, they had low selectivity during PCR replication, leading to pairing with the natural bases and eventual elimination from the system.
More recently, Benner and colleagues have developed a synthetic base-pair called P-Z with improved selectivity during PCR amplification. In a paper published by the Journal of the American Chemical Society in September 2011, Benner and colleagues reported that their Z and P nucleotides were capable of amplification in a six-letter genetic system with a theoretical 99.8% selectivity (1).
But the design of the nucleotides was only one of three optimized variables that led to this success; the others were the polymerase and the PCR conditions. “If you look at these 20 years of papers, you’re going to see all three of these things,” says Benner. “It’s a three-part optimization.” With every incarnation of their synthetic base-pair, Benner’s team had to test their new bases with a large library of polymerases and variety of PCR conditions.
In the end, Benner and his team have developed what he calls the first Darwinian six-letter synthetic genetic system that is capable not only of maintaining the unnatural bases during replication efficiently but also of incorporating them as mutations during replication because of favorable mismatches with natural bases by the polymerase. “Now the only question that remains is the level of loss and gain in this particular system.”
A Larger Step Away
Meanwhile, at the Scripps Research Institute in La Jolla, CA, Floyd Romesberg was following in Benner’s footsteps while taking a slightly larger step away from nature. “We’ve taken a completely different approach than his lab has, but he was the first to popularize the idea and work on it,” says Romesberg.
The inspiration for Romesberg’s base-pairs came when Eric T. Kool and Barbara A. Schweitzer published a paper in 1995 in the Journal of the American Chemical Society (2) where they reported that when the oxygen molecules of a T base were replaced by fluorine, the modified base was still matched with A by a polymerase during DNA replication, even in the absence of hydrogen bonding. “This is one of the most interesting observations in chemical biology that have ever been made,” says Romesberg. “That those hydrogen bonds that we’re all so fond of, that we all draw when we draw DNA, are not absolutely required for replication.”
So, for the past 12 years, Romesberg’s group has been developing their own unique base-pairs that do not rely on hydrogen bonding in their chemistry. The idea is that the elimination of this common feature of the four natural base-pairs would improve the selectivity of the unnatural base-pairs within a genetic system. While the candidate synthetic base-pairs that they developed over the past decade have continuously shown high selectivity, replication has been another story. For example, the replication of the team’s 5SCIS-MMO2 and 5SCIS-NaM base-pairs, which was described in a paper published in Chemistry (3), was dependent upon sequence context.
But now, in a paper published in the Proceedings of the National Academy of Sciences in July 2012, Romesberg and colleagues reported that their unnatural base-pair d5SICS-dNaM could be amplified by PCR with high efficiency and greater than 99.9% fidelity (4). In addition, the replication is sequence-independent, so no matter where the unnatural base-pairs are in the DNA sequence and no matter what bases surround them, they will be replicated under optimized conditions.
In the end, they showed that their unnatural base-pair actually has less sequence bias during replication than natural base-pairs have. “In contrast, the Benner lab has never looked at sequence bias, they only reported work in a single sequence context, so it’s simply unknown whether they’d have sequence bias or not,” says Romesberg.
Now, after 12 years of optimizing the chemistry of his synthetic base-pair, Romesberg is excited to get started on the applications. “I didn’t let anyone in my group work on applications until very recently because I didn’t want to work on an application until we had something that was worth developing. Essentially, our PNAS paper is the final demonstration of what we have now, which is at least first-order suitable,” says Romesberg.
Specifically, Romesberg’s lab is splitting their application development into two broad categories: in vivo and in vitro. Both of these efforts are geared toward leveraging the ability of their synthetic base-pair to modify DNA in a site-specific manner. Their in vivo efforts aim to optimize their system to replicate in DNA in living cells to in order to encode unnatural proteins. “You could also do a whole bunch of things like site-specific modify chromosomal DNA with a histone or a fluorophore or site-specific modify RNA for RNA trafficking experiments,” says Romesberg. Meanwhile, their in vitro efforts include applying their unnatural bases to materials, diagnostic arrays, and other molecular biology techniques that are reliant upon PCR amplification.
An Even Bigger Step
Since 1997, Ichiro Hirao and his group at RIKEN Systems and Structural Biology Center in Tokyo, Japan, have been developing unnatural base-pairs for diagnostic and therapeutic applications. At first, the Hirao group took a similar approach to Benner’s group by reorganizing the hydrogen bonds of a natural base-pair. But Hirao’s lab also leveraged the steric hindrance effect to improve selectivity by attaching a large dimethylamino group to one of the unnatural bases, which prevented reactivity and thus mismatches with T.
Unfortunately, this approach never achieved a synthetic base-pair with high selectivity. So, in a dramatic change, Hirao’s group decided to remove the hydrogen bonds from their candidate base-pair, similar to Romesberg’s approach. As a result, their base-pair represents an even greater step away from nature and required further modifications for recognition by the polymerase and highly optimized conditions for replication.
In a paper published in Nucleic Acids Research in November 2011 (5), Hirao and colleagues reported that their base-pair Px-Ds demonstrated high amplification efficiency and fidelity. Their reported selectivity was higher than 99.9%, and their products remained even after 100 cycles of PCR.
But their system is far from a Darwinian system that Benner’s group has developed. “Hirao will never have a case where a copy will see the introduction of his synthetic base. So, this results in a unidirectional loss,” says Benner.
Another downside of their accomplishment is the bias of their system in DNA amplification. When it’s next to certain nucleotides, replication is sequence dependant. “Unfortunately, Hirao’s base-pairs are extremely biased,” says Romesberg. “They cannot include a G next to one of their unnatural nucleotides. It just doesn’t amplify. That dramatically limits their ability to use this base-pair.”
“We’re taking a very small step away from natural base-pairs. Our Z-P pair still presents a chemically functional groove that the polymerase can form a hydrogen bond with,” says Benner. “The bigger these steps away from nature you take, the more the polymerase is going to worry about this stuff and more severe the context becomes.”
1. Yang, Z., F. Chen, J. B. Alvarado, and S. A. Benner. 2011. Amplification, mutation, and sequencing of a six-letter synthetic genetic system. J. Am. Chem. Soc. 133(38):15105-15112.
2. Schweitzer, B. A., and E. T. Kool. 1995. Hydrophobic, Non-Hydrogen-bonding bases and base pairs in DNA. Journal of the American Chemical Society 117(7):1863-1872.
3. Malyshev, D. A., D. A. Pfaff, S. I. Ippoliti, G. T. Hwang, T. J. Dwyer, and F. E. Romesberg. 2010. Solution structure, mechanism of replication, and optimization of an unnatural base pair. Chem. Eur. J. 16(42):12650-12659.
4. Malyshev, D. A., K. Dhami, H. T. Quach, T. Lavergne, P. Ordoukhanian, A. Torkamani, and F. E. Romesberg. 2012. Efficient and sequence-independent replication of DNA containing a third base pair establishes a functional six-letter genetic alphabet. Proceedings of the National Academy of Sciences of the United States of America 109(30):12005-12010.
5. Yamashige, R., M. Kimoto, Y. Takezawa, A. Sato, T. Mitsui, S. Yokoyama, and I. Hirao. 2011. Highly specific unnatural base pair systems as a third base pair for PCR amplification. Nucleic Acids Research 40(6):2793-2806.