Hachimoji DNA and RNA: A Genetic System with Eight Building Blocks

  • Researchers Seek to Expand A, G, C, and T Genetic Coding to Additional Nucleobase “Letters”
  • Steven A. Benner Led Expansion from Four to Six Letter-Coding in 2015
  • In 2019, Coding Uses Eight (“Hachi”) Letters (“Moji”)

According to the principles attributed to the early writings of Francis Crick in 1958, The Central Dogma of Molecular Biologystates that H-bonding between A/T and C/G base pairs underlies the storage of genetic information. This information is in “Watson-Crick” DNA for replication, transcribed into RNA, and finally decoded into protein. Exploring the expansion of such H-bonding (shown here) to include synthetic analogs of these four natural nucleobases has been of interest for theoretical and evolutionary reasons, and could have utility for many hybridization-based applications, as well as for storage of information.

Readers interested in an overview these subjects can consult a 2017 review in Acc. Chem. Res.by Richards and Georgiadis, titled Toward an Expanded Genome: Structural and Computational Characterization of an Artificially Expanded Genetic Information System(AEGIS). The pioneering work of Steven A. Benner on expanding the genetic code from four to six building-block letters is reviewed therein. This blog will highlight Benner’s 2019 Sciencepublication (Hoshika et al.) on AEGIS, which reports further expansion to eight building-block letters. This recently expanded system is appropriately named “hachimoji” DNA and RNA:  Hoshika et al. coined this term by combining the Japanese words for eight (“hachi”) and letters (“moji”).

Hachimoji DNA

At the outset, I should point out that there is a YouTube video lecture by Benner that is well worth watching to fully appreciate the rationale behind investigating AEGIS, and the experimental approaches explored.

Benner’s previously published work (Zhang et al.) on the evolution of a functional six nucleotide genetic code included the two new nucleotides, Z and P, which are shown below for DNA; dR is replaced with R in RNA. The H-bonding in these structures is between oppositely positioned donor (red) and acceptor (blue) atoms. These Z and P nucleotides were shown to undergo enzymatic copying, PCR amplification, and successive transcriptions, first into six-letter RNA, and then back into six-letter DNA.

Taken from Hoshika et al. Science363, 884-887. Copyright © 2019, American Association for the Advancement of Science, with permission.

Expansion from six letters to eight letters was investigated using two additional nucleotides, S and B, shown here. The nucleotides Z, P, S and B along with A, G, C and T were each incorporated into 94 different 8-mer sequences of hachimoji DNA oligonucleotides by use of otherwise conventional phosphoramidite chemistry for solid-phase chain-assembly. Duplexes of these GACTZPSB-containing hachimoji 8-mers were then used to measure melting temperature (Tm) values under a set of standard conditions. These experimental Tmvalues were then compared to predicted melting temperatures derived from state-of-the art thermodynamic parameterization of nearest-neighbor base-pair dimers, as described by Hoshika et al.

Plots of experimental versus predicted free-energy change (ΔG°37) (A) and experimental versus predicted melting temperature (Tm) (B) shown here indicate that, on average, Tmis predicted to within 2.1°C for the 94 GACTZPSB hachimoji duplexes, and ΔG°37 is predicted to within 0.39 kcal/mol. These errors were said to be similar to those observed with nearest-neighbor parameters for standard DNA:DNA duplexes, which was interpreted as meaning that “GACTZPSB hachimoji DNA reproduces, in expanded form, the molecular recognition behavior of standard 4-letter DNA. It is an informational system.”

Taken from Hoshika et al. Science363, 884-887. Copyright © 2019, American Association for the Advancement of Science, with permission.

High-resolution crystal structures were determined for three different hachimoji duplexes assembled from three self-complementary 16-mer sequences: 5’-CTTATPBTASZATAAG, 5’-CTTAPCBTASGZTAAG, and 5’-CTTATPPSBZZATAAG. These duplexes were crystallized with Moloney murine leukemia virus reverse transcriptase to give a “host-guest” complex with two protein molecules (host) bound to each end of a 16-mer duplex (guest). With interactions between the host and guest limited to the ends, the intervening 10 base pairs were free to adopt a sequence-dependent structure.

The hachimoji DNA in all three structures adopted a B-form with 10.2 to 10.4 base pairs per turn, similar to natural B-DNA shown here. The major and minor groove widths for hachimoji DNA were similar to one another and to the DNA duplex 5’-CTTATGGGCCCATAAG, but not to the DNA duplex 5’-CTTATAAATTTATAAG.

Despite these and other differences in structure (i.e. propeller and buckle angles), the structural parameters for the individual pairs and the dinucleotide steps of the hachimoji DNA were said to fall well within the ranges observed for natural 4-letter DNA, consistent with hachimoji DNA being a “mutable information storage system” like natural DNA, according to Hoshika et al. I should interject and state that these researchers use the term “mutable” with reference to Schrödinger, who theorized in 1943 that regularity in size was necessary for nucleobase pairs to fit into what he called an “aperiodic crystal,” which he proposed as necessary for reliable molecular information storage and faithful information transfer.

Hachimoji RNA

T7 RNA polymerase bound to DNA and RNA.

With the information storage and mutability properties shown for hachimoji DNA, Hoshika et al. then asked whether hachimoji information DNA could also be transmitted to give hachimoji RNA. To investigate whether native T7 RNA polymerase (pictured here) is capable of transcribing hachimoji DNA, they started with four model sequences that each contained a single nonstandard hachimoji component, B, P, S, or Z, each followed by a single cytidine. To analyze hachimoji RNA products, they labeled transcripts with [α-32P]cytidine 5´-triphosphate; digestion with ribonuclease T2 then generated the corresponding hachimoji 3′-phosphates. These were resolved in thin-layer chromatography (TLC) systems and compared with synthetic authentic nonstandard 3′-phosphates.

These experiments showed that native T7 RNA polymerase incorporates riboZTP opposite template dP, riboPTP opposite template dZ, and riboBTP opposite template dS. However, incorporation of riboSTP opposite template dB was not seen with native RNA polymerase. This observation was attributed to an absence of electron density in the minor groove from the aminopyridone heterocycle on riboSTP. Polymerases are believed to recognize such density, as it is presented by all other triphosphate substrates.

A, G, C, and U 2′-O-Methyl-Nucleotides.

Hoshika et al. therefore searched for T7 RNA polymerase variants able to transcribe a complete set of hachimoji nucleotides. One variant (Y639F H784A P266L, “FAL”) was especially effective at incorporating riboSTP opposite template dB. Interestingly, FAL was originally developed as a thermostable polymerase to accept 2′-O-methyl triphosphates, pictured right.

High-performance liquid chromatography (HPLC) analysis of its transcripts showed that 1.2 ± 0.4 riboSTP nucleotides were incorporated opposite a single template dB. FAL also incorporated the other nonstandard components of the hachimoji system into transcripts.


The findings reported by Hoshika et al. have been lauded by experts, and they represent a significant advance in synthetic biology with the availability of a further expanded, mutable genetic system built from eight different building blocks: four natural (stars) and four synthetic (circles). By continued investigations, additional synthetic building blocks will perhaps lead to further expansion of genetic coding. Intrigued by this possibility, I found that the Japanese word for ten is “juu,” so “juumoji” DNA and RNA might be next.

In any event, with currently increased information density over natural DNA and predictable duplex stability across all 8nsequences of lengthn, Hoshika et al. concluded that hachimoji DNA has potential applications in sequence-based bar-coding and combinatorial tagging, retrievable information storage, and self-assembling nanostructures. I have covered DNA-based information storage and self-assembling DNA nanostructures (aka origami), in some of my previous blogs.

Hoshika et al. also concludedthat structural differences among three different hachimoji duplexes are not larger than the differences between various standard DNA duplexes, making this system potentially able to support molecular evolution. Furthermore, the ability to have structural regularity independent of sequence shows the importance of inter-base H-bonding in such mutable informational systems. Thus, in addition to its technical applications, this work expands the scope of the structures that might be encountered in search for life in the cosmos, which Benner has written about here.

As usual, your comments are welcomed.

Clinical Efficacy Found: Two Oligonucleotide Drugs for Transthyretin (TTR)-Related Amyloidosis (ATTR)

  • ATTR Is a Class of Life-Threatening and Progressive Genetic Diseases    
  • ATTR Results from Misfolding of TTR, a Blood Protein Carrier of Retinol (Vitamin A), and Thyroxine (T4)
  • TTR mRNA Knockdown by an Antisense Oligonucleotide (ASO) or a Short-Interfering RNA (siRNA) Leads to Clinical Efficacy for ATTR

A recent blog here lauded the beginning of American and European clinical studies of gene-editing by CRISPR, which is enabled by sequence-specific interaction of synthetic guide RNAs with genomic DNA targets. At the end of that post, I mentioned that future clinical successes with CRISPR may encounter the challenge of unforeseen drug development issues, as has been the case for antisense oligonucleotide (ASO) and short-interfering RNA (siRNA) approaches. It took a couple of decades for FDA-approved ASO drugs to progress from “bench to bedside,” and the first siRNA drug has yet to be approved.

After opining the above just a few weeks ago, I am pleased to say that another ASO drug and the first siRNA drug have been reported to both be on the brink of FDA approval, based on back-to-back publications of the results of Phase 3 clinical studies in the highly respected Journal of New England Medicine. Interestingly, both of these mechanistically different oligonucleotide drugs are aimed at the same clinical indication, namely, transthyretin (TTR)-related amyloidosis (ATTR). This blog will serve as an introduction to ATTR, as well as a brief summary of both of these very promising clinical studies.

Introduction to ATTR

As detailed elsewhere, ATTR represents a class of life-threatening and progressing diseases that is associated with the misfolding of TTR, a major blood protein that is a carrier of retinol (vitamin A) and thyroxine (thyroid hormone T4). TTR is primarily (>95%) produced in the liver as a tetramer, as depicted below. Amyloid aggregation is believed to be a result of decreased tetramer stability, resulting in dissociation of TTR into monomers. These monomers are prone to unfolding, and can self-assemble into oligomers and amyloid fibrils found in peripheral neurons, the gastrointestinal tract, and heart.

Taken from link.springer.com

The three major forms of ATTR disease are described in the NIH Genetic Home Reference (GHR), which also provides information about their underlying genetics and diagnosis by, for example, DNA sequencing. In hereditary ATTR with polyneuropathy [aka familial amyloid polyneuropathy (FAP)], the peripheral nerves are primarily affected, while in cardiomyopathy-related ATTR [aka familial amyloid cardiomyopathy (FAC)], neuropathy is usually less prominent.

According to the GHR, ATTR among Americans of European descent is estimated to affect one in 100,000 people, while FAC is far more frequent among people with African ancestry. For example, it is estimated that this form affects between three to four in 100 African Americans. Although the reasons for this disproportionately higher incidence are unknown, suggestions on how to raise awareness on this issue are discussed in a review I found while researching FAC among African Americans.

For the most part, patients develop a severe disease and die within 5 to 15 years after onset. While these forms of ATTR can be ascribed to a dominant expression of the TTR gene variants, (their number is ~100, as depicted here in an artistic rendering), only wild-type TTR is expressed in senile systemic amyloidosis (SSA), a type of amyloidosis frequently found in elderly people.

Taken from pharmaceuticalintelligence.com

The ATTR pathway above depicts the possibility of using oligonucleotides that interfere with mRNA function to block the synthesis of TTR. While blocking synthesis of a specific mutant of TTR is possible, biochemical and clinical evidence suggests that the wild-type TTR can also significantly contribute to the disease. In 2016, results obtained from an in vitro model system were jointly published by Ionis Pharmaceuticals and Alnylam Pharmaceuticals. The study supports the notion that ASOs or siRNAs can block TTR synthesis, and can therefore represent a possible therapy for ATTR. These findings paved the way for each company to independently carry out individual clinical development programs for oligonucleotide drug candidates, as outlined below.

ASO Drug Inotersen

Ionis’ inotersen (formerly IONIS-TTRRx/ISIS 420915), which is a 2′-O-methoxyethyl (MOE)– and phosphorothioate (PS)-modified RNA (orange)-DNA (black) “gapmer” ASO shown here, is an inhibitor of the hepatic production of TTR. Readers interested in ASO chemistry and mechanistic details for gapmers can consult an excellent review by Shen and Corey.

Taken from Shen and Corey Nucleic Acids Research (2018)

Taken from Buxbaum Journal of New England Medicine (2018)

As shown here, the genome of patients with hereditary amyloidotic polyneuropathy contains one copy of wild-type TTR and one copy with a base substitution (mutation), resulting in a change in the amino acid sequence. In hepatocytes, both copies are equally transcribed and translated. Inotersen binds to wild-type and mutant TTR mRNA transcripts, resulting in their degradation by ribonuclease H.

In healthy volunteers, inotersen showed dose-dependent and sustained reductions of circulating TTR levels. Ionis then conducted a randomized, double-blind, placebo-controlled, Phase 3 trial (NEURO-TTR) to determine the efficacy and safety of inotersen treatment in patients with hereditary transthyretin amyloidosis with polyneuropathy, in the presence or absence of cardiomyopathy. According to Benson et al., eligible patients were randomly assigned (in a 2:1 ratio) to receive 300 mg of inotersen or placebo. Patients 18 to 82 years of age received three subcutaneous injections during the first week to achieve near steady-state drug levels, followed by a once-weekly subcutaneous injection for the next 64 weeks.

A total of 172 patients (112 in the inotersen group and 60 in the placebo group) received at least one dose of a trial regimen, and 139 (81%) completed the intervention period. In the inotersen group, reductions in circulating TTR reached steady-state levels by week 13 and were sustained through the end of the intervention period. From week 13 to week 65, decreases in serum TTR from baseline levels in the inotersen group reached a median nadir of 79%.

Primary efficacy assessments favored inotersen. These improvements were independent of disease stage, mutation type, or the presence of cardiomyopathy. The most frequent serious adverse events in the inotersen group were glomerulonephritis [in 3 patients (3%)] and thrombocytopenia [in 3 patients (3%)]. It was concluded that “inotersen improved the course of neurologic disease and quality of life in patients with hereditary transthyretin amyloidosis” and that “[t]hrombocytopenia and glomerulonephritis were managed with enhanced monitoring.”

siRNA Drug Patisiran

According to the above cited review by Shen and Corey, and as shown here, Alnylam’s patisiran is a double-stranded siRNA comprised of 3’-end DNA (black), RNA (pink), and 2’-O-MOE RNA (blue) moieties. Partisiran is delivered in complex with a lipid nanoparticle (LNP), (also depicted here), and is described in detail elsewhere by Acuitas Therapeutics, an LNP drug-delivery company that partnered with Alnylam.

Taken from Shen and Corey Nucleic Acids Research (2018)

Taken from acuitatx.com

The mechanism of action depicted here for the siRNA patisiran involves the formation of an RNA-induced silencing complex (RISC) with both wild-type and mutant TTR mRNA transcripts and subsequent mRNA degradation.

Taken from Buxbaum Journal of New England Medicine (2018)

Adams et al. have now reported results from a Phase 3 clinical trial by Alnylam wherein patients were randomly assigned (in a 2:1 ratio) to receive patisiran (0.3 mg per kilogram of body weight) or placebo intravenously over a period of approximately 80 minutes, once every 3 weeks for 18 months. A total of 225 patients 24 to 83 years of age were randomly assigned to receive patisiran (148 patients) or placebo (77). In the patisiran group, the reduction in serum TTR levels was rapid and sustained over a period of 18 months, with a median reduction of 81% that remained similar across age, sex, or genotype.

Primary efficacy endpoints favored patisiran and approximately 20% of the patients who received patisiran and 10% of those who received placebo had mild or moderate infusion-related reactions; the overall incidence and types of adverse events were similar in the two groups. It was concluded that “patisiran improved multiple clinical manifestations of hereditary transthyretin amyloidosis.”

One Expert’s Opinion

Joel Buxbaum, an MD and Professor Emeritus in the Department of Molecular Medicine at the Scripps Research Institute, accompanied these back-to-back publications about inotersen and patisiran in the Journal of New England Medicine with an Editorial. Based on his publications in this field, Buxbaum is an expert on the molecular basis of acquired and hereditary human disease related to human amyloidosis. After summarizing the results of each clinical trial, Buxbaum posed the following scientifically intriguing questions, which had not occurred to me after reading these two reports.

“Although each of the trials unequivocally shows a therapeutic effect, there remain questions. Would a single patient have a response to each of the therapeutics to the same degree? If so, it is in the patient’s interest to use the least expensive therapy. If not, N-of-1 trials of sequential treatments, although cumbersome, would be required to identify the best treatment for each patient, conditional on the availability of a rapidly responsive, validated surrogate marker of disease.” In regard to the currently trending concept of N-of-1 trials, Buxbaum refers the reader to a commentary in Nature by Nicholas Schork titled Personalized medicine: Time for one-person trials, from which I borrowed this illustration that visually coveys Schork’s point.

Taken from Schork Nature (2015). Illustration by Greg Clarke

Buxbaum continues his expert commentary as follows:

“It is also possible that a combination of interventions would elicit a more pronounced, durable therapeutic effect. Although there remains much work to be done, the trials by Adams et al. and Benson et al. represent a landmark: together, they show that the rate of progression of a peripheral neurologic disease can be slowed, and perhaps ameliorated, through the use of oligonucleotide drugs that are administered systemically.”

In my humble opinion, the promising outcomes of these two trials are indeed great news for people afflicted with ATTR, who can be future recipients of this oligonucleotide therapy. The results are likewise great news for families and friends. The many researchers, who collectively contributed—directly or indirectly—to achieving this milestone, are also undoubtedly pleased by these results.

As usual, your comments are welcomed.

60th Anniversary of the Discovery of DNA’s Double Helix Structure…Diamond Jubilee for the “Monarch of Molecules”

Welcome to my inaugural blog post!  My intention is to provide timely nucleic acid-related scientific content that is informative and, hopefully, will be of interest to a broad readership. New posts will be published biweekly so please check back often.  I encourage thoughtful commentary as well as constructive suggestions.  To find out more about me and my relationship with TriLink Biotechnologies, please visit the ‘About Jerry’ tab at the top of the page.

Considering TriLink’s focus on providing nucleic acid-based products, it seemed appropriate that this inaugural post feature the upcoming 60th anniversary of Watson & Crick’s proposed structure for DNA published in Nature on April 25, 1953. It is widely acknowledged that insights provided therein had a fundamentally transformative impact on science and society. Anyone working with DNA should take a bit of time to read this historically important publication that is freely available at Nature.com. Doing so reveals an extremely brief report—one page and one figure—with perhaps the most understated—and similarly brief, one sentence!—conclusion suggesting the structural basis for genetic encoding by DNA sequence and its replication:

watson-crick“It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material.”

Those interested in an annotated and illustrated autobiographical account of this making of a scientific revolution are referred to a recent book by James D. Watson, Alexander Gann and Jan Witkowski, The Annotated and Illustrated Double Helix that has already received numerous excellent reviews.  There is also a YouTube video of a delightful presentation by Dr. Watson that is well worth watching to see and hear the story of events leading to the 1953 publication and subsequent recognition that DNA is transcribed into mRNA for translation into protein. This new knowledge coupled with the availability of various molecular biological “tools,” and initial advances in oligonucleotide synthesis, set the stage for an exciting race during the late 1970s to synthesize for the first time a medically important human gene—insulin—involving three main groups—Harvard University, University of California at San Francisco, and City of Hope National Medical Center. This fascinating story involving science, egos, and the beginning of bio-venture “start-up” financing that collectively spawned Genentech and Biogen—and subsequent “genetic engineering” companies—is engagingly chronicled by Stephen S. Hal in the book Invisible Frontier-The Race to Synthesize a Human Genome through interviews with numerous individuals who were directly involved.

Through many advances in DNA Sanger sequencing methods and automated instrumentation during 1990-2003, base-by-base sequence determination of the entire human genome was realized by parallel public (government funded) and private (Celera Corporation) efforts.1 During the next decade, taking us to 2013, there has been stunning progress in developing various methods of massively parallel sequencing2 (aka “next-generation sequencing”) that, in part, has enabled further elucidation of epigenomics and RNA-mediated regulation as well as factor-mediated reprogramming of cells and pursuit of regenerative medicine. Such topics were discussed at a recently held Cold Spring Harbor Laboratory meeting entitled “From Base Pair to Body Plan – Celebrating 60 Years of DNA.” Below are just a few selected author names (in alphabetical order) and titles of presentations taken from the list of nearly 50 talks or posters found at the website for this event.3

Baylin, S. Celebrating the discoveries of the “hard drive” of DNA and its “software package,” the epigenome—Basic and translational implications
Young, R.A. Control of gene expression programs
Zaret, K.S. Programming and reprogramming cell fate

In closing this inaugural blog post, one can only wonder what new and exciting aspects of life sciences, diagnostics and medicine—including regenerative—will be realized during the next 10 years on the way to the 70th Anniversary of the Discovery of DNA’s Double Helix Structure. My view of future advances based on DNA fall into three interrelated areas:

  • “faster, better, cheaper” synthesis to create useful DNA-directed, self-assembling  “smart materials”
  • “bigger and deeper” experimentation aimed at complete molecular-level models for organisms and memory
  • “bio-factories” for Green Manufacturing

These thoughts about the next decade for DNA will be elaborated in my future posts.  What do you see happening during this time?



1. http://en.wikipedia.org/wiki/Human_Genom_Project

2. http://en.wikipedia.org/wiki/Massive_parallel_sequencing

3. http://meetings.cshl.edu/abstracts/dna602013_absstat.html