Structural basis for functional properties
Antisense oligos act to block the function of their targeted RNA transcripts. To be useful as research tools and therapeutics they need to be stable in biological systems for at least hours, but preferably days to weeks or longer. While short strands of DNA or RNA may appear to be the obvious choice for use as antisense oligos, because they are degraded by enzymes in the blood and within cells in just a few minutes, it is necessary to design in a means to substantially slow or prevent their enzymatic degradation.
Described below are the ways by which resistance to degradation is achieved in the four antisense structural types which dominate the antisense therapeutics field. It is important to realize that the means by which resistance to enzymatic degradation is achieved in a given structural type strongly affects key functional properties, including: stability; sequence specificity; and, off-target effects.
S-DNAs and 2'-modified S-RNAs (phosphorothioates)
One popular way to achieve moderate stability (hours to days) is to replace the anionic oxygen with an anionic sulfur at some or all of the phosphorodiester inter-subunit linkages in DNA and 2'-modified RNA antisense oligos - as shown below.
siRNAs (short-interfering RNAs)
A “natural” way to achieve moderate stability (hours to days) of short strands of RNA is to pair two complementary RNA strands, as shown below. With such structures one of the RNA strands will ultimately associate with several proteins to form the RISC structure that provides the antisense function, while the other RNA strand will be degraded.
Complete resistance to enzymatic degradation can be achieved by designing an antisense oligo with non-ionic inter-subunit linkages between the subunits. The most successful non-ionic antisense oligos are the Morpholinos, which have non-ionic phosphorodiamidate inter-subunit linkages instead of the anionic phosphorodiester inter-subunit linkages of conventional antisense oligos, and also have morpholine backbone rings instead of conventional ribose or deoxyribose rings. These are shown below.
Four antisense types, described above and shown below, dominate the antisense therapeutics field. Drugs of three of these structural types have been approved for clinical use ( S-DNAs, 2'-modified S-RNAs, and Morpholinos ) and drugs of the fourth type, siRNAs, are in clinical trials.
The properties of these antisense structural types are described below in relation to their respective structures. (reviewed in: Cur. Top. in Med. Chem. 7: 651-660 (2007)).
The key structural elements of phosphorothioates (S-DNAs and 2'-modified S-RNAs) are their anionic sulfurs which serve to slow enzymatic degradation. However, those sulfurs also cause strong binding to a wide range of proteins in the blood, on cell surfaces, and within cells. Binding to proteins in the blood does slow excretion of
S-DNAs by the kidneys, but such protein binding is also responsible for multiple non-antisense effects that plague this structural type. A few such effects include:
a) activation of the complement cascade, which can cause convulsions and death within minutes;
b) CpG activation of the innate immune system; and,
c) G-quartet complexes which can cause multiple effects that can be mistaken for true antisense effects.
The anionic sulfurs in the backbones of S-DNAs also substantially reduce the binding affinities of such antisense oligos for their targeted RNA transcripts. This reduced binding affinity seriously compromises the ability of
S-DNAs to invade secondary structures which are ubiquitous in nearly all RNA transcripts. This makes it difficult to predict suitable target sequences for S-DNAs.
Because S-DNA oligos closely resemble DNA oligos, an S-DNA oligo paired to a complementary RNA transcript is recognized and cleaved by RNase H within cells. The problem this presents is that RNase H recognizes and cleaves paired S-DNA-RNA duplexes as short as 8 base-pairs. Thus, while an S-DNA antisense oligo can often recognize and mediate cleavage of its targeted sequence in a selected RNA transcript, that same S-DNA will also typically recognize and mediate cleavage of thousands of RNA sequences one did not wish to cleave and is unaware of having inadvertently cleaved.
Like S-DNAs, the anionic sulfurs of 2'-modified S-RNAs also bind to a wide range of proteins in the blood, on cell surfaces, and within cells. As with S-DNAs, such binding to proteins in the blood slows excretion of the antisense oligo by the kidneys, but also slows passage of the antisense oligo from the vascular bed to the extra-vascular interstitial space where most cells reside.
In contrast to the case for S-DNAs, the 2'-modified S-RNAs apparently bind proteins with a lower affinity than do S-DNAs and so cause fewer and less severe non-antisense effects due to protein binding. Probably this lower-affinity binding to proteins is due to greater steric crowding around the anionic sulfur by the 2'-modifying group of the ribose.
Another advantage of the 2'-modified S-RNAs over the S-DNAs is that the 2'-modified S-RNAs do not mediate RNase H cleavage of duplexes formed between 2'-modified S-RNAs and their complementary and partially complementary sequences in RNA transcripts. As a consequence, 2'-modified S-RNAs are not plagued by the thousands of undesired RNase H-mediated cleavages of partially-complementary sequences that are inadvertently cleaved by most S-DNAs.
The key structural element of an siRNA is its short base-paired RNA-RNA duplex. A small number of such duplex sequences have been found to induce the innate immune system and so must be avoided when designing siRNAs.
A far more serious challenge in designing an siRNA to target a selected RNA transcript comes from the fact that if the antisense strand of the RISC structure is a perfect, or near-perfect, complement to the targeted RNA transcript sequence then that RISC structure (which incorporates the antisense component) will cleave and thereby destroy the targeted RNA transcript, as desired. However, that same RISC structure can also partially pair with dozens to hundreds of other partially-complementary sequences in other RNA transcripts and act to suppress the function of those partially-complementary RNA transcripts (ie., the RISC structure can act in a microRNA mode wherein the function of an RNA transcript is blocked, but that RNA transcript is not cleaved). Because only about 8 to 12 bases need be complementary to an RNA transcript to mediate this microRNA-type activity, such “non-cleavage” blocking activity typically leads to inadvertent blocking of dozens to hundreds of “non-targeted” transcripts.
While inadvertent blocking of some RNA transcripts by this microRNA-type activity can be partially reduced by a massive informatics search of the transcriptome for each prospective target sequence, even with such a search there typically are still dozens of “non-targeted” RNA transcripts which will be blocked to some extent by any given siRNA. Alternatively, blocking of “non-targeted” RNA transcripts by microRNA-type activity can be reduced by lowering the concentration of the original siRNA - but this comes at the cost of reducing the potency of that siRNA - possibly to the point of being inadequate for achieving any reasonable biological result.
Key structural elements of Morpholinos are their non-ionic inter-subunit linkages and their novel morpholino backbone ring structures. Because proteins interact with nucleic acids in large part via ionic interactions, the lack of ionic charges on the Morpholino backbone avoids significant interactions with proteins. This lack of interaction provides complete resistance to degradative enzymes in blood and within cells. The lack of interaction with proteins probably also accounts for why Morpholinos do not activate the complement cascade, are not involved in CpG-mediated activation of the innate immune system, do not induce interferon, and do not form biologically active G-quartet complexes - all of which are problems which can plague more conventional poly-anionic antisense oligos..
Their lack of ionic charge probably also accounts for why Morpholinos freely pass between the cytosol and nucleus of the cell - which allows their versatile use for both splice modification in the nucleus and translational blocking in the cytosol.
Because Morpholinos are not dependent on cellular machinery (such as RNase H for S-DNAs, or RISC structure for siRNAs) in order to achieve blocking of their targeted RNA transcripts, Morpholino antisense activity tends to be fast, straight forward, and reliable. Also because their antisense activity is independent of cellular machinery, they are quite effective for invading RNA secondary structures. And that facile invasion of RNA secondary structures results in unsurpassed targeting predictability.
Finally, an unusually long run of 14 or 15 contiguous bases of a Morpholino must bind to a complementary sequence in its targeted RNA transcript in order for the Morpholino to effectively achieve its antisense effect. Because of this stringent binding requirement, Morpholinos provide by far the highest sequence specificity of all antisense types.
Morpholinos’ exquisite sequence specificity, combined with their other beneficial properties, have made them the preferred antisense tools for use in the most demanding of all antisense applications - the study of developing embryos wherein intricate cascades of gene activations and deactivatrions are precisely controlled with respect to both time and position in the rapidly maturing embryo. Since the year 2000 Morpholinos have been the essential tools for most researchers in developmental biology, and their use has revolutionized that very challenging field. Scientists using Morpholinos have published over 8,000 research papers (searchable at: pubs.gene-tools.com ), with over 5,000 of those publications being in the developmental biology field.
The following table provides a qualitative comparison of the four dominant antisense structural types in regard to properties needed for high-specificity and long-duration therapeutic activity.
|Stability in blood and cells||limited||limited||limited||Complete|
|General lack of off-target effects||no||no||no||Yes|