The nuts and bolts of the genius behind aptamers - the future for drug delivery

Aptamers  have created a lot of research interest for the use of biosensors and novel methods of drug delivery. Aptamers are short, single-stranded nucleic (or peptide) molecules that can be created to bind very selectively to their targets, which can be very different in size, shape and chemistry. So how do aptamers work? How can any polymer do what aptamers can? This article is exploring the nuts and bolts of the tiny environment that binding reactions occur at.

There’s no such thing as ‘just a protein’:

Proteins, as you may remember, are made up of specific sequences of amino acids. While every amino acid has their backbone groups, which allow them to join to each other, each type of amino acid has its own specific chemical properties on their side chains– some side-chains are very large, some are small, some are negatively charged, some are positively charged, some have polar groups (which interact with other polar groups, such as water) and some have hydrophobic groups (which interact with hydrophobic groups, such as fats).  It’s a bit like having beads on a necklace: many different shapes and colours of beads (the side chains), all joined together by a single thread (the backbone).

Very few proteins, however, are exactly linear. Most of them fold to create very specific three-dimensional shapes after being synthesised by cells, some spontaneously, others with a little help from other cellular components. The precise shape of a protein has a huge influence on how it behaves and what reactions it participates in. The exact sequence of amino acids in a protein (the “primary structure”) puts the various side-groups of the amino acids close to one-another. These amino acid side-groups can interact with each other: negatively-charged groups can bind to positively-charged groups, hydrophobic side-chains can bind to one-another, and so on. The interaction of these side-chains allows areas of the protein to form certain shapes, like spring-shaped helices and flat beta sheets (“secondary structures”). All of these structures within the protein then associate with one another to produce the final complicated three-dimensional shape of the protein, the “tertiary structure”. From this, we can see that the exact sequence of amino acids in a protein has an enormous effect on the finished product. Altering just one amino acid in the sequence of a protein can prevent an important secondary structure to form and change the entire shape of the protein.

It’s not just the shape of the protein that affects their function. Many of the side-chains are left exposed on the surface of the protein, which creates tiny pockets of different chemistry along the outside of the protein, the microenvironment. Each protein is therefore made of a unique combination of both its shape and the microenvironment, which together provide the protein with the ability to function within its intended environment. Understanding this allows us to see why a protein can bind to a target in the selective manner that it does – the shape and microenvironment can provide a perfect lock into which the target (itself having a unique shape / microenvironment) fits into. This is how enzymes recognise the molecules that it catalyses, or antibodies recognise the pathogen in the bloodstream.

 

 

 

 

 

 

Figure 1: The exact shape and microenvironment of a protein. The different colours on the protein all indicate different types of amino acids present on the surface of the protein. Red areas contain positively charged amino acid residues, blue are negatively-charged, green are hydrophobic and white groups are polar. The protein in this example is streptavidin (accession number: 1MEP.pdb)

Oligonucleotides: Revenge of the double-helix

In the same way that proteins are created of chains of amino acids, nucleic acids (such as RNA, or DNA) are made of chains of nucleotides. Like amino acids, nucleotides have portions of their molecular structure that lets them join to another one (their backbone) and side-groups of their own (their “bases”). The bases allow nucleotides to bind to one-another at specific areas. Guanosine (a type of nucleotide) binds to cytosine nucleotides, adenine nucleotides bind to thymine nucleotides. The reason that two strands of DNA ends up forming a double helix is because the two molecules line up head-to-tail and allow the nucleotides to bind to their counterparts.  That’s why DNA requires two strands of DNA with precisely opposite sequences of nucleotides to form. If there is an adenine nucleotide at the top of one strand, there needs to be a thymine at the bottom of the other strand (remember, they line up head-to-tail with each other) in order for the area to bind together, and so on.

Just like proteins, however, it doesn’t require two separate strands of DNA, or RNA to make a complex, three-dimensional shape from nucleic acids. The precise sequence of nucleotides (its primary structure) allows much more complicated shapes to arise, in much the same way that protein’s primary structure allows. Each of these nucleotides has its own unique chemistry, so just like the different amino acids in proteins, there are unique little pockets of chemistry all along this shape.

Aptamers, made of short, single-stranded chains of nucleotides, operate on this principle in order to bind their targets. Since a short sequence of DNA, or RNA can fold itself into a specific shape having many specific microenvironments, a known sequence of DNA can fold itself into a known shape with a known microenvironment. We can take advantage of this behaviour to produce aptamers that bind, or fit to, their target compounds in the same manner that proteins do.

This implication is huge for two important reasons. The first is that scientific research has become very good at manipulating nucleic acids in the lab – they’re much easier (and cheaper) to make than proteins – there are companies that specialise in just that and a number of easy-to-use protocols to make some in the lab. And we can generate aptamers to bind many things. If you’re trying to bind a compound using a protein, you need a protein that can interact with it in such a way that a signal is generated. These are often very difficult, if not impossible, to either find in nature, or make in a lab for all the things that we’d like to engineer binding reactions for.

So, the big question is how do we engineer nucleic acids to selectively bind to target molecules that we’d like to sense, or target? The answer is by using the laws of probability and a few rules of evolution.

 

SELEX: Evolving aptamers to targets

SELEX, short for the Systematic Evolution of Ligands by EXponential enrichment is a very cunning strategy. Usually, it is performed by throwing a huge amount of different sequences against that target you’d like to make an aptamer for, and seeing what sticks. Seriously.

10E15 is a big number. It’s 10,000,000,000,000,000. It’s also about the number of distinct nucleic acid sequences that you start out with when you’re starting your SELEX, give or take a few ,000s. That’s 10E15 different shapes and microenvironments you’re flinging at the target. Something’s bound to...err...bind, really. SELEX  (Figure 2) is evolution in miniature – (1) you start with a target population (your nucleic acid sequences), (2) put them in a selective environment (can they bind to a target?), (3) remove the losers (the many sequences that don’t bind) and (4) separate the winners (the few sequences that can bind to the target) from the target. Once you have a small pool of winning sequences (5), you can easily make more of them and put them back into the same conditions and repeat this cycle a few times until you end up with a few sequences that turn out to be really, really, good at binding to the target.

Figure 2: The first cycle of SELEX. Inset diagram (dashed circle) shows the 3-D shape of the nucleic acid sequence compared to the schematic that we use for the purposes of the diagram. A more in-depth explanation of the process is in the above paragraph.

The amazing thing about this is that, just by altering the conditions of the target, you can end up with at least one sequence that can perform under your required conditions. Need an aptamer for a sugar in acidic conditions? You can probably get that. Need it for something as big as a whole cell? It’s been done. Need an RNA aptamer, instead of DNA? Just run it with RNA molecules at the start. Need an aptamer that binds selectively to Compound 1, but not that really similar Compound 2 over there? Just get the sequences that bind to Compound 1, react them with Compound 2 and remove the ones that bind to both.

That, in a nutshell is the mechanism that gives aptamers their current promise in a lot of therapeutic and sensor work.

*Dr Ronen Fogel has a PhD in Biotechnology