#protein science

[moo-cyte]

I can be the substrate

[lifesciencepotluck]

Figuring out structure

Proteins are like molecular machines. Unlike actual machines, proteins can fold into intricate 3-D structures on their own. Wouldn't it be convenient if your bookshelf from IKEA could do that? If we want to understand how proteins work, we need a blueprint. Some parts of the protein are essential for stability; other parts are crucial for getting things done through binding or catalysis. We need to understand how different parts of the protein fit together; this is where Structural Biology comes in. Getting the 3-D structure is one of the main steps in characterising new anemone toxins. I will cover some of the techniques we use to investigate the toxin structure.

The hierarchy of protein structure goes like this: primary structure-->secondary structure-->tertiary structure-->quaternary structure. For anemone toxins, we don't have to worry about the quaternary structure. Usually, anemone toxins have one polypeptide, not multiple polypeptides.

Figure 1. Structures of anemone toxins (Source: Norton, R. S, 2009)

Primary structure:

It's the amino acid sequence or 'code' for the protein. You can get the amino acid sequence from DNA, mRNA or protein sequencing. Protein sequencing just gives you the sequence for your protein. However, if we use DNA or mRNA sequencing, then we have to figure out the 'mature' peptide sequence. DNA and mRNA code for the 'first draft' of the peptide. After translating mRNA to protein, the signal peptide and pro-peptide are attached to the 'mature peptide.' Those two parts of the protein get cut out before our mature peptide can go out in the world and decimate prey.

Secondary structure:

Our flat polypeptide chain will fold into different secondary structures like alpha-helices, beta-sheets, beta turns, etc. Circular dichroism and Fourier transform infrared radiation (FT-IR) can tell us about the secondary structure of the toxin. CD and FT-IR are commonly used to predict the amount of each secondary structure in proteins. For CD, we measure how much of circularly polarised light in the far-UV region is absorbed by the sample. We use curve fitting programs to determine which combination of secondary structures fits the spectra the best (e.g. 50% beta-sheet, 10% alpha-helix, 20 % random coil and 20 % beta-turn). FT-IR works similarly, but we use infrared radiation instead of polarised light. FT-IR is a better option for investigating the interactions between pore-forming toxins and lipid membranes. The lipid vesicle or 'sacs' scatter polarised light, which it makes challenging to use CD for the same purpose.

One of the best things about spectroscopic techniques is that they are not as time-consuming as, X-ray crystallography. You can have multiple CD, and FT-IR runs in 1-2 hours and get the processed data before lunch. However, CD and FT-IR are low-resolution techniques. They don't tell you the exact position and orientation of tryptophan 32. We need atomic resolution structures to get an in-depth understanding of how anemone toxins work.

Tertiary structure:

The full 3-D structure of proteins will have secondary structure plus the disulphide linkages, salt bridges, hydrophobic interactions. There are two main ways of getting atomic resolution structures: X-ray crystallography and NMR. X-ray crystallography can give high-resolution 3-D structures—if you manage to please the crystal gods. If you don't have the ideal crystallisation conditions—no crystal for you. I know PhD students who spent 12-18 months on crystallising their protein. However, once you get past the crystallisation bottleneck, the diffraction experiments and data processing doesn't take as much time. In theory, you can get better resolution with X-ray crystallography than NMR. Let's assume you don't have time for crystallisation and you go with NMR. NMR is excellent for finding the structure of smaller proteins, but for larger proteins, you have to use isotopic labelling. Some proteins that are too big for NMR. In X-ray crystallography, the protein is a 'frozen' state. However, NMR also lets you study proteins in solution to understand the dynamics of the protein. You can observe how the protein changes when it binds to something.

We can also predict the 3-D structure of the toxin-based on the sequence alone.

How?

Computer magic !

With the protein data bank and structure prediction tools like SWISS-MODEL, you can predict the 3-D structure of a new toxin. If there is an anemone toxin in the database with an experimentally determined 3-D structure and it has high sequence similarity with your new toxin, you can use homolog modelling. If you can't use homology modelling, you can use de novo structure prediction like I-TASSER. In de novo modelling, the program predicts the structure for smaller sections of the protein and uses those bits to build the full predicted model. Model prediction isn't 100% accurate, and we still have to determine the 3-D structure experimentally. However, predicted 3-D structures act as a guide.

Every technique has its strengths and weaknesses. A good researcher tries to use the most suitable method for their project to its full potential. You have to realise the power that's inside.

So, what can you do once you have the 3-D structure of a toxin?

I am going to use APETx1 from the clonal anemone as an example. APETx1 is a potassium channel toxin, which targets the Ether-a-gogo channels in heart cells. Since APETx1 is a small potassium channel toxin, Chagot and company could use NMR. They found that APETx1 belong to the Defensin family. A variety of anti-bacterial proteins and ion channel toxins belong to this family. Proteins with Defensin-4 domain, have double or triple-stranded antiparallel b-sheets linked to a short a-helix by two disulphide bridges. Since NMR can reveal disulphide linkages, Chagot found APETx1 has an inhibitor cysteine knot (ICK). The inhibitor cysteine is a very popular protein fold because it's so stable that it makes proteins resilient against heat stress and proteolysis. Getting the structure of a toxin allows us to classify the toxin and compare it with structurally similar toxins. Since APETx1 has structurally similarity with two other toxins BeKm1 and CnErg1, you can infer it would have a similar mechanism for blocking potassium channels. Chagot suggested that five key amino acids on APETx1 (i.e. Tyr 5, Tyr 32, Phe33, Lys8 and Lys 18) 'block' the Ether-a-gogo channels through electrostatic interactions.

Figure 2. 3-D structure for APETx1 (Source: Chagot et al, 2005)

In short, we can use a variety of biophysical techniques to get structural information for new toxins  and the structure provides insight into the function. Protein structure and function are different sides of the same coin. Next time, I will flip the coin by covering how we figure out function.

Citations:

Chagot, B., Diochot, S., Pimentel, C., Lazdunski, M. & Darbon, H. Solution structure of APETx1 from the sea anemone Anthopleura elegantissima: A new fold for an HERG toxin. Proteins Struct. Funct. Genet. 59, 380–386 (2005).

Jouiaei, M. et al. Ancient venom systems: A review on cnidaria toxins. Toxins (Basel). 7, 2251–2271 (2015).

Norton, R. S. Structures of sea anemone toxins. Toxicon 54, 1075–1088 (2009).

Norton, R. S. & Pallaghy, P. K. The cystine knot structure of ion channel toxins and related polypeptides. Toxicon 36, 1573–1583 (1998).

Dauplais, M. et al. On the Convergent Evolution of Animal Toxins. J. Biol. Chem. 272, 4302–4309 (1997).

Belmonte, G. et al. Primary and secondary structure of a pore-forming toxin from the sea anemone, Actinia equina L., and its association with lipid vesicles. BBA - Biomembr. 1192, 197–204 (1994).

Honma, T. & Shiomi, K. Review Article Peptide Toxins in Sea Anemones: Structural and Functional Aspects. Mar. Biotechnol. 8, 1–10 (2006).

Bernard Gilquin, Judith Racape, Anja Wrisch, Violeta Visan, Alain Lecoq, Stephan Grissmer, Andre´Menez, and S. G. Structure of the BgK-Kv1.1 Complex based on distance restraints identified by double mutant cycles: Molecular basis for convergent evolution of Kv1 channel blockers. J. Biol. Chem. 277, 37406–37413 (2002).

[lifesciencepotluck]

Hunting for anemone toxins

Maybe some of you read my posts in Anemone toxins 101 and asked yourself, "How do people find anemone toxins and how do they know what these toxins do?"

Well, my friends, I am going to cover the arduous process of discovering and characterizing anemone toxins in Anemone Toxins 201. Once people figure out the sequence, structure and function of an anemone toxin, there are a lot of amazing you can do with anemone toxins. We can use anemone toxins to develop therapeutic proteins. For Ecologists, investigating the toxin profile of an anemone can provide a lot of information on its physiology, behaviour, predator-prey relationships, etc. As a Biophysics nerd, I was drawn to anemone toxins because they are a great tool for learning about ion channel biophysics.

In general, there are three main avenues for finding new anemone toxins: genomics, transcriptomics and proteomics. For genomics and transcriptomes, we extract the DNA and RNA respectively from the anemone's tissues for sequencing. For proteomic approach, you can just extract the protein and sequence it but there one major problem with this approach. The relative abundance of anemone toxins is much lower than all the other housekeeping proteins. A technique like MS/MS may not detect anemone toxins due to their lower relative abundance.

So we have to make our anemone smoothie and purify proteins with size exclusion or reverse phase chromatography, before trying to use protein sequencing for specific fractions. Or you use SDS PAGE for separating proteins by size. If you do MS/MS after isolating the anemone toxin, you have a much better chance of detecting and identifying the toxin.

Once we have our sequence, from the anemone transcriptome, genome or proteome we just dump it into BLAST. The gist of BLAST is that it uses multiple sequence alignments to see if there are other sequences (RNA, DNA or protein) in the database which are similar to our sequence. We kind of inferring what our new toxin is based on sequence similarity. The protein sequence dictates protein structure and function. So, if your new sequence is highly similar to a well-characterised anemone toxin, there is a decent chance that it will have a similar structure and function as the known toxin.

 Okay, so you have your BLAST output. What does it mean?

If you have read my 'Misadventures in undergrad research' post, you will know that the biggest challenge for me in my second project was figuring if my sequence was an actual anemone toxin.  

You know how to use this cool bioinformatics tool! Yay! Do you know how to interpret the data?

 Here are a few things to look out for:

 Conserved amino acids: Over time, as proteins evolve you’ll have some mutations. Sometimes, changing some minor amino acid is no big deal for the protein. However other amino acids are so important for folding, structure or function that the protein can't afford to lose them. These amino acids are conserved and you will see them being conserved in a variety of structurally or functionally similar proteins. For example, sometimes amino acids which are important for binding to potassium channel will be conserved from potassium channels toxins from different anemones.

Conserved cysteine residues: Cysteine residues are important for forming disulphide linkages and these linkages stabilise the entire 3-D structure. Sometimes ditching a disulphide linkage can impact the structure and function of a protein.

Conserved domains: Domains are functional or structural units which make up the protein. Just like conserved amino acids, conserved domains are important for function and structure. For anemone toxins, if you see that your protein has a conserved Kunitz domain, that tells you that your toxins probably acts as a serine protease inhibitor. It doesn't matter where you find your cytolysin, it could be a carpet anemone or a hellfire anemone, but almost all cytolysins will have a conserved cytotoxic domain because it required for penetrating the cell membrane.

Key residues which are different: Maybe the toxin is incredibly similar to a known toxin—expect for this one important amino acid. Sometimes minor differences in the amino acid sequence can have huge implications for the function. APETx1 and APETx2 are highly similar toxins from the same anemone, but they block completely different channels. APETx1 blocks potassium channels, whereas APETx2 block acid-sensing ion channels. It's all because of differences in a few key amino acids.

 Figure 1- The power of bioinformatics. Structure and MSA of two potassium channel toxins : ShK and BgK ( Jouiaei, M. et al, 2015) 

There is a minor problem with this approach because you can miss a lot of hidden gems. Some anemone toxins are so special, that you won't get a match with another anemone toxin. Maybe your hypothetical anemone toxin has similarity to spider or scorpion toxin. That doesn't mean that the anemone got bitten by a radioactive spider and now it makes spider toxins. Due to nature of evolution a lot of proteins with different functions in different organisms can similar motifs, binding sites and domains. Keep in mind that multiple sequence alignments only tell you what the toxin could be, not exactly what it is and what it does.For all you know the hypothetical protein does nothing.  It is important to synthesize the toxin and characterise the function.

You can go the microbial route and use E. coli to express your protein. Or you can do what I did as an undergrad: peptide synthesis. Peptide synthesis is more suitable synthesizing smaller toxins. Either way, you have to go through the process of purifying your protein with liquid chromatography. If you have the protein sequence, you can predict its isoelectric point, molecular weight and hydrophobicity. Once you have this information you can decide what's the best purification method e.g. ion exchange, size exclusion or reverse-phase.

So that's the gist of the first step to finding a novel toxin. Next time I will cover how to figure out the structure of anemone toxins.

Citations:

Anderluh, G. & Mac, P. Cytolytic peptide and protein toxins from sea anemones (Anthozoa : Actiniaria ). Toxicon 40, (2002).

Diochot, S., Baron, A., Rash, L. D., Deval, E., Escoubas, P., Scarzello, S., Lazdunski, M. (2004). A new sea anemone peptide, APETx2, inhibits ASIC3, a major acid-sensitive channel in sensory neurons. The EMBO journal, 23(7), 1516–1525. DOI:10.1038/sj.emboj.760017

Jouiaei, M. et al. Ancient venom systems: A review of cnidaria toxins. Toxins (Basel). 7, 2251–2271 (2015).

Norton, R. S. Sea Anemone Venom Peptides. Handbook of Biologically Active Peptides (Elsevier Inc., 2006). DOI:10.1016/B978-012369442-3/50056-8

Prentis, P. J., Pavasovic, A. & Norton, R. S. Sea anemones: Quiet achievers in the field of peptide toxins. Toxins (2018). doi:10.3390/toxins10010036

[lifesciencepotluck]

The most important relationship ever

Previously on Anemone Toxins 201:

Last time I covered how we discover novel anemone toxins, determine the structure and figure out the function. Now, everything comes together because we're looking at the most important relationship ever: the structure-function relationship.

We know the structure of our anemone toxin; we know what the toxin binds to—but how does an anemone toxin bind to its target? What kind of exciting interactions are happening between a toxin and its target? Protein structure and function are like soulmates: protein function relies on the structure, and protein structure adapts to a protein's functional needs. It's essential to understand the structure-function relationship of animal toxins, before developing therapeutic proteins. Once we understand which amino acids are critical for binding to the receptor, it's easier to tweak the toxin to improve its stability, specificity, and affinity.

I won't cover every single approach for exploring the structure-function relationship of anemone toxins, but I want to highlight some of the most common strategies. 

Chemical Modifications

It's funny how Biologists learn about living things by screwing things up and seeing what happens. Let's change a couple of critical amino acids in this protein. Let's knock out a possibly important gene. Let's expose to rat to radiation! What could possibly go wrong? The 'screw it up and see' approach is the basis for chemically modifying amino acids or mutating specific amino acids.

Amino acids aren't forever: they are prone to degrading and forming weird side products. For chemical modifications, we don't change the protein sequence. Instead, we use a reagent or stress condition to modify the side chains. For mutagenesis, we introduce substitutions: usually, the target amino acid is replaced with alanine, and then we express the protein. Turk et al. (1989), chemically modified arginine with 2,3 butane-dione and they modified tyrosine with tetranitromethane. They found tyrosine was important for Equinatoxin II's haemolytic activity. The toxin with the modified tyrosine residues had lower haemolytic activity compared to the native form. However, using chemical modifications gives very general information about the type of amino acid involved in binding. It's hard to determine which specific modified amino acid, directly binds to the receptor, does nothing, or maintains the structure.

Figure 1: 3-D structure of Equinatoxin II. Tyrosine side chains are in red (PDB ID: 1IAZ)

Mutagenesis

A lot of recent papers use site-directed mutagenesis to target one amino acid at a time: it's precise and also laborious. Moran et al. used mutagenesis to figure out which amino acids in the sodium channel toxin, Av2 were involved in binding to  the Nav1.5 channel. They found that six residues were vital for binding and toxicity: valine-2, leucine-5, aspartic acid-9, asparagine-16, leucine-18, and isoleucine-41. Mutagenesis is excellent for pinpointing specific amino acids in a toxin. However, it's harder to figure out what kinds of biophysical interactions are happening between the toxin and receptor. Are there hydrogen bonds? Are there electrostatic interactions? Are there hydrophobic interactions?

X-ray crystallography

Sometimes, we can crystallise the toxin-receptor complex. The 3-D for the complex gives information about toxin-receptor interactions at an atomic level. Celie et al. managed to get the 3-D structure of a-conotoxin from cone snails (PnIA) bound to the nicotinic acetylcholine receptor (AChBP). They found that when PnIA binds to AChBP, PnIA shifts a specific loop (the C-loop).But, there's no significant conformational change in AChBP. The crystallised complex also showed that hydrophobic interactions were the main driver for PnIA-AChBP binding.

Sometimes, crystallising the toxin-receptor complex is not feasible. If you have the crystal structures of the toxin and receptors separately, you can mash the two puzzle pieces together by using molecular docking. The gist of molecular docking involves figuring how the two partners in a complex are orientated together. Tools like PatchDock, use the complementary shapes of two partners to predict the docked complex. Lanigan et al. (2002) used a combination of mutagenesis and docking to figure out how an analogue for ShK (ShK-Dap22) binds to potassium channel (Kv1.3). They found that ShK and ShK-Dap22 bind to Kv1.3 channel in different ways. ShK has a conserved, positively lysine residue, which blocks the negatively charged pore in the potassium channel. However, ShK-Dap22 doesn't bind to the same site as ShK. In fact, ShK-Dap22 binds closer to the mouth of the Kv1.3 channel.

Figure 2: 3-D structure of a-conotoxin-AChBP. The bound PnIA are color coded as brick red (PDB ID: 2BR7)

NMR

The crystal structure of a protein is like a photograph which captures one conformation—but proteins are dynamic and their conformations are in flux. NMR is suitable for monitoring protein dynamics, since it can pick up on changes in the environment of amino acids, e.g. polarity, hydrogen bonding, orientation, etc. Sometimes, we want to see if there are conformational changes in a receptor after the toxin binds to the receptor. Are there any changes in NMR absorption for specific amino acids within the receptor, after binding? Lange et al. used solid-state NMR to explore how kalotoxin (KTX) from Fattail scorpions blocks Kv1.3 potassium channels. They found that both KTX and Kv1.3 went through conformational changes after binding. KTX binding impacted amino acids in the pore and the selectivity filter of the Kv1.3 channel.

Figure 3: General structure for potassium channels (Source: https://commons.wikimedia.org/wiki/File:Potassium_channel.jpg) 

All of the approaches I've covered are not only useful for exploring the structure-function relation of anemones toxins, but proteins in general. I have gushed about the therapeutic potential of anemone toxins in my posts. So, next time, I will cover a real-life example of how drugs are developed from an anemone toxins.

[drlucyalford]

In the latest University of Glasgow MCSB podcast, Anir Pandit and I got to talk to Professor Neil Bulleid who is the Institute Director and Professor of Cell Biology. Not only did I learn all about Neil’s exciting research into protein folding, but also that Neil has incredibly good taste in music, if I do say so myself!

[drlucyalford]

Anir Pandit and I have recently taken over running the podcast series at the University of Glasgow’s Institute of Molecular, Cell and Systems Biology. Our first guest was Professor Lynne Regan who holds a Chair in Interdisciplinary Science at the Centre for Synthetic & Systems Biology at the University of Edinburgh. Her research interests include protein-protein interactions and the ways in which the affinity and specificity of such interactions can be manipulated. I particularly enjoyed the quick fire question round where I got to find out more about Lynne’s interests and hobbies outside of science (and her love of The Rolling Stones. What can I say? Lynne most definitely has taste!)

[eclecctica]

This is it! My video for the final round of Chemistry Champions of the American Chemical Society. Learn about the Green Fluorescent Protein in less than three minutes and if you like it, please share!

[sarrahmorgan86-blog]

Also want these pretty website previews?

Current Protein & Peptide Science is a high impact factor journal and publishes reviewed articles on specific aspects involving proteins and peptides science.