Scramblase

Pickings have been slim lately, but here’s a great paper and a puzzle for you chemists out there. Most chemists (and biologists) know what a lipid bilayer is. It’s basically a soap bubble, with water loving (hydrophilic) groups on the outside of both sides of the bilayer, and hydrocarbon chains within. If the hydrocarbon chains are all stretched out the distance between carbons 1 and 3 is 2.66 Angstroms, and you have an 18 carbon fatty acid (stearic acid) it should be 8 * 2.66 + 1.33 Angstroms long (22.6 Angstroms). Double this for the bilayer and you have a thickness of 45 Angstroms. It’s probably less because carbon chains aren’t extended, partially because of entropy and largely because of cholesterol which breaks up any chance of such order (which maybe an important function for it). Sitting on either side of the lipid bilayer are phosphates esterified to one of the 3 hydroxyls of glycerol, with fatty acids of at least 16 – 18 carbons esterified to the other two. Hanging off the phosphates are a variety of things, but mostly serine and choline, forming phosphatidyl serine (PS) and phosphatidyl choline (PC). Here’s a picture — https://en.wikipedia.org/wiki/Lipid_bilayer.

Scramblases are enzymes which move phospholipids from one side of the lipid bilayer essentially randomizing their composition. They undo the action of other enzymes (called flippases believe it or not) which make the lipid composition of the two leaflets of the lipid bilayer rather different. This isn’t trivial, and is behind an elegant mechanism to show scavenger cells that a cell is dead. FLippases work to put phosphatidyl serine (PS) on the side of the lipid bilayer (the leaflet) facing the cytoplasm. This, of course takes energy, and when a cell lacks energy, entropy takes its course and PS appears on the outer leaflet, telling scavenger cells (phagocytes) to eat (phagocytose) the cell.

So how does an enzyme drag phosphatidyl choline (PC) https://en.wikipedia.org/wiki/Phosphatidylcholine or phosphatidyl serine (PC) across the lipid bilayer — scrambling the compositional asymmetry. Can you figure out a mechanism for a membrane protein to do this without looking at Proc. Natl. Acad. Sci. vol. 113 pp. 140149 – 14054 ’16? Chemists think they’re smart, and if you can design a protein to do this you’re smarter than I am because I’ve always wondered (ineffectually) how this was done for a long time.

The authors describe the structure of a fungal scramblase. It functions as a dimer with each subunit containing a hydrophilic groove containing polar and charged amino acid side chains facing the dimer interface. The protein itself does something unusual — it twists the sheet of the membrane, and decreases the thickness of the membrane from 29 to 18 Angstroms (remember the maximum possible thickness of the lipid bilayer was 45 Angstroms, but isn’t that thick for the reasons given above).

Phosphatidyl choline is a zwitterion (e.g. it contains both negative and positive charges although overall electrically neutral). The charges are separated in space forming a dipole. On the cytoplasmic side of the bilayer the scramblase has some amino acid side chains also forming a dipole, and right near the channel formed by the two hydrophilic grooves of the dimer. So it attracts the head group of PC (phosphate plus choline) as one dipole does to another which is then further attracted to the hydrophilic groove entering it — its hydrocarbon tail remains in the lipid part of the membrane. Then another PC joins the fun, pushing PC #1 farther into the groove, so that a chain of PCs fills the groove, wagging their lipid tails behind them (a la Little Bo Peep).

Clever no?

All is not perfect as the model doesn’t explain how phosphatidyl serine (which isn’t a zwitterion) moves across, but it’s an incredible start.

Not physical organic chemistry but organic physical chemistry

This post is about physical chemistry with organic characteristics in the sense that capitalism in China is called socialism with Chinese characteristics. A lot of cell biology is also involved.

I remember the first time I heard about Irving Langmuir and the two dimensional gas he created. It even followed a modified perfect gas law (PA = nRT where A is area). He did this by making a monolayer of long chain fatty acids on water, with the carboxyl groups binding to the water, and the hydrocarbon side chain sticking up into the air. I thought this was incredibly neat. It was the first example of organic physical chemistry. He published his work in 1917 and won the Nobel in Chemistry for it in 1932.

Fast forward to our understanding of the membrane encasing our cells (the technical term is plasma membrane to distinguish from the myriad other membranes inside our cells. To a first approximation it’s just two Langmuir films back to back with the hydrocarbon chains of the lipids dissolving in each other, and the hydrophilic parts of the membrane lipids binding to the water on either side. This is why it’s called a lipid bilayer.

Most of the signals going into our cells must pass through the plasma membrane, using proteins spanning it. As a neurologist I spent a lot of time throwing drugs at them — examples include every known receptor for neurotransmitters, reuptake proteins for them (think the dopamine transporter), ion channels. The list goes on and on and includes the over 800 G protein coupled receptors (GPCRs) with their 7 transmembrane segments we have in our genome [ Proc. Natl. Acad. Sci. vol. 111 pp. 1825 – 1830 ’14 ].

Glypiated proteins (you heard right) also known as PIGtailed proteins (you heard that right too) don’t follow this pattern. They are proteins anchored in the outer leaflet of the plasma membrane lipid bilayer by covalently linked phosphatidyl inositol. https://en.wikipedia.org/wiki/Phosphatidylinositol — the picture shows you why — inositol is a sugar, hence crawling with hydroxyl groups, while the phosphatidic acid part has two long hydrocarbon chains which can embed in the outer leaflet. We have 150 of them as of 2009 (probably more now). Examples of PIGtailed proteins include alkaline phosphatase, Thy-1 antigen, acetyl cholinesterase, lipoprotein lipase, and decay accelerating factor. So most of them are enzymes working on stuff outside the cell, so they don’t need to signal.

Enter the lipid raft. [ Cell vol. 161 pp. 433 – 434, 581 – 594 ’15 ] It’s been 18 years since rafts were first proposed, and their existence is still controversial (with zillions of papers saying they exist and more zillions saying they don’t). What are they — definitions vary (particularly about how large they are). Here’s what Molecular Biology of the Cell 4th edition p. 589 had to say about them — Rafts are small (700 Angstroms in diameter). Rafts are rich in sphingolipids, glycolipids and cholesterol. The hydrocarbon chains are longer and straighter than those of most membrane lipids, rafts are thicker than other parts of bilayer. This allows them to better accomodate ‘certain’ membrane proteins, which accumulate there. [ Proc. Natl. Acad. Sci. vol. 100 p. 8055 – 7’03 ] These include glycosylphosphatidylinositol anchored proteins (glypiated proteins), cholesterol linked and palmitoylated proteins such as Hedgehog, Src family kinases and the alpha subunits of G proteins, cytokine receptors and integrins.

Biochemical analysis shows that rafts consist of cholesterol and sphingolipids in the exoplasmic leaflet (outer layer of the plasma membrane) of the lipid bilayer and cholesterol and phospholipids with saturated fatty acids in the endoplasmic leaflet (layer facing the cytoplasm). The raft is less fluid than surrounding areas of the membrane. So if they in fact exist, rafts contain a lot of important cellular players.

The Cell paper introduced synthetic fluorescent glypiated proteins into the outer plasma membrane leaflet of Chinese hamster ovary cells and was able to demonstrate nanoClustering on scales under 1,000 Angstroms (way too small to see with visible light, accounting for a lot of the controversy concerning their existence).

How can the authors make such a statement? The evidence was a decrease in fluorescence anisotropy due to Forster resonance energy transfer effects. Forster energy transfer is interesting in that it doesn’t involve molecule #1 losing energy by emitting a photon which is absorbed by molecule #2 increasing its energy. It works by molecule #1 inducing a dipole in molecule #2 (by a Van der Waals effect). Obviously, to do this, the molecules must be fairly close, and transfer efficiency falls off as the inverse 6th power of the distance between the two molecules.

In Fluorescence Resonance Energy Transfer (FRET), one fluorophore (the donor) transfers its excited state energy to a different fluorophore (the acceptor) which emits fluorescence of a different color. For more details see — https://en.wikipedia.org/wiki/Förster_resonance_energy_transfer — its interesting stuff. Again an example of physical chemistry with organic characteristics (and pretty good evidence for the existence of lipid rafts to boot).

Now it gets even more interesting. Nanoclustering is dependent on the length of the acyl chain forming the GPI anchor (at least 18 carbons must be present). NanoClustering diminishes on cholesterol depletion in actin depleted cell blebs and mutant cell lines deficient in the inner leaflet lipid — phosphatidylserine (PS) — which has two long chain fatty acids hanging off the glycerol. So it looks as if the saturated acyl chains of the glypiated proteins of the outer leaflet interdigitate with those of PS in inner leaflet. The effect is also enhanced on expression of proteins specifically linking PS to the actin cytoskeleton of the cortex. Binding of PS to the cortical actin cytoskeleton determines where and when the clusters will be stabilized. The coupling can work both ways — if something immobilizes and stabilizes the glypiated proteins extracellularly, than PS lipids can form correlated patches.

This might be a mechanism for information transfer across the plasma membrane (and acrossother membranes to boot). This could also serve as a way to couple many outer leaflet membrane lipids such as gangliosides and other sphingolipids with events internal to the cell. Cholesterol can stabilize the local liquid ordered domain over a length scale that is large than the size of the immobilized cluster. A variety of membrane associated proteins inside the cell (spectrin, talin, caldesmon) are able to bind actin. “The formation of the contractile actin clusters then determine when and where the domains may be stabilized, bringing the generation of membrane domains in live cells under control of the actomyosin signaling network.’

So just like the integrins which can signal from outside the cell to inside and from inside the cell to outside, glypiated proteins and the actin cytoskeleton may form a two way network for signaling. No one should have to tell you how important the actomyosin cytoskeleton is in just about everything the cell does. Truly fascinating stuff. Stay tuned.

Curioser and curioser

Curious Wavefunction alluded to the first example of a protein which stands everything we thought we knew about them on its head. At the end of this post you’ll find another equally counterintuitive example.

We all know that proteins fold into a relatively dry core where hydrocarbon side chains and other hydrophobic elements hide out. This was one of Walter Kauzmann’s many contributions to chemistry and biology. He also wrote one of the first books on quantum chemistry, as did his PhD advisor Henry Eyring at Princeton (I was lucky enough to take PChem from him). The driving force for the formation of globular proteins according to him, was pretty much entropic, with hydrocarbon side chains solvating each other so water wouldn’t have to form an elaborate (hence structured) cage to do so.

Which brings us to the wonderfully named fish Pseudopleuronectes Americanus which lives in frigid polar waters. To keep ice crystals from forming in their cells, arctic fish have evolved proteins to prevent it. It is a fascinating example of evolution solving a problem different ways, because by 1996 at least 4 different types of antifreeze proteins were known [ PNAS vol. 93 pp. 6835 – 6840 ’96 ].

The new protein is a 3 kiloDalton alanine rich helix bundle 145 Angstroms long.
Amazingly the helices surround a core of 400 water molecules (surround as in the water is on the inside of the protein, not the outside). The water molecules inside the protein are arranged as pentagons (not hexagons as they would be in ice) — so they form a clathrate. The pentagonal arrangement of water was predicted on theoretical grounds 50 years ago by Scheraga ( J. Biol. Chem. vol. ?? pp. 2506 – 2508 1962 ).

The protein has an amino acid periodicity of 11 amino acids, which nicely comes out to 3 turns of the alpha helix. There is a threonine at position i, alanine at position i + 4 and alanine a position i + 8. All of these bind water — not surprising for threonine, but alanine is a hydrocarbon. The evolving fish clearly didn’t listen to protein chemists. However, most of carbonyl groups of the protein backbone are involved in hydrogen bonding to water.

Not to be outdone, a freeze tolerant beetle (Upis cermaboides — don’t you love these names) has an antifreeze molecule made mostly of sugar and lipid.

Well even if we don’t know what we thought we knew about proteins, at least we understand biologic membranes and the proteins that go through them. Don’t we?

Apparently not. [ Proc. Natl. Acad. Sci. vol. 111 pp. 2425 – 2430 ’14 ] studied the alpha-hemolysin of staphylococci. We know that the membrane of our cells is made of a double layer of molecules which a charged head which binds water and a long (16 + carbons) hydrocarbon tail. So the hydrocarbon core is 30 Angstroms across, and the lipid head groups are about 40 Angstroms away from each other on either side of the membrane.

We also know how proteins fit into the membrane — one model is the G Protein Coupled Receptor (GPCR) for which we have at least 800 human genes, and which is the target for 30% of all drugs approved by the FDA [ Science vol. 335 pp. 1106 – 1110 ’12 ]. These all have 7 alpha helices arranged like a stack of logs extending across the membrane. The amino acids here are usually hydrophobic. Another model is the beta barrel — used mostly by bacteria — these have beta strands arranged across the membrane (like the staves of a barrel — get it). I’m not sure what the record is for the number of strands, but one from the gonococcus has 16 of them. They surround a large pore.

Back to the alpha hemolysin of staphylococci It’s designed to kill its target by forming a hole in the membrane. And so 7 of them get together to do so. However, instead of the running back and forth across the 30 Angstroms of the anhydrous part of the membrane, the heptamers put their heads together forming the hole (like skydivers holding hands), with their hydrocarbon like parts sticking out into the membrane and the water filled hole in the center. How do they know? They studied truncated mutants of the hemolysin, which weren’t long enough to span the 30 Angstroms across the membrane, and they still formed holes. An entirely new (to me) protein arrangement.