Cells are not bags of cytoplasm

How Ya Gonna Keep ’em Down on the Farm (After They’ve Seen Paree) is a song of 100+ years ago when World War I had just ended. In 1920, for the first time America was 50/50 urban/rural. Now it’s 82%.

What does this have to do with cellular biology? A lot. One of the first second messengers to be discovered was cyclic adenosine monophosphate (CAMP). It binds to an enzyme complex called protein kinase A (PKA), activating it, making it phosphorylate all sorts of proteins changing their activity. But PKA doesn’t float free in the cell. We have some 47 genes for proteins (called AKAPs for protein A Kinase Anchoring Protein) which bind PKA and localize it to various places in the cell. CAMP is made by an enzyme called adenyl cyclase of which we have 10 types, each localized to various places in the cell (because most of them are membrane embedded). We also have hundreds of G Protein Coupled Receptors (GPCRs) localized in various parts of the cell (apical, basal, primary cilia, adhesion structures etc. etc.) many of which when activated stimulate (by yet another complicated mechanism) adenyl cyclase to make CAMP.

So the cell tries to keep CAMP when it is formed relatively localized (down on the farm if you will). Why have all these ways of making it if its going to run all over the cell after all.

Actually the existence of localized signaling by CAMP is rather controversial, particularly when you can measure how fast it is moving around. All studies previous to Cell vol. 182 pp. 1379 – 1381, 1519 – 1530 ’20 found free diffusion of CAMP.

This study, found that CAMP (in low concentrations) was essentially immobile, remaining down on the farm where it was formed.

The authors used a fluorescent analog of CAMP which allowed them to use fluorescence fluctuation spectroscopy which gives the probability distribution function of an individual molecule occupying a given position in space and time (SpatioTemporal Image correlation Spectroscopy — STICS).

Fascinating as the study is, it is ligh tyears away from physiologic — the fluorescent CAMP analog was not formed by anything resembling a physiologic mechanism (e.g. by adenyl cyclase). A precursor to the fluorescent CAMP was injected into the cell and broken down by ‘intracellular esterases’ to form the fluorescent CAMP analog.

Then the authors constructed a protein made of three parts (1) a phosphodiesterase (PDE) which broke down the fluorescent CAMP analog and (2) another protein — the signaler — which fluoresced when it bound the CAMP analog. The two were connected by (3) a flexible protein linker e.g. the ‘ruler’ of the previous post. The ruler could be made of various lengths.

Then levels of fluorescent CAMP were obtained by injecting it into the cell, or stimulating a receptor.

If the sensor was 100 Angstroms away from the PDE, it never showed signs of CAMP, implying the the PDE was destroying it before it could get to the linker implying that diffusion was quite slow. This was at low concentrations of the fluorescent CAMP analog. At high injection concentrations the CAMP overcame the sites which were binding it in the cell and moved past the signaler.

It was a lot of work but it convincingly (to me) showed that CAMP doesn’t move freely in the cell unless it is of such high concentration that it overcomes the binding sites available to it.

They made another molecule containing (1) protein kinase A (2) a ruler (3) a phophodiesterase. If the kinase and phosphodiesterase were close enough together, CAMP never got to PKA at all.

Another proof that phosphodiesterase enzymes can create a zone where there is no free CAMP (although there is still some bound to proteins).

Hard stuff (to explain) but nonetheless impressive, and shows why we must consider the cell a bunch of tiny principalities jealously guarding their turf like medieval city states.

*****

A molecular ruler

Time to cleanse your mind by leaving the contentious world of social issues and entering the realm of pure thought with some elegant chemistry. 

You are asked to construct a molecular ruler with a persistence length of 150 Angstroms. 

Hint #1: use a protein

Hint #2; use alpha helices

Spoiler alert — nature got there first. 

The ruler was constructed and used in an interesting paper on CAMP nanoDomains (about which more on the next post).

It’s been around since 2011 [ Proc. Natl. Acad. Sci. vol. 108 pp. 20467 – 20472 ’11 ] and I’m embarrassed to admit I’d never heard of it.

It’s basically a run of 4 negatively charged amino acids (glutamic acid or aspartic acid) followed by a run of 4 positively charged amino acids (lysine, arginine). This is a naturally occurring motif found in a variety of species. 

My initial (incorrect) thought was that this couldn’t work as the 4 positively charged amino acids would bend at the end and bind to the 4 negatively charged ones. This can’t work even if you make the peptide chain planar, as the positive charges would alternate sides on the planar peptide backbone.

Recall that there are 3.5 amino acids/turn of the alpha helix, meaning that between a run of 4 Glutamic acid/Aspartic acids and an adjacent run of 4 lysines/arginines, an ionic bond is certain to form between the side chains (and not between adjacent amino acids on the backbone, but probably one 3 or 4 amino acids away)

Since a complete turn of the alpha helix is only 5.4 Angstroms, a persistence length of 150 means about 28 turns of the helix using 28 * 3.5 = 98 amino acids or about 12 blocks of ++++—- charged amino acids. 

The beauty of the technique is that by starting with an 8 amino acid ++++—- block, you can add length to your ruler in 12 Angstrom increments. This is exactly what Cell vol. 182 pp. 1519 – 1530 ’20 did. But that’s for the next post. 

A molecular ruler

Time to cleanse your mind by leaving the contentious world of social issues and entering the realm of pure thought with some elegant chemistry.

You are asked to construct a molecular ruler with a persistence length of 150 Angstroms.

Hint #1: use a protein

Hint #2; use alpha helices

Spoiler alert — nature got there first.

The ruler was constructed and used in an interesting paper on CAMP nanoDomains (about which more on the next post).

It’s been around since 2011 [ Proc. Natl. Acad. Sci. vol. 108 pp. 20467 – 20472 ’11 ] and I’m embarrassed to admit I’d never heard of it.

It’s basically a run of 4 negatively charged amino acids (glutamic acid or aspartic acid) followed by a run of 4 positively charged amino acids (lysine, arginine). This is a naturally occurring motif found in a variety of species.

My initial (incorrect) thought was that this couldn’t work as the 4 positively charged amino acids would bend at the end and bind to the 4 negatively charged ones. This can’t work even if you make the peptide chain planar, as the positive charges would alternate sides on the planar peptide backbone.

Recall that there are 3.5 amino acids/turn of the alpha helix, meaning that between a run of 4 Glutamic acid/Aspartic acids and an adjacent run of 4 lysines/arginines, an ionic bond is certain to form between the side chains (and not between adjacent amino acids on the backbone, but probably one 3 or 4 amino acids away)

Since a complete turn of the alpha helix is only 5.4 Angstroms, a persistence length of 150 means about 28 turns of the helix using 28 * 3.5 = 98 amino acids or about 12 blocks of ++++—- charged amino acids.

The beauty of the technique is that by starting with an 8 amino acid ++++—- block, you can add length to your ruler in 12 Angstrom increments. This is exactly what Cell vol. 182 pp. 1519 – 1530 ’20 did. But that’s for the next post.

How a chemical measuring stick actually works

The immune system knows something is up when a foreign peptide fragment is presented to it.  Here’s the hand holding the peptide — https://www.researchgate.net/figure/Overall-structure-of-HLA-peptide-complex_fig1_26490512.

There it sits, lying on top of a bed of beta sheets, with two side rails of alpha helices.  Proteins are big, way too big to fit into the hand, so the fragments must be chopped up into peptides no longer than 9 amino acids long (see the picture of it lying in state).

So the class assignment for today is to figure out how to design a protein which takes peptides from 10 – 16 amino acids long, and shortens them to 9 amino acids.

Obviously a trick question, because the actual amino acids making up the peptide don’t really matter much.  So somehow the protein is reacting to length rather than chemistry.

Tricky no?

ERAP1 (Endoplasmic Reticulum aminopeptidase associated with Antigen Processing has figured it out [ Proc. Natl. Acad. Sci. vol. 116 pp. 22709 – 22715 ’19 ].  It is a huge protein (948 amino acids) with four domains forming a large cavity (which it must have to accomodate a 19 amino acid paptide).  The peptide is chopped up from the amino terminal, stopping when the length reaches 9 amino acids.  The active site is at one end of the cavity, and at the other end there is a site which looks like it should cleave the carboxyterminal amino acid, but it doesn’t because the site is inactive.  However, even catalytically inactive enzymatic sites have enough structure left so they bind the substrate.

So binding of the carboxy terminal amino acid to the back site causes conformational changes transmitted through various alpha helices to the active enzyme at the other end.  It munches away removing amino acid after amino acid until the peptide gets short enough (translation 9 amino acids) so that it doesn’t push on the back site.

Incredibly clever, even though it hurts me as a chemist to see the enzyme essentially ignoring the chemistry of its substrate.

I far prefer this to politics where data is ignored.  Two examples

l. From a review of a book by Paul Krugman in the Jan/Feb 2020 Atlantic

“Krugman is substantively correct on just about every topic he addresses.” Yes except Peak Oil in 2010, Stock Market collapse in Nov 2016 and the coming recession in an article April 2019

2. Former Secretary of Labor Robert Reich in the Guardian 22 Dec ’19 — “How Trump has betrayed the working class” — by employing them and raising their wages no doubt.

Of what use is an inactive enzyme?

Why should a cell take the trouble make an enzyme protein with no enzymatic activity? It takes metabolic energy to store the information for a protein in DNA, transcribe the DNA into RNA and then translate the RNA into protein. Is this junk protein a la junk DNA? Not at all — and therein lies a tale.

All sorts of nasty bugs inveigle their way into cells, among them viruses (such as influenza) whose genome is made of RNA, rather than DNA. Not only that, but in many virus their genome is not single stranded (like mRNA) but double stranded with two RNA strands base paired to each other (just like DNA, except for an extra oxygen on the ribose sugars in the backbone).

Nucleated cells don’t contain much double stranded RNA (dsRNA) outside the nucleus, so it almost always means trouble. An extremely elegant mechanism exists to find and respond to such RNA. Recall that double helix molecules can reach enormous lengths.The 3.2 billion base pairs of our genome, if stretched out, would be more than a yard.

Well we have at least 4 genes which bind dsRNA and then signal trouble. They all make a molecule called 2′ – 5′ oligoadenyic acid (2-5A) from ATP, so they are called OligoAdenylate Syntheses (OASs). The 2-5A, once made wanders about the cell until it finds another enzyme called RNAase L. 2-5A binds to RNAase L causing it to dimerize and become active. RNAase L then destroys all the RNA in the cell, killing it along with the invading virus. Pretty harsh, but it’s one way to stop the virus from spreading and killing more cells.

A recent paper http://www.pnas.org/content/112/13/3949.full concerns OAS3, which has 3 catalytic modules rather than just one like most enzymes. Even worse, 2 of the 3 catalytic modules can’t make 2-5A (but they still can bind dsRNA). OAS3 is a large protein (over 1,000 amino acids), so it has some length to it. The 3 catalytic modules are spread out along OAS3 with the active catalytic module at one end and one of the inactive modules at the other.

The modules at both ends bind dsRNA, but only the active module makes 2-5A when it does. Interestingly, the inactive module binds dsRNA much more strongly than the active one.

OK, you’ve got the picture — what possible use is this rather Byzantine set up?

See if you can figure it out.

It’s incredibly clever and elegant, and shows the danger to regarding anything within the cell as functionless (or junk). Teleology rides supreme in molecular and cellular biology.

Give up?

OAS3 essentially acts as a molecular ruler making 2-5A only when long dsRNA (e.g. over 50 nucleotides long) binds to it. The inactive module gloms onto longish dsRNA, holding it tightly until till Brownian motion brings it to the other end of OAS3 activating the catalytic module to make 2-5A. This is good as the cell normally contains all sorts of shorter RNA duplexes (the binding of microRNAs to the 3′ end of mRNAs come to mind — but they are much shorter (22 nucleotides at most).