Tag Archives: moonlighting effect

Why drug discovery is so hard: Reason #27 Moonlighting effects.

Well, we all know what heat shock proteins (Hsps) do — they bind to proteins which have lost their shape due to heat (or other stressors), cuddle them hydrolyze ATP and nurse them back to health. But what  if some of them do other things? The phenomenon is called moonlighting.

The case of Hsp70 is instructive. Some background first. The Hsp70 chaperone transiently associates with its substrates in a manner controlled by its ATPase cycle. ATP binding to the amino terminal nucleotide binding domain (NBD) induces a conformational change in the carboxy terminal substrate binding domain (SBD) which results in low affinity for substrate. Hydrolysis of ATP converts the Hsp70 to the ADP state, which binds substrates with higher affinity. Exchange of ADP for ATP releases substrate completing the cycle. The hydrolysis of ATP is stimulated by J-domain containing cochaperones. These are the nucleotide exchange factors.  Back and forth Hsp70 and the damaged protein go through the cycle until the protein is nursed back to normal or, failing this, is destroyed.

The Hsp70 family acts early in protein synthesis by binding to a small stretch of hydrophobic amino acids on a protein’s surface. Aided by a set of smaller Hsp40 proteins (also known as J proteins), a hsp70 monomer binds to its target protein and then hydrolyzes ATP to ADP, undergoing a conformational change that causes the hsp70 to clamp down very tightly on the target. After the hsp40 dissociates (see below), the dissociation of the hsp70 protein is induced by the rapid rebinding of ATP after ADP release. Repeated cycles of hsp protein binding and release help the target protein to refold.

Enter [ Proc. Natl. Acad. Sci. vol. 112 pp. E3327 – E3336 ’15 ] This work shows Hsp70 is methylated on arginine #469 by Coactivator Associated aRginine Methyltransferase 1/Protein aRginine MethylTransferase 4 (CARM1/PRMT4) and demethylated by JuMonJi Domain containing 6 (JMJD6) — hideous acronyms shortening even more hideous names. Methylated Hsp70 then functions in transcription as a ‘regulator’ of Retinoid Acid Receptor beta 2 (RARbeta2) transcriptional acitivty. R468Mmethylated Hsp70 mediates the interaction between Hsp70 and TFIIH (Transcription Factor IIH).

The regulation of gene transcription is an entirely novel and unsuspected function for a heat shock protein. A classic example of moonlighting.

Drug chemists and pharmacologists are always concerned about off-target effects. For an interesting example please see https://luysii.wordpress.com/2011/02/02/medicinal-chemists-do-you-know-where-your-drug-is-and-what-it-is-doing/.  Off-target effects occur when their drug hits something else in the cell producing an unexpected (and usually untoward) effect.

If you are unaware that your target of choice is doing a little something else on the side (e.g. moonlighting) you can get an off target effect even when you hit your desired target. It’s a tough business. How many more moonlighters are out there that we don’t know about?

Hsp70 is a good example. Here are two more — no background provided, so you’re on your own — except to point out that glucocorticoids are a widely used class of drug.

[ Proc. Natl. Acad. Sci. vol. 112 pp. E1540 – 1549 ’15 ] Amazingly, the glucocorticoid receptor (GR)plays a role in mRNA degradation by acting as an RNA binding protein. When loaded onto the 5′ UnTranslated Region (5′ UTR) of a target mRNA, the GR recruits UPF1 through Proline-rich Nuclear Receptor Coregulatory protein 2 (PNRC2) in a ligand (of itself?) dependent manner to cuase rapid mRNA degradation. They call this GMD (Glurocorticoid receptor Mediated Decay). Along with Staufen Mediated mRNA Decay (SMD) and Nonsense Mediated mRNA Decay (NMD), they share UPF1 (Upstream Frameshift 1) and PNRC2.

[ Science vol. 323 pp. 723 – 724, 793 – 797 ’09 ] Stat3 proteins represent the canonical mediators of signals elicited by cytokines binding to type I cytokine receptors. However, GRIM19 (Gene associated with Retinoid Interferon Mortality 19), a mitochondrial protein, interacts with Stat3 and inhibits its transcriptional activity (where?). This work shows that Stat3 associates with GRIM19 containing complexes I and II (components of the electron transport chain) in mouse liver and muscle mitochondria. Levels of Stat3 in mitochondria are 10% of cytosolic levels.

Cells lacking Stat3 show decreased activity of mitochondrial complexes I and II. Effects on complex I and II don’t require Stat3’s DNA binding domain, the dimerization motif, or the tyrosine phosphorylation site controlling Stat3 nuclear localization and transcriptional activity — so this is a ‘moonlighting’ role for State3 having nothing to do with gene transcription. The serine phosporylation site on Stat3 is important. So Stat3 is required to maintain normal mitochondrial function.


When the active form of a protein is intrinsically disordered

Back in the day, biochemists talked about the shape of a protein, influenced by the spectacular pictures produced by Xray crystallography. Now, of course, we know that a protein has multiple conformations in the cell. I still find it miraculous that the proteins making us up have only relatively few. For details see — https://luysii.wordpress.com/2010/08/04/why-should-a-protein-have-just-one-shape-or-any-shape-for-that-matter/.

Presently, we also know that many proteins contain segments which are intrinsically disordered (e.g. no single shape).The pendulum has swung the other way — “estimations that contiguous regions longer than 50 amino acids ‘may be present” in ‘up to’ 50% of proteins coded in eukaryotic genomes [ Proc. Natl. Acad. Sci. vol. 102 pp. 17002 – 17007 ’05 ]

[ Science vol. 325 pp. 1635 – 1636 ’09 ] Compared to ordered regions, disordered regions of proteins have evolved rapidly, contain many short linear motifs that mediate protein/protein interactions, and have numerous phosphorylation sites compared to ordered regions. Disordered regions are enriched in serine and threonine residues, while ordered sequences are enriched in tyrosines — this highlights functional differences in the types of phosphorylation. Interestingly tyrosines have been lost during evolution.

What are unstructured protein segments good for? One theory is that the disordered segment can adopt different conformations to bind to different partners — this is the moonlighting effect. Then there is the fly casting mechanism — by being disordered (hence extended rather than compact) such proteins can flail about and find partners more easily.

Given what we know about enzyme function (and by inference protein function), it is logical to assume that the structured form of a protein which can be unstructured is the functional form.

Not so according to this recent example [ Nature vol. 519 pp. 106 – 109 ’15 ]. 4EBP2 is a protein involved in the control of protein synthesis. It binds to another protein also involved in synthesis (eIF4E) to suppress a form of translation of mRNA into protein (cap dependent translation if you must know). 4EBP2 is intrinsically disordered. When it binds to its target it undergoes a disorder to ordered transition. However eIF4E binding only occurs from the intrinsically disordered form.

Control of 4EBP2 activity is due, in part, to phosphorylation on multiple sites. This induces folding of amino acids #18 – #62 into a 4 stranded beta domain which sequesters the canonical YXXXLphi motif with which 4EBP2 binds eIF4E (Y stands for tyrosine, X for any amino acid, L for leucine and phi for any bulky hydrophobic amino acid). So here we have an inactive (e.g. nonbonding) form of a protein being the structured rather than the unstructured form. The unstructured form of 4EBP2 is therefore the physiologically active form of the protein.