Category Archives: Chemistry (relatively pure)

Lactic acid, the mitotic spindle killer

Nature vol. 616 pp. 790 – 797 ’23 is one of the most interesting papers I’ve read in the past year, both for its contents and for the two very large issues it raises (which the authors don’t really discuss).

Simply stated, the rise in cellular lactic acid levels from 6  milliMolar at mitosis onset, to 15 – 20 when mitosis is nearly over is what ’causes’ the breakdown of the mitotic spindle.

It’s now 100 years since Otto Warburg noted that tumors metabolize glucose by glycolysis producing 2 molecules of ATP per glucose (and two molecules of lactic acid) when, with plenty of oxygen around, they could get 38 molecules of ATP using their mitochondria.   This is called aerobic glycolysis.

Tumors are said to be energy hungry, so why do they use aerobic glycolysis? Simply because using oxygen to chew up glucose gives you lots of ATP along with CO2 and water, leaving you nothing to build new tumor cells with.  All 6 carbons remain present after glycolysis

The last stage of mitosis is called anaphase, where the mitotic spindle (made of microtubules) is broken down, among other things such as reformation of the nuclear membrane, and separation of the two daughter cells.

Well protein breakdown immediately brings ubiquitin to mind which, when added to most proteins, targets them to the proteasome, a huge molecular complex which breaks proteins down completely to their constituent amino acids.

APC/C is another huge multiprotein complex (at least 13 different protein subunits with a molecular mass of 1.2 megaDaltons) which acts to add ubiquitin to components of the mitotic  spindle (made mostly of microtubules).  So APC/C is a ubiquitin ligase, a dangerous thing to have around most of the time, which it is why it is usually inhibited so the cell doesn’t destroy itself.

One APC/C subunit is APC4, which has ubiquitinLike molecules (SUMO) attached to two of its lysines (#722 and #798) to activate the ubiquitin ligase activity of APC/C.    APC4 is held in check by yet another enzyme, SENP1, which removes the SUMOs.

Where does lactic acid fit in to all this?  It binds to the active site of SENP1 when coordinated with zinc ions, inhibiting SENP1’s ability to remove SUMO.

Byzantine enough for you?  Lactic acid inhibits SENP1 which inhibits APC4 allowing uninhibited APC4 to activate APC/C which breaks down the mitotic spindle.

Lactic acid, if thought of at all, was regarded as an important part of cellular metabolism, not an enzyme inhibitor.   This is an example of moonlighting, a lot of which goes on in the cell.  with its links will get you started.

Here is one of the larger issues the paper raises — how events in the cell at all levels of structure are linked to each other.  Phillip Anderson famously said “More is Different”.  The paper shows how something very small (lactic acid fits into a 5 Angstrom (.5 nanoMeters) sphere) and yet  is responsible for breaking down something 40,000 – 100,000 times larger  (the length of a microtubule in the mitotic spindle).

Here is the other (even larger) issue — Lactic acid was found as a player in cell metabolism, e.g., it is a member of the metabolome.  I was amazed to find out how large it is — some 42,000 for in the Human Metabolome DataBase — for details please see  Not only do we not know what they are doing, we don’t even know the structure of most of them. State of the art untargeted metabolomics studies still report ‘up to’ 40% unidentified, but potentially important metabolitcs which can be detected reproducibly. The unknown metabolites are only rarely characterized because of the extensive work required for de novo structure determination..

The science behind Cassava Sciences (SAVA) — the latest as of 23 April ’23

It’s time for an update on the science  behind Cassava Sciences’ anti-Alzheimer drug, Simufilam.  It is  based on an older post of mine and a review of the published literature and my decades of experience as a clinical neurologist.

Disclaimer:  My wife and I have known Lindsay Burns, one of the Cassava Sciences principals since she was a teenager and we were friendly with her parents when I practiced neurology in Montana.

But as H. L. Mencken said, “A Professor must have a theory as a dog must have fleas”, and the reason I’m excited about Simufilam has nothing to do with the theory of the science behind it.  Simply put, the results of Cassava’s open label trial have never  been seen with Alzheimer’s patients.  10% improved by nearly 50% at 1 year, and over half did not deteriorate.  As a clinical neurologist with decades of experience seeing hundreds of demented people, I never saw anything like this, especially significant improvement after a year).  For more detail please see

Here is the science behind the drug.  We’ll start with the protein the drug is supposed to affect — filamin A, a very large protein (2,603 amino acids to be exact).  I’ve known about it for years because it crosslinks actin in muscle, and I read everything I could about it, starting back in the day when I ran a muscular dystrophy clinic in Montana.

Filamin binds actin by its amino terminal domain.  It forms a dimerization domain at its carboxy terminal end.  In between are 23 repeats of 96 amino acids which resemble immunoglobulin — forming a rod 800 Angstroms long.  The dimer forms a V with the actin binding domain at the two tips of the V, making it clear how it could link actin filaments together.

Immunoglobulins are good at binding things and 90 different proteins are known to which filamin A binds.  This is an enormous potential source of trouble.

As one might imagine, filamin A could have a lot of conformations in addition to the V, and the pictures shown in

One such altered (from the V) conformation binds to the alpha7 nicotinic cholinergic receptor on the surface of neurons and Toll-Like Receptor 4 (TLR4) inside the cell.

Abeta42, the toxic peptide, has been known for years to bind tightly to the alpha7 nicotinic receptor — they say in the femtoMolar (10^-15 Molar) range, although I have my doubts as to whether such tiny concentration values are meaningful.  Let’s just say the binding is tight and that femtoMolar binding is tighter than picoMolar is tighter than nanoMolar is tighter than microMolar  binding etc., etc.

When aBeta42 binds to alpha7 on the outside of the neuronal plasma membrane  filamin A binds to alpha 7 on the inside making  aBeta42 binding even tighter.

The tight binding causes signaling inside the cell  to hyperphosphorylate the tau protein forming the neurofibrillary tangle, which is more directly correlated with dementia in Alzheimer’s disease than the number of senile plaques.

In more detail, the high affinity aBeta42-alpha7 nicotinic cholinergic receptor binding activates the MAPK cascade (Mitogen Activated Protein Kinase cascade), ending in activation of the protein kinases ERK2, and JNK1.  Activated protein kinases catalyze the addition of phosphate to proteins forming an ester with the free hydroxyl groups of serine and/or threonine.  Activating ERK2 and JNK1 allows them to phosphorylate the tau protein leading to the neurofibrillary tangle of  Alzheimer’s disease (which is just a mess of hyperphosphorylated tau protein).

But there is still more about the mechanism which isn’t clear.  Recall that MAPK stands for Mitogen Activated Protein Kinase where a mitogen binds to a receptor on the cell surface, and a mitogen is nowhere in sight here, so there are still a few missing steps between aBeta42 binding to the alpha7 nicotinic cholinergic receptor and MAPK activation.  The references do show that MAPK signaling, ERK2 and JNK1 are activated when aBeta42 binds to the alpha7 nicotinic acetyl choline receptor.

Also the mechanism is radical in the extreme. The nicotinic acetyl choline receptor is a receptor all right but for acetyl choline. It is an ion channel and   looks nothing like the receptors that proteins and peptides bind to which are usually G Protein Coupled Receptors (GPCRs) or Receptors with Tyrosine Kinase activity (RTKs).  Also aBeta42 is not a mitogen.

So what does Sumifilam actually do — it changes the ‘altered’ conformation of filamin A getting it away from the alpha7 acetyl choline receptor and “indirectly reducing the high femtoMolar binding affinity of aBeta42 for alpha7” (and however this binding triggers tau hyperphosphorylation)  How do they know the conformation of filamin A has changed?  They haven’t done cryoEM or Xray crystallography on the protein.  The only evidence for a change in conformation, is a change in the electrophoretic mobility (which is pretty good evidence, but I’d like to know what conformation is changed to what).

So there you have it, after a fairly deep dive into protein chemistry, cellular physiology and biochemistry, the current thinking of how Simufilam works.

But even if the theory is completely wrong, the data in the link above must be regarded with respect.  Clinical blinded studies are ongoing, and the soon to be released Cognition Maintenance Study should  give us more information –

How herpes viruses use the cell’s machinery to shut themselves off

Herpes viruses (simplex — for fever blisters, Kaposi’s sarcoma, herpes zoster — shingles) persist in the body in a latent state where they don’t don’t reproduce, don’t make many of their proteins and don’t make trouble.   Every now and then they reproduce and cause disease, as anyone with recurrent fever blisters will tell you.  Staying quiet allows them to avoid the immune system and essentially act as selfish DNA.

A recent paper [ PNAS vol. 120 e2212864120 ’23 ] shows that Kaposi’s Sarcoma Herpes Virus (KSHV) uses a circular RNA (circRNA) derived from a human oncogene called RELL1.  The circular RNA they induce is called hsa_circ-0001400.  In general circular RNAs are formed by back splicing of a 5′ splice to an upstream 3′ splice site.   One of their functions is to act as sponges for microRNAs as some contain multiple binding sites for them.  Some cells contain 25,000 of them.

Viruses are known to hijack cellular proteins to use for their own ends.  It isn’t clear how the herpes viruses stimulate formation of hsa_circ_0001400, but use it they do, as it promotes viral latency,  cell cycle genes and inhibits apoptosis.

This another example of the RNA world which supposedly existed before the DNA world, like DOS under the Windows operating system (forgotten but not gone)

Tubulin needs a lot of help from its friends

Our neurons (and us) would be the size of amoebas if weren’t for tubulin which forms the superhighways (microtubules) along which cargo is shipped to the end of axons.   Your average NBA player has axons over 3 feet long going from his sacral spinal cord to his calf muscles.   Split the difference and call it a meter.  Diffusion is way too slow to get anything that far. The trucks schlepping things back and forth on the microtubular highway are called Kinesin and dynein. I think in terms of nanoMeters (10^-9 meters).  Each tubulin dimer is 80 nanoMeters long, and K & D essentially jump from one to the other in 80 nanoMeter steps.

How many jumps do Kinesin and Dynein have to make to go a meter? Just 10^9/80 — call it 10,000,000. Kinesin and Dynein also have to jump from one microtubule to another, as the longest microtubule in our division is at most 100 microns (.1 milliMeter).  So even in the best of cases they have to make at least 10,000 transfers between microtubules.  It’s a miracle they get the job done at all.

To put this in perspective, consider a tractor trailer (not a truck — the part with the motor is the tractor, and the part pulled is the trailer — the distinction can be important, just like the difference between rifle and gun as anyone who’s been through basic training knows quite well).  Say the trailer is 48 feet long, and let that be comparable to the 80 nanoMeters Kinesin and Dynein have to jump. That’s 10,000,000 jumps of 48 feet or 90,909 miles.  It’s amazing they get the job done.

Now that you’re sufficiently impressed with tubulin’s importance, it’s time to see why it needs help.  First a bit of history.  Christian Anfinsen was a Swarthmore football player who happened to win the Nobel prize 50 years ago for his work on the protein ribonuclease, an enzyme.  If you heat it, enzymatic activity is lost (the protein is said to be denatured).  This is because the exact 3 dimensional path of the protein backbone forming the catalytic site of ribonuclease was lost. However if you leave the denatured protein alone (under the proper conditions) it folds back up to the correct 3 dimensional shape.  His point was that the amino acid sequence of the protein was all that was needed to determine ‘the’ three dimensional shape of the protein.  This was at a time when we didn’t know that most proteins have a variety of shapes not just one.

Unfortunately tubulin does not fold up to the shape found in microtubules.  It needs significant help from two friends, prefoldin and TRiC.  TRiC is a monster conglomerate of 2 copies each of 8 different proteins with a molecular mass over 1,000,000 Daltons (e.g. a megaDalton).  What is one Dalton — it’s the mass of a hydrogen atom.   TRiC is made of two back to back rings (with built in lids) each ring consisting of 8 different but related proteins).  Each of the proteins has a domain which binds ATP and a domain which binds the protein to be folded.  There is a central cavity 90 x 90 x 50 Angstroms in size.  Since each hydrogen atom is about 1 Angstrom in diameter, it can fit 405,000 hydrogen atoms inside, or about 200,000 carbons, hydrogens, oxygens and nitrogens — enough room for most proteins.

Prefoldin is equally amazing.  It basically looks like a Portuguese man o’ war —  It is made of 2 copes of one protein and 4 of another.  The tentacles are long alpha helices projecting down from the body.

The tentacles interact with tubulin, carrying it in an unstructured form, thrusting one of its tentacles into the central chamber of TRiC carrying unstructured tubulin with it.   ATP addition leads to lid closure and tubulin encapsulation in the chamber.

A magnificent paper [ Cell vol. 185 pp. 4770 – 4787 ’22 ] describes what happens to tubulin in the TRiC chamber at near atomic resolution.  They are literally watching tubulin fold as it passes from one of the 8 different proteins making up the TRiC ring to another.  The disordered carboxy terminal chains of TRiC are postulated to function as a tethered solvent allowing the intially disordered amino acid sequence of tubulin, to slither into their correct positions more easily.

I’m sure it’s behind a paywall, but if you can look at the figures in the paper, you’ll be bound to be impressed.

So Anfinsen turned out to be wrong, and some 10%  of newly translated proteins turn our to need TRiC’s help.  And yet he wasn’t, because AlphaFold uses only the amino acid sequence of proteins to predict their three dimensional structure.

One further point.  The ancestral bacterial protein for tubulin is called FtsZ.  It happily folds to the correct structure by itself.  However tubulin developed new domains, some of which are for the motor proteins Dynein and Kinesin, and others are for microtubule associated proteins such as tau, the major component of the neurofibrillary tangle of Alzheimer’s disease. These domains are on the surface of the protein, making it harder to fold by itself.

All this information would have been impossible to get 10 years ago, and it’s all due to the sharpening of our technological tools.

A few Thanksgiving thankyou’s

The following was published 5 years ago, but with time and ever more research our organization seems even more miraculous (see last paragraph).  It’s amazing that it lasts as long as it does, and for that we should be thankful.   Call this prayer if you wish.

As CEO of a very large organization, it’s time to thank those unsung divisions that make it all possible.  Fellow CEOs should take note and act appropriately regardless of the year it’s been for them.

First: thanks to the guys in shipping and receiving.  Kinesin moves the stuff out and Dynein brings it back home.  Think of how far they have to go.  The head office sits in area 4 of the cerebral cortex and K & D have to travel about 3 feet down to the motorneurons in the first sacral segment of the spinal cord controlling the gastrocnemius and soleus, so the boss can press the pedal on his piano when he wants. Like all good truckers, they travel on the highway.  But instead of rolling they jump.  The highway is pretty lumpy being made of 13 rows of tubulin dimers.

Now chemists are very detail oriented and think in terms of Angstroms (10^-10 meters) about the size of a hydrogen atom. As CEO and typical of cell biologists, I have to think in terms of the big picture, so I think in terms of nanoMeters (10^-9 meters).  Each tubulin dimer is 80 nanoMeters long, and K & D essentially jump from one to the other in 80 nanoMeter steps.  Now the boss is shrinking as he gets older, but my brothers working for players in the NBA have to go more than a meter to contract the gastrocnemius and soleus (among other muscles) to help their bosses jump.  So split the distance and call the distance they have to go one Meter.  How many jumps do Kinesin and Dynein have to make to get there? Just 10^9/80 — call it 10,000,000. The boys also have to jump from one microtubule to another, as the longest microtubule in our division is at most 100 microns (.1 milliMeter).  So even in the best of cases they have to make at least 10,000 transfers between microtubules.  It’s a miracle they get the job done at all.

To put this in perspective, consider a tractor trailer (not a truck — the part with the motor is the tractor, and the part pulled is the trailer — the distinction can be important, just like the difference between rifle and gun as anyone who’s been through basic training knows quite well).  Say the trailer is 48 feet long, and let that be comparable to the 80 nanoMeters K and D have to jump. That’s 10,000,000 jumps of 48 feet or 90,909 miles.  It’s amazing they get the job done.

Second: Thanks to probably the smallest member of the team.  The electron.  Its brain has to be tiny, yet it has mastered quantum mechanics because it knows how to tunnel through a potential barrier.   In order to produce the fuel for K and D it has to tunnel some 20 Angstroms from the di-copper center (CuA) to heme a in cytochrome C oxidase (COX).  Is the electron conscious? Who knows?  I don’t tell it what to do.   Now COX is just a part of one of our larger divisions, the power plant (the mitochondrion).

Third: The power plant.  Amazing to think that it was once (a billion years or more ago) a free living bacterium.  Somehow back in the mists of time one of our predecessors captured it.  The power plant produces gas (ATP) for the motors to work.  It’s really rather remarkable when you think of it.   Instead of carrying a tank of ATP, kinesin and dynein literally swim in the stuff, picking it up from the surroundings as they move down the microtubule.  Amazingly the entire division doesn’t burn up, but just uses the ATP when and where needed.  No spontaneous combustion.

There are some other unsung divisions to talk about (I haven’t forgotten you ladies in the steno pool, and your incredible accuracy — 1 mistake per 100,000,000 letters [ Science vol. 328 pp. 636 – 639 ’10 ]).  But that’s for next time.

To think that our organization arose by chance, working by finding a slightly better solution to problems it face boggles this CEO’s mind (but that’s the current faith — so good to see such faith in an increasingly secular world).

Call the thankfulness of the CEO prayer if you wish.

Addendum 29 November ’22 — from a friend “We also have to thank all the tau molecules that stabilize the microtubules— until the misbehavior of ERK and JNK1 overdecorate them with holiday lighting (phosphates) and they fall apart. So after Thanksgiving, be careful not to overcommercialize the holiday season.”

The cryoEM work of the last 5 years has shown us the structure of large molecular machines made of multiple proteins, DNAs and RNAs which are even more impressive (to me) than single protein structure.   One example [ Nature vol. 609 pp. 630 – 639 ’22 ] shows the Holliday junction which allows strand migration between the strands of two DNA duplexes.   Pictured is the complex from bacteria which is confined in a rectangle with sides 220 and 120 Angstroms (not sure how thick it is).  The complex contains a molecular motor which slides the junction.  You could spend your life just studying this one structure.  It’s hard for me to see how it arose.

Why trying to remove aBeta was plausible

The recent collapse of the latest attempt to remove the main constituent of the Alzheimer plaque, the aBeta peptide (gantenerumab from Roche) is just the latest in a long sad story.

Monoclonal after monoclonal antibody targeting aBeta has failed.  It certainly is time to move on and try new approaches.

The companies pursuing monoclonals were not stupid.  Their approach was (but no longer is) quite reasonable in view of the clinical and experimental evidence implicating the aBeta peptide as causative of Alzheimer’s  Before moving on, here are some of the reasons why.

First (and probably the best) is the mutation that protects against Alzheimer’s disease.  As most of you know, the aBeta peptide (39 to 42 amino acids) is part of a much larger protein the Amyloid Precursor Protein (APP) which contains 639 to 770  amino acids.  This means that enzymes must  cut it out.  Such enzymes (called proteases) are finicky, cutting only between certain amino acids.  In what follows A673T stands for the 673rd position which normally has amino acid Alanine (A) there.  Instead there is amino acid Threonine (T).   The enzyme cleaving at 673  is Beta Secretase 1 (BACE1).

       [ Nature vol. 487 pp. 153 ’12 ] A mutation in APP protects against Alzheimer’s disease.   First the genome sequence APP of 1,795 Icelanders  were studied to look for low frequency variants.  They found a mutation A673T adjacent to the site that is cleaved by beta secretase 1 (BACE1) which doesn’t vary — it’s gamma secretase which cleaves at variable sites leading to Abeta40, Abeta42 formation.  The mutation is at position 2 in Abeta.  The mutation results in a 40% reduction in the formation of amyloidogenic peptides  in vitro (293T cells transfected with variant and normal APP). Amazingly, a different variant at 673 (A673V — V stands for the amino acid Valine) — increases Abeta formation.    Because BACE1 can’t cleave APP containing the A673T mutation, alternative processing of APP at another site the alpha site (which is within aBeta preventing formation of the full 39 – 43 amino acid peptide).
So if you can’t make the full aBeta peptide you don’t get Alzheimer’s (or have less chance of getting it).
Then there are the mutations in the part of APP which code for the aBeta peptide which increase the risk of Alzheimer’s.  They cause the different familial Alzheimer’s disease.   Now that we know the actual structure of the aBeta amyloid fiber, we can understand how they cause Alzheimer’s disease.  This is more strong evidence that the aBeta peptide is involved in the causation of Alzheimer’s disease.
You’ll need some protein chemistry chops to understand the following

Recall that in amyloid fibrils the peptide backbone is flat as a flounder (well in a box 4.8 Angstroms high) with the amino acid side chains confined to this plane.  The backbone winds around in this plane like a snake.  The area in the leftmost loop is particularly crowded with bulky side chains of glutamic acid (single letter E) at position 22 and aspartic acid (single letter D) at position 23 crowding each other.  If that wasn’t enough, at the physiologic pH of 7 both acids are ionized, hence negatively charged.  Putting two negative charges next to each other costs energy and makes the sheet making up the fibril less stable.

The marvelous paper (the source for much of this) Cell vol. 184 pp. 4857 – 4873 ’21 notes that there are 3 types of amyloid — pathological, artificial, and functional, and that the pathological amyloids are the most stable. The most stable amyloids are the pathological ones.  Why this should be so will be the subject of a future post, but accept it as fact for now

In 2007 there were 7 mutations associated with familial Alzheimer’s disease (10 years later there were 11). Here are 5 of them.

Glutamic Acid at 22 to Glycine (Arctic)

Glutamic Acid at 22 to Glutamine (Dutch)

Glutamic Acid at 22 to Lysine (Italian)

Aspartic Acid at 23 to Asparagine (Iowa)

Alanine at 21 to Glycine (Flemish)

All of them lower the energy of the amyloid fiber.

Here’s why

Glutamic Acid at 22 to Glycine (Arctic) — glycine is the smallest amino acid (side chain hydrogen) so this relieves crowding.  It also removes a negatively charged amino acid next to the aspartic acid.  Both lower the energy

Glutamic Acid at 22 to Glutamine (Dutch) — really no change in crowding, but it removes a negative charge next to the negatively charged Aspartic acid

Glutamic Acid at 22 to Lysine (Italian)– no change in crowding, but the lysine is positively charged at physiologic pH, so we have a positive charge next to the negatively charged Aspartic acid, lowering the energy

Aspartic Acid at 23 to Asparagine (Iowa) –really no change in crowding, but it removes a negative charge next to the negatively charged Glutamic acid next door

Alanine at 21 to Glycine (Flemish) — no change in charge, but a reduction in crowding as alanine has a methyl group and glycine a hydrogen.

As a chemist, I find this immensely satisfying.  The structure explains why the mutations in the 42 amino acid aBeta peptide are where they are, and the chemistry explains why the mutations are what they are.

It’s time to look elsewhere.  The best this class of drug (monoclonal antibodies against aBeta) offers is lecanemab which slows the rate of decline by a measly 27%.   This is very small beer

While big pharma was far from stupid to intensively (and expensively) to give the monoclonals the old college try in the past (for the reasons cited above), they would be incredibly stupid to continue this line of attack.

Why you should never see a doctor who couldn’t pass organic chemistry

Because Maitland Jones is front page stuff in the New York Times and CNN, being fired from NYU because his tests in organic chemistry are too hard, it is worth republishing a post of 13 years ago about why passing organic chemistry is so important in weeding out people unable to muster the type of thinking they will need in practice.

Disclaimer:  I spent a pleasant hour with Dr. Jones over 15 years ago in his office chatting about the Princeton Chemistry department of the late 50s, and the Harvard Chemistry department of the early 60s.  He seemed like a reasonable guy.  I’ve also read and own his textbook of organic chemistry which is excellent.

For those new to the blog, I was a practicing clinical neurologist from 1967 to 2000. Prior to that I did some graduate work in Chemistry picking up a Masters (Harvard ’62)

Here is the post of 2009.

Why premeds should be required to take (and pass) organic chemistry

This post is to be mentioned in the 2 Nov C&EN. I’m reposting it so people can find it. The original came out 1 Sep.

Back when I was posting on “The Skeptical Chymist”, the editor (Stuart Cantrill), told me that noises were being made about dropping organic chemistry from the pre-med curriculum and asked me to comment. I didn’t because the idea seemed so ridiculous. There is no possibility of really understanding anything about cellular biology, drug action, molecular biology etc. etc. without a firm grounding in organic chemistry. You simply must have some idea what vitamins, proteins, DNA and RNA and the drugs you’ll be using look like and how they chemically interact — which is what organic chemistry gives you the background for. Not that you can stop there — but all medical schools teach biochemistry — which starts at organic chemistry and takes off from there. Organic certainly helped me follow molecular biology as it exploded starting in the 60s.

Cynics might say that docs don’t synthesize things or crystallize the drugs they use. Knowing what’s going on under the hood is just esthetic filigree. Just tell them what ‘best practice’ is, and let them follow it like robots. Who cares if they know the underlying science. People drive cars without really understanding what a carburator or a manifold does (myself included).

It wasn’t until I got about 400 pages into the magnificent textbook of Organic Chemistry by Clayden, Greeves, Warren and Wothers (only 1100 action packed pages to go !) that the real answer became apparent. The stuff is impossible to memorize. Only assimilating principles and applying them to novel situations will get you through — exactly like the practice of medicine.

Let us suppose you have an eidetic memory, and know the best treatment for every condition. You wouldn’t have to know any science at all, would you?

What’s wrong with this picture? First of all, there isn’t a best treatment known for every condition. Second, every doc will see conditions and problems that simply aren’t in the books. When I first started out, I was amazed at how much of this there was. I asked an excellent internist who’d been in practice for 30 years if he’d seen it all. He thought he saw something completely new each week. Third, conditions occur in combinations, and many patients (and nearly all the elderly) have many more than one problem. The conditions and treatments interact in a highly nonlinear fashion. The treatment for one problem might make another much worse (see below).

Here is a concrete example using a familiar person (Sonia Sotomayor) and a disorder which should be known to all (the new Swine Flu which swept America and the world this spring). Let’s say that you’re that lucky soul with the perfect memory who knows all the best treatments (well those that exist anyway) and as such you’ve been given the responsibility of taking care of her.

It is public knowledge (e.g. Wikipedia) that Justice Sotomayor has had diabetes since age 8, requiring insulin since that time. Pictures show, that like many diabetics, she is overweight — depending on how tall she is I’d guess by 25 – 45 pounds. Influenza is usually a disease of the fall and winter, and the new Swine Flu is now down in South America, but it’s likely to sweep back up here this fall. We know it’s extremely infectious, but so far fortunately rather benign. There is no guarantee that it will stay that way. Suppose that while down in S. A. it mutated and has become more virulent (a possibility that the CDC takes extremely seriously).

What if she gets the new Swine flu next month? At this point there is no ‘best treatment’ known. Diabetics don’t do well with infections — they get more of them, and have more complications when they do. Her diabetes is certainly going to get worse. What if some think the ‘best treatment’ is corticosteroids (which is often used for severe lung infections) — which will really raise hell with her diabetes? Should you give it? Recall that corticosteroid use during the Asian SARS epidemic (another serious lung infection) seemed to hurt rather than help (Journal of Infection, Volume 51, Issue 2, Pages 98-102). There is no data to help you here and you and your patient don’t have the luxury of waiting for it. Don’t forget that her father died at 42 of heart disease. That could be relevant to what you do. Suppose, like many overweight diabetics she has high blood pressure and elevated lipids as well. How will that affect her management?

Your perfect eidetic memory of medicine will not be enough to help you with her management — you are going to have to think, and think hard and apply every principle of medicine you know to a new and unfamiliar situation with very little data to help you.

Sounds like Organic Chemistry doesn’t it? Anyone without the particular type of mind that is able absorb and apply multiple and (often) conflicting principles doesn’t belong in medicine. A hardnosed mathematician I audited a course from a few years ago, said that people would come up to him saying that if they couldn’t pass Calculus, they wouldn’t get into medical school. He felt that if they couldn’t, he didn’t want them in medical school (I’m not sure he told them this — probably he did). The same thing holds in spades for Organic Chemistry.

4 diseases explained at one blow said the protein chemist — part 2 — TDP43

A brilliant paper [ Science vol. 377 eabn5582 pp. 1 –> 20 ’22 ] explains how changing a single amino acid (proline) to another  can cause 4 different diseases, depending on the particular protein it is found in (and which proline of many is changed).

There is so much in this paper that it will take several posts to go over it all.  The chemistry in the paper is particularly fine.  So it’s back to Biochemistry 101 and the alpha helix and the beta sheet.

A lot of the paper concerns TDP43, a protein familiar to neurologists because it is involved in FTD-ALS (FrontoTemporal Dementia — Amyotrophic Lateral Sclerosis) and ALS itself.

I actually saw a case early in training.  I had been taught that ALS patients remained cognitively intact until the end (certainly true in my experience — think of Stephen Hawking), so here was this ALS case who was mildly demented.  My education, deficient at that time, so I’d never heard of FTD-ALS, had me writing in the chart “we’re missing something here”.  These were calmer times in the medical malpractice world.

TDP43 is a protein with a lot of different parts in its 414 amino acids.  There are two regions which bind to RNA (Rna Recognition Motifs { RRMs } ), and a glycine rich low complexity domain at the carboxy terminal end.

TDP43 proteins are found in the neuronal inclusions of ALS (interestingly, these weren’t recognized when I was in training).  The low complexity domain of TDP43 aggregates and form fibers.  Some 50 different mutations have been found here in patients.

Just this year the cryoEM structure of TDP43 aggregates from two patients with FTD-ALS were described [ Nature vol. 601 pp. 29 – 30, 139 – 143 ’22 ].  It appears to be a typical amyloid structure with all 79 amino acids (from # 282 Glycine to #360  Glutamine) in a single plane.  Here’s a link to the actual paper —  It is likely behind a paywall, but if you can get it, look at figure 2 p. 140, which has the structure.  Who would have ever thought that a protein could flatten out this much.

Both structures were from TDP-43 with none of the 24 mutations known to cause FTD-ALS.

But that’s far from the end of the story.  The same area of TDP43 can also form liquid droplets (perhaps the precursor of the fibers).  But that’s where the brilliant chemistry of [ Science vol. 377 eabn5582 pp. 1 –> 20 ’22 ] comes in.

That’s for next time.  After that, I should be finished with Needham and will have time to write about 6 or so of the interesting papers I’ve run across in the past 6 months.

We interrupt this program . . .

I’ll interrupt the series of posts on the brilliant article [ Science vol. 377 eabn5582 pp. 1 –> 20 ’22 ] to talk about working with the very frightening diazo methane 61 years ago.

I was able to convince Woodward to let me work on an idea of mine to show that carbenes were generated by photolysis of a diazo compound (this was suspected but not known at the time).

Here’s the idea

l. Condense acrylic acid with cyclopentadiene by a Diels Alder reaction.  Because of steric effects the acid points below the ring

2. Form the acyl chloride

3. React with diazoMethane to form the diazocarbonyl (no change in the orientation of the carbonyl relative to the ring.

4. Photolyze — if  a carbene is formed, it’s in perfect position to form a cyclopropane on the other side of the ring which if formed would pretty much prove the point.

Diazomethane was known to be quite explosive, and I spent a lot of time tiptoing around the lab when working with it.  Combine this with the worst lab technique in the world and I couldn’t get things to work. Subsequently the idea was shown to be correct, and an enormous amount of work has been done on carbenes.

So why interrupt the flow of posts about the brilliant  [ Science vol. 377 eabn5582 pp. 1 –> 20 ’22 ] ?

Because Science vol. 377 pp. 649 – 654 ’22 reports a simple (and nonexplosive) way to form carbenes from aldehydes.  Here’s what they say

“Common aldehydes are readily converted (via stable a-acyloxy halide intermediates) to electronically diverse (donor or neutral) carbenes to facilitate >10 reaction classes. This strategy enables safe reactivity of nonstabilized carbenes from alkyl, aryl, and formyl aldehydes via zinc carbenoids. Earth-abundant metal salts [iron(II) chloride (FeCl2), cobalt(II) chloride (CoCl2), copper(I) chloride (CuCl)] are effective catalysts for these chemoselective carbene additions to s and p bonds.”

How I wished I had this back then.

4 diseases explained at one blow said the protein chemist — part 1

A brilliant paper [ Science vol. 377 eabn5582 pp. 1 –> 20 ’22 ] explains how changing a single amino acid (proline) to another  can cause 4 different diseases, depending on the particular protein it is found in (and which proline of many is changed).

There is so much in this paper that it will take several posts to go over it all.  The chemistry in the paper is particularly fine.  So it’s back to Biochemistry 101 and the alpha helix and the beta sheet.

Have a look at this

If you can tell me how to get a picture like this into a WordPress post please make a comment.

The important point is that hydrogen bonds between the amide hydrogen of one amino acid and the carbonyl group of another hold the alpha helix and the beta pleated sheet together.

Enter proline : p//  Proline when not embedded in a protein has a hydrogen on the nitrogen atom in the ring.  When proline is joined to another amino acid by a peptide bond in a protein, the hydrogen on the nitrogen is no longer present.  So the hydrogen bond helping to hold the two structures (alpha helix and beta sheet) is no longer present at proline, and alpha helices and beta sheets containing proline are not has stable.  Prolines after the fourth amino acid of the alpha helix (e. g. after the first turn of the helix) produce a kink.  The proline can’t adopt the alpha helical configuration of the backbone and it can’t hydrogen bond.

But it’s even worse than that (and this observation may even be original).  Instead of a hydrogen bonding to the free electrons of the oxygen in the carbonyl group you have the two electrons on the nitrogen jammed up against them.  This costs energy and further destabilizes both structures.

Being a 5 membered ring which contains the alpha carbon of the amino acid, proline in proteins isn’t as flexible as other amino acids.

This is why proline is considered to be a helix breaker, and is used all the time in alpha helices spanning cellular membranes to cause kinks, giving them more flexibility.

There is much more to come — liquid liquid phase separation, prion like domains, low complexity sequences, frontotemporal dementia with ALS, TDP43, amyloid, Charcot Marie Tooth disease and Alzheimer’s disease.

So, for the present stare at the link to the diagram above.