Easy Reads About Evolution From a Biochemist
See a Problem?
Thanks for telling us about the problem.
Friend Reviews
Community Reviews
Reading level: Undergrad for biology majors. Graduate for general STEM majors.
Chap 1 is a basic over
Written by one of the world experts on enzyme kinetics, it is a masterful overview of biochemistry in light of evolution, because nothing in biology makes sense except in light of evolution! It really allowed me to grok biochemistry, and encouraged me to buy some glycine because it turns out glycine deficiency is well-nigh universal for large terrestrial adult animals! (I'll expand on this below).Reading level: Undergrad for biology majors. Graduate for general STEM majors.
Chap 1 is a basic overview of physics, chemistry, and biology. If you are going to read this book, it's probably too elementary for you.
Chap 2 is an overview of evolution, to support later arguments about what is naturally selected vs what is unselected. Example: Competitive inhibition by two similar molecules? Not necessarily selected, since it can easily happen by chance. Allosteric regulation? Selected, since allosteric regulation is unlikely to happen by chance. Enzyme cooperativity? Definitely selected, since human engineers, despite trying very hard, has not designed a single catalyst that has cooperativity.
In short: if it's rare and mainly appears in places where it's functionally advantageous than a random alternative, it's probably selected.
Also a fun fact:
> Two Asian species of deer, the Chinese muntjac (Muntiacus reevesi) and the Indian muntjac (Muntiacus muntjac), provide an extreme illustration of the danger of regarding chromosome numbers as absolutely fixed. The two species are quite similar in appearance (much more similar than chimpanzees and humans) and can form viable hybrids, but **whereas the Chinese muntjac has 46 chromosomes, the Indian muntjac has six in females and seven in males.**
Chap 3 is a long list of examples of selection and non-selection, as well as correct and wrong explanations.
Sperm whale myoglobin: selected for long dives
Horse liver alcohol dehydrogenase: for dealing with gut fermentation
Polar bear liver vitamin A concentration: adaptation to tolerate high vitamin A concentration in fish, not because they need all the vitamin.
Insulin variation in pigs, dogs, and humans: neutral drift.
Left-right brain inversion: probably a frozen accident.
Arginine has more codons than aspartate: complicated frozen accident. In fact, new codons can evolve when there's less pressure for it to stay the same! For example, some mitochondria evolved new codons, because they have only a few hundred proteins.
Chap 4 and 5 provide a satisfying interpretation for two seemingly arbitrary metabolic pathways: the pentose phosphate pathway, and the Calvin cycle, regeneration of ribulose phase
In short, both pathways are essentially about moving carbon atoms around carbohydrates, using only two kinds of enzymes: transketolase can move 2 carbons at a time, and transaldolase can move 3 at a time.
There is no enzyme that can move 1 carbon at a time, for somewhat complicated reasons. From what I gather, it's because there is a limit to enzyme specificity. Distinguishing between a 4-ose and a 5-ose is just too hard, compared to distinguishing 4-ose and 6-ose.
Finally, add another constraint: we can't have a carbohydrate with 1 or 2 carbons. 1-carbon would be formaldehyde, extremely toxic. 2-carbon would be glycoaldehyde, not as toxic, but still used in any known metabolic pathways (I searched Google Scholar for a reason, but can't find any.).
Under these constraints, it turns out that the two pathways are indeed optimal (give or take a few isomers and epimers)! Optimal, in the sense that they use the least number of steps, and least number of intermediate metabolite species. That is, they are the *minimal* metabolic networks that can do their job.
Chap 6 shows why glycogen is optimal.
This chapter is based on The Fractal Structure of Glycogen: A Clever Solution to Optimize Cell Metabolism (1999)
The structure of glycogen is optimal under a particular metabolic constraint model. In detail, the glycogen structure is the optimal design that maximizes a fitness function based on maximizing three quantities: the number of glucose units on the surface of the chain available for enzymic degrading, the number of binding sites for the degrading enzymes to attach to, the total number of glucose units stored; and minimizing one quality: total volume.
If each chain has 0 or 1 branch points, we obtain essentially a long chain, not a sphere, and it would occupy too big a volume with only a few terminal glucose units for degrading. If each chain has 3 branch points, the glycogen would fill up too quickly. The balance-point is 2.
With that branch number 2, the chain length needs to be at least 4. As modelled by Meléndez et al, the fitness function reaches maximum at 13, then declines slowly.
Empirically, the branch number is 2 and the chain length ranges 11-15 for most organisms ranging from vertebrates to bacteria and fungi. The only significant exception is oyster, with glycogen chain length ranging 2-30, averaging 7.
The author speculates that this is what doomed many species: they are outcompeted by species with more efficient biochemical substances. Those that didn't become sluggish underachievers like oyster.
Chap 7 is basically a retelling of the author's paper Enthalpy—entropy compensation: a phantom phenomenon (2002).
It shows that the "entropy–enthalpy compensation" is "too good to be true", and on scrutiny, turns out to be a mathematical tautology, not a biological insight, and certainly not a result of natural selection. The paper is just 5 pages long, so you can look it up yourself for details.
Chap 8 develops a hydraulic analogy for enzyme kinetics (the author's specialty). Basically, think of a metabolic network as a network of water tanks connected by valved tubes. The water height in each tank represents the metabolite concentration, and how big the tubes are open represents how much enzyme is available for that step of metabolism.
Alternatively, you can think of a electric network, with electric potentials being the metabolite concentrations and the resistor conductance as enzyme concentration.
In a typical metabolic network with N nodes, decreasing each edge by half only decreases the total flux by 1/(2N). As a result, many enzymes seem to be "not doing much", as decreasing their concentration by half doesn't show any visible effect. However, this is not a sign of "failure of evolution".
Halving the enzyme would save 1/(2N) in enzyme synthesis, and that, at the price of decreasing the total flux by 1/(2N)... a perfect balance! Similarly, doubling the enzyme is also useless. This allows considerable neutral drift.
Why AMP became a signaling substance? Because it is "useless" in metabolism, and also sensitive to [ATP].
Under biological conditions, [ATP] = 10 mmol/L, while ADP = 1 mmol/L. Then, a 10% drop in [ATP] translates to a 4.4-time increase in [AMP], since the reaction constant K = [ATP][AMP]/[ADP]^2 = 0.5.
> Because it is not a substrate or product of many metabolic reactions, a fourfold change in its concentration is unlikely to produce any problems. Instead, it can be used as a signal to enzymes that would "like" to be able to respond to small changes in ATP concentration.
Chap 9 explains how recessivity works.
Observation: Almost all mutations are recessive. Why?
The theory of modifier genes, R. A. Fisher (1928): through natural selection, other genes come to suppress dominancy together.
Sewell's theory (1934): in a polyploidal organism, most enzyme-coding genes are recessive, simply because a big metabolic network is insensitive to halving the rate of any particular step. No natural selection needed.
In other words, recessivity is an example of unselected feature (see chap 2) due to the nature of metabolic network (see chap 8).
"A test of Fisher's theory of dominance" (1991) conclusively disproved Fisher's theory, while supporting Sewell's theory. It showed that yeast, even though spending most of its life cycles in haploploid form, still has most genes being recessive when in diploid form.
Humans apparently have many dominant genes, but that's only because we have studied humans with particular closeness.
> The idea of a "symptomless disease" might seem absurd, but it is applied in all seriousness to conditions like maturity-onset diabetes of the young ("MODY"), in which small differences in blood-sugar levels allow a prediction that diabetes may develop later. If Mendel had measured the exact amounts of pigment in peas from different plants, he would have been able to detect differences that are not obvious to the unaided eye.
Chap 10: the economic laws of cellular homeostasis (NOT homeorhesis).
Suppose you want to design a metabolic network for a cell, what should you keep within safe limits? You could keep the flowrate (homeorhesis), or the concentration (homeostasis). Cells, it turns out, uses homeostasis, NOT homeorhesis. This is actually quite understandable: if the concentration is too high or too low, generally you have severe osmotic troubles. If the flowrate is too high or too low, you don't get any problem (except perhaps overheating, which happens very rarely).
In order to keep homeostasis, several network motifs appeared in cellular metabolism.
Negative feedback ("feedback inhibition"): often long-range, at the first committed step, often enhanced by allosterism and cooperativity.
Allosteric long-range control at the first committed step: Think back to the toy model. If you notice a pile-up at the last water tank, you should tell not just the last valve to tighten up, but all the way to the first valve. Otherwise, you would get pile-ups at the second-to-last water tank, then the third-to-last, etc... Thus, we need long-range feedback at the first committed step.
Short-range feedback is easy. Consider an enzyme E: X->X'. Usually, X' and X look similar, and are thus competitive agonists to each other on the same binding site. However, long-range feedback is hard, and requires allosteric control.
Cooperativity: Hemoglobins are supposed to ferry oxygen between atmospheric oxygen level and muscular oxygen levels. The muscles work best if they have a high oxygen level. So the hemoglobin is selected for picking up almost all oxygen at a slightly higher concentration, and dumping almost all oxygen at a slightly lower concentration -- thus cooperativity.
Positive feedforward: mostly used in detoxification metabolic pathways.
Negative feedforward and positive feedbacks do not exist in metabolic pathways for obvious reasons: they are destabilizing, breaking homeostasis.
Biotechnology applications of homeostasis: There is no "rate-limiting step" in metabolic pathways, because there is no homeorhesis. Money wasted!
> The idea is that once you identify the enzyme that catalyzes the rate-limiting step, you can overexpress the enzyme so that its activity in the living organism is higher than it would normally be; then the organism will make more product and everyone will be happy. The major problem with this approach is that it does not work.
> The crucial experiments were done in Germany in the 1980s, early enough in the story for vast amounts of money to have been saved if the results had been taken seriously. One of the things that "everyone knows"—everyone who has followed a standard course of biochemistry, in this case—is that phosphofructokinase controls glycolysis... Its activity responds to an impressive variety of metabolic signals that indicate the immediate requirements of the organism. This behavior is certainly crucial for regulating glycolysis, but it is translated in an excessively simple-minded way into the naive idea that phosphofructokinase is the rate-limiting enzyme and that increasing its activity will inevitably result in a higher glycolytic rate. When the activity of phosphofructokinase in fermenting yeast was increased by a factor of 3., however, there was no detectable effect on the amount of alcohol produced.
As an example of using homeostasis the right way, consider glutamate, used in MSG.
> during the 1950s by growing the bacterium Corynebacterium glutamicum (at that time called Micrococcus glutamicus) on pure cane sugar as a carbon source, a cheap source of sugar imported in large quantities from Cuba. The bacteria were deliberately deprived of biotin, a vitamin essential for making cell membranes, with the result that their membranes leaked glutamate, which could then be harvested by the food company.
> At the end of the 1950s, the cheap source of pure sugar was no longer available [due to the 1953 Cuban revolution] and molasses had to be used instead. Although this was sufficiently cheap, it was much less pure and contained large amounts of biotin, which would have been far too expensive to remove. The bacteria no longer made low-quality membranes, which no longer leaked, so glutamate ceased being lost to the growth medium. To overcome the problem, a different way of making the membranes leaky was needed, and this was achieved by growing the bacteria in the presence of penicillin (Figure 10.), which acts by interfering with the building of membranes.
Chap 11. A biochemical evolutionary "mistake": a metabolic network problem causes near-universal glycine deficiency in large terrestrial animals.
Glycine is the simplest amino acid, and it is a major component of the body, mostly in the collagen. The collagen is a distinguishing feature of animals, and it is 1/3 glycine. Moreover, collagen turnover is very fast, on the order of 40 days. Thus, it is surprising that there is a subtle rate-limiting flaw in the main glycine synthesis pathway: the tetrahydrofolate cycle (THF cycle).
The THF cycle has 2 phases.
- Phase 1: serine gives its -CH2OH group to THF, to become glycine. THF becomes THF-C1, where "C1" stands for the -CH2OH group because it has 1 carbon atom.
- Phase 2: THF-C1 gives away its C1 to someone else, returning to THF.
Now, we immediately see a problem: if nobody wants a C1, then THF-C1 cannot return to THF, and the cycle grinds to a halt.
Now, a reasonable way to fix this is to construct an extra pathway: Serine -> THF. However, what we get is Glycine -> THF-C1, which doesn't fix the problem at all. This mistake is universal in all animals. Plant's don't have it because they don't use collagen.
"Fortunately", the mistake only causes problems for large, terrestrial, adult animals.
Why large? Scaling laws. Metabolic rate scales as 0.75 (Kleiber's law), but collagen need scales as 1.1. Thus, it is about 7 times as difficult for a human to produce enough collagen as it is for a rat.
Why terrestrial? Because they need more and stronger bones.
Why adult? Because in a growing animal, there are many acceptors of C1, so the THF cycle won't grind to a halt.
A weak link in metabolism: the metabolic capacity for glycine biosynthesis does not satisfy the need for collagen synthesis (2009)
> Detailed assessment of all possible sources of glycine shows that synthesis from serine accounts for more than 85% of the total, and that the amount of glycine available from synthesis, about 3 g/day, together with that available from the diet, in the range 1.5–3.0 g/day, may fall significantly short of the amount needed for all metabolic uses, including collagen synthesis by about 10 g per day for a 70 kg human. This result supports earlier suggestions in the literature that glycine is a semi-essential amino acid and that it should be taken as a nutritional supplement to guarantee a healthy metabolism.
Branch point stoichiometry can generate weak links in metabolism the case of glycine biosynthesis (2008):
> Osteoarthritis is, in fact, one of the rare diseases that are common in the wild: it has been detected in the skeletons of large dinosaurs from the Jurassic Age, 100-150 million years ago (Rothschild 1987, 1993); it occurs in present-day large mammals, such as elephants (Weissengruber et al 2006) and rhinoceroses (Bonard 1987; Wallach 1967). In general, osteoarthritis has been found in a broad variety of present-day mammals, both in the wild and in captivity (Greer et al 1977), and specifi cally in great apes, such as chimpanzees, gorillas and bonobos (Jurmain 2000). It has also been found in fossil Hominidae, including Neanderthals (Straus and Cave 1957), in Upper Paleolithic and Neolithic human fossils (Ackernecht 1953), and in human fossils from the Middle Pleistocene in Atapuerca (Spain) (Pérez and Martínez 1989).
Chap 12: how unlikely is it to evolve the biochemistry of life?
Proteins are not as hard to evolve as they seem. Some proteins are highly preserved, but many are not.
The histone H4 protein is almost perfectly conserved across the entire range of life, but its sequence is not. Even humans have several different DNA sequences coding for the same H4 protein.
DNA shuffle experiment (1997) shows that even random shuffling of protein-coding genes can generate a large number of functional enzyme variants, some even better.
> The idea is to take a set of genes from different strains of a bacterium that code for forms of an enzyme that are all different, but which are all capable of catalyzing the same reaction. If the genes are broken into fragments, mixed together, and then reassembled in a way that allows each reassembled gene to contain a random collection of fragments from different sources, the result is a large collection of genes coding for related but different protein sequences. When these genes are expressed as proteins, many of them turn out to be nonfunctional, but, more interestingly, large numbers of novel proteins do act as catalysts, some of them better than the natural ones.
Chap 13 talks about cancer, 14 about origin of life, and 15 about creationism. I find them less illuminating.
...more
News & Interviews
Welcome back. Just a moment while we sign you in to your Goodreads account.
Source: https://www.goodreads.com/book/show/30225163-biochemical-evolution
0 Response to "Easy Reads About Evolution From a Biochemist"
Enregistrer un commentaire