I want to address one particular point in that article.
For decades enzymologists have recognized that certain enzymes are catalytically perfect—meaning that they process reactant molecules as rapidly as these molecules can reach them by diffusion. [1] That hinted at a principle of physical perfection in biology, but no one anticipated its breadth until recently.The point of the article is that some things in biology are "perfect" but this presents a problem for "Darwinism." According to the "scientists" at the Biologic Institute. perfection is beyond the reach of a "Darwinian mechanism." The fact that we observe perfection in biology is evidence for design.
ERV picked up on this theme in a 2009 posting: Having your Ford Pinto and Crashing it Too. That posting seemed to concede the point that enzymes are perfect.
That's the issue I want to discuss.
There's no such thing as a perfect enzyme. It's true that there are enzymes whose rate is so fast that it is only limited by the rate of diffusion of substrates into the active site. The diffusion-controlled rate is about 108 or 109 M-1 s-1. Several enzymes have second order rate constants (kcat/Km) that are close to this maximum rate.
Many textbooks claim that these enzymes have attained "catalytic perfection." For example, here's what Voet & Voet (4th edition) say (page 490) ...
Some Enzymes Have Attained Catalytic Perfection
Thus, enzymes with such values of kcat/Km must catalyze a reaction every time they encounter a substrate molecule. ... several enzymes ... have achieved this state of virtual catalytic perfection.
The "Perfect" Enzyme?There are also many enzymes that occasionally catalyze incorrect reactions. For example, several amino acid synthetases have appreciable rates of incorrectly attaching the wrong amino acid to a given tRNA molecule. The error rate of the DNA replication complex is well known, it's responsible for most mutations. I suspect there are very few enzymes that are so specific that they never make mistakes.
Much of our understanding of the mechanism of triose phosphate isomerase (TPI) comes from the lab of Jeremy Knowles at Harvard University (Cambridge, MA, USA). He points out that the enzyme has achieved catalytic perfection because the overall rate of the reaction is limited only by the rate of diffusion of substrate into the active site. TPI can’t work any faster than this!
This has led many people to declare that TPI is the “perfect enzyme” because it has evolved to be so efficient. However, as Knowles and his coworkers have explained, the “perfect enzyme” isn’t necessarily one that has evolved the maximum reaction rate. Most enzymes are not under selective pressure to increase their rate of reaction because they are part of a metabolic pathway that meets the cell’s needs at less than optimal rates.
Even if it would be beneficial to increase the overall flux in a pathway (i.e., produce more of the end product per second), an individual enzyme need only keep up with the slowest enzyme in the pathway in order to achieve “perfection.” The slowest enzyme might be catalyzing a very complicated reaction and might be very efficient. In this case, there will be no selective pressure on the other enzymes to evolve faster mechanisms and they are all “perfect enzymes.”
In all species, triose phosphate isomerase is part of the gluconeogenesis pathway leading to the synthesis of glucose. In most species, it also plays a role in the reverse pathway where glucose is degraded (glycolysis). The enzyme is very ancient, and all versions—bacterial and eukaryotic—have achieved catalytic perfection. The two enzymes on either side of the reaction pathway, aldolase and glyceraldehyde 3-phosphate dehydrogenase (Section 11.2), are much slower. Thus, it is by no means obvious why TPI works as fast as it does.
The important point to keep in mind is that the vast majority of enzymes have not evolved catalytic perfection because their in vivo rates are “perfectly” adequate for the needs of the cell.
I think we should stop talking about "perfect" enzymes. Not only does it encourage the Intelligent Design Creationists, it's also scientifically misleading if you want students to understand biochemistry from an evolutionary perspective.
[See also: Fixing Carbon: Building a Better Rubisco and Finding the "perfect" enzyme]
35 comments :
While I haven't gone back to the literature, I assume the reaction rates of TPI were determined using purified protein and assays done in cell free conditions. If this assumption is correct, then the intelligent designer (hallowed by his name) designed these enzymes several billion years ago to be 'perfect' in vitro once biochemistry developed into its own field in the 20th century.
Nice explanation!
Is the "perfect enzyme" even really that much of a surprise? Many enzymes "simply" have a binding pocket that matches the shape of the reactants within an environment which reduces the activation energy of the resulting reaction (i.e. hydrogen bond-mediated stabilization of intermediary products). In the case of these basic enzymes, which don't require much in terms of conformational change, etc, is it really that big of a surprise that they are efficient, and limited primarily by the diffusion rate of the reactants/products in/out of the enzyme? Why you would expect anything less of what is little more than a passive catalyst* is a bit of a mystery to me.
*obviously, many enzymes's function is much more complex than this, but to my limited knowledge, most (all?) of the "perfect" enzymes are these sort of "basic catalysts", rather than more complex enzymes requiring conformation changes & other slower processes
What Axe was saying is that catalytic rate can be mapped to a fitness landscape with lots of valleys and peaks of different height. So, starting from most positions on the landscape one would travel up the closest peak to a faster rate, but it would probably not be the highest possible rate. Therefore most enzymes should not be 'perfect' if they've evolved. Hes assuming the fitness landscape for reaction rate will be very rugged, and hes also assuming that most enzymes will be 'perfect' - both unfounded assumptions I believe.
This leads me to some general obsevations/questions which perhaps Larry will indulge. It seems to me that some reactions can be carried out by more than one mechanism- some reactions are carried out by different classes of enzymes that have a fundamentally different mechanism. Also, for any particular mechanism there are probably many configurations of active site that would get the job done. Now my question is this- is there any reason to believe that all of these possibilities could reach catalytic 'perfection' or will some be inherently limited in how fast they could catylize?
RodW
Is the "perfect enzyme" even really that much of a surprise?
Yes.
Many enzymes "simply" have a binding pocket that matches the shape of the reactants within an environment which reduces the activation energy of the resulting reaction (i.e. hydrogen bond-mediated stabilization of intermediary products)
But one would still expect a geometric factor to come into play - substrate needs to reach the enzyme from the "right side" in order to bind. Working at exactly diffusion limit implies that this geometric factor does not exist - the enzyme "conveys" substrate toward/into active site no matter how they bumped into each other. I've seen references to the enzymes exceeding diffusion limits but, personally, I do not understand how it is possible. The "that's because the active site is surrounded by charged residues that attract oppositely charged substrate molecules!" does not cut it because it implies that the substrate is moving directionally (and faster than diffusion) under the influence of coulombic force, which is, I think, too weak for that.
How much diffusion actually goes on in a cell anyway? These are crowded environments with scaffold proteins and all sorts of protein complexes that pass substrates around or position them appropriately.
I agree, though, that "perfection" in biology is almost always misleading, not just with enzymes. Most systems have trade-offs, such as speed versus accuracy, or different optima in different environments/situations. In order to know that something had a supernatural level of "perfection", you would probably need a supernatural level of omniscience...
Finally, even if one could imagine a situation with a single unidirectional selection pressure and no pleiotropic effects, why could this not evolve? If "perfection" is better than "imperfection" (and presumably it must be to be "perfection") then would there not be a positive selection pressure for things that are more "perfect"? It's just a stupid argument all round. (No surprises there, I guess.)
There might be another molecule actively delivering the substrate, or increasing the local concentration.
This is just about whether a particular enzyme activity is possible though. Once you have a mechanism, whatever that is, there is no a priori reason why it could not have evolved. If it is physically possible, evolution has a chance of finding that solution (if it is beneficial or, at worst, neutral (or very slightly deleterious)). A designer is still surplus to requirements unless you can demonstrate that this solution appeared from nowhere.
I am reminded of videos of molecular activity in cells in which substrate molecules do not diffuse, instead they go directly to the exact place they are needed, exactly when they are needed (examples: The Inner Life of the Cell, and animation which appeared in the movie Expelled). The activity in these videos more closely resembles a factory assembly line than a living cell, and is far more "perfect" than reality.
but one would still expect a geometric factor to come into play - substrate needs to reach the enzyme from the "right side" in order to bind
I think you've mis-understood what is meant by "catalytically perfect" enzymes; it relates to an enzyme where the catalytic rate is limited not by the rate of the chemical reaction, but rather by the rate of diffusion. Meaning the limit of these enzymes is exactly the limit you mention - the rate at which substrate can reach the catalytic site. "Imperfect" enzymes are limited by the rate of chemical reaction rather than diffusion of the substrate(s) - i.e. the maximum speed of the enzyme is limited by the enzymatic speed itself, rather than by diffusion.
I've seen references to the enzymes exceeding diffusion limits but, personally, I do not understand how it is possible. The "that's because the active site is surrounded by charged residues that attract oppositely charged substrate molecules!" does not cut it because it implies that the substrate is moving directionally (and faster than diffusion) under the influence of coulombic force, which is, I think, too weak for that.
In this, you are simply wrong. I have authored papers where we've directly measured the effect of charged residues (in my case, charged phospholipids) on concentrating ions, charged proteins, etc. These "weak" interactions are capable of producing local increases in concentration several magnitudes in order. The same occurs with proteins and ions. The increased catalytic rate is a direct result of this, as higher local concentrations decrease the average time it'll take diffusion to deliver reactants to the catalytic site - no directionality is required, merely conventional brownian diffusion.
How much diffusion actually goes on in a cell anyway? These are crowded environments with scaffold proteins and all sorts of protein complexes that pass substrates around or position them appropriately.
A lot - we measure it daily in my lab. It is far slower than you'd expect of a free protein (ion, etc) in water, but its far from static. Typical rates are 0.5-5um^2/s; in water similar sized proteins diffusion at 10-15x that (i.e. GFP = 87um^2/s). There are generalities; diffusion rate cn vary hugely for proteins of a similar rate, presumably due to differences in interactions with membranes, cytoskeleton, etc.
Oops, where I wrote diffusion rate cn vary hugely for proteins of a similar rate, it should have read:
diffusion rate can vary hugely for proteins of a similar size
1. Diffusion says absolutely nothing about the rate at which one molecule reaches particular site on another molecule. The only thing diffusion tells you is a rate at which two spheres that approximate two molecules bump into each other. Because proteins are large, the becomes important.
2. I said nothing about concentration of anything. Nothing at all. The relevant issue is mobility. You add molecule B to the molecule A. At time zero, they are at a distance L from each other and local concentration of A near B is zero. If A and B are able to react faster than an average time required to travel the distance L by diffusion, the reaction rate is not diffusion-limited. This is the only situation where one might be justified claiming that reaction rate is not diffusion-limited.
Your scenario involves changes in local concentration, in which case kcat/Km that "exceeds diffusion limit" is only apparent, not real - because we did not/cannot measure Km properly (i.e., assumption of equality of bulk and local concentration is no longer correct). Claiming that the rate of reaction exceeds diffusion limit in this case makes zero physical sense and such claims are absurd. At the very least, a word "apparent" must be used.
Many will never reach the diffusion-controlled rate because the substrates are large and the orientation is important. This is especially true for bimolecular reactions where two different substrates have to be aligned properly.
Most of the reactions catalyzed by fast enzymes are first order reactions.
For small molecules (i.e. most substrates of chemical reactions) the diffusion rate inside the cell is about 25% of the rate in pure water.
The effect of molecular crowding is more of a problem for proteins. For small proteins, like myoglobin, the diffusion rate inside the cell will be about 10% of the rate in pure water.
The videos are proof of intelligent design.
I think we should stop talking about "perfect" enzymes. Not only does it encourage the Intelligent Design Creationists, it's also scientifically misleading if you want students to understand biochemistry from an evolutionary perspective. LM
Yes. It would be possible to compile a lexicon of unwisely chosen words, phrases and metaphors in science that obscure or mislead more than they clarify. "Perfect" enzymes might be a perfect example of those.
Having been critical of the substitution of evidence with an ideologically created simulation of evidence among materialists, this looks like like the same thing from the other side.
My understanding of biochemistry is very poor (but maybe adequate for most of what I do) but I wonder about the statement
Even if it would be beneficial to increase the overall flux in a pathway (i.e., produce more of the end product per second), an individual enzyme need only keep up with the slowest enzyme in the pathway in order to achieve “perfection.”
How does that fit into the work of my former landlord, and a delightful person, the late Henrik Kacser? In papers like this:
Kacser, H., and J. A. Burns, 1973. The control of flux. Symposia of the Society for Experimental Biology 27: 65-104.
The message of that paper is that changes in enzyme activity in a linear pathway can always increase the flux through the pathway at least a little. There is no one perfectly limiting step, though one step could be nearly-limiting. This result has been called "molecular democracy". It implies that one is always a bit short of perfection.
This is your field and not mine, but it implies to me that natural selection would always be at work on all the enzymes of a pathway. However for the more-perfect among them the selection coefficients of improvements could get so small that further improvement would stall in the face of deleterious mutation, particularly in small populations (that part is my field). Arguments like this are used by Michael Lynch, applied to selection on genomes, in his recent and important work The Otigins of Genome Architecture.
Diffusion says absolutely nothing about the rate at which one molecule reaches particular site on another molecule
Of course not, but it does say a lot about the probability of a population of molecules (reactants) entering a specific space (i.e. the active site). And since we're dealing with populations of molecules, not individual ones, you can use the basic principals of diffusion to approximate the rate of reactant entry (and product exit) from an enzymes active site, at the level of the population.
Your scenario involves changes in local concentration, in which case kcat/Km that "exceeds diffusion limit" is only apparent, not real
Exactly. While the rate is above the theoretical rate for the bulk solution, it is exactly what you'd expect given the local concentration.
Larry is dead on; I should have specified that I was talking about "average" proteins (80kDa-ish). Ions are much less restricted, while lipids are much more restricted.
Bryan
This is your field and not mine, but it implies to me that natural selection would always be at work on all the enzymes of a pathway.
Why would you assume that? I imagine that almost all common metabolic pathways have already evolved to the "good enough" state. There's no selective pressure to get any faster.
However for the more-perfect among them the selection coefficients of improvements could get so small that further improvement would stall in the face of deleterious mutation, particularly in small populations (that part is my field).
That's an interesting point. Let's take triose phosphate isomerase (TPI) as an example. There's no obvious reason why mutations that make the enzyme somewhat slower don't accumulate since this reaction is not a rate-limiting step in gluconeogenesis or glycolysis. And even if those mutations were slightly deleterious, you might expect some of them to become fixed. I wonder how many TPI'S from different species have actually been looked at?
Arguments like this are used by Michael Lynch, applied to selection on genomes, in his recent and important work The Otigins of Genome Architecture.
Lynch is trying to explain how complexity (large complicated genomes) can arise by non-adaptive means. What I'm trying to do here is understand why highly efficient enzymes can evolve in the apparent absence of selection.
It might be instructive to look at human diseases associated with catalytic deficiency in these types of enzymes. "The impairment of TPI activity apparently does not affect the energy metabolism at system level; however, it results in accumulation of dihydroxyacetone phosphate followed by its chemical conversion into the toxic methylglyoxal, leading to the formation of advanced glycation end products." (F. Orosz, J. Olah, J. Ovadi, Triosephosphate isomerasedeficiency: New insights into an enigmatic disease, Biochim Biophys Acta, 1792 (2009), pp. 1168–1174)
I think we are talking past one another. Kacser and Burns's work implies that there is actually no such thing as a rate-limiting step: every enzyme in a linear pathway (they assume no feedback too) has some effect on the throughput of the pathway. So there is some selection on all of them, though maybe much less on some enzymes than on others.
Likewise with "good enough state". Natural selection does not magically disappear when the enzymes become "good enough". What does happen is that as the selection coefficients get small, genetic drift and deleterious mutation start to become more influential. That is the kind of argument Michael Lynch makes. You are right that he is not talking about enzymes -- I brought him up to show an example of a population genetic framework that should be used instead of "good enough". We need to use that kind of framework and not rely on notions of the "rate-limiting" step or "good enough adaptation".
You were posing the question as possibly including cases where it would be beneficial to increase the overall flux in a pathway. The "good enough" argument is shorthand for the fitness approaching an asymptote. But it is never all the way there.
You're right. We are probably talking past each other to some extent.
Natural selection does not magically disappear when the enzymes become "good enough". What does happen is that as the selection coefficients get small, genetic drift and deleterious mutation start to become more influential.
I suppose there are many way to look at this. To my way of thinking, the selection coefficient is zero, to all intents and purposes, once the enzyme has evolved to be good enough. You seem to think that it's still possible for natural selection to make a better enzyme but the beneficial mutation is unlikely to be fixed.
The "good enough" argument is shorthand for the fitness approaching an asymptote. But it is never all the way there.
I define the asymptote as the rate at which no further amino acid substitutions will be beneficial. They will all be either neutral or detrimental. Why will evolution never get to that asymptote?
Is the fact that this is reversible, but with a strong bias, potentially relevant? There is a 20:1 preference for the DHAP direction over GAP (per Wikipedia ... :0). I'm not clear if the enzyme is regarded as 'perfect' in both directions ... but GAP synthesis proceeds because downstream reactions are continually reducing its concentration, 'sucking' GAP out of TPI against the unbalanced equilibrium. Therefore, in vitro TPI could produce DHAP from GAP with seemingly extravagant efficiency simply because it has to produce a sufficiency of GAP in the reverse direction on cellular demand ... ?
The standard Gibbs free energy change for the reaction is about 7.6 kJ per mole. It's a near-equilibrium reaction inside the cell so the concentration of DHAP will be about 20 times greater than the concentration of GAP at equilibrium (ΔG = 0).
The rate of the reaction (both directions) is very fast but this has nothing to do with the thermodynamics or the equilibrium constant.
Flux can occur readily in either direction, as with any near-equilibrium reaction. GAP is produced during gluconeogenesis so that any small increase in GAP concentration will be quickly restored to equilibrium values by making more DHAP. The reverse situation occurs during glycolysis.
Larry, we are still on different wavelengths here. If one enzyme in the pathway has the strongest effect, so it looks to be the rate-limiting one, another enzyme will still have some small effect. If fitness is higher the higher the throughput, then there will be some small positive selection coefficient in favor of an increase in enzyme activity at that second enzyme. I am not claiming that it will be effective -- it may not, in the face of genetic drift and mutation.
But you are saying something else -- that at this point there is no selection favoring improvement of the second enzyme. I disagree.
a geometric factor to come into play - substrate needs to reach the enzyme from the "right side" in order to bind. Working at exactly diffusion limit implies that this geometric factor does not exist
Not the case. Working at exactly the diffusion limit means that the reaction is completed and the products are cleared before the next substrate molecule is encountered. There is no requirement that the reaction is completed instantly when the substrate arrives.
While TPI may appear 'too fast for glycolysis/gluconeogenesis', I wonder whether its position at a biochemical crossroads may also be relevant - to supply or consume from the pentose pathway and lipid metabolism for example. These pathways may not all be in simultaneous operation, but TPI (itself unregulated) needs to be quick enough and responsive enough (by having no appreciable lag time) to maintain steady levels of DHAP/GAP regardless of which pathway(s) are producing each substrate - a spinning flywheel, reversible with no notice.
Dimeric catalysis also makes it more than twice as fast as a similarly configured monomer (being geometrically bidirectional). There may not be a great deal of latitude between fast (how long does it take to move a proton anyway!) and deleterious - a rugged fitness landscape.
'Nuff armchair guesswork!
No! Working at diffusion limit means that *every bump is productive*. But with large macromolecules, an active site is not and cannot be at any and all points of the enzyme's surface. So the geometric factor has to exist (and it does, of course - rough calculations from experimental values for, say, high affinity calcium binders, show that affinity = diffusion corrected by geometry).
Joe is right. In a real metabolic pathway there is no such thing as ‘THE’ rate-limiting enzyme. All enzymes contribute to the control of the flux, although the quantitative contribution of individual enzymes (technically, their control coefficient) differ. This is standard stuff and has been well understood for about 30 years, starting (as Joe mentioned) with the work of Kacser and Burns.
I’m slightly surprised that Athel hasn’t popped up to explain this since he has made major contributions to metabolic control analysis.
If you have a super-efficient enzyme, wouldn't that mean the cell could get by with less of it, and that would be a metabolic benefit? For example, if you have substrate (enzyme A) - intermediate (enzyme B) - product, and enzyme B had 100 times the reaction rate of enzyme A , you'd think the cell would evolve to have a much higher enzyme A:enzyme B ratio(compared to a situation where the two enzymes had similar reaction rates, concentration of substrate, etc.)
It would be interesting to do the experiment - Break (deoptimize) the gene for one of the superfast enzymes, see what happens over a few hundred or thousand generations in copy number and enzyme concentration.
If you have a super-efficient enzyme, wouldn't that mean the cell could get by with less of it, and that would be a metabolic benefit?
Good point. I forgot all about that possibility.
It's certainly true that some very inefficient enzymes have to be present at enormous concentrations—Rubisco is a classic example.
However, in most cases, a slight improvement in the rate of a reaction isn't likely to lead directly to a reduction in the amount of enzyme being produced. In the absence of such a direct connection, there doesn't seem to be an obvious benefit to improvements in the rate of the reaction.
On the other hand, you could probably make a random genetic drift argument that makes sense. The chance increase in the rate makes it possible for otherwise deleterious mutations affecting the synthesis/degradation of the enzyme to accumulate. Once that happens, the rate improvement allele becomes essential.
"On the other hand, you could probably make a random genetic drift argument that makes sense. The chance increase in the rate makes it possible for otherwise deleterious mutations affecting the synthesis/degradation of the enzyme to accumulate. Once that happens, the rate improvement allele becomes essential."
That sounds like a 'just so story' to me !
I agree, and I'm thrilled that you have learned how to recognize them.
Hmmm catalase k_cat/K_M in Voet, Voet & Pratt 4ed. (table 12.1 p362) is an order of magnitude out: 10^7 / (2.5 x 10^-2) = 4 x 10^8... The same figures are given in table 14.1 (p489, 4e) of V&V
I think the numbers you've put up are the ones I've seen before though... 4 x 10^7 is also the figure given in table 6.4 of Horton 4ed., and Berg, Tymoczko & Stryer (table 8.7, 7e, p.243), the latter sources its numbers to Fersht's Structure & Mechanism in Protein Sci (1999).
http://i.imgur.com/WZkkncm.jpg
Post a Comment