Gauger: Answering Art Hunt on Real Time Evolution

[swamidass (S. Joshua Swamidass)]

It is my pleasure to welcome back @Agauger to Peaceful Science, after her brief absence.

She wanted to respond in detail to @Art Hunt’s explanation of an important paper that directly demonstrates an important evolutionary mechanism in the laboratory. This paper was quoted in our science review of Darwin Devolves (Darwin Devolves: The End of Evolution?).

Behe and Gauger’s Response to This Study

http://science.sciencemag.org/content/338/6105/384

1. Our discussion of Behe’s response is here: Behe's Trainwreck Response to Science

2. Our discussion of @Agauger’s response is here: https://discourse.peacefulscience.org/t/gauger-realtime-evolution-by-innovation-amplification-and-diversification/4828

Regarding this reference, Behe writes:

Let me emphasize: in reviewing a book expressly advocating intelligent design, Lenski et al. can’t seem to distinguish between experiments where investigators keep their hands off and those where investigators actively manipulate a system. Perhaps they can’t see the difference.

At this point four people from DI (including Behe and @Agauger) have objected on the same grounds that this is a “designed” experiment. I do not understand this objection. It is an experiment designed to test an evolutionary mechanism. The fact it is designed does not erase what it tells us about this evolutionary mechanism. This seems to me (and most biologists) as an implausible objection.

On side note, it seem Behe is unfamiliar how Lenski’s LTEE experiment works. The researchers are very hands on with these bacteria. Setting that aside, in our review we write:

modern evolutionary theory provides a coherent set of processes—mutation, recombination, drift, and selection—that can be observed in the laboratory and modeled mathematically and are consistent with the fossil record and comparative genomics.

Gene duplication followed by divergence is a process that we can observe directly in the laboratory (in systems engineered to test it) and is consistent with the fossil record and comparative genomics (outside the laboratory). It explains how new functional information arises.

@Art’s Explanation

One another thread (Leisola: Cited to Attack Darwin Devolves, Study Devolves on Close Inspection - #19), @art offered this excellent explanation. This is the explanation to which @Agauger wants to respond now.

@Art

Nasvall et al. developed and set out to test a model for the evolution of new genes that they call the “innovation-amplification-divergence (IAD)” model:

We propose the innovation-amplification-divergence (IAD) model (Fig. 1A), which allows the evolution of new genes to be completed under continuous selection that favors maintenance of the functional duplicate copies and divergence of the extra copy from the parental allele (5). The IAD model proposes that the ancestral gene has a weak secondary activity (innovation) (6, 7), and when a change in conditions makes this activity useful, selection favors increased gene dosage (amplification), resulting in two or more copies of the parent allele. The increased copy number pro- vides multiple targets for beneficial mutations and buffers any negative effects a new mutation may have on the original activity. During continuous growth under conditions that select for both the original activity and the new activity, beneficial mutations will accumulate (divergence) in the copies. Any improved copy can be further amplified, whereas less functional copies, including the parental gene, can be lost. Ultimately, this results in a gene duplication in which one gene copy encodes the parent activity and another copy provides an improved, new activity.

To test this model, they broke it down into its three components and evaluated each step separately. For this, they focused on two enzymes that catalyze chemically-related reactions, but that are in pathways for histidine or tryptophan biosynthesis, respectively.

To experimentally test the IAD model, we examined a histidine biosynthetic enzyme (HisA), and through continuous selection we created, by duplication and divergence, a new gene that catalyzes a step in tryptophan synthesis. The original HisA and TrpF enzymes both catalyze isomerization of a phosphoribosyl compound, but each acts on different substrates in the biosynthesis of the amino acids histidine and tryptophan (Fig. 1B). HisA and TrpF enzyme activities are selectable by growth in minimal media lacking histidine and tryptophan. In addition, the enzymes are structurally related and evolved from a common ancestor (8).

Thus, briefly, they first asked whether they could identify a HisA variant that could catalyze the TrpF reaction; this was simple, accomplished by screening trpF mutants that could grow on media lacking tryptophan (and, obviously, histidine). They isolated bacteria able to do this and confirmed that they had indeed isolated strains in which HisA had acquired TrpF activity (and was thus bifunctional). This is a positive experimental confirmation for the “innovation” part of their model. (Note that, other than conduct the experiment with a mutant strain, these researchers did nothing to promote or guide the outcome.)

The researchers then set out to test the second step of the model, and specifically to ask if gene amplification could improve the growth of cells that carry the bifunctional gene. I won’t describe the set-up, other than to say that they placed the bifunctional gene into a genetic context that allowed for easy generation and assay of gene amplification, and then performed selections in strains in which the bifunctional strain was the only source of HisA and TrpF activity. They identified variants that indeed grew better, and confirmed that the improved growth was due to gene amplification. This outcome is positive experimental support for the second part of their model.

To test the third step of the model, they then allowed lineages to continue to evolve under the same selection (no his or trp) for thousands of generations. They succeeded in isolating many better-growing variants. Different of these had bifunctional enzymes with further alterations; collectively, these alterations define many (most?) of the different mechanisms by which evolution might improve the enzyme:

As predicted from the IAD model, we observed the appearance of a diverged gene copy with improved activity, relaxed selection for maintenance of the unimproved copies in the amplified array, loss of the unimproved copies, and, in some cases, reduction in the total gene copy number.

Compare what Nasvall did with Ann’s description:

The story: Näsvall et al. created a gene that encoded an enzyme that was able to carry out two functions, but very poorly.

No, they did not “create” a bifunctional gene. Random mutation and natural selection was the mechanism, not design.

They placed that gene in a strain of Salmonella that lacked the genes for those functions. Then they cultured the bacteria under conditions where they needed to carry out those functions to grow. Guess what? The bacteria duplicated the genes so as to make more of the poorly functioning enzymes.

Which was the exact point of the assay – to test the hypothesis that duplication can contribute to enzyme evolution. The answer, it turns out, is yes. (It seems to me that Ann’s objection here is quite akin to claiming that enzyme assays are not valid because the biochemist first makes extracts (and purified enzymes) and then conducts tests with non-natural substrates. In other words, the objection fails.)

Over time, the genes acquired mutations (remember, we are talking about a lot of bacteria) and the ones that helped the most gave the best growth rate, and… after thousands of generations they had evolved separate enzymes for each function — meaning not that they had evolved new enzymes, but that they had divided up the pre-existing dual-functional gene into two genes.

Again, not correct. Nasvall et al. noted:

The evolved genes fell into three classes: (i) specialized genes with strongly improved HisA activity and loss of TrpF activity, (ii) specialized genes with strongly improved TrpF activity and loss of HisA activity, and (iii) generalist genes whose encoded enzyme showed a moderate increase in both activities …

Note that, among the classes of enzymes are ones in which HisA has now become TrpF (for all intents and purposes). This is a transition that Gauger and Axe say cannot occur (from the abstract):

Considering that Kbl2 and BioF2 are judged to be close homologs by the usual similarity measures, this result and others like it challenge the conventional practice of inferring from similarity alone that transitions to new functions occurred by Darwinian evolution.

Note that the only “tools” Nasvall et al. used were random mutation and natural selection - exactly what Gauger and Axe claim are not tenable means for generating new functions.

The bottom line - Ann, @Wayne_Rossiter, Behe, and others object to Nasvall et al. for reasons that are completely baseless.

Our Goal is Understanding, Not Agreement

Before returning to this conversation, I want to be clear what our goals are and what we can expect.

@swamidass

If you must dispute mainstream science, it is not our aim to change your mind. Rather, we hope that you might understand what you reject, and that, perhaps, we might understand you too. We might also find ways to see how your most deeply held beliefs are not threatened.

The goal here is understanding. @Art’s explanation of this paper is lucid and (it seems to me) correct.

In the past, @Agauger has also explained what her limits are too. She is an employee of the Discovery Institute, and a colleague of Behe and Axe. Even if @art has not made any errors, she may not be willing to concede any of Behe or Axe’s arguments. We should all respect these limitations.

Nonetheless, I hope we can understand why DI thinks this is a valid objection, and they can understand why most scientists reject their objection as invalid. The goal here is understanding, not agreement.

[Agauger (Ann Gauger)]

[@moderators note: this email was conveyed to us from @Agauger by @Wayne_Rossiter]

@Art Hunt has posted a deceptive rebuttal to my description of the Nasvall paper. He lightly glosses over some major technical feats that were required to get the experiment to go. If either of you is willing, could you make these points for me?

1. Doug and I have always granted that it is possible to recruit a protein to a new function for which it already has a low level of activity, in other words, to recruit a promiscuous enzyme to one of its minor reactions. This is likely to be the case if the chemistry of the reaction (the kinds of bonds broken and formed and their relative positions and energy requirements) and/or the substrates are similar. THIS WE HAVE ALWAYS SAID IS POSSIBLE. HisA and TrpF have been classic cases of promiscuous enzymes for a while. See below.

Jürgens C, Strom A, Wegener D, Hettwer S, Willmans M and Sterner R (2000) Directed evolution of a (βα)8-barrel enzyme to catalyze related reactions in two different metabolic pathways - PMC. Proc Natl Acad Sci 97, 9925-9930

In this paper they converted the HisA enzyme (from Thermotoga) to carry out TrpF function by directed evolution on limiting trp medium. They recovered 3 mutant strains that could grow slowly (after several days). 2 strains were identical and had four amino acid changes (4 base changes). The other strain had 3 amino acid changes (3 base changes). When all seven changes were tested individually by site directed mutagenesis, only one had a measurable effect on TrpF function [Asp127Val (A380T)]. Kcat/Km of converted enzyme was 10-5 lower than wild type. All 3 aa changes improved function four fold.

The chemistry for HisA and TrpF are the same, but the substrates have different R groups (see below). Conversion results in loss of original function, and poor new function. The original substrate ProFar for HisA is quite large-

The new substrate PRA is much smaller, but has many structural similarities.

2. When Art says they used evolution to create the trpF hisA mutants he left out a bit. Here is what it took (from the Methods and Materials of the Nasvall paper) to make the double mutant they started with. To say it happened by evolution is a half-truth at best. If this is evolution then all of genetics is evolution, and geneticists should just sit back and relax and let evolution do the work. This required a sophisticated set-up, and follow through. What you see below is just the set up.

Deletion of the trpF and hisA genes

In Salmonella enterica and many other bacteria, the trpF gene is fused to the trpC gene, resulting in a bifunctional fusion protein (TrpCF). Hereafter we refer to the 3’ part of the trpCF gene (encoding aminoacids 257-452 of the fusion protein) as trpF. The hisA and trpF genes were replaced with FLP-recombinase target (FRT) site flanked kanamycin resistance (KanR ) cassettes derived from plasmid pKD4, using the lambda-red recombineering functions from plasmid pKD46 (24). The following primers were used for PCR amplification of the resistance cassettes: Sal_hisA_del_Fwd: cgcggcgggcgcacagttgctgaaaaacttcctggagatgtaatGTGTAGGCTGGAGCTGCTTC Sal_hisA_del_Rev: accatcacgaacatccagacacggaattatacgttttgccagCATATGAATATCCTCCTTAG Sal_trpF_del_Fwd: gacgatcttaacgccgccgtccgtcgcgtgctgcttgccgaaaattaatGTGTAGGCTGGAGCTGCTTC G Sal_trpCF_del_Rev2: catgccgccgaattcaccaaagtaggggttgagaagtgttgtCATATGAATATCCTCCTTAGTTCC The kanamycin cassettes in the resulting mutants, ΔhisA::kan and ΔtrpF::kan, were used as selectable markers for strain constructions, but were later removed by expression of the FLP recombinase from plasmid pCP20 (26) to generate the ΔhisA::FRT and ΔtrpF::FRT mutations.

Selection for bifunctional HisA mutants

We plated several independent cultures of a ΔtrpF::FRT, hisO1242 strain (DA15428) on M9+glycerol+histidine plates to select any mutant that could produce tryptophan without a TrpF enzyme. We included histidine in the medium because previous studies have demonstrated that mutations that confer TrpF activity to Thermotoga maritima HisA results in loss of HisA activity (27). The expression of the his operon (including hisA) is regulated by an attenuation mechanism that regulates the amount of read-through of a transcriptional terminator before the first structural gene of the operon according to the availability of charged histidinyl-tRNA (28). As this results in very low expression of hisA in medium containing histidine, we included a mutation (hisO1242) (29) which removes the transcriptional terminator, thereby leading to derepressed transcription of the his operon even in the presence of histidine. This selection resulted in very few mutants (an estimated mutation frequency of less than 10-11 / cfu). All of the isolated mutants were auxotrophic for histidine, but based on growth on M9+glycerol+histidine plates they fell in two different classes. One class was relatively fast-growing, forming colonies in 2-3 days while the other class grew more slowly, forming colonies in 4-5 days. Sequencing hisA in six independent clones from the faster growing class revealed two very similar mutations; three mutants had a duplication of nt. 37-45 and three had a duplication of nt. 36-44, resulting in duplication of the same three amino acids (V13, V14 and R15). For the rest of the experiments we have used a mutant with duplication of nt. 37-45 (GTGGTGCGT), and refer to it as hisA(dup13-15). Sequencing hisA in five independent isolates of the slower growing class of mutants revealed that they all have the same CUG (Leu) to CGG (Arg) mutation at codon 169 (L169R). As all of the isolated mutants had completely lost the original hisA activity we performed a second selection, starting with hisA(dup13-15) and hisA(L169R) mutants and selecting for growth in the absence of both histidine and tryptophan. As histidine was omitted from the medium, this selection did not require the hisO1242 mutation. Selections starting with hisA(dup13-15) resulted in relatively high frequencies of mutants (ranging between 3×107 /cfu and 6×107 /cfu in different experiments). Sequencing hisA in 18 independent mutants revealed two different mutations (in addition to the dup13-15 mutation), 15 of the mutants had a mutation resulting in a Gly to Asp substitution at amino acid 11 (G11D), while three had an Asp to Gly substitution at amino acid 10 (D10G). The same selection starting with the hisA(Leu169R) mutant also resulted in His+, Trp+ mutants (approximately 1x10-8 /cfu), but none of the tested mutants had any additional mutation in hisA. To map the mutations, a pool of approximately 20,000 random Tn10dTet insertions was generated in one of the mutants (DA16721). A phage P22 lysate grown on the transposon pool was used to transduce the parent strain (DA16608), selecting tetracycline resistance and histidine prototrophy (His+ phenotype). We found four different Tn10dTet insertions that were genetically linked to the mutation that caused the His+ phenotype. Seven more independent mutants were tested for cotransduction between one of the isolated Tn10 insertions (miaB::Tn10dTet) and the mutation that caused the His+ phenotype. All mutations were linked to the same location. Sequencing outwards from the Tn10s revealed that they were close to the metT tRNA operon. Sequencing the metT operon revealed that all eight tested mutants had a mutation in either glnX (3 mutants) or glnV (5 mutants), converting a tRNAGln to a missense suppressor tRNA with an anticodon complementary to CGG (Arg) codons. Thus, a fraction of the HisA molecules will have a L169Q substitution instead of L169R, apparently restoring at least some of the original activity. The mutation in glnX is referred to as glnXCCG (where CCG is the anticodon) in the text.

NOW HERE’S THE IRONIC THING. NASVALL ET AL DEMONSTRATED THAT PROMISCUOUS ENZYMES CAN BE SHIFTED TO A NEW FUNCTION USING MUTATION AND SELECTION JUDICIOUSLY APPLIED, AND THAT DUPLICATION AND DIVERGENCE CAN LEAD TO THE SEPARATION OF THE TWO FUNCTIONS INTO SEPARATE GENES. YOU START WITH HIS A, YOU MAKE HIS A TRP F, THEN YOU SEE IF YOU CAN GET THE GENES TO DUPLICATE AND DIVERGE. AND THEY DO! WAS IT AN EVOLUTIONARY PROCESS? IT WAS A GENETIC TOUR DE FORCE, AN EVOLUTIONARY PROCESS DESIGNED AND GUIDED BY 4 GENETICISTS. ONE OF THEM THE BEST IN HIS FIELD.

So why was I shouting? Because this is ludicrous. Evolutionary biologists are using a perfect example of evolution by design as their argument against Michael Behe. I am also yelling because apparently Art Hunt doesn’t understand anything about bacterial genetics, and how much work that would be, and he doesn’t understand anything I have written about the limits of protein evolution, or he hasn’t read it, or he is misrepresenting it. This HisA Trp F conversion is something we would say could happen. If it can be reached within a few mutations, i.e. in a genetic selection like here, then, hey, it potentially could evolve. Without investigator interference? Probably not. But potentially. If the need was great. And apparently has in different species of bacteria.

Stop with the misrepresentations already. @Art Hunt You have people reading you who don’t know the material and depend on you for the truth. So don’t twist it or play word games. Whether or not it is plausible trpF and HisA could have evolved one from the other, the fact is the Nasvall paper was an exercise in intelligently designed directed evolution. Not evolution.

[Rumraket (Mikkel R.)]

Sorry, again, that’s all complete and utter nonsense what Ann Gauger writes there. It was NOT a bifunctional enzyme to begin with. It was NOT a promiscuous enzyme used to begin the experiment. ONLY after multiple spontaneous mutations in the hisA gene could it catalyze BOTH reactions.

They did NOT start with a promiscuous enzyme capable of doing both reactions. They started with a His+ capable enzyme (hisA) in an organism lacking the TrpF enzyme, and selected for the presence of Trp+ function in the hisA gene. But because they knew that mutations for Trp+ in hisA genes typically eradicate the His+ function, histidine was included in the agar.

They plated bacteria lacking the ability to make tryptophan on agar containing histidine but not tryptophan, so only if a spontaneous mutant occurred capable of Trp+ would colonies form. They kept doing this until such colonies were observed. They found that initially, as expected, the hisA mutants capable of growth on medium lacking tryptophan but containing histidine have lost their His+ activity.

They then selected these Trp+ colonies for further plating and selection experiments(meaning they cultured the picked colonies for a time so they had more bacteria with spontaneous mutations to plate with) and kept plating them for a time looking for new mutants. But this time they plated on agar lacking both histidine and tryptophan, so if a colony should form, it would be from a spontaneous mutant capable of both His+ and Trp+. Such colonies were detected, meaning such spontaneous mutations in fact occurred. Nobody forced the requisite mutations to occur or “engineered” them. What was provided was the means to detect them (by a colony forming on agar), so they could be selected for further experiments. That’s it.

Nobody made the mutations come into existence. They merely set up an environment that would make it possible for them to detect when and if the requisite mutations they were interested in, occurred, by plating bacteria on plates lacking particular substrates. If a colony is observed on such a plate, it means the bacteria have evolved some way to make the substrate themselves. Unless of course it’s God intervening in the experiment, but I hear he’s stopped doing that and all his interventions are relegated to the ancient geological past and he refuses to do test-tube interventions. Must be a kind of social contract I guess.

There was no promiscuous enzyme to begin with, and a promiscuous enzyme was not created, it evolved by spontaneous mutation by plating bacteria on agar lacking the nutrients the researchers wanted to see if the bacteria were capable of evolving the ability to biosynthesize. And they were.

@Ann Gauger

Doug and I have always granted that it is possible to recruit a protein to a new function for which it already has a low level of activity, in other words, to recruit a promiscuous enzyme to one of its minor reactions.

Sure, but this is not what happened. What happened here is something you say CAN’T happen.

This HisA Trp F conversion is something we would say could happen.

No it isn’t. You have an entire paper dedicated to arguing that this can’t possibly take place by evolution:
Gauger AK, Axe DD (2011) The evolutionary accessibility of new enzyme functions : a case study from the biotin pathway. BIO-Complexity 2011(1):1-17.

In this paper you take two highly divergent, non-promiscuous enzymes , fail to convert the function of one into the function of the other by intelligently picking residues from one and replacing the aligned corresponding amino acids in the other. After you fail converting the function of A to the function of B by rational design, you amazingly conclude evolution couldn’t ever be expected to perform a conversion of one enzyme function into another even closely related one.

So now that a multifunctional enzyme has been evolved from a single-function one by spontaneous mutation and selection by plating on agar lacking particular substrates, you are trying to spin it as a case of intelligent design. Amazingly.

When it fails (for two proteins with ~280 amino acid differences out of about 400), it means evolution is effectively impossible and ID is required, and when it succeeds (for two proteins differing by 4-6 amino acids), that’s a success of ID and couldn’t have happened by evolution.

C’mon.
Edit: for some spelling/grammar.

[Faizal_Ali]

I’m having a hard time following your line of reasoning as a whole, Ann.

The work that you and Axe did, which you are often citing, involved deliberately manipulating proteins in the lab and trying to produce proteins with new or improved functions.

From the failure of such functions to arise in your experiment, you conclude that evolution is unable to produce such functional improvements.

Now, when other researchers are actually able to produce such improvements, you claim this does not demonstrate such processes can occur in the natural world, because they took place in the lab.

But, if that is the case, then what was the point of your experiments? Were they not intended to test whether evolutionary processes can produce novel functions, and is that not exactly how you have interpreted its results?

If such laboratory experiments cannot, in the view of ID researchers, test this hypothesis, then why are you wasting your time running them in attempt to test the hypothesis?

Could you please clear up my confusion?

[swamidass (S. Joshua Swamidass)]

@Rumraket

When it fails (for two proteins with ~280 amino acid differences out of about 400), it means evolution is effectively impossible and ID is required, and when it succeeds (for two proteins differing by 4-6 amino acids), that’s a success of ID and couldn’t have happened by evolution.

To be clear, isn’t the native TrpF much more different than hisA? How much more different is it? This is an important detail to clarify.

[Rumraket (Mikkel R.)]

@swamidass

To be clear, isn’t the native TrpF much more different than hisA? How much more different is it? This is important detail to clarify.

That appears to be true. In the article @Agauger cites:
Jürgens C, Strom A, Wegener D, Hettwer S, Willmans M and Sterner R (2000) Directed evolution of a (βα)8-barrel enzyme to catalyze related reactions in two different metabolic pathways. Proc Natl Acad Sci 97, 9925-9930

The tHisA protein has a ~10% identity to the tTrpF protein:

But in that same paper we also see that the wildtype tHisA enzyme appears to have no measurable Trp+ activity in vitro (and does not support growth in media lacking Trp in vivo).

[Argon]

Just an aside: If you’ve got enough time to wait and especially if you go through enough different substrates, every enzyme is ‘promiscuous’.

[Rumraket (Mikkel R.)]

To substantiate what I wrote above about the non-activity of Trp+ from wt hisA, the authors of Näsvall et al 2015write:

In a strain lacking trpF , we selected a spontaneous hisA mutant of Salmonella enterica that maintained its original function (HisA) but acquired a low level of TrpF activity sufficient to support slow growth on a medium lacking histidine and tryptophan, representing the innovation of the IAD model (see table S1 for strains). Two mutations were required for this innovation: First, an internal duplication of codons 13 to 15 (dup13-15) gave a weak TrpF activity but led to a complete loss of HisA activity. A subsequent amino acid substitution [Asp10→Gly10 (D10G)] restored some of the original HisA activity (10).

If a mutation gave a weak TrpF activity, it wasn’t there to begin with, and it completely abolished the HisA activity. I’d say that means it’s not a bifunctional enzyme to begin with.

Similar things are stated in the supplementary materials:

We plated several independent cultures of a ΔtrpF::FRT, hisO1242 strain (DA15428) on M9+glycerol+histidine plates to select any mutant that could produce tryptophan without a TrpF enzyme. We included histidine in the medium because previous studies have demonstrated that mutations that confer TrpF activity to Thermotoga maritima HisA results in loss of HisA activity (27).

And:

As all of the isolated mutants had completely lost the original hisA activity we performed a second selection, starting with hisA(dup13-15) and hisA(L169R) mutants and selecting for growth in the absence of both histidine and tryptophan.

Suffice it to say that an enzyme that has lost the His+ activity is not bifunctional towards His+ and Trp+.

Also, from reference 27(Jürgens C et al 2000, linked above) we see this:

The Trp+ activity of tHisA is at the level of noise (zig-zagging black line in at the top of graph A)

[Rumraket (Mikkel R.)]

And with respect to @Agauger and Axe’s 2011 paper in bio-complexity, here’s part of their discussion:

@Axe&Gauger 2011

Still, we will proceed with the lower requirement of five amino-acid substitutions in order to arrive at a lower-bound estimate of the time needed for functional conversion in a natural population. That number must be incremented slightly in order to estimate the minimum number of nucleotide substitutions for achieving conversion. Since an average of 1.5 nucleotide substitutions is needed per amino-acid substitution11, seven is a reasonable lower-bound estimate of the specific nucleotide substitutions required for conversion. Although this is not exceptionally high compared to the extent of change used in other attempted conversions (see Introduction), it nonetheless places the Kbl→BioF conversion outside the bounds of what can be achieved by the Darwinian mechanism. Specifically, using a population model described previously [18] with the following assumptions:

• A1– Duplications of kbl that are suitable starting points for paralogous innovation occur at a rate of 10^-8 to 10^-3 per cell,
• A2– The metabolic cost of carrying one such duplicate allele (unconverted) decreases fitness by 0.01% to 16% relative to the wild-type, 12
• A3– Conversion to bioF function can be achieved with seven base changes, and
• A4– The converted gene confers an overall growth advantaga of 1%, averaged over the full range of environments encountered by the species,

we estimate that some 10^30 or more generations would elapse before a bioF-like innovation that is paralogous to kbl could become established (Figure 11). This places the innovation well beyond what can be expected within the time that life has existed on earth, under favorable assumptions. In fact, even the unrealistically favorable assumption that kbl duplicates carry no fitness cost leaves the conversion just beyond the limits of feasibility (Figure 11). Using appropriate caution, we conclude not that paralogous evolution absolutely cannot have accomplished a functional jump like the one examined here, but rather that there is now a scientific case for doubting this particular jump to be evolutionarily feasible. At first glance, this claim may seem so modest as to verge on insignificance. Indeed, it could be interpreted as nothing more than a curious exception if the standard dogma—that paralogous evolution readily explains most examples of smallscale innovation—were well supported by the evidence. But as we have discussed, the many attempts to confirm that dogma have left it in question.

[T_aquaticus]

@Rumraket

But in that same paper we also see that the wildtype tHisA enzyme appears to have no measurable Trp+ activity in vitro (and does not support growth in media lacking Trp in vivo).

Did they run enzyme assays using the native or recombinant proteins?

[T_aquaticus]

@Argon

Just an aside: If you’ve got enough time to wait and especially if you go through enough different substrates, every enzyme is ‘promiscuous’.

That is worth repeating. It is rare to find an enzyme that is exclusive to a single molecule. In fact, I don’t even know of one.

“Activity” is also a moving target. If you put in enough enzyme and and have high sensitivity it is possible to find extremely low level, and probably biologically irrelevant, activity.

[Rumraket (Mikkel R.)]

Yes:
According to this table, at least within the detection limits of the method they could not detect any activity for Trp+ in the wild-type tHisA enzyme. That is not to say it is absolutely zero in reality, hard to know. But the reaction probably has some non-zero spontaenous rate of occurrence even in the absence of the enzyme.

[T_aquaticus]

@Agauger

NOW HERE’S THE IRONIC THING. NASVALL ET AL DEMONSTRATED THAT PROMISCUOUS ENZYMES CAN BE SHIFTED TO A NEW FUNCTION USING MUTATION AND SELECTION JUDICIOUSLY APPLIED, AND THAT DUPLICATION AND DIVERGENCE CAN LEAD TO THE SEPARATION OF THE TWO FUNCTIONS INTO SEPARATE GENES. YOU START WITH HIS A, YOU MAKE HIS A TRP F, THEN YOU SEE IF YOU CAN GET THE GENES TO DUPLICATE AND DIVERGE. AND THEY DO! WAS IT AN EVOLUTIONARY PROCESS? IT WAS A GENETIC TOUR DE FORCE, AN EVOLUTIONARY PROCESS DESIGNED AND GUIDED BY 4 GENETICISTS. ONE OF THEM THE BEST IN HIS FIELD.

I have to echo @Faizal_Ali sentiments. When @Agauger and Douglas Axe used the same or similar techniques they considered them valid techniques for testing evolutionary pathways. Gauger and Axe even used site directed mutagenesis which is a step more “designed” than naturally occurring random mutations used in the paper under discussion. The authors did not direct mutations to specific bases or genes. Rather, they looked for a trp+ phenotype and then determined which mutations in which genes were responsible for the new phenotype.

The criticism quoted above is way off base.

[Art (Arthur Hunt)]

The problems with @Agauger’s description of the study by Näsvall et al. starts with her description of the very first steps in the process. @Agauger asserts:

The story: Näsvall et al. created a gene that encoded an enzyme that was able to carry out two functions , but very poorly.

This sentence conveys a very different message from the experiment that was actually done. @Agauger, in claiming that a gene was created, implies that the authors performed the sorts of directed mutagenesis she is familiar with to design, assemble, and express the gene encoding the bifunctional enzyme (actually, several genes). The truth of the matter is that the authors did no such thing. Random mutation and natural selection were used to isolate a bifunctional enzyme. The authors played no role in directing mutations to specific loci, or otherwise designing the novel bifunctional enzyme.

@Agauger may wish to argue that I am mis-interpreting the word “create”, but, in standard lab parlance, this term means exactly what I say in the preceding paragraph. @Agauger wants readers to think that some sort of design went into the origination of the bifunctional enzyme, but there was no design in determining the nature of this enzyme. Just random mutation.

I also find @Agauger’s angry reaction to the way I described the study - specifically, a systematic testing of three mechanistic hypotheses - curious. My description of the study was spot-on, reasonable, and depicts the scientific approaches we take to the testing of models such as those put forth by Näsvall et al.

[swamidass (S. Joshua Swamidass)]

@Art

I also find @Agauger’s angry reaction to the way I described the study - specifically, a systematic testing of three mechanistic hypotheses - curious.

We need to be clear about this @Art. This is an example of Poisoning the Well executed by @Agauger (Poisoning the Well). Her reaction included several unwarranted ad homimems. Including a charge that you are intentionally misrepresenting the paper.

• “Hunt has posted a deceptive rebuttal”

• “Stop with the misrepresentations already. @ArtHunt You have people reading you who don’t know the material and depend on you for the truth. So don’t twist it or play word games.”

These are serious charges of intentional deception of an unknowledgeable audience. Unless @Agauger really means to accuse @Art of intentional deception, I hope she walks back those accusations. Disagreements are not deception. In this particular case too, @Art’s response seems spot on to several scientists who have read the paper. We do know the material. If he was lying, I would point it out.

@Art

My description of the study was spot-on, reasonable, and depicts the scientific approaches we take to the testing of models such as those put forth by Näsvall et al.

@Agauger also calls @art’s response “ludicrous.” @Art’s response is not ludicrous. It is exactly how most scientists reading the paper will understand the DI’s response. If @art (and the rest of us) missed something, I want to know. It certainly is not obvious to us. Some one needs to show us what we missed.

[Mercer (John Mercer)]

@T_aquaticus

Argon: Just an aside: If you’ve got enough time to wait and especially if you go through enough different substrates, every enzyme is ‘promiscuous’.

That is worth repeating. It is rare to find an enzyme that is exclusive to a single molecule. In fact, I don’t even know of one.

“Activity” is also a moving target. If you put in enough enzyme and and have high sensitivity it is possible to find extremely low level, and probably biologically irrelevant, activity.

I third this. This aspect of enzymology is well understood.

[Faizal_Ali]

Not by everyone, it would appear. Not even by everyone doing research on enzymology.

[Agauger (Ann Gauger)]

@swamidass @Art I have limited time to reply. I don’t know if my extensive quote from the methods section of the Nasvall paper was posted here or another thread. I am tol it was. These were the experimental manipulations used to carry out the experiment. The paper as it appeared in the journal did not include the methods in the main paper. They were in the Appendix. I assumed Art knew that and knew that saying they placed a bifunctional gene in a strain lacking his A and trpF function left out a whole bunch of stuff they did indeed create that bifunctional enzyme. It didn’t happened without som careful experimental work. To say it evolved is a very partial truth.
He and I, I am guessing, know what setting up the strains for the big evolutionary experiment involved. I was angry because at no point did Art acknowledge the engineering and experimental sophistication involved. For the non-geneticist observers, reading what he wrote would be misleading.
They would think that the bifunctional gene and the strain in which it was placed was completely natural, and that perhaps that all that was done was tochange the media. Not true.
@Art I apologize for my derogatory and excessive language. I was angry at what I perceived as deliberate intent to hide the necessary investigator involvement.

[swamidass (S. Joshua Swamidass)]

@Agauger

I don’t know if my extensive quote from the methods section of the Nasvall paper was posted here or another thread. I am tol it was.

It is here at the beginning of the thread. I will edit your comment to quote it out so it is clear.

@Agauger

@Art I apologize for my derogatory and excessive language. I was angry at what I perceived as deliberate intent to hide the necessary investigator involvement.

Thank you @Agauger. I appreciate this. I hope @art sees this too.

I do not believe @Art is lying. Nor do I believe you are lying either. This is a real gap in our understand that can be bridged if we can hash this out. @Art did not hide the investigator involvement. Rather, your interpretation of the investigator’s involvement and manipulation does not yet make sense to us. You are going to have to explain why if you want us to understand you.