Existence of RNA Genome or Fertility Selection?

| | 30 Comments | 3 TrackBacks

A new paper by Susan Lolle, Robbert Pruitt, and collegues at Purdue Univeristy is making some press. The NY Times covered it in a story entitled, Startling Scientists, Plant Fixes Its Flawed Gene. Carl Zimmer has discussed it on his blog: Move Over, Mendel (But Don’t Move Too Far). The paper is entitled, “Genome-wide non-mendelian inheritance of extra-genomic information in Arabidopsis.” It was quickly published in Nature this week, and is now available as an advanced access publication. I think Nature in its rush to publish cutting edge science has (once again) published a paper of fantastic claims that could use a little more support. I will explain below.

In a 1998 paper, the two main author identified and studied the genes that are involved in organ fusion in Arabidopsis thaliana. They subsequently noticed that one of their genes, hothead (HTH/hth) showed an extremely high rate of reversion from null (hth) to wild-type (HTH). The rate of reversion for several different null alleles, 0.040 - 0.082, was very much higher than expected from typical mutation rates.

They initially suspected that their stocks had been contaminated with wild-type seeds or that their plants were out-crossing with wild-type pollen. However, they genotyped embryos and demonstrated the existence of revertant embryos, which ruled out seed contamination. They next ruled out pollen contamination by demonstrating that pollen from an hth/hth plant can transmit the HTH allele.

Sequencing ruled out the involvement of transposons and repeated sequences. They also tested whether the gene had a higher mutation rate but ruled it out due to lack of silent nucleotide substitutions. They ruled out gene conversion through non homologous recombination because they found no suitable donor sequence in the genome. They also found that molecular markers were genetically unstable (i.e. high mutation rate) in hth/hth plants but not in HTH/HTH plants. This argued for a genome wide effect.

With all these options investigated and ruled out, the authors proposed that genetic information was transmitted not only in the genome but in an additional way. This secondary information was then restoring the HTH allele in some hth/hth individuals. The researchers proposed (double-stranded) RNA as the possible vehicle for this secondary storage.

I think that Lolle et al. missed one important mechanism when they were considering possible explanations: selection. Before I continue, I want to point out what else I think is missing in the study: direct sequencing of male gametophytes to determine their haplotypes, some statistical tests, hth alleles not due to point mutations, knowledge of the rate of silent mutation rates for wild-type hothead genes, and probably a few other things.

I prefer selection over the explanation proposed by the authors, since it is a well established mechanism for changing allele frequencies and it can explain why the reversions only occur in pollen. The selection I’m proposing is the result of pollen competition; i.e. HTH pollen fertilizes ovules better than hth pollen. With pollen having a much larger population size than ovules, there is a greater chance at producing a revertant through mutation, at which point selection can act upon it. HTH could also be important for ovule success, but the lower population size of ovules and less chances for competition would make the effect less likely to show up.

What about the fact that genetic instability was found in other markers in hth/hth individuals? Well, I suspect that the loss of a functioning hothead gene increases the global mutation rate, which affects those markers too. Perhaps this is the reason why HTH pollen is selected for; it puts the mutation rate back in order.

Why didn’t the researchers find silent mutations in revertant HTH? Well, if the revertant HTH is selected because it fixes mutation rates, then individuals that revert sooner than later may be preferred. The fastest way to produce revertant HTH is to reverse the point mutation that produced hth. The researchers might not have had the power to detect silent mutations in revertant HTH because they only sequenced three lines.

What about the fact that HTH/hth heterozygotes produce hth/hth individuals at nearly 25%? Doesn’t that exclude pollen selection? Well, the advantage of HTH pollen over hth pollen may be a paternally influenced trait that only shows up from hth/hth pollen donors. The advantage may also be frequency dependent and only occur when HTH is rare.

The paper by Lolle et al. involves a really interesting system, but I seriously doubt that their proposed explanation will hold up. I think selection, which isn’t discounted by their results, is a better explanation. It is really going to be interesting as more papers and other labs try to figure out what is going on in this system.


  • Lolle SJ et al. (2005) Nature 434: 505-509.
  • Lolle SJ et al. (1998) Genetics 149: 607-619.

3 TrackBacks

I am reminded why I don't read wingnut weblogs. It's not just that extremist Republican politics is vile, but that the politics is so often accompanied by other forms of insanity. Case in point: I discover via the World Wide Rant this remarka... Read More

Mendel overturned, or just more Darwin? Read More

I guess I'm blogging too much when I see a http://www.facultyof1000.com/article/15785770/evaluationhot paper mentioned at http://www.nationalreview.com/thecorner/05_03_20_corner-archive.asp#059172the Corner before I've heard about it from my colleagues... Read More


I think the paper discusses “reversion” frequency. In no place do the authors’ identify increased mutation freqency at any particular site. Additionally, I’m not sure the model proposed here can explain that the paper describes much higher reversion frequency for non-hth mutations in the hth background (they looked an allele of another gene, as well as silent SNPs). I think this rules out selection for HTH over hth.

The authors’ discussion of mutation is incomplete. They only tested whether HTH revertants were caused by a higher mutation rate. They discounted this due to a lack of mutations in the hth revertants. But there are some problems with this analysis.

1) They based their expectations on the high frequency of revertants being based solely on a high mutation rate. This would not discount an increased mutation rate and selection working together.

2) They only sequenced three HTH revertants. This gives them a low power to test for high mutation rates in hth if hth itself is the cause of the high mutation rates.

3) There was no HTH/HTH control for silent mutations. They don’t tell us what the wild-type rate of sequence mutation is for this gene.

4) They did not analyze the sequences of the unstable non-hth markers for silent mutations. Thus they did not discount that instability in their non-hth markers was due to higher mutation rates.

Wouldn’t be the first time that selection turned out to be behind an apparent violation of the normal rules of heredity.

How many pollen grains does an Arabidopsis flower produce? Say, that only 100 out of every 10^6 hth pollen grains from an hth/hth flower produced viable pollen. Further, let’s say that 1 out of every 10^6 hth grains 1 was a revertant to HTH. This would make it appear like HTH revertants appeared with a frequency of 1/100.

If hth pollen from HTH/hth flowers was basically normal, this would explain the 25% rate of production of hth/hth from HTH/hth - HTH/hth crosses.

Would the fact that Arabidopsis engages in selfing have any impact?

They only observed reversion, even at supposedly phenotypically silent SNPp. They never observed a de novo mutation, or did not state that they did. I think this is a very strong argument against selection. Spotaneous mutation coupled with incredibly strong selection is certainly normally high enough to revert a single base in the genome, but to revert basically other alleles outside of hth, as well as other SNPs (and only revert to one background or another, not generate new mutations) really argues that this is not selection and mutation acting together. I just don’t think you can get around it.


As far as I can tell from the paper, the authors didn’t analyze the sequence of the non-hth markers. They are not very clear in the paper on how these markers work, and I have never dealt with them nor had the time recently to research them. I suspect that these markers are not detected through sequencing but through blotting, giving a polymorphism of presence versus absense. Such analysis will be ill suited to detect whether the markers reverted their sequences to the exact sequence of Col or Ler.

They did detect HTH/HTH embryos from hth/hth plants even though they were rare. The selection argument doesn’t really work for ovules because there isn’t much competition between them for pollen.

Average pollen grains/flower = 156 Science, Vol 274, Issue 5292, 1535-1537 , 29 November 1996 That gets you ~50 seeds/fruit.

The markers are all CAPS markers: you PCR up a band and then cut with a restriction enzyme to get a different banding pattern. So you are basically looking at either (1) the frequency of mutation of the 6 base restriction site, or (2) reversion to the grandparent allele. If it’s (1) you are looking at a screamingly high mutation rate and you would expect to see more point mutations in the hth gene when they went back to sequence the revertants.

They actually sequenced 9 plants, 3 each for 3 separate lines. “To test this we determined the complete DNA sequence of the coding region of the HTH gene after three independent reversion events in each of three independently derived mutant hth alleles (hth-4, hth-8 and hth-10).”

I don’t really buy the “pollen competition in the hth/hth background would be invisible in the hth/HTH background” argument. I think you would see an effect like you see with this pollen mutant. … http://www.pnas.org/cgi/content/full/101/25/9502

and you should be able to see something in the +/- lines.

So, in summary, I don’t think your hypothesis is more likely than theirs. But your suggested experiments are all good except that it would be mega hard to sequence the individual male gametophytes. But more sequencing of other loci would be good. Additionally, I have to say that Nature & Science really make it hard to evaluate papers because of the size limitations of their articles. They should have sent this to PLOS Biology. It really is the way of the future.

I have to say it’s weird to be posting about science on a blog. I have taken a lot longer to write this than my usual blog comments, but a LOT shorter than anything I would write for a science publication. It makes me a little nervous, so if I have made a big boo-boo somewhere in the above, please forgive me. I would take more time, but I really should be working on that conference abstract. :)


Ivan Wrote:

They did detect HTH/HTH embryos from hth/hth plants even though they were rare. The selection argument doesn’t really work for ovules because there isn’t much competition between them for pollen.

It is true that selection couldn’t explain HTH/HTH embryos. I wish I knew what the rate was, but I don’t remember them saying it except to point out that it was rare. Actually, HTH/HTH embryos from hth/hth crosses would be a very good piece of data since it would be useful in dissecting the effects of mutation from the effects of pollen selection.

If it’s (1) you are looking at a screamingly high mutation rate and you would expect to see more point mutations in the hth gene when they went back to sequence the revertants.

Not really. That expectation assumes that hth is not the cause of the high mutation rate. Once hth reverts to wild-type, the mutation rate settles down to the wild-type rate, and the gene non longer experiences high mutation rates. If revertants are selected for, then the fewer mutations that hit that gene the better. This is where the experimental design of the paper suffers. They didn’t control for effects of selection. What they should have done is also sequence non revertants to see how they are accumulating silent mutations.

I don’t really buy the “pollen competition in the hth/hth background would be invisible in the hth/HTH background” argument.

What don’t you like about “paternal effect” or frequency dependent selection? I would really like to see some studies in which they varied the pollen proportion of HTH in the pollen population. That could help establish if it is a frequency dependent effect.

I think you would see an effect like you see with this pollen mutant . …

Can you be a little more specific?

So, in summary, I don’t think your hypothesis is more likely than theirs.

I think my main point is that the authors need to have considered more complex (i.e. realistic) evolutionary genetic situations before they proposed an alternative method of molecular heredity. I think the most important data that is missing is the actual biochemical function of the hothead gene. Knowing that would really help to figure out what is occurring in the double nulls.

I do agree with the latest comment. I think that the paper is provocative enough that I am glad they published it, though. I feel that this kind of paper is interesting enough that by being published it can convince people to look into results that don’t necessarily make sense. The papers that are published by bigshots that are completely crap (I’m thinking about a particular journal family- one can guess) are the ones that bug me the most. This paper might end up being new biology or a really interesting artifact, either way I will be keeping tuned.

“Existence” not “Existance.”

We have obtained similar results on Arabidopsis with Muller’s mutation in 2003


Results was published: П.П.Даряев, Е.А.Леонова, 2003, Странный мир волновой генетики. Журнал «Сознание и физическая реальность», т.8, №6, с.27-40. Peter Gariaev, PhD. Moscow, Inst. Quantum Genetics. http://www.self-managing.net/genetica/


Any new development on this? No more comments since Apr 1?

I haven’t been watching the literature. (I’m very bad about that.) I have gone back to read some earlier studies on the gene and it does seem likely to me that it serves to break down a chemical that is in a class of chemicals that are mutagens.

Dr Cartwright

I’m still trying to figure out the Lolle et al paper myself, but I see one flaw in your argument for selection:

Where are you getting the wild type allele that is being selected for?

I have no problem with selection being used to explain an increase in the frequency of the allele at the level of pollen health, but you seem to have forgotten to explain where the selected allele came from.

Is your problem with the reversion event itself, or their explanation?

Granted, they are missing some very important controls, like you and several others have pointed out. However, reversion has been noted in other situations (see citations in paper, and the post by Peter Gariaev), so that adds some credibility, even if their methods were not perfect.

Other than the authors’ suggestion, how could the allele revert from mutant to wild type? I agree, they didn’t screen broadly enough (or at least they didn’t show and discuss the data) for other silent mutations, and they don’t know the normal frequency of silent mutations in that gene. They also never looked at the non-reverted individuals to see if they had changed to nucleotides other than the mutant or revertant (ie. if there were non-reverting mutations at the point mutation). If there were (and if so, there would probably be other silent mutations), that would completely explain the entire problem.

I believe that we have to take things in the ‘high impact’ journals with a grain of salt - consider the papers claiming that hematopoietic stem cells (and other stem cell types) ‘transdifferentiated’ into functional mycoardium…this was later disproven when the recipient and donor mice each had distinguishable markers (GFP in one, lacZ in the other), showing that the HSCs were fusing with existing cardiomyocytes (Nygren JM, Jovinge S, Breitbach M, Sawen P, Roll W, Hescheler J, Taneera J, Fleischmann BK, Jacobsen SE. Bone marrow-derived hematopoietic cells generate cardiomyocytes at a low frequency through cell fusion, but not transdifferentiation. Nat Med. 2004 May;10(5):494-501. Epub 2004 Apr 25.) (see paper for examples of the papers claiming transdifferentiation)

At the same time, we skeptic scientists have to open up our minds to possibilities that would necessitate rewriting textbooks. How many times has the model of the atom been changed? Don’t jump up and start rewriting now, wait until the theory has been proven, or disproven. But do be open.

Dana, I’m not a Dr., just a PhD student.

The wild type allele is the product of mutation. The system is more complicated than I first thought when I proposed fertility selection.

I think one of the functions of the hth gene is to break down an aromatic chemical that causes mutations. When hth is broken, this chemical builds up during some life stage and causes problem. I suspect that the life stage may be tied to pollen development, but I can’t say for sure without more knoweldege of the system.

I, also am a PhD student. I did my undergraduate work in plant genetics/plant breeding, and worked in canola breeding for several summers. I’m now working on developmental cardiac genetics.

I have read a bit of Petr Gariaev’s work, via a rough translation/abstract he supplied me with, and some figures with a bit of explanation. Essentially, the Russian work is suggesting that DNA has quantum mechanics-compatible qualitites - ie. electromagnetic signals, in addition to the actual nucleotide sequence. Like the Lolle et al work, this seems pretty far fetched, and I’m not at all familiar with quantum mechanics physics, so I really can’t evaluate the claim. But the idea behind it would suggest that the reversion events that Lolle et al are seeing would be due to this electromagnetic memory of what nucleotide should be in a particular position.

So this would be an alternate explanation to the RNA cache. I don’t know if it’s any more plausible than an RNA cache or mutation. Perhaps it’s more of a directed mutation.

Have you heard of the ‘wave genetics’ theory before? I’m hesitant to make much out of a topic when i can’t read the primary literature, as most of the early work seems to be in Russian. On the other side of the coin, there seems to be so many questions that we aren’t finding answers for by simply looking at the genetic code, and some of the answers we do find are completely unexpected. Perhaps we need another depth to DNA/genetics. I don’t know.

A note about the Lolle et al paper:

The chimaeric RNA-DNA oligonucleotide-mediated site-directed mutagenesis: papers have been rather torn apart in the literature. Can’t be replicated consistently, or can’t be replicated at all. See Science 298:2116 (The Strange Case of Chimeraplasty).

The comments on this article show up in a PubMed search, why would they cite something that has been discredited to that extent, and so obviously???

me, a phd student too. Have been discussing with buddies a lot on the paper. I guess many of us would propose selection model at a first glance but may hesitate doing so after a while. I am not really familiar with the field. However, what i wish to throw in here: recent progresses on RNA world may well remind us to be fully open-minded and prepared to embrace what might be astonishing or revolutionary. It seems the mother nature is way smarter than we have seen and is determined to surprise us again and again and again. At least central dogma is no longer as true as it was. By the way, I am currently hunting for a postdoc position and RNA world is one territory I like to explore. The more I read about the RNA world, the more I think this paper may be telling us something even though I would still be as surprised when follow-up tells us they were right! I’ll be watching on this and really look forward to follow-ups–true or false.

humm, i think its really the best paper and best narrations of this topic too.

humm, i think its really the best paper and best narrations of this topic too.

About this Entry

This page contains a single entry by Reed A. Cartwright published on March 23, 2005 8:44 PM.

Second Paper was the previous entry in this blog.

HOTHEAD is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.


Powered by Movable Type 4.37