This is the third post in a series covering my lab’s recent virus datamining efforts. Here are the first two posts.
In 1809, Jean-Baptiste Lamarck proposed an evolutionary model in which parents pass acquired traits to their offspring. Classic examples are the idea of a weightlifter passing down bigger muscles to his sons, or the idea that each generation of giraffe ancestors spent its days stretching its neck to reach low-hanging tree leaves - such that each generation of giraffes developed progressively longer necks. Now that we understand how DNA works, Lamarckism sounds flatly ridiculous. Everybody knows Darwin was right and natural selection runs on the substrate of random mutations that, by chance, happen to incrementally improve fitness.
But do we really know how DNA works? Some of the little demons I’ve been battling down in the datamines have got me wondering whether there might be some cracks in the foundation of our understanding. Cracks that could make room for Lamarckism to sneak back in a little.
Trigger warning: this article is sheer speculation. If your natural reflex is to think that sounds unscientific, I urge you to read “Speculation is not a Dirty Word.” The hypothesis I’m wrestling with here offers experimentally testable predictions. The point isn’t to convince you the hypothesis is true, the point is to convince you the experiments are worth doing.
Before we dive into basic molecular biology, let’s first zoom out to an interesting macro phenomenon known as domestication syndrome. I recommend Peter Gray’s excellent coverage of the topic over on his mandatory must-read ‘stack,
. The story begins in 1959, when Dmitri Belyaev and Lyudmila Trut began a series of experiments in which captive foxes were only allowed to breed if they were willing to approach a gloved human hand without biting it. Within six generations the experiment produced friendly tame foxes that shared a full suite of distinctive physical traits with domestic dogs (floppy ears, rounded skull, curly tail, mottled fur, etc). It’s a bit surprising that the change in behavioral characteristics tracked with seemingly unrelated somatic traits1, but the truly amazing thing about Belyaev and Trut’s results is the simple fact that a major behavioral shift occurred in only six generations. If evolution depends entirely on random events like some stray cosmic ray hitting some particular spot in the genome, the process should have taken way longer than six years.Belyaev’s foxes started out as a bottlenecked inbred lineage that can be traced back to perhaps as few as two founding breeding pairs. The lack of genetic diversity makes it unreasonable to invoke the standard Darwinian concept of standing genetic variation. In other words, it’s unlikely the experiment was simply expanding a rare outlier genotype that already existed in the starting population of 130 captive animals. The less orthodox (but much simpler) explanation is that a single uniform genotype rapidly mutated toward tameness. Astonishingly rapid evolution has also been documented in inbred rats and fruit flies.
One area where Lamarckism has already attempted to sneak back in a little is a field called epigenetics - in which a specific DNA sequence is decorated with reversible methyl groups that regulate gene expression. While epigenetic effects are important in the somatic tissues of postnatal animals, I’ve always found the idea of heritable epigenetic effects pretty far-fetched. There’s a stage during early embryonic development where nearly all epigenetic marks in the genome are erased. I’m not aware of any heritable epigenetic effects in mammals being worked out in mechanistic detail - meaning a map of exactly which epigenetic mediators are activated in the parental germline cells and exactly how the epigenetic marks persist in the face of early-embryo erasure. The field seems to rest on the starting assumption that effects across a single generation can’t possibly be genetic, so nobody’s really looking for ordinary genetic explanations. I’m worried we’ve been drilling a dry well.
My hope is that understanding the molecular biology of simpler organisms like viruses can provide us with some alternative hypotheses to test. In particular, I’ve been contemplating an obscure family called anelloviruses. In contrast to cellular DNA, which is double-stranded, anelloviruses have a single-stranded DNA genome. Anellovirus genomes are so small they don’t have room to encode any proteins known to be involved in DNA replication. For geeks like me, having no idea how anelloviruses replicate their genomes is an abominable mystery.
The #1 dumbest blunder of my scientific career was the summer when I lazily neglected to check back in with Tom Cech to find out whether his lab might have an opening for me to pursue my undergraduate research. Aside from missing out on the opportunity to have a Nobel laureate on my CV, I also missed out on learning firsthand about Cech’s amazing work showing that RNA strands can fold in a way that makes them capable of catalyzing basic chemical reactions, including synthesizing new RNA strands. I occasionally kick myself for the missed brush with greatness - and a useful effect of the kicking is that it helps keep Cech’s general ideas in working memory. The new thinking for the case at hand is simply to wonder whether single-stranded DNA might be doing some of the same things as Cech’s original RNA “ribozymes.” For anelloviruses, the idea is that the viral DNA might auto-catalyze a strand break that could help prime its own replication. DNA-based site-specific “nicking” ribozymes of the sort I’m invoking have previously been developed in the lab. If humans can develop DNA ribozymes from scratch on a scale of years, then surely nature must have at least occasionally hit upon such mechanism over geologic time scales2?
The concept of a self-manipulating DNA molecule opens the door to new classes of inducible genetic switches. One example of an inducible switch is found in bacteria, which use a class of proteins called invertases to reversibly rearrange certain promoters3. In principle, a bacterium can sense an environmental stimulus and then, in response, activate a specific invertase that selectively mutates a specific patch of the genome in a way that stably gives the bacterium, and all its progeny, a predictable new phenotype. It’s Lamarck’s giraffes in microcosm. Although rapidly evolving fruit flies develop specific chromosomal inversions in response to climate variations, mammals aren’t known to have any invertase-like genetic switches or other protein-mediated site-directed mutagenesis mechanisms. If we invoke the idea that hidden ribozymes are doing these types of jobs I doubt anybody would have noticed. I’ve certainly never heard anybody speculate about it. To invert Pasteur’s famous saying, chance doesn’t favor the unprepared mind.
Now that the little demons of the datamines have prepared my mind, the initial experiments testing the hypothesis suddenly seem incredibly simple. I’m gearing up to test the prediction that single-stranded anellovirus DNAs can cleave themselves in a test tube. Colleagues are teaching me the art of long-read whole genome sequencing, which should enable us to detect the inversions or site-specific mutations the ribozyme hypothesis predicts we’d find in inbred mammals undergoing rapid evolution.
Scientists, like journalists, are famous for worrying about getting scooped. There’s none of that here. I’ll honestly feel a little relieved if somebody who already has appropriate samples and/or stronger bioinformatic skills can solve this puzzle first. Because I can see an insidious dark underbelly to the whole line of thinking4. If we find out there are inducible genetic switches that can shift placental mammal lineages from feral to domestic behavior/morphology in six generations then all hell’s gonna break loose. Humans are placental mammals. The discovery wouldn’t just resurrect Lamarck, it could accidentally resurrect phrenology. The genetically definable existence of feral-shifted and domestic-shifted individuals would have profound implications for everything from psychology to geopolitics. My advice to anybody fixin’ to scoop me is to please also start honing your social media skills. You’re gonna need ‘em.
The idea that changes in body morphology coincide with selection for tame behavior isn’t necessarily all that surprising. Darwin noted the suite of distinctive morphological traits commonly observed in domesticated animals, and there’s now a solid body of evidence establishing a plausible mechanistic explanation that ties everything back to how neural crest stem cells migrate from place to place in early embryos. Here’s some popular coverage of that idea. I can’t think of any reasons why floppy ears might benefit a friendly fox, but it’s possible some of the other changes associated with domestication syndrome are evolutionarily adaptive. Friendly/egalitarian/domestic bonobos are more gracile than mean/hierarchical/feral chimps. The correlation makes teleological sense. Why waste lots of energy growing bulging biceps if all you’re gonna be doing is sharing fruit and playing with your friends all day.
Cech’s work inspired the “RNA world” hypothesis, in which the precursors of living organisms consisted of self-replicating RNAs with a variety of enzymatic activities. But what if it was actually DNA all along? Maybe RNA ribozymes only became necessary after DNA became double-stranded (thus occluding its original enzymatic functions)? Maybe anelloviruses are a distant echo of DNA’s pre-biotic heyday?
Promoters are patches of the genome that regulate the expression of nearby genes. The interesting thing about “invertons” is that bacteria can use them to reversibly toggle the regulatory region on and off like a light switch.