Showing my age a bit, a young Chris Rock in a 1988 movie amusingly asks: “How much for one rib?” Given that the crafting of a single protein plays a central role in this post, and ribs are a source of protein, the association was too much for me to pass up in the title.

I’ve pointed out before that our most elaborate inventions absolutely pale in comparison to even the simplest form of Life. Our gizmos can’t self-replicate, heal wounds, feed themselves, stave off pathogens, or self-evolve. Even though both gadgets and Life appear to be based on atoms and the same fundamental interactions, the level of complexity in Life is far beyond our means to create. At best, we bootstrap and copy.

To make the point, we’ll embark on a well-funded thought experiment that is able to assemble the top talent from around the world in a team given one mission: generate the genetic coding that would carry out a specific novel function by way of synthesizing a novel protein specific to that task. We stipulate a novel function that hasn’t arisen in any lifeform, otherwise the open-book (Google-connected) nature of this test would instantly result in “cheating” off a billion-year heritage.

Let’s see how they do.

Major Caveat

The whole premise of this thought experiment suffers a major flaw, in that the world’s experts could not be tricked into pursuing such a mission. They know better, for all the reasons sketched below (and more). So, I want to be clear that this post isn’t at all “hey, look how dumb the experts are” as much as “we don’t possess the most rudimentary capability next to biological evolution.” The real experts are not in disagreement.

Basic Idea

Getting on with the fantasy, I should first provide a crude primer on the crucial role of proteins, to my limited knowledge. Proteins provide specialized structures, defense, communication, mobility, and facilitation of chemical reactions (catalysis). For the purposes of this post, I’ll focus on the last function, in which case we call the proteins enzymes. These magical beasts promote and accelerate reactions that would either proceed too slowly or not happen at all on energetic grounds without substantial assistance. Some enzymes break down certain molecules—like lactose, for instance. Other enzymes build molecules out of parts, which I select here as the main example for the post. Similar arguments would apply to all other functions as well.

Proteins achieve their primary functions based on shape (how the long polypeptide chains fold and curl) and electrostatic affinities of their exposed surfaces. The building enzymes are configured to attract certain molecules like puzzle pieces fitting into customized pockets of the perfect shape and size. A new molecule is synthesized by the protein grabbing two (or more) different components out of the cytoplasm, holding them in the appropriate orientation relative to each other, and putting them next to each other so that they can attach.

In this sense, enzyme proteins act like a romantic matchmaking service: identifying compatible pieces who might otherwise never find each other (or if they did, might be too shy to talk to each other) and putting them together in a happy couple. Other enzymes, of course, trigger divorce.

DNA contains genes, and genes perform a similar trick to proteins, although in an unbelievably-elaborate multi-stage dance. Particular 3-letter chunks of …GATTACACCAGACTA… are matched to amino acids (via transfer RNA adapters), so that they build complex proteins by sitting amino acids in chairs right next to each other until they kiss. We’ve impressively reverse-engineered DNA/RNA to this point: we can appreciate how they build the proteins they do, and we can appreciate how a particular protein selects particular components and puts them together (all physics-based). Likewise, I can understand how a symphony works by watching the bowstrings and puffs while hearing the unique instrument contributions. Asking me to compose and play a symphony from scratch is an entirely different story. Bring ear plugs.

Ground Rules

Okay, so we’re going to ask our imaginary crack team of scientists to design a stretch of DNA (constrained to its four base pair “letters”) capable of assembling a novel enzymatic protein (constrained to a small set of amino acid building blocks) not yet encountered in any biological context. Maybe its job is to stick ammonia onto methane. But this is a poor example, because Life already figured this one out. To steer well clear of what Life has already conjured, let’s have it attach yellowcake (H₈N₂O₇U₂) onto dipotassium titanium hexafuoride (F₆K₂Ti). I don’t know: these may not be “stickable,” but the exact pairing is not crucial to define in this thought experiment: something could work that Life has not stumbled upon or needed yet. We just want two molecules that current protein libraries don’t assemble to prevent cheating off Life. That way, we can make the test “open book” without worry.

Actually, scratch that prescriptive approach. To be more Life-relevant, we start by defining a novel function the cell is to accomplish. Maybe the cell needs to import aluminum atoms through the cell wall, so we need to come up with a molecule (or set of molecules) that would accomplish that targeted task—and therefore require the protein(s) that could synthesize that molecule(s) out of available materials. This is probably already too complex as an example, requiring multiple interacting genes. But you get the point: we want a particular functional outcome at the cellular level that requires a particular molecule/structure that in turn requires a specialized protein (or set of them) that requires a stretch of DNA to carry out the protein synthesis. This is what Life does: establishes bridges across the multiple gaps between final function and construction of the various assemblies that execute the plan.

No Go

Life has done this trick millions of times. It’s old hat. How will our well-paid team do on just one example?

Well, the task would be hopeless to design based on knowledge/understanding/brain-work (which is why no experts would cooperate in real life). The first major hurdle is figuring out what novel structure would carry out the intended task. The next barrier is how to assemble that structure out of available compounds. Then one needs to define what relative orientation of the two components would stick in such a way as to accomplish the desired function. We still need to devise a new protein structure that would attach to some part of component A (in an appropriate place) and the same for component B. Next, it has to twist in the right way—possibly triggered by the successful acquisition of A and B—so-as to stick them together in a compatible way. Of course, this protein operates under the serious constraint of having to use a restricted menu of available amino acids as its building blocks.

It’s pretty awesome what Life achieves! And I’ve glossed over tons of intricate twists in the story. If reading about the process for the first time as part of a sci-fi novel, I wouldn’t believe it to be possible: outlandish fantasy.

Once down to amino acid configurations, figuring out the DNA sequence to assemble those amino acids is basically already in hand. But even this is a way of cheating: bootstrapping off an already wildly-complex mechanism for synthesizing proteins. If relying on basic first-principles understanding, just designing that one step would be a a hundred bridges too far.

That’s Just One

Consider that this unachievable task is just to build one protein. Billions of dollars couldn’t do it. All the king’s horses and all the king’s “men”…

Now, humans are based on 20,000–25,000 genes. The range of uncertainty alone tells us something, right? Even a “simple” amoeba has 13,000–15,000 genes.

If we can’t make a single gene to construct a novel protein that catalyzes a particular reaction to create a molecule that carries out a specific task, think how impossibly hard it would be to make thousands from scratch! Granted, each would not be as hard as the first, but new challenges would continually crop up all along the way.

And They Interact

Of course, the task of creating a viable living being from scratch is not as “simple” as churning through thousands of novel genes. The functions of the proteins interact with each other. The products of the proteins interact. The environment interacts. Other organisms interact. Just consider that two of your novel proteins might stick to each other unintentionally, incapacitating both.

I’m having trouble even finding verbiage that remotely captures how enormous this complexity becomes. Every gene serves some adaptive purpose that has been worked out by long trial in the full, unredacted context of all these interactions. The number of relevant interactions between ten thousand proteins and their products easily stretches into the millions. It would be impossible for us to anticipate all the goofy interactions that take place in a real system of this complexity if starting from a blank slate—many of which would cripple or kill the resulting organism, resulting in non-viability.

Basically, for every intended consequence (function, via proteins), we’ll find numerous unintended consequences. These cascade and self-amplify as complexity increases. Only patient trial and error—not brains—can sort it all out, step by step.

Regulating

As if not hard enough already, only 2% of DNA is dedicated to genes (protein-coding). The vast majority contributes to regulation of those genes. If the cell was constantly running full-throttle, churning out every protein in its library willy-nilly, it would be absolute pandemonium (and not a viable lifeform; the stomach wouldn’t know what to do with all those retinal cells!). It would be like every tool in a giant hardware store performing “work” on a car at the same time. Nothing good comes of that. Get those saws away from the vacuum hoses!

The cell needs to selectively turn on or off specific combinations of protein generation depending on what it’s trying to accomplish in that place and time. This is a sort of decision-tree programming. How does the cell know when to seek food? How does it know when to initiate mitosis? How does it know when to slow metabolism? I don’t know enough to fill out the thousands of questions a microbe must answer based on its varying circumstances—let alone the correct answers or how to program them (they’re geniuses, I tell you!). But all of this “knowledge” is written in DNA, mediated by various sensors and structures that the DNA knows how to build and maintain. The complexity is truly staggering, but that’s what billions of years of patient and open-ended experimentation gets you. Don’t foolishly try to fit it in your brain.

Thus, even if we were able to generate DNA sequences capable of coding tens of thousands of proteins, each with a specific functional duty that in some cases is a few steps removed (e.g., through the synthesized products of the proteins), and each interacting in compatible ways, the next step of regulating all this activity is itself immensely difficult and well beyond our capability.

Absolutely Hopeless

To summarize, the best talent in the world could not concoct a genetic sequence to make a single novel molecule in service of a particular functional goal. This is just one of over ten thousand genes needed for something like an amoeba. The genes and their protein products and the products of the protein products interact with each other and with the external world (including DNA/proteins/products in other organisms). All of this is regulated by genetic code that dwarfs the protein coding portions. And the result is a viable lifeform capable of self-replication, metabolism, food acquisition, evasion, repair, defense, and lots more, I’m sure—all in relation to many other organisms being their own irrepressibly-quirky selves.

Thus, I claim there’s no way that we could replicate even the barest of Life’s achievements. Our artificial gadgets and machines—no matter how sophisticated they appear to us and how proud we are to have invented them—are light years away from even the simplest living being. (They tend to employ just a few straightforward tricks, brute-force replicated millions of times.) Even working backwards surrounded by complete and functioning templates, we’re scrambling to understand the most basic aspects of Life.

Yet, we find that we do possess the asymmetric power to destroy life and permanently erase species from the planet. What took millions (and really billions) of years to hone is undone in decades. At least that’s something to make us proud.

What Doesn’t Follow

Clearly, I believe Life to be amazing and that it’s far, far, far beyond our capabilities to create anything even remotely as incredible. Keep that in mind when I say that I also believe living beings to be arrangements of atoms adapted to the universe, using the same physics everything else does, including our pathetic machines. To me, it’s all the more amazing that Life managed to solve this enormous list of difficult problems employing the same atoms and physics as everything else: just lying around and available for use.

I mean, we can at least catch glimpses of how atomic arrangements and electromagnetic interactions conspire to store information, build proteins, regulate protein activation, and use protein shapes to catalyze reactions—all without requiring new physics. We immediately get lost in complexity after that, though. All the same, we can appreciate the presence of the complexity, as the receipts are present in the form of DNA coding and the menagerie of resultant functional proteins. It’s just that interactions quickly overwhelm our cognitive capabilities, even when allowed to write stuff down [the subject of next week’s post].

The goal, here, is deep humility, a form of awe, and a sense of universal connectedness. As stupendously breathtaking as Life is, everything uses the same stuff and the same rules. We share a deep kinship with all matter, then. We are as dirt, and if that’s not grounding, I’m not sure what would be.

I thank Nigel Goldenfeld for taking a look at a draft and offering useful comments as an expert in biophysics.

Views: 280