05
end is content in the presence of water, while the other wants to avoid it
06
entirely.
07
The lipid’s search for contentment is a metaphorical way of talking
08
about the fact that the system evolves so as to minimize free energy. En-
09
tropy increases, which suggests to us a certain emergent vocabulary, in
10
which the molecules “want” to find a state with low free energy. The arrow
11
of time leads us to speak a language of purpose and desire, even though
12
we’re only talking about molecules obeying the laws of physics.
13
The one thing that the hydrophobic carbon tails can do is to seek com-
14
fort in the company of their own kind. The lipids can line up next to one
15
another, so that their tails are all surrounded by other, equally hydrophobic
16
compatriots, rather than by water. There are a few different ways this can
17
happen. The simplest is for the lipids to form a little ball, called a micel e, 18
with the hydrophilic heads on the outer surface, exposed to water, while the
19
hydrophobic chains are bundled up with one another.
20
21
Fatty Acid
Phospholipid
22
23
Hydrophilic
24
Head
25
26
27
Hydrophobic
28
Hydrocarbon
Micelle
29
Tail
30
31
32
33
Bilayer
34
S35
N36
255
Big Picture - UK final proofs.indd 255
20/07/2016 10:02:49
T H E B IG PIC T U R E
01
There is another option: a bilayer— two sheets of lipids, each one of
02
which has the hydrophilic heads pointing in the same direction, with the
03
hydrophobic tails of the two sheets clinging together. That way the heads
04
get to enjoy the water they seek out, while the tails are completely shielded
05
from it.
06
In an aqueous ( water- containing) solution, lipids will spontaneously or-
07
ganize themselves into one of these types of structures. Which one depends
08
on the circumstances: on what kind of lipid we’re dealing with, and on
09
other properties of the solution, especially whether it is more acidic (likes
10
to give away protons and accept electrons) or alkaline (the opposite).
11
Examples of lipids include fatty acids, which are relatively simple, and
12
phospholipids, which are a bit more complicated. Fatty acids are every-
13
where in biochemistry— they are one of the fuel sources that mitochondria
14
can use to make ATP, for example. Phospholipids consist of two fatty acids
15
joined together by a phosphate group (a compound of phosphorous, carbon,
16
oxygen, nitrogen, and hydrogen).
17
The cellular membranes in organisms living on Earth today are made
18
from bilayers of phospholipids. These molecules very naturally self- organize
19
into bilayers, but not into micelles, because their double tails are too thick
20
to easily fit into the ball- like micelle configuration. The bilayer membranes
21
then fold into themselves to form spherical enclosures, known as vesicles.
22
That’s the easiest part of assembling a cell.
23
•
24
25
One problem with phospholipids, as far as the origin of life is concerned, is
26
that the bilayers they construct are just too good at their job. They are fairly
27
impenetrable, with only water and some other small molecules able to pass
28
from one side to another. It therefore seems likely that the earliest form of
29
cellular membranes were actually made of fatty acids rather than phospho-
30
lipids. Once they are put in place, evolution can set about improving them.
31
Fatty acids can self- assemble into bilayers, but only under the right con-
32
ditions. In highly alkaline solutions, fatty acids prefer to form micelles; in
33
highly acidic conditions, they glom together into big oily drops. At inter-
34
mediate levels of acidity, their favorite configuration is a bilayer. It’s a phase
35S
transition, governed by the acidity of the surrounding medium.
36N
These bilayers of fatty acids don’t relax into long two- dimensional
256
Big Picture - UK final proofs.indd 256
20/07/2016 10:02:49
S P O n t A n E Ou S OR g A n I z A t I O n
sheets, like a piece of paper. Rather, they quickly pinch off and form little
01
spheres. That’s the configuration with the lowest free energy in that envi-
02
ronment. It’s another manifestation of how, rather than smooshing every-
03
thing into featureless goo, the second law helps create the kind of organized
04
structures that are useful for life.
05
Fatty acids are relatively simple molecules, so it wouldn’t be hard to find
06
them in appropriate environments on the prebiotic Earth. What’s more, the
07
membranes they form are more permeable than those made of phospholip-
08
ids. That’s good news for early life. In a mature organism, you don’t want
09
chemicals leaking willy- nilly into and out of your cell; embedded in the
10
membranes are very specific structures (like ATP synthase) that guide nu-
11
trients and energy sources in and out as appropriate. Early on, before such
12
dedicated mechanisms have evolved, what you’re looking for is some-
13
thing that can do a fairly good job of compartmentalizing the chemical
14
precursors of life, but not such a good job that they are isolated from the
15
outside world and essentially choked to death. Fatty acids seem just right
16
for the task.
17
18
•
19
From the perspective of a poetic naturalist, one of the most interesting fea-
20
tures of spontaneous compartmentalization is how it lends itself readily to
21
an emergent description of the system. Without compartments and mem-
22
branes, we’re faced with a soupy mess of compounds, energy sources, and
23
reactions. Once a boundary forms between different kinds of stuff, we can
24
readily talk about the “object” (inside the boundary) and its environment
25
(everything outside). The boundary— whether it’s literally a cell membrane,
26
or the skin or exoskeleton of a multicellular organism— both helps the
27
structure take advantage of the free energy around it and helps us talk
28
about it in useful, computationally efficient ways.
29
Karl Friston, a British neuroscientist, has suggested that the function of
30
biological membranes can be understood in terms of a Markov blanket, a
31
term coined by statistician Judea Pearl in the context of machine learning.
32
Imagine we have a network: a collection of “nodes” connected by lines. A
33
“Bayesian network” is a graph formed from nodes that can send, receive,
34
and process information, like computers on the Internet or neurons in a
S35
brain. If we pick out any one node, its Markov blanket consists of all the
N36
257
Big Picture - UK final proofs.indd 257
20/07/2016 10:02:49
T H E B IG PIC T U R E
01
nodes that can directly influence it (its “parents”), plus all the nodes it can
02
directly influence (its “children”), plus all the nodes that can also influence
03
its children (its “spouses,” of which there may be many).
04
This complicated- sounding construction captures a simple idea: given
05
some part of the network, the Markov blanket captures everything
06
you need to know about its input and output. There may be an enormous
07
number of possible internal states of the nodes, but all that matters for the
08
operation of the network is what gets filtered through the Markov blanket.
09
A cell membrane, argues Friston, can be thought of as a Markov blan-
10
ket. Many intricate processes go on inside the cell, and many things are
11
happening all the time in the environment outside. But communication
12
between the two is mediated through the cell membrane. Under these con-
13
ditions, the system evolves toward a configuration in which the cell mem-
14
brane is robust— it maintains its configuration, even in the presence of
15
( not- too- large) perturbations from inside or out.
16
17
External
18
Environment
19
20
21
22
Markov
Internal
23
Blanket
States
24
25
26
27
Stimuli & Responses
28
29
30
This theory was originally developed not for individual cells but as a
31
way of thinking about how brains interact with the outside world. Our
32
brains construct models of their surroundings, with the goal of not being
33
surprised very often by new information. That process is precisely Bayesian
34
reasoning— subconsciously, the brain carries with it a set of possible things
35S
that could happen next, and updates the likelihood of each of them as new
36N
data comes in. It is interesting that the same mathematical framework
258
Big Picture - UK final proofs.indd 258
20/07/2016 10:02:49
S P O n t A n E Ou S OR g A n I z A t I O n
might apply to systems on the level of individual cells. Keeping the cell
01
membrane intact and robust turns out to be a kind of Bayesian reasoning.
02
As Friston puts it:
03
04
The internal states (and their blanket) will appear to engage
05
in active Bayesian inference. In other words, they will appear to
06
model— and act on— their world to preserve their functional
07
and structural integrity, leading to homeostasis [preserving
08
stable internal conditions] and a simple form of autopoiesis
09
[maintaining structure through self- regulation].
10
11
This is a speculative and new set of ideas, not an established picture of
12
how we should think about the function of cells and membranes. It’s worth
13
remarking on because it shows how the concepts we’ve been talking
14
about— Bayesian reasoning, emergence, the second law— come together to
15
help explain the appearance of complex structures in a world governed by
16
simple, unguided laws of nature.
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
S35
N36
259
Big Picture - UK final proofs.indd 259
20/07/2016 10:02:49
01
02
03
32
04
05
The Origin and Purpose of Life
06
07
08
09
10
11
12
13
14
15
16
17
On a crowded flight to a conference in Bozeman, Montana, I was
reading some research papers on the connection between statistical
physics and the origin of life. The man sitting next to me glanced
over at them, curiously. “Oh, yes,” he offered, “I know that work well.”
18
Over the course of a career as a physicist, you run into people who have
19
theories of how the universe works, and are eager to share them with you.
20
Those theories are rarely very promising. Presumably the study of life at-
21
tracts similar numbers of garrulous enthusiasts. But we had a long flight
22
ahead of us; I asked him what his thoughts were on the matter.
23
“That’s easy,” he replied with a nod. “The purpose of life is to hydroge-
24
nate carbon dioxide.”
25
It wasn’t the answer I was expecting. I had been fortuitously seated next
26
to Michael Russell, a geochemist at NASA’s Jet Propulsion Laboratory,
27
close to my own home institution of Caltech. It wasn’t a complete
28
accident— he and I were both traveling to give talks at the same conference.
29
Russell, it turns out, is a leading (if somewhat iconoclastic) figure in the
30<
br />
study of life’s origin, and one whose approach is especially physics- friendly.
31
We got along fine.
32
Russell is one of the leaders of a faction in the origin-of-life debates who
33
believes that the first crucial step was the appearance of metabolism. This
34
camp imagines that the crucial event was the appearance of a complex net-
35S
work of chemical reactions that took advantage of free energy in the
36N
2 60
Big Picture - UK final proofs.indd 260
20/07/2016 10:02:49
t h E OR I g I n A n d P u R P O S E O F l I F E
environment of the young Earth, which could then be used to power repli-
01
cation once it began. There is also a replication- first camp, which currently
02
enjoys wider popularity in the community. They tend to think that energy
03
sources are relatively plentiful and unproblematic, and the important leap
04
in the development of life was the synthesis of an information- bearing mol-
05
ecule (presumably RNA, ribonucleic acid) that could duplicate itself and
06
pass down its genetic information.
07
We won’t be adjudicating this disagreement: these are hard questions to
08
which we simply don’t yet know the answers. But they are not hopeless
09
questions. Progress toward understanding abiogenesis has been made on
10
multiple fronts, both theoretically and experimentally. Whatever order me-
11
tabolism and replication appeared in, they are both necessary, and part of
12
the scientific fun will be in figuring out how all the ingredients fit together
13
into the final recipe.
14
15
•
16
If you want to understand how life began, it makes sense to begin by look-
17
ing for features that are shared by existing forms of life. One such feature
18
seems to be the proton- motive force involved in chemiosmosis, as we dis-
19
cussed in chapter 30. Cell membranes collect energy from photons or from
20
compounds like sugar, and use that energy to expel electrons outside the
21
cell, leaving an excess of protons inside. The mutual repulsion of the pro-
22
tons creates a force that can be used to do useful things like creating ATP.
23
Where did life ever get that idea from? It’s not exactly the obvious way
24
for a cell to manipulate energy. When the chemiosmotic process was worked
25
out by Peter Mitchell and Jennifer Moyle in the 1960s, they were met with
26
enormous skepticism in the biology community, until the experimental
27
evidence became definitive. The fact that nature finds this technique so
The Big Picture Page 44