The Intelligence of Water - Aquaporins and the Medium That Participates in the Computation
The Intelligence of Water: What Aquaporins Reveal
The argument that water possesses a form of molecular intelligence begins with its unusual physical properties. Water molecules respond to their surroundings, interact cooperatively, and continuously change their arrangements. Their positive and negative electrical regions allow them to form hydrogen bonds that break and reform within trillionths of a second. These interactions allow water molecules to respond rapidly to nearby molecules, electrical fields, and changing chemical conditions. Aquaporins demonstrate how these properties of water become useful biological information.
Each aquaporin contains four protein subunits, each subunit being a narrow channel through which water molecules pass in single file through the membrane either out of the cell or into it. The channel's precise shape and electrical environment allow water to pass extraordinarily rapidly while excluding ions and preventing protons from traveling along the hydrogen-bonded chain. Water molecules continually adjust their orientation and hydrogen bonds in response to the channel's structure. The protein supplies the selective environment, while water supplies the rapid molecular responses that make selective transport possible. Neither can perform this function in the same way without the other.
Water participates in a larger system that senses conditions, regulates its behavior, and corrects disturbances. When a cell loses or gains water, its volume, internal concentrations, and membrane tension change. Cellular sensors detect these changes, signaling pathways respond, and the cell can adjust the number or activity of aquaporins in its membrane. In the kidneys, for example, the hormone vasopressin causes additional aquaporin-2 channels to enter the cell membrane, allowing more water to be recovered. As conditions change, the channels can be removed again. This creates a feedback loop: water movement changes cellular conditions; cells detect those changes; signaling systems regulate aquaporins; and aquaporins alter subsequent water movement.
Although water’s hydrogen-bond network rapidly reorganizes in bulk water rather than storing information over long periods, water is not a passive background: its physical properties are essential to the system's sensing, selectivity, and response. Water's is part of molecular intelligence being an active, indispensable participant in a distributed information-processing system.
Summary
The case that water takes part in intelligent behavior, explained step by step
Water being intelligent refers to a system that does five things: it notices a condition, weighs it together with other information signals, responds in a fitting way, keeps some record of what happened, and adjusts over time. Scientists use this plain test for any system to be intelligent, whether a brain, a cell, or a swarm of ants.
Blood is mostly water with salt dissolved in it. If you get dehydrated, you lose water but not salt, so the blood becomes saltier. Water naturally moves toward the saltier side of a thin barrier, a process called osmosis. Some brain cells are surrounded by fluid that gets saltier along with your blood, so water seeps out of those cells and they shrink very slightly. Those cells can sense their own shrinking, and they signal the body to release a hormone, which is a chemical signal carried in the blood.
The body also checks your blood pressure and volume, so the hormone’s release reflects several signals together. The hormone travels to the kidneys, where it activates an enzyme, a molecule that does chemical jobs inside cells. The enzyme attaches a small chemical tag to aquaporin protein channel molecules. An aquaporin is a tiny tunnel, made of protein, that sits in a cell’s outer membrane and lets water through. Before the tag is attached, the tunnels are stored inside the cell in storage vesicles. The signal that is sent by adding the tag sends the channel to the outer membrane, where they let water return to the blood. The blood becomes less salty, the hormone fades, and the tunnels are taken back inside. This is a feedback loop, meaning the result of the action changes the very condition that set it off.
Water is at the center of that loop in two ways. The thing being measured, saltiness and cell size, exists only because water moves. And the channel works only because of how water behaves. Water molecules link to each other in chains, using weak attractions called hydrogen bonds. That is how they can travel, connected via these bonds, through the narrow tunnel in single file about three billion per second through each tunnel.
But the same chain property creates a danger. Protons, which are tiny positively charged particles, can hop along a chain of water molecules the way a bucket brigade passes buckets. Cells keep an electrical charge difference across their membrane boundaries, much like a battery, and protons leaking through would drain it. So, the protein channel is constructed to reorient each water molecule in the middle of the tunnel, which breaks the single file bonded chain of waters exactly at that point. The protein uses water’s chain-forming nature to move water fast, then interrupts that same property to stop the protons.
When a system’s success depends on the specific properties of a material, and not just on the material being present, that material is a working part of the system, not just a background. Water’s chain-forming is not incidental here. The tunnel is designed around it.
Water hyrodrogen bond links break and re-form in about a trillionth of a second, so water keeps no record, and it cannot adjust by itself. The sensors, hormones, and enzymes do the remembering and adjusting. What is established is that water is a necessary working part of a system that behaves intelligently. Water is part of the distributed intelligence of the system.
The Question
What would it mean for water to be intelligent? The phrase invites two mistakes. One is to dismiss it at once. The other is to embrace it too quickly and credit water with memory, intention or consciousness on the strength of anecdote and wishful physics. This post asks a narrower question. Does water, as it actually behaves inside living cells, take part in the kind of information handling that meets the minimal criteria for intelligence-like behavior: acquiring information about conditions, integrating it, acting selectively, retaining some record of what has happened, and adjusting? Aquaporins, the channels through which water crosses membranes, are the test case, because they sit exactly where the passive medium meets the regulated machinery.
My thesis is that water is not the inert background it is taught to be, and that the intelligence-like behavior of the aquaporin system cannot be separated from the physical properties of the liquid that flows through it. Whether that makes water a participant in the computation is the real question.
A Molecule Built for Cooperation
A water molecule is bent, with the two hydrogens separated by an angle of about 104.5 degrees. Oxygen draws electrons toward itself, so the molecule carries a partial negative charge at one end and partial positive charges at the other. This polarity lets each molecule form up to four hydrogen bonds, two as donor and two as acceptor, arranged roughly in a tetrahedron. In the liquid the average is a little under four. Each bond is individually weak. Bonds break and re-form on a timescale of picoseconds, so the network is in constant motion, yet at any instant it is highly structured over short distances.
The most important property of this network is cooperativity, that is, each bond affects how others operate. A hydrogen bond to one molecule alters that molecule’s electron distribution and makes its other bonds slightly stronger, so a molecule in a chain behaves differently from a molecule on its own, and a disturbance at one point is felt by its neighbors almost instantly. Many of water’s unusual properties follow from this collective character. Ice is less dense than the liquid, so it floats and lakes freeze from the top. The high heat capacity buffers temperature in organisms, the high dielectric constant lets water screen and dissolve ions, and the surface tension is unusually high. These are properties of the network, not of the isolated molecule, and they are the first sense in which water behaves as a system rather than a substance.
Water Inside the Cell
Water makes up roughly seventy percent of a cell’s mass and the great majority of its molecules by number. The cytoplasm is not a dilute solution but a crowded mixture in which most water lies within a few nanometers of a macromolecular surface. Four functions of that water go beyond being a passive solvent.
The first is the hydrophobic effect. Nonpolar surfaces cannot form hydrogen bonds with water, so the network around them is forced into more ordered arrangements, and the system lowers the cost of the reorganizations by pushing the nonpolar parts together. This is a central driver of protein folding, in which oily side chains collapse into a core, and of the self-assembly of lipid bilayers. The very membranes in which aquaporins sit exist because of water’s collective behavior.
The second is the hydration shell. Water molecules at a protein surface are slowed and ordered, and the exchange of bound water is part of how proteins recognize one another. Two kinds of water matter when another molecule binds. Ordered waters sitting in a binding site can be displaced when a drug arrives and releasing them to the bulk liquid contributes to the strength of binding. Other hydrated water molecules stay put and serve as hydrogen-bonded bridges between protein and binding molecule, acting as part of the complex’s architecture. When a drug attaches to a protein, some water molecules in the pocket are pushed out, which helps the drug hold on, while others stay and act as bridges between the protein and the drug. Designers have to take both into account.
The third is dynamical coupling. The motions of a protein are coupled to fluctuations in the surrounding water, and changing the hydration of a protein changes its dynamics.
The fourth is chemical. Water is the agent that triggers the reaction in hydrolysis, that is, the large number of vital reactions that use a water molecule to break down a larger target molecule. Many of the steps described in the ubiquitin posts are hydrolyses in which a water molecule is the final donor of electrons: the proteasome’s ATPase motors split ATP, the proteasome cleaves peptide bonds, and deubiquitinases cut ubiquitin from its substrates. The language of the cell is written with chemistry powered by adenosine nucleotides posting tags on molecules and then these tags are erased by hydrolysis, which consumes water.
The Aquaporin: Passage Under Control
The discovery of the aquaporins showed how a channel in a membrane can pass water at extraordinary speed while excluding almost everything else. An aquaporin consists of four identical protein tubes, and each tube forms its own water pore through the membrane. Each tube is built from six membrane-spanning helices and two short half-helices that meet in the middle of the membrane. A single tube passes on the order of three billion water molecules per second, in single file, driven entirely by osmotic or hydrostatic gradients. The channel does not pump water and does not choose its direction. It simply makes passage easy.
Selectivity depends on two features. At the narrowest point is an amino acid-based filter, the aromatic/arginine filter (called the ar/R constriction). At the filter, the pore is about 2.8 angstroms across, barely wider than one water molecule. The positively charged arginine at the filter repels positively charged ions, such as an isolated H+ proton ion. Deeper in the channel, two short amino acid sequences of asparagine-proline-alanine, place their two asparagine amino acids at the center of the channel.
These asparagine amino acids forcefully reorient each passing water from the way it has been oriented to travel through the channel hydrogen bonded in a chain of water molecules. The new orientation of the water molecule breaks the bonding of the chain and points the oxygen toward the middle and its hydrogens point outward toward the two ends. Carbonyl oxygens in amino acids along the pore wall provide hydrogen bonds that compensate for losing the bonds from the chain.
This design also solves a problem that water’s own physics creates. In bulk water, protons do not travel as individual particles. They hop along chains of hydrogen-bonded molecules, called proton wires, which makes liquid water an exceptionally good proton conductor. A channel filled with water in single file is therefore a potential proton wire, and a membrane that leaked protons would lose the voltage and pH gradients on which cellular energy conversion depends.
The aquaporin prevents this. As mentioned, the ar/R (aromatic/arginine) constriction zone forms a barrier by narrowing the channel and forcing a hydrogen-bond interruption in the water chain. The Asn-Pro-Ala motif region located near the center of the pore, generates a strong positive electrostatic field. As water molecules pass through, the partial charges force them to flip or reorient. This reorientation breaks the continuous donor-acceptor alignment of the water chain. But, it also imposes a major electrostatic penalty that blocks positive protons from crossing.
The walls of the channel at this exact turning point have positive electrical charges from asparagine. Because like charges repel, this positive zone creates a massive energy barrier. A positive proton, H+ trying to pass through will be powerfully pushed back by the positive charges of the channel walls. In short, the channel acts like a turnstile that physically flips the water molecules so they can't pass protons down the line, while simultaneously using a positive electrical shield to repel the protons away, ensuring only neutral water gets through.
A channel that exploits water’s cooperative network to move water quickly also deliberately interrupts the same network to stop protons. In that sense the protein edits a property of water rather than simply transmitting it.
Who Decides: Regulation of the Channel
The channel is passive, but the number of channels in a membrane and their activity are not. Cells use two main regulatory tools. The first is gating, in which the channel stays embedded in the membrane but a signal changes its shape so that the pore is open or closed. Gating is best understood in plants, where a loop can fold into the pore like a cap, and pH, calcium and phosphorylation determine whether it does, which lets a plant close its channels within seconds under flooding or drought. Gating has also been reported for some mammals, but it is not the dominant mechanism.
The second tool is the dominant mechanism in mammals. Here the cell adds more aquaporin channels to the membrane by bringing more vesicles to the membrane to place channels there and then removes them by the opposite function of having the membrane constrict around the channel, placing them in a a new vesicle that pulls from the membrane to return to storage. This is case described here in the kidney. These replacement channels work on a timescale of minutes rather than seconds.
The kidney aquaporin channel story is a complete regulatory loop. When the concentration of particles in the blood rises, osmosensitive neurons in the hypothalamus, which respond to their own shrinkage, trigger release of vasopressin from the pituitary. With a signaling cascade it activates protein kinase A, which phosphorylates the aquaporin on serine 256, a modification required for the channels to accumulate in the membrane. The kinase also loosens the actin network under the membrane that would otherwise hold storage vesicles back. Vesicles carrying aquaporin move to the cell membrane and fuse with it through a complex signaling cascade. This leaves the aquaporin as a channel embedded in the cell membrane.
Water from the urine then flows into the front side of the kidney epithelial cell through the aquaporin channel and leaves through other channels on the other side of the special kidney cell, so the blood is diluted. Water moves down its osmotic gradient from the dilute/less concentrated tubule fluid into the interior of the cell. Once inside the cell, water immediately crosses the opposite side where and quickly passes through the cell and enters the surrounding tissue. From there, it is absorbed into nearby blood vessels, increasing the blood's water volume and diluting it.
As the osmolality falls, vasopressin falls, and aquaporin channels are removed from the membrane and sent to its storage vesicle. Over longer times, vasopressin also increases activation of the aquaporin gene.
The general structure of the loop is a sensor measuring a condition. A signaling pathway integrating it. An enzyme modifying the channel. The channel changes how water moves, which changes the condition that was being measured. This is a negative feedback loop, and it is where the decisions are made.
Where Water Itself Computes
If the story ended there, water would simply be the quantity being controlled. But there are several places where water’s own physics does work that we would otherwise credit to a protein, and these are the places where the claim of its being part of the distributed intelligence becomes serious.
Cell volume is a signal.
Water acts here as carrier, a medium and a measurable quantity all at once. As carrier, it moves across membranes and physically swells or shrinks the cell. As medium, it sets the dilution of ions and osmolytes inside and outside. As measurable quantity or component its movement changes the size of the cell, the tension in its membrane and the crowding of its cytoplasm.
A cell cannot count water molecules. It reads the consequences of water movement, above all its own volume, and acts on them. Swelling activates volume-regulated anion channels that release ions and osmolytes so that the cell shrinks back, and shrinking activates transporters that bring ions in so that water follows. The hypothalamic neurons that start the aquaporin loop work the same way, since they respond to their own shrinkage.
Water is a wire. In cytochrome c oxidase, protons are pumped across the inner mitochondrial membrane through channels lined with ordered water. In bacteriorhodopsin, light-driven proton transfer depends on chains of water molecules in the protein interior. In photosystem II, the oxygen-evolving complex splits water to supply electrons for photosynthesis, and the protein around the manganese-calcium cluster includes channels that deliver water in and remove protons. In each case a protein has been shaped to harness water’s capacity to conduct protons by hopping, and to gate that conduction. Gate means to control, regulate, or turn on and off the flow of protons through the protein. Water supplies the physical process, and the protein supplies the control.
Water shapes the organization of the cell. The condensates described in earlier posts, droplet-like assemblies of disordered proteins and RNAs, form because the interactions among protein segments, RNAs, and between those segments and water, balance in a particular way, and small changes in hydration or salt can push the system across a phase boundary. Some organisms have evolved antifreeze proteins that bind ice crystals and prevent them from growing, manipulating the structure of water directly in order to survive freezing. Proteins are not only embedded in water. They are tuned to its collective behavior, and its behavior is in turn altered by them.
Testing the Water Intelligence Criteria
Water Acquires information. In a thin sense. Osmotic and mechanical conditions alter water’s distribution and activity, and proteins sense the consequences. Water carries the variable, and the sensors are proteins.
Water Integrates. Over molecular distances, water’s cooperative network couples events almost instantly, which is a real physical integration. Over the scale of a cell, integration is done by signaling pathways.
Water Acts selectively. Always through the structures that shape it. The aquaporin passes water in single file while its internal chemistry interrupts the chain that would carry protons. A potassium channel is lined with carbonyl oxygens that mimic the shell of water around a potassium ion, so the ion sheds its water coat and slips through, whereas a sodium ion is rejected because it fits that layout poorly. Condensates assemble because of how their protein and RNA components interact with water and with each other, and they pull in some molecules while excluding others. In the condensate case, the specificity comes mainly from the macromolecules, with water acting as the solvent that makes their interactions possible.
Water Retains a Record. Hydrogen bonds in the bulk liquid rearrange within picoseconds, so the arrangement of the liquid forgets almost immediately. Waters in hydration shells are slower, with residence times from tens of picoseconds to nanoseconds, and a few buried structural waters can stay in place for much longer. A strong record is held in the cellular structures that water helps to hold together, not in the water. Water acts as the mortar and the building holds the data.
Water Adapts. Through the regulatory loops around it. The body maintains water balance, cell volume and blood pressure through sensors, hormones, kidneys and pumps. Water is the passive element, and the loops do the adapting.
What meets all five criteria is the system of water, protein and regulatory loop together, with water contributing to that distributed intelligence system.
For water to count as a constitutive part of the computation there must be cases in which the dynamical state of the water itself carries information that other components read and use. Here we are describing specific cases of these. The orientation and connectivity of water in a proton wire is the signal that a protein gates. Hydration state and cell volume are read by sensors as the variable of interest. Displacement of ordered waters from a binding site is part of how recognition is achieved.
Each of these can be redescribed as a protein exploiting a physical property of its solvent, in the way an engineer exploits the physics of a transistor.
The Aquaporin in Miniature
The aquaporin is a good place to end because it contains the whole argument in miniature. A protein of about 270 residues builds a pore just wide enough for water, lines it with the chemistry that makes passage easy, and breaks the cooperative network of the liquid at the center so that protons cannot follow. The cell then wraps this passive device in a loop of sensors, hormones, kinases and ubiquitin tags that decide how many channels are in the membrane at any moment.
Inside the narrowest part of the channel, water molecules cannot pass one another. They sit in a single file, hydrogen-bonded to their neighbors and to the wall, so the movement of one is coupled to the movement of others. This does not mean they advance as a rigid train. They keep fluctuating, exchanging hydrogen bonds and moving in either direction, and the net flow is set by differences in water chemical potential, usually established by osmotic gradients. No ATP is spent on the water itself.
The process works at several connected levels. Individual water molecules respond to hydrogen bonds and electric fields. The protein creates a selective transport environment, with its shape and chemical groups influencing the water, while the water’s interactions and fluctuations influence the energetics of the channel. Cells regulate the abundance and location of the channels, and hormones coordinate millions of cells to regulate the body’s water balance.
Water supplies the physics, the protein supplies the filter, and the signaling system supplies the judgment. None of the three contributes passively, and the outcome resides in none of them alone. The intelligence emerges from their continual interaction.
Water Supplies the Physics -- Water in Distributed Intelligence
Water is not a passive fluid being pushed around. Its inherent, natural physical laws and properties drive the movement itself. While the protein acts as the mechanical gatekeeper and the cellular signals act as the manager, water brings its own intrinsic forces to the table to make the system work:
Water molecules are in constant, chaotic motion. They possess their own kinetic energy, meaning they don't need to be "pumped" through the channel; their natural thermodynamic drive causes them to diffuse through on their own.
Water molecules are highly polar and stick to one another in single-file chains via hydrogen bonds. When one water molecule moves through the channel, it pulls the next one along with it like links in a chain.
Because water has a positive and negative end, it responds dynamically to the electrostatic charges inside the protein channel. As it passes through, the water molecules physically flip and reorient themselves to navigate past the protein's chemical groups, preventing unwanted particles (like protons) from leaking through.
The protein channel doesn't provide the engine. It only provides the track. Water's own molecular physics supply the actual locomotive force and behavioral traits that allow it to stream through the cell.
The aquaporin system with its regulators demonstrates distributed intelligence in which water is both the medium and the quantity controlled.
