Cell Signaling · Unit 2: The parts — switches and wiring

Phosphorylation: Kinases and Phosphatases

Why do cells use a phosphate group as their most common switch, and which enzymes flip it? This lesson explains what a phosphate does to a protein, why kinases and phosphatases have to work as a pair, and what the size of each family says about how it is controlled. You will finish able to plan the experiments that show a protein is phosphorylated, at which residue, and to what effect.

Watch first A Charge You Can Add and Remove · 2:54

Your genome holds about 518 genes for protein kinases, and nearly every one of those enzymes runs the same reaction: it takes the end phosphate off ATP and fastens it to a protein. Why do cells rely so heavily on one chemical group, what does it do once it is on, which enzymes take it off, and how would you prove that a protein is controlled this way?

The idea: a phosphate is a charge you can add and take away

Start with a question from the lecture: if we have a protein, what are some of the ways we could alter its function? You could make more of it, which is slow, or bind a small molecule to it. Or an enzyme could attach a chemical group to one of its side chains after the protein is made. That last route is covalent modification, and it always comes as a pair, modifications and modifiers: the mark, and the enzymes that write and erase it.

Of the many marks cells use, signaling pathways use phosphate most. A protein kinase is an enzyme that moves the end (gamma) phosphate of ATP onto a protein. A protein phosphatase is an enzyme that removes that phosphate with water.

protein-OH + ATP → protein-O-PO3(2−) + ADP (kinase)

protein-O-PO3(2−) + H2O → protein-OH + Pi (phosphatase)

In animals the phosphate goes onto a side chain that ends in a hydroxyl group: serine, threonine or tyrosine. Nearly all of it sits on serine and threonine; tyrosine carries less than 1%, which kept it hidden until 1979.

Four properties make phosphate a good switch.

Charge. Near pH 7 a phosphate on a protein carries close to two negative charges, on a spot that had none. A folded protein is held in shape by charges attracting and repelling each other, so few changes this small disturb it this much.

Reversibility. The bond holds until an enzyme removes it, and removal leaves the protein exactly as it was.

A separate writer and eraser. The phosphatase does not run the kinase reaction backwards; it runs a different reaction, and each one goes downhill. A whiteboard works because the marker and the eraser are different tools that you can hand to different people. The cell likewise regulates adding and removing independently, at the price of one ATP per round trip.

Supply. ATP sits at millimolar concentration in the cell, so a kinase never waits for its donor.

What a phosphate does: four results of modification

The phosphate has no activity of its own. What matters is how the protein and its neighbours respond to the new charge. The lecture lists four results.

It changes shape, and with it activity. Every copy of a protein flickers among several arrangements, some working and some not. A phosphate makes the arrangements that accommodate its charge cheaper to occupy, so more copies sit in them. In muscle glycogen phosphorylase the phosphate lands on serine 14, near the start of the chain and far from the catalytic site, and it shifts the enzyme toward its active form. The same trick can turn a protein off: glycogen synthase, which builds glycogen, is inhibited by phosphorylation.

It creates a docking site. A phosphorylated side chain is a surface that was not there a moment ago, and some proteins carry pockets built to recognise one (a later lesson covers them). A phosphate can therefore bring two proteins together without changing either one’s catalytic rate.

It changes location. NFAT is a transcription factor, a protein that switches genes on. It waits in the cytoplasm carrying many phosphates, which hide the sequence that would send it into the nucleus. When the phosphatase calcineurin strips them off, NFAT enters the nucleus. Here the erasing enzyme carries the signal.

It primes another modification. Some kinases accept a substrate only if a phosphate is already nearby. Glycogen synthase kinase 3 is the standard example: it prefers a serine or threonine that lies four residues before one that another kinase has already phosphorylated. Priming lets a protein wait for two inputs before it responds.

The modifiers: 518 kinases and the phosphatases that answer them

Of the 518 or so human kinases, about 90 are tyrosine kinases, and 58 of those are receptors, proteins that span the plasma membrane with the kinase on the inside. Most of the rest phosphorylate serine and threonine.

Why does a cell need phosphatases at all? Think of a sink with the tap running and the drain open. The water level depends on both, and you can change it at either end. The same holds for a pool of protein acted on by a kinase and a phosphatase at simple first-order rates:

fraction phosphorylated at steady state = k(kinase) / (k(kinase) + k(phosphatase))

A signal can raise the phosphorylated fraction by activating the kinase or by inhibiting the phosphatase, and cells do both. The time the level takes to settle after a change is set by the sum of the two rates, so a pathway that must shut off in seconds keeps an active phosphatase running against its kinase and pays for the speed in ATP.

On the tyrosine side the lecture slide reads “107 PTPs to counteract 90 PTKs” (PTP, protein tyrosine phosphatase; PTK, protein tyrosine kinase). The eraser family is about as large as the writer family, and many erasers are built like membrane receptors. The 107 counts the whole PTP family, though: about 38 are classical enzymes that read only phosphotyrosine, and some relatives, such as PTEN, prefer substrates other than proteins.

On the serine/threonine side the counts are lopsided: hundreds of kinases against a few dozen genes for phosphatase catalytic subunits. Specificity comes from partners: PP1, a major catalytic subunit, is encoded by just three genes yet pairs with well over a hundred regulatory subunits, which decide where the enzyme goes and what it touches.

The counts suggest that tyrosine dephosphorylation is steered by the choice of enzyme, and serine/threonine dephosphorylation by the company the enzyme keeps. That is an inference from numbers, not a proof.

How we know: one muscle enzyme, then all the rest

Watch How a Phosphate Was Found · 1:36

1943. Carl and Gerty Cori and their colleague Arda Green, in St. Louis, crystallized glycogen phosphorylase, the muscle enzyme that releases glucose units from glycogen, and found it in two forms. Phosphorylase a was active alone; phosphorylase b worked only when AMP was present. They also had an enzyme that turned a into b and assumed it clipped off a bound AMP. It was in fact a phosphatase.

1954. Burnett and Kennedy reported a liver enzyme that moves phosphate from ATP onto the milk protein casein, the first protein kinase described.

1955. Edmond Fischer and Edwin Krebs, in Seattle, reported what had started as an effort to make phosphorylase a by the Coris’ recipe: they kept getting b, unless they filtered the extract through paper. The paper was leaching calcium, and extracts left standing ran out of ATP. Over the next few years the conversion of b to a proved to be an enzyme transferring one phosphate from ATP to one serine on each chain of the enzyme. The overview lesson on signal transduction tells this story in full.

1979 to 1980. Tony Hunter saw a radioactive spot run between the phosphoserine and phosphothreonine markers, and a phosphotyrosine standard he made ran to the same place. He and Bartholomew Sefton then showed that Src, the cancer-causing protein of Rous sarcoma virus, phosphorylates tyrosine. Within the year H. Ushiro and Stanley Cohen, using membranes from a tumour cell line and no living cells, found the same activity in the receptor for epidermal growth factor (EGF), a cell’s own protein with no virus involved.

1988. Nicholas Tonks, working with Fischer, purified the first protein tyrosine phosphatase, PTP1B, from human placenta. Its sequence resembled part of CD45, a surface protein of white blood cells, so a tyrosine phosphatase could be built like a receptor too.

Design problem

You have a protein that you suspect is phosphorylated. Show that it is, find where, and show what the phosphate does.

One way to do it

Show that it is. Run the protein from cells on an SDS gel, which sorts proteins by size; a phosphorylated protein often runs a little slower. Treat half the sample with a purified phosphatase. If the slow band collapses into the fast one, and a phosphatase inhibitor prevents the collapse, the difference was phosphate. Confirm it by growing cells with radioactive phosphate (32P), pulling the protein out with an antibody, and checking that it is labelled. A site found only in a test tube must still be confirmed in cells.

Find where. Break the labelled protein down to amino acids and separate them to learn whether serine, threonine or tyrosine carries the label. Then change each candidate serine or threonine to alanine, or each tyrosine to phenylalanine, which leaves no hydroxyl, and see which change removes the label or the gel shift. Mass spectrometry does the same job faster: a phosphate adds 80 Da to a peptide, and fragmenting the peptide shows which residue holds it.

Show what it does. Compare the alanine mutant, which can never be phosphorylated, with an aspartate mutant, whose negative charge imitates a phosphate that never leaves. Ask which of the four results the mimic shows. Aspartate carries one charge where a phosphate carries two, and it is smaller, so a mimic that does nothing proves little. Check the answer by inhibiting or activating the kinase: the normal protein should respond and the alanine mutant should not.

To find which kinase does the writing, inhibitors narrow the list, and Kevan Shokat’s analog-sensitive kinases, engineered to accept an ATP analog that normal kinases cannot use, label only their direct targets.

In lecture, Prof. Deppmann works this problem with a transcription factor from his 2003 paper: a phosphatase-sensitive gel shift, serine and threonine sites, and a serine-to-aspartate mimic.

When the switch breaks: kinases in disease and in the clinic

One lecture slide carries the title “Nearly every kinase has been associated with disease.” A common pattern in cancer is a gain-of-function mutation, one that makes a protein more active: a kinase, often a receptor tyrosine kinase, is stuck on and tells the cell to grow when no signal has arrived. Losing a phosphatase can do similar damage, because the brake is gone.

That makes kinases drug targets, and a second slide says why they are good ones: “Highly conserved: target for lots of drugs.” Kinases bind ATP in a cleft between two lobes, a snug, mostly greasy pocket of the kind a small drug molecule can fill. Imatinib, approved in 2001 for chronic myeloid leukemia, sits in that pocket in the kinase that drives the disease: BCR-ABL, an overactive fusion protein made when a chromosome rearrangement joins two genes. Gefitinib (2003) and erlotinib (2004) do the same in the EGF receptor.

The conservation that makes the pocket a good target also makes selectivity hard. A molecule shaped to fit one ATP pocket tends to fit some of the other 517, and each extra kinase it blocks is a possible side effect. The drug also has to outcompete the cell’s own millimolar ATP. Chemists find selectivity at the edges of the pocket, where kinases differ, or in shapes that only some kinases adopt: imatinib binds the Abl kinase of BCR-ABL in an inactive conformation, which helps explain why it spares most other kinases.

How we measure it

Radioactive labelling with gamma-32P ATP

You incubate a kinase and its substrate with ATP whose end phosphate is radioactive, or grow cells in radioactive phosphate, then separate the proteins on a gel and expose it to film. A dark band shows which protein received phosphate. Only a label in the gamma position works in the test tube, because that is the phosphate the kinase transfers.

Phosphatase-sensitive mobility shift

Many phosphorylated proteins run more slowly on an SDS gel. If treating the sample with a purified phosphatase, such as calf intestinal alkaline phosphatase, collapses the slow band into the fast one, the difference was phosphate. It is quick and needs no radioactivity, though some phosphates cause no shift at all.

Phosphoamino acid analysis

You break a 32P-labelled protein down to its amino acids in acid and separate them by electrophoresis on a thin-layer plate, beside phosphoserine, phosphothreonine and phosphotyrosine standards. The position of the radioactive spot tells you which kind of residue carries the phosphate. This is the method that revealed phosphotyrosine in 1979.

Site-directed mutagenesis: alanine and aspartate substitutions

You change the suspected residue to alanine, which cannot accept a phosphate, or to aspartate, whose negative charge imitates one. Loss of labelling in the alanine mutant maps the site, and the behaviour of the aspartate mutant suggests what the phosphate does. A carboxylate is only a rough copy of a phosphate, so the result needs a second line of evidence.

Finish line

Test yourself

8 questions

Practice and MCAT-style, with explained answers.

Next · Lesson 7

GTPase Switches: The Timer

A hormone leaves the cell and the response stops. What keeps time?

Continue →

← Lesson 5 · Circuits: Amplification, Feedback, Adaptation and Switches

Glycogen phosphorylase a — the first enzyme known to be switched by a phosphate
Glycogen phosphorylase a — the first enzyme known to be switched by a phosphate ● LIVE · RCSB PDB 1GPA
open on RCSB PDB ↗
Loading live structure 1GPA from RCSB PDB…

Find the phosphoserine near the start of each chain (Ser14). It lies at the contact between the two subunits, far from the catalytic site where the cofactor pyridoxal phosphate sits.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Protein kinase A catalytic subunit with ATP — the kinase fold and its ATP pocket
Protein kinase A catalytic subunit with ATP — the kinase fold and its ATP pocket ● LIVE · RCSB PDB 1ATP
open on RCSB PDB ↗
Loading live structure 1ATP from RCSB PDB…

ATP sits in the cleft between the small lobe and the large lobe. The inhibitor peptide lies where a substrate would, with an alanine in the place of the serine that would receive the phosphate.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
PTP1B — the first tyrosine phosphatase to be purified
PTP1B — the first tyrosine phosphatase to be purified ● LIVE · RCSB PDB 2HNP
open on RCSB PDB ↗
Loading live structure 2HNP from RCSB PDB…

Find cysteine 215 at the bottom of the active-site pocket. The pocket is deep enough for a phosphotyrosine to reach the cysteine, while the shorter phosphoserine and phosphothreonine side chains cannot.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
c-Src in its off state — a kinase held shut by its own phosphotyrosine
c-Src in its off state — a kinase held shut by its own phosphotyrosine ● LIVE · RCSB PDB 2SRC
open on RCSB PDB ↗
Loading live structure 2SRC from RCSB PDB…

Look for the phosphotyrosine near the end of the chain. The field calls it Tyr527, the number it has in chicken Src; in the human sequence it is Tyr530. It is tucked against another part of the same protein, which holds the kinase lobes in an inactive arrangement. Here a phosphate switches a kinase off.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Abl kinase with imatinib — a drug in the ATP pocket
Abl kinase with imatinib — a drug in the ATP pocket ● LIVE · RCSB PDB 1IEP
open on RCSB PDB ↗
Loading live structure 1IEP from RCSB PDB…

Imatinib fills the cleft where ATP would bind and reaches into a neighbouring pocket that opens only when the kinase is in an inactive shape. That second contact is one source of its selectivity.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Edwin G. Krebs

Edwin G. Krebs 1918–2009

With Edmond Fischer, showed that glycogen phosphorylase is switched on when a kinase moves a phosphate from ATP onto one serine. His laboratory later described the cyclic AMP-dependent protein kinase.

Krebs trained as a physician, served as a Navy medical officer, and then worked as a postdoctoral fellow with Carl and Gerty Cori in St. Louis, where he learned to handle glycogen phosphorylase. When he joined the biochemistry department at the University of Washington in 1948 he made sure he was licensed to practise medicine in the State of Washington, as insurance in case research did not work out. Soon after Fischer joined the department in 1953, the two of them set out to learn why phosphorylase b needs AMP and phosphorylase a does not. They never answered that question. Following the Coris' recipe, they found that b became a only when ATP and calcium were present, and they traced the change to an enzyme that moves a phosphate from ATP onto a single serine on each chain of phosphorylase. In 1968 his laboratory described the cyclic AMP-dependent protein kinase, the enzyme through which a hormone's second messenger reaches its targets. He shared the 1992 Nobel Prize in Physiology or Medicine with Fischer. His work supplied links in the chain, pieced together by several laboratories over years, that runs from a hormone outside the cell to a change in enzyme activity inside it.

✓ verbatim Krebs, Nobel Lecture, 8 December 1992; Fischer & Krebs, J. Biol. Chem. 216:121 (1955).
Edmond H. Fischer

Edmond H. Fischer 1920–2021

With Edwin Krebs, showed that a removable phosphate controls an enzyme. The first protein tyrosine phosphatase was later purified in his laboratory.

Fischer was born in Shanghai, trained as a chemist in Geneva, and arrived at the University of Washington in the fall of 1953. He and Edwin Krebs soon shared a project and a routine. By Fischer's account they planned each day's experiment the evening before and ran it together, so that either could carry on if the other was called away. Their two 1955 papers showed that phosphorylase b becomes phosphorylase a in muscle extracts, and within a few years they showed that an enzyme transfers a phosphate from ATP to the protein. For years the result was treated as a peculiarity of glycogen metabolism, and Fischer later wrote that they did not know at the time whether it was a rare event. His laboratory kept working on the switch, and in 1988 the first protein tyrosine phosphatase was purified there. The 1992 Nobel citation named 'reversible protein phosphorylation as a biological regulatory mechanism.' The work made a removable phosphate something to look for on any protein.

✓ verbatim Fischer, Nobel Lecture, 8 December 1992; Krebs & Fischer and Fischer & Krebs, J. Biol. Chem. 216:113 and 216:121 (1955).
Tony Hunter

Tony Hunter b. 1943

Found that proteins can be phosphorylated on tyrosine, and that the cancer-causing Src protein of Rous sarcoma virus is a tyrosine kinase.

On 14 June 1979, at the Salk Institute in La Jolla, Hunter was checking which amino acid picked up phosphate in a kinase reaction on a protein from polyomavirus. The method separates phosphoserine from phosphothreonine by electrophoresis at pH 1.9. His radioactive spot landed between the two markers, where nothing should run. By his later account the cause was a reused buffer whose pH had drifted to about 1.7, enough to pull apart two compounds that travel together at 1.9. Suspecting the third amino acid with a hydroxyl group, he made his own phosphotyrosine standard, and it ran to the same place. That autumn he and Bartholomew Sefton showed that Src, the cancer-causing protein of Rous sarcoma virus, also puts its phosphate on tyrosine; the paper appeared in 1980. An earlier report had read the Src product as threonine; Hunter's explanation is that phosphotyrosine and phosphothreonine travel together at pH 1.9. Within the year the receptor for epidermal growth factor was shown to do the same. The finding opened the search that turned up about 90 human tyrosine kinases and, two decades later, the first tyrosine kinase inhibitor drugs, beginning with imatinib in 2001.

✓ verbatim Hunter & Sefton, Proc. Natl. Acad. Sci. USA 77:1311 (1980); Hunter, 'Discovering the first tyrosine kinase', Proc. Natl. Acad. Sci. USA 112:7877 (2015).
Nicholas K. Tonks

Nicholas K. Tonks work of 1988; at Cold Spring Harbor Laboratory from 1990

Purified the first protein tyrosine phosphatase, PTP1B, and helped show that the enzymes that erase phosphotyrosine are a regulated family that includes receptor-like proteins.

Tonks, who had just finished his doctorate with Philip Cohen in Dundee, joined Edmond Fischer's laboratory in Seattle as a postdoctoral fellow. By then a growing list of tyrosine kinases was known, and no one had purified an enzyme that removes phosphate from tyrosine. Tonks built an affinity column from a chemically modified lysozyme carrying a thiophosphate on one tyrosine, a group a phosphatase can bind and cannot easily remove. Starting from human placenta, Tonks, Diltz and Fischer purified two activities many thousand-fold and published them in 1988 as PTP1A and PTP1B. When sequence from PTP1B came in, it resembled the inside portion of CD45, a surface protein that immunologists had long used to mark white blood cells. A familiar receptor-like protein was a phosphatase. The match showed that removing a phosphate could be as regulated as adding one, and it made the tyrosine phosphatases a family to map.

✓ verbatim Tonks, Diltz & Fischer, J. Biol. Chem. 263:6722 and 263:6731 (1988); Charbonneau, Tonks, Walsh & Fischer, Proc. Natl. Acad. Sci. USA 85:7182 (1988).
Test yourself
Question 1 of 8 Score 0 / 0

Which amino acid side chains carry regulatory phosphate groups in animal proteins, and what chemical feature do they share?
Sources & Integrity

Every structure, slide, simulation, and clip on this page is listed below with the source it came from. None of the 3D structures is drawn by hand or by AI: each one loads by accession ID from RCSB PDB or PubChem. This panel is generated from the lesson's own data file, so what it lists is what the page loads.

Molecular structures live data

  • Glycogen phosphorylase a — the first enzyme known to be switched by a phosphate — RCSB PDB 1GPA · coordinates CC0 · Rabbit muscle enzyme, phosphorylated form; Barford, Hu & Johnson, J. Mol. Biol. (1991)
  • Protein kinase A catalytic subunit with ATP — the kinase fold and its ATP pocket — RCSB PDB 1ATP · coordinates CC0 · Mouse catalytic subunit with MnATP and a peptide inhibitor; Zheng et al., Acta Crystallogr. D (1993)
  • PTP1B — the first tyrosine phosphatase to be purified — RCSB PDB 2HNP · coordinates CC0 · Human catalytic domain; Barford, Flint & Tonks, Science (1994)
  • c-Src in its off state — a kinase held shut by its own phosphotyrosine — RCSB PDB 2SRC · coordinates CC0 · Human c-Src with the ATP analogue AMP-PNP; Xu et al., Mol. Cell (1999)
  • Abl kinase with imatinib — a drug in the ATP pocket — RCSB PDB 1IEP · coordinates CC0 · Mouse Abl kinase domain with imatinib (STI-571); Nagar et al., Cancer Res. (2002)

Scientist profiles & quotes

  • Edwin G. Krebs — Krebs, Nobel Lecture, 8 December 1992; Fischer & Krebs, J. Biol. Chem. 216:121 (1955).
  • Edmond H. Fischer — Fischer, Nobel Lecture, 8 December 1992; Krebs & Fischer and Fischer & Krebs, J. Biol. Chem. 216:113 and 216:121 (1955).
  • Tony Hunter — Hunter & Sefton, Proc. Natl. Acad. Sci. USA 77:1311 (1980); Hunter, 'Discovering the first tyrosine kinase', Proc. Natl. Acad. Sci. USA 112:7877 (2015).
  • Nicholas K. Tonks — Tonks, Diltz & Fischer, J. Biol. Chem. 263:6722 and 263:6731 (1988); Charbonneau, Tonks, Walsh & Fischer, Proc. Natl. Acad. Sci. USA 85:7182 (1988).

Quotes are verbatim from the cited source. Profile text is original to this site and was checked against the sources cited.

Animated explainer

A Charge You Can Add and Remove · narration: an AI narrator voice (Gemini, Charon) reading narration written from this lesson's text · animation: AI-generated in code (Manim) by Claude, a schematic with no drawn molecular structures, plots computed from their equations · captions from the narration script.

Section films

  • How a Phosphate Was Found · narration: an AI narrator voice (Gemini, Charon) reading narration written from this lesson's text · animation AI-generated in code (Manim) by Claude, a schematic with no drawn molecular structures · captions from the narration script.

Images & figures

This lesson carries no raster figures. Every structure is rendered live from RCSB PDB or PubChem, and every plot is drawn from its equation in the browser.

integrity Machines check the mechanics. The build confirms that every local file exists, that every question's answer key points at a real choice, and that any AI-generated scientific figure names the person who checked it; a weekly job re-resolves every accession ID against its database. Questions are original; MCAT-style items are written for this site and are not real AAMC questions. Banned sources (the Tymoczko textbook, publisher decks, copyrighted figures) are not used.

Like I'm five

Imagine a toy that only works when it has a special sticker on it. One helper, called a kinase, puts the sticker on, and the toy switches on. Another helper, called a phosphatase, peels the sticker off, and the toy switches off. Your cells have hundreds of these helpers, so they can turn the right proteins on and off at the right moment.
High school

Proteins do most of the work in a cell, and the cell has to turn them on and off quickly. One common way is to attach a phosphate group, a small cluster of atoms with a negative charge. Enzymes called kinases attach the phosphate, taking it from ATP, and enzymes called phosphatases remove it. The added charge changes the protein's shape or lets other proteins hold onto it, and that changes what the protein does. Because adding and removing are done by different enzymes, the cell can control each direction separately. Humans have more than 500 kinases, and faulty ones are a common cause of cancer.
College

Protein phosphorylation is a reversible covalent modification. A protein kinase transfers the gamma phosphate of ATP to the hydroxyl of a serine, threonine or tyrosine side chain; a protein phosphatase hydrolyses the phosphate ester and releases inorganic phosphate. The phosphate adds about two negative charges, which can shift the protein's conformational equilibrium (as at Ser14 of glycogen phosphorylase), create a binding site for another protein, change where the protein sits in the cell, or prime a second modification. The phosphorylated fraction of any protein is a steady state set by the opposing kinase and phosphatase rates. Humans encode about 518 kinases, about 90 of them tyrosine kinases. The tyrosine phosphatase family is about as large as the tyrosine kinase family, while a small set of serine/threonine phosphatase catalytic subunits gains specificity from many regulatory subunits. To show that a site matters you detect the phosphate, map the residue, and then test alanine and aspartate mutants.
Graduate

A phosphorylation site behaves as a futile cycle. For first-order kinase and phosphatase rate constants, the steady-state occupancy is k(kinase) / (k(kinase) + k(phosphatase)) and the relaxation rate is their sum, so a fast response is bought with ATP turnover. When both enzymes are saturated by substrate the same cycle becomes switch-like (the zero-order ultrasensitivity of Goldbeter and Koshland, 1981), which is one reason site occupancy cannot be read as kinase activity alone. Eukaryotic protein kinases share a two-lobed fold with ATP in the cleft and are often themselves switched by phosphorylation of an activation loop. Classical tyrosine phosphatases use a cysteine nucleophile and a phosphocysteine intermediate, whereas the major serine/threonine phosphatases are metalloenzymes that activate water. The count of 107 PTP genes includes dual-specificity enzymes and members such as PTEN whose preferred substrate is a lipid, so the comparison with 90 tyrosine kinases overstates the number of dedicated phosphotyrosine erasers.

An open question is specificity on the phosphatase side. For most of the tens of thousands of phosphorylation sites catalogued by mass spectrometry, neither the responsible phosphatase holoenzyme nor any function for the site is known, and what fraction of sites is regulatory, as opposed to incidental, remains unsettled.
Who found the first protein kinase?

Krebs and Fischer are often credited with discovering protein kinases. Krebs did not accept the credit. In his Nobel lecture he pointed to Burnett and Kennedy, who in 1954 described a liver enzyme that phosphorylates the milk protein casein, and said they 'were the first to describe a protein kinase.' What Krebs and Fischer added a year later was a reason to care: a phosphate that changed what an enzyme did. The prize citation is correspondingly narrow. It names reversible phosphorylation as a regulatory mechanism and does not mention discovering the enzymes.

— Krebs, Nobel Lecture, 8 December 1992; Burnett & Kennedy, J. Biol. Chem. 211:969 (1954).
A growth factor in a dish with no cells

Stanley Cohen expected, as a biochemist, that a growth factor would control some chemical reaction, and he wanted to see it in a system with no living cells. Because phosphorylation and dephosphorylation control so many metabolic pathways, the test was whether EGF changed the phosphorylation of membrane proteins. At Vanderbilt at the end of the 1970s his laboratory used membranes from A-431 cells, a human tumour line with many receptors for epidermal growth factor (EGF). When they added EGF, phosphate began to move onto protein. Cohen's group first reported threonine, the same misreading that had been made for Src, because the two phosphoamino acids ran together in the electrophoresis system used. H. Ushiro then rechecked and found tyrosine, and the EGF receptor turned out to be the kinase itself.

— Ushiro & Cohen, J. Biol. Chem. 255:8363 (1980); told here from Prof. Deppmann's graduate signaling module.
It pays to be last

PTP1B became important because of what had already been sequenced. CD45 was a well-known protein on the surface of white blood cells, used for years as a marker, with no known enzyme activity. When sequence from the placenta phosphatase turned out to resemble the inside portion of CD45, a familiar marker became a receptor-like phosphatase. Fischer told his Nobel audience that if the phosphatase had been sequenced before CD45, nobody would have paid much attention. In this business, he said, 'it really pays to be last.'

— Fischer, Nobel Lecture, 8 December 1992; Charbonneau, Tonks, Walsh & Fischer, Proc. Natl. Acad. Sci. USA 85:7182 (1988).

 ↑ Top