Nuclear Receptors: Receptors That Are Transcription Factors

Inject a rat with radioactive estradiol, and almost every tissue lets the hormone go again. The uterus and the vagina keep it. Elwood Jensen reported that result from the University of Chicago in 1958. It meant that the tissues estrogen acts on contain something built to catch it. Every pathway so far in this course started at a protein in the cell membrane. This one starts inside the cell and has almost no middle, because the protein that catches the hormone is the same protein that switches the genes on.

The idea: the hormone goes in, and its receptor reads DNA

Estradiol, cortisol and testosterone are steroids: four fused carbon rings made from cholesterol (their synthesis is in the lesson on lipid synthesis). A steroid dissolves in fat, so it crosses the lipid bilayer on its own. Nothing on the outer surface of the cell has to receive it. The receptor field guide set out the receptor architectures and why speed follows architecture. This is the slow one, and here is how it works.

Jensen’s radioactive label, tritium (a heavy form of hydrogen), was still attached to unchanged estradiol when he recovered it. That was the awkward part. Many biochemists then expected estradiol to work by being oxidized and reduced over and over, passing hydrogen between coenzymes. An intact hormone held in place pointed to a binding protein instead. In 1966 David Toft and Jack Gorski, at the University of Illinois, extracted an estrogen-binding protein from rat uterus. In 1980 Jensen and Geoffrey Greene made monoclonal antibodies against the human receptor (antibodies of a single kind, each binding one spot on its target). In March 1986 the laboratories of Pierre Chambon in Strasbourg and of Greene in Chicago published the full sequence of its cDNA (a DNA copy of a messenger RNA) within weeks of each other.

The sequence, and the experiments it made possible, showed one chain with two grips. One region closes around the hormone. Another takes hold of a specific short stretch of DNA. No second messenger is made and no kinase cascade runs. A transcription factor is a protein that binds DNA and controls whether a gene is copied into RNA. In the lesson on TGF-β and Smads the messenger was the transcription factor. Here the receptor is, and the family is called the nuclear receptors.

Two cautions keep this exact. Thyroid hormone, retinoic acid (made from vitamin A) and the active form of vitamin D use receptors of the same family, but thyroid hormone carries charges and enters cells through transporter proteins. A transporter is a door; the receptor is still inside. And some steroid effects appear within seconds to minutes. They do not run through new RNA, and this lesson leaves them aside.

One plan, forty-eight receptors

Once receptors had sequences, the sequences could be compared. Ronald Evans’s laboratory published the sequence of the human glucocorticoid receptor in 1985, and the estrogen receptor, months later, resembled it. Both resembled v-erbA, a cancer gene carried by a chicken virus. By the end of 1986 the cell’s normal version of that gene was known to encode a receptor for thyroid hormone. The human genome holds 48 genes of this family, and they share one plan:

• A variable front end. It differs most between receptors. It carries one surface for switching genes on, called AF-1, that can work without the hormone-binding domain.
• A DNA-binding domain of about 70 amino acids, folded around two zinc ions. Four cysteines hold each zinc. The fold sets one helix into the major groove of DNA, where it reads six bases.
• A hinge.
• A hormone-binding domain: about twelve helices packed around a pocket lined with hydrophobic side chains. It also carries the surface that pairs two receptors, and a second switching surface, AF-2.

Most of these receptors work in pairs, so their binding site on DNA, called a response element, has two six-base halves. For the estrogen receptor:

AGGTCA nnn TGACCT (two halves that read the same on opposite strands, three bases apart; n is any base)

Most of the 48 were found by their resemblance to known receptors before anyone knew what they bind, and were named orphan receptors. Several have since been matched to a ligand (the small molecule a receptor binds): the PPARs bind fatty acids, LXR binds oxidized cholesterol and FXR binds bile acids. So this family also reports on the cell’s own lipids. Others still have no agreed ligand.

Two ways to wait for a hormone

What does a receptor do before its hormone arrives? The family has two answers.

Waiting with a chaperone. The glucocorticoid and androgen receptors wait mostly in the cytoplasm, held by a group of chaperones built around Hsp90. A chaperone is a protein that helps other proteins hold their shape. Here it keeps the pocket ready for hormone and keeps the receptor off DNA. When hormone binds, the receptor is released, moves into the nucleus, pairs with a second copy of itself and binds its response elements. The estrogen receptor is held by the same chaperones, but experiments in 1984, on enucleated cells and with antibody staining, placed most of it in the nucleus even before hormone arrives.

Waiting on the DNA. The receptors for thyroid hormone, retinoic acid and vitamin D, and the PPARs, stay in the nucleus. They bind DNA paired with a partner from the same family called RXR. Their two half-sites point the same way, and the gap between them helps decide which pair fits: one base for PPAR, three for vitamin D, four for thyroid hormone, five for retinoic acid. Without hormone, the thyroid hormone and retinoic acid receptors hold co-repressors, proteins that keep the gene off. Hormone releases the co-repressor and brings in its place a co-activator, a protein that helps switch the gene on. An empty receptor of this kind is an active “off”: a gene carrying an empty thyroid hormone receptor is quieter than the same gene with no receptor at all. This repression is best documented for those two receptors, so do not assume it for every partner of RXR.

Co-repressors and co-activators act largely on chromatin, the packaged form of DNA. DNA in the nucleus is wound around histone proteins, and tightly packed DNA is hard to transcribe. Co-activators bring enzymes that add acetyl groups to histones, which loosens the packing and helps RNA polymerase start; co-repressors bring enzymes that take the acetyl groups off.

What the hormone changes: one helix closes

The moving part is the last helix of the hormone-binding domain, helix 12. With an agonist (a molecule that switches the receptor on) in the pocket, helix 12 folds across the mouth of the pocket like a lid. In that position it completes a shallow groove on the surface of the domain. Co-activators carry a short helix with the sequence LXXLL, where L is leucine and X is any amino acid, and that helix fits the groove. The groove is the hormone’s whole message.

An antagonist, a molecule that occupies the receptor without switching it on, fills the pocket and leaves the helix in the wrong place. The liver converts tamoxifen to forms that bind the estrogen receptor much more tightly, such as 4-hydroxytamoxifen. That molecule fits the pocket, but one side chain is too long to stay inside and sticks out where the lid should close. In a 1998 crystal structure from the laboratories of Greene and David Agard, helix 12 lies in the co-activator groove itself. The groove is taken, so the co-activator has nowhere to bind.

Design problem

Tamoxifen blocks the action of estrogen in the breast and imitates it in bone and in the uterus. One receptor and one drug give opposite answers. Propose what could differ between the tissues to explain this, using the parts in this lesson. Then design one experiment that would test your proposal.

A drug of this kind is called a selective estrogen receptor modulator. It is neither an agonist nor an antagonist until you name the tissue. Raloxifene is a second example: it acts like estrogen in bone, does not stimulate the uterus, and is used against osteoporosis.

Why it takes hours, and where it meets medicine

A signal that works through new RNA and new protein cannot be fast. The first new messenger RNAs appear within tens of minutes of the hormone, and the change in the cell’s proteins builds over hours. Adrenaline acting through cyclic AMP (the lesson on cAMP and PKA) takes seconds. The slow route has its own strength: the new proteins remain after the hormone has gone, and one receptor can reset many genes at once.

This family meets medicine in many places. Here are three.

Glucocorticoids. Prednisone and dexamethasone are synthetic relatives of cortisol, used to calm inflammation. The hormone-bound receptor can bind NF-κB (from the lesson on Toll-like receptors and NF-κB) at inflammatory genes and blunt it without binding DNA itself, and in some cells it also raises production of the inhibitor IκB. Nearly every cell has this receptor, so long treatment also raises blood glucose and thins bone.

PPAR-γ. This receptor drives the making of fat cells. The thiazolidinediones, such as pioglitazone, are agonists for it and make the body more sensitive to insulin. The drugs were found by their effect in diabetic animals, and in 1995 their target turned out to be PPAR-γ.

Androgen insensitivity. The gene for the androgen receptor is on the X chromosome. A person with one X and one Y whose only copy cannot work has testes that make testosterone and cells that cannot read it. In the complete form of the condition the body develops along female lines. Blood testosterone is normal or high. Measure the hormone and nothing looks wrong. The fault is in the receptor, and in this pathway the receptor is nearly the whole pathway.