Gs, Gi, Gq and G12/13: The Four G-Protein Subfamilies Explained

A human cell can carry roughly 800 different G-protein-coupled receptors on its surface. Behind all of them sit just sixteen G-alpha genes. Those sixteen sort into exactly four subfamilies — Gs, Gi/o, Gq/11 and G12/13 — and which family a receptor engages determines what the cell does with the signal. If you're reading mechanism papers on a compound labeled for research use only, that family name is usually the first real piece of information on the page: it tells you which assay produced the numbers. We'll walk each of the four, then the shared switch underneath them, then why the neat four-way map gets messy in practice.

What a Heterotrimeric G Protein Actually Is

The short version: a three-part relay on the inside face of the cell membrane that turns a receptor's shape change into a chemical signal.

Three subunits, one switch

Heterotrimeric G proteins are built from alpha, beta and gamma subunits, with the beta and gamma pair behaving as a single functional unit. The alpha subunit binds and hydrolyzes GTP — it's the part that carries the family name. Ligand binding changes the receptor's shape, and that change drives GDP-for-GTP exchange on the alpha subunit. The GTP-loaded alpha subunit then dissociates from the beta-gamma dimer, and both halves go on to engage effectors of their own.

Why four families and not sixteen

Classification runs on alpha-subunit sequence similarity, which tracks effector choice closely. Grouped that way, the sixteen human G-alpha genes fall into four families with four distinct outputs: Gs pushes cyclic AMP up, Gi/o pulls it down, Gq/11 routes to calcium and protein kinase C, and G12/13 routes to the cytoskeleton. Everything below elaborates those four sentences.

Gs — The Adenylyl Cyclase Accelerator

In one line: Gs is the family that makes a cell produce more cyclic AMP.

Members and effector

Gs is the smallest of the four — just Gαs and Golf, encoded by GNAS and GNAL. Its effector is adenylyl cyclase, and activation converts ATP into cyclic adenosine monophosphate. Golf's restricted expression in olfactory tissue is a reminder that these families are specialized by location as much as by effector.

Downstream: cyclic AMP and PKA

Cyclic AMP's main destination is protein kinase A, described in the literature as a tetrameric enzyme consisting of two catalytic subunits that phosphorylates downstream targets. We covered second messengers like cAMP, IP3, DAG and calcium separately.

Receptors that run through Gs

Two well-characterized examples: the GLP-1 receptor in pancreatic beta-cells, where the predominant pathway involves Gαs coupling to adenylyl cyclase, and the beta-adrenergic receptors ADRB1, ADRB2 and ADRB3. In practical terms, Gs coupling gets read out in cell culture as a cyclic AMP accumulation assay — which is why so much receptor literature leads with a cAMP EC50 value.

Gi/o — The Same Dial, Turned Down

In one line: Gi/o is the brake on the exact enzyme Gs accelerates.

The largest family

Gi/o spans eight genes and a range of specialized proteins: Gαi1 through Gαi3, Gαo, transducin, gustducin and Gαz, from GNAI1, GNAI2, GNAI3, GNAO1, GNAT1, GNAT2, GNAT3 and GNAZ. Transducin carries the signal in photoreceptors, gustducin in taste cells. This family is tuned by tissue, not just by effector.

Two mechanisms, not one

The canonical action is inhibition of adenylyl cyclase, which lowers cyclic AMP while also engaging PI3K and MAPK signaling. But Gi/o has a second route the Gs family doesn't: the released beta-gamma dimer opens potassium channels and closes calcium channels directly, with no second messenger involved.

Opposed inputs in the same cell

Because Gs and Gi converge on one enzyme, a cell expressing both can titrate its own cyclic AMP level. The beta2-adrenergic receptor shows this inside a single receptor: it can interact with Gαi proteins, which antagonize the Gαs pathway by inhibiting adenylyl cyclase. That bears directly on how you read a negative result. A compound that looks inert on a cyclic AMP readout may simply be engaging the inhibitory arm.

Gq/11 — Phospholipase C, Calcium and PKC

In one line: Gq skips cyclic AMP entirely and splits a membrane lipid into two signals instead.

Members

The Gq/11 subfamily comprises four related proteins — Gq, G11, G14 and G15/16 — encoded by GNAQ, GNA11, GNA14 and GNA15. They share structural similarity and have overlapping but non-identical roles.

The PLC-beta cascade

Activated Gαq stimulates phospholipase C-beta, which hydrolyzes membrane phosphatidylinositol 4,5-bisphosphate into two second messengers at once: inositol 1,4,5-trisphosphate and diacylglycerol. IP3 diffuses through the cytoplasm and mobilizes calcium from intracellular stores. DAG stays membrane-bound and switches on protein kinase C isoforms. The resulting calcium mobilization and PKC activation subsequently activate mitogen-activated protein kinases, which carry the signal to nuclear transcription of target genes.

Receptors on this branch

The Gq roster is long: muscarinic M1, M3 and M5, alpha-1 adrenergic receptors, the angiotensin II type 1 receptor, endothelin and vasopressin receptors, the gonadotropin-releasing hormone receptor, and several chemokine receptors. The practical consequence is that Gq coupling is typically measured as a calcium-flux assay — a different instrument, time course and data shape than a cyclic AMP experiment.

G12/13 — RhoGEFs and the Cytoskeletal Branch

In one line: G12/13 changes a cell's shape rather than its second-messenger chemistry.

Two proteins, two genes

This family is just Gα12 and Gα13, two 43 kDa proteins encoded by GNA12 and GNA13. Neither produces a diffusible second messenger, which is the main reason it gets described apart from the other three.

The RhoGEF route

Instead the signal passes to small GTPases through guanine nucleotide exchange factors. Three canonical RhoGEFs do the work — p115-RhoGEF, PDZ-RhoGEF and LARG (leukemia-associated RhoGEF) — each catalyzing GDP-GTP exchange on Rho. A non-canonical factor, Ric-8A, binds Gα13 directly and potentiates Rho activity, so the route depends on cellular context.

Rho and ROCK outputs

Activated Rho engages Rho-associated kinase. From there, the commonly reported downstream events are actin cytoskeleton modification and induction of the JNK signaling pathway, along with activation of the Serum Response Element; moesin phosphorylation is one concrete cytoskeletal step. The two family members aren't interchangeable. Gα13 knockout in mice is embryonically lethal because blood vessels fail to form, and Gα12 doesn't compensate.

Regulated by different machinery

Their control layers differ too. Gα13 is governed post-transcriptionally by miR-182, miR-200a and miR-31, which suppress translation without changing mRNA levels, while Gα12 is governed transcriptionally through c-Jun/AP-1 at the GNA12 promoter plus miR-564.

The Shared GDP/GTP Switch Underneath All Four Families

In one line: whatever the output, all four families run on the same timer.

The switching loop

Receptor activation drives GDP-to-GTP exchange on the alpha subunit, which adopts an active conformation and engages its effector. Termination is built in: intrinsic GTPase activity hydrolyzes GTP back to GDP, causing reassociation of the inactive heterotrimer and terminating the signal. That self-limiting design allows rapid, precise temporal control — and it relates closely to why receptor signaling fades in cell culture.

RGS proteins set the clock

The speed of that shutoff is tunable. Regulators of G protein signaling act as GTPase-activating proteins, accelerating hydrolysis on the alpha subunit. They're the built-in off switch for every family rather than a fifth pathway.

Toxins and hotspot mutations confirm the model

Two bacterial toxins pin the mechanism to specific residues. Cholera toxin ADP-ribosylates R201 of Gαs, permanently blocking its GTPase activity and locking the subunit active; pertussis toxin ADP-ribosylates Gαi so it can no longer inhibit adenylyl cyclase. Tumor-biology data converges on the same residues from another direction — activating Gα11 or Gαq mutations appear in more than 90 percent of uveal melanoma, primarily at Q209 — and the engineered Q231L substitution yields constitutively active, GTPase-deficient Gα12/13 variants used widely as laboratory tools. These are observations from cell and tumor biology literature, not statements about any product.

One Receptor, Several G Proteins — Why Coupling Is Not One-to-One

In one line: the four-family map is a reading guide, not a lookup table.

The arithmetic

Roughly 800 human GPCRs and about 1,000 known ligands funnel into sixteen G-alpha genes. That asymmetry makes overlap inevitable: one G protein has to serve many receptors, and many receptors engage more than one G protein.

Promiscuous coupling and biased agonism

Some serotonin receptors activate both Gi and Gq, while certain dopamine receptor subtypes are far more restricted. Beyond that, the same receptor can activate different G-protein subtypes or beta-arrestin signaling in a ligand-dependent manner — biased agonism. GPR40 is the textbook case: the partial agonist fasiglifam is beta-arrestin-biased, while the full agonist AM-1638 enables dual coupling to both Gαq/calcium and Gαs/cyclic AMP. Structural work makes these distinctions visible, as in the cryo-EM structures of GLP-1 receptor Gq and Gi coupling.

What decides the pairing

Selectivity appears to be encoded in a surprisingly small set of residues. The G-alpha "G-Red" selectivity sector comprises 28 residues, concentrated in the beta1-beta2 loops, the C-terminal portions of the alpha3 and alpha4 helices, and the alpha5 helix — and 24 of the 28 overlap previously mapped selectivity barcodes. Some sit in the alpha-helical domain, which never contacts the receptor, implying selectivity is partly allosteric rather than purely interface fit.

The Four Families Are Not Four Separate Wires

In one line: remove one family and the others change.

Crosstalk runs across the boundaries. Gαq, the classic phospholipase C activator, can also activate Rho signaling proteins classically assigned to G12/13 signaling. Abundance is reciprocally regulated too: siRNA knockdown of Gq upregulated Gi subunits, with basal Gi-1 roughly 2-fold higher in Gq/11-null mouse embryonic fibroblasts but not in G12/13-null ones. In the same work, cells lacking Gq/11 or G12/13 showed roughly 8-fold and 20-fold higher active beta-catenin than wild-type fibroblasts, implying G12/13 restrains beta-catenin about 2.5-fold more strongly. The takeaway is specific: a knockout result describes a rewired cell, not a cell with one protein missing.

Why the Four-Family Map Matters When You Read Peptide Literature

If you're researching a compound sold for laboratory use, the G-protein family named in a paper's first figure does more work than it looks like. It tells you which assay generated the numbers: cyclic AMP accumulation points to Gs or Gi, calcium flux points to Gq, and migration or cytoskeletal readouts point to G12/13. Matching a reported observation back to the assay that produced it is the most direct way to judge whether a claim rests on solid ground — the same reasoning we applied when comparing receptor pharmacology between two related ligands. Every observation described here comes from cell-culture or animal-model literature.

Frequently Asked Questions

How many G protein subtypes are there in humans?

Sixteen distinct G-alpha subunit genes, sorting by sequence similarity into exactly four subfamilies: Gs, Gi/o, Gq/11 and G12/13. Gi/o is the largest with eight genes; Gs has two (GNAS, GNAL), Gq/11 has four (GNAQ, GNA11, GNA14, GNA15) and G12/13 has two (GNA12, GNA13). Those sixteen transducers serve roughly 800 GPCRs, which is why coupling patterns overlap rather than pairing one-to-one.

What is the difference between Gs and Gi signaling?

They act on the same enzyme in opposite directions. Gs stimulates adenylyl cyclase, raising cyclic AMP and activating protein kinase A; Gi/o inhibits it and lowers cyclic AMP. Because they converge on one target, a cell expressing both can titrate its cyclic AMP level by balancing the two inputs. Gi/o also has a mechanism Gs lacks: its released beta-gamma dimer opens potassium channels and closes calcium channels directly.

What does Gq do that Gs and Gi do not?

Gq/11 bypasses the cyclic AMP system entirely. Activated Gαq stimulates phospholipase C-beta, which cleaves membrane PIP2 into two second messengers at once: IP3, which releases calcium from intracellular stores, and DAG, which activates protein kinase C in the membrane. That gives Gq a calcium-plus-PKC output, and both branches feed into MAP kinase signaling.

Why is G12/13 usually described separately from the other three?

Because its output is structural rather than metabolic. Gα12 and Gα13 don't generate a diffusible second messenger. They activate RhoGEF proteins, which switch on Rho GTPases, which signal through Rho-associated kinase to reorganize the actin cytoskeleton. The readouts are cell shape, migration and smooth-muscle contractility rather than a concentration of cyclic AMP or calcium, so G12/13 experiments use different assays from the other three families.

Can one receptor couple to more than one G protein subtype?

Yes, and many do. Some serotonin receptors activate both Gi and Gq. The beta2-adrenergic receptor engages Gαs but can also interact with Gαi, which antagonizes its own cyclic AMP output. Which transducer a receptor engages can even depend on which ligand is bound — biased agonism. This is why careful published work specifies both the receptor and the pathway the result was read out on.

Putting It All Together

Four families, four effector routes: adenylyl cyclase up, adenylyl cyclase down, phospholipase C, and Rho — all running on one GDP/GTP switch, with RGS proteins setting the shutoff speed. Held in mind, that map earns its keep every time you open a mechanism paper. It tells you which assay a given number came from, where crosstalk might complicate the interpretation, and which claims rest on a single family's readout when the biology involves more than one. For the chemistry one layer down, our explainer on cAMP, IP3, DAG and calcium picks up where this one leaves off.

For research use only. Not for human or animal consumption of any kind. The information in this article is for educational purposes only and is not intended to diagnose, treat, cure, or prevent any disease. The statements made have not been evaluated by the U.S. Food and Drug Administration. These products are NOT FDA APPROVED. Please consult with a licensed healthcare professional before making any decisions regarding your health or research.

Optides LLC is a chemical supplier. Optides LLC is not a compounding pharmacy or chemical compounding facility as defined under 503A of the Federal Food, Drug, and Cosmetic Act. Optides LLC is not an outsourcing facility as defined under 503B of the Federal Food, Drug, and Cosmetic Act.