10.1: Membrane and Membrane Proteins

Search Fundamentals of Biochemistry

Learning Goals (ChatGPT o3-mini, 2/2/25)

(Learning goals written by Claude, Sonnet 4.6, Anthropic)

Classification and Structural Features of Membrane Proteins

• Distinguish among peripheral, lipid-anchored, and integral (transmembrane) membrane proteins based on their mode of membrane association, depth of membrane penetration, and ease of experimental removal, and explain how each class is stabilized by specific noncovalent or covalent interactions with the bilayer.
• Compare the structural organization of single-pass (biotopic) and multi-pass (polytopic) alpha-helical transmembrane proteins and beta-barrel transmembrane proteins, explaining how the secondary structure used in membrane crossing reflects the physical demands of a hydrophobic bilayer interior.

Lipid-Anchored and Peripherally Associated Proteins

• Describe the four major types of covalent lipid modifications that anchor proteins to membranes — N-myristoylation, S-palmitoylation, isoprenylation (farnesyl and geranylgeranyl), and GPI anchors — identifying the lipid attached, the linkage chemistry, the location in the membrane, and a specific biological example of each.
• Explain how calcium-dependent membrane binding of prothrombin through its GLA domain — requiring γ-carboxylation of glutamate residues by a Vitamin K-dependent enzyme — illustrates how peripheral proteins can acquire membrane affinity through post-translational modification and metal ion coordination.

Transmembrane Proteins: Structure and Biological Function

• Explain how single-pass transmembrane proteins such as cadherins and receptor tyrosine kinases (e.g., the insulin receptor) connect extracellular signals or adhesion events to intracellular cytoskeletal or signaling networks, and describe the structural basis for ligand-induced dimerization and autophosphorylation in receptor tyrosine kinases.
• Describe the structural and functional properties of G protein-coupled receptors (GPCRs) as seven-pass alpha-helical polytopic proteins, using the cannabinoid receptor as an example to explain how ligand binding in the transmembrane domain is transduced to conformational changes that activate cytoplasmic G proteins.
• Contrast the alpha-helical and beta-barrel (BBF) folds used by polytopic transmembrane proteins, citing the voltage-dependent anion channel (VDAC) and bacterial outer membrane factor (OMF) as examples of beta-barrel pores, and explain how the gating mechanism of VDAC links membrane potential to metabolite flux across the outer mitochondrial membrane.

Membrane Mimetic Systems for Structural Studies

• Compare the use of detergent micelles, bicelles, nanodiscs, and SMALPs as membrane-mimetic environments for NMR and structural studies of membrane proteins, explaining how each system balances the need to preserve native lipid bilayer properties with the practical requirements of solution-phase spectroscopy.

Introduction

One easily understandable function of membrane bilayers is to separate the inside and outside of the cell or intracellular organelles. Yet, as we mentioned before, such barriers cannot be so rigid or impenetrable that they prevent the movement of materials across the membrane. Also, all cells must sense and respond to their environment through signal transduction. We have already discussed lipid molecules involved in signaling. Now, let's turn our attention to membrane-associated proteins that confer additional functionality. Figure \(\PageIndex{1}\) reviews some of the features of membranes we've discussed before and shows a simple bilayer (top) to the complicated membrane/cell wall of bacteria.

Figure \(\PageIndex{1}\): Membrane from simple to complex

Types of Membrane Proteins

Although we presented this image earlier, Figure \(\PageIndex{2}\) reviews the details that should now be clearer. In this section, we will explore membrane proteins in more detail.

Figure \(\PageIndex{2}\): A cell membrane with peripheral and integral membrane proteins. https://commons.wikimedia.org/wiki/F...Components.jpg. This file is licensed under the Creative Commons Attribution 4.0 International license.

Proteins can be loosely associated with the membrane (peripheral or extrinsic), embedded deeply, or pass through the membrane and become a transmembrane (also called integral or intrinsic) protein. Sometimes they pass through a single alpha helix, while others pass through multiple times (for example, 7 times in G-protein-coupled receptors). They can also be classified based on the number of membrane leaflets they cross, as shown in Figure \(\PageIndex{3}\).

Figure \(\PageIndex{3}\): Peripheral and biotopic, monotopic, and polytopic integral membrane proteins

Peripheral Proteins

The proteins interact with the membrane via protein-lipid headgroup interactions but may slightly penetrate the membrane. Those that do would be classified as monotopic peripheral. Peripheral proteins are generally easy to remove from a membrane in vitro by changing the solution ion concentration, as the interactions are often of an ion-ion nature. The first model below shows a matrix Metalloproteinase (MPP) 12 binding to a lipid bilayer. This protein is involved in inflammation, wound healing, arthritis, cardiovascular disease, and neural synapse remodeling, suggesting a broad role in recovery from cellular and tissue aberrations. Macrophages secrete MMP-12, so it is considered a water-soluble (aqueous) protein. It travels to viral cells and appears to be active not in aqueous solution but near membranes, suggesting that the enzyme is activated by binding to the bilayer.

Studies show the catalytic domain of MMP 12 can bind bilayers through both α- and β-secondary structure regions of the protein. Figure \(\PageIndex{4}\) shows an interactive iCn3D model of the protein and its interaction with the membrane through the alpha-helical region. Once bound to the membranes, catalytic activity increases.

Figure \(\PageIndex{4}\): Catalytic domain of peripheral membrane Protein MMP 12 interacting with a bilayer. (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...wyF6WvRMx9JP6A

Other MMPs localize to membranes in other ways. MMP-7 interacts with heparan sulfate proteoglycans (CD44) and lipid rafts. Others bind transmembrane proteins like integrins.

Other examples of peripheral proteins include many precursor forms of protein clotting factors. Clotting is initiated when the serine protease thrombin cleaves fibrinogen to form fibrin, which self-associates to form a fibrin clot, or when thrombin activates platelet receptors. The soluble precursor of thrombin, prothrombin, a zymogen, is activated on membrane binding through interactions with several proteins assembled on a negatively charged phospholipid (like phosphatidylserine) bilayer in the prothrombinase complex. How does the precursor zymogen interact with the membrane? It requires calcium ions, which bind to a series of gamma-carboxylated glutamic acid (GLA) residues on the zymogen. The enzyme that carboxylates the zymogen depends on Vitamin K.

Figure \(\PageIndex{5}\) shows an interactive iCn3D model of bovine prothrombin Fragment 1 (N terminal) bound to a bilayer through its GLA domain (1NL2).

Figure \(\PageIndex{5}\): Bovine prothrombin Fragment 1 (N terminal) bound to a bilayer through its GLA domain (1NL2). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...SG6qByvpAsvwe6

The Gla sidechains are shown in CPK-colored sticks and interact with Ca²⁺ ions (gray spheres). Click on this link to see a zoomed view of just the calcium ions and Gla sidechains: https://structure.ncbi.nlm.nih.gov/i...xydtj8a4WHgdb7

The Gla domain, in the absence of calcium ions, is disordered. Upon binding, an ordered linear alignment of bound calcium ions forms, stabilizing the Gla domain's ordered structure and enabling interaction with the membrane. Three nonpolar amino acid side chains, Phe 5, Leu 6, and Val 9, are now clustered and exposed, allowing penetration of this hydrophobic patch partway into the membrane. They are represented in cyan spacefill just underneath the surface of the red dots in the model above (the red dots are dummy atoms that represent the outer bilayer leaflet). Given this level of penetration, this protein domain would be considered monotopic.

What the model does not show is the role of negatively charged phosphatidylserine. Studies have shown that the head group of serine in lysophosphatidylserine (which has only one acyl group) provides additional ion-ion interactions with the Ca²⁺ ions that bind Gla residues 17 and 21. Arg 10 and Arg 16 also interact with the phosphatidyl serine head group. Phosphatidylcholine could also spatially fit into the active site, but electrostatic interactions would prevent it. Why?

Lipid-Anchored Proteins

We have studied lipids, proteins, and carbohydrates. Although phospholipids can spontaneously form bilayers, the actual structure of biological membranes is made much more complicated by adding protein and carbohydrate substituents to the membrane. Soluble proteins can be made to insert into bilayers by adding nonpolar attachments. Localization to a membrane alters the protein's functional expression. Several examples of such attachments are described below.

Fatty acid linkers

Two common covalent modifications of proteins are N-myristoylation (attached myristic acid - 14:0 - through an amide link) and S-palmitoylation (attached palmitic acid - 16:0 - through a thioester link with a Cys).

Myristoylation is usually a cotranslational modification of eukaryotic and viral proteins, occurring after cleavage of the N-terminal methionine. Figure \(\PageIndex{6}\) shows an image of the serine/threonine phosphatase 2C (1A6Q) with its N-terminal glycine myristoylated. It should be obvious how this post-translationally modified protein interacts with a membrane.

Figure \(\PageIndex{6}\): The serine/threonine phosphatase 2C (1A6Q) with its N-terminal glycine myrisoylated

This modification is a key part of initiating immune system signal transduction pathways. The modification is catalyzed by N-myristoyltransferase (NMT), which uses myristoyl-coenzyme A as the fatty acid acyl donor. This partly activates the protein's function by reducing the dimensionality of substrate diffusion to the protein to a 2D surface on the membrane, rather than a 3D search in the cytoplasm. NMT acylates a protein at this consensus sequence: G¹X²X³X⁴S/T⁵X⁶R⁷R⁸.

Likewise, many signaling proteins are palmitoylated, thereby recruiting them to membranes. Small G proteins, such as Ras, Rho, and the alpha subunit of heterotrimeric G proteins, are often palmitoylated. This modification can also be found in transmembrane proteins where localization is not an issue (see example xx below). In such circumstances, the modification might help target proteins to membrane rafts. Palmitic acid is saturated, and attaching it to a protein might target the modified protein to more ordered membrane regions with cholesterol and sphingolipids within rafts.

Isoprenoids linkers:

The isoprenoids farnesyl (15C) or geranylgeranyl (20C) are added to a CAAX carboxy-terminal sequence in a target protein, such as RAS, where C is Cys, A is aliphatic, and X is any amino acid, which helps target proteins to the membrane. The enzymes used for these modifications are farnesyltransferase (FTase) and protein geranylgeranyltransferase I (GGTase I), respectively. This and the other modifications have the potential to do more than target proteins to the membrane. The modification can also modulate protein-ligand interactions and protein stability. Ras, a key signaling protein, is a target of prenylation.

Ras and other small G proteins are involved in many human cancers. As the G protein Ras has a somewhat billiard-ball-like surface with obvious sites for targeting drugs that would affect its aberrant function in cancers, efforts have been made to target the prenyltransferases that anchor it to the membrane.

In humans, there are three genes in the Ras family, H-Ras and N-Ras, whose gene products localize to both plasma and Golgi membranes, and K-Ras, which localizes predominantly to the plasma membrane. These and other G proteins bind GTP and possess GTPase activity. The GTP-bound form is active, while the GDP form is inactive. Point mutations that attenuate or prevent GTP cleavage leave the protein continually activated, contributing to oncogenesis.

KRas has two predominant isoforms, 4A (the canonical form, also called 2A) and 4B (also called 2B), arising from alternative splicing of the primary RNA transcript. The C-terminal protein sequences of isoforms 4A and 4B differ significantly.

Isoform 4A: QYRLKKISKEEKTPGCVKIKKCIIM

Isoform 4B: KHKEKMSKDGKKKKKKSKTKCVIM

The farnesylation motif site containing the modified Cys is highlighted in yellow above. The same cysteine is also often carboxymethylated. The Cys six residues from the farnesylated Cys in isoform 4A are also often palmitoylated.

Figure \(\PageIndex{7}\) shows isoform KRas 4B bound to a membrane bilayer through its farnesylated tail (PDB file provided by Gorfe, Nair, and McCammon). The tail is essential for its function at the plasma membrane, where KRAS-mediated signaling events occur. Phosphodiesterase-δ (PDEδ) binds to KRAS4b and is important in targeting it to cellular membranes. Note that the farnesyl attachments only penetrate part of the upper leaflet.

Figure \(\PageIndex{7}\): Isoform KRas 4B bound to a membrane bilayer through its farnesylated tail

Glycosyl-phosphatidylinositol linkers

Soluble cytosolic proteins can normally attach to membranes by adding a glycosylphosphatidylinositol (GPI) anchor. The attachment usually contains a conserved tetrasaccharide core of three mannoses (Man) and one unacetylated glucosamine (GlcN) linked to the carboxy terminus of the protein. The GPI can be further modified with additional galactoses and mannoses, as well as modifications to the PI group, which help secure the protein in the membrane. Figure \(\PageIndex{8}\) shows the common backbone for GPI anchors. Note the additions of the phosphoethanolamines to the core polysaccharide.

Figure \(\PageIndex{8}\): Common backbone of GPI anchors

GPIs in eukaryotic cells link many surface antigens, adhesion molecules, and hydrolases to the membrane. GPIs from Plasmodium falciparum, the malarial parasite that kills about 2 million people each year, appear to act as toxins and are the most common CHO modification of the parasite protein. Mice immunized against the GPI sequence, NH₂-CH₂-CH₂-PO₄-Man (α1-2) 6Man (α1-2) Man (α1-6) Man (α1-4) GlcNH2 (α1-6) myo-inositol-1,2-cyclic-phosphate, were substantially protected from malarial symptoms and death after they were exposed to the actual parasite.

Figure \(\PageIndex{9}\) shows a cross-section of a membrane (with cholesterol, PE, SM) containing the glycosylated form of the human complement regulatory protein CD59 protein (1cdr) with a GPI anchor attached at its C-terminus. Note that the middle part of the anchor (glycan) holds the protein well above the top of the lipid bilayer. The soluble protein is also glycosylated. The protein binds to complement proteins C8 and/or C9, which are effector immune proteins that assemble on the surface of a cell undergoing lysis.

Figure \(\PageIndex{9}\): Cross section of a membrane containing the glycosylated form of the human complement regulatory protein CD59 protein (1cdr) with a GPI anchor attached at its C-terminus

The GPI anchor is shown in spacefill. Note that it only extends halfway into the bilayer, as expected from the size of the fatty acids attached to the phosphatidyl inositol. The glycan part of the GPI is shown in spacefill between the lipid and its protein attachment site. The protein is also glycosylated in the extracellular domain.

Something new!

An unexpected type of glycosylated molecule has been found in the outer leaflet of mammalian cells - a glycosylated RNA, as shown in Figure \(\PageIndex{10}\). This adds RNA to lipids and proteins, making them targets for glycosylation. These surface glycoRNAs interact with antibodies against dsRNA and the Siglec lectin family. They are found in cells in vivo and cultured cells in vitro.

Figure \(\PageIndex{10}\): A glycoRNA - a small noncoding RNAs with sialylated glycans. Park. https://doi.org/10.14348/molcells.2021.0178
www.molcells.org. Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/3.0/.

Flynn et al., Small RNAs are modified with N-glycans and displayed on the surface of living cells, Cell (2021), https://doi.org/10.1016/j.cell.2021.04.023

Transmembrane (Integral) Proteins

These proteins pass through the membrane either in a single pass, usually with a single alpha-helix, or in multiple passes, with many membrane-spanning helices. For example, G protein-coupled receptors, often called serpentine receptors, cross the membrane seven times. There are three types based on the number of times the protein crosses the membrane and the type of secondary structure used in crossing: biotopic (single pass), alpha-helical polytopic, and beta-barrel. These proteins are found in all types of membranes and have many types of functions, from receptors, receptor ligands, structural, adhesion, transport, gene regulation, and transport.

Transmembrane Biotopic - Single Pass Proteins

There are four types of single-pass transmembrane proteins:

• Type I: N-terminal outside of the cell (extracellular), and the precursor signal sequence on the N-terminus, which is a localization sequence, is removed
• Type II: N-terminal intracellular, with the transmembrane domain close to the N-terminus
• Type III: N-terminus extracellular and no signal sequence in precursor protein
• Type IV: N-terminus intracellular and the transmembrane domain close to the C-terminus

The transmembrane domain of single-pass integral membrane proteins consists of a single alpha-helix, with nonpolar side chains extending outward from the helical axis and interacting with the nonpolar lipid regions of the membrane. These nonpolar sides are more stable in nonpolar environments.

To study such proteins in a less complex environment, membranes are often "dissolved" in nonpolar, single-chain amphiphilic detergents. These single-chain amphiphiles form micelles without membrane proteins, but form mixed micelles in which the nonpolar part of the protein is surrounded by the detergent's nonpolar acyl chains.

Figure \(\PageIndex{11}\) (top) below shows just the transmembrane and juxtamembrane (next to the membrane) domains of the single-pass Notch protein, which is critical in many signal transduction pathways.

Figure \(\PageIndex{11}\): Transmembrane and juxtamembrane domains of the single-pass Notch protein in a bilayer and an octylglucoside micelle

The top images in the figure above show different ways to represent the protein in the bilayer, with the right-hand image showing a cross-section through the membrane to better illustrate how the protein traverses the bilayer.

The bottom images in the figure above show the protein after excess detergent, in this case, octylglucoside, is added to the protein-containing bilayer.

 

Studying membrane proteins in lipid aggregates - NMR applications

Recent Updates: 9/26/24

Various discrete lipid aggregates are used to conduct NMR studies of membrane proteins in a more realistic biological environment.  Examples are shown in the Figure \(\PageIndex{12A}\) below. 

Figure \(\PageIndex{12A}\): Schematic diagram of different membrane mimetic systems used in NMR studies: detergent micelle, bicelle, nanodisc, liposome, amphipols, salipro, and styrene–malic acid co-polymer lipid particles (SMALPS). The membrane protein is represented as a blue cartoon block. Yeh, V.; Goode, A.; Bonev, B.B. Membrane Protein Structure Determination and Characterization by Solution and Solid-State NMR. Biology 2020, 9, 396. https://doi.org/10.3390/biology9110396.  Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).  For an alternative image see https://www.cell.com/chem/fulltext/S...2822%2900092-4. 

An alternative description and use for general membrane protein purification and studies include a general polymer form and virus-like particles, as shown in Figure \(\PageIndex{12B}\) below. 

Figure \(\PageIndex{12B}\): Alternative descriptions and uses in membrane protein solubilization, purification, and characterization.  Creative BioLabs.  MAGICTM MEMBRANE PROTEIN SERVICES & PRODUCTS.  With permission (10/20/24)

Micelles: We've studied these in Chapter 10.2: Lipid Aggregates, and in the above descriptions. They form from the self-assembly of single-chain amphiphiles (detergents, such as octylglucoside and [mono]dodecylphosphocholine) as well as double-chain amphiphiles with short alkyl chains, such as hexanol-phosphatidylcholine (DHPC, also abbreviated C6DH).

Bicelle: Bicelles are made of phospholipids (such as dimyristoyl phosphatidylcholine - DMPC) that can form a bilayer and a detergent (a single-chain amphiphile that can form a micelle.  The detergent fills in the hydrophobic "edges" of the bilayer part to form the final structure. Bicelle size varies with a Q value, the [phospholipid]/[detergent] ratio.  When Q = 0 (high detergent), a micelle forms.

A phospholipid with short acyl chains such as dihexyl phosphatidylcholine (DHPC or C6DH) can replace the detergent. The structures of these two phospholipids (DMPC and DHPC) are shown in Figure \(\PageIndex{13}\) below.

Figure \(\PageIndex{13}\):  Structures of dimyristoyl phosphatidylcholine (DMPC) and dihexyl phosphatidylcholine (DHPC or C6DH)

Figure \(\PageIndex{14}\) below represents a cross-sectional view of a bicelle made from the above lipids.  The central white/gray chains represent the alkyl chains of the DMPC phospholipid bilayer.  The cyan spheres represent the carbon atoms of the smaller DHPC "detergent" that fill the edges of the central bilayer.  

Figure \(\PageIndex{14}\): Cross-sectional view of a bicelle made with DMPC (white carbons) and DHPC (cyan carbons) which are seen around the periphery.

Figure \(\PageIndex{15}\) below shows an iCn3D model of the same bicelle.  The black lines represent the alkyl chains of the central DMPC bilayers.  The cyan sticks represent the alkyl chains of the DHPCs that fill the edges.  Other atoms are shown in their own CPK colors.

Figure \(\PageIndex{15}\): Bicelle

Open iCn3D and follow these prompts to get an interactive model of this bicelle:  

• Download this file to your computer
• in iCn3D, File, Open File, and then open the iCn3D PNG Image

Nanodiscs:  Nanodiscs are like bicelles that use two copies of an amphiphilic protein (membrane scaffold protein (MSP) from high-density lipoprotein (HDL) instead of a detergent or small double-chain amphiphile to protect the exposed hydrophobic edges of a small section of a bilayer.  Figure \(\PageIndex{16}\) shows a cross-sectional view of a nanodisc composed of DMPC and two copies of the protein MSP.

Figure \(\PageIndex{16}\): Cross-sectional view of a nanodisc made with DMPC (white carbons) and the membrane scaffold protein (MSP, red helices). Carbon atoms are shown in cyan.  Image made with Pymol.

Figure \(\PageIndex{17}\) below shows an iCn3D model of the same nanodisc.  The cyan atoms represent the carbons in the alkyl chains of the DMPCs.  Other atoms are shown in their own CPK colors.  The two helical amphiphilic proteins are shown in cartoon form in red.

Figure \(\PageIndex{17}\):  A nanodisc with DMPC and the protein MSP.

Open iCn3D and follow these prompts to get an interactive model of this nanodisc:  

• Download this file to your computer
• in iCn3D, File, Open File, and then open the iCn3D PNG Image

Follow these instructions to view the hydrophobicity of the two alpha helices.

With the full model in the iCn3D window:

• Analysis, Defined Sets, choose Proteins
• View, View Selection,
• Style, SideChains, Sticks
• Color, Hydrophobicity, Normalized

Rotate the structure to see a top-down view, and you will observe that the hydrophobic side chains are all pointed inward in a fashion that would stabilize the exposed nonpolar alkyl chains of the phospholipids.

SMALPS (styrene–malic acid co-polymer lipid particles):  These are nanodiscs wrapped with this amphiphilic polymer instead of an amphiphilic protein.  The structure of SMALPS is shown in Figure \(\PageIndex{18}\) below.

Figure \(\PageIndex{18}\)

Figure \(\PageIndex{19}\) below shows an iCn3D model of SMALPS.  Rotate the image to confirm its that it is amphiphilic.

Figure \(\PageIndex{19}\):  SMALPS (styrene–malic acid co-polymer lipid particles)

Here are some examples of bitopic single-pass transmembrane proteins

Cadherins

All structures need support and connections. At the macro level, the skeleton supports the mass and organization of organs and tissues in whole organisms. Within an organ, how can cells hold together? How do they adhere to each other? Certainly not through outer leaflet lipid contacts, as the outer surface of the leaflet is typically charged. The extracellular matrix does provide some of the glue that holds cells together. At a more detailed level, transmembrane proteins are involved. One class of adhesive proteins is cadherins, a calcium-dependent cell adhesion molecule. There are over 100 human cadherins. They are mostly ditopic, single transmembrane-pass proteins. Their cytoplasmic domains interact with proteins such as catenin, which then bind to the cytoskeletal network composed of actin and other proteins. This provides a mechanism for the intracellular region to regulate the cell's extracellular interactions.

The extracellular domain is composed of five repeating "cadherin" domains, each approximately 110 amino acids in length, that can fold independently. Calcium ions bind at the domain interfaces. A cadherin can interact with other cadherin domains on other cadherins on other cells, leading to cell adhesion. Essentially, the receptor cadherin on one cell binds to the ligand cadherin on the other. As metastatic tumor cells lose their adhering feature and leave the primary tumor site, you would expect mutations in cadherins to be often involved. They may also be involved in cell sorting during morphogenesis, regulation of tight and gap junctions, and in the control of intercellular spacing.

Figure \(\PageIndex{20}\) shows a "constructed" image of cadherin-1 (1L3W) interacting with cytoplasmic β-catenin (1I7X) through a modeled transmembrane helix (amino acids QIPAILGILGGILALLILILLLLLFLRR, amino acids 706-731). No full-length membrane-bound cadherin structure is available.

Figure \(\PageIndex{20}\): A constructed" image of cadherin-1 (1L3W) interacting with cytoplasmic β-catenin (1I7X) through a modeled transmembrane helix

Membrane Protein Kinases

Kinases are enzymes that phosphorylate substrates. Hexokinase is a protein enzyme that catalyzes the phosphorylation of a hexose substrate, such as glucose. A protein kinase is a protein enzyme that phosphorylates a protein substrate. That protein could be another copy of itself or another protein. We will see in Chapter 12 that many protein kinases are involved in cell signaling. Many tyrosine protein kinases are bitopic, single-pass, integral membrane proteins that become active upon ligand binding. Typically, upon binding an extracellular ligand, two monomeric copies of the kinase form a dimer in the membrane, activating the kinase's cytoplasmic domain, which typically phosphorylates (using ATP as a substrate) the other member of the dimer in an "autophosphorylation" reaction. Sometimes, the dimers are held together by disulfide bonds.

Figure \(\PageIndex{21}\) shows a "constructed" image of the human dimeric insulin receptor.

Figure \(\PageIndex{21}\): aA "constructed" image of the human dimeric insulin receptor

One of the monomers is shown in gray. The other monomer is shown in colors corresponding to the protein's domain organization. Each extracellular dimer (6PXV) contains two insulin molecules (yellow spacefill). The intracellular domains (1IR3) are activated on insulin binding. No full-length structure of the full insulin receptor in a membrane is available.

Almost half of all helical membrane proteins in humans are bitopic, compared to 20-25% in prokaryotes. Humans have 10-20 fold more bitopic proteins than E.Coli. There appear to be about 196 bitopic proteins in E. coli (located in the inner membrane ) and 70 in M. jannaschii (Archaea in the plasma membrane). In humans, 57% are in the plasma membrane, with the remaining 43% distributed among the Golgi, ER, nuclear, mitochondrial, and chloroplast membranes. In single-celled yeast, only 8% are in the plasma membrane.

Beta-Dystroglycan

This protein is another bitopic protein with a single membrane-spanning alpha-helix. Dystroglycan is a dimer of alpha- and beta-subunits. Alpha-dystroglycan is a peripheral protein that binds beta-dystroglycan, a transmembrane protein. Alpha dystroglycan also binds the glycoproteins of Lassa virus and lymphocytic choriomeningitis virus, serving as viral receptors. It also binds the protein dystrophin, which is missing in Duchenne muscular dystrophy and affects 1 out of 5000 live male births. As an integral transmembrane protein, beta-dystroglycan connects the extracellular matrix to the cytoskeleton through dystrophin. Alpha- and beta-dystroglycan share the same gene, which codes one long protein proteolyzed post-translationally to form the alpha (N-terminal end) and beta subunits (C-terminal end).

Figure \(\PageIndex{22}\) shows a schematic outline of dystrophin and the dystrophin-associated glycoprotein complex (DAGC).

Figure \(\PageIndex{22}\): Schematic outline of dystrophin and the dystrophin-associated glycoprotein complex (DAGC). Dystrophin contains N-terminal (NT), middle rod, cysteine-rich (CR), and C-terminal (CT) domains. The middle rod domain is composed of 24 spectrin-like repeats (numbers in the cartoon, positively charged repeats are marked in white color) and four hinges (H1, H2, H3, and H4). Dystrophin has two actin-binding domains at NT and repeats 11-15, respectively. Repeats 1-3 interact with the negatively charged lipid bilayer. Repeats 16 and 17 form the neuronal nitric oxide synthase (nNOS)-binding domain. Dystrophin interacts with microtubules through repeats 20-23. Part of H4 and the CR domain binds to the β-subunit of dystroglycan (βDG). The CT domain of dystrophin interacts with syntrophin (Syn) and dystrobrevin (Dbr). Dystrophin links cytoskeleton components (actin and microtubules) to laminin in the extracellular matrix. Sarcoglycans and sarcospan do not interact with dystrophin directly, but they strengthen the entire DAGC, which consists of dystrophin, DG, sarcoglycans, sarcospan, Syn, Dbr, and nNOS.Disease Models & Mechanisms (2015) doi:10.1242/dmm.018424vailable via license: CC BY 3.0

Transmembrane - Alpha-helical polytopic

There are so many intriguing examples of these proteins. We'll illustrate just two.

Rhodopsin-like receptors and pumps

These proteins are involved in cell signaling and are the target of most pharmaceutical drugs. G protein-coupled receptors (GPCRs) are incredibly important, and we will discuss them extensively in the capstone chapter (28.2: At the cell membrane- receptors and receptor enzymes) for Vol 1.  

GPCRs are cell receptors that span the membrane seven times in a serpentine fashion. They bind ligands (neurotransmitters, hormones, etc.) in the extracellular or internal membrane domains (the latter for hydrophobic ligands) and, through propagated conformational changes, alter the cytoplasmic domain where they functionally interact with a heterotrimeric G protein.

Figure \(\PageIndex{23}\) shows an interactive iCn3D model of the human cannabinoid receptor with bound cholesterol and Δ ⁹ -tetrahydrocannabinol (Δ⁹ -THC) in spacefill (5xra). The red dummy atoms represent the outer leaflet, and the blue the inner.

Figure \(\PageIndex{23}\): Human cannabinoid receptor with bound cholesterol and Δ ⁹ -tetrahydrocannabinol (Δ⁹ -THC). (5xra). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...PaTFrkGFgKeP2A

Δ ⁹ -THC is a partial agonist and tunes the response of the receptor. The active site is conformationally somewhat flexible or plastic. Other ligands bound to it act as antagonists instead of agonists and must do so by eliciting inactive conformations.

ABC Transporter

Figure \(\PageIndex{24}\) shows an interactive iCn3D model of the P-glycoprotein multidrug resistance transporter protein (6nf1).

Figure \(\PageIndex{24}\): P-glycoprotein multidrug resistance transporter protein (6nf1). (Copyright; author via source). Click the image for a popup or use this external link:https://structure.ncbi.nlm.nih.gov/i...zzXduCZjhCMaw9

The spacefill ligands represent Zosuquidar, which binds with high affinity to P-glycoprotein and inhibits its activity, making it a cancer treatment by preventing chemotherapeutic drugs that have entered the cell from being pumped out. The protein chain interacting with it on the cytoplasmic face is an antibody fragment that stabilizes P-glycoprotein, allowing crystals to form.

Transmembrane Beta-barrel Fold

We will focus on two of these proteins.

Outer Membrane Factor (OMF) - Gram-negative bacteria

Figure \(\PageIndex{25}\) shows an interactive iCn3D model of a beta-barrel transmembrane protein OPRM - Outer Membrane Factor (4y1k) from Pseudomonas aeruginosa that acts as a pore.  The transmembrane part of the protein consists of a beta-barrel fold (BBF). The protein also has a palmitoyl fatty acid in a thioester linkage to Cys 1 for extra, but unnecessary, anchorage.

Figure \(\PageIndex{25}\): Beta-barrel transmembrane protein OPRM - Outer Membrane Factor (4y1k) from Pseudomonas aeruginosa (4y1k). (Copyright; author via source). Click the image for a popup or use this external link:https://structure.ncbi.nlm.nih.gov/i...xrAuW1oFtY5FL7

Use this link for another view: https://structure.ncbi.nlm.nih.gov/i...FtY5FL7&t=4Y1K (OPM) in iCn3D

This protein is part of a large complex spanning both the inner and outer membranes of the Gram-negative bacterial cell wall (e.g., E. coli and Pseudomonas aeruginosa). Unfortunately for humans, this protein complex provides a conduit to the outside of the cell for drugs toxic to bacteria (such as antibiotics) that are pumped out of the cell by inner membrane multidrug efflux pumps. The OPRM acts as the outer passageway or duct for the pumped molecules. The Bacterial Outer Membrane Factor (OMF) protein differs in sequence, but they all form the beta-barrel duct. The E. Coli version of OMF has a triacylated lipid modification of the N-terminus. The N-terminal lipid modification might be necessary for the initial attachment of the protein to a membrane before the insertion of the beta-barrel. As such, the enzymes involved in tail attachment could be targets for new antibiotics.

Voltage-dependent anion channel (VDAC) - mouse

This protein, which also has a beta-barrel fold, regulates the movement of molecules between the cytoplasm and the mitochondrial matrix across the outer mitochondrial membrane. VDAC also serves as a docking site or scaffold for assembling molecules into a complex that regulates mitochondrial function. The protein's conformation and function are regulated by changes in the transmembrane potential, which we will explore in the following sections. Hence, the protein and its function are voltage-dependent. Figure \(\PageIndex{26}\) shows an interactive iCn3D model of mouse VDAC with a beta-barrel fold formed by 19 beta-strands (3emn).

Figure \(\PageIndex{26}\): Mouse voltage-dependent anion channel (VDAC) (3emn). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...WxUqWaGPxgHH79 

Note the N-terminal alpha-helix resident inside the channel opening. This helical section moves in response to changes in membrane potential, gating open and regulating the flow of metabolites and ions across the membrane through the pore. The conductance is high at a low transmembrane potential (10 mV) as the channel is in the open state. When the potential increases to 30 mV (either + or -), conductance drops as the protein forms the closed state.

An alternative beta-barrel fold is found in the protein enzyme Triose phosphate IsoMerase (TIM), which we will see in Chapter x. That fold has eight α-helices connecting 8 parallel β-strands that form the barrel. The TIM beta-barrel fold (TBF) is found in up to 10% of protein enzymes.  The transmembrane pore proteins above superficially resemble the TIM barrels, but they have loops, not helices, connecting the β-strands, which are antiparallel rather than parallel as in TIM. They represent a different fold called the beta-barrel fold (BBF).

Membrane proteins containing the beta-barrel fold (BBF) can now be designed de novo (see Chapter 4.14:Predicting Structure from Sequence and Sequence from Structure/Function) to create nanopores of predefined sizes.  These could serve as designed sensors.  The links below lead to iCn3D models of designed proteins with TIM barrel folds (TBF).

• 8UZL: Designed Transmembrane beta-barrel- TMB10_163 (10-stranded)
• 9FDG: Solution structure of a de novo designed 12-stranded transmembrane beta-barrel in LDAO micelles

Now that you understand the structure of membrane proteins, let's explore a key function of a subset of integral membrane proteins in the next chapter section: the movement of molecules/ions across the membrane.

Summary

(Summary written by Claude, Sonnet 4.6, Anthropic)

This chapter examines the structural diversity of membrane-associated proteins, which, together with the lipid bilayer, constitute the functional membrane. Membrane proteins are classified by the nature and extent of their association with the bilayer into three broad categories: peripheral proteins, lipid-anchored proteins, and integral (transmembrane) proteins.

Peripheral proteins associate with the membrane surface primarily through electrostatic interactions with charged lipid head groups and may partially penetrate the bilayer. Two illustrative examples highlight the functional importance of membrane binding. Matrix metalloproteinase 12 (MMP-12) is a soluble protease secreted by macrophages that binds the bilayer surface through both α-helical and β-strand regions, gaining catalytic activity only upon membrane association — a striking example of how dimensionality reduction from three-dimensional diffusion to a two-dimensional membrane surface can activate enzymatic function. Prothrombin (the zymogen precursor of the clotting protease thrombin) binds negatively charged membranes through its GLA domain, which contains multiple γ-carboxylated glutamate residues installed by a Vitamin K-dependent enzyme. Calcium ions coordinate these residues into an ordered structure that exposes a hydrophobic patch — Phe 5, Leu 6, and Val 9 — allowing partial membrane penetration (monotopic association) and positioning the zymogen for activation within the prothrombinase complex.

Lipid-anchored proteins are soluble proteins rendered membrane-associated through covalent attachment of a lipid moiety. N-myristoylation attaches the 14-carbon saturated fatty acid myristic acid via an amide bond to an N-terminal glycine, often co-translationally, and is critical for recruiting signaling proteins to membranes. S-palmitoylation attaches palmitic acid (16:0) to cysteine via a thioester bond, a reversible modification that can target proteins to cholesterol-enriched raft domains. Isoprenylation — the addition of farnesyl (C15) or geranylgeranyl (C20) chains to a C-terminal CAAX motif — anchors small GTPases such as Ras to the inner leaflet; because oncogenic Ras mutations that impair GTP hydrolysis require membrane localization for their transforming activity, prenyltransferases have been explored as anticancer drug targets. GPI anchors attach proteins to the outer leaflet through a conserved glycan core (three mannoses and glucosamine) linked to phosphatidylinositol, holding proteins well above the bilayer surface and targeting them preferentially to cholesterol- and sphingolipid-enriched raft domains. The complement regulatory protein CD59 illustrates this arrangement, and GPI-like structures from Plasmodium falciparum have been identified as malaria toxins whose immunization can protect against disease.

Integral membrane proteins are further organized by the secondary structure used to cross the bilayer and the number of membrane-spanning segments. Biotopic (single-pass) proteins cross once via a single alpha-helix with outward-facing nonpolar side chains. Cadherins use this architecture to connect extracellular calcium-dependent adhesion (through five tandem extracellular cadherin domains) to the intracellular cytoskeleton via β-catenin, coupling tissue organization to intracellular signaling. Receptor tyrosine kinases such as the insulin receptor are also single-pass proteins; ligand binding drives extracellular dimerization, activating the intracellular kinase domains through trans-autophosphorylation. Polytopic alpha-helical proteins cross the membrane multiple times. G protein-coupled receptors (GPCRs), including the cannabinoid receptor, traverse the membrane seven times in a serpentine arrangement; hydrophobic ligands such as Δ9-THC bind within the transmembrane bundle and propagate conformational changes to the cytoplasmic surface, where they engage heterotrimeric G proteins. The P-glycoprotein ABC transporter illustrates another polytopic helical protein, acting as an ATP-driven efflux pump that expels chemotherapeutic drugs and confers multidrug resistance in cancer cells. Beta-barrel transmembrane proteins, found predominantly in the outer membranes of Gram-negative bacteria and mitochondria, form pores from antiparallel β-strands connected by loops. The bacterial outer membrane factor (OMF) assembles into a trimeric beta-barrel duct that exports antibiotics as part of a tripartite efflux complex spanning both bacterial membranes, contributing to antibiotic resistance. The mitochondrial voltage-dependent anion channel (VDAC) forms a 19-stranded beta-barrel in the outer mitochondrial membrane; an internal N-terminal helix gates the channel in a voltage-dependent manner, regulating metabolite exchange between the cytoplasm and mitochondrial intermembrane space.

The chapter closes by addressing the practical challenge of studying membrane proteins in vitro. Because integral membrane proteins require a hydrophobic environment to maintain their native fold and function, a range of membrane-mimetic systems has been developed for structural studies, particularly NMR spectroscopy. Detergent micelles are the simplest but least biomimetic system. Bicelles — discoidal structures composed of a central phospholipid bilayer with short-chain lipid or detergent filling the exposed hydrophobic edges — provide a more bilayer-like environment. Nanodiscs replace the detergent rim with an amphipathic membrane scaffold protein (MSP) derived from HDL, encapsulating a small disc of phospholipid bilayer. SMALPs (styrene-maleic acid lipopolymer particles) use an amphiphilic synthetic copolymer to wrap a bilayer patch directly from a native membrane. Each system represents a trade-off among native-like lipid environment, protein solubility, and compatibility with downstream structural methods, and each has enabled structural insights into membrane proteins that were previously inaccessible.