8.1 Solubilising a protein that was not soluble #
The study of Solubilising a protein that was not soluble provides fundamental principles governing cellular architecture, molecular interactions, and physiological regulation.
Protein Ribbon (Nih Bioart 727 791019)
8.1.1 Detergent and delipidation #
Lipid covers the hydrophobic belt. Remove the bilayer and the protein aggregates. Detergent replaces the belt with a micelle above the critical micelle concentration. Nanodiscs, amphipols, and a lipid cubic phase keep or substitute a lipid environment in different ways. No concentrations or screens are given here. Detergent is not lipid. It can strip annular lipids and structural lipids that a gate or a site required. A map of the detergent particle is a map of that particle; activity in a defined bilayer is what says the job survives. Cryo-EM changes the menu and does not abolish the problem. The early archive under-represented channels and receptors because they would not crystallise, not because they were unimportant.
8.2 Two folds #
8.2.1 Transmembrane helical bundles #
Plasma membranes and bacterial inner membranes are mostly Ξ±-helical bundles. A hydrophobic thickness of about 30 Γ is about twenty to twenty-five residues of helix. Hydropathy plots look for those spans; the algorithms are Bioinformatics, and they miss tilted helices, breaks, and re-entrant loops that enter and leave the same side. Such loops line the pores of potassium channels and aquaporins. The positive-inside rule enriches lysine and arginine in cytoplasmic loops of bacterial inner-membrane proteins. It checks topology. It does not find a binding site.
8.2.2 Ξ² barrels #
Outer membranes of Gram-negative bacteria, mitochondria, and chloroplasts contain Ξ² barrels. A transmembrane strand is short, often about ten residues, with side chains alternating between lipid and pore. Hydrogen bonds between the edge strands close the barrel and satisfy backbone polar groups. Bacterial barrels are commonly even-stranded; mitochondrial VDAC, with nineteen strands, shows that “always even” is false. Porins combine a stable scaffold with a constriction that sets selectivity. BAM and SAM insert them; those pathways are Microbiology and Cell Biology. A helix hydropathy plot will not find this sequence pattern, and this fold belongs to a different membrane.
8.3 A potassium channel #
8.3.1 The selectivity filter #
KcsA, solved in MacKinnon’s laboratory, is the worked example: four subunits, two transmembrane helices each, plus a pore helix. The filter is the signature TVGYG. Potassium is dehydrated. Backbone carbonyls, and a threonine hydroxyl at one end, replace the waters; the signature side chains mostly brace those carbonyls from behind. Four sites lie in a row. In the conducting picture two ions occupy the filter at once, in alternating configurations, and repulsion helps the queue. Rates reach about 10^8 ions per second, with selectivity over sodium often greater than a thousand-fold. Whether water always separates the ions is still argued. The carbonyl cage is not. At very low potassium the filter changes shape. The conducting geometry is organised around the ion. It is not an empty sieve.
8.3.2 Why potassium and not sodium #
Sodium is smaller, about 0.95 Γ against about 1.33 Γ for potassium on the Pauling scale. A narrow hole would exclude the larger ion. The channel does the opposite. Dehydration is costly, and the smaller ion holds water more tightly. The braced cage pays that cost back only at the potassium distance. Sodium would need shorter contacts, which this rigid geometry does not provide, so bare sodium is not compensated for the water it lost. Selectivity is that balance, not size exclusion. A mutation can ruin it by weakening the brace without deleting a carbonyl. Gating is separate. Below the filter, the inner helices cross and open at a glycine hinge. During one form of inactivation the filter itself pinches shut, so the separation of gate and filter is a principle with an exception. An open filter in a picture is not an open probability.
8.3.3 The general lesson #
Bacteriorhodopsin and a class A G-protein-coupled receptor both have seven helices and are not the same machine. Bacteriorhodopsin, from Henderson’s electron crystallography, pumps protons with a retinal Schiff base. It is not a homologue of the receptor. Activation of the receptor moves the cytoplasmic end of helix 6 outward to admit a G protein, the change from the work of Kobilka and Lefkowitz. KcsA conducts K+ through a carbonyl cage while a lower gate opens. Counting helices predicted none of this. A fold class says which physical problem was solved, such as unsatisfied hydrogen bonds in a bilayer. The chemical site, and the change between functional states, say what the protein does. These notes therefore stop at one channel.
8.4 Key terms, summary, and self-check #
Key terms. Detergent; delipidation; transmembrane helix; positive-inside rule; re-entrant loop; Ξ² barrel; selectivity filter; TVGYG; gating.
Summary. Detergent or a lipid disc replaces the bilayer imperfectly, and activity has to be shown, not assumed from a map. Helical bundles occupy plasma and inner membranes; Ξ² barrels occupy bacterial and mitochondrial outer membranes. KcsA selects potassium because a braced carbonyl cage pays for dehydration at the potassium distance and not at the sodium distance, although sodium is smaller. The bundle crossing gates. Seven helices are not a mechanism.
Self-check.
1. Why can a membrane protein be inactive in a detergent micelle even when the map looks intact?
2. How does the sequence pattern of a transmembrane helix differ from that of a barrel strand?
3. Why is “the hole is too small for sodium” the wrong account?
4. What braces the filter, and which part of KcsA gates rather than selects?
5. Why do seven helices not make bacteriorhodopsin and a G-protein-coupled receptor the same machine?
—
—
—