Rice University

Department of Biochemistry and Cell Biology




Kevin R. MacKenzie

Laboratory Members

Publications

Research

Address

NMR of membrane proteins

General considerations: solids or liquids?

NMR spectroscopy can provide information about the structure and dynamics of membrane proteins that are dissolved in detergent micelles or reconstituted into lipid bilayers.  Detergent-solubilized membrane proteins can tumble freely on a nanosecond timescale, allowing the application of solution NMR methods to the collection and assignment of NMR spectra.  However, the detergent environment may alter the conformation of the protein, so sample conditions need to be correlated with other measures of stability or activity.  A lipid bilayer provides the protein with a physical and chemical environment most similar to a natural membrane, but the large size and slow tumbling of lipid bilayer vesicles require the use of solid-state NMR methods.  While the upper limit on the size of the macromolecular complexes to which solution NMR methods can presently be applied has been pushed well past 50 kD, solid state methods for structure determination of membrane proteins are more technically challenging and are still being developed.  The following reviews provide perspectives on the state-of-the-art in these endeavors.

"Solid-state nuclear magnetic resonance investigation of protein and polypeptide structure"
R Fu and TA Cross Annu Rev Biophys Biomol Struct (1999) 28, 235-68

"NMR structural studies of membrane proteins"
FM Marassi and SJ Opella Curr Opin Struct Biol (1998), 8(5), 640-8

"NMR and membrane proteins"
SJ Opella Nat Struct Biol (1997), 4(5) Suppl, 845-8



Solution NMR of a-helical membrane proteins

Standard triple- or quadruple-resonance solution NMR methods enable the assignment of resonances, the determination of backbone and sidechain geometries using J couplings, and the measurement of inter-atomic distances using the nuclear Overhauser effect (NOE).  These methods have been applied to several membrane protein systems, including the TM domain of glycophorin A, which forms a detergent-resistant parallel dimer of helices.  While chemical shift dispersion of both backbone and sidechain resonances is poor and linewidths are generally broad, resulting in peak overlap, complexes containing at least four transmembrane helices should be analyzable by solution NMR if favorable detergent conditions can be identified. 

The process of determining the structure of a complex of transmembrane helices might involve the following steps:

  •  expression of the sequence of interest in E. coli
  •  purification of 15N labeled protein in milligram quantities
  •  reconstitution into detergent micelles
  •  collection of 2D 1H-15N correlation spectrum
  •  optimization of detergent environment
  •  collection of 15N separated 3D NOESY-HSQC spectra
  •  purification of 15N,13C labeled protein in milligram quantities
  •  collection of 3D triple resonance HNCA, HNCACB (etc.) spectra
  •  assignment of backbone resonances
  •  collection of 3D 13C TOCSY spectra
  •  assignment of sidechain resonances
  •  collection of 13C separated 3D NOESY-HSQC spectra
  •  collection of J coupling data
  •  structure calculations using NOEs and J couplings


    Sample preparation

    Biophysical investigations of macromolecules usually rely on the availability of a highly purified sample.  The strongly hydrophobic nature of transmembrane domains of helical proteins necessitates the use of detergents or organic solvents; reverse-phase HPLC in organic solvents is a particularly useful purification method.  The final samples will preferably be studied in detergent micelles.  The choice of detergent and the path by which the protein is transferred from organic solvents into detergent micelles can greatly affect the structure and aggregation state of the TM domains, and hence their suitability for NMR structure determination. 

    Labeling the protein with stable isotopes 15N, 13C and/or 2H can greatly improve the quality of the NMR data that can be obtained.  This is usually accomplished by expression in bacteria grown in labeled media, since chemical synthesis of uniformly isotopically labeled peptides is extremely expensive.  However, chemical synthesis can provide a convenient route to the production of peptides with labels incorporated at discrete sites.