Heterologous gene expression in a membrane-protein-specific system

Tuning microbial hosts for membrane protein production

The last four years have brought exciting progress in membrane protein research. Finally those many efforts that have been put into expression of eukaryotic membrane proteins are coming to fruition and enable to solve an ever-growing number of high resolution structures. In the past, many skilful optimization steps were required to achieve sufficient expression of functional membrane proteins. Optimization was performed individually for every membrane protein, but provided insight about commonly encountered bottlenecks and, more importantly, general guidelines how to alleviate cellular limitations during microbial membrane protein expression. Lately, system-wide analyses are emerging as powerful means to decipher cellular bottlenecks during heterologous protein production and their use in microbial membrane protein expression has grown in popularity during the past months.

This review covers the most prominent solutions and pitfalls in expression of eukaryotic membrane proteins using microbial hosts (prokaryotes, yeasts), highlights skilful applications of our basic understanding to improve membrane protein production. Omics technologies provide new concepts to engineer microbial hosts for membrane protein production.

The bacteriorhodopsin carboxyl-terminus contributes to proton recruitment and protein stability

We examined functional and structural roles for the bacteriorhodopsin (bR) carboxyl-terminus. The extramembranous and intracellular carboxyl-terminus was deleted by insertion of premature translation stop codons. Deletion of the carboxyl-terminus had no effect on purple membrane (PM) lattice dimensions, sheet size, or the electrogenic environment of the ground-state chromophore. Removal of the distal half of the carboxyl-terminus had no effect on light-activated proton pumping, however, truncation of the entire carboxyl-terminus accelerated the rates of M-state decay and proton uptake approximately 3.7-fold and severely distorted the kinetics of proton uptake. Differential scanning calorimetry (DSC) and SDS denaturation demonstrated that removal of the carboxyl-terminus decreased protein stability. The DSC melting temperature was lowered by 6 degrees C and the calorimetric enthalpy reduced by 50% following removal of the carboxyl-terminus. Over the time range of milliseconds to hours at least 3 phases were required to describe the SDS denaturation kinetics for each bR construction. The fastest phases were indistinguishable for all bR's, and reflected PM solubilization. At pH 7.4, 20 degrees C, and in 0.3% SDS (w/v) the half-times of bR denaturation were 19.2 min for the wild-type, 12.0 min for the half-truncation and 3.6 min for the full-truncation. Taken together the results of this study suggest that the bR ground state exhibits two "domains" of stability: (1) a core chromophore binding pocket domain that is insensitive to carboxyl-terminal interactions and (2) the surrounding helical bundle whose contributions to protein stability and proton pumping are influenced by long-range interactions with the extramembranous carboxyl-terminus.

Structural genomics and drug discovery

Structure determination has already proven useful for lead optimization and direct drug design. The number of high-resolution structures available in public databases today exceeds 30,000 and will definitely aid in structure-based drug design. Structural genomics approaches covering whole genomes, topologically similar proteins or gene families are great assets for further progress in the development of new drugs. However, membrane proteins representing 70% of current drug targets are poorly characterized structurally. The problems have been related to difficulties in obtaining large amount of recombinant membrane proteins as well as their purification and structure determination. Structural genomics has proven successful in developing new methods in areas from expression to structure determination by studying a large number of target proteins in parallel.

Structural Genomics

Drug discovery based on structural knowledge has proven useful as several structure-based medicines are already on the market. Structural genomics aims at studying a large number of gene products including whole genomes, topologically similar proteins, protein families and protein subtypes in parallel. Particularly, therapeutically relevant targets have been selected for structural genomics initiatives. In this context, integral membrane proteins, which represent 60–70% of the current drug targets, have been of major interest. Paradoxically, membrane proteins present the last frontier to conquer in structural biology as some 100 high resolution structures among the 30,000 entries in public structural databases are available. The modest success rate on membrane proteins relates to the difficulties in their expression, purification and crystallography. To facilitate technology development large networks providing expertise in molecular biology, protein biochemistry and structural biology have been established. The privately funded MePNet program has studied 100 G protein-coupled receptors, which resulted in high level expression of a large number of receptors at structural biology compatible levels. Currently, selected GPCRs have been purified and subjected to crystallization attempts

In vivo enzyme immobilization by use of engineered polyhydroxyalkanoate synthase

This study demonstrated that engineered polyhydroxyalkanoate (PHA) synthases can be employed as molecular tools to covalently immobilize enzymes at the PHA granule surface. The beta-galactosidase was fused to the N terminus of the class II PHA synthase from Pseudomonas aeruginosa. The open reading frame was confirmed to encode the complete fusion protein by T7 promoter-dependent overexpression. Restoration of PHA biosynthesis in the PHA-negative mutant of P. aeruginosa PAO1 showed a PHA synthase function of the fusion protein. PHA granules were isolated and showed beta-galactosidase activity. PHA granule attached proteins were analyzed and confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and matrix-assisted laser desorption ionization-time of flight mass spectrometry. Surprisingly, the beta-galactosidase-PHA synthase fusion protein was detectable at a high copy number at the PHA granule, compared with PHA synthase alone, which was barely detectable at PHA granules. Localization of the beta-galactosidase at the PHA granule surface was confirmed by enzyme-linked immunosorbent assay using anti-beta-galactosidase antibodies. Treatment of these beta-galactosidase-PHA granules with urea suggested a covalent binding of the beta-galactosidase-PHA synthase to the PHA granule. The immobilized beta-galactosidase was enzymologically characterized, suggesting a Michaelis-Menten reaction kinetics. A Km of 630 microM and a Vmax of 17.6 nmol/min for orthonitrophenyl-beta-D-galactopyranoside as a substrate was obtained. The immobilized beta-galactosidase was stable for at least several months under various storage conditions. This study demonstrated that protein engineering of PHA synthase enables the manufacture of PHA granules with covalently attached enzymes, suggesting an application in recycling of biocatalysts, such as in fine-chemical production.

G-protein-coupled receptor domain overexpression in Halobacterium salinarum: long-range transmembrane interactions in heptahelical membrane proteins

The aminergic alpha(2b)-adrenergic receptor (alpha(2b)-AR) third intracellular loop (alpha(2b)-AR 3i) mediates receptor subcellular compartmentalization and signal transduction processes via ligand-dependent interaction with G(i)- and G(o)- proteins. To understand the structural origins of these processes we engineered several lengths of alpha(2b)-AR 3i into the third intracellular loop of the proton pump bacteriorhodopsin (bR) and produced the fusion proteins in quantities suitable for physical studies. The fusion proteins were expressed in the Archaeon Halobacterium salinarum and purified. A highly expressed fusion protein was crystallized from bicelles and diffracted to low resolution on an in-house diffractometer. The bR-alpha(2b)-AR 3i(203-292) protein possessed a photocycle slightly perturbed from that of the wild-type bR. The first half of the fusion protein photocycle, correlated with proton release, is accelerated by a factor of 3, whereas the second half, correlated with proton uptake, is slightly slower than wild-type bR. In addition, there is a large decrease in the pK(a), (from 9.6 to 8.3) of the terminal proton release group in the unphotolyzed state of bR-alpha(2b)-AR 3i as deduced from the pH-dependence of the M-formation. Perturbation of a cytoplasmic loop has thus resulted in the perturbation of proton release at the extracellular surface. The current work indicates that long-range and highly coupled intramolecular interactions exist that are capable of "transducing" structural perturbations (e.g., signals) across the cellular membrane. This gene fusion approach may have general applicability for physical studies of G-protein-coupled receptor domains in the context of the bR structural scaffold.

Aminoacylation of an unusual tRNA(Cys) from an extreme halophile

The extreme halophile Halobacterium species NRC-1 overcomes external near-saturating salt concentrations by accumulating intracellular salts comparable to those of the medium. This raises the fundamental question of how halophiles can maintain the specificity of protein-nucleic acid interactions that are particularly sensitive to high salts in mesophiles. Here we address the specificity of the essential aminoacylation reaction of the halophile, by focusing on molecular recognition of tRNA(Cys) by the cognate cysteinyl-tRNA synthetase. Despite the high salt environments of the aminoacylation reaction, and despite an unusual structure of the tRNA with an exceptionally large dihydrouridine loop, we show that aminoacylation of the tRNA proceeds with a catalytic efficiency similar to that of its mesophilic counterparts. This is manifested by an essentially identical K(m) for tRNA to those of the mesophiles, and by recognition of the same nucleotide determinants that are conserved in evolution. Interestingly, aminoacylation of the halophile tRNA(Cys) is more closely related to that of bacteria than eukarya by placing a strong emphasis on features of the tRNA tertiary core. This suggests an adaptation to the highly negatively charged tRNA sugar-phosphate backbone groups that are the key elements of the tertiary core.

Downstream coding region determinants of bacterio-opsin, muscarinic acetylcholine receptor and adrenergic receptor expression in Halobacterium salinarum

The aim of this work is to develop a prokaryotic system capable of expressing membrane-bound receptors in quantities suitable for biochemical and biophysical studies. Our strategy exploits the endogenous high-level expression of the membrane protein bacteriorhodopsin (BR) in the Archaeon Halobacterium salinarum. We attempted to express the human muscarinic acetylcholine (M(1)) and adrenergic (a2b) receptors by fusing the coding region of the m1 and a2b genes to nucleotide sequences known to direct bacterio-opsin (bop) gene transcription. The fusions included downstream modifications to produce non-native carboxyl-terminal amino acids useful for protein identification and purification. bop mRNA and BR accumulation were found to be tightly coupled and the carboxyl-terminal coding region modifications perturbed both. m1 and a2b mRNA levels were low, and accumulation was sensitive to both the extent of the bop gene fusion and the specific carboxyl-terminal coding sequence modifications included. Functional a2b adrenergic receptor expression was observed to be dependent on the downstream coding region. This work demonstrates that a critical determinant of expression resides in the downstream coding region of the wild-type bop gene and manipulation of the downstream coding region of heterologous genes may affect their potential for expression in H. salinarum.

Specific lipid requirements of membrane proteins--a putative bottleneck in heterologous expression

Membrane proteins are mostly protein-lipid complexes. For more than 30 examples of membrane proteins from prokaryotes, yeast, plant and mammals, the importance of phospholipids and sterols for optimal activity is documented. All crystallized membrane protein complexes show defined lipid-protein contacts. In addition, lipid requirements may also be transitory and necessary only for correct folding and intercellular transport. With respect to specific lipid requirements of membrane proteins, the phospholipid and glycolipid as well as the sterol content of the host cell chosen for heterologous expression should be carefully considered. The lipid composition of bacteria, archaea, yeasts, insects,Xenopus oocytes, and typical plant and mammalian cells are given in this review. A few examples of heterologous expression of membrane proteins, where problems of specific lipid requirements have been noticed or should be thought of, have been chosen.

Semliki Forest virus vectors for rapid and high-level expression of integral membrane proteins

Semliki Forest virus (SFV) vectors have been applied for the expression of recombinant integral membrane proteins in a wide range of mammalian host cells. More than 50 G protein-coupled receptors (GPCRs), several ion channels and other types of transmembrane or membrane-associated proteins have been expressed at high levels. The establishment of large-scale SFV technology has facilitated the production of large quantities of recombinant receptors, which have then been subjected to drug screening programs and structure-function studies on purified receptors. The recent Membrane Protein Network (MePNet) structural genomics initiative, where 100 GPCRs are overexpressed from SFV vectors, will further provide new methods and technologies for expression, solubilization, purification and crystallization of GPCRs.

High-throughput screening of bacteriorhodopsin mutants in whole cell pastes

A high-throughput screening method has been developed which enables functional analysis of bacteriorhodpsin in whole cell pastes. Reflectance spectra, from as little as 5 ml of Halobacterium salinarum cells, show close correspondence to that obtained from the purified purple membrane (PM), containing bacteriorhodopsin (BR) as the sole protein component. We demonstrate accurate quantification of BR accumulation by ratiometric analysis of BR (A(max) 568) and a membrane-bound cytochrome (A(max) 410). In addition, ground-state light- and dark-adapted (LA and DA, respectively) spectral differences were determined with high accuracy and precision. Using cells expressing the BR mutant D85N, we monitored transitions between intermediate-state homologues of the reprotonation phase of the light-activated proton pumping mechanism. We demonstrate that phenotypes of three mutants (D85N/T170C, D85N/D96N, and D85N/R82Q) previously characterized for their effect on photocycle transitions are reproduced in the whole cell samples. D85N/T170C stabilizes accumulation of the N state while D85N/D96N accumulates no N state. D85N/R82Q was found to have perturbed the pK(a) of M accumulation. These studies illustrate the correspondence between pH-dependent ground-state transitions accessed by D85N and the transitions accessed by the wild-type protein following photoexcitation. We demonstrate that whole cell reflectance spectroscopy can be used to efficiently characterize the large numbers of mutants generated by engineering strategies that exploit saturation mutagenesis.

Subdomains in the F and G helices of bacteriorhodopsin regulate the conformational transitions of the reprotonation mechanism

We have performed cysteine scanning mutagenesis of the bacteriorhodopsin mutant D85N to explore the role of individual amino acids in the conformational transitions of the reprotonation mechanism. We have used whole-cell reflectance spectroscopy to evaluate the spectral properties of the 59 mutants generated during a scan of the entire F and G helices and the intervening loop region. Cys mutants were grouped into one of six phenotypes based on the spectral changes associated with their M <--> N <--> O intermediate-state transitions. Mutations that produced similar phenotypes were found to cluster in discrete molecular domains and indicate that M, N, and O possess distinct structures and that unique molecular interactions regulate the transitions between them. The distribution of these domains suggests that 1) the extramembranous loop region is involved in the stabilization of the N and M intermediates, 2) lipid-protein interactions play a key role in the accumulation of N, and 3) the amino acid side-chain interactions in the extracellular portion of the interface between helices G and A participate in the accumulation of M.

Strategies for prokaryotic expression of eukaryotic membrane proteins

High-level heterologous expression of integral membrane proteins at full-length is a useful tool for their structural and functional characterization. Here, systems that have previously been used for efficient bacterial expression of eukaryotic membrane proteins are reviewed and novel vectors consisting of a modular fusion moiety based on nuclease A from Staphylococcus aureus are presented.