Biotechnol Lett (2006) 28:767–778 DOI 10.1007/s10529-006-9005-y
REVIEW PAPER
New tools for G-protein coupled receptor (GPCR) drug discovery: combination of baculoviral expression system and solid state NMR Venkata R. P. Ratnala
Received: 4 January 2006 / Accepted: 14 February 2006 / Published online: 23 May 2006 Springer Science+Business Media B.V. 2006
Abstract Biotechnology using molecular biology, biochemistry, biophysics, and computational approaches provides an alternative approach for classical pharmacological screening to look at ligand-receptor interactions and receptor specificity, which should support the design of selective drugs based on detailed structural principles. This review addresses specific approaches to study function, structure and relevance of a major pharmaceutical target, namely the G-Protein Coupled Receptors (GPCRs). The main aim of this review has been to exploit and combine GPCR over-expression in a baculoviral expression system with solid-state MAS NMR (ssNMR) approaches for the elucidation of electronic structures of the coordinating ligands/drugs and their modes of interactions with the GPCRs. This review summarizes the approaches, possible future experiments and developments using the above combination of tools for GPCR drug discovery.
V. R. P. Ratnala (&) Leiden Institute of Chemistry, Leiden University, Einsteinweg 55, P.O. Box 9502, 2300 RALeiden, The Netherlands e-mail:
[email protected] V. R. P. Ratnala Stanford University, 157 Beckman Center, 279 Campus Drive, Stanford, CA, USA
Keywords Drug Discovery Æ Functional reconstitution Æ GPCRs Æ Histamine H1 receptors Æ Ligand binding Æ Magic angle spinning overexpression Æ Solid state NMR
Introduction The power of utilizing genetic engineering to produce recombinant proteins for a variety of purposes has revolutionized many areas in biotechnology. The GPCRs currently represent one of the most important classes of drug-targets in the pharmaceutical industry. It is therefore apparent that a determined approach is required to study this class of membrane proteins. Approaches need not be based on any single technique but on a range of low-resolution methods and high-resolution techniques such as ssNMR, fluorescence spectroscopy etc. On the one hand the advances in molecular biology and biophysics for structure–function studies are fast evolving. These experimental approaches need to be supplemented and aided by computational and bioinformatics approaches. Bioinformatics and computational approaches, on the other hand, are very promising and are rapidly developing in structure prediction. These predicted structural models need experimental verification by biophysical approaches, in particular by ssNMR approach. This review describes the potential of
123
768
Biotechnol Lett (2006) 28:767–778
baculovirus expression system for over-expression of the GPCRs and ssNMR methodologies to study ligand receptor interactions of these important classes of drug-targets.
Receptor over-expression for structural studies Structural studies require reasonably large amounts of protein, often more than milligram quantities of a receptor in a fully functional state. For this, efficient expression systems are essential for studies on ligand-receptor interactions or for structure determinations by heteronuclear ssNMR. This is not a trivial problem for membrane proteins because their over-expression often causes lysis and cell death. Most ssNMR studies of membrane proteins to date have been carried out with proteins that were expressed in Escherichia coli, recovered from inclusion bodies and subsequently refolded. In some cases, signal sequences were deleted and purification tags (histidine tags, glutathione S-transferase, staphylococcal nuclease, maltose-binding protein, and so on, with appropriate proteolytic cleavage sites) were engineered into the expression vectors. Purification can occur in detergent or, in the case of b-barrel membrane proteins, in denaturants such as urea or guanidinium chloride. Refolding conditions have to be carefully monitored for each membrane protein. In many cases, the lack of an appropriate refolding protocol has become a major obstacle in the structure determination of membrane proteins. One of the aims of this review is to present a framework for efficient expression and purification of GPCRs in their
native form by the use of a baculovirus expression system avoiding the often-cumbersome refolding steps when using E. coli. Using such an approach, it is possible to produce GPCRs in large scale for structural studies (Ratnala et al. 2004). Heterologous expression with the baculovirus– insect cell system typically ensures production of good protein quantities in cells performing posttranslational modifications essentially identical to those occurring in naturally producing tissues. This opens the way for high-level expression of recombinant GPCRs with characteristics almost identical to their native counterparts, making them ideal for subsequent structural studies. The functional receptor have been successfully expressed with production levels of 30–60 pmole/ 106 cells or 18–35 · 106 copies/cell, which corresponds to the final yields of 5–7 mg of functional receptor per liter (Table 1) (Ratnala et al. 2004).
Improving receptor expression and purification for large scale productions Proteolysis is a common problem during the production of a recombinant protein in a heterologous expression system. During the expression of human histamine H1 receptor, limited fragmentation of the receptor was observed due to proteolysis, without the loss of ligand binding properties. However proteolysis could be efficiently suppressed by including the protease inhibitor leupeptin during cell culture and all subsequent manipulations (Ratnala et al. 2003a, 2004). Several attempts have been made to improve GPCR expression with the baculovirus
Table 1 Comparative table showing the functional expression levels and the ease of scaling up of GPCRs using different expression systems Expression system
Total expression levels (mg/l)
Functional percentage (%)
Scale-up
Insect/Baculovirus (Ratnala et al. 2004) E. coli Yeast (Pichia pastoris) Insect/Baculovirus Mammalian/Semliki forest virus
10–15 0.1–10 0.1–5 2–7 0.5–10
90 <5 < 10 20–90 10–80
+++ +++ ++ +++ +
Scaling up is shown by the ‘‘+’’ symbol, ease of scaling up is shown by increase in the number of signs (Grisshammer and Tate 1995; Lundstrom 2003; Reilander and Weiss 1998).
123
Biotechnol Lett (2006) 28:767–778
system. Membrane proteins, like secreted soluble polypeptides, travel through the cell secretory pathway to reach the cell outer membrane. Hence, improvements brought to address secretion or folding problems of secreted proteins are often of use for membrane polypeptides. Fusions to a variety of signal sequences from baculoviral, prokaryotic or eukaryotic origin successfully increased the amount of soluble secreted proteins (Allet et al. 1997; Golden et al. 1998; Mroczkowski et al. 1994; Murphy et al. 1993). The N-terminus of all GPCRs is predicted to be on the extracellular surface of the plasma membrane, so attempts were therefore made to use a similar strategy to increase functional GPCR production (Robeva et al. 1996). Engineering of the insect cells could also be undertaken to try and prevent the formation of misfolded receptor and therefore optimize GPCR expression. Co-expression of chaperones such as calreticulin or the ER membrane-bound calnexin appeared an attractive strategy that could be applied to the GPCR family (Ailor and Betenbaugh 1999). Expression of receptors is also dependent upon the culture conditions: Sf9 cells adapted to serumfree medium expressed the human l-opioid receptor fused to the gp64 signal sequence at about twice the level as in Sf9 cells grown in the presence of serum and at levels comparable to those obtained in T. ni cells adapted to serum-free medium (Massotte et al. 1999). This latter observation suggests that higher expression levels observed in T. ni cells compared to Sf9 are due to serum-free culture conditions rather than to a cell line effect (Massotte et al. 1999). High levels of expression (20 nmol/l) were also obtained for the human b2-adrenergic receptor fused to the hemagglutinin signal sequence and a C-terminal (6xHis) tag in Sf9 cells adapted to serum-free medium (Kobilka 1995) and production of bovine rhodopsin was increased in Sf9 cells adapted to serum-free medium (Klaassen et al. 1999). Efforts have also been made to facilitate downstream purification of the receptors. A hexahistidine (6xHis) tag has often been added to the protein that allows the use of an affinity step on an ion chelating resin for an efficient purification step. A systematic study on rhodopsin, human histamine receptors, human opioid receptors were examined
769
on evaluating the influence of this tag on the receptor production. The authors showed that addition of a (6xHis) tag did not interfere with ligand binding whether located at the N or C terminus of the receptor (Massotte et al. 1997, 1999; Ratnala et al. 2004). The cDNA encoding a C-terminally histidine-tagged (10xHis) human histamine H1 receptor could achieve expression at high levels (30–40 pmol/106 cells), which is comparable with other GPCRs expressed with (6xHis) tag expressed under identical conditions in Sf9 cells. Ligand affinities of the Sf9 cell expressed his-tagged H1 receptor were also similar to those of the untagged receptor expressed in COS-7 cells. So it could be generally concluded that there was a not major effect in expression with a 10xHis tag, which assists in enhancing the purification strategies (Ratnala et al. 2004). Clearly, other GPCRs could show a different effect and the tagging strategy may well have a detrimental effect. The amounts of GPCR produced per cell or liter of culture with the baculovirus system represent a great improvement compared to natural expression yields (at least several 100-fold increase), but structural studies are very demanding and large-scale cultures are likely to be required. Insect cells are semi-adherent, allowing both growth under attached conditions or in suspension cultures. Both attached and suspensiongrown cells present advantages though the latter is often favored because of its simplicity and the long-standing experience from microbial fermentation. However, the cells are also very sensitive to shear damage resulting from gas bubbling. This can be reduced by different additives such as Pluronic F-68, a tensioactive compound that prevents the cells from attaching to the gas– medium interface (Wang et al. 1993). Bioreactors allow not only large culture volumes but also control of optimized conditions for protein production. Oxygen proves to be a crucial ratelimiting component because its consumption increases drastically following viral infection. Fermentation volumes of 60 l or more are used in industry but, at the laboratory scale, bioreactors are rarely larger than 5–10 l. Over the years, numerous protocols for culture and infection in spinner flasks or bioreactors have been
123
770
established (Kioukia et al. 1995a, b; Massotte et al. 1999; Pereira et al. 2001; Rice et al. 1993; Wang et al. 1993; Yang et al. 1996). The systematic study performed on rhodopsin production in Sf9 cells showed that the optimal multiplicity of infection (MOI) was lower, when cells were infected in the early-mid-exponential phase of growth, the highest levels of recombinant rhodopsin production per culture volume were obtained with the lowest MOI value (Klaassen et al. 1999). This is of importance if one considers a large-scale production. Because the virus is lytic, attention has also to be paid when deciding for optimal MOI and very often the use of a semicontinuous system or repeated batches was preferred over continuous production. During the course of cell culture, an inevitable loss of production happens, probably due to a high proportion of defective viral particles (Agathos 1996). Despite the variety of bioreactor designs that have been studied, no single system has proven to be universally superior in productivity and ease of scale-up. A new disposable type of bioreactor (Wave bag) from Wave Biotech AG (Tagelswangen, Switzerland) was also tested (Ratnala et al. 2004). This innovative design claims better mixing and oxygen transfer and lower shear stress, and offers a broad range of culture volumes (from 2 up to 500 l). Preliminary results with the wave bag technology gave approximately 50 mg of functional receptor from a 10 l culture and this approach may hold promise as an alternative for bioreactor culture, which is quite timeconsuming in maintenance and preparation for sterilization. Post-translational modifications that are known to play important roles in GPCR function would not be predicted to occur normally in bacteria. Post-translational modifications like palmitoylation, glycosylation, methylation, acetylation and phosphorylation play a major role in obtaining functional receptors on a large scale. Overexpression in insect cells with the baculovirus system proved to be very helpful because insect cells perform post-translational modifications identical to mammalian ones with the exception of glycosylation (Hawtin et al. 2001; Horstmeyer et al. 1996; Okamoto et al. 1997; Ponimaskin et al. 2002). Most proteins present at the extracellular
123
Biotechnol Lett (2006) 28:767–778
surface of mammalian cells are glycosylated. In a study on the human histamine H1 receptor, no evidence for N-linked glycosylation was obtained (Ratnala et al. 2004). The inability of insect cells to perform complex glycosylation is apparently due to the absence or insufficient amounts of the terminal glycosyltransferases that are required to convert N-linked side chains to complex forms (Jarvis et al. 1998). Strict control of growth and infection parameters, especially dissolved oxygen also play a major role in glycosylation (Zhang et al. 2002). The baculovirus/insect cell system is seen as a promising expression system for obtaining authentic post-translational modification compared to the mammalian and prokaryotic systems. Furthermore, proteins expressed using this system have largely been found to maintain their normal characteristics, including specific post-translational modifications, such as fatty acid acylation and phosphorylation. However, it has been observed that no N-linked protein glycosylation in human histamine H1 receptor protein expressed in Sf9 cells is occurring. It is now known that Sf9 cells can synthesize and transfer oligomannosidic oligosaccharides, can trim N-glycans to trimannosyl cores and can process these by the addition of fucose; thus, the complex oligosaccharides found in vertebrate cells are replaced by small truncated side chains (Kuroda et al. 1990). Nevertheless, this change in the glycosylation profile was not reported to affect the functional properties of proteins such as GPCRs. Membrane lipid composition also influences functionality to a certain extent. Compared to mammalian cells, the Sf9 cell plasma membrane contains higher amounts of unsaturated lipids and lower amounts of cholesterol to ensure membrane fluidity at the organism growth temperature (27–28C) (Reilander et al. 1991). Moreover, baculovirus infection changes the lipid composition of the membrane through enrichment in phosphatidylcholine (Reilander et al. 1991). Several reports have pointed out the influence of membrane lipid composition on GPCR activity, showing that acyl chain saturation (Banerjee et al. 1993), negatively charged lipids and/or cholesterol (Emmerson et al. 1999; Lagane et al. 2000; Litman and Mitchell 1996) constitute parameters that sometimes influence receptor functionality.
Biotechnol Lett (2006) 28:767–778
For example, the histamine H2 (Beukers et al. 1997) receptors were expressed in a low affinity state in Sf9 cells; addition of cholesterol by incubation of the cells with cholesterol-b-cyclodextrin complexes led to an increased amount of high affinity binding sites. Thus, in some cases, overexpression increases the amount of receptors produced in a non-active conformation by reducing the overall availability of less abundant membrane components. Stable expression in Sf9 insect cells can be achieved using the baculovirus ‘‘immediate-early’’ promoters that are recognized and used by insect cell RNA polymerases without requiring any additional viral encoded proteins. During the normal viral replication cycle, these promoters are the first to be switched on following viral entry in the cell. Immediate-early promoters have been widely used to engineer cell lines with improved post-translational processing, but very few attempts have been made to express membrane proteins under their control. Though significantly lower levels of expression were obtained under the control of ‘‘immediate-early’’ promoters when compared to the very late polyhedrin promoter, stably transfected insect Sf9 cells may represent a valuable alternative for GPCR large-scale expression. They grow to high density in suspension and adapt easily to serum-free medium but they do not require the production of large-scale high-titre viral stocks and allow continuous culture in a bioreactor. Moreover, compared to mammalian cells, insect cells are considerably less expensive and time consuming to grow in large volumes and reach higher cell densities. Selection of the proper detergent, and searching for appropriate stabilizing components, played an essential role in obtaining solubilized functional H1 receptor (Ratnala et al. 2003b, 2004). Solubilization is a critical step in membrane protein purification. On the one hand, it is essential for purification, while on the other hand it destabilizes the receptor, resulting in a timedependent loss of functional properties. The kinetics of this loss of activity depends on the type of detergent and on buffer composition (DeGrip et al. 1998). Solubilization of functional GPCRs has in many cases been achieved with the very mild detergent digitonin (Banerjee et al. 1995;
771
DeGrip et al. 1998; Rinken et al. 1994; Toll and Snyder 1982). However, impurities and batch variations make this natural product unsuitable for reproducible purification and reconstitution. For the solubilization of the H1 receptor best results were obtained with dodecylmaltoside, a mild detergent that is commercially available in high purity and also provides good results with other membrane proteins (DeGrip 1982; Grisshammer and Tate 1995; Klaassen et al. 1999; Rosenbusch 2001). Solubilization of the H1 receptor in a micellar dodecylmaltose solution increased to 70–90% at higher ionic strength. This may reflect more dense detergent packing, that mimics the lipid bilayer in a better way than at low ionic strength (Rosenbusch 2001). Analogous to the situation in visual pigments (Klaassen and DeGrip 2000) it is possible to considerably stabilize the H1 receptor by adding the high-affinity inverse agonist, mepyramine (Haltia and Freire 1995; Ratnala et al. 2003a, 2004) and took advantage of this ligand stabilization further by using radiolabeled [3H]mepyramine. The low Koff of mepyramine (Wallace 1983) is convenient for protocol development. On the other hand, removal or exchange of H1 receptor bound mepyramine from the final proteoliposomal preparation is difficult at temperatures below 20C. Therefore, saturation with an alternative, more easily washable ligand was exploited for routine production. The low-affinity agonist histamine was observed to reduce the stability of the solubilized receptor, while the high affinity inverse agonist tripelennamine behaved quite similar to mepyramine. In agreement with its higher Koff, tripelennamine (Bakker et al. 2001) could be more easily washed away from the final preparation. Thus the entire procedure was performed in the presence of 2-lM tripelennamine, which is sufficient to fully saturate the receptor (Kd ~ 4.2 nM) (Bakker et al. 2001). In this way functional receptor could be easily identified and traced by scintillation counting. This dual effect both stabilizing and estimating functional receptors greatly facilitated finding optimal conditions for solubilization, purification and reconstitution. Probably the inverse agonists reduce the dynamics and flexibility of the protein, while an agonist has the opposite effect, making the receptor more
123
772
vulnerable for detergent destabilization. The presence of inverse agonist and a 20% level (w/v) of the renowned protein stabilizing agent glycerol, rendered the H1 receptor sufficiently stable in micellar solution to allow purification with good recovery (Table 1) (Ratnala et al. 2004). The procedure adapted for the reconstitution by selective detergent extraction of membrane protein into functional proteoliposomes using detergent extraction by means of inclusion complex formation with cyclodextrins worked out very well in case of human H1 receptor purification (Ratnala et al. 2004). Detergents can also be removed by binding to cyclodextrins. This procedure works out to be a single step process for achieving both the goals of removing the detergents and obtaining the functional proteoliposomes fractions. Solubilized receptors had to be reconstituted in proteoliposomes and exposure to detergents should be kept to a minimum, to avoid denaturation of the receptors, a fast procedure for the reconstitution is very essential. It is evident that, instead of a natural lipid mixture, the commercially available (Sigma) soybean extract asolectin containing a mixture of several lipids, was found to be most suitable and also produces fully functional proteoliposomes with many receptors at a molar lipid to protein ratio of 100–150. Overall, the ligand binding data unequivocally demonstrates the ability to successfully purify and reconstitute the H1 receptor in a stable form, with full preservation of functional integrity.
Solid state NMR to study membrane bound targets ssNMR methods provide a number of powerful biophysical tools for mapping structure and dynamics in biomolecules. These tools are especially promising approaches for the study of important systems that are inaccessible to crystallography or solution NMR, such as membrane proteins. Using ssNMR techniques, there are two approaches that can be used to deal with the anisotropic interactions, which lead to NMR spectral broadening and permit molecular structural information to be extracted. One is to exploit the spectral anisotropy in oriented sam-
123
Biotechnol Lett (2006) 28:767–778
ples to give molecular orientations in static samples, and the second is to use random dispersions in magic angle spinning (MAS) NMR. In the second case, both resolution and sensitivity are improved dramatically by spinning the sample at a magic angle (Fu and Cross. 1999) (Fig. 1), in which orientational information may be lost, but can be regained from analysis of spinning sidebands. Rather than exploiting sample orientation, MAS experiments average orientation-dependent interactions as an alternate approach to achieve narrow resonances and thus increase resolution and sensitivity in the spectrum (Fig. 2). In addition, dipolar couplings can be recoupled to yield distance constraints, as well as chemical shift information and quadrupolar interactions to define local environmental characteristics. Proton detection in ssNMR remains challenging due to the strong dipolar interaction, so most studies focus on 13C or 15N labeled samples. Biophysical applications of MAS experiments have expanded significantly with the development of a number of methods for reintroducing and measuring specific dipolar interactions to obtain structural information (Griffin 1998; Tycko 2001). Resolving the interactions of the ligands and its functional receptor protein has been demonstrated by the use of ssNMR methodologies. This is a powerful non-perturbing approach to gaining insights into information that is less readily obtained by other approaches. Similarly, there is an extensive range of GPCRs which can be studied using similar methods, when they become available from expression methods, such as that used for the human histamine H1 receptor now expressed using a in baculoviral expression system at amplified levels. Drug design requires both
Fig. 1 Showing the ‘‘magic’’ angle of 54.74 of the rotor with respect to the magnetic field (B0) applied
Biotechnol Lett (2006) 28:767–778
773
Improving effective technologies for sample preparation
Fig. 2 Solid-state 13C NMR spectra of tyrosine shown in ppm. The spectrum shown at the top of the figure was acquired from a stationary sample. The spectrum below was obtained with magic angle spinning (MAS) at trot = 12.5 kHz
structure and definition of binding sites and their binding characteristics, usually referred to as pharmacophores (electronic configuration, conformational distortions, etc. NMR is well placed to provide this information, as already shown for a range of soluble drug targets, but now possible using ssNMR methods for membrane bound targets (De Groot 2000). In addition to helping in resolving structural features for bound ligands, electronic environments can be defined and as shown for retinal whilst at its binding site within rhodopsin (Creemers et al. 2002). These ligand binding sites can be resolved from chemical shift perturbations recorded from ssNMR spectra for bound ligand, which indicate uniquely the important regions of a ligand required for binding and interaction with the protein binding site. Further studies by site-directed mutagenesis, sitespecific modification, epitope mapping with photoactive ligands, studies using other biophysical techniques like Fluorescence spectroscopy, FTIR spectroscopy, AFM, EPR and modeling studies could prove to be the next step for obtaining the further information on the ligand interactions to further map the binding site.
Sample preparation is generally straightforward for soluble proteins or protein domains that can be over-expressed in bacterial expression systems typically in E. coli. However, the requirement to prepare relatively large amounts of material can be a serious limitation for studies of protein containing post-translational modifications that are critical for function and in studies of membrane proteins and GPCRs. There is a clear need to devise more efficient methods to over-express proteins in eukaryotic expression systems to allow studies of proteins with the full complement of post-translational modifications as well as to study membrane proteins. Membrane proteins in lipid membranes are ideal for ssNMR studies and reconstitution technology for membrane proteins is a very active area of research in its own right. Any newly produced membrane protein or a GPCR that needs to be studied functionally, or by biophysical methods, is usually reconstituted into bilayers. The choice of detergents and methods for protein–lipid reconstitution represent a major activity in and of itself. It may be some time before general methods are available, and this aspect of membrane protein studies needs to be developed for ssNMR methods as well as for any other biophysical approach to membrane protein biology. Sensitivity for ssNMR of membrane proteins is not dissimilar to that for solution NMR. A few lmoles of peptide or protein are required, although in studies of ligand–receptor interactions, where cross polarization can enhance sensitivity, nano moles of protein binding sites are sufficient for efficient detection. For small peptides and larger proteins, this translates into the requirement to prepare a few tens of milligrams of the protein of interest. Expression methods work well for small peptides, but for larger proteins, these amounts are still only just within reach for amplified expression systems. For example, human histamine H1 receptor has been expressed at 5–7 mg of functional receptor (Ratnala et al. 2004), rhodopsin expression has been scaled up to 10 mg and protein labeling has been achieved
123
774
using a recombinant baculovirus expression system (Klaassen et al. 1999).
Approaches for probing structure and function by isotopic labeling Isotopic labeling methods have been indispensable in aiding assignments of resonances, and in obtaining the structural information that forms the basis for NMR structure determination both in solution and in the solid state. Complete proteins, ligands and specific residues have been labeled, whenever the synthesis has been possible. Uniform isotope labeling of proteins is possible when a good expression system is available, and effective biochemical methods exist to label all residues of a certain type. However, incorporation of a single or a few NMR isotopes into a polypeptide chain or ligand is generally far from achievable, especially if multiple but specific (rather than uniform) labeling is required. These requirements are expected to become ever more pressing as larger biological systems become amenable to NMR investigation, because as the number of resonances increase, even uniform or residue-selective labels are not sufficient to provide the necessary spectral dispersion. There have been a number of important technical developments in the last few years, these developments promise to provide universal tools to incorporate labels at virtually any site of choice in a soluble or membrane protein. Namely, in vitro protein synthesis allows production of large amounts of proteins with very specific labeling patterns (Schindler et al. 1999) and protein splicing methods that allow the segmental labeling of proteins at all domains or sub-domain of interest (Xu et al. 1999). With isotopic labeling of specific sites, local structure can be measured in peptides and proteins by using MAS methods to measure chemical shifts, distances, and torsion angles. Analysis of the active site can be achieved by addition of labeled substrates, inhibitors, or cofactors. The active site can thus be labeled uniquely, often with a chemically distinct species, which can therefore be resolved and assigned in the spectrum. Similar experiments have led to a model for the binding
123
Biotechnol Lett (2006) 28:767–778
of an inhibitor to H+/K+ ATPase (Watts et al. 2001). Structural measurements beyond active sites typically focus on measurements of distances between specifically labeled backbone atoms in synthetic peptides to determine the backbone conformation. For example, rotational resonance (Raleigh et al. 1988) measurements of 13C–13C distances have characterized the structure of melittin in lipid bilayers, and report that the angle between the two helical segments is larger than that observed in the crystal structure or micelles (Lam et al. 2001). Innovative labeling and/or assignment strategies are needed to probe sites beyond the active site in larger proteins produced biosynthetically. For widespread labeling, an elegant method employing the bacterial metabolic pathways and [2-13C]glycerol as the sole carbon source during growth was used by Hong and coworkers to incorporate 13C at many Ca and other carbons of 10 amino acids types throughout the pore forming toxin colicin Ia (Huster et al. 2002). Protein expression in media containing isotopically labeled amino acids typically results in multiple labeled sites. NMR methods can be used to select a specific site of interest in biosynthetically labeled proteins for residues that occur as unique dipeptides in the protein. A biochemical method for targeting the NMR experiment to a specific site of interest in an expressed protein employs site-directed mutagenesis to introduce unique residues for biosynthetic labeling (Murphy et al. 2001). This general approach is applicable to any site where a functional mutant incorporating a unique residue can be made, allowing measurements of secondary and tertiary structure at sites throughout large proteins.
Advances in ssNMR technology Extremely high quality structural and dynamic information on biomolecules and their complexes are now available through NMR studies with unprecedented detail. This has been made possible by the tremendous technical progress in all aspects of NMR spectroscopy in the last 10– 15 years. Despite these tremendous advancements, sensitivity remains one of the main
Biotechnol Lett (2006) 28:767–778
limitations of the NMR approach to structural analysis, which implies that large amounts of material are required. Limitations in solubility or the tendency of biomolecules to aggregate at high concentration, sometimes make it impossible to study systems of great biological importance. In addition to increasing magnetic field strength, reductions in thermal, electronic and radio frequency noise would contribute tremendously to an enhancement of sensitivity. Some of the major improvements that should be aggressively pursued are continued development of more sensitive probes, particularly of so called cryoprobes and receiver units capable of operating at low temperatures to improve the signal to noise ratio. Development of static and spinning probes, methodology and wide bore magnets for highfield ssNMR. Development of special probes and robotic sample handling methods (Edwards et al. 2001) to allow high throughput screening and more effective use of NMR as an analytical tool. Further increase in magnetic field strengths to allow the production of spectrometers that operate at 1 GHz and beyond. The development of instrumentation and technology must be paralleled by a development of new theoretical approaches. Understanding of the fundamental principles of the NMR techniques, and the discovery of new experimental protocols and physical phenomena will improve the quality and reliability of biological structures and dynamics and to discover entirely new information about the relationship between molecular structure and dynamics and biological function. Methods of data analysis and post acquisition data processing must also keep place with the development in hardware and in NMR methods to allow these new sources of information to be used in structure determination. On the one hand, the baculovirus–insect cell system has proven extremely effective for the production of large amounts of many GPCRs with properties comparable to those of their native counterparts. Insect cells are indeed able to perform post-translational modifications crucial for GPCR function such as palmitoylation and phosphorylation. Receptors produced in insect cells constitute, therefore, an excellent source of
775
material for structural studies. The expression yields obtained for some GPCRs are high enough to produce the milligram quantities required for ssNMR or 3D crystallization trials. However, for many GPCRs, expression levels are still too low for such ambitious goals or would at least require very large culture volumes. GPCR expression and purification, on the other hand, have been improved, but some parameters designed to optimize expression remain elusive and do not always act in a predictable manner. Nevertheless, a lot of information has been gathered on the system that will definitely help future studies and efforts are still being made to develop new approaches and optimize current procedures. Advantages of the baculovirus–insect cell system include its simplicity to set up in any laboratory, the fact that it does not require extensive biosafety measures, and that it benefits from over 15 years of experience and of well-established protocols. As a consequence, it represents a valuable tool for GPCR studies whether the ultimate goal is to gain structural information or to better understand the regulation of the first steps of the signaling cascade. On the other hand ssNMR methods provide a promising biophysical approach for many important and challenging biological systems. One of the lessons that were learnt from this 13C and 15N study was that, using CP methods and reasonable MAS rates some high quality spectra could be obtained. The application of the ssMAS NMR dipolar correlation techniques for structure refinement and structure function study of ligand binding to receptors is a valuable asset for future rational drug design studies. At the level of the chemical shift assignment structural information could be obtained. It is seen that the changes in the chemical shifts induced by binding of the agonist histamine indicated the probable residues that reside in the functional binding sites. A comparison between chemical shifts of the bound and the unbound complex was used to map out the probable binding sites with in the receptor in its active conformation. The chemical shift assignment can be assessed to identify intermolecular correlations, which provide distance constraints. Even a limited set of such constraints will substantially reduce the number of possibilities to
123
776
assemble the ligand–receptor complex in the next modeling step. Prospects are good that complete structure determination of uniformly labeled samples will be possible for membrane proteins. In combination with biochemical methods such as segmental labeling (Cowburn and Muir 2001), it should now be possible to determine structures of portions of larger proteins, such as specific helices of multi span membrane proteins. Studies of selectively labeled proteins continue to provide important insights into structure and mechanism. The combination of the methodologies ranging from molecular biology, biochemistry and biophysics described in this review set a stage for a new development towards drug screening and rational drug design strategies for drugs acting on GPCRs. Most importantly, the present methodologies should provide a working model for designing small molecule antagonists and receptor decoys as therapeutics for GPCR-related diseases. Acknowledgements I acknowledge Prof. Dr. W. J. de Grip, Prof. Dr. H. J.M. de Groot and Prof. Dr. Rob Leurs for their support during my research at Leiden University. Dr. S. Kiihne and J. Hollander are acknowledged for their technical support on ssNMR.
References Agathos SN (1996) Insect cell bioreactors. Cytotechnology 20:173–189 Ailor E, Betenbaugh MJ (1999). Modifying secretion and post-translational processing in insect cells. Curr Opin Biotechnol 10:142–145 Allet B, Bernard AR, Hochmann A, Rohrbach E, Graber P, Magnenat E, Mazzei GJ, Bernasconi L (1997). A bacterial signal peptide directs efficient secretion of eukaryotic proteins in the baculovirus expression system. Protein Expr Purif 9:61–68 Bakker RA, Schoonus SBJ, Smit MJ, Timmerman H, Leurs R (2001). Histamine H-1-receptor activation of nuclear factor-kappa B: roles for G beta gamma- and G alpha(q/11)-subunits in constitutive and agonistmediated signaling. Mol Pharmacol 60:1133–1142 Banerjee P, Dasgupta A, Chromy BA, Dawson G (1993) Differential solubilization of membrane–lipids by detergents—coenrichment of the sheep brain-serotonin 5-ht(1a) receptor with phospholipids containing predominantly saturated fatty- acids. Arch Biochem Biophys 305:68–77 Banerjee P, Joo JB, Buse JT, Dawson G (1995). Differential solubilization of lipids along with membraneproteins by different classes of detergents. Chem Phys Lipids 77:65–78
123
Biotechnol Lett (2006) 28:767–778 Beukers MW, Klaassen CHW, DeGrip WJ, Verzijl D, Timmerman H, Leurs R (1997) Heterologous expression of rat epitope-tagged histamine H-2 receptors in insect Sf9 cells. Br J Pharmacol 122:867– 874 Cowburn D, Muir TW (2001) Segmental isotopic labeling using expressed protein ligation. Nuclear Magnetic Resonance of Biological Macromolecules, Pt B 339: 41–54 Creemers AFL, Kiihne S, Bovee-Geurts PHM, DeGrip WJ, Lugtenburg J, de Groot HJM (2002) H-1 and C-13 MAS NMR evidence for pronounced ligand– protein interactions involving the ionone ring of the retinylidene chromophore in rhodopsin. Proc Natl Acad Sci USA 99:9101–9106 De Groot HJM (2000) Solid-state NMR spectroscopy applied to membrane proteins. Curr Opin Struct Biol 10:593–600 DeGrip WJ (1982) Baculovirus expression system for expression and characterization of functional recombinant visual pigments. Meth Enzymol 81:256–265 DeGrip WJ, VanOostrum J, Bovee-Geurts PHM (1998) Selective detergent-extraction from mixed detergent/ lipid/protein micelles, using cyclodextrin inclusion compounds: a novel generic approach for the preparation of proteoliposomes. Biochem J 330:667–674 Edwards SM, Kote-Jarai Z, Hamoudi R, Eeles RA (2001). An improved high throughput heteroduplex mutation detection system for screening BRCA2 mutations—Fluorescent mutation detection (F-MD). Hum Mutat 17:220–232 Emmerson PJ, Clark MJ, Medzihradsky F, Remmers AE (1999) Membrane microviscosity modulates mu-opioid receptor conformational transitions and agonist efficacy. J Neurochem 73:289–300 Fu RQ, Cross TA (1999) Solid-state nuclear magnetic resonance investigation of protein and polypeptide structure. Annu Rev Biophys Biomol Struct 28:235–268 Golden A, Austen DA, van Schravendijk MR, Sullivan BJ, Kawasaki ES, Osburne MS (1998). Effect of promoters and signal sequences on the production of secreted HIV-1 gp120 protein in the baculovirus system. Protein Expr Purif 14:8–12 Griffin RG (1998) Dipolar recoupling in MAS spectra of biological solids. Nature Struct Biol 5:508–512 Grisshammer R, Tate CG (1995) Over-expression of integral membrane-proteins for structural studies. Q Rev Biophys 28:315–422 Haltia T, Freire E (1995) Forces and factors that contribute to the structural stability of membrane-proteins. Biochim Biophys Acta Bioenerg 1228:1–27 Hawtin SR, Tobin AB, Patel S, Wheatley M (2001) Palmitoylation of the vasopressin V-1a receptor reveals different conformational requirements for signaling, agonist-induced receptor phosphorylation, and sequestration. J Biol Chem 276:38139–38146 Horstmeyer A, Cramer H, Sauer T, MullerEsterl W, Schroeder C (1996) Palmitoylation of endothelin receptor A—Differential modulation of signal transduction activity by post-translational modification. J Biol Chem 271:20811–20819
Biotechnol Lett (2006) 28:767–778 Huster D, Yao YL, Jakes K, Hong M (2002) Conformational changes of colicin Ia channel-forming domain upon membrane binding: a solid-state NMR study. Biochim Biophys Acta Biomemb 1561:159– 170 Jarvis DL, Kawar ZS, Hollister JR (1998) Engineering Nglycosylation pathways in the baculovirus–insect cell system. Curr Opin Biotechnol 9:528–533 Kioukia N, Alrubeai M, Zhang Z, Emery AN, Nienow AW, Thomas CR (1995a) A study of uninfected and baculovirus-infected spodoptera–frugiperda cells in T-flasks and Spinner-flasks. Biotechnol Lett 17:7– 12 Kioukia N, Nienow AW, Emery AN, Alrubeai M (1995b) Physiological and environmental-factors affecting the growth of insect cells and infection with baculovirus. J Biotech 38:243–251 Klaassen CHW, Bovee-Geurts PHM, DeCaluwe GLJ, DeGrip WJ (1999) Large-scale production and purification of functional recombinant bovine rhodopsin with the use of the baculovirus expression system. Biochem J 342:293–300 Klaassen CHW, DeGrip WJ (2000) Baculovirus expression system for expression and characterization of functional recombinant visual pigments. Vertebrate Phototransduction and the Visual Cycle 315:12–29 Kobilka BK (1995) Amino and carboxyl-terminal modifications to facilitate the production and purification of a G-protein-coupled receptor. Anal Biochem 231:269–271 Kuroda K, Geyer H, Geyer R, Doerfler W, Klenk HD (1990) The oligosaccharides of influenza-virus hemagglutinin expressed in insect cells by a baculovirus vector. Virology 174:418–429 Lagane B, Gaibelet G, Meilhoc E, Masson JM, Cezanne L, Lopez A (2000) Role of sterols in modulating the human mu-opioid receptor function in Saccharomyces cerevisiae. J Biol Chem 275:33197–33200 Lam YH, Wassall SR, Morton CJ, Smith R, Separovic F (2001) Solid-state NMR structure determination of melittin in a lipid environment. Biophys J 81:2752– 2761 Litman BJ, Mitchell DC (1996) A role for phospholipid polyunsaturation in modulating membrane protein function. Lipids 31:S193–S197 Lundstrom K (2003). Semliki forest virus vectors for rapid and high-level expression of integral membrane proteins. Biochim Biophys Acta Biomemb 1610:90–96 Massotte D, Baroche L, Simonin F, Yu L, Kieffer B, Pattus F (1997) Characterization of delta, kappa, and mu human opioid receptors overexpressed in baculovirus-infected insect cells. J Biol Chem 272:19987– 19992 Massotte D, Pereira CA, Pouliquen Y, Pattus F (1999) Parameters influencing human mu opioid receptor over-expression in baculovirus-infected insect cells. J Biotech 69:39–45 Mroczkowski BS, Huvar A, Lernhardt W, Misono K, Nielson K, Scott B. (1994) Secretion of thermostable DNA-polymerase using a novel baculovirus vector. J Biol Chem 269:13522–13528
777 Murphy CI, McIntire JR, Davis DV, Hodgdon H, Seals JR, Young E (1993) Enhanced expression, secretion, and large-scale purification of recombinant Hiv1 Gp120 in insect cells using the baculovirus Egt and P67 signal peptides. Protein Expr Purif 4:349–357 Murphy OJ, Kovacs FA, Sicard EL, Thompson LK (2001). Site-directed solid-state NMR measurement of a ligand-induced conformational change in the serine bacterial chemoreceptor. Biochemistry 40:1358–1366 Okamoto Y, Ninomiya H, Tanioka M, Sakamoto A, Miwa S, Masaki T (1997) Palmitoylation of human endothelin(B)—Its critical role in G protein coupling and a differential requirement for the cytoplasmic tail by G-protein subtypes. J Biol Chem 272:21589–21596 Pereira CA, Pouliquen Y, Rodas V, Massotte D, Mortensen C, Sogayar MC, Menissier-de Murcia J (2001) Optimized insect cell culture for the production of recombinant heterologous proteins and Baculovirus particles. Biotechniques 31:1262–1269 Ponimaskin EG, Heine M, Joubert L, Sebben M, Bickmeyer U, Richter DW, Dumuis A (2002) The 5hydroxytryptamine(4a) receptor is palmitoylated at two different sites, and acylation is critically involved in regulation of receptor constitutive activity. J Biol Chem 277:2534–2546 Raleigh DP, Levitt MH, Griffin RG (1988) Rotational resonance in solid-state NMR. Chem Phys Lett 146:71–76 Ratnala VRP, de Groot HJM, DeGrip WJ (2003a) Largescale expression, purification, and reconstitution of a GPCR for SSNMR studies. Biophys J 84:496A–496A Ratnala VRP, Hulsbergen F, de Groot HJM, DeGrip WJ (2003b) Analysis of histamine and modeling of ligandreceptor interactions in the histamine H(1) receptor for magic angle spinning NMR studies. Inflamm Res 52:417–423 Ratnala VRP, Swarts HGP, VanOostrum J, Leurs R, DeGroot HJM, Bakker RA, DeGrip WJ (2004). Large-scale overproduction, functional purification and ligand affinities of the His-tagged human histamine H1 receptor. Eur J Biochem 271:2636–2646 Reilander H, Boege F, Vasudevan S, Maul G, Hekman M, Dees C, Hampe W, Helmreich EJM, Michel H (1991) Purification and functional-characterization of the human-beta-2-adrenergic receptor produced in baculovirus-infected insect cells. Febs Lett 282:441–444 Reilander H, Weiss HM (1998) Production of G-proteincoupled receptors in yeast. Curr Opin Biotechnol 9:510–517 Rice JW, Rankl NB, Gurganus TM, Marr CM, Barna JB, Walters MM, Burns DJ (1993) A comparison of large-scale Sf9 insect-cell growth and proteinproduction—stirred vessel Vs airlift. Biotechniques 15:1052–1059 Rinken A, Kameyama K, Haga T, Engstrom L (1994) Solubilization of muscarinic receptor subtypes from baculovirus-infected Sf9 insect cells. Biochem Pharmacol 48:1245–1251 Robeva AS, Woodard R, Luthin DR, Taylor HE, Linden J (1996) Double tagging recombinant A(1)- and A(2A)adenosine receptors with hexahistidine and the FLAG
123
778 epitope—Development of an efficient generic protein purification procedure. Biochem Pharmacol 51:545– 555 Rosenbusch JP (2001) Stability of membrane proteins: relevance for the selection of appropriate methods for high-resolution structure determinations. J Struct Biol 136:144–157 Schindler PT, Macherhammer F, Arnold S, Reuss M, Siemann M (1999) Investigation of translation dynamics under cell-free protein biosynthesis conditions using high-resolution two-dimensional gel electrophoresis. Electrophoresis 20:806–812 Toll L, Snyder SH (1982) Solubilization and characterization of histamine H1 receptors in brain. J Biol Chem 257:13593–13601 Tycko R (2001) Biomolecular solid state NMR: advances in structural methodology and applications to peptide and protein fibrils. Annu Rev Phys Chem 52:575–606 Wallace RM (1983) Temperature dependence of the binding of [3H]mepyramine and related compounds to the histamine H1 receptor. Mol Pharmacol 23:60– 66
123
Biotechnol Lett (2006) 28:767–778 Wang MY, Kwong S, Bentley WE (1993) Effects of oxygen glucose glutamine feeding on insect-cell baculovirus protein expression—a study on epoxide hydrolase production. Biotechnol Prog 9:355–361 Watts JA, Watts A, Middleton DA (2001) A model of reversible inhibitors in the gastric H+/K+-ATPase binding site determined by rotational echo double resonance NMR. J Biol Chem 276:43197–43204 Xu R, Ayers B, Cowburn D, Muir TW (1999) Chemical ligation of folded recombinant proteins: segmental isotopic labeling of domains for NMR studies. Proc Natl Acad Sci USA 96:388–393 Yang JD, Gecik P, Collins A, Czarnecki S, Hsu HH, Lasdun A, Sundaram R, Muthukumar G, Silberklang M (1996) Rational scale-up of a baculovirus-insect cell batch process based on medium nutritional depth. Biotechnol Bioeng 52:696–706 Zhang FM, Saarinen MA, Itle LJ, Lang SC, Murhammer DW, Linhardt RJ (2002) The effect of dissolved oxygen (DO) concentration on the glycosylation of recombinant protein produced by the insect cell-baculovirus expression system. Biotechnol Bioeng 77:219–224
