EMBO Practical course on protein expression, purification and crystallization’ – PEPC5, 2006
Expression of membrane proteins for structure determination
Reinhard GrisshammerNINDS, NIH, Bethesda MD, USA
Department of Health and Human Services
objectives
• expression of membrane proteins for structure determination (not for ‘functional’ experiments)– large amounts (milligram quantities)– functional, correctly-folded membrane protein– source: membrane– alternative source: inclusion bodies → refolding to
obtain functional membrane protein• bacterial vs. eukaryotic membrane proteins
topics
• general considerations, mechanism of membrane insertion, topology
• examples of overexpression of bacterial and eukaryotic membrane proteins in various hosts systems
• functional expression of G-protein-coupled receptors (GPCRs) in E. coli
• factors influencing the expression levels and stability of GPCRs
• functional expression of GPCRs in eukaryotic hosts
topics
• general considerations, mechanism of membrane insertion, topology
• examples of overexpression of bacterial and eukaryotic membrane proteins in various hosts systems
• functional expression of G-protein-coupled receptors (GPCRs) in E. coli
• factors influencing the expression levels and stability of GPCRs
• functional expression of GPCRs in eukaryotic hosts
structure determination of membrane proteins
• membrane proteins are encoded by about 30% of all genes
• 218 coordinate sets for membrane proteins deposited, over 38,000 structure entries of soluble proteins in the Protein Data Bank (Aug 2006)
• Hartmut Michel http://www.mpibp-frankfurt.mpg.de/michel/public/memprotstruct.html
• Martin Caffrey http://www.lipidat.chemistry.ohio-state.edu/MPDB/index.asp or http://www.mpdb.ul.ie\par
• Steve White http://blanco.biomol.uci.edu/Membrane_Proteins_xtal.html
structure determination of membrane proteins – 3D / X-ray
• Photosynthetic reaction centers, light-harvesting complexes• Cytochrome c oxidase• Cytochrome bc1 complexes, bc1 complex with cytochrome c, cytochrome b6f• Potassium channels (KcsA …), chloride channel from E. coli (ClC)• Mechanosensitive ion channels (MscL, MscS)• Bacteriorhodopsin, halorhodopsin, sensoryrhodopsin II, SRII-transducer
complex• Fumarate reductase / succinate dehydrogenase• F1Fo-ATPase (F1 and c subunits)• Sarcoplasmic reticulum calcium-ATPase• Rhodopsin• Aquaporin AQP1, glycerol facilitator GlpF• Photosystems I and II• Lipid A transporter from E. coli (MsbA), vitamin B12 uptake transporter from E.
coli (BtuCD)• Formate dehydrogenase-N from E. coli (Fdn-N)• AcrB multidrug exporter from E. coli• Lac permease (LacY), glycerol-3-phosphate transporter (GlpT)• Sec protein-conducting channel from Methanococcus jannaschii• …
structure determination of membrane proteins – 3D / X-ray
• FepA / FhuA / FecA / BtuB outer membrane iron transporters• BtuB bound with colicin• Porins (OmpF, PhoE, LamB, ScrY)• Alpha-hemolysin• Outer membrane phospholipase (OMPLA)• Outer membrane protease OmpT, OpcA• TolC channel• Outer membrane proteins OmpA, OmpX
• Prostaglandin H2 synthase• Squalene-hopene cyclase• …
structure determination of membrane proteins – 3D / X-ray
• mostly bacterial / archae-bacterial membrane protein structures
• few eukaryotic membrane protein structures– natural source (rhodopsin from retina, Ca-ATPase
from muscle)– even fewer recombinant eukaryotic membrane
protein structures
structures of recombinant eukaryotic membrane proteins
• rat voltage-gated potassium channel (methylotrophicyeast Pichia pastoris) Long, MacKinnon 2005
• crystals of the rabbit calcium ATPase (S. cerevisiae) Jidenko, Nissen 2005
• spinach aquaporin (Pichia pastoris) Törnroth-Horsefield, Kjellbom 2006
• rat aquaporin (insect cell – baculovirus) Hiroaki, Fujiyoshi 2006
integral membrane protein
▫ hydrophobic → detergents▫ expression is often toxic for the host▫ no universal expression system for all membrane proteins
insertion of membrane proteins into the membrane
topology of integral membrane proteins
• Kim, von Heijne, PNAS 103: 11142, 2006, yeast membrane proteome
• Cout ~ 20%• Cin ~ 80%, even number of TMDs predominate
topology of integral membrane proteins
++
+ + + ++
+
• positive-inside rule (von Heijne, Nature 341: 456, 1989)• hydrophobicity of TMD• length of TMD
recognition of transmembrane helices by the endoplasmic reticulum translocon
• Hessa, White, von Heijne: Nature 433: 377-381, 2005
• in vitro assay to quantify the efficiency with which designed transmembranesegments insert into dog pancreas rough microsomes
recognition of transmembrane helices by the endoplasmic reticulum translocon
derived from H-segments with the indicated amino acid placed in the middle of the 19-residue hydrophobic stretch
recognition of transmembrane helices by the endoplasmic reticulum translocon
• ‘biological’ hydrophobicity scale, much in common with hydrophobicity scales derived from biophysical measurements
• implies that direct protein-lipid interactions are involved in the recognition of TM helices by the translocon
• strong dependence on sequence position of aromatic and charged residues in TM segments
• can one ‘optimize’ membrane protein expression by sequence analysis ?
structure of translocon
• v. d. Berg, Rapoport, Nature 427: 36, 2004 Methanococcus jannaschii
topics
• general considerations, mechanism of membrane insertion, topology
• examples of overexpression of bacterial and eukaryotic membrane proteins in various hosts systems
• functional expression of G-protein-coupled receptors (GPCRs) in E. coli
• factors influencing the expression levels and stability of GPCRs
• functional expression of GPCRs in eukaryotic hosts
expression systems
• mammalian cells• stable, transient transfection• Semliki Forest Virus
• insect cells• stable (Sf9, Drosophila Schneider S2)• baculovirus system
• yeast• chromosomal integration• plasmids
• Escherichia coli, Lactococcus lactis
which expression system is best for high-level production of functional receptors ?
• no universal expression system for all membrane proteins
• trial and error approach
• but: membrane protein overexpression is possible !!
requirements for correct folding and function of membrane protein
• post-translational modifications• N-glycosylation• disulphide bond formation
• lipid composition of host membrane• molecular chaperones• …
practical aspects
• maintenance of cell line• may be difficult for stable mammalian cell lines• easy for E. coli and yeast
• scale-up of expression• yes: stable mammalian cells, insect cells /
baculovirus, yeast, E. coli• biological safety aspects• cell breakage at large scale
• problematic in case of yeast ?
problems with expression systems
• mammalian cells• incorrect folding has been shown
• insect cells• membrane protein may not be N-glycosylated• high-mannose type glycosylation• membranes with low levels of cholesterol• large proportion of membrane protein may be incorrectly folded
• yeast• proteolysis• targeting of membrane protein to vacuole• ergosterol (no cholesterol)
• Escherichia coli• proteolysis• inclusion body formation (refolding necessary)• no N-glycosylation• lack of cholesterol
why is membrane protein overexpressiontoxic to host cells ?
• function of membrane protein itself (e.g. ion channel)• pGluR: potassium selective glutamate receptor from
Synechocystis (M1-P-M2, N-out) cannot be expressed in E. coli
• but: KcsA potassium channel from Streptomyces lividans(M1-P-M2, N-in) can be overexpressed in E. coli– Schrempf, EMBO J. 14: 5170, 1995– Heginbotham, Biochem. 36: 10335, 1997; synthetic
gene, Eco high codon usage, N-terminal His tag, C-terminal Strep tag, pASK75 (tet promoter)
– Cortes, Biochem. 36: 10343, 1997; N-terminal His tag, pQE vector
why is membrane protein overexpressiontoxic to host cells ?
• constitutive activity of membrane protein can promote intracellular signaling– GPCR β2-adrenergic receptor in stable CHO cells
at 200 pmol/mg (20 Mio R/cell), constitutive expression (Lohse, Naunyn-Schmiedeberg’s Arch. Pharmacol. 345, 444, 1992) → cell line died
– but: GPCR calcitonin receptor in stable MEL cells, integration-independent, erythroid-specific expression from β-globin promoter, differentiation / induction with DMSO → 60 pmol/mg (2 Mio R/cell) (Needham, PEP 6, 124, 1995)
why is membrane protein overexpressiontoxic to host cells ?
• presence of large amounts of membrane protein could disturb membrane– but: overexpression of fumarate reductase in E.
coli leads to additional intracellular membrane systems containing ordered arrays of Frd (Weiner, J. Bacteriol. 158: 590, 1984)
why is membrane protein overexpressiontoxic to host cells ?
• T7 RNA polymerase expression system / BL21(DE3) (Miroux and Walker, JMB 260: 289, 1996)– protein overexpression is limited or prevented by
cell death– selection procedure to allow high-level expression
of target proteins (deposited as inclusion bodies)– high cell density, no toxic effect
• transcription, translation, membrane insertion or inclusion body formation cannot be considered separately
• no general cloning strategy that guarantees overexpression of a given prokaryotic membrane protein in E. coli
• Gunn, Tate, Henderson: sugar-H+ symport protein FucP, Mol. Microbiol. 12: 799-809, 1994
topics
• general considerations, mechanism of membrane insertion, topology
• examples of overexpression of bacterial and eukaryotic membrane proteins in various hosts systems
• functional expression of G-protein-coupled receptors (GPCRs) in E. coli
• factors influencing the expression levels and stability of GPCRs
• functional expression of GPCRs in eukaryotic hosts
G-protein-coupled receptors
(Palczewski et al., 2000)(Okada et al., 2002)
(Li et al., 2004)
rat neurotensin receptor (NTR, NTS-1)
• 424 aa, 47 kDa• rat brain cDNA library• Tanaka et al., Neuron 4, 847, 1990• cDNA from S. Nakanishi
• neurotensinGlp-Leu-Tyr-Glu-Asn-Lys-Pro-Arg-Arg-Pro-Tyr-Ile-LeuNTR modulates dopaminergic neuronsNTR is involved in pancreatic cancer
expression of NTR in eukaryotic cells
• insect cells / baculovirus (transient, NTR-H5M)• 188000 R/cell or 0.02 mg/L
• MEL cells (stable, NTR-H6F)• 455000 R/cell or 0.04 mg/L• proteolysis of C-terminus
• CHO, pCyt-TS (stable, NTR-H10F)• 360000 R/cell or 0.03 mg/L
expression of seven-helix G-protein coupled receptors in Escherichia coli
practical aspects• maintenance of “cell line” is easy• scale-up of expression is possible• cell breakage at large scale not problematic
tools for assessing expression levels
• functional receptors → ligand binding analysis• total receptor protein → Western blot (‘tag’)
expression as maltose-binding protein fusion
influence of tag on expression
Tucker & Grisshammer, 1996
expression in E. coli of the neurotensinreceptor fusion protein
expression levels of NTR fusion protein
• 1000 receptors/cell ([3H]NT)• 3-5 nmol/L of culture (0.3 - 0.5 mg/L)• 9 pmol/mg of total solubilized protein• 24 pmol/mg of membrane protein
expression levels of NTR fusion protein
• monitor some parameters during expression
• pH, OD600 vs. radio-ligandbinding assay
• source of media
general applicability of expression system
• Cannabinoid CB1 receptor: No (degraded)• Cannabinoid CB2 receptor: 17-39 pmol/mg
(MBP-CB2-HF) (Calandra et al., 1997)
• Substance K receptor: 7 pmol/mg(MBP-SKR-HMTX) (Grisshammer et al., 1994)
• Neurotensin receptor: 24 pmol/mg(MBP-T43NTR-TrxA-H10) (Grisshammer & Tucker, 1997)
• Adenosine A2a receptor: 17-34 pmol/mg(MBP-A2aTr316-H10) (Weiß & Grisshammer, 2002)
topics
• general considerations, mechanism of membrane insertion, topology
• examples of overexpression of bacterial and eukaryotic membrane proteins in various hosts systems
• functional expression of G-protein-coupled receptors (GPCRs) in E. coli
• factors influencing the expression levels and stability of GPCRs
• functional expression of GPCRs in eukaryotic hosts
expression of GPCRs in P. pastoris, insect cells, SFV
• Andre, Pattus, Michel, Reinhart, Protein Science 15: 1115, 2006: 20 GPCRs in Pichia pastoris
• Akermoun, Gearing, PEP 44: 65, 2005: 16 GPCRs in 3 insect cell lines
• Hassaine, Lundstrom PEP 45: 343, 2006: SFV 101-GPCRs
• even closely related proteins behave differently• Western blot analysis vs. ligand binding: Western blot
signals do not allow any correlation with amount of correctly folded receptors
summary
membrane protein overexpression is possiblemembrane proteins show individualitybacterial targets best made in E. colieukaryotic targets best made in eukaryotic hosts (exception: some GPCRs …)analysis mode for functionality