Use of six chimeric proteins to investigate the role of intramolecular interactions in determining the kinetics of carnitine palmitoyltransferase I isoforms

The two isoforms of carnitine palmitoyltransferase I (CPT I; muscle (M)- and liver (L)-type) of the mitochondrial outer membrane have distinct kinetic characteristics with respect to their affinity for one of the substrates (l-carnitine) and the inhibitor malonyl-CoA. Moreover, they differ markedly in their hysteretic behavior with respect to malonyl-CoA and in their response to changes in the in vivo metabolic state. However, the two proteins are 62% identical and have the same overall structure. Using liver mitochondria, we have previously shown that the protein is polytopic within the outer membrane, comprising a 46-residue cytosolic N-terminal sequence, two transmembrane segments (TM1 and TM2) separated by a 27-residue loop, and a large catalytic domain (also cytosolic) (Fraser, F., Corstorphine, C. G., and Zammit, V. A. (1997) Biochem. J. 323, 711-718). We have now conducted a systematic study on six chimeric proteins constructed from combinations of three linear segments of rat L- and M-CPT I and on the two parental proteins to elucidate the effects of altered intramolecular interactions on the kinetics of CPT activity. The three segments were (i) the cytosolic N-terminal domain plus TM1, (ii) the loop plus TM2, and (iii) the cytosolic catalytic C-terminal domain. The kinetic properties of the chimeric proteins expressed in Pichia pastoris were studied. We found that alterations in the combinations of the N-terminal plus TM1 and C-terminal domains as well as in the N terminus plus TM1/TM2 pairings resulted in changes in the K(m) values for carnitine and palmitoyl-CoA and the sensitivity to malonyl-CoA of the L-type catalytic domain. The changes in affinity for malonyl-CoA and palmitoyl-CoA occurred independently of changes in the affinity for carnitine. The kinetic characteristics of the M-type catalytic domain and, in particular, its malonyl-CoA sensitivity were much less susceptible to influence by exchange of the other two segments of the protein. The marked difference in the response of the two catalytic domains to changes in the N-terminal domain and TM combinations explains the previously observed differences in the response of L- and M-CPT I to altered physiological state in intact mitochondria and to modulation of altered lipid molecular order of the mitochondrial outer membrane in vivo and in vitro.     

Copper transfer to the N-terminal domain of the Wilson disease protein (ATP7B): X-ray absorption spectroscopy of reconstituted and chaperone-loaded metal binding domains and their interaction with exogenous ligands

The copper-transporting ATPases are 165-175 kDa membrane proteins, composed of 8 transmembrane segments and two large cytosolic domains, the N-terminal copper-binding domain and the catalytic ATP-hydrolyzing domain. In ATP7B, the Wilson disease protein, the N-terminal domain is made up of six metal-binding sub-domains containing the MXCXXC motif which is known to coordinate copper via the two cysteine residues. We have expressed the N-terminal domain of ATP7B as a soluble C-terminal fusion with the maltose binding protein. This expression system produces a protein which can be reconstituted with copper without recourse to the harsh denaturing conditions or low pH reported by other laboratories. Here we describe the reconstitution of the metal binding domains (MBD) with Cu(I) using a number of different protocols, including copper loading via the chaperone, Atox1. X-ray absorption spectra have been obtained on all these derivatives, and their ability to bind exogenous ligands has been assessed. The results establish that the metal-binding domains bind Cu(I) predominantly in a bis cysteinate environment, and are able to bind exogenous ligands such as DTT in a similar fashion to Atox1. We have further observed that exogenous ligand binding induces the formation of a Cu-Cu interaction which may signal a conformational change of the N-terminal domain.     

A heteromeric complex of the two nucleotide binding domains of cystic fibrosis transmembrane conductance regulator (CFTR) mediates ATPase activity

The cystic fibrosis transmembrane conductance regulator (CFTR) protein is a member of the ABC superfamily of transporter proteins. Recently, crystal structures of intact, prokaryotic members of this family have been described. These structures suggested that ATP binding and hydrolysis occurs at two sites formed at the interface between their nucleotide binding domains (NBDs). In contrast to the prokaryotic family members, the NBDs of CFTR are asymmetric (both structurally and functionally), and previous to the present studies, it was not clear whether both NBDs are required for ATP hydrolysis. In order to assess the relative roles of the two NBDs of human CFTR, we purified and reconstituted NBD1 and NBD2, separately and together. We found that NBD1 and NBD2 by themselves exhibited relatively low ATPase activity. Co-assembly of NBD1 and NBD2 exhibited a 2-3-fold enhancement in catalytic activity relative to the isolated domains and this increase reflected enhanced ATP turnover (V(max)). These data provide the first direct evidence that heterodimerization of the NBD1 and NBD2 domains of CFTR is required to generate optimal catalytic activity.     

Three-dimensional structure of P-glycoprotein: the transmembrane regions adopt an asymmetric configuration in the nucleotide-bound state

Multidrug resistance of cancer cells and pathogens is a serious clinical problem. A major factor contributing to drug resistance in cancer is the over-expression of P-glycoprotein, a plasma membrane ATP-binding cassette (ABC) drug efflux pump. Three-dimensional structural data with a resolution limit of approximately 8 A have been obtained from two-dimensional crystals of P-glycoprotein trapped in the nucleotide-bound state. Each of the two transmembrane domains of P-glycoprotein consists of six long alpha-helical segments. Five of the alpha-helices from each transmembrane domain are related by a pseudo-2-fold symmetry, whereas the sixth breaks the symmetry. The two alpha-helices positioned closest to the (pseudo-) symmetry axis at the center of the molecule appear to be kinked. A large loop of density at the extracellular surface of the transporter is likely to correspond to the glycosylated first extracellular loop, whereas two globular densities at the cytoplasmic side correspond to the hydrophilic, nucleotide-binding domains. This is the first three-dimensional structure for an intact eukaryotic ABC transporter. Comparison with the structures of two prokaryotic ABC transporters suggests significant differences in the packing of the transmembrane alpha-helices within this protein family.     

Molecular structure of the glibenclamide binding site of the beta-cell K(ATP) channel.

We have investigated the structure of the glibenclamide binding site of pancreatic beta-cell ATP-sensitive potassium (K(ATP)) channels. K(ATP) channels are a complex of four pore-forming Kir6.2 subunits and four sulfonylurea receptor (SUR1) subunits. SUR1 (ABCC8) belongs to the ATP binding cassette family of proteins and has two nucleotide binding domains (NBD1 and NBD2) and 17 putative transmembrane (TM) sequences. Co-expression in a baculovirus expression system of two parts of SUR1 between NBD1 and TM12 leads to restoration of glibenclamide binding activity, whereas expression of either individual N- or C-terminal part alone gave no glibenclamide binding activity, confirming a bivalent structure of the glibenclamide binding site. By using N-terminally truncated recombinant proteins we have shown that CL3 - the cytosolic loop between TM5 and TM6 - plays a key role in formation of the N-terminal component of the glibenclamide binding site. Analysis of deletion variants of the C-terminal part of SUR1 showed that CL8 - the cytosolic loop between TM15 and TM16 - is the only determinant for the C-terminal component of the glibenclamide binding site. We suggest that in SUR1 in the native K(ATP) channel close proximity of CL3 and CL8 leads to formation of the glibenclamide binding site.     

Split Na+-Ca2+ exchangers. Implications for function and expression

The Na(+)-Ca(2+) exchanger has nine transmembrane segments, with a large cytoplasmic loop between the fifth and sixth transmembrane segments. The protein was split within the cytoplasmic loop into two domains consisting of the first five transmembrane segments and the last four transmembrane segments, respectively. The two domains were either expressed individually or coexpressed. Each of the two domains with different lengths of the cytoplasmic loop was fused to green fluorescent protein. We show that coexpression of both domains is required for proper membrane targeting and for expression of functional exchange activity. Fusion to green fluorescent protein does not alter biophysical properties of the exchange process. In addition, truncation of a large portion of the cytoplasmic loop does not alter important properties of the exchanger such as Na(+)-dependent inactivation, activation by chymotrypsin, or exchanger inhibitory peptide (XIP) sensitivity.     

Atrial natriuretic factor-receptor guanylate cyclase signal transduction mechanism

Atrial natriuretic factor (ANF) receptor guanylate cyclase (ANF-RGC), like the other members of the membrane guanylate cyclase family, is a single transmembrane-spanning protein. The transmembrane domain separates the protein into two regions, extracellular and intracellular. The extracellular region contains the ANF-binding domain and the intracellular region the catalytic domain located at the C-terminus of the protein. Preceding the catalytic domain, the intracellular region is comprised of the following functional domains: juxtaposed 40 amino acids to the transmembrane domain is the ATP-regulated module (ARM) domain [also termed the kinase homology domain (KHD)], and the putative dimerization domain. The ANF-RGC signaling is initiated by hormone, ANF, binding to its extracellular binding site. The binding signal is transduced through the transmembrane domain to the intracellular portion where ATP binding to the ARM domain partially activates the cyclase and prepares it for subsequent steps involving phosphorylation and attaining the fully activated state. This chapter reviews the signaling modules of ANF-RGC.     

An isoform-specific interaction of the membrane anchors affects mammalian adenylyl cyclase type V activity.

The nine membrane-bound mammalian adenylyl cyclases (ACs) contain two highly diverged membrane anchors, M1 and M2, with six transmembrane spans each and two conserved cytosolic domains which coalesce into a pseudoheterodimeric catalytic unit. Previously, the catalytic segments, bacterially expressed as soluble proteins, were characterized extensively whereas the function of the membrane domains remained unexplored. Using the catalytic C1 and C2 domains of AC type V we employed the membrane anchors from type V and VII ACs for construction of enzymes with duplicated, inverted, fully swapped and chimeric membrane anchors. Further, in the M1 membrane domain individual transmembrane spans were removed or exchanged between type V and VII ACs. The constructs were expressed in HEK293 cells, the expression levels and membrane localization was assessed by Western blotting. Cell-free basal, forskolin-, GTP gamma S-and G(s alpha)/GTP gamma S-stimulated AC activities were determined. The results demonstrate that enzymatic activities were only maintained when the M1 and M2 membrane domains were derived from either AC V or VII. Constructs with chimeric membrane domains, i.e. M1 from type V and M2 from type VII AC or vice versa, were essentially inactive although the expression levels and membrane localization appeared to be normal. The data indicate a functionally important interaction of the membrane domains of ACs in that they seem to interact in a pair-like, isoform delimited manner. This interaction directly impinges on the formation of the catalytic interface. We propose that protein-protein interactions of the AC membrane domains may constitute another, yet unexplored level of AC regulation.     

The inhibition of degradation of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase by sterol regulatory element binding protein cleavage-activating protein requires four phenylalanine residues in span 6 of HMG-CoA reductase transmembrane domain

3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGR) is the rate-limiting enzyme in the cholesterol biosynthetic pathway. This endoplasmic reticulum membrane protein contains a cytosolic catalytic domain and a transmembrane domain with eight membrane spans that are necessary for sterol-accelerated degradation. Competition experiments showed that wild-type transmembrane domains of HMGR and sterol regulatory element binding protein cleavage-activating protein (SCAP) blocked sterol-accelerated degradation of intact HMGR and HMGal, a model protein containing the membrane domain of HMGR linked to Escherichia coli beta-galactosidase. However, mutant transmembrane domains of HMGR and SCAP whose sterol-sensing functions were abolished did not inhibit sterol-accelerated degradation of HMGR and HMGal. In addition, our mutagenesis studies on HMGal indicated that four Phe residues conserved in span 6 of HMGR and the sterol-sensing domains of other sterol-related proteins are required for the regulated degradation of HMGR. These results suggest that HMGR and SCAP compete for binding to a sterol-regulated regulator protein, and this binding may need the four Phe residues.     

A new experimental approach to detect long-range conformational changes transmitted between the membrane and cytosolic domains of LmrA,a bacterial multidrug transporter

LmrA confers multidrug resistance to Lactococcus lactis by mediating the extrusion of antibiotics, out of the bacterial membrane, using the energy derived from ATP hydrolysis. Cooperation between the cytosolic and membrane-embedded domains plays a crucial role in regulating the transport ATPase cycle of this protein. In order to demonstrate the existence of a structural coupling required for the cross-talk between drug transport and ATP hydrolysis, we studied specifically the dynamic changes occurring in the membrane-embedded and cytosolic domains of LmrA by combining infrared linear dichroic spectrum measurements in the course of H/D exchange with Trp fluorescence quenching by a water-soluble attenuator. This new experimental approach, which is of general interest in the study of membrane proteins, detects long-range conformational changes, transmitted between the membrane-embedded and cytosolic regions of LmrA. On the one hand, nucleotide binding and hydrolysis in the cytosolic nucleotide binding domain cause a repacking of the transmembrane helices. On the other hand, drug binding to the transmembrane helices affects both the structure of the cytosolic regions and the ATPase activity of the nucleotide binding domain.     

FRET analysis indicates that the two ATPase active sites of the P-glycoprotein multidrug transporter are closely associated

Members of the ABC superfamily carry out the transport of various molecules and ions across cellular membranes, powered by ATP hydrolysis. Substantial evidence indicates that the two catalytic sites of the nucleotide binding domains function in a highly cooperative, alternating sites mode, which suggests the possibility that they interact with each other physically. In this study, fluorescence energy transfer experiments were used to estimate the distance between two fluors, each covalently linked to a highly conserved Cys residue (Cys428 and Cys1071) within the Walker A motif of the catalytic site. The vanadate.ADP.Mg(2+) complex was trapped in one catalytic site of membrane-bound or highly purified P-glycoprotein, and the other site was labeled with MIANS [2-(4'-maleimidylanilino)naphthalene-6-sulfonic acid]. Following loss of the trapped vanadate complex, the newly vacant site was then labeled with NBD-Cl (7-chloro-4-nitrobenzo-2-oxa-1,3-diazole). The fluorescence properties of the singly labeled P-glycoproteins showed that no energy transfer occurred between MIANS (the donor) and NBD (the acceptor) when they were simply mixed together. On the other hand, the fluorescence emission of the MIANS group in doubly labeled P-glycoprotein was highly quenched as a result of energy transfer to NBD, leading to an estimate of a donor-acceptor separation distance of approximately 16 A for P-glycoprotein labeled in the native plasma membrane and approximately 22 A for P-glycoprotein labeled in detergent solution. The separation of the two fluorophores is compatible with the recently reported crystal structure of the Rad50cd dimer, but not with that of the HisP dimer. These results suggest that the two catalytic sites of the P-glycoprotein nucleotide binding domains are relatively close together, which would facilitate cooperation between them during the catalytic cycle.     

Role of the diacylglycerol kinase alpha-conserved domains in membrane targeting in intact T cells

Diacylglycerol kinase (DGK) phosphorylates diacylglycerol to phosphatidic acid, modifying the cellular levels of these two lipid mediators. Ten DGK isoforms, grouped into five subtypes, are found in higher organisms. All contain a conserved C-terminal domain and at least two cysteine-rich motifs of unknown function. DGKalpha is a type I enzyme that acts as a negative modulator of diacylglycerol-based signals during T cell activation. Here we studied the functional role of the DGKalpha domains using mutational analysis to investigate membrane binding in intact cells. We show that the two atypical C1 domains are essential for plasma membrane targeting of the protein in intact cells but unnecessary for catalytic activity. We also identify the C-terminal sequence of the protein as essential for membrane binding in a phosphatidic acid-dependent manner. Finally we demonstrate that, in the absence of the calcium binding domain, receptor-dependent translocation of the truncated protein is regulated by phosphorylation of Tyr(335). This functional study provides new insight into the role of the so-called conserved domains of this lipid kinase family and demonstrates the existence of additional domains that confer specific plasma membrane localization to this particular isoform.     

Protein modules as organizers of membrane structure.

Investigations conducted over the past 18 months have shed new light on how modular protein-binding domains, in particular PDZ domains, co-ordinate the assembly of functional plasma membrane domains. Members of the MAGUK (membrane-associated guanylate kinase) protein family, like PSD-95, use multiple domains to cluster ion channels, receptors, adhesion molecules and cytosolic signaling proteins at synapses, cellular junctions, and polarized membrane domains. Other PDZ proteins, like the Drosophila protein INAD and the epithelial Na(+)/H(+) regulatory factor (NHERF), organize cellular signaling by localizing transmembrane and cytosolic components to specific membrane domains and assembling these components into functional complexes. The organization of these proteins into discreet structures has functional consequences for downstream signaling.     

Review. ATP hydrolysis-driven gating in cystic fibrosis transmembrane conductance regulator

Proteins belonging to the ATP-binding cassette superfamily couple ATP binding and hydrolysis at conserved nucleotide-binding domains (NBDs) to diverse cellular functions. Most superfamily members are transporters, while cystic fibrosis transmembrane conductance regulator (CFTR), alone, is an ion channel. Despite this functional difference, recent results have suggested that CFTR shares a common molecular mechanism with other members. ATP binds to partial binding sites on the surface of the two NBDs, which then associate to form a NBD dimer, with complete composite catalytic sites now buried at the interface. ATP hydrolysis and gamma-phosphate dissociation, with the loss of molecular contacts linking the two sides of the composite site, trigger dimer dissociation. The conformational signals generated by NBD dimer formation and dissociation are transmitted to the transmembrane domains where, in transporters, they drive the cycle of conformational changes that translocate the substrate across the membrane; in CFTR, they result in opening and closing (gating) of the ion-permeation pathway.     

A tweezers-like motion of the ATP-binding cassette dimer in an ABC transport cycle

The ATPase components of ATP binding cassette (ABC) transporters power the transporters by binding and hydrolyzing ATP. Major conformational changes of an ATPase are revealed by crystal structures of MalK, the ATPase subunit of the maltose transporter from Escherichia coli, in three different dimeric configurations. While other nucleotide binding domains or subunits display low affinity for each other in the absence of the transmembrane segments, the MalK dimer is stabilized through interactions of the additional C-terminal domains. In the two nucleotide-free structures, the N-terminal nucleotide binding domains are separated to differing degrees, and the dimer is maintained through contacts of the C-terminal regulatory domains. In the ATP-bound form, the nucleotide binding domains make contact and two ATPs lie buried along the dimer interface. The two nucleotide binding domains of the dimer open and close like a pair of tweezers, suggesting a regulatory mechanism for ATPase activity that may be tightly coupled to translocation.     

Molecular Disruption of the Power Stroke in the ATP-binding Cassette Transport Protein MsbA

ATP-binding cassette transporters affect drug pharmacokinetics and are associated with inherited human diseases and impaired chemotherapeutic treatment of cancers and microbial infections. Current alternating access models for ATP-binding cassette exporter activity suggest that ATP binding at the two cytosolic nucleotide-binding domains provides a power stroke for the conformational switch of the two membrane domains from the inward-facing conformation to the outward-facing conformation. In outward-facing crystal structures of the bacterial homodimeric ATP-binding cassette transporters MsbA from Gram-negative bacteria and Sav1866 from Staphylococcus aureus, two transmembrane helices (3 and 4) in the membrane domains have their cytoplasmic extensions in close proximity, forming a tetrahelix bundle interface. In biochemical experiments on MsbA from Escherichia coli, we show for the first time that a robust network of inter-monomer interactions in the tetrahelix bundle is crucial for the transmission of nucleotide-dependent conformational changes to the extracellular side of the membrane domains. Our observations are the first to suggest that modulation of tetrahelix bundle interactions in ATP-binding cassette exporters might offer a potent strategy to alter their transport activity.     

Cytoplasmic tail of phospholemman interacts with the intracellular loop of the cardiac Na+/Ca2+ exchanger

Phospholemman (PLM), a member of the FXYD family of small ion transport regulators, inhibits cardiac Na+/Ca2+ exchanger (NCX1). NCX1 is made up of N-terminal domain consisting of the first five transmembrane segments (residues 1-217), a large intracellular loop (residues 218-764), and a C-terminal domain comprising the last four transmembrane segments (residues 765-938). Using glutathione S-transferase (GST) pull-down assay, we demonstrated that the intracellular loop, but not the N- or C-terminal transmembrane domains of NCX1, was associated with PLM. Further analysis using protein constructs of GST fused to various segments of the intracellular loop of NCX1 suggest that PLM bound to residues 218-371 and 508-764 but not 371-508. Split Na+/Ca2+ exchangers consisting of N- or C-terminal domains with different lengths of the intracellular loop were co-expressed with PLM in HEK293 cells that are devoid of endogenous PLM and NCX1. Although expression of N-terminal but not C-terminal domain alone resulted in correct membrane targeting, co-expression of both N- and C-terminal domains was required for correct membrane targeting and functional exchange activity. NCX1 current measurements indicate that PLM decreased NCX1 current only when the split exchangers contained residues 218-358 of the intracellular loop. Co-immunoprecipitation experiments with PLM and split exchangers suggest that PLM associated with the N-terminal domain of NCX1 when it contained intracellular loop residues 218-358. TM43, a PLM mutant with its cytoplasmic tail truncated, did not co-immunoprecipitate with wild-type NCX1 when co-expressed in HEK293 cells, confirming little to no interaction between the transmembrane domains of PLM and NCX1. We conclude that PLM interacted with the intracellular loop of NCX1, most likely at residues 218-358.     

Overlapping effects of S3 stalk segment mutations on the affinity of Ca2+-ATPase (SERCA) for thapsigargin and cyclopiazonic acid

Chimeric exchanges and mutations were produced in the Ca(2+)-ATPase (SERCA) to match (in the majority of cases) corresponding sequences of the Na(+),K(+)-ATPase. The effects of these mutations on the concentration dependence of the specific Ca(2+)-ATPase inhibition by thapsigargin (TG) and cyclopiazonic acid (CPA) were then determined. Extensive chimeric mutations on the large cytosolic loop, on the S4 stalk segment, and on the M3 transmembrane segments produced little or no modification of the Ca(2+)-ATPase sensitivity to either inhibitor. On the other hand, the presence of a six amino acid Na(+), K(+)-ATPase sequence within the S3 stalk segment of the Ca(2+)-ATPase raised 60-fold the apparent K(i) for TG and 250-fold the apparent K(i) for CPA. More limited mutations within the same S3 segment, however, affected differently the concentration dependence of the Ca(2+)-ATPase inhibition by TG or CPA. Specifically, single mutation of Phe256 to Val increased 20-fold the apparent K(i) for TG, while having very little effect on the apparent K(i) for CPA. These findings indicate significant overlap of the TG and CPA binding domains within the S3 stalk segment of the Ca(2+)-ATPase, where the contribution of each protein residue is dependent on the structures of the two inhibitors. Saturating concentrations of either or both TG and CPA produce an identical reduction of the affinity of the ATPase for ATP, suggesting that only one inhibitor can bind at any time due to significant overlap of their binding domains. It is suggested that perturbations produced by binding of either inhibitor within the stalk segment interfere with the long-range functional linkage between ATP utilization in the ATPase cytosolic region and Ca(2+) binding in the membrane-bound region.     

Role of third extracellular domain of plasma membrane Ca2+-Mg2+-ATPase based on the novel inhibitor caloxin 3A1

The plasma membrane Ca2+ pump (PMCA) is a Ca2+-Mg2+-ATPase that expels Ca2+ from cells to help them maintain low concentrations of cytosolic Ca2+ ([Ca2+]i). It contains five putative extracellular domains (PEDs). Earlier we had reported that binding to PED2 leads to PMCA inhibition. Mutagenesis of residues in transmembrane domain 6 leads to loss of PMCA activity. PED3 connects transmembrane domains 5 and 6. PED3 is only five amino acid residues long. By screening a phage display library, we obtained a peptide sequence that binds this target. After examining a number of peptides related to this original sequence, we selected one that inhibits the PMCA pump (caloxin 3A1). Caloxin 3A1 inhibits PMCA but not the sarcoplasmic reticulum Ca2+-pump. Caloxin 3A1 did not inhibit formation of the 140 kDa acylphosphate intermediate from ATP or its degradation. Thus, PEDs play a role in the reaction cycle of PMCA even though sites for binding to the substrates Ca2+ and Mg-ATP2-, and the activator calmodulin are all in the cytosolic domains of PMCA. In endothelial cells exposed to low concentration of a Ca2+-ionophore, caloxin 3A1 caused a further increase in [Ca2+]i proving its ability to inhibit PMCA pump extracellularly. Thus, even though PED3 is the shortest PED, it plays key role in the PMCA function.     

Targeting of a Nicotiana plumbaginifolia H+ -ATPase to the plasma membrane is not by default and requires cytosolic structural determinants

The structural determinants involved in the targeting of multitransmembrane-span proteins to the plasma membrane (PM) remain poorly understood. The plasma membrane H+ -ATPase (PMA) from Nicotiana plumbaginifolia, a well-characterized 10 transmembrane-span enzyme, was used as a model to identify structural elements essential for targeting to the PM. When PMA2 and PMA4, representatives of the two main PMA subfamilies, were fused to green fluorescent protein (GFP), the chimeras were shown to be still functional and to be correctly and rapidly targeted to the PM in transgenic tobacco. By contrast, chimeric proteins containing various combinations of PMA transmembrane spanning domains accumulated in the Golgi apparatus and not in the PM and displayed slow traffic properties through the secretory pathway. Individual deletion of three of the four cytosolic domains did not prevent PM targeting, but deletion of the large loop or of its nucleotide binding domain resulted in GFP fluorescence accumulating exclusively in the endoplasmic reticulum. The results show that, at least for this polytopic protein, the PM is not the default pathway and that, in contrast with single-pass membrane proteins, cytosolic structural determinants are required for correct targeting.