Portal de Periódicos da CAPES

Phospholipid-assisted protein folding: phosphatidylethanolamine is required at a late step of the conformational maturation of the polytopic membrane protein lactose permease

1998; Springer Nature; Volume: 17; Issue: 18 Linguagem: Inglês

10.1093/emboj/17.18.5255

1460-2075

Mikhail Bogdanov, William Dowhan,

Bacterial Genetics and Biotechnology

Article15 September 1998free access Phospholipid-assisted protein folding: phosphatidylethanolamine is required at a late step of the conformational maturation of the polytopic membrane protein lactose permease Mikhail Bogdanov Mikhail Bogdanov Department of Biochemistry and Molecular Biology, University of Texas-Houston, Medical School, Houston, TX, 77225 USA Search for more papers by this author William Dowhan William Dowhan Department of Biochemistry and Molecular Biology, University of Texas-Houston, Medical School, Houston, TX, 77225 USA Search for more papers by this author Mikhail Bogdanov Mikhail Bogdanov Department of Biochemistry and Molecular Biology, University of Texas-Houston, Medical School, Houston, TX, 77225 USA Search for more papers by this author William Dowhan William Dowhan Department of Biochemistry and Molecular Biology, University of Texas-Houston, Medical School, Houston, TX, 77225 USA Search for more papers by this author Author Information Mikhail Bogdanov1 and William Dowhan1 1Department of Biochemistry and Molecular Biology, University of Texas-Houston, Medical School, Houston, TX, 77225 USA The EMBO Journal (1998)17:5255-5264https://doi.org/10.1093/emboj/17.18.5255 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Previously we presented evidence that phosphatidylethanolamine (PE) acts as a molecular chaperone in the folding of the polytopic membrane protein lactose permease (LacY) of Escherichia coli. Here we provide more definitive evidence supporting the chaperone properties of PE. Membrane insertion of LacY prevents its irreversible aggregation, and PE participates in a late step of conformational maturation. The temporal requirement for PE was demonstrated in vitro using a coupled translation–membrane insertion assay that allowed the separation of membrane insertion from phospholipid-assisted folding. LacY was folded properly, as assessed by recognition with conformation-specific monoclonal antibodies, when synthesized in the presence of PE-containing inside-out membrane vesicles (IOVs) or in the presence of IOVs initially lacking PE but supplemented with PE synthesized in vitro either co- or post-translationally. The presence of IOVs lacking PE and containing anionic phospholipids or no addition of IOVs resulted in misfolded or aggregated LacY, respectively. Therefore, critical folding steps occur after membrane insertion dependent on the interaction of LacY with PE to prevent illicit interactions which lead to misfolding of LacY. Introduction Newly synthesized polypeptides fold into their native structures via partially folded intermediates that display hydrophobic regions transiently on their surfaces. These regions can interact with one another, resulting in aggregates that may become insoluble. The high concentrations of interacting protein surfaces inside cells result in a need for molecular chaperones (Hendrick and Hartl, 1995; Ellis and Hartl, 1996; Hartl, 1996) to prevent incorrect interactions between these surfaces. A well-established class of such molecular chaperones is the proteins that recognize and bind transiently to these surfaces, thus preventing their premature or inappropriate interaction which lead to misfolding and self-aggregation. Molecular chaperones can bind to proteins either co-translationally (Fedorov and Baldwin, 1997) or post-translationally (Hendrick and Hartl, 1993), and thereby assist in correct folding at multiple stages of protein assembly. Although there is significant progress in our understanding of protein folding for water-soluble proteins, there is little information available for the folding of multispanning polytopic integral membrane proteins. Large hydrophobic membrane proteins in Escherichia coli are synthesized in the cytoplasm and may interact with cytoplasmic molecular chaperones (Bochkareva et al., 1996) which prevent aggregation. These proteins are delivered to the membrane where the final stages of folding take place. During membrane assembly, integral membrane proteins must interact with other proteins within the membrane, the phospholipid bilayer as solvent and individual phospholipid molecules either as permanent ligands or transiently during folding, which collectively results in the correct folding of a native membrane protein. The role of the phospholipid bilayer as an important structure-forming environment has been established (Selinsky, 1992; In't Veld et al., 1993; Marsh, 1993); however, the role of individual phospholipid molecules as specific participants in the process of protein assembly has been largely ignored. Can phospholipids act as specific permanent or transient ligands during the assembly of a membrane protein to either prevent unproductive folding or direct the proper folding of a membrane protein? Conceptually, there is considerable similarity between interaction of protein folding intermediates with either amphipathic phospholipid molecules or protein molecular chaperones (Zardeneta and Horowitz, 1992). What is lacking is a large body of evidence supporting such a role for phospholipids. In order to determine the role that individual phospholipids play in the folding of membrane proteins in E.coli, we have employed LacY (lactose permease) as a model for studying the folding and assembly of such proteins. LacY is particularly suited for this purpose because of the extensive molecular genetic and structural information available (Varela and Wilson, 1996; Kaback et al., 1997) as well as its homology with the major facilitator superfamily of transport proteins (Marger and Saier, 1993). We have presented experimental evidence that the phospholipid phosphatidylethanolamine (PE) fulfills the minimum requirements to be a non-protein molecular chaperone in the folding of LacY (Bogdanov et al., 1996). PE corrects in vitro a LacY folding defect caused by in vivo assembly in PE-deficient membranes and, once properly folded, PE is no longer required to maintain the 'native' conformation. These results were obtained using a novel blotting technique (Eastern–Western) employing phospholipid during renaturation of proteins on solid supports in the standard Western blotting procedure. Although the refolding of denatured full-length protein provides important information on the mechanism of protein folding, it is not an ideal model, since the refolding of the protein starts from the complete polypeptide chain, while the folding of the nascent peptide in vivo begins during translation (Fedorov and Baldwin, 1997). In addition, folding steps or intermediates observed in vitro may not lie along the in vivo folding pathway. Molecular chaperones have been shown to interact differently with denatured proteins than with their newly synthesized counterparts (Hartl, 1996). In this study, we used translation coupled with membrane insertion/assembly (Ahrem et al., 1989; Bochkareva et al., 1996; van Klompenburg et al., 1997) of LacY in vitro either during or before cell-free biosynthesis of phospholipid (Sparrow et al., 1984) to answer the following questions. Can an integral membrane protein, such as LacY, be folded and assembled into a native-like conformation in vitro? Are membrane phospholipid and/or their composition a prerequisite for correct membrane folding? Are protein–lipid interactions non-specific or selective with respect to phospholipid type? Do specific phospholipids affect protein folding at an early or late stage of conformational maturation within the membrane? Results In vivo basis for design of a novel assay for in vitro assembly of LacY Monoclonal antibodies (mAbs) have been used successfully to monitor conformational changes and detect different immunoreactive folding intermediates of proteins (Goldberg, 1991). Since mAbs generally are directed against limited regions of proteins, they are particularly useful in detection of the appearance of local native-like structures along the folding pathway of proteins. Most useful are conformation-sensitive mAbs which have been used successfully in detecting the folding intermediates formed by coupled transcription–translation synthesis of E.coli tryptophan synthase β subunit (Fedorov et al., 1992). Based on past success in utilizing mAbs to study protein folding, we employed two such antibodies directed against distinct epitopes of native LacY. mAb4B1 inhibits all reactions involving net H+ translocation (active transport) during LacY function (Frillingos and Kaback, 1996; Sun et al., 1996) and recognizes a continuous epitope formed by extramembrane periplasmic domain P7 (Frillingos and Kaback, 1996; Sun et al., 1996, 1997b) (Figure 1) only when LacY is folded in the presence of PE (Bogdanov et al., 1996). LacY assembled in the absence of PE is also severely compromised in active but not facilitated transport of lactose (Bogdanov and Dowhan, 1995). mAb4B11 recognizes a discontinuous epitope composed of the extramembrane cytoplasmic domains C8 and C10 (Sun et al., 1997a) and is thus dependent on proper folding of the permease to bring these two domains into close proximity. Therefore, we have used recognition of LacY by these antibodies to establish the close association of cytoplasmic domains C8 and C10, which are located on the same side of the membrane, and the formation of native-like structure of periplasmic domain P7. Figure 1.Topological model of LacY. The 12 TM-spanning domains are indicated by Roman numerals. Periplasmic domain P7 is recognized by mAb4B1, and the amino acids required for recognition are numbered (Sun et al., 1997b). mAb4B11 recognizes a discontinuous epitope formed by the two cytoplasmic domains C8 and C10 (Sun et al., 1997a). Download figure Download PowerPoint When LacY is highly overproduced by expression from the T7 RNA polymerase-driven promoter, it is found integrated into the membrane as a monomer as well as both in the cytoplasm and in association with the membrane in a non-integrated form as dimers and trimers. These 'soluble' cytoplasmic and non-integrated membrane-associated forms apparently lack tightly bound lipid (Roepe and Kaback, 1989). The latter form can be 'stripped' from the membrane by urea while the fully assembled and membrane-integrated monomeric form cannot. Using a polyclonal antibody (pAb) which can detect all forms of LacY described above, we detected all three forms of LacY (i.e. monomeric, dimeric and trimeric) in the membrane fraction (Figure 2, lane 1), but the supernatant fraction (lane 2) only contained dimers and trimers and no monomers. Washing the membranes displayed in lane 1 with urea removed the aggregated forms (lane 4), but did not release the monomeric form (lane 3). In Western blot analysis using the pAb, all these forms of LacY (lanes 1–4) were recognized in all fractions. However, mAb4B1 only recognized the tightly membrane-associated 33 kDa monomer of LacY (lanes 5–8), indicating that the loosely membrane-associated and soluble aggregates (lanes 2 and 4, and 6 and 8) do not form the native continuous epitope recognized by this antibody. Identical results were seen for lanes 5–8 with mAb4B11 (data not shown), indicating that the discontinuous epitope recognized by this antibody is only present in membrane-integrated LacY. Unlike mAb4B1 (Bogdanov et al., 1996), mAb4B11 recognizes LacY assembled in both PE-containing and PE-deficient cells (lanes 9 and 10, respectively). Therefore, the formation of epitope 4B1 not only requires expression of LacY in PE-containing cells but the integration of LacY into the membrane where association with PE must occur. In addition, integration into the membrane is required for the proper folding of domains C8 and C10 (defining epitope 4B11) but in a PE-independent manner. These findings are consistent with different conformations of LacY depending on whether it is embedded in the lipid bilayer of the membrane (either with or without PE) or present in the lipid-free cytoplasm. Based on these results for in vivo expressed LacY, these two conformation-specific antibodies were used to determine whether membranes with and without PE are required co- or post-translationally for the proper assembly of LacY in vitro. Figure 2.Antibody analysis of LacY expressed and assembled in vivo. Synthesis of LacY from plasmid pT7-5 was induced from the T7 RNA polymerase-dependent strain JF618/pT7-5 (lanes 1–8) or from lacOP in strains AD93/pDD72/pT7-5 (+PE) (lane 9) and AD93/pT7-5 (−PE) (lane 10). Crude cell lysates were separated by centrifugation (150 000 gmax for 45 min) into a membrane (lane 1) and supernatant (lane 2) fraction. The membrane fraction (lane 1) was suspended in 5 M urea and separated again into a pellet (lane 3) and supernatant (lane 4) fraction. In lanes 9 and 10, only IOVs prepared as described in Materials and methods were used. Each fraction was subjected to SDS–PAGE, electroblotted and immunostained either with pAb (lanes 1–4), mAb4B1 (lanes 5–8) or mAb4B11 (lanes 9–10). Download figure Download PowerPoint In vitro translation and insertion assay of LacY A translation system for the in vitro expression of LacY either in the presence or absence of inside-out membrane vesicles (IOVs) was developed using a plasmid-borne copy of the lacY gene under T7 promoter control. In order to control the make up and PE content of added IOVs, membrane- and phosphatidylserine (PS) synthase-free S30 fractions were prepared, as described in Materials and methods, from E.coli strain AD905 (pss93::kan degP) carrying the null allele of the pssA gene. This strain cannot synthesize PE due to the lack of PS synthase (DeChavigny et al., 1991) which in wild-type cell lysates is almost exclusively bound to ribosomes (Dutt and Dowhan, 1977). Addition of an S30 fraction from wild-type cells containing PS synthase would result in the formation of PE in the IOVs which contain PS decarboxylase (Dowhan, 1997). The IOVs tested in these experiments were prepared either from PE-containing [strain AD93 (pss93::kan) carrying plasmid pDD72 (pss+)] or PE-deficient (strain AD93 lacking plasmid pDD72) cells grown in the absence of isopropyl-β-D-thiogalactopyranoside (IPTG), which resulted in no detectable LacY, as previously described (Bogdanov and Dowhan, 1995). LacY synthesis was initiated from the transcript of plasmid pT7-5 (carrying the lacY gene) in the presence of the above S30 fraction. After 5 min, the reaction mixture was either supplemented (co-translational membrane assembly) or not supplemented with the indicated IOVs. Synthesis was continued for an additional 60 min, followed by analysis of the translation–membrane insertion mix by SDS–PAGE and immunoblotting either with pAb, mAb4B11 or mAb4B1 (Figure 3) Figure 3.Assembly of in vitro synthesized LacY into IOVs. Transcription of the lacY gene using the T7 RNA polymerase-dependent promoter of plasmid pT7-5 was followed by in vitro translation using a membrane- and PS synthase-free S30 extract either in the absence (lanes 1 and 6) or in the presence of IOVs (lanes 2–5 and 7–14) prepared either from PE-containing (+PE, AD93/pDD72) or PE-deficient (−PE, AD93) cells grown in the absence of IPTG. Each reaction mixture was analyzed by SDS–PAGE followed by immunodetection with either pAb, mAb4B11 or mAb4B1 as indicated. Download figure Download PowerPoint If LacY was synthesized using an S30 extract without co-translational addition of IOVs, then no monomeric form was detected and only multimeric forms were detected by the site-directed pAb (Figure 3, lane 1 versus lanes 2–5). If IOVs were present co-translationally, then a diffuse immunoreactive band was also observed in the 31–33 kDa range that corresponds to lactose permease (Dunten et al., 1993); the basis for the diffuse nature of these bands is discussed below. An apparently stronger signal was detected for monomeric LacY assembled in vitro in PE-deficient (lanes 4 and 5) IOVs versus PE-containing IOVs (lanes 2 and 3), as we had observed previously for LacY assembled in vivo in the absence or presence of PE (Bogdanov et al., 1996). The last five residues in the C-terminal portion of transmembrane (TM) helix XII are necessary for proper folding of LacY (Roepe et al., 1989). The pAb used here and mAb4A1OR share a common overlapping epitope in the C-terminus of LacY. In the latter case, the mAb recognition of LacY on Western blots is sensitive to conformational changes in the C-terminus brought about by a mutation in LacY resulting in energy uncoupling of lactose transport (Herzlinger et al., 1985). LacY assembled in PE-deficient membranes is uncoupled similarly (Bogdanov and Dowhan, 1995), which might explain the differences in sensitivity of LacY to pAb dependent on the presence or absence of PE during its assembly. More important was the observation that mAb4B11 (lanes 6–10) only recognized the monomeric form of LacY, as observed for in vivo assembly (Figure 2), with the same apparent sensitivity independent of the presence of PE. This supports a correct in vitro co-translational membrane assembly of LacY at least with respect to domains C8 and C10 independent of the presence of PE during assembly. Site-directed pAb was not suitable for comparing LacY insertion yields between PE-containing and PE-deficient IOVs (Figure 3, lanes 2–5). However, mAb4B11 can be used as an immunological probe to discriminate between inserted and uninserted LacY molecules as well as an indicator of comparable membrane insertion efficiency of LacY whether or not PE is present. In this co-translational assay, 15–18% of the in vitro synthesized LacY was fully integrated into IOVs from wild-type cells as judged by quantitative Western blot analysis (Bogdanov and Dowhan, 1995). pAb and 35S-labeled protein A were used to estimate the percentage of integrated LacY (lanes 2) relative to total LacY (lane 1) (data not shown). Direct binding studies and in vitro refolding experiments had provided independent evidence for the preferential recognition by conformation-sensitive mAb4B1 of the P7 periplasmic domain of LacY assembled in vivo in the cells containing PE (Bogdanov et al., 1996). A similar dependence on PE was observed in vitro when mAb4B1 recognition of LacY co-translationally assembled in PE-containing IOVs (Figure 3, lanes 11 and 12) was compared with LacY assembled into PE-deficient IOVs (Figure 3, lanes 13 and 14). Probing with mAb4B11 after removal of the mAb4B1 verified the presence of the same amount of LacY in lanes 11–14 (data not shown). Therefore, these results confirm the in vivo assembly data indicating the dependence on PE for the formation of the native conformation of epitope 4B1, and also indicate that the insertion efficiency of LacY into the membrane is not dependent on PE. In our experiments, LacY displayed diffuse multiple species with different mobilities. These properties of LacY have been seen numerous times (Kaback, 1986), but the high salt content of the in vitro system resulted in even broader bands. Since LacY maintains a high degree of secondary structure in SDS (Bogdanov et al., 1996) and displays an aberrant mobility on SDS gels for a protein of Mr = 46 000 (Kaback, 1986), the multiple bands most likely represent different conformations of LacY due to incomplete denaturation under the conditions of SDS–PAGE. PE corrects co-translationally the misfolding of LacY assembled in vitro in IOV without PE We next wished to determined whether addition of PE to LacY assembled in vitro in PE-deficient IOVs could result in the formation of native epitope 4B1. Previously we had shown that addition of a dispersion of PE to LacY assembled in membranes in vivo in the absence of PE prior to solubilization with SDS did not result in recognition of LacY by mAb4B1 after Western blot analysis (Bogdanov et al., 1996). Therefore, we employed a cell-free system of phospholipid biosynthesis that utilizes intact inner membranes of E.coli (Sparrow et al., 1984) which should reflect more accurately in vivo conditions for delivery of PE to membranes. We took advantage of the fact that the absence of PE in PE-deficient cells results in an elevated level of acidic phospholipids including phosphatidic acid (PA) of up to 8% of total phospholipid composition. As previously shown (Sparrow et al., 1984), the addition of CTP and either L-serine or sn-glycerol-3-phosphate to PA-enriched E.coli membranes results in the formation of PE or phosphatidylglycerol (PG), respectively. Membranes lacking PS synthase, as is the case for IOVs from PE-deficient mutants, can be reconstituted by the addition of partially purified PS synthase which normally is associated with ribosomes in cell lysates of wild-type cells (Dutt and Dowhan, 1977; Louie et al., 1986). A membrane-free PS synthase-enriched fraction (PSS) (see Materials and methods) converted nearly all of the PA (pre-labeled with 32PO4 by in vitro exchange reaction with radioactive ATP) in IOVs lacking PE to PE in 1 h, dependent on the addition of CTP and L-serine (Figure 4); E.coli membranes contain PS decarboxylase which is necessary for this conversion (Dowhan, 1997). These membranes also convert PA to PG and cardiolipin (CL) if CTP and sn-glycerol-3-phosphate are added (data not shown). Figure 4.Cell-free phospholipid biosynthesis using IOVs isolated from PE-deficient cells. IOVs derived from strain AD93 lacking PE but containing [32P]PA were supplied with PS synthase-enriched fraction, CTP and dCTP, and serine as indicated. After 1 h of incubation, the lipids were isolated and analyzed by TLC. Download figure Download PowerPoint LacY was expressed in vitro co-translationally as described in Figure 3 either in the presence of PE-deficient (−PE) IOVs, without or with in vitro synthesis of PE (Figure 5, lanes 1 and 3, respectively), or in the presence of PE-containing (+PE) IOVs (Figure 5, lanes 4). After Western blot analysis, monomeric LacY was recognized by mAb4B1 only from membranes containing PE originally or from IOV supplied with PE by in vitro biosynthesis (lanes 4 and 3, respectively); no significant detection of LacY assembled in PE-deficient IOVs was observed (lane 1). Coupling in vitro biosynthesis of either zwitterionic phospholipids (PE and PS) or anionic phospholipids (PG and CL) with assembly of LacY resulted in no change in recognition of LacY assembled in IOVs previously containing PE (data not shown). Furthermore, in vitro biosynthesis of anionic phospholipids (lane 2) in IOVs originally lacking PE did not result in the recognition of monomeric LacY by mAb4B1. Addition of all components necessary for in vitro PE synthesis except PSS as a control showed no appearance of the native epitope 4B1 (Figure 5, lane 5). Finally the use of a PSS-like fraction (PSS*) lacking PS synthase (Figure 5, lane 6) or the absence of added serine and CTP (Figure 5, lane 7 versus 8) resulted in no LacY detectable by mAb4B1. These results confirm the direct dependence on PE biosynthesis and the absence of PE-containing membranes in the PSS fraction. When mAb4B1 was stripped from the nitrocellulose sheet and the sheet probed again with mAb4B11, the same amount of LacY was detected in all cases (data not shown), indicating that PE is not a determinant for the yield of insertion. Therefore, when in vitro synthesized LacY was inserted into PE-containing IOVs or assembled in PE-deficient membranes during co-synthesis in vitro of PS and PE (but not PG and CL), the permease could be recognized by mAb4B1, indicating the proper folding of domain P7. Figure 5.Assembly of in vitro synthesized LacY into IOVs coupled with in vitro synthesis of phospholipid. A membrane- and PS synthase-free S30 extract from strain AD905 and membrane-free PS synthase-enriched fraction (PSS) were used in this experiment. IOVs made from either PE-containing (+PE) or PE-deficient (−PE) cells (as in Figure 3) were added co-translationally to the in vitro LacY expression system described in Figure 3. Each reaction mixture was supplemented co-translationally with components necessary for in vitro PE biosynthesis (see Figure 4) or PG biosynthesis (sn-glycerol-3-phosphate, G-3-P) as indicated. For the results in lanes 7 and 8, the transcription reaction was stopped by using addition of LiCl to remove CTP prior to initiation of translation without or with PE synthesis. All samples were analyzed as described in Figure 3, using mAb4B1. Download figure Download PowerPoint PE corrects post-translationally the misfolding of LacY assembled either in vitro or in vivo without PE In vitro protein synthesis was carried out in the presence of IOVs either with (Figure 6, lanes 1–2) or without (lanes 3–4) PE until membranes were saturated with newly made LacY before in vitro PE synthesis was initiated. mAb4B1 recognized monomeric LacY assembled in IOVs containing PE (lane 1) but not in those lacking PE (lanes 3–4). Introduction of PE by in vitro synthesis had no effect on the level of mAb4B1 recognition of LacY assembled in IOVs containing PE (lane 2), but had a dramatic effect on recognition of LacY from IOVs originally lacking PE (lanes 5 and 6). Similar effects of in vitro PE synthesis on the formation of native epitope 4B1 were observed using LacY assembled initially in vivo in either PE-containing (Figure 7, lanes 1 and 2 versus lanes 3 and 4, no effect) or PE-deficient (Figure 7, lanes 5 and 6 versus 7 and 8, large effect) membranes. When mAb4B1 was stripped from the sheet and the sheet probed again with mAb4B11, the same amount of LacY was detected in all lanes in Figures 6 and 7 (data not shown), indicating that membrane insertion was not dependent on PE. Figure 6.Post-translational reconstitution of in vitro assembled LacY with in vitro synthesized PE. LacY was inserted co-translationally into PE-containing (+PE) or PE-deficient (−PE) IOVs as described in Figure 3. After 60 min, the reaction mixtures were supplemented as indicated (see Figure 5) and phospholipid synthesis was continued for 60 min. Samples were analyzed using mAb4B1. Download figure Download PowerPoint Figure 7.Reconstitution of in vivo assembled LacY with in vitro synthesized PE. IOVs in which LacY had been expressed in vivo were isolated from PE-containing (+PE) and PE-deficient (−PE) cells. The indicated supplements were added as described in Figure 6 for in vitro phospholipid biosynthesis, and samples were analyzed using mAb4B1; duplicate samples are shown. Download figure Download PowerPoint Therefore, PE can bring about the proper folding of LacY first assembled either in vivo or in vitro in membranes lacking PE even when added via cell-free biosynthesis after assembly into the membrane bilayer. These results suggest that PE can promote the transition of fully integrated LacY during a late step in assembly from a state approximating its final tertiary structure to the native state. PE acts as molecular chaperone In order to classify a biological molecule as a molecular chaperone, the molecule must assist protein folding through a transient non-covalent interaction and not be required to maintain the final native structure. The possibility that residual PE is bound to LacY from PE-containing cells after Western blotting analysis, and thereby is essential to maintain the conformation of epitope 4B1, was re-investigated in a more quantitative manner than previously reported (Bogdanov et al., 1996). Strain AD93/pDD72/pT7-5 induced by IPTG for LacY expression was labeled by growth in the presence of 32Pi (10 μCi/ml) in minimal media. Total membranes were prepared, 25 μg of membrane protein were subjected to SDS–PAGE and the protein was transferred to nitrocellulose sheets by electroblotting. Beta imaging (90 min) using an Instant Imager® and autoradiography (12 h) of the nitrocellulose sheet detected no radioactivity associated either with the position of LacY or anywhere on the nitrocellulose sheet. Phospholipids were extracted from the cells subjected to Western blot analysis, and serial dilutions were separated by thin-layer chromatography (TLC) and subjected to autoradiography and beta imaging for the same times as indicated above to determine the minimum radioactivity detectable with these methods. Beta imaging and autoradiography were easily able to detect 10 c.p.m. (as determined by direct scintillation counting of lipids eluted from plates) of [32P]PE in 90 min and 12 h, respectively (data not shown). The specific activity of PE (based on chemical analysis of phosphate present) was 25 000 c.p.m./nmol. These same membranes (5–25 μg of protein) were again subjected to SDS–PAGE and the protein transferred to nitrocellulose sheets by electroblotting. In the same experiment, known amounts of purified LacY were treated in a similar manner over a range similar to the amount present in the membrane samples. These blots were subjected to Western blot analysis using mAb4A10R that recognizes the C-terminus of LacY independently of structure (Carrasco et al., 1982). Treatment with [35S]protein A was used to quantify the amount of LacY in the membrane samples relative to the known amounts of pure LacY (Bogdanov and Dowhan, 1995). Both homogeneous LacY and the LacY in the membrane samples showed a linear relationship between amount of sample analyzed and radiolabeled protein A bound. Based on this analysis, 5% of the membrane protein was LacY, which is consistent with estimates of 3–5% previously reported (Newman et al., 1981) (H.R.Kaback, personal communication). Thus the 25 μg of membrane protein subjected to SDS–PAGE and Western blot analysis co