Membrane binding and substrate access merge in cytochrome P450 7A1, a key enzyme in degradation of cholesterol.

To study membrane topology and mechanism for substrate specificity, we truncated residues 2-24 in microsomal cytochrome P450 7A1 (P450 7A1) and introduced conservative and nonconservative substitutions at positions 214-227. Heterologous expression in Escherichia coli was followed by investigation of the subcellular distribution of the mutant P450s and determination of the kinetic and substrate binding parameters for cholesterol. The results indicate that a hydrophobic region, comprising residues 214-227, forms a secondary site of attachment to the membrane in P450 7A1 in addition to the NH(2)-terminal signal-anchor sequence. There are two groups of residues at this enzyme-membrane interface. The first are those whose mutation results in more cytosolic P450 (Val-214, His-225, and Met-226). The second group are those whose mutation leads to more membrane-bound P450 (Phe-215, Leu-218, Ile-224, and Phe-227). In addition, the V214A, V214L, V214T, F215A, F215L, F215Y, L218I, L218V, V219T, and M226A mutants showed a 5-12-fold increased K(m) for cholesterol. The k(cat) of the V214A, V214L, V219T, and M226A mutants was increased up to 1.8-fold, and that of the V214T, F215A, F215L, F215Y, L218I, and L218V mutants was decreased 3-10.5-fold. Based on analysis of these mutations we suggest that cholesterol enters P450 7A1 through the membrane, and Val-214, Phe-215, and Leu-218 are the residues located near the point of cholesterol entry. The results provide an understanding of both the P450 7A1-membrane interactions and the mechanism for substrate specificity.     

Putative helix F contributes to regioselectivity of hydroxylation in mitochondrial cytochrome P450 27A1.

On the basis of alignment with structurally characterized cytochromes P450 (P450s), we have identified the putative F and G helices of mitochondrial P450s 27A1 and 11A. We introduced substitutions at Phe-207, Ile-211, and Phe-215 within putative helix F and at Trp-235 and Tyr-238 within putative helix G in P450 27A1 and compared wild type and mutants with respect to catalytic activity, product pattern, substrate binding, formation of hydrogen peroxide, and interaction with redox partner. Results indicate that the mutated residues are important for delivery of the correctly oriented substrate to the P450 active site. The I211K and F215K mutations, for example, affected the regioselectivity of P450 27A1-dependent hydroxylation reactions and conferred the P450 capacity to cleave the C-C bond of the substrate during the catalytic cycle. Studies of P450 11A1 indicate that Phe-202 has functions similar to those of its counterpart in P450 27A1 (Phe-215). We propose that putative helices F and G form the sides of the substrate-access channel, thus providing the additional mechanism to control regioselectivity of hydroxylation in mitochondrial P450s.     

The structure of residues 7-16 of the A alpha-chain of human fibrinogen bound to bovine thrombin at 2.3-A resolution.

The tetradecapeptide Ac-D-F-L-A-E-G-G-G-V-R-G-P-R-V-OMe, which mimics residues 7f-20f of the A alpha-chain of human fibrinogen, has been co-crystallized with bovine thrombin from ammonium sulfate solutions in space group P2(1) with unit cell dimensions of a = 83.0 A, b = 89.4 A, c = 99.3 A, and beta = 106.6 degrees. Three crystallographically independent complexes were located in the asymmetric unit by molecular replacement using the native bovine thrombin structure as a model. The standard crystallographic R-factor is 0.167 at 2.3-A resolution. Excellent electron density could be traced for the decapeptide, beginning with Asp-7f and ending with Arg-16f in the active site of thrombin; the remaining 4 residues, which have been cleaved from the tetradecapeptide at the Arg-16f/Gly-17f bond, are not seen. Residues 7f-11f at the NH2 terminus of the peptide form a single turn of alpha-helix that is connected by Gly-12f, which has a positive phi angle, to an extended chain containing residues 13f-16f. The major specific interactions between the peptide and thrombin are 1) a hydrophobic cage formed by residues Tyr-60A, Trp-60D, Leu-99, Ile-174, Trp-215, Leu-9f, Gly-13f, and Val-15f that surrounds Phe-8f; 2) a hydrogen bond linking Phe-8f NH to Lys-97 O;3) a salt link between Glu-11f and Arg-173; 4) two antiparallel beta-sheet hydrogen bonds between Gly-14f and Gly-216; and 5) the insertion of Arg-16f into the specificity pocket. Binding of the peptide is accompanied by a considerable shift in two of the loops near the active site relative to human D-phenyl-L-prolyl-L-arginyl chloromethyl ketone (PPACK)-thrombin.     

Structural determinants of the substrate and stereochemical specificity of phosphotriesterase

Bacterial phosphotriesterase (PTE) catalyzes the hydrolysis of a wide variety of organophosphate nerve agents and insecticides. Previous kinetic studies with a series of enantiomeric organophosphate triesters have shown that the wild type PTE generally prefers the S(P)-enantiomer over the corresponding R(P)-enantiomers by factors ranging from 1 to 90. The three-dimensional crystal structure of PTE with a bound substrate analogue has led to the identification of three hydrophobic binding pockets. To delineate the factors that govern the reactivity and stereoselectivity of PTE, the dimensions of these three subsites have been systematically altered by site-directed mutagenesis of Cys-59, Gly-60, Ser-61, Ile-106, Trp-131, Phe-132, His-254, His-257, Leu-271, Leu-303, Phe-306, Ser-308, Tyr-309, and Met-317. These studies have shown that substitution of Gly-60 with an alanine within the small subsite dramatically decreased k(cat) and k(cat)/K(a) for the R(P)-enantiomers, but had little influence on the kinetic constants for the S(P)-enantiomers of the chiral substrates. As a result, the chiral preference for the S(P)-enantiomers was greatly enhanced. For example, the value of k(cat)/K(a) with the mutant G60A for the S(P)-enantiomer of methyl phenyl p-nitrophenyl phosphate was 13000-fold greater than that for the corresponding R(P)-enantiomer. The mutation of I106, F132, or S308 to an alanine residue, which enlarges the small or leaving group subsites, caused a significant reduction in the enantiomeric preference for the S(P)-enantiomers, due to selective increases in the reaction rates for the R(P)-enantiomers. Enlargement of the large subsite by the construction of an H254A, H257A, L271A, or M317A mutant had a relatively small effect on k(cat)/K(a) for either the R(P)- or S(P)-enantiomers and thus had little effect on the overall stereoselectivity. These studies demonstrate that by modifying specific residues located within the active site of PTE, it is possible to dramatically alter the stereoselectivity and overall reactivity of the native enzyme toward chiral substrates.     

The conserved glycine-rich segment linking the N-terminal fusion peptide to the coiled coil of human T-cell leukemia virus type 1 transmembrane glycoprotein gp21 is a determinant of membrane fusion function

Retroviral transmembrane proteins (TMs) contain an N-terminal fusion peptide that initiates virus-cell membrane fusion. The fusion peptide is linked to the coiled-coil core through a conserved sequence that is often rich in glycines. We investigated the functional role of the glycine-rich segment, Met-326 to Ser-337, of the human T-cell leukemia virus type 1 (HTLV-1) TM, gp21, by alanine and proline scanning mutagenesis. Alanine substitution for the hydrophobic residue Ile-334 caused an approximately 90% reduction in cell-cell fusion activity without detectable effects on the lipid-mixing and pore formation phases of fusion. Alanine substitutions at other positions had smaller effects (Gly-329, Val-330, and Gly-332) or no effect on fusion function. Proline substitution for glycine residues inhibited cell-cell fusion function with position-dependent effects on the three phases of fusion. Retroviral glycoprotein fusion function thus appears to require flexibility within the glycine-rich segment and hydrophobic contacts mediated by this segment.     

The resonant recognition model (RRM) predicts amino acid residues in highly conserved regions of the hormone prolactin (PRL)

The resonant recognition model (RRM) is a model which treats the protein sequence as a discrete signal. It has been shown previously that certain periodicities (frequencies) in this signal characterise protein biological function. The RRM was employed to determine the characteristic frequencies of the hormone prolactin (PRL), and to identify amino acids ('hot spots') mostly contributing to these frequencies and thus proposed to mostly contribute to the biological function. The predicted 'hot spot' amino acids, Phe-19, Ser-26, Ser-33, Phe-37, Phe-40, Gly-47, Gly-49, Phe-50, Ser-61, Gly-129, Arg-176, Arg-177, Cys-191 and Arg-192 are found in the highly conserved amino-terminal and C-terminus regions of PRL. Our predictions agree with previous experimentally tested residues by site-direct mutagenesis and photoaffinity labelling.     

Prostaglandin endoperoxide synthase. The aspirin acetylation region.

Aspirin selectively acetylates Ser-530 of prostaglandin endoperoxide (PGH) synthase-1. This causes inactivation of the cyclooxygenase activity of the enzyme, but does not appreciably affect its peroxidase activity. Although the aspirin-acetylated enzyme is inactive, we found that PGH synthase-1 in which Ser-530 had been replaced with an alanine was catalytically active; accordingly, we proposed that aspirin inhibits cyclooxygenase activity by placing a larger than normal side chain at position 530 thereby interfering with arachidonate binding (DeWitt, D.L., El-Harith, E. A., Kraemer, S. A., Andrews, M. J., Yao, E. F., Armstrong, R. L., and Smith, W. L. (1990) J. Biol. Chem. 265, 5192-5198). As a further test of this hypothesis we have used site-directed mutagenesis and transient expression in cos-1 cells to prepare and characterize five additional substitutions of Ser-530. Consistent with our proposal, the presence of amino acids with bulky side chains at position 530 inhibited cyclooxygenase activity and decreased the apparent affinity of the enzyme for arachidonate. In related work, we characterized a series of mutant PGH synthases-1 having substitutions at residues adjoining Ser-530, including Phe-529, Leu-531, Lys-532, and Gly-533, in order to evaluate the contributions of each residue to cyclooxygenase catalysis. The most significant conclusion of this part of the study is that residues 529-533 all are important for the peroxidase activity as well as the cyclooxygenase activity of PGH synthase-1. Phe-529, in particular, was found to be critical for PGH synthase-1 structure and catalysis; some substitutions at this position led to the production of proteins lacking about 100 amino acids from their COOH termini.     

Probing the structure-activity relationship of Escherichia coli LT-A by site-directed mutagenesis

Computer analysis of the crystallographic structure of the A subunit of Escherichia coil heat-labile toxin (LT) was used to predict residues involved in NAD binding, catalysis and toxicity. Following site-directed mutagenesis, the mutants obtained could be divided into three groups. The first group contained fully assembled, non-toxic new molecules containing mutations of single amino acids such as Val-53 → Glu or Asp, Ser-63 → Lys, Val-97 → Lys, Tyr-104 → Lys or Asp, and Ser-14 → Lys or Glu. This group also included mutations in amino acids such as Arg-7, Glu-110 and Glu-112 that were already known to be important for enzymatic activity. The second group was formed by mutations that caused the collapse or prevented the assembly of the A subunit: Leu-41 → Phe, Ala-45 → Tyr or Glu, Val-53 → Tyr, Val-60 → Gly, Ser-68 → Pro, His-70 → Pro, Val-97 → Tyr and Ser-114 → Tyr. The third group contained those molecules that maintained a wild-type level of toxicity in spite of the mutations introduced: Arg-54 → Lys or Ala, Tyr-59 → Met, Ser-68 → Lys, Ala-72 → Arg, His or Asp and Arg-192 → Asn. The results provide a further understanding of the structure–function of the active site and new, non-toxic mutants that may be useful for the development of vaccines against diarrhoeal diseases.     

Isolation of four novel tachykinins from frog (Rana catesbeiana) brain and intestine

In a survey for unknown bioactive peptides in frog (Rana catesbeiana) brain and intestine, we isolated four novel peptides that exhibit potent stimulant effects on smooth muscle preparation of guinea pig ileum. By microsequencing and synthesis, these peptides were identified as Lys- Pro- Ser- Pro- Asp- Arg- Phe- Tyr- Gly- Leu- Met- NH2 (ranatachykinin A), Tyr- Lys- Ser- Asp- Ser- Phe- Tyr- Gly- Leu- Met- NH2 (ranatachykinin B), His- Asn- Pro- Ala- Ser- Phe- Ile- Gly- Leu- Met- NH2 (ranatachykinin C) and Lys- Pro- Ans- Pro- Glu- Arg- Phe- Tyr- Ala- Pro- Met- NH2 (ranatachykinin D). Ranatachykinin (RTK) A, B and C conserve the C- terminal sequence, Phe- X- Gly- Leu- Met- NH2, which is common to known members of the tachykinin family. On the other hand, RTK-D has a striking feature in its C-terminal sequence, Phe- Tyr- Ala- Pro- Met- NH2, which has never been found in other known tachykinins, and may constitute a new subclass in the tachykinin family.     

Identification of residues involved in a conformational change accompanying substitutions for glutamate-43 in staphylococcal nuclease

A recent paper from our laboratories [Hibler, D. W., Stolowich, N. J., Reynolds, M. A., Gerlt, J. A. Wilde, J. A., & Bolton, P. H. (1987) Biochemistry 26, 6278] described the generation of site-directed substitutions for the putative general base Glu-43 in the active site of Staphylococcal nuclease (SNase) and the use of 1H NMR spectroscopy to characterize the effect of the substitutions on the conformations of the mutant proteins. The replacements for Glu-43 (Asp, Gln, Asn, Ser, and Ala) both decreased the catalytic efficiency and changed the one- and two-dimensional NMR spectral properties of the mutant enzymes. We have prepared and studied the NMR spectral properties of several samples of deuteriated wild-type SNase that allow sequence-specific resonance assignments for several aromatic and aliphatic amino acid side chains that experience changes both in normal one-dimensional spectra and in two-dimensional NOESY spectra. Due to severe spectral congestion of resonances in the one- and two-dimensional spectra of protiated SNase, the assignments would have been difficult, if not impossible, to obtain without deuteriation of selected amino acids. The spectra we have obtained demonstrate that changes in NOE intensities involve a valine residue that is spatially adjacent to two phenylalanine residues; given the X-ray structure for SNase [Cotton, F. A., Hazen, E. E., & Legg, M. J. (1979) Proc. Natl. Acad. Sci. U.S.A. 76, 2551], these residues must be Val-74, Phe-34, and Phe-76. In addition, a leucine residue experiencing changes in NOE intensities spatially adjacent to Val-74 and Phe-34 can be assigned to Leu-25.(ABSTRACT TRUNCATED AT 250 WORDS)     

Structural insights into activation of phosphatidylinositol 4-kinase (Pik1) by yeast frequenin (Frq1)

Yeast frequenin (Frq1), a small N-myristoylated EF-hand protein, activates phosphatidylinositol 4-kinase Pik1. The NMR structure of Ca2+-bound Frq1 complexed to an N-terminal Pik1 fragment (residues 121-174) was determined. The Frq1 main chain is similar to that in free Frq1 and related proteins in the same branch of the calmodulin superfamily. The myristoyl group and first eight residues of Frq1 are solvent-exposed, and Ca2+ binds the second, third, and fourth EF-hands, which associate to create a groove with two pockets. The Pik1 peptide forms two helices (125-135 and 156-169) connected by a 20-residue loop. Side chains in the Pik1 N-terminal helix (Val-127, Ala-128, Val-131, Leu-132, and Leu-135) interact with solvent-exposed residues in the Frq1 C-terminal pocket (Leu-101, Trp-103, Val-125, Leu-138, Ile-152, and Leu-155); side chains in the Pik1 C-terminal helix (Ala-157, Ala-159, Leu-160, Val-161, Met-165, and Met-167) contact solvent-exposed residues in the Frq1 N-terminal pocket (Trp-30, Phe-34, Phe-48, Ile-51, Tyr-52, Phe-55, Phe-85, and Leu-89). This defined complex confirms that residues in Pik1 pinpointed as necessary for Frq1 binding by site-directed mutagenesis are indeed sufficient for binding. Removal of the Pik1 N-terminal region (residues 8-760) from its catalytic domain (residues 792-1066) abolishes lipid kinase activity, inconsistent with Frq1 binding simply relieving an autoinhibitory constraint. Deletion of the lipid kinase unique motif (residues 35-110) also eliminates Pik1 activity. In the complex, binding of Ca2+-bound Frq1 forces the Pik1 chain into a U-turn. Frq1 may activate Pik1 by facilitating membrane targeting via the exposed N-myristoyl group and by imposing a structural transition that promotes association of the lipid kinase unique motif with the kinase domain.     

Site-directed mutagenesis of the GDP binding domain of bacterial elongation factor Tu

The tertiary structure model of EF-Tu predicts that the amino acid sequence Val-Asp-His-Gly-Lys-Thr-Thr-Leu (residues 20-27) forms a pocket that binds the pyrophosphate group. To test this model we used site-directed mutagenesis to produce forms of EF-Tu altered in this region. The following mutations were constructed: Gly-20, Val-23, Glu-24, Ile-25, and Pro-27. Each protein was labeled with [35S]Met and was tested for its ability to interact with guanosine nucleotides and EF-Ts. The in vivo activity of each altered protein was tested by determining its ability to confer aurodox sensitivity to a resistant host. Mutations at residues 23, 24, 25, and 27 eliminated the ability of EF-Tu to interact with either guanosine nucleotides or EF-Ts in vitro, and these forms were also inactive in vivo. In contrast, the Gly-20 form was nearly as active as wild-type EF-Tu in vitro and in vivo. This mutation is theoretically equivalent to reversion of the Gly to Val transforming mutation of the cellular form of the ras gene product p21, a protein proposed to be structurally similar to EF-Tu in the GDP binding domain. In contrast to its effect in the ras gene, the Val to Gly conversion did not affect the endogenous GTPase of EF-Tu. We conclude that the tertiary structure model is correct in its assignment of the pyrophosphate binding site to residues 23-27; however, there are likely to be some significant differences between the configurations of the GTPases of EF-Tu and p21.     

Cholesterol binding to cytochrome P450 7A1, a key enzyme in bile acid biosynthesis

The conversion of cholesterol to 7alpha-hydroxycholesterol catalyzed by cytochrome P450 7A1 (CYP7A1) initiates the major pathway for cholesterol elimination in mammals. In the present work we focused on identification of determinants of the CYP7A1 substrate specificity inside the active site using a homology model with a novel P450-fold, site-directed mutagenesis, and substrate-binding and kinetic studies. Forty-one mutants, encompassing twenty-six amino acid residues, were generated and characterized, and of these, seven residues appear to determine cholesterol binding in the active site. In addition, four cholesterol derivatives were used as active site probes in the wild type and the seven mutant enzymes, and the spectral binding constants and products were analyzed. It was concluded that Asn288 in the I helix plays a key role in the P450-cholesterol contacts by hydrogen bonding to the steroid 3beta-hydroxyl, while Val280 and Ala284 are beside and the Trp283 is above the steroid nucleus orienting the cholesterol molecule. Leu360 and Ala358 between the K helix and the beta1-4 strand and Leu485 in the beta4 sheet-turn appear to define the size of the active site over the heme pyrrole ring A, thus limiting the orientation and size of the substrate at the steroid A ring. Additionally, the A358V mutant was found to form two new products, one being 7beta-hydroxycholesterol. Our data indicate that a tight fit of cholesterol in the enzyme active site is in part responsible for the high efficiency of cholesterol turnover by CYP7A1.     

Structural motif-based homology modeling of CYP27A1 and site-directed mutational analyses affecting vitamin D hydroxylation

Human CYP27A1 is a mitochondrial cytochrome P450, which is principally found in the liver and plays important roles in the biological activation of vitamin D(3) and in the biosynthesis of bile acids. We have applied a systematic analysis of hydrogen bonding patterns in 11 prokaryotic and mammalian CYP crystal structures to construct a homology-based model of CYP27A1. Docking of vitamin D(3) structures into the active site of this model identified potential substrate contact residues in the F-helix, the beta-3 sheet, and the beta-5 sheet. Site-directed mutagenesis and expression in COS-1 cells confirmed that these positions affect enzymatic activity, in some cases shifting metabolism of 1alpha-hydroxyvitamin D(3) to favor 25- or 27-hydroxylation. The results suggest that conserved hydrophobic residues in the beta-5 hairpin help define the shape of the substrate binding cavity and that this structure interacts with Phe-248 in the F-helix. Mutations directed toward the beta-3a strand suggested a possible heme-binding interaction centered on Asn-403 and a structural role for substrate contact residues Thr-402 and Ser-404.     

Structural determinants of ubiquitin-CXC chemokine receptor 4 interaction

Ubiquitin, a post-translational protein modifier inside the cell, functions as a CXC chemokine receptor (CXCR) 4 agonist outside the cell. However, the structural determinants of the interaction between extracellular ubiquitin and CXCR4 remain unknown. Utilizing C-terminal truncated ubiquitin and ubiquitin mutants, in which surface residues that are known to interact with ubiquitin binding domains in interacting proteins are mutated (Phe-4, Leu-8, Ile-44, Asp-58, Val-70), we provide evidence that the ubiquitin-CXCR4 interaction follows a two-site binding mechanism in which the hydrophobic surfaces surrounding Phe-4 and Val-70 are important for receptor binding, whereas the flexible C terminus facilitates receptor activation. Based on these findings and the available crystal structures, we then modeled the ubiquitin-CXCR4 interface with the RosettaDock software followed by small manual adjustments, which were guided by charge complementarity and anticipation of a conformational switch of CXCR4 upon activation. This model suggests three residues of CXCR4 (Phe-29, Phe-189, Lys-271) as potential interaction sites. Binding studies with HEK293 cells overexpressing wild type and CXCR4 after site-directed mutagenesis confirm that these residues are important for ubiquitin binding but that they do not contribute to the binding of stromal cell-derived factor 1α. Our findings suggest that the structural determinants of the CXCR4 agonist activity of ubiquitin mimic the typical structure-function relationship of chemokines. Furthermore, we provide evidence for separate and specific ligand binding sites on CXCR4. As exogenous ubiquitin has been shown to possess therapeutic potential, our findings are expected to facilitate the structure-based design of new compounds with ubiquitin-mimetic actions on CXCR4.     

Molecular basis for the omega-regiospecificity of the CYP4A2 and CYP4A3 fatty acid hydroxylases

The CYP4A fatty acid omega-hydroxylases are involved in important physiological processes such as the regulation of vascular pressure. A previous study of chimeras of the rat CYP4A2 and CYP4A3 enzymes established that the regiochemistry of fatty acid hydroxylation is determined by the first 119 amino acid residues (Hoch, U., Zhang. Z. P., Kroetz, D. L., and Ortiz de Montellano, P. R. (2000) Arch. Biochem. Biophys. 373, 63-71). The role of the individual amino acid differences in this region has therefore been examined by site-specific mutagenesis to determine which residues actually control the omega- versus (omega-1)-regiospecificity. The results indicate that regiospecificity is controlled by the presence or absence of a three-residue insert (Ser-114, Gly-115, Ile-116) in CYP4A3 and by the residue at position 119 (CYP4A3 numbering). Furthermore, analysis of the absolute stereochemistry of the 11-hydroxylauric acid product indicates that this stereochemistry is not very sensitive to changes in the residues that line the substrate access channel. These results define a model of the specificity determinants of an important class of cytochrome P450 enzymes.     

Structural and thermodynamic consequences of 1-(4-chlorophenyl)imidazole binding to cytochrome P450 2B4

The crystal structure of P450 2B4 bound with 1-(4-chlorophenyl)imidazole (1-CPI) has been determined to delineate the structural basis for the observed differences in binding affinity and thermodynamics relative to 4-(4-chlorophenyl)imidazole (4-CPI). Compared with the previously reported 4-CPI complex, there is a shift in the 1-CPI complex of the protein backbone in helices F and I, repositioning the side chains of Phe-206, Phe-297, and Glu-301, and leading to significant reshaping of the active site. Phe-206 and Phe-297 exchange positions, with Phe-206 becoming a ligand-contact residue, while Glu-301, rather than hydrogen bonding to the ligand, flips away from the active site and interacts with His-172. As a result the active site volume expands from 200 A3 in the 4-CPI complex to 280 A3 in the 1-CPI complex. Based on the two structures, it was predicted that a Phe-206-->Ala substitution would alter 1-CPI but not 4-CPI binding. Isothermal titration calorimetry experiments indicated that this substitution had no effect on the thermodynamic signature of 4-CPI binding to 2B4. In contrast, relative to wild-type 1-CPI binding to F206A showed significantly less favorable entropy but more favorable enthalpy. This result is consistent with loss of the aromatic side chain and possible ordering of water molecules, now able to interact with Glu-301 and exposed residues in the I-helix. Hence, thermodynamic measurements support the active site rearrangement observed in the crystal structure of the 1-CPI complex and illustrate the malleability of the active site with the fine-tuning of residue orientations and thermodynamic signatures.     

Membrane topography of ColE1 gene products: the immunity protein.

The topography of the colicin E1 immunity (Imm) protein was determined from the positions of TnphoA and complementary lacZ fusions relative to the three long hydrophobic segments of the protein and site-directed substitution of charged for nonpolar residues in the proposed membrane-spanning segments. Inactivation of the Imm protein function required substitution and insertion of two such charges. It was concluded that the 113-residue colicin E1 Imm protein folds in the membrane as three trans-membrane alpha-helices, with the NH2 and COOH termini on the cytoplasmic and periplasmic sides of the membrane, respectively. The approximate spans of the three helices are Asn-9 to Ser-28, Ile-43 to Phe-62, and Leu-84 to Leu-104. An extrinsic highly charged segment, Lys-66 to Lys-74, containing seven charges in nine residues, extends into the cytoplasmic domain. The specificity of the colicin E1 Imm protein for interaction with the translocation apparatus and the colicin E1 ion channel is proposed to reside in its peripheral segments exposed on the surface of the inner membrane. These regions include the highly charged segment Lys-66 to Lys-83 (loop 2) and the short (approximately eight-residue) NH2 terminus on the cytoplasmic side, and Glu-29 to Val-44 (loop 1) and the COOH-terminal segment Gly-105 to Asn-113 on the periplasmic side.