The peptidoglycan-binding protein FimV promotes assembly of the Pseudomonas aeruginosa type IV pilus secretin.

The Pseudomonas aeruginosa inner membrane protein FimV is among several proteins of unknown function required for type IV pilus-mediated twitching motility, arising from extension and retraction of pili from their site of assembly in the inner membrane. The pili transit the periplasm and peptidoglycan (PG) layer, ultimately exiting the cell through the PilQ secretin. Although fimV mutants are nonmotile, they are susceptible to killing by pilus-specific bacteriophage, a hallmark of retractable surface pili. Here we show that levels of recoverable surface pili were markedly decreased in fimV pilT retraction-deficient mutants compared with levels in the pilT control, demonstrating that FimV acts at the level of pilus assembly. Levels of inner membrane assembly subcomplex proteins PilM/N/O/P were decreased in fimV mutants, but supplementation of these components in trans did not restore pilus assembly or motility. Loss of FimV dramatically reduced the levels of the PilQ secretin multimer through which pili exit the cell, in part due to decreased levels of PilQ monomers, while PilF pilotin levels were unchanged. Expression of pilQ in trans in the wild type or fimV mutants increased total PilQ monomer levels but did not alter secretin multimer levels or motility. PG pulldown assays showed that the N terminus of FimV bound PG in a LysM motif-dependent manner, and a mutant with an in-frame chromosomal deletion of the LysM motif had reduced motility, secretin levels, and surface piliation. Together, our data show that FimV's role in pilus assembly is to promote secretin formation and that this function depends upon its PG-binding domain.

Positional dependence of non-native polar mutations on folding of CFTR helical hairpins.

Mutations of the cystic fibrosis transmembrane conductance regulator (CFTR) cause CF disease by altering the biosynthesis, maturation, folding and ion conductance of this protein. Our laboratory has focused on expression and structural analysis of the CFTR transmembrane (TM) domains using two-TM segments (i.e., helix-loop-helix constructs) which we term 'helical hairpins'; these represent the minimal model of tertiary contacts between two helices in a membrane. Previous studies on a library of TM3/4 hairpins of the first CFTR TM domain suggested that introduction of non-native polar residues into TM4 can compromise CFTR function through side chain-side chain H-bonding interactions with native Q207 in TM3 [Choi, M. Y., Cardarelli, L., Therien, A. G., and Deber, C. M. Non-native interhelical hydrogen bonds in the cystic fibrosis transmembrane conductance regulator domain modulated by polar mutations, Biochemistry 43 (2004) 8077-8083]. In the present work, we combine gel shift assays with a series of NMR experiments for comparative structural characterization of the wild type TM3/4 hairpin and its mutants V232D, I231D, Q207N/V232E. Over 95% of the backbone resonances of a 15N,13C-labelled V232D-TM3/4 construct in the membrane-mimetic environment of perfluorooctanoate (PFO) micelles were successfully assigned, and the presence and boundaries of helical segments within TM3 and TM4 were defined under these conditions. Comparative analysis of 15N and 1H chemical shift variations among HSQC spectra of WT-, V232D-, I231D- and Q207N/V232E-TM3/4 indicated that hairpin conformations vary with the position of a polar mutation (i.e., V232D and I231D vs. WT), but remain similar when hairpins with identically-positioned polar partners are compared (i.e., V232D vs. Q207N-V232E). The overall findings suggest that a polar mutation in a TM helix can potentially distort native interfacial packing determinants in membrane proteins such as CFTR, with consequences that may lead to disease.

Water-miscible organic cosolvents enhance phosphatidylinositol-specific phospholipase C phosphotransferase as well as phosphodiesterase activity.

Phosphatidylinositol-specific phospholipase C (PI-PLC) from Bacillus thuringiensis catalyzes the hydrolysis of phosphatidylinositol (PI) in a Ca(2+)-independent two-step mechanism: (i) an intramolecular phosphotransferase reaction to form inositol 1,2-(cyclic)-phosphate (cIP), followed by (ii) a cyclic phosphodiesterase activity that converts cIP to inositol 1-phosphate (I-1-P). Moderate amounts of water-miscible organic solvents have previously been shown to dramatically enhance the cyclic phosphodiesterase activity, that is, hydrolysis of cIP. Cosolvents [isopropanol (iPrOH), dimethylsufoxide (DMSO), and dimethylformamide (DMF)] also enhance the phosphotransferase activity of PI-PLC toward PI initially presented in vesicles, monomers, or micelles. Although these water-miscible organic cosolvents caused large changes in PI particle size and distribution (monitored with pyrene-labeled PI fluorescence, 31P NMR spectroscopy, gel filtration, and electron microscopy) that differed with the activating solvent, the change in PI substrate structure in different cosolvents was not correlated with the enhanced catalytic efficiency of PI-PLC toward its substrates. PI-PLC stability was decreased in water/organic cosolvent mixtures (e.g., the T(m) for PI-PLC thermal denaturation decreased linearly with added iPrOH). However, the addition of myo-inositol, a water-soluble inhibitor of PI-PLC, helped stabilize the protein. At 30% iPrOH and 4 degrees C (well below the T(m) for PI-PLC in the presence of iPrOH), cosolvent-induced changes in protein secondary structure were minimal. iPrOH and diheptanoylphosphatidylcholine, each of which activates PI-PLC for cIP hydrolysis, exhibited a synergistic effect for cIP hydrolysis that was not observed with PI as substrate. This behavior is consistent with a mechanism for cosolvent activation that involves changes in active site polarity along with small conformational changes involving the barrel rim tryptophan side chains that have little effect on protein secondary structure.

Role of the extracellular loop in the folding of a CFTR transmembrane helical hairpin.

The folding of membrane-spanning domains into their native functional forms depends on interactions between transmembrane (TM) helices joined by covalent loops. However, the importance of these covalent linker regions in mediating the strength of helix-helix associations has not been systematically addressed. Here we examine the potential structural impact of cystic fibrosis-phenotypic mutations in the extracellular loop 2 (ECL2) on interactions between the TM3 and TM4 helices of the cystic fibrosis transmembrane conductance regulator (CFTR) in constructs containing CFTR residues 194-241. When the effects of replacements in ECL2 (including the CF-phenotypic mutants E217G and Q220R) were evaluated in a library of wild-type and mutant TM3-ECL2-TM4 hairpin constructs, we found that SDS-PAGE gel migration rates differed over a range of nearly 40% +/- the wild-type position and that decreased migration rates correlate with increasing hairpin alpha-helical content as measured by circular dichroism spectra in sodium dodecyl sulfate micelles. The decreased mobility of TM3/4 constructs by introduction of non-native residues is interpreted in terms of an elongation or "opening" of the helical hairpin and concomitant destabilization of membrane-based helix-helix interactions. Our results support a role for short loop regions in dictating the stability of membrane protein folds and highlight the interplay between membrane-embedded helix-helix interactions and loop conformation in influencing the structure of membrane proteins.

Dimer structure of an interfacially impaired phosphatidylinositol-specific phospholipase C.

The crystal structure of the W47A/W242A mutant of phosphatidylinositol-specific phospholipase C (PI-PLC) from Bacillus thuringiensis has been solved to 1.8A resolution. The W47A/W242A mutant is an interfacially challenged enzyme, and it has been proposed that one or both tryptophan side chains serve as membrane interfacial anchors (Feng, J., Wehbi, H., and Roberts, M. F. (2002) J. Biol. Chem. 277, 19867-19875). The crystal structure supports this hypothesis. Relative to the crystal structure of the closely related (97% identity) wild-type PI-PLC from Bacillus cereus, significant conformational differences occur at the membrane-binding interfacial region rather than the active site. The Trp --> Ala mutations not only remove the membrane-partitioning aromatic side chains but also perturb the conformations of the so-called helix B and rim loop regions, both of which are implicated in interfacial binding. The crystal structure also reveals a homodimer, the first such observation for a bacterial PI-PLC, with pseudo-2-fold symmetry. The symmetric dimer interface is stabilized by hydrophobic and hydrogen-bonding interactions, contributed primarily by a central swath of aromatic residues arranged in a quasiherringbone pattern. Evidence that interfacially active wild-type PI-PLC enzymes may dimerize in the presence of phosphatidylcholine vesicles is provided by fluorescence quenching of PI-PLC mutants with pyrene-labeled cysteine residues. The combined data suggest that wild-type PI-PLC can form similar homodimers, anchored to the interface by the tryptophan and neighboring membrane-partitioning residues.

Cross-linking phosphatidylinositol-specific phospholipase C traps two activating phosphatidylcholine molecules on the enzyme.

Bacillus thuringiensis phosphatidylinositol-specific phospholipase C (PI-PLC), a bacterial model for the catalytic domain of mammalian PI-PLC enzymes, was cross-linked by 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride to probe for the aggregation and/or conformational changes of PI-PLC when bound to activating phosphatidylcholine (PC) interfaces. Dimers and higher order multimers (up to 31% of the total protein when cross-linked at pH 7) were observed when the enzyme was cross-linked in the presence of PC vesicles. Aggregates were also detected with PI-PLC bound to diheptanoyl-PC (diC(7)PC) micelles, although the fraction of cross-linked multimers (19% at pH 7) was lower than when the enzyme was cross-linked in the presence of vesicles. PI-PLC cross-linked in the presence of a diC(7)PC interface exhibited an enhanced specific activity for PI cleavage. The extent of this cross-linking-enhanced activation was reduced in PI-PLC mutants lacking either tryptophan in the rim (W47A and W242A) of this (betaalpha)(8)-barrel protein. The higher activity of the native protein cross-linked in the presence of diC(7)PC correlated with an increased affinity of the protein for two diC(7)PC molecules as detected by matrix-assisted laser desorption-ionization time-of-flight mass spectrometry. In contrast to wild type protein, W47A and W242A had only a single diC(7)PC tightly associated when cross-linked in the presence of that activator molecule. These results indicate that (i) each rim tryptophan residue is involved in binding a PC molecule at interfaces, (ii) the affinity of the enzyme for an activating PC molecule is enhanced when the protein is bound to a surface, and (iii) this conformation of the enzyme with at least two PC bound that is stabilized by chemical cross-linking interacts more effectively with activating interfaces, leading to higher observed specific activities for the phosphotransferase reaction.

Role of tryptophan residues in interfacial binding of phosphatidylinositol-specific phospholipase C.

The phosphatidylinositol-specific phospholipase C (PI-PLC) from Bacillus thuringiensis exhibits several types of interfacial activation. In the crystal structure of the closely related Bacillus cereus PI-PLC, the rim of the active site is flanked by a short helix B and a loop that show an unusual clustering of hydrophobic amino acids. Two of the seven tryptophans in PI-PLC are among the exposed residues. To test the importance of these residues in substrate and activator binding, we prepared several mutants of Trp-47 (in helix B) and Trp-242 (in the loop). Two other tryptophans, Trp-178 and Trp-280, which are not near the rim, were mutated as controls. Kinetic (both phosphotransferase and cyclic phosphodiesterase activities), fluorescence, and vesicle binding analyses showed that both Trp-47 and Trp-242 residues are important for the enzyme to bind to interfaces, both activating zwitterionic and substrate anionic surfaces. Partitioning of the enzyme to vesicles is decreased more than 10-fold for either W47A or W242A, and removal of both tryptophans (W47A/W242A) yields enzyme with virtually no affinity for phospholipid surfaces. Replacement of either tryptophan with phenylalanine or isoleucine has moderate effects on enzyme affinity for surfaces but yields a fully active enzyme. These results are used to describe how the enzyme is activated by interfaces.

Investigating the interfacial binding of bacterial phosphatidylinositol-specific phospholipase C.

The interactions of PI-PLC with nonsubstrate zwitterionic [phosphatidylcholine (PC)] and anionic [phosphatidylmethanol (PMe), phosphatidylserine, phosphatidylglycerol, and phosphatidic acid] interfaces that affect the catalytic activity of PI-PLC have been examined. PI-PLC binding is strongly coupled to vesicle curvature and is tighter at acidic pH for all of the phospholipids examined. PI-PLC binds to small unilamellar vesicles (SUVs) of anionic lipids with much higher affinity (K(d) is 0.01-0.07 microM for a site consisting of n = 100 +/- 25 lipids when analyzed with a Langmuir adsorption isotherm) than to zwitterionic PC SUVs (K(d) is 5-20 microM and n = 8 +/- 3). The binding to PC surfaces is dominated by hydrophobic interactions, while binding to anionic surfaces is dominated by electrostatic interactions. The contributions of specific cationic side chains and hydrophobic groups at the rim of the alpha beta-barrel to zwitterionic and anionic vesicle binding have been assessed with mutagenesis. The results are used to explain how PC activates the enzyme for both phosphotransferase and cyclic phosphodiesterase activities.