The Role of Specific Lysine Residues in the Passage of Anions through the Pseudomonas aeruginosa Porin OprP*

When grown under phosphate-limiting conditions Pseudomonas aeruginosa expresses the phosphate-spe-cific porin OprP. In order to determine whether any of the lysine residues located in the amino-terminal half of the protein play a role in the transport of anions through the channels, the first nine amino-terminal lysine residues of OprP were substituted with glutamates. The mutant proteins were purified and the channels they formed were characterized by reconstituting the purified porins in planar lipid membranes. In comparison to the wild-type protein, the Lys 74 , Lys 121 , and Lys 126 mutants all displayed reduced levels of conductance at KCl concentrations below 1 M , and the Lys 74 and Lys 121 mutants no longer exhibited a saturation of conduct- ance at high anion concentrations. In addition, the ability of phosphate ions to inhibit the conductance of Cl (cid:50) ions through the channels formed by the Lys 121 mutant was greatly reduced, while their ability to inhibit the Cl (cid:50) conductance of the Lys 74 mutant was reduced by approximately 2-fold. To clarify the roles that Lys 74 , Lys 121 , and Lys 126 play in regulating the channel characteristics of OprP, these amino acids were replaced with either glycine or glutamine residues. Analysis of these mutants suggested that both Lys 74 and Lys 126 may serve to funnel anions toward the binding site, but only the presence of Lys 121 is required for the formation of the inorganic phosphate-specific binding site of OprP. The acquisition of inorganic phosphate (P i ) and phosphoryl-ated compounds is an essential function of growing microor-ganisms. Many bacteria have been shown to possess a group of phosphate starvation-inducible an Applied Biosystems DNA synthesizer. Western Immunoblotting— Whole cell lysates of E. coli CE1248 expressing the mutant forms of OprP were loaded on to 12% polyacryl- amide gels and subjected to SDS-polyacrylamide gel electrophoresis. The gels were transferred to nitrocellulose membranes and blotted with monomer-specific anti-OprP rabbit serum as described previously (3). Purification of Mutant Proteins— Overnight cultures of E. coli CE1248 expressing the mutant forms of OprP were pelleted, resus-pended in 20% sucrose containing 50 (cid:109) g/ml DNase, and broken by through a French press. Outer membranes were isolated using a two-step sucrose density gradient as described previously isolated outer membranes were subjected to a stepwise solubilization with octyl-polyoxyethylene and the detergent-purified mutant pro- teins were loaded on to preparative SDS-polyacrylamide gels and elec-trophoresed. nondenatured proteins were excised and eluted over- night at 4 °C into 10 m M Tris-HCl (pH 8.0) containing 0.1% SDS. Analysis of the channel charac- teristics of the OprP mutant proteins was accomplished using planar lipid bilayer techniques as previously (4, 17). Membranes composed of 2% oxidized cholesterol. conductance the Lys Lys Lys proteins was determined using the planar bilayer method. salt were used unbuffered phosphate-induced inhibition of chloride conductance was determined by measuring the single-channel conductance of the mutant proteins in chloride

The acquisition of inorganic phosphate (P i ) and phosphorylated compounds is an essential function of growing microorganisms. Many bacteria have been shown to possess a group of phosphate starvation-inducible genes whose expression result in enhanced P i uptake. This group of genes is often referred to as the Pho regulon (1). One member of the Pseudomonas aeruginosa Pho regulon is the gene that encodes the P i -specific porin OprP (2).
Like the analogous phosphate starvation-inducible Escherichia coli porin PhoE, OprP has been proposed to exist in the outer membrane as a trimer of three identical subunits, each of which traverse the membrane as a 16-stranded ␤-barrel (3). However, unlike PhoE, the P. aeruginosa porin forms channels that contain a saturable P i -binding site (K d ϭ 30 mM/Cl Ϫ ; 0.3 mM/P i ) (2,4). Chemical modification studies suggested that the P i specificity of OprP is due in part to the presence of one or more lysine residues that may take part in the formation of the binding site (4,5).
It has been demonstrated previously through both combinatorial and site-directed mutagenesis that the amino acids responsible for determining the ion selectivities of the highly homologous E. coli porins OmpF, OmpC, and PhoE are located exclusively in the amino-terminal halves of these proteins (6 -8). One specific PhoE residue (Lys 125 ) located in the third surface-exposed loop appeared to be critical for ion transport. The substitution of glutamate for this residue resulted in a cation-rather than an anion-selective channel (8). Additionally, mutagenesis of the maltose-specific porin LamB has shown that residues important for the substrate specificity of this protein are concentrated in the amino-terminal end (9,10).
Of the 23 lysine residues found in OprP, only nine are located in the amino-terminal half of the protein; one is located in the vicinity of the proposed second surface-exposed loop and two are found in the proposed third surface-exposed loop (3). In order to determine whether any of the lysines contained in the amino-terminal half of OprP are necessary for the transport of anions a PCR 1 -based site-directed mutagenesis protocol was used to individually mutate specific lysine residues. In this study we report the results of the mutagenesis of the nine amino-terminal lysine residues of OprP and the effect these mutations had on the single-channel conductance and P i binding of this porin.

EXPERIMENTAL PROCEDURES
Chemicals-KCl, K 2 HPO 4 , and KH 2 PO 4 were purchased from Fisher Canada. KCl was used unbuffered (pH 6.0) while equal molar concentrations of K 2 HPO 4 and KH 2 PO 4 were mixed to achieve a pH of 8.0.
Bacterial Strains and Growth Conditions-E. coli DH5-␣ was used for all procedures involved in creating the oprP mutant plasmids. Strain CE1248 was utilized in all expression experiments (11). Cells were grown overnight at 37°C in LB broth supplemented with ampicillin (50 g/ml), and in the case of cells grown for the purpose of porin purification, 0.4% glucose.
General Molecular Techniques-Restriction endonucleases, Vent DNA polymerase, and T4 DNA ligase purchased from Life Technologies, Inc. and New England Biolabs Inc. were used in accordance with the accompanying literature. Cells were transformed using the CaCl 2 method (12).
Site-directed Mutagenesis-The oprP substitution mutants were created using a recombinant PCR method (13,14) with the oprP containing plasmid pAS27 (3) used as the template. Mutagenic oligonucleotides contained mismatches that corresponded to a substitution mutation in the encoded amino acid sequence (Table I). The mutagenized fragments of oprP were subcloned back into plasmid pAS27 and sequenced. Oligonucleotides were synthesized on a model 392 Applied Biosystems DNA synthesizer (Applied Biosystems Canada, Mississauga, Ontario, Canada).
DNA Sequencing-Plasmid DNA was sequenced using an Applied Biosystems model 373 fluorescent sequencer and PCR protocols provided by the manufacturer. Template DNA was prepared by the polyethylene glycol precipitation method (12). Primers were synthesized on * This work was supported by a grant from the Medical Research Council of Canada. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
an Applied Biosystems DNA synthesizer.
Western Immunoblotting-Whole cell lysates of E. coli CE1248 expressing the mutant forms of OprP were loaded on to 12% polyacrylamide gels and subjected to SDS-polyacrylamide gel electrophoresis. The gels were transferred to nitrocellulose membranes and blotted with monomer-specific anti-OprP rabbit serum as described previously (3).
Purification of Mutant Proteins-Overnight cultures of E. coli CE1248 expressing the mutant forms of OprP were pelleted, resuspended in 20% sucrose containing 50 g/ml DNase, and broken by passage through a French press. Outer membranes were isolated using a two-step sucrose density gradient as described previously (15). The isolated outer membranes were subjected to a stepwise solubilization with octyl-polyoxyethylene (16), and the detergent-purified mutant proteins were loaded on to preparative SDS-polyacrylamide gels and electrophoresed. The nondenatured proteins were excised and eluted overnight at 4°C into 10 mM Tris-HCl (pH 8.0) containing 0.1% SDS.
Planar Lipid Bilayer Experiments-Analysis of the channel characteristics of the OprP mutant proteins was accomplished using planar lipid bilayer techniques as described previously (4,17). Membranes were composed of 2% oxidized cholesterol.

RESULTS
Site-directed Mutagenesis of oprP.-To assess the role that the first nine amino-terminal lysine residues of OprP play in determining the channel characteristics of this protein, these residues were replaced with glutamates using a PCR-based site-directed mutagenesis method as described under "Experimental Procedures." In addition, Lys 74 , Lys 121 , and Lys 126 were replaced with glycines, Lys 121 and Lys 126 were replaced with glutamines, and a triple mutant in which Lys 74 , Lys 121 , and Lys 126 were all replaced with glutamates was also created. The secondary PCR products were digested with either HindIII (Lys 13 -Lys 74 ) or EcoRV/SphI (Lys 109 -Lys 181 ) and were ligated to similarly digested and gel-purified plasmid pAS27. Recombinant plasmids having the appropriate restriction enzyme digestion patterns were sequenced to verify the presence of the desired mutation (Table I). No errors in the coding sequence of the mutagenized fragments were detected.
Expression and Purification of Mutant Proteins-The OprP substitution mutant plasmids were transformed into the porindeficient strain CE1248, and overnight cultures were used to prepare whole cell lysates. After heating at 100°C for 10 min the samples were electrophoresed, transferred to nitrocellulose membranes, and blotted with anti-OprP antiserum. All the mutant proteins were expressed at levels comparable to that of the wild-type protein (data not shown).
Single-channel Conductance of Lys 3 Glu Mutant Proteins-In order to determine whether the substitution of indi-vidual lysine residues had an effect on the conductance saturation of OprP, the single-channel conductance of each of the Lys 3 Glu mutant proteins was assessed at various salt concentrations (Table II). The average conductance of six of the mutant proteins (Lys 13 , Lys 15 , Lys 25 , Lys 30 , Lys 109 , and Lys 181 ) was similar to that of wild-type OprP at all tested salt concentrations. In contrast, three of the mutant proteins displayed distinctly altered channel characteristics. In 1 M KCl, the Lys 74 and Lys 126 mutants exhibited levels of conductance that were approximately one-half of that of the wild-type protein, while the Lys 121 mutant possessed a conductance of approximately one-third of that of wild-type OprP (Fig. 1).
The conductance of the channels formed by these three mutant proteins as well as wild-type OprP was plotted as a function of increasing salt concentration (Fig. 2). As shown previously, the single-channel conductance of wild-type OprP plateaus as the salt concentration approaches 1 M. However, both the Lys 74 and the Lys 121 mutant proteins formed channels that displayed linear concentration-conductance relationships at up to 3 M KCl. The Lys 126 mutant channel conductance, although significantly lower than wild-type OprP, tended to follow the same pattern as the wild-type protein, plateauing as the salt concentration approached 1 M.
While it might be expected that mutations which so profoundly affected the channel conductance might also have an effect on anion selectivity, measurements of the ion selectivity of these mutants revealed no significant differences from the wild-type protein (data not shown).
Phosphate-induced Inhibition of Single-channel Conductance of Lys 3 Glu Mutant Proteins-In order to determine whether any of the Lys 3 Glu substitutions had an effect on the P i -binding site of OprP, the ability of phosphate ions to inhibit the single-channel conductance of the mutant proteins was measured. The single-channel conductance of each mutant protein in 0.1 M KCl was determined prior to the addition of P i (Table II). Increasing amounts of potassium phosphate were added to the bathing solutions, and the resultant channel conductances were measured. These data were then used to calculate the percent inhibition and the I 50 concentration of the added phosphate ions (Table III).
The majority of the mutant proteins exhibited degrees of conductance inhibition similar to or greater than the wild-type protein, which displayed a 74% decrease in conductance after the addition of 3.3 mM potassium phosphate. The Lys 74 mutant had a slightly lowered affinity for P i with a maximum inhibition of 58% and an I 50 concentration of 1.95 mM compared to 0.96 mM for wild-type OprP. The Lys 121 substitution had a  profound effect on the ability of the protein to bind phosphate ions. This mutant showed a maximum inhibition of 30%, and while the I 50 for this mutant could not be measured under the conditions used to examine the other mutant proteins, additional experiments revealed that it was above 10 mM. The Lys 126 mutant channel conductances, although greatly reduced compared to those of the wild-type protein, appeared to be inhibited by the presence of phosphate ions to a similar degree. Fig. 3 shows the channel conductances of the Lys 74 , Lys 121 , and Lys 126 mutant proteins along with wild-type OprP plotted as a function of increasing phosphate ion concentrations. Channel Characteristics of Lys 3 Gly, Lys 3 Gln, and Lys 74, 121, 126 3 Glu Mutant Proteins-To further examine the roles Lys 74 , Lys 121 , and Lys 126 play in determining the electrochemical nature of the channels formed by OprP, these amino acids were substituted with either a Gly or a Gln residue, and the single-channel conductance and phosphate-induced inhibition of chloride conductance was determined for each of these mutant proteins (Table IV). In addition, a triple mutant with Lys 74 , Lys 121 , and Lys 126 all substituted with glutamates was also created and analyzed. Substituting Lys 74 with Gly resulted in a channel with a conductance comparable to the wild-type protein in 1 M KCl. However, the channel conductance of this mutant in 0.1 M KCl was similar to that of the Lys 74 3 Glu mutant protein. The phosphate-induced conductance inhibition of the Lys 74 3 Gly mutant was comparable to that of the wild-type protein.
Substituting the Lys 126 residue with either Gly or Gln resulted in channels that had reduced levels of conductance in comparison to the wild-type protein. In the case of the Gly substitution, the channel conductance at both 0.1 M and 1 M KCl was lower than that of the Lys 126 3 Glu mutant. Substituting Lys 121 with either Gly or Gln resulted in channels with reduced conductance at both 0.1 M and 1 M KCl. These mutant proteins also formed channels that were as severely impaired in their ability to bind phosphate ions as the initial Lys 121 3 Glu mutant protein.
The single-channel conductance of the Lys 74, 121, 126 3 Glu triple mutant was somewhat lower than any of the single mutants in both 0.1 M and 1 M KCl; however, the phosphate-

DISCUSSION
It has been demonstrated previously that the channel characteristics of several general diffusion porins are dependent on the presence of one or more amino acids located in their aminoterminal domains (6 -8). In this study we have identified three amino-terminal lysine residues in OprP that play a role in defining the channel characteristics of this porin.
In 1 M KCl, the Lys 121 3 Glu mutant channels displayed a 3-fold decrease in conductance, while the Lys 74 3 Glu and Lys 126 3 Glu mutants displayed 2-fold and 2.5-fold decreases, respectively. This is in contrast with the 10-fold reduction in conductance exhibited by chemically modified forms of OprP (18). Even a triple mutant in which all three of these lysine residues had been substituted with glutamates displayed only a 5-fold decrease in conductance. These findings suggest that there may be certain lysine residues contained in the carboxylterminal end of OprP that also play a role in determining the channel conductance. Alternately, the severe reduction in conductance displayed by the chemically modified forms of OprP may have been due to factors unrelated to the actual loss of the positive charges of the modified lysine residues (e.g. the presence of the modifying groups within and/or around the channel).
Lysine-specific acetylation of OprP was shown to produce channels with conductances that were no longer saturated at high anion concentrations (18). This result was explained to be due to a modification of specific residues which are involved in forming the anion-binding site (4). Of the eight Lys 3 Glu mutant forms of OprP created during the course of this study, only the Lys 74 and Lys 121 mutants exhibited losses in the ability to saturate at KCl concentrations above 1 M. The Lys 126 3 Glu mutant displayed saturation kinetics similar to that of wild-type OprP despite the fact that the conductance of these channels was as severely affected at low salt concentrations as that of the Lys 74 and Lys 121 mutant porins. The conductance patterns of the other six Lys 3 Glu mutants did not differ significantly from that of the wild-type protein. Apparently only the Lys 74 and Lys 121 substitutions had a detrimental effect on the anion-binding site.
The phosphate-induced inhibition of channel conductance of the Lys 74 3 Glu mutant was approximately 2-fold lower than that of the wild-type protein. Substituting this lysine residue with a glycine resulted in a protein with a phosphate-induced inhibition of conductance that was similar to that of the wildtype protein. This result can be explained if Lys 74 is assumed to occupy a space proximate to the P i -binding site. The positive charge of this residue would not be required for the formation of the binding site; however, the placement of a negatively charged residue at this location may have indirectly affected the interaction of phosphate (and chloride) ions with the binding site. According to our recent OprP topological model, Lys 74 is located at the top of the fourth ␤-strand and would presumably face the interior of the channel (3).
Substitution of Lys 121 with glutamate, glycine, or glutamine residues resulted in proteins with channel conductances that were severely impaired in their abilities to be inhibited by the presence of phosphate ions. This particular residue is located in the third surface-exposed loop according to a recently published topological model of OprP (3). The placement of this residue in the third loop is in agreement with the role of this loop in constricting the interior of the channels formed by several bacterial porins (19 -21). The equivalent lysine residue in PhoE (Lys 125 ) that has been established as being responsible for determining the anion selectivity of this porin (8) was also shown to be located in the third surface-exposed loop (19). Substituting Lys 126 with glutamate in OprP had no apparent effect on the P i -binding site, despite the fact that this residue is also predicted to be located in the third surface-exposed loop.
The roles of Lys 74 and Lys 126 in OprP appear to be to form an electrostatic funnel that serves to focus the flow of anions toward the binding site. The Lys 121 residue seems to render a more critical function in this porin. Not only does the presence of this residue serve to increase the flow of anions through the channel, but it appears that this particular amino acid is required for the formation of the anion/P i -binding site. Whether this is the only residue involved in maintaining the P i specificity of OprP remains to be seen.