Protein Engineering, Vol. 15, No. 10, 799-804,
October 2002
© 2002 Oxford University Press
Asymmetric conductivity of engineered porins
Institut für Organische Chemie und Biochemie, Albert-Ludwigs-Universität, Albertstrasse 21, D-79104 Freiburg im Breisgau, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResults and discussionComparison of membrane...Conductivity changes of membrane...ConclusionReferences |
|---|
Positively charged peptide segments of 16 and 18 residues were inserted at a periplasmic turn of the porin from Rhodobacter blasticus in order to form an electric field-dependent plug. The X-ray diffraction analysis of a mutant confirmed that the structure of the porin had remained intact and that the insert was mobile. Incorporation experiments of single molecules into lipid bilayers showed that the distribution of electric conduction increments depended on the field polarity. The observed distributions are explained if the porin molecules enter the bilayer preferentially with their periplasmic surface first. Furthermore, the conduction of membrane-incorporated porin mutants changed reproducibly on field reversal showing asymmetries of
15%, while the wild-type remained constant. This asymmetry is most likely caused by the electric field pressing the charged insert onto the pore eyelet in one field direction and removing it from the eyelet in the other. The results encourage attempts to improve the inserts in order to eventually reach diode characteristics.
Keywords: black lipid membrane/ion channels/ion current gating/peptide insert/X-ray diffraction
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionComparison of membrane...Conductivity changes of membrane...ConclusionReferences |
|---|
Porins form channels across the outer membrane of Gram-negative bacteria that allow for the passive permeation of small polar solutes with an upper Mr limit of
600 Da (Jap and Walian, 1990
; Nikaido, 1994; Schulz, 1996). The permeation characteristics can be assessed by incorporating single trimeric porin molecules into lipid bilayer membranes and measuring the resulting ion current increments (Montal and Mueller, 1972; Benz and Bauer, 1988; Van Gelder et al., 2000). Numerous mutants of such porins have been produced and analyzed with respect to conductivity and ion selectivity (Schmid et al., 1998; Saxena et al., 1999; Liu et al., 2000). It is conceivable that such a channel can be blocked in one of the two field directions by fastening a charged mobile plug at one of the channel ends (Figure 1). This should result in an asymmetric conductivity, in an ideal case showing the diode characteristics known from plasma membrane channels (Zhou et al., 2001). Here we inserted positively charged peptide segments of 16 and 18 residues at a periplasmic turn of the porin from Rhodobacter blasticus (Butz et al., 1993; Kreusch et al., 1994; Schmid et al., 1998) and demonstrated that they caused reproducible conductivity differences on electric field reversal, in contrast to the wild-type that showed identical conductivity in both field directions.
|
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionComparison of membrane...Conductivity changes of membrane...ConclusionReferences |
|---|
Design and production of mutants
All inserts were designed to start after Asn203. They were flanked by glycines in order to make them mobile (Van Gelder et al., 1997). Proline with its high 
-helix initiation potential was always used at the first non-glycine position, and Leu, Met, Glu, Lys, Ala, Arg with high -helix propensities for the following nine positions, hoping that the inserts would form -helices.
First, gene insB was produced by two PCR runs with two internal and two external primers starting from the plasmid pET-3b-por (Schmid et al., 1996) that had NdeI and BamHI sites before and after the porin gene. The two internal PCR primers coded for the insert, carried a XhoI restriction site, and overlapped with the porin gene for 19 bases at the 3'-end and 31 bases at the 5'-end. The external primers, each 35 bases long, were located beyond the NdeI and BamHI sites, respectively. The PCR products were cut with NdeI and XhoI, or XhoI and BamHI, respectively, and isolated from a 1% agarose gel using a gel extraction kit (Qiagen, Hilden, Germany). Together, the resulting two DNA fragments were ligated into a pUC18 vector that had been opened with NdeI and BamHI. The correct product was selected from heat-shock-transformed Escherichia coli XL1-Blue cells (Hanahan, 1983) and transferred into the production vector pET-3b-insB. The genes of mutants InsA through InsE (Figure 1A
) were then produced by a cassette mutagenesis approach using three further restriction sites, NotI, ApaI and MunI, respectively, which were introduced within and near the insert.
The mutant-carrying plasmids were expressed into inclusion bodies in E.coli BL21(DE3)pLysS. All porin mutants carried the additional mutation Glu1
Met that is required for the cytosolic production. The porins were isolated as described (Schmid et al., 1996), except that urea and superfluous detergent were removed by dialysis against a 10-fold volume of buffer containing 10 mM Tris–HCl pH 8.0, 20 mM NaCl and 0.2 mM EDTA in the (re)naturation procedure. The resulting porin was then purified using a Q-Sepharose-FF column. The procedure of Pautsch et al. (Pautsch et al., 1999) was equally efficient. The protein yields were 20 mg per liter of culture.
Crystallization and X-ray diffraction analysis
Crystals of mutant InsA grew from 50 mM NaOAc pH 4.75, 100 mM (NH4)2SO4, 5.5% PEG-4000, 5 mM octyltetraoxyethylene (C8E4), 12.5 mM decanoyl sucrose and 5 mg/ml of the protein. Diffraction data were collected at room temperature from one crystal using a rotating anode (Rigaku, model RU200B) and an image plate (MARresearch, 30 cm). The data were processed with the programs MOSFLM (Leslie, 1999) as well as SCALA and TRUNCATE (CCP4, 1994). The crystal was not merohedrally twinned (Yeates, 1997). Calculated phases were taken from the wild-type (PDB accession code 1PRN). The structure was refined with a fixed temperature factor using program CNS (Brünger et al., 1998). The coordinates and structure factors are deposited in the Protein Data Bank under accession code 1H6S.
Conductivity measurements
The purified protein was incorporated into a planar lipid bilayer of diphytanoyl phosphatidyl choline (Avanti Polar Lipids, Alabaster, AL, USA) separating two chambers each filled with 5 ml of buffer (20 mM MOPS titrated with 11 mM Tris to pH 7.2) (Figure 1D). In order to be consistent with the crystallization conditions, mutant InsA was also measured at pH 4.8 buffered with 20 mM LiOAc. In all experiments, the conducting electrolyte was added to both chambers. A further 10–40 µM Triton X-100 applied to the cis chamber (containing the porins) facilitated the incorporation into the bilayer. Each incorporated porin molecule gave rise to an ion current increment of 50 pA that was usually stable and reported electronically.
In our electric field reversal experiments we always used the MOPS buffer at pH 7.2. The electrode connections were reversed with a double switch in such a manner that the current amplifier and the detection apparatus remained unaffected. The electrodes themselves showed an asymmetry of 1 mV which was taken into account when calculating conductivities.
|
Results and discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionComparison of membrane...Conductivity changes of membrane...ConclusionReferences |
|---|
Asn203 in the mobile turn T5 at the periplasmic barrel end was chosen as the insert position, because a disturbance at this turn should not affect the porin structure. Moreover, this position is close to the channel and far from any crystal packing contact (Kreusch and Schulz, 1994
) so that crystals isomorphous with the wild-type could still be formed. Since there was no appropriate DNA restriction site near turn T5, the first mutant gene was produced in two pieces which were then joined by means of a newly introduced XhoI site within the insert to yield gene insB. For the construction of the other inserts given in Figure 1A, three more restriction sites were introduced in the same region. The porin preparations were pure as demonstrated by gel electrophoresis (Figure 2). Moreover, the presence of the insertions in the mutant porins was confirmed by appropriate retardations.
|
In contrast to the well diffracting wild-type crystals grown at pH 7.2 (Kreusch and Schulz, 1994
), the insertion mutants formed no or only very small crystals at this pH. However, X-ray grade crystals from mutant InsA that were isomorphous with those of wild-type porin were obtained at pH 4.8 (Table I). This demonstrates that the insertion did not interfere with any packing contact, as it was designed to do. The structure was elucidated by difference-Fourier methods and subsequently refined. The resulting refined structure was identical to the wild-type structure (Kreusch and Schulz, 1994) apart from a chain rupture after position 203 at the insert which is shown in Figure 3. Since the solvent region around residue 203 contained no electron density, we conclude that the insert followed our design in as far as it was mobile and did not affect the native porin structure. It is not clear, however, if it formed the envisaged -helix sketched in Figure 1C.
|
|
(solid) and –3 |
Comparison of membrane incorporation out of anode versus cathode chamber |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionComparison of membrane...Conductivity changes of membrane...ConclusionReferences |
|---|
In two series of experiments we analyzed the behavior of the porin mutants on incorporation into the lipid bilayer. First, we added porin to the chamber with the more positive potential, i.e. the `anode' chamber (cis+) and measured the ionic current steps caused by the incorporation of single porin molecules into the membrane (Figure 1D
). Then, we applied the porin to the `cathode' chamber (cis–) and repeated the measurements. The statistics of all step sizes are depicted in Figure 4. Wild-type porin was symmetric, showing essentially the same histogram of conduction events when applied to the anode or the cathode chamber. For the five insertion mutants, however, the histograms differed significantly. In the case of cis+ incorporation, the distribution was approximately like that of the wild-type, whereas the cis– incorporation gave rise to broad distributions containing numerous small but also large conduction events (Figure 4, right side).
|
Obviously, membrane incorporation is affected by the applied electric field. Wild-type porin, and in particular the mutants, contain an electric dipole (Figure 1C
) which probably pre-orients the porin near the membrane. It is likely that on cis+ incorporation the periplasmic porin surface enters the membrane first so that the field forces the insert away from the incorporated porin (Figure 1D), resulting in a normal conduction behavior (Figure 4, left side). During cis– incorporation, on the other hand, the highly charged external surface enters the bilayer first. If this happened in an orderly manner, the pore in the membrane would behave normally. The numerous small and also large conduction events indicate, however, that this is not the case (Figure 4, right side). Possibly, the porin disintegrates or is clogged if it tries to enter with its external side first, causing only a small conduction event. The large conduction events with approximately doubled conductivity values may be caused by the incorporation of (trimeric) porins that have associated to dimers which are likely to be more resistant to disintegration.
In similar experiments, the cis+ incorporation of wild-type porin and mutants InsA and InsB was repeated with lithium acetate as the electrolyte instead of KCl. The resulting histograms in Figure 5 show a reduction by a factor of five with respect to the KCl conductivity, contrasting the factor of two that is expected from the ion mobility differences in free solution. This deviation reflects the geometric influence of a narrow pore eyelet on the large acetate and hydrated Li+ ions compared with the small K+ and Cl- ions.
|
|
Conductivity changes of membrane-incorporated porins on electric field reversal |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionComparison of membrane...Conductivity changes of membrane...ConclusionReferences |
|---|
In a further series of experiments we added porins to the anode chamber (cis+), waited for the incorporation of as few molecules as possible and then reversed the electric field across the bilayer. The reversal was repeated several times in order to check reproducibility. The cis+ option was applied because in this case the membrane incorporation behavior was much more normal than with cis– (Figure 4
). Since large ions are more easily stopped by a leaky plug than small ones, all experiments were performed with the electrolyte lithium acetate.
Using wild-type porin we observed no conductivity differences on field reversal in any of our experiments. The results for a single incorporated porin molecule (as judged from the observed conductivity of 0.7 nS which fits the distribution of Figure 5A) are shown in Figure 6A. Apart from wild-type we used the mutant InsA in a number of trials. Most of them showed clear and reproducible conductivity differences on field reversal. A measurement with more than 10 reversals is shown in Figure 6B. A further experiment with a reproducible two-step conductivity decrease on field reversal is given in Figure 6C. Such double steps occurred in several measurements. Asymmetries of similar magnitude were observed in experiments using the other mutants and other salts (data not shown).
|
The observed conductivity differences were plotted against the average conductivity in Figure 7
. Given the uncertainty of a conceivable anti-parallel incorporation of porins, only small conduction events can be interpreted with sufficient confidence. Therefore, we chose a 3.5 nS cut-off that corresponds to five or less incorporated porins as derived from the distribution of Figure 5B. The data fit reasonably well to a straight line showing an asymmetry of 15%. The differences were overwhelmingly positive supporting the suggested dipole-oriented cis+ membrane incorporation which places the positive insert into the initial cathode chamber (Figure 1D) so that field reversal presses the insert onto the pore and reduces the conductivity.
|
The spread of the conductivity differences shown in Figure 7
is rather broad and the lower conductivity values of these differences are still appreciably above zero. Accordingly, the mobile insert does not plug the pore completely but only disturbs the ion flow by assuming a binding position at the pore. In this picture, the observed two steps of Figure 6C result from two different binding positions in the `open' state when the electric field pushes the insert away from the pore, whereas the third binding position of the `closed' state is uniform. Since such pairs of `open' states occurred in several experiments with InsA, we suggest that the insert shows an appreciable binding affinity to the rigid porin core structure which, after field reversal, is overcome by the electric field for a couple of seconds. |
Conclusion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionComparison of membrane...Conductivity changes of membrane...ConclusionReferences |
|---|
The designed conduction asymmetry was in fact achieved, although it remained rather small. The experiments suggest a dipole-oriented incorporation of the porins into the lipid bilayer with a preference for the less strongly charged periplasmic surface entering first. The reported asymmetries were observed at low voltages of 20–53 mV and therefore differ from the asymmetric voltage gating effects of natural or slightly mutated porins occurring at higher voltages (Morgan et al., 1990
; Dahan et al., 1994; Gokce et al., 1997; Mathes and Engelhardt, 1998; Samartzidou and Delcour, 1998), which are likely to be caused by conformational changes of the bulk porin. However, they do resemble channel conductance asymmetries observed in the presence of large ions (Rink et al., 1994; delaVega and Delcour, 1995). We expect that insert optimization will lead to higher asymmetries, although it may be difficult to reach the diode characteristics known from numerous natural channels (Zhou et al., 2001).
|
Notes |
|---|
1 To whom correspondence should be addressed. E-mail: schulz{at}bio.chemie.uni-freiburg.de
|
Acknowledgments |
|---|
The conductivity measurements were performed in the laboratory of J.Weckesser. We thank R.Harwardt and J.Weckesser for their help and for discussions. The project was supported by contract BEO-BMBF no. 0310898.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionComparison of membrane...Conductivity changes of membrane...ConclusionReferences |
|---|
Benz,R. and Bauer,K. (1988) Eur. J. Biochem., 176, 1–19.[Web of Science][Medline]
Brünger,A.T., Adams,P.D., Clore,G.M., DeLano,W.L., Gros,P. Grosse-Kunstleve,R.W., Jiang,J.-S., Kuszewski,J., Nilges,M., Pannu,N.S. et al. (1998) Acta Crystallogr. Sect. D, 54, 905–921.[CrossRef][Medline]
Butz,S., Benz,R., Wacker,T., Welte,W., Lustig,A., Plapp,R. and Weckesser,J. (1993) Arch. Microbiol., 159, 301–307.
CCP4: Collaborative Computational Project, Number 4. (1994) Acta Crystallogr. Sect. D, 50, 760–763.[CrossRef][Medline]
Dahan,D., Vachon,V., Laprade,R. and Coulton,J.W. (1994) Biochim. Biophys. Acta, 1189, 204–211.[Medline]
delaVega,A.L. and Delcour,A.H. (1995) EMBO J., 14, 6058–6065.[Web of Science][Medline]
Gokce,I., Bainbridge,G. and Lakey,J.H. (1997) FEBS Lett., 411, 201–205.[CrossRef][Web of Science][Medline]
Hanahan,D. (1983) J. Mol. Biol., 166, 557–580.[Web of Science][Medline]
Jap,B.K. and Walian,P.J. (1990) Q. Rev. Biophys., 23, 367–403.[Web of Science][Medline]
Kreusch,A. and Schulz,G.E. (1994) J. Mol. Biol., 243, 891–905.[CrossRef][Web of Science][Medline]
Kreusch,A., Neubüser,A., Schiltz,E., Weckesser,J. and Schulz,G.E. (1994) Protein Sci., 3, 58–63.[Web of Science][Medline]
Leslie,A.G.W. (1999) Acta Crystallogr. Sect. D, 55, 1696–1702.[CrossRef][Medline]
Liu,N., Samartzidou,H., Lee,K.W., Briggs,J.M. and Delcour,A.H. (2000) Protein Eng., 13, 491–500.
Mathes,A. and Engelhardt,H. (1998) Biophys. J., 75, 1255–1262.[Medline]
Montal,M. and Mueller,P. (1972) Proc. Natl Acad. Sci. USA, 69, 3561–3566.
Morgan,H., Lonsdale,J.T. and Alder,G. (1990) Biochim. Biophys. Acta, 1021, 175–181.[Medline]
Nikaido,H. (1994) J. Biol. Chem., 269, 3905–3908.
Pautsch,A., Vogt,J., Model,K., Siebold,C. and Schulz,G.E. (1999) Proteins, 34, 167–172.[CrossRef][Web of Science][Medline]
Rink,T., Bartel,H., Jung,G., Bannwarth,W. and Boheim,G. (1994) Eur. Biophys. J., 23, 155–165.[CrossRef][Web of Science][Medline]
Samartzidou,H. and Delcour,A.H. (1998) EMBO J., 17, 93–100.[CrossRef][Web of Science][Medline]
Saxena,K., Drosou,V., Maier,E., Benz,R. and Ludwig,B. (1999) Biochemistry, 38, 2206–2212.[CrossRef][Medline]
Schmid,B., Krömer,M. and Schulz,G.E. (1996) FEBS Lett., 381, 111–114.[CrossRef][Web of Science][Medline]
Schmid,B. Maveyraud,L. Krömer,M. and Schulz,G.E. (1998) Protein Sci., 7, 1603–1611.[Web of Science][Medline]
Schulz,G.E. (1996) Curr. Opin. Struct. Biol., 6, 485–490.[CrossRef][Web of Science][Medline]
Van Gelder,P., Saint,N., van Boxtel,R., Rosenbusch,J.P. and Tommassen,J. (1997) Protein Eng., 10, 699–706.
Van Gelder,P., Dumas,F. and Winterhalter,M. (2000) Biophys. Chem., 85, 153–167.[CrossRef][Web of Science][Medline]
Yeates,T.O. (1997) Methods Enzymol., 276, 344–358.[Web of Science][Medline]
Zhou,M., Morais-Cabral,J.H., Mann,S. and MacKinnon,R. (2001) Nature, 411, 657–661.[CrossRef][Medline]
Received September 27, 2001; revised June 25, 2002; accepted July 3, 2002.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||