Biosensors & Bioelectronics 6 (1991) 233-237
Biosensors based on membrane transport proteins Hans Kiefer, Barbara Klee, Edgar John, York-Dieter Stierhof & Fritz Jahnig Max-Planck-Institut fur Biologie, Corrensstrasse 38, D7400 Tubingen, FRG (Received 4 May 1990 ; revised version received 24 September 1990 ; accepted 26 September 1990)
Abstract: We propose a novel class of biosensors based on membrane bound receptors or transport proteins as the sensing element The protein is incorporated in a planar lipid bilayer which covers the transducer . The transducer may detect an electric current, a voltage, or a change in fluorescence . A prototype lactose sensor is presented which consists of a quartz slide covered by a lipid membrane containing the protein lactose permease from Escherichia coli . This protein is a lactose/H+ cotransporter, hence lactose in the external medium initiates lactose/H+ cotransport across the lipid membrane . This leads to a rise in proton concentration in the small volume between the lipid membrane and the quartz surface which can be detected by a pH-sensitive fluorescence dye. Keywords: biosensor, lactose/H+ cotransport, supported membrane, TIRF .
INTRODUCTION
function in nature can also be used as receptors in an artificial system . If the solute is present on one side of the membrane, cotransport will start and produce an ionic current which can be detected as described above. This is why these transport proteins are subsumed under membrane receptors in this paper . For practical use, the following problem must be considered . If a current across the membrane should be detected, the membrane must separate an external compartment containing the analyte from an `internal' compartment in contact with the transducer . This separation has to be leakagefree compared to the current to be measured. Three artificial lipid membrane systems are commonly used for the measurement of ionic currents: (i) the vesicle system is restricted to spectroscopic observations of dyes entrapped in the inner volume, since this volume is much too small to contain any macroscopic detection
Receptors are defined as proteins which produce a signal in the cell upon binding an external substrate . This signal can be the release of a second messenger, the activation of a G-protein, or an ionic current across the membrane induced by opening a gated ion channel . The latter case suggests the use of such proteins as sensing elements for biosensors, because an ionic current can be detected by various transducers . The current may be measured directly (amperometric detection), the resulting membrane potential may be detected (potentiometric detection) or, finally, a change in ion concentration may be measured with ion-sensitive electrodes, ISFETs or spectroscopically by ion-sensitive probes. It should be noted in this context that ion/ solute cotransporters that have a totally different 233 Biosensors & Bioelectronics
0956-5663/91/$03 .50 0
1991 Elsevier Science Publishers Lid, England . Printed in Great Britain
234 device; (ii) the black lipid membrane (BLM) system can be used for any form of detection, since the two compartments that are separated by the BLM can be chosen as large as desired . However, a BLM is mechanically very instable which restricts its use for a biosensor drastically; (iii) a supported planar bilayer (SPB) seems most promising, since it separates two macroscopic volumes on the one hand and, on the other hand, is extremely stable even over a long time . However, as the distance between membrane and support is only about 2 nm, the entrapped space is too small to contain a macroscopic detection device. Therefore, the SPB must be spread directly on the transducer surface. A prototype lactose sensor is currently tested which uses the lactose permease from Escherichia coli as the sensing element. This protein is a cotransporter for lactose and H +. The purified and reconstituted protein is incorporated in a SPB spread on the surface of a quartz slide . Changes of protein concentration in the space between SPB and quartz surface are monitored by measuring the fluorescence of a pH-sensitive probe entrapped in this volume.
H. Kiefer et al.
For measurements on the vesicle system, vesicles were passed over a sephacryl S-100 HR (Pharmacia) column to remove external pyranine. Valinomycin (Serva, 50 nM), was added . Measurements were carried out on a Perkin-Elmer MPF-3 spectrofluorimeter connected to a chart recorder. The excitation wavelength was 460 nm; the emission wavelength, 516 nm . pH gradients were produced by adding a small amount of 100 mm H2 SO4 to the solution . Lactose transport was initiated by adding various amounts of 0. 5 M lactose dissolved in the above buffer . For SPB measurements, vesicles were prepared in the presence of pyranine and spread on the surface of a pre-treated quartz slide (Hellma) yielding a planar bilayer as described in McConnell et al. (1986). The slide was part of a flow-through cell shown schematically in Fig. 1 . External pyranine was removed by passing buffer through the cell . Fluorescence of pyranine entrapped between the slide and the planar bilayer was excited by the 458 nm line of an argon ion laser (Spectra Physics, Model 2016). The laser beam was guided inside the slide under total
PM
MATERIALS AND METHODS All chemicals not further specified were from Merck (Darmstadt, F.R.G.). E. coli phospholipid was exctracted from wild-type E. coli K12 as described in Radin (1981) and purified over a silica gel column using a chloroform/methanol gradient Lactose permease was purified from the overproducing strain E. coli T206 and reconstituted into vesicles of E. coli lipid at a lipid/protein molar ratio of 4000 as described by Wright and Overath (1983). Vesicles were freeze-thawed three times in 10 mm potassium phosphate buffer pH 7 . 3, 100 mm K2 SO4, I mm DTT (Sigma), 1 mm pyranine (8-hydroxipyrene-1,3,6-trisulfonic acid, Molecular Probes Eugene, OR, U.S.A.). Subsequently, they were brought to a homogeneous size by the extrusion method (Hope et al., 1985). The pore sizes of the nucleopore filters were 400, 200, 100 and 50 nm . This method provides predominantly unilamellar vesicles . Control vesicles without lactose permease were made starting from a lipid-detergent mixture equal to the mixture used for protein reconstitution, treating this mixture in the same way .
L LM
Fig. 1. Schematic illustration ofthe experimental set-up for the TIRF measurements. The beam of a laser (L) is coupled into the quartz slide (S) via a mirror and a prism (P). The prism is optically connected to the slide by immersion oil. The quartz slide is clamped on a block made of Plexiglass (B). The 0-ring (R) between quartz slide and Pleriglass serves to seal a volume of about 0 .1 ml providing the sample cell. The inlet of the cell is connected to a peristaltic pump . The lipid membrane (LM) establishes on the surface of the quartz slides upon ,hushing through the cell a suspension of vesicles. The fluorescence light is detected by an objective lense (O) with 10-fold magnification, a monochromator (M) and a photomultiplier (PM). Both monochromator and photomultiplieraremounted on a conventional microscope (Leitz) which also allows visual inspection of the probe.
235
Biosensors based on membrane transport proteins
internal reflection and only the evanescent wave was used for excitation (TIRF). The emitted light passed through a monochromator set at 510 nm and was detected by a photomultiplier as shown in Fig. 1 . For electron microscopy, the glass coverslips were frozen in liquid propane and freeze-dried in a Balzers BAF300. After shadowing with platinum! carbon (Pt/C), the replicas were floated and cleaned on hydrofluoric acid (6%) and distilled water, mounted on grids and viewed in a Philips 201 electron microscope. RESULTS AND DISCUSSION At first, a series of experiments was performed on vesicles with the pH-probe pyranine entrapped. Unilamellar vesicles of homogeneous size were prepared by the extrusion technique and showed a low permeability for protons. As seen from Fig. 2, an imposed pH gradient decays with a half-life time of about 5 min . Addition of the proton ionophore gramicidin induces a rapid breakdown of the pH gradient . These results agree with those of Grzesiek and Dencher (1986) . There was no significant difference in the decay of a pH gradient between vesicles with and without incorporated lactose permease . This indicates that in the absence of lactose the protein does not transport protons . Addition of lactose to vesicles with incorporated lactose permease leads to a fast pH decrease
ON Lactose
Grew
I
Y
4 5 TIME Loin)
Fig. 3. The time course of the fluorescence intensity after addition of lactose to vesicles containing reconstituted lactose permease and loaded with pyranine. Cotransport leads to a rapid pH decrease passive proton leakage to a slow increase in pH and addition of gramicidin to a f the initial pH. The amplitude of the steady state signal depends on the lactose concentration.
0
CD V 0.
Y
5
10
15 LACTOSE CONCENTRATION (mM)
4
20
Fig. 4. Dependence of the fluorescence signal shown in Fig. 3 on the lactose concentration. Lactose can be measured in a rangefi+om 0.1 mm up to 20 mu See textfor explanation.
0
z
G 5 TIME {min)
9
Fig. 2. The time course of the fluorescence inten 'ty after imposition of a pH gradient. Lipid vesicles without protein were prepared as described and loaded with pyranine The external pH was lowered by a2 units and the fluorescence intensity recorded. Rapid pH equlibration is induced by addition of I pu gram icidin D.
e vesicles as sho ig. 3. The steady state at the end of the transport process was reached within I min, hence, the half-life time of the transport process was even shorter . Upon addition of gramicidin the initial pH is restored . The amplitude of the initial pH change increased monotonically with the lactose concentration, as is shown in Fig . 4. The reason for the nonlinearity of the curve is as follows : the pH gradient resulting from lactose/proton cotransport is opposed to the lactose gradient which drives this transport. Transport stops at the moment where these gradients are equal, which means that the
236 concentration ratios of lactose and protons are reciprocal. At this moment the total driving force for cotransport is zero, but there is still more lactose outside than inside the vesicles . The higher the lactose concentration, the higher the opposing pH gradient and the lower the percentage of lactose transported into the vesicles. The specificity of lactose permease reconstituted in lipid vesicles is the same as in whole cells. A variety of a- and f -galactosides are transported by the protein, while other sugars such as glucose, maltose or saccharose are transported either much more slowly or not at all (Wright et al., 1981). Reconstituted lactose permease is very stable, probably because the membrane protects most of the protein from noxious water soluble substances such as proteases or heavy metal ions . As long as microbial growth is suppressed, there is no measurable loss of transport activity after one week at room temperature . Longer times have not yet been tested . Encouraged by these results experiments were started with SPBs . Figure 5 shows the electron microscopic view of a SPB which starts to detach from the glass surface at one edge . This detachment possibly occurs during processing for electron microscopy. The photograph shows that the otherwise structureless surface seen over large areas is not just the glass surface but indeed a lipid bilayer . pH jump experiments on these SPBs exhibit a slow kinetic as shown in Fig. 6. One can exclude that vesicles are just adsorbed to the quartz
H. Kiefer et al. I• a
4n
- c •
•
T C
pH7.9
o a Y W .u c • •
C O
0
pH 7.3 0
2
4
6
8
10
12
14
16
18
TIME (min)
Fig. 6. The time course of fuorescence ofpyranine entrapped between a SPB without protein and a quartz surface. The pH in theflow-through cell shown in Fig. 1 was raisedfrom 7.3 to 7.9 at the time indicated.
surface since this would have been seen in the electron micrographs . So it is concluded that the SPB is sufficiently impermeable for protons to allow its future use for lactose measurements . Preliminary experiments with SPBs containing lactose permease show a pH decrease upon addition of lactose. This indicates that such a biosensor for lactose based on a membrane transport protein may indeed function . CONCLUSIONS The principle of this sensor is not limited to lactose but can be extended to any substrate for which a cotransport protein exists . This is the case for many sugars, amino acids, neurotransmitters and small metabolites . The advantages of this sensor type are: (i) the stability of the membrane and the incorporated proteins, especially with respect to proteases ; (ii) the easy immobilization procedure ; and (iii) the independence of the signal on the amount of active protein . A disadvantage is that not many transport proteins have been sufficiently characterized so far, nor are they available in such large quantities as in the case of lactose permease . Much work remains to be done in this field . ACKNOWLEDGMENT This work was supported by the German Federal Ministry for Scientific Research and Technology (BMFT). REFERENCES
Fig 5. Vesicles spread on glass coverslips were freeze-dried and shadowed with Pt/C.
Grzesiek, S . & Dencher, N. A (1986) . Dependency of ApH-relaxation across vesicular membranes on
Biosensors based on membrane transport proteins
the buffering power of bulk solutions and lipids. Biophys. J., 50, 265-76.
Hope, M. J., Bally, M . B., Webb, G. & Cullis, P . R. (1985) . Production of large unilamellar vesicles by a rapid extrusion procedure . Characterization of size distribution, trapped volume and ability to maintain a membrane potential . Biochim . Biophys. Acta, 812, 55-65 .
McConnell, H. M., Watts, T. H., Weis, R. M. & Brian, A A (1986). Supported planar membranes in studies of cell-cell recognition he immune
237
system. Biochim . Biophys. Acta 864, 95-106. Radin, N. S. (1981). Extraction of tissue lipids with a solvent of low toxicity . Meth. EnzymoL, 75, 5-8. Wright, I K & Overath, P . (1983). Purification of the lactose: H+ carrier ofEscherichia coil and characterization of galactoside binding and transport. Eur. J Biochem., 138, 497-508 .
Wright, J. K, Riede, I . & Overath, P . (1981) . Lactose carrier protein of Escherichia coif: Interaction with galactosides and protons . Biochemistry, 20, 640415.
Biosensors based on membrane transport proteins Hans Kiefer, Barbara Klee, Edgar John, York-Dieter Stierhof & Fritz Jahnig Max-Planck-Institut fur Biologie, Corrensstrasse 38, D7400 Tubingen, FRG (Received 4 May 1990 ; revised version received 24 September 1990 ; accepted 26 September 1990)
Abstract: We propose a novel class of biosensors based on membrane bound receptors or transport proteins as the sensing element The protein is incorporated in a planar lipid bilayer which covers the transducer . The transducer may detect an electric current, a voltage, or a change in fluorescence . A prototype lactose sensor is presented which consists of a quartz slide covered by a lipid membrane containing the protein lactose permease from Escherichia coli . This protein is a lactose/H+ cotransporter, hence lactose in the external medium initiates lactose/H+ cotransport across the lipid membrane . This leads to a rise in proton concentration in the small volume between the lipid membrane and the quartz surface which can be detected by a pH-sensitive fluorescence dye. Keywords: biosensor, lactose/H+ cotransport, supported membrane, TIRF .
INTRODUCTION
function in nature can also be used as receptors in an artificial system . If the solute is present on one side of the membrane, cotransport will start and produce an ionic current which can be detected as described above. This is why these transport proteins are subsumed under membrane receptors in this paper . For practical use, the following problem must be considered . If a current across the membrane should be detected, the membrane must separate an external compartment containing the analyte from an `internal' compartment in contact with the transducer . This separation has to be leakagefree compared to the current to be measured. Three artificial lipid membrane systems are commonly used for the measurement of ionic currents: (i) the vesicle system is restricted to spectroscopic observations of dyes entrapped in the inner volume, since this volume is much too small to contain any macroscopic detection
Receptors are defined as proteins which produce a signal in the cell upon binding an external substrate . This signal can be the release of a second messenger, the activation of a G-protein, or an ionic current across the membrane induced by opening a gated ion channel . The latter case suggests the use of such proteins as sensing elements for biosensors, because an ionic current can be detected by various transducers . The current may be measured directly (amperometric detection), the resulting membrane potential may be detected (potentiometric detection) or, finally, a change in ion concentration may be measured with ion-sensitive electrodes, ISFETs or spectroscopically by ion-sensitive probes. It should be noted in this context that ion/ solute cotransporters that have a totally different 233 Biosensors & Bioelectronics
0956-5663/91/$03 .50 0
1991 Elsevier Science Publishers Lid, England . Printed in Great Britain
234 device; (ii) the black lipid membrane (BLM) system can be used for any form of detection, since the two compartments that are separated by the BLM can be chosen as large as desired . However, a BLM is mechanically very instable which restricts its use for a biosensor drastically; (iii) a supported planar bilayer (SPB) seems most promising, since it separates two macroscopic volumes on the one hand and, on the other hand, is extremely stable even over a long time . However, as the distance between membrane and support is only about 2 nm, the entrapped space is too small to contain a macroscopic detection device. Therefore, the SPB must be spread directly on the transducer surface. A prototype lactose sensor is currently tested which uses the lactose permease from Escherichia coli as the sensing element. This protein is a cotransporter for lactose and H +. The purified and reconstituted protein is incorporated in a SPB spread on the surface of a quartz slide . Changes of protein concentration in the space between SPB and quartz surface are monitored by measuring the fluorescence of a pH-sensitive probe entrapped in this volume.
H. Kiefer et al.
For measurements on the vesicle system, vesicles were passed over a sephacryl S-100 HR (Pharmacia) column to remove external pyranine. Valinomycin (Serva, 50 nM), was added . Measurements were carried out on a Perkin-Elmer MPF-3 spectrofluorimeter connected to a chart recorder. The excitation wavelength was 460 nm; the emission wavelength, 516 nm . pH gradients were produced by adding a small amount of 100 mm H2 SO4 to the solution . Lactose transport was initiated by adding various amounts of 0. 5 M lactose dissolved in the above buffer . For SPB measurements, vesicles were prepared in the presence of pyranine and spread on the surface of a pre-treated quartz slide (Hellma) yielding a planar bilayer as described in McConnell et al. (1986). The slide was part of a flow-through cell shown schematically in Fig. 1 . External pyranine was removed by passing buffer through the cell . Fluorescence of pyranine entrapped between the slide and the planar bilayer was excited by the 458 nm line of an argon ion laser (Spectra Physics, Model 2016). The laser beam was guided inside the slide under total
PM
MATERIALS AND METHODS All chemicals not further specified were from Merck (Darmstadt, F.R.G.). E. coli phospholipid was exctracted from wild-type E. coli K12 as described in Radin (1981) and purified over a silica gel column using a chloroform/methanol gradient Lactose permease was purified from the overproducing strain E. coli T206 and reconstituted into vesicles of E. coli lipid at a lipid/protein molar ratio of 4000 as described by Wright and Overath (1983). Vesicles were freeze-thawed three times in 10 mm potassium phosphate buffer pH 7 . 3, 100 mm K2 SO4, I mm DTT (Sigma), 1 mm pyranine (8-hydroxipyrene-1,3,6-trisulfonic acid, Molecular Probes Eugene, OR, U.S.A.). Subsequently, they were brought to a homogeneous size by the extrusion method (Hope et al., 1985). The pore sizes of the nucleopore filters were 400, 200, 100 and 50 nm . This method provides predominantly unilamellar vesicles . Control vesicles without lactose permease were made starting from a lipid-detergent mixture equal to the mixture used for protein reconstitution, treating this mixture in the same way .
L LM
Fig. 1. Schematic illustration ofthe experimental set-up for the TIRF measurements. The beam of a laser (L) is coupled into the quartz slide (S) via a mirror and a prism (P). The prism is optically connected to the slide by immersion oil. The quartz slide is clamped on a block made of Plexiglass (B). The 0-ring (R) between quartz slide and Pleriglass serves to seal a volume of about 0 .1 ml providing the sample cell. The inlet of the cell is connected to a peristaltic pump . The lipid membrane (LM) establishes on the surface of the quartz slides upon ,hushing through the cell a suspension of vesicles. The fluorescence light is detected by an objective lense (O) with 10-fold magnification, a monochromator (M) and a photomultiplier (PM). Both monochromator and photomultiplieraremounted on a conventional microscope (Leitz) which also allows visual inspection of the probe.
235
Biosensors based on membrane transport proteins
internal reflection and only the evanescent wave was used for excitation (TIRF). The emitted light passed through a monochromator set at 510 nm and was detected by a photomultiplier as shown in Fig. 1 . For electron microscopy, the glass coverslips were frozen in liquid propane and freeze-dried in a Balzers BAF300. After shadowing with platinum! carbon (Pt/C), the replicas were floated and cleaned on hydrofluoric acid (6%) and distilled water, mounted on grids and viewed in a Philips 201 electron microscope. RESULTS AND DISCUSSION At first, a series of experiments was performed on vesicles with the pH-probe pyranine entrapped. Unilamellar vesicles of homogeneous size were prepared by the extrusion technique and showed a low permeability for protons. As seen from Fig. 2, an imposed pH gradient decays with a half-life time of about 5 min . Addition of the proton ionophore gramicidin induces a rapid breakdown of the pH gradient . These results agree with those of Grzesiek and Dencher (1986) . There was no significant difference in the decay of a pH gradient between vesicles with and without incorporated lactose permease . This indicates that in the absence of lactose the protein does not transport protons . Addition of lactose to vesicles with incorporated lactose permease leads to a fast pH decrease
ON Lactose
Grew
I
Y
4 5 TIME Loin)
Fig. 3. The time course of the fluorescence intensity after addition of lactose to vesicles containing reconstituted lactose permease and loaded with pyranine. Cotransport leads to a rapid pH decrease passive proton leakage to a slow increase in pH and addition of gramicidin to a f the initial pH. The amplitude of the steady state signal depends on the lactose concentration.
0
CD V 0.
Y
5
10
15 LACTOSE CONCENTRATION (mM)
4
20
Fig. 4. Dependence of the fluorescence signal shown in Fig. 3 on the lactose concentration. Lactose can be measured in a rangefi+om 0.1 mm up to 20 mu See textfor explanation.
0
z
G 5 TIME {min)
9
Fig. 2. The time course of the fluorescence inten 'ty after imposition of a pH gradient. Lipid vesicles without protein were prepared as described and loaded with pyranine The external pH was lowered by a2 units and the fluorescence intensity recorded. Rapid pH equlibration is induced by addition of I pu gram icidin D.
e vesicles as sho ig. 3. The steady state at the end of the transport process was reached within I min, hence, the half-life time of the transport process was even shorter . Upon addition of gramicidin the initial pH is restored . The amplitude of the initial pH change increased monotonically with the lactose concentration, as is shown in Fig . 4. The reason for the nonlinearity of the curve is as follows : the pH gradient resulting from lactose/proton cotransport is opposed to the lactose gradient which drives this transport. Transport stops at the moment where these gradients are equal, which means that the
236 concentration ratios of lactose and protons are reciprocal. At this moment the total driving force for cotransport is zero, but there is still more lactose outside than inside the vesicles . The higher the lactose concentration, the higher the opposing pH gradient and the lower the percentage of lactose transported into the vesicles. The specificity of lactose permease reconstituted in lipid vesicles is the same as in whole cells. A variety of a- and f -galactosides are transported by the protein, while other sugars such as glucose, maltose or saccharose are transported either much more slowly or not at all (Wright et al., 1981). Reconstituted lactose permease is very stable, probably because the membrane protects most of the protein from noxious water soluble substances such as proteases or heavy metal ions . As long as microbial growth is suppressed, there is no measurable loss of transport activity after one week at room temperature . Longer times have not yet been tested . Encouraged by these results experiments were started with SPBs . Figure 5 shows the electron microscopic view of a SPB which starts to detach from the glass surface at one edge . This detachment possibly occurs during processing for electron microscopy. The photograph shows that the otherwise structureless surface seen over large areas is not just the glass surface but indeed a lipid bilayer . pH jump experiments on these SPBs exhibit a slow kinetic as shown in Fig. 6. One can exclude that vesicles are just adsorbed to the quartz
H. Kiefer et al. I• a
4n
- c •
•
T C
pH7.9
o a Y W .u c • •
C O
0
pH 7.3 0
2
4
6
8
10
12
14
16
18
TIME (min)
Fig. 6. The time course of fuorescence ofpyranine entrapped between a SPB without protein and a quartz surface. The pH in theflow-through cell shown in Fig. 1 was raisedfrom 7.3 to 7.9 at the time indicated.
surface since this would have been seen in the electron micrographs . So it is concluded that the SPB is sufficiently impermeable for protons to allow its future use for lactose measurements . Preliminary experiments with SPBs containing lactose permease show a pH decrease upon addition of lactose. This indicates that such a biosensor for lactose based on a membrane transport protein may indeed function . CONCLUSIONS The principle of this sensor is not limited to lactose but can be extended to any substrate for which a cotransport protein exists . This is the case for many sugars, amino acids, neurotransmitters and small metabolites . The advantages of this sensor type are: (i) the stability of the membrane and the incorporated proteins, especially with respect to proteases ; (ii) the easy immobilization procedure ; and (iii) the independence of the signal on the amount of active protein . A disadvantage is that not many transport proteins have been sufficiently characterized so far, nor are they available in such large quantities as in the case of lactose permease . Much work remains to be done in this field . ACKNOWLEDGMENT This work was supported by the German Federal Ministry for Scientific Research and Technology (BMFT). REFERENCES
Fig 5. Vesicles spread on glass coverslips were freeze-dried and shadowed with Pt/C.
Grzesiek, S . & Dencher, N. A (1986) . Dependency of ApH-relaxation across vesicular membranes on
Biosensors based on membrane transport proteins
the buffering power of bulk solutions and lipids. Biophys. J., 50, 265-76.
Hope, M. J., Bally, M . B., Webb, G. & Cullis, P . R. (1985) . Production of large unilamellar vesicles by a rapid extrusion procedure . Characterization of size distribution, trapped volume and ability to maintain a membrane potential . Biochim . Biophys. Acta, 812, 55-65 .
McConnell, H. M., Watts, T. H., Weis, R. M. & Brian, A A (1986). Supported planar membranes in studies of cell-cell recognition he immune
237
system. Biochim . Biophys. Acta 864, 95-106. Radin, N. S. (1981). Extraction of tissue lipids with a solvent of low toxicity . Meth. EnzymoL, 75, 5-8. Wright, I K & Overath, P . (1983). Purification of the lactose: H+ carrier ofEscherichia coil and characterization of galactoside binding and transport. Eur. J Biochem., 138, 497-508 .
Wright, J. K, Riede, I . & Overath, P . (1981) . Lactose carrier protein of Escherichia coif: Interaction with galactosides and protons . Biochemistry, 20, 640415.
