[65]
SENSITIZED PHOTOOXIDATION
561
due modified in this inactivation process has a I)K~ greater than 9.0. Dramatic differences are observed in the substrate protection profile for 5b vs 5a. Of particular interest is the observation that the enzyme can be partially protected from inactivation by 5b when S-adenosylmethionine is included in the preincubation mixture. This result is in sharp contrast to the rate enhancement effect that S-adenosylmethionine has on enzyme inactivation by 5a. The S-adenosylmethionine protection by 5b suggests that when this ligand binds to the enzyme there is either a physical protection of the nucleophile or a conformational change of the enzyme that reduces accessibility of the protein nucleophile to the affinity labeling reagent. These basic differences in the properties of the affinity labeling reagents 5a and 5b would suggest that the two classes of affinity labeling reagents are modifying different nucleophilic residues on the enzyme. The different modes of binding of these reagents (5a and 5b) to catechol-Omethyltransferase, as depicted in Fig. 1, are probably responsible for the interaction of these ligands with different protein nucleophiles.
Acknowledgments This investigation was supported by a Research Grant from the National Institute of Neurological Diseases and Stroke (NS-10918). RTB gratefully acknowledges support by the American Heart Association for an Established Investigatorship.
[65] Affinity Labeling of Binding Sites in Proteins by Sensitized Photooxidation By JOHNNY BRANDT Affinity labeling of ligand-binding sites in proteins is an important tool in the study of their structure-function relationship. Precise localization of amino acid residues near or in the active site can in this way be obtained by degrading the protein and identifying the labeled fragments. Several review articles have been devoted to this subject, and a large number of different active-site-directed reagents have been described. '-3 A new approach to such affinity labeling of binding sites in proteins is to use photooxidative coupling reactions for the attachment of S. J. Singer, Adv. Protein Chem. 22, 1 (1967). "~E. Shaw, in "The Enzymes" (P. Boyer, ed.), 3rd ed., Vol. 1, p. 91. Academic Press, New York, 1970. ~J. R. Knowles, Ace. Chem. Res. 5, 155 (1972). See also this volume [8].
562
ENZYMES, ANTIBODIES, AND OTHER PROTEINS
[55]
the ligand. This technique seems to offer considerable promise. 4 This report presents the principles involved, the general procedure, and some results obtained in studies with a model system.
Principles Dye-sensitized photooxidation ~,~ is a well-known method which has been used to modify certain amino acid residues in proteins. In this technique, a mixture of a dye and the protein under study is irradiated in the presence of oxygen at wavelengths at which only the dye can absorb light. In this way one can, under special conditions, accomplish very specific chemical modification of one or several amino acid residues in proteins. By using a ligand as the sensitizer, modification can be obtained preferentially at the ligand-binding site of the protein. 7 The amino acid modifications previously described using sensitized photooxidation are, however, relatively minor and include such systems as the oxidation of methionine to methionine sulfoxide or of tryptophan to kynurenine. Such modification can be difficult to detect and result in difficulties in subsequent separation required to locate the modified residues. However, we have found that by using a high concentration of dye and a rather high light intensity one obtains not only some modification of amino acid residues, but also covalent coupling of the dye to the protein. This observation suggests a new, simple method for affinity labeling of protein ligand-binding sites. The method can easily be applied to the detection of dye-binding sites in proteins, but can also be extended to other ligands by using dye-ligand conjugates as affinity-labeling reagents. In Fig. 1 are shown the structural formulas of some of the dyes that have been successfully coupled to biomolecules using sensitized photooxidation. I stress that the coupling reaction is not restricted to proteins but can occur as well with ribonucleic acid and carbohydrate. The presence of oxygen is not necessary for the coupling of dye but strongly enhances the reaction. Although detailed information about the mechanism of the dye-protein coupling reaction is unavailable, a general outline can be given. When irradiated by light the dye undergoes electronic excitation followed by radical formation. The radicals formed can react and couple to the protein, probably as the result of radical-radical reactions. By using a *J. Brandt, M. Fredriksson, and L.-O. Andersson, Biochemislry 13, 4758 (1974). 5 W. J. Ray, Jr., this series, Vol. 11, p. 490 (1967). E. W. Westhead, this series, Vol. 25, p. 401 (1972). ' G. Gennari, G. Jori, G. Galiazzo, and E. Scoffone, J. Am. Chem. Soc. 92, 4140 (1970).
[65]
SENSITIZED PHOTOOXIDATION
563
HO~ 0 - ~ O Fluorescein _~..~- COOH Br
Bromphenol blue
Acridine orange
blue
Br
B r ~.~,~,~. C~ , ~ . , . ~ B r
~_S03H
(CH3)2N~N(CH3)2
b'2N' S " CH3
C,
FIa. 1. Structural formulas of some dyes effective in photooxidative coupling reactions. dye or a dye-ligand conjugate as a sensitizer with affinity for a protein binding site, coupling to amino acid residues in the binding site should be favored, because of the localization and high concentration of the dye or dye-ligand conjugate in the binding site. Specific labeling of binding sites for a dye or a dye-ligand conjugate in proteins should thus be accomplished. Methods
Photooxidative coupling is fairly easy to perform. The only special equipment required is a suitable irradiation source that gives a high intensity at those wavelengths that are readily absorbed by the dye. For the model experiment presented here, a high-pressure mercury lamp (Gates 420-U2, 250 watts) was used, but a tungsten lamp would have been satisfactory. A typical experimental arrangement is illustrated in Fig. 2. The sample solution is contained in an open, flat-bottom beaker (d = 15 cm) at a distance of 20 cm from the irradiation source. A filter is inserted between the lamp and the sample solution to absorb the infrared irradiation and the wavelengths absorbed by the protein; a flat-bottomed beaker of glass containing 0.5 M sodium nitrite is suitable (lower wavelength cutoff is about 400 rim). The temperature of the sample is controlled by a water bath connected to a thermostat, or more simply, by keeping the sample in an ice bath. A stirring motor is placed under the water bath for a gentle driving of a long magnetic stirrer that
564
ENZYMES, ANTIBODIES, .AND OTHER PROTEINS
[55]
Irradiation source
Reaction vessel
~ter bath
Motor l u r ~ J__ magnetic stirrer FIG. 2. Schematic illustration of simple experimental arrangements suitable for the photooxidative coupling method. is in the sample solution. When the volume of sample is 50 to 100 ml, the degree of coupling will be about 1 mg of dye per gram of protein after irradiation for 1 hr. Longer irradiation periods should be avoided because the protein denaturation that often accompanies modification by photooxidation processes increases the risk of unspecific labeling. The degree of coupling of dye can be determined after removal of excess dye from a known amount of protein by gel filtration on a Sephadex G-25 column. Usually dyes are strongly adsorbed to the gel and a rapid and efficient separation of free from protein-bound dye is easily achieved. The concentration of protein-bound dye is then determined spectrophotometrically in the visible region. It is worth noting that the strong adsorption of dyes to dense Sephadex gels greatly facilitates the isolation of the labeled peptides after fragmentation of the protein. Small labeled peptides are easily fractionated and separated from unlabeled peptides by Sephadex G-25 because of the strong adsorption of the dye moiety to the gel.
Model System For our pilot studies of the specificity of the dye-sensitized photooxidative coupling reaction we have, for several reasons, chosen the coupling of fluorescein to bovine serum albumin. The first advantage of this system is that the photochemistry of fluorescein s-ll is well known. s L. Lindqvist, Ark. Kemi 16, 79 (1960). ° L. Lindqvist, J. Phys, Chem. 67, 1701 (1963). loV. Kasche and L. Lindqvist, d. Phys. Chem. 68, 817 (1964). "V. Kasche, Photochem. Photobiol. 6, 643 (1967).
[55]
SENSITIZED PHOTOOXIDATION
565
However, the main reason for our choice was that the binding of fluorescein to bovine seurm albumin L° has been extensively studied in our laboratory and we have considerable experience in the isolation of peptides from digests of bovine serum albumin.
Application Preparation o] Reagents Fluorescein (G. T. Gurr, London) is purified by precipitation from 0.01 M NaOH solution with acetic acid. The procedure is repeated three times. A stock solution of 6 mM fluorescein is then prepared at p H l l . Bovine serum albumin (Statens Bakteriologiska Laboratorium, Stockholm), about 100 mg/ml in 0.1 M Tris chloride at pH 8.0, is extensively dialyzed against water. A 5% stock solution is prepared.
Affinity Labelinq A 1% albumin solution containing 1 mM fluorescein is prepared by mixing 10 ml of 6 mM fluorescein and 12 ml of 5% albumin with 38 ml of 30 mM sodimn phosphate at pH 7.7 to give a final pH of 8.0. The mixture is illuminated for 60 rain at 20 ° using the apparatus described above. Excess fluorescein is removed by gel filtration on a Sephadex G25 column (3.2 X 30 era) in 0.1 M ammonia-ammonium acetate at pH 9.2. The green color of the free dye changes to red upon coupling. The deep red color of the protein fraction eluting at the void volume of the column thus serves as a visible check on successful labeling. The degree of coupling, based upon spectrophotometric measurement, is about 0.20 mole of dye per mole of protein. The labeled protein fraction is concentrated by ultrafiltration after adjustment to pH 8 with neutral 1 M sodium phosphate.
Isolation of Labeled Peptides The labeled albumin fraction (600 mg in 8.5 ml) is extensively digested with trypsin (36 rag) at pH 7.7 to 7.9. The tryptic digest is subjected to gel filtration on a Sephadex G-50 column (3.2 X 90 cm) in 0,1 M ammonia-ammoniunl acetate at pH 9.2. The main part of the labeled peptides is eluted at the total volume of the column. This peptide fraction is further fraetionated by ion-exchange chromatography on a SE1,~L.-O. Andersson, A. RehnstrSm, and D. L. Eaker, Eur. J. Biochem. 20, 371 (1971).
566
ENZYMES, ANTIBODIES, AND OTHER PROTEINS
[55]
Sephadex column (1.6 X 40 cm) using pH gradient elution. The gel is equilibrated with 0.1 M acetic acid-0.1 M formic acid buffer, pH 3.0, and the peptides are eluted by increasing the pH of the eluent with 0.4 M ammonia. One main peptide labeled with fluorescein is eluted at pH 4.3. A complete purification of the labeled peptide is performed by gel filtration on a Sephadex G-25 column (1.5 X 40 cm) in 0.1 M ammoniaammonium acetate at pH 9.2.
Characterization o] Labeled Peptide Amino acid analysis after acid hydrolysis of the pure peptide yields leucine and tyrosine in a 1:1 molar ratio. The peptide material (100 nmoles) is digested with carboxypeptidase A (10 ~g) at pH 8.3 and 37 ° for 2 hr. The digestion mixture is then passed through a Sephadex G-25 column (1.5 X 30 cm) in 0.1 M ammonia-ammonium acetate at pH 9.2. A complete separation of the peptide fragments is obtained. On the basis of N-terminal and amino acid analysis data on the peptide fragments, and from the known sequence data of bovine serum albumin, 13 it has been concluded ~ that the labeled peptide is the Tyr-Leu-Tyr sequence of residues 137-139 with the fluorescein molecule attached to Tyr-137.
Control Experiment The specificity of the coupling reaction can be checked in a control experiment by labeling the protein in denaturing medium. The conditions were the same as in the model experiment described above except that the solution was 6 M in urea during irradiation. The degree of coupling of fluorescein to albumin was about the same as that observed in the absence of urea. The labeled protein was digested with trypsin and gel filtered on Sephadex G-150, the low molecular weight peptides were separated by ion-exchange chromatography on SE-Sephadex as described above. The elution curve obtained showed that labeling is much less specific under denaturing conditions and that no major labeled peptide was found.
Conclusions The studies on the model system described show that the photooxidarive coupling reaction results in a covalent bond between fluorescein and albumin. The labeled protein has a spectral absorption maximum at 505 is j . R. Brown, Fed. Proc., Fed. Am. Soc. Exp. Biol. 34, 591 (1975).
[65]
BROMOCOLCHICINE FOR TUBULIN
567
nm compared to 499 nm for the fluorescein adduct obtained after degradation. This spectral shift suggests that the fluorescein molecule is in a hydrophobic region within the intact albumin molecule. We also know from earlier studies'" that bovine serum albumin has hydrophobic binding sites for fluorescein that cause a red shift in the absorption maximum upon binding. Moreover, the separation of peptides after degradation of labeled protein shows selective labeling of a single peptide. The specificity for labeling of this peptide is lost if coupling is performed under denaturing conditions. It thus seems probable that the main labeled peptide obtained derives from a fluorescein-binding site in the albumin molecule. Comments
Although the pilot studies described in this report indicate that the principle is valid, affinity labeling of binding sites in proteins using sensitized photooxidation cannot yet be considered a standard method. The approach may form a valuable complement to existing affinity labeling methods, although more information is needed. Further developmental work is now in progress with protein-dye systems for which detailed knowledge of the active-site structure of the protein is available.
[65] B r o m o c o l c h i c i n e a s a L a b e l f o r T u b u l i n B y DAPHNE ATLAS and HENRI SCHMITT
Microtubules are found in all eukaryotic cells, wherein they participate in a wide variety of functions including mitosis, cell shaping, secretion, motility, axonal growth, and transport. 1,~ Microtubules are polymers of a protein called tubulin, which is itself a dimer composed of two similar but not identical subunits (o~, fl), each of molecular weight 55,0002 Drugs, such as colchicine, that inhibit mitosis and axonal function through disruption of microtubules have been shown to interact in vitro with tubulin isolated from several sources. 4 Knowledge of the precise localization and properties of these drug receptors would contribute to the better understanding of the mechanism of tubulin assembly to form microtubules. 1j. B. Olmsted and G. G. Borisy, Annu. Rev. Biochem. 42, 507 (1973). M. L. Shelanski and H. Feit, in "The :Structure and Function of Nervous Tissue" (G. H. Bourne, ed.), ¥ol. 6, pp. 47-80. Academic Press, New York, 1972. 3R. E. Stephens, in "Subunits in Biological Systems" (S. N. Timasheff and " G. D. Fasman, eds.), Part A, pp. 355--391. Dekker, New York, 1971. 4L. Wilson and J. Bryan, Adv. Cell Mol. Biol. 3, 21 (1974).
SENSITIZED PHOTOOXIDATION
561
due modified in this inactivation process has a I)K~ greater than 9.0. Dramatic differences are observed in the substrate protection profile for 5b vs 5a. Of particular interest is the observation that the enzyme can be partially protected from inactivation by 5b when S-adenosylmethionine is included in the preincubation mixture. This result is in sharp contrast to the rate enhancement effect that S-adenosylmethionine has on enzyme inactivation by 5a. The S-adenosylmethionine protection by 5b suggests that when this ligand binds to the enzyme there is either a physical protection of the nucleophile or a conformational change of the enzyme that reduces accessibility of the protein nucleophile to the affinity labeling reagent. These basic differences in the properties of the affinity labeling reagents 5a and 5b would suggest that the two classes of affinity labeling reagents are modifying different nucleophilic residues on the enzyme. The different modes of binding of these reagents (5a and 5b) to catechol-Omethyltransferase, as depicted in Fig. 1, are probably responsible for the interaction of these ligands with different protein nucleophiles.
Acknowledgments This investigation was supported by a Research Grant from the National Institute of Neurological Diseases and Stroke (NS-10918). RTB gratefully acknowledges support by the American Heart Association for an Established Investigatorship.
[65] Affinity Labeling of Binding Sites in Proteins by Sensitized Photooxidation By JOHNNY BRANDT Affinity labeling of ligand-binding sites in proteins is an important tool in the study of their structure-function relationship. Precise localization of amino acid residues near or in the active site can in this way be obtained by degrading the protein and identifying the labeled fragments. Several review articles have been devoted to this subject, and a large number of different active-site-directed reagents have been described. '-3 A new approach to such affinity labeling of binding sites in proteins is to use photooxidative coupling reactions for the attachment of S. J. Singer, Adv. Protein Chem. 22, 1 (1967). "~E. Shaw, in "The Enzymes" (P. Boyer, ed.), 3rd ed., Vol. 1, p. 91. Academic Press, New York, 1970. ~J. R. Knowles, Ace. Chem. Res. 5, 155 (1972). See also this volume [8].
562
ENZYMES, ANTIBODIES, AND OTHER PROTEINS
[55]
the ligand. This technique seems to offer considerable promise. 4 This report presents the principles involved, the general procedure, and some results obtained in studies with a model system.
Principles Dye-sensitized photooxidation ~,~ is a well-known method which has been used to modify certain amino acid residues in proteins. In this technique, a mixture of a dye and the protein under study is irradiated in the presence of oxygen at wavelengths at which only the dye can absorb light. In this way one can, under special conditions, accomplish very specific chemical modification of one or several amino acid residues in proteins. By using a ligand as the sensitizer, modification can be obtained preferentially at the ligand-binding site of the protein. 7 The amino acid modifications previously described using sensitized photooxidation are, however, relatively minor and include such systems as the oxidation of methionine to methionine sulfoxide or of tryptophan to kynurenine. Such modification can be difficult to detect and result in difficulties in subsequent separation required to locate the modified residues. However, we have found that by using a high concentration of dye and a rather high light intensity one obtains not only some modification of amino acid residues, but also covalent coupling of the dye to the protein. This observation suggests a new, simple method for affinity labeling of protein ligand-binding sites. The method can easily be applied to the detection of dye-binding sites in proteins, but can also be extended to other ligands by using dye-ligand conjugates as affinity-labeling reagents. In Fig. 1 are shown the structural formulas of some of the dyes that have been successfully coupled to biomolecules using sensitized photooxidation. I stress that the coupling reaction is not restricted to proteins but can occur as well with ribonucleic acid and carbohydrate. The presence of oxygen is not necessary for the coupling of dye but strongly enhances the reaction. Although detailed information about the mechanism of the dye-protein coupling reaction is unavailable, a general outline can be given. When irradiated by light the dye undergoes electronic excitation followed by radical formation. The radicals formed can react and couple to the protein, probably as the result of radical-radical reactions. By using a *J. Brandt, M. Fredriksson, and L.-O. Andersson, Biochemislry 13, 4758 (1974). 5 W. J. Ray, Jr., this series, Vol. 11, p. 490 (1967). E. W. Westhead, this series, Vol. 25, p. 401 (1972). ' G. Gennari, G. Jori, G. Galiazzo, and E. Scoffone, J. Am. Chem. Soc. 92, 4140 (1970).
[65]
SENSITIZED PHOTOOXIDATION
563
HO~ 0 - ~ O Fluorescein _~..~- COOH Br
Bromphenol blue
Acridine orange
blue
Br
B r ~.~,~,~. C~ , ~ . , . ~ B r
~_S03H
(CH3)2N~N(CH3)2
b'2N' S " CH3
C,
FIa. 1. Structural formulas of some dyes effective in photooxidative coupling reactions. dye or a dye-ligand conjugate as a sensitizer with affinity for a protein binding site, coupling to amino acid residues in the binding site should be favored, because of the localization and high concentration of the dye or dye-ligand conjugate in the binding site. Specific labeling of binding sites for a dye or a dye-ligand conjugate in proteins should thus be accomplished. Methods
Photooxidative coupling is fairly easy to perform. The only special equipment required is a suitable irradiation source that gives a high intensity at those wavelengths that are readily absorbed by the dye. For the model experiment presented here, a high-pressure mercury lamp (Gates 420-U2, 250 watts) was used, but a tungsten lamp would have been satisfactory. A typical experimental arrangement is illustrated in Fig. 2. The sample solution is contained in an open, flat-bottom beaker (d = 15 cm) at a distance of 20 cm from the irradiation source. A filter is inserted between the lamp and the sample solution to absorb the infrared irradiation and the wavelengths absorbed by the protein; a flat-bottomed beaker of glass containing 0.5 M sodium nitrite is suitable (lower wavelength cutoff is about 400 rim). The temperature of the sample is controlled by a water bath connected to a thermostat, or more simply, by keeping the sample in an ice bath. A stirring motor is placed under the water bath for a gentle driving of a long magnetic stirrer that
564
ENZYMES, ANTIBODIES, .AND OTHER PROTEINS
[55]
Irradiation source
Reaction vessel
~ter bath
Motor l u r ~ J__ magnetic stirrer FIG. 2. Schematic illustration of simple experimental arrangements suitable for the photooxidative coupling method. is in the sample solution. When the volume of sample is 50 to 100 ml, the degree of coupling will be about 1 mg of dye per gram of protein after irradiation for 1 hr. Longer irradiation periods should be avoided because the protein denaturation that often accompanies modification by photooxidation processes increases the risk of unspecific labeling. The degree of coupling of dye can be determined after removal of excess dye from a known amount of protein by gel filtration on a Sephadex G-25 column. Usually dyes are strongly adsorbed to the gel and a rapid and efficient separation of free from protein-bound dye is easily achieved. The concentration of protein-bound dye is then determined spectrophotometrically in the visible region. It is worth noting that the strong adsorption of dyes to dense Sephadex gels greatly facilitates the isolation of the labeled peptides after fragmentation of the protein. Small labeled peptides are easily fractionated and separated from unlabeled peptides by Sephadex G-25 because of the strong adsorption of the dye moiety to the gel.
Model System For our pilot studies of the specificity of the dye-sensitized photooxidative coupling reaction we have, for several reasons, chosen the coupling of fluorescein to bovine serum albumin. The first advantage of this system is that the photochemistry of fluorescein s-ll is well known. s L. Lindqvist, Ark. Kemi 16, 79 (1960). ° L. Lindqvist, J. Phys, Chem. 67, 1701 (1963). loV. Kasche and L. Lindqvist, d. Phys. Chem. 68, 817 (1964). "V. Kasche, Photochem. Photobiol. 6, 643 (1967).
[55]
SENSITIZED PHOTOOXIDATION
565
However, the main reason for our choice was that the binding of fluorescein to bovine seurm albumin L° has been extensively studied in our laboratory and we have considerable experience in the isolation of peptides from digests of bovine serum albumin.
Application Preparation o] Reagents Fluorescein (G. T. Gurr, London) is purified by precipitation from 0.01 M NaOH solution with acetic acid. The procedure is repeated three times. A stock solution of 6 mM fluorescein is then prepared at p H l l . Bovine serum albumin (Statens Bakteriologiska Laboratorium, Stockholm), about 100 mg/ml in 0.1 M Tris chloride at pH 8.0, is extensively dialyzed against water. A 5% stock solution is prepared.
Affinity Labelinq A 1% albumin solution containing 1 mM fluorescein is prepared by mixing 10 ml of 6 mM fluorescein and 12 ml of 5% albumin with 38 ml of 30 mM sodimn phosphate at pH 7.7 to give a final pH of 8.0. The mixture is illuminated for 60 rain at 20 ° using the apparatus described above. Excess fluorescein is removed by gel filtration on a Sephadex G25 column (3.2 X 30 era) in 0.1 M ammonia-ammonium acetate at pH 9.2. The green color of the free dye changes to red upon coupling. The deep red color of the protein fraction eluting at the void volume of the column thus serves as a visible check on successful labeling. The degree of coupling, based upon spectrophotometric measurement, is about 0.20 mole of dye per mole of protein. The labeled protein fraction is concentrated by ultrafiltration after adjustment to pH 8 with neutral 1 M sodium phosphate.
Isolation of Labeled Peptides The labeled albumin fraction (600 mg in 8.5 ml) is extensively digested with trypsin (36 rag) at pH 7.7 to 7.9. The tryptic digest is subjected to gel filtration on a Sephadex G-50 column (3.2 X 90 cm) in 0,1 M ammonia-ammoniunl acetate at pH 9.2. The main part of the labeled peptides is eluted at the total volume of the column. This peptide fraction is further fraetionated by ion-exchange chromatography on a SE1,~L.-O. Andersson, A. RehnstrSm, and D. L. Eaker, Eur. J. Biochem. 20, 371 (1971).
566
ENZYMES, ANTIBODIES, AND OTHER PROTEINS
[55]
Sephadex column (1.6 X 40 cm) using pH gradient elution. The gel is equilibrated with 0.1 M acetic acid-0.1 M formic acid buffer, pH 3.0, and the peptides are eluted by increasing the pH of the eluent with 0.4 M ammonia. One main peptide labeled with fluorescein is eluted at pH 4.3. A complete purification of the labeled peptide is performed by gel filtration on a Sephadex G-25 column (1.5 X 40 cm) in 0.1 M ammoniaammonium acetate at pH 9.2.
Characterization o] Labeled Peptide Amino acid analysis after acid hydrolysis of the pure peptide yields leucine and tyrosine in a 1:1 molar ratio. The peptide material (100 nmoles) is digested with carboxypeptidase A (10 ~g) at pH 8.3 and 37 ° for 2 hr. The digestion mixture is then passed through a Sephadex G-25 column (1.5 X 30 cm) in 0.1 M ammonia-ammonium acetate at pH 9.2. A complete separation of the peptide fragments is obtained. On the basis of N-terminal and amino acid analysis data on the peptide fragments, and from the known sequence data of bovine serum albumin, 13 it has been concluded ~ that the labeled peptide is the Tyr-Leu-Tyr sequence of residues 137-139 with the fluorescein molecule attached to Tyr-137.
Control Experiment The specificity of the coupling reaction can be checked in a control experiment by labeling the protein in denaturing medium. The conditions were the same as in the model experiment described above except that the solution was 6 M in urea during irradiation. The degree of coupling of fluorescein to albumin was about the same as that observed in the absence of urea. The labeled protein was digested with trypsin and gel filtered on Sephadex G-150, the low molecular weight peptides were separated by ion-exchange chromatography on SE-Sephadex as described above. The elution curve obtained showed that labeling is much less specific under denaturing conditions and that no major labeled peptide was found.
Conclusions The studies on the model system described show that the photooxidarive coupling reaction results in a covalent bond between fluorescein and albumin. The labeled protein has a spectral absorption maximum at 505 is j . R. Brown, Fed. Proc., Fed. Am. Soc. Exp. Biol. 34, 591 (1975).
[65]
BROMOCOLCHICINE FOR TUBULIN
567
nm compared to 499 nm for the fluorescein adduct obtained after degradation. This spectral shift suggests that the fluorescein molecule is in a hydrophobic region within the intact albumin molecule. We also know from earlier studies'" that bovine serum albumin has hydrophobic binding sites for fluorescein that cause a red shift in the absorption maximum upon binding. Moreover, the separation of peptides after degradation of labeled protein shows selective labeling of a single peptide. The specificity for labeling of this peptide is lost if coupling is performed under denaturing conditions. It thus seems probable that the main labeled peptide obtained derives from a fluorescein-binding site in the albumin molecule. Comments
Although the pilot studies described in this report indicate that the principle is valid, affinity labeling of binding sites in proteins using sensitized photooxidation cannot yet be considered a standard method. The approach may form a valuable complement to existing affinity labeling methods, although more information is needed. Further developmental work is now in progress with protein-dye systems for which detailed knowledge of the active-site structure of the protein is available.
[65] B r o m o c o l c h i c i n e a s a L a b e l f o r T u b u l i n B y DAPHNE ATLAS and HENRI SCHMITT
Microtubules are found in all eukaryotic cells, wherein they participate in a wide variety of functions including mitosis, cell shaping, secretion, motility, axonal growth, and transport. 1,~ Microtubules are polymers of a protein called tubulin, which is itself a dimer composed of two similar but not identical subunits (o~, fl), each of molecular weight 55,0002 Drugs, such as colchicine, that inhibit mitosis and axonal function through disruption of microtubules have been shown to interact in vitro with tubulin isolated from several sources. 4 Knowledge of the precise localization and properties of these drug receptors would contribute to the better understanding of the mechanism of tubulin assembly to form microtubules. 1j. B. Olmsted and G. G. Borisy, Annu. Rev. Biochem. 42, 507 (1973). M. L. Shelanski and H. Feit, in "The :Structure and Function of Nervous Tissue" (G. H. Bourne, ed.), ¥ol. 6, pp. 47-80. Academic Press, New York, 1972. 3R. E. Stephens, in "Subunits in Biological Systems" (S. N. Timasheff and " G. D. Fasman, eds.), Part A, pp. 355--391. Dekker, New York, 1971. 4L. Wilson and J. Bryan, Adv. Cell Mol. Biol. 3, 21 (1974).
