The role of active site residue arginine 218 in firefly luciferase bioluminescence

Firefly luciferase catalyzes the highly efficient emission of yellow-green light from substrate firefly luciferin by a sequence of reactions that require Mg-ATP and molecular oxygen. We had previously developed a working model of the luciferase active site based on the X-ray structure of the enzyme without bound substrates. In our model, the side chain guanidinium group of Arg218 appears to be located in close proximity to the substrate's hydroxyl group at the bottom of the luciferin binding pocket. A similar role for Arg337 also has been proposed. We report here the construction, purification, and characterization of mutant luciferases R218A, R218Q, R218K, R337Q, and R337K. Alteration of the Arg218 side chain produced enzymes with 15-20-fold increases in the Km values for luciferin. The contrasting near-normal Km values for luciferin determined with the Arg337 enzymes support our proposal that Arg218 (and not Arg337) is an essential luciferin binding site residue. Bioluminescence emission studies indicated that in the absence of a positively charged group at position 218, red bioluminescence was produced. Based on this result and those of additional fluorescence experiments, we speculate that Arg218 maintains the polarity and rigidity of the emitter binding site necessary for the normal yellow-green emission of P. pyralis luciferase. The findings reported here are interpreted in the context of the firefly luciferase X-ray structures and computational-based models of the active site.     

Delicate balance of electrostatic interactions and disulfide bridges in thermostability of firefly luciferase

The wild type Photinus pyralis luciferase does not have any disulfide bridge. Disulfide bridges are determinant in inherent stability of protein at moderate temperatures. Meanwhile, arginin is responsible for thermostability at higher temperatures. In this study, by concomitant introduction of disulfide bridge and a surface arginin in a mutant (A296C-A326C/I232R), the contribution of disulfide bridge introduction and surface hydrophilic residue on activity and global stability of P. pyralis luciferase is investigated. In addition to the mentioned mutant; I232R, A296C-A326C and wild type luciferases are characterized. Though addition of Arg caused stability against proteolysis but in combination with disulfide bridge resulted in decreased thermal stability compared to A296C-A326C mutant. In spite of long distance of two different mutations (A296C-A326C and I232R) from each other in the three-dimensional structure, combination of their effects on the stability of luciferase was not cumulative.     

Dominant negative mutants of ornithine decarboxylase.

Conserved lysines of mouse ornithine decarboxylase were individually mutated to arginines. The mutations at amino acid residues 69, 115, and 169 greatly reduced or abolished enzymatic activity. Lysine 69 is the site of Schiff base formation with the cofactor pyridoxal phosphate; the functional role of the other two lysines essential for activity is not known. Coexpression of wild type ornithine decarboxylase along with the lysine 115 to arginine mutant reduced the activity of the former without diminishing the amount of wild type protein. This form of negative complementation was seen when wild type and mutant protein were coexpressed either by in vitro translation or in bacteria. The data are consistent with the conclusion that a wild type and mutant subunit form a heterodimer that is enzymatically inactive.     

Histidine residues of zinc ligands in beta-lactamase II.

On the basis of the chemical and structural features of the amino acid sequences in the vicinities of phosphorylatable hydroxyamino acid residues in several of the well-known protein substrates for skeletal-muscle cyclic AMP-dependent protein kinase, it is hypothesized that the phosphorylatable residue at position i and arginine residue at position i-3 of these protein substrates are located on a peptide turn on the hydrophilic protein surface. It is further hypothesized that there is an arginine-recognition site near the active centre on the protein kinase. This site is essential for the function of cyclic AMP-dependent protein kinase, for, not only does it recognize specifically the exposed arginine residue of the protein substrate, but, more importantly, via the interaction with arginine-(i--3), it may help to steer the topologically adjacent serine-i into proper orientation on the nearby active centre for phosphorylation. Model-building and kinetic data that provide support for the proposed hypotheses are presented.     

Site-Directed Mutagenesis Evidence for Arginine-384 Residue at the Active Site of Maize Branching Enzyme II

Essential arginine residues are suggested to be located at the active sites of maize branching enzymes (BE) based on the evidence that two arginine residues are conserved in all BE from various species and that as little as one arginine residue is located at the active site of maize BE by phenylglyoxal (PGO) modification. To determine the exact location of the active-site arginine residue in BE, we employed peptide mapping and site-directed mutagenesis approaches. A single trypsin-digested, [¹⁴C]PGO-labeled peptide was purified from maize BEII by two rounds of HPLC separation, but we failed to obtain amino acid sequencing information. Site-directed mutagenesis was then used to create one mutant (arginine-384 to alanine-384), R384A. Immunoblotting result showed that BEII protein was expressed at a similar level in the wild type and the R384A mutant. However, BE activity in the R384A mutant was only 1.4% of the wild type. These results support the conclusion that the conserved arginine-384 residue is important in BEII catalysis.     

A single amino acid substitution in a histidine-transport protein drastically alters its mobility in sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

Mutation hisJ5625 has altered the histidine-binding protein J of Salmonella typhimurium such that histidine transport is impaired, even though binding of histidine by the J protein is unimpaired [Kustu, S.G., & Ames, G.F. (1974) J. Biol. Chem. 249, 6976--6983]. We have determined by protein analytical methods that the only effect of this mutation has been the substitution of a cysteine residue for an arginine at a site in the interior of the polypeptide chain. This arginine residue is therefore potentially essential for the transport function of the protein. The mutant protein migrates in sodium dodecyl sulfate-polyacrylamide gel electrophoresis more slowly than the wild type protein, as if its molecular weight were greater by as much as 2000. Since this behavior is apparently due to a single amino acid replacement, a molecular weight difference even between two closely related proteins should not be inferred solely on the basis of sodium dodecyl sulfate gel electrophoresis.     

Alanine-scanning mutagenesis of HM-1 killer toxin and the essential residues for killing activity

Each of the aromatic, acidic and basic amino acid residues in HM-1 were separately substituted with alanine by site-directed mutagenesis. The mutant genes were successfully expressed in HM-1 resistant Saccharomyces cerevisiae. HM-1 gene analogues corresponding to the aromatic substitutions resulted in lower production of HM-1 analogues. In the case of the acidic amino acid residue and basic amino acid residue substitutions, some analogues were produced in the same amount as and exhibited similar killing activity to that of the wild type HM-1. But the H35A HM-1 analogue had completely lost the killing activity, and D44A, K21A, K46A, R82A, R85A and R86A HM-1 showed highly decreased killing activities. These results strongly indicate the importance of histidine-35, aspartic acid-44, lysine-21, lysine-46, and C-terminal arginine residues in HM-1 for the killing activity.     

Thermostable red and green light-producing firefly luciferase mutants for bioluminescent reporter applications

Light emission from the North American firefly Photinus pyralis, which emits yellow-green (557-nm) light, is widely believed to be the most efficient bioluminescence system known, making this luciferase an excellent tool for monitoring gene expression. We present studies on the production of a set of thermostable red- and green-emitting luciferase mutants with bioluminescent properties suitable for dual-color reporter assays, biosensor measurements with internal controls, and imaging techniques. Starting with the luciferase variant Ser284Thr, we introduced the mutations Thr214Ala, Ala215Leu, Ile232Ala, Phe295Leu, and Glu354Lys to produce a new red-emitting enzyme with a bioluminescence maximum of 610 nm, narrow emission bandwidth, favorable kinetic properties, and excellent thermostability at 37 degrees C. By adding the same five changes to luciferase mutant Val241Ile/Gly246Ala/Phe250Ser, we produced a protein with an emission maximum of 546 nm, providing a set of thermostable enzymes whose bioluminescence maxima were separated by 64 nm. Model studies established that the luciferases could be detected at the attomole level and six orders of magnitude higher. In microplate luminometer format, mixtures containing 1.0 fmol total luciferase were quantified from measurements of simultaneously emitted red and green light. The results presented here provide evidence that it is feasible to monitor two distinct activities at 37 degrees C with these novel thermostable proteins.     

The influence of insertion of a critical residue (Arg356) in structure and bioluminescence spectra of firefly luciferase

The firefly bioluminescence reaction, which uses luciferin, Mg-ATP, and molecular oxygen to yield an electronically excited oxyluciferin, is carried out by luciferase and visible light is emitted. The bioluminescence color of firefly luciferases is determined by the luciferase structure and assay conditions. Among different beetle luciferases, those from Phrixothrix railroad worm emit either yellow or red bioluminescence colors. Sequence alignment analysis shows that the red-emitter luciferase from Phrixothrix hirtus has an additional Arg residue at 353, which is absent in firefly luciferases. We report here the construction and purification of a mutant at residue Arg(356), which is not conserved in beetle luciferases. By insertion of an additional residue (Arg(356)) using site-specific insertion mutagenesis in a green-emitter luciferase (Lampyris turkestanicus) the color of emitted light was changed to red and the optimum temperature of activity was also increased. Insertion of this Arg in an important flexible loop showed changes of the bioluminescence color and the luciferase reaction took place with relatively retention of its basic kinetic properties such as Km and relative activity. Comparison of native and mutant luciferases using homology modeling reveals a significant conformational change of the flexible loop in the red mutant. Movement of flexible loop brought about a new ionic interaction concomitant with a change in polarity of the emitter site, thereby leading to red emission. It is worthwhile to note that the increased optimum temperature and emission of red light might make mutant luciferase a suitable reporter for the study of gene expression and bioluminescence imaging.     

K+ binding sites and interactions between permeating K+ ions at the external pore mouth of an inward rectifier K+ channel (Kir2.1).

The arginine at position 148 is highly conserved in the inward rectifier K+ channel family. Increases of external pH decrease the single-channel conductance in mutant R148H of the Kir2.1 channel (arginine is mutated into histidine) but not in the wild type channel. Moreover, in 100 mM external K+, varying external pH induced biphasic changes of open channel noise, which peaks at around pH 7.4 in the R148H mutant but not in the wild type channel. The maximum single-channel conductances are higher in the wild type channel and R148H mutant at pH 6.0 than those in the R148H mutant at pH 7.4. However, the maximal conductance is achieved with much lower external [K+] for the latter. Interestingly, the single-channel conductances and open channel noise of the wild type channel at pH 6. 0 and the R148H mutant at pH 6.0 and 7.4 become the same in [K+] = 10 mM. These results indicate that the residue at position 148 is accessible to the external H+ and probably is involved in the formation of two K+ binding sites in the external pore mouth. Effective repulsion between permeating K+ ions in this area requires a positive charge at position 148, and such K+-K+ interaction is the essential mechanism underlying high K+ conduction rate through the Kir2.1 channel pore.     

New properties of Bacillus subtilis succinate dehydrogenase altered at the active site. The apparent active site thiol of succinate oxidoreductases is dispensable for succinate oxidation.

Mammalian and Escherichia coli succinate dehydrogenase (SDH) and E. coli fumarate reductase apparently contain an essential cysteine residue at the active site, as shown by substrate-protectable inactivation with thiol-specific reagents. Bacillus subtilis SDH was found to be resistant to this type of reagent and contains an alanine residue at the amino acid position equivalent to the only invariant cysteine in the flavoprotein subunit of E. coli succinate oxidoreductases. Substitution of this alanine, at position 252 in the flavoprotein subunit of B. subtilis SDH, by cysteine resulted in an enzyme sensitive to thiol-specific reagents and protectable by substrate. Other biochemical properties of the redesigned SDH were similar to those of the wild-type enzyme. It is concluded that the invariant cysteine in the flavoprotein of E. coli succinate oxidoreductases corresponds to the active site thiol. However, this cysteine is most likely not essential for succinate oxidation and seemingly lacks an assignable specific function. An invariant arginine in juxtaposition to Ala-252 in the flavoprotein of B. subtilis SDH, and to the invariant cysteine in the E. coli homologous enzymes, is probably essential for substrate binding.     

Analysis of recognition elements for mitochondrial processing peptidase using artificial amino acids: roles of the intervening portion and proximal arginine.

We recently demonstrated, using synthetic peptides modeled on the extension peptide of malate dehydrogenase, that amino acid residues present at the proximal and distal positions relative to the cleavage site are critical determinants for the recognition of substrates by mitochondrial processing peptidase [Niidome et al. (1994) J. Biol. Chem. 269, 24719-24722). While the proximal arginine is unexceptionally located at the -2 position, the position of the distal residue varies among mitochondrial precursor proteins. Between the proximal and distal residues, proline and/or glycine are present in most mitochondrial precursor proteins, and they are considered to play a role in the specific recognition of a substrate by the peptidase. To elucidate the role of the intervening portion, we introduced a non-natural amino acid [2-(2-aminoethoxy)acetic acid] between the distal and proximal residues. We also analyzed the functional elements in the proximal arginine by replacing the residue with various arginine or lysine analogs. The results of kinetic studies indicated that the intervening portion should be flexible for efficient processing, and that the guanidino group of the proximal arginine is recognized by the peptidase through hydrogen and ionic bonds.     

Analysis of the catalytic mechanism of juvenile hormone esterase by site-directed mutagenesis.

1. Juvenile hormone esterase (JHE) is a serine hydrolase selective for hydrolysis of the conjugated methyl esters of insect juvenile hormones. 2. We have investigated the mechanism of catalytic action of this enzyme by site-directed mutagenesis of the cloned enzyme and expression of the mutants in a baculovirus system. 3. A series of individual mutations of JHE were made to residues possibly involved in catalysis of juvenile hormones, and which are highly conserved in both esterases and lipases. 4. Mutation of the serine residue at position 201 to glycine (S201G), or aspartate 173 to asparagine (D173N), or histidine 446 to lysine (H446K), removed all detectable activity and these mutagenized enzymes were determined to be at least 10(6)-fold less active than wild type JHE. 5. Mutation of arginine 47 to histidine (R47H) decreased but did not abolish activity, with Km essentially unchanged at 66 nM for R47H compared to 34 nM for wild type JHE. 6. The kcat for R47H was decreased from 103 min-1 for wild type JHE to 1.9 min-1. 7. In addition, glutamate residue 332 was altered to glutamine (E332Q) and expressed in an Escherichia coli system. 8. This mutation was also found to remove all detectable activity. 9. From the results presented in this study and by comparison of JHE to other serine esterases and lipases, we predict that JHE possesses a Ser201-His446-Glu332 catalytic triad. 10. In addition, aspartate 173 and arginine 47 are essential for the efficient functioning of JHE.     

Primary structure and properties of an inactive mutant aspartate transcarbamoylase.

A mutant form of Escherichia coli aspartate transcarbamoylase (ATCase) which lacks catalytic activity has been purified and characterized (Wall, K.A., Flatgaard, J.E., Gibbons, I., and Schachman, H.K. (1979) J. Biol. Chem 254, 11910-11916). Peptide mapping of the mutant and wild type catalytic chains followed by the determination of the amino acid sequence of the one altered peptide in the mutant indicated that a glycyl residue was replaced by aspartic acid. This substitution is located at position 125 in the tentative sequence kindly provided by W. Konigsberg (personal communication). The mutant protein has an overall secondary structure similar to that of the wild type as indicated by circular dichroism spectroscopy. However, marked changes in the reactivity of several amino acid residues were demonstrated. Lysyl residue 84 which in the wild type subunits reacts specifically with pyridoxal 5'-phosphate is only slightly reactive in the mutant even though the peptide containing that residue was not altered in amino acid composition. Another residue, cysteinyl 46, which is thought to be in the active site, is much more reactive toward p-hydroxymercuribenzoate in the mutant subunit than in the wild type protein. Finally, tyrosyl residue 213, which according to recent crystallographic studies is not near the active site and which exhibits an unusually low pK (9.1) in the wild type catalytic subunits, appears to have its pK shifted to 10.5 or higher as a result of the mutation. The evidence indicates that the substitution of an aspartyl for a glycyl residue at a region of the amino acid sequence remote from those residues in the active site causes sufficient modification of the tertiary structure to cause the loss of enzyme activity and to affect the reactivity of other residues in the protein. Moreover, the quaternary structure of the intact enzyme is altered as well since the subunit interactions are greatly weakened.     

Site-directed mutagenesis of Escherichia coli ornithine transcarbamoylase: role of arginine-57 in substrate binding and catalysis

In the carbamoyl-transfer reaction catalyzed by ornithine transcarbamoylase, an arginine residue in the active site of the Escherichia coli enzyme has been suggested to bind the phosphate moiety of the substrate carbamoyl phosphate. With the application of site-specific mutagenesis, the most likely arginine residue among three candidates at the binding site of carbamoyl phosphate, Arg-57, has been replaced with a glycine. The resultant Gly-57 mutant enzyme is drastically inefficient in catalysis. In the synthesis of L-citrulline from carbamoyl phosphate and L-ornithine with the release of inorganic phosphate, the turnover rate of the mutant is 21,000-fold lower than that of the wild type. However, the mutation of Arg-57 affects only moderately the binding of carbamoyl phosphate; the dissociation constant of this substrate, measured under steady-state turnover condition, is increased from 0.046 to 3.2 mM by the mutation. On the other hand, ornithine binding is substantially affected as estimated by the change in the dissociation constant of its analogue L-norvaline. The dissociation constant of L-norvaline increases about 500-fold from 54 microM for the wild type to 25 mM for the mutant. Since Arg-57 is expected to be distal from the ornithine site and the amino acid (both ornithine and norvaline) binds only after carbamoyl phosphate in the wild-type reaction, the poor norvaline affinity to the mutant suggests that Arg-57 is involved in interactions essential for productive addition of the amino acid. This interpretation is supported by difference ultraviolet absorption spectra which show that the conformational changes induced in the wild type by carbamoyl phosphate upon binding are absent in the mutant. Furthermore, steady-state kinetic data reveal that the ordered binding mechanism of the wild-type enzyme is transformed into a random binding mechanism in the mutant. Thus, the presence of carbamoyl phosphate in the mutant active site is no longer a requisite for ornithine binding. In the 5-50 degrees C temperature range, transcarbamoylation catalyzed by either the wild type or the mutant observes the Arrhenius rate law with almost identical enthalpies of activation, 11 and 10 kcal/mol, respectively. The entropy of activation is -5.5 eu for the wild-type reaction and -29 eu for the mutant reaction, accounting for a loss of 6-7 kcal/mol in the rate-determining step of the enzymic reaction.(ABSTRACT TRUNCATED AT 400 WORDS)     

Structure and activity of insulin, XV[1-5]. Further evidence for the importance of arginine residue B22 in the activity of insulin. Semisyntheses of despentapeptide-(B23 - 30)-insulins varied in B22 using desnonapeptide-(B22 - 30)-insulin and tetrapeptides

Insulin hexamethyl ester was digested by trypsin. The resulting desoctapeptide-(B23 - 30)-insulin pentamethyl ester was purified. This compound was digested by carboxypeptidase B to remove the arginine residue B22 at the end of the B chain. Then the N-terminal amino groups of the remaining desnonapeptide-(B22 - 30)-insulin pentamethyl ester were protected with the Boc residue. The free carboxyl group of the glutamic acid residue B21 of this product was coupled to the following synthetic tetrapeptide esters: Arg-Gly-Phe-Phe-OMe, Lys(Boc)-Gly-Phe-Phe-OMe, Orn(Boc)-Gly-Phe-Phe-OMe, Cit-Gly-Phe-Phe-OMe, Ala-Gly-Phe-Phe-OMe and Gly-Gly-Phe-Phe-OMe. The syntheses of these peptide esters are described. After removal of all protecting groups, despentapeptide-insulin (B22-Arg) and analogues of this product with variation in position B22 could be obtained. They were purified by column chromatography. The biological activities of these components were determined by the mouse fall test. In the case of despentapeptide insulin (C-terminus Arg-Gly-Phe-Phe), the activity rose to the expected value of 34%. The insulin variants with amino acid residues other than arginine in position B22 had much lower activities: with lysine 13%, with ornithine 12%, with citrulline 9%, with alanine 8% and with glycine 6%. Desnonapeptide-insulin by itself posses an activity of 3%. These results demonstrate once more the essential nature of arginine residue B22 for insulin activity.     

A single carboxyl mutant of the multidrug transporter EmrE is fully functional

EmrE, a multidrug transporter from Escherichia coli removes toxic compounds from the cell in exchange with protons. Glu-14 is the only charged residue in the putative membrane domains and is fully conserved in more than 50 homologues of the protein. This residue was shown to be an essential part of the binding site, common to protons and substrate. EmrE bearing a single carboxylic residue, Glu-14, shows uptake and binding properties similar to those of the wild type. This suggests that a small protein bearing only 110 amino acids with a single carboxyl in position 14 is the most basic structure that shows ion-coupled transport activity. The role of Glu-14 in substrate binding was examined by using dicyclohexylcarbodiimide, a hydrophobic carbodiimide that is known to react with carboxyls. Tetraphenylphosphonium binding to both wild type and the single carboxyl mutant is inhibited by dicyclohexylcarbodiimide in a dose-dependent manner. Ethidium and other substrates of EmrE prevent this inhibition with an order of potency in accord with their apparent affinities. This suggests that dicyclohexylcarbodiimide binding is sterically prevented by the substrate, supporting the contention that Glu-14, the reactive residue, is part of the substrate-binding site.     

Relationship between activating and editing functions of the adenylation domain of apo-tyrocidin synthetase 1 (apo-TY1)

Tyrocidine synthetase 1 (TY1), the initial monomodular constituent of the tyrocidine biosynthetic system, exhibits relaxed substrate specificity, however an efficient editing of the mis-activated amino acid provides for fidelity of product formation. We chose to analyse the consequence of single amino acid substitutions, in the amino acid activation site of apo-TY1, on the editing functions of the enzyme. Discrimination between L-Phe and D-Phe by apo-TY1 depends primarily on the editing reaction. Distraction of unnatural amino acid substrates, such as L-PheSer, implies that editing is not designated to select a specific mis-activated amino acid, but instead to discriminate all mis-activated amino acid analogues. It was shown that active site residues which interact with the adenylate are essential for both activation and editing. Substitution of Lys186 with arginine substantially reduces the editing capacity of the protein. Loss of amino acid discrimination ability by the apo-K186T and apo-R416T mutant proteins suggests a role of active site residues in maintaining the structural determinants for substrate selection. Inadequate conformational changes, induced by non-cognate amino acid substrates, promote ATP breakdown yielding P(i) and ADP. Replacement of residue Lys186 or Arg416 enhances ATP hydrolysis implying a role in binding or adjusting of the triphosphate chain for adenylate formation and pyrophosphate cleavage.     

Thermostability of firefly luciferases affects efficiency of detection by in vivo bioluminescence

Luciferase from the North American firefly (Photinis pyralis) is a useful reporter gene in vivo, allowing noninvasive imaging of tumor growth, metastasis, gene transfer, drug treatment, and gene expression. Luciferase is heat labile with an in vitro halflife of approximately 3 min at 37 degrees C. We have characterized wild type and six thermostabilized mutant luciferases. In vitro, mutants showed half-lives between 2- and 25-fold higher than wild type. Luciferase transfected mammalian cells were used to determine in vivo half-lives following cycloheximide inhibition of de novo protein synthesis. This showed increased in vivo thermostability in both wild-type and mutant luciferases. This may be due to a variety of factors, including chaperone activity, as steady-state luciferase levels were reduced by geldanamycin, an Hsp90 inhibitor. Mice inoculated with tumor cells stably transfected with mutant or wild-type luciferases were imaged. Increased light production and sensitivity were observed in the tumors bearing thermostable luciferase. Thermostable proteins increase imaging sensitivity. Presumably, as more active protein accumulates, detection is possible from a smaller number of mutant transfected cells compared to wild-type transfected cells.     

The beta subunit of tryptophan synthase. Clarification of the roles of histidine 86, lysine 87, arginine 148, cysteine 170, and cysteine 230

Our studies, which are aimed at understanding the catalytic mechanism of the beta subunit of tryptophan synthase from Salmonella typhimurium, use site-directed mutagenesis to clarify the functional roles of several putative active site residues. Although previous chemical modification studies have suggested that histidine 86, arginine 148, and cysteine 230 are essential residues in the beta subunit, our present findings that beta subunits with single amino acid replacements at these positions have partial activity show that these 3 residues are not essential for catalysis or substrate binding. These conclusions are consistent with the recently determined three-dimensional structure of the tryptophan synthase alpha 2 beta 2 complex. Amino acid substitution of lysine 87, which forms a Schiff base with pyridoxal phosphate in the wild type beta subunit, yields an inactive form of the beta subunit which binds alpha subunit, pyridoxal phosphate, and L-serine. We also report a rapid and efficient method for purifying wild type and mutant forms of the alpha 2 beta 2 complex from S. typhimurium from an improved enzyme source. The enzyme, which is produced by a multicopy plasmid encoding the trpA and trpB genes of S. typhimurium expressed in Escherichia coli, is crystallized from crude extracts by the addition of 6% poly(ethylene glycol) 8000 and 5 mM spermine. This new method is also used in the accompanying paper to purify nine alpha 2 beta 2 complexes containing mutant forms of the alpha subunit.