Bioluminescence of the Ca2"-Binding Photoprotein Aequorin After

Home , Aequorin

Proc. Nadl. Acad. Sci. USA Vol. 86, pp. 80-84, January 1989 Biochemistry Bioluminescence of the Ca2"-binding photoprotein aequorin after cysteine modification (protein modification/site-specific mutagenesis/protein active-site/oxygenase/protein conformation) KouICHI KUROSE*, SATOSHI INOUYE*t, YOSHIYUKI SAKAKI*, AND FREDERICK 1. TSUjit§¶ *Research Laboratory for Genetic Information, Kyushu University, Fukuoka 812, Japan; tOsaka Bioscience Institute, 6-2-4 Furuedai, Suita, Osaka 565, Japan; and WMarine Biology Research Division, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093 Communicated by Martin D. Kamen, September 26, 1988

ABSTRACT Aequorin is a monomeric Ca2+-binding pro- experimentally demonstrated. Further, the changes that tein (Mr, 21,400) that emits light upon reacting with Ca2+. The accompany the regeneration of aequorin are not known. To protein has three Ca2+-binding sites, three cysteine residues, better understand structure-function relationships in ae- and a noncovalently bound chromophore that consists of quorin, we have prepared large quantities of the protein for coelenterazine and molecular oxygen. Light is emitted via an x-ray crystallographic studies, using recombinant DNA intramolecular reaction in which coelenterazine is oxidized by methods (12). the bound oxygen. After light emission, aequorin may be In the past, structural changes in aequorin have been regenerated by incubating the protein with coelenterazine, studied by electron paramagnetic resonance (13) and proton dissolved oxygen, EDTA, and 2-mercaptoethanol. To under- NMR (14), and these studies have revealed that aequorin is stand structure-function relationships in this protein, we used easily destabilized structurally when a spin-label is attached the technique of site-specific mutagenesis to replace the three to an essential sulfhydryl group or when coelenterazine is cysteine residues with serine. Six of the seven modified aequor- autoxidized. Site-specific mutagenesis studies have also ins had reduced luminescence activity, whereas the seventh shown that when the protein is modified, there is a loss of with all three cysteines replaced by serine had luminescence luminescence activity (15). Such results, while interesting, activity equal to or greater than that of the wild-type aequorin. are not always fully informative because it is difficult to Further, the time required for the regeneration of the triply decide whether a particular change is due to specific inter- substituted aequorin was substantially increased compared to ference with the binding of a ligand or to a change in the the time required for the regeneration of the wild-type ae- tertiary structure of the protein. To focus more narrowly on quorin. The results suggest that cysteine plays an important this problem, we have replaced the three cysteine residues of role in the regeneration of aequorin but not in its catalytic aequorin with serine, a close structural analogue, in the hope activity. that the substitutions will not perturb the tertiary structure of the protein significantly. The results of the changes in The bioluminescent jellyfish Aequorea Victoria from Friday luminescence activity are reported in the present paper. Harbor, WA, possesses in the outer margin of its umbrella a Ca2l-binding protein, aequorin (Mr, 21,400), which emits MATERIALS AND METHODS light upon reacting with Ca2+ (1-3). The photoprotein consists of two components: an apoprotein (apoaequorin) and a Enzymes and Chemicals. Restriction enzymes, Escherichia chromophore (3). The chromophore is made up of coelenter- coli T4 DNA ligase, Klenow fragment, T4 polynucleotide azine and molecular oxygen, with the oxygens attached to the kinase, and E. coli alkaline phosphatase were obtained from C-2 carbon of coelenterazine in the form of a peroxide or Takara Shuzo (Kyoto, Japan) and Nippon Gene (Toyama, hydroperoxide (4). The nature of the binding of the chro- Japan). Radiolabeled compounds were purchased from Am- mophore to apoaequorin is not known (5). When Ca2' binds ersham and New England Nuclear. Oligonucleotides were to aequorin, an intramolecular reaction takes place in which synthesized by the phosphoramidite method (16) by using coelenterazine is oxidized to coelenteramide by the bound an Applied Biosystems (Foster City, CA) model 380A oxygen, yielding as products light (Ama, 470 nm), C02, and DNA synthesizer. Coelenterazine, 2-(p-hydroxybenzyl)- a blue fluorescent protein (6, 7). The blue fluorescent protein 6-(p-hydroxyphenyl)-3 ,7-dihydroimidazo[1 ,2-a]pyrazine- consists of coelenteramide attached to apoaequorin, and the 3-one, was chemically synthesized (17). All other reagents excited-state coelenteramide is the emitter in the reaction (8). were of the highest quality available and were obtained from The coelenteramide can be dissociated from the blue fluo- Wako Pure Chemicals (Osaka, Japan) or Nakarai Chemicals rescent protein with acid or ether or by gel filtration (8). (Kyoto, Japan). Aequorin may be regenerated from apoaequorin by incuba- Bacterial Strain and Plasmids. The host strain was E. coli tion with coelenterazine, dissolved oxygen, EDTA, and D1210 carrying lacP and lacy (12, 18), and the plasmid used 2-mercaptoethanol (9). was the previously described piQ9-2HE (12, 19). In the Recently, it has been shown that aequorin contains three regeneration experiments with glutathione, dithioerythritol, Ca2+-binding sites, which are homologous to the Ca2+- dithiothreitol, and 2-mercaptoethanol, the plasmid used was binding sites of calmodulin (10, 11). The binding of Ca2+ to piQ5 (12). these sites presumably induces a conformational change in Modification of Cysteine Residues. The three cysteine the protein, causing an active site to be formed, which residues ofaequorin are located at positions 145, 152, and 180 catalyzes the oxidation of the bound coelenterazine. How- (10, 11) and are designated 1, 2, and 3, respectively. Con- ever, neither a conformational change nor a conversion of a struction of the mutant cDNAs of apoaequorin (Fig. 1) was region in the aequorin molecule to an active site has been carried out by oligonucleotide-directed site-specific muta-

The publication costs of this article were defrayed in part by page charge tPresent address: Chisso Corporation, Tokyo 100, Japan. payment. This article must therefore be hereby marked "advertisement" $To whom reprint requests should be addressed at the Osaka in accordance with 18 U.S.C. §1734 solely to indicate this fact. Bioscience Institute.

80 Downloaded by guest on September 29, 2021 Biochemistry: Kurose et al. Proc. Nati. Acad. Sci. USA 86 (1989) 81 compared by using glutathione, dithiothreitol, and dithioerythritol in place of 2-mercaptoethanol. TQ 100 ,u of a piQ5/D1210 cell extract, 2 ug of coelenterazine (1 ,g/,ul of absolute methyl alcohol), and 2 ul of one of the 0.104 M solutions of each of the four compounds dissolved in 30 mM Tris-HCl, pH 7.60/10 mM EDTA was added and mixed thoroughly. The mixture was allowed to stand in an ice bath for 2 hr. Fifty microliters of the regeneration mixture was assayed for aequorin activity. Regeneration of Aequorin from Apoaequorin. In all other cases studied, regeneration of aequorin from apoaequorin was carried out by mixing a measured volume of each of the OligoI =CAGAAGATTCCGAGGAA,Cysl45-Ser following in a 1.5-ml Eppendorf tube: piQ9-2HE/D1210 cell OligoIl =CAGAGTGTCCGATATTG,CA'ys152Ser extract, coelenterazine, either 2-mercaptoethanol or dithio- OligoHlI=TC C TG C T TCCG A AA AGCCys 18O-Ser threitol, and 30 mM Tris-HC1, pH 7.60/10 mM EDTA; then after thorough mixing, the mixture was allowed to stand in an ice bath for the required time. The regeneration mixture FIG. 1. Scheme used in constructing cDNAs for apoaequorin and minus coelenterazine or cell extract served as the control. a list of synthetic oligonucleotides used in site-specific mutagenesis. Details are presented in the text. An asterisk indicates the position of nucleotide substitution. Assay for Aequorin Activity. Aequorin luminescence activity was determined by transferring a measured volume of the genesis as described (15, 20). The cDNA for C1+3S was incubation mixture to a glass reaction cell and injecting 1.5 ml constructed by using E. coli T4 DNA ligase to ligate the the BamHI/Sca I 2.26-kilobase (kb) fragment from CiS and the of 30 mM CaC12/30 mM Tris HCl, pH 7.60 into cell, as in Table 1. The initial maximal light BamHI/Sca I 1.01-kb fragment from C3S. All mutant cDNAs except noted otherwise intensity was recorded with either a Labo Science (Tokyo) were sequenced to confirm the presence of the mutation by model TD-4000 photometer with a Hamamatsu R268 bialkali a dideoxynucleic acid sequencing method using modified or a (22) with (21). photomultiplier Mitchell-Hastings photometer a Hamamatsu 1P21 Sb-Cs photomultiplier and a Soltec (Sun Growth of Oacteria and Preparation of Ceil Extracts. Ten Mitchell- milliliters of LB medium supplemented with 50 ,ug of ampi- Valley, CA) model S-4201 strip-chart recorder. The Hastings photometer was calibrated with a carbon-14 light per was 100 of an overnight cillin ml inoculated with ,Al light intensity was culture ofthe transformed E. coli and incubated with shaking standard (23, 24), and the initial maximal converted to quanta per sec to serve as a measure ofactivity. at 37°C for 2 hr (cell density= 115-125 Klett colorimeter units with red filter 66; 30 Klett units 2 108 cells per ml). At the end of 2 hr, isopropyl-p-D-thiogalactopyranoside was RESULTS added to a final concentration of 1 mM, and the incubation was continued for an additional 2 hr (cell density = 300-350 Luminescence Activity ofModified Aequorins. Table 1 gives Klett colorimeter units). The cells were harvested by cen- the relative luminescence activities of modified aequorins trifugation at 12,000 x g for 5 min at 4°C and were washed regenerated from the seven apoaequorins in the presence and once by mixing in 2 ml ofM9 salt solution. After repeating the absence of 2-mercaptoethanol, expressed as the percentage centrifugation, the cells were resuspended in 2.5 ml of30 mM of the wild-type piQ9-2HE aequorin regenerated for 3 hr in Tris'HCl, pH 7.60/30 mM EDTA and were disrupted by using the presence of2-mercaptoethanol. The Klett readings for the a Branson model 200 sonicator (for 30 sec twice) with the cell cell densities were nearly the same, indicating that about the holder immersed in an ice bath. The suspension was centri- same number of cells were disrupted in preparing each cell fuged at 12,000 x g for 10 min at 4°C, and the supernatant extract. When regeneration was carried out with 2- (designated "cell extract") was diluted with 5.0 ml of 30 mM mercaptoethanol, luminescence activity ranged from 105 for Tris'HCl (pH 7.60) to give a final EDTA concentration of 10 the piQ9-2HE aequorin (27 hr) to 0.6 for the C1+2S aequorin mM. The extracts were used immediately or stored at -70°C (27 hr), whereas in the absence of 2-mercaptoethanol, the for a few days before being used. activity ranged from 0.2 for C1+2S aequorin (3 hr and 27 hr) Comparison of Reducing Agents in the Regeneration of to 116 for the C1+2+3S aequorin (27 hr). For the C1S, C2S, Aequorin. Regeneration of aequorin from apoaequorin was and C3S aequorins, the results showed the same pattern as in Table 1. Bioluminescence activities of aequorins after cysteine replacement and regeneration for 3 and 27 hr with and without 2-mercaptoethanol Activity after 3 hr regeneration* Activity after 27 hr regeneration* Position(s) of Cys With Without With Without Aequorin replaced by Ser HSCH2CH2OH HSCH2CH2OH HSCH2CH2OH HSCH2CH2OH piQ9-2HE 100 1 105 0.8 CiS 145 61 3 59 1 C2S 152 46 9 22 4 C3S 180 13 9 11 14 C1+2S 145, 152 1 0.2 0.6 0.2 C1+3S 145, 180 68 28 89 39 C2+3S 152, 180 17 14 15 19 C1+2+3S 145, 152, 180 21 49 8 116 *Composition ofregeneration mixtures: 500 bl ofcell extract and 5 .ug ofcoelenterazine (1 jtg/pli ofabsolute methyl alcohol) with or without 10 /1 of 2-mercaptoethanol. Assay of the mixtures was performed in duplicate on 50-A1 aliquots injected with 0.6 ml of 30 mM CaCI2/30 mM Tris HCI, pH 7.60. Activities are expressed as the percentage of piQ9-2HE activity or 10.2 x 109 quanta per sec (after a 3-hr regeneration with 2-mercaptoethanol). Downloaded by guest on September 29, 2021 82 Biochemistry: Kurose et al. Proc. Natl. Acad. Sci. USA 86 (1989)

the previous study (15). Low or no activity was consistently 18.0 c ^ found with C1+2S aequorin and with piQ9-2HE aequorin regenerated in the absence of 2-mercaptoethanol. The ClS and C2S aequorins also had relatively low activity when regenerated in the absence of2-mercaptoethanol. The highest 14.4 _- 14.4==_ __ activity was observed with wild-type piQ9-2HE and C-)UL) C1+2+3S aequorins regenerated in the presence and absence u of 2-mercaptoethanol, respectively. Controls gave no light. Effectiveness of Reducing Agents in Regenerating Aequorin. - 10.8 .__ 10.8 0 In comparing the effectiveness of reducing agents in regen- C erating aequorin, a number of reducing agents were mixed 3 with piQ5/D1210 cell extracts at a concentration of 2 mM. a The light intensities triggered by Ca2+ from the regenerated 7.2 piQ5 aequorins, in relative light units, were as follows: > 2-mercaptoethanol, 100; glutathione (reduced form), 9; di- Cl- < .0__ 1_ thiothreitol, 128; and dithioerythritol, 128. 3.6 __TIM _ec I= Time Course of Regeneration of Aequorin. The findings (i) that the C1+2+3S aequorin regenerated in the absence of 2-mercaptoethanol showed higher activity at 27 hr than at 3 hr and (ii) that the C1+2+3S aequorin regenerated for 27 hr o had luminescence activity greater than that of the wild-type 0 5 piQ9-2HE aequorin regenerated with 2-mercaptoethanol (Table 1) prompted us to carry out a time course study in FIG. 3. Recordings of a Ca2+-triggered flash. (A) C1+2+3S which the regeneration of C1+2+3S and piQ9-2HE aequor- aequorin regenerated in the absence of either 2-mercaptoethanol or ins were compared. Unlike the aequorins that had one or two dithiothreitol. (B) Flash of piQ9-2HE aequorin regenerated in the of their cysteine residues replaced with serine, which all presence ofdithiothreitol. Regeneration time was 51 hr. Composition showed lower activity (Table 1), the C1+2+3S aequorin ofthe regeneration mixture and method oftriggering ofthe flash were regenerated in the complete absence ofa reducing agent (Fig. the same as in Fig. 2. Chart speed of the recorder (Soltec, model 2, curve a) had activity greater than that of the piQ9-2HE S-4201) was 12 cm/min. aequorin regenerated in the presence ofdithiothreitol (Fig. 2, curve c) and 2-mercaptoethanol (Fig. 2, curve d). This DISCUSSION showed that the C1+2+3S aequorin was capable of regenerating in the absence of a reducing agent, and the near The aequorin molecule is made up of 189 amino acid residues identity of the two C1+2+3S aequorin regeneration curves arranged in a single polypeptide chain (10, 11). The primary (Fig. 2, curves a and b) indicated that dithiothreitol did not structure, the three Ca2+-binding sites, and the biolumines- influence the regeneration of C1+2+3S aequorin. In con- cent reaction of aequorin have been described (15). Accord- trast, the regeneration of C1+2+3S aequorin was strongly ing to our current knowledge, aequorin is a protein-sub- inhibited by 2-mercaptoethanol (Fig. 2, curve e). Interest- strate-02 complex requiring only the binding ofthree Ca2+ to ingly, although it took piQ9-2HE aequorin about 6 hr to reach convert the protein moiety into a catalytically active enzyme full activity, it took C1+2+3S aequorin almost 18 hr to reach (oxygenase). Once activated, the protein catalyzes the oxi- peak activity. As found with piQ5 aequorin, luminescence dation ofprebound coelenterazine as in an enzyme-catalyzed activity ofpiQ9-2HE aequorin was greater when regeneration reaction, with a single turnover. The function of cysteine was carried out with dithiothreitol than with 2-mercapto- residues is not clear, but because they are rarely found in ethanol. Ca2+-binding proteins (15), it may be assumed that they are Ca2 -Triggered Flash of C1+2+3S Aequorin. Recordings involved in the bioluminescence of aequorin in some way, ofthe Ca2+-triggered flash ofC1 +2+3S aequorin regenerated either in a catalytic role or in the regeneration reaction. in the absence of a reducing agent and piQ9-2HE aequorin Five components are essential for the regeneration of regenerated in the presence ofdithiothreitol are shown in Fig. aequorin (9). EDTA is required to chelate Ca2+. The purpose 3. The flashes are seen to be virtually identical to each other. of 2-mercaptoethanol is not clear, but it has been found

Q FIG. 2. Time course of regeneration of piQ9- O.)01) 2HE and C1+2+3S aequorins in the presence and 0) absence of 2-mercaptoethanol and dithiothreitol. 0 Curves: a, C1+2+3S alone; b, C1+2+3S and x dithiothreitol; c, piQ9-2HE and dithiothreitol; d, a piQ9-2HE and 2-mercaptoethanol; e, C1+2+3S and 2-mercaptoethanol. piQ9-2HE alone had min- a- imal to no activity (not shown). Composition of the H> regeneration mixture at time zero: 1200 Al of cell extract, 12 ,g ofcoelenterazine (1 ,ug/,ul ofabsolute C) methyl alcohol), and 24 ,ul of 2-mercaptoethanol or 48 p.1 of 52 mM dithiothreitol/30 mM Tris HCI, pH 7.60/10 mM EDTA or no reducing agent but addi- tion of 30 mM Tris HCI, pH 7.60/10 mM EDTA to attain a final volume of 1260 ,ul. Assay of mixtures was performed in duplicate on 100-pl aliquots removed at designated time intervals from regen- REGENERATION TIME (hr) eration mixtures kept in an ice bath. Downloaded by guest on September 29, 2021 Biochemistry: Kurose et al. Proc. Natl. Acad. Sci. USA 86 (1989) 83 necessary for regeneration of full activity; presumably, it aequorin (Fig. 3). The fact that the regeneration of C1+2+3S reduces disulfide bonds. Coelenterazine is believed to be aequorin is sharply inhibited by 2-mercaptoethanol suggests bound noncovalently to apoaequorin because during purifi- that the SH group of 2-mercaptoethanol somehow interferes cation of native aequorin, it is not dissociated through with the OH group of serine during the regeneration of successive steps in gel filtration (7, 25). Coelenterazine is also apoaequorin, which is not evident with dithiothreitol (Fig. 2). not extracted by organic solvents (6), becoming dissociable This may explain why the luminescence activities of piQ9- only after the completion of the light reaction (8). Molecular 2HE aequorin (Fig. 2) and piQ5/D1210 aequorin are lower oxygen is bound tightly because the luminescence reaction when the apoaequorins are regenerated with 2-mercapto- can take place in the complete absence of air. While the ethanol than with dithiothreitol. A previous proton NMR nature of the noncovalent binding of coelenterazine to study (14) has shown that native aequorin undergoes com- apoaequorin is not known (5), it has been shown recently that plete unfolding during slow, Ca2+-independent light emission an ionic bond between a negatively charged phenolic hy- (autoxidation), suggesting that apoaequorin may be in the droxyl of coelenterazine and a positively charged group in form of a random coil prior to regeneration. The long apoaequorin is not required in regenerating active aequorin regeneration time for apoaequorin suggests that the protein (26). It is possible that cysteine may be involved in coelen- undergoes a folding reaction to form a binding site for terazine binding because treatment of aequorin with the coelenterazine and molecular oxygen. For piQ9-2HE apoae- sulfhydryl reagent N-ethylmaleimide leads to 95% loss of quorin, this may involve reduction of disulfide bonds fol- Ca2+-triggered activity in 15 min (6) and labeling the SH lowed by folding, whereas in the case of C1+2+3S apoae- group of aequorin with a maleimide spin-label or fluorescent quorin, it would involve only folding. label (Acrylodan) causes the release of coelenterazine from the molecule, accompanied by complete loss of activity (13). This work was supported in part by a research grant (DMB The wide variation in luminescence activity found in Table 85-17578) from the National Science Foundation (to F.I.T.) and by 1 for modified aequorins with one of the three cysteine a grant-in-aid (62105001) from the Ministry of Education, Culture, residues replaced (CiS, C2S, C3S) and with two of the three and Science of Japan (to Y.S.). cysteine residues replaced (C1+2S, C1+3S, C2+3S) is not easily explainable. The SH group of cysteine and the OH 1. Shimomura, O., Johnson, F. H. & Saiga, Y. (1962) J. Cell. group of serine differ only slightly from each other in Comp. Physiol. 59, 223-240. geometry, and replacing the cysteine residues in aequorin 2. Shimomura, O., Johnson, F. H. & Saiga, Y. (1963) Science 140, with be to cause a 1339-1340. serine may expected only small perturba- 3. Shimomura, 0. & Johnson, F. H. (1976) Symp. Soc. Exp. Biol. tion in the tertiary structure ofthe protein. The van der Waals 30, 41-54. radius for sulfur is 1.85 A, and that for oxygen is 1.40 A. The 4. Musicki, B., Kishi, Y. & Shimomura, 0. (1986) J. Chem. Soc. C-C-S bond angle for the cysteine SH differs little from the Chem. Commun., 1566-1568. C-C-O bond angle of the serine OH, and the CH-O-H 5. Shimomura, O., Johnson, F. H. & Morise, H. (1974) Biochem- bond angle ofthe serine OH is about 1080 compared to around istry 13, 3278-3286. 980 for the C-S-H bond angle of the cysteine SH (27-30). 6. Shimomura, 0. & Johnson, F. H. (1978) Proc. Natl. Acad. Sci. Chemically, the SH group of cysteine is a more effective USA 75, 2611-2615. nucleophile than the OH group of serine, and the serine OH 7. Shimomura, 0. & Johnson, F. H. (1978) Methods Enzymol. 57, is able to form a much bond A 271-291. stronger hydrogen (31, 32). 8. Shimomura, 0. & Johnson, F. H. (1973) Tetrahedron Lett. similar study has been reported only once previously in 2963-2966. which replacement of the active-site cysteine of tyrosyl- 9. Shimomura, 0. & Johnson, F. H. (1975) Nature (London) 256, tRNA synthetase by serine resulted in reduced binding of 236-238. ATP substrate (33). The loss in aequorin activity also cannot 10. Inouye, S., Noguchi, M., Sakaki, Y., Takagi, Y., Miyata, T., be attributed to protease digestion or to a difference in Iwanaga, S., Miyata, T. & Tsuji, F. I. (1985) Proc. Natl. Acad. apoaequorin concentration. Previously it was shown that Sci. USA 82, 3154-3158. extracts ofapoaequorin prepared from E. coli D1210 carrying 11. Charbonneau, H., Walsh, K. A., McCann, R. O., Prendergast, the aequorin plasmid piQ5 have activities significantly F. G., Cormier, M. J. & Vanaman, T. C. (1985) Biochemistry than of extracts 24, 6762-6771. greater those prepared from protease- 12. Inouye, S., Sakaki, Y., Goto, T. & Tsuji, F. I. (1986) Biochem- deficient E. coli BNN103 carrying the piQ5 plasmid (12). In istry 25, 8425-8429. the present study, the luminescent activity of the modified 13. Kemple, M. D., Ray, B. D., Jarori, G. K., Nageswara Rao, C1+2+3S aequorin was also observed to be greater than that B. D. & Prendergast, F. G. (1984) Biochemistry 23, 4383-4390. of the wild-type piQ9-2HE aequorin. Further, apoaequorin 14. Ray, B. D., Ho, S., Kemple, M. D., Prendergast, F. G. & has been found to be very stable chemically (12, 34). Klett Nageswara Rao, B. D. (1985) Biochemistry 24, 4280-4287. values for cell densities were nearly the same, indicating that 15. Tsuji, F. I., Inouye, S., Goto, T. & Sakaki, Y. (1986) Proc. approximately the same number ofcells were present in each Nati. Acad. Sci. USA 83, 8107-8111. culture medium and presumably the same concentration of 16. Caruthers, M. H. (1987) in Synthesis and Application ofDNA was in the extracts. andRNA, ed. Narang, S. A. (Academic, New York), pp. 47-94. apoaequorin present 17. Inoue, S., Sugiura, S., Kakoi, H., Hasizume, K., Goto, T. & In the experiments described, the regeneration time for Iio, H. (1975) Chem. Lett. 141-144. C1+2+3S aequorin in the absence of either dithiothreitol or 18. deBoer, H. A., Comstock, L. J. & Vasser, M. (1983) Proc. 2-mercaptoethanol was almost 3 times longer (18 hr) than the Natl. Acad. Sci. USA 80, 21-25. regeneration time for the wild-type piQ9-2HE aequorin (6 hr) 19. Vieira, J. & Messing, J. (1982) Gene 19, 259-268. in the presence of dithiothreitol or 2-mercaptoethanol (Fig. 20. Inouye, S. & Inouye, M. (1987) in Synthesis andApplication of 2). However, after 27 hr, the activity of the C1+2+3S DNA and RNA, ed. Narang, S. A. (Academic, New York), pp. aequorin reached a value equal to or greater than the activity 181-206. of the 2 and Table This suggests 21. Hattori, M. & Sakaki, Y. (1986) Anal. Biochem. 152, 232-238. wild-type aequorin (Fig. 1). 22. Mitchell, G. W. & Hastings, J. W. (1971) Anal. Biochem. 39, that the modification ofthe three cysteine residues affects the 243-250. regeneration reaction rather than the catalytic reaction. This 23. Hastings, J. W. & Weber, G. (1963) J. Opt. Soc. Am. 53, 1410- view is also supported by the fact that the initial maximal light 1415. intensity and shape of the curve of the C1+2+3S aequorin 24. Hastings, J. W. & Weber, G. (1965) Photochem. Photobiol. 4, flash triggered by Ca2' are virtually identical to those of the 1049-1050. flash triggered by Ca2+ when mixed with wild-type piQ9-2HE 25. Blinks, J. R., Mattingly, P. H., Jewell, B. R., van Leeuwen, Downloaded by guest on September 29, 2021 84 Biochemistry: Kurose et al. Proc. Nati. Acad. Sci. USA 86 (1989)

M., Harrer, G. C. & Allen, D. G. (1978) Methods Enzymol. 57, 30. Wilkinson, A. J., Fersht, A. R., Blow, D. M. & Winter, G. 292-328. (1983) Biochemistry 22, 3581-3586. 26. Shimorium, O., Musicki, B. & Kishi, Y. (1988) Biochem. J. 31. Jocelyn, P. C. (1972) Biochemistry ofthe SHGroup (Academic, 251, 405-410. New York), pp. 47-62. 27. Kerr, K. A., Ashmore, J. P. & Koetzle, T. F. (1975) Acta 32. Crampton, M. R. (1974) in Chemistry ofthe SHGroup, ed. Pati, Crystallogr. Sect. B 31, 2022-2026. S. (Wiley-Interscience, New York), pp. 379-415. 28. Frey, M. N., Lehmann, M, S., Koetzle, T. F. & Hamilton, 33. Winter, G., Fersht, A. R., Wilkinson, A. J., Zoller, M. & W. C. (1973) Acta Crystallogr. Sect. B 29, 876-884. Smith, M. (1982) Nature (London) 299, 756-758. 29. Kistenmacher, T. J., Rand, G. A. & Marsh, R. E. (1974) Acta 34. Shimomura, 0. & Shimomura, A. (1981) Biochem. J. 199, 825- Crystallogr. Sect. B 30, 2573-2578. 828. Downloaded by guest on September 29, 2021