- Email: [email protected]
A site-directed mutagenesis analysis of tNOX functional domains
Biochimica et Biophysica Acta 1594 (2002) 74^83 www.bba-direct.com
A site-directed mutagenesis analysis of tNOX functional domains Pin-Ju Chueh a , Dorothy M. Morre¨ b , D. James Morre¨ a
a;
*
Department of Medicinal Chemistry and Molecular Pharmacology, Purdue University, West Lafayette, IN 47907, USA b Department of Foods and Nutrition, Purdue University, West Lafayette, IN 47907, USA Received 17 May 2001; received in revised form 11 September 2001; accepted 21 September 2001
Abstract Constitutive NADH oxidase proteins of the mammalian cell surface exhibit two different activities, oxidation of hydroquinones (or NADH) and protein disulfide-thiol interchange which alternate to yield oscillatory patterns with period lengths of 24 min. A drug-responsive tNOX (tumor-associated NADH oxidase) has a period length of about 22 min. The tNOX cDNA has been cloned and expressed. These two proteins are representative of cycling oxidase proteins of the plant and animal cell surface. In this report, we describe a series of eight amino acid replacements in tNOX which, when expressed in Escherichia coli, were analyzed for enzymatic activity, drug response and period length. Replacement sites selected include six cysteines that lie within the processed plasma membrane (34 kDa) form of the protein, and amino acids located in putative drug and adenine nucleotide (NADH) binding domains. The latter, plus two of the cysteine replacements, resulted in a loss of enzymatic activity. The recombinant tNOX with the modified drug binding site retained activity but the activity was no longer drug-responsive. The four remaining cysteine replacements were of interest in that both activity and drug response were retained but the period length for both NADH oxidation and protein disulfide-thiol interchange was increased from 22 min to 36 or 42 min. The findings confirm the correctness of the drug and adenine nucleotide binding motifs within the tNOX protein and imply a potential critical role of cysteine residues in determining the period length. ß 2002 Elsevier Science B.V. All rights reserved. Keywords: NADH oxidase; Protein disul¢de-thiol interchange; Tumor-associated NADH oxidase; Cycling oxidase; Capsaicin; Site-directed mutagenesis; Growth; Biological clock
1. Introduction A recently described family of cell surface proteins
Abbreviations: DTDP, dithiodipyridine ; NOX, plasma membrane-located NADH (hydroquinone) oxidase with protein disul¢de-thiol interchange activity; tNOX, tumor-associated NOX; H-NOX, recombinant truncated tNOX; CNOX, constitutive NOX; CLOX, cycling oxidase; capsaicin, 8-methyl-N-vanillyl-6noneamide; LY181984, N-(4-methylphenylsulfonyl-NP-(4-chlorophenyl) urea * Corresponding author. Fax: +1-765-494-4007. E-mail address: [email protected] (D.J. Morre¨).
that oxidize hydroquinones or NADH (NADH oxidase = NOX) is characterized uniquely by the capacity to carry out two distinct biochemical activities not simultaneously but in series [1]. The two activities, hydroquinone (NADH) oxidation and protein disul¢de-thiol interchange alternate to give rise to a period length of 24 min for the constitutive NOX activity (CNOX) [2]. The cDNA for a tumor-associated NOX protein (tNOX) has been cloned and expressed in COS cells and Escherichia coli (GenBank accession No. AF207881 [3]). tNOX activities also alternate to generate a pattern of oscillations but the period
0167-4838 / 02 / $ ^ see front matter ß 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 8 3 8 ( 0 1 ) 0 0 2 8 6 - 2
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
Fig. 1. NOX activities with assays based on disappearance of NADH or cleavage of DTDP measured at 340 nm for wild type recombinant truncated tNOX protein (ttNOX-1) expressed in E. coli. (A) Oxidation of NADH. (B) Cleavage of DTDP. In A, maxima in NADH oxidation labeled 1 and 2 recur at 21.5 (approx. 22) min intervals. Minor in£ections labeled 3, 4 and 5 also recur at 21.5 (approx. 22) min intervals. With protein disul¢de-thiol interchange measured in parallel in B, the major in£ections of A align with minor in£ections labeled 1 and 2 and the minor in£ections of A align with maxima labeled 3, 4 and 5 to generate the recurrent 2+3 patterns of activity alternations that appear to characterize all CLOX proteins. These data are representative of triplicate determinations and recapitulate the 21.5 (approx. 22) min period length of a processed tNOX form of the HeLa cell surface [4].
length of about 22 min is shorter than that of CNOX [4]. The NOX oscillations are not simple sine functions but, rather, represent complex 2+3 oscillatory patterns [5,6]. In the oxidizing portion of the cycle, hydroquinone or NADH oxidation is represented by two maxima followed by three minor oscillations. When protein disul¢de-thiol interchange is measured, the two maxima of NADH oxidation are the minor oscillations whereas the three minor oscillations of NADH oxidation now become major oscillations [5]. These relationships are illustrated in Fig. 1 for the wild type tNOX of HeLa cells.
75
Each of the components of the 3+2 pattern of oscillations repeats with a precise and temperaturecompensated (independent of temperature) period length of approx. 22 min [6]. NOX proteins at the cell surface or in solution function synchronously [6]. Autosynchrony of the protein is achieved in solution if two asynchronous preparations are mixed, for example [7]. In this paper we report the response of the activities of recombinant truncated tNOX (ttNOX-1) to amino acid replacements within putative functional domains as well as the response of period length to cysteine to alanine replacements. The truncated form of the protein was utilized as being representative of the processed 34 kDa form of the protein found at the cell surface of cancer cells and in the sera of cancer patients [1]. In those cysteine to alanine replacements where activity was retained, period length was modi¢ed in a manner suggesting that cysteines of the NOX molecule contribute to the determination of period length. 2. Methods 2.1. Mutagenic oligonucleotides and site-directed mutagenesis Eight sets of oligonucleotides listed in Table 1 were designed to replace amino acid residues potentially involved in tNOX activity by site-directed mutagenesis according to Braman et al. [8]. Cysteines C505, C510, C558, C569, C575, and C602 were replaced by alanines. A methionine of the putative drug binding site was replaced by alanine (M396A). A glycine in the potential adenine binding site was replaced by valine (G592V). High ¢delity thermostable Pfu DNA polymerase, low cycle number, and primers designed only to copy the parental strand in a linear fashion were used to minimize unwanted second site mutations. Double-stranded, super-coiled expression plasmid pET11ttNOX (40 ng) and mutagenic sense and antisense primers (100 ng) were combined in a 50 Wl reaction mixture containing deoxyribonucleotides, bu¡er, and Pfu DNA polymerase according to the manufacturer's protocol (Stratagene, La Jolla, CA, USA). The cycling parameters were 95³C for 30 s,
BBAPRO 36514 24-1-02
76
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
Table 1 The mutagenic oligonucleotides for site-directed mutagenesisa Amino acid replacement
Oligonucleotides
C505A
5P-GAAAAGGAAAGCGCCGCTTCTAGGCTGTGTGCC-3P (forward) 5P-GGCACACAGTCTAGAAGCGGCGCTTTCCTTTTC-3P (reverse) 5P-GCTTCTAGGCTGGCCGCCTCAAACCAGGATAGCG-3P (forward) 5P-CGCTATCCTGGTTTGAGGCGGCCAGCCTAGAAGC-3P (reverse) 5P-GCAAGCATTGAATACATCGCTTCCTACTTGCACCGTCTTG-3P (forward) 5P-CAAGACGGTGCAAGTAGGAAGCGATGTATTCAATGCTTGC-3P (reverse) 5P-CGTCTTGATAATAAGATCGCCACCAGCGATGTGGAGTG-3P (forward) 5P-CACTCCACATCGCTGGTGGCGATCTTATTATCAAGACG-3P (reverse) 5P-CCAGCGATGTGGAGGCCCTCATGGGTAGACTCC-3P (forward) 5P-GGAGTCTACCCATGAGGGCCTCCACATCGCTGG-3P (reverse) 5P-GAAAAGAGATGGAAATTCGCTGGCTTCGAGGGCTTGAAG-3P (forward) 5P-CTTCAAGCCCTCGAAGCCAGCGAATTTCCATCTCTTTTC-3P (reverse) 5P-GTCTGATGATGAAATAGAAGAAGCGACAGAAACAAAAGAAACTGAGG-3P (forward) 5P-CCTCAGTTTCTTTTGTTTCTGTCGCTTCTTCTATTTCATCATCAGAC-3P (reverse) 5P-CAGGAAATGACTGGAGTTGTGGCCAGCCTGGAAAAGAG-3P (forward) 5P-CTCTTTTCCAGGCTGGCCACAACTCCAGTCATTTCCTG-3P (reverse)
C510A C558A C569A C575A C602A M396A G592V a
Nucleotide replacements are in bold.
55³C for 1 min, and 68³C for 12.8 min, 16 cycles. The linear ampli¢cation product was treated with endonuclease DpnI (10 units/Wl) for 1 h to eliminate the parental template. Subsequently, 4 Wl of this reaction mixture containing the double-nicked mutated plasmid were used for the transformation of supercompetent E. coli XL-1 Blue cells (Stratagene). DNA sequencing was utilized to con¢rm that the correct replacements were obtained in all mutants. 2.2. Protein expression in E. coli The replacements were expressed in bacteria and puri¢ed as recombinant truncated tNOX forms. The recombinant tNOX protein designated ttNOX-1, comprised of amino acid residues 220^610 of tNOX protein (391 amino acids), retained full functional activity. NdeI and BamHI restriction sites were incorporated to accommodate subcloning into the protein expression vector pET-11b. The ampli¢cation was performed as follows: 94³C, 1 min 30 s/55³C, 1 min 30 s/72³C, 1 min 30 s/29 cycles/72³C, 5 min. The open reading frame of truncated tNOX cDNA was ampli¢ed by PCR. The PCR product was digested with NdeI plus BamHI and ligated into the pET-11b vector, which had been digested by NdeI and BamHI and dephosphorylated by calf intestine alkaline phosphatase. The resultant vector was transformed into E. coli (XL-1Blue). Plasmid pET11-
ttNOX was puri¢ed and transformed into E. coli (BL21(DE3)). The transformed bacteria were grown on ampicillin plates. Single colonies were used to inoculate LB medium containing 100 Wg/ml ampicillin. Cells were grown for 16^20 h at 25^37³C. Since plasmid pET11-ttNOX is a leaky plasmid, IPTG was not required to induce expression. The cells were harvested by centrifugation (6000 rpm 5 min, Sorvall SS-34 rotor). The cell pellet was then washed in 20 mM Tris^HCl, pH 8.0 and ¢nally resuspended in 20 mM Tris^HCl, pH 8.0 containing 10 mM DTT and protease inhibitors (1 mM PMSF, 1 mM benzamidine and 1 mM 6-aminocaproic acid). The cells were then lysed with a French press (20 000 psi, three passages) and the cell lysates were centrifuged at 9000^11 000 rpm for 20 min (Sorval SS-34 rotor). The supernatant was collected and saved as the source of solubilized ttNOX-1 protein. The ttNOX-1 was then precipitated by ammonium sulfate 20% of saturation. Once precipitated, ttNOX-1 was not soluble in 20 mM Tris^HCl, pH 8.0. Refolding of ttNOX-1 was according to the procedure provided by Novagen (Madison, WI, USA). Brie£y, the pellet was washed with 1% Triton X-100 and the recombinant ttNOX proteins were dissolved in 50 mM CAPS, pH 11, containing 0.3% N-laurylsarcosine (sodium salt). Proteins were then dialyzed in two changes of dialysis bu¡er (20 mM Tris^HCl, pH 8.5 containing 0.5 mM cysteamine and 0.05 mM
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
cystamine) over a period of 14 h followed by dialysis in two changes of equilibration bu¡er (20 mM Tris^ HCl, pH 7.5, 0.1 mM DTT). After dialysis, ttNOX-1 protein was stored at 320³C until further use. Expression of ttNOX-1 was analyzed by SDS^PAGE with silver staining and Western blotting. Western blotting used anti-tNOX monoclonal and polyclonal antibodies. The antibody binding was detected using an alkaline phosphate conjugated to anti-mouse IG antibody or anti-rabbit IG antibody. 2.3. Spectrophotometric assay of tNOX activities NADH oxidase activity was determined as the disappearance of NADH measured at 340 nm in a reaction mixture containing 25 mM Tris^MES bu¡er (pH 7.2), 1 mM KCN to inhibit low levels of mitochondrial oxidase activity, and 150 WM NADH at 37³C with temperature control ( þ 0.5³C) and stirring [9]. Activity was measured using paired Hitachi U3210 spectrophotometers. Assays were initiated by addition of NADH. Assays were for 1 min and were repeated on the same sample every 1.5 min for the times indicated. A millimolar extinction coe¤cient of 6.22 was used to determine speci¢c activity. Protein disul¢de-thiol interchange activity associated with the tNOX protein was measured using a dipyridyldithio substrate, dithiodipyridine (DTDP) [10]. The assay was in 50 mM Tris^MES bu¡er (pH 7.0). The assay was initiated by the addition of 0.5 Wmole DTDP in 5 Wl of DMSO. The reaction was monitored from the increase in absorbance at 340 nm using the pyridinethionine product. The speci¢c activity was calculated using a millimolar absorption coe¤cient of 6.2 cm31 . Proteins were estimated by the bicinchoninic acid method [11] with bovine serum albumin as standard. 3. Results Expression of recombinant truncated tNOX (ttNOX-1) was veri¢ed by Western blot analysis for all eight amino acid replacements compared to wild type. Assays used bacterial extracts with approximately equivalent amounts of recombinant protein as determined from intensities of the ttNOX-1 protein band of the Western blots.
77
The oscillation of NADH oxidase activity of recombinant ttNOX-1 exhibited maxima at approx. 22 min intervals for both NADH oxidation (Fig. 1A) and DTDP cleavage (Fig. 1B). The precise length over three periods was 64.5 min or an average single period length of 21.5 min. As is customary, the even possibility was used as the convention for rounding o¡, hence a period length of 22 min. The oscillations were not simple sine waves but varied with a complex, recurring pattern (Fig. 1). When both enzymatic activities, measured in parallel using two di¡erent spectrophotometers on equal portions of the same preparation, were analyzed in greater detail, a total of ¢ve oscillatory components were resolved within each 22 min period (Fig. 1). With NADH as substrate for the oxidative activity (Fig. 1A), the oscillatory components labeled 1 and 2 were most evident. With DTDP as substrate for the protein disul¢de-thiol interchange activity (Fig. 1B), the oscillatory components labeled 3, 4 and 5 were most evident. As reported previously for recombinant tNOX protein [1], the mean times of the dominant oscillatory components of the two activities were 12 min out of phase with each other. In Fig. 1, with NADH as substrate, the two main oscillations were at 18, 40, 61 and 81 min and 20, 42, 60 and 85 min. In addition, there were three minor oscillations labeled 3, 4 and 5 to complete each 22 min period. With DTDP as substrate, the major oscillations were at 6, 26, 48 and 69 min, at 9, 31, 51 and 72 min and at 12, 34, 54 and 74 min. These major oscillations with DTDP as substrate corresponded exactly to the minor oscillations of the oxidative activity with NADH as substrate. The minor oscillations with the DTDP substrate were at 18, 40, 61 and 81 min and at 20, 42, 64 and 85 min which corresponded to the times of the major oscillations in NADH oxidation. Thus, the two major oscillations of NADH oxidation aligned with the two minor oscillations of DTDP cleavage. Similarly, the three major oscillations of DTDP cleavage aligned with the three minor oscillations of NADH oxidation. This repeating 3+2 pattern of activity occurred reproducibly and has been validated by time series decomposition analyses [5]. The M396A replacement in the putative drug binding site (see Section 4) did not eliminate either of the two enzymatic activities associated with the
BBAPRO 36514 24-1-02
78
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
Fig. 2. Assay as in Fig. 1 of the M396A replacement within the putative drug binding site. An activity pattern similar but not identical to that of Fig. 1 was observed but capsaicin no longer inhibited. (A) NADH oxidation. (B) DTDP cleavage.
ttNOX-1 (Fig. 2). Additionally, the period length remained unchanged at approx. 22 min (Table 2). However, neither activity was any longer inhibited by capsaicin. Even after 1 h or more, the activity in the presence of capsaicin was equivalent to that in the absence of capsaicin (Fig. 2). The period length of 21.5 (approx. 22) min was reproduced in each of three separate experiments. The M396A replacement did appear to alter the pattern of oscillations, however. In three experiments, component 2 of NADH oxidation was reduced (Fig. 2). This reduction in NADH oxidative activity was compensated in part by an overall
broadening of the maxima labeled 1 and by a possible enhancement of peak 2 with DTDP cleavage. The G592V replacement in the putative adenine binding domain (see Section 4) resulted in a loss of enzymatic activity (Fig. 3). With this and the other replacements lacking activity, capsaicin was without e¡ect as well. When the ttNOX proteins with six di¡erent cysteine to alanine replacements were expressed in E. coli, those containing C505A or C569A no longer exhibited NADH oxidation or cleavage of DTDP (Fig. 3). The four other cysteine replacements retained both NADH oxidase and protein disul¢dethiol interchange activities but the period lengths were altered (Table 1). For the C510A replacement (Fig. 4) as well as for the C575A and the C602A replacements (Fig. 5), the period lengths were increased to 36 min. For the C558A replacement, the period length was 42 min (Fig. 5A). The increased period lengths were observed with both the oxidase activity with NADH as substrate and for the protein disul¢de-thiol interchange activity with DTDP as substrate. The expressed proteins carrying the cysteine replacements yielding period lengths of 36 or 42 min recapitulated the complex 2+3 pattern of activity oscillations. With the 36 min periods, the oxidative activity exhibited two major activity maxima separated by about 12 min (1 and 2) and three minor in£ections (3, 4 and 5). The protein disul¢de-thiol interchange activity, on the other hand, exhibited two minor in£ections and three maxima. The two maxima of the NADH oxidation and the two minor in£ections of the protein disul¢de-thiol interchange
Table 2 E¡ects of site-directed mutagenesis of ttNOX on NADH oxidase activity, period length and inhibition by capsaicin Amino acid replacementa
Enzymatic activity (nmoles/min/mg protein)
Period length (min)
Inhibition by 1 WM capsaicin
None C505A C510A C558A C569A C575A C602A M396A G592V
0.9 þ 0.1 None 0.85 þ 0.05 0.9 þ 0.15 None 0.8 þ 0.05 0.8 þ 0.05 0.85 þ 0.1 None
22
Complete
36 42
Complete Complete
36 36 22
Complete Complete None
a
Resequencing con¢rmed the expected DNA sequences for each of the amino acid replacements indicated.
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
79
coincided when assayed in parallel. Similarly the three minor in£ections of the NADH oxidation coincided with the three maxima of the protein disul¢de-thiol interchange when assayed in parallel. Both enzymatic activities of each of the four cysteine replacements with enzymatic activity were also inhibited completely by capsaicin (Figs. 4 and 5). For each of the four replacements, assays were continued for 80 min in separate experiments (not shown). After approx. 20 min, both NADH oxidation and DTDP cleavage activities reached baseline in the presence of capsaicin and remained at baseline for the duration of the experiment.
Fig. 4. NOX activities assayed as in Fig. 1 for the C510A replacement. (A) Oxidation of NADH. (B) Cleavage of DTDP. In A, maxima labeled 1 and 2 spaced about 12 min apart recur at 36 min intervals. Also recurrent at 36 min intervals were minor oscillations labeled 3, 4 and 5. With the protein disul¢dethiol interchange measured in parallel (B), the minor in£ections of A aligned with maxima labeled 3, 4 and 5 to generate the recurrent 2+3 pattern of activity alternations seen in Fig. 1 but with the recurrent maxima separated at intervals of 36 min rather than intervals of 22 min. Additionally, the ¢rst of the three maxima labeled 3 determined from cleavage of DTDP appears between maxima labeled 1 and 2 instead of after maxima labeled 1 and 2 as in Fig. 1. The values shown are representative of duplicate determinations for each activity.
Fig. 3. Assay as in Fig. 1 of the C505A, C569A and G592V replacements where functional activities were lost. (A) NADH oxidation. (B) Cleavage of DTDP. No signi¢cant regular periodicities above background were noted. Capsaicin was without e¡ect (not shown). Data are representative of duplicate (B) or triplicate (A) determinations for each replacement.
While each of the three cysteine replacements with period lengths of 36 min exhibited overall close similarities, i.e. the recurrent 2+3 pattern of activity oscillations within a 36 min period, unique features were revealed for each which appeared reproducible. Compared to the C510A replacement (Fig. 4), the C575A replacement exhibited more pronounced activity components labeled 4 and 5 with the DTDP substrate compared to activity components labeled 3 (Fig. 5). The C602A replacement exhibited much more closely spaced components designated as 1 and 3 compared to the other cysteine replacements (Fig. 5).
BBAPRO 36514 24-1-02
80
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
4. Discussion Previous studies have identi¢ed an unusual NADH oxidase activity at the external cell surface of the
plasma membrane of both plant and animal cells [1]. The physiological function of the oxidative portion of the NOX cycle appears to be that of a hydroquinone oxidase [12] even though NADH oxidation provides a convenient measure of the enzymatic activity [9]. These proteins are unique not only from their plasma membrane location but also from their oscillating activities that provide for potential roles as time-keeping proteins [13] and a relationship between the oscillatory enzymatic activity and the enlargement phase of cell growth [1,6]. The two di¡erent enzymatic activities of the NOX proteins, hydroquinone oxidation and protein disul¢de-thiol interchange, alternate. With the normal constitutive (CNOX) protein, an approx. 24 min period is generated as a result [1,5,6]. With tNOX forms present together with CNOX at the surfaces of cancer cells, three period lengths of 24 þ 0.3, 22.5 þ 0.6 min and 21.5 þ 0.6 min were observed for HeLa (human cervical carcinoma) cells [4]. The latter period length corresponds exactly with the 21.5 min period length of the recombinant ttNOX reported here. Both tNOX and CNOX appear to be members of a family of NADH oxidases with activities that cycle with regular period lengths independent of temperature (CLOX family of cell surface cycling oxidases). While several NOX forms belonging to this CLOX family may exist, the drug-responsive tNOX is the ¢rst to be cloned (GenBank accession No. AF207881). tNOX di¡ers from the constitutive CNOX form primarily in its sensitivity to capsaicin [14], (3)-epigallocatechin 3-gallate (EGCg) [15], several anticancer drugs including adriamycin [16] and to thiol reagents [17]. The response of tNOX activity to capsaicin was used to guide puri¢cation of the processed tNOX protein from sera of cancer patients
6
Fig. 5. NOX activities assayed as in Fig. 1 for the C558A, C575A and C602A replacements all of which displayed patterns similar to those of the C510A replacement of Fig. 4. (A) Oxidation of NADH. (B) Cleavage of DTDP. Maxima generating the 2 (NADH)+3 (DTDP) pattern recurred at intervals of about 36 min for C575A and C602A replacements and at intervals of 42 min for the C558A replacement. The experiments shown are representative of duplicate or triplicate determinations for each activity.
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
and as the basis for the monoclonal antibody selection used to obtain the cloned cDNA [3]. A number of motifs within the tNOX sequence account for the range of its functional activities [1]. A NADH binding site was indicated earlier from binding of photoa¤nity-labeled [32 P]NAD(H). A putative adenine nucleotide binding sequence motif (GX-G-X-X-G) with downstream remote acidic amino acid residues D or E [18] was represented by T589-GV-G-A-S-L and E605 near the C-terminus and compared closely with the sequence T-G-V-G-A-G-V-G from mitochondrial ATP synthase subunit 9 from Chondrus crispus [19]. The NOX protein binds the antitumor sulfonylurea LY181984 [20]. Activity is inhibited or stimulated by LY181984 depending on the redox environment of the protein [21]. The protein oxidized reduced coenzyme Q [12] and substances such as capsaicin [14], EGCg [15] and adriamycin [16], which inhibit the activity, are considered to occupy quinone sites. Ubiquinone competes with both LY181984 binding and inhibition of enzymatic activities. By analogy with several quinone binding proteins of the photosystem II complex of chloroplasts, a site containing a methionine-histidine pair has been equated with the quinone binding site of pyruvate oxidase [22]. Such sites are targets for all known urea and sulfonylurea herbicide inhibitors of photosystem II [23]. Apparently arginine can substitute for the critical histidine. A consensus sequence for the amino acids surrounding the charged residues critical to sulfonylurea and quinone binding site was determined to be A-M-H-G or a closely related sequence based on the above considerations. The putative quinone binding site of the D1 protein of a cyanobacterium, Synechococcus, contains the sequence E-T-MR-E. A sequence similar to E-T-M-R-E is present in the NADH-ubiquinone oxidoreductase of chloroplasts. We found a sequence E394-E-M-T-E as a potential quinone site having neither H nor R in the 4th position but still with considerable similarity to other putative quinone and/or sulfonylurea binding sites. The correct identi¢cation of E394-E-M-T-E as the drug and/or quinone binding site of tNOX is supported by the present ¢ndings. The recombinant protein with the M396A replacement retained NADH oxidase activity but the activity was no longer inhibited by capsaicin (Table 1, Fig. 2).
81
Initially, the protein disul¢de-thiol interchange activity of the tNOX protein was based on the restoration of activity to reduced, denatured and oxidized (scrambled) yeast RNase through reduction and refolding under non-denaturing conditions and reoxidation to form a correct secondary structure stabilized by internal disul¢de bonds [24]. The activity was similar to that catalyzed by protein disul¢de isomerases of the endoplasmic reticulum [25] yet a different protein motif appeared to be involved since the activity was not altered by the presence of two di¡erent antisera to protein disul¢de isomerases [24]. One of the antisera was a mouse monoclonal (SPA891) (StressGen Biotechnologies) to protein disul¢de isomerase (PDI) from bovine liver (cross-reactive with PDI from human, monkey, rat, mouse and hamster cell lines). The other was a peptide antibody directed to the C-X-X-C motif common to most, if not all, members of the protein disul¢de isomerase family of proteins [26]. This motif is also present in thioredoxin reductase and related proteins where it appears to catalyze the transfer of electrons in conjunction with bound £avin [27^30]. A C-X-X-C motif is absent in tNOX and tNOX lacks thioredoxin reductase activity. tNOX appears not to contain bound £avin nor is its activity dependent on added £avin (FAD or FMN). While the two C-X-X-X-X-C motifs characteristic of £avoproteins are missing from tNOX, the redox active disul¢de of thioredoxin reductase from the malaria parasite Plasmodium falciparum contains a motif C88-X-X-X-X-C93 [29] similar to one found in tNOX. Together with a downstream His509, the motif was shown to be a putative proton donor/acceptor. Either the C88A or the C93A replacement resulted in complete loss of enzymatic activity [29]. A C535-X-X-X-X-C540 motif in the same protein was shown to be involved in substrate coordination and/or electron transfer [30]. A C535A replacement did result in diminution of enzymatic activity but the C540A replacement did not [30]. Thus, either or both of the two comparable motifs present in tNOX, C505-X-X-X-X-C510 or C569-X-X-X-X-X-C575, alone or together with downstream histidines provide potential active sites for protein disul¢de-thiol interchange. With tNOX, the C505A and C569A replacements lost activity as with the C535A replacement above for the P. falciparum protein but the C510A
BBAPRO 36514 24-1-02
82
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
and C575A did not as with the above C540A replacement for the P. falciparum protein. For the four cysteine replacements with enzymatic activity, NADH oxidation and protein disul¢de-thiol interchange were a¡ected in parallel. E. coli cells transfected with ttNOX-1 cDNA with C510A, C558A, C575A or C602A replacement generated expressed ttNXO-1 proteins where the 2+3 pattern characteristic of the cell surface NADH oxidase was still evident. The change from a 22 min period to a 36 or 42 min period was re£ected in the entire 2+3 NOX activity pattern and not just in the principal maxima. Both of the two NADH oxidase maxima were a¡ected equally and were spaced approx. 12 min apart in the mutants compared to about 6 min apart for the unmodi¢ed truncated tNOX. Also oscillating with 36 or 42 min periods were the three maxima of the 2+3 NOX activity pattern associated with the protein disul¢de-thiol interchange activity that alternates with NADH (hydroquinone) oxidation. The precision and relative immutability of the pattern of NOX activity oscillations, together with a lack of dependence of period length on temperature (temperature compensation) [7,32] and entrainability by light [33], make the NOX proteins candidates as cellular time-keeping proteins. Expanding on preliminary work [31], experiments to transfect appropriate cell lines with the cDNAs carrying cysteine replacements having a period length of 36 or 42 min to determine e¡ects on higher order periodic responses including circadian metabolic oscillations normally with period lengths of 24 h are in progress. Acknowledgements We thank Lian-Ying Wu for assistance with cell culture, Andrew Brightman, Xiaoyu Tang and Zensui Tian for access to unpublished data, Dongjiao Zhao and Brett Carnahan for technical assistance and Peggy Runck for manuscript preparation. Work supported in part by NASA NAG2-1334, NIH 5RO1CA75461-03 and NIH 1P50AT00477-01 (Botanical Center for Age-Related Diseases).
References [1] D.J. Morre¨, in: H. Asard, A. Be¨rczi, R.J. Caubergs (Eds.), Plasma Membrane Redox Systems and their Role in Biological Stress and Disease, Kluwer Academic Publishers, Dordrecht, 1998, pp. 121^156. [2] D. Sedlak, D.M. Morre¨, D.J. Morre¨, Arch. Biochem. Biophys. 386 (2001) 106^116. [3] M. Kelker, C. Kim, P.-J. Chueh, R. Guimont, D.M. Morre¨, D.J. Morre¨, Biochemistry 40 (2001) 7351^7354. [4] S. Wang, R. Pogue, D.M. Morre¨, D.J. Morre¨, Biochim. Biophys. Acta 1539 (2001) 192^204. [5] P. Sun, D.J. Morre¨, D.M. Morre¨, Biochim. Biophys. Acta 1498 (2000) 52^63. [6] R. Pogue, D.M. Morre¨, D.J. Morre¨, Biochim. Biophys. Acta 14662 (2000) 1^8. [7] D.J. Morre¨, J. Lawler, S. Wang, T.W. Keenan, D.M. Morre¨, Biochim Biophys. Acta (2001) in press. [8] J. Braman, C. Papworth, A. Greener, Methods Mol. Biol. 57 (1996) 31^44. [9] A.O. Brightman, J. Wang, R.K. Miu, I.L. Sun, R. Barr, F.L. Crane, D.J. Morre¨, Biochim. Biophys. Acta 1105 (1992) 109^117. [10] D.J. Morre¨, M.L. Gomez-Rey, C. Schramke, O. Em, J. Lawler, J. Hobeck, D.M. Morre¨, Mol. Cell. Biochem. 200 (1999) 7^13. [11] P.K. Smith, R.I. Krohn, G.T. Hermanson, A.K. Mallia, F.D. Gartner, M.D. Provenzano, E.K. Fugimoto, N.M. Goeke, B.J. Olson, D.C. Klenk, Anal. Biochem. 150 (1985) 706.76. [12] T. Kishi, D.M. Morre¨, D.J. Morre¨, Biochim. Biophys. Acta 1412 (1999) 66^77. [13] S. Wang, H. Cutter, C. Martinez, N. Windberg, D.J. Morre¨, D.M. Morre¨, FASEB J. 12 (1998) A519. [14] D.J. Morre¨, P.-J. Chueh, D.M. Morre¨, Proc. Natl. Acad. Sci. USA 92 (1995) 1831^1835. [15] D.J. Morre¨, A. Bridge, L.-Y. Wu, D.M. Morre¨, Biochem. Pharmacol. 60 (2000) 937^946. [16] D.J. Morre¨, C. Kim, M. Paulik, D.M. Morre¨, W.P. Faulk, J. Bioenerg. Biomembr. 29 (1997) 269^280. [17] D.J. Morre¨, D.M. Morre¨, J. Bioenerg. Biomembr. 27 (1995) 137^144. [18] T. Yano, S.S. Chu, V.D. Sled, T. Ohnishi, T. Yagi, J. Biol. Chem. 272 (1997) 4201^4211. [19] C. Leblanc, C. Boyen, O. Richard, G. Bonnard, J.M. Grienenberger, B. Kloareg, J. Mol. Biol. 250 (1995) 484^ 495. [20] D.J. Morre¨, L.-Y. Wu, D.M. Morre¨, Biochim. Biophys. Acta 1240 (1995) 11^17. [21] D.J. Morre¨, L.-Y. Wu, D.M. Morre¨, Biochim. Biophys. Acta 1369 (1998) 185^192. [22] C. Grabau, J.E. Cronan, Nucleic Acids Res. 14 (1986) 5449. [23] S.O. Duke, Health Perspect. 87 (1990) 263. [24] D.J. Morre¨, E. Jacob, M. Sweeting, R. de Cabo, D.M. Morre¨, Biochim. Biophys. Acta 1325 (1997) 117^125.
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83 [25] M.E. Shinina¨, P. Carlini, F. Polticelli, F. Zappacosta, F. Bossa, L. Calabrese, Eur. J. Biochem. 237 (1996) 433^ 439. [26] R.B. Freedman, Cell 57 (1989) 1069^1072. [27] M. Russel, P. Model, J. Biol. Chem. 263 (1988) 9015^ 9019. [28] K. Ohnishi, Y. Niimura, M. Hidaka, H. Masaki, H. Suzuki, T. Uozumi, T. Nishino, J. Biol. Chem. 270 (1995) 5812^ 5817.
83
[29] T.-W. Gilberger, R.D. Walter, S. Mu«ller, J. Biol. Chem. 272 (1997) 29584^29589. [30] T.-W. Gilberger, B. Bergmann, R.D. Walter, S. Mu«ller, FEBS Lett. 425 (1998) 407^410. [31] D.J. Morre¨, J. Pletcher, X. Tang, P.-J. Chueh, L.-Y. Wu, D.M. Morre¨, Mol. Biol. Cell II (Suppl.) (2000) 449a. [32] D.J. Morre¨, D.M. Morre¨, Plant J. 16 (1998) 279^284. [33] D.J. Morre¨, D.M. Morre¨, C. Penel, H. Greppin, Int. J. Plant Sci. 160 (1999) 855^860.
BBAPRO 36514 24-1-02
A site-directed mutagenesis analysis of tNOX functional domains Pin-Ju Chueh a , Dorothy M. Morre¨ b , D. James Morre¨ a
a;
*
Department of Medicinal Chemistry and Molecular Pharmacology, Purdue University, West Lafayette, IN 47907, USA b Department of Foods and Nutrition, Purdue University, West Lafayette, IN 47907, USA Received 17 May 2001; received in revised form 11 September 2001; accepted 21 September 2001
Abstract Constitutive NADH oxidase proteins of the mammalian cell surface exhibit two different activities, oxidation of hydroquinones (or NADH) and protein disulfide-thiol interchange which alternate to yield oscillatory patterns with period lengths of 24 min. A drug-responsive tNOX (tumor-associated NADH oxidase) has a period length of about 22 min. The tNOX cDNA has been cloned and expressed. These two proteins are representative of cycling oxidase proteins of the plant and animal cell surface. In this report, we describe a series of eight amino acid replacements in tNOX which, when expressed in Escherichia coli, were analyzed for enzymatic activity, drug response and period length. Replacement sites selected include six cysteines that lie within the processed plasma membrane (34 kDa) form of the protein, and amino acids located in putative drug and adenine nucleotide (NADH) binding domains. The latter, plus two of the cysteine replacements, resulted in a loss of enzymatic activity. The recombinant tNOX with the modified drug binding site retained activity but the activity was no longer drug-responsive. The four remaining cysteine replacements were of interest in that both activity and drug response were retained but the period length for both NADH oxidation and protein disulfide-thiol interchange was increased from 22 min to 36 or 42 min. The findings confirm the correctness of the drug and adenine nucleotide binding motifs within the tNOX protein and imply a potential critical role of cysteine residues in determining the period length. ß 2002 Elsevier Science B.V. All rights reserved. Keywords: NADH oxidase; Protein disul¢de-thiol interchange; Tumor-associated NADH oxidase; Cycling oxidase; Capsaicin; Site-directed mutagenesis; Growth; Biological clock
1. Introduction A recently described family of cell surface proteins
Abbreviations: DTDP, dithiodipyridine ; NOX, plasma membrane-located NADH (hydroquinone) oxidase with protein disul¢de-thiol interchange activity; tNOX, tumor-associated NOX; H-NOX, recombinant truncated tNOX; CNOX, constitutive NOX; CLOX, cycling oxidase; capsaicin, 8-methyl-N-vanillyl-6noneamide; LY181984, N-(4-methylphenylsulfonyl-NP-(4-chlorophenyl) urea * Corresponding author. Fax: +1-765-494-4007. E-mail address: [email protected] (D.J. Morre¨).
that oxidize hydroquinones or NADH (NADH oxidase = NOX) is characterized uniquely by the capacity to carry out two distinct biochemical activities not simultaneously but in series [1]. The two activities, hydroquinone (NADH) oxidation and protein disul¢de-thiol interchange alternate to give rise to a period length of 24 min for the constitutive NOX activity (CNOX) [2]. The cDNA for a tumor-associated NOX protein (tNOX) has been cloned and expressed in COS cells and Escherichia coli (GenBank accession No. AF207881 [3]). tNOX activities also alternate to generate a pattern of oscillations but the period
0167-4838 / 02 / $ ^ see front matter ß 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 8 3 8 ( 0 1 ) 0 0 2 8 6 - 2
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
Fig. 1. NOX activities with assays based on disappearance of NADH or cleavage of DTDP measured at 340 nm for wild type recombinant truncated tNOX protein (ttNOX-1) expressed in E. coli. (A) Oxidation of NADH. (B) Cleavage of DTDP. In A, maxima in NADH oxidation labeled 1 and 2 recur at 21.5 (approx. 22) min intervals. Minor in£ections labeled 3, 4 and 5 also recur at 21.5 (approx. 22) min intervals. With protein disul¢de-thiol interchange measured in parallel in B, the major in£ections of A align with minor in£ections labeled 1 and 2 and the minor in£ections of A align with maxima labeled 3, 4 and 5 to generate the recurrent 2+3 patterns of activity alternations that appear to characterize all CLOX proteins. These data are representative of triplicate determinations and recapitulate the 21.5 (approx. 22) min period length of a processed tNOX form of the HeLa cell surface [4].
length of about 22 min is shorter than that of CNOX [4]. The NOX oscillations are not simple sine functions but, rather, represent complex 2+3 oscillatory patterns [5,6]. In the oxidizing portion of the cycle, hydroquinone or NADH oxidation is represented by two maxima followed by three minor oscillations. When protein disul¢de-thiol interchange is measured, the two maxima of NADH oxidation are the minor oscillations whereas the three minor oscillations of NADH oxidation now become major oscillations [5]. These relationships are illustrated in Fig. 1 for the wild type tNOX of HeLa cells.
75
Each of the components of the 3+2 pattern of oscillations repeats with a precise and temperaturecompensated (independent of temperature) period length of approx. 22 min [6]. NOX proteins at the cell surface or in solution function synchronously [6]. Autosynchrony of the protein is achieved in solution if two asynchronous preparations are mixed, for example [7]. In this paper we report the response of the activities of recombinant truncated tNOX (ttNOX-1) to amino acid replacements within putative functional domains as well as the response of period length to cysteine to alanine replacements. The truncated form of the protein was utilized as being representative of the processed 34 kDa form of the protein found at the cell surface of cancer cells and in the sera of cancer patients [1]. In those cysteine to alanine replacements where activity was retained, period length was modi¢ed in a manner suggesting that cysteines of the NOX molecule contribute to the determination of period length. 2. Methods 2.1. Mutagenic oligonucleotides and site-directed mutagenesis Eight sets of oligonucleotides listed in Table 1 were designed to replace amino acid residues potentially involved in tNOX activity by site-directed mutagenesis according to Braman et al. [8]. Cysteines C505, C510, C558, C569, C575, and C602 were replaced by alanines. A methionine of the putative drug binding site was replaced by alanine (M396A). A glycine in the potential adenine binding site was replaced by valine (G592V). High ¢delity thermostable Pfu DNA polymerase, low cycle number, and primers designed only to copy the parental strand in a linear fashion were used to minimize unwanted second site mutations. Double-stranded, super-coiled expression plasmid pET11ttNOX (40 ng) and mutagenic sense and antisense primers (100 ng) were combined in a 50 Wl reaction mixture containing deoxyribonucleotides, bu¡er, and Pfu DNA polymerase according to the manufacturer's protocol (Stratagene, La Jolla, CA, USA). The cycling parameters were 95³C for 30 s,
BBAPRO 36514 24-1-02
76
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
Table 1 The mutagenic oligonucleotides for site-directed mutagenesisa Amino acid replacement
Oligonucleotides
C505A
5P-GAAAAGGAAAGCGCCGCTTCTAGGCTGTGTGCC-3P (forward) 5P-GGCACACAGTCTAGAAGCGGCGCTTTCCTTTTC-3P (reverse) 5P-GCTTCTAGGCTGGCCGCCTCAAACCAGGATAGCG-3P (forward) 5P-CGCTATCCTGGTTTGAGGCGGCCAGCCTAGAAGC-3P (reverse) 5P-GCAAGCATTGAATACATCGCTTCCTACTTGCACCGTCTTG-3P (forward) 5P-CAAGACGGTGCAAGTAGGAAGCGATGTATTCAATGCTTGC-3P (reverse) 5P-CGTCTTGATAATAAGATCGCCACCAGCGATGTGGAGTG-3P (forward) 5P-CACTCCACATCGCTGGTGGCGATCTTATTATCAAGACG-3P (reverse) 5P-CCAGCGATGTGGAGGCCCTCATGGGTAGACTCC-3P (forward) 5P-GGAGTCTACCCATGAGGGCCTCCACATCGCTGG-3P (reverse) 5P-GAAAAGAGATGGAAATTCGCTGGCTTCGAGGGCTTGAAG-3P (forward) 5P-CTTCAAGCCCTCGAAGCCAGCGAATTTCCATCTCTTTTC-3P (reverse) 5P-GTCTGATGATGAAATAGAAGAAGCGACAGAAACAAAAGAAACTGAGG-3P (forward) 5P-CCTCAGTTTCTTTTGTTTCTGTCGCTTCTTCTATTTCATCATCAGAC-3P (reverse) 5P-CAGGAAATGACTGGAGTTGTGGCCAGCCTGGAAAAGAG-3P (forward) 5P-CTCTTTTCCAGGCTGGCCACAACTCCAGTCATTTCCTG-3P (reverse)
C510A C558A C569A C575A C602A M396A G592V a
Nucleotide replacements are in bold.
55³C for 1 min, and 68³C for 12.8 min, 16 cycles. The linear ampli¢cation product was treated with endonuclease DpnI (10 units/Wl) for 1 h to eliminate the parental template. Subsequently, 4 Wl of this reaction mixture containing the double-nicked mutated plasmid were used for the transformation of supercompetent E. coli XL-1 Blue cells (Stratagene). DNA sequencing was utilized to con¢rm that the correct replacements were obtained in all mutants. 2.2. Protein expression in E. coli The replacements were expressed in bacteria and puri¢ed as recombinant truncated tNOX forms. The recombinant tNOX protein designated ttNOX-1, comprised of amino acid residues 220^610 of tNOX protein (391 amino acids), retained full functional activity. NdeI and BamHI restriction sites were incorporated to accommodate subcloning into the protein expression vector pET-11b. The ampli¢cation was performed as follows: 94³C, 1 min 30 s/55³C, 1 min 30 s/72³C, 1 min 30 s/29 cycles/72³C, 5 min. The open reading frame of truncated tNOX cDNA was ampli¢ed by PCR. The PCR product was digested with NdeI plus BamHI and ligated into the pET-11b vector, which had been digested by NdeI and BamHI and dephosphorylated by calf intestine alkaline phosphatase. The resultant vector was transformed into E. coli (XL-1Blue). Plasmid pET11-
ttNOX was puri¢ed and transformed into E. coli (BL21(DE3)). The transformed bacteria were grown on ampicillin plates. Single colonies were used to inoculate LB medium containing 100 Wg/ml ampicillin. Cells were grown for 16^20 h at 25^37³C. Since plasmid pET11-ttNOX is a leaky plasmid, IPTG was not required to induce expression. The cells were harvested by centrifugation (6000 rpm 5 min, Sorvall SS-34 rotor). The cell pellet was then washed in 20 mM Tris^HCl, pH 8.0 and ¢nally resuspended in 20 mM Tris^HCl, pH 8.0 containing 10 mM DTT and protease inhibitors (1 mM PMSF, 1 mM benzamidine and 1 mM 6-aminocaproic acid). The cells were then lysed with a French press (20 000 psi, three passages) and the cell lysates were centrifuged at 9000^11 000 rpm for 20 min (Sorval SS-34 rotor). The supernatant was collected and saved as the source of solubilized ttNOX-1 protein. The ttNOX-1 was then precipitated by ammonium sulfate 20% of saturation. Once precipitated, ttNOX-1 was not soluble in 20 mM Tris^HCl, pH 8.0. Refolding of ttNOX-1 was according to the procedure provided by Novagen (Madison, WI, USA). Brie£y, the pellet was washed with 1% Triton X-100 and the recombinant ttNOX proteins were dissolved in 50 mM CAPS, pH 11, containing 0.3% N-laurylsarcosine (sodium salt). Proteins were then dialyzed in two changes of dialysis bu¡er (20 mM Tris^HCl, pH 8.5 containing 0.5 mM cysteamine and 0.05 mM
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
cystamine) over a period of 14 h followed by dialysis in two changes of equilibration bu¡er (20 mM Tris^ HCl, pH 7.5, 0.1 mM DTT). After dialysis, ttNOX-1 protein was stored at 320³C until further use. Expression of ttNOX-1 was analyzed by SDS^PAGE with silver staining and Western blotting. Western blotting used anti-tNOX monoclonal and polyclonal antibodies. The antibody binding was detected using an alkaline phosphate conjugated to anti-mouse IG antibody or anti-rabbit IG antibody. 2.3. Spectrophotometric assay of tNOX activities NADH oxidase activity was determined as the disappearance of NADH measured at 340 nm in a reaction mixture containing 25 mM Tris^MES bu¡er (pH 7.2), 1 mM KCN to inhibit low levels of mitochondrial oxidase activity, and 150 WM NADH at 37³C with temperature control ( þ 0.5³C) and stirring [9]. Activity was measured using paired Hitachi U3210 spectrophotometers. Assays were initiated by addition of NADH. Assays were for 1 min and were repeated on the same sample every 1.5 min for the times indicated. A millimolar extinction coe¤cient of 6.22 was used to determine speci¢c activity. Protein disul¢de-thiol interchange activity associated with the tNOX protein was measured using a dipyridyldithio substrate, dithiodipyridine (DTDP) [10]. The assay was in 50 mM Tris^MES bu¡er (pH 7.0). The assay was initiated by the addition of 0.5 Wmole DTDP in 5 Wl of DMSO. The reaction was monitored from the increase in absorbance at 340 nm using the pyridinethionine product. The speci¢c activity was calculated using a millimolar absorption coe¤cient of 6.2 cm31 . Proteins were estimated by the bicinchoninic acid method [11] with bovine serum albumin as standard. 3. Results Expression of recombinant truncated tNOX (ttNOX-1) was veri¢ed by Western blot analysis for all eight amino acid replacements compared to wild type. Assays used bacterial extracts with approximately equivalent amounts of recombinant protein as determined from intensities of the ttNOX-1 protein band of the Western blots.
77
The oscillation of NADH oxidase activity of recombinant ttNOX-1 exhibited maxima at approx. 22 min intervals for both NADH oxidation (Fig. 1A) and DTDP cleavage (Fig. 1B). The precise length over three periods was 64.5 min or an average single period length of 21.5 min. As is customary, the even possibility was used as the convention for rounding o¡, hence a period length of 22 min. The oscillations were not simple sine waves but varied with a complex, recurring pattern (Fig. 1). When both enzymatic activities, measured in parallel using two di¡erent spectrophotometers on equal portions of the same preparation, were analyzed in greater detail, a total of ¢ve oscillatory components were resolved within each 22 min period (Fig. 1). With NADH as substrate for the oxidative activity (Fig. 1A), the oscillatory components labeled 1 and 2 were most evident. With DTDP as substrate for the protein disul¢de-thiol interchange activity (Fig. 1B), the oscillatory components labeled 3, 4 and 5 were most evident. As reported previously for recombinant tNOX protein [1], the mean times of the dominant oscillatory components of the two activities were 12 min out of phase with each other. In Fig. 1, with NADH as substrate, the two main oscillations were at 18, 40, 61 and 81 min and 20, 42, 60 and 85 min. In addition, there were three minor oscillations labeled 3, 4 and 5 to complete each 22 min period. With DTDP as substrate, the major oscillations were at 6, 26, 48 and 69 min, at 9, 31, 51 and 72 min and at 12, 34, 54 and 74 min. These major oscillations with DTDP as substrate corresponded exactly to the minor oscillations of the oxidative activity with NADH as substrate. The minor oscillations with the DTDP substrate were at 18, 40, 61 and 81 min and at 20, 42, 64 and 85 min which corresponded to the times of the major oscillations in NADH oxidation. Thus, the two major oscillations of NADH oxidation aligned with the two minor oscillations of DTDP cleavage. Similarly, the three major oscillations of DTDP cleavage aligned with the three minor oscillations of NADH oxidation. This repeating 3+2 pattern of activity occurred reproducibly and has been validated by time series decomposition analyses [5]. The M396A replacement in the putative drug binding site (see Section 4) did not eliminate either of the two enzymatic activities associated with the
BBAPRO 36514 24-1-02
78
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
Fig. 2. Assay as in Fig. 1 of the M396A replacement within the putative drug binding site. An activity pattern similar but not identical to that of Fig. 1 was observed but capsaicin no longer inhibited. (A) NADH oxidation. (B) DTDP cleavage.
ttNOX-1 (Fig. 2). Additionally, the period length remained unchanged at approx. 22 min (Table 2). However, neither activity was any longer inhibited by capsaicin. Even after 1 h or more, the activity in the presence of capsaicin was equivalent to that in the absence of capsaicin (Fig. 2). The period length of 21.5 (approx. 22) min was reproduced in each of three separate experiments. The M396A replacement did appear to alter the pattern of oscillations, however. In three experiments, component 2 of NADH oxidation was reduced (Fig. 2). This reduction in NADH oxidative activity was compensated in part by an overall
broadening of the maxima labeled 1 and by a possible enhancement of peak 2 with DTDP cleavage. The G592V replacement in the putative adenine binding domain (see Section 4) resulted in a loss of enzymatic activity (Fig. 3). With this and the other replacements lacking activity, capsaicin was without e¡ect as well. When the ttNOX proteins with six di¡erent cysteine to alanine replacements were expressed in E. coli, those containing C505A or C569A no longer exhibited NADH oxidation or cleavage of DTDP (Fig. 3). The four other cysteine replacements retained both NADH oxidase and protein disul¢dethiol interchange activities but the period lengths were altered (Table 1). For the C510A replacement (Fig. 4) as well as for the C575A and the C602A replacements (Fig. 5), the period lengths were increased to 36 min. For the C558A replacement, the period length was 42 min (Fig. 5A). The increased period lengths were observed with both the oxidase activity with NADH as substrate and for the protein disul¢de-thiol interchange activity with DTDP as substrate. The expressed proteins carrying the cysteine replacements yielding period lengths of 36 or 42 min recapitulated the complex 2+3 pattern of activity oscillations. With the 36 min periods, the oxidative activity exhibited two major activity maxima separated by about 12 min (1 and 2) and three minor in£ections (3, 4 and 5). The protein disul¢de-thiol interchange activity, on the other hand, exhibited two minor in£ections and three maxima. The two maxima of the NADH oxidation and the two minor in£ections of the protein disul¢de-thiol interchange
Table 2 E¡ects of site-directed mutagenesis of ttNOX on NADH oxidase activity, period length and inhibition by capsaicin Amino acid replacementa
Enzymatic activity (nmoles/min/mg protein)
Period length (min)
Inhibition by 1 WM capsaicin
None C505A C510A C558A C569A C575A C602A M396A G592V
0.9 þ 0.1 None 0.85 þ 0.05 0.9 þ 0.15 None 0.8 þ 0.05 0.8 þ 0.05 0.85 þ 0.1 None
22
Complete
36 42
Complete Complete
36 36 22
Complete Complete None
a
Resequencing con¢rmed the expected DNA sequences for each of the amino acid replacements indicated.
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
79
coincided when assayed in parallel. Similarly the three minor in£ections of the NADH oxidation coincided with the three maxima of the protein disul¢de-thiol interchange when assayed in parallel. Both enzymatic activities of each of the four cysteine replacements with enzymatic activity were also inhibited completely by capsaicin (Figs. 4 and 5). For each of the four replacements, assays were continued for 80 min in separate experiments (not shown). After approx. 20 min, both NADH oxidation and DTDP cleavage activities reached baseline in the presence of capsaicin and remained at baseline for the duration of the experiment.
Fig. 4. NOX activities assayed as in Fig. 1 for the C510A replacement. (A) Oxidation of NADH. (B) Cleavage of DTDP. In A, maxima labeled 1 and 2 spaced about 12 min apart recur at 36 min intervals. Also recurrent at 36 min intervals were minor oscillations labeled 3, 4 and 5. With the protein disul¢dethiol interchange measured in parallel (B), the minor in£ections of A aligned with maxima labeled 3, 4 and 5 to generate the recurrent 2+3 pattern of activity alternations seen in Fig. 1 but with the recurrent maxima separated at intervals of 36 min rather than intervals of 22 min. Additionally, the ¢rst of the three maxima labeled 3 determined from cleavage of DTDP appears between maxima labeled 1 and 2 instead of after maxima labeled 1 and 2 as in Fig. 1. The values shown are representative of duplicate determinations for each activity.
Fig. 3. Assay as in Fig. 1 of the C505A, C569A and G592V replacements where functional activities were lost. (A) NADH oxidation. (B) Cleavage of DTDP. No signi¢cant regular periodicities above background were noted. Capsaicin was without e¡ect (not shown). Data are representative of duplicate (B) or triplicate (A) determinations for each replacement.
While each of the three cysteine replacements with period lengths of 36 min exhibited overall close similarities, i.e. the recurrent 2+3 pattern of activity oscillations within a 36 min period, unique features were revealed for each which appeared reproducible. Compared to the C510A replacement (Fig. 4), the C575A replacement exhibited more pronounced activity components labeled 4 and 5 with the DTDP substrate compared to activity components labeled 3 (Fig. 5). The C602A replacement exhibited much more closely spaced components designated as 1 and 3 compared to the other cysteine replacements (Fig. 5).
BBAPRO 36514 24-1-02
80
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
4. Discussion Previous studies have identi¢ed an unusual NADH oxidase activity at the external cell surface of the
plasma membrane of both plant and animal cells [1]. The physiological function of the oxidative portion of the NOX cycle appears to be that of a hydroquinone oxidase [12] even though NADH oxidation provides a convenient measure of the enzymatic activity [9]. These proteins are unique not only from their plasma membrane location but also from their oscillating activities that provide for potential roles as time-keeping proteins [13] and a relationship between the oscillatory enzymatic activity and the enlargement phase of cell growth [1,6]. The two di¡erent enzymatic activities of the NOX proteins, hydroquinone oxidation and protein disul¢de-thiol interchange, alternate. With the normal constitutive (CNOX) protein, an approx. 24 min period is generated as a result [1,5,6]. With tNOX forms present together with CNOX at the surfaces of cancer cells, three period lengths of 24 þ 0.3, 22.5 þ 0.6 min and 21.5 þ 0.6 min were observed for HeLa (human cervical carcinoma) cells [4]. The latter period length corresponds exactly with the 21.5 min period length of the recombinant ttNOX reported here. Both tNOX and CNOX appear to be members of a family of NADH oxidases with activities that cycle with regular period lengths independent of temperature (CLOX family of cell surface cycling oxidases). While several NOX forms belonging to this CLOX family may exist, the drug-responsive tNOX is the ¢rst to be cloned (GenBank accession No. AF207881). tNOX di¡ers from the constitutive CNOX form primarily in its sensitivity to capsaicin [14], (3)-epigallocatechin 3-gallate (EGCg) [15], several anticancer drugs including adriamycin [16] and to thiol reagents [17]. The response of tNOX activity to capsaicin was used to guide puri¢cation of the processed tNOX protein from sera of cancer patients
6
Fig. 5. NOX activities assayed as in Fig. 1 for the C558A, C575A and C602A replacements all of which displayed patterns similar to those of the C510A replacement of Fig. 4. (A) Oxidation of NADH. (B) Cleavage of DTDP. Maxima generating the 2 (NADH)+3 (DTDP) pattern recurred at intervals of about 36 min for C575A and C602A replacements and at intervals of 42 min for the C558A replacement. The experiments shown are representative of duplicate or triplicate determinations for each activity.
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
and as the basis for the monoclonal antibody selection used to obtain the cloned cDNA [3]. A number of motifs within the tNOX sequence account for the range of its functional activities [1]. A NADH binding site was indicated earlier from binding of photoa¤nity-labeled [32 P]NAD(H). A putative adenine nucleotide binding sequence motif (GX-G-X-X-G) with downstream remote acidic amino acid residues D or E [18] was represented by T589-GV-G-A-S-L and E605 near the C-terminus and compared closely with the sequence T-G-V-G-A-G-V-G from mitochondrial ATP synthase subunit 9 from Chondrus crispus [19]. The NOX protein binds the antitumor sulfonylurea LY181984 [20]. Activity is inhibited or stimulated by LY181984 depending on the redox environment of the protein [21]. The protein oxidized reduced coenzyme Q [12] and substances such as capsaicin [14], EGCg [15] and adriamycin [16], which inhibit the activity, are considered to occupy quinone sites. Ubiquinone competes with both LY181984 binding and inhibition of enzymatic activities. By analogy with several quinone binding proteins of the photosystem II complex of chloroplasts, a site containing a methionine-histidine pair has been equated with the quinone binding site of pyruvate oxidase [22]. Such sites are targets for all known urea and sulfonylurea herbicide inhibitors of photosystem II [23]. Apparently arginine can substitute for the critical histidine. A consensus sequence for the amino acids surrounding the charged residues critical to sulfonylurea and quinone binding site was determined to be A-M-H-G or a closely related sequence based on the above considerations. The putative quinone binding site of the D1 protein of a cyanobacterium, Synechococcus, contains the sequence E-T-MR-E. A sequence similar to E-T-M-R-E is present in the NADH-ubiquinone oxidoreductase of chloroplasts. We found a sequence E394-E-M-T-E as a potential quinone site having neither H nor R in the 4th position but still with considerable similarity to other putative quinone and/or sulfonylurea binding sites. The correct identi¢cation of E394-E-M-T-E as the drug and/or quinone binding site of tNOX is supported by the present ¢ndings. The recombinant protein with the M396A replacement retained NADH oxidase activity but the activity was no longer inhibited by capsaicin (Table 1, Fig. 2).
81
Initially, the protein disul¢de-thiol interchange activity of the tNOX protein was based on the restoration of activity to reduced, denatured and oxidized (scrambled) yeast RNase through reduction and refolding under non-denaturing conditions and reoxidation to form a correct secondary structure stabilized by internal disul¢de bonds [24]. The activity was similar to that catalyzed by protein disul¢de isomerases of the endoplasmic reticulum [25] yet a different protein motif appeared to be involved since the activity was not altered by the presence of two di¡erent antisera to protein disul¢de isomerases [24]. One of the antisera was a mouse monoclonal (SPA891) (StressGen Biotechnologies) to protein disul¢de isomerase (PDI) from bovine liver (cross-reactive with PDI from human, monkey, rat, mouse and hamster cell lines). The other was a peptide antibody directed to the C-X-X-C motif common to most, if not all, members of the protein disul¢de isomerase family of proteins [26]. This motif is also present in thioredoxin reductase and related proteins where it appears to catalyze the transfer of electrons in conjunction with bound £avin [27^30]. A C-X-X-C motif is absent in tNOX and tNOX lacks thioredoxin reductase activity. tNOX appears not to contain bound £avin nor is its activity dependent on added £avin (FAD or FMN). While the two C-X-X-X-X-C motifs characteristic of £avoproteins are missing from tNOX, the redox active disul¢de of thioredoxin reductase from the malaria parasite Plasmodium falciparum contains a motif C88-X-X-X-X-C93 [29] similar to one found in tNOX. Together with a downstream His509, the motif was shown to be a putative proton donor/acceptor. Either the C88A or the C93A replacement resulted in complete loss of enzymatic activity [29]. A C535-X-X-X-X-C540 motif in the same protein was shown to be involved in substrate coordination and/or electron transfer [30]. A C535A replacement did result in diminution of enzymatic activity but the C540A replacement did not [30]. Thus, either or both of the two comparable motifs present in tNOX, C505-X-X-X-X-C510 or C569-X-X-X-X-X-C575, alone or together with downstream histidines provide potential active sites for protein disul¢de-thiol interchange. With tNOX, the C505A and C569A replacements lost activity as with the C535A replacement above for the P. falciparum protein but the C510A
BBAPRO 36514 24-1-02
82
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83
and C575A did not as with the above C540A replacement for the P. falciparum protein. For the four cysteine replacements with enzymatic activity, NADH oxidation and protein disul¢de-thiol interchange were a¡ected in parallel. E. coli cells transfected with ttNOX-1 cDNA with C510A, C558A, C575A or C602A replacement generated expressed ttNXO-1 proteins where the 2+3 pattern characteristic of the cell surface NADH oxidase was still evident. The change from a 22 min period to a 36 or 42 min period was re£ected in the entire 2+3 NOX activity pattern and not just in the principal maxima. Both of the two NADH oxidase maxima were a¡ected equally and were spaced approx. 12 min apart in the mutants compared to about 6 min apart for the unmodi¢ed truncated tNOX. Also oscillating with 36 or 42 min periods were the three maxima of the 2+3 NOX activity pattern associated with the protein disul¢de-thiol interchange activity that alternates with NADH (hydroquinone) oxidation. The precision and relative immutability of the pattern of NOX activity oscillations, together with a lack of dependence of period length on temperature (temperature compensation) [7,32] and entrainability by light [33], make the NOX proteins candidates as cellular time-keeping proteins. Expanding on preliminary work [31], experiments to transfect appropriate cell lines with the cDNAs carrying cysteine replacements having a period length of 36 or 42 min to determine e¡ects on higher order periodic responses including circadian metabolic oscillations normally with period lengths of 24 h are in progress. Acknowledgements We thank Lian-Ying Wu for assistance with cell culture, Andrew Brightman, Xiaoyu Tang and Zensui Tian for access to unpublished data, Dongjiao Zhao and Brett Carnahan for technical assistance and Peggy Runck for manuscript preparation. Work supported in part by NASA NAG2-1334, NIH 5RO1CA75461-03 and NIH 1P50AT00477-01 (Botanical Center for Age-Related Diseases).
References [1] D.J. Morre¨, in: H. Asard, A. Be¨rczi, R.J. Caubergs (Eds.), Plasma Membrane Redox Systems and their Role in Biological Stress and Disease, Kluwer Academic Publishers, Dordrecht, 1998, pp. 121^156. [2] D. Sedlak, D.M. Morre¨, D.J. Morre¨, Arch. Biochem. Biophys. 386 (2001) 106^116. [3] M. Kelker, C. Kim, P.-J. Chueh, R. Guimont, D.M. Morre¨, D.J. Morre¨, Biochemistry 40 (2001) 7351^7354. [4] S. Wang, R. Pogue, D.M. Morre¨, D.J. Morre¨, Biochim. Biophys. Acta 1539 (2001) 192^204. [5] P. Sun, D.J. Morre¨, D.M. Morre¨, Biochim. Biophys. Acta 1498 (2000) 52^63. [6] R. Pogue, D.M. Morre¨, D.J. Morre¨, Biochim. Biophys. Acta 14662 (2000) 1^8. [7] D.J. Morre¨, J. Lawler, S. Wang, T.W. Keenan, D.M. Morre¨, Biochim Biophys. Acta (2001) in press. [8] J. Braman, C. Papworth, A. Greener, Methods Mol. Biol. 57 (1996) 31^44. [9] A.O. Brightman, J. Wang, R.K. Miu, I.L. Sun, R. Barr, F.L. Crane, D.J. Morre¨, Biochim. Biophys. Acta 1105 (1992) 109^117. [10] D.J. Morre¨, M.L. Gomez-Rey, C. Schramke, O. Em, J. Lawler, J. Hobeck, D.M. Morre¨, Mol. Cell. Biochem. 200 (1999) 7^13. [11] P.K. Smith, R.I. Krohn, G.T. Hermanson, A.K. Mallia, F.D. Gartner, M.D. Provenzano, E.K. Fugimoto, N.M. Goeke, B.J. Olson, D.C. Klenk, Anal. Biochem. 150 (1985) 706.76. [12] T. Kishi, D.M. Morre¨, D.J. Morre¨, Biochim. Biophys. Acta 1412 (1999) 66^77. [13] S. Wang, H. Cutter, C. Martinez, N. Windberg, D.J. Morre¨, D.M. Morre¨, FASEB J. 12 (1998) A519. [14] D.J. Morre¨, P.-J. Chueh, D.M. Morre¨, Proc. Natl. Acad. Sci. USA 92 (1995) 1831^1835. [15] D.J. Morre¨, A. Bridge, L.-Y. Wu, D.M. Morre¨, Biochem. Pharmacol. 60 (2000) 937^946. [16] D.J. Morre¨, C. Kim, M. Paulik, D.M. Morre¨, W.P. Faulk, J. Bioenerg. Biomembr. 29 (1997) 269^280. [17] D.J. Morre¨, D.M. Morre¨, J. Bioenerg. Biomembr. 27 (1995) 137^144. [18] T. Yano, S.S. Chu, V.D. Sled, T. Ohnishi, T. Yagi, J. Biol. Chem. 272 (1997) 4201^4211. [19] C. Leblanc, C. Boyen, O. Richard, G. Bonnard, J.M. Grienenberger, B. Kloareg, J. Mol. Biol. 250 (1995) 484^ 495. [20] D.J. Morre¨, L.-Y. Wu, D.M. Morre¨, Biochim. Biophys. Acta 1240 (1995) 11^17. [21] D.J. Morre¨, L.-Y. Wu, D.M. Morre¨, Biochim. Biophys. Acta 1369 (1998) 185^192. [22] C. Grabau, J.E. Cronan, Nucleic Acids Res. 14 (1986) 5449. [23] S.O. Duke, Health Perspect. 87 (1990) 263. [24] D.J. Morre¨, E. Jacob, M. Sweeting, R. de Cabo, D.M. Morre¨, Biochim. Biophys. Acta 1325 (1997) 117^125.
BBAPRO 36514 24-1-02
P.-J. Chueh et al. / Biochimica et Biophysica Acta 1594 (2002) 74^83 [25] M.E. Shinina¨, P. Carlini, F. Polticelli, F. Zappacosta, F. Bossa, L. Calabrese, Eur. J. Biochem. 237 (1996) 433^ 439. [26] R.B. Freedman, Cell 57 (1989) 1069^1072. [27] M. Russel, P. Model, J. Biol. Chem. 263 (1988) 9015^ 9019. [28] K. Ohnishi, Y. Niimura, M. Hidaka, H. Masaki, H. Suzuki, T. Uozumi, T. Nishino, J. Biol. Chem. 270 (1995) 5812^ 5817.
83
[29] T.-W. Gilberger, R.D. Walter, S. Mu«ller, J. Biol. Chem. 272 (1997) 29584^29589. [30] T.-W. Gilberger, B. Bergmann, R.D. Walter, S. Mu«ller, FEBS Lett. 425 (1998) 407^410. [31] D.J. Morre¨, J. Pletcher, X. Tang, P.-J. Chueh, L.-Y. Wu, D.M. Morre¨, Mol. Biol. Cell II (Suppl.) (2000) 449a. [32] D.J. Morre¨, D.M. Morre¨, Plant J. 16 (1998) 279^284. [33] D.J. Morre¨, D.M. Morre¨, C. Penel, H. Greppin, Int. J. Plant Sci. 160 (1999) 855^860.
BBAPRO 36514 24-1-02