The NADP+-binding site of ferredoxin-NADP+ reductase. Sequence ... ing site of the reductase. With the ..... However, in our hands it appeared to make Edman.
Eur. J. Biochem. 146,295-299 (1985) 0FEBS 1985
The NADP+-bindingsite of ferredoxin-NADP+reductase Sequence of the peptide containing the essential lysine residue Dante CIDARIA, Pier Antonio BIONDI, Giuliana ZANETTI, and Severino RONCHI Department of General Physiology and Biochemistry and Institute of Veterinary Physiology and Biochemistry, University of Milano (Received May 30/0ctober 3, 1984) - EJB 840569
The flavoprotein ferredoxin-NADP' reductase is inactivated and loses its ability to bind NADP' during covalent modification of a lysine by 5-dimethylaminonaphthalene-I-sulfonylchloride (dansyl chloride) [Zanetti, G. (1976) Biochim. Biophys. Acta 445, 14-24]. The substrate NADP' gives almost complete protection against inactivation and modification. These observations are extended in this report by the characterization of an octapeptide containing the dansyl-lysine which was isolated by high-performance liquid chromatography from tryptic digests of protein modified with radiolabeled reagent. The amount of this peptide was severely reduced in protein modified in the presence of NADP'. The sequence of the dansyl-peptide, only partially obtained by Edman degradation, was completed by analysis of the fragments resulting from thermolysin digestion of the purified tryptic dansyl-peptide. Thus, the octapeptide containing the essential lysine residue has the following sequence : H2N-Ser-Val-Ser-Leu-Cys-Val-Lys-Arg-COOH. A comparison with corresponding sequences of other known NADP +-dependent dehydrogenases is attempted. Ferredoxin-NADP' reductase is the terminal electroncarrier of the photosynthetic transport-chain. The enzyme from spinach leaves has been resolved to 0.37-nm resolution by X-ray crystallography [I], whereas its sequence is still under investigation [2]. Thus, chemical modification studies are the only means to assign specific function to specific residues in this protein. Accordingly, the NADP+-binding domain of the reductase has been exaustively explored by this technique. It has been shown that essential lysine [3], arginine [4], carboxyl [5] and histidine [6] residues are located in the NADP+-binding site of the reductase. With the aim to further characterize this region of the protein, we decided to isolate the peptide containing the essential lysine residue [3]. Chemical modification with dansyl chloride, yielding a covalent fluorescent group attached to the residue, facilitates the recognition and the isolation of such a peptide. Furthermore, the selectivity of this labeling was well established [3], since a differential incorporation of only 1 - 1.5 mol dansyl/mol FAD was found between free and NADP+-protected enzyme. Here we report the sequence of an octapeptide thus isolated. A comparison is made with other NADP+-dependent dehydrogenases.
specific activity in the range 10000- 12000 dpm/nmol. Ferredoxin-NADP reductase was prepared and assayed as previously described [7]. Acetonitrile was Lichrosolv grade from Merck; all other chemicals were analytical reagent grade. +
Modification of ferredoxin-NADP+ reductase by dansyl chloride
The enzyme was inactivated by dansyl chloride essentially as described [3]. 50-100 nmol of protein in 0.1 M Tris/HCl, pH 8.8, were adjusted to a final protein concentration of 35 pM. [Me-14C]Dansyl chloride (in acetone) was added in a 13-fold molar excess over enzyme-bound FAD and the reaction was allowed to proceed for about 100 min at 0°C in the dark to reach 85 - 90% inactivation. Excess reagent and its hydrolysis product were removed by gel filtration on a Sephadex G-25 column equilibrated in 50 mM sodium phosphate, pH 8. The protein solution was concentrated by a Millipore Immersible CX-10 unit. The amount of bound dansyl was determined by radioactivity measurements in a Packard-Prias liquid scintillation counter. The protein was denatured in 6 M guanidine/HCl and reduced by addition of a 30-fold molar excess dithiothreitol under nitrogen at 30 "C MATERIALS AND METHODS for 2 h. Then alkylation of the thiol groups was performed by incubation for 40 min with iodoacetate at a twofold molar NADP', Tris, trypsin, thermolysin and dansyl chloride excess over the total -SH concentration. The protein was were obtained from Sigma; (Me-'4C)dansyl chloride was from Amersham. An aliquot of this solution was taken to dialyzed extensively against 50 mM sodium phosphate, pH 8, containing 2 M urea and digested for 6 h at 37°C with dryness and redissolved in unlabeled dansyl chloride to a diphenylcarbamoylchloride-trypsin (5%, w/w). A control sample was prepared in the same way, except that during Abbreviations. Dansyl chloride, 5-dimethylaminonaphtalene-l- inactivation by dansyl chloride 2 mM NADP+ was present. sulfonyl chloride; HPLC, high-performance liquid chromatography; TLC, thin-layer Chromatography. Enzymes. Ferredoxin-NADP' reductase (EC 1.18.1.2); trypsin (EC 3.4.21.4); thermolysin (EC 3.4.24.4); carboxypeptidase Y (EC 3.4.17. -); dihydrofolate reductase (EC 1.5.1.3); glutathione reductase [NAD(P)H] (EC 1.6.4.2).
Peptide isolation and characterization Analytical peptide mapping and peptide isolation were performed by reverse phase HPLC. 2 nmol of protein digest
296 were analyzed on a Waters Associates pBondapak C18 column (0.39 x 30 cm) at room temperature using a linear gradient from 0.1% CF3COOH (solvent A) to 50% mixture with 0.075% CF,COOH in acetonitrile (solvent B) over 64 min at a flow rate of 1 ml/min. Peptides were detected by their absorbance at 220 nm; the fluorescence of the dansyl derivatives was monitored by a Jasco P-I 10 spectrofluorometer (excitation 334 nm, emission 506 nm). Radioactivitiy was measured by liquid scintillation counting on the collected fractions. Isolation of the dansyl-peptide was also performed on a semipreparative C18 column (0.78 x 30 cm) using the same chromatographic conditions mentioned above. Thermolysin digestion of the purified dansyl-peptide was performed as follows: 30 nmol peptide in 1% NH4HC03were incubated with thermolysin (2.5%, w/w) at 37°C for 6 h. The reaction was terminated by addition of 1% CF3COOH. The peptides thus obtained, were separated by reverse phase HPLC and analyzed as described above.
Acid precipitation of the modified protein After the incubation with dansyl chloride and gel filtration, the enzyme was precipitated by 10% CC13COOH and recovered by filtering the solution on a nitrocellulose filter (Millipore HAWP, 0.45 pM). The filter, after extensive washing, was dissolved in 1ml 2(methoxy)-ethanol; following the addition of the scintillation fluid (Packard Instagel), the protein-bound radioactivity was counted. Amino acid analysis was performed on a Carlo Erba amino acid analyzer. Edman degradation was made according to the manual micromethod recently introduced by Tarr [S]. RESULTS
Modijication of ferredoxin-NADP' reductase with dansyl chloride The enzyme was inactivated at pH 8.8 (O'C) with a 13fold molar excess of [Me-'4C]dansyl chloride in the presence and in the absence of 2 mM NADP'. Inactivation reached 90% in 80- 100 min in the unprotected samples whereas only 3 - 10% activity was lost in the presence of NADP+. In two typical experiments, 7.5 x lo6 cpm and 4.1 x lo6 cpm were added to 68.8 nmol and 34.4 nmol of flavoprotein, respectively. After removal of excess reagent by gel filtration, in the protein peak 1 x lo6 cpm and 0.55 x lo6 cpm for 63.7 nmol and 32 nmol flavoprotein were found, respectively. In the presence of NADP', the radioactivity used was 4.6 x lo6 cpm (34.4 nmol of flavoprotein) and that bound to protein was 0.17 x lo6 cpm (32 nmol of flavoprotein). These data indicated that 1.9- 2.0 mol dansyl residues had been incorporated per mol FAD in the unprotected samples; in the presence of NADP' the incorporation was 0.5 mol dansyl residues/ rnol FAD. Theses results are consistent with those previously reported [3]. HPLC tryptic peptide maps of the modified protein are shown in Fig. 1. In Fig. 1A the elution pattern of the peptides of the unprotected enzyme as monitored at 220 nm and for the fluorescence of the dansyl residues is reported; Fig. 2B is a histogram of the radioactivity associated with the collected fractions both of the free and the NADP'complexed enzyme. Only one major radiolabeled peak was observed in the tryptic map of the enzyme modified in the absence of NADPC. In the protected sample, only minor radioactive peaks were found: the amount of the corresponding peptide was consistent with the small quantity of inactivat-
ed enzyme. The peptide relative to the higher peak of radioactivity contained 40% of the radioactivity eluted; since the enzyme inactivated to 90% had incorporated 2.0 mol dansyl/ mol FAD, if modification of a single specific residue/FAD results in enzyme inactivation, the main labeled peak should have contained 45% of the total radioactivity. Thus, the value found represents the 89% of the theoretical value.
Isolation and characterization of dansyl-peptide The major radioactive peptide was isolated from 60100 nmol of the tryptic digest and further purified by isocratic reverse phase HPLC on a pBondapak C18 column equilibrated in 0.1% CF3COOH containing 29% acetonitrile. The amino acid composition of the dansyl-peptide is reported in Table 1. The lysine found (0.48 residues) should derive from &-dansyl-lysine;the low yield of free lysine in amino acid analysis was presumably due to only partial breakdown of the derivative. The peptide thus consisted of seven residues plus one E-dansyl-lysine. The stoichiometry of the labeled dansyl bound to the octapeptide, as determined by specific radioactivity and amino acid analysis was 0.85: 1. As expected [3], only lysine was labeled: TLC analysis [9] showed a radioactive spot corresponding to &-dansyl-lysine.It is worthwhile to notice that the peptide contains also an arginine residue, which should be the C-terminal amino acid of the peptide. The dansyl-peptide was then subjected to the manual Edman degradation giving the following sequence: H2N-Ser-Val-SerLeu-. Probably due to the presence of the dansyl group, the yields decreased drastically at each cycle, thus preventing further sequencing of the peptide beyond the fourth residue. As a consequence, C-sequence of the petide was attempted with carboxypeptidase Y. This analysis could only confirm that arginine was the carboxy-terminal amino acid; in fact, arginine was the only amino acid released in stoichiometric amount with respect to the digested peptide. To obtain the complete sequence of the dansyl-peptide, another specific hydrolysis was performed using thermolysin. The digest was analyzed by reverse phase HPLC and the resulting peaks collected for further analysis. The chromatograms (Fig. 2) show that three peptide peaks (2-4) were highly fluorescent, with peak 4 eluting at the same retention time of the original dansyl-peptide. The amino acid composition of peaks 1 - 3 is reported in Table 1. Analysis of the three peptides by dansyl method [9] showed that serine, valine and leucine were the N-terminal amino acids, respectively. Thus, thermolysin (Th) has hydrolyzed the Ser-Leu and Cys-Val bonds of the dansylpeptide, as expected from its known specificity. These data, together with the results reported above, allowed to align the sequence of the dansyl-peptide which is the following: Th Th 1 J +
HIN-Ser-Val-Ser-Leu-Cys-Val-Lys-Arg-COOH.
Due to the presence of a cysteine residue in the same peptide close to the lysine residue and being aware that a transfer of the dansyl group from a cysteine to another residue has been reported to occur in alkaline conditions [lo], it was attempted to modify the enzyme at a lower pH. Incubation of the protein with dansyl chloride for 3 h at 25°C in 0.05 M phosphate, pH 6.5, failed to give inactivation. Then, an intermediate value of pH (0.1 M Tris/HCl, pH 7.5) was selected: enzyme inactivation reached 80% in 80 min at 25°C. After gel filtration,
291
-.
,.-...pi-
'
- ...
~
..
E
.
C
.o
I
0 u)
+ 0 0,
u C Q)
u v)
0) L
0
-3 Y
$
.
0
-
cy cy
-
c
0
i~
I1 UI!
a) U
C
0
n L
51
n Q
c
BI 7-
1
n
I
0
50
70
n
.
9c
110
Fraction number
Fig. 1. High-performance liquid chromatograms of tryptic digests of ferredoxin-NADP' reductase mod@ed with [Me-'4C]dansyl chloride in the presence and the absence of NADP'. (A) Elution profile, monitored at 220 nm and for the dansyl fluorescence, of the enzyme modified in the absence of NADP'. (B) Histograms of the radioactivity found in the eluates, as determined after collection of fractions (0.6 ml) and scintillation counting, of the protein modified in the presence (shaded area) and in the absence of 2 mM NADP'
298
E C
.o 0 v) c
0 Qi U
C al U G?
a, L
0 3 U -
s-
1
50
30
10
70
E l u t i o n time ( m i n )
Fig. 2. High-performance liquid chromutogram of thermolysin digest of the tryptic dunsyl-peptide f r o m ferredoxin-NADP' reductase modified with dunsyl chloride. The elution was followed both by the absorbance at 220 nm and by monitoring the fluorescence at 506 nm
Table 1. Amino acid composition of the tryptic dunsyl-peptide and of frugments released by thermolysin digestion of the dansyl-peptide, purified f r o m ferredoxin-NADP+ reductase modified with dansyl chloride Residue
Dansyl-peptide
Themolysin-peptide 1
2
3
0.86 1.00 0.26
0.87
mol residue/mol peptide
Cys" Ser Val Leu '4%
LYS a
0.85 2.00 1.85 1.25 1.05 0.48
0.75 2.05 0.87 1.OO -
-
0.67 1.OO 1.01 0.35
Estimated as carboxymethylcysteine.
2.22 mol radioactive dansyl were bound per mol FAD. After acid treatment of the protein, which should labilize the Ssulfonyl bond, more than 80% of the radioaetivity incorporated remained protein-bound: 1.83 residues per FAD over 2.22 residues modified at pH 7.5, were acid-stable dansylderivatives. Thus, these experiments did not make it possible to discriminate the first target for the dansyl reagent. DISCUSSION A lysine residue has been showed by chemical modification with dansyl chloride [3] to be located in the NADP+-binding site of ferredoxin-NADPf reductase. We have found that only one peptide was highly fluorescent and highly radioactive in the tryptic digest of the modified protein. The stoichiometry
calculated from the specific radioactivity and amino acid analysis was approximately one. The dansyl group has been a very useful tool in view of the isolation and characterization of the labeled peptide, both for the sensitivity of its fluorescent moiety and for the stability of the corresponding lysine derivative. However, in our hands it appeared to make Edman sequencing difficult, probably due to loss of residual peptide during the extraction cycles. Moreover, obtaining the C-sequence of the peptide using carboxypeptidase Y was also arduous because of the low yield of the amino acids released and of difficulties in quantifying the dansyl-lysine by amino acid analysis. However, qualitative analysis by TLC showed a release of e-dansyl-lysine. It is worthwhile to mention that an identical sequence has been found by J. R. Herriot et al., who are determining the primary structure of spinach ferredoxin-NADP' reductase (Karplus, P. A,, Herriot, J. R., and Walsh, K. A., personal communication). Lysine has been reported to be the only amino acid quantitatively labeled by dansyl chloride in the whole protein [3]: these data have been confirmed by direct identification of dansyl-lysine in the hydrolyzate of the isolated peptide. The presence of a modified lysine is also demonstrated by the appearance of a Lys-Arg bond which has not been cleaved by trypsin because the modified lysine side-chain would not be recognized by the proteinase. Nevertheless, there are other nucleophiles in the peptide that could potentially become sulfonylated, namely thiol and hydroxyl groups. In this hypothesis, these nucleophiles may be the first target of the dansyl chloride, then, being at alkaline pH, a migration of the dansyl group to the e-amino group of the lysine residue could occur. However, the experiments at acidic pH showed no -SH or -OH involvement in the dansyl reaction. Nevertheless, supposing that a migration could actually happen in our experimental conditions, we would like to point out that the
299 Table 2. Sequences around the essential basic residue in ferredoxin-NADP' reductase, glutathione reductase and dihydrofolate reductuse isolated from several species Protein (source) Ferredoxin-NADP
Sequence
+
reductase (spinach)
Reference
+
(Lys)-Ser-Val-Ser-Leu-Cys-Val-Lys-Arg Arg
this report [I41
Glutathione reductase (human erythrocytes)
Lys-Thr-
Dihydrofolate reductase (Streptococcus fuecium)
Lys -1le-Leu-
-Val-Met-Gly-Arg-Lys
~ 3 1
Dihydrofolate reductase (Lactobacillus cusei)
Lys -1le-Met-
-Val-Val-Gly-Arg-Arg
[11,131
-Ser-Leu-Met-Ile-Arg-His
Dihydrofolate reductase (Escherichia coli)
Lys -Pro-Val-
-1le-Met-Gly-Arg-His
~131
Dihydrofolate reductase (chicken liver)
Asn-Ala-Val-
-1le-Met-Gly-Lys-Lys
[I 31
Dihydrofolate reductase (bovine, pig liver)
Asn-Leu-Val-
-1le-Met-Gly-Arg-Lys
I131
enzyme which has the lysine dansylated is inactive and unable to bind NADP' in the usual manner [3]. It was suggested from experiments with coenzyme analogs that the essential lysine interacts most likely with the 2'phosphate moiety of NADP' [3]. Possibly, both positive charges of the vicinal lysine and arginine residues of the peptide could be involved in binding the coenzyme; in fact, an essential arginine residue has been shown to be protected in the NADP+-enzyme complex [4]. To our knowledge, the binding site for the pyridine nucleotide of only two NADP(H)dependent dehydrogenases have been resolved by X-ray crystallography, namely dihydrofolate reductase [ 111 and glutathione reductase [12]. The guanidinium group of Arg-43 in Lactobacillus casei dihydrofolate reductase has been shown to be involved in binding the 2'-phosphate of NADPH [ll]. Notably, this residue is conserved in all the enzymes so far sequenced, except for the chicken liver reductase where it has been replaced by lysine [33]. In glutathione reductase the 2'phosphate group of NADPH is bound to Arg-218, His-219 and Arg-224 [14]. In both enzymes, the Arg or Lys residue implicated in binding the 2'-phosphate of NADPH are followed in the sequence by an other basic residue (Arg, Lys or His). The same situation is present in the peptide from ferredoxin-NADP reductase here described, where we found the sequence: Lys-Arg. Furthermore, another basic residue is found in these three enzymes (except for the vertebrate dihydrofolate reductase, in which Asn is substituting) 7 or 6 positions before the essential basic residue (Table 2). A comparison of these peptides is attempted, although the shortness precludes more meaningful conclusions. If minimum single-base changes are included, considering that the proteins and organisms are quite distantly related, the homology is 67-78%, with a minimum base change per codon ranging from 0.875-1.125. According to Haber and Koshland [15], a better method to evaluate sequences which are important in establishing a particular environment at an active site is to compare peptide sequences in terms of amino acid properties. By this procedure, we find seven correspondences over nine residues of the spinach enzyme after the introduction of one gap, with the glutathione reductase peptide and an average of five correspondences with the peptide of the dihydrofolate +
reductase from various sources. These values seem to indicate a high degree of relatedness among these peptides. However, until the three-dimensional structure of ferredoxin-NADP' reductase is fully resolved, a more general interpretation of chemical modification studies cannot be excluded. Namely, that a large conformational change is involved in differentiating the reactivity of the essential lysine in free and coenzyme-bound enzyme. This work was supported by grants from Minister0 della Pubblica Istruzione of Italy. We thank Dr J. R. Herriot for providing us with unpublished results.
REFERENCES 1. Sheriff, S. & Herriot, J. R. (1981) J . Mol. Biol. 145, 441 -451. 2. Karplus, P. A. & Herriot, J. R. (1 982) in Flavins and Flavoproteins (Massey, V. & Williams, C. H., Jr, eds) pp. 28-31, Elsevier/ North-Holland, New York. 3. Zanetti, G. (1976) Biochim. Biophys. Acta 445, 14-24. 4. Zanetti, G., Gozzer, C., Sacchi, G. & Curti, B. (1979) Biochim. Biophys. Acta 569, 127- 134. 5. Carillo, N., Arana, J. L. & Vallejos, R. H. (1981) J . Bid. Chem. 256,6823 - 6828. 6. Carrillo, N. & Vallejos, R. H. (1983) Biochemistry 22, 58895897. 7. Zanetti, G. & Curti, B. (1980) Methods Enzymol. 69, 250-255. 8. Tarr, G. E. (1982) in Methods in Protein Sequence Analysis (Elzinga, M., ed.) pp. 223-232, Humana Press, Clifton, New Jersey. 9. Gray, W. R. (1972) Methods Enzymol. 25B, 121- 138. 10. Dallocchio, F., Signorini, M. & Rippa, M. (1978) Arch. Biochem. Biophys. 185, 57-60. 11. Filman, D. J., Bolin, T. J., Matthews, D. A. & Kraut, J. (1982) J. Biol. Chem. 257, 13663 - 13672. 12. Pai, E. F. & Schulz, G. E. (1983) J . Biol. Chem. 258, 1752- 1757. 13. Volz, K. W., Matthew, D. A., Alden, R. A,, Freer, S. T., Hansch, C., Kaufman, B. T. & Kraut, J. (1982) J . Biol. Chem. 257, 2528 - 2536. 14. Krauth-Siegel, R. L., Blatterspiel, R., Saleh, M., Schiltz, E., Schirmer, R. H. & Untucht-Grau, R. (1982) Eur. J. Biochem. 121,259-267. 3 5. Haber, J. E. & Koshland, D. E., Jr (1970) J. Mol. Bid. 50, 617639.
D. Cidaria and G. Zanetti, Dipartimento di Fisiologia e Biochimica Generali, Via Celoria 26, 1-20133 Milano, Italy
P. A. Biondi and S. Ronchi, Istituto di Fisiologia Veterinaria e Biochimica, Via Celoria 10, 1-20133 Milano, Italy
