Binding of Nickel(II) and Copper(II) to the N-Terminal Sequence of

Aug 18, 1997 - A potentiometric and spectroscopic (UV/vis and CD) study of Cu(II) and Ni(II) binding to the N-terminal pentadecapeptide of human prota...
2 downloads 0 Views 247KB Size
Download PDF
906

Chem. Res. Toxicol. 1997, 10, 906-914

Binding of Nickel(II) and Copper(II) to the N-Terminal Sequence of Human Protamine HP2 Wojciech Bal,‡,§ Małgorzata Jez˘ owska-Bojczuk,† and Kazimierz S. Kasprzak*,‡ Laboratory of Comparative Carcinogenesis, National Cancer Institute, FCRDC, Frederick, Maryland 21702, and Faculty of Chemistry, University of Wroclaw, Wroclaw, Poland Received February 24, 1997X

A potentiometric and spectroscopic (UV/vis and CD) study of Cu(II) and Ni(II) binding to the N-terminal pentadecapeptide of human protamine HP2 (HP21-15) was performed. The results indicate that the N-terminal tripeptide motif Arg-Thr-His is the exclusive binding site for both metal ions at a metal to HP21-15 molar ratio not higher than 1. The very high value of protonation-corrected stability constant (log *K) for Ni(II)-HP21-15 complex, -19.29, indicates that HP2 has the potential to sequester Ni(II) from other peptide and protein carriers, including albumin. The same is likely for Cu(II) (log *K ) -13.13). The CD spectra of Cu(II) and Ni(II) complexes of HP21-15 indicate that the N-terminal metal binding affects the overall conformation of the peptide that, in turn, may alter interaction of HP2 with DNA. These results imply HP2 as a likely target for the toxic metals Ni(II) and Cu(II).

Introduction During spermatogenesis in eutherian mammals, the histones in chromatin are replaced by smaller protamines (1). This replacement results in much tighter compaction of DNA (2), as well as suppression of transcription, replication, and repair processes (3, 4). Protamine HP2,1 a 57-amino acid peptide, constitutes ca. 50-70% of human protamines (5, 6). Its sequence is given below:

RTHGQ-SHYRR-RHCSR-RRLHR-IHRRQ-HRSCRRRKRR-SCRHR-RRHRR-GCRTR-KRTCR-RH Even at a glance, this molecule offers a plethora of potential binding sites for any metal ions with affinity to nitrogen (His) or sulfur (Cys) donors (underlined). Such binding may alter biological functions of the protamine. Indeed, formation of a zinc finger in the central part of the molecule had been postulated on the basis of far-UV CD spectra (7) and Co(II) substitution (8). Convincing structural studies have not, however, yet been performed. Another striking feature of HP2, one that has not been yet noted, is the presence of the N-terminal motif ArgThr-His-. N-Terminal sequences of the X-X-His- type are found in many serum albumins, including human albumin (9, 10), in other blood-circulating proteins (11), and in some neurohormones (12, 13). Peptides of this sequence bind Ni(II) and Cu(II) with the terminal amino, imidazole, and two intervening amide nitrogen (Figure 1) (14, 15). For both metal ions the stabilities of such complexes are several orders of magnitude higher than in any other oligopeptide system (16-19). The apparent * Address correspondence to this author at NCI-FCRDC, Bldg 538, Rm 205, Frederick, MD 21702-1201; tel, 301-846-5738; fax, 301-8465946. ‡ National Cancer Institute. § On leave from the University of Wroclaw. † University of Wroclaw. X Abstract published in Advance ACS Abstracts, August 1, 1997. 1 Abbreviations: HP2, human protamine 2; HP2 1-5, Arg-Thr-HisGly-Gln-amide; HP21-15, Arg-Thr-His-Gly-Gln-Ser-His-Tyr-Arg-ArgArg-His-Cys-Ser-Arg-amide; L, peptide with side chains and amino terminus deprotonated; CT, charge transfer.

S0893-228x(97)00028-3

Figure 1. Structure of the four-nitrogen complex of X-X-Hispeptides: M, Cu(II) or Ni(II); R1, R2, side chains of amino acids in positions 1 and 2 (14, 15).

formation constants for Cu(II) complexes with various albumins possessing the X-X-His- terminal sequence are 1011-1013 M (20-22). Studies of peptide models indicate that amino acid substitutions in positions 1 and 2 do not affect the binding mode but may further increase the complex stability by ionic or hydrophobic interactions (23, 24). As a consequence, serum albumins serve as Cu(II) and Ni(II) carriers in blood plasma (25-29). By analogy, copper and/or nickel involvement in the biological activity of other peptides sharing such sites was proposed (24, 30). These facts indicate a distinct possibility that human protamine HP2 and its mouse analog should bind Cu(II) and Ni(II) in vivo. Such binding could contribute to the known toxicity of Ni(II) in the mouse testes (31). Ni(II) compounds are carcinogenic (32), acting, at least in part, through oxidative damage mechanisms (33). One of the major exposure routes is inhalation of nickelcontaining welding fumes (32). Epidemiology indicates that paternal exposure to these fumes may contribute to higher incidence of cancer in the progeny (34). Nickelrelated sperm DNA damage and/or nickel transport to the embryo with sperm seems to be an intriguing, although unproven, possibility for an explanation of this phenomenon. Also, evidence is emerging that copper may be a powerful mutagen and an oxidative damage catalyst in vivo (33, 35, 36). Protamine HP2 might be involved in the mechanisms of sperm DNA damage by providing binding sites for the metals close to DNA and facilitating their redox activity. Binding of Ni(II) or Cu-

This article not subject to U.S. Copyright.

Published 1997 by the American Chemical Society

Ni(II) and Cu(II) Binding to Human Protamine

(II) to the N-terminal sequence of HP2 may also result in abnormal packaging of DNA with possible epigenetic effects. The above facts encouraged us to investigate the coordinative properties of HP2 toward Ni(II) and Cu(II) and, subsequently, DNA damage in the presence of the resulting complexes (37). A direct attempt of a thorough thermodynamic and spectroscopic study of metal binding to the whole HP2 molecule, with its nine histidines and five cysteines, would be difficult. We therefore targeted the N-terminal sequence of this protamine. Two peptides were selected for the study on the basis of the presence of the Arg-ThrHis- motif, and of another potential Ni(II)- and Cu(II)binding site, comprised of the neighboring His and Cys residues:

HP21-15: RTHGQ-SHYRR-RHCSR-amide HP21-5: RTHGQ-amide To confirm the apparent lack of -SH involvement in metal binding by the HP21-15 peptide, which emerged during experiments, we also synthesized and studied the respective “midpeptide” HP210-15 (Ac-RHCSR-amide). This paper presents the results of the studies of Cu(II) and Ni(II) complexes of these peptides by pH-metry and UV/ vis and CD spectroscopies. The potential of promoting DNA damage by HP21-15 complexes is presented in the following paper (37).

Experimental Section Materials.2

Sodium and potassium phosphates, NaOH, and HCl were purchased from Fisher Scientific (Pittsburgh, PA). CuCl2, 99.999% purity, NiCl2, 99.9999% purity, and KNO3 were purchased from Aldrich Chemical Co. Inc. (Milwaukee, WI). The peptides were custom synthesized by QBC, Inc. (Hopkinton, MA). Their purity was verified by HPLC, mass spectrometry, and potentiometry to be >99%. Potentiometry. The protonation constants of the peptides and the stability constants of their Cu(II) and Ni(II) complexes in the presence of 0.1 M KNO3 were determined at 25 °C using pH-metric titrations over the pH range 3-11.5 (Molspin automatic titrator, Molspin Ltd., Newcastle-upon-Tyne, U.K.) with NaOH as titrant. Changes in pH were monitored with a combined glass-calomel electrode calibrated daily in hydrogen concentrations by HNO3 titrations (38). Sample volumes of 1.5 mL, concentrations of peptides of 0.7-1 mM, and peptide-tometal ion molar ratios of 1.1:1 were used. All measurements were performed under argon. The data were analyzed using the SUPERQUAD program (39). UV/Vis Spectra. The spectra of Cu(II) and Ni(II) complexes were recorded on a Beckman DU-640 spectrophotometer over the range of 250-1100 nm in 1 cm cuvettes, for 0.5-1 mM concentrations of peptides and 1.1:1 peptide-to-metal ion ratios. All samples were prepared under argon. CD Spectra. The spectra were recorded for 0.2-1 mM peptides and complexes in 0.02-1 cm cuvettes over the range of 190-750 nm, at 25 °C, using a JASCO-720 spectropolarimeter for far-UV measurements and a JASCO J-500A spectropolarimeter for near-UV and visible spectral ranges. Spectra are expressed in terms of ∆ ) l - r, where l and r are molar absorption coefficients for left and right circularly polarized 2 Certain commercial equipment and materials are identified in this paper in order to specify adequately the experimental procedure. Such specification does not imply recommendation or endorsement by the National Cancer Institute (National Institutes of Health, Department of Health and Human Services, U.S. Government), nor does it imply that the materials or equipment identified are necessarily the best available for the purpose.

Chem. Res. Toxicol., Vol. 10, No. 8, 1997 907 light, respectively. All operations with the samples were performed under argon. Combined Spectrophotometric and pH-Metric Determination of Stability Constants of Ni(II) Complexes. Stock solutions containing 0.55 mM peptide (HP21-5 or HP21-15), 0.5 mM NiCl2, and 0.1 M KNO3 were acidified with HNO3 to pH 4. Samples of HP21-5 were stored at 4 °C under argon, and samples of HP21-15 were stored frozen at -70 °C. Single portions of 0.1 M NaOH were added under argon to individual 0.4 mL samples of stock solutions, and resulting mixtures were monitored by UV/vis at 420 nm until equilibrium was achieved (typically within 3 h for HP21-5, 3-6 h for HP21-15). Spectra of equilibrated samples were recorded and their final pH values measured. For HP21-5, absorptions at three different wavelengths, 480, 466, and 423 nm, were used as a measure of concentration of a Ni(II) complex at a given pH between 5.4 and 5.9, using the average of absorption of samples at pH 7.0-7.9 as 100% complex formation. Alkali consumption at these pH values corresponded to neutralization of 4 equiv of H+. The shape of the visible range absorption band was constant in the pH range used, indicating the formation of a single complex species. The concentration stability constant (β) of the NiH-2L complex was calculated with thus obtained concentration data. For HP21-15, in addition to absorption, also sums of magnitudes of CD bands at 480 and 411 nm were taken to calculate the stability constants. The data at pH 5.6-6.4 were used for calculations, with those at pH 6.6-7.0 representing 100% complexation. The shapes of absorption and CD bands were unchanged in the 5.6-7.0 pH range, indicating a single coordination mode. Stability constants (β) for NiHL and NiL complexes were calculated from those data. Stoichiometries of these complexes were confirmed by the amounts of alkali consumed by complex formation. The pH titration curve, obtained by combining the pH readings for individual samples, yielded the pKa value for the NiHL f NiL reaction, consistent with β values.

Results The results of pH-metric titrations and the analysis of UV/vis and CD spectra allowed for determination of the protonation and metal complex stability constants and the assignment of pKa values to specific amino acid residues. Data were obtained for all three peptides. However, the complexation modes of both Cu(II) and Ni(II) to HP210-15 were completely different from the ones seen with the other two peptides. Also, Cu(II), but not Ni(II), rapidly oxidized HP210-15 to the corresponding disulfide dimer. CT characteristics for square-planar Ni(II)-sulfur bands were seen in Ni(II)-HP210-15 complexes at pH g 8. The stabilities of Cu(II) complexes with (HP210-15)2 and Ni(II) complexes with HP210-15 were 6 orders of magnitude lower than those seen for the other two peptides. All this indicated that the HP210-15 amino acid motif did not participate in the formation of metal complexes in the complete HP21-15 peptide. Therefore, except for ligand protonation constants, the data concerning HP210-15 are not included in this paper. Protonation Equilibria of Free Peptides. Protonation constants for the peptides studied are presented in Table 1A. CD spectra of HP21-5 and HP21-15 are presented in Figure 2A. These spectra exhibit a strong negative band below 200 nm and another negative band or shoulder above 230 nm. An intervening positive band is seen in HP21-5 as well as in HP21-15 at acidic and neutral pH. All these bands originate from amide chromophores. The presence of the aromatic chromophore of Tyr does not perturb this pattern below pH 10. Deprotonation of the Tyr phenolic oxygen causes a

908 Chem. Res. Toxicol., Vol. 10, No. 8, 1997

Bal et al.

Figure 2. Far-UV CD spectra: (A) uncoordinated peptides, HP21-15 at pH 4.2 (s), 6.1 (- - -), 7.4 (‚‚‚), and 11.7 (-‚-) and HP21-5 at pH 7.4 (-‚‚-); (B) Cu(II) complexes of HP21-15 at pH 4.0 (s), 5.0 (- - -), 7.4 (‚‚‚), and 12.2 (-‚-) and of HP21-5 at pH 7.4 (-‚‚-); (C) Ni(II) complexes of HP21-15 at pH 5.9 (s), 7.0 (- - -), 7.4 (‚‚‚), and 11.7 (-‚-) and of HP21-5 at pH 7.4 (-‚‚-). Table 1. Protonation and Stability Constants (log β valuesa) of HP2 Peptides and Their Cu(II) and Ni(II) Complexes A. Protonation of Peptides H6L

log β

H5L

H4L

H3L

HL

B. Stability Constants of Cu(II) Complexes CuH2L

CuHL

CuL

CuH-1L CuH-2L CuH-3L

-0.96(4) -10.34(6) HP21-5 HP21-15 30.06(6) 24.67(5) 18.38(6) 9.13(6) -1.11(6) C. Stability Constants of Ni(II) Complexes log β HP21-5 HP21-15

NiHL 18.65(9)b

NiL 12.35(9)b

A. Peptides pKa

H2L

13.276(2) 7.246(2) HP21-5 HP210-15 14.20(2) 8.22(2) HP21-15 43.19(2) 37.94(2) 32.07(2) 25.68(2) 18.49(1) 9.95(1)

log β

Table 2. Stepwise Protonation Constants (pKa valuesa) Calculated from the Data in Table 1

NiH-1L

NiH-2L

3.89(2)c

-5.95(7)b -5.95(2)c

a For peptides, β(H L) ) [H L]/([L][H+]n); for complexes, β(MH L) n n n ) [MHnL]/([M][L][H+]n), M ) Cu(II) or Ni(II). Values in parentheses are standard deviations of the last digits of constants. b Constant determined by spectroscopy. c Constant determined by potentiometry, with correction from spectroscopy.

red shift and intensification of the bands of the aromatic ring (40). Accordingly, the La band of Tyr chromophore is observed at 248 nm at pH 11.67. A weak band at 250 nm, unchanged throughout the acidic and neutral pH range, can be attributed to the thiol chromophore. This band broadens and blue-shifts slightly at pH 9.04. Cys thiolate is expected to deprotonate near this pH. Such broad CD bands around 240-250 nm are usually assigned to disulfide bridges (40). However, we did not detect any peptide dimerization in the course of the CD

HP21-5 HP210-15 HP21-15

H6L

5.25b

H5L

5.87b

H4L

6.39b

H3L

H 2L

HL

7.17c

6.030b 5.98b 8.54d

7.246c 8.22d 9.95e

B. Cu(II) Complexes pKa

CuH2L

CuHL

CuL

CuH-1L

HP21-5 HP21-15

5.41b

6.29b

9.25d

10.24e

CuH-2L 9.38f

C. Ni (II) Complexes pKa

NiHL

NiL

NiH-1L

HP21-15

6.3b

8.46d

9.84e

a For peptides, pK (H L) ) log β(H L) - log β(H a n n n-1L); for complexes, pKa(MHnL) ) log β(MHnL) - log β(MHn-1L), M ) b c d Cu(II) or Ni(II). His imidazole N-3. Terminal amine. Cys thiol. e Tyr phenol. f His imidazole N-1 (?).

experiment. Deprotonated thiol exhibits a strong transition at 240 nm in absorption spectra (41), and it is very likely that this transition gives rise to the subtle change in the CD spectra. The above features, as well as the comparison of protonation constants with literature data (19) and peptide sequences, allowed for an overall assignment of pKa values, as given in Table 2. Cu(II) Binding. Stability constants for Cu(II) complexes of HP21-5 and HP21-15 calculated from pH-metric titrations are presented in Table 1B. Spectral parameters of these complexes derived from UV/vis and CD

Ni(II) and Cu(II) Binding to Human Protamine

Chem. Res. Toxicol., Vol. 10, No. 8, 1997 909

Table 3. Spectroscopic Parameters of Cu(II) Complexes of HP21-5 and HP21-15 species

vis: λ ()

CuH-2L

HP21-5 527 (114)

CuH-3L

527 (116)

CuH2L

HP21-15 529 (85)

CuHL

529 (95)

CuL

CuH-1L

CuH-2L

529 (112)

526 (124)

526 (138)

CD: λ (∆) 566 (-1.02)a 484 (+0.49)a 312 (+2.10)b 271 (-4.39)c 243 (+1.45)d 221 (-11.8)e 196 (-12.0)e 568 (-1.01)a 487 (+0.48)a 311 (+2.02)b 271 (-4.33)c 242 (+1.75)d 221 (-12.1)e 196 (-12.8)e 567 (-0.63)a 486 (+0.33)a 312 (+1.34)b 270 (-4.40)c 239 (-1.3)d 222 (+7.7)e 197 (-69)e 567 (-0.70)a 486 (+0.36)a 312 (+1.38)b 270 (-3.60)c 245 (+1.4)d sh 220 (-12)e 197 (-58)e 567 (-0.82)a 486 (+0.41)a 312 (+1.70)b 273 (-3.80)c 247 (+2.8)d sh 220 (-12)e 198 (-53)e 568 (-0.82)a 486 (+0.41)a 313 (+1.68)b 273 (-3.80)c 248 (+3.0)d sh 219 (-23)e 200 (-60)e 569 (-0.90)a 485 (+0.45)a 310 (+1.72)b 271 (-4.2)c 245 (+6.6)f sh 218 (-12)e 200 (-59)e

a d-d transition; sh denotes a shoulder on the spectrum. b Nf Cu(II) CT with possible contribution of N (imidazole) f Cu(II) CT. c NH2 f Cu(II) CT with possible contribution of N (imidazole) f Cu(II) CT. d Intraligand (amide and sulfur-centered) transitions with possible contribution of N (imidazole) f Cu(II) CT. e Intraligand (amide and aromatic) transitions. f La transition in the Tyr ring with contributions of amide and sulfur-centered transitions and N (imidazole) f Cu(II) CT.

spectra with the aid of potentiometrically derived speciations are presented in Table 3. Figure 3 presents species distribution curves calculated for various potentiometric models of the Cu(II)/HP21-5 system. Addition of minor forms, CuHL, CuL, and CuH-1L in particular, improved the numerical fit. These three complexes were not, however, detected by spectroscopy. Wavelengths of CD extrema and ratios of their intensities are constant in the pH range of 4.3-6.4, indicating the presence of a single complex species. Concentrations of this complex at various pH values, derived from the CD spectra, are marked in Figure 3 over speciation curves derived from

Figure 3. Comparison of various potentiometric models of speciation in the Cu(II)/HP21-5 system with spectroscopic data: CCu(II) ) CL ) 0.5 mM; Cu2+ (‚‚‚), CuL (-‚‚-), CuH-1L (-‚-), CuH-2L (s), CuH-3L (- - -), CuH-2L by spectroscopy (b); (A) 1N to 4N complexes included; (B) 3N and 4N complexes included; (C) only 4N complexes included.

stability constants. A perfect match is seen only with the bottom model in Figure 3. The stability constants from this particular model are included in Table 1B. Spectroscopic parameters of 4N complexes, CuH-2L and CuH-3L, are typical for X-X-His binding. The formation of CuH-3L could be attributed to the deprotonation of the N-1 of the imidazole ring (16, 17). On the other hand, however, the pKa value of 9.4 is too low for this process; alternatives, like deprotonation of a complex-associated water molecule, cannot be excluded. In contrast to the HP21-5, the best-fitting potentiometric model for HP21-15 only included complexes having four or more protons dissociated from the peptide. Spectroscopic data showed the presence of exclusively 4N complexes (formation of which requires displacement of four protons by the metal ion). A very good match between the two methods allowed calculation of the spectra of individual complexes (Table 3). The initial CuH2L complex undergoes four consecutive deprotonations with increasing pH (Figure 4A). The corresponding pKa values (Table 2) match deprotonations of two of the His, the Cys, and the Tyr residues in the free peptide. These deprotonations alter the Cu(II)-related spectral bands only slightly, and so the complex structure remains basically unchanged throughout the pH range studied. The imidazole N-1 deprotonation was not seen in titrations performed up to pH 11, consistent with the studies of simpler peptides (16, 17). CD spectra provide the evidence of Tyr deprotonation as the final step, analogously to the free peptide (Table 3). Figure 5 presents visible-range CD spectra of Cu(II) complexes of HP21-5

910 Chem. Res. Toxicol., Vol. 10, No. 8, 1997

Figure 4. Species distribution diagrams for HP21-15 complexes: CCu(II) or CNi(II) ) CL ) 0.5 mM; (A) Cu(II) complexes; (B) Ni(II) complexes. Profiles of formation of corresponding MH-2L complexes of HP21-5 are shown with dotted lines.

Bal et al.

and HP21-15 at pH 7.4, illustrating the similarity of the binding modes. Figure 2B presents the far-UV CD spectra of Cu(II) complexes of HP21-5 and HP21-15. Cu(II) complexation apparently affects the 15-peptide conformation, as seen in the change of magnitude and position of the short wavelength minimum, as well as introduction of a negative band at 220 nm in the neutral pH region. The latter effect is also apparent in the 5-peptide complex. Ni(II) Binding. Stability constants of Ni(II) complexes with the HP2 peptides are presented in Table 1C. The slow rate of reaction between Ni(II) and HP21-5 made the potentiometric study impossible, because the pH change resulting from Ni(II) binding could not be discerned from the electrode drift during titrations. Instead, absorption spectroscopy was used to determine the stability constant of the first square-planar complex in a batch approach, as described in the Experimental Section. The NiH-2L stoichiometry, equivalent to a 4N complex, was confirmed by the mass balance of added alkali, by self-consistency of calculations, and by a characteristic CD spectrum (Figure 5B, Table 4), typical of this class of complexes (Table 5). Prolonged incubation of the Ni(II)/HP21-5 mixtures in neutral pH resulted in complex decomposition, manifested by changes in its CD spectra, and prevented detection of a hypothetical NiH-3L species. Potentiometric titrations of HP21-15/Ni(II) mixtures were successful in providing a set of stability constants. However, a significant discrepancy was noted between potentiometry-derived speciation and spectroscopic measurements, in which samples were monitored for the equilibrium. Therefore, the combined spectroscopic and pH-metric determination of stability constants was performed. It was found that one proton was released from

Figure 5. Visible range of CD spectra of complexes of HP21-15 (s) and HP21-5 (‚‚‚) at pH 7.4: (A) Cu(II) species; (B) Ni(II) species.

Ni(II) and Cu(II) Binding to Human Protamine

Chem. Res. Toxicol., Vol. 10, No. 8, 1997 911

Table 4. Spectroscopic Characterization of Ni(II) Complexes species

vis: λ ()

NiH-2L

HP21-5 sh 480 (85) 423 (160)

CD: λ (∆)

NiHL

HP21-15 sh 480 (80) 424 (154)

NiL

sh 480 (80) 424 (154)

NiH-1L

428 (174)

NiH-2L

428 (202)

480 (-2.71)a 411 (+1.48)a 260 (+3.8)b 235 (-3.0)c 211 (+3.9)d >190 (>-9)d 480 (-2.57)a 411 (+1.38)a sh 275 (+1.9)e 260 (+3.1)e 231 (-13.0)c 196 (-61)d 480 (-2.57)a 411 (+1.38)a sh 275 (+1.9)e 260 (+3.1)e 231 (-15.5)c 198 (-56)d 477 (-2.60)a 409 (+1.18)a sh 272 (+1.6)e 259 (+2.9)e 236 (-6.0)c 198 (-70)d 477 (-2.83)a 409 (+1.11)a sh 272 (+1.6)e 259 (+4.8)e 226 (-15.0)f 198 (-60)d

a d-d transition; sh denotes a shoulder on the spectrum. b Nf Ni(II) CT with possible contribution of N (imidazole) f Ni(II) CT. c NH2 f Ni(II) CT with possible contributions of N (imidazole) f Ni(II) CT and intraligand (amide and sulfur-centered) transitions. d Intraligand (amide and aromatic) transitions. e Mixture of N- f Ni(II) CT and N (imidazole) f Ni(II) CT. f Intraligand amide with contribution from NH2 f Ni(II) CT.

Table 5. Comparison of CD Spectra of 4N Ni(II) Complexes of X-X-His- Peptides peptide

λext I

∆ I

λext II

∆ II

Gly-Gly-Hisa Val-Ile-His-Asnb Asp-Thr-His-c Asp-Ala-His-d His-Arg-His-e Arg-Thr-His-f Arg-Thr-His-g

497 468 470 477 478 480 480

-1.31 -1.73 -2.45 -2.43 -2.54 -2.71 -2.57

413 402 405 412 412 411 411

+1.77 +0.78 +1.79 +1.48 +2.15 +1.48 +1.38

a Reference 42. b Reference 24. c Human serum albumin (W. Bal and P. J. Sadler, unpublished results). d Bovine serum albumin (43). e pNiXa peptide (42). f This work; NiH-2L complex of HP21-5. g This work; NiL complex of HP2 1-15.

the peptide at pH below the onset of complexation and up to 4.8 equiv were consumed in the course of complexation under pH 7. These facts indicate that the lowestpH Ni(II) complex of HP21-15 has NiHL stoichiometry, and in the course of Ni2+ binding, also a NiL complex is formed. Calculations with spectroscopic data provided stability constants for these complexes, and the analysis of the pH profile of complexation yielded the pKa value of NiL formation in excellent agreement with these constants (Tables 1C, 2C). The Ni(II) binding reaction was very slow below pH 6.5 but increased rapidly above that pH, reducing the time of equilibration in reaction mixtures from several hours to a few minutes. As a result, the upper parts of potentiometric titration curves, above pH 8, gave reliable values for deprotonation constants of NiL and NiH-1L complexes (Table 2C).

Stability constants were back-calculated from these pKa values using the log β value for NiL (Table 1C). Figure 4B presents the speciation curves for the Ni(II)/HP21-15 system; the formation of the NiH-2L complex of HP21-5 is superimposed for comparison. As in the case of Cu(II) complexes, the stepwise deprotonation processes can be assigned to particular amino acid side chains. Figure 5B shows the visible part of CD spectra of Ni(II) complexes with HP21-15 and HP21-5 at pH 7.4. Table 4 presents the parameters of all Ni(II) complexes derived from CD and UV/vis spectra.

Discussion Modes of Binding of Cu(II) and Ni(II) to HP2 Peptides. The results clearly show that the N-terminal tripeptide Arg-Thr-His- is the primary binding site for both Cu(II) and Ni(II) in HP21-15. The binding involves four nitrogen donors arranged equatorially around the metal ion in the typical X-X-His- manner (Figure 1). A transition from the lowest-pH forms, CuH2L and NiHL, to highest-pH forms, CuH-2L and NiH-2L, involves deprotonations of the two His, the Cys, and finally the Tyr side chains. These deprotonations do not alter the binding mode but have some influence on the spectral parameters of the metal binding. Intensities of d-d bands in CD spectra of Cu(II) complexes of HP21-15 increase gradually with the decrease of the electrostatic charge of the peptide, but the ratio of intensities of component transitions at 567 and 486 nm remains constant. Maximum intensity, achieved in the CuH-2L complex, equals 90% of the corresponding value for the CuH-2L complex of HP21-5. These relations indicate that the geometry of the binding site is identical in these complexes. The differences of CD intensities result from changes in conformational flexibility of the binding site. Higher mobility leads to the decrease of CD intensity (44). Such plasticity, resulting from electrostatic interactions with charged side chains, is typical for Cu(II) complexes. The Ni(II) binding geometry is not sensitive to the His deprotonation (NiHL f NiL), and the sum of intensities of CD d-d bands is ca. 3 times higher compared to Cu(II) complexes. This is consistent with the properties of square-planar Ni(II) complexes, which are not susceptible to deformations of the coordination plane. Stability of Complexes. Stabilities of Cu(II) and Ni(II) complexes with HP21-5 and HP21-15 are very high. The comparison of *K constants, describing competition of the metal ions with protons for the peptide binding, M(II) + HnL f MHn-jL + jH+, clearly indicates that stabilities of both Cu(II) and Ni(II) complexes are 2-3 orders of magnitude higher than those of the “generic” Gly-Gly-His species (Table 6). Ni(II) complexes of HP2 peptides are even more stable than the corresponding complex of Val-Ile-His-Asn (24), the most stable Ni(II)peptide complex within the X-X-His- family, known previously! This is a surprising result for such hydrophilic and highly positively charged peptides. Beside similarities between Cu(II) and Ni(II) complexation, there is also one subtle difference. While the affinity of Ni(II) to both HP21-5 and HP21-15 is practically identical, the CuH2L complex of HP21-15 is more stable from the corresponding CuH-2L complex of HP21-5 by 1 log unit (Figure 4). There are two major factors influencing the stabilities of X-X-His complexes. One of them is the differential affinity for protons and a given metal ion,

912 Chem. Res. Toxicol., Vol. 10, No. 8, 1997

Bal et al.

Table 6. Comparison of log β and log *K Values for Selected 4N Complexes of X-X-His- Peptides with Cu(II) and Ni(II) Cu(II) peptide Gly-Gly-Hisb Gly-Gly-Hisc Gly-Gly-histd Val-Ile-His-Asne HP21-5f HP21-15g

Ni(II)

log β

log

*Ka

log β

log *Ka

-1.73 -1.55 -2.48

-16.43 -16.33 -17.14

-6.93

-21.81

-1.11 -0.96

-14.24 -13.13

-7.99 -5.39 -5.95 -5.95

-22.65 -19.75 -19.23 -19.29

a log *K ) log β(CuH b c n-jL) - log β(HnL). Reference 46. Reference 17. d Reference 47; hist stands for histamine. e Reference 24. f This work. g This work; CuH2L and NiHL complexes, respectively.

resulting from the electronic effects in the peptide. The other is the extent of sterical shielding of the metal-bound amide nitrogens in the complex molecule from the bulk of solution (45). Such shielding, exerted, for example, by amino acid side chains, slows down the hydrogen ioncatalyzed complex dissociation. The electronic mechanism certainly operates in HP2 complexes. It can be seen in the lowering of pKa values of both Arg1 and His3 by 0.6-0.8 log unit vs Gly-Gly-His (17, 46). But this effect is almost identical in HP21-5 and HP21-15, and so the steric mechanism is likely to be involved in the stabilization of Cu(II)-HP21-15 complexes. The question of why it affects copper, but not nickel, binding is addressed in the section describing peptide conformation, below. The difference of binding affinities between Ni(II) complexes of HP21-15 and HP210-15 is ca. 6 orders of magnitude in favor of the former. Similar difference, between HP21-15 and (HP210-15)2, can be estimated for Cu(II). At a molar ratio of HP21-15 to Cu(II) or Ni(II) higher than 1, there is no involvement of Cys sulfur, or any other groups beyond the Arg-Thr-His motif, in the metal binding. Clearly, in a biologically feasible situation, where the excess of protamine over toxic metal is assured, there will be no binding to the -His12-Cys13motif. Effects of Cu(II) and Ni(II) on the Conformation of HP21-15. The CD spectral pattern observed for free HP21-15 (Figure 2A) is typical for unordered peptides (41) assuming a range of rapidly interconverting conformations with no rigid structures (like R-helices or β-turns) present. However, the positive band at 220 nm has been associated with the presence of significant amounts of the left-handed 3-fold helix (48). Such pattern has been observed for charged amino acid homopolymers. The 220 nm band vanishes gradually, and the intensity of the negative band at 196 nm increases, as peptide deprotonates, suggesting the loss of helicity. Complexation of HP21-15 with Cu(II), together with partial deprotonation of nonbinding histidine residues, induces a moderately strong negative band at 220 nm and an apparent red shift in the negative maximum below 200 nm. These effects are not introduced by the sole fact of Cu(II) coordination. Planar arrangement of the peptide chain around Cu(II) does not give rise to any CD bands (18), and all Cu(II)-nitrogen CT bands are located above 250 nm (49). Difference spectra (Figure 6) reveal a pattern strikingly resembling a mixture of R-helix and parallel β-sheet (40, 50). Curiously, a similar but not identical pattern is produced by the complex of HP21-5. Similar but less pronounced changes also result from Ni(II) coordination to HP21-15. In this case the

Figure 6. Difference CD spectra at pH 7.4: CuL complex of HP21-15 minus free HP21-15 (s); CuH-2L complex of HP21-5 minus free HP21-5 (‚‚‚); NiL complex of HP21-15 minus free HP21-15 (- - -); NiH-2L complex of HP21-5 minus free HP21-5 (-‚-).

pattern is partly obscured by the presence of the NH2Ni(II) CT band around 230 nm. The HP21-5-Ni(II) complex, however, does not exhibit any effects of this kind. These results indicate that both N-terminally bound Cu(II) and Ni(II) introduce partial ordering, possibly short stretches of R-helix, in HP21-15 but located in different parts of the molecule. The formation of a β-sheet would require peptide aggregation, which is improbable for a peptide carrying an electronic charge of +6 or more and was in fact not observed. The Cu(II) complexation affects the conformation of residues 4 and 5 (neighboring the binding site), but in the Ni(II) complex these residues seem not to be affected. These effects may help rationalize the stability difference between HP21-5 and HP21-15 complexes of individual metals. The conformational arrangement of residues 4 and 5, next to the Cu(II)-binding site, can position further residues so that they shield the coordination plane from the solvent. Amino acids in positions 4 and 5, however, cannot participate in this effect directly. Hence, the spectral properties of Cu(II) complexes of HP21-5 and HP21-15 are similar, but the stabilization is only present in the latter. In the Ni(II) complex, ordering occurs away from the binding site, and there is no shielding and no stability gain. It is interesting to notice that the Cu(II)- and Ni(II)-imposed changes in CD spectra of HP21-15 are similar to those produced in the whole HP2 by Zn(II) (7), interpreted as formation of multiple β-turns. Further studies are necessary to test whether there is a common phenomenon behind these spectral effects.

Conclusion The results of the present study show that the binding affinity of Ni(II) and Cu(II) for the N-terminal motif of HP2 is at least as high as that for serum albumin, the major metal carrier in blood. Therefore HP2 may be an important target for toxicity of both metals. One possibly

Ni(II) and Cu(II) Binding to Human Protamine

pathogenic effect of Ni(II) and Cu(II) on HP2 would be partial ordering of its molecule that may affect proper assembly of HP2 with DNA. Another one includes mediation by the bound metals of oxidative damage to both HP2 and associated DNA, reported in the following paper (37). Technical Note: Reliability of Potentiometric Determinations of Stability Constants. In the course of our studies we had to refine one potentiometric model of speciation using spectroscopic data and completely revise another one. These observations reinforce the need of the multimethod approach in elucidating coordination equilibria in biologically relevant systems.

Acknowledgment. The authors wish to thank Dr. Jan Lukszo for help with verification of purity of the peptides, Drs. Anthony Dipple, Dong Xie, and John W. Erickson for making their CD spectrometers available to us, and Drs. Aloka Srinivasan and Lucy M. Anderson for critical comments on the manuscript.

References (1) McKay, D. J., Renaux, B. S., and Dixon, G. H. (1986) Human sperm protamines. Amino-acid sequences of two forms of protamine P2. Eur. J. Biochem. 156, 5-8. (2) Pogany, G. C., Corzett, M., Weston, S., and Balhorn, R. (1981) DNA and protein content of mouse sperm. Exp. Cell Res. 136, 127-136. (3) Chandley, A. C., and Kofman-Alfaro, S. (1971) “Unscheduled” DNA synthesis in human germ cells following UV irradiation. Exp. Cell Res. 69, 45-48. (4) Sega, G. A. (1974) Unscheduled DNA synthesis in the germ cells of male mice exposed in vivo to the chemical mutagen ethyl methanesulfonate. Proc. Natl. Acad. Sci. U.S.A. 71, 4955-4959. (5) Ammer, H., Henschen, A., and Lee, C.-H. (1986) Isolation and amino-acid sequence analysis of human sperm protamines P1 and P2. Occurrence of two forms of protamine P2. Biol. Chem. HoppeSeyler 367, 515-522. (6) Arkhis, A., Martinage, A., Sautiere, P., and Chevaillier, P. (1991) Molecular structure of human protamine P4 (HP4), a minor basic protein of human sperm nuclei. Eur. J. Biochem. 200, 387-392. (7) Gatewood, J. M., Schroth, G. P., Schmid, C. W., and Bradbury, E. M. (1990) Zinc-induced secondary structure transitions in human sperm protamines. J. Biol. Chem. 265, 20667-20672. (8) Bianchi, F., Rousseaux-Prevost, R., Sautiere, P., and Rousseaux, J. (1992) P2 protamines from human sperm are zinc-finger proteins with one Cys2/His2 motif. Biochem. Biophys. Res. Commun. 182, 540-547. (9) Peters, T., Jr., and Blumenstock, F. A. (1967) Copper-binding properties of bovine serum albumin and its amino-terminal peptide fragment. J. Biol. Chem. 242, 1574-1578. (10) Predki, P. F., Harford, C., Brar, P., and Sarkar, B. (1992) Further characterization of the N-terminal copper(II)- and nickel(II)binding motif of proteins. Biochem. J. 287, 211-215. (11) Hilgenfeldt, U. (1988) Half-life of rat angiotensinogen: influence of nephrectomy and lipopolysaccharide stimulation. Mol. Cell. Endocrinol. 56, 91-98. (12) Minamino, N., Kangawa, K., and Matsuo, H. (1984) Neuromedin B is a major bombesin-like peptide in rat brain: regional distribution of neuromedin B and neuromedin C in rat brain, pituitary and spinal cord. Biochem. Biophys. Res. Commun. 124, 925-932. (13) Del Rio, M., and De la Fuente, M. (1994) Chemoattractant capacity of bombesin, gastrin-releasing peptide and neuromedin C is mediated through PKC activation in murine peritoneal leukocytes. Regul. Pept. 49, 185-193. (14) Camerman, N., Camerman, A., and Sarkar, B. (1967) Molecular design to mimic the copper(II) transport site of human albumin. The crystal and molecular structure of copper(II)-glycylglycyl-Lhistidine-N-methyl amide monoaquo complex. Can. J. Chem. 54, 1309-1316. (15) Bal, W., Djuran, M. I., Margerum, D. W., Gray, E. T., Jr., Mazid, M. A., Tom, R. T., Nieboer, E., and Sadler, P. J. (1994) Dioxygeninduced decarboxylation and hydroxylation of [NiII(Glycyl-GlycylL-Histidine)] occurs via NiIII: X-ray crystal structure of [NiII(Glycyl-Glycyl-R-hydroxy-D,L-Histamine)]·3H2O. J. Chem. Soc., Chem. Commun. 1889-1890.

Chem. Res. Toxicol., Vol. 10, No. 8, 1997 913 (16) Sigel, H., and Martin, R. B. (1982) Coordinating properties of the amide bond. Stability and structure of metal ion complexes of peptides and related ligands. Chem. Rev. 82, 385-426. (17) Farkas, E., Sovago, I., Kiss, T., and Gergely, A. (1984) Studies on transition-metal-peptide complexes. Part 9. Copper(II) complexes of tripeptides containing histidine. J. Chem. Soc., Dalton Trans. 611-614. (18) Bal, W., Kozlowski, H., Kupryszewski, G., Mackiewicz, Z., Pettit, L. D., and Robbins, R. (1993) Complexes of Cu(II) with Asn-SerPhe-Arg-Tyr-NH2; an example of metal ion-promoted conformational organization which results in exceptionally high complex stability. J. Inorg. Biochem. 52, 79-87. (19) IUPAC and Academic Software. (1995) IUPAC Stability Constants Database, release 2. (20) Giroux, E., and Schoun, J. (1981) Copper and zinc binding by bovine, dog, and rat serum albumins. J. Inorg. Biochem. 14, 359362. (21) Zgirski, A., and Frieden, E. (1990) Binding of Cu(II) to nonprosthetic sites in ceruloplasmin and bovine serum albumin. J. Inorg. Biochem. 39, 137-148. (22) Masuoka, J., and Saltman, P. (1994) Zinc(II) and copper(II) binding to serum albumin. J. Biol. Chem. 269, 25557-25561. (23) Sadler, P. J., Tucker, A., and Viles, J. H. (1994) Involvement of a lysine residue in the N-terminal Ni2+ and Cu2+ binding site of serum albumins. Comparison with Co2+, Cd2+, and Al3+. Eur. J. Biochem. 220, 193-200. (24) Bal, W., Chmurny, G. N., Hilton, B. D., Sadler, P. J., and Tucker, A. (1996) Axial hydrophobic fence in highly-stable Ni(II) complex of des-angiotensinogen N-terminal peptide. J. Am. Chem. Soc. 118, 4727-4728. (25) Neumann, P. Z., and Sass-Kortsak, A. (1967) The state of copper in human serum: evidence for an amino acid-bound fraction. J. Clin. Invest. 46, 646-658. (26) Nieboer, E., Tom, R. T., and Sanford, W. E. (1988) Nickel metabolism in man and animals. In Metal Ions in Biological Systems, vol. 23, Nickel and Its Role in Biology (Sigel, H., and Sigel, A., Eds.) pp 91-121, Marcel Dekker, New York. (27) Sunderman, F. W., Jr. (1977) A review of the metabolism and toxicology of nickel. Ann. Clin. Lab. Sci. 7, 377-398. (28) Bligh, S. W., Boyle, H. A., McEwen, A. B., Sadler, P. J., and Woodham, R. H. (1992) 1H NMR studies of reactions of copper complexes with human blood plasma and urine. Biochem. Pharmacol. 43, 137-145. (29) Patel, S. U., Sadler, P. J., Tucker, A., and Viles, J. H. (1993) Direct detection of albumin in human blood plasma by 1H NMR spectroscopy. Complexation of Nickel2+. J. Am. Chem. Soc. 115, 9285-9286. (30) Harford, C., and Sarkar, B. (1995) Neuromedin C binds Cu(II) and Ni(II) via the ATCUN motif: implications for the CNS and cancer growth. Biochem. Biophys. Res. Commun. 209, 877-882. (31) Xie, J., Funakoshi, T., Shimada, H., and Kojima, S. (1995) Effects of chelating agents on testicular toxicity in mice caused by acute exposure to nickel. Toxicology 103, 147-155. (32) International Agency for Research on Cancer. (1990) IARC Monographs on the Evaluation of Carcinogenic Risk to Humans. Vol. 49. Chromium, Nickel and Welding, IARC, Lyon, France. (33) Kasprzak, K. S. (1995) Possible role of oxidative damage in metalinduced carcinogenesis. Cancer Invest. 13, 411-430. (34) Bunin, G. R., Noller, K., Rose, P., and Smith, E. (1992) Carcinogenesis. In Occupational and Environmental Reproductive Hazards: A Guide for Clinicians (Paul, M., Ed.) pp 76-88, Williams and Wilkins, Baltimore, MD. (35) Tkeshelashvili, L. K., McBride, T., Spence, K., and Loeb, L. A. (1991) Mutation spectrum of copper-induced DNA damage. J. Biol. Chem. 266, 6401-6406. (36) Aruoma, O. I., Halliwell, B., Gajewski, E., and Dizdaroglu, M. (1991) Copper-ion-dependent damage to the bases in DNA in the presence of hydrogen peroxide. Biochem. J. 273, 601-604. (37) Bal, W., Lukszo, J., and Kasprzak, K. S. (1997) Mediation of oxidative DNA damage by nickel(II) and copper(II) complexes with the N-terminal sequence of human protamine HP2. Chem. Res. Toxicol. 10, 915-921. (38) Irving, H., Miles, M. G., and Pettit, L. D. (1967) A study of some problems in determining the stoichiometric proton dissociation constants of complexes by potentiometric titrations using a glass electrode. Anal. Chim. Acta 38, 475-488. (39) Gans, P., Sabatini, A., and Vacca, A. (1985) SUPERQUAD: An improved general program for computation of formation constants from potentiometric data. J. Chem. Soc., Dalton Trans. 11951199.

914 Chem. Res. Toxicol., Vol. 10, No. 8, 1997 (40) Woody, R. W. (1994) Circular dichroism of peptides and proteins. In Circular Dichroism. Principles and Applications (Nakanishi, K., Berova, N., and Woody, R. W., Eds.) pp 473-496, VCH Publishers, New York. (41) Kortemme, T., and Creighton, T. E. (1995) Ionisation of cysteine residues at the termini of model R-helical peptides. Relevance to unusual thiol pKa values in proteins of the thioredoxin family. J. Mol. Biol. 253, 799-812. (42) Sunderman, F. W., Jr., Varghese, A. H., Kroftova, O. S., GrbacIvankovic, S., Kotyza, J., Datta, A. K., Davis, M., Bal, W., and Kasprzak, K. S. (1996) Characterization of pNiXa, a serpin of Xenopus Laevis oocytes and embryos, and its histidine-rich, Ni(II) binding domain. Mol. Reprod. Dev. 44, 5307-5324. (43) Laurie, S. H., and Pratt, D. E. (1986) A spectroscopic study of nickel(II)-bovine serum albumin binding and reactivity. J. Inorg. Biochem. 28, 431-439. (44) Moffit, W., and Moscovitz, A. (1959) Optical activity in absorbing media. J. Chem. Phys. 30, 648-660. (45) Raycheba, J. M. T., and Margerum, D. W. (1980) Effect of noncoordinative axial blocking on the stability and kinetic behavior of ternary 2,6-lutidine-nickel(II)-oligopeptide complex. Inorg. Chem. 19, 837-843.

Bal et al. (46) Hay, R. W., Hassan, M. M., and You-Quan, C. (1993) Kinetic and thermodynamic studies of the copper(II) and nickel(II) complexes of glycylglycyl-L-histidine. J. Inorg. Biochem. 52, 17-25. (47) Gajda, T., Henry, B., Aubry, A., and Delpuech, J.-J. (1996) Proton and metal ion interactions with glycylglycylhistamine, a serum albumin mimicking pseudopeptide. Inorg. Chem. 35, 586-593. (48) Woody, R. W. (1992) Circular dichroism and conformation of unordered polypeptides. Adv. Biophys. Chem. 2, 37-79. (49) Pettit, L. D., Bal, W., Bataille, M., Cardon, C., Kozlowski, H., Leseine-Delstanche, M., Pyburn, S., and Scozzafava, A. (1991) A thermodynamic and spectroscopic study of the complexes of the undecapeptide Substance P, of its N-terminal fragment and of model pentapeptides containing two prolyl residues with copper ions. J. Chem. Soc., Dalton Trans. 1651-1656. (50) Perczel, A., Park, K., and Fasman, G. D. (1992) Deconvolution of the circular dichroism spectra of proteins: the circular dichroism spectra of the antiparallel β-sheet in proteins. Proteins 13, 57-69.

TX970028X