Copper(II) complexes of some N-substituted bis(aminomethyl)phosphinate ligands. An integrated EPR...

JOURNAL OF

www.elsevier.com/locate/jinorgbio

Journal of Inorganic Biochemistry 98 (2004) 1655–1666

InorganicBiochemistry

Copper(II) complexes of some N-substitutedbis(aminomethyl)phosphinate ligands. An integrated EPR study of

microspeciation and coordination modes by thetwo-dimensional simulation method

Nora Veronika Nagy a, Terezia Szabo-Planka a,*, Gyula Tircso b, Robert Kiraly b,Zsuzsanna Arkosi c, Antal Rockenbauer c, Ern}o Brucher b

a Department of Physical Chemistry, University of Szeged, P.O. Box 105, H-6701 Szeged, Hungaryb Department of Inorganic and Analytical Chemistry, University of Debrecen, P.O. Box 21, H-4010 Debrecen, Hungary

c Chemical Research Center, Institute of Chemistry, Hungarian Academy of Sciences, P.O. Box 17, H-1525 Budapest, Hungary

Received 21 April 2004; received in revised form 30 June 2004; accepted 2 July 2004

Available online 28 August 2004

Abstract

Copper(II) complexes of bis(aminomethyl)phosphinic acid (L1), bis(N-glycino-N-methyl)phosphinic acid (L2), bis(N-benzylgly-

cino-N-methyl)phosphinic acid (L3), bis(LL-prolino-N-methyl)phosphinic acid (L4) and bis(iminodicarboxymethyl-N-methyl)phos-

phinic acid (L5) were studied in aqueous solution by pH-potentiometric and electron paramagnetic resonance (EPR)

spectroscopic methods. The EPR spectrum packages recorded at various ligand-to-metal concentration ratios and pH�s were ana-

lyzed (after matrix rank analysis by the method of residual intensities as a complementary method) by the two-dimensional com-

puter simulation method, which simultaneously determines the formation constants and the EPR parameters of the various

(micro)species. L1 forms mono and bis complexes in different protonation states; for the other ligands, the mono complexes are

always prevalent. For steric reasons, the formation of CuL is shifted to increasingly higher pH regions in the sequence L2, L3

and L4. CuLH was identified for L3, L4 and L5, and also CuLH2 for L4 and L5. Cu2L2 was found in small amounts for L3

and L4, while it predominates at pH > 4 for L5. For L5, Cu2L2H2 was also detected. For the ligands that form dimeric metal com-

plexes in equimolar solution or at a ligand excess, Cu2L is formed at a metal ion excess. Ligation of the phosphinate O was suggested

by indirect proofs in the protonated complexes of L1. For the ligands L2, L3 and L4, the copper(II) coordination in various species

in different protonation states is reminiscent of that in the mono and bis complexes of simple amino acids. For the bis(amino-

methyl)phosphinates, however, the cis positions of the amino groups in CuL are ensured by the structure of the ligand, and the iso-

mers differ from each other in the (equatorial or axial) position of the second carboxylate group.

� 2004 Elsevier Inc. All rights reserved.

Keywords: Electron paramagnetic resonance; Copper(II) complexes; N-Substituted bis(aminomethyl)phosphinic acids; Microspeciation; Coordina-

tion modes

1. Introduction

Our recently developed two-dimensional (2D) elec-tron paramagnetic resonance (EPR) simulation

0162-0134/$ - see front matter � 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.jinorgbio.2004.07.001

* Corresponding author. Tel.: +36-62-546-368; fax: +36-62-544-652.

E-mail address: [email protected] (T. Szabo-Planka).

method [1] offers data on both the speciation and

the coordination modes in paramagnetic equilibrium

systems from the simultaneous analysis of series ofEPR spectra recorded at various metal and ligand

concentrations and pHs. In this approach, the EPR

intensity is considered to be a function of two

independent variables: the magnetic field and the

mailto:[email protected]

L1: R1 = H, R 2 = H

L2: R1 = H, R 2 = CH2COO−

L3: R1 = CH2C6H5, R2 = CH2COO−

L4: R1R2N =

L5: R1 = CH2COO−, R2 = CH2COO−

R1N P P

R1

O−

O

R2 R2

NCO O −

Fig. 1. Structures of the ligands studied.

1656 N.V. Nagy et al. / Journal of Inorganic Biochemistry 98 (2004) 1655–1666

concentration, and in the iteration procedure the for-

mation constants and the individual EPR parameters

for the various species are adjusted simultaneously.

As the EPR parameters rather sensitively reflect the

changes in the first coordination sphere, there is a

good chance of identifying the isomers of variouscomplexes (microspecies) and of detecting oligomeri-

zation.

In consequence of their biological activities, many

aminophosphinate ligands with the phosphinate

groups in terminal positions have been studied. In re-

cent years, a number of c-aminobutyric acidB antago-

nists have been developed, phosphinic acids occupying

a prominent position among them [2–4]. The coordi-nation abilities of the phosphinate and carboxylate

groups have been found to be similar, though the sta-

bility constants of the aminophosphinate complexes

are lower than those of the corresponding aminocarb-

oxylate complexes [5], the phosphinic acid being more

acidic than the carboxylic acid. When a peptide bond

is substituted by the phosphinate moiety, the resulting

molecule furnishes a convenient mimic of a substratein the transition state for some hydrolytic enzymes

[6–9]. Multidentate ligands containing in-chain phos-

phinate groups exhibit biological activity and a strong

coordination ability: they are applied as inhibitors of a

number of bacterial enzymes [6–9], and symmetric

phosphinic acids have been found to be powerful

inhibitors of Human Immunodeficiency Virus protease

[10]; bis(LL-prolino-N-methyl)phosphinic acid andbis(iminodicarboxymethyl-N-methyl)phosphinic acid,

ligands studied in the present paper, have been shown

to have plant growth-regulating properties [11–14].

For ligands of the latter type, essentially amino acid-

like behavior is expected because of the aminocarb-

oxylate donor groups. One interesting question

that arises is the role of the phosphinate group in

the complexation.We report here on an EPR study of the copper(II)

complexes of five N-substituted bis(aminomethyl)phos-

phinic acids, formed in aqueous solution in the pH range

2–10. We set out to characterize all detectable

(micro)species, by identifying their coordinating donor

groups, to seek information on the coordination ability

of the phosphinate moiety, and to clarify the roles of

various coordinating and noncoordinating substituentsin the complexation process. As the 2D spectrum

decomposition demands a knowledge of the formation

constants of the proton complexes, and also good start-

ing values for the parameters of the metal complexes,

the above systems were additionally studied by

pH-potentiometry, and the number of independent

EPR-active species also was checked by a model-free,

purely mathematical analysis of spectrum packages:matrix rank analysis (MRA) by the method of residual

intensities [15].

2. Experimental

2.1. Reagents and solutions

The following ligands (denoted by L in their deprot-

onated forms) were synthesized via Mannich-typereactions: bis(aminomethyl)phosphinic acid (L1),

bis(N-glycino-N-methyl)phosphinic acid (L2), bis

(N-benzylglycino-N-methyl)phosphinic acid (L3), bis(LL-

prolino-N-methyl)phosphinic acid (L4) and bis(iminodi-

carboxymethyl-N-methyl)phosphinic acid (L5). The

solvents and reagents used in the preparative work were

commercially available chemicals (Aldrich, Sigma, Flu-

ka or Spektrum 3D), and were used without furtherpurification. The purity of the ligands was controlled

by 1H, 13C and 31P NMR spectroscopy. The concentra-

tions of the stock solutions of the ligands were deter-

mined pH-potentiometrically. The details of the

syntheses of L4 and L5 were published earlier [16]. L1

was synthesized by the debenzylation of bis[(dibenzyl-

amino)methyl]phosphinic acid in aqueous acetic acid

[14]. The physical data on the product are in good agree-ment with the data published by Maier [13]. L2 and L3

were prepared by a modification of the procedure re-

ported by Dhansay et al. [11]. The structures of the lig-

ands are shown in Fig. 1. (For simplicity, the charges on

the various complexes are neglected throughout the text

and in the tables.)

2.2. pH-potentiometric measurements

The titrations were carried out at 298.0 K, at an ionic

strength I = 1.0 M tetramethylammonium chloride, un-

der an argon atmosphere. The ligand concentrations

were 2 or 4 mM, depending on the ligand-to-metal con-

centration ratio (which was 1:1, 2:1, 3:1 or 1:2). 0.20 M

tetramethylammonium hydroxide solution was used as

titrant. A Radiometer pHM 85 precision pH-meterequipped with a Metrohm 6.0234.100 combined glass

electrode and an ABU 80 autoburet were utilized for

N.V. Nagy et al. / Journal of Inorganic Biochemistry 98 (2004) 1655–1666 1657

the measurements. The protonation constants of the lig-

ands and the stability constants of the complexes were

computed with the program PSEQUAD [17]. The liq-

uid-junction potential of the electrode was taken into

consideration via the method proposed by Irving et al.

[18].

2.3. EPR Measurements

The copper(II) concentration was 5 mM, while the

ligand-to-metal concentration ratios were 1:2, 1:1 and

5:1, except for the case of L4, where the ligand excess

was 10-fold. The initial pH was adjusted with 0.2 M

HCl solution. 0.2 M NaOH solution was then addedfrom a Metrohm 765 Dosimat apparatus to the 10-cm3

sample to adjust the chosen pH, which was measured

to an accuracy of 0.01 pH unit with a Radiometer

PHN 240 potentiometer equipped with a Radiometer

GK2401C combined glass electrode. The electrode was

calibrated with IUPAC Standard Buffers from Radiom-

eter. During the titration, the sample was mixed by bub-

bling gaseous argon through the solution. A MasterflexCL peristaltic pump ensured the circulation (14 cm3

min�1) of the solution through the capillary tube in

the cavity. The EPR spectra were recorded after circula-

tion for 2 min at room temperature (291 K) on an up-

graded JEOL JES-FE3X spectrometer with 100-kHz

field modulation, using manganese(II)-doped MgO

powder for the calibration of g. All the details of the

measurements are given in a previous paper [19].For each of the various systems, 20–32 EPR spectra

were recorded. At a 2-fold metal excess, precipitation

occurred at about pH 6 in each system. At equal metal

and ligand concentrations, the pH of the appearance

of the precipitate was shifted to 6.5, 8.5 and 10.7 for

L1, L2 and L3, respectively. For L4, hydrolysis occurred

at pH > 10. In the case of L5, precipitation was observed

at ligand-to-metal concentration ratios of both 1:1 and5:1 (at pH 8.5 and 10.6, respectively).

2.4. Preparation of spectra for evaluation

The spectral analysis was preceded by elimination of

the background signal, which contained the vanadium

signal of the Pyrex tube and the Mn peaks of the

Mn:MgO external standard. To compensate the minorfrequency shifts in the course of measurements, the field

scale was shifted to secure a perfect fit for the Mn lines

of the external standard.

2.5. Matrix rank analysis

The number of independent EPR-active species was

determined by calculating the residual intensity curveswith the program MRA [15].

2.6. Decomposition of spectra

The series of spectra were evaluated with the program

2D_EPR [1]. Each component curve was described by

the parameters go, the copper and ligand hyperfine cou-

pling constants Ao and aNo, and the relaxation parame-ters a, b and c, which are related to the widths of the

copper lines as rMI ¼ aþ bMI þ cM2I (MI is the mag-

netic quantum number of copper nuclei). Since the cop-

per(II) (as chloride) used to make the stock solution was

a natural mixture of the isotopes, the spectrum of each

species was calculated as the sum of spectra containing63Cu and 65Cu weighted by their abundances in nature.

The copper and ligand coupling constants and the relax-ation parameters with field dimensions are given in units

of gauss (G) throughout the paper; 1 G = 10�4 T.

For each spectrum, the computer program provides

the noise-corrected regression parameter (Rj for the jth

spectrum) derived from the average square deviation

(SQD) between the experimental and the calculated

spectrum. For the series of spectra, the fit is character-

ized by the overall regression coefficient R, calculatedfrom the overall average SQD (the sum of the SQD val-

ues for the various curves in the series). To compare

alternative models, the program also furnishes the criti-

cal value of the difference in overall regression coeffi-

cients (DRcrit), which shows whether the species in

question, included in or omitted from the model, causes

a significant change in the spectral fit or not. The details

of the statistical analysis were published previously [1].

3. Results and discussion

3.1. Equilibrium models in various systems

Matrix rank analysis [15], a model-free mathematical

analysis of the spectrum package, can help to build upthe equilibrium model, as it furnishes the number of

the independent EPR-active species represented in the

spectra. We illustrate the results of MRA in the case

of the copper(II)–L1 system, for which the series of

experimental EPR spectra is depicted in Fig. 2. Fig. 3

shows the gradual reduction of the residual intensities

at various fields, obtained for the above spectrum pack-

age by MRA, as more and more EPR-active species areassumed. The residual intensity reaches the value of

noise at 6 independent complexes. (The disappearance

of the last small peak near 3000 G can be better seen

in the enlarged part (b) of Fig. 3.) MRA resulted in 2,

4, 6 and 5 independent species for L2, L3, L4 and L5,

respectively. This can be a good starting point for the

subsequent 2D-analysis, though MRA cannot distin-

guish those species which are present in constant con-centration ratio in all spectra (e.g., isomers at constant

(a)

(b)

Fig. 2. Series of EPR spectra for the copper(II)–L1 system; (a)

TCu = TL = 5 mM, (b) TCu = 5 mM, TL = 25 mM.

(a)

(b)

Fig. 3. Results of matrix rank analysis on the overall series of spectra

for the copper(II)–L1 system; (a) residual intensity curves for different

numbers of EPR-active species, and (b) residual intensity curves

enlarged for 3–7 species near 3000 G.


temperature, or monomer-dimer pairs at constant metal

ion concentration and temperature).

For the copper(II)–L1 equilibrium system, MRA

indicated more independent species than the pH-

potentiometric model. Accordingly, CuL2H2 was in-cluded in addition to Cu, CuLH, CuL, CuL2H and

CuL2, and we then also obtained a satisfactory spectral

fit in the moderately acidic region at a ligand excess.

Moreover, the spectral fit improved significantly when

each of the spectra of CuL2H2 and CuL2 was described

as the sum of two component curves which can be as-

signed to two different structures (isomers or microspe-

cies). The formation constants obtained by EPRanalysis for the various metal complexes are given in Ta-

ble 1, together with the pH-potentiometric formation

constants for the proton and metal complexes. The cor-

responding constants obtained by the two different

methods agree well: their differences do not exceed the

usual differences in the literature data when the same

systems are studied under similar conditions by different

researchers. The concentration distribution curves cal-culated from the formation constants determined by

2D-EPR simulation are depicted in Fig. 4.

Though L2, L3 and L4 possess the same donor

groups, their coordination properties are quite different.

The most striking feature of the copper(II)–L2 system is

the extremely high stability of CuL: this molecule al-

ready predominates at pH 1.5. Its EPR spectrum can

be described as the sum of two component curves, sug-

gesting the coexistence of two microspecies. For L3 and

L4, the protonated complex CuLH too can be demon-

strated, as the formation of CuL is shifted to considera-bly higher pH (the deprotonation pK of CuLH is 2.44

and 3.10 for L3 and L4, respectively). Moreover, for

L4, the complexes CuLH2 and CuL2H3 were also neces-

sary for a satisfactory description of the corresponding

spectrum package. The 2D-simulation allowed a signifi-

cant refinement of the equilibrium model at pH > 5 as

well: we had to assume two isomeric species for CuL,

and the spectral fit further improved when the dimercomplex Cu2L2 was also taken into consideration. The

maximum concentrations of the dimer for the ligands

L3 and L4 were 3% and 15%, respectively. The corre-

sponding formation constants obtained by pH-potenti-

ometry or EPR likewise agree well for these systems

(Table 1).

The large differences in complexation for the above

ligands do not correlate with the differences in basicities

Table 1

Formation constants as log ba for the proton and copper(II) complexes of the various bis(aminomethyl)phosphinate ligands

Complex Ligand

L1 L2 L3 L4 L5

EPRb pH-pot. EPRb pH-pot. EPRb pH-pot. EPRb pH-pot.c EPRb pH-pot.

LH 9.05(2) 9.00(1) 9.22(2) 10.92(2) 10.01(1)

LH2 16.50(2) 15.53(2) 14.78(3) 18.27(3) 16.51(1)

LH3 18.06(2) 16.91(6) 20.27(6) 19.19(2)

LH4 19.72(3) 18.3(2) 21.57(6) 21.36(2)

LH5 22.74(6)

Cu2L 16.13(6) 16.25(9) 18.32(4) 23.03(8) 23.0(1)

CuLH2 20.30(2) 23.13(4)

CuLH 13.65(2) 13.6(1) 15.94(1) 15.92(6) 18.50(1) 18.44(9) 21.38(3) 22.1(2)

Cu2L2H2 45.24(6)

CuL 8.48(1) 8.30(6) 17.30d 17.4(2) 13.50d 13.44(6) 15.40d 15.36(9) 17.5(1) 19.3(2)

Isomer 1 17.28(1) 13.40(1) 15.28(1)

Isomer 2 16.62(2) 12.82(1) 14.78(2)

Cu2L2 27.62(7) 32.19(1) 40.0(1)

CuL2H3 38.95(1)

CuL2H2 26.86d

Isomer 1 26.66(7)

Isomer 2 26.42(8)

CuL2H 21.69(1) 21.53(7)

CuL2 14.88d 14.65(7)

Isomer 1 14.72(1)

Isomer 2 14.37(1)

a The confidence intervals (3r) of the last digit at a significance level of 99.7% are given in parentheses.b For the proton complexes, the pH-potentiometric formation constants were used in the EPR analysis, too.c Data taken from [14] (I = 1.0 M KNO3).d log b = log (bisomer1 + bisomer2).

(a)

(b)

Fig. 4. Concentration distribution in the copper(II)–L1 system,

calculated from the EPR spectroscopic formation constants at (a)

TCu = TL = 5 mM, and (b) TCu = 5 mM, TL = 25 mM.


of their donor groups. The pK values of 9.00, 6.53 and

9.22, 5.56 for the N atoms of L2 and L3, respectively

(Table 1), reflect moderate basicities, while for L4, the

donor groups are much more basic (with pK values of10.92 and 7.35, Table 1). At the same time, an extremely

high stability of CuL can be observed for L2, while for

L3, and particularly for L4, this complex is formed at

much higher pH (see above), pointing to significant

steric hindrance caused by the noncoordinating

substituents (L3) or the rigid structure: for the proline

derivative (L4), this effect suppresses even the fairly high

basicities of the donors.The presence of the additional carboxymethyl substit-

uents on the N atoms in L5 considerably enhances the

tendency of the complexes to dimerize: the monomer–di-

mer pairs CuLH and Cu2L2H2 could be identified in the

strongly acidic region, as could CuL and Cu2L2 at

pH > 4; in the latter case, the dimer became prevalent.

Here, larger differences can be observed between the

EPR and pH-potentiometric formation constants (Table1), which may be attributed to the fact that dimer for-

mation was not considered in the latter method. The dis-

tribution of copper(II) amongst the various complexes

at a 5-fold ligand excess is shown in Fig. 5, and that

at a 2-fold metal excess in Fig. 6. It is noteworthy that,

in parallel with the increasing tendency of CuL to form

Table 2

Ratio of DR to DRcritical in the various copper(II)-bis(amino-

methyl)phosphinate systems for different equilibrium models

Neglected species Ligand

L1 L2 L3 L4 L5

[CuLH2]a 46.7 9.2

[Cu2L2H2]a 4.9

[CuL]a 1.5

[CuL]b 10.0 42.9 44.9

[Cu2L2]a 8.9 534.9

[CuL2H3]a 698.5

[CuL2H2]b 3.4

[CuL2H2]a 20.7

[Cu2L]a 122.6 28.3 808.7

a The given species was not taken into consideration at all.b One or other of the isomers of the given species was neglected.

(a)

(b)

(c)

(d)

Fig. 5. Concentration distribution calculated from the EPR spectro-

scopic formation constants at a 5-fold ligand excess in the (a)

copper(II)–L2, (b) copper(II)–L3, (c) copper(II)–L4 and (d) cop-

per(II)–L5 systems.

(a)

(b)

(c)

Fig. 6. Concentration distribution calculated from the EPR spectro-

scopic formation constants at a 2-fold metal ion excess in the (a)

copper(II)–L3, (b) copper(II)–L4 and (c) copper(II)–L5 systems.


dimers (Fig. 5), the maximum concentration of the binu-

clear Cu2L increases in the sequence L3, L4, L5 (Fig. 6).

3.2. Verification of the equilibrium models

The best model for each system results in a good

spectral fit: the overall regression coefficient R for each

system lies between 0.99017 and 0.99528. When any,

even minor species involved in the best model is ne-

glected, the decrease in R significantly exceeds the criti-

cal value (Table 2). (The only exception is the complex

CuL of L5, where the decrease in R is only slightly high-

er than the critical value.) Fig. 7 illustrates the deteriora-

(a)

(b)

(c)

(d)

Fig. 7. Experimental and calculated EPR spectra in the copper(II)–L4

system for the best model (upper spectra) and for models when one or

other of the complexes was neglected (lower spectra), near the

maximum concentration of the species in question (a) [CuLH2], pH

1.76; (b) [CuL]�, pH 9.00; (c) [CuL2H3]�, pH 2.89; and (d) [Cu2L]

+, pH

2.96. The metal and ligand concentrations were: (a)–(c) Tcu = 5 mM,

TL = 50 mM; (d) Tcu = TL = 5 mM.


tion in the fit for some individual spectra in the cop-

per(II)–L4 system when species featuring in the best

model are neglected.

3.3. Confidence levels of parameters

The EPR parameters for the various complexes are

presented in Table 3, while the component spectra calcu-lated from them are depicted in Fig. 8. The low confi-

dence level intervals (3r at a significance level of

99.7%) reveal the good sensitivity of the fit to the param-

eters go and Ao. The higher 3r values for Cu2L,

Cu2L2H2 and Cu2L2 can be attributed to the broad,

poorly-resolved spectra.

3.4. EPR parameters and coordination modes

When stronger ligand field donors replace the weak O

atoms of water molecules in the equatorial coordination

around the copper(II) ion, the increase in the energy of

the dxy dx2�y2 electronic transition decreases gi, andthereby go (go = (gi + 2g^)/3) (g^ is affected to a much

lower extent) [20]. The go shift is significantly higher

for an equatorial amino N (about 0.030 for the first,

and 0.015–0.020 for the subsequent ones) than for a

carboxylate O (approximately 0.005–0.01) [21]. Axial

coordination of any donor group to copper(II) in the

dx2�y2 ground state has only minor effects on go and

Ao. Accordingly, the go values inform reliably on thenumber of amino N atoms in equatorial positions, and

the equatorial ligation of carboxylate O atoms also

can be detected by the comparison of go for the various

species.

The decrease in go is accompanied by an increase in

Ao for copper(II) complexes with effective D4h symmetry

[22–24]. An opposite trend is indicative of rhombic dis-

tortion [22–26]. There was a trend for the linewidths todecrease together with go, i.e., we obtained better-re-

solved spectra for species with stronger ligand fields (a

lower go reflects a stronger ligand field; see above). This

trend can be explained by the fact that the major relax-

ation parameter a is proportional to the square of the

anisotropy of the g tensor, and the value of go generally

(though not necessarily) varies in parallel with the aniso-

tropic contributions of the Zeeman interaction. Unusu-ally broad lines for complexes with strong ligand fields

suggest an enhanced magnetic dipole interaction be-

tween the paramagnetic centers as a result of oligomeri-

zation [22–26]. Based on the comparison of go, Ao, and

linewidth data, with regard to the above considerations,

we suggest the following coordination modes for the

various species.

3.5. The copper(II)–L1 system

The gradual decrease in go (Table 3) demonstrates

that the stepwise proton loss from the mono and bis

complexes is accompanied by the equatorial binding of

the corresponding amino group, in accordance with

expectations (except isomer 2 of CuL2). The question

arises of whether the phosphinate group takes part inthe coordination or not. Though below pH 3 no phosph-

inate binding could be detected by EPR, it likely occurs

in strongly acidic solutions: Kubicek et al. [27] could

crystallize the complex all-trans-[CuCl2(H2L)2-

(H2O)2]Cl2 in the range pH 2–3, where the amino groups

are protonated, and the crystallographic studies [27]

showed monodentate phosphinate coordination. Proba-

bly, the ligation of this weak donor induces only negligi-ble changes in the ligand field, and consequently in the

magnetic parameters of the aqua complex, and thus

Table 3

EPR parameters for the copper(II) complexes of the various bis(aminomethyl)phosphinate ligandsa

Complex Parameters Ligand

L1 L2 L3 L4 L5

Cu2+ go 2.1967(2) 2.1956(3) 2.1968(9) 2.1954(5) 2.196(3)

Ao/G 34.3(3) 34.8(1) 34(3) 35.2(6) 36(1)

[Cu2L] go 2.149(3) 2.153(3) 2.1486(5)

Ao/G 34(3) 36(2) 36.4(5)

[CuLH2] go 2.1911(9) 2.1421(6)

Ao/G 37.8(7) 49.7(6)

aN/G 11(1)

[CuLH] go 2.1646(5) 2.1457(4) 2.1496(2) 2.1468(4)

Ao/G 44.9(5) 59.6(4) 56.9(2) 47.9(4)

aN/G 6(1) 10.9(6) 11.7(3) 9.8(6)

[Cu2L2H2] go 2.138(1)

Ao/G 48(1)

[CuL] isomer1 go 2.1463(3) 2.1249(1) 2.1228(1) 2.1206(1) 2.125(6)

Ao/G 60.2(3) 73.1(3) 73.4(1) 76.8(4) 40(3)

aN/G 10.7(6), 11.9(5) 10.5(1), 10.5(1) 9.6(1), 9.6(1) 10.9(1), 10.9(1) 11(3), 11(3)

Isomer 2 go 2.1199(1) 2.1095(4) 2.1178(1)

Ao/G 71.9(2) 73.3(3) 71.2(1)

aN/G 10.6(2), 10.6(2) 6.0(4), 6.0(4) 10.7(1), 10.7(1)

[Cu2L2] go 2.153(1) 2.1508(7) 2.1463(9)

Ao/G 31(1) 35.4(6) 41.9(8)

[CuL2H3] go 2.1500(2)

Ao/G 56.3(1)

aN/G 10.9(2), 10.9(2)

[CuL2H2] isomer 1 go 2.1398(7)

Ao/G 39(2)

aN/G 10(5), 8(5)

Isomer 2 go 2.1430(6)

Ao/G 62(2)

aN/G 8(2), 8(2)

[CuL2H] go 2.1249(2)

Ao/G 64.4(2)

aN/G 10.2(4),

9.2(6),

9.2(6)

[CuL2] isomer 1 go 2.1122(1)

Ao/G 68.8(2)

aN/G 9.0(2), 9.0(2)

8.2(2), 8.2(2),

Isomer 2 go 2.1246(2)

Ao/G 69.2(2)

aN/G 9.2(2), 9.2(2),

8.2(2)

a The confidence intervals of the last digit (3r) at a significance level of 99.7% are given in parentheses.


we cannot distinguish the corresponding spectra. A re-

lated ligand provides a further example of weak ligation

of the in-chain phosphinate group: bis(morpholinato-N-

methyl)phosphinic acid forms a complex of very low

stability with copper(II) (log KCuL = 1.6) in which

monodentate phosphinate binding is assumed, since

the ligation of the N atoms is strongly hindered [28].

At above pH 3, two experimental facts seem to suggest

aminophosphinate-type coordination for the monoprot-

onated ligand L1 (Fig. 9). First, the go values for the iso-

Fig. 8. EPR spectra calculated from the data in Table 3 at 9.4 GHz in the copper(II)–L1, copper(II)–L2, copper(II)–L3, copper(II)–L4 and

copper(II)–L5 systems.


mers of CuL2H2 are slightly lower than the go for CuL,

though the go values reflect on two amino groups in

equatorial positions for all three species. The amino

groups can form a chelate ring with the metal ion in

CuL, which lends enhanced stability [21] to this mole-

cule as compared with the monodentate coordination

of two amino groups, and this chelate-type binding

mode would result in a lower go of CuL if the phosphi-nate groups were not bound to copper(II) in CuL2H2. In

contrast, the lower go values for CuL2H2 indicate

bis(aminophosphinate) coordination with two chelate

rings in the latter case. Second, isomerism could be dem-

onstrated for CuL2H2 on the assumption that the proc-

esses of interconversion of the various species are slow

on the EPR time-scale. This condition is readily fulfilled

when the motion of the donor groups in question isinhibited by their participation in chelate rings, while

the faster rearrangement of the monodentate ligands

would result in an averaged spectrum instead of the

superimposed component spectra.

For CuL2H2, identical equatorial donor atom sets,

and hence cis–trans isomerism, are suggested by the

nearly identical go values of the isomers (Table 3). The

low Ao for the major species indicates rhombic distor-tion, and therefore we assign the corresponding set of

EPR parameters to the cis isomer of lower symmetry

(Fig. 9). This type of isomerism is typical for a number

of copper(II)–a-amino acid [29,30] and copper(II)–b-amino acid [31] systems. However, the go values are

higher for CuL2H2 of bis(aminophosphinate)-type coor-

dination than for the amino acid complexes involving

bis(aminocarboxylate) coordination.

In the case of CuL2, the go values for the isomers dif-

fer significantly. The minor isomer has a go as large as

that of CuL2H with 3N atoms coordinated equatorially.

In this isomer, one of the ligands is probably coordi-

nated by both N atoms in equatorial positions, whilethe other ligand occupies one equatorial and one axial

position (Fig. 9).

3.6. The copper(II)–L2, copper(II)–L3 and copper(II)–

L4 systems

These ligands have a carboxymethyl and a noncoor-

dinating substituent in each of their terminal parts.The differences in the nature of the noncoordinating

substituent are manifested in significantly different coor-

dinating abilities, as discussed in detail in the previous

section, whereas the nearly identical EPR parameters

suggest similar structures for the analogous complexes

(Table 3). For these ligands, competition may occur be-

tween aminocarboxylate-type and aminophosphinate-

like ligation, due to the carboxymethyl substituents onthe N atoms. The go values for CuLH of L3 and L4,

and for CuL of L2, L3 and L4, are lower than for the

corresponding complexes of L1 (Table 3). At the same

time, these parameters are nearly identical to the go val-

ues of the analogous amino acid complexes formed by

[CuLH]2+ [CuL]+

[CuL2H2]2+

[CuL2]

NH 2

P

NH 2

Cu2+O −

O

NH 2

P

NH 2

O −

O

NH2

P

NH2

Cu2+O−

O

NH2

P

NH2

O−

O

OH2

NH2

H2O Cu2+

OH2

P

O−

O

NH3+

NH2

OO−

PNH2

H2O Cu2+

OH2

NH2

PO−

O

NH3+

NH3+

O

O−

P

Cu2+

NH2

NH2

Cu2+

P

O-

O

NH3+

NH3+

O

O−

P NH 2

[CuL2H]+

NH2O

O−

P

NH2

NH3+

O

O−

P

Cu2+

NH2

isomer 1 isomer 2

isomer 1 isomer 2

Fig. 9. Coordination modes proposed for the various complexes in the

copper(II)–L1 system.

O

O

O−

NH +O

O−

N

Cu2+

P

O−

OH2

OH 2

C O−

O−

N

H2C

O

P CH2

Cu2+

O−O

C

N

OC

O−

O−

N

H2C

O

P CH2

Cu2+

O−O

C

N

O

NPO−O

N

O

O−

Cu2+

Cu2+

O

O−

O−

O

O−

O

N

OO-

PN

OH2 H2O

[CuLH]

isomer 1 isomer 2[CuL]−

[Cu2L2]6−

Fig. 10. Coordination modes proposed for some complexes in the



mono- and bis(aminocarboxylate) coordination [29–31].

These facts indicate that it is the carboxylate groups that

form chelate rings, together with the amino groups in-

stead of the phosphinate moiety in the cases of L2, L3

and L4. As bis(aminocarboxylate) coordination can oc-

cur through the tetradentate binding of the ligand in

CuL, this excludes the cis–trans isomerism characteristic

of bis(amino acid) complexes [29–31]. L4 is probablybound equatorially by four donor atoms in the minor

isomer (with lower go), while in the major isomer (with

higher go) one of the carboxylate groups occupies an ax-

ial position. For L3 and L4, the spectrum of a third

complex in the range of formation of CuL was assigned

to the dimer Cu2L2, its broad bands suggesting dipole

interactions between the metal ions. The relatively high

go (Table 3) corresponds to the equatorial coordinationof one amino group (and one or two carboxylate

groups), while the very low Ao points to a distorted

structure. Overall, therefore, we assume that one half

of the first ligand is bound in equatorial positions to

the first copper(II) via the amino and carboxylate

groups, while it occupies an equatorial and an axial

position around the second metal ion via the carboxy-

late and the amino groups of its second half. The other

ligand is coordinated diequatorially to the second cop-

per(II) ion through the amino and carboxylate groups,

while it is bound in equatorial and axial positions to

the first metal ion by the carboxylate and amino groups

from the other end. The coordination modes proposedfor the complexes L2, L3 and L4 are depicted on the

example of L4 in Fig. 10.

For L4, CuLH2 too was identified. Its high go corre-

sponds to the ligation of an O atom-containing group

(either carboxylate or phosphinate, the data not allow-

ing a distinction). The coordination around the cop-

per(II) ions in Cu2L is most likely of

aminocarboxylate type, as suggested by the agreementof the go values of Cu2L and CuLH (Table 3). In the

copper(II)–L4 system, a further protonated complex,

CuL2H3, could be detected. Here, the aminocarboxy-

late-type binding of the monoprotonated ligand and

the coordination of the diprotonated ligand via an O do-

nor are most probable.

3.7. The copper(II)–L5 system

The close similarity of the EPR parameters of CuLH2

for L5 to those of CuLH of L3 and L4 (Table 3) points

to aminocarboxylate-like equatorial coordination in the


former species as well. Its lower go and Ao, however,

indicate a somewhat stronger and more distorted ligand

field. This suggests the equatorial binding of the amino

and both carboxylate groups (Fig. 11). The following

deprotonation step may affect either the protonated

amino or the carboxylic group of the other half of theligand. As compared with CuLH2, the small increase

(0.0047) in go for CuLH (Table 3) reveals similar equa-

torial coordination and an additional strong axial bond

in the latter complex. This bond is most probably

formed by the amino N atom from the other end of

the ligand, because it is a strong donor, and a six-mem-

bered chelate ring can be formed. (A nine-membered

chelate ring with participation of the weak carboxylategroup from the other end of the ligand would be less

favorable.) It can be concluded from go that the latter

carboxylate group, probably for steric reasons, remains

noncoordinated (its equatorial binding would reduce go)

(Fig. 11). For the dimer Cu2L2H2, the somewhat lower

go indicates a slightly stronger ligand field than in CuLH

(Table 3), suggesting that the fourth equatorial position

here is also occupied by a carboxylate group. Both lig-ands are most likely to be bound to both metal ions:

the amino and two carboxylate donors from the first

half of the chain occupy three equatorial positions

around the first metal ion, while the amino N and carb-

oxylate O atoms from the other half of the chain are

[CuLH]2− [Cu2L2H2]4−

[CuL]3− [Cu2L2]6−

[CuLH2]-

O

O−Cu2+

O−

O

NN H+

O

O−

OH

O

P

O−

O

H2O

N

OO−

Cu2+

O−

O

P

O

N

O−

OO−

O

OH

H2O

PO O−

O−O

OO−

N

OO-

Cu2+

O− O

NOO−

O

O−

NP

O−

O

N

O

O−

O−

O−

O

Cu2+

Cu2+O−

O

NO

O-

PN

O−O

O−

O

O

O

HO

O

O−O

NP

O−O

N

O

O−

O

O−

O−

O

Cu2+

Cu2+O−O

N

OO−

PN

O−

O

OH

Fig. 11. Coordination modes proposed for some complexes in the


bound in axial and equatorial positions, respectively,

to the second copper(II) ion, and vice versa (Fig. 11).

For CuL of L5, go is nearly as low as for L2, L3 and

L4, pointing to two N and two O atoms in the equato-

rial positions. At the same time, the value of Ao is much

lower than for the analogous complexes of the other lig-ands (Table 3). This suggests a strong rhombic distor-

tion, which can be attributed to the stress caused in

the structure by the axial ligation of two additional carb-

oxymethyl groups (Fig. 11). The difference in go between

Cu2L2H2 and Cu2L2 (Table 3) indicates that the proton

loss from the last carboxylate groups induces rearrange-

ment of the donor atoms and the formation of addi-

tional strong axial bonds in the latter complex. All theamino and carboxylate groups can undergo coordina-

tion if both ends of the ligands form one axial and

two equatorial bonds: the first half is ligated by the

amino group and one carboxylate group in equatorial

positions, while its other carboxylate group occupies

an axial site around the same copper(II). The second

half is bound to the other metal ion by two carboxylate

groups in equatorial positions and the amino group inan axial position (Fig. 11).

In the systems studied, we did not find any direct

evidence for the binding of the phosphinate group to

the metal ion. For L1 containing one phosphinate

and two amino donors, indirect proofs suggest that

in moderately acidic media the non-protonated amino

group and the phosphinate moiety act as chelating do-

nors. In neutral and alkaline solutions, however, thesecond amino group replaces the phosphinate group.

For the complexes of the ligands with carboxymethyl

substituents, the EPR parameters support that the

aminocarboxylate-type coordination dominates over

the aminophosphinate and diamine-like binding modes.

The most likely explanation for the low coordinating

ability of the phosphinate group in the ligands studied

is its very low basicity. More basic phosphinate groupscan act similarly to carboxylate donors: e.g., it was

found for some dipeptide-analogue ligands with phos-

phinic groups in terminal positions that the phosphi-

nate group stabilized the tridentate coordination of

the ligand, just as carboxylate group does in dipeptides,

and this effect varied parallel with its basicity [32].

4. Conclusions

In the event of nonsubstituted terminal amino groups

(ligand L1), various mono and bis complexes are

formed, where the deprotonated ligand is bound in a

diamin-like way, while aminophosphinate-type coordi-

nation probably occurs for the ligand with one amino

group protonated. When each amino group bears acarboxymethyl substituent, mono complexes and the

aminocarboxylate binding mode are favored (ligands


L2, L3 and L4). Simultaneous coordination of all termi-

nal donors appreciably enhances the stability of the

complex CuL, particularly for L2. At the same time,

the steric effect of the noncoordinating substituents on

the N atoms (L3), or the rigidity of the ligand with the

amino group built into a ring (L4), in part counterbal-ance the former effect: at a ligand excess dimerization

occurs, while at a metal excess the binuclear complex

Cu2L is formed in significant amounts. Two carboxy-

methyl substituents on each N atom (L5) make triden-

tate coordination of the terminal parts favored. This

binding mode is observed mainly in the binuclear and di-

mer species, and is accompanied by considerable distor-

tion of the complexes.

Acknowledgement

We gratefully acknowledge the financial support of

the Hungarian Scientific Research Fund OTKA (Grants

T-032929, T-038364 and T-046953). We are indebted

to Dr. David Durham for stylistic correction of themanuscript.

Appendix A. Supplementary material

The full set of EPR parameters for all species, includ-

ing relaxation parameters. This material is available via

the Internet at http://www.staff.u-szeged.hu/~szabot/su-

pinf.html. Supplementary data associated with this arti-

cle can be found, in the online version, at doi:10.1016/

j.jinorgbio.2004.07.001.

References

[1] A. Rockenbauer, T. Szabo-Planka, Zs. Arkosi, L. Korecz, J. Am.

Chem. Soc. 123 (2001) 7646–7654.

[2] J. Kehler, B. Ebert, O. Dahl, P. Krogsogaard-Larsen, Tetrahe-

dron 55 (1999) 771–780.

[3] M. Chebib, G.A.R. Johnston, J. Med. Chem. 43 (2000) 1427–

1447.

[4] G.A.R. Johnston, Curr. Top. Med. Chem. 2 (2002) 903–913.

[5] P. Hermann, I. Lukes, P. Vojtisek, I. Cisarova, J. Chem. Soc.

Dalton Trans. (1995) 2611–2618.

[6] M. Collinsova, I. Jiracek, Curr. Med. Chem. 7 (2000) 629–647.

[7] I.A. Natchev, Liebigs Ann. Chem. (1988) 861–867.

[8] M.A. Hussain, M.S.L. Lim, K.S. Raghavan, N.J. Rogers, R.

Hidalgo, C.A. Kettner, Pharm. Res. 9 (1992) 626–628.

[9] N. Valiaeva, D. Bartley, T. Konno, J.K. Coward, J. Org. Chem.

66 (2001) 5146–5154.

[10] A. Peyman, K. Budt, J. Spanig, B. Stowasser, D. Ruppert,

Tetrahedron Lett. 33 (1992) 4549–4552.

[11] M.A. Dhansay, P.W. Linder, R.G. Torrington, T.A. Modro, J.

Phys. Org. Chem. 3 (1990) 248–254.

[12] M.A. Dhansay, P.W. Linder, J. Coord. Chem. 28 (1993) 133–145.

[13] L. Maier, J. Organomet. Chem. 178 (1979) 157–169.

[14] Gy. Tircso, A. Cs. Benyei, E. Brucher, I. Lazar, R. Kiraly, R. Pal,

in preparation.

[15] G. Peintler, I. Nagypal, A. Jancso, I.R. Epstein, K. Kustin, J.

Phys. Chem. A 101 (1997) 8013–8020.

[16] T. Varga, Synth. Commun. 27 (1997) 2899–2902.

[17] L. Zekany, I. Nagypal, in: D.J. Legett (Ed.), Computational

Methods for Determination of Formation Constants, Plenum,

New York, 1985, p. 291.

[18] H.M. Irving, M.G. Miles, L. Pettit, Anal. Chim. Acta 38 (1967)

475–488.

[19] T. Szabo-Planka, A. Rockenbauer, L. Korecz, Magn. Reson.

Chem. 37 (1999) 484–492.

[20] T. Szabo-Planka, G. Peintler, A. Rockenbauer, M. Gyor, M.

Varga-Fabian, L. Institorisz, L. Balazspiri, J. Chem. Soc. Dalton

Trans. (1989) 1925–1932.

[21] N.V. Nagy, T. Szabo-Planka, A. Rockenbauer, G. Peintler, I.

Nagypal, L. Korecz, J. Am. Chem. Soc. 125 (2003) 5227–5235.

[22] A.H. Maki, B.R. McGarvey, J. Chem. Phys. 29 (1958) 31–34.

[23] D. Kivelson, R. Neiman, J. Chem. Phys. 35 (1961) 149–157.

[24] B.A. Goodman, J.B. Raynor, Adv. Inorg. Chem. Radiochem. 13

(1970) 135–362.

[25] J. Peisach, W.E. Blumberg, Arch. Biochem. Biophys. 165 (1974)

691–708.

[26] A. Rockenbauer, J. Magn. Reson. 35 (1979) 429–438.

[27] V. Kubicek, P. Vojtisek, V. Rudovsky, P. Hermann, I. Lukes, J.

Chem. Soc. Dalton Trans. (2003) 3927–3938.

[28] T.R. Varga, R. Kiraly, E. Brucher, V. Hietapelto, Models Chem.

(1999) 431–439.

[29] B.A. Goodman, D.B. McPhail, J. Chem. Soc. Dalton Trans.

(1985) 1717–1723.

[30] T. Szabo-Planka, A. Rockenbauer, L. Korecz, Polyhedron 18

(1999) 1969–1974.

[31] Zs. Arkosi, T. Szabo-Planka, A. Rockenbauer, N.V. Nagy, L.

Lazar, F. Fulop, Inorg. Chem. 42 (2003) 4842–4848.

[32] M. Lukas, M. Kyvala, P. Hermann, I. Lukes, D. Sanna, G.

Micera, J. Chem. Soc. Dalton Trans. (2001) 2850–2857.

http://www.staff.u-szeged.hu/~szabot/supinf.html

http://www.staff.u-szeged.hu/~szabot/supinf.html

http://dx.doi.org/

http://dx.doi.org/

Copper(II) complexes of some N-substituted bis(aminomethyl)phosphinate ligands. An integrated EPR...

Documents

Transcript of Copper(II) complexes of some N-substituted bis(aminomethyl)phosphinate ligands. An integrated EPR...