Bioorganometallic Chemistry. 27. Synthetic, X-ray Crystallographic

Apr 22, 2014 - Bioorganometallic Chemistry. 27. Synthetic, X-ray Crystallographic, and Competitive .... Alba Collado , Mar Gómez-Gallego , Miguel A. ...
0 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/Organometallics

Bioorganometallic Chemistry. 27. Synthetic, X‑ray Crystallographic, and Competitive Binding Studies in the Reactions of Nucleobases, Nucleosides, and Nucleotides with [Cp*Rh(H2O)3](OTf)2, as a Function of pH, and the Utilization of Several Cp*Rh−DNA Base Complexes in Host−Guest Chemistry David P. Smith,† Hong Chen,† Seiji Ogo,†, ‡ Ana I. Elduque,§ Miriam Eisenstein,∥ Marilyn M. Olmstead,*,⊥ and Richard H. Fish*,† †

Lawrence Berkeley National Laboratory, University of California, Berkeley, California 94720, United States Department of Chemistry and Biochemistry, Graduate School of Engineering, Kyushu University, 744 Moto-oka, Nishi-ku, Fukuoka, 819-0395, Japan § Department of Inorganic Chemistry, Faculty of Science-ICMA, University of Zaragoza-CSIC, 5009 Zaragoza, Spain ∥ Department of Chemical Research Support, Weizmann Institute of Science, Rehovot 76100, Israel ⊥ Department of Chemistry, University of California, Davis, California 95616, United States ‡

S Supporting Information *

ABSTRACT: The reactions of the air- and water-stable tris(aqua) complex [Cp*Rh(H2O)3](OTf)2 (1; OTf = trifluoromethanesulfonate) with nucleobases and nucleosides that included 9-methyladenine (9-MA), 9-ethylguanine (9-EG), 9methylhypoxanthine (9-MH), 9-ethylhypoxanthine (9-EH), 1methylcytosine (1-MC), 1-methylthymine (1-MT), adenosine (Ado), and guanosine (Guo) provided new bonding modes, all as a function of pH. The 9-MA nucleobase provided a novel cyclic trimer, at pH 6, characteristic for all Ado complexes: [Cp*Rh(μ2-η1(N1):η2(N6,N7)-9-MA/Ado)]3(OTf)3. The Cp*Rh(9-EG) and Cp*Rh(Guo) complexes showed N7 and 6-CO binding modes in water, [Cp*Rh((η2(N7,O6)-9-EG/Guo)(OH)](OTf), and no cyclic trimer products, due to a pronounced steric effect of the 2-amino group. This was shown convincingly by the results with 9-MH and 9-EH, which did form cyclic trimers at pH 6.1, [Cp*Rh(μ2-η1(N1):η2(N7,O6)-9-MH/9-EH)]3(OTf)3, with a structure similar to that of 9-EG, but with no 2amino group available. At pH 10.2, the pKa of the 9-MH’s NH1 hydrogen dictated the structure, providing a μ-hydroxy dimer, trans-[Cp*Rh(η1(N1)-9-MH)(μ-OH)]2(OTf)2, while in methanol the same reaction provided a mononuclear complex, [Cp*Rh(η1(N7)-9-MH)(MeOH)2](OTf)2. The reaction of 1 and 1-MC, at pH 5.4, provided another μ-hydroxy dimer with intramolecular H bonding of the O and H atoms of the μ-OH groups (H-acceptor and H-donor, respectively), trans[Cp*Rh(η1(N3)-1-MC)(μ-OH)]2(OTf)2, while in acetone, the product was a monomeric complex, [(η5-Cp*Rh)(η1(N3)-1MC)(η2(O2,N3)-1-MC)](OTf)2. The reaction of 1 and 1-MT at pH 10 showed the initial complex 1 being converted to its equilibrium complex, [(Cp*Rh)2(μ-OH)3]+, and this led to two components being formed. The anionic component was a linear [(η1(N3)-MT)−RhI−(η1(N3)-MT)]− (12e RhI center) assembly, formed via a presumed reductive elimination of Cp*OH, and included an orthogonal array of two thymine planes. The cationic component was [(Cp*Rh)2(μ-OH)3]+, with its Cp* moiety being π−π stacked with thymine rings, as well as the π−π interactions of two thymine rings: {[RhI(η1(N3)-1MT)2]2[(Cp*Rh)2(μ-OH)3]3}OH. The competitive order of nucleoside reactivity was Ado ≫ Guo, while for the nucleotides it was GMP > AMP ≫ CMP ≈ TMP. Finally, we also discuss several examples of the utilization of these unique Cp*Rh−DNA base complexes, as aqueous hosts for molecular recognition of aromatic amino acids and as NMR shift reagents for many organic compounds.

■

INTRODUCTION

as useful probes for molecular biology, as well as regulators of gene expression.1 Surprisingly, similar in-depth studies in aqueous solution of the aforementioned biologically important ligands with organometallic complexes had not been as

For historical reasons, the reactions of inorganic complexes, principally late-transition-metal complexes, with DNA/RNA nucleobases, nucleosides, nucleotides, and oligonucleotides have been extensively studied in order to determine the binding mode, the mode of action of these metal complexes as a consequence of their antitumor activity, and to employ them © 2014 American Chemical Society

Received: January 29, 2014 Published: April 22, 2014 2389

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

tions.9a−e Thus, in this full account, we present the complete synthetic, X-ray structural, and spectroscopic studies on the reactions of the tris(aqua) complex 1 with the nucleobases 9methyladenine (9-MA), 9-ethylguanine (9-EG), 9-methylhypoxanthine (9-MH), 9-ethylhypoxanthine (9-EH), 1-methylcytosine (1-MC), and 1-methylthymine (1-MT) and including the nucleosides adenosine (Ado) and guanosine (Guo); see Chart 1 for structures. We have also determined the selectivity/

pervasive during the 1970s and 1980s, even though the area of bioorganometallic chemistry represented a new and exciting direction for chemists interested in carbon−metal-bonded complexes for applications to biological problems.2 More importantly, in the past, the reactions of organometallic complexes with biologically significant ligands were not fully focused on water as a reaction medium.3 Many of the early pioneering studies with organometallic complexes and biological ligands by Beck et al. were conducted in methanol, an organic solvent as close to water as possible, but in which pH could not be an accurate factor for structural variations and future biological applications.4 When we started our very initial bioorganometallic chemistry program on bioorganotin chemistry in the early 1970s, few studies in water had been reported, due to the thought that these organometallic complexes were not stable in aqueous solution.5 Early in the 1990s with the inception of the NIH/DOE genome mapping and sequencing program, we discovered, along with an LBNL biophysical chemist colleague, Marcos Maestre, a novel method for mapping and sequencing large DNA strands, such as the 50000 base paired λ-DNA, using the (η5pentamethylcyclopentadienyl)rhodium tris(aqua) complex [Cp*Rh(H2O)3](OTf)2 (1; OTf = trifluoromethanesulfonate) and a 20-mer of 10 adenine bases and 10 bases that complemented the end bases of λ-DNA, to tether the λ-DNA to the glass stage containing gold electrodes of an epifluorescence microscope.6 However, even though our mapping and sequencing method was successful and was patented, it was the proverbial black box in that we did not understand the binding modes of complex 1 to the 20-mer or, more pertinently, even to individual nucleobases, nucleosides, or nucleotides. In the early 1990s, no reactions had been initiated with the tris(aqua) complex 1 and DNA bioligands in water, a more compatible solvent for biological relevance. However, Marks et al. had conducted, during the late 1980s and early 1990s, reactions of nucleobases and nucleotides in water with metallocene dihalides, and these were important studies, since we could compare our results with the tris(aqua) complex, 1, which had three available coordination sites, with the metallocene dihalides V, Ti, and Mo, with two available coordination sites.7 Furthermore, other groups have made contributions to this field with several important organometallic complexes, for instance, organoruthenium/iridium/rhodium− DNA base complexes, etc., over the intervening years,8a−j but few presented new bonding modes of the organometallic− DNA base structures.8f−j For example, Sheldrick and coworkers8f−i extended our nucleobase studies with the Cp*Ir aqua complex and guanine, adenine, and hypoxanthine; these have no N9 substituents and therefore are not representative of DNA base derivatives. Thus, they reported the X-ray structure of the Cp*Ir tetranuclear complex [Cp*Ir(μ-η1(N7,N9)guaninyl)]44+, while adenine provided the μη1(N9):η2(N6,N7) bonding mode. Furthermore, Yamanari and co-workers replaced the NH2 group on the adenosine nucleus with a CS group, an unconventional DNA base derivative, and reacted this with the Cp*Rh aqua complex, which provided by X-ray analysis a cyclic hexamer rather than the cyclic trimer we found with adenosine,9a having a μη1(N1):η2(S6,N7) bonding mode.8j Therefore, in the early 1990s, we started our studies on the reactions of complex 1 with biologically important nucleobases and nucleosides, reported in several preliminary communica-

Chart 1. Structures of Nucleobases, Nucleosides, and Nucleotides in Reactions with [Cp*Rh(H2O)3](OTf)2 (1)

reactivity order of several of the structurally different nucleosides and nucleotides with complex 1 in competition experiments. Thus, we will also discuss, for the first time, competitive binding studies with the nucleosides adenosine (Ado) and guanosine (Guo), as well as competitive reactions of the nucleotides guanosine 5′-monophosphate (GMP), adeno2390

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

Figure 1. (left) X-ray structure of the trication of [Cp*Rh(μ2-η1(N1):η2(N6,N7)-9-MA)]3(OTf)3 (2). Selected bond lengths (Å) and angles (deg): Rh−N6, 2.181(16); Rh−N7; 2.08(2); Rh-Nl′, 2.10(2); Rh−C15, 2.22(3); N6−Rh−N7, 79.5(7); N6−Rh−N1′, 88.7(8); N7−Rh−N1′, 84.2(9); Rh−N6−C6, 113.4(13); Rh−N7−C5, 108.3(15); C6−C5−N7, 122(2); C5−C6−N6, 117(2); N7−Rh−C15, 119.8(10). The dihedral angle between the two 9-MA planes is 82.3°. (right) 1H NMR structure of the trication of [Cp*Rh(μ2-η1(N1):η2(N6,N7)-Ado)]3(OTf)3 (3).

nucleotide and was in the range of ∼4 Å. Previously reported studies have verified a N6−N7, five-membered-ring chelate of 9-MA with a Cp2Mo center7 and N1 and N7 bridging of 9-MA between metal centers (e.g., Rh2+, Ag+, Co2+);10,11 however, no metal−9-MA complex had been reported, at that time, incorporating both bonding features in one molecule, as is shown in 2. Clearly, the formation of the 12-membered-ring cyclic trimer resulted from the ability of the Cp*Rh2+ cation to act as an azaphile, with the favorable geometry of N1 poised to form the third bond to a Cp*Rh of an adjacent [Cp*Rh(η2(N6,N7)-9-MA)](OTf) moiety: [Cp*Rh(μ2η1(N1):η2(N6,N7)-9-MA)]3(OTf)3. The Cp*Rh(Ado) complex 3, from the reaction of 1 and Ado at pH 7.1, also indicated from 1H NMR studies (DMSO-d6) dramatic downfield H8 and upfield H2 chemical shifts similar to those of complex 2 at 8.92 and 7.62 ppm, respectively, for [Cp*Rh(μ2-η1(N1):η2(N6,N7)Ado)]3(OTf)3, which was also suggestive of a cyclic trimer structure (Figure 1). Reactions of [Cp*Rh(H2O)3](OTf)2 with the Nucleobase 9-Ethylguanine and the Nucleoside Guanosine in Methanol and Water. The reactions of the Cp*Rh tris(aqua) complex 1 with 9-EG in methanol and Guo in water provided monomers 4 and 5, via 1H NMR, MS, and elemental analysis, with η1(N7) and η2(N7,O6) binding, respectively. However, no evidence was found for cyclic trimer formation, as was previously shown for the Cp*Rh complexes of 9-MA and Ado. Thus, the reaction of 1 in MeOH to generate [Cp*Rh(MeOH)3](OTf)2 in situ with 9-EG gave [Cp*Rh(η1(N7)-9-EG)(MeOH)2](OTf)2 (4), including typical N7 binding, via 1H NMR in DMSO-d6, with an H8 downfield shift in comparison to free 9-EG of Δδ = 0.70 ppm and the CH3OH ligands at 3.15 ppm; see complex 6 (vide infra).9c Moreover, the reaction of 1 in water with Guo at pH 5.4 provided, by FAB-MS (m/z 670.1, [Cp*Rh(Guo)(OTf)]; m/z 556.1, [Cp*Rh(Guo)(H2O)2(H)]) and elemental analysis, a monomer with the formula [Cp*Rh(Guo)(OH)](OTf)·4H2O (5). The structure of 5 was also elucidated by 500 MHz 1H NMR spectroscopy in DMSO-d6 to show a substantial downfield shift for H8 at 8.93 ppm (Δδ = 1.02 ppm), which was consistent with classical N7 binding to the Guo nucleus, while the NH1 group was also shifted downfield by Δδ = 0.53 ppm, and this was interpreted as being indicative of the 6-CO group interacting with the Cp*Rh metal center to provide the

sine 5′-monophosphate (AMP), cytidine 5′-monophosphate (CMP), and thymidine 5′-monophosphate (TMP). Finally, we will also discuss the utility of several of these Cp*Rh−DNA base complexes as hosts in molecular recognition studies with, for example, aromatic amino acids, as well as novel NMR shift reagents for di- and tetrapeptides, among other biorelated compounds.

■

RESULTS AND DISCUSSION Reactions of [Cp*Rh(H2O)3](OTf)2 with the Nucleobase 9-Methyladenine and the Nucleoside Adenosine in Water. The tris(aqua) complex [Cp*Rh(H2O)3](OTf)2 (1), prepared in situ,3a was reacted with 9-methyladenine (9-MA) in D2O at pD 7.2 (pD = pH + 0.4) to provide evidence, by 1H NMR spectroscopy, for the formation of the Cp*Rh(9-MA) complex 2, with dramatic chemical shifts for both H2 and H8 in comparison to those for free 9-MA, at 8.51 and 7.62 ppm. By utilizing 9-MA-d8, which was selectively deuterated at H8, we were able to unequivocally assign the chemical shifts to each proton of 2 and show that H8 was shifted downfield, Δδ = 0.75 ppm, from free 9-MA, while H2 was shifted upfield, Δδ = −0.47 ppm, in DMSO-d6. It was interesting to observe that formation of complex 2 occurred over the pD range of 6−9 and that it was found to be stable for over 1 week at ambient temperature (1H NMR, pD 7.2). Complex 2 was isolated and purified by recrystallization from methanol to yield an orange solid (26%) and was found by FAB/MS (m/z 1456.7; M − OTf), single-crystal X-ray crystallography, and elemental analysis to have, at that time, the unusual and unprecedented structure of a cyclic trimer (Figure 1).9a Cyclic trimer 2 was found to be a racemic mixture from CD studies, while a single-crystal X-ray structure of an enantiomer of 2 (crystal picking) showed that it had a triangular domelike supramolecular structure, with three Cp* groups stretching out from the top of the dome, three Me groups pointing to the bottom, three adenine planes forming the surrounding shell, and three Rh atoms embedded in the top of the dome. This molecule also possessed a C3 axis, which passed from the top of the dome to the bottom. The distance between the adjacent methyl groups at the bottom of the dome; i.e., at the opening of this potential molecular receptor, was about 7.5 Å, while the cavity depth was a consequence of the substituent on N9 of the nucleobase, nucleoside, or 2391

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

plausible structure [Cp*Rh((η2(N7,O6)-Guo)(OH)](OTf) (5) from the 1H NMR, FAB-MS, and elemental analysis data.9a

Reactions of [Cp*Rh(H2O)3](OTf)2 with the Nucleobases 9-Methylhypoxanthine and 9-Ethylhypoxanthine in Methanol and Water. From the aforementioned results with 9-EG and Guo, it was shown that these two bioligands, nucleobase/nucleoside, did not provide a cyclic trimer, as did 9MA and Ado. Our supposition was that the difference between these nucleobases/nucleosides was the 2-amino group on the guanine nucleus, which could generate a steric effect that prevented cyclic trimer formation. Thus, we reacted 1 with 9MH and 9-EH in methanol and in water, at various pH values, in order to answer the following questions. Does the 2-amino group on the guanine nucleus prevent cyclic trimer formation, since the substituted hypoxanthine nucleobase has no 2-amino group? Does pH have a profound effect on the structure of Cp*Rh−nucleobase/nucleoside complexes, in comparison to methanol? We found that the structure of a monomeric complex, [Cp*Rh(η1(N7)-9-MH)(MeOH)2](OTf)2 (6), obtained from the reaction of 1 equiv of 9-MH with in situ generated [Cp*Rh(MeOH)3](OTf)2 in MeOH, also showed exclusive N7 binding, by 500 MHz 1H NMR spectroscopy in DMSO-d6, via substantial downfield shifts for H8 of Δδ = 0.53 ppm and H2 of Δδ = 0.15 ppm.9c An X-ray crystal structure of 6 corroborated the N7 binding regime, while also showing interand intramolecular hydrogen bonding for O9···H1 (1.86(1) Å) and O6···H2 (1.83(1) Å), respectively (Figure 2); a similar 1H NMR structure was elucidated for [Cp*Rh(η1(N7)-9-EH)(MeOH)2](OTf)2 (6a). An 1H NMR pD titration profile of 6 in D2O clearly defined the role of pD in the conversion of this monomer 6 to cyclic trimeric and dimeric complexes (Figure 3).9d By observing the H8 and H2 regions of the 1H NMR spectra at various pD values (pD = pH + 0.4), we found that 6 remained unchanged from pD 2.45 to 5.13 (η2(N7,O6) binding by 1H NMR, 7; the signal broadened slightly at the latter pD); however, an equilibrium process occurred when the pD values were increased to 6.45, 7.96, and 10.50. At pD 6.45, the 1H NMR spectrum showed the dramatic chemical shifts we have found to be diagnostic for the formation of cyclic trimers,9a while at pD 10.50, a new complex was formed with significant upfield shifts of H8 and H2 that were reminiscent of [Cp*Rh(L)(μ-(OH)]2 dicationic, dimeric complexes, where for example L = quinoline,12 1-MC.9b This was basically the paradigm for Cp*Rh−DNA base structures we solved by single-crystal X-ray analysis, with pH dictating the structure of the complexes that formed. Thus,

Figure 2. Structure of the dication of [Cp*Rh(η1(N7)-9-MH)(MeOH)2](OTf)2 (6; corrected CSD structure, where N3 was mislabeled as a C−H in ref 9c). Selected bond lengths (Å) and angles (deg): Rh1−C (average), 2.13; Rh1−N7, 2.127(4); Rh1−O1, 2.172(5); Rh1−O2, 2.172(5); H2−O6, 1.83(1); H1−O9, 1.86(1); H1a−O5′, l.89(1); O3′···O10″, 2.734(1); N7−Rh1−O2, 82.9(2); N7−Rh1−O2, 86.3(2); O1−Rh1−O2, 83.0(2).

different structures, based on the pKa of the hypoxanthine ligand, were obtained: complexes 7−9 (Figure 4).9d The cyclic trimer 8, a racemic mixture, was obtained as yellow crystals from MeOH/H2O, and the cation of one enantiomer (crystal picking) has the X-ray crystal structure shown in Figure 4. A key feature was the simultaneous η1(N1) and η2(N7,O6) ligation of 9-EH, which was unprecedented in comparison to a great majority of N7 monodentate complexes and the few examples of η2(N7,O6) binding modes of 9substituted guanine and hypoxanthine derivatives found in the literature.13 It is interesting to note, for comparison, the η2(N7,N6) and η1(N1) ligation of 9-MA with Cp*Rh in this paper. Two other structural features of 8 merit some comments, and they include the trication, a triangular domelike cavity, with three Cp* groups stretching out from the top of the dome, three ethyl groups pointing to the bottom, three hypoxanthine planes form the surrounding shell, and three Rh atoms were embedded on the top of the dome for the superstructure. The view shown in Figure 4 is from the bottom of this cavity along the C axis, which passes from the top of the dome to the bottom. Second, the C6−O6 bond distance of 1.30(2) Å fell between the single-bond length of 1.42 Å found in an alcohol 14 and the double-bond lengths of 1.233(4) and 1.230(7) Å in inosine15 and the η1(N7) monodentate monomer 6, respectively. This result suggested that a significant 2392

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

omatic nitrogen ligands (cis-L = quinoline, trans-L = nucleobase, 1-MC; vide infra, complex 11); however, some slight differences merit some comment. The planar Rh2(μOH)2 fragment of 9 has an O1−Rh−O1′ angle of 77.9(1)°, which was between that of trans-[Cp*Rh(η1(N3)-1-MC)(μOH)] 2 (OTf) 2 (11; 78.8(2)°) and cis-[Cp*Rh(η 1 -N1quinoline)(μ-OH)]2(OTf)2 (75.1(1), 74.8(1)°), and an Rh1− O1−Rh1′ angle of 102.1(1)°, which was larger than that of 11 (101.2(2)°) yet smaller than that of [Cp*Rh(η1(N1)quinoline)(μ-OH)]2(OTf)2, (105.0(1)°). These structural features are consistent with the intermediate Rh1···Rh1′ distance of 3.309(1) Å for 9, relative to those observed for trans-[Cp*Rh(η1(N1)-quinoline)(μ-OH)]2(OTf)2 (3.322(1) Å)12 and 11 (3.290(2) Å). These slight structural differences were presumably caused by the different numbers of intramolecular H bonds in these three complexes. Complex 9 has two intramolecular H bonds in which the Rh2(μ-OH)2 core acted only as H donors, whereas complex 11 has four intramolecular H bonds in which the Rh2(μ-OH)2, core acted as both a H donor and a H acceptor, while [Cp*Rh(η1(N1)quinoline)(μ-OH)]2(OTf)2 had no intramolecular H bonds. The four intramolecular H bonds in 11 possibly caused the two μ-OH groups to be farther apart, thus providing the largest O− Rh−O angle and shortest Rh···Rh distance among these three dimeric complexes. Although the exact mechanism of the formation of 9 has still not been fully elucidated, it is conceivable that the deprotonation of NH1 (pK, = 9.5) at pH 10.2 made the N1 site much more nucleophilic than N7 or O6 and, in addition, the intramolecular H bonding between O2 and the μ-OH group stabilized the dimeric arrangement of 9. Moreover, at pH 6.1 the construction of the cyclic trimer 8 might emanate from the more favorable η2(N7,O6)-bidentate five-membered-ring complex, since N1 is more nucleophilic at this pH than NH1 and presumably this η2(N7,O6) coordination was more thermodynamically favored. The facile deprotonation of NH1 was followed by formation of the third binding site to an adjacent Cp*Rh moiety, a [Cp*Rh(η2(N7,O6)-9-MH)](OTf)2 intermediate complex, since the NH1 would be more acidic upon the formation of the η2(N7,O6)-bidentate five-membered ring. Reactions of [Cp*Rh(H2O)3](OTf)2 with 1-Methylcytosine in Acetone and Water. As we have seen with nucleobases/nucleosides with exocyclic NH2 groups, 9-MA, Ado, 9-EG, and Guo, the first two DNA bases provided cyclic trimers, while the last two did not. In order to further understand the scope of this condensation reaction with nucleobases/nucleosides that have exocyclic NH2 groups, we studied the reaction of the tris(aqua) complex, 1, with 1methylcytosine (1-MC). We found that two different Cp*Rh(1-MC) complexes could be isolated, depending on the solvent used in the reaction. For example, the reaction in acetone with 2 equiv of 1-MC for 36 h gave an orange precipitate (67%) of 10, whose 1H NMR spectrum in DMSO-d6 revealed only one set of 1-MC and Cp*Rh resonances and considerable downfield chemical shifts for H5 and H6 in comparison to those of free 1-MC (Δδ(H5) = 0.25 ppm, Δδ(H6) = 0.26 ppm). In addition, two broad N(4)H2 resonances were also shifted downfield upon coordination (7.70 and 8.50 ppm, in comparison to free 1-MC at 6.91 ppm). The FAB/MS data were consistent with a monomeric species (m-nitrobenzyl alcohol; m/z 512, [Cp*Rh(1-MC)(OTf)]) and provided no evidence for dimer formation. The ’H NMR data (one set of 1-

Figure 3. 400 MHz 1H NMR spectra of 6, at various pD values.

amount of multiple-bond character still existed in the C6−O6 bond of 8. The X-ray structure of the orange dimer 9, isolated from its aqueous reaction mixture at pH 10.2, is shown in Figure 4.9d The main structural features of interest were the unique η1(N1) rather than η1(N7) or η2(N7,O6) binding mode of 9-MH and the intramolecular hydrogen bonding between the μ-OH groups and O2 of this nucleobase. These binding arrangements were in sharp contrast to those in 7 and 8, as well as other reported metal complexes of 9-substituted guanine and hypoxanthine derivatives, but were precedented, as shown by the studies of Lippert and co-workers on the η1(N1) binding of 9-methylguanine (9-MG) to PtII.16 However, this (η1(N1)-9MG)PtII complex was not formed directly but was prepared via the (μ-η1(N7):η1(N1)-9-MG)Pt2II complex by selective cleavage of the PtII atom at N7 with CN−. As shown with Guo, the tris(aqua) complex 1 formed a mononuclear complex with η2(N7,O6) coordination rather than a cyclic trimer such as 9EH and thus, unlike PtII, the Cp*Rh group was too sterically demanding to form a η1(N1) complex with the guanine nucleus; i.e., the adjacent NH2 group at C2 prevented this type of bonding (vide supra). Dimeric complex 9 represented a very rare example in which the μ-OH group is involved in intramolecular hydrogen bonding.9b The C4−O2 bond of 1.259(7) Å was slightly longer than those observed in inosine (1.233(4) Å) and a typical aldehyde (1.22 Å),14 which indicated the primary double-bond character of this C4−O2 group. The elongation of this bond was presumably due to the intramolecular H bond between O2 and the μ-OH group, as well as the intermolecular H bond between O2 and the H2O molecules present in the crystal lattice. The core structure of 9 was similar to those of cis- and trans[Cp*Rh(η1-L)(μ-OH)]22+, dimeric complexes with heteroar2393

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

Figure 4. Cp*Rh(9-methyl-/9-ethylhypoxanthine) structures 7 (1H NMR postulated dicationic structure), 8, and 9, dictated by pH. X-ray structure of the trication of [Cp*Rh(μ2-η1(N1):η2(N7,O6)-9-EH)]3(OTf)3 (8). Selected bond lengths (Å) and angles (deg): Rh1−Cl5, 2.12(3); Rh1−N1′, 2.135(14); Rh1−O6, 2.240(14); Rh1−N7, 2.133(16); O6−Rh1−N7, 80.2(6); O6−Rh1−Nl′, 88.1(6); N7−Rh1−N1′, 85.0(7); Rh−O6−C6, 106.9(14); Rh1−N7−C5, 106.0(14); N7−Rh1−C15, 116.8(9); O6−C6−C5, 122(2); C6−C5−N7, 123.6(17); C2−N1−C6, 115(2); N1−C6−C5, 114.1(18); N1−C6−O6, 124(2); C5−N7−C8, 102.2(17). X-ray structure of the dication of trans-[Cp*Rh(η1(N1)-9-MH)(μ-OH)]2(OTf)2 (9). Selected bond lengths (Å) and angles (deg): Rh1−O1, 2.136(4); Rh1−N1, 2.148(5); O1−Rh1−N1, 86.3(2); Rh1−O1−Rh1′, 102.1(1); Rh1−N1− C4, 118.4(4); Rh1−N1−C1, 119.8(4); C1−N1−C4, 121.1(5); N1−C4−C3, 113.2(5); N1−C4−O2, 121.2(6); O2−C4−C3, 125.5(5); C4−C3− N3, 130.8(5); C3−N3−C5, 103.1(5); O1···O2, 2.76(2).

MC signals for 10, down to −90 °C in CD3OD) supported the solution structure of 10 as [(η5-Cp*Rh)(η1(N3)-1-MC)2(S)](OTf)2. An X-ray structural determination of complex 10 recrystallized from methanol, a weakly coordinating solvent, clearly showed in the solid state that one of the 1-MC ligands bonded via a four-membered-ring chelate, N3−Rh−OC2 (Rh−N3a, 2.143(7) Å; Rh−O2a, 2.251(6) Å), with the other ligand bound through the expected N3 site (Rh−N3b, 2.126(8) Å) (Figure 5). Several structurally characterized examples of cytosineN3,O2 metal semichelates exist and exhibit longer M−O bond lengths (M = Cu, 2.76 Å; M = Cd, 2.64, 2.56, and 2.89 Å; M = Hg, 2.84 Å) in comparison to those observed in compound 10. The structural consequences of N3,O2 chelation, versus N3 coordination alone, were manifested in complex 10 by a longer C−O bond length (C2a−O2a, 1.264 (10) Å; C2b−O2b, 1.209(11) Å), a shorter C2−N3 bond length (C2a−N3a, 1.350(10) Å; C2b−N3b, 1.402(11) Å), a shorter C4−N4 bond length (N4a−C4a, 1.308 Å; N4b−C4b, 1.349(13) Å), and a smaller O2−C2−N3 bond angle (O2a−C2a−N3a, 117.1(7)°; O2b−C2b−N3b, 121.6(8)°). The N4a···O2b through-space

distance (3.175 Å) precluded any intramolecular hydrogen bonding between these sites. Upon recrystallization of complex 10 from H2O at pH 5.1, we isolated complex 11. Complex 11 was also isolated by the dropwise addition of a deoxygenated, aqueous solution of 1 equiv of l-MC to an aqueous solution of 1 that was adjusted to pH 5−6 by addition of NaOH. The ’H NMR spectrum of 11 in DMSO-d6 showed upfield chemical shifts for H5 (5.27 ppm) and H6 (7.45 ppm) of Δδ = −0.32 and −0.09 ppm, respectively, in comparison to free l-MC, while one set of exocyclic NH2 protons at 7.18 ppm was shifted downfield by Δδ = 0.27 ppm. The other exocyclic NH2 signal was not observed and was apparently broadened into the baseline. FABMS verified the dimeric nature of 11 (m-nitrobenzyl alcohol; m/z 766, [(η5-Cp*Rh)2(1-MC)(μ-OH)](OTf); m/z 659, [(η5Cp*Rh)2(μ-OH)2](OTf)), while a single-crystal X-ray analysis confirmed the trans stereochemistry and the extensive intramolecular hydrogen-bonding regime of the μ-OH groups with the CO and the exocyclic NH2 groups (Figure 6). This hydrogen-bonding network created a hydrophobic environment around the metal centers and appeared to be a plausible reason that 11 was found to be insoluble in H2O at pH 5.1. 2394

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

Figure 6. X-ray structure of the dication of trans-[Cp*Rh(η1(N3)-1MC)(μ-OH)]2(OTf)2 (11), showing the intramolecular H-bonding modes of OH···OC and HO···HN. The two OTf− counterions are not shown. Selected bond lengths (Å) and angles (deg): Rh1−N3, 2.181(6); Rh1−O10, 2.138(5); Rh1−OlO′, 2.118(6); C(2)−O(2), 1.232(10); C(4)−N(4), 1.32(9); H4b−O1O′, 1.93(1); H10−O2, 1.96(1); N3−Rh1−O10, 87.9(2); Rh1−O10−Rh1′, 101.2(2); O2− C(2)−N3, 123.1(6); N4−H4b−O10′, 149(1); O10−H10−O2, 155(1).

complex [Cp*Rh(H2O)(μ-OH)]22+, the equilibrium complex between 1 and [(Cp*Rh)2(μ-OH)3]3+, obtained previously, indicative of a reactive dimeric core, [Cp*Rh(H2O)(μOH)]22+.3a Condensation reactions between the exocyclic NH2 of 1-MC and M−OH centers to form either four-membered-ring chelates (η2(N3,N4); Cp2Mo2+, PtIV) or μ-1-MC complexes (PtII) have been well documented.7d,17 However, we were not able to induce a similar condensation reaction with 11 between the exocyclic NH2 and the μ-OH group. Mild thermolysis (70 °C) for 16 h of 11 in DMSO-d6 solution resulted in overall decomposition, with no evidence of a condensation reaction. The fact that we observed no apparent condensation reaction of the exocyclic NH2 group with the μ-OH group could be indicative of the pronounced stability of the extensive intramolecular hydrogen-bonding regime shown in 11 or could be simply a manifestation of the instability of fourmembered-ring chelates. Reaction of [Cp*Rh(H2O)3](OTf)2 with 1-Methylthymine in Water. We have clearly shown that pH had a truly profound effect on the reactions of nucleobase/nucleoside with [Cp*Rh(H2O)3](OTf)2 (1). This was a consequence of the pKa of the nucleobase NH groups and, as well, the structural pH dependence of the tris(aqua) complex 1. Therefore, we extended these pH structural scenarios to the nucleobase 1-MT (Chart 1), with N3−H and pKa = 9.7. We had previously described the synthesis and structural characterization of the first example of a novel, linear, two-coordinate RhI anionic amide complex, [RhI(η1(N3)-1-MT)2]− (1-MT was deprotonated at N3−H), from the reaction of 1-MT with in situ generated [(Cp*Rh)2(μ-OH)3]+ from 1 performed at pH 10.9e This unusual coordination around the RhI complex was presumed to be stabilized by three factors: an organometallic hydrophobic cavity, generated from 1.5 molecules of [(Cp*Rh)2(μ-OH)3]+ to provide the interaction of two 1-MT ligands with the adjacent Cp* groups by a classical π−π molecular recognition process, an electrostatic interaction of the anionic component with the cationic component, and a

Figure 5. Structure of the dication of [(η5-Cp*Rh)(η1(N3)-1MC)(η2(O2,N3)-1-MC)](OTf)2 (10). View A shows the cation of structure 10, with an emphasis on the two different modes of nucleobase coordination. View B is a side-on view of the cation. Hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and angles (deg): Rh−N3a, 2.143(7); Rh−N3b, 2.126(8); Rh−O2a, 2.251(6); N3a−Rh−N3b, 90.0(3); N3a−Rh−O2a, 61.0(2); O2a− Rh−N3b, 88.0(2).

The core structure of 11 was found to be quite similar to those of 9 and cis-[Cp*Rh(η1(N1)-quinoline)(μ-OH)]2(OTf)2. The planar Rh2(μ-OH)2 fragment of 11 showed a wider O10− Rh1−O10′ angle of 78.8(2)°, in comparison to the cisquinoline derivative (76.1(1), 74.8(1)°). Moreover, a smaller Rh1−O10−Rh1′ angle (101.2(2)°) for 11 was observed in comparison to 104.7(1) and 105.0(1)° for the cis-quinoline derivative. This resulted in a Rh···Rh through-space distance of 3.290(2) Å for 11, in comparison to 3.322(1) Å for the cisquinoline derivative.12 This slight deformation was presumably caused by the unusual binding of the l-MC ligand, in which the (Rh-μ-OH)2 core acted as a covalent electrophile, a hydrogen bond donor, and a hydrogen bond acceptor. This system was found to be a rare example of a DNA base complex in which a μ-hydroxy ligand exhibited simultaneous hydrogen bond donor and acceptor capabilities.9b Although other metal complexes of 1-MC have shown extensive intermolecular hydrogen bonding, we were unaware of any system that recognized the bonding capabilities of this ligand so readily in an intramolecular fashion. In addition, the formation of 11 provided some evidence for the pH-dependent, putative structure of the intermediate aqua 2395

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

possible shielding of the RhI center to nucleophilic attack by four sets of oxygen lone pair electrons contained in the 1-MT ligands. Therefore, reaction of [Cp*Rh(H2O)3](OTf)2, (1) with 1 equiv of 1-MT in H2O (degassed once) at pH 10 afforded a yellow solution; the rapid formation of [(Cp*Rh)2(μ-OH)3]+ at pH 10 has been extensively studied.3a After it was stirred at 25 °C overnight and at 60 °C for 2 h, to drive the reaction to completion, the solution turned orange, and when the volume of this reaction mixture was reduced, the complex crystallized with 1.5 equiv of [(Cp*Rh)2(μ-OH)3]+ at 4 °C to form the adduct {[RhI(η1- (N3)-(1-MT)2]2−·[(Cp*Rh)2(μ-OH)3]3+} (12) as orange plates (20%). A small amount ( AMP ≫ CMP ≈ TMP, while for nucleosides Ado ≫ Guo. Utilization of the Cp*Rh Cyclic Trimer Complex of 2′DeoxyAdo 13 and the Cp*Rh-1-MC Dimer Complex 11 as Hosts for Molecular Recognition Studies. While the preceding Cp*Rh−DNA base complexes were found to be structurally interesting in themselves, we have always strived to

Figure 11. (left) Dreiding molecular model of the host [Cp*Rh(μ2η1(N1):η2(N6,N7)-2′-deoxyAdo)]3(OTf)3 (13) and (right) the CPK model.

analysis of these host−guest noncovalent interactions was accomplished via 1H NMR studies, which showed selective hydrogen upfield shifts via π−π, H-bonding, and hydrophobic interactions, with the inner 2′-deoxyadenine groups that make up the cyclic trimer structure; the Cp*Rh moiety had no role in the noncovalent interactions of host−guests but stabilized the domelike supramolecular structure of the cyclic trimer complex 13. From the X-ray structure of the cyclic trimer 2, the depth of the cavity inside, which was also a consequence of the N9 substituents, for example, nucleobase, nucleoside, or nucleotide, was ∼4 Å and the width of the molecular receptor opening was ∼7.5 Å. The docking of the aromatic amino acid L-trytophan with the host 13 (Figure 12, left), clearly corroborated the 1H NMR studies, which showed the selective, complexation-induced chemical shifts (CICS) of protons designated in Figure 12 for L-tryptophan (right), specifically, protons a, a′, b, and c, via π−π interactions with the inner 2′-deoxyadenine groups that make up the cyclic trimer structure, with upfield shifts for a and a′ of Δδ = −0.45 ppm, while for b and c, Δδ = −0.19 ppm. Proton d

Figure 12. (left) Host−guest docking of 13−L-tryptophan (left) and (right) the structure of the guest L-tryptophan with proton designations. 2398

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

and the amino acid −CH2C*H− protons e−g were minimally affected with shifts of Δδ = −0.01−0.02 ppm; the NMR estimated association constant, Ka, is 607 M−1, and the free energy for complexation, ΔG°, is −3.8 kcal/mol. Furthermore, we have also utilized host 13 as an NMR shift reagent, using aromatic and aliphatic carboxylic acids and diand tetrapeptides with terminal L-tryptophan and L-phenylalanine groups.27 Even at 500 MHz, some 1H NMR spectra were not first order, and thus, [Cp*Rh(μ2-η1(N1):η2(N6,N7)2′-deoxyAdo)]3(OTf)3 (13) can be useful in this situation, since it provided a global influence on the chemical shifts, as opposed to the use of lanthanide shift reagents (LSRs), which only bond to heteroatoms and cause contact or pseudocontact shifts of protons in the proximity of the magnetic anisotropy of the LRSs; line broadening was also a critical problem with the use of LRSs with increasing concentrations for NMR spectral analysis. As an example of the global nature of the influence of protons, via diamagnetic anisotropy effects, we analyzed the sleep-inducing tetrapeptide L-Trp-L-Met-L-Asp-L-Phe with 13.27 The continual addition of several 0.1 equiv portions of 13 to this tetrapeptide allowed almost complete identification of all 10 aromatic hydrogens, while aliphatic protons (not shown) were not affected by the shift reagent (Figure 13). Thus, we see in Figure 13 the powerful nature of this NMR shift reagent, with resolution of protons c, f, and h from spectrum A to spectrum D. Complex 11, trans-[Cp*Rh(η1(N3)-1-MC)(μ-OH)]2(OTf)2, also presented opportunities for host−guest docking experiments, with the thought that the H-bonding regime could support guest aromatic amino acids through similar noncovalent interactions.26 Moreover, we further surmised that an intermolecular recognition process also based on H bonding to the μ-OH groups and the cytosine NH 2 and CO functionalities might be feasible with the aromatic amino acid NH3+ and COO− groups, without seriously disrupting the intramolecular hydrogen-bonding regime shown in Figure 6. We again utilized 1H NMR techniques to discern the complexation-induced 1H NMR chemical shifts (CICS) for the host and the guest, in aqueous solution.26 Table 1 shows the results with the guest L-tryptophan (see the right-hand side of Figure 12 for proton designation), in the presence of host 11. What was dramatically evident for the guest L-tryptophan were the CICS for Hd (Δδ = −0.34), He (Δδ = −0.15), Hf (Δδ = −0.07), and Hg (Δδ = −0.12), which were diametrically opposite to the previously reported Cp*Rh-2′-deoxyadenosine cyclic trimer molecular recognition studies with L-tryptophan, where no upfield CICS for these designated protons were observed; in that process, the indole phenyl group was found inside the hydrophobic receptor, while the hydrophilic aromatic amino acid NH3+ and COO− groups were outside in the water media and the chiral C−H attached to these groups, as well as the adjacent asymmetric CH2, were not affected by the magnetic anisotropy of the inner shell of the host adenosine ligands.21 More importantly, we also observed two sets of signals for the host ligand, 1-MC, bound to Cp*Rh: the N−CH3, H5, and H6 protons. The CICS for one of the now apparently asymmetrical 1-MC ligands were similar to those of complex 11 alone, while the other had CICS upfield shifts for the N− CH3 (Δδ = −0.41), H5 (Δδ= −0.31), and H6 (Δδ = −0.18) (Table 2). Clearly, the CICS for one of the 1-methylcytosine ligands were affected by the noncovalent interactions with the indole ring of L-trytophan and vice versa. Thus, it appeared

Figure 13. 500 MHz 1H NMR spectra of a tetrapeptide, L-Trp-L-MetL-Asp-L-Phe, with added amounts of 13 in D2O at pH 9.4: (A) 0 equiv; (B) 0.1 equiv; (C) 0.2 equiv; (D) 0.3 equiv; (E) 0.4 equiv.

Table 1. CICS Shifts upon Host (11)−Guest (L-Tryptophan) Recognitiona free tryptophan (δ) 7.13 7.05 7.58 7.39 7.16 3.15 3.34 3.90

(a) (a′) (b) (c) (d) (e) (f) (g)

with host 11 (δ)

Δδ

7.02 6.89 7.16 7.14 6.82 3.00 3.26 3.77

−0.11 −0.16 −0.42 −0.25 −0.34 −0.15 −0.07 −0.12

a1

H NMR shifts at pH 7.0, 300 MHz, 1/1 host/guest ratio, 0.01/0.01 mmol.

plausible that the primary host−guest interaction of 11 with Ltryptophan was from a H-bonding process of the NH3+ and COO− groups with 11, enhancing noncovalent interactions of the 1-MC ligand with L-tryptophan. In order to better understand these H-bonding and noncovalent interactions between host and guest, we conducted computer docking experiments to provide the energy minimized, space-filling/ball and stick model of 11 with a ball and stick model of the guest L-tryptophan, as shown in Figure 14. The left view in Figure 14 demonstrated the H bonding of the NH3+ group to one μ-OH and to the CO group of one of 2399

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

Table 2. Host 11 1H NMR Data with Guest L-Trytophana

appeared in the docking experiment to be somewhat orthogonal to one of the Cp* ligands (Figure 14).

■

CONCLUSIONS In this full account of the synthesis of our novel structures of Cp*Rh nucleobase and nucleoside complexes, we have found that pH has played a critical role in product formation. Moreover, we have also shown the dramatic differences between the use of an organic solvent and water, with regard to the structure of the Cp*Rh−DNA base complexes; in organic solvents, we only observed mononuclear complexes. We have also clearly shown that highly stabilized Cp*Rh−DNA base complexes were the thermodynamic sink, generating cyclic trimers with 9-MA, Ado, 9-MH, and 9-EH at pH ∼6 and intramolecular H-bonding, μ-OH dimers for 1-MC in water at pH 5−6 and 9-MH/9-EH at pH 10.2, while 9-EG and Guo only provided mononuclear complexes in water, due to the pronounced steric effect of the NH2 at the 2-position. Furthermore, the nucleobase 1-MT has provided in its reaction with the Cp*Rh tris(aqua) complex 1 at pH 10 one of the most unusual and unique metal−DNA base complexes ever reported. This molecule has anionic and cationic components, with the first example of a linear 12e [RhI(η1(N3)-1-MT)2]− complex, trapped by the π−π interactions of Cp* groups of the cation [(Cp*Rh)2(μ-OH)3]+ with the 1-MT ligands, complex 12. It is interesting to note that the Cp*Ir(1-MT) analogue was also synthesized using the same procedure (unpublished results). Future experiments with 12 will be concentrated on its reactions with small molecules, such as CH4, CO, and H2, with the focus on the coordinatively unsaturated RhI center, in order to provide evidence for potential chemical reactivity by this novel bioorganometallic enzyme model. In competitive binding experiments, the cyclic trimer dominated with the reaction of [Cp*Rh(η2(N7-O6)-Guo)(OH)](OTf) (5) with 1 equiv of Ado in D2O at pD 7.3, providing the Ado cyclic trimer complex 3 and displacement of the Guo ligand; reactivity order Ado ≫ Guo. Thus, the thermodynamic stability of the cyclic trimer 3 prevailed in competition with mononuclear complex 5. In a reversal of competitive binding order in the Ado and Guo scenario, 5′GMP replaced 5′-AMP in the Cp*Rh(5′-AMP) complex, to provide [Cp*Rh((η1(N7)-5′-GMP)(H2O)2](OTf)2, reflecting possible differences in interaction of the Cp*Rh metal center with the P−O− group. In other competition experiments with 1, both 5′-CMP and 5′-TMP showed broadening, but no 1H NMR shifts, which allowed the order of reactivity to be 5′-GMP > 5′-AMP ≫ 5′-CMP ≈ 5′-TMP. We have also demonstrated the usefulness of several Cp*Rh−DNA base complexes, including 13, as a supramolecular host for guest molecules that encompass aromatic amino acids, as well as aromatic and aliphatic carboxylic acids; this benefited from predominately π−π and hydrophobic interactions with the 2-deoxyadenosine inner core of 13. Alternatively, complex 11, as a host for aromatic amino acids, utilized the amino and carboxylate groups with H bonding between NH of the 1-MC ligand of host and the COO− of the guest L-tryptophan, while the Rh-μ-OH and CO of the host H bond to the NH3+ of the guest. Both host examples, 11 and 13, provided convincing evidence of their ability for molecular recognition of biologically important guests, via noncovalent π−π, H-bonding, and hydrophobic interactions. We believe these to be the first examples of this type of molecular recognition with bioorganometallic DNA complexes, and this

guest L-tryptophan (δ) Δδ

free host 11 (δ) 3.24 7.44 5.83 1.46

(N−CH3) (H6) (H5) (Cp*)

2.83 7.26 5.52 1.35

−0.41 −0.18 −0.31 −0.12

Δδ 3.22 7.41 5.81 1.35

−0.02 −0.03 −0.02 −0.12

a1

H NMR shifts at pH 7.0, 300 MHz, 1/1 host/guest ratio, 0.01/0.01 mmol.

Figure 14. A novel bioorganometallic molecular recognition process with the host trans-[Cp*Rh(η1(N3)-1-MC)(μ-OH)]2(OTf)2 (11) docking with the guest L-tryptophan (left) and host−guest on the left turned 90° (right). The selective H bonds are designated between NH the 1-MC ligand of the host and the COO− of the guest Ltryptophan, while Rh-μ-OH and CO of the host H bond to the NH3+ of the guest (left).

the 1-MC ligands, while the COO− group H bonds to an NH2 group of the 1-MC ligand. This H-bonding scheme of 11 with L-tryptophan then provides that the remaining structure of the guest was fixed in relation to the host, as shown in the left and right (the structure on the left is turned 90°) views in Figure 14. Therefore, the indole group is shown in Figure 14 to be positioned orthogonal to the plane of the 1-MC ligands in host 11, while selectively affecting one of the two 1-MC groups, accounting for this ligand’s asymmetry and the upfield shifts observed in the 1H NMR CICS values (Table 1). Figure 14 also shows the plausible reason that Hb was appreciably shifted upfield due to its proximity to the CO group of 1-MC, while also noting the asymmetric CH2 hydrogens, where He is more affected by the CICS effects then Hf (Table 1). It is also interesting to note the appreciable upfield shift for Hd (Δδ = −0.34), which is also shown in Figure 14, and we attribute this to the proximity to one of the Cp* ligands via a plausible CH−π noncovalent interaction. Moreover, the potentially asymmetric Cp*Rh groups are coincident in the NMR (only one signal), even though the nitrogen ring of the indole nucleus 2400

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

200 to 530 nm and showed no optical activity. Thus, complex 2 must be a racemic mixture of enantiomers. [Cp*Rh(Ado)]3(OTf)3 (3)..9a,21 [Cp*RhCl2]2 (101 mg, 0,.163 mmol) and Ag(OTf) (148 mg, 0.616 mmol) were slurried in 15 mL of H2O. After 3 h, the yellow slurry was filtered thru cellulose, which was then washed twice with 5 mL of H2O. To the resulting filtrate was added Ado (88 mg, 0.33 mmol). After the Ado was dissolved, the pH of the yellow solution was adjusted to 7.3, by the addition of 0.1 N NaOH. The solution was then degassed for 20 min with argon and then capped and stirred for 16 h. The reaction mixture was then stripped in vacuo, and the resulting yellow solid was slurried in 3 mL of methanol and filtered. The off-white precipitate (20 mg) was found to be free Ado by 1H NMR. The resulting filtrate was layered with 5 mL of Et2O and stored at −15 °C for recrystallization and yielded 15 mg of a deep orange solid that still contained some small amount of Ado by NMR analysis. 1H NMR (DMSO-d6, ppm): 8.92 (s, 1H, H8), 7.62 (s, 1 H, H2), 5.86 (d, JHH = 6.2 Hz, 1 H, H1′), (b, 1H, −NH), 1.80 (s, 15H, Cp*). [(η5-Cp*)Rh(η1(N7)-9-EG)(MeOH)2](OTf)2 (4).9c To a solution of [Cp*RhCl2]2 (1.01 g, 3.27 mmol of Rh) in anhydrous CH2Cl2 (45 mL) in a glovebox was added AgOTf (1.68 g, 6.52 mmol). The resulting orange slurry was stirred in the absence of light for 19 h and filtered through Celite, and the resulting orange filtrate was stripped to yield 1.60 g (2.98 mmol, 91%) of orange [Cp*Rh(OTf)2]x. [Cp*Rh(OTf)2]x (0.102 g, 0.19 mmol of Rh) was dissolved in 5 mL of MeOH and the solution stirred overnight to form [Cp*Rh(MeOH)3](OTf)2 in situ. The solution was filtered, and 27 mg of 9EG (0.15 mmol) was added. This orange solution was stirred for 18 h, and the volume was reduced to 3 mL. Then 3 mL of Et2O was added and 61 mg (0.09 mmol, 57%) of fine orange needles precipitated after cooling the solution to −30 °C overnight. 1H NMR (500 MHz, DMSO-d6, referenced to TMS, ppm): 11.19 (s, 1H, NH1), 7.99 (s, 1H, H8), 6.85 (s, 2H, NH2), 4.11 (q, 2H, CH2), 3.15 (s, 6H, CH3OH), 1.47 (s, 15H, Cp*), and 1.33 (t, 3H, CH3). Anal. Calcd for C19H24F6N5O7RhS2: C, 31.90; H, 3.38; N, 9.79. Found: C, 31.58; H, 3.37; N, 9.54. It should be noted that the elemental analysis was consistent with [Cp*Rh(9-EG)](OTf)2, indicating that the weakly bound methanol ligands were removed under vacuum. [(η5-Cp*)Rh(η1(N7,O6)-Guo)(OH)](OTf)·4H 2O (5).9a [Cp*Rh(OTf)2]x (0.200 g, 0.373 mmol Rh) and Guo (0.108 g, 0.381 mmol) were slurried in 15 mL of deoxygenated H2O for 16 h. The reaction mixture was stripped and then slurried in 20 mL of CH2Cl2 and filtered to yield 0.240 g (0.316 mmol, 85%) of an orange solid. The analytically pure compound was obtained by recrystallization from minimal methanol. 1H NMR (500 MHz, DMSO-d6, referenced to TMS, ppm): 11.15 (b, 1 H, −NH), 8.94 (s, 1H, H8), 6.92 (b, 2H, −NH2), 5.79 (d, 1H, H1′), 4.37 (m, 1 H, H2′), 4.14 (m, 1H, H3′), 3.93 (m, 1 H, H4′), 3.69 and 3.57 (m, 2H, H5′ and H5″), 1.71 (s, 15H, Cp*). FAB/MS (%): [{Cp*Rh(Guo)}(OTf)]+ (58), [{Cp*Rh(Guo)(H2O)2}(OTf) + H]+ (14), [{Cp*Rh(Guo)} − H]+ (74). Anal. Calcd for C21H37F3N5O13RhS: C, 33.20; H, 4.87; N, 9.22; S, 4.22. Found: C, 33.43; H, 4.27; N, 9.28; S, 4.25. [(η 5 -Cp*)Rh(η 1 (N7)-9-MH)(MeOH) 2 ](OTf) 2 ·CH 3 OH (6).. 9c,d [Cp*Rh(OTf)2]x (0.206 g, 0.38 mmol Rh) was dissolved in 10 mL of MeOH and the solution stirred overnight to form [Cp*Rh(MeOH)3](OTf)2 in situ. The solution was filtered, and 54 mg (0.36 mmol) of 9-MH was added. This orange solution was stirred for 18 h and the volume reduced to 3 mL. Then 3 mL of Et2O was added and the product precipitated out as an orange oil after the solution was cooled to −30 °C overnight. The supernatant was decanted off, and the oil was slurried in 10 mL of Et2O overnight. The slurry was filtered to give 30 mg of an orange-yellow solid, while the Et2O filtrate yielded 123 mg of orange blocks after storing overnight at −30 °C. Both products (153 mg, 0.20 mmol, 56%) were found to be the N7-bound adduct by 1H NMR. 1H NMR (500 MHz, DMSO-d6, referenced to TMS, ppm): 12.65 (s, 1H, NH1), 8.47 (s, 1H, H8), 8.45 (s, 1H, H2), 4.11 (s, 3H, CH 3 ), 1.71 (s, 15H, Cp*). Anal. Calcd for C20H29F6N4O9RhS2: C, 32.01; H, 3.89; N, 7.47. Found: C, 32.31; H, 3.53; N, 7.78.

bodes well for their future use as probes for drug and protein noncovalent interactions.

■

EXPERIMENTAL SECTION

Materials. All organometallic compounds were handled under N2 or Ar using Schlenk or glovebox techniques. All chemicals (highest purity available), including 9-methyladenine, were purchased from commercial vendors and used as received. Physical Measurements. Elemental analyses and FAB/MS measurements were carried out by the Microanalytical Laboratory and Mass Spectrometer Laboratory of the Chemistry Department at University of California, Berkeley. The 1H NMR spectra were recorded on AM400 and AM500 spectrometers. GC-MS experiments were conducted on the HP 5890 Series II GC system with an HP 5971A mass selective detector. Synthesis. [Cp*Rh(OTf)2]x.9a [Cp*Rh(OTf)2]x was prepared in a Vacuum Atmospheres glovebox by the following procedure: [Cp*RhCl2]2 (1.01 g, 3.27 mmol of Rh), which was dried in vacuo at 65 °C for 16 h, was dissolved in 45 mL of CH2Cl2 (twice distilled from CaH2), and anhydrous AgOTf (1.44 g, 6.52 mmol) was added. The resulting orange slurry was stirred in the absence of light for 19 h and filtered through Celite, and the resulting orange filtrate was stripped to yield 1.60 g (91%, 2.98 mmol) of orange [Cp*Rh(OTf)2]x. This material was sufficiently pure for further reactions with water to provide 1 or with the solvents methanol and acetone to give [Cp*Rh(S)3](OTf)2 complexes. The analytically pure material [Cp*Rh(OTf)2]x was obtained by recrystallization from dry CH2Cl2. 1 H NMR (CD2Cl2): 1.59 (s, Cp*). The anhydrous complex was difficult to analyze, because of its hydroscopic nature. Anal. Calcd for C12H15F6O6RhS2·H2O: C, 26.00; H, 3.09. Found: C, 25.90; H, 4.36. Cp*Rh(H2O)3](OTf)2 (1).3a [Cp*Rh(H2O)3](OTf)2 (1) was prepared by the following procedure: [Cp*RhCl2]2 (0.214 g, 0.69 mmol Rh) was dissolved in 20 mL of CH2Cl2, and AgOTf (0.306 g, 1.39 mmol) was added. The resulting orange slurry was stirred for 2 h and filtered through Celite, and the resulting orange filtrate was stripped to a 10 mL total volume. To this orange solution was added 50 μL of H2O (2.78 mmol), and the reaction mixture was stirred for 1 h, over which time a yellow solid precipitated out of solution to provide 0.217 mg (0.38 mmol, 55%) after filtration. When the solution was not stirred and left alone for 24 h, complex 1 crystallized out of the CH2Cl2 solution, in a near-quantitative yield. Alternatively, complex 1 could be generated in situ by reacting [Cp*RhCl2]2 and 4 equiv of AgOTf in H2O for 3 h, filtering through Celite, and removing all volatiles from the filtrate. This method allows in situ generated [Cp*Rh(H2O)3](OTf)2 to be used in reactions. 1H NMR (DMSO-d6, ppm): 1.59 (s, Cp*). 1H NMR (D2O, ppm, referenced to Me4NOH, 3.18 ppm), 1.61 (s, Cp*). 1H NMR (CD2Cl2, ppm): 1.69 (s, Cp*). FAB-MS (glycerol/H2O; m/z (relative intensity)): 688.9 (43) [Cp*Rh(OH)2]2(OTf) − 4H; 540.9 (15) [Cp*Rh(OH)2]2 − 4H; 386.9 (100) [Cp*Rh](OTf). [Cp*Rh(9-MA)]3(OTf)3·6H2O (2).9a [Cp*RhCl2]2 (110 mg, 0.178 mmol) and Ag(OTf) (148 mg, 0.670 mmol) were slurried in 15 mL of H2O. After 3 h, the yellow slurry was filtered thru cellulose, which was then washed twice with 5 mL of H2O. To the resulting filtrate was added 9-MA (54 mg, 0.362 mmol). After the 9-MA was dissolved, the pH of the yellow solution was adjusted to 7.1 by the addition of 0.1 N NaOH. The solution was then degassed for 20 min with argon and then capped and stirred for 16 h. The reaction mixture was then stripped in vacuo, and the resulting orange solid (NMR analysis provided a 75% yield) was recrystallized from 5 mL of methanol to yield 74 mg (0.13 mmol, 26%) of 2. Crystals of 2 were obtained from methanol at −30 °C. 1H NMR (500 MHz, D2O, ppm): 8.83 (s, 1H, H8), 7.67 (s, 1 H, H2), 4.50 (bs, 1H, −NH), 3.64 (s, 3 H, Me), 1.80 (s, 15H, Cp*). FAB-MS (nitrobenzyl alcohol, MS 50; m/z (relative intensity)): 1456.7 (4) [Cp*Rh(9-MA)]3(OTf)3 − OTf; 1307.4 (3) [Cp*Rh(9-MA)] 3 (OTf) 3 − 2OTf; 1158.5 (1) [Cp*Rh(9MA)]3(OTf}3 − 3OTf; 921.3 (10) [Cp*Rh(9-MA)]2(OTf)2 − OTf; 772.3 (16) [Cp*Rh(9-MA)]2(OTf)2 − 2OTf; 386.2 (100) [Cp*Rh(9MA)]. The circular dichroism (CD) spectrum was run in H2O from 2401

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

− MC − H2O]+ (10), ; [(Cp*Rh(μ-OH))2(OTf)]+ (23), [(Cp*Rh(μOH))2(OTf) − H2O]+ (13), [Cp*Rh(MC)(OTf)]+ (20), [Cp*Rh(MC)]+ (82), [Cp*Rh − H]+ (25), [OTf]+ (100), [MC + H]+ (30). Anal. Calcd for C32H54F6N6Rh2O14S2: C, 33.99; H, 4.81; N, 7.43. Found: C, 33.53; H: 4.40; N, 7.82. Orange parallelepipeds of 11 were obtained from water at room temperature for X-ray analysis. [RhI(η1(N3)-1-MT)2]2·[(η5-Cp*)Rh(μ-OH)3Rh(η5-Cp*)]3(OH) (12).9e To a solution of [Cp*RhCl2]2 (0.10g, 0.16 mmol) in H2O (15 mL, degassed once) was added AgOTf (0.175 g, 0.68 mmol). The reaction mixture was stirred at ambient temperature for 3 h, and then it was filtered. The resulting filtrate was treated with 1-MT (0.045 g, 0.32 mmol); after all the 1-MT had dissolved, the pH was adjusted to 10 by the addition of 0.1 N NaOH. The final reaction mixture was degassed and stirred at 25 °C overnight and at 60 °C for 2 h to drive the reaction to completion. The solution turned orange, and by reduction of the volume of this reaction mixture to ∼3 mL, complex 12 crystallized at 4 °C as orange plates (20%). 1H NMR (400 MHz, CD3OD, ppm): 7.24 (s, 1H, C6−H), 3.29 (s, 3H, N1−Me), 1.85 (s, 3H, C5−Me), 1.62 (s, 22.5H, Cp*). Anal. Calcd for 12·10.5H2O (C84H128N8O18Rh8·10.5H2O): C, 39.5; H, 5.84; N, 4.39. Found: C, 39.3; H, 5.65; N, 4.55. The formation of [(η5-Cp*)Rh(μ-OH)3Rh(η5Cp*)](OH) was observed only when the solution was supersaturated; i.e., reduced volume on crystallization of the product, {[RhI(η1(N3)-1MT)2]2·[(η5-Cp*Rh)2(μ-OH)3]3}(OH) (12). The product was found to be unstable in MeOH-d4 after 2 h at room temperature. Identification of Cp*OH via GC-MS. The aforementioned orange reaction mixture was distilled at 50 °C in vacuo, the distillate was passed through a C-18 cartridge, and finally the Cp*OH was eluted off the cartridge with a small amount of MeOH. The distillate of the reaction mixture was analyzed by GC-MS techniques and provided information that Cp*OH was formed (m/z 151 and 135 for [M − H]+ and [M − OH]+) during the reaction, which further provided clear evidence for the loss of the Cp* ligand from Rh3+. 1 H NMR Experiments on the Reaction of [Cp*Rh(H2O)3](OTf)2 and 1-Methylthymine in D2O, as a Function of pD: Complex 12. Thus, 18 mg (0.031 mmol) of [Cp*Rh(H2O)3](OTf)2 and 6.1 mg (0.043 mmol) of 1-methylthymine (1-MT) were dissolved in 300 μL of D2O at pD 10.45. The pD was again measured and found to be pD 8.65, after the two reactants were dissolved, providing a drop of 1.8 pD units. The 1H NMR spectrum showed that no reaction had taken place with free 1-MT at 7.24 ppm (C6−H),and 3.15 ppm (N-Me), with the CpRh* signal being at 1.67 ppm. The [Cp*Rh2(μ-OH)3]+ dimer signal was at 1.41 ppm. When the pD was raised to 13.6 with NaOD and the mixture heated for 2 h at 60 °C, as before, Cp*OH precipitated and was filtered, while there was a pronounced upfield shift of the C6−H to 7.10 ppm, or an upfield shift of Δδ = −0.14 ppm, while N−Me also moved upfield to 3.10 (Δδ = −0.05 ppm) and C5Me was now at 1.63 ppm (Δδ = −0.04 ppm), as expected from the 1MT−Cp* π−π interactions shown by the X-ray structure. The [(Cp*Rh)2(μ-OH)3]OH dimer signal was at 1.38 ppm (from 1.41 ppm, Δδ = −0.03 ppm), which verified that the solid-state and solution structures of complex 12 were similar. Control Experiment: Reaction of 1-MT with [Rh(H2O)3](OTf)3. [Rh(H2O)3](OTf)3 was formed from the reaction of Rh(Cl)3·3H2O with 3 equiv of AgOTf in water, while AgCl was filtered from the solution. Reaction of [Rh(H2O)3](OTf)3 with 1-MT at pH 10 did not provide complex 12, via 1H NMR spectroscopy, but Rh(OH)3 was formed as a yellow precipitate and was identified via elemental analysis. Anal. Calcd for Rh(OH)3·1.2H2O: Rh, 58.6; H, 3.08. Found: Rh, 58.1; H, 2.71. [Cp*Rh(μ2-η1(N1):η2(N7,N6)-2′-deoxyAdo)]3(OTf)3 (13).21 To a solution of [Cp*RhCl2]2 (0.20 g, 0.324 mmol) in H2O (20 mL, degassed once) was added AgOTf (0.30 g, 1.2 mmol). The reaction mixture was stirred at ambient temperature for 3 h, and then it was filtered. 2′-Deoxyadenosine (2′-deoxyAdo, 0.15g, 0.60 mmol) was added to the filtrate, and after it had dissolved, the pH was adjusted to 7.3 by the addition of 0.1 N NaOH. The final reaction mixture was degassed and stirred overnight. The reaction mixture was then stripped in vacuo, and the yellow residue was slurried in MeOH (8 mL). The slurry was filtered and the filtrate was treated with diethyl ether (12

[(η5-Cp*)Rh(η2(N7,O6)-9-MH)(H2O)2](OTf)2 (7).9d The reaction of 1 and 9-MH (1/1) was conducted in an NMR tube at pHs from 2.45 to 5.13. Complex 7 was identified by 1H NMR and showed η2(N7,O6) binding with H8 (8.48 ppm, Δδ = 0.44) and H2 (8.35 ppm, Δδ = 0.17) from free 9-MH, H8 8.04 ppm and H2 8.18 ppm, with selective deuteration of H8 in refluxing D2O verifying the H8, H2 chemical shift designations. [(η5-Cp*)Rh(μ2-η1(N1):η2(N7,O6)-9-MH)]3(OTf)3 (8).9d To a solution of [Cp*RhCl2]2 (0.10 g, 0.16 mmol) in H2O (14 mL, degassed once, pump−freeze−thaw) was added AgOTf (0.17 g, 0.66 mmol). The reaction mixture was stirred at ambient temperature for 3 h, and then it was filtered. To the filtrate was added 9-MH (0.050 g, 0.32 mmol), and after all the 9-MH was dissolved, the pH was adjusted to 6.1 by the addition of 0.10 N NaOH. The final reaction mixture was degassed and stirred at 25 °C overnight and at 50 °C for 45 min. The reaction mixture was then stripped in vacuo, and the yellow residue was slurried in MeOH (5 mL). The product was precipitated out of MeOH, collected on a frit, washed with a small amount of MeOH, and dried in vacuo to give yellow microcrystals (70 mg, 37.5%). 1H NMR (500 MHz, MeOH-d4): δ 8.77 (s, 1H, H8), 7.81 (s, 1H, H2), 3.76 (s, 3H, CH 3 ), and 1.89 (s, 15H, Cp*). Anal. Calcd for C51H60F9N12O12Rh3S3·8H2O: C, 34.9; H, 4.34; N, 9.59. Found: C, 34.6; H, 3.99; N, 9.68. [(η5-Cp*)Rh(μ2-η1(N1):η2(N7,O6)-9-EH)]3(OTf)3 (8a).9d The same procedures as the preparation of 8 were used, except that the starting material 9-EH was employed (43.8%). Single crystals were grown from MeOH/H2O. 1H NMR (400 MHz, MeOH-d4, ppm): 8.86 (s, 1H, H8), 7.80 (s, 1H, H2), 4.20 (q, 2H, CH2), 1.89 (s, 15H, Cp*), 1.44 (t, 3H, CH3). FAB/MS (%): [M + H − OTf]+ (15.2), [{Cp*Rh(9EH)}2(OTf)2 − OTf]+ (20), [{Cp*Rh(9-EH)}2(OTf)2 − 2OTf]+ (45%). Anal. Calcd for C54H66F9N12O12Rh3S3·2H2O: C, 38.4; H, 4.15; N, 9.97. Found: C, 38.1; H, 3.97; N ,9.71. [(η5-Cp*)Rh-(η1(N1)-(μ-OH)-9-MH)]2 (9).9d The same procedures as for the preparation of 8 were used, except that the pH was adjusted to 10.2 and X-ray-quality single crystals were obtained by concentrating the aqueous reaction mixture (yield 60%). 1H NMR (400 MHz, MeOH-d4, ppm): 8.17 (s, 1H, H8), 7.81 (s, 1H, H2), 3.70 (s, 3H, CH3), 1.76 (s, 15H, Cp*). Anal. Calcd for C32H42N8O4Rh2·3H2O: C, 44.5; H, 5.57; N, 12.99. Found: C, 44.3; H, 5.29; N, 12.56. [(η 5 -Cp*)Rh(η 1 (N3)-1-MC)(η 2 (O2,N3)-1-MC)](OTf) 2 (10). 9b [Cp*RhCl2]2 (0.214 g, 0.69 mmol of Rh) was dissolved in 20 mL of CH2Cl2, and AgOTf (0.357 g, 1.39 mmol) was added. The resulting orange slurry was stirred for 2 h and filtered through Celite, and the resulting orange filtrate was stripped to 10 mL. To this orange solution was added 50 μL of H2O (2.78 mmol), and the reaction mixture was stirred for 1 h, over which time a yellow solid precipitated out of solution to provide 0.217 g of [Cp*Rh(H2O)3](OTf)2 (0.38 mmol, 55%) after filtration. This material was sufficiently pure for further reactions. [Cp*Rh(H2O)3](OTf)2 (0.206 g, 0.36 mmol) and 1-MC (0.091 g, 0.73 mmol) were slurried for 36 h in 20 mL of acetone. The solvent of the orange solution was stripped in vacuo, and the remaining orange solid was slurried in 20 mL of CH2Cl2 for 2 h and then filtered and dried to give 0.195 g (67%) of complex 10. Analytically pure 10 was obtained by recrystallization from minimal methanol. 1H NMR (400 MHz, DMSO-d6, ppm): 8.52 (b, 1H, NH), 7.82 (d, 1H, H6), 7.72 (b, 1H, NH), 5.84 (d, 1H, H5), 3.28 (s, 3H, Me), 1.71 (s, 15H, Cp*). FAB/MS (%): [M − OTf − MC]+ (52), [Cp*Rh(OTf)]+ (18), [Cp*Rh(MC)]+ (100), [Cp*Rh − H]+ (30), [MC + H]+ (40). Anal. Calcd for C22H29F6N6RhO8S2: C, 33.60; H, 3.72; N, 10.67. Found: C, 33.35; H, 3.75; N, 10.33. Orange needles of 10·1.5MeOH were obtained from a methanol/Et2O solution at −30 °C, under an inert atmosphere. [(η5-Cp*)Rh(η1(N3)-1-MC)(μ-OH)]2(OTf)2 (11).9d Complex 10 (0.043 g) was dissolved in 5 mL of deoxygenated water (pH 5.1), and the solution was stirred for 36 h, resulting in precipitation of a yellow-orange solid. The volume of the reaction mixture was reduced to ca. 2 mL, and the orange precipitate (0.023 g, 81%) was collected by filtration and dried for 18 h under vacuum. 1H NMR (400 MHz, DMSO-d6, ppm): 7.45 (d, 1H, H6), 7.18 (b, 1H, NH), 5.27 (d, 1H, H5), 3.18 (s, 3H, CH3), 1.78 (s, 15H, Cp*). FAB/MS (%): [M − OTf 2402

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

mL) to precipitate the product. The supernatant was discarded and the product was stripped in vacuo to give a yellow solid (59% yield). 1H NMR (400 MHz, D2O, referenced to Me4NOH, 3.180 ppm): 8.77 (d, 1H, H8), 7.65 (d, 1H, H2), 6.24 (m, 1H, H1′), 4.58 (m, 1H, H3′), 4.10 (m, 1H, H4′), 3.74 (m, 2H, H5′ and H5″), 2.67 (m, 1H, H2′), 2.49 (m, 1H, H2″), 1.85 (s, 15H, Cp*). FAB/MS (%): [M − OTf]+ (5.4), [M − OTf − dAdo]+ (dAdo = 2′-deoxyadenosine, 3.1), [M − OTf − 2dAdo]+ (1.7), {[Cp*Rh(dAdo − H)]2(OTf)2 − OTf}+ (100). Anal. Calcd for C63H81F9N15O18Rh3S3·8H2O: C, 36.8; H, 4.7; N, 10.2. Found: C, 36.9; H, 4.5; N, 9.8. 1 H NMR Competition Studies. 1H NMR competition studies of nucleosides and nucleotides for the Cp*Rh tris(aqua) complex 1 in D2O were performed at pH 7 by using a phosphate buffer. A typical sample preparation is as follows: appropriate amounts of nucleosides, for example, Ado and Guo (0.0071 mmol each), and [Cp*Rh(D2O)3](OTf)2 (0.0071 mmol) in a 5 mm NMR tube were dissolved in 1.0 mL of D2O. To this was added 20 μL of a 3 M phosphate buffer (pH 7) solution in D2O and 5 μL of a 6 × 10−2 M Me4NOH solution in D2O, as the internal reference, with the methyl proton resonance set at 3.180 ppm. Similar experiments with [Cp*Rh((η2(N7-O6)Guo)(OH)](OTf) (5) with 1 equiv of Ado in D2O at pD 7.3 provided the Ado cyclic trimer complex 3 and displacement of the Guo ligand. A control experiment showed that there was no apparent interaction between [Cp*Rh(D2O)3](OTf)2 and the phosphate buffer at pH 7. The results indicated that the Ado cyclic trimer prevailed, with the stipulation that Guo only forms a mononuclear and not a cyclic trimer, due to the presumed steric effect of the 2-NH2 group. In several other competitive binding studies with the aforementioned parameters, 1H and 31P NMR experiments were conducted individually with [Cp*Rh(D2O)3](OTf)2 (0.0071 mmol) in a 5 mm NMR tube, along with the nucleotides (0.0071 mmol) 5′-guanosine monophosphate (5′-GMP), 5′-cytidine monophosphate (5′-CMP), and 5′-thymidine monophosphate (5′-TMP) at pD 7.2, dissolved in 1.0 mL of D2O. The 1H and 31P NMR spectra showed broadened and shifted signals at pD 7.2 for 5′-GMP but only broadened 31P NMR spectra for 5′-CMP and 5′-TMP, and no shifts in the 1H NMR spectra. The broadened and only slightly shifted 31P NMR spectra may be consistent with a partial weak interaction of the P−O− group with a Cp*Rh metal center, as shown in Figure 11. Alternatively, 5′-GMP showed an 1H NMR spectrum with a downfield shift for H8 of Δδ = 0.6 ppm, indicative of N7 bonding, as well as a broadened, slightly shifted 31P NMR signal for a similar weak interaction of the Cp*Rh metal center with the P−O− group, as observed for 5′-AMP. Moreover, 5′-GMP competed with Cp*Rh-5′-AMP, and this provided [Cp*Rh((η1(N7)-5′-GMP)(H2O)2](OTf)2, with displacement of 5′AMP. Furthermore, 5′-CMP or 5′-TMP did not alter the 1H NMR spectrum of the Cp*Rh(5′-AMP) complex. 1 H NMR Analysis and Molecular Recognition Docking Experiments with Complexes 11 and 13 as Hosts.21,26 Methods. An initial structure of the host−guest complex 11−L-Trp (Figure 14) was obtained via rigid-body docking of several conformers of L-Trp to the organometallic host 11. We used the program Molfit, employing a small grid interval of 0.78 Å, which was appropriate for small-molecule docking experiments. The docking results for the conformers were combined and sorted by the complementarity score. All of the results were statistically analyzed, providing unique values for all of the solutions. The best docking solutions were energy minimized, restricting the host complex 11 to its initial X-ray structure (the Rh ions were omitted) and allowing free movement only of the host molecule L-Trp. We used the CVFF force field in the Discover module of the MSI package. The rigid-body docking produced an interesting model structure, which was statistically highly unique. In this structure, the carboxyl terminus of L-Trp was at a hydrogen bond distance from the amino group of one of the 1-MC molecules in the host complex, while the amino terminus of L-Trp was at a hydrogen bond distance from the carbonyl group of the second methylcytosine molecule in the host complex. A hydrogen bond could also be formed between the amino terminus of L-Trp and the μ-OH moiety in the host complex 11. By preserving these hydrogen bond interactions, the docking produced several host−guest complexes with different conformers of

L-Trp, which were energy minimized. The structure presented in Figure 14 was one of two similar lowest-energy structures. Similar molecular docking experiments were conducted with host 13 and LTrp (Figure 12, left). The 1H NMR techniques for selective, noncovalent hydrogen chemical shifts can be found in refs 21, 26, and 27.

■

ASSOCIATED CONTENT

S Supporting Information *

CIF files giving crystallographic data for the crystal structure determinations. This material is available free of charge via the Internet at http://pubs.acs.org. Original structural information was previously deposited with the Cambridge Crystallographic Database. REFCODES are as follows: compound 2, JUTCIK; compound 6, PIBNET; compound 8, ZAPRIR; compound 9, ZAPROX; compound 10, WAKZEN; compound 11, WAKZIR; compound 12, ZEKYET. In the Supporting Information, compound 2 was renumbered to correspond to the original numbering scheme. Compounds 11 and 12 were re-refined.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail for M.M.O.: [email protected] (for all X-ray structures). *E-mail for R.H.F.: rhfi[email protected] (for bioorganometallic chemistry). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The studies at LBNL were generously supported by LBNL Laboratory Directed Research and Development Funds to R.H.F. and the Department of Energy under Contract No. DEACO2-05CH11231. We also acknowledge the reviewers for insightful comments that were helpful in the writing of the manuscript and the drawing of the figures.

■

REFERENCES

(1) (a) Tullius, T. D. In Metal-DNA Chemistry; Tullius, T. D., Ed.; American Chemical Society: Washington, DC, 1989; ACS Symposium Series 402, Chapter 1 and references therein. (b) Barton, J. K. Comments Inorg. Chem. 1985, 3, 321 and references therein. (c) Pyle, A. M.; Barton, J. K. in Progress in Inorganic Chemistry, Bioinorganic Chemistry, Lippard, S. J., Ed.; J. Wiley and Sons: New York, 1990, vol 38, p 413 and references therein. (d) Marzilli, L. G. In Progress in Inorganic Chemistry; Lippard, S. J., Ed.; Wiley: New York, 1977; Vol. 23, p 255, and references therein. (e) Howe-Grant, M. E.; Lippard, S. J. Met. Ions Biol. Syst. 1980, 11, 63 and references therein. (f) RaudaschlSieber, G.; Schollhorn, H.; Thewalt, U.; Lippert, B. J. Am. Chem. Soc. 1985, 107, 3591. (g) Reily, M. D.; Hambley, T. W.; Marzilli, L. G. J. Am. Chem. Soc. 1988, 110, 2999. (h) Alink, M. A.; Nakahara, H.; Hirano, T.; Inagaki, K.; Hakanishi, M.; Kidani, Y.; Reedijk, J. Inorg. Chem. 1991, 30, 1236 and references therein. (i) Reily, M. D.; Marzilli, L. G. J. Am. Chem. Soc. 1986, 108, 8299. (j) Qu, Y.; Farrell, N. J. Am. Chem. Soc. 1991, 113, 4851 and references therein. (k) Caradonna, J. P.; Lippard, S. J.; Gait, M. J.; Singh, M. J. Am. Chem. Soc. 1982, 104, 5793. (l) Marcelis, A. T. M.; den Hartog, J. H. J.; Reedijk, J. J. Am. Chem. Soc. 1982, 104, 2664. (m) Sherman, S. E.; Gibson, D.; Wang, A. H.-J.; Lippard, S. J. J. Am. Chem. Soc. 1988, 110, 7368. (n) Mukundan, S., Jr; Xu, Y.; Zon, G.; Marzilli, L. G. J. Am. Chem. Soc. 1991, 113, 3021. (o) Pneumatikakis, G.; Hadjiliadis, H. J. Chem. Soc., Dalton Trans. 1979, 596. (p) Rainen, L.; Howard, R. A.; Kimball, A. P.; Bear, J. L. Inorg. Chem. 1975, 14, 2752. (q) Torres, L. M.; Marzilli, L. G. J. Am. Chem. Soc. 1991, 113, 4678. (r) Scheller, K. H.; Mitchell, P. R.; Prijs, B.; Sigel, H. J. Am. Chem. Soc. 1981, 103, 247. (s) Sigel, H. In 2403

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404

Organometallics

Article

Metal-DNA Chemistry; Tullius, T. D., Ed.; American Chemical Society: Washington, DC, 1989; ACS Symposium Series 402, Chapter 11 and references therein. (t) Hodgson, D. J. In Progress in Inorganic Chemistry; Lippard, S. J., Ed.; Wiley: New York, 1977; Vol. 23, p 211. (u) Muller, J. G.; Zheng, P.; Rokita, S. E.; Burrows, C. J. J. Am. Chem. Soc. 1996, 118, 2320 and references therein. (2) (a) Fish, R. H.; Jaouen, G. Organometallics 2003, 22, 2166 and references therein. (b) Fish, R. H. Aus. J. Chem. 2010, 63, 1505 and references therein. (3) (a) Eisen, M. S.; Haskel, A.; Chen, H.; Olmstead, M. M.; Smith, D. P.; Maestre, M. F.; Fish, R. H. Organometallics 1995, 14, 2806 and references therein. (b) Ogo, S.; Chen, H.; Olmstead, M. M.; Fish, R. H. Organometallics 1996, 15, 2009. (c) Ogo, S.; Buriez, O.; Kerr, J. B.; Fish, R. H. J. Organomet. Chem. 1999, 589, 66. (4) (a) Kramer, R.; Polborn, K.; Robl, C.; Beck, W. Inorg. Chim. Acta 1992, 198−200, 415. (b) Severin, K.; Bergs, R.; Beck, W. Angew. Chem., Int. Ed. 1998, 37, 1086 and references therein. (c) Beck, W.; Kottmair, N. Chem. Ber. 1976, 109, 970. (d) Singh, M. M.; Rosopulos, Y.; Beck, W. Chem. Ber. 1983, 116, 1364. (e) Krämer, R.; Polborn, K.; Beck, W. J. Organomet. Chem. 1991, 410, 111. (5) (a) Fish, R. H.; Kimmel, E. C.; Casida, J. E. J. Organomet. Chem. 1975, 93, C1. (b) Fish, R. H.; Kimmel, E. C.; Casida, J. E. J. Organomet. Chem. 1976, 118, 41. (c) Fish, R. H.; Casida, J. E.; Kimmel, E. C. ACS Symp. Ser. 1978, 82, 82. (d) Fish, R. H.; Price, R. T. Organometallics 1989, 8, 225. (6) Parvin, B. A.; Maestre, M. F.; Fish, R. H.; Johnston, W. E. A Method and Apparatus for Accurately Manipulating an Object During Microelectrophoresis. U. S. Patent No. US5671086(A), University of California, September 23, 1997. (7) (a) Toney, J. H.; Marks, T. J. J. Am. Chem. Soc. 1985, 107, 947. (b) Kuo, L. Y.; Kanatzdis, M. G.; Marks, T. J. J. Am. Chem. Soc. 1987, 109, 7207. (c) Toney, J. H.; Brock, C. P.; Marks, T. J. J. Am. Chem. Soc. 1986, 108, 7263. (d) Kuo, L. Y.; Kanatzdis, M. G.; Sabat, M.; Tipton, A. L.; Marks, T. J. J. Am. Chem. Soc. 1991, 113, 9027. (8) (a) Ang, W. H.; Daldini, E.; Scolaro, C.; Scopelliti, R.; JuilleratJeannerat, L.; Dyson, P. J. Inorg. Chem. 2006, 45, 9006. (b) Allardyce, C. S.; Dyson, P. J.; Ellis, D. J.; Heath, S. L. Chem. Commun. 2001, 1396. (c) Scolaro, C.; Bergamo, A.; Brescacin, L.; Delfino, R.; Cocchietto, M.; Laurenczy, G.; Geldbach, T. J.; Sava, G.; Dyson, P. J. J. Med. Chem. 2005, 48, 4161. (d) Scolaro, C.; Chaplin, A. B.; Hartinger, C. G.; Bergamo, A.; Cocchietto, M.; Keppler, B. K.; Sava, G.; Dyson, P. J. Dalton Trans. 2007, 43, 5065. (e) Dorcier, A.; Hartinger, C. G.; Scopelliti, R.; Fish, R. H.; Keppler, B. K.; Dyson, P. J. Inorg. Biochem. 2008, 102, 1066 and references therein. (f) Annen, P.; Schildberg, S.; Sheldrick, W. S. Inorg. Chim. Acta 2000, 307, 115. (g) Korn, S.; Sheldrick, W. S. Inorg. Chem. Acta 1997, 254, 85. (h) Korn, S.; Sheldrick, W. S. J. Chem. Soc., Dalton Trans. 1997, 2191. (i) Sheldrick, W. S.; Hagen-Eckard, H. S.; Heeb, S. Inorg. Chim. Acta 1993, 206, 15. (j) Yamanari, K.; Yamamoto, S.; Ito, R.; Kushi, Y.; Fuyuhiro, A.; Kubota, N.; Fukuo, T.; Arakawa, R. Angew. Chem., Int. Ed. 2001, 40, 2268. (9) For preliminary communications see: (a) Smith, D. P.; Baralt, E.; Morales, B.; Olmstead, M. M.; Maestre, M. F.; Fish, R. H. J. Am. Chem. Soc. 1992, 114, 10647. (b) Smith, D. P.; Olmstead, M. M.; Noll, B. C.; Maestre, M. F.; Fish, R. H. Organometallics 1993, 12, 593. (c) Smith, D. P.; Griffin, M. T.; Olmstead, M. M.; Maestre, M. F.; Fish, R. H. Inorg. Chem. 1993, 32, 4677. (d) Chen, H.; Olmstead, M. M.; Smith, D. P.; Maestre, M. F.; Fish, R. H. Angew. Chem., Int. Ed. Engl. 1995, 34, 1514. (e) Chen, H.; Olmstead, M. M.; Maestre, M. F.; Fish, R. H. J. Am. Chem. Soc. 1995, 117, 9097. (10) Gagnon, P. C.; Beauchamp, A. L. Acta Crystallogr., Sect. B 1977, 33, 1448. (11) De Meester, P.; Goodgame, D. M. L.; Skapski, A. C.; Warkne, Z. Biochim. Biophys. Acta 1973, 324, 301. (12) Fish, R. H.; Kim, H.-S.; Babin, J. E.; Adams, R. D. Organometallics 1988, 7, 2250. (13) (a) Dunbar, K. R.; Matonic, J. H.; Saharan, V. P.; Crawford, C. A.; Christou, G. J. Am. Chem. Soc. 1994, 116, 2201 and references

therein. (b) Chifotides, H. T.; Koshlap, K. M.; Pérez, L. M.; Dunbar, K. R. J. Am. Chem. Soc. 2003, 125, 10703 and references therein. (14) Johnson, D. A. Some Thermodynamic Aspects of Inorganic Chemistry; Cambridge University Press: Cambridge, U.K., 1982. (15) Munns, A. R. I.; Tollin, P. Acta Crystallogr., Sect. B 1970, 26, 1101. (16) Frommer, G.; Mutikainen, I.; Pesch, F. J.; Hillgeris, E. C.; Preut, H.; Lippert, B. Inorg. Chem. 1992, 31, 2429. (17) (a) Faggiani, R.; Lippert, B.; Lock, C. J. L.; Speranzini, R. A. J. Am. Chem. Soc. 1981, 103, 1111. (b) Schollhorn, H.; Thewalt, U.; Lippert, B. J. Am. Chem. Soc. 1986, 108, 3680. (18) Kosturko, L. D.; Folzer, C.; Stewart, R. F. Biochemistry 1974, 13, 3949. (19) Raudaschi-Sieber, G.; Lippert, B. Inorg. Chem. 1985, 24, 2426. (20) (a) Kickham, J. E.; Loeb, S. J.; Murphy, S. L. J. Am. Chem. Soc. 1993, 115, 7031 and references therein. (b) Kickham, J. E.; Loeb, S. J. Inorg. Chem. 1995, 34, 5656. (21) Chen, H.; Ogo, S.; Fish, R. H. J. Am. Chem. Soc. 1996, 118, 4993. (22) (a) Power, P. P. Comments Inorg. Chem. 1989, 8, 177 and references therein. (b) Bartlett, R. A.; Chen, H.; Power, P. P. Angew. Chem., Int. Ed. Engl. 1989, 28, 316. (c) Chen, H.; Bartlett, R. A.; Dias, H. V. R.; Olmstead, M. M.; Power, P. P. J. Am. Chem. Soc. 1989, 111, 4338. (d) Chen, H.; Bartlett, R. A.; Olmstead, M. M.; Power, P. P.; Shoner, S. C. J. Am. Chem. Soc. 1990, 112, 1048. (e) Bartlett, R. A.; Power, P. P. J. Am. Chem. Soc. 1987, 109, 7563. (f) Andersen, R. A.; Faegri, K., Jr.; Green, J. C.; Haaland, A.; Lappert, M. F.; Leung, W.-P.; Rypdal, K. Inorg. Chem. 1988, 27, 1782. (23) Lappert, M. F.; Power, P. P.; Sanger, A. R.; Srivastava, R. C. Metal and Metalloid Amides; Ellis-Horwood: Chichester, U.K., 1980. (24) Paneque, M.; Maitlis, P. J. Chem. Soc., Chem. Commun. 1989, 105. (25) Smith, D. P.; Kohen, E.; Maestre, M. F.; Fish, R. H. Inorg. Chem. 1993, 32, 4119. (26) For a preliminary communication using 11 as a host: Elduque, A. I.; Carmona, D.; Oro, L. A.; Eisenstein, M.; Fish, R. H. J. Organomet. Chem. 2003, 668, 123. (27) Ogo, S.; Nakamura, S.; Chen, H.; Isobe, K.; Watanabe, Y.; Fish, R. H. J. Org. Chem. 1998, 63, 7151.

■

NOTE ADDED AFTER ASAP PUBLICATION This paper was published on the Web on April 22, 2014, with an error in the name of compound 5. The corrected version was reposted on April 23, 2014.

2404

dx.doi.org/10.1021/om500106r | Organometallics 2014, 33, 2389−2404