RESPONSE

A response to this article has been published. To view the response, click here.

Rebuttal to the article Pathological crystal structures

A section in the Acta Crystallographica Section C article by Raymond & Girolami [Acta Cryst. (2023), C79, 445–455] stated that the product of the reaction of [(Cp*Rh)₂(μ-OH)₃]⁺ (Cp* is 1,2,3,4,5-penta­methyl­cyclo­penta­diene) with 1-methyl­thymine (1-MT) at pH 10 and 60 °C, to synthesize the anionic com­ponent [Rh^I(η¹-N³-1-MT)₂]⁻, was not an Rh^I com­plex, but rather an Ag^I com­plex, due to the use of silver triflate (AgOTf) to remove Cl⁻ from [Cp*RhCl₂]₂ to synthesize [Cp*Rh(H₂O)₃](OTf)₂, a water-soluble crystalline com­plex. We will clearly show that this premise, as stated, is invalid, while the authors have simply avoided several important facts, including that Cp*OH, a reductive elimination product, at pH 10 and 60 °C, was unequivocally identified, thus leading to the Rh^I anionic com­ponent [Rh^I(η¹-N³-1-MT)₂]⁻. More importantly, AgOH, from the reaction of NaOH at pH 10 with any potentially remaining AgOTf, after the AgCl was filtered off, would be insoluble in water. Furthermore, a control experiment with the inorganic com­plex Rh(OH)₃, reacting with 1-methyl­thymine at pH 10, provided no product, and this bodes well for a similar fate with AgOTf and 1-methyl­thymine, i.e. at pH 10, AgOTf would again be converted to the water-insoluble AgOH; therefore, no reaction would occur! Finally, a ¹H NMR spectroscopy experiment was carried out with synthesized and crystallized [Cp*Rh(H₂O)₃](OTf)₂ in D₂O at various pD values; at pD 8.65 no reaction took place, while at pD 13.6, and at 60 °C for 2 h, a reductive elimination reaction caused the precipitation of Cp*OH. The subsequent ¹H NMR spectrum clearly demonstrated, in the absence of any Ag^I com­plexes, that the solution structure and the X-ray crystals in D₂O were similar. A postulated mechanism for this novel anionic com­ponent structure, as published previously [Smith et al. (2014). Organometallics, 33, 2389–2404], will be presented, along with the experimental data, to insure the credibility of our results. We will also answer the comments in the response of Drs Raymond and Girolami to this rebuttal.

1. Results of our bioorganometallic chemistry reaction with [Cp*Rh(H₂O)₃](OTf)₂ and 1-methyl­thymine in water at 60 °C and pH 10

Raymond and Girolami (2023) apparently did not fully com­prehend that the AgOTf was used to form [Cp*Rh(H₂O)₃](OTf)₂, in situ, in this example, while this aqua com­plex was used for the reactions with nucleobases, nucleosides, and nucleotides, as a function of pH (Chen et al., 1995; Smith et al., 2014). Moreover, the formed AgCl was not soluble in water; it precipitated out of the aqueous reaction mixture and was filtered off. These authors have also presented no definitive information to justify their premise, since AgOH, a water-insoluble Ag^I com­plex at pH 10, would be inactive for any reaction with 1-methyl­thymine, or even a plausible mechanism for Ag^I replacing Rh^I in this bioorganometallic chemistry reaction (Raymond & Girolami, 2023).

Furthermore, the formed [Cp*Rh(H₂O)₃](OTf)₂ (pH 2–5) has been unequivocally shown to have a pH-dependent equilibrium, and provides a μ-OH com­plex, i.e. [(CpRh)₂(μ-OH)₂(H₂O)₂](OTf)₂, at pH 5–7, and [(CpRh)₂(μ-OH)₃](OTf/OH) at pH 7–10 (Eisen et al., 1995). Thus, the pH was raised to pH 10 and provided [(CpRh)₂(μ-OH)₃](OTf/OH), which was reacted with 1-methyl­thymine, and eventually, by raising the temperature to 60 °C, this caused a reductive elimination reaction to occur that gave Cp*OH, identified by GC–MS (gas chromatography–mass spectrometry), after purification. The ironical aspect of this important finding of Cp*OH from a reductive elimination reaction was totally ignored by Ray­mond & Girolami (2023) and totally negated their supposition of a linear Ag^I versus the linear Rh^I characterization of this aspect of the X-ray structure (Fig. 1) (Chen et al., 1995; Smith et al., 2014).

Moreover, they stated that Cp*Rh com­plexes do not, in general, reductively eliminate the Cp* group from Cp*Rh. However, we found a precedent in the literature, in that a Cp*RhH₃(SiEt₃) com­plex in the presence of an excess of Me₃P gave Cp*H, i.e. reductive elimination to provide an Rh^I com­plex (Paneque & Maitlis, 1989).

The unique stabilization provided by the π–π inter­actions of the cationic portion, [(Cp*Rh)₂(μ-OH)₃]₃OH, which pro­tec­ted the 12e Rh^I atom from reactions with H₂O, was to emphasize the special structural aspects of these bioorganometallic complexes. Furthermore, it was our supposition, in this particular case, that the electrostatic inter­action between the anionic and cationic com­ponents, i.e. [Rh^I(η¹-N³-1-MT)₂]⁻ and [(Cp*Rh)₂(μ-OH)₃]⁺, further stabilized adduct formation (Fig. 2).

The putative mechanism for the reductive elimination of Cp*OH is shown in Fig. 3. This mechanism provided a logical pathway for the Rh^I com­plex [Rh^I(η¹-N³-1-MT)₂]⁻.

2. Experiments to provide further proof of the structure of the anionic com­ponent [Rh^I(η¹-N³-1-MT)₂]⁻

In order to provide more information on the [Rh^I(η¹-N³-1-MT)₂]⁻ structure, in contention with the Raymond and Girolami premise of Ag^I rather than Rh^I, we had conducted several control experiments, never mentioned by these authors (Smith et al., 2014). The first experiment included the identification of Cp*OH, the reductive elimination product, via GC–MS (see Experimental, Section 5.3).

Secondly, the reaction of pre-synthesized Rh(OH)₃ with 1-methyl­thymine at pH 10 did not provide any 1-methyl­thymine com­plexes, except Rh(OH)₃ had precipitated from the aqueous solution. This result bodes well for a similar result with AgOH, and again, further solidifies the invalid premise of Ag^I versus Rh^I replacement put forward by Raymond and Girolami (see Experimental, Section 5).

The most revealing experiment for the Rh^I anionic com­ponent was a ¹H NMR spectroscopy study utilizing the purified and crystalline aqua com­plex [Cp*Rh(H₂O)₃](OTf)₂, i.e. no Ag^I was present. This experiment was conducted in D₂O and at pD 8.65, with no reaction taking place according to ¹H NMR spectroscopy; however, when the pD was raised to 13.6, and the reaction mixture heated to 60 °C for 2 h, Cp*OH again precipitated, and was filtered off. The pronounced upfield shifts were observed for π–π inter­actions, in concert with the X-ray crystal structure (see Experimental for details).

3. Conclusions

It was unfortunate that Marilyn Olmstead could not defend her X-ray structure of [Rh(η¹-N³-1-MT)₂]₂[(Cp*Rh)₂(μ-OH)₃]₃OH, 12 (Fig. 2); however, we have been able to present data that clearly invalidate the premise of Ag^I instead of Rh^I in the anionic com­ponent [Rh^I(η¹-N³-1-MT)₂]⁻.

The salient data, not noted in the Raymond and Girolami article, concerning our X-ray structure of the anionic com­ponent is as follows: no Ag^I com­plex, such as AgOH, would be soluble at pH 10 in water to react with 1-methyl­thymine, as was the case with Rh(OH)₃; since the authors stated the similarity of Ag^I with Rh^I, this single fact also eliminates their Ag^I for Rh^I premise.

Further corroborating evidence was the formation and identification of the reductive elimination product, Cp*OH, when the pH and the temperature were raised to 10 and 60 °C, respectively. To reiterate, this seminal fact solidifies the Rh^I thesis and totally discounts the premise of an Ag^I com­plex for the anionic com­ponent [Rh^I(η¹-N³-1-MT)₂]⁻ (Chen et al., 1995; Smith et al., 2014).

Finally, the most significant experiment was the ¹H NMR study in D₂O, using the [Cp*Rh(H₂O)₃](OTf)₂ that was puri­fied and crystallized, with no Ag^I contamination. No reaction occurred until the pD was raised to 13.6, and the reaction mixture heated for 2 h at 60 °C. These reaction pa­ram­eters caused Cp*OH to precipitate out of solution and it was filtered off. The ¹H NMR analysis clearly showed significant upfield shifts of the C6-H signal 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 C5-Me was now at 1.63 ppm (Δδ = −0.04 ppm), which was from the 1-MT-Cp* π–π inter­actions that were 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 with a Δδ = −0.03 ppm, which further verified that the solid-state and solution structures of com­plex 12 were very similar (Smith et al., 2014).

In this rebuttal to Rh^I versus Ag^I for the anionic com­ponent [Rh^I(η¹-N³-1-MT)₂]⁻, we have presented a convincing argument that reflected extensive experiments conducted to clearly de­fine Rh^I as the metal for the entire com­plex. What Raymond and Girolami were focused on was their feelings about the Ag^I structure, without any other viable chemical facts. Furthermore, to reiterate, they apparently did not thoroughly read the Chen et al. (1995) and Smith et al. (2014) articles, since they totally missed the significance of the reductive elimination product, Cp*OH, clearly to provide an Rh^I anionic com­ponent. Moreover, the ¹H NMR study was very revealing in that the pD (pD = pH + 4) and temperature were critical to the Rh^I designation, while totally dismissing the Ag^I scenario. We are con­fi­dent that the Raymond/Girolami suggestion of Ag^I rather than Rh^I for the anionic com­ponent [Rh^I(η¹-N³-1-MT)₂]⁻ is clearly incorrect, as our data have demonstrated.

4. Answering the Response of Raymond and Girolami to this rebuttal

The Response of Raymond & Girolami (2024) to this rebuttal included comments that were contradictory to what we stated; for example, `The rebuttal claims that compound 1 [12 in Smith et al. (2014)] can be synthesized from the `purified and crystalline aqua complex [Cp*Rh(H₂O)₃](OTf)₂' and that therefore 1 [12 in Smith et al. (2014)] could not possibly have contained silver. If this claim were true, it would be significant, but in fact the Experimental section given in the rebuttal (and in the original articles) describes an in situ preparation of this triflate salt, as described in our next point' (Raymond & Girolami, 2024). The latter statement was patently false, in that the ¹H NMR experiments in D₂O, which incidentally Raymond and Girolami characterized as irrelevant, were conducted with purified crystals of [Cp*Rh(H₂O)₃](OTf)₂ (no Ag^I present) at various pD values (see Experimental) and noting the pD values, where no reaction took place until the pD was raised to 13.6 and the temperature was raised to 60 °C to form Cp*OH, that was further purified via distillation and the use of a C-18 cartridge, followed by identification by GC–MS (Smith et al., 2014); note that pD = pH + 4.

Drs Girolami and Raymond has further questioned the ¹H NMR spectra results concerning the chemical shifts, but failed to realize that the ¹H NMR spectra of 12 in Chen et al. (1995) was conducted in CD₃OD, while the ¹H NMR experiment in Smith et al. (2014) with purified [Cp*Rh(H₂O)₃](OTf)₂ was conducted in D₂O, and thus the disparity in chemical shifts he mentioned were solvent driven. The ¹H NMR of the crystals of 12 and the ¹H NMR solution studies were compared in D₂O, and were similar. Thus, his premise that Ag^I must be present in the X-ray results in Chen et al. (1995), because of the disparity between the ¹H NMR chemical shift results in Chen et al. (1995) and Smith et al. (2014), is invalid.

The only relevant statement made by the authors in their Response was a caveat that said the following in their conclusion statement: `Consequently, we fully acknowledge that the additional evidence we reported above does not prove that 1 [12 in Smith et al. (2014)] contains silver…' To reiterate, if Raymond and Girolami read the Smith et al. (2014) article, as they stated, they would also see structures not seen previously with inorganic–DNA complexes!

Moreover, I now remember Dr Girolami sending me an email that asked the question, Ag^I versus Rh^I, and that led to further experiments published in the Smith et al. (2014) article, as described in this rebuttal. One last thought about this scenario of Ag^I versus Rh^I is that this type of speculation can cause many colleagues who are inter­ested in our bioorganometallic chemistry with many bioligands to possibly question our published X-ray structural studies in this area of research, even in the written presence of Raymond and Girolami's caveat contending that they were only questioning the X-ray structure of the anionic component [Rh^I(η¹-N³-1-MT)₂]⁻. Therefore, we stand firm in that with all their X-ray analysis and caveats, we believe we have shown clearly with the added experiments in Smith et al. (2014) that the Ag^I postulate is not valid and that the anionic component [Rh^I(η¹-N³-1-MT)₂]⁻ contains an Rh^I metal-ion center!

Furthermore, we will attempt, in the future, to resynthesize complex 12 with pure [Cp*Rh(H₂O)₃](OTf)₂ and obtain an Rh analysis to resolve the Ag^I versus Rh^I contention; however, our ¹H NMR experiment in D₂O did show the same chemical shifts, in the absence of any Ag^I contamination, as the ¹H NMR in D₂O for the crystals utilized for an X-ray analysis, which cannot be called irrelevant. To reiterate, the most critical data would be the Rh/Ag elemental analysis to definitively answer the question of Rh^I and not Ag^I for the anionic component of complex 12.

Figure 1 The anionic com­ponent [Rh^I(η^1-N^3-1-MT)2]^− (Chen et al., 1995; Smith et al., 2014). Reprinted with permission from Smith et al. (2014). Copyright (2014) American Chemical Society.

Figure 2 The partial X-ray crystal structure of the anionic and cationic com­ponents of [Rh^I(η^1-N^3-1-MT)2]2[(Cp*Rh)2(μ-OH)3]3OH, demonstrating the im­por­tant π–π inter­actions. Reprinted with permission from Smith et al. (2014). Copyright (2014) American Chemical Society.

Figure 3 Putative mechanism of the reductive elimination of Cp*Rh-OH to provide the anionic com­ponent [Rh^I(η^1-N^3-1-MT)2]^−. Reprinted with permission from Smith et al. (2014). Copyright (2014) American Chemical Society.

5. Experimental

5.1. Synthesis of [Rh^I(η¹-N³-1-MT)₂]₂[(Cp*Rh)₂(μ-OH)₃]₃OH, 12

To a solution of [Cp*RhCl₂]₂ (0.10 g, 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 was then filtered. The resulting filtrate was treated with 1-MT (0.045 g, 0.32 mmol) and, after all the 1-MT was 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 reducing the volume of this reaction mixture to ∼3 ml, com­plex 12 was crystallized at 4 °C as orange plates (20% yield). ¹H NMR (400 MHz, CD₃OD): δ 7.24 (s, 1H, C6-H), 3.29 (s, 3H, N1-Me), 1.85 (s, 3H, C5-Me), 1.62 (s, 22.5H, Cp*). Elemental analysis calculated (%) 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.

5.2. Reductive elimination product Cp*OH

The mechanism for the formation of the anionic com­ponent can be tentatively rationalized by the following observations. 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]⁺, respectively) during the reaction, which provided clear evidence for the loss of the Cp* ligand from Rh^III to Rh^I.

5.3. 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 and provided information that Cp*OH was formed (m/z 151 and 135 for [M − H]⁺ and [M − OH]⁺, respectively) during the reaction, which further provided clear evidence for the loss of the Cp* ligand from Rh^III.

5.4. ¹H NMR experiments on the reaction of [Cp*Rh(H₂O)₃](OTf)₂ and 1-methyl­thymine in D₂O, as a function of pD (pD = pH + 4): 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 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 (C6-H) and 3.15 ppm (N-Me), with the CpRh* signal being at 1.67 ppm. The [(Cp*Rh)₂(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 out, and was filtered off, while there was a pronounced upfield shift of the C6-H signal to 7.10 ppm, or an upfield shift of Δδ = −0.14 ppm, while N-Me also moved upfield to 3.10 ppm (Δδ = −0.05 ppm) and C5-Me was now at 1.63 ppm (Δδ = −0.04 ppm), as expected from the 1-MT-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 ¹H NMR in D₂O for the solid-state crystals and solution structures of com­plex 12 were similar.

5.5. Control experiment: reaction of 1-MT with [Rh(H₂O)₃](OTf)₃

[Rh(H₂O)₃](OTf)₃ was synthesized by the reaction of RhCl₃·3H₂O with 3 equivalents 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, as evidenced via ¹H NMR spectroscopy, but Rh(OH)₃ was formed as a yellow precipitate and was identified via elemental analysis. Analysis calculated (%) for Rh(OH)₃·1.2H₂O: Rh 58.6, H 3.08; found: Rh 58.1.

Footnotes

‡Present address: Bio-Rad Laboratories, Hercules, CA 94547, USA

§Dr Marilyn Olmstead, a world class X-ray crystallography expert, unfortunately passed away on September 30, 2020.