Fluorescence Changes of RCO and DEB-CO Upon Addition of a CO Surrogate, CORM-3, and CO Gas

In examining the effect of CO on the fluorescent intensity of RCO and DEB-CO, we conducted three lines of work: (1) reproduction of the original experimental work by using CORM-3 as the CO source, (2) quantitative analyses of what the fluorescent results might mean, and (3) assessment of the probe fluorescent response to gaseous CO.

Using CORM-3 as the Source of CO

Over the last 20 years, several metal/borane-carbonyl complexes have been widely used as CO surrogates for studying CO biology. They are named as CO-releasing molecules (abbreviated as CORMs) with CORM-2, CORM-3, CORM-401, and CORM-A1 being commercially available and thus most widely used. In the original study of using RCO to detect CO, CORM-3 (a ruthenium–carbonyl complex) was used as the source of CO for assessment. Though it is named as a CO-releasing molecule, it mostly releases CO₂ instead of CO in aqueous buffer unless there is an added nucleophile, such as sodium dithionite.^29,33−35 Furthermore, CORM-3 has extensive chemical reactivity as a reducing agent, catalase mimic, and an electrophile.^27,28,31,36 In the original report, RCO (10 μM) was incubated with 1 equiv of CORM-3 in 10% dimethyl sulfoxide (DMSO) and 90% phosphate base saline (PBS) for 30 min, leading to an obvious fluorescent intensity increase. We conducted the same experiments and observed the same results, indicating reproducibility of the original experimental observation. Specifically, Figure ​Figure11A shows time-dependent fluorescent intensity increases at 580 nm (λ_ex = 530 nm). Then, there is the question of how to explain the fluorescent results in the context of the NMR studies described in the preceding paragraph. In the next section, we conduct a quantitative assessment of the fluorescence intensity changes observed to shed light on this issue.

Open in a separate window

Figure 1

Effects of CORM-3 on the fluorescence of RCO solution (10 μM, in 10% DMSO, 90% PBS solution). (A) Fluorescence spectra after CORM-3 (10 μM) addition to the RCO solution (RCO only was used as a control). (B) Comparison of the fluorescence intensity of the CORM-3 (10 μM)-RCO solution (10 μM), rhodamine B (50 nM), and RCO probe only (10 M) (n = 3, mean ± SD, λ_ex = 530 nm, λ_em = 580 nm bandwidth = 5 nm).

What the Fluorescence Increases Mean in a Quantitative Sense

There are good reasons that fluorescence has been widely used for the sensitive detection of analytes because it is highly sensitive. In contrast, NMR does not allow for reliable detection below a fraction of 1%. Could the inconsistency between the NMR results and fluorescence results be due to this sensitivity difference? For this, we measured the fluorescence of rhodamine B, the proposed product after RCO’s reaction with CO (Scheme 1B). As shown in Figure ​Figure11B, the fluorescence intensity of rhodamine B (positive control) at 50 nM is far higher than that of the reaction mixture between RCO and CORM-3 at 10 μM each. Such results mean that if rhodamine B was indeed formed in the reaction as proposed, the yield would be less than 0.5% based on fluorescence intensity (Figure S4 shows standard curves for rhodamine B in different solvents). Furthermore, there was no characterization of the fluorescent species in the original publication. Therefore, the nature of the reaction is unknown.

Considering all of the results thus far, the highly improbable nature of uncatalyzed CO insertion, the known complex nature of CORM-3 chemistry, the lack of structural characterization of the product between RCO and CORM-3, and the meniscal amount of fluorescent product, we can say with a high degree of certainty that there is no evidence to support CO insertion into hydrazone double bond as a way to sense CO by RCO. Further, it has been shown before that sensing a CORM does not equate to sensing CO,³⁷ and CORM-3 does not release a meaningful amount of CO unless there is a strong nucleophile.^34,38 We did not expend extra efforts to examine the mechanistic details because of the known complexity of the Ru-carbonyl complexes in terms of its chemical reactions including redox properties, catalase-like activities, and reactivity with nucleophiles. However, it should be noted that there are reports of enhanced fluorescence intensity of fluorophores through complexation with Ru(II).^39,40 Further, a Ru(II) complexation with hydrazone has been reported.⁴¹ The fluorescence changes described in the original papers are likely due to Ru-mediated events, which further support the theme of this study: uncatalyzed CO insertion does not happen. The scope of this paper remains focused on examining whether CO insertion into hydrazone double bond was the reason for the reported fluorescent turn-on and do not wish to make this a ruthenium chemistry project, which would go beyond the scope of this study.

With the demonstration of the minuscule level of fluorescent intensity changes from RCO and CORM-3, we sought to examine a second case (DEB-CO) briefly just to see whether we would have similar experimental observations. When CORM-3 (3 equiv) was added to 10 μM DEB-CO solution in HEPES buffer and incubated at 37 °C, the solution showed a weak fluorescence intensity increase after 30 min of incubation. The fluorescence intensity was lower than that of 1 nM rhodamine B (the original paper used rhodamine B as their quantum yield standard, although the proposed end product is only a derivative of rhodamine B). When we compared rhodamine B to the reaction mixture of DEB-CO (10 μM) and CORM-3 (30 μM), the fluorescent intensity indicated the generation of a negligible percentage (<0.01%) of fluorophoric compound from the parent DEB-CO (Figure ​Figure22). The results from DEB-CO were less than that of the original publication and the identity of the weakly fluorescent species from DEB-CO after incubation with CORM-3 is not clear.

Open in a separate window

Figure 2

Comparison of the fluorescence intensity of the DEB-CO solutions (10 μM, in 5% DMSO, 95% HEPES solution) incubated with CORM-3 (30 μM) for 30 min at 37 °C; DEB-CO were probed with CO gas by bubbling CO gas for 5 min, then incubating at 37 °C for 30 min (DEB-CO only and 1 nM rhodamine B were used as a control, n = 3, mean ± SD, λ_ex = 580 nm, λ_em = 630 nm, bandwidth = 10 nm).

With all of the findings that are inconsistent with uncatalyzed CO insertion into a hydrazone double bond to a meaningful degree, we sought to conduct additional confirmation experiments using pure CO gas.

Experiments with CO Gas

To lay any claim to a new CO-insertion reaction, it must be able to undergo the same reaction with CO gas. Thus, pure CO gas was used to assess the true CO-sensing ability of RCO and DEB-CO. Briefly, for RCO, CO gas was bubbled into a RCO solution (10% DMSO, 90% PBS solution) for 5 min. As shown in Figure ​Figure33A, there was no meaningful difference in fluorescence intensity between the CO group and the control group (RCO 10 μM), after 30 min of incubation. In comparison, 1 equiv CORM-3 (10 μM) was able to cause a small increase in the fluorescence of RCO within 30 min. These results show that RCO only responded to CORM-3, but not CO. It should be noted that in the original publication, experiments were also conducted using CO gas, which was said to increase the fluorescence of RCO by an estimated 6-fold after 12 h of incubation (Figure S52). We conducted similar experiments using 1 mM CO in solution. Briefly, RCO solution (10 μM) was sealed in a vial. Then pure CO gas was bubbled into the RCO solution for 20 min followed by incubation at 37 °C for 18 h. No fluorescence intensity change was observed (Figure ​Figure33B). We have no explanation of the different experimental observations in using CO gas.

Open in a separate window

Figure 3

Effects of CORM-3 and CO gas on the fluorescence of RCO solution (10 μM, in 10% DMSO, 90% PBS solution). (A) RCO solutions were probed with CO gas by bubbling CO gas for 5 min, then incubating at 37 °C for 30 min. (B) RCO solutions were probed with CO gas by bubbling CO gas for 20 min, then incubating at 37 °C for 18 h (probe only was used as control) (n = 3, mean ± SD, λ_ex = 530 nm, λ_em = 580 nm bandwidth = 5 nm).

For DEB-CO, we conducted similar studies by using pure CO gas. Briefly, pure CO gas was bubbled into a DEB-CO solution (5% DMSO, 95% HEPES solution) for 5 min followed by incubation at 37 °C for 30 min. We saw no indication that DEB-CO possessed meaningful ability to sense CO gas (Figure ​Figure22, column 3). Such results indicate that DEB-CO did not respond to or react with CO and is not a probe for CO. Comparative studies were carried out using CORM-3, leading to marginal fluorescence intensity increases. Importantly, such results are different from what was reported in the original publication.²⁴ Again, we have no explanation of the different experimental observations.

Overall, our results do not support an uncatalyzed CO-insertion reaction to lead to fluorescence response of RCO and DEB-CO to CO in a meaningful fashion. To our knowledge, there is no known chemical pathway for the originally proposed CO insertion into a hydrazone bond without metal catalysis. The establishment of such a novel uncatalyzed CO-insertion reaction will require much more rigorous evidence than presented in the original publication.

We thank the National Institutes of Health for supporting our CO work (R01DK119202 in support of CO and colitis and R01DK128823 in support of CO and kidney injury). We would also like to thank Georgia Research Alliance for an Eminent Scholar Endowment (BW), the Dr. Frank Hannah Chair endowment (BW), and GSU internal sources for financial support including a GSU B&B fellowship to NB. Mass spectrometric work was conducted by the GSU Mass Spectrometry Facilities, which are partially supported by an NIH grant (S10OD026764). Graphic abstract were created with biorender.com.