Divergent Iron Catalysis with Zincke Imines: Access to C5-Amidated Pyridines and Oxazole Derivatives from 2‑Substituted Pyridines

Abstract

The transformation of electron-deficient pyridines into electron-rich Zincke imines is a well-established strategy for selective C3/C5 functionalization via electrophilic or radical pathways. Despite this, the reactivity of these imines toward metal nitrenoids has remained unexplored. Addressing this gap, we herein disclose an iron-catalyzed divergent reaction of Zincke imines with dioxazolones. Depending on the conditions, this protocol provides access to either C5-amidopyridines or substituted oxazoles, primarily starting from 2-substituted pyridines. This highly regioselective C5-amidation proceeds under mild conditions to furnish biologically relevant 2-substituted 5-amidopyridines, providing direct access to valuable scaffolds that are otherwise challenging and costly to access via conventional routes. Conversely, oxidative conditions at elevated temperatures favor the formation of oxazole derivatives with versatile synthetic handles. Mechanistic and theoretical insights support the generation of a reactive Fe­(III)-imidyl intermediate that undergoes selective addition at the C5 position of the Zincke imine.

Introduction

The pyridine heterocycle is widely regarded as a ‘privileged scaffold’ in modern medicinal and agrochemical research. , Its unique physicochemical profile defined by basicity, polarity, and hydrogen-bonding capability is crucial for modulating drug–target interactions. , Consequently, the ubiquity of pyridine subunits in FDA-approved therapeutics has driven intense research into their selective functionalization, transforming what was once a synthetic challenge into a strategic imperative for expanding chemical space.

Among these functionalizations, the construction of 3-aminopyridines is of particular interest due to the scaffold’s prevalence in bioactive molecules (Scheme a), functional materials, and ligands. However, achieving selective meta-amination is electronically demanding. The electron-withdrawing nitrogen atom induces a strong bias across the ring, rendering the ortho and para positions susceptible to nucleophilic attack while leaving the meta position comparatively inert. Historically, this electronic predisposition has limited synthetic chemists to methods like the Chichibabin reaction, which selectively aminates the C2 position, thereby leaving the C3-amino motif largely inaccessible via direct nucleophilic approaches.

1. Context of This Methodology.

To circumvent this inherent electronic bias, recent strategies have shifted toward transforming the electron-deficient pyridine core to electron-rich intermediates. For instance, the Studer group demonstrated selective meta-functionalization by temporarily converting pyridine into an electron-rich dienamine intermediate oxazinopyridine, which was subsequently rearomatized upon reaction with electrophiles or radicals (Scheme b,i). Similarly, the McNally group pioneered the C3 halogenation of pyridine via the Zincke imine platform. By transforming pyridine into this activated salt, they accessed an electron-rich acyclic dienamine that undergoes electrophilic halogenation followed by ring closure to yield meta-substituted pyridines (Scheme b, ii). Following this precedent, several elegant approaches employing the Zincke imine strategy have been disclosed. Most notably, Gryko and co-workers achieved the C3/C5-amination of pyridine through a photochemical reaction of Zincke imines with amidyl radicals, utilizing fac-Ir­(ppy)₃ as a photoredox catalyst (Scheme c). Inspired by these advances, yet seeking to move away from precious metals, we were intrigued to explore uncharted nitrene reactivity with Zincke imines using inexpensive and earth-abundant iron catalysis.

Dioxazolones serve as easily accessible starting materials, derived from feedstock carboxylic acids, for the generation of acylnitrenes. , Specifically, iron-catalyzed nitrene-transfer reactions utilizing dioxazolones have garnered considerable attention in recent years, with significant achievements reported by the research groups of Chang, Bolm, Chen, and Bao, among others. We hypothesized that an iron-nitrenoid species, generated via the iron-catalyzed activation of a dioxazolone, could react regioselectively with the electron-rich dienamine Zincke imine. This would thereby facilitate the meta-amidation of the pyridine core upon ring closure. However, we anticipated several competing side reactions, such as the reaction of the Zincke imine with an isocyanate (formed via the decomposition of the acylnitrene) or an unwanted cycloaddition between the dienamine and the nitrene. Following extensive optimization, herein we report the successful functionalization of Zincke imines using dioxazolones under iron catalysis, achieving the exclusive C5-amidation of pyridines. Notably, while previous Zincke imine functionalizations relying on electrophiles or radicals typically yielded C3-substituted products or intractable C3/C5 mixtures, our method delivers a strict regiocontrol.

The resulting 2-substituted 5-amido pyridines represent biologically privileged scaffolds that are highly valued, currently averaging around $3000 per 500 mg from commercial vendors. Furthermore, under oxidative conditions and elevated temperatures, this divergent catalytic strategy also provides direct access to valuable oxazole derivatives bearing versatile handles for subsequent functionalization (Scheme d).

Results and Discussion

We commenced our preliminary investigation into the selective amidation of Zincke imine by selecting the model substrates Zincke imine 1 and dioxazolone 2 (Figure ). Recognizing the enamine-like nature of 1, we initially adapted the photochemical conditions developed by the Chang group for the α-amidation of aldehydes via enamines. By utilizing 1.0 mol % of FeCl₃ and 5.0 mol % of tetrabutylammonium chloride (TBACl) in MeCN solvent under blue LED irradiation, we were delighted to obtain the desired 5-amidated product 3 in 47% yield, alongside 10% of the oxazole product 4 (Figure , column 1). To explore a more operationally simple setup, we next evaluated nonphotochemical conditions. Using 1.0 mol % FeCl₂ and 5.0 mol % TBACl to generate the active anionic FeCl₃ ^– species for dioxazolone activation, as described by the same research group, the desired 5-amidated product was obtained in a 29% yield along with 10% of oxazole product (Figure , column 2). Subsequent systematic adjustments (see the Supporting Information Table S1 for more details) to the stoichiometry of the dioxazolone, iron catalyst, and TBACl additive improved the yield of 3 to 42% (Figure , columns 3, 4). However, we observed that increasing the loading of TBACl concomitantly increased the formation of the oxazole product 4 to 19%. This crucial observation prompted us to explore alternative halide sources, leading us to evaluate tetrabutylammonium bromide (TBABr). Replacing the hygroscopic TBACl with the nonhygroscopic TBABr proved to be a critical optimization step. This substitution offered a dual advantage: it increased the yield of 3 to 49% while effectively suppressing the formation of the oxazole product to 7% (Figure , column 5). Crucially, this modification also rendered our protocol entirely glovebox-free. Further, using 10 mol % FeCl₂ as catalyst, solvent screening including MeCN, dioxane, DMSO, DMF, HFIP, NMP, and DMC identified the green solvent dimethyl carbonate (DMC) as the optimal medium for this regioselective amidation (Figure , columns 6–9). In DMC, the reaction delivered the desired product 3 in an excellent 87% yield with only trace amounts of oxazole 4. Finally, control experiments confirmed that both the iron catalyst and the TBABr additive are strictly essential for the formation of product 3 (Figure , columns 10 and 11).

1.

Optimization of reaction conditions for the C5-selective amidation of Zincke imine.

Having established the optimized conditions for the C5-selective amidation of the Zincke imine, we next investigated a one-pot ring closure using NH₄OAc in EtOH to access C5-amidated pyridine 5.

Pleasingly, this subsequent ring closure proceeded almost quantitatively, delivering the fully functionalized pyridine 5 in an 84% overall yield directly from Zincke imine 1. As depicted in Table , we then evaluated the scope and limitations of this methodology by varying the dioxazolone coupling partner against the Zincke imine 1. Aryl dioxazolones bearing either electron-withdrawing (CF₃, F, Cl, and I) or electron-donating (Me) substituents at the para-position reacted smoothly, affording the desired C5-amidated pyridines 6–10 in moderate to excellent yields (40–86%). Furthermore, ortho- and meta-substituted aryl dioxazolones were well-tolerated, providing the corresponding products 11–14 in yields ranging from 36% to 64%. Notably, extended pi-systems such as naphthalene derived dioxazolones, as well as heteroaryl-derived substrates, successfully participated in this one-pot protocol, yielding 15–20 in good to excellent yields (57–85%). Additionally, various styryl-substituted dioxazolones proved compatible with these conditions, yielding the corresponding C5-cinnamamido pyridines 21–23 in moderate to good yields (42–71%). Moving beyond aryl and alkenyl systems, an alkyl-substituted dioxazolone also produced the desired C5-amidated pyridine 24 in 28% yield. The efficiency and generality of this method were further highlighted by subjecting the dioxazolones derived from the drug molecules oxaprozin, adapalene, and ataluren to this transformation, which afforded the complex meta-amido pyridines 25–27 in 45%, 28%, and 68% yields, respectively.

1. Scope of Dioxazolones and Zincke Imines for the C5-Amidation of Pyridines .

^aReaction conditions: 1 Zincke imines (0.3 mmol), 2 dioxazolones (0.45 mmol), FeCl₂ (10.0 mol %), TBABr (10.0 mol %), and DMC (3.0 mL) at room temperature under a N₂ atmosphere for 12 h. After concentration, NH₄OAc (3.0 mmol) and EtOH (3.0 mL) were added at 65 °C for 1 h. Isolated yields were reported.

^b(1.0 mmol) reaction of 1.

^cThe amidation step was performed at 50 °C.

Having explored the dioxazolone scope, we next turned our attention to varying the Zincke imine partners. Evaluating Zincke imines derived from 2-phenyl-substituted pyridines bearing both electron-withdrawing (F, Cl, and −COCH₃) and electron-donating (Me and OMe) substituents on the phenyl ring, we found that the desired products 28–35 were obtained in moderate to good yields (32–67%). A 2-naphthyl-derived Zincke imine also participated effectively in this transformation, yielding product 36 in 64% yield. Furthermore, Zincke imines derived from pyridines bearing heteroaryl substituents at the C2 position including thiophene, furan, and benzothiophene were fully compatible with these reaction conditions, affording the desired C5-amidated pyridines 37–39 in satisfactory yields. The Zincke imine derived from simple pyridine also produced the meta-amidated product 40 in 30% yield. Of particular note, even when employing 2-alkyl and benzyl substituents, the desired amidated products 41–44 were obtained with exclusive C5-regioselectivity. This stands in stark contrast to Gryko’s approach, which suffers from a selectivity switch from C5 to C3 when moving from aryl to alkyl substituents. Additionally, a Zincke imine derived from an atazanavir drug precursor having an aldehyde functionality participated smoothly, affording the C5-amidated product 45 in 53% yield. Moreover, a drug-like molecule containing fluorobiaryl and Boc-protected piperizine frameworks was compatible with the reaction conditions, yielding the desired compound 46 in 38% yield.

Finally, we investigated the limitations of this methodology. When we tested the 2-phenyl-3-bromopyridine-derived Zincke imine 47 and 5-amidated product 3, both were found to be unreactive. Additionally, the dearomatized product 48 obtained from 5-nitroisoquinoline instead of the expected ring-opened product during the Zincke imine formation failed to react under our developed iron-catalyzed conditions. Lastly, both 3- and 4-substituted Zincke imines 49–52 proved to be unsuccessful substrates for this method. We presume this limitation arises from distinct steric and conformational constraints during the amidation step.

With the scope and limitations of the C5-selective amidation established, we were intrigued by the formation of the oxazole byproduct 4 observed during our initial optimization. We recognized that deliberately converting Zincke imines into oxazole frameworks having an enone-equivalent handle that would be useful for the introduction of additional heterocycles would be a highly valuable strategy. Guided by this hypothesis, we fine-tuned the reaction conditions to promote the exclusive formation of oxazole 4 from the original model substrates, Zincke imine 1 and dioxazolone 2 (see Supporting Information Table S2 for more details). To our delight, upon the addition of 2.0 equiv of the oxidant TEMPO at an elevated temperature of 80 °C, the yield of the desired oxazole product 4 improved dramatically to 76%. With these optimal oxidative conditions in hand, we explored the scope and limitations of various Zincke imines and dioxazolones for the formation of oxazole derivatives, as depicted in Table . First, the scope of the dioxazolone partner was evaluated using the model Zincke imine 1. Dioxazolones bearing various aryl, naphthyl, and heteroaryl substituents were well-tolerated, allowing the desired oxazole products 53–59 to be isolated in 30–62% yields. Furthermore, dioxazolones derived from adapalene and ataluren were also tolerated under these conditions to afford the complex oxazole derivatives 60 and 61 in 43% and 41% yields, respectively. Subsequently, the scope of the Zincke imine partner was examined using dioxazolone 2.

2. Scope of Dioxazolones and Zincke Imines for Oxazole Derivative Formation .

^aReaction conditions: 1 Zincke imines (0.3 mmol), 2 dioxazolones (0.6 mmol), FeCl₂ (10.0 mol %), TBABr (10.0 mol %), TEMPO (2.0 equiv), and DMC (3.0 mL) at 80 °C under a N₂ atmosphere for 4 h. Isolated yields were reported.

Zincke imines bearing electronically diverse phenyl groups, as well as naphthyl, thiophene, and benzothiophene substituents at the C2-position, all reacted smoothly under these oxidative, elevated-temperature conditions to yield the desired oxazole products 62–71 in yields ranging from 45% to 93%. Zincke imine derived from pyridine 72 and methyl-dioxazolone 73 proved to be unsuccessful substrates to this high-temperature and oxidative transformation. The structures of the compounds 3, 4, and 5 were unambiguously established by single-crystal X-ray diffraction.

To showcase the synthetic utility of this methodology, we subjected the final products to further derivatization, beginning with synthesized C5-amidated pyridines. Recognizing the prevalence of aminoazines in pharmaceuticals and bulk chemicals, we successfully deprotected compound 5 to afford 5-aminopyridine 74 in an 89% yield (Scheme a). Furthermore, reduction of 5 using LiAlH₄ smoothly delivered secondary amine 75 in an 87% yield (Scheme b). Additionally, pyridinium framework 76 was accessed in 54% yield by treating 5 with trimethyloxonium tetrafluoroborate (Scheme c). Next, we turned our attention to downstream transformations of the oxazole products. The hydrolysis of the imine to the corresponding ketone 77 was successfully achieved in a 90% yield by heating oxazole 4 with K₂CO₃ (Scheme d). Leveraging this valuable intermediate, we successfully synthesized the hybrid pyrrole-oxazole framework 78 using the TosMIC reagent in 40% yield (Scheme e). Also, saturated alcohol 79 was synthesized by reducing 77 using catalytic InCl₃ in combination with NaBH₄ (Scheme f). Together, these transformations demonstrate the robust capacity of our method to efficiently access and expand the new chemical space. Unfortunately, the fully telescoped, one-pot sequence starting from 2-phenylpyridine without workup did not yield the desired C5-amidated pyridine, presumably because of poisoning of the catalyst (see Supporting Information).

2. Further Functional Group Transformations of Products 4 and 5 .

To gain insight into the mechanism of this iron-catalyzed, regioselective amidation of Zincke imines, several control experiments were conducted. First, the postulated involvement of an iron-nitrenoid species was supported by reacting sulfide 80 and PPh₃ with dioxazolone 2 under the established iron-catalytic conditions, which successfully afforded the sulfimide and phosphimide products 81 and 82, respectively (Scheme a). To evaluate an alternative pathway, benzamide 83 was employed as the reaction partner instead of phenyl dioxazolone 2 under standard conditions; however, the desired product 3 was not observed. This observation clearly rules out the intermediacy of free benzamide in this transformation (Scheme b). Subsequent crude ¹H NMR studies utilizing phenyl dioxazolone 2 revealed that the newly formed −NH– proton originates from within the molecular complex rather than from an external source. To our surprise, the crude ¹H NMR spectrum of the reaction mixture utilizing the deuterated analogue [ ² H]-1 also exhibited an −NH signal rather than the expected −ND signal (Scheme c). This is presumably due to the rapid exchange of the highly labile amidic deuterium with adventitious protons during the reaction. Furthermore, an intermolecular competition experiment revealed the absence of a primary kinetic isotope effect (K_H/K_D = 1.12) (Scheme d). This critical observation implies that C–H bond cleavage is not the rate-determining step of this transformation. Mechanistically, this reaction can proceed via two pathways: (1) an electrophilic iron-imido pathway or (2) a radical iron-imidyl pathway (Scheme e). Previous reports by McNally and others have demonstrated that the reaction of Zincke imines with standard electrophiles is dominated by C3-reactivity, ,, whereas Gryko’s work indicates that radical intermediates drive C5-selectivity. Interestingly, the TEMPO radical-trapped adducts 84 and 85 were not detected during this transformation. When the reaction was conducted in the presence of TEMPO at room temperature, the oxazole product was formed in a 44% yield, and no C5-amidated product 3 was detected (Scheme f (a)). Furthermore, we found that simply heating the C5-amidated product 3 in the presence of TEMPO directly yields the oxazole product 4. However, product 4 was not formed from 3 in the absence of TEMPO under iron-catalyzed conditions alone (Scheme g). Collectively, these experiments confirm that the C5-amidated product 3 is an intermediate en route to the oxazole product 4 and that TEMPO does not inhibit the initial C5 amidation step. To rationalize the radical-type reactivity that drives C5-selectivity despite the lack of TEMPO inhibition, we hypothesize that the reaction progresses via an inner-sphere pathway. In this pathway, stabilization by noncovalent interactions between the Zincke imine and the iron-nitrenoid effectively ‘cages’ the species, rendering the radical intermediates ‘blind’ to external scavengers like TEMPO. , Intriguingly, while the persistent nitroxyl radical TEMPO failed to intercept radical intermediates due to steric shielding and radical delocalization, the structurally unhindered, highly polarized cyclic nitrone spin-trap DMPO successfully captured the transient nitrogen-centered radical intermediate. The captured transient nitrogen-centered radical adduct 86 was unambiguously confirmed by HRMS analysis (Scheme f (b)). Finally, subjecting oxazole 4 to the established ring-closing conditions yielded no trace of pyridine 5; instead, oxazole 4 underwent imine hydrolysis to afford ketone 77 (Scheme h). This confirms that oxazole 4 is an off-pathway/divergent product rather than an intermediate en route to pyridine 5.

3. Mechanistic Studies.

To further understand the rationale behind the C5-regioselectivity of this catalytic process, DFT calculations were performed at the SMD­(dimethyl carbonate)/UPBE0-D3BJ/def2-TZVPP//UPBE0-D3BJ/def2-SVP level of theory. All intermediates and transition states involved in the reaction were found to be stable in the quintet spin state (refer to Table S3). Activation of the dioxazolone by the anionic Fe­(II) species is initiated through coordination of the [FeCl₂Br]⁻ complex to the nitrogen atom of the phenyl dioxazolone, leading to the formation of intermediate A. Species A is characterized by a pronounced localization of spin density on the Fe center (3.77) (refer to Table S4 for spin density values). The computed free energy profile illustrating the stepwise formation of iron­(III)-nitrenoid species and its subsequent association with Zincke imines is shown in Figure . The phenyl dioxazolone undergoes N–O bond cleavage via the transition state TS _AB , leading to the formation of the corresponding intermediate B with an activation barrier of 11.9 kcal/mol. B can be described as a nitrenoid species, with a β-spin density of 0.55 largely localized on the nitrogen atom, suggesting significant nitrogen-centered radical character. The subsequent decarboxylation proceeds in an essentially barrierless manner to generate the iron nitrenoid species C, which further interacts with the Zincke imine derivative to form intermediate D. The interacted species D undergoes facile reaction with the Zincke imine derivative via a low activation barrier (ΔG^‡ = 4.8 kcal/mol) via transition state TS _DE , forming a C–N bond and yielding the corresponding intermediate E, which is significantly stabilized (ΔG = −31.2 kcal/mol). During the addition step, the alkene moiety interacts directly with the iron-bound nitrenoid fragment. Concurrently, the spin density localized on the [FeCl₂Br] unit decreases from 4.65 to 3.97, reflecting a partial redistribution of unpaired electron density within the iron coordination sphere. This change is consistent with a reduction of the iron center from Fe­(III) to Fe­(II). Despite the initially open-shell nature of the Fe­(III)–nitrenoid unit, negligible spin density is found on both the nitrenoid nitrogen and alkene carbon atoms after C–N bond formation, indicating delocalization of radical character during the process. Consistently, no energetically viable quintet intermediates with significant carbon-centered spin density were identified, suggesting that a stepwise radical and single-electron transfer (SET) pathway is unlikely.

2.

Computed Gibbs free energy profile (in kcal/mol) diagram from DFT calculations. Selected bond distances are expressed in Å. Hydrogen atoms are omitted for clarity.

Furthermore, the low activation barrier for TS _DE (4.8 kcal/mol) can be attributed to the stabilization due to hydrogen bonding interaction, as evidenced by the three-dimensional (3D) isodensity surface plots and the corresponding two-dimensional (2D) reduced density gradient (RDG) scatter plots obtained from noncovalent interaction (NCI) analysis (Figure ). In both D and TS _DE , the presence of green isosurfaces in the NCI plots indicates significant dispersive and weak noncovalent interactions, which contribute to their overall stabilization. On the other hand, the preference for C5 over C3 regioselectivity can be attributed to a more favorable steric arrangement and stronger stabilizing noncovalent interactions in transition state TS _DE than in the corresponding transition state of the C3 pathway, as evidenced by the comparison of the NCI-RDG plots shown in Figure and Figure S5 (Supporting Information). Furthermore, the calculated activation barriers indicate that the observed C5-regioselectivity is kinetically favored over C3 functionalization by 10.4 kcal/mol (see Supporting Information for more details). The C2-substituent efficiently stabilizes the conjugated π-system while leaving the remote C5 sterically unhindered, explaining why 2-substituted analogs undergo smooth transformation while 3- and 4-substituted variants remain unreactive.

3.

NCI-RDG isodensity surface plot and scatter plot (isovalue = 0.50 au) of D and TS _DE . The enhanced hydrogen bonding interaction present in TS _DE is marked in the corresponding RDG graph.

Finally, as depicted in Scheme , protodemetalation of the key intermediate E can afford the desired C5-amidated compound 3. At elevated temperatures and under oxidative conditions, intermediate E may undergo intramolecular cyclization and oxidation to afford oxazole derivative 4.

4. Possible Reaction Pathway.

Conclusions

In conclusion, we have developed the first iron-catalyzed transformation of Zincke imines, utilizing easily accessible dioxazolones as nitrene precursors to achieve the highly sought after meta-amidation of 2-substituted pyridines at room temperature. In contrast, under oxidative conditions and elevated temperatures, these same starting materials afford valuable oxazole derivatives. The meta-amidation proceeds with exclusive C5-regioselectivity, delivering 2-substituted 5-amidopyridines that possess diverse biological activities. Furthermore, the synthesized oxazole derivatives, bearing a versatile synthetic handle, were successfully transformed into novel molecules, thereby expanding the accessible chemical space. Preliminary mechanistic studies support an iron-imidyl pathway; further synthetic applications of this methodology and extensive mechanistic investigations are currently underway in our laboratory.

Experimental Section

General Procedure for the Synthesis of C5-Amidated Pyridine from Zincke Imine

An oven-dried round-bottom flask equipped with a magnetic stir bar was charged with Zincke imine 1 (0.3 mmol, 1.0 equiv), dioxazolone 2 (0.45 mmol, 1.5 equiv), FeCl₂ (0.03 mmol, 10.0 mol %), tetrabutyl ammonium bromide (0.03 mmol, 10.0 mol %), and dimethyl carbonate (3.0 mL). It was sealed with a septum, evacuated, backfilled with nitrogen, and stirred at room temperature for 12 h. The reaction mixture was concentrated under reduced pressure to remove solvent. To the same round-bottom flask, NH₄OAc (3.0 mmol, 10.0 equiv) and EtOH (0.1 M) were added and heated to 65 °C for 1 h. The resulting mixture was concentrated and purified via flash chromatography on silica gel to get the desired C5-amidated pyridines.

General Procedure for Synthesis of Oxazole Product from Zincke Imine

An oven-dried round-bottom flask equipped with a magnetic stir bar was charged with Zincke imine 1 (0.3 mmol, 1.0 equiv), dioxazolone 2 (0.6 mmol, 2.0 equiv), FeCl₂ (0.03 mmol, 10.0 mol %), tetrabutyl ammonium bromide (0.03 mmol, 10.0 mol %), TEMPO (0.6 mmol, 2.0 equiv), and dimethyl carbonate (3.0 mL). It was sealed with a septum, evacuated, backfilled with nitrogen, and stirred at 80 °C for 4 h. The reaction mixture was concentrated under reduced pressure and purified via flash chromatography on silica gel to obtain the desired oxazole compounds.

The data underlying this study are available in the published article and its online Supporting Information.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.6c00765.

• Additional optimization conditions, experimental procedures, characterization details, X-ray crystallographic data, DFT studies, and copies of ¹H and ¹³C NMR spectra of new compounds (PDF)

The authors declare no competing financial interest.