Deciphering Cu(II) EPR in Biological Buffers: Implications for Copper Protein Studies

Abstract

Copper proteins are essential for many biological processes, mediating electron transfer and redox catalysis in biochemistry and playing key roles in human health. Electron paramagnetic resonance (EPR) spectroscopy is a powerful tool to probe Cu(II) centers in these proteins, but its application can be complicated by spurious signals arising from adventitious Cu(II) binding to buffers or affinity tags. Here, we systematically examined Cu(II) EPR signals from common biological buffers (Tris, HEPES, MOPS) over pH 6.0–8.0, as well as from a His affinity tag. We show that Tris coordinates Cu(II) across this pH range, whereas HEPES binds Cu(II) weakly at near neutral pH and MOPS essentially does not chelate Cu(II) and rather precipitates Cu(II). These buffer-dependent Cu(II) complexes produce distinct EPR signatures that can interfere with spectral interpretation and spin quantifications. Using two case studies—a bacterial multicopper oxidase (MnxEFG) and Drosophila lysyl oxidase (DmLOX), we show how buffer choices and purification tags influence the observed Cu EPR spectra. In MnxEFG, switching from HEPES to Tris eliminated adventitiously bound Cu(II) and revealed the true copper content and sites. In DmLOX, we identified an EPR signature of Cu(II) bound to the protein’s His-tag, which could be minimized by appropriate refolding conditions. These findings highlight the importance of controlling solution chemistry during sample preparation and suggest practical strategies to obtain cleaner, more interpretable spectra. Overall, this work provides guidance for minimizing nonspecific Cu(II) binding and improving the reliability of EPR analysis in studies of copper proteins in biological systems.

Graphic Abstract

Adventitiously bound Cu2+ signals in biological buffers are studied by EPR spectroscopy. The moderate copper-binding capacity of Tris is useful to clean up adventitiously bound Cu in protein samples.

1. Introduction

Copper is a pivotal transition metal in biology, serving as a redox-active cofactor in numerous enzymes and proteins.[1] Copper proteins participate in fundamental processes such as electron transfer and oxidative catalysis in respiration and metabolism.[2–5] For example, cytochrome c oxidase, a multi-copper enzyme in mitochondria, catalyzes the four-electron reduction of O₂ to H₂O in the final step of the respiratory electron transport chain.[6] Other copper enzymes include superoxide dismutase (antioxidant defense),[7] tyrosinase (pigmentation chemistry),[8] and laccases or multi-copper oxidases involved in oxidative metabolism.[4] Given copper’s role in biology, disruptions in copper homeostasis leads to severe medical conditions.[9] Menkes syndrome, caused by defective copper transport, results in systemic copper deficiency and is typical lethal early in life.[10] Conversely, Wilson’s disease leads to toxic copper accumulation—particularly in the liver and brain—resulting in severe neurodegenerative symptoms.[11, 12] Beyond these classical disorders, copper imbalance is implicated in neurodegeneration such as Alzheimer’s disease, where Cu–amyloid-β interactions may influence protein aggregation and oxidative stress.[13–17] These examples underscore the broad biological and biomedical significance of copper proteins and the necessity of understanding their structure and function.

EPR spectroscopy is especially well-suited to studying Cu(II) centers because Cu(II) (3d⁹, S = 1/2) provides a characteristic and information-rich signal.[18] In an approximately octahedral ligand field, the Cu(II) ground state is an E_g term that undergoes Jahn-Teller distortion to a tetragonally elongated geometry with a singly-occupied d_x2–y2 state. The resulting axial EPR spectrum shows distinct parallel (‖) and perpendicular (⊥) features, with anisotropic g-tensor (g_‖ > g_⊥> g_e, where g_e ~ 2.0023 for the free electron) arising from spin-orbit coupling. At X-band (~9 GHz), a hallmark of Cu(II) EPR signal is the quartet hyperfine lines in the parallel region due to coupling to the ^63,65Cu nucleus (I = 3/2 for both isotopes). The parallel hyperfine splitting A_‖ is a sensitive gauge of unpaired spin density on the copper ion: smaller A_‖ generally indicates greater delocalization of the electron onto ligands (i.e. more covalent Cu-L), whereas a larger A_‖ suggests a more localized Cu(II) with primarily N/O ligand environment. Likewise, the g-values are also sensitive to the ligand field strength and geometry. These empirical correlations—commonly summarized in Peisach–Blumberg plots—enable inference of metal–ligand identity and geometry directly from EPR parameters, providing a practical “fingerprint” for copper sites.[19]

Copper centers in proteins are classically grouped (by structure and spectroscopy) as Type I, Type II, or Type III, alongside a unique Type 0,[20] the specialized Cu_B, binuclear Cu_A, and tetranuclear Cu_Z sites.[18] Type I (“blue”) copper centers are mononuclear electron-transfer sites characterized by a strong Cu-S(Cys) covalent bond within a distorted trigonal/tetrahedral environment.[21] They display an intense ~600 nm S→Cu charge-transfer band. In EPR spectroscopy, Type I sites are notable for their small A_‖ due to the covalent Cu-thiolate bond delocalizing the unpaired electron. Typically, Type I copper EPR signals have g_‖ ~2.20–2.25 with A_‖ on the order of ~170–250 MHz—much smaller than simple Cu(II) compounds. Type II copper centers, by contrast, are “normal” mononuclear Cu(II) sites frequently found in enzyme active sites involving oxygen activation. A Type II copper is often coordinated in a tetragonal geometry by N/O ligands. They lack the intense 600 nm electronic band and typically exhibit a classic axial EPR spectrum with larger A_‖, commonly g_‖ ~2.25–2.30 and A_‖ ~350–550 MHz, consistent with a more localized Cu(II) environment. Type III copper centers consist of antiferromagnetically coupled binuclear Cu(II) pairs (as in the O₂-binding sites of tyrosinase and the trinuclear clusters of multicopper oxidases) and are EPR-silent. The Cu_A center (found in cytochrome c oxidase and nitrous oxide reductase) is a mixed-valence binuclear cluster that together behave as a single S = 1/2 center delocalized over two coppers, giving an EPR signal distinct from Type I or II (with very small A_‖ ~120 MHz and unique g_‖ ~2.18).[22] Finally, the Cu_Z site found in nitrous oxide reductase is a unique tetranuclear copper center [Cu⁺₃Cu²⁺S] which is a delocalized mixed-valent cluster.[23] This rich variety of copper sites illustrates the diverse coordination chemistry of copper in proteins and the correspondingly diverse EPR signatures that can be observed.

Despite its advantages, EPR analysis of copper proteins can be complicated by background Cu(II) signals arising from adventitious binding of Cu(II) to other components in a sample. As a borderline hard/soft Lewis acid, Cu(II) binds readily with a wide range of ligands (O, N, or S donors) in buffers, chelators, protein surfaces, and affinity tags. If a copper protein sample contains excess Cu(II) or is prepared under suboptimal conditions, adventitiously bound Cu(II) species can form and produce their own EPR signals, which overlay those of the intended protein site(s). This is especially problematic when the protein’s Cu concentration is low or when multiple Cu species need to be assigned. For instance, common biological buffers like Tris and HEPES, while often considered innocuous, can coordinate Cu(II) to form complexes. Previous studies have shown that Tris buffer forms several Cu(II) complexes depending on pH and buffer:metal ratio.[24] HEPES, another widely used buffer, was also found to weakly bind Cu(II) at neutral to slightly acidic pH.[25] Stop-flow kinetics experiments has also been performed to monitor the formation of complexes between copper ion and Tris, HEPES and MOPS buffers.[26] In practice, buffer concentrations in protein preparations are usually much higher than the copper or protein concentration, so even weak binding may competitively sequester a significant fraction of Cu(II). This can lead to background signals or even loss of Cu from the protein via precipitation of Cu–buffer complexes. Strong chelators such as EDTA is often used to remove metal ions. While EDTA effectively binds free Cu(II), its presence can also strip copper from metalloprotein sites if not used cautiously, potentially altering the metal content of the sample during preparation. In addition, polyhistidine affinity tags (His-tags) used for recombinant protein expression can be a source of nonspecific Cu(II) binding. A His-tag (typically six consecutive histidines) strongly chelates many metal ions, including Cu(II), via the imidazole groups. If the His-tag is not removed, Cu(II) can bind to the tag rather than the native active site. This issue is particularly acute for studies of metal transporters or chaperones with intrinsically weak metal-binding sites. In sum, without careful control of experimental conditions, one can observe EPR signals from Cu–buffer complexes, Cu–chelator complexes, or Cu–His-tag adducts, all of which complicate spectral simulation, spin quantification, and the assignment of signals to the true active-site copper.

In this work, we systematically examine the background Cu(II) EPR signals arising from commonly used buffers and tags, and demonstrate methods to mitigate their interference. We first present a survey of Cu(II) EPR spectra in three widely used buffers (Tris, HEPES, MOPS) at varying pH (6.0–8.0), alongside EDTA for comparison, to characterize the complexes and precipitates these buffers form with Cu(II). We then discuss two case studies from our recent research: (1) a bacterial multicopper oxidase complex (MnxEFG) involved in Mn(II) biomineralization, and (2) a eukaryotic mononuclear copper enzyme, Drosophila lysyl oxidase (DmLOX), implicated in cancer development. In each case, we show how the choice of buffer and the presence of a His-tag affected the Cu(II) EPR spectra, and how we optimized the conditions to isolate the spectra of the true metalloprotein centers. Finally, we provide practical guidelines—such as choosing appropriate buffer systems and handling affinity tags—for purifying and reconstituting copper proteins with minimal adventitious Cu(II) binding. By highlighting and addressing the issue of background Cu(II) signals, this study serves as a reference to improve the accuracy of EPR characterization in bioinorganic chemistry and to aid in the correct assignment of copper sites in complex metalloproteins.

2. Experimental details

General consideration.

All chemicals were obtained from Sigma-Aldrich (Milwaukee, WI) unless otherwise noted. Buffer solutions were prepared at 20 mM concentration and adjusted to the desired pH with NaOH or HCl as needed. Ultrapure water and analytical-grade reagents were used to minimize metal contamination. Protein samples were expressed and purified by standard protocols (details for each specific protein are provided below).

EPR sample preparation.

Each EPR sample contained 100 μM Cu(II) (as CuSO₄), 20 mM buffer (Tris, HEPES, MOPS, or other specified buffer) and 100 mM NaCl. Samples were prepared at four pH values (6.0, 7.0, 7.5, 8.0) for each buffer. Ethylene glycol was added as a cryoprotectant to 20% (v/v) final concentration. After mixing, ~200 μL of each sample was loaded into a quartz EPR tube (Wilmad, 4 mm OD, 2.8 mm ID) and flash-frozen in liquid nitrogen. For Cu(II)–EDTA reference samples, CuSO₄ (100 μM) was pre-mixed with equimolar EDTA in water and then combined with 20 mM buffer as above. For protein samples, Mnx and LOX proteins were adapted from previous studies. CcoA sample was kindly provided by Professor Fevzi Daldal (UPenn).

EPR spectroscopy.

X-band (9.4 GHz) continuous-wave EPR spectra were performed on a Bruker (Billerica, MA) EleXsys E-500 spectrometer with a super-high Q resonator (ER4122SHQE) and an ESR900 cryostat with a temperature controller (Oxford Instruments ITC503). Copper (II) EPR spectra were recorded under slow-passage, non-saturating conditions with the following parameters: temperature, 15 K; microwave power, 0.02 mW; modulation amplitude: 0.5 mT; modulation frequency, 100 kHz. Spin quantification was performed by comparing the double integration of baseline-corrected EPR spectra to that of the EPR spectra of CuSO₄ solutions (50 μM, 100 μM, 200 μM) in 20% ethylene glycol. The EPR spectra and calibration curve are shown in Figure S1. Spectral simulation was performed in Matlab 2025a with easyspin package 6.0.10 package.[27]

3. Results

Cu(II) EPR signals in common buffers.

We first measured the EPR spectra of Cu(II) in three commonly used buffering agents—Tris, HEPES and MOPS—to characterize their interactions with Cu(II) across a range of pH. Phosphate buffer was not tested because Cu(II) is known to precipitate with phosphate as insoluble Cu₃(PO₄)₂. In addition, we included EDTA (a strong metal chelator) as a comparative benchmark for a tightly bound Cu complex. Each buffer (20 mM) was mixed with Cu(II) (100 μM) at pH 6.0, 7.0, 7.5, and 8.0. For CW-EPR measurements, a microwave power of 0.02 mW was used to ensure that the Cu(II) signals were not saturated, and the spin count can be accurately quantified. Figure 1 (A–D) displays the X-band CW-EPR spectra of these samples, and the parameters of the resolved species are summarized in Table 1.

Figure 1.

X-band CW EPR spectra (15 K, 0.02 mW) of Cu²⁺ in different biological buffers at different pH.

Table 1.

EPR parameters (g_‖ and A_‖) of Cu(II) species shown in Figure 1.

Copper chelator	pH	g ‖	A‖ (Cu), MHz	% total CuII	Notes
Tris	6.0	2.419; 2.339	410; 452	105	two species (1:1)
	7.0	2.238	580	56
	7.5			63
	8.0			61
HEPES	6.0	2.419	410	92	free Cu2+
	7.0	2.257	556	25
	7.5			31
	8.0			33
MOPS	6.0	2.419	410	104	free Cu2+
	7.0	-	-	17	signals too weak
	7.5			16
	8.0			13
EDTA	6.0	2.335; 2.294	435; 465	96	two species (1:2.5)
	7.0			93
	7.5			98
	8.0			97
His-tag	7.5	2.256	559	~100

Previous study indicated that Tris can form a variety of complexes with Cu(II) at different pH and ligand to metal ratio.[24] At pH = 6, Tris is mostly protonated (TrisH⁺, pKa ~ 8.1) and expected to have low affinity for metal ions. The resulted EPR spectra contains two species with nearly equal ratio, and a total spin count close to 100% of total Cu(II) (Figure 1A, see Figure S2 for simulation), indicating that essentially all Cu(II) is EPR-detectable without significant precipitation or EPR-silent species formed at acidic pH. The first species has g_‖ = 2.419 and A_‖ = 410 MHz, and matches the parameters of unbound [Cu(H₂O)₆]²⁺ (see Figure S3 for a comparison to the EPR of [Cu(H₂O)₆]²⁺ signal from a CuSO₄ solution). The second species (g_‖ = 2.339 and A_‖= 452 MHz) is assigned to a [CuTris]²⁺ complex according to prior study.[24] At pH 7–8, the behavior in Tris changed markedly. In neutral and alkaline Tris buffer, deprotonated Tris (Tris-H) can act as a stronger ligand. The EPR spectra at pH 7.0, 7.5, and 8.0 all showed a dominant species with g_‖ = 2.238 and A_‖ = 580 MHz, corresponding to either [CuTris₂-H]⁺ or [CuTris₂-2H] complexes.[24] The decreased g_‖ and increased A_‖ compared to the [CuTris]²⁺ species is consistent with less positive charge and increased equatorial ligand field of the deprotonated Tris ligand. Notably, the spin quantification for the pH 7–8 Tris samples recovered only ~55–60% of the total Cu(II) in the EPR signal. This suggests that a significant fraction of the Cu(II) became EPR-silent. A likely explanation is the formation of μ-O bridged binuclear [Cu₂Tris₂-2H]²⁺ and [Cu₂Tris₂-3H]⁺ complexes, in which the two Cu centers are antiferromagnetically coupled. Formation of di-copper species at higher pH and the percentage (~40%) is consistent with previous report.[24]

HEPES is often described as a “non-coordinating” buffer, but prior work indicates it can weakly bind Cu(II) under certain conditions.[25] Our EPR results support this conclusion (Figure 1B). At pH 6, the Cu/HEPES sample shows ~100% unbound [Cu(H₂O)₆]²⁺ signal, consistent with the “non-coordinating” nature of HEPES at acidic condition (Figure S2). At pH 7–8 where HEPES is (partially) deprotonated (pKa ~ 7.5), it forms a complex with Cu(II), with g_‖ = 2.257 and A_‖ = 556 MHz, which we assign as a [CuHEPES-H]⁺ complex. These parameters are similar to the [CuTris-H]⁺ complex but suggest a slightly weaker equatorial ligand field consistent with HEPES being a weaker ligand. Importantly, the EPR spin quantifications for HEPES at pH 7–8 recovered only ~25–30% of the total copper in the form of EPR-visible Cu(II). The majority (~70% or more) of the Cu(II) in these samples was missing from the spectra, which we attribute to the formation of a precipitate or EPR-silent aggregate. As noted in a previous study, Cu²⁺ forms complexes with HEPES and MOPS rapidly at pH = 7.4; however, Cu(OH)₂ precipitations were eventually formed in both buffers.[26] At similar or higher pH, we expect the formation of similar precipitations. However, the total amount of Cu(OH)₂ (0.098 mg per ml solution) might be too little to be observed even after centrifugation. Formation of soluble but EPR-active species such as dimers is also a possible explanation for the loss of EPR-detectable Cu(II) signals.

For MOPS buffer (pKa ~ 7.2), even less interaction with Cu(II) could be observed. At pH 6, the Cu/MOPS sample again shows ~100% free Cu²⁺ signals (Figure S2). As the pH increases, the Cu(II) EPR signal intensity drops dramatically. By pH 7.5–8.0, only ~15% or less of the Cu(II) added is EPR-detectable (and these residual signals are too weak and broadened to extract reliable g/A values). The rest of the Cu(II) presumably forms precipitate or EPR-silent species, similar to the scenario in HEPES. Unlike HEPES, MOPS at neutral pH does not even produce a distinct complex signal—it appears that MOPS does not significantly bind Cu(II) at all, and raising the pH mainly causes Cu(II) to form hydroxide precipitates. We conclude that MOPS has essentially no appreciable Cu(II) chelation ability under these conditions, consistent with its reputation as a non-coordinating buffer. Any Cu(II) present with MOPS at pH 7–8 will likely either remain free (and get hydrolyzed/precipitated) or bind to other available sites (such as the protein or impurities if present).

Finally, we examine EDTA as a control. The Cu(II)-EDTA complexes are stable from pH 6 to 8 and contain two species with g_‖ and A_‖ of 2.335, 435 MHz and 2.294, 465 MHz, respectively (see Figure S4 for simulation). According to previous study, the two species correspond to two different coordination modes—2 N, 3 COO⁻ and 2 N, 4 COO⁻—respectively.[28] The g_‖ and A_‖ values match well with reported numbers. The relative populations of these two species are affected by pH. At pH ~ 3, the N2O3 form predominates, whereas at pH 4–10 both forms coexist. Within the limited pH range examined in our experiments, the ratio between these two coordination isomers remains nearly constant. Notably, the Cu–EDTA samples showed very high spin recoveries (~95–100% of Cu was EPR-detectable at all pH), and no signs of precipitation.

Collectively, these buffer results demonstrate a spectrum of Cu(II) binding affinities: Tris > HEPES > MOPS (with EDTA binding even stronger than Tris). Tris can significantly compete for Cu(II) and even remove Cu from other sites at physiological pH, HEPES binds Cu weakly and mostly at neutral pH (with substantial loss of Cu to precipitation), and MOPS does not meaningfully bind Cu. These findings are important because they show how buffer choice can influence the “background” EPR signals in a copper protein sample. Interestingly, the Cu(II)-chelating ability of Tris can be turned to an advantage, as we explore next, by using Tris to selectively remove adventitiously bound Cu from protein samples.

Effects of buffer choice in multicopper oxidase MnxEFG complex.

Next, we show an example to illustrate the effects of different buffers in our study of the Bacillus MnxE₃F₃G complex. This enzyme complex contains multiple copper sites including Type I, Type II and Type III centers, akin to the laccase/MCO families. It is responsible for the biomineralization of manganese oxide in the oceanic bacteria.[29–31] Due to its large size (210 kDa) and moderate solubility, Mnx complex sample is typically obtained at low concentration (~50 μM). In early stage of our studies, MnxE₃F₃G was purified and dialyzed in HEPES buffer. The resulted protein sample exhibited an EPR spectrum shown in Figure 2A, blue trace. The parallel region of the spectrum contained multiple species that were inconsistent from different batches of protein preps. Moreover, a spin quantification indicated that there were 10–15 Cu(II) per protein complex, far exceeding the possible copper binding sites according to bioinformatics analysis. This inconsistency suggests that many of the Cu(II) ions in the sample were not tightly bound in the intended protein sites, but rather adventitiously associated.

Figure 2.

(A) X-band EPR spectra of MnxE₃F₃G dialyzed with different buffers. (B) X-band EPR spectra of DmLOX prepared at different conditions. EPR conditions are the same with those in Figure 1.

We hypothesized that switching to a buffer with moderate Cu-binding capacity, such as Tris, might help strip away the loosely bound, adventitious copper from MnxE₃F₃G. Indeed, when the MnxE₃F₃G complex was dialyzed against 20 mM Tris buffer, a more consistent spectrum was obtained (Figure 2A, green trace), and the spin quantification showed that 6 Cu(II) are present in each complex that can be assigned into two species. The improved consistency suggested that the majority of extraneous Cu(II) had been removed, leaving primarily the Cu ions intrinsically associated with MnxE₃F₃G’s structured metal sites. When the complex was further dialyzed against a higher concentration of Tris, 50 mM, the spectrum was simplified into a single species with 3 Cu(II) per complex (Figure 2A, red trace). This dialysis step also removed the characteristic blue color of the T1-Cu center. Our study verified that that the remained species is the Cu(II) site in the MnxE₃F₃ subunit, whereas the species that is lost during the second dialysis step is the Cu(II) sites associated with the MnxG subunit, which apparently is a Cu-binding site with weaker affinity.

These EPR experiments helped distinguish the different Cu sites in the MnxE₃F₃G complex, together with biochemical and later cyro-EM structural studies.[29, 31] A single MnxE₃F₃G complex contains 6 EPR-detectable Cu(II) sites: the MnxE₃F₃ hexamer harbors three T2-Cu sites with uncertain function. The MnxG subnit contains one T1-Cu site, one tri-nuclear Cu cluster and an additional T2-Cu site with of unclear role. Within the tri-nuclear site, a pair of Cu²⁺ ions are antiferromagnetically coupled with S_tot = 0, while the remaining Cu²⁺ give rise to T2-Cu EPR signal. In the MnxG subunit, both the T1-Cu (2 His, 1 Cys, 1 weak axial Met) and the tri-nuclear Cu sites (3 His) adopt active site geometries typical of multicopper oxidases. In MnxEF, the exact coordination environment of Cu(II) and the function of these sites are still unknown. A possible explanation for the higher Cu binding affinity in MnxEF is that MnxEF forms more tightly bound Cu sites for structural roles, whereas the MnxG subunit host more accessible Cu centers for substrate binding (O₂ and Mn²⁺ in this case) and catalysis.

These results also highlight a critical point: buffer-mediated removal of copper can be tuned to selectively strip adventitious metal while retaining native metal sites, if done carefully. In the case of MnxE₃F₃G, HEPES was ineffective at removing any nonspecific Cu (hence the over-loaded, messy spectra), whereas Tris at 20 mM removed the bulk of loosely bound Cu, and 50 mM Tris removed even the moderately bound Cu which turned out to be a genuine but weak site on MnxG. One must, however, use caution: too strong a chelator or too high a concentration can begin to extract metal from sites that are part of the protein’s functional complement. In our study, 50 mM Tris was slightly too effective, stripping a site intended to keep occupied. The good news is that this process is reversible—when the MnxE₃F₃G complex with only 3 Cu(II) (red trace) is supplemented with ~3x Cu²⁺ in 20 mM Tris, full Cu loading in the MnxE₃F₃G complex can be regained (green trace). This indicates that the additional Cu-sites in MnxG is re-incorporated without introducing any adventitious Cu(II) signals. This also demonstrated an interesting utility of Tris: the Cu–Tris complex itself can serve as a gentle copper donor during reconstitution. Since Tris binds Cu(II) moderately well, it keeps Cu in solution (preventing precipitation) but also readily hands it off to higher-affinity sites on the protein when available while minimizing nonspecific binding.

Histag-bound Cu(II) and its effect in lysyl oxidase (DmLOX).

Recombinantly over-expressed proteins are usually attached to affinity tags to facilitate purification, His-tag among the most frequently used. Non-specific binding of various metal ions to His-tag could be a problem in studying metalloproteins. This problem is particularly important in studying metal transporter proteins since they generally have less metal binding affinities than do metalloenzymes. Our second case study involves Drosophila melanogaster lysyl oxidase (DmLOX), a mononuclear Type II copper enzyme (a copper amine oxidase) that catalyzes the oxidative crosslinking of collagen and elastin in the extracellular matrix.[32] The human analog of this enzyme is involved in cancer metastasis, making it of biomedical interest.[33] DmLOX is challenging to work with due to its poor solubility and tendency to misfold or aggregate when expressed recombinantly. In our expression system, we produced DmLOX with an N-terminal His₆-tag to facilitate affinity purification. We examined how the His-tag and the conditions of reconstitution (folding in presence of urea) affected the Cu(II) EPR spectra of DmLOX.

When the DmLOX protein was reconstituted at the presence of 2 M urea and dialyzed with 20 mM Tris, the EPR spectrum of the resulted sample contained two species due to pH speciation in DmLOX (Figure 2B, blue trace). Notably, the spectrum remained essentially the same if the sample was further dialyzed against EDTA instead of Tris (Figure 2B, green trace), again justifying the effectiveness of using Tris to remove adventitious Cu(II). This sample also showed high activity in an enzymatic assay, confirming the formation of properly folded protein. In contrast, attempts to reconstitute DmLOX at lower urea concentration, 0.2 M, resulted in a new species in EPR (Figure 2B, red trace), with a Cu(II) incorporation percentage of ~ 100 %. The EPR spectrum of this species, however, has virtually the same lineshape, with g_‖ = 2.256 and A_‖ = 559 MHz, to that of another sample consisted of Cu(II) bound to Rhodobacter capsulatus CcoA protein, which is a Cu transporter containing a His-tag for affinity purification (Figure 2B, magenta trace). When the His-tag is removed, CcoA did not show any obvious Cu(II) binding, indicating that the Cu(II) species in the magenta trace is most likely a His-tag bound Cu(II). This further suggests that the dominant Cu(II) species formed in the low-urea DmLOX preparation (red trace) was Cu(II) bound to the His-tag, rather than Cu in the enzyme’s active site. The absence of any other species in this sample suggests that essentially every DmLOX molecule ended up with a Cu(II) at the His-tag and not in the native site, an undesirable outcome for functional studies.

Even in the 2 M urea samples, careful inspection of the EPR spectrum revealed subtle features (indicated by dash lines and the g_⊥ symbol in Figure 2B) corresponding to the His-tag bound Cu signal. This implies that a small fraction of DmLOX had Cu bound to the tag even under high-urea conditions, though it was a minor component compared to the native-site Cu. Collectively, these results highlight how strongly a His-tag can compete for Cu(II) binding. If a protein is not perfectly folded or if the active site is not immediately available to capture Cu, the His-tag will scavenge the metal, especially for metalloproteins that have lower affinity for copper, such Cu transporters or enzymes that cycle through Cu(I)/Cu(II) states.[34] In these cases, any EPR (or other spectroscopic) characterization might predominantly observe the tag-bound Cu(II) and mistakenly attribute it to the protein’s active site. In the case of DmLOX, recognizing the His-tag EPR signature allowed us to avoid such misassignment.

4. Discussion

Our results underscore that sample preparation is a critical factor in spectroscopic studies of metalloproteins, particularly for techniques such as EPR that are sensitive to even minor paramagnetic impurities (~μM). By systematically examining how buffers and affinity tags interact with Cu(II), we have revealed some common pitfalls and developed strategies to obtain more reliable EPR data for copper proteins.

Our work highlights how easily adventitious Cu(II) can bind to components of a biochemical sample, leading to misleading observations. For example, the initial HEPES-buffered MnxEFG samples appeared to have ~10–15 Cu per complex with multiple distinct EPR signals. The true nature of Cu-binding in MnxEFG was only revealed after switching buffers. Similarly, in the case of DmLOX, a researcher might have observed the strong g_‖= 2.256 signal and erroneously concluded it represents the native Cu site, whereas in reality it was a His-tag artifact. These scenarios mirror challenges in structural characterization and metal site assignment in metalloenzymes. In X-ray crystallography or cryo-EM, adventitiously bound metals can also appear in electron density maps, potentially leading to confusion over which metal ions are functionally relevant. Our results suggest that coupling biochemical preparation techniques with spectroscopic validation (EPR) is a powerful approach to ensure that the metals present in a sample are truly those intended and not a byproduct of preparation.

Our findings also have implications for metalloenzyme reconstitution and metalloprotein engineering. We observed that a moderate chelator (Tris) was extremely useful both in removing loosely bound metals and in gently reloading a metal site. In the case of Cu proteins, the as-isolated protein usually contains adventitiously bound copper, which can be removed by dialyzing the protein against a metal chelating reagent. EDTA has been frequently used for such a role. Occasionally, EDTA can be used to distinguish various Cu species in the protein sites, as seen with the red copper protein nitrosocyanin,[35] where the as-isolated protein contained two Cu(II) species, with g_‖ of 2.245, 2.224 and A_‖ of 432 and 629 MHz, respectively, and the latter was removed by treating with EDTA followed by gel filtration. This “labile” Cu^II site was suggested to be at the N-terminus of nitrosocyanin where a His residue is present.[35] However, the strong chelating ability of EDTA might affect the copper binding in loosely bound copper sites. For instance, if the MnxEFG complex is dialyzed against EDTA, only ~0.2 Cu²⁺ remained to be bound to each complex unit, while in the case of DmLOX, the spin counts was ~0.1 Cu²⁺/protein after dialyzing against EDTA. Instead, Tris buffer can serve as a weak Cu chelating agent that effectively removes non-specific Cu²⁺. With careful tuning, this approach could offer more advantages than the traditional approach using strong chelators.

With regard to affinity tags, His-tag could be problematic when expressing Cu, or many other metalloproteins due to the strong binding of Cu to the tag, as seen in the cases of the Cu transporter CcoA and DmLOX.[32] Although the contribution of His-tag bound Cu(II) to the EPR of DmLOX was diminished at the presence of 2 M urea, such conditions are not typical for most proteins. Previous study have shown that Cu(II) binds to His₆ polypeptides primarily through two or more imidazole groups and several backbone amides under near-physiological conditions, resembling the interactions seen with Cu(II) and amyloid-β peptides in Alzheimer’s disease.[36, 37] Moreover, some endogenous proteins, usually with high His contents, could be co-purified from a His-tag chromatography,[38] many of which have metal binding affinities and introduce additional metal-binding impurities, complicating EPR spectra. Instead, Strep-tag proves to be a better purification tag and no obvious Cu binding affinity is observed in our study. Beyond purification strategy, proper in vivo incorporation of copper is often critical for correct protein folding and activity. Previous studies have shown that Cu accumulates inside E. coli cells at microaerobic conditions,[39, 40] and Cu incorporation into several multicopper oxidases is greatly improved if a microaerobic incubation stage is included during cell growth for protein over-expression.[41] We found this strategy equally effective for the MnxEFG complex.

Finally, our work demonstrates that the use of EPR as a diagnostic tool during sample prep can greatly facilitate the identification and removal of unwanted metal species. Since Cu(II) EPR is very sensitive and species-specific (each type of copper environment has a distinct signature), even minor contaminants can be detected. by recording spectra after each purification or treatment step, the removal of adventitious Cu can be directly monitored, as demonstrated for the MnxEFG complex. Our work shows that EPR is not only a probe of active sites but also a valuable aid in refining sample preparation. The broader lesson for bioinorganic chemistry is that careful attention to buffer composition, pH, and purification tags, combined with iterative spectroscopic verification, can greatly improve the reliability of data on metal sites. This approach will aid in surmounting challenges in metalloprotein research, such as assigning metals to the correct sites and ensuring that what we observe truly reflects the biology of the system rather than artifacts of preparation.

5. Conclusion

In summary, we have characterized the background EPR signals arising from Cu(II) interacting with common biochemical buffers and demonstrated their relevance in real-world samples of copper proteins. Tris, HEPES, and MOPS buffers show distinctly different capacities to bind Cu(II), with Tris>HEPES>MOPS. These differences proved crucial when studying copper-containing enzymes. By using Tris buffer in a controlled manner, we successfully eliminated adventitious Cu(II) from a multicopper enzyme complex (MnxEFG), thereby clarifying its EPR spectrum and copper stoichiometry. We also identified a signature EPR signal for Cu(II) bound to a polyhistidine tag and showed how this can confound the spectral analysis of a His-tagged copper enzyme (DmLOX) if not addressed. Based on our findings, we provided practical guidelines for the preparation of copper protein samples: including recommendations on buffer selection, use of mild chelators, management or removal of His-tags, and verification of metal content via spectroscopy. Overall, these results are of general interest to the bioinorganic community as a reference for avoiding and recognizing spurious Cu(II) signals. We anticipate that the strategies outlined here will be applicable to many other metalloenzymes, helping to ensure that what we observe in the EPR samples faithfully represents the biology of the metalloprotein under investigation.

Highlights.

• Biological buffers show distinct Cu(II) binding behaviors and background EPR signals.

• Buffer selection and His-tags influence observed Cu(II) spectra in copper proteins.

• Practical guidelines are provided for minimizing nonspecific Cu(II) binding.