Combining Native Mass Spectrometry and Proteomics to Differentiate and Map the Metalloform Landscape in Metallothioneins

Within the intricate landscape of the proteome, approximately 30% of all proteins bind metal ions. This repertoire is even larger when considering all the different forms of a protein, known as proteoforms. Here, we propose the term “metalloforms” to refer to different structural or functional variations of a protein resulting from the binding of various hetero- or homogeneous metal ions. Using human Cu(I)/Zn(II)-metallothionein-3 as a representative model, we developed a chemical proteomics strategy to simultaneously differentiate and map Zn(II) and Cu(I) metal binding sites. In the first labeling step, N-ethylmaleimide reacts with Cysteine (Cys), resulting in the dissociation of all Zn(II) ions while Cu(I) remains bound to the protein. In the second labeling step, iodoacetamide is utilized to label Cu(I)-bound Cys residues. Native mass spectrometry (MS) was used to determine the metal/labeling protein stoichiometries, while bottom-up/top-down MS was used to map the Cys-labeled residues. Next, we used a developed methodology to interrogate an isolated rabbit liver metallothionein fraction containing three metallothionein-2 isoforms and multiple Cd(II)/Zn(II) metalloforms. The approach detailed in this study thus holds the potential to decode the metalloproteoform diversity within other proteins.


■ INTRODUCTION
Approximately 30% of human genes encode proteins that bind metal ions as cofactors to catalyze reactions, stabilize protein structures, or regulate total or mobile metal concentrations. 1 Those estimations consider only a single existing state of a protein, whereas in reality, a protein can exist in multiple states, collectively known as proteoforms. 2Among all essential metals in mammals, Zn(II), Fe(II)/(III), and Cu(I)/(II) are critical for enormous physiological functions. 3The proper regulation of these metal ions is crucial to the overall homeostatic stability of an organism.Mammalian metallothioneins (MTs) play a crucial role in maintaining the homeostatic control of Zn(II) and Cu(I) ions in cells, ensuring the proper balance and regulation of these metal ions within cellular environments. 4,5−12 As a result of this biological process, MT3 exists in multiple heterogeneous and homogeneous Cu(I)/Zn(II) states, referred to as metalloforms.
−19 The detection and analysis of cysteine residues often rely on using thiol-specific probes that react with the nucleophilic sulfur of cysteine. 20,21These probes exhibit different reactivity profiles, allowing for the differentiation of various forms and states of cysteine.Commonly employed protein thiol probes include iodoacetamide (IAM), iodoacetic acid (IAA), N-ethylmaleimide (NEM), methylmethanethiosulfonate (MMTS), Cys-reactive mass tag (cys-TMT), and p-benzoquinone (Bq). 22,23IAM and NEM are widely used probes that covalently react with sulfhydryl groups using different mechanisms. 13,24IAM undergoes S N 2 nucleophilic substitution, while NEM proceeds via nucleophilic Michael addition.NEM exhibits a faster reaction rate with thiols than IAM and is effective over a wider pH range, including acidic conditions. 24In contrast, IAM requires neutral or basic pH. 20However, NEM shows lower specificity, potentially leading to side reactions with histidine and lysine residues at high excess or basic pH.IAM is preferred due to its formation of stable thioether bonds, while NEM may undergo partial ring hydrolysis.−35 This method involves the initial blocking of a reduced free thiol using one alkylator, followed by a reduction step and subsequent labeling using a different alkylator.By employing this approach, cysteine redox modifications can be identified and characterized, providing insights into the redox status of cysteine residues.These chemical methodologies are usually coupled with highresolution mass spectrometry (MS).Our previous research presented a differential labeling approach capable of distinguishing between free Cys and Zn(II)-protected Cys residues within zinc sites, exhibiting nanomolar and picomolar metal affinities. 36his methodology has been used to demonstrate the absence of the Zn(II) ion in alcohol dehydrogenase 1 in a murine model of alcohol-associated liver disease. 19ere, we developed an MS-based differential alkylation strategy to distinguish and map metalloform landscapes.As a representative model protein, we selected heterogeneous Cu(I)/Zn(II)-MT3, which has been previously demonstrated to exist as an air-stable Cu(I) 4 Zn(II) 3−4 MT3 complex when isolated from mammalian brains. 37,38MT3 folds in a dumbbellshaped structure comprising 68 amino acid residues and encompassing two domains. 39,40Spectroscopic studies have unveiled the interaction of MT3 with divalent metal ions (M), resulting in the formation of an M 3 Cys 9 metal cluster in the Nterminal β-domain and an M 4 Cys 11 metal cluster in the Cterminal α-domain. 8,41,42−46 Interestingly, no preference for a specific domain was found when using Cu(I) titrations on Zn 7 MT3. 47,48Using this protein, we demonstrated that NEM could be employed in the initial labeling step to dissociate all Zn(II) ions while Cu(I) remains bound due to its higher affinity for Cys.In the second labeling step, an excess of IAM is utilized to label Cu(I)-bound Cys residues, leading to Cu(I) dissociation.Native MS is used to monitor the reaction at each step, and bottom-up/top-down MS techniques are utilized to map the Cys-IAM and Cys-NEM labeled residues, allowing for the determination of the locations of Cu(I) and Zn(II) within the protein, respectively.The developed methodology was then tested on an isolated rabbit liver fraction containing Nacetylated metallothionein-2a/b/c isoforms.NEM was used to dissociate Zn(II) ions, while Cd(II) remained bound to the protein.Next, IAM was employed to dissociate tightly bound Cd(II) ions and label the Cys residues.Combining native MS with proteomics approaches unveiled the location of Zn(II) and Cd(II) ions in all metallothionein-2 isoforms.The described methodology has the potential to be expanded for the investigation of metalloforms.

■ EXPERIMENTAL SECTION
Metallothionein-3 (MT3) (Addgene plasmid ID 105710) was overexpressed and purified in a bacterial system as described in the Supporting Information.Rabbit liver metallothionein-2 (MT2) was purchased from Santa Cruz Biotechnology, Inc., USA.The preparation of the samples for the experiments is described in the Supporting Information.Electronic spectroscopy experiments were performed on a JASCO V-750 spectrophotometer, at 25 °C, in a 1 cm quartz cuvette, as described in the Supporting Information.
MS experiments were carried out on a Synapt XS HDMS equipped with nanoelectrospray ionization (Waters Corp., UK) using a 500−5000 m/z range.5−10 μL of the sample (10−20 μM in 200 mM ammonium acetate) was loaded into borosilicate glass capillaries (O.D. 1.2 mm, I.D. 0.9 mm, World Precision Instruments, Stevenage, UK) produced in-house using a Flaming/Brown P-1000 micropipette puller (Sutter Instrument Co., Novato, CA, USA) and ions were produced by applying a positive potential of 0.9−1.4kV via a platinum wire (Goodfellow).
Native top-down collision-induced dissociation (CID) MS experiments were performed by applying 20−60 V of trap collision energies of quadrupole-selected ions with argon as the collision gas.Detailed calculation description of the survival yield (SY) curves and data analysis can be found in the Supporting Information.Top-down electron transfer dissociation (ETD) MS experiments were performed by introducing the sample via syringe with a 2−3 μL•min −1 flow rate and a spray voltage of 2−2.5 kV.The glow discharge was tuned to obtain an ETD reagent (1,4-dicyanobenzene) current of ∼1 × 10 6 counts/ s for charge reduction.The anions were accumulated in the trap collision cell for 100 ms using a refill interval of 1 s.The reaction was started by lowering the wave height from 1.5 to 0.2−0.3V, using a wave velocity of 300 ms −1 .
To perform bottom-up MS experiments, single or dual labeled MT3 proteins were buffer exchanged to 100 mM ammonium bicarbonate.Trypsin (Sigma-Aldrich) was added at 1:1 (w/w), followed by the addition of 0.1% RapiGest SF (Waters Corp., UK) and overnight digestion at 37 °C.Samples were acidified to 0.5% formic acid (FA) to quench digestion, centrifuged, and desalted on Stage Tips tips eluting the peptides with 80:20 ACN/H 2 O 0.1% FA.After removing ACN (acetonitrile) by speed-vac, the peptides were resuspended in 0.1% FA.Samples were analyzed via LC−MS using a Waters Acquity UPLC mclass system coupled to a Synapt XS HDMS operated in a positive ion/resolution mode.Peptides were first trapped on a Waters Acquity BEH C18 1.7 μm VANGUARD column and then separated on a Waters Acquity UPLC BEH C18 1.7 μm, 1.0 × 100 mm.Mobile phase A consisted of 0.1% FA (v/v) in Milli-Q water, while mobile phase B consisted of 0.1% FA (v/v) in ACN.The LC gradient was supplied at 40 μL•min −1 over a 22 min gradient (5−35% B).MS analysis was performed using MS e data independent acquisition, which used a collision energy ramp from 20 to 45 V in a 50−2000 m/z range for the high energy scans.Raw LC-MS e data files were processed using a ProteinLynx Global Server 3.0.3(Waters Corp., UK) and searched against the MT3 fasta protein sequence (UniProt Journal of Proteome Research P25713).The search was performed using 15 ppm as a precursor mass tolerance and a fragment mass tolerance of 20 ppm with three minimum fragment ion matches per peptide.Trypsin was set as the protease with a maximum of three missed cleavage allowed.Cysteine carbamidomethyl, cystine, and Cysteine Nethylmaleimide were set as variable modifications.Leu-Enkephalin (556.27m/z) was used as a lock-mass.Peak lists were exported as text files and analyzed using custom scripts in Python 3.5.Native MS and top-down MS experiments were analyzed using MassLynx v4.2 (Waters Corp., UK) and custom scripts in Python 3.5 (available at https://github.com/ManuelPerisDiaz/Cu-Zn-MT3) as described in the Supporting Information.

■ RESULTS AND DISCUSSION
In order to develop a differential labeling strategy that may selectively target Cys residues in proteins that bind Zn(II) and Cu(I) metal ions, we first prepare the heterogeneous Cu(I) 4 Zn-(II) 4 MT3 ox complex.Addition of four CuCl 2 molar equivalent (mol Eq) to Zn 7 MT3 red (reduced and fully Zn(II)-loaded MT3) results in the formation of the Cu(I) 4 Zn(II) 4 MT3 ox along with other metal complexes.
We observed at least seven metalloforms (Cu  .The mechanism of the reaction involves a Zn(II)-to-Cu(I) exchange by reducing Cu(II) to Cu(I) through the formation of two intramolecular disulfide bonds (indicated here as "ox") (Figures 1 and S5).To precisely annotate the ions to particular complexes, multiple theoretical isotopic distributions for the protein were generated, and the one demonstrating the optimal fit was selected as outlined in Figure S1.We observed all series from tetra-to heptametallic complexes, which suggests that the reaction proceeds by first Zn(II) clusters disassembly (Zn 3 S 9 and Zn 4 S 11 ) and then reassembly following a probability distribution.Monitoring intensities of the ligand-to-metal charge transfer (LMCT) transition in the middle-to-far UV range also showed that Zn 7 MT3 red is able to bind and reduce Cu(II) to Cu(I) with concomitant Zn(II) displacement and disulfides' formation (Figure S2).

Profiling Reactive Cysteine Residues in Cu(I)/Zn(II)-MT3 ox Species by NEM and IAM
Once we had prepared the heterogeneous Cu(I) 4 Zn(II) 4 MT3 ox complex, we studied the kinetic and thermodynamic lability of the metal−thiolate bonds by using the cysteine alkylators IAM and NEM, which follow a different reaction chemistry.While IAM follows a nucleophilic substitution S N 2, NEM reacts following a nucleophilic addition.To do so, the metal-protein complex was incubated with increasing concentrations of both alkylation reagents for different periods of time, and the reaction was monitored by native MS.The addition of IAM to the cocktail of Cu(I)/Zn(II)-MT3 complexes formed upon Cu(II) reaction with Zn 7 MT3 red results in a mixture of Cu(I) x Zn-(II) y IAM z MT3 ox species, coexisting simultaneously (Figure S3A).In our previous paper, we reported that IAM is a very good choice not only for labeling free Cys residues but also for mapping Zn(II) binding sites.Its low reactivity prevents it from competing with Zn(II) ions for Cys residues. 36As we have partially metal-loaded species upon Cu(II) reaction with Zn 7 MT3 red, IAM reacts with free Cys residues in the Cu 4 Zn 1−3 MT3 ox complex as well as Cu 4 MT3 ox .Then, the use of IAM leads to the stepwise dissociation of four Zn(II) ions in Cu 4 Zn 4 MT3 ox and Cu 5−6 Zn 4 MT3 ox , and finally, at high concentrations (100 mM), IAM can react with Cu(I)-Cys bound residues, dissociating the Cu(I) ions (Figure S3B).However, at different combinations of IAM concentration and reaction time, we have multiple and heterogeneous Cu(I)/ Zn(II)-MT3 complexes, which can be very difficult to interrogate by using a well-established bottom-up proteomics approach.As the main aim is to develop a differential labeling strategy to map Cu(I)/Zn(II) binding sites in proteins, we then decided to use NEM.We have shown that NEM reacts cooperatively and fast with Cys residues that had bound Zn(II) with weak and moderate binding affinities.Even at low NEM concentrations, we clearly observe selective dissociation of most of the Zn(II) ions while keeping Cu(I) bound to the protein (Figures 1 and S4 and S5).Increasing NEM concentration results in complete dissociation of Zn(II), obtaining products that only contain four and five Cu(I) ions, and a different extent of NEM modification.Such results highlight that NEM produces homogeneous Cu(I)-MT3 ox complexes and could  S5).The proteins (10 μM) were sprayed in ease the identification and conclusions about NEM moieties using a bottom-up MS approach (Scheme 1).We could observe that the modification profile was centered at Cu-(I) 5 NEM 7−9 MT3 ox and Cu(I) 4 NEM 8−12 MT3 ox species, and these remain the most intense ions upon doubling the NEM concentration.The difference in two NEM moieties between both proteoform families arises from the coordination chemistry, as Cu(I) tends to bind to Cys residues forming a linear geometry.The NEM experiments revealed that out of 20 Cys residues in the MT3 protein, 13 Cys residues could be modified by NEM, while the four Cu(I) ions were bound to MT3 (Figures 1 and S4).

CID Experiments on Cu(I)/Zn(II)-NEM x MT3 ox Species
To identify the ligands bound to MT3 and confirm the stoichiometries inferred from the native MS experiments, we employed a native top-down MS approach.First, we massselected the Cu(I) 4 Zn(II) 4 MT3 ox 5+ ions and subjected them to CID.The CID spectra contain an m/z region, with ions corresponding to the precursor and the precursor with the loss of some metal ions, thus providing direct information about the number and nature of metal ions bound to the protein (Figure S6).Under collisional activation (CA) conditions, we observed a stepwise dissociation of the Zn(II) ions from Cu(I) 4 Zn-(II) 4 MT3 ox 5+ to form an intermediate of Cu(I) 4 MT3 ox 5+ ions (Figure 2A).The signals were isotopically resolved and therefore permitted us to observe that increasing CA leads to a cooperative dissociation of the four Cu(I) ions from Cu(I) 4 MT3 ox 5+ to form Cu(I) 0 MT3 ox 5+ (Figure S7).This experiment demonstrates that CA can be used to induce metal ion dissociation and, therefore, identify the ligands that bind to proteins.Moreover, it is evident here that applying CA to intact metal-protein complexes is not the best choice for mapping metal binding sites, as we obtain partial metal ion dissociation.One strategy to avoid this issue is the use of Cys alkylators.However, the unambiguous determination of alkylator and metal protein stoichiometries is not always trivial due to the close masses.By applying the CID approach, we can further identify the precise Cu(I):NEM:MT3 stoichiometries (Figure 2B).As NEM remains bound to the protein upon CID activation, we could use a native top-down MS approach to localize which Cys residues are modified and, therefore, infer where Zn(II) ions were previously bound in the Cu(I) 4 Zn(II) 4 MT3 ox complex (Scheme 1).To this end, we mass-selected and fragmented each of the Cu(I) 4 NEM x MT3 ox 5+ (x = 8−12) ions as the NEM modification profile was centered around these species (Figures 1 and S11 and S12).The CID spectrum revealed three regions: a metal-free b/y fragment ion in the 300−800 m/z region, metalfree y-fragment ions in the 800−1200 m/z region, and yfragment ions with four Cu(I) ions bound, observed at 1700− 2000 m/z (Figure S12).To compare all fragmentation spectra, we calculated a correlation matrix based on ion intensities (Figure 3A).We observed a high correlation between the fragmentation patterns of all Cu(I) 4 NEM x MT3 ox 5+ (x = 8−12) complexes, with the similarity decreasing as the NEM stoichiometry increased.As a control to ensure that NEM alkylation did not induce Cu(I) migration to non-native metal binding sites, we obtained a CID spectrum of Cu(I) 4 MT3 ox 5+ (Figure 3B).A low correlation coefficient (R = 0.08−0.12)between the spectra of Cu(I) 4 MT3 ox 5+ and all Cu-(I) 4 NEM x MT3 ox 5+ (x = 8−12) complexes was obtained (Figure 3A).
A more comprehensive analysis of the fragmentation spectrum that avoids using peak intensities was conducted using an UpSet plot analysis (Figure 3C).This analysis allowed us to calculate the percentage of detected fragment ions common among all fragmentation spectra.In agreement with the correlation analysis, the Cu(I) 4 NEM x MT3 ox 5+ (x = 8−12) complexes shared ca.70−80% of the detected peaks and displayed high similarity.By employing this analysis, we demonstrated that ca.30% of the ions were shared between Cu(I) 4 MT3 ox 5+ and/or Cu(I) 4 Zn(II) 4 MT3 ox 5+ complexes with Cu(I) 4 NEM x MT3 ox 5+ (x = 8−12) complexes.In the next step, we annotated the previously detected peaks and observed that 75% of b-fragment ions were shared between both protein species.For Cu(I) 4 MT3 ox 5+ , the Cu(I)-thiolate protects against generation of b-fragment ions, resulting in the annotation of only b2 to b8 fragments.The metal-free α-domain is easily fragmented upon CA, generating a series of y-fragment ions.For Cu(I) 4 NEM 10 MT3 ox 5+ , we also annotated unmodified b2 to b5 fragment ions as well as specific b6 + 1NEM, b7 + 1NEM, and b8 + 2NEM fragment ions.
Similarly to Cu(I) 4 MT3 ox 5+ , the metal-free α-domain was completely fragmented but generated NEM-labeled y-fragment ions (Figure 3B).This fragmentation pattern indicates that Cu(I) binding remains at the native sites after the NEM reaction.A close examination of the annotated fragments allowed us to conclude that the α-domain is metal-free, and the four Cu(I) ions are bound in the β−domain (Figure 3D).When using Cu(I) as a metal source rather than Cu(II), no domain preference was observed. 47,48Therefore, these results indicate that the intramolecular disulfide formation formed upon Cu(II) reduction to Cu(I) dictates the domain preference and not the nature of the metal ion.
Due to the low sequence coverage in the N-terminus β− domain, covering until the b8 fragment ion, one might suspect that this is due to metal protection by the Cu(I) ions.However, we have previously shown that even in apoMT3, complete fragmentation of the β−domain is not possible using CA. 50Our conclusions are based on obtaining almost a complete sequence coverage of y-ions spanning the α-domain, which exhibited varying degrees of NEM modification, up to 7 NEM moieties (y35 + 7NEM).Additionally, when considering the b8 + 2NEM fragment ion, we accounted for 9 out of the 10 NEM modifications observed in the Cu(I) 4 NEM 10 MT3 ox parent ion.These results indicate that Cys6 and Cys8 in the β−domain weakly interacts with Cu(I) ions, as they could be labeled by NEM.Two b-type fragment ions were found, matching with two proteoforms.A proteoform with Cys5 unmodified and a second proteoform with Cys5 and Cys7-NEM modified.Similarly, the α-domain does not bind Cu(I) ions either.By using this native top-down MS approach, we were able to determine the locations of several Cys-NEM labeled residues, such as Cys66, Cys67, and Cys64.In addition, we reconstructed the metal and NEM stoichiometries of the parent ion Cu(I) 4 NEM 10 MT3 ox by combining two fragment ions: y62 + 9NEM and b6 + 1NEM.In another attempt to localize the remaining Cys-NEM labeled residues, we performed top-down ETD (Figure S13).However, instead of obtaining ETD fragmentation, we observed a nondissociative electron transfer dissociation (ETnoD). 51o determine the locations of all NEM moieties in the Cu(I) 4 NEM x MT3 ox 5+ (x = 8−12) complexes, we initially prepared the Cu(I)/Zn(II)-MT3 as described above.Afterward, we added 25 mM NEM to the resulting mixture, and the sample was digested into peptides.These resulting peptides were then analyzed by bottom-up liquid chromatography−MS (nanoLC-MS/MS).The data revealed complete sequence coverage of Cys-NEM labeled residues from peptides covering the β− domain (residues 1−31) and α-domain (residues 32−68) (Figure S14A).This can be explained by the observation that, even though NEM produces metal homogeneous Cu-(I) 4 NEM x MT3 ox 5+ (x = 8−12) complexes, portions of NEM 16−20 MT3 ox 5+ ions remain present upon reaction (Figure 1).Additionally, in a bottom-up MS, the peptide-to-proteoform connectivity is lost, and we cannot infer the origin of the peptide. 52Based on our recent findings that the Cu(I) 4 -thiolate cluster remains stable even under denaturing conditions, we decided to skip any desalting/purification step after enzymatic digestion, and the resulting peptide mixture was directly infused into the mass spectrometer. 53The recorded peptide mass fingerprint similarly displayed all Cys-NEM labeled residues, but we were also able to observe three peptides that covered the full β-domain (residues 1−31) and were bound to four Cu(I) ions (Figure S14B).Remarkably, these peptides also had two NEM moieties bound, consistent with our top-down MS analysis, which revealed that the β-domain binds four Cu(I) ions and that the Cys6 and Cys8 residues are labeled by NEM.Previous lowtemperature (77 K) luminescence spectroscopy data identified 5−6 thiolate-Cu(I) bonds in the Cu(I) 4 Zn(II) 4 MT3 complex, suggesting the formation of a Cu(I) 4 Cys 5 cluster and two intramolecular disulfides. 44Our high-resolution native MS data support the presence of two intramolecular disulfides in the Cu(I) 4 Zn(II) 4 MT3 complex (inset of Figure 1A), but they also reveal a population with only one disulfide.This matches our observations of two b-type proteoforms.Considering that four Cu(I) ions are bound in the β-domain, which contains 9 Cys residues, and that Cys6 and Cys8 are found both modified and unmodified, we conclude that Cu(I) 4 Cys 5 cluster and two disulfides, involving Cys6 and Cys8, are formed.The location of second disulfide remains to be elucidated.

Journal of Proteome Research Native Top-Down and Bottom-Up MS of NEM x IAM y MT3 ox Species
In order to further strengthen our conclusions, we extended the single to a double labeling approach (Scheme 1).Briefly, in this step, NEM-labeled Cu(I)-MT3 ox complexes were incubated with 100 mM IAM to react with Cys-Cu(I) binding residues.This process involved IAM alkylating Cys residues, resulting in subsequent Cu(I) dissociation.The recorded native mass spectrum shows a vast number of partially overlapping signals originating from multiple IAM stoichiometries for a particular NEM-MT3 complex (Figure 4A).For instance, NEM 12 IAM 3−7 MT3 ox or NEM 11 IAM 1−7 MT3 ox products were annotated (Figure 4B).We observed a repetitive pattern where three complexes could be annotated within small m/z windows, following this rule: NEM x IAM y MT3 ox , NEM x+1 IAM y−2 MT3 ox , and NEM x+2 IAM y−4 MT3 ox , where x and y are the NEM and IAM stoichiometry (Figure S15).For example, in the most intense 1730 m/z ion region, we annotated three complexes: N E M 1 1 I A M 6 M T 3 o x , N E M 1 2 I A M 4 M T 3 o x , a n d NEM 13 IAM 2 MT3 ox .To localize which Cys residues were modified with IAM and therefore infer where Cu(I) ions were previously bound in the Cu(I) 4 NEM x MT3 ox complex, we selected several m/z windows and employed top-down MS (Scheme 1).We observed that the fragmentation spectra of these complexes shared most of the ions (approximately 90%), and their intensities showed a high correlation (R = 0.97−0.99)(Figure S16).The CID fragmentation map of the 1730 m/z ions, comprising the above-mentioned complexes, revealed the presence of multiple b-fragment ions labeled with varying extents of IAM modifications (Figure 4C).
As in the case of the single NEM labeling approach, the bottom-up MS results were ambiguous as many Cys residues were found to be labeled by both IAM and NEM due to the sample heterogeneity (Figure S17).In this case, the top-down MS approach proves to be beneficial as the dissociation of metals and the alkylation of Cys residues induce protein unfolding, resulting in enhanced fragmentation and achieving high sequence coverage with overlapping b-and y-fragment ions.
In the previous experiment (Figure 3), the β-domain was metal-protected by the Cu(I)-thiolate cluster against CID fragmentation, and two b-type ions corresponding to two proteoforms were identified.We link these findings to our native MS data, revealing the existence of a proteoform with one disulfide and a second proteoform with two disulfides.Here, employing a double labeling approach, we observed the appearance of multiple Cys-IAM labeled b-fragment ions covering the β−domain.These findings clearly indicate that Cu(I) is bound to Cys residues in the β−domain of the Cu 4 NEM x MT3 ox complexes.The CID fragmentation map also indicates the existence of two proteoforms: one with a single disulfide and a second one with two disulfides, involving Cys5 and Cys7.

Application to Rabbit Liver MT2
Here, we sought to demonstrate the applicability of the proposed methodology to a mixture of metallothionein isoforms obtained from rabbit liver.The native mass spectra shows two metalloforms for each MT isoform: N-acetylated (N-Ac) MT2a, MT2b, and MT2c isoforms in their apo-form and with one Zn(II) ion bound (Figures 5A and S18A).Purified rabbit liver metallothionein has been found natively binding Zn(II) and Cd(II) ions.Therefore, we prepared these metalloforms in vitro by first adding 4 mol equiv of Cd(II) resulting in seven metalloforms (Figure 5B).This was followed by adding 3 mol equiv of Zn(II), which led to the formation of five metalloforms (Figures 5C and S18B,C).We observed multiple mixed Zn(II)/ Cd(II) stoichiometries for each isoform with a maximum of 7 metal ion bound to one protein (Figure 5C).Incubation of that sample with 50 mM NEM leads to N-Ac Cd 4 NEM 6−11 MT2 proteoforms (Figure 6A,B).
The N-Ac Cd 4 NEM 10 MT2 is the most dominant form, revealing that 10 Cys residues are weakly bound to three Zn(II).Considering MT2 has 20 Cys residues, the remaining 10 Cys residues are coordinating 4 Cd(II) ions.Performing a top-down MS shows that the β-domain is mostly modified by NEM, and Cd(II) ions are bound in the α-domain (Figure 6C).The data also show that the C-terminal Cys residues interact weakly with

■ CONCLUSIONS
The simultaneous mapping of different metal ions bound to cysteine residues presents a significant challenge.In this study, we developed a differential labeling strategy to map Cu(I)/ Zn(II) and Cd(II)/Zn(II) metal binding sites within proteins.To accomplish this, we utilized NEM and IAM as Cys labeling probes and Cu(I)/Zn(II) metallothionein-3 as a model protein.Therefore, the methodology presented is limited to map metalbinding Cys residues.Metallothionein-3 plays a crucial role regulating copper and zinc levels in the CNS. 4,42Previous studies have shown that MT3 can be isolated from mammalian brains in the form of an air-stable complex known as Cu(I) 4 Zn(II) 3−4 MT3 ox . 37,38−46 This reaction has been consistently observed to facilitate the binding and subsequent reduction of Cu(II) to Cu(I), leading to the dissociation of Zn(II) ions and the formation of two intramolecular disulfide bonds.
Through our investigations, we prepared a heterogeneous Cu(I) 4 Zn(II) 4 MT3 ox complex, incubated it with increasing concentrations of NEM and IAM, and monitored the reactions using native MS.Our findings demonstrate that NEM undergoes cooperative nucleophilic addition with Cys residues, leading to the dissociation of all Zn(II) ions.In contrast, due to the higher affinity of Cu(I) for Cys compared to Zn(II), Cu(I) remains stably bound to the protein.These results emphasize the remarkable capability of NEM to generate homogeneous Cu(I) x NEM y MT3 ox complexes and illustrate that, under carefully controlled conditions, NEM selectively dissociates all Zn(II) ions while preserving the binding of Cu(I).However, the use of a chemical reaction also introduces a level of complexity, as the alkylation process follows a statistical distribution.For instance, NEM labeling dissociates all four Zn(II) ions coordinated by 11 Cys residues.However, instead of observing a single peak with 11 Cys-NEM labeled residues, we find a statistical distribution ranging from 2 to 13 Cys-NEM modifications.
Determining the precise stoichiometries of alkylator and metal proteins based on only mass information can be challenging due to the presence of isobaric species. 52However, our study has demonstrated that through CID experiments, we overcame this challenge and determined the exact stoichiometry between Cu(I) ions and NEM moieties in MT3 ox complexes.The CID spectra obtained contained an m/z region where ions corresponding to the precursor and the precursor with the loss of certain metal ions were observed.This enabled us to directly obtain information regarding the number and type of metal ions bound to the system.In addition, we observed that the Cu(I) 4 NEM y MT3 ox complexes exhibited greater gas-phase stability than NEM y MT3 ox , providing further evidence for the stabilizing role of Cu(I) in the protein.
Subsequently, we employed both bottom-up and top-down MS to precisely identify the sites of Cys-NEM modifications in the Cu(I) 4 NEM y MT3 ox complexes, allowing us to deduce the specific locations where Zn(II) ions were originally bound within the Cu(I) 4 Zn(II) 4 MT3 ox complex.Our observations from the peptide-centric data yielded ambiguous results, as all Cys residues within the protein were found to be labeled with NEM modifications.It can be attributed to the presence of a portion of NEM 16−20 MT3 ox 5+ ions alongside the formation of homogeneous Cu(I) 4 NEM x MT3 ox 5+ complexes (x = 8−12), as NEM reacts with the protein.The implementation of a bottomup MS approach in this strategy becomes infeasible due to its larger dynamic range compared to native MS.
Alternatively, through native top-down MS, we successfully mapped all the residues labeled with Cys-NEM modifications.Our findings revealed that the α-domain of the protein was indeed labeled with NEM, while the β-domain demonstrated binding of four Cu(I) ions.However, it is important to note that we did not observe fragmentation within the protein region where Cu(I) is bound.
In order to get insights into the Cu(I) binding sites, we introduced a second labeling step in our strategy using IAM.Following the initial labeling with NEM, NEM-labeled Cu(I)-MT3 ox complexes were incubated with IAM.IAM reacted with the Cys residues involved in Cu(I) coordination, resulting in their alkylation and subsequent dissociation of Cu(I) from the protein.The native mass spectrum we recorded displayed numerous partially overlapping signals originating from multiple IAM stoichiometries for a specific NEM-MT3 ox complex.Similar to previous findings, the bottom-up MS results were inconclusive due to the sample heterogeneity, but the top-down MS approach yielded valuable information.Examining the CID fragmentation map of the most intense m/z window region (1730 m/z ions), we annotated multiple b-fragment ions labeled with varying extents of IAM modifications.These findings provide clear evidence that Cu(I) is bound to Cys residues, specifically within the β-domain of the Cu 4 NEM x MT3 complexes.Finally, we have applied the proposed methodology to infer the location of Cd(II) and Zn(II) ions in the rabbit liver metallothionein fraction, which contains multiple MT isoforms and post-translational modifications.
To conclude, the described approach holds promise for further exploration of the Cu(I)/Zn(II) and Cd(II)/Zn(II) metalloforms' diversity within other proteins and expand its spectrum to other metal ions.

* sı Supporting Information
The Supporting Information is available free of charge at https://pubs.acs.org/doi/

Figure 1 .
Figure 1.Native mass spectra of the Cu(I)/Zn(II)-MT3 ox complexes obtained after incubation of Zn 7 MT3 red with 4 CuCl 2 mol eq at increasing concentration of NEM.The m/z region corresponds to 5+ ions (full spectra in Figure S5).The proteins (10 μM) were sprayed in 200 mM ammonium acetate (pH 6.8)."ox" subscript refers to oxidized (two intramolecular disulfides) MT3.Peak series corresponding to Cu 4 NEM 0−13 MT3 ox are colored in orange, while peak series corresponding to Cu 5 NEM 0−12 MT3 ox are colored in violet.The asterisks indicate m/z signals that correspond to 4+ ions.

Figure 2 .
Scheme 1. Overview of the Presented Differential Labeling Strategy for Labeling Zn(II)-and Cu(I)-Cys Metal Binding Sites

Figure 4 .
Figure 4. Native mass spectra of NEM-and IAM-Cys labeled residues of the Cu(I)/Zn(II)-MT3 complexes obtained after incubation of Zn 7 MT3 with 4 CuCl 2 mol eq (A).Native top-down CID MS for quadrupole-selected 1730 m/z ions (B).CID fragmentation map for the 1730 m/z ions (C).

Figure 5 .
Figure 5. Native mass spectra of rabbit apo-MT (A), incubated with 4 Cd(II) mol eq (B), and after addition of 3 Zn(II) mol eq (C).The m/z region corresponds to 4+ ions.For the full spectra, see Figure S18.The metalloforms were assigned based on isotopic distribution fitting (see the inset in (A) and Figure S1).

Figure 6 .
Figure 6.Native mass spectra of rabbit apo-MT after addition of 4 Cd(II) and 3 Zn(II) mol eq. and incubated with 50 mM NEM (A).Zoom in the 4+ ions showing the multiple iso-and proteoforms that differ in the NEM moieties content.These species were isotopically resolved as shown in inset B (B).The N-Ac Cd 4 NEM 10 MT2a 4+ ions were quadrupole-selected and subjected to top-down CID MS.CID fragmentation map for the Cd 4 NEM 10 MT2a 4+ ions (C).