Are Native Conformations of Proteins Retained Throughout Laser Ablation Capture?

Abstract

Combining ion mobility-mass spectrometry (IM-MS) with laser ablation (LA) is an attractive approach for characterizing protein higher-order structure in native ambient MS (NAMS). Recently, we demonstrated that proteins ablated from native deposits are detected in their native higher-order structure. However, it remains unclear whether the proteins are ablated and captured in their native form or denatured by ablation and refolded after capture in the solvent. We used refolding properties of bovine carbonic anhydrase II (CA) and bovine serum albumin (BSA) to investigate the stabilities of these proteins during the LA and solvent capture. The denatured BSA refolded to a native structure when captured in a native solvent, yielding ESI mass spectra and collision-induced unfolding (CIU) curves comparable to those from native protein deposits. In contrast, denatured CA sample deposits dissolved in native solvent did not produce ESI mass spectra. However, matrix-assisted laser desorption ionization (MALDI) MS and bottom-up proteomics analysis of the same denatured deposit extracted samples confirmed the presence of CA, suggesting aggregate formation rather than refolding. The inability of denatured CA to refold upon redissolving in the ammonium acetate solvent was leveraged to investigate protein stability in LA NAMS. Observed charge-state distributions (CSDs) from native MS of intact and surface-deposited laser-ablated CA were comparable. However, minor variations in their CIU plots were observed. These findings, along with structural characterization of laser-ablated bovine hemoglobin, as well as weakly associated concanavalin A-mannose carbohydrate and CA-sulfanilamide complexes highlight the potential of LA as a promising surface sampling method for investigating protein structures.

Introduction

Many important biological processes for cellular functions rely on noncovalent interactions among macromolecules. For instance, vital cellular activities such as transcriptional regulation, replication, and DNA repair depend on interactions of DNA with proteins. Also, functions and structures of membrane proteins can be altered by noncovalent interactions, such as those with lipids. Therefore, it is imperative to characterize these interconnections to elucidate cellular functions at the molecular level.

In native mass spectrometry (NMS), noncovalent interactions of biomolecules are preserved; this method offers a valuable opportunity to gain structural insights into intact protein structures and their interactions in biological environments. In NMS, protein solutions are introduced into the mass spectrometer via electrospray ionization (ESI) using a pH-adjusted nondenaturing solvent (e.g., ammonium acetate (NH₄OAc)) to retain noncovalent interactions and protein foldings. Electrosprayed protein ions in NMS exhibit lower charge-states than those generated in conventional ESI from denaturing solvents; therefore, charge-state distributions (CSDs) are frequently used as a probe to confirm the preservation of native structure. However, relying solely on charge-state is insufficient to confidently confirm that the electrosprayed ions accurately represent the native solution configuration. Consequently, it is desirable to utilize other complementary analytical techniques, such as ion mobility spectrometry (IMS), to characterize physical parameters, including ion mobility (IM) drift time and rotationally averaged collisional cross-section (CCS), that can provide additional structural details. In addition, collision-induced unfolding (CIU) experiments can be performed to assess gas-phase unfolding characteristics of the ions upon collisional activation.

There is a growing interest in combining IMS with surface analysis for acquiring spatially resolved information on protein higher-order structure and for imaging in disease diagnosis. Conventional surface sampling techniques for probing intact proteins are either solvent- or laser-based. For example, desorption electrospray ionization (DESI) − and liquid extraction surface analysis (LESA) are among the reported solvent extraction-based approaches for analysis of proteins and protein assemblies.

In addition to solvent extraction methods, laser desorption and laser ablation (LA) provide alternative options for surface sampling, offering numerous advantages for investigating structural modifications in biological tissue. For instance, because of its “soft” ionization characteristics, MALDI is extensively employed for spatially resolved characterization of intact molecules and tissue imaging. Although retention of noncovalent interactions has been achieved using MALDI, it is not the most prevalent method for NAMS. The use of MALDI to probe noncovalent interactions presents critical difficulties as matrix solution may disrupt the innate noncovalent interactions, cause acid-induced protein denaturing, and promote analyte migration. In contrast, LA and laser capture microdissection (LCM) are surface sampling methods that do not require a solvent for localized substrate sampling.

Laser ablation sampling utilizes a focused pulsed laser beam, commonly in infrared (IR) or ultraviolet (UV) wavelength ranges, to remove material from the surface for off-line analysis. Laser ablation has been widely used in analytical studies, enabling two-dimensional (2D) and three-dimensional (3D) imaging MS with high resolution across the tissue surface and depth, , as well as in situ MS analysis of metabolites in single cells. IR laser has been used for ablation of labile macromolecules such as DNA, RNA, and an enzyme with conserved activity and without fragmentation. Recently, we reported that proteins deposited on a surface in their native configuration retain their native structure after being captured in and electrosprayed from a native solution. Comparable results from laser-ablated and directly injected proteins supported the idea that LA did not change protein conformation. However, the generality of this assertion and the possibility of denatured ablated proteins reverting to their folded forms in the capture solvent were not addressed. Thus, it is necessary to verify that the observed higher-order structures of laser-ablated proteins truly represent the sampled proteins and not the denatured proteins that have refolded into native-like structures.

The propensity of a denatured protein to refold to its native structure varies among proteins. Refolding and aggregate formation of proteins are often divergent pathways, sharing “molten-globule” intermediate states that can be influenced by physicochemical conditions, chemical agents, and protein structures. It has been shown that unfolded proteins such as bovine serum albumin (BSA) and cytochrome c can refold to their native forms. However, bovine carbonic anhydrase (CA), a zinc-bound protein, has unique refolding characteristics and has been used to investigate protein intermediate refolding states. , For instance, depending on the physicochemical conditions, CA can either refold or form aggregates. Employing ESI MS, Nabuchi et al. exploited the formation of apo- and holo-CA in aqueous solutions to gain insight into CA refolding kinetics. It was reported that denatured CA, redissolved in ammonium acetate solvent, did not revert to its original conformation; instead, after losing its zinc cofactor, it aggregated. This distinctive aggregation, rather than refolding, characteristic of CA can be used to check whether it unfolds during the LA process.

Here, we employ NAMS to explore unfolding and refolding of BSA and CA, two proteins with varied propensities for refolding and aggregation, , and examine their conformational stabilities after LA. Moreover, we probe the structural integrity of laser-ablated bovine hemoglobin (BHb), as a model protein assembly, and two well-characterized noncovalent complexes, including concanavalin A (ConA)-mannose carbohydrate and CA-sulfanilamide, through analysis of CSDs and CIU patterns.

Experimental Section

Reagents

BHb, BSA, CA (C2624), jack bean concanavalin A (ConA), 3α,6α-mannopentaose, sulfanilamide (SLFA), and 7.5 M NH₄OAc solution were purchased from Sigma-Aldrich (St. Louis, MO, USA); unfolding and refolding characteristics of BSA and CA have been studied previously. , BSA (MW ∼ 66.4 kDa), a globular single-chain protein, is known for its ability to undergo refolding after unfolding and has been used as a model in protein unfolding and aggregation studies. CA (MW ∼ 29.1 kDa) can form aggregates; unique properties of CA have been exploited to use it as a model protein for ligand-protein studies. Sinapic acid (SA) and trifluoroacetic acid (TFA) were purchased from Sigma-Aldrich and used for MALDI analysis. All other solvents and reagents for LC/MS and bottom-up proteomics studies were purchased from commercial sources; a detailed list of chemical reagents and a single-pot solid-phase-enhanced sample preparation (SP3) tryptic digest procedures for the bottom-up studies are available in Supporting Information and Methods. Amino acid sequence and a list of tryptic digest peptides of CA are summarized in Table S1.

Sample Preparation

For conventional ESI experiments, each native protein stock solution was prepared by dissolving the lyophilized protein powder in the native solvent, containing aqueous 200 mM ammonium acetate. The pH values for 200 mM ammonium acetate solutions were measured separately before and after the addition of each protein (to yield 750 μM solutions), which were all within 6.81 to 6.88, requiring no additional pH adjustment. For direct injections, each protein stock solution was diluted to 25 μM, and 4 μL was injected for IM-MS analysis, providing over an hour of stable electrospray for multiple CIU runs. Proteins were acid-denatured in a mixture solution of water, acetonitrile, and formic acid (49.5%: 49.5%: 1% v/v/v).

For preparation of CA-SLFA complex native solution, a stock solution of SLFA (25 mg mL^–1) was prepared in dimethyl sulfoxide (DMSO, Millipore Sigma, Burlington, MA) and subsequently diluted with water to yield a 3 mM solution. The complex was prepared by mixing 100 μL of 750 μM CA with 25 μL of the 3 mM SLFA solution (molar ratio of 1:1), followed by incubation at room temperature for 48 h.

For preparation of the ConA-mannose carbohydrate complex native solution, 500 μM stock solutions of mannose and ConA were prepared separately in ultrapure water and 200 mM ammonium acetate, respectively. Equal volumes (50 μL) of each stock solution were mixed at a molar ratio of 1:1 and incubated at room temperature for 2 h.

For the LA experiments, either 4 μL (for BSA and CA), 2 μL (for BHb), or 12 μL (for noncovalent complexes) of the native protein stock solution was spotted on a borosilicate glass microscope slide and air-dried at room temperature. Details of the IR LA system and the ablation method have been described previously. , For the presented work, the laser was focused to a 6 μm beam diameter and operated at a fluence of approximately 50 kJ/m². The ablated material was collected in a 200-μL microcentrifuge tube that contained 20 μL of capture solvent (95% 200 mM NH₄OAc:5% ACN; pH ∼ 6.85); to sustain 60 min of stable ESI spray, 4 μL of the capture solution was used. Extracted (control) samples were prepared by depositing 4 μL (for BSA and CA), 2 μL (for BHb), or 12 μL (for noncovalent complexes) of the native stock solutions on the slide and allowing them to air-dry. Each dried spot was washed by applying 10 μL of capture solvent and immediately withdrawing it with a pipet. The washing was repeated for each dried spot, using an additional 10 μL of solvent to yield a combined volume of 20 μL extracted protein solution, collected in a 200-μL centrifuge tube. Triplicate capture tubes, containing the extracted or laser-ablated proteins, were vortexed for 30 s and shipped on dry ice to Baylor University (BU) for IM and MS analyses.

For the protein refolding experiments, separate denaturing protein solutions were made to yield 200 μM of either BSA or CA. For each protein, 5 μL of 200 μM stock denatured solution was spotted on a glass slide, air-dried, and washed from the slide by adding 35 μL of capture solvent, which was allowed to stand for 30 s, then withdrawn and delivered into an Eppendorf tube for MS analysis. For protein refolding studies, the denatured deposit extracted CA and BSA samples were analyzed after varying incubation time in the native capturing solvent. To assess the sensitivity of CIU patterns for detecting potential thermal-induced structural changes, protein dry deposits were examined before and after heating. For these thermal stress experiments, 5 μL of a 100 μM BHb native solution was deposited onto a glass slide and air-dried at room temperature. The dried deposits were either directly extracted into 20 μL of native solvent or first incubated in a preheated oven at 45, 63, 140, 150, 153, 155, 160, 170, and 200 °C for 20 min, followed by extracting into 20 μL of native solvent. After identifying the maximum temperature at which tetrameric and dimeric species BHb still remained intact, the effect of incubation time duration at this threshold temperature was further examined in the range of 20–60 min.

MALDI mass spectra were recorded using a time-of-flight mass spectrometer (Rapiflex, Bruker, Bremen, Germany). Mass spectra were obtained in linear mode as the sum of 100 laser shots. A stock solution of CA at 200 μM was prepared in the denaturing solvent. Two microliters of the CA solution were deposited on a microscope slide and allowed to air-dry at room temperature. The dried deposit was brought into solution with 2 μL of 200 mM NH₄OAc deposited on the surface for 10 s and withdrawn. The resuspended protein was mixed in a 1:1 (v/v) ratio with 2 μL of a SA matrix solution at a concentration of 10 mg/mL in a 3:7 ratio of ACN and 0.1% aqueous TFA (v/v) in water. A 1 μL volume of the matrix-analyte mixture was deposited on a stainless-steel target and air-dried for MALDI analysis.

ESI Mass Spectrometry

Details of borosilicate glass nanoelectrospray emitter production and characterization are described in Supporting Information and Methods. MS and CIU data were recorded in sensitivity positive-ion mode on a Synapt G2-S traveling wave ion mobility spectrometry quadrupole time-of-flight (TWIMS-q-ToF) IM-MS system (Waters). MassLynx software V4.1 (Waters) was used for data acquisition and controlling the instrumental parameters. All native mass spectra were acquired at respective ESI capillary and pre-IM trap collision voltages of 600 and 10 V. Tandem mass spectra were acquired at a pre-IM mobility collision voltage of 20 V; all other IM-MS instrumental parameters are summarized in Table S2.

The high-resolution ESI mass spectra were acquired with an Orbitrap Q-Exactive Focus mass spectrometer (Thermo Scientific, Waltham, MA) under denaturing conditions using 10 μM CA. All instrumental parameters for the Orbitrap are summarized in Table S3.

Bottom-up mass spectra were acquired using a Synapt G2-S HDMS (Waters) in positive-ion resolution mode (Table S2). For MS^E proteomic analysis, the tryptic peptides were eluted with a binary solvent gradient for 45 min. Additional details on LC separation settings are provided in Table S4.

Data Analysis

The raw data from MassLynx were exported to OriginLab V2023 (Northampton, MA) for further processing. For all reported ESI mass spectra, acquired data from triplicate runs were averaged and smoothed via 10-point adjacent-averaging windows. The open-source software CIUSuite 2 was used to generate CIU and root-mean-squared deviation (RMSD) plots from the acquired IMS data for comparisons. Two-dimensional (2D)-Savitsky-Golay function, with a window size of five and two iterations, , was used for IM data smoothing. All displayed CIU plots were averages of their respective triplicate runs. Standard feature detection was carried out using a minimum feature length of 2 steps, an allowed feature drift time width of 0.75, and a maximum collision voltage (CV) gap length of zero. CIU50 values were then assigned using max spectral centroiding mode with a transition region padding of 15 CV and a maximum CV gap length of 0.

Results and Discussion

ESI and MALDI mass spectrometry, along with bottom-up proteomics and IM experiments, were carried out to characterize higher-order structures of BSA, CA, BHb, and weakly associated complexes of ConA-mannose and CA-SLFA before and after LA. The CSDs from ESI mass spectra and drift times from IM spectra were utilized to probe protein structures from native, denatured, and laser-ablated samples.

Refolding and Laser Ablation of Serum Albumin

Generally, the CSDs for proteins in native and denatured forms are significantly different, and these patterns are often utilized to confirm native spray conditions. For instance, the positive-ion mode ESI mass spectrum and IM drift time vs m/z drift-scope plot of BSA under native conditions yield predominant charge states of +13 to +16 (Figures S1 and S2); conversely, under denaturing conditions, the most prominent charge states range from +21 to +62 (Figure S1). Thus, the presence of lower charge state BSA (i.e., charge states +13 to +16) can be utilized to assess its unfolding and the occurrence of natively structured species. ,, Previously, we showed that there were no detectable differences in the ESI mass spectra of BSA collected prior to and following laser ablation. However, there is the possibility that unfolded laser-ablated proteins refold to native-like structures after capture in ammonium acetate, which can complicate data interpretation. To assess whether an unfolded or denatured BSA can refold to its native-like structure, mass spectrometry and ion mobility data were acquired and examined under three different sample preparation conditions. Specifically, native or denatured BSA (in ammonium acetate or denaturing solvent, respectively) was deposited on a microscope slide, air-dried, and either extracted or laser-ablated into ammonium acetate solvent in separate Eppendorf vials.

Figure shows representative positive-ion mode direct infusion ESI mass spectra (m/z range 2000–6000) for BSA from the (A) native deposit extracted, (B) native deposit laser-ablated, and (C) denatured deposit extracted samples. Similar to Figure S1A, all three ESI mass spectra in Figure A–C show BSA low charge-states +13 to +16; the calculated average charge-state from triplicate runs at the 95% confidence level (C.L.) for BSA from native deposit extracted, native deposit laser-ablated, and denatured deposit extracted samples were 14.9 ± 1.0, 14.7 ± 0.9, and 14.8 ± 0.6, respectively. At the 95% C.L., these average charge-states are statistically indistinguishable, and charge-state +15 is the most abundant ion in all three native mass spectra in Figure A–C. Thus, all three mass spectra indicate the presence of native-like structures.

1.

Positive-ion mode ESI mass spectra of BSA from (A) native deposit extracted, (B) native deposit laser-ablated, and (C) denatured deposit extracted samples.

Although high-charge states of BSA (>+20, m/z range ∼2000–3324) are present as low-abundance species (<5%) in all three mass spectra shown in Figure , the majority of the observed ions correspond to lower charge-states associated with folded conformations (see Figure S1 and associated text description). The presence of low charge-state ions (+13 to +16) in Figure C suggests that the unfolded BSA refolds in the native capture solvent, yielding a mass spectrum that resembles the native deposit mass spectrum in Figure A. BSA is known to exhibit complex refolding kinetics. Using fluorescence spectroscopy, Damodaran monitored the formation of hydrophobic regions to gain insight into BSA refolding kinetics. These findings revealed that BSA refolding in Tris buffer begins with slow subdomain folding, followed by a faster phase involving intersubdomain interactions, and a final slow phase to form a compact 3D structure.

Because protein folding is a solvent-dependent process, we also investigated BSA refolding in ammonium acetate, acquiring triplicate CIU plots 5 min, 4, 24, and 144 h after redissolving the denatured form of BSA in ammonium acetate capture solvent. These experiments indicated that the majority of refolding of the unfolded BSA occurred within the first 5 min (Figure S3), which could complicate mass spectrometry data interpretation: unfolded BSA is rapidly converted to folded BSA in a native solvent. It is worth noting that laser ablation did not perturb BSA refolding. For instance, mass spectra and CIU plots of denatured deposit extracted and denatured deposit laser-ablated BSA were comparable (Figures S4 and S5). Moreover, consistent with previous reports, , protein fragment ions were not observed in Figure A–C mass spectra. Comparison between the drift-scope plots for the solvent blank, direct native, laser-ablated, and CID fragments of quadrupole isolated electrosprayed BSA (charge state +15 at m/z 4450) confirmed that the peaks below m/z ∼1000 were not protein fragments (Figures S6 and S7). These findings indicate that LA does not generate any detectable fragment ions from either BSA native or denatured samples.

While the similar mass spectra for the native deposit extracted (Figure A) and laser-ablated BSA captured in the native solvent (Figure B) indicate native structures, the spectrum in Figure C prevents a definitive conclusion regarding laser ablation unfolding. For instance, analogous to the BSA refolding observed in Figure C, a protein that unfolds during laser ablation could subsequently refold in the native capture solvent. Such a refolding event prevents a clear distinction between folded and unfolded BSA using either mass spectra or ion mobility data. For instance, the t-test results on the averaged (n = 3) IM-MS data suggest that drift times for the IM features corresponding to charge state +14–16 of BSA for denatured deposit extracted and native deposit laser-ablated samples are not significantly different (Table S5). Protein folding and unfolding have been detected using rapid mixing in theta-glass emitters, providing additional examples to suggest protein refolding might occur during the electrospray. Also, Karki et al. have reported that some portion of the protein population folds upon interaction of a laser vaporized acid-denatured protein with electrospray droplets containing ammonium acetate. Therefore, it is crucial to verify that the protein conformation after LA and capture in a native solvent accurately represents surface molecules.

To further characterize configurations of the native deposit extracted and denatured deposit extracted BSA, their CIU plots were acquired. Figure shows the average (n = 3) CIU plots for BSA charge states +14 to +16 from the native deposit extracted (A−C) and denatured deposit extracted (D−F) samples for identical m/z extraction windows (Table S6); Figure G–I contain corresponding RMSD comparison plots. Computed values for all replicate CIU data sets for all charge states of BSA electrosprayed from control and denatured deposit extracted samples yielded RMSD values <7.5%, indicating reproducibility of CIU plots for both samples. The comparison RMSD values of 16.08, 11.35, and 12.94% for charge-states +16, +15, and +14 of BSA, respectively, suggests potential minor structural differences between the native deposit extracted and the denatured deposit extracted BSA, presumably due to partial refolding of BSA. Additionally, the t-test results indicated that the calculated average drift times from triplicate runs at the 95% C.L. for almost all observed “features” were statistically different (Table S7). However, small variations in the capture conditions (e.g., minor solvent composition differences, room temperature variations, protein concentration, and incubation time in capturing solvent as the protein refolds) may lower the comparison RMSD values, thereby making it difficult to differentiate between the native deposit extracted and denatured deposit extracted BSA samples. Therefore, the inherent propensity of denatured BSA to refold in the capturing solvent makes it unsuitable for probing potential laser-induced structural changes, such as protein unfolding.

2.

Averaged CIU curves for charge states +16 to +14 of BSA directly sprayed from its native deposit extracted (A–C) and its denatured deposit extracted (D–F) samples along with their corresponding RMSD plots (G−I).

Thus, although concluding that LA does not yield fragment ions is straightforward, , based on the BSA data alone, it is impossible to confidently assert that LA does not alter protein higher-rider structures. Therefore, to properly examine protein unfolding upon LA, it is necessary to utilize a protein whose denatured form cannot revert to its native conformation.

Aggregation of Acid-Induced Denatured Carbonic Anhydrase: MALDI and Bottom-up Characterization

Once unfolded, carbonic anhydrase might aggregate rather than refold. , Figure shows the positive-ion mode ESI mass spectra (m/z region 500–4000) for direct infusion of CA from native solvent (Figure A), from denaturing solvent (Figure B), and after depositing its denatured form onto a microscope slide, drying, and redissolving in the capture solvent (Figure C). Consistent with the previous reports, , the ESI mass spectrum of CA from a native solution in Figure A yielded charge-states +8 to +10. The slightly lower average charge-state in Figure A (+9 most abundant) compared to previous reports for CA (+10 most abundant) is likely due to the slightly more acidic ammonium acetate solution employed in the earlier studies (viz., pH 6.8) or increased acidity within the electrospray droplets. Under the native MS conditions (Figure A), using Gaussian-fit peak centroids for mass-resolved charge-states +9 and +10 from four different runs (Figure S8), yielded an experimental average MW of 29090.7 ± 3.1 Da, (at the 99% C.L.), which encompasses the theoretical average MW for the acetylated holo-CA (29089.6529 Da). The N-terminus acetylation of CA has been previously reported, , and is supported by top-down MS analysis of the same protein with an identical vendor product number, as well as our Orbitrap data (Figure S9).

3.

ESI mass spectra for CA from direct infusion of (A) native solution (25 μM) (B) denatured form (25 μM), and (C) denatured deposit extracted.

Consistent with a prior study, under denaturing conditions, CA yielded higher charge-states from +41 to +18 (Figure B). Figure B inset shows the expanded region for the most abundant charge-states +36 to +30 of CA in the m/z range 800 to 1000. CA is a metalloenzyme and can change from holo- to apo- form when zinc is released. The transition from holo- to apo- form of CA can also occur under low pH, which is encountered in conventional electrospray.

Under denaturing conditions, the experimentally calculated average MW for the presumed acetylated apo-CA, obtained from a single run using Gaussian-fit peak centroids across all high charge-states ranging from +17 to +40 (total of 24 peaks) (Figure B), yielded 29024.5 ± 1.2 Da. This value is consistent with the theoretical average MW for the acetylated apo-CA (29024.2751 Da) and reported values in the literature. Corroborating previous findings, the mass difference between the average MW of the acetylated holo- and apo- forms of CA in Figure is consistent with the loss of Zn²⁺ (Figure S9). However, neither apo- nor holo-forms of CA were detected in the ESI mass spectrum of denatured deposit extracted samples (Figure C).

Refolding of CA is a three-stage process. The “double-jump” experiments reported by Semisotnov et al. showed that the initial stage involves a rapid compaction of the protein chain with half-time (t _1/2) of ∼ 0.03 s at 25 °C. The second stage (t _1/2 ∼ 140 s) involves desolvation of hydrophobic clusters and formation of a rigid native-like hydrophobic core, and the final stage (t _1/2 ∼ 600 s) culminates in the formation of the native protein structure. To investigate the time dependence of CA refolding in ammonium acetate, we acquired ESI mass spectra of CA 5 min, 12 min, and a week after extracting its denatured deposit form into ammonium acetate solvent. However, consistent with a previous study on CA refolding (where the denatured protein spot was sampled under native conditions for 60 s), the ESI mass spectrum of the denatured deposit extracted CA did not show any intact protein ion (similar to data in Figure C). The absence of CA peaks suggests that, under our experimental conditions, acid-induced unfolded CA does not refold in native solvent, likely forming large aggregates beyond the instrument mass range. This observation is consistent with a previous report on refolding and aggregation of CA during liquid extraction, as well as a capillary electrophoresis study on the influence of Zn²⁺ on CA refolding and aggregation after treatment with sodium dodecyl sulfate. Gudiksen et al. suggested that the lack of Zn²⁺ in the hydrophobic core of apo-CA may cause this region of the protein to become less ordered, resulting in the aggregation of a partially folded intermediate. Favoring aggregate formation over reversion to native conformation in the native solvent makes CA suitable for determining protein stabilities after LA: if laser ablation unfolds the protein, it cannot refold in the native solvent.

To further investigate whether the aggregate formation was the reason for not detecting CA ions in ESI mass spectrum of denatured deposit extracted sample (Figure C), MALDI and bottom-up proteomics experiments were carried out. Both MALDI MS (Figure S10) and bottom-up proteomics (Table S1) confirmed the presence of CA in denatured deposit extracted samples. These results suggest that once the denatured CA is redissolved in the native solution, it forms aggregates.

Laser Ablation of Carbonic Anhydrase

The inability of denatured CA to revert to its native structure makes it a suitable probe to examine potential laser-induced protein unfolding; thus, the higher-order structural stability of CA after laser ablation was examined. The positive-ion mode ESI mass spectra from native deposit extracted and native deposit laser-ablated CA were comparable (Figure ), suggesting no laser-induced unfolding.

4.

ESI mass spectra for direct infusion of CA from its (A) native deposit extracted and (B) native deposit laser-ablated samples.

Although minor variations were observed in the CIU plots, the IM data largely agreed with the MS findings regarding post laser ablation structural stability of CA, while also revealing subtle differences.

The CIU plots in Figure , corresponding to native deposit extracted and native deposit laser-ablated of CA, yielded replicate RMSD values <6% across all charge states, indicating the reproducibility of CIU data. Comparison of RMSD values (Figure G–I) between the CIU plots for native deposit extracted and native deposit laser-ablated CA suggests minor differences between control and laser-ablated samples. For instance, the t-test results at the 95% C.L. indicate that the average (n = 3) CIU50 values for transition one (in charge state +10) of control and laser-ablated CA are statistically different (Table S8). Also, average (n = 3) IM drift times at the 95% C.L. for the ion mobility feature one of charge state +10 and feature two of charge state +9 CA were slightly higher for the laser-ablated sample (Table S9). It should be noted that protein concentrations in control and laser-ablated samples can be slightly different, which can increase the RMSD variations (Figures S11 and S12). Moreover, the stability of signals and small variations in the position of the emitter from the instrument inlet may influence the reproducibility of CIU plots. Nevertheless, general CIU trends and specific energies (or voltages) at which protein unfolding transitions occur (Figure ) verify the similarities between control and laser-ablated samples. All control and laser-ablated samples reported here were analyzed under identical instrumental and acquisition conditions. However, despite these efforts, we acknowledge that intrinsic method variability associated with the IM-MS experiments cannot be eliminated.

5.

Averaged CIU plots for charge states +10 to +8 of CA from direct infusion of native deposit extracted and native deposit laser-ablated solutions (top and middle rows, respectively) along with RMSD plots.

Although BSA results were inconclusive, IM-MS data for CA revealed that protein structures remained intact after laser ablation. In addition to CA, we evaluated LA for surface sampling of protein assemblies using BHb, a model protein previously studied in DESI and hydrogen/deuterium exchange MS (HDX-MS) experiments.

Laser Ablation of Hemoglobin (Protein Assembly)

BHb is a protein assembly consisting of four subunits (two α-chains and two β-chains), each bound to a heme group. These subunits interact noncovalently to form a tetrameric structure. Figure shows representative positive-ion mode direct infusion ESI mass spectra (m/z region 1500–6000) for BHb from the (A) native deposit extracted and (B) native deposit laser-ablated samples. The native ESI mass spectra of control and native deposit laser-ablated BHb show similar patterns dominated by the charge state +11 of dimeric heme-bound (α^hβ^h) ions (at m/z ∼ 2931.9), accompanied by other lower intensity ions such as charge state +7 (at m/z ∼ 2239.4), and +8 (at m/z ∼ 1959.6) of holo-α monomer (α^h). As shown in the Figure insets (m/z range 3700–4500), and consistent with previous native MS reports, intact tetrameric (α^hβ^h)₂ species of BHb (charge states +17–15) are also observed in both control and laser ablation native mass spectra. The detection of dimers and tetramers of BHb before and after laser ablation indicates that laser interaction does not disrupt the structure of BHb and allows the preservation of stoichiometry. Moreover, the CIU plots of BHb tetrameric species, corresponding to the control native deposit extracted (Figure A–C) and native deposit laser-ablated (Figure D–F) samples, exhibit indistinguishable gas-phase unfolding patterns for all three +17, +16, and +15 charge-states of (α^hβ^h)₂, with respective RMSD values of 5.47, 5.47, and 5.34, suggesting stability of laser-ablated (BHb)₄. Moreover, the t-test results indicated that the average (n = 3) CIU50 values for nearly all transitions of (BHb)₄ charge state +16 and +15 were statistically indistinguishable for native deposit extracted and native deposit laser-ablated samples at the 95% C.L. (Table S8). These CIU comparisons, characterized by low RMSD values and closely aligned CIU50 values, further strengthen the conclusion that no unfolding of this protein assembly is detected within the sensitivity of the CIU analysis. In other words, partial unfolding, local structural perturbations, or transient unfolding events might be challenging to fully characterize by CIU MS experiments alone.

6.

ESI mass spectra of BHb from (A) native deposit extracted and (B) native deposit laser-ablated samples. The notation α^h, D, and T represent the monomeric holo-α subunit, dimeric α^hβ^h, and tetrameric (α^hβ^h)₂ ions, respectively.

7.

Average CIU curves for charge state +17 to +15 of BHb tetramer (BHb)₄ directly electrosprayed from its native deposit extracted (A–C) and its native deposit laser-ablated (D–F) samples along with their corresponding RMSD plots (G–I).

To determine whether protein conformational changes would be detectable by CIU under conditions comparable to laser ablation, which involves localized heating, we performed controlled thermal stress experiments as a complementary approach. Although the charge state distributions of BHb electrosprayed from the extracted sample dried at 153 °C (a temperature that was determined to be a threshold for disappearance of BHb tetramers) was comparable to those dried at room temperature (Figure S13); CIU comparison suggests that there are minor structural changes caused by heat stress (Figures S14–S15). These observations suggest that CIU analysis is sensitive to detect heat-induced conformational changes and, therefore, would be expected to detect potential laser-induced changes.

Laser Ablation of Weakly Associated Complexes

Protein–ligand interactions are key to maintaining protein structure and biological function in cells. Thus, in addition to globular proteins, we also examined the effect of laser ablation on conformational stabilities of ConA-mannose carbohydrate and CA-SLFA as model noncovalent complexes.

Concanavalin A-Mannose Carbohydrate Complex

Concanavalin A (Con A) is a homotetramer protein (∼102.4 kDa), and its carbohydrate-binding characteristics have been extensively studied. For instance, ConA-3α,6α-mannopentaose complex has a reported low association constant (K _a) of (6.6 ± 0.4) × 10⁵ M^–1, rendering it a suitable candidate for our study. Detection of mannose-bound ConA dimers before and after laser ablation (Figure S16) suggests that laser interaction does not disrupt the noncovalent interactions of carbohydrate and ConA. Consistent with a previous report on ConA-mannose complex, addition of the carbohydrate to ConA enhances the stability of the assembly, as evidenced by higher CIU50–1 values for charge state +13 of ConA dimer (Figure S17). Moreover, the CIU plots of mannose-bound and unbound forms of ConA dimeric and tetrameric species before and after laser ablation exhibit similar gas-phase unfolding patterns across all charge-states (Figures S18–S19); however, RMSD comparisons (>23%) suggest minor variations. One plausible explanation for high RMSD values is that analyte molecules, acting as a sacrificial matrix, are incorporated into the native capturing solvent during the laser ablation, which can slightly alter solvent composition. In addition to the ConA-carbohydrate complex, we evaluated the preservation of protein–ligand complexes using CA-SLFA, a model complex previously studied in LESA and IR-MALDESI experiments.

CA-SLFA Complex

Sulfanilamide (SLFA) is a weak binding inhibitor of carbonic anhydrase (dissociation constant, K _d ∼ 3 μM), , making it a suitable probe for evaluating the preservation of noncovalent protein–ligand interactions during the laser ablation and capture process. Detection of CA-bound SLFA species before and after laser ablation (Figure S20) suggests that this enzyme–inhibitor interaction is retained after laser ablation. However, some weakly bound nonspecific SLFA adducts are disrupted by laser ablation, which are also lost at low collision voltages (25 V) in the control sample (Figures S20C and S21). The CIU plots of unbound and SLFA-bound forms of CA before and after laser ablation exhibit similar gas-phase unfolding patterns across all charge-states with low comparison RMSD values (Figure S22), suggesting the survival of the CA-SLFA complex after LA.

Laser Ablation Mechanisms

A key question in the LA of biological samples is how labile proteins remain intact despite the high-energy absorption of the sample. Previous studies have established that proteins can be sampled by laser ablation without fragmentation. ,, The native MS data presented here demonstrate that the higher-order structures of laser-ablated proteins are also preserved. Various mechanisms have been proposed to explain energy deposition and material ejection in laser ablation. In general, transfer of proteins from the condensed phase by ablation for capture in a buffer can occur through three possible mechanisms, namely, the ejection of either free protein molecules, solvated proteins, or small clusters and particles. The ejection of free proteins is unlikely because full desolvation of the protein in the gas phase will lead to loss of native conformation. It is unlikely that desolvated CA or BHb will retain their native structure after entering the electrospray droplet. Furthermore, the native mass spectra of ablation captured proteins have broad peaks, suggesting formation or retention of adducts and solvent clusters. Although no size measurement was performed in this study, previous particle size analyses of surface ablated materials under the conditions employed in this study have shown that a large fraction of the material ejected is in the form of small particulate with a bimodal distribution comprising small clusters tens of nanometers in diameter and coarse particulate ranging up to tens of μm in diameter.

The transport of biomolecules by LA of particles has been suggested as a speculative mechanism for their removal from surfaces and transport through an ambient atmosphere without fragmentation, and should be explored in future studies.

An additional issue is the nature of the energy absorber. The proteins are deposited as a solid sample on a glass surface and most likely exist as solid hydrated proteins. The likely absorbers of the laser energy are residual protein waters of hydration or the OH and NH groups of the protein itself. It has been shown that pure proteins deposited on a surface can be ablated and ionized directly or by electrospray postionization. Photochemical decomposition of some of the proteins on the surface to form gaseous products may be a factor driving the ablation process. Alternatively, energy absorbed by residual water in the protein deposit may facilitate the ablation. Regardless of the mechanisms involved, the NMS data presented here demonstrate that LA leads to the detection of intact native molecules.

Conclusions

To probe higher-order structures of laser-ablated proteins, we used BSA, CA, BHb as well as complexes of ConA-mannose carbohydrate and CA-SLFA, leveraging their divergent propensity for aggregation and refolding. This study allowed us to demonstrate that proteins in their native surface-bound conformations can be laser-ablated with no fragmentation or significant unfolding detected, within the sensitivity of the methods used. However, following laser ablation of CA-SLFA complex, some weak interactions of nonspecific SLFA adducts were disrupted. Observed laser ablation results are consistent with a mechanism whereby residual water in the dried deposit or some of the protein molecules absorb the laser energy, undergo a rapid phase change or dissociation, forming gas molecules that drive the ablation process. The ejected material is likely in the form of particles, clusters, and solvated protein molecules rather than desolvated free protein. It is also possible that a small fraction of surface proteins serves as a sacrificial matrix to enable expulsion of intact proteins. Regardless of the exact mechanism(s) involved in laser ablation experiments, the results presented here suggest that intact proteins can be observed in laser ablation experiments. Although a significant portion of the laser-ablated BSA, CA, BHb, ConA-mannose carbohydrate complex and CA-SLFA complex remained intact, structural survivability is a protein-dependent phenomenon and should be further investigated.

The ability to ablate and capture native proteins from solid samples presents a powerful surface analysis tool for characterization of intact proteins and their noncovalently formed complexes by native MS. Future studies will explore the capabilities of the LA NAMS method and extend its application to tissue using IR and other laser wavelengths for improved tissue sampling and spatially resolved characterization of weakly associated biomolecules.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.analchem.5c07180.

• Supporting Information and Methods; ESI mass spectra for native and denatured BSA (S1); IM drift time vs m/z drift-scope plots for solvent blank and native BSA (S2); averaged CIU plots for charge states +16 to +14 of BSA after 5 min, 4, 24, and 144 h of redissolving the denatured form of BSA in capturing solvent (S3); ESI mass spectra for denatured deposit extracted and denatured deposit laser-ablated BSA (S4); average CIU plots for BSA from denatured deposit extracted and denatured deposit laser-ablated BSA (S5); three-dimensional IM drift time vs m/z of solvent blank and BSA samples (S6); ESI mass spectrum of solvent blank for m/z < 1000, MS isolation of solvent cluster peaks, its MS/MS spectrum, and drift-scope plots of the corresponding MS/MS spectra (S7); Gaussian fit to CA mass spectrum (S8); high-resolution mass spectrum of denatured CA (S9); the MALDI mass spectra for CA (S10); dependency of replicate RMSD values on BSA concentration (S11); RMSD comparison between the average CIU plots corresponding to 12 μM BSA and its lower concentrations (S12); ESI mass spectra of BHb from native deposit extracted samples collected at (A) room temperature and (B) 153 °C (S13); averaged CIU plots for charge states +12–10 of BHb dimer from native deposit extracted samples at room temperature and 153 °C (S14); Averaged CIU plots for charge states +17–15 of BHb tetramer from native deposit extracted samples at room temperature and 153 °C (S15); ESI mass spectra of ConA-mannose complex from (A) native deposit extracted and (B) laser-ablated samples (S16); Averaged CIU plots for charge states +13 of ConA dimer bound to mannose in two stoichiometries (S17); averaged CIU plots for charge states +15–13 of ConA dimer incubated with carbohydrate (S18); Averaged CIU plots for charge states +19 and +18 of ConA tetramer incubated with mannose (S19); ESI mass spectra of CA-SALF complex from native deposit (A) extracted, (B) laser-ablated, and extracted at pre-IM voltage of 4, 4, and 25 V (S20); Normalized peak area for CA-SLFA complexes vs collision voltages (S21); averaged CIU plots for charge states +11–9 of CA-SLFA complex (S22); Sequence coverage and peptide list of CA (Table S1); istrumental parameters for MS experiments (Table S2); Instrumental parameters for high-resolution MS experiments (Table S3); UPLC 45 min separation gradient program (Table S4); average drift time for ion mobility features corresponding to charge states +16–14 of BSA from denatured deposit extracted (after 5 min refolding period) and native deposit laser-ablated samples (Table S5); selected m/z windows for IM data extraction (Table S6); average drift time for the ion mobility features corresponding to charge states +16 to +14 of BSA from native deposit extrcated and denatured deposit extracted (after 5 min period) samples (Table S7); average CIU50 for ion mobility features corresponding to charge states +16–14 of BSA, +10–8 of CA, and +17–15 of (BHb)₄ (Table S8); average IM drift times for IM features corresponding to charge states +10–8 of CA from native deposit extracted and native deposit laser-ablated samples (Table S9) (PDF)

The manuscript was written through the contributions of all authors. T.S. and K.K.M. conceived and coordinated the experiments. N.F., with assistance from K.B.H., carried out all native MS and bottom-up experiments at BU. MALDI and sample LA were conducted by B.C.E. and K.O.H. at LSU. N.F., B.F.S., K.K.M, and T.S. assisted with the acquisition, processing, and analysis of IM-MS data. N.F. and T.S. wrote the manuscript, and all authors provided feedback and suggestions. J.M.C. assisted with bottom-up proteomics experiments. All authors reviewed the manuscript and approved the final version of the manuscript.

The authors declare the following competing financial interest(s): K.K.M. is the owner of Laser Bioanalytics LLC.