High-throughput Proteomics Enabled by a Photocleavable Surfactant

Abstract

Mass spectrometry (MS)-based proteomics provides unprecedented opportunities for understanding the structure and function of proteins in complex biological systems; however, protein solubility and sample preparation before MS analysis remain a bottleneck preventing high-throughput proteomics. Herein, we report for the first time a high-throughput bottom-up proteomics method enabled by a newly developed MS-compatible photocleavable surfactant, 4-hexylphenylazosulfonate (Azo)^[1] that facilitates robust protein extraction, rapid enzymatic digestion (30 min compared to overnight), and subsequent MS-analysis following UV degradation. Moreover, we developed an Azo-aided bottom-up method for efficient analysis of integral membrane proteins, which are key drug targets and are generally underrepresented in global proteomic studies. Furthermore, we demonstrated the unique ability of Azo to serve as an “all-in-one” MS-compatible surfactant for both top-down and bottom-up proteomics, with streamlined workflows for high-throughput proteomics amenable to clinical applications.

Graphical Abstract

Photocleavable Surfactant. In their Communication, Ge, Jin, and co-workers demonstrated the unique ability of 4-hexylphenylazosulfonate (Azo) to serve as an “all-in-one” MS-compatible surfactant for both high-throughput bottom-up and top-down proteomics.

Mass spectrometry (MS)-based proteomics allows large-scale, quantitative measurements of proteins, and provides unprecedented insights into the structure and function of proteins in complex biological systems.^{[2, 3]} Currently, there are two major MS-based proteomic approaches: the bottom-up approach, where proteins are enzymatically digested and the resulting peptides measured by MS to infer the identity, quantity, and modification status of the proteins; and the top-down approach, where intact proteins are measured and their sequences determined for complete characterization of their proteoforms.^[4–7] Bottom-up proteomics remains the most widespread proteomic method due to the well-established experimental and computational tools^[8] developed over the last two decades, but it has limitations for comprehensive analysis of proteoforms.^{[5, 7]} On the other hand, significant strides have recently been made to advance top-down proteomics, which provides comprehensive analysis of proteoforms, owing to the recent successful developments of high-resolution mass spectrometers, protein separations, and user-friendly software for data analysis.^{[5, 6, 9]} However, further method developments are urgently needed to increase the throughput in both bottom-up and top-down approaches, especially for time-sensitive clinical applications.

Sample preparation remains a bottleneck for obtaining high-quality MS data in a high-throughput fashion for both top-down and bottom-up proteomics.^[9–12] Typically, surfactants (also known as detergents) are added to effectively solubilize proteins, especially membrane proteins, in cells and tissues.^{[3, 8, 11, 13, 14]} However, conventional surfactants that can solubilize cells and tissue well, such as sodium dodecyl sulfate (SDS), are incompatible with MS and need be removed before MS analysis, which is not only time consuming but may result in protein loss and degradation.^{[13, 15, 16]} An MS-compatible surfactant that can solubilize and denature proteins with comparable efficacy to SDS without the need for time-consuming sample clean-up is attractive for streamlining sample preparation for proteomics.^[12] To address this problem, MS-compatible surfactants that can be degraded into innocuous non-surfactant by-products have been developed.^{[15, 17–19]} Particularly, acid-labile ionic surfactants such as RapiGest™^{[18, 20]}, ProteaseMax™^[19], and MaSDeS^[17] can significantly improve the protein solubility and facilitate protein digestion. However, these acid-labile surfactants are degraded in highly acidic conditions post-digestion and often require an offline clean-up step, precluding automation.^[15] Moreover, strong acidic degradation may cause the loss of important post-translational modifications that are unstable in acidic conditions (i.e. acid-labile phosphorylations^[21] and glycosylations^[22]). Furthermore, some of these acid-labile surfactants generate amine products following degradation that interfere with the isobaric amine-reactive tandem mass tag labeling reagents, such as tandem mass tag (TMT) and isobaric tags for relative and absolute quantitation (iTRAQ).^[15] Most importantly, none of these acid-labile surfactants are compatible with top-down proteomics for the direct analysis of intact proteins, because they greatly suppressed intact protein MS signal as reported previously.^[1] Therefore, we aim to develop a high-throughput sample preparation method that is amenable to both top-down and bottom-up proteomics and overcome the limitations of the conventional acid cleavable surfactants.

Recently, we developed a photo-cleavable surfactant, 4-hexylphenylazosulfonate (Azo), for top-down proteomics.^[1] Azo is straightforward to synthesize (requiring only two-step synthesis and simple purification),^[23] effectively solubilizes proteins (including membrane proteins), and can be rapidly degraded before MS analysis.^[1] Here, for the first time, we demonstrate that Azo is fully compatible with bottom-up proteomics and uniquely serves as an “all-in-one” MS-compatible surfactant for both bottom-up and top-down proteomics, which greatly facilitates high-throughput sample handling before MS-analysis (Scheme 1). A direct comparison of the Azo and a leading acid-labile surfactant, RapiGest™ (RG, also known as ALS),^{[18, 20]} clearly demonstrated Azo is compatible with direct-infusion electrospray MS analysis of intact proteins and peptides whereas RG greatly suppressed MS signal (Figure S1). Moreover, we demonstrate the ability of Azo to facilitate protein extraction, enable rapid enzymatic digestion, and rapidly degrade by UV before MS analysis for a streamlined bottom-up proteomics approach (Figure 1a).

Scheme 1.

Scheme for azo-enable high-throughput top-down and bottom-up proteomics.

Figure 1.

Enhanced enzymatic digestion and MS analysis using a photocleavable surfactant, Azo. (a) Scheme for Azo-aided bottom-up proteomics. (b) Degradation of Azo by UV irradiation. (c) UV-Vis spectrum for Azo highlighting maximal absorbance at 305 nm. (d) UV-Vis spectrum monitoring the rapid degradation of Azo (0.1% in water) as a function of irradiation time with a 100 W mercury lamp. (d) Digestion of myoglobin [0–1 h, Azo (0–0.2%)] was monitored by SDS-PAGE, stained with Coomassie Blue, and (f) LC-MS analysis using a Q-TOF mass spectrometer. Relative abundances of the extracted ion chromatograms was normalized to 5.5 E8.

First, we demonstrated the rapid degradation of Azo by UV-Vis spectroscopy. The intact surfactant has a maximal absorbance at 305 nm^{[1, 24]} (Figure 1b–c); thus, degradation of Azo can be achieved using a mercury lamp (305 nm) in less than 5 min ^{[1, 25]} (Figure 1d and S2). This facile surfactant degradation method makes Azo ideal for rapid sample processing.

Next, we evaluated Azo-aided in-solution digestion. Myoglobin, a trypsin resistant protein,^[18] was digested in-solution without (0%, control) or in the presence of Azo (0.05%, 0.1%, or 0.2%). The efficiency of digestion was assessed by visualization of the remaining intact myoglobin through SDS-polyacrylamide gel electrophoresis (SDS-PAGE) analysis (Figure 1e). After 0.5 h of trypsin digestion in the absence of surfactant, the intact protein band at ~17 kDa was clearly observed, suggesting poor digestion efficiency. In contrast, only a faint intact protein band was observed for sample digested in the presence of 0.05% Azo, and no intact protein band was observed in the presence of 0.1% and 0.2% Azo after 0.5 h of digestion indicating that Azo greatly improved digestion efficiency. Subsequently, we evaluated a longer digestion time (1 h) in the presence or absence of Azo. Even after 1 h digestion, a predominant intact protein was still observed without Azo, whereas only a minute amount of intact protein was visible in the presence of 0.05% Azo and no intact protein was detected for 0.1% or 0.2% Azo (Figure 1ef). This is conceivable, since without a denaturing agent (e.g. an anionic surfactant) the enzyme has limited access to the protein backbone, precluding digestion. These results indicate Azo facilitates denaturation of the protein, providing efficient digestion. Furthermore, we analyzed the Azo-aided myoglobin digest (0.1% Azo, 1 h) via direct electrospray ionization for peptide mass fingerprinting analysis without an additional clean-up procedure and observed 89% sequence coverage (Figure S3, bottom panel). On the other hand, analysis of the myoglobin sample digested in the absence of surfactant (0% Azo, 1 h) yielded minimal detectable peptides, and the MS signal was largely dominated by the intact protein (Figure S3, top panel). Next, we performed a direct comparison between Azo and RG for trypsin digestion of myoglobin and direct-infusion MS analysis, observing lower peptide signal for the sample with RG than that with Azo (Figure S4). Finally, to further confirm the rapid rate of enzymatic digestion in the presence of Azo, we digested a standard mixture of insulin, myoglobin, and carbonic anhydrase with (0.1% or 0.2%) and without Azo. We observed a similar improvement in digestion with Azo (Figure S5). Taken together, these results indicate proteins (even those resistant to trypsin digestion) can be rapidly and efficiently digested using Azo.

After demonstrating rapid enzymatic digestion using standard proteins, we further investigated Azo-aided digestion of complex cell lysates. A single-step Azo protein extraction was performed using human embryonic kidney 293T (HEK293T) cells to generate whole cell lysate. SDS-PAGE analysis demonstrated reproducible extractions for three HEK293T samples (Figure 2a). The proteins were digested for 1 h or overnight to determine whether digestion time, the most time-consuming step in the bottom-up proteomics workflow, could be shortened significantly using Azo while maintaining reproducibility. After surfactant degradation by UV, a brief centrifugation step, and liquid chromatography (LC)-MS/MS analysis, the data was processed using MaxQuant ^[26] to determine the number of protein groups identified in each sample and to perform label-free quantitation (Table S1). Overall, 910 and 938 proteins were identified in all three extraction replicates using overnight digestion (Figure 2b) and 1 h digestion (Figure 2c), respectively. Combined, we identified 1557 proteins using an overnight digestion and 1572 proteins using a 1 h digestion with a 60 min LC gradient before MS analysis. This corresponds to an 81% overlap between the two conditions (Figure 2d), confirming 1 h is sufficient for effective protein digestion in the presence of Azo. Next, we compared the performance of Azo-aided digestion (30 min and overnight) with a gold-standard protocol for bottom-up proteomics using SDS, Filter Assisted Sample Preparation (FASP) developed by Mann and co-workers (overnight digestion).^[11] Using three digestion replicates (proteins were extracted with Azo or SDS respectively) and a 60 min LC-MS analysis, we observed the greatest number of identifications using the Azo digestion (overnight) followed by the FASP (overnight) and Azo digestion (30 min) method (Figure S6 and Table S2). Significantly, we observed only 33 fewer protein identifications using a 30 min digestion protocol with Azo when compared to FASP, demonstrating a very robust, reproducible method that greatly reduces the sample preparation time by eliminating the multiple centrifugation and offline desalting steps (SI Discussion)

Figure 2.

Rapid and reproducible extraction, enzymatic digestion, and LC-MS/MS analysis of proteins from human embryonic kidney 293T (HEK293T) cells (a) SDS-PAGE analysis demonstrates consistent extraction profiles across three extractions (marker (M), 1, 2, 3). Aliquots were taken from each exaction and digested overnight or for 1 h. After LC-MS/MS and MaxQuant identification, a high degree of overlap was observed for both the (b) overnight and (c) 1 h samples. Similarly (d) 1403 of the 1726 combined identified protein groups were observed in both digestion methods. Data collected using Q-TOF.

Next, we demonstrate Azo’s ability to aid in the analysis of integral membrane proteins, which are hydrophobic and generally expressed at lower levels, thus are underrepresented in broad proteomics studies.^{[15, 27]} Here, we focused on improving the bottom-up workflow for membrane proteomics by implementing a cloud-point extraction, using Triton X-114, to enrich membrane proteins from HEK293T cells with a single extraction (SI Discussion).^[28] However, Triton surfactants are highly incompatible with MS analysis^[29] and require removal before MS analysis. Subsequently we used Azo to solubilize the precipitated protein pellets enriched with integral membranes for a more effective analysis of this important class of proteins. We observed more than a 6-fold increase in integral membrane protein identified using this method when compared to whole cell extraction experiment (conceivably because highly abundant proteins that traditionally suppress the detection of integral membrane proteins were depleted) (Figure 3a and Table S3). Similarly, a 4-fold increase in protein groups identified was achieved using 30 min digestions times for the enriched sample compared to the whole cell lysate (Figure 3a and Table S3). Importantly, we detected several proteins with over 10 transmembrane domains, demonstrating the ability of Azo to solubilize highly hydrophobic proteins and enable their MS analysis (Figure 3b). The results indicate longer digestion improves the identification of integral membrane proteins even with Azo, which can be rationalized by the lower abundance of lysine and arginine residue particularly in the transmembrane regions.^[27] Finally, we performed a global PTM discovery search using MetaMorephus^[30] to localize modification on membrane proteins (Figure S7 and Table S3).

Figure 3.

Analysis of integral membrane protein enabled by Azo. Proteins were extracted and membrane proteins enriched using a Triton x-114 cloud point extraction procedure,²⁸ then precipitated, and solubilized again in Azo for in-solution digestion. Additionally, Azo was used for whole cell protein extraction. Using an (a) overnight digestion a significantly higher number of integral membrane proteins were identified using the phase separation enrichment followed by Azo-aided digestion. Similarly a 30 min digestion was performed demonstrating a high-throughput membrane protein analysis methods. (b) A table of representative membrane proteins with a high number of transmembrane domains that were solubilized using Azo. Data represent a single LC-MS experiment.

Here, we sought to demonstrate the unique application of Azo as an “all-in-one” MS-compatible surfactant for both top-down and bottom-up proteomics. Proteins were extracted, using Azo, from left ventricle heart tissue (after aqueous buffer extractions) and an aliquot was taken for bottom-up analysis. We observed that many of the protein identifications were involved in cellular and metabolic processes or localization (Figure S8a and Table S4). Moreover, we demonstrated the potential of this method to investigate disease-relevant proteins (e.g. sarcomeric proteins) and their interacting partners^[31] which is critical for gaining a better understanding of their role in disease mechanisms (Figure S8b). Furthermore, an aliquot was taken for top-down proteomics which provides additional information about existing proteoforms (Figure 4 and Figure S9), thus improving the biological information that can be obtained from a sample. Overall LC-MS analysis of cardiac tissue illustrated the complexity of the Azo-extracted samples with many proteins and their respective proteoforms eluting simultaneously (Figure S10), demonstrating the promise of Azo for future applications in global top-down proteomics studies. Importantly, Azo also enables the extraction and analysis of intact integral membrane proteins (Figure S11).

Figure 4.

Azo-enabled top-down proteomics analysis of cardiac tissue. (a-d) Intact mass spectra of proteoforms analyzed by LC-MS. Data collected using Maxis II Q-TOF.

In summary, we developed a novel approach using an anionic photocleavable surfactant, Azo, to enable high-throughput bottom-up proteomics. Azo can effectively extract and solubilize proteins reproducibly, enables rapid enzymatic digestion, and is amenable to MS analysis without an additional desalting step which improves the throughput, permitting bottom-up analyses for a wide range of applications. Moreover, we have developed an Azo-aided bottom-up proteomics workflow for effective membrane protein solubilization. Furthermore, we have demonstrated the unique capability of Azo as an “all-in-one” MS-compatible surfactant for both top-down and bottom-up proteomics with great potential to provide a streamlined strategy for high throughput proteomics with far-reaching applications to fully realize the impact of proteomics in clinical diagnosis.