Introduction

Implementations of desorption electrospray ionization mass spectrometry (DESI-MS) have been disproportionately in favor of direct/ambient analysis of smaller molecules such as metabolites and lipids, since analysis of larger molecules such as proteins by DESI-MS has been challenging [1, 2]. However, with the continuous efforts towards improving DESI-MS of proteins, this technique is rapidly becoming a powerful tool for direct analysis of large proteins (> 25 kDa) from complex mixtures.

Solvent additives such as ammonium bicarbonate [3] and serine [4], or delayed-desorption-DESI [5] and combinations of these approaches have aimed to address supposed problems with the slow kinetics of protein dissolution during the analysis of proteins by DESI.

Another powerful approach entails coupling of DESI-MS to ion mobility which now allows for imaging of small proteins and peptides directly from tissue samples [6, 7].

The addition of polar organic vapors into the spray chamber or curtain gas during ESI analysis was shown to enhance electrospray ionization of proteins and peptides. Under such conditions, a reduction of alkali metal adduction was observed together with changes in protein charge states typically to lower charge values [8,9,10]. The addition of polar organic vapors such as acetonitrile, acetone, ethyl acetate, water, and small alcohols helped remove metal adducts presumably via ion evaporation. It was suggested that the effectiveness of these vapors in removing the metal species comes from their ability to lower the activation energy required for metal ion evaporation. Vapors that have a greater impact in lowering the activation energy of ion evaporation of the metal will be more beneficial in terms of removing adducts from protein complexes [8].

Additionally, an enclosed commercial ionization source was shown to increase the charge states of tryptic peptides when ionized in an atmosphere enriched in acetonitrile vapors [11].

Given the similarities between ESI and DESI, it is likely that the same treatment could positively affect DESI-MS analysis of proteins. Despite the differences in the initial sample phase, after dissolution and desorption, DESI follows similar ionization mechanisms as ESI [12]. Therefore, successful approaches to improve protein analysis by ESI have often been applicable to DESI as well.

The application of vapor additives in DESI requires enclosure of the DESI desorbing and ionizing plume to contain the vapors. An enclosed DESI source was previously described [13]. Here we introduce polar organic vapors of acetone, acetonitrile, ethyl acetate, methanol, and water to the gas phase through a semi-enclosed DESI system.

Experimental Section

Materials

Equine cytochrome c (cyt c, 12.3 kDa), bovine erythrocyte carbonic anhydrase isozyme II (CAII, 30.0 kDa), and bovine serum albumin (BSA, 66 kDa) were purchased from Sigma-Aldrich (St. Louis, MO). Bovine myoglobin (Myo, 16.9 kDa) was purchased from Protea (Morgantown, WV). HPLC-MS grade methanol, acetone, and acetonitrile were purchased from Sigma-Aldrich (St. Louis, MO). Extra dry (water < 50 ppm) ethyl acetate was purchased from ACROS Organics (Geel, Belgium). LC-MS grade formic acid was purchased from Fluka Analytical (Morris Plains, NJ). Ultrapure water was supplied from Thermo-Barnstead Water Polisher. Porous-polyethylene surfaces (PE) with average pore size of 15–45 μm (POREX-4900) were purchased from Interstate Specialty Products (Sutton, MA).

Sample Preparation

Each protein standard was prepared and analyzed independently. Lyophilized proteins were first dissolved in milliQ water to prepare stock solutions. Protein samples were made from the stock solutions by further dilution with the milliQ water to reach a final concentration of 80 μM. The protein solutions were spray-deposited into a PE surface to yield sample lines with an estimated surface concentration of 20–25 pmol/mm2. For each experiment, at least 3 sample lines were scanned. Error bars represent ± standard deviation.

Instrumentation

The DESI-MS analysis was performed with a linear ion trap mass spectrometer (LTQ, Thermo Scientific, Waltham, MA, USA) combined with a 3-dimensional translational stage (Purdue University, West Lafayette, IN, USA). A house-built electrosonic spray ionization source (ESSI) was used for generating pneumatically assisted electrospray using two co-axial fused silica capillaries inside a T-piece. An 80% methanol solution containing 0.1% formic acid was delivered through the inner silica capillary (I.D. 50 μm, O.D. 220 μm) with the flow rate of 5 μL/min and nebulizing nitrogen gas was delivered through the outer silica capillary (I.D. 320 μm, O.D. 450 μm, length 1.5 cm) at 100 psi. For generating charged solvent droplets, 4.0 kV was applied to the syringe delivering the DESI spray solvent. The MS inlet temperature was set at 250 °C. LTQ ion optics voltages were optimized for each protein individually. Tube lens voltage and ion transfer capillary voltage were optimized between 110–130 V and 20–45 V, respectively.

DESI Parameters and Enclosure

The sprayer to MS inlet distance and sprayer to surface distance were set at 4 mm and 1 mm, respectively, and the incident spray angle was adjusted to 54–55°. The plastic enclosure was cut from a 1-mL plastic pipette tip which fitted tightly around the front ring of the 1/16″ Swagelok nut that secures the nebulizing gas capillary of the ESSI sprayer assembly, as shown in Figure 1. The enclosure was specifically cut for the desorption sprayer so that attaching the enclosure would make minimum change in the desorption spray geometry. To fit the extended MS inlet inside the enclosure, a small opening was cut in the front rim of the plastic enclosure. Vapors were delivered to the enclosure cavity by 1/8″ PTFE tubing that was connected to a Schott® bottle half-filled with solvent. The N2 inlet tube protruded into the bottle to a position close to the solvent surface and below that of the vapor exit tube. The vapor exit tube entered the enclosure through a small hole positioned behind the DESI sprayer drilled into the back of the plastic tip. The nitrogen flow rate was controlled with a needle valve and optimized at 1 L/min or 50 mL/min for less or more confining enclosures, respectively, as discussed below. Reagent vapors investigated were acetone, acetonitrile, ethyl acetate, methanol, and water.

Figure 1
figure 1

Photo of enclosed DESI sprayer and vapor addition setup

Full size image

Results and Discussion

Enclosure Considerations

The effects resulting from vapor addition during DESI analysis are the consequences of two facets: physical effects and chemical effects. Physically attaching the enclosure to the DESI sprayer can affect the performance of the DESI source, most notably through subtle changes in DESI sprayer geometry which influences droplet dynamics. Moreover, the enclosure’s physical parameters such as the shape, position of vapor delivery inlet, and whether the surface and DESI sprayer that are completely confined within the enclosure cavity can affect the gas flow dynamics. For example, a less confining enclosure with space between the surface and the enclosure optimizes to higher vapor flow rates (1 L/min), whereas in a tightly confined enclosure cavity the flow dynamics are completely disturbed with such high vapor flow rates, and as a result, signal deteriorates. In the following experiments, the enclosure was made to maintain the optimized non-enclosed DESI spray geometry as much as possible. The back of the enclosure was also raised slightly above the surface to provide a less restricted cavity which allows using higher vapor flow rates.

Aside from the physical effects, the chemical effects of each solvent vapor on the primary electrospray droplets leaving the ion source, on the sample, on the surface, or on the secondary droplets after desorption, are related to the properties of the solvent molecules such as polarity, gas-phase proton affinity, and dipole moment. Optimal conditions, therefore, will depend on an intricate balance between the shape of the enclosure, DESI sprayer parameters, vapor flow rate, and possibly also the chemical identity of the vapor.

Effects of Different Solvent Vapors on Different Proteins

To survey the effects of different vapors on protein signal, initially two model proteins (myoglobin and bovine serum albumin) were analyzed by DESI using an array of vapor additives as summarized in Figure 2. Nitrogen gas by itself as control, or doped with methanol, acetone, acetonitrile, water, and ethyl acetate vapors were each separately introduced to the semi-enclosed DESI at a flow rate of 1 L/min. The effect of the vapor additives on protein signal was analyzed in positive ion mode.

Figure 2
figure 2

Effect of different vapors on signal intensity of natively deposited proteins (a) myoglobin and (b) bovine serum albumin when analyzed by DESI-MS using 80% methanol containing 0.1% formic acid as the solvent

Full size image

The signal intensities of the highest intensity charge states (HICS) and the total protein intensities for myoglobin and the HICS for bovine serum albumin were analyzed under nitrogen gas enriched with the different vapors. When ethyl acetate vapor was supplied under these conditions, the HICS signal intensity of both proteins increased approximately 4 times as shown in Figures 2, 3a, and 3b. However, other vapor additives only mildly affected the HICS intensities and caused the signal to decrease or remain relatively unchanged. The effects of the various vapors on the total protein signal for myoglobin shown in Figure S1 mirror the observations in Figure 2a. Unfortunately, the total protein signal intensity for BSA was difficult to obtain due to heavy adduction and the low mass resolution of the mass spectrometer used. These observations were also similar in magnitude to the results previously published when vapors were introduced into the curtain gas during ESI-MS of holo-myoglobin [9].

Figure 3
figure 3

Comparison of spectra and absolute signal intensities for different proteins (a) myoglobin, (b) bovine serum albumin, (c) cytochrome c, and (d) carbonic anhydrase II when exposed to N2 vapor (top spectrum) and when exposed to ethyl acetate vapor (bottom spectrum). Vapor flow rate was set at 1 L/min

Full size image

The addition of ethyl acetate as a fraction directly into the desorption spray solvent reduced the signal intensity dramatically. Figure S2 shows data for CAII analyzed with and without ethyl acetate mixed as a fraction into the desorption spray solvent. Ethyl acetate is miscible in water up to 10% and totally miscible in 50% methanol water solutions. When 10% ethyl acetate was added as a fraction into the desorption spray solvent, the signal intensity reduced by over an order of magnitude. When the fraction of ethyl acetate was further increased to 20% it was hard to detect any protein peaks. Considering the 5 μL/min desorption spray flow rate, even at 20%, the total moles of ethyl acetate delivered per unit time was still significantly lower than when ethyl acetate vapors were supplied as a dopant into the auxiliary gas, where the ethyl acetate consumption measured at the supply bottle was approximately 500 μL/min.

The remarkable effect of ethyl acetate vapors on signal intensity was also observed with other proteins such as cytochrome c (Figure 3c) and carbonic anhydrase II (Figure 3d). Figure 3 shows the effect of ethyl acetate vapor on signal intensities and mass spectra for 4 different proteins compared with nitrogen. Here, when envelopes were deconvoluted, the summed signal intensities for each protein increased by factors of 6, 5, and 3 times each for myo, cyt c, and CAII, respectively. (The unresolved envelope of BSA could not be deconvoluted.)

In an attempt to further optimize the conditions, the desorption sprayer and surface were more tightly enclosed, i.e., the plastic tip completely touched the sample surface and enclosed the DESI sprayer. This setup was not tolerant of the 1 L/min vapor flow rate, and the vapor doped nitrogen auxiliary gas flow rate optimized at 50 mL/min. With the more confining setup, ethyl acetate vapors increased the signal of cytochrome c (Figure 4a) and carbonic anhydrase II (Figure 4b) even more, to approximately 6 times when compared with nitrogen vapor. This illustrates that further improvements to signal intensity will be possible with a carefully designed and optimized enclosure geometry and operating conditions. The chemical effect of ethyl acetate on signal intensity can be observed regardless of the physical complications of the addition of an enclosure and the introduction of auxiliary N2 gas, both which affects the droplet dynamics of the DESI processes.

Figure 4
figure 4

Comparison of spectra and signal intensity for (a) cytochrome c and (b) carbonic anhydrase II when enclosure area was more restricted. Vapor flow rate was set at 50 mL/min

Full size image

A noteworthy observation from comparing the spectra for all proteins exposed to ethyl acetate vapors was an overall shift towards higher charge states. For example, the addition of ethyl acetate vapors to CAII caused a shift in the HICS to slightly higher charge states (z = + 33) compared with nitrogen controls (z = + 30). The HICS of cyt c increased by 2 when comparing nitrogen vapor with ethyl acetate. The highest observed charge state (HOCS) for cyt c exposed to nitrogen was z = + 19 and this increased to z = + 21 when exposed to ethyl acetate vapor. The addition of ethyl acetate leads to a complicated envelope, as if bimodal ion populations are created. Interestingly, organic vapors in ESI and nano-ESI showed an overall charge reduction for native proteins when vapors were introduced through the curtain gas [8, 10]. The opposite effect can be seen here with all proteins showing an overall increase to higher charge states after ethyl acetate vapor interaction. This is possibly a consequence of the denaturing solvent used in the DESI desorbing spray, even as the proteins were deposited in native state. Previously, it was shown that peptides and proteins analyzed in ESI under denaturing conditions responded differently to vapors in the gas phase compared with when analyzed under native state preserving conditions [14]. This observation can be explained through the possibility of a different ionization mechanism recently proposed for denatured proteins, known as the chain ejection model (CEM) [15, 16].

In conclusion, exposure of the DESI spray plume to organic vapors demonstrated the ability to change the charge state distributions, and in the case of ethyl acetate, also to increase signal intensities obtained for proteins. This effect appears to be independent of protein characteristics, such as size or isoelectric point values. The magnitude of this effect is however dependent on the enclosure setup and vapor flow rate. The physical parameters are also interdependent, and in addition to geometrical complexities, determine the amount of vapor that can be delivered to the spray plume. Therefore, detailed optimization of enclosure parameters and vapor flow rates are necessary. The promising observation was that regardless of physical parameters of the process, the improvement in signal intensity was observed for multiple proteins, including proteins larger than 25 kDa, which are challenging to analyze by DESI-MS.