Research Article Received: 9 April 2014
Revised: 13 May 2014
Accepted: 13 May 2014
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673 (wileyonlinelibrary.com) DOI: 10.1002/rcm.6946
Transmission geometry laser ablation into a non-contact liquid vortex capture probe for mass spectrometry imaging Olga S. Ovchinnikova, Deepak Bhandari†, Matthias Lorenz and Gary J. Van Berkel* Organic and Biological Mass Spectrometry Group, Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6131, USA RATIONALE: Capture of material from a laser ablation plume into a continuous flow stream of solvent provides the means for uninterrupted sampling, transport and ionization of collected material for coupling with mass spectral analysis. Reported here is the use of vertically aligned transmission geometry laser ablation in combination with a new non-contact liquid vortex capture probe coupled with electrospray ionization for spot sampling and chemical imaging with mass spectrometry. METHODS: A vertically aligned continuous flow liquid vortex capture probe was positioned directly underneath a sample surface in a transmission geometry laser ablation (355 nm, 10 Hz, 7 ns pulse width) set up to capture into solution the ablated material. The outlet of the vortex probe was coupled to the Turbo V™ ion source of an AB SCIEX TripleTOF 5600+ mass spectrometer. System operation and performance metrics were tested using inked patterns and thin tissue sections. Glass slides and slides designed especially for laser capture microdissection, viz., DIRECTOR® slides and PEN 1.0 (polyethylene naphthalate) membrane slides, were used as sample substrates. RESULTS: The estimated capture efficiency of laser-ablated material was 24%, which was enabled by the use of a probe with large liquid surface area (~2.8 mm2) and with gravity to help direct ablated material vertically down towards the probe. The swirling vortex action of the liquid surface potentially enhanced capture and dissolution not only of particulates, but also of gaseous products of the laser ablation. The use of DIRECTOR® slides and PEN 1.0 (polyethylene naphthalate) membrane slides as sample substrates enabled effective ablation of a wide range of sample types (basic blue 7, polypropylene glycol, insulin and cyctochrome c) without photodamage using a UV laser. Imaging resolution of about 6 μm was demonstrated for stamped ink on DIRECTOR® slides based on the ability to distinguish features present both in the optical and in the chemical image. This imaging resolution was 20 times better than the previous best reported results with laser ablation/liquid sample capture mass spectrometry imaging. Using thin sections of brain tissue the chemical image of a selected lipid was obtained with an estimated imaging resolution of about 50 μm. CONCLUSIONS: A vertically aligned, transmission geometry laser ablation liquid vortex capture probe, electrospray ionization mass spectrometry system provides an effective means for spatially resolved spot sampling and imaging with mass spectrometry. Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
Recently, several groups have demonstrated that laser ablation (LA) with subsequent liquid capture of the ablated material is a valuable method for atmospheric pressure (AP) surface sampling/ionization and imaging with mass spectrometry (MS).[1–7] With this approach, material from a LA plume is collected into hanging droplets or a continuous flow stream of solvent. Sample collection is followed by continuous infusion, flow injection, or a post-sampling process such as high-performance liquid chromatography (HPLC)[1,4] or capillary electrophoresis[6] separation of the captured components, and then ionization and detection by mass spectrometry (MS). As an alternative, LA and capture * Correspondence to: G. J. Van Berkel, Organic and Biological Mass Spectrometry Group, Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6131, USA. E-mail: [email protected] †
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673 Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
1665
Current address: Environmental Health, Division of Laboratory Sciences, Centers for Disease Control and Prevention, Atlanta, GA 30341, USA.
directly into a formed probe-to-surface liquid microjunction (versus, e.g., a non-contact hanging droplet above the surface) has been shown to provide 100% sampling efficiency of surface material that is not directly soluble or that dissolves slowly in the extraction solvents on the time frame of the experiment.[7] The use of an AP secondary, or post-sampling, ionization process such as electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) with LA can sometimes be limited by ablation and formation of particulates rather than the desired gas-phase molecules.[8] A major fraction of these particles can be in a size range unsuitable for ionization by these techniques, given the limited gas-phase particulate and charged droplet/reagent ion interaction times (i.e., milliseconds).[9] Capture of laser-ablated material into an appropriate solvent droplet or flow stream provides a greater time period for the particulates or agglomerates to dissolve or otherwise dissociate forming molecular species suitable for ionization. Once the captured material is dissolved in solution, ESI, APCI, or other liquid introduction AP ion source might be used to efficiently ionize it.
O. S. Ovchinnikova et al. Our group has demonstrated that the capture of material from a LA plume into a continuous flow stream of solvent using a non-contact probe provides uninterrupted sampling, transportation and ionization of collected material, affording the means for mass spectrometry imaging (MSI).[2] We showed this possibility using horizontally aligned transmission geometry LA combined with a self-aspirating liquid microjunction surface sampling probe/electrospray ionization (LMJ-SSP/ESI) emitter.[10,11] The analysis of inked lines, letters and fingerprints on glass slides using a 210 μm laser spot size and oversampling demonstrated the ability to image with a spatial resolution approaching 100 μm, based on a comparison of the distinguishable features in the optical and chemical images of the surfaces analyzed. The LMJ-SSP/ESI emitter used in that work was an openair system that replaced the normal ion source on the mass spectrometer. The probe was built using a stainless steel tee, a 10-cm-long inner sampling/emitter capillary with a 254 μm o.d. and a 127 μm i.d., an outer tube on the sampling end with 635 μm o.d. and 327 μm i.d., and a nebulizer tube on the spray side. The small size of the liquid surface at the sampling end (ca 0.6 mm diameter) and volume of that probe (ca 1.3 μL) provided a rapid response time (3.8 s at 20 μL/min) and minimal sample dilution. However, the small size of the liquid surface area combined with the relatively large laser spot size required careful alignment for optimal capture of the ablated material. Even when optimized, we have shown that the basic geometry of operation provided an ablation plume capture efficiency of no more than 10%.[7] In this paper, we report on a vertically aligned transmission geometry LA setup combined with a non-contact liquid vortex capture probe and ESI-MS for spatially resolved spot sampling and MSI. Compared with our previous horizontal design, the capture efficiency was more than doubled by the use of a probe with an order of magnitude larger liquid surface area (2.8 mm2 versus 0.28 mm2) and with gravity to help direct ablated material vertically down towards the probe. The large liquid surface of the present probe also made the exact on-axis LA spot-to-probe alignment less critical, simplifying the operation of the system. In addition, the continuous flow liquid capture sampling probe was connected directly to the existing ion source of the mass spectrometer keeping in place all the safety features and the optimized ion source components of the commercial system. The use of DIRECTOR® and PEN 1.0 (polyethylene naphthalate) membrane slides, both designed for laser capture microdissection, as sample substrates permitted a broad range of materials to be ablated using a UV laser (355 nm) without photodamage. An MSI resolution of 6 μm was demonstrated using stamped ink patterns on glass and a resolution of about 50 μm was estimated when imaging a selected lipid in a thin-tissue section.
EXPERIMENTAL Chemicals and materials
1666
LC-MS Optima grade methanol, acetonitrile and water with 0.1% formic acid were purchased from Fisher Scientific (Pittsburg, PA, USA). Glass slides were purchased from VWR International (West Chester, PA, USA), PEN 1.0
wileyonlinelibrary.com/journal/rcm
(polyethylene naphthalate) membrane coated slides were from Leica (Buffalo Grove, IL, USA) and DIRECTOR® slides from Expression Pathology (Rockville, MD, USA). Horse heart cytochrome c and bovine insulin side chain B were purchased from Sigma-Aldrich (St. Louis, MO, USA). Polypropylene glycol (PPG) tune solution was acquired from AB Sciex (Framingham, MA, USA). Blue fine and standard point permanent markers (Sharpie®, Sanford, Oak Brook, IL, USA) containing the dye basic blue 7 were used to prepare ink layers and patterns on the slides. The stamped grid pattern was prepared on a DIRECTOR® slide by applying the blue Sharpie marker ink to one side of a TEM grid (G100HS fine square mesh with 19 μm hole openings and 6 μm wide lines, SPI Supplies, West Chester, PA, USA) and then pressing the grid onto a DIRECTOR® slide. Coronal mouse brain tissue sections (12-μm-thick) were thaw mounted on PEN 1.0 (polyethylene naphthalate) membrane slides. Optical images of the surfaces analyzed were obtained using a Nikon Biophot upright microscope (Tokyo, Japan) and associated software. Experimental setup The overall experimental setup is illustrated in Fig. 1(a). A Nd:YAG MiniLight II laser (10 Hz, 355 nm, 7 ns pulse width; Continuum, Santa Clara, CA, USA) was used for all experiments. A UV fused-silica right-angle prism was used to turn the beam 90° to be directly on-axis with the vertically mounted liquid vortex capture probe. Using a LP-1-XYZ 3-axis lens positioner (Newport, Irvine, CA, USA) and a Newport M-10X (10X) microscope objective, the laser beam was focused directly on the sample side of the support (Fig. 1(b)). The lens positioner assembly was mounted onto model 430 and 429 translation stages (Newport) to give x-y positioning control for laser alignment and for fine adjustment of the z-positioning of the objective. A SH05 beam shutter with a SC10 shutter controller (Thorlabs Inc., Newton, NJ, USA) was used to control the laser beam. The laser pulse energy was measured at the exit from the glass slides using a FieldMaxII laser power meter (Coherent, Santa Clara, CA, USA). Microscope slides were positioned in the slide holder of a MS2000 x-y robotic platform (Applied Scientific Instrumentation Inc., Eugene, OR, USA) that was used to manipulate the sample surface relative to the stationary laser beam. The HandsFree Surface Analysis© control software used to position the sample surface has been described in previous work.[12] Schematic details of the LA liquid capture probe used in these experiments are shown in Fig. 1(b). The probe used a co-axial tube design on the sampling end. The outer tube was a stainless steel tube (~1.91 mm i.d. × ~3.18 mm o.d. × 5 cm long) connected to the mass spectrometer electrical ground. The inner tube was a fused-silica capillary (254 μm i.d. × 361 μm o.d. × ~24 cm long). Each tube was secured in a PEEK tee so that solvent could be delivered by an HPLC pump (HP 1090, Agilent Technologies, Santa Clara, CA, USA) into the annulus region of the two tubes where it flowed to the top of the vertically mounted probe. An HPLC column (250 mm × 4.6 mm, 3 μm C18; Imtakt USA, Philadelphia, PA, USA) in the flow path from the pump was used to stabilize the solvent flow. From there the solvent was aspirated down the inner capillary into the Turbo V™ ion source of an AB SCIEX TripleTOF 5600+ mass spectrometer
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
Laser ablation into liquid vortex capture probe
Figure 1. Vertically aligned transmission geometry laser ablation in combination with a non-contact liquid vortex capture probe coupled with ESI-MS. (a) Schematic illustration of complete experimental setup, (b) an enlarged view showing the details of the transmission geometry LA region including the sample mounting, the liquid vortex capture probe, and coupling to Turbo V™ ion source, and (c) topview photograph of the sampling probe showing the liquid vortex flow into the inner capillary.
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
probe were replaced by equivalent parts with larger internal diameters. The ESI emitter capillary used here had an i.d. of 150 μm. The i.d. of the nebulizer capillary was 530 μm. The position of the probe relative to the surface was adjusted using a LS-Series linear stage (Applied Scientific Instrumentation Inc.) on which the probe was mounted. The probe-to-sample distance was monitored using a 50 mm f/2.8 25.5 mm diameter zoom lens (Tamron, Saitama, Japan) connected to a SI-C400N CCD camera (Costar Inc., Anaheim, CA, USA).
wileyonlinelibrary.com/journal/rcm
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
1667
and electrosprayed. The nebulizing gas (nitrogen) in the ion source was used to adjust the solvent aspiration rate through the inner tube to roughly match the delivery rate of solvent to the probe from the HPLC system. In doing so, a stable vortex, or whirlpool, drain was maintained at the sampling end of the probe to capture into the liquid laser-ablated material from the sample mounted above the probe. To increase the accessible self-aspiration flow rate range of the system, the standard ESI emitter and nebulizer capillaries of the ESI
O. S. Ovchinnikova et al. Data collection for imaging The data collection procedure for the imaging experiments was similar to that we have utilized for other atmospheric pressure surface sampling/ionization techniques.[7,13,14] Briefly, at the beginning of an imaging experiment the probe was moved to a position ca 500 μm below the sample, and positioned centered on the vertical axis defined by the laser beam traversing down through the sample to the probe. The x-y sample stage was manually moved so that one corner of the area of interest to be imaged was moved into the path of the laser beam. After that, x-y stage movements were conducted under computer control. The first lane was scanned by moving the surface parallel to the x axis at a forward surface scan rate, with the laser triggered to fire at the beginning of the lane. At the end of the first lane the laser was turned off, and the surface moved to the beginning of the first lane. When the beginning of the first lane was reached, the surface was moved parallel to the y-axis to achieve the required lane spacing distance. Subsequent lanes were scanned in a similar fashion. Data for each lane scan was collected into individual data files. Movement of the stage was synchronized with laser firing and the corresponding mass spectral data by triggering the start of the data collection at the beginning of a lane scan using the stage control software (HandsFree Surface Analysis©).[12] Imaging data was collected in full scan and enhanced product ion mode. Mass spectrometric data files were converted into a TissueView© (AB Sciex) compatible format using another software developed in-house (TissueView Converter for WIFF Data©). Chemical images were rendered using the image visualization module of the TissueView© software package.
RESULTS AND DISCUSSION Basic operation
1668
A schematic illustration of the experimental setup used in this work is shown in Fig. 1(a). The LA system, sample and liquid vortex capture probe were placed on top of the mass spectrometer inside a light-tight, interlocked enclosure for safety. The grounded capture probe (detailed in Fig. 1(b)) was positioned under the sample in vertical alignment with the laser beam traveling down through the sample. The probe was constructed to be connected directly to the existing ESI emitter in the ion source of the mass spectrometer via a length of fused-silica capillary. This kept in place all the safety features and the optimized ion source components of the commercial system. Replacement of the standard nebulizer capillary and the emitter capillary with larger inner diameter parts supplied by the instrument manufacturer provided the ability to use the nebulizing gas and resulting venturi effect from the pneumatically assisted ESI source to aspirate the capture solvent through the probe at rates up to 200 μL/min. When this aspiration flow rate was matched by the solvent flow into the probe supplied by an HPLC pump, a stable vortex drain to capture the laser-ablated material was maintained at the sampling end of the probe (photograph in Fig. 1(c)). On the basis of the total calculated volume of the inner capillary of the probe and the ESI emitter capillary (15.8 μL), and the
wileyonlinelibrary.com/journal/rcm
maximum aspiration rate of 200 μL/min, the system flush time was relatively short (ca 4.7 s). This short transport time minimized sample dilution via diffusion during transport to the ion source. Nonetheless, the vortex action of the drain and finite transit time provided an opportunity for the laser-ablated material (small particles, clusters, and possibly gaseous material) to dissolve in the solvent before reaching the electrospray emitter. Optimization of probe-to-surface distance and laser spot-to-probe vertical alignment for maximum capture efficiency Compared with our previous horizontal design,[2] the present capture probe provided an order of magnitude larger liquid surface area (2.8 mm2 vs 0.28 mm2) and it was positioned vertically under the sample to let gravity help direct ablated material down towards the capture liquid. We expected and found that this design improved capture efficiency and also made less critical the exact on-axis LA spot-to-probe alignment as well as the probe-to-surface separation, thereby simplifying the setup and operation of the system. The effect of the probe-to-surface separation distance on the efficiency of laser-ablated material capture was determined by measuring the mass spectral signal for the cationic dye basic blue 7 (m/z 478) recorded when scanning five hand-drawn blue ink lines on glass through the laser beam (Fig. 2(a)). The highest average signal level for the five lines was observed with the probe positioned just below the surface (0.2 mm). However, this positioning was found to sometimes cause instability in the liquid vortex owing to interaction of the LA plasma with the liquid/air boundary causing signal drop-outs or splashing of the liquid on the sample. Between 0.3 and 0.8 mm, the signal level plateaued at about 95% of this maximum value. Beyond a spacing of 0.8 mm, the signal level began to drop significantly and the variation in the signal increased. At a spacing of 2.0 mm the signal levels were only 50% of the maximum value. On the basis of these data, all additional work was carried out with a probe-to-surface spacing of about 0.5 mm. This distance provided high capture efficiency and good signal reproducibility, but was far enough from the sample to avoid accidental or LA-induced splashing of the sample surface with the capture solvent. The effect of laser beam-to-probe vertical alignment on capture efficiency was determined by measuring the mass spectral signal for basic blue 7 (m/z 478) recorded from LA spots in a blue ink thin film with the probe at different positional offsets from the vertical center line (Fig. 2(b)). The plot in Fig. 2(b) shows that mass spectral signals were relatively consistent over a distance of about 1.9 mm, a distance roughly equivalent to the diameter of the liquid vortex in the probe. Beyond this distance, in either direction, the signal level rapidly declined. These data show that precise positioning of the probe center and the LA spot was not critical in this case. A positioning offset of as much as about 0.8 mm could be tolerated without a significant signal loss. The absolute efficiency of the LA capture process was estimated by comparing mass spectral signal intensities recorded under conditions optimized for LA liquid capture and those for a direct liquid extraction with the same probe. Comparison was made between the integrated signal intensities recorded for basic blue 7 generated by LA liquid
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
Laser ablation into liquid vortex capture probe
(a)
Rel. Area (%)
100
50
0 0.0
0.5
1.0
2.0
1.5
Probe-to-Surface Distance (mm)
(b)
Rel. Area (%)
100
50
1.9 mm i.d. of probe
0 -1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
Probe Offset from Vertical Center Line (mm)
Figure 2. Optimization of liquid vortex capture probe positioning. (a) Plot of normalized mass spectral peak area as a function of probe-to-surface distance obtained while scanning through the laser beam five ink lines containing basic blue 7 (m/z 478 → m/z 434, CE = 50 eV) that were drawn freehand on a glass microscope slide. The surface scan speed was 50 μm/s and the TOF accumulation time 250 ms (m/z 100–900). The probe was centered vertically with the laser beam. Error bars represent ±1 standard deviation calculated from the five replicate measurements. (b) Plot of normalized mass spectral peak area (m/z 478 → m/z 434) obtained as a function of vertical alignment position of the probe relative to the center line defined by the location of the fixed laser beam. At each probe position six replicate measurements from single laser shots were made at different locations in a blue ink line drawn on a glass microscope slide. Probe-to-surface distance was 0.5 mm. TOF accumulation time 250 ms (m/z 100–900). Error bars represent ±1 standard deviation calculated from the six replicate measurements. Capture solvent: methanol/0.1% by volume formic acid flowing at 200 μL/min. Laser: 355 nm, Nd:YAG, 10 Hz, 66 μm spot size, 100 μJ pulse energy measured on the sample side of the surface.
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
Samples and analytes applicable to analysis Successful analysis by this sampling and ionization strategy requires first that the sample can be examined in an upsidedown position. This is possible for many spotted samples, inked samples, thin films of materials such as polymers, or thaw mounted tissue sections. Furthermore, the support surface needs to be transparent to the laser light and the laser beam is required to ’punch through’ and ablate the sample from the surface. For ablation to take place the laser fluence must be sufficient, but the sample or a matrix of some type on the sample side of the support must absorb the particular
wileyonlinelibrary.com/journal/rcm
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
1669
capture from LA spots 44.5 μm in diameter and those by liquid extraction of inked rectangles (44.9 μm × 39.0 μm) etched from a thin-film layer of blue Sharpie marker. Normalizing for the small area differences in the absolute areas sampled, and subtracting out the signal levels recorded for direct liquid extraction from a slide from which the blue ink was completely removed by laser etching, the capture efficiency for LA into the liquid vortex probe was determined to be about 24% (Supplementary Fig. S1, see Supporting Information). This assumes 100% extraction efficiency in the case of the direct liquid extraction, and this is a reasonable assumption because the spot sampled (ca 50 μm square) was much smaller than the size of the liquid junction (at least 1.9 mm) and because the blue ink was very soluble in the methanol extraction solvent. This estimated capture efficiency was more than double the efficiency that we have reported with LA into a horizontally configured non-contact continuous flow LMJ-SSP/ESI emitter probe[2] or with LA into a suspended droplet.[15]
Figure 3. Spatially resolved sampling of spotted materials. (a) Full scan mass spectrum from polypropylene glycol (PPG) mixture containing PPG 425 (1 μL of 100 μM solution (50/49.8/0.1/0.1/0.027 v/v/v/v/v methanol/water/formic acid/acetonitrile/ ammonium acetate) = 0.1 nmol) spotted on a DIRECTOR® slide, where nx = [M + H]+, n*x = [M + Na]+, and n^x = [M + NH4]+. Capture solvent: 80/20/0.1 (v/v/v) methanol/water/formic acid at 200 μL/min (m/z 100–2000). (b) Full scan mass spectrum from bovine insulin side chain B (3494 Da, 1 μL of 300 μM solution (50/50/0.1 (v/v/v) methanol/water/formic acid) = 0.3 nmol) spotted on a DIRECTOR® slide. Capture solvent: 50/50/0.1 (v/v/v) acetonitrile/water/formic acid flowing at 200 μL/min (m/z 100–3000). (c) Full scan mass spectrum from horse heart cytochrome c (12360.2 Da, 1 μL of 81 μM solution (50/50/0.1 (v/v/v) methanol/water/formic acid) = 81 pmol) spotted on a DIRECTOR® slide. Capture solvent: 50/50/0.1 (v/v/v) methanol/water/formic acid flowing at 100 μL/min (m/z 100–2000). Each spectrum shown was acquired from a single laser shot resulting in a laser ablation spot approximately 50 μm in diameter. Each spectrum was independently normalized to the most abundant peak observed in the spectrum. TOF accumulation time was 1.0 s in all cases. Laser: 355 nm, Nd: YAG, 10 Hz, 100 μJ pulse energy on the surface.
O. S. Ovchinnikova et al. laser wavelength in use. With the 355 nm laser light used in this case, colored materials such as inks can act as their own matrix, but in doing so some photodamage to the sample can occur.[16] As a simple means of extending the range of material types applicable to analysis, we used as sample supports PEN 1.0 (polyethylene naphthalate) membrane coated slides and DIRECTOR® slides, both of which were designed for the preservation of fragile materials during transmission geometry laser capture microdissection. The use of UV-absorbing PEN polymer on glass to facilitate the gentle removal of material from a surface in transmission geometry LA is a well-studied phenomenon in material science and it is commonly used for the precise removal of material from surfaces.[17] The DIRECTOR® slides use a proprietary energy transfer coating bonded on the glass slide that converts the energy from the UV laser into kinetic energy, vaporizing the coating and thus causing the material on the surface to be propelled from it.[18] The final requirements for successful analysis are that the captured material dissolves in the capture solvent and that the material is amenable to analysis by ESI-MS. The spectra shown in Fig. 3 were each acquired from standard samples of different material types spotted and dried onto a DIRECTOR® slide. A laser ablation spot size of approximately 50 μm in each of the materials gave rise to the spectra shown. Figure 3(a) shows a portion of the full scan mass spectrum obtained from a sample of mixed polypropylene glycols (PPGs) spotted from a solution containing ammonium acetate. In the region of the spectrum shown, a distribution of polymers n-mers (from n = 5–11) was observed as a series of protonated molecules, sodiated adducts, and ammonium adducts. The spectrum from LA/liquid capture of bovine insulin shown in Fig. 3(b) was dominated by the two peaks corresponding to the [M + 3H]3+ and [M + 4H]4+ ions. The spectrum from the analysis of cytochrome c in Fig. 3(c) shows the expected charge state distribution, in this case extending from [M + 8H]8+ to [M + 20H]20+. Note that in each of the separate examples shown, a different capture solvent composition was used. The capture solvent can be chosen to optimize dissolution and ESI of the material of interest.
Imaging of a stamped ink grid pattern The quality of the chemical images and the limits of the achievable imaging resolution possible with the current system were evaluated using a blue ink grid pattern stamped on a DIRECTOR® slide. The technique of sampling was used to achieve a spatial sampling resolution smaller than the present laser spot size of 50 μm.[14,19,20] With oversampling, all the sample material of interest needs to be removed during each lane scan to ensure that only new material is accessed by each sequential lane step so that there are no signal contributions from previous lanes. Exploratory studies showed that lane steps sizes as small as 2.5 μm (defining the pixel size in the lane step direction) could be used and a good-quality mass spectral signal from inked lines was still obtained. Lane scan speeds (10 μm/s) and full scan mass spectral data acquisition rate (250 ms) were chosen to provide the same 2.5 μm pixel size in the lane scan direction. Figures 4(a) and 4(b) show the optical and mass spectral chemical images, respectively, of a portion of the grid pattern analyzed. The inked lines in the grid, as measured from the optical image, were approximately 6 μm wide while the ink-free spaces were about 16 μm square. Correlation between the optical and chemical images was good with both the stamped ink lines and the blank spaces clearly distinguishable. The image pixel size was 2.5 μm × 2.5 μm and the imaging resolution, based on the size of the smallest feature distinguished in both the optical and the chemical images (i.e., the grid lines), was estimated to be about 6 μm. This imaging resolution was 20 times better than our previous best reported results with laser ablation/liquid sample capture MSI.[2]
Imaging the distribution of a distinct lipid in a mouse brain thin-tissue section Mouse brain tissue provides a real sample that has optically distinguishable features on the 10s of micrometer scale as well as chemically specific distributions of lipids, among other molecular species, within those features.[21] Figure 5(a) shows the lipid-rich section (m/z 720–840) of
1670
Figure 4. Imaging of a stamped ink grid pattern. (a) Optical image of a blue ink grid pattern stamped on top of a DIRECTOR® slide. (b) Corresponding chemical image for m/z 478 (basic blue 7) extracted from the full scan mass spectra obtained in 51 lane scans with 2.5 μm spacing using a surface scan speed of 10 μm/s and a 250 ms TOF accumulation time (m/z 100–1500) (2.5 μm × 2.5 μm pixel size). The probe-to-surface distance was 0.5 mm. Capture solvent: methanol/0.1% by volume formic acid flowing at 200 μL/min. Laser: 355 nm, Nd:YAG, 10 Hz, 50 μm spot size, 100 μJ pulse energy on the surface.
wileyonlinelibrary.com/journal/rcm
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
Laser ablation into liquid vortex capture probe the full scan mass spectrum obtained from a single shot laser ablation (ca 50 μm laser ablation spot)/liquid capture analysis in a random area of the brain section.
Figure 5. Spot sampling of thin-tissue sections. (a) Full scan mass spectrum obtained from a single laser shot on a mouse brain thin-tissue section mounted on a PEN 1.0 (polyethylene naphthalate) membrane slide (m/z 150–1500, TOF accumulation time = 250 ms). (b) Product ion spectrum of the base peak ion observed in (a) (m/z 760.6 → ○, CE = 45 eV) obtained a single laser shot at a near-by but separate location on the mouse brain thin-tissue section (m/z 100–796, TOF accumulation time = 1.0 s). Capture solvent: methanol + 0.1% by volume formic acid flowing at 200 μL/min. The probe-to-surface distance was 0.5 mm. Laser: 355 nm, Nd:YAG, 10 Hz, 50 μm spot size, 100 μJ pulse energy on the surface.
The most abundant ion observed (m/z 760.6) was tentatively identified as being from a phosphatidylcholine (PC)-type lipid, possibly the protonated PC (34:1), on the basis of the measured mass and the product ion spectrum. The base peak in the product ion spectrum of this species corresponded to the characteristic choline head group ion (m/z 184.06, Fig. 5(b)). Imaging the distribution of this same lipid in a different, small sub-region of the same brain section was performed using enhanced product ion mode (i.e., m/z 760.6 → m/z 184.06) detection. Figure 6(a) shows the optical bright field image of the area imaged with distinct regions of the brain labeled and color coded for easy identification. The corresponding mass spectral chemical image shown in Fig. 6(b) was rendered from the abundance of m/z 184.06 measured in 151 separate lane scans spaced 20 μm apart. The surface scan rate of 50 μm/s and a data acquisition time of 250 ms resulted in an image pixel size of 12 μm × 20 μm. This larger pixel size relative to the images of the inked grids above was required to achieve signal levels suitable for quality images to be rendered. The chemical image shows that this targeted lipid, although present in both the arbor vitae and molecular and granular layers, was more highly concentrated in the arbor vitae part of the brain. This is consistent with other reports on the analysis of lipids in mouse brain by MS showing a different concentration between the white (arbor vitae) and grey matter (molecular and granular layers) regions in the mouse brain.[22] While the image pixel size was 12 μm × 20 μm, the image resolution, based on the smallest feature distinguished in both the optical and the chemical images, was estimated to be about 50 μm (see scale bars in Fig. 6).
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
wileyonlinelibrary.com/journal/rcm
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
1671
Figure 6. Imaging the distribution of a distinct lipid in a mouse brain thintissue section. (a) Optical image of mouse brain thin-tissue section mounted on a PEN 1.0 (polyethylene naphthalate) membrane coated slide. The colored regions highlight three anatomical regions found in the mouse brain section. (b) Corresponding chemical image generated from the ion intensity recorded for the transition m/z 760.6 → m/z 184.06, CE = 45 eV) acquired in enhanced product ion mode (m/z 150–200, TOF accumulation time = 250 ms). The image was rendered from 151 lane scans obtained with a surface scan speed of 50 μm/s and a lane spacing of 20 μm. The probe-tosurface distance was 0.5 mm. Capture solvent: methanol/0.1% by volume formic acid flowing at 200 μL/min. Laser: 355 nm, Nd:YAG, 10 Hz, 50 μm spot size, 100 μJ pulse energy on the surface.
O. S. Ovchinnikova et al.
CONCLUSIONS
1672
We have reported herein on the design and use of vertically aligned transmission geometry laser ablation in combination with a non-contact liquid vortex capture probe coupled with ESI for spatially resolved spot sampling and imaging with mass spectrometry. In comparison with our previous work with non-contact continuous flow liquid capture probes,[2] the large liquid surface of the present probe made less critical the exact on-axis laser ablation spot-to-probe alignment and surface-to-probe spacing, simplifying the setup and operation of the system. The estimated 24% capture efficiency of laser-ablated material was double what we had previously reported and was made possible by the large liquid surface area and with gravity to help direct ablated material vertically down towards the probe. The vortex action of the capture liquid may have contributed to both the enhanced capture efficiency and the dissolution of captured material. Although not discussed, this ablation ’down’ sampling geometry was also expected to limit contamination of the unanalyzed portion of the sample via deposition of ablated material back onto the surface compared with an ablation ’up’ geometry in either a transmission or a reflection configuration. To extend the range of material types applicable to analysis using a UV laser, we used as sample supports PEN 1.0 (polyethylene naphthalate) membrane coated slides and DIRECTOR® slides, both of which were designed for the preservation of fragile materials during transmission geometry laser capture microdissection. An imaging resolution of about 6 μm was estimated from the examination of both the optical and the mass spectral data images of stamped ink grids on a DIRECTOR® slide. This imaging resolution was 20 times better than that obtained in our previous best results with laser ablation/liquid sample capture mass spectrometry imaging. Using thin sections of brain tissue, the mass spectral image of a selected lipid was obtained with an estimated imaging resolution of about 50 μm. Future improvements in this system might include a smaller laser spot size (
Revised: 13 May 2014
Accepted: 13 May 2014
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673 (wileyonlinelibrary.com) DOI: 10.1002/rcm.6946
Transmission geometry laser ablation into a non-contact liquid vortex capture probe for mass spectrometry imaging Olga S. Ovchinnikova, Deepak Bhandari†, Matthias Lorenz and Gary J. Van Berkel* Organic and Biological Mass Spectrometry Group, Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6131, USA RATIONALE: Capture of material from a laser ablation plume into a continuous flow stream of solvent provides the means for uninterrupted sampling, transport and ionization of collected material for coupling with mass spectral analysis. Reported here is the use of vertically aligned transmission geometry laser ablation in combination with a new non-contact liquid vortex capture probe coupled with electrospray ionization for spot sampling and chemical imaging with mass spectrometry. METHODS: A vertically aligned continuous flow liquid vortex capture probe was positioned directly underneath a sample surface in a transmission geometry laser ablation (355 nm, 10 Hz, 7 ns pulse width) set up to capture into solution the ablated material. The outlet of the vortex probe was coupled to the Turbo V™ ion source of an AB SCIEX TripleTOF 5600+ mass spectrometer. System operation and performance metrics were tested using inked patterns and thin tissue sections. Glass slides and slides designed especially for laser capture microdissection, viz., DIRECTOR® slides and PEN 1.0 (polyethylene naphthalate) membrane slides, were used as sample substrates. RESULTS: The estimated capture efficiency of laser-ablated material was 24%, which was enabled by the use of a probe with large liquid surface area (~2.8 mm2) and with gravity to help direct ablated material vertically down towards the probe. The swirling vortex action of the liquid surface potentially enhanced capture and dissolution not only of particulates, but also of gaseous products of the laser ablation. The use of DIRECTOR® slides and PEN 1.0 (polyethylene naphthalate) membrane slides as sample substrates enabled effective ablation of a wide range of sample types (basic blue 7, polypropylene glycol, insulin and cyctochrome c) without photodamage using a UV laser. Imaging resolution of about 6 μm was demonstrated for stamped ink on DIRECTOR® slides based on the ability to distinguish features present both in the optical and in the chemical image. This imaging resolution was 20 times better than the previous best reported results with laser ablation/liquid sample capture mass spectrometry imaging. Using thin sections of brain tissue the chemical image of a selected lipid was obtained with an estimated imaging resolution of about 50 μm. CONCLUSIONS: A vertically aligned, transmission geometry laser ablation liquid vortex capture probe, electrospray ionization mass spectrometry system provides an effective means for spatially resolved spot sampling and imaging with mass spectrometry. Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
Recently, several groups have demonstrated that laser ablation (LA) with subsequent liquid capture of the ablated material is a valuable method for atmospheric pressure (AP) surface sampling/ionization and imaging with mass spectrometry (MS).[1–7] With this approach, material from a LA plume is collected into hanging droplets or a continuous flow stream of solvent. Sample collection is followed by continuous infusion, flow injection, or a post-sampling process such as high-performance liquid chromatography (HPLC)[1,4] or capillary electrophoresis[6] separation of the captured components, and then ionization and detection by mass spectrometry (MS). As an alternative, LA and capture * Correspondence to: G. J. Van Berkel, Organic and Biological Mass Spectrometry Group, Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6131, USA. E-mail: [email protected] †
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673 Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
1665
Current address: Environmental Health, Division of Laboratory Sciences, Centers for Disease Control and Prevention, Atlanta, GA 30341, USA.
directly into a formed probe-to-surface liquid microjunction (versus, e.g., a non-contact hanging droplet above the surface) has been shown to provide 100% sampling efficiency of surface material that is not directly soluble or that dissolves slowly in the extraction solvents on the time frame of the experiment.[7] The use of an AP secondary, or post-sampling, ionization process such as electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) with LA can sometimes be limited by ablation and formation of particulates rather than the desired gas-phase molecules.[8] A major fraction of these particles can be in a size range unsuitable for ionization by these techniques, given the limited gas-phase particulate and charged droplet/reagent ion interaction times (i.e., milliseconds).[9] Capture of laser-ablated material into an appropriate solvent droplet or flow stream provides a greater time period for the particulates or agglomerates to dissolve or otherwise dissociate forming molecular species suitable for ionization. Once the captured material is dissolved in solution, ESI, APCI, or other liquid introduction AP ion source might be used to efficiently ionize it.
O. S. Ovchinnikova et al. Our group has demonstrated that the capture of material from a LA plume into a continuous flow stream of solvent using a non-contact probe provides uninterrupted sampling, transportation and ionization of collected material, affording the means for mass spectrometry imaging (MSI).[2] We showed this possibility using horizontally aligned transmission geometry LA combined with a self-aspirating liquid microjunction surface sampling probe/electrospray ionization (LMJ-SSP/ESI) emitter.[10,11] The analysis of inked lines, letters and fingerprints on glass slides using a 210 μm laser spot size and oversampling demonstrated the ability to image with a spatial resolution approaching 100 μm, based on a comparison of the distinguishable features in the optical and chemical images of the surfaces analyzed. The LMJ-SSP/ESI emitter used in that work was an openair system that replaced the normal ion source on the mass spectrometer. The probe was built using a stainless steel tee, a 10-cm-long inner sampling/emitter capillary with a 254 μm o.d. and a 127 μm i.d., an outer tube on the sampling end with 635 μm o.d. and 327 μm i.d., and a nebulizer tube on the spray side. The small size of the liquid surface at the sampling end (ca 0.6 mm diameter) and volume of that probe (ca 1.3 μL) provided a rapid response time (3.8 s at 20 μL/min) and minimal sample dilution. However, the small size of the liquid surface area combined with the relatively large laser spot size required careful alignment for optimal capture of the ablated material. Even when optimized, we have shown that the basic geometry of operation provided an ablation plume capture efficiency of no more than 10%.[7] In this paper, we report on a vertically aligned transmission geometry LA setup combined with a non-contact liquid vortex capture probe and ESI-MS for spatially resolved spot sampling and MSI. Compared with our previous horizontal design, the capture efficiency was more than doubled by the use of a probe with an order of magnitude larger liquid surface area (2.8 mm2 versus 0.28 mm2) and with gravity to help direct ablated material vertically down towards the probe. The large liquid surface of the present probe also made the exact on-axis LA spot-to-probe alignment less critical, simplifying the operation of the system. In addition, the continuous flow liquid capture sampling probe was connected directly to the existing ion source of the mass spectrometer keeping in place all the safety features and the optimized ion source components of the commercial system. The use of DIRECTOR® and PEN 1.0 (polyethylene naphthalate) membrane slides, both designed for laser capture microdissection, as sample substrates permitted a broad range of materials to be ablated using a UV laser (355 nm) without photodamage. An MSI resolution of 6 μm was demonstrated using stamped ink patterns on glass and a resolution of about 50 μm was estimated when imaging a selected lipid in a thin-tissue section.
EXPERIMENTAL Chemicals and materials
1666
LC-MS Optima grade methanol, acetonitrile and water with 0.1% formic acid were purchased from Fisher Scientific (Pittsburg, PA, USA). Glass slides were purchased from VWR International (West Chester, PA, USA), PEN 1.0
wileyonlinelibrary.com/journal/rcm
(polyethylene naphthalate) membrane coated slides were from Leica (Buffalo Grove, IL, USA) and DIRECTOR® slides from Expression Pathology (Rockville, MD, USA). Horse heart cytochrome c and bovine insulin side chain B were purchased from Sigma-Aldrich (St. Louis, MO, USA). Polypropylene glycol (PPG) tune solution was acquired from AB Sciex (Framingham, MA, USA). Blue fine and standard point permanent markers (Sharpie®, Sanford, Oak Brook, IL, USA) containing the dye basic blue 7 were used to prepare ink layers and patterns on the slides. The stamped grid pattern was prepared on a DIRECTOR® slide by applying the blue Sharpie marker ink to one side of a TEM grid (G100HS fine square mesh with 19 μm hole openings and 6 μm wide lines, SPI Supplies, West Chester, PA, USA) and then pressing the grid onto a DIRECTOR® slide. Coronal mouse brain tissue sections (12-μm-thick) were thaw mounted on PEN 1.0 (polyethylene naphthalate) membrane slides. Optical images of the surfaces analyzed were obtained using a Nikon Biophot upright microscope (Tokyo, Japan) and associated software. Experimental setup The overall experimental setup is illustrated in Fig. 1(a). A Nd:YAG MiniLight II laser (10 Hz, 355 nm, 7 ns pulse width; Continuum, Santa Clara, CA, USA) was used for all experiments. A UV fused-silica right-angle prism was used to turn the beam 90° to be directly on-axis with the vertically mounted liquid vortex capture probe. Using a LP-1-XYZ 3-axis lens positioner (Newport, Irvine, CA, USA) and a Newport M-10X (10X) microscope objective, the laser beam was focused directly on the sample side of the support (Fig. 1(b)). The lens positioner assembly was mounted onto model 430 and 429 translation stages (Newport) to give x-y positioning control for laser alignment and for fine adjustment of the z-positioning of the objective. A SH05 beam shutter with a SC10 shutter controller (Thorlabs Inc., Newton, NJ, USA) was used to control the laser beam. The laser pulse energy was measured at the exit from the glass slides using a FieldMaxII laser power meter (Coherent, Santa Clara, CA, USA). Microscope slides were positioned in the slide holder of a MS2000 x-y robotic platform (Applied Scientific Instrumentation Inc., Eugene, OR, USA) that was used to manipulate the sample surface relative to the stationary laser beam. The HandsFree Surface Analysis© control software used to position the sample surface has been described in previous work.[12] Schematic details of the LA liquid capture probe used in these experiments are shown in Fig. 1(b). The probe used a co-axial tube design on the sampling end. The outer tube was a stainless steel tube (~1.91 mm i.d. × ~3.18 mm o.d. × 5 cm long) connected to the mass spectrometer electrical ground. The inner tube was a fused-silica capillary (254 μm i.d. × 361 μm o.d. × ~24 cm long). Each tube was secured in a PEEK tee so that solvent could be delivered by an HPLC pump (HP 1090, Agilent Technologies, Santa Clara, CA, USA) into the annulus region of the two tubes where it flowed to the top of the vertically mounted probe. An HPLC column (250 mm × 4.6 mm, 3 μm C18; Imtakt USA, Philadelphia, PA, USA) in the flow path from the pump was used to stabilize the solvent flow. From there the solvent was aspirated down the inner capillary into the Turbo V™ ion source of an AB SCIEX TripleTOF 5600+ mass spectrometer
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
Laser ablation into liquid vortex capture probe
Figure 1. Vertically aligned transmission geometry laser ablation in combination with a non-contact liquid vortex capture probe coupled with ESI-MS. (a) Schematic illustration of complete experimental setup, (b) an enlarged view showing the details of the transmission geometry LA region including the sample mounting, the liquid vortex capture probe, and coupling to Turbo V™ ion source, and (c) topview photograph of the sampling probe showing the liquid vortex flow into the inner capillary.
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
probe were replaced by equivalent parts with larger internal diameters. The ESI emitter capillary used here had an i.d. of 150 μm. The i.d. of the nebulizer capillary was 530 μm. The position of the probe relative to the surface was adjusted using a LS-Series linear stage (Applied Scientific Instrumentation Inc.) on which the probe was mounted. The probe-to-sample distance was monitored using a 50 mm f/2.8 25.5 mm diameter zoom lens (Tamron, Saitama, Japan) connected to a SI-C400N CCD camera (Costar Inc., Anaheim, CA, USA).
wileyonlinelibrary.com/journal/rcm
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
1667
and electrosprayed. The nebulizing gas (nitrogen) in the ion source was used to adjust the solvent aspiration rate through the inner tube to roughly match the delivery rate of solvent to the probe from the HPLC system. In doing so, a stable vortex, or whirlpool, drain was maintained at the sampling end of the probe to capture into the liquid laser-ablated material from the sample mounted above the probe. To increase the accessible self-aspiration flow rate range of the system, the standard ESI emitter and nebulizer capillaries of the ESI
O. S. Ovchinnikova et al. Data collection for imaging The data collection procedure for the imaging experiments was similar to that we have utilized for other atmospheric pressure surface sampling/ionization techniques.[7,13,14] Briefly, at the beginning of an imaging experiment the probe was moved to a position ca 500 μm below the sample, and positioned centered on the vertical axis defined by the laser beam traversing down through the sample to the probe. The x-y sample stage was manually moved so that one corner of the area of interest to be imaged was moved into the path of the laser beam. After that, x-y stage movements were conducted under computer control. The first lane was scanned by moving the surface parallel to the x axis at a forward surface scan rate, with the laser triggered to fire at the beginning of the lane. At the end of the first lane the laser was turned off, and the surface moved to the beginning of the first lane. When the beginning of the first lane was reached, the surface was moved parallel to the y-axis to achieve the required lane spacing distance. Subsequent lanes were scanned in a similar fashion. Data for each lane scan was collected into individual data files. Movement of the stage was synchronized with laser firing and the corresponding mass spectral data by triggering the start of the data collection at the beginning of a lane scan using the stage control software (HandsFree Surface Analysis©).[12] Imaging data was collected in full scan and enhanced product ion mode. Mass spectrometric data files were converted into a TissueView© (AB Sciex) compatible format using another software developed in-house (TissueView Converter for WIFF Data©). Chemical images were rendered using the image visualization module of the TissueView© software package.
RESULTS AND DISCUSSION Basic operation
1668
A schematic illustration of the experimental setup used in this work is shown in Fig. 1(a). The LA system, sample and liquid vortex capture probe were placed on top of the mass spectrometer inside a light-tight, interlocked enclosure for safety. The grounded capture probe (detailed in Fig. 1(b)) was positioned under the sample in vertical alignment with the laser beam traveling down through the sample. The probe was constructed to be connected directly to the existing ESI emitter in the ion source of the mass spectrometer via a length of fused-silica capillary. This kept in place all the safety features and the optimized ion source components of the commercial system. Replacement of the standard nebulizer capillary and the emitter capillary with larger inner diameter parts supplied by the instrument manufacturer provided the ability to use the nebulizing gas and resulting venturi effect from the pneumatically assisted ESI source to aspirate the capture solvent through the probe at rates up to 200 μL/min. When this aspiration flow rate was matched by the solvent flow into the probe supplied by an HPLC pump, a stable vortex drain to capture the laser-ablated material was maintained at the sampling end of the probe (photograph in Fig. 1(c)). On the basis of the total calculated volume of the inner capillary of the probe and the ESI emitter capillary (15.8 μL), and the
wileyonlinelibrary.com/journal/rcm
maximum aspiration rate of 200 μL/min, the system flush time was relatively short (ca 4.7 s). This short transport time minimized sample dilution via diffusion during transport to the ion source. Nonetheless, the vortex action of the drain and finite transit time provided an opportunity for the laser-ablated material (small particles, clusters, and possibly gaseous material) to dissolve in the solvent before reaching the electrospray emitter. Optimization of probe-to-surface distance and laser spot-to-probe vertical alignment for maximum capture efficiency Compared with our previous horizontal design,[2] the present capture probe provided an order of magnitude larger liquid surface area (2.8 mm2 vs 0.28 mm2) and it was positioned vertically under the sample to let gravity help direct ablated material down towards the capture liquid. We expected and found that this design improved capture efficiency and also made less critical the exact on-axis LA spot-to-probe alignment as well as the probe-to-surface separation, thereby simplifying the setup and operation of the system. The effect of the probe-to-surface separation distance on the efficiency of laser-ablated material capture was determined by measuring the mass spectral signal for the cationic dye basic blue 7 (m/z 478) recorded when scanning five hand-drawn blue ink lines on glass through the laser beam (Fig. 2(a)). The highest average signal level for the five lines was observed with the probe positioned just below the surface (0.2 mm). However, this positioning was found to sometimes cause instability in the liquid vortex owing to interaction of the LA plasma with the liquid/air boundary causing signal drop-outs or splashing of the liquid on the sample. Between 0.3 and 0.8 mm, the signal level plateaued at about 95% of this maximum value. Beyond a spacing of 0.8 mm, the signal level began to drop significantly and the variation in the signal increased. At a spacing of 2.0 mm the signal levels were only 50% of the maximum value. On the basis of these data, all additional work was carried out with a probe-to-surface spacing of about 0.5 mm. This distance provided high capture efficiency and good signal reproducibility, but was far enough from the sample to avoid accidental or LA-induced splashing of the sample surface with the capture solvent. The effect of laser beam-to-probe vertical alignment on capture efficiency was determined by measuring the mass spectral signal for basic blue 7 (m/z 478) recorded from LA spots in a blue ink thin film with the probe at different positional offsets from the vertical center line (Fig. 2(b)). The plot in Fig. 2(b) shows that mass spectral signals were relatively consistent over a distance of about 1.9 mm, a distance roughly equivalent to the diameter of the liquid vortex in the probe. Beyond this distance, in either direction, the signal level rapidly declined. These data show that precise positioning of the probe center and the LA spot was not critical in this case. A positioning offset of as much as about 0.8 mm could be tolerated without a significant signal loss. The absolute efficiency of the LA capture process was estimated by comparing mass spectral signal intensities recorded under conditions optimized for LA liquid capture and those for a direct liquid extraction with the same probe. Comparison was made between the integrated signal intensities recorded for basic blue 7 generated by LA liquid
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
Laser ablation into liquid vortex capture probe
(a)
Rel. Area (%)
100
50
0 0.0
0.5
1.0
2.0
1.5
Probe-to-Surface Distance (mm)
(b)
Rel. Area (%)
100
50
1.9 mm i.d. of probe
0 -1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
Probe Offset from Vertical Center Line (mm)
Figure 2. Optimization of liquid vortex capture probe positioning. (a) Plot of normalized mass spectral peak area as a function of probe-to-surface distance obtained while scanning through the laser beam five ink lines containing basic blue 7 (m/z 478 → m/z 434, CE = 50 eV) that were drawn freehand on a glass microscope slide. The surface scan speed was 50 μm/s and the TOF accumulation time 250 ms (m/z 100–900). The probe was centered vertically with the laser beam. Error bars represent ±1 standard deviation calculated from the five replicate measurements. (b) Plot of normalized mass spectral peak area (m/z 478 → m/z 434) obtained as a function of vertical alignment position of the probe relative to the center line defined by the location of the fixed laser beam. At each probe position six replicate measurements from single laser shots were made at different locations in a blue ink line drawn on a glass microscope slide. Probe-to-surface distance was 0.5 mm. TOF accumulation time 250 ms (m/z 100–900). Error bars represent ±1 standard deviation calculated from the six replicate measurements. Capture solvent: methanol/0.1% by volume formic acid flowing at 200 μL/min. Laser: 355 nm, Nd:YAG, 10 Hz, 66 μm spot size, 100 μJ pulse energy measured on the sample side of the surface.
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
Samples and analytes applicable to analysis Successful analysis by this sampling and ionization strategy requires first that the sample can be examined in an upsidedown position. This is possible for many spotted samples, inked samples, thin films of materials such as polymers, or thaw mounted tissue sections. Furthermore, the support surface needs to be transparent to the laser light and the laser beam is required to ’punch through’ and ablate the sample from the surface. For ablation to take place the laser fluence must be sufficient, but the sample or a matrix of some type on the sample side of the support must absorb the particular
wileyonlinelibrary.com/journal/rcm
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
1669
capture from LA spots 44.5 μm in diameter and those by liquid extraction of inked rectangles (44.9 μm × 39.0 μm) etched from a thin-film layer of blue Sharpie marker. Normalizing for the small area differences in the absolute areas sampled, and subtracting out the signal levels recorded for direct liquid extraction from a slide from which the blue ink was completely removed by laser etching, the capture efficiency for LA into the liquid vortex probe was determined to be about 24% (Supplementary Fig. S1, see Supporting Information). This assumes 100% extraction efficiency in the case of the direct liquid extraction, and this is a reasonable assumption because the spot sampled (ca 50 μm square) was much smaller than the size of the liquid junction (at least 1.9 mm) and because the blue ink was very soluble in the methanol extraction solvent. This estimated capture efficiency was more than double the efficiency that we have reported with LA into a horizontally configured non-contact continuous flow LMJ-SSP/ESI emitter probe[2] or with LA into a suspended droplet.[15]
Figure 3. Spatially resolved sampling of spotted materials. (a) Full scan mass spectrum from polypropylene glycol (PPG) mixture containing PPG 425 (1 μL of 100 μM solution (50/49.8/0.1/0.1/0.027 v/v/v/v/v methanol/water/formic acid/acetonitrile/ ammonium acetate) = 0.1 nmol) spotted on a DIRECTOR® slide, where nx = [M + H]+, n*x = [M + Na]+, and n^x = [M + NH4]+. Capture solvent: 80/20/0.1 (v/v/v) methanol/water/formic acid at 200 μL/min (m/z 100–2000). (b) Full scan mass spectrum from bovine insulin side chain B (3494 Da, 1 μL of 300 μM solution (50/50/0.1 (v/v/v) methanol/water/formic acid) = 0.3 nmol) spotted on a DIRECTOR® slide. Capture solvent: 50/50/0.1 (v/v/v) acetonitrile/water/formic acid flowing at 200 μL/min (m/z 100–3000). (c) Full scan mass spectrum from horse heart cytochrome c (12360.2 Da, 1 μL of 81 μM solution (50/50/0.1 (v/v/v) methanol/water/formic acid) = 81 pmol) spotted on a DIRECTOR® slide. Capture solvent: 50/50/0.1 (v/v/v) methanol/water/formic acid flowing at 100 μL/min (m/z 100–2000). Each spectrum shown was acquired from a single laser shot resulting in a laser ablation spot approximately 50 μm in diameter. Each spectrum was independently normalized to the most abundant peak observed in the spectrum. TOF accumulation time was 1.0 s in all cases. Laser: 355 nm, Nd: YAG, 10 Hz, 100 μJ pulse energy on the surface.
O. S. Ovchinnikova et al. laser wavelength in use. With the 355 nm laser light used in this case, colored materials such as inks can act as their own matrix, but in doing so some photodamage to the sample can occur.[16] As a simple means of extending the range of material types applicable to analysis, we used as sample supports PEN 1.0 (polyethylene naphthalate) membrane coated slides and DIRECTOR® slides, both of which were designed for the preservation of fragile materials during transmission geometry laser capture microdissection. The use of UV-absorbing PEN polymer on glass to facilitate the gentle removal of material from a surface in transmission geometry LA is a well-studied phenomenon in material science and it is commonly used for the precise removal of material from surfaces.[17] The DIRECTOR® slides use a proprietary energy transfer coating bonded on the glass slide that converts the energy from the UV laser into kinetic energy, vaporizing the coating and thus causing the material on the surface to be propelled from it.[18] The final requirements for successful analysis are that the captured material dissolves in the capture solvent and that the material is amenable to analysis by ESI-MS. The spectra shown in Fig. 3 were each acquired from standard samples of different material types spotted and dried onto a DIRECTOR® slide. A laser ablation spot size of approximately 50 μm in each of the materials gave rise to the spectra shown. Figure 3(a) shows a portion of the full scan mass spectrum obtained from a sample of mixed polypropylene glycols (PPGs) spotted from a solution containing ammonium acetate. In the region of the spectrum shown, a distribution of polymers n-mers (from n = 5–11) was observed as a series of protonated molecules, sodiated adducts, and ammonium adducts. The spectrum from LA/liquid capture of bovine insulin shown in Fig. 3(b) was dominated by the two peaks corresponding to the [M + 3H]3+ and [M + 4H]4+ ions. The spectrum from the analysis of cytochrome c in Fig. 3(c) shows the expected charge state distribution, in this case extending from [M + 8H]8+ to [M + 20H]20+. Note that in each of the separate examples shown, a different capture solvent composition was used. The capture solvent can be chosen to optimize dissolution and ESI of the material of interest.
Imaging of a stamped ink grid pattern The quality of the chemical images and the limits of the achievable imaging resolution possible with the current system were evaluated using a blue ink grid pattern stamped on a DIRECTOR® slide. The technique of sampling was used to achieve a spatial sampling resolution smaller than the present laser spot size of 50 μm.[14,19,20] With oversampling, all the sample material of interest needs to be removed during each lane scan to ensure that only new material is accessed by each sequential lane step so that there are no signal contributions from previous lanes. Exploratory studies showed that lane steps sizes as small as 2.5 μm (defining the pixel size in the lane step direction) could be used and a good-quality mass spectral signal from inked lines was still obtained. Lane scan speeds (10 μm/s) and full scan mass spectral data acquisition rate (250 ms) were chosen to provide the same 2.5 μm pixel size in the lane scan direction. Figures 4(a) and 4(b) show the optical and mass spectral chemical images, respectively, of a portion of the grid pattern analyzed. The inked lines in the grid, as measured from the optical image, were approximately 6 μm wide while the ink-free spaces were about 16 μm square. Correlation between the optical and chemical images was good with both the stamped ink lines and the blank spaces clearly distinguishable. The image pixel size was 2.5 μm × 2.5 μm and the imaging resolution, based on the size of the smallest feature distinguished in both the optical and the chemical images (i.e., the grid lines), was estimated to be about 6 μm. This imaging resolution was 20 times better than our previous best reported results with laser ablation/liquid sample capture MSI.[2]
Imaging the distribution of a distinct lipid in a mouse brain thin-tissue section Mouse brain tissue provides a real sample that has optically distinguishable features on the 10s of micrometer scale as well as chemically specific distributions of lipids, among other molecular species, within those features.[21] Figure 5(a) shows the lipid-rich section (m/z 720–840) of
1670
Figure 4. Imaging of a stamped ink grid pattern. (a) Optical image of a blue ink grid pattern stamped on top of a DIRECTOR® slide. (b) Corresponding chemical image for m/z 478 (basic blue 7) extracted from the full scan mass spectra obtained in 51 lane scans with 2.5 μm spacing using a surface scan speed of 10 μm/s and a 250 ms TOF accumulation time (m/z 100–1500) (2.5 μm × 2.5 μm pixel size). The probe-to-surface distance was 0.5 mm. Capture solvent: methanol/0.1% by volume formic acid flowing at 200 μL/min. Laser: 355 nm, Nd:YAG, 10 Hz, 50 μm spot size, 100 μJ pulse energy on the surface.
wileyonlinelibrary.com/journal/rcm
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
Laser ablation into liquid vortex capture probe the full scan mass spectrum obtained from a single shot laser ablation (ca 50 μm laser ablation spot)/liquid capture analysis in a random area of the brain section.
Figure 5. Spot sampling of thin-tissue sections. (a) Full scan mass spectrum obtained from a single laser shot on a mouse brain thin-tissue section mounted on a PEN 1.0 (polyethylene naphthalate) membrane slide (m/z 150–1500, TOF accumulation time = 250 ms). (b) Product ion spectrum of the base peak ion observed in (a) (m/z 760.6 → ○, CE = 45 eV) obtained a single laser shot at a near-by but separate location on the mouse brain thin-tissue section (m/z 100–796, TOF accumulation time = 1.0 s). Capture solvent: methanol + 0.1% by volume formic acid flowing at 200 μL/min. The probe-to-surface distance was 0.5 mm. Laser: 355 nm, Nd:YAG, 10 Hz, 50 μm spot size, 100 μJ pulse energy on the surface.
The most abundant ion observed (m/z 760.6) was tentatively identified as being from a phosphatidylcholine (PC)-type lipid, possibly the protonated PC (34:1), on the basis of the measured mass and the product ion spectrum. The base peak in the product ion spectrum of this species corresponded to the characteristic choline head group ion (m/z 184.06, Fig. 5(b)). Imaging the distribution of this same lipid in a different, small sub-region of the same brain section was performed using enhanced product ion mode (i.e., m/z 760.6 → m/z 184.06) detection. Figure 6(a) shows the optical bright field image of the area imaged with distinct regions of the brain labeled and color coded for easy identification. The corresponding mass spectral chemical image shown in Fig. 6(b) was rendered from the abundance of m/z 184.06 measured in 151 separate lane scans spaced 20 μm apart. The surface scan rate of 50 μm/s and a data acquisition time of 250 ms resulted in an image pixel size of 12 μm × 20 μm. This larger pixel size relative to the images of the inked grids above was required to achieve signal levels suitable for quality images to be rendered. The chemical image shows that this targeted lipid, although present in both the arbor vitae and molecular and granular layers, was more highly concentrated in the arbor vitae part of the brain. This is consistent with other reports on the analysis of lipids in mouse brain by MS showing a different concentration between the white (arbor vitae) and grey matter (molecular and granular layers) regions in the mouse brain.[22] While the image pixel size was 12 μm × 20 μm, the image resolution, based on the smallest feature distinguished in both the optical and the chemical images, was estimated to be about 50 μm (see scale bars in Fig. 6).
Rapid Commun. Mass Spectrom. 2014, 28, 1665–1673
wileyonlinelibrary.com/journal/rcm
Published in 2014. This article is a U.S. Government work and is in the public domain in the USA.
1671
Figure 6. Imaging the distribution of a distinct lipid in a mouse brain thintissue section. (a) Optical image of mouse brain thin-tissue section mounted on a PEN 1.0 (polyethylene naphthalate) membrane coated slide. The colored regions highlight three anatomical regions found in the mouse brain section. (b) Corresponding chemical image generated from the ion intensity recorded for the transition m/z 760.6 → m/z 184.06, CE = 45 eV) acquired in enhanced product ion mode (m/z 150–200, TOF accumulation time = 250 ms). The image was rendered from 151 lane scans obtained with a surface scan speed of 50 μm/s and a lane spacing of 20 μm. The probe-tosurface distance was 0.5 mm. Capture solvent: methanol/0.1% by volume formic acid flowing at 200 μL/min. Laser: 355 nm, Nd:YAG, 10 Hz, 50 μm spot size, 100 μJ pulse energy on the surface.
O. S. Ovchinnikova et al.
CONCLUSIONS
1672
We have reported herein on the design and use of vertically aligned transmission geometry laser ablation in combination with a non-contact liquid vortex capture probe coupled with ESI for spatially resolved spot sampling and imaging with mass spectrometry. In comparison with our previous work with non-contact continuous flow liquid capture probes,[2] the large liquid surface of the present probe made less critical the exact on-axis laser ablation spot-to-probe alignment and surface-to-probe spacing, simplifying the setup and operation of the system. The estimated 24% capture efficiency of laser-ablated material was double what we had previously reported and was made possible by the large liquid surface area and with gravity to help direct ablated material vertically down towards the probe. The vortex action of the capture liquid may have contributed to both the enhanced capture efficiency and the dissolution of captured material. Although not discussed, this ablation ’down’ sampling geometry was also expected to limit contamination of the unanalyzed portion of the sample via deposition of ablated material back onto the surface compared with an ablation ’up’ geometry in either a transmission or a reflection configuration. To extend the range of material types applicable to analysis using a UV laser, we used as sample supports PEN 1.0 (polyethylene naphthalate) membrane coated slides and DIRECTOR® slides, both of which were designed for the preservation of fragile materials during transmission geometry laser capture microdissection. An imaging resolution of about 6 μm was estimated from the examination of both the optical and the mass spectral data images of stamped ink grids on a DIRECTOR® slide. This imaging resolution was 20 times better than that obtained in our previous best results with laser ablation/liquid sample capture mass spectrometry imaging. Using thin sections of brain tissue, the mass spectral image of a selected lipid was obtained with an estimated imaging resolution of about 50 μm. Future improvements in this system might include a smaller laser spot size (
