We recently demonstrated that ultraviolet laser desorption ionization orthogonal time-of-flight mass spectrometry (UV-LDI o-TOF MS) could be used for the matrix-free analysis of cuticular lipids (unsaturated aliphatic and oxygen-containing hydrocarbons and triacylglycerides) directly from individual Drosophila melanogaster fruit flies (Yew, J. Y.; Dreisewerd, K.; Luftmann, H.; Pohlentz, G.; Kravitz, E. A., Curr. Biol. 2009, 19, 1245–1254). In this report, we show that the cuticular hydrocarbon, fatty acid, and triglyceride profiles of other insects and spiders can also be directly analyzed from intact body parts. Mandibular pheromones from the jaw of a queen honey bee are provided as one example. In addition, we describe analytical features and examine mechanisms underlying the methodology. Molecular ions of lipids can be generated by direct UV-LDI when non-endogenous compounds are applied to insect wings or other body parts. Current sensitivity limits are in the 10 pmol range. We show also the dependence of ion signal intensity on collisional cooling gas pressure in the ion source, laser wavelength (varied between 280–380 nm and set to 2.94 μm for infrared LDI), and laser pulse energy.
Key wordsUV-LDI MS FALDI Cuticular hydrocarbons Triglycerides Drosophila melanogaster Arthropods Pheromones
The standard method for the analysis of hydrocarbons (HCs) is gas chromatography coupled with mass spectrometry (GC/MS). This method has been used extensively to characterize lipids expressed on the cuticular surface of insects . Some of these exogenous lipids serve as communication cues (pheromones) and can greatly influence behaviors such as mate choice and aggregation . In addition, pheromones have been manipulated in order to control the population growth of agricultural pests and disease-bearing insects . Recently, we showed that ultraviolet laser desorption ionization orthogonal time-of-flight mass spectrometry (UV-LDI o-TOF MS) allows analysis of cuticular hydrocarbons and triglycerides directly from the external cuticle of individual fruit flies [3–5]. Direct UV-LDI MS provides spatially-resolved analysis of HCs without chemical extraction or the addition of a matrix to the cuticle. Moreover, low volatility, higher molecular weight polar compounds [e.g., long-chain aliphatic HCs, oxygen-containing HCs, fatty acids (FA), and triacylglycerides (TAGs)] are amenable to analysis. Using this method, several new oxygen-containing hydrocarbons and a novel sex pheromone were characterized from D. melanogaster . One disadvantage of the LDI MS method is that saturated aliphatic HCs (alkanes) are not detectable.
In this article, we show that chemical profiles of cuticles from other species of insects and spiders can also be analyzed by UV-LDI mass spectrometry. Furthermore, we demonstrate that lipids are detected with picomole sensitivity when exogenously applied to insect wings. Thus, the cuticle itself can be utilized as a substrate to facilitate UV-LDI MS analysis. To better characterize the mechanisms contributing to the phenomenon of fly-assisted laser desorption/ionization (FALDI), we investigated the influence of the laser wavelength, laser fluence, and buffer gas pressure in the ion source of the o-TOF mass spectrometer.
Z-11-hexadecen-1-ol (C16H32O), Z-9-hexadecenal (C16H30O), and Z-9-hexadecen-1-yl acetate (C18H34O2) were from ISCA Technologies (Riverside, CA, USA). All other commercial chemicals were from Sigma-Aldrich (München, Germany). Crude extracts of Drosophila melanogaster cuticle were prepared by placing 20 flies of both sexes in 200 μL of hexane or chloroform/methanol (2:1) for 10 min at room temperature. The supernatant was removed, evaporated using a gentle stream of N2, and stored at −20 °C. Samples were reconstituted in 20 μL of hexane or chloroform/ methanol (2:1) prior to analysis.
Drosophila melanogaster (Canton S) husbandry was performed as described . Bees were kind gifts from local bee keepers. Other arthropods were serendipitous findings; these animals were not classified with respect to sex and species.
2.3 Sample Preparation
Whole insects or parts of the animals were prepared on glass coverslips using adhesive (G304; Plano, Wetzlar, Germany). The coverslips were mounted into a milled-out (1.8 mm depth), custom-made sample plate holder using 200 μm thick sticky pads. FALDI MS analyses were performed with wings from D. melanogaster bees or wasps. To reduce the chemical background from endogenous compounds, fly wings were partially washed by soaking in hexane for 10 min. Analyte solutions, dissolved in hexane, were applied onto wings using pipettes. Standard matrix-assisted laser desorption ionization (MALDI) sample preparations were produced by mixing 2,5-dihydoxybenzoic acid (DHB) matrix (10 mg/mL) and analyte solutions on glass coverslips followed by air-drying.
2.4 UV/IR-MALDI-o-TOF Mass Spectrometer
The o-TOF mass spectrometer has been described in detail previously . This instrument is a custom-made hybrid mass spectrometer that consists of an oMALDI 2 ion source, as used in QSTAR mass spectrometers (AB SCIEX, Concord, CA, USA), and a prOTOF-like (Perkin Elmer, Waltham MA, USA) quadrupole transfer-mass analyzer unit. A few measurements were also made with a commercial MALDI QSTAR instrument (pulsar-I, AB SCIEX; see below). In addition to the standard N2 UV-laser (337ND-Sx; Spectra Physics, Mountain View, CA, USA; λ = 337 nm; τpulse ~3 ns; maximum repetition rate, 30 Hz), a Nd:YAG laser (Surelite II; Continuum, Darmstadt, Germany; λ = 532 nm; τpulse ~6 ns) pumped dye laser (FL 2001; Lambda Physik, Göttingen, Germany; repetition rate, 10 Hz) and an infrared (IR) Er:YAG laser (Speser; Spektrum Laser GmbH, Berlin, Germany; λ = 2.94 μm; τpulse ~150 ns; operated at 2 Hz) are available on the o-TOF instrument for UV- and IR-LDI MS, respectively. Using a set of dyes and a BBO crystal for second harmonic generation, the dye laser can be tuned continuously between 280 and 380 nm. In one experiment, a fundamental of the dye laser of 560 nm was employed for comparison. The emission wavelength was monitored using a StellarNet GREEN-Wave spectrometer (STE-GW-UV; Laser 2000, Wessling, Germany). The UV laser beams were delivered to the sample via an optical fiber system. A standard silica fiber with 200 μm core diameter and ~2 m length was used for delivery of the beam of the nitrogen laser ; for delivery of the dye laser beam, a second fiber with a core diameter of 200 μm that was 10 m in length (SFS200/220y, NA 0.12; Fiberguide Industries, Stirling, NY, USA) was used. The fiber end surfaces were imaged 1:1 onto the sample using a custom-made telescope (f1,2 = 15 mm, ϕlenses = 10 mm) producing an approximately “flat-top” intensity profile of about 200 × 230 μm2 in the focal plane ; the angle of incidence of the laser beam is ~30° relative to the ion optical axis. A custom-made opto-mechanical adjustment unit served for compensation of chromatic aberration (i.e., the wavelength-dependent focal lengths). Laser pulse energies were determined after the beam passed the glass fiber and the telescope unit using a commercial energy meter. After correction for reflection losses at the CaF2 vacuum window these values correspond to the laser energy supplied to the sample. A custom-made pyroelectric detector was furthermore used in combination with a beam splitter to monitor on-line the stability of the dye laser output.
The beam of the IR laser was focused onto the sample with a plano-convex infrasil lens to a spot of ~200–300 μm in diameter (1/e2-definition); the focal intensity profile of the IR laser beam was near-Gaussian; the angle of incidence relative to the ion optical axis was ~30°. The three lasers exhibited pulse-to-pulse energy variations (standard deviations) of ~2% (N2 laser), 2%–10% (dye laser, depending on the emission wavelength), and ~10%–20% (Er:YAG laser). Laser fluences were adjusted with dielectric attenuators (Er:YAG and dye laser) and a graduated neutral density filter (N2 laser). The ion extraction voltage applied between sample plate and extraction cone (the latter having a central aperture of 2 mm) was set to 45 V. Argon was used as buffer gas for collisional cooling and focusing [6, 8]. Pressure in the ion source and in the interface region (that houses the collisional focusing transfer quadrupole) was controlled via two gas inlets. The pressure in the ion source was monitored with a capacitive pressure gauge (DVR 5; Vaccubrand, Wertheim, Germany) as described previously . The pressure in the quadrupole ion guide was measured with a Pirani gauge. This pressure was kept between 1 × 10–3 and a few 10–2 mbar. After passing the differentially-pumped interface region, ions were accelerated orthogonally using a TOF voltage of 10 kV. Typically, a pusher plate frequency of 7.5 kHz and a bin width for the time-to-digital conversion of 132 ps were applied. Depending on the sample, a few hundred to a few thousand laser pulses were typically applied to accumulate a mass spectrum. Mass spectra were acquired using TOFMA software (developed by V. Spicer and W. Ens at the University of Manitoba, Winnipeg, CA). Data files were processed with MoverZ software (ver. 2001.02.13; Genomics Solutions, Ann Arbor, MI, USA). Mass resolution of the analysis was about 10,000 (FWHM) and mass accuracy better than 50 ppm when an external calibration was applied using ions of DHB matrix and a small peptide (e.g., bradykinin fragment 1–7). Using internal calibration mass accuracy could be improved to about 10 ppm.
Experiments with the QSTAR instrument allowed detection of HCs after adjusting the buffer gas pressure in the ion source to 4–5 mbar of argon and ion extraction conditions. The first adjustment required four modifications to the ion source: (1) replacing the standard ion extraction cone with a central aperture of 4 mm with one of 1.5 mm diameter; (2) shutting off a vacuum leak between ion source and quadrupole region (the leaks were intended for standard operation of the oMALDI 2 source); (3) closing a second leak in the curtain gas tube, and (4) adding an additional gas supply line for the Q0 quadrupole region. For HC analysis, “soft” ion extraction conditions were established by setting the “focusing potential” to not more than 20 V, “declustering potential 2” to values between 5 and 7 V, and the “quadrupole rod offset” to 6–14 V.
Elemental composition assignments of detected ions were made based on the assumptions that theoretical and experimental mass values correspond within 50 ppm (using external calibration) and that ions are comprised of C, O, H, Na, and K. For the queen bee, identification of mandibular pheromones was based on previously published literature (see text for references).
2.5 UV Absorption Spectroscopy
Absorption spectroscopy was performed with a 2101-PC spectrometer (Shimadzu, Duisburg, Germany). Whole wings from female D. melanogaster were attached to the surface of a UV quartz cuvette.
3 Results and Discussion
3.1 General Features of the UV-/IR-LDI MS Analysis of Endogenous Compounds
3.2 Influence of the Source Pressure
The buffer gas pressure in the ion source contributes significantly to the detection of HCs. While intense HC ion signals are recorded at elevated gas pressures above ~1.2 mbar (e.g., the spectra shown in Figure 1a and b were recorded at 2 mbar of argon), HC signals are not detected at pressures below 1 mbar (Figure 1c). More polar oxygenated HCs are occasionally detected with signal intensities that are on the order of the chemical background noise. Normally, pressure settings just below 1 mbar, along with an extraction voltage in the 50 V range, provide an optimum for standard MALDI o-TOF MS analysis of compounds like peptides and polar lipids [9, 10]. The general HC signal intensity–pressure dependence is displayed in Figure 1d, demonstrating a dramatic increase in HC ion yield above a threshold pressure around 1.2 mbar, before a saturation plateau is reached at ~1.6 mbar. In addition to argon, nitrogen and methane were also tested as buffer gases and found to provide qualitatively similar results. The data illustrate the importance of establishing sufficient collisional cooling [8, 11, 12] conditions for the stabilization of labile hydrocarbon-cation complexes. Notably, signal intensities of triglyceride ions are much less affected by the source pressure. These species are detected at pressures below 1 mbar (Figure 1c).
Tests with a commercial o-TOF instrument (QSTAR) allowed successful detection of HCs after adjusting the buffer gas pressure in the ion source to 4–5 mbar or argon and that of some ion extraction conditions (see the Experimental section). Before these modifications were made, only very low and less reproducible HC ion intensities were obtained for some major HC compounds (data not shown). In the absence of sufficient pressure conditions within the ion source, measurements with other o-TOF and axial instruments from different suppliers also produced only low intensity signals for the most abundant oxygen-containing HCs (data not shown).
3.3 Features of IR-LDI MS Versus UV-LDI MS
Analysis of insect cuticles with an Er:YAG IR laser for direct tissue analysis  did not provide advantages for the analysis of HC compounds (see Supplemental Figure S1). As with UV analysis, IR-LDI MS analysis of male flies at 2 mbar of argon produced signals corresponding to two major acetylated sex pheromones, 11-cis-vaccenyl acetate (cVA) and 3-O-acetyl-1,3-dihydroxy-octacosa-11,19(Z,Z)-diene (CH503; ). TAGs, FAs, and FA methyl esters (FAME) were also detected. Low intensity signals corresponding to oxygen-containing HCs and dienes were observed using both methods though overall signal intensity was higher using the UV laser.
3.4 Spatial Resolution
Assuming that ions are emitted solely from irradiated areas, the spatial resolution of the UV-LDI MS analysis is approximately 200 μm, corresponding to the laser spot diameter . In experiments where the focal laser spot was reduced to a diameter of 100 μm by utilizing a second optical fiber with smaller core diameter, no noticeable loss of sensitivity was observed (data not shown). We assume that a spatial resolution in the low 10 μm range could be achieved if further modifications were made (e.g., by employing fibers with core diameters in the 10 μm-range ). This could allow spatially-resolved probing of fine features of the insect body like glands. Automated MS imaging of tissue is not possible in the current o-TOF mass spectrometer system. Nevertheless, profiling by UV-LDI MS is able to detect distinct chemical profiles from different anatomical regions. For example, the anogenital region of male flies shows high local expression of sex pheromones that is distinct from other regions on the fly body . In the future, fine-scale MS imaging experiments may allow us to detect microdomains of specialized HC expression.
3.5 Detection of Endogenous Hydrocarbons and Triglycerides from Other Arthropods
Interestingly, some insects yielded sparse mass spectra, displaying only low HC ion abundances. For example, UV-LDI mass spectra acquired from worker bees displayed only few signals (data not shown). The likely reason for this observation is that the cuticular HC composition of this species is mostly comprised of alkanes , compounds that are not ionized under UV- or IR-LDI conditions. Queen bees express a richer chemical profile near pheromone-producing glands in the mandible and abdomen . Accordingly, signals corresponding to three fatty acids comprising part of the queen mandibular pheromone (QMP)  are detected from the jaw of a queen bee: oxodecenoic acid, hydroxydecenoic acid, hydroxydecanoic acid (Figure 3a; Supplemental Table S1). Small amounts of hydroxydecenoic acid are also found when the head of a worker bee was analyzed (data not shown). Aromatic components hydroxymethoxyphenylethanol (Figure 3a, inset) and methyl-p-hydroxybenzoate (theor. m/z 191.01; not displayed in Figure 3a), two other major components of the QMP, are detected on the queen bee mandible with low intensity, in addition to geranic and/or isobaric nerolic acid, known pheromones involved in bee aggregation behavior . Other signals (e.g., a series of ions between 400 and 500 Da exhibiting low mass increments) have so far not been identified.
3.6 Fly-Assisted Laser Desorption Ionization (FALDI) of Exogenous Compounds
Figure 4b shows that crude cuticular extract from Drosophila can also be analyzed when spotted onto wings of other insects. All major classes of lipids detected by UV-LDI MS endogenously on Drosophila (mono- and di-saturated hydrocarbons, oxygen-containing HCs, and triglycerides) were observed after application of crude fly extract on a wasp wing. In addition to providing a larger surface area, bee and wasp wings are particularly advantageous for facilitating FALDI MS analysis since these parts generally produce little endogenous signal and do not require washing. As with detection of endogenous compounds by direct UV-LDI MS, successful FALDI MS analysis of HCs requires elevated pressure settings.
3.7 Comparison with MALDI MS
When cuticular extract from the equivalent of 1 fly was analyzed using DHB as a MALDI matrix, [M + K]+ signals of two of the most abundant oxygen-containing HCs could be detected but with at least 10-fold lower intensity compared to analysis with an insect wing (data not shown). Monoenes and dienes without oxygen atoms were not detected. When the HC standards were analyzed by MALDI MS, compounds were detected as sodium adduct ions, albeit with intensities that were at least a factor of 10 lower than those of the [M + K]+ signals recorded using fly wings (data not shown).
MALDI MS analysis of cuticular HCs from extract has previously been demonstrated with lithiated dihydroxybenzoate matrix using a standard “axial”-MALDI-TOF mass spectrometer operated with high vacuum in the ion source . Notably, this approach allowed the analysis of saturated HCs, detected in form of lithium adducts, albeit with sensitivity limits that appear to fall into the low μg-range . Using the same matrix for “MALDI imaging” of Drosophila wings the two major dienes, hexacosadiene and nonacosadiene, found in female fly cuticle appear to be detected just above noise level . Other laser-based approaches utilize adduct formation of HC molecules with silver ions [23, 24]. However, the occurrence of two natural silver isotopes at 107 and 109 u and their tendency to form oligomers upon laser desorption can produce rather complex mass spectra and obscure resolution of HCs differing by two hydrogen atoms.
3.8 FALDI MS of Fatty Acids, Fatty Acid Methyl Esters, and Triglycerides
We tested whether FALDI MS is suitable for the analysis of FAs, FAMEs, and TAGs and examined the dependence of signal intensities on buffer gas pressure values. Signals corresponding to ions from synthetic octadecenoic acid and octadecenoic acid methyl ester were detected using insect wing as a substrate down to amounts in the low 10 pmol range (Supplemental Figure S3). The limit of detection is approximately 10 pmol for the FA and FAME standards (Supplemental Figure S3, inset). At 0.6 mbar, signal intensities were 6–8 times lower (data not shown). Mass spectra acquired from MALDI preparations with DHB matrix showed signals for these two compounds as well, though the sensitivity limit was at 200 pmol in this case (data not shown). TAGs can also be detected using FALDI. A drop of commercial olive oil applied to a bee wing produced a spectrum dominated by a distribution of triglycerides, the most abundant constituents in the oil (Figure 4c).
3.9 FALDI MS of Peptides
Small peptides (approximately up to 1 μmol were applied to wings) were also tested as analytes. Approximately up to 1 μmol were applied to wings and did not produce any detectable signal at any pressure setting. It is likely that there is a different interaction of these hydrophilic analytes with the hydrophobic surface and different ionization pathways. While lipids are detected as alkali-adducts, peptides are most typically detected by MALDI MS as protonated species [M + H]+. The lacking proton donor in the matrix-less LDI preparation may account for the failure to analyze peptides by FALDI MS. However, further factors must be involved since also [M+Na]+ or [M+K]+ signals of the peptides are not detected.
3.10 Possible Use of FALDI MS as an Analytical Method
Chemical matrices such as DHB and 4-α-cyano-hydroxycinnamic acid (CHCA) are commonly used for the analysis of polar lipids  and other biomolecules by MALDI MS. While small molecules and, in particular, metabolites can be analyzed by MALDI MS [26, 27], the presence of matrix-associated signals in the lower mass region of the spectra can heavily interfere with analyte-derived signals of small molecules. To circumvent this problem, a number of “matrix-free” preparation protocols have been developed. For example, etched silica surfaces for desorption/ionization on silicon (DIOS ) and nanowire-assisted LDI (NALDI ) have been introduced, to name only a few. These materials share a strong optical absorption at the excitation laser wavelength and provide a large specific surface area. However, these substrates are relatively costly and require care regarding proper storage and handling in order to prevent contamination by adsorbing compounds. Insect wings could provide a biomimetic model for the production of synthetic materials exhibiting the analytical features seen in these initial experiments.
3.11 Signal Intensities of Endogenous Hydrocarbon Ions as a Function of Laser Wavelength and Pulse Energy
To better understand the physicochemical mechanisms underlying the UV-LDI MS analysis of fly cuticle, we studied the contribution of laser wavelength and optical properties of the cuticle. The dorsal surfaces of female flies were irradiated and the excitation laser wavelength varied between 280 and 380 nm. “Threshold” pulse energies necessary to provoke an overall ion count/se of 1000 ± 300 were recorded as a function of wavelength in order to identify whether a specific endogenously-expressed chromophore may facilitate deposition of the laser pulse energy and desorption of cuticular compounds. If this were the case, enhanced signal intensities are likely to be observed within a limited wavelength range, corresponding to the peak absorption of the chromophore. Hydrocarbon and triglycerides are nonabsorbing, i.e., transparent throughout the probed laser wavelength range.
A few wavelength-dependent fine features were notable. For example, for wavelengths below 300 nm, the cuticle appears to be disrupted by fewer applied laser pulses, giving rise to an earlier onset of the detection of subsurface TAGs. This observation might be attributed to an enhanced photochemistry at high photon energies. The best results in terms of the reproducibility, ease of spectrum generation, and minimal destruction to the cuticles were produced with wavelengths of 340 nm and higher. Despite similar “threshold” pulse energies of approximately 4–5 μJ at 350–380 nm and 280–300 nm, no detectable damage to fly cuticles was observed for the longer wavelengths.
In one experiment, we also tested a fundamental wavelength of the dye laser at 560 nm. Desorption/ionization of HC and TAG ions was possible with this visible radiation but required a threshold pulse energy of ~66 μJ, about a factor of 10 higher than in the probed UV region. The elevated threshold is likely to be attributed to a reduced absorption of laser light by the cuticle.
3.12 Optical Properties of Fly Cuticle
Iridescent coloring of insect cuticle and wings is a well-studied phenomenon not only because of its intrinsic aesthetic appeal but also because of applications to material science and detection technology that may arise from understanding the underlying optical processes. Insect cuticle consists of lipids, phenolic compounds, and chitin filaments that are embedded in a protein matrix [31–33]. Depending on the species, coloring can be caused by chromophores and/or by “structural coloring” imparted from microstructures . For example, pigmentation in flies, bees, and wasps is attributed to the presence of aromatic compounds heavily involved in cross-linking of the extracellular matrix, as well as to the occurrence of melanin in black-colored cuticle . In the case of moth and butterfly wings, electron microscopy has revealed that cuticular structures are arrayed in multilayers or contain photonic crystals , both of which produce optical interference and can contribute to iridescent coloring.
To begin to investigate the optical properties of fly cuticle, we compared the light transmission properties of dissected fly wings to those of crude cuticle extract. An absorption spectrum recorded from whole Drosophila wings in transmission mode is displayed in Figure 5b. The data reveal a local maximum in the optical absorption at approximately 280 nm, likely corresponding to the peak absorption of small aromatic molecules like phenols in the cuticle and to aromatic tryptophan/tyrosine residues in proteins. The data were not corrected for scattering which, in addition to absorption by melanin, is likely to partially account for the monotonous increase in absorption with decreasing wavelength. One should note, however, that the optical properties of uncolored wings may only partially reflect those of the pigmented black and tan-colored cuticle. Cuticular chloroform/methanol extract of whole Drosophila displayed a blue-shift of the peak absorption to 272 nm (dashed line in Figure 5b). The different absorption curves of fly wing and chloroform/methanol extract indicate that only some chromophores were extracted. In particular, compounds cross-linked to proteins are unlikely to be dissolved. The possibility that cross-linked compounds could contribute to the optical properties of cuticle is supported by observations that wings that are sonicated for 30 min in methanol, placed in boiling water for 5 min, or microwaved for several minutes still retain FALDI properties (data not shown). Comparison of spectroscopic data with the wavelength dependence of the MS threshold pulse energies shows some correlation for the short wavelength side in that the increased optical absorption below ~300 nm (Figure 5a) appears to be reflected by a decreased energy required to reach the ion count threshold. However, the fact that almost identical threshold pulse energies are observed between 350 and 380 nm is not reflected in the spectroscopic data. Evidence for the contribution of structural coloring, which presumably would lead to more pronounced absorption-wavelength modulations, is not found in the UV or in the visible wavelength range (data not shown).
Future experiments will seek to identify the molecules found on the cuticle that serve as a matrix. In addition, it will be interesting to characterize the wavelength-dependent MS properties of insect wings that exhibit notable structural coloring (e.g., butterfly wings) and to examine whether structural properties may also enhance MS-related attributes.
3.13 Ion Signal Intensity as a Function of Laser Pulse Energy
A second parameter that could reveal fine features underlying FALDI processes is the dependence of HC signal intensity on laser pulse energy. To examine this property, the dorsal cuticles of female flies were irradiated using a N2 laser (λ = 337 nm) at various pulse energies above ion generation threshold and the intensity of signals corresponding to two prominent female HC compounds, hexacosadiene and nonacosadiene, were evaluated (Figure 5c). HC signals display a strong exponential increase with laser pulse energy that can be fit to an empirical function of the form \(Y = E_p^m\), where Y is the ion yield, E p the laser pulse energy, and m a fit parameter. A best fit is obtained with m ~10. Notably, essentially the same ion yield-pulse energy dependencies are found in UV-MALDI experiments. Data analysis, including desorbed neutral molecules, showed that this dependence can be attributed to a quasi-thermal evaporation of molecules from near-surface elements convoluted with a gas phase ionization process [7, 15].
3.14 Desorption Ionization Mechanisms
Taking the different experimental findings together, we hypothesize that thermal processes are likely to dominate the generation of ions of endogenous and exogenous lipid compounds from insect cuticle by pulsed UV laser desorption/ionization, facilitated by absorption of laser energy by endogenously expressed chromophores. Collisions of neutral HCs and alkali cations in the expanding material plume probably form the basis for the necessary ionization step . These labile cation-bound complexes are stabilized by rapid collisional cooling in the elevated pressure ion source.
3.15 Comparison to GC, ESI, and DART MS
Compared with GC/MS, the standard method for the analysis of cuticular HCs, UV-LDI MS analysis extends the accessible mass range and allows detection of long-chain and more polar HCs such as of hydroxylated and acetylated HCs . Due to the only ns-long excitation laser pulse, gaseous molecules are desorbed as intact cationized molecular ions. It is likely that these compounds would fragment upon evaporation under the quasi-continuous heating conditions found in a GC oven. In contrast to GC/MS, saturated HCs cannot currently be detected by UV-LDI MS. Ultimately, a combination of GC/MS and UV-LDI MS analyses would provide a more comprehensive picture of the cuticular HC profile of insects and could enable new insights into chemical communication and behavior.
Many of the compounds that are detected by direct UV-LDI mass spectrometry of fly cuticle can also be analyzed from solution using electrospray ionization (ESI) mass spectrometry and by direct desorption in real time (DART) MS [3, 38]. One advantage of these methods is that because protonated molecules can be produced, collisional induced dissociation (CID) can more readily be employed for structural elucidation and detection of functional groups like hydroxyls and acetates in oxygen-containing HCs. In addition, emission of ionic cuticular compounds was recently also observed from live flies when exposed at atmospheric pressure (AP) to a strong electrical field applied between the insects and the inlet cone of an ESI/AP-MALDI mass spectrometer . Future work will address the possibility of whether UV-LDI at AP conditions can be used for the analysis of live flies.
The authors gratefully acknowledge Alexandre Loboda for excellent technical support, Clemens Pöppelmann and Helge Karch for providing worker and queen bee samples, Malcolm Clench and Gottfried Pohlentz for helpful discussions, Edward Kravitz, Franz Hillenkamp, Stefan Berkenkamp, and Sequenom GmbH for support of the project. J.Y.Y. acknowledges a research fellowship by the Alexander von Humboldt foundation and a short-term fellowship by the Human Science Frontier Program Organization. The authors further acknowledge support for this work by the fund Innovative Medical Research of the University of Münster Medical School (grants DR520805 and DR511003), the Singapore National Research Foundation, and the Deutsche Forschungsgemeinschaft (grant DR 416/8-1).