Anal Bioanal Chem DOI 10.1007/s00216-015-8499-3
RESEARCH PAPER
Fragmentation studies for the structural characterization of marine dissolved organic matter Nuria Cortés-Francisco & Josep Caixach
Received: 6 October 2014 / Revised: 28 December 2014 / Accepted: 15 January 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract High-resolution tandem mass spectrometry by collision-induced dissociation with a linear ion trap-Orbitrap has been performed on marine dissolved organic matter (DOM). Product ion spectra of selected precursor ions (m/z 359–375) have been acquired to obtain structural information, after method development. To evaluate the performance of the method, the Suwannee River fulvic acid (SRFA) reference standard was also analyzed. By reconstructing the individual product ion spectrum of marine DOM, several fragments were assigned to the different precursor ions indicating the presence of carboxyl, hydroxyl, lactones, quinones, esters, and structures more similar to lignin-degraded molecules. On the basis of these findings, coastal marine DOM molecules, although structurally homogeneous, might be more rich in diversity of functional groups than previously described. Keywords High-resolution mass spectrometry . Marine DOM . Product ion spectra . Fragmentation studies . Orbitrap
Introduction The application of electrospray ionization high-resolution mass spectrometry has transformed the comprehension of dissolved organic matter (DOM) at molecular level. Several studies have assigned molecular formulas to DOM extracts from different origins, providing information on natural or affected Electronic supplementary material The online version of this article (doi:10.1007/s00216-015-8499-3) contains supplementary material, which is available to authorized users. N. Cortés-Francisco : J. Caixach (*) Mass Spectrometry Laboratory/Organic Pollutants, IDAEA-CSIC, Jordi Girona 18, 08034 Barcelona, Spain e-mail: [email protected]
environments and degradation/transformation processes. However, DOM is very complex and each elemental formula could represent millions or more constitutional isomers depending on the mass [1]. Nowadays, structural information from DOM is still missing, due to the fact that DOM separation into individual compounds by conventional techniques such as liquid chromatography or electrophoresis is impossible. Several attempts have been done to obtain structural information using different techniques such as nuclear magnetic resonance (NMR), Fourier transform-infrared spectroscopy (FT-IR), and tandem mass spectrometry. For instance, NMR analysis has revealed that phenolic carbons are not the major components of humic substances [2] and FT-IR analysis has shown presence of aromatic, aliphatic, and carboxylic groups [3]. Mass spectrometry has also been successfully applied. Low-resolution tandem mass spectrometry has been applied in most of the studies, and it has been used to analyze the Suwannee River fulvic acid (SRFA) standard, deep ocean DOM, and freshwater fulvic acids. Few studies have used high-resolution tandem mass spectrometry for the characterization of SRFA and organosulfates in particulate organic matter (see Table 1) [4–13]. The presence of carboxyl and hydroxyl groups has been confirmed by several of these studies. Tandem mass spectrometry has been successfully applied to elucidate the structure of some of the molecular formulas assigned, and some tentative structures have been proposed. In this respect, the introduction of hybrid linear ion trap-Orbitrap has provided a new approach for structural characterization of DOM with enhanced sensitivity and resolution, thus providing improved selectivity and mass measurement accuracy for tandem mass spectrometry data. The purpose of this study was to develop a method to acquire product ion spectra of marine DOM to obtain structural information, which is still missing. The method has been based on the use of high-resolution tandem
N. Cortés-Francisco, J. Caixach Table 1 Previous studies (from the last 15 years) where fragmentation studies of DOM from different origins have been performed Dissolved organic matter analyzed
References
Deep ocean DOM
Reemtsma et al. [4]
Suwannee River standards (IHSS)
Stenson et al. [5, 6]
details on the extraction procedure performance can be found elsewhere [14]. LTQ Orbitrap mass spectrometry
Acidic metabolites in fulvic acids from groundwater Jobelius et al. [7] Fulvic acids from different origins
Plancque et al. [8]
SRFA standard (IHSS)
Witt et al. [9]
River and ocean DOM
Liu et al. [10]
SRFA (IHSS)
Leenher et al. [11]
Standard soil and peat fulvic acid standards (IHSS) McIntyre et al. [12] Organosulfates in aerosols
Lin et al. [13]
mass spectrometry by collision-induced dissociation (CID) experiments with an LTQ Orbitrap. High resolved product ion spectra and good mass accuracies in accurate mass measurements have been used to find new functionalities of marine DOM.
Materials and methods Chemicals All reagents were of analytical or high-performance liquid chromatographic grade. Dichloromethane, methanol, and hydrochloric acid were purchased from Merck (Darmstadt, Germany). Isopropyl alcohol was from Carlo Erba (Milan, Italy) and formic acid was from Panreac (Barcelona, Spain). Highpurity water produced with a Milli-Q Organex-Q System Millipore (Millipore Corp., Bedford, MA) was used. Suwannee River fulvic acid (SRFA) (1S101F) standard from International Humic Substances Society (MN, USA) has been diluted in methanol and analyzed to evaluate the performance of the method.
Sample collection and preparation A composite water sample from the seawater intake system of a pilot desalination plant installed in the coast of Barcelona (Spain) was collected in Pyrex borosilicate amber glass bottles. The seawater has been previously characterized as part of a whole study [14]. Briefly, 1.5 L from the sample was acidified with 10 % hydrochloric acid to pH 2 and extracted with 2×100 mL of dichloromethane/isopropyl alcohol (90:10 v/v) and the extract was concentrated down to 250 μL at 40 °C under nitrogen. It should be highlighted that as any other isolation method, fractionation of DOM may occur. Further
The analysis of the standard and the sample was carried out with an LTQ Orbitrap XL (Thermo Fisher Scientific, Bremen, Germany) equipped with a nano-ESI source (Nanospray Flex Ion Source, Thermo Fisher Scientific, Bremen, Germany). Optimization was carried out analyzing an SRFA standard by infusion in the Orbitrap in full-scan negative ionization mode. Optimized parameters were capillary voltage, skimmer voltage, tube lens voltage, and capillary temperature. Finally, all mass spectra were acquired in negative ionization mode produced by capillary voltage −35 V, tube lens −90 V, and capillary temperature 300 °C. The automatic gain control was used to consistently fully fill the C-trap and gain mass accuracy and resolution [15]. High resolution defined as R 100,000 (m/z 400, full width at half maximum) was set. For tandem mass experiments, ions in the mass range between m/z 359 and 375 (Fig. 1) have been selected in the LTQ before they were pulsed through the trap to the collision cell where they were fragmented by CID, using increasing voltage increments up to 30 eV. It was only possible to isolate ions in the 0.7-Da mass window, and therefore, tandem mass experiments include several precursor ions. The same m/z were selected for both SRFA standard and marine DOM extracts. Post-acquisition calibration had to be performed using reference homologous series described in DOM before by the RecalOffline Xcalibur application. The accuracy of the precursor and product ions was always better than 2 ppm. Data analysis The mass peaks were exported to peak lists and from these lists feasible elemental formulas were generated. Different restrictive criteria were set to generate reliable elemental formulas, depending on the precursor ion selected (Table 2). This way, the possible assignments for each product ion were not so numerous and the assignation of the product ions was more easy and reliable. The molecular formula calculation was performed with Xcalibur 2.1 (Thermo Fisher Scientific, Bremen, Germany), and the posterior analysis of the data was done using our own developed Excel macros. Data filtering of the assigned formulas was done by applying exact mass differences/neutral losses (Table 3), to correlate each precursor ion with the possible ion fragments. The methodology developed for fragment assignment relies on a not broken fragmentation pathway, where the differences between fragment ions could be attributable to a known neutral loss or functional group fragmentation.
Structural characterization of marine dissolved organic matter Fig. 1 Suwannee River fulvic acid full-scan spectrum of the precursor ions chosen for the posterior fragmentation experiments
369.11918
100
367.10384
90
Relative Abundance
80
365.08799
70
363.07240 371.09873
60
361.05693
50 40 30
359.04097
373.05672
375.03601
20 368.07084
10 362.05152
364.07559
370.12274
366.12801
372.10196
374.04691
0 360
362
364
366
368
370
372
374
m/z
In silico fragmentation In order to build the puzzle obtained from the tandem experiments for the marine DOM and the SRFA, the different HR product ion spectra obtained after data filtering were used in MetFrag application [16] to relate the neutral losses and product ions obtained to a known structure from databases. The search was performed in PubChem and ChemSpider databases.
Results and discussion Method development In addition to optimizing the experimental parameters for efficient DOM ionization, in the development of the method, the collision energy for each precursor ion had to be optimized. It was interesting to observe that the collision energy had to be increased from a lower m/z (m/z 359) to higher ones (m/z 375) by 15 eV. The precursor ions were fragmented, maintaining Table 2
>20 % of the ion intensity in the product ion spectra. However, as one precursor ion selected is in reality several different compounds (see Fig. 2), not for all the ions was it possible to maintain a peak intensity >20 %, but always higher than 5 %. Once the mass spectra were acquired, all the fragment ions had to be assigned to the corresponding precursor ion. Due to the fact that soft fragmentation was applied (collision energy
RESEARCH PAPER
Fragmentation studies for the structural characterization of marine dissolved organic matter Nuria Cortés-Francisco & Josep Caixach
Received: 6 October 2014 / Revised: 28 December 2014 / Accepted: 15 January 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract High-resolution tandem mass spectrometry by collision-induced dissociation with a linear ion trap-Orbitrap has been performed on marine dissolved organic matter (DOM). Product ion spectra of selected precursor ions (m/z 359–375) have been acquired to obtain structural information, after method development. To evaluate the performance of the method, the Suwannee River fulvic acid (SRFA) reference standard was also analyzed. By reconstructing the individual product ion spectrum of marine DOM, several fragments were assigned to the different precursor ions indicating the presence of carboxyl, hydroxyl, lactones, quinones, esters, and structures more similar to lignin-degraded molecules. On the basis of these findings, coastal marine DOM molecules, although structurally homogeneous, might be more rich in diversity of functional groups than previously described. Keywords High-resolution mass spectrometry . Marine DOM . Product ion spectra . Fragmentation studies . Orbitrap
Introduction The application of electrospray ionization high-resolution mass spectrometry has transformed the comprehension of dissolved organic matter (DOM) at molecular level. Several studies have assigned molecular formulas to DOM extracts from different origins, providing information on natural or affected Electronic supplementary material The online version of this article (doi:10.1007/s00216-015-8499-3) contains supplementary material, which is available to authorized users. N. Cortés-Francisco : J. Caixach (*) Mass Spectrometry Laboratory/Organic Pollutants, IDAEA-CSIC, Jordi Girona 18, 08034 Barcelona, Spain e-mail: [email protected]
environments and degradation/transformation processes. However, DOM is very complex and each elemental formula could represent millions or more constitutional isomers depending on the mass [1]. Nowadays, structural information from DOM is still missing, due to the fact that DOM separation into individual compounds by conventional techniques such as liquid chromatography or electrophoresis is impossible. Several attempts have been done to obtain structural information using different techniques such as nuclear magnetic resonance (NMR), Fourier transform-infrared spectroscopy (FT-IR), and tandem mass spectrometry. For instance, NMR analysis has revealed that phenolic carbons are not the major components of humic substances [2] and FT-IR analysis has shown presence of aromatic, aliphatic, and carboxylic groups [3]. Mass spectrometry has also been successfully applied. Low-resolution tandem mass spectrometry has been applied in most of the studies, and it has been used to analyze the Suwannee River fulvic acid (SRFA) standard, deep ocean DOM, and freshwater fulvic acids. Few studies have used high-resolution tandem mass spectrometry for the characterization of SRFA and organosulfates in particulate organic matter (see Table 1) [4–13]. The presence of carboxyl and hydroxyl groups has been confirmed by several of these studies. Tandem mass spectrometry has been successfully applied to elucidate the structure of some of the molecular formulas assigned, and some tentative structures have been proposed. In this respect, the introduction of hybrid linear ion trap-Orbitrap has provided a new approach for structural characterization of DOM with enhanced sensitivity and resolution, thus providing improved selectivity and mass measurement accuracy for tandem mass spectrometry data. The purpose of this study was to develop a method to acquire product ion spectra of marine DOM to obtain structural information, which is still missing. The method has been based on the use of high-resolution tandem
N. Cortés-Francisco, J. Caixach Table 1 Previous studies (from the last 15 years) where fragmentation studies of DOM from different origins have been performed Dissolved organic matter analyzed
References
Deep ocean DOM
Reemtsma et al. [4]
Suwannee River standards (IHSS)
Stenson et al. [5, 6]
details on the extraction procedure performance can be found elsewhere [14]. LTQ Orbitrap mass spectrometry
Acidic metabolites in fulvic acids from groundwater Jobelius et al. [7] Fulvic acids from different origins
Plancque et al. [8]
SRFA standard (IHSS)
Witt et al. [9]
River and ocean DOM
Liu et al. [10]
SRFA (IHSS)
Leenher et al. [11]
Standard soil and peat fulvic acid standards (IHSS) McIntyre et al. [12] Organosulfates in aerosols
Lin et al. [13]
mass spectrometry by collision-induced dissociation (CID) experiments with an LTQ Orbitrap. High resolved product ion spectra and good mass accuracies in accurate mass measurements have been used to find new functionalities of marine DOM.
Materials and methods Chemicals All reagents were of analytical or high-performance liquid chromatographic grade. Dichloromethane, methanol, and hydrochloric acid were purchased from Merck (Darmstadt, Germany). Isopropyl alcohol was from Carlo Erba (Milan, Italy) and formic acid was from Panreac (Barcelona, Spain). Highpurity water produced with a Milli-Q Organex-Q System Millipore (Millipore Corp., Bedford, MA) was used. Suwannee River fulvic acid (SRFA) (1S101F) standard from International Humic Substances Society (MN, USA) has been diluted in methanol and analyzed to evaluate the performance of the method.
Sample collection and preparation A composite water sample from the seawater intake system of a pilot desalination plant installed in the coast of Barcelona (Spain) was collected in Pyrex borosilicate amber glass bottles. The seawater has been previously characterized as part of a whole study [14]. Briefly, 1.5 L from the sample was acidified with 10 % hydrochloric acid to pH 2 and extracted with 2×100 mL of dichloromethane/isopropyl alcohol (90:10 v/v) and the extract was concentrated down to 250 μL at 40 °C under nitrogen. It should be highlighted that as any other isolation method, fractionation of DOM may occur. Further
The analysis of the standard and the sample was carried out with an LTQ Orbitrap XL (Thermo Fisher Scientific, Bremen, Germany) equipped with a nano-ESI source (Nanospray Flex Ion Source, Thermo Fisher Scientific, Bremen, Germany). Optimization was carried out analyzing an SRFA standard by infusion in the Orbitrap in full-scan negative ionization mode. Optimized parameters were capillary voltage, skimmer voltage, tube lens voltage, and capillary temperature. Finally, all mass spectra were acquired in negative ionization mode produced by capillary voltage −35 V, tube lens −90 V, and capillary temperature 300 °C. The automatic gain control was used to consistently fully fill the C-trap and gain mass accuracy and resolution [15]. High resolution defined as R 100,000 (m/z 400, full width at half maximum) was set. For tandem mass experiments, ions in the mass range between m/z 359 and 375 (Fig. 1) have been selected in the LTQ before they were pulsed through the trap to the collision cell where they were fragmented by CID, using increasing voltage increments up to 30 eV. It was only possible to isolate ions in the 0.7-Da mass window, and therefore, tandem mass experiments include several precursor ions. The same m/z were selected for both SRFA standard and marine DOM extracts. Post-acquisition calibration had to be performed using reference homologous series described in DOM before by the RecalOffline Xcalibur application. The accuracy of the precursor and product ions was always better than 2 ppm. Data analysis The mass peaks were exported to peak lists and from these lists feasible elemental formulas were generated. Different restrictive criteria were set to generate reliable elemental formulas, depending on the precursor ion selected (Table 2). This way, the possible assignments for each product ion were not so numerous and the assignation of the product ions was more easy and reliable. The molecular formula calculation was performed with Xcalibur 2.1 (Thermo Fisher Scientific, Bremen, Germany), and the posterior analysis of the data was done using our own developed Excel macros. Data filtering of the assigned formulas was done by applying exact mass differences/neutral losses (Table 3), to correlate each precursor ion with the possible ion fragments. The methodology developed for fragment assignment relies on a not broken fragmentation pathway, where the differences between fragment ions could be attributable to a known neutral loss or functional group fragmentation.
Structural characterization of marine dissolved organic matter Fig. 1 Suwannee River fulvic acid full-scan spectrum of the precursor ions chosen for the posterior fragmentation experiments
369.11918
100
367.10384
90
Relative Abundance
80
365.08799
70
363.07240 371.09873
60
361.05693
50 40 30
359.04097
373.05672
375.03601
20 368.07084
10 362.05152
364.07559
370.12274
366.12801
372.10196
374.04691
0 360
362
364
366
368
370
372
374
m/z
In silico fragmentation In order to build the puzzle obtained from the tandem experiments for the marine DOM and the SRFA, the different HR product ion spectra obtained after data filtering were used in MetFrag application [16] to relate the neutral losses and product ions obtained to a known structure from databases. The search was performed in PubChem and ChemSpider databases.
Results and discussion Method development In addition to optimizing the experimental parameters for efficient DOM ionization, in the development of the method, the collision energy for each precursor ion had to be optimized. It was interesting to observe that the collision energy had to be increased from a lower m/z (m/z 359) to higher ones (m/z 375) by 15 eV. The precursor ions were fragmented, maintaining Table 2
>20 % of the ion intensity in the product ion spectra. However, as one precursor ion selected is in reality several different compounds (see Fig. 2), not for all the ions was it possible to maintain a peak intensity >20 %, but always higher than 5 %. Once the mass spectra were acquired, all the fragment ions had to be assigned to the corresponding precursor ion. Due to the fact that soft fragmentation was applied (collision energy
