Molecular characterization of dissolved organic matter in a North Brazilian mangrove porewater and mangrove-fringed estuaries by ultrahigh resolution Fourier Transform-Ion Cyclotron Resonance mass spectrometry and excitation/emission spectroscopy

Abstract

Electrospray ionization (ESI) combined with ultrahigh resolution Fourier Transform-Ion Cyclotron Resonance mass spectrometry (FT-ICR MS) at 9.4 Tesla has been shown to be an ideal tool for the molecular characterization of dissolved natural organic matter (DOM). Here, we apply this technique for tracing DOM as it moves from mangrove porewaters into a Brazilian estuary and assess the changes that occur in the initial stages of outwelling. Comparisons of molecular features of DOM from the porewater and two sites within the adjacent estuary revealed that mass spacing patterns were far more pronounced in the mangrove porewater DOM, and the average molecular weights of estuarine DOM were lower. Calculation of double bond equivalents and Kendrick mass analysis normalized for both methylene (-CH₂−) and oxygen (O) substitution indicated that the lower molecular weights observed in the estuarine DOM were due primarily to loss of high molecular weight, highly unsaturated and/or aromatic components observed in the porewater DOM. Photodegradation is suggested as one mechanism that removes these high molecular weight components once mangrove-derived DOM reaches the estuary. The mass analyses were supported by UV–Vis and 3-D synchronous scanning fluorescence spectroscopy measurements to aid in the characterization of DOM in this unique ecosystem.

Introduction

Mangrove forests are highly productive ecosystems that form the fringe areas of 60–75% of tropical coasts (Clough, 1998). Up to half of mangrove net-primary production is exported as leaves, detritus or dissolved organic matter via mangrove creeks to adjacent waters (Robertson et al., 1992, Dittmar and Lara, 2001a, Dittmar and Lara, 2001b). This export greatly affects the food webs in coastal waters (Alongi, 1990). Coastal zones are heavily impacted by human activity, which in turn affects food webs of entire shelves, resulting in widespread economic consequences (Alongi et al., 1989, Alongi, 1990, Chong et al., 2001). In order to properly assess the extent of human impact in the coastal zone on food webs and, ultimately, global carbon cycling, we must better define the role intertidal zones play in global elemental cycles. The mangroves in Northern Brazil dramatically illustrate the importance of coastal outwelling, especially of dissolved solutes (Dittmar et al., 2006). The high productivity of these intertidal areas combined with efficient material exchange and the virtual lack of physicochemical removal mechanisms make them important sources for dissolved organic carbon (DOC) flux to oceans (Dittmar et al., 2001, Dittmar and Lara, 2001a, Dittmar and Lara, 2001b, Dittmar and Lara, 2001c). Contrary to suspended particles and coarse detritus, DOC can be distributed over significant distances, affecting the ocean on a large scale. Approximately a quarter of the total organic matter exchange in the North Brazilian tidal creeks occurs in dissolved form (Dittmar et al., 2001). Outwelling of DOC occurs through tidal flushing of porewater that flows in considerable amount into tidal creeks and the ocean. The contribution of mangrove-derived DOC to the estuarine organic carbon pool greatly exceeds the riverine input in Northern Brazil. On a global level, more than 10% of the land–ocean fluxes of DOC are derived from mangroves, although only about 0.1% of the continents are covered by these intertidal forests (Dittmar et al., 2006).

Chemical tracers (lignin and stable carbon isotopes) have been applied to identify the fate and source of DOM in coastal environments (e.g. Meyers-Schulte and Hedges, 1986, Lobbes et al., 2000, Dittmar et al., 2006). To date lignin is the only established molecular tracer for terrigenous DOM in the ocean. It is an unambiguous tracer for vascular plants and can even distinguish vegetation types. Lignin has been widely used to trace the fate and transport of terrestrial organic matter in rivers and marine environments (e.g. Sarkanen and Ludwig, 1971, Hedges and Mann, 1979, Dittmar and Lara, 2001c). Benner et al. (1990) found that lignin-derived phenols were leached from mangrove leaves during early diagenesis. However, exposure to sunlight also decreases lignin concentrations within just a few days, making the quantitation of terrigenous DOM in the ocean difficult (Opsahl and Benner, 1998).

Mass spectrometry shows great promise for the characterization of high molecular weight compounds typically found in DOM. Although high resolution techniques (mass resolving power ⋍ 10,000) such as time of flight (TOF) have had some limited success in distinguishing DOM molecules at molecular weights less than about 300 Da (These et al., 2004), ultrahigh resolution mass spectrometry (mass resolving power > 100,000) is currently the only analytical technique capable of fully resolving individual species in complex DOM mixtures up to masses of ⋍ 1100 (Stenson et al., 2003). The introduction of soft ionization methods, in particular electrospray ionization (ESI), has further advanced the applicability of mass spectral analysis of natural organic matter (Fievre et al., 1997, Solouki et al., 1999, Persson et al., 2000, Brown and Rice, 2000, Alomary et al., 2000, Ikeda et al., 2000, Plancque et al., 2001, Kujawinski et al., 2002). ESI produces largely intact ions from biochemically-derived molecules in aqueous solution at atmospheric pressure, although the ESI stability of molecules in DOM has not been absolutely verified. Stenson et al. (2002) carried out a series of experiments on Suwannee River fulvic acids and concluded most if not all the covalently-bound core structures of these fulvic acids survived ES ionization intact. Conversely, Reemtsma and These (2003) coupled size exclusion chromatography (SEC) with ESI-MS and observed molecular weights of the same material that did not agree as determined by the two complimentary techniques. These authors suggest that fragmentation in the ESI source as one possible reason for the discrepancy. It should be noted, however, that the use of size exclusion chromatography to determine mass distributions of DOM mixtures has not been thoroughly validated. Large aggregates held together by ionic interactions which are disrupted in the ESI process may result in overestimation of actual covalent molecular weights by SEC.

Electrospray ionization and ultrahigh resolution mass spectrometry can distinguish individual elemental compositions in DOM and define the fundamental molecular differences between them. Recently this technique was successfully applied to the molecular characterization of oceanic DOM (Koch et al., 2005). The repeating mass spacing patterns observed in the mass spectra of humic and fulvic acids can also provide insight into the bulk composition of humic substances (Stenson et al., 2002, Stenson et al., 2003). For example, the frequently observed spacing of 0.0364 Da represents the replacement of a CH₄ group for an O atom. That replacement and the resulting mass spacing pattern is consistent with the complex series of chemical changes that accompany the microbial degradation of lignin (Filley et al., 2002, Stenson et al., 2003).

A characteristic of DOM mixtures is the natural fluorescence of a portion of the molecules. Chromophoric dissolved organic matter (CDOM) is the absorbing fraction of DOM and a major contributor to the optical properties of natural waters. CDOM is also a source for reactive oxygen in marine waters (Cooper et al., 1989, Kieber et al., 2003), but high concentrations may inhibit primary production by blocking light penetration (Mopper and Kieber, 2002). The highest concentration of CDOM is found in coastal waters and in semi-enclosed areas where terrestrial organic matter is the primary source (Morel and Prieur, 1977).

Conventional fluorescence has been used to distinguish marine and freshwater CDOM (Willey and Atkinson, 1982, Ferrari and Dowell, 1998, Ferrari, 2000). However, a more recent technique, Excitation Emission Matrix (EEM) fluorescence spectroscopy (3-D fluorescence), can provide more detail about CDOM (Coble, 1996, Kowalczuk et al., 2003, Jaffe et al., 2004, Chen et al., 2004). 3-D spectra are obtained by acquiring emission spectra at a series of successively longer excitation wavelengths. Emission spectra are then superimposed to display fluorescence intensity as a function of excitation and emission wavelengths. These synchronous scanning fluorescence technique results in EEM spectra that provide highly detailed information about fluorescing compounds in complex mixtures. EEM spectra can be used to follow changes in CDOM resulting from biological or physical processing of the material. 3-D fluorescence spectroscopy is thus a complementary tool for illuminating the biogeochemical mechanisms responsible for the molecular formula patterns observed by FT-ICR MS in mangrove-derived DOM.

Before fluorescence and mass spectrometry analyses of DOM can be carried out, it is first necessary to concentrate and desalt the DOM samples (King et al., 2000, Kujawinski et al., 2002). Traditional methods are ultrafiltration and solid phase extraction with XAD resins (Leenheer, 1981, Thurman and Malcolm, 1981, Burba et al., 1995, Kim et al., 2003). Ultrafiltration isolates DOM molecules based on their ability to penetrate a porous membrane, and that penetration depends on the molecular hydrodynamic diameter and thus ultimately on molecular weight. However, the most retentive membranes available have molecular weight cutoffs of ∼ 1000 Dalton (Da), and, since much of DOM is composed of molecules of molecular weight less than 1000 Da, the resulting molecular weight distribution and composition may not be truly representative. While solid phase extraction with hydrophobic resins is also selective, higher extraction efficiencies are usually obtained compared to ultrafiltration. Traditional XAD resin extractions require highly extensive cleaning before adsorption. C₁₈ solid phase extraction (SPE) has been used as an alternative and more practical method for concentrating and desalting organic matter from various sources (Kim et al., 2003). C₁₈ SPE of DOM is reportedly more efficient than XAD extraction by 24–84% (Amador et al., 1990), and was thus the method of choice for this work.

In this study, we have characterized the molecular composition of DOM in a Brazilian mangrove porewater and two sites within an adjacent estuary by FT-ICR MS to assess the changes that occur in the initial stages of outwelling. We have also used patterns in molecular formulas to infer degradation pathways and identify potential molecular tracers of mangrove-derived DOM in the ocean. The mass analyses were supported by UV–Vis and 3-D synchronous scanning fluorescence spectroscopy measurements that yield insights into the changes in spectral properties that accompany changes in mass distributions. To assess the degree of selective extraction via the SPE procedure, UV–Vis spectra of the porewater and estuarine water samples were obtained before and after solid phase extraction.