Insights on Dissolved Organic Matter Production Revealed by Removal of Charge-Transfer Interactions in Senescent Leaf Leachates

: Dissolved organic matter (DOM) is a critical part of the global carbon cycle. Currently, it is understood that at least a portion of the chromophoric DOM (CDOM) character can be described through an electronic interaction of charge transfer (CT) complexes. While much work has been done to understand the inﬂuence of CT on soil and aquatic reference standard DOM, little is known about the inﬂuence of CT in fresh terrestrially derived DOM. In this study, leaf litter leachates from three tree species were treated (reduced) with sodium borohydride to determine the contribution of CT on a source of fresh terrestrial DOM. Leaf litter was sampled four times through decomposition under natural (ﬁeld) conditions to determine the inﬂuence of degradation on response to borohydride treatment. Leaf litter CDOM displayed a unique loss of UVB absorption following borohydride treatment, as well as a homogenizing e ﬀ ect on ﬂuorescence emission character. Humiﬁcation index (HIX) di ﬀ erentiated Elliot Soil Humic Acid and Suwannee River Fulvic Acid from leaf litter leachates. However, biological index (BIX), and spectral slope metrics were not able to di ﬀ erentiate leaf leachates from these reference standards. Apparent quantum yields were similar in magnitude between leaf leachates and reference standards, although leaf leachate spectra displayed features not evident in reference standards. These results help understand the origins of DOM optical properties and associated quantitative indices in freshly sourced terrestrial material. Overall, these results suggest that even at the initial stages of decomposition, terrestrial CDOM exhibits optical characteristics and responses to removal of electron accepting ketones and aldehydes, through borohydride treatment, similar to more processed CDOM.


Introduction
Dissolved organic matter (DOM) in freshwater systems is a complex mixture of organic molecules largely sourced from plant materials undergoing continuous biological and chemical processing across the terrestrial/aquatic interface [1]. The degree to which DOM from individual sources is processed largely determines the molecular structure and bioavailability of DOM in aquatic environments [2,3], but quantifying the processing of DOM has proven difficult [4]. Optical methods for characterization of DOM, namely absorption and fluorescence spectroscopy, offer high throughput and detailed characterization of the chromophoric DOM (CDOM) in aquatic environments [5][6][7]. However, there are still unresolved interactions among seemingly unique CDOM components [4] which make linkages of distinct chromophores to molecular structures difficult [5][6][7].
Intramolecular electronic interactions (EI), defined as charge transfer (CT) interactions, has been proposed as a model to explain the distinct exponential loss of absorption with increasing wavelength in CDOM [8]. Recent studies probing origins of aquatic DOM under the premise of the EI model have employed reduction of natural samples with sodium borohydride. Sodium borohydride causes selective reduction of carbonyl containing ketones and aldehydes, removing the electron acceptors of CT interactions [9][10][11][12][13][14][15][16]. The effects of borohydride treatment on the CDOM optical properties are two-fold: (i) the removal of CT interactions through the reduction of select electron acceptors decreases absorbance, and (ii) increases fluorescence by electron donors that are no longer quenched by CT relaxation pathways [13]. Electron accepting moieties responsible for CT also play an important role in the ability of DOM to act as a terminal electron acceptor during anaerobic respiration in aquatic environments [17,18].
A bulk of the work probing the importance of CT has been centered in aquatic environments which represent "downstream" environments of CDOM originating from terrestrial sources that has undergone various degrees of humification. For example, Suwannee River Fulvic Acid (SRFA), which is often used as a terrestrial reference material [10,12,14], originates from a collection point at the Suwanee River's headwaters, and represents a highly degraded DOM pool, which has likely undergone much microbial, and some photochemical processing, compared to terrestrial vegetation inputs [19]. The terrestrial/aquatic interface "upstream" of where SRFA is collected has been shown to be a hotspot in DOM processing [20]. Therefore, we hypothesize that fresh terrestrial DOM, yet to reach this interface, and containing relatively high concentrations of ketones and aldehydes in partially oxidized phenolic moieties, may display a greater sensitivity to CT reductions. This sensitivity is likely caused by the prevalence of electron accepting carbonyl moieties, as well as presumed donors such as polyphenols and substituted aromatic carboxylic acids in relatively unprocessed DOM [21].
Fundamental differences in photochemical properties between aquatic and soil reference substances have been demonstrated using reference standards from the International Humic Substances Society (IHSS) [22]. As such, SRFA has been used as an aquatic reference and Elliot Soil Humic Acid (ESHA) as a soil reference [22]. This work has suggested three main differences in electrochemical and photochemical properties between DOM in terrestrial (e.g., soil) and aquatic environments: (1) Soil HA displayed higher electron accepting capacity (EAC) and lower electron donating capacity (EDC) than aquatic DOM. (2) Lower wavelength absorption tracked changes in EAC in soil HA, to a greater degree than aquatic DOM. (3) Lower wavelength absorption appeared to track EDC during photodegradation of aquatic DOM, indicating that this absorption is influenced by oxidative removal of electron donors, such as phenols, more in aquatic environments than in soils.
The mechanism behind these findings is unclear because electron donors, such as low molecular weight phenolics, are likely delivered from the terrestrial environment, and therefore would be expected to contribute more in upstream (e.g., soil) than further downstream (e.g., aquatic) environments. Similar humic-like features have been observed in degradation of lignin-free phytoplankton DOM [14], suggesting that the humic-like fluorescence character is not exclusive to lignin derivatives acting as terrestrially-derived electron donors. Rather, it appears that microbial processing transfers the initial optical properties of source material, along with a modified humic-like signal that can significantly contribute to CT interactions in the produced DOM. In this way, the attribution of optical properties to a particular source can be confused by ubiquitous microbial fingerprints superimposed across terrestrial and aquatic environments [23].
Understanding CT interactions in natural waters warrants an understanding of the source of DOM and the role of microbes in altering its signal, particularly for understanding processes that span the terrestrial/aquatic interface. The main DOM inputs to both soils and aquatic environments in forested ecosystems are from groundwater and precipitation mediated leaching of plant materials, or throughfall [20]. Much of the leachable DOM is removed from leaves during the autumn season [24,25]. In forested watersheds, DOM leached by throughfall is stored in soils, and some is transported directly to the aquatic environment by leaf litter falling into the stream channel [26]. Regardless of its fate, freshly fallen leaf litter represents one of the youngest and most constant pulses of terrestrial CDOM delivered to aquatic environments from the terrestrial environment.
This study investigated CT influence on optical properties of leaf litter (O i soil horizon) leachates from three different tree species collected in situ spanning a time frame from directly after leaf fall to the following spring. We compared the CT influence in leaf leachates to CT characteristics in soil and aquatic reference standards to accurately characterize a source of fresh terrestrial DOM and the associated early stages of decomposition to both soil and aquatic, terrestrially sourced, ecosystems. The main aim of this study was to optically characterize young, terrestrially sourced DOM and observe how degradation modifies the source DOM composition, as observed through CT interactions.

Experimental Design
Leaf leachates were made from three species, Sugar maple (Acer saccharum Marsh), Paper birch (Betula papyrifera Marsh), and Red oak (Quercus rubra L.), herein referred to as maple, birch, and oak, respectively. A total of 10 g of leaves from each species were randomly collected from unmanipulated plots adjacent to the USDA Forest Service Rhizotron Facility in Houghton, MI (https://www.nrs.fs. fed.us/research/facilities/rhizotron/). To assess changes in leaf leachates with biological processing, we harvested leaves on four occasions, from November to June. For each experiment, 5.0 g of homogenized fresh leaves cut into 2.5 cm 2 sections were soaked in 250 mL MilliQ water for 4 h, (similar to [25]). Leaf leachates were then vacuum filtered through 0.7 µm GF/F filters followed by 0.45 µm nylon filters and added to 500 mL pre-combusted glass bottles. Samples were stored in the dark at 4 • C, to minimize additional microbial processing after extraction.
Reference standards SRFA and ESHA were chosen as representative aquatic and soil reference materials and to compare our results with recent studies [4,8]. 5.0 mg of SRFA and 2.0 mg of ESHA were solubilized overnight in 100 mL MilliQ water adjusted using high purity NaOH (Fisher) to pH 9. These stock solutions were then adjusted to pH 7 and diluted with MilliQ to desired concentrations. Absorption and fluorescence spectra were measured for SRFA and ESHA at 25.0 mg L −1 and 6.7 mg L −1 respectively, which gave an absorbance of 0.6 at excitation 250 nm.

Optical Measurements
Absorption and fluorescence spectra were measured on a Horiba Aqualog in a 1 cm quartz cell (Starna Cells, Inc., Atascadero, CA, USA). Absorption spectra were run from 240 nm to 800 nm at 1 nm resolution. Leachates were diluted to ensure measurements fell into the linear range of the instrument if necessary. Fluorescence spectra were recorded at 1 nm excitation wavelengths from 240 to 800 nm, and emission was recorded at 3 nm resolution from 240 to 640 nm. Fluorescence spectra were converted to Raman units using Raman scattering from a sealed MilliQ cuvette (Starna) (R.U.) and quinine sulfate equivalents (Q.S.E.) using the fluorescence at 350 nm (ex 350, em 290-600) of 10 ppb quinine sulfate (NIST) in 0.105 M HClO 4 (VWR). Absorbance values were converted to Naperian absorption coefficients (a λ ) [27]: where L is the pathlength, in meters and A λ is wavelength specific absorbance. Absorption spectra were blank subtracted with MilliQ for untreated spectra and a procedural blank for borohydride treated samples. All measurements were taken at pH 7. Inner filter effects, Raman scattering, and Raleigh scattering for fluorescence EEMs were removed using the PLS toolbox (v782; Eigenvector) in Matlab (2014a) (The MathWorks Inc., Natick, MA, USA). Derived parameters from absorption spectra were E2:E3 and spectral slopes. E2:E3 was defined as: where a 250 is absorption at 250 nm and a 365 is absorption at 365 nm. S 275-295 and S 350-400 values were calculated as the slopes of linear regressions of log-transformed absorption spectra from 275-295 nm and 350-400 nm respectively [7].
Derived parameters from the fluorescence spectra were the biological index (BIX) [28] and humification index (HIX) [29,30] and were defined as: where F 380 and F 430 are emission at 380 nm and 430 nm, respectively, for excitation at 310 nm.
where ΣF is the integrated emission over the indicated wavelength ranges for excitation 255 nm. Absorption changes following borohydride treatment are reported as: where a(0) is absorption before borohydride treatment and a(t) is absorption after borohydride treatment. Apparent quantum yields (Φ app ) were calculated at each excitation wavelength from 240 to 600 nm, relative to a 10 ppb solution of quinine sulfate [27]: where 0.60 accounts for 10 ppb quinine sulfate in 0.105 M perchloric acid [31]. I s,λ is the integrated sample fluorescence emission intensity at the excitation wavelength (λ). I QS,350 is the fluorescence emission intensity of quinine sulfate at excitation 350 nm. a QS,350 is the absorption coefficient of quinine sulfate at excitation 350 nm and a s,λ is the integrated sample absorption coefficient the excitation wavelength.

Borohydride Treatments
Leaf leachates were adjusted to pH 7 and pH 10 using high purity NaOH (Fisher, Hampton, NH, USA), and the optical properties were recorded for each pH. A total of 17.0 mg of sodium borohydride (Acros Organics, Hampton, NH, USA) was added to 17.0 mL of leaf tea, and the reactions were allowed to proceed for 48 h [16]. Leachates were then filtered through 0.45 µm nylon filters and optical properties were again recorded at pH 10, pH 7 and natural pH. Additions of HClO 4 were used to adjust pH (VWR). Results are reported at a neutral pH of 7, as recommended by [16]. All experiments were run with three analytical replicates. Under visual inspection, some experiments displayed unique variation in absorption spectral shape for individual replicates before treatment. To ensure representativeness, the average of the two most similar experimental replicates was reported for all experiments (see also, NASA optical measurement protocol; [32]).

Absorption Change after Borohydride Treatments
Following borohydride treatment, all leaf leachates at all time points displayed a loss of absorption in the UV (particularly the UVB (280-315 nm)). The normalized difference spectra ( Figure 1) show that approximately 20-60% of the original absorption was lost from 250-400 nm, generally increasing with wavelength. Longer wavelength loss was similar for all three of the autumn samples. The spring samples (May 2019) displayed a smaller loss in absorption from the borohydride treatment. Those samples had significantly lower absorption at wavelengths greater than 400 nm ( Figure S1). The magnitude of this loss for IHSS reference standards was similar to that of birch and maple. Oak showed a greater fraction of absorption loss than IHSS standards, consistently displaying the highest absorption loss of the leachates (Figure 1). A key feature, a peak around 280 nm, became more pronounced after borohydride treatment for all leaf leachates ( Figure S1). Spectral changes between species and among the autumn samples were not as large as the changes from December to May (Figure 1). Oak DOM showed a greater sensitivity to borohydride treatment than birch and maple. Oak and birch displayed two broad regions of absorption that were less impacted by borohydride treatment, centered around 280 and 370 nm, suggesting the presence of unaltered chromophore groups in these spectral regions. For CDOM only participating in charge-transfer, spectra in Figure 1 should present Oak DOM showed a greater sensitivity to borohydride treatment than birch and maple. Oak and birch displayed two broad regions of absorption that were less impacted by borohydride treatment, centered around 280 and 370 nm, suggesting the presence of unaltered chromophore groups in these spectral regions. For CDOM only participating in charge-transfer, spectra in Figure 1 should present a monotonic increase from shorter to longer wavelengths, representing the increasing contribution of CT to total absorption at longer wavelengths.
For all leaf leachates, there is a consistent decrease in response to borohydride treatment from 13 November 2018 to 29  When spectra are characterized as S 350-400 and S 275-295 , leaf leachate absorption is within the range of reference standards and therefore not easily distinguished, even following borohydride treatment ( Figure 2). Both S 350-400 and S 275-295 increased for each species following borohydride treatment, similar to SRFA. However, the range of oak leachates overlapped with the borohydride treated maple and birch leachates. When spectra are characterized as S350-400 and S275-295, leaf leachate absorption is within the range of reference standards and therefore not easily distinguished, even following borohydride treatment ( Figure 2). Both S350-400 and S275-295 increased for each species following borohydride treatment, similar to SRFA. However, the range of oak leachates overlapped with the borohydride treated maple and birch leachates.

Fluorescence Response to Borohydride Treatment
Fluorescence response to borohydride is best represented through Φapp spectra [27]. Magnitudes of Φapp of all treated and untreated samples were similar to IHSS reference standards (Figure 3). The most evident defining characteristics in contrast to IHSS reference standards was the emergence of two local maxima between 300-400 nm following borohydride treatment (Figure 3). Oak was least changed through time (Figure 3). The leachates collected 5 June 2019 displayed less structure than the earlier times as well as no increase in Φapp in response to borohydride treatment at greater than 500 nm (Figure 3).
When fluorescence was characterized as BIX, leaf leachate CDOM absorption was within the range of the reference standards, and is therefore not easily distinguished. On the other hand, HIX

Fluorescence Response to Borohydride Treatment
Fluorescence response to borohydride is best represented through Φ app spectra [27]. Magnitudes of Φ app of all treated and untreated samples were similar to IHSS reference standards (Figure 3). The most evident defining characteristics in contrast to IHSS reference standards was the emergence of two local maxima between 300-400 nm following borohydride treatment (Figure 3). Oak was least changed through time (Figure 3). The leachates collected 5 June 2019 displayed less structure than the earlier times as well as no increase in Φ app in response to borohydride treatment at greater than 500 nm (Figure 3). increased in all species following borohydride treatment, yet HIX was unresponsive for Oak. SRFA displayed a unique decreasing response in HIX, following borohydride treatment. This response was not observed in leaf litter leachates or ESHA. HIX increased through time in maple and oak leachates. Birch leachates also increased in HIX although there was a slight decrease between the first two time points (Table 1). BIX did not change through time.  When fluorescence was characterized as BIX, leaf leachate CDOM absorption was within the range of the reference standards, and is therefore not easily distinguished. On the other hand, HIX displayed separation between IHSS reference standards and leaf litter leachates (Figure 4). BIX increased in all species following borohydride treatment, yet HIX was unresponsive for Oak. SRFA displayed a unique decreasing response in HIX, following borohydride treatment. This response was not observed in leaf litter leachates or ESHA.
HIX increased through time in maple and oak leachates. Birch leachates also increased in HIX although there was a slight decrease between the first two time points (Table 1). BIX did not change through time.

Discussion
Following borohydride treatment, carbonyls, specifically ketones and aldehydes, are removed from DOM [13]. As these chromophoric moieties are converted to colorless alcohols, absorption decreases. Additionally, borohydride likely removes electron transfer pathways that quench fluorescence, thereby increasing fluorescence intensities. Plant tissues such as lignin derivatives, higher in carbonyls than lignin itself, are therefore expected to exhibit a significant response to borohydride treatment. The findings presented herein are in support of this hypothesis.

Absorption Response to Borohydride
Absorption decreased in all leaf leachates following borohydride treatment. Shorter wavelength absorption (UVB) decreased in autumn leaf litter more than SRFA and ESHA (Figure 1). This supports the findings of [4], that electron acceptors play a larger role in UVB absorption in fresh relative to aged CDOM. Interestingly, unlike leaf litter leachates, fresh planktonic DOM does not exhibit decreases in UVB absorption following borohydride treatment [14]. This supports the hypothesis that partially oxidized products of degradation play an important role as CT donors in increasingly fresh terrestrial DOM, as these moieties are expected to be present and contribute significantly in leaf litter; such degradation products may not be present in phytoplankton derived DOM.
Changes in the visible region due to borohydride treatment of the autumn leachates were consistent with soil and aquatic HA (Figure 1, Figure S1). However, the most evident change in leaf leachate absorption through time occurred in leachates from leaves that had aged under the snow over the winter. These samples showed a loss of the longer wavelength response to borohydride which was dramatically different from SRFA and ESHA (Figures 1 and 2). This loss extends the findings of Sharpless and Blough [4], by showing the sensitivity of fresh terrestrial CDOM to removal of electron acceptors following borohydride treatment. However, it is unclear why the earlier leaf leachates tracked reference standards better than the more aged leaf matter. More aged DOM that is more condensed (lower H:C) and/or more aromatic should result in higher electron accepting capacity [33,34]. One hypothesis is that these electron acceptors, with relatively high quinone contents [13], are preferentially removed through leaching during snowmelt, due to their higher molecular weight. Interestingly, the higher UVB absorption losses are conserved into the spring for all leachates. This suggests that UVB absorption response to borohydride may not be an indicator of DOM processing, but rather a descriptor of source, and is further supported by the unique increase in UVB absorption of phytoplankton derived DOM to borohydride treatment [14].
E2:E3, S 275-295 , and S 350-400 all increased following borohydride treatment (Table 1). These increases contrast fresh and degraded phytoplankton DOM, which was reported to decrease in spectral slope following borohydride treatment [14]. This significant deviation between algal and terrestrial DOM inputs is likely driven by the absorption peak observed around 350 nm. These findings highlight the utility of E2:E3 as a robust indicator of EAC, even in terrestrial materials [17,22]. S 275-295 did not appear to differentiate leaf litter leachates from IHSS reference standards. Furthermore, S 275-295 increased following borohydride treatment, displaying its sensitivity to processes other than changes in DOM molecular weight, for which it is often employed. Overall, absorption characteristics changed in response to borohydride treatment, indicating that the metrics reported here may not be independent of DOM processing, and therefore are not recommended to infer DOM source.

Fluorescence Response to Borohydride
The fluorescence of leaf leachates showed multiple features in the Φ app spectra with considerable variability between samples producing more structured Φ app spectra in leaf leachates than SRFA and ESHA ( Figure 3). Specifically, the leaf leachates generally displayed two unique maxima between 300-400 nm following borohydride treatment. An important note is that these maxima appear to be blue shifted or red shifted from the original spectra depending on sample. These findings suggest that the initial maxima may not be directly related to the maxima of the borohydride treated sample (e.g., blue shifting of a fluorophore's peak fluorescence), even though they exhibit similar spectral signatures.
Multiple insights can be gained simply through the interpretation of leaf leachate fluorescence as HIX and BIX. For one, leaf leachates display lower HIX than reference standards, and are similar to planktonic CDOM, presumably indicating similar amounts of conjugation between these source materials. Contrarily, BIX does not differentiate leaf leachates from either SRFA or ESHA, however, leaf leachates appear to exhibit lower BIX than planktonic CDOM [14]. Interestingly, all of these patterns are preserved following borohydride treatment, indicating the uniqueness of each source's fluorescent character when represented as HIX and BIX. However, HIX and BIX did not differentiate between species of leaf leachates.
Another interesting finding is the removal of the fluorescent peak at (excitation 280 nm/emission 350 nm), often referred to as tryptophan-like, following borohydride treatment. This change is most evident in the 29 November 2018 Birch leachate ( Figure S2). The occurrence of this peak only in select leachates suggests that this peak is an intermediate product of DOM processing, consistent with [19]. In addition, this peak can be removed when treated by borohydride ( Figure S2). Pearl et al. [35] showed that relatively minor modifications to small molecules exhibiting protein-like fluorescence can dramatically shift the observed fluorescence signal. The peak observed here (excitation 280 nm/emission 350 nm) exhibits tryptophan-like fluorescence, yet is removed by borohydride treatment, which would not be expected for tryptophan. This suggests that the tryptophan-like fluorescence observed here is produced by partially oxidized organic compounds, which exhibit protein-like fluorescence (perhaps catechins [36]), yet becomes less fluorescent or emission shifted in response to borohydride treatment. Murphy et al. [37] observed a similar tryptophan-like fluorescent feature in pharmaceutically derived sodium salicylate, providing strong evidence for compounds other than proteins to produce protein-like fluorescence. Our results provide additional evidence that forest-derived phenolics may be responsible for this fluorescent feature in many aquatic environments. Further work employing a combination of mass spectroscopy with borohydride treatment is warranted, to understand the unique molecules contributing to this fluorescent feature.

Tree Species Differences
The variation in borohydride response through time outweighs species differences between birch and maple in both absorption and fluorescence efficiencies (Figures 1 and 3), suggesting that microbial and fungal degradation of leaf litter can manifest into a common optical signature from a variety of terrestrial source material. This is consistent with recent work suggesting a homogenization of source signals into a complex, non-unique signal in aquatic DOM, largely attributed to photo-and microbial transformations of source material [23]. Oak displays a much more structured Φ app spectra compared to the other species and is more consistently changed over time (Figure 3). Oak leaf chemistry is known to inhibit decomposition [38], and therefore, our results are likely tracking a prolonged process of decomposition not captured for the other species, with a reduced microbial and fungal decomposition signal. However, at 5 June 2019, all three species had lost the long wavelength absorption, responsive to borohydride treatment, found in the previous time points and present in SRFA and ESHA. The longer wavelength absorption is stable through the autumn, so this absorption may represent a recalcitrant fraction of fresh DOM, removed from leaf litter by leaching.

Conclusions
Our findings displayed an overall similar response of fresh leaf litter leachates and IHSS reference standards to borohydride treatment, with largely similar percentages of absorption loss as well as fluorescence gained among samples following borohydride treatment. However, our findings highlight the complex and highly variable signatures associated with a fresh terrestrial DOM among deciduous broadleaf tree species and through time, even following borohydride treatment. Further work is warranted to understand if these responses are consistent for coniferous tree species, as well as woody and herbaceous shrubs. Although HIX was able to differentiate IHSS standards from leaf leachates both before and after borohydride treatment, we stress the variability of end-member signatures, even after removal of the CT "haziness", which make inferences to DOM source not recommended. These results confirm that absorption and fluorescence indices from fresh DOM sources-particularly protein-like fluorescence-are sensitive to redox environment, and therefore redox conditions need to be considered when comparing optical data from different electrochemical environments.