Differences in the composition of dissolved organic matter (DOM) among samples provide an indication of sample origins and reflect the refractory or labile nature of DOM [1
]. Therefore, understanding the composition of DOM is an essential requirement in studies of carbon and nitrogen cycling, e.g., [3
]. Amino acids (AAs) account for a small fraction of the bulk DOM, but they offer a powerful tool for determining the composition of organic matter and are widely used in contemporary organic matter studies, e.g., [1
]. AAs are derived mainly from living organisms, and variations in both their total amounts and relative abundances are tightly coupled to the metabolism of organisms. Therefore, both the concentrations and relative abundances of AAs are dependent on the status or progress of their metabolism (in the short term, such as within a day) and upon their source (in the longer term, such as seasonally).
In addition to the diel rhythm of photosynthesis (day and night) in phytoplankton, bacteria are periodically active in surface waters, even though their abundances are temporally similar [9
]. The coupling of phytoplankton production, organic matter generation, and bacterial production in oligotrophic surface waters on a scale of hours has been proposed [10
]. Such variations in microbial activity over a diel cycle have been repeatedly reported and confirmed, e.g., [11
]. With respect to DOM, a labile fraction of bulk dissolved organic carbon (DOC) fluctuates over hours to days, as reviewed by Hansell (2013) [13
]. In addition to the bulk DOC concentrations, the composition of DOM also shows temporal variability. The dissolved free AAs varied from 46 nM to 160 nM within 24 h at an inner Baltic Sea site [14
]. Indeed, the DOM composition is frequently sampled and monitored in studies of diazotrophs and their fixation of nitrogen [15
]. These studies have shown that by releasing dissolved organic nitrogen (DON), diazotrophs in oligotrophic waters affect DOM over very short time scales. However, despite frequent observations of dissolved free AAs and other free small molecules [18
], total hydrolysable dissolved AAs (THDAAs) are rarely reported in diel observations of DOM. Though THDAAs and their relevant proxy variations over days to thousands of years are well revealed [20
], the high-frequent variation degree of marine DOM within 1 day remains surprisingly unclear, which further interferes with our understanding of DOM composition over seasonal or even longer time scales.
Predominantly composed of combined AAs, THDAAs usually turn over much more slowly than dissolved free AAs or other small organic molecules. However, some THDAAs are also active and show changes in both field observations [21
] and laboratory cultures [22
], although the time scale (sampling interval) is usually longer than 1 day. The extent or degree of the THDAA changes within 1 day remains largely unclear. Such changes can be blurred due to their spatially heterogeneous distribution (in the case of field sampling) and/or by analytical errors, but should be attributable, in large part, to variations in microbial metabolism. It is also helpful to understand the temporal variability in THDAAs within 24 h, because such short-term changes allow us to more confidently clarify the changes in molecular indicators over longer timescales.
The northern slope of the South China Sea (SCS) is an oligotrophic open sea (Figure 1
), usually characterized by nitrogen limitation [23
]. In autumn and winter, vertical mixing in the upper water column is stronger, relative to spring and summer. Occasional eddy events further enhance the vertical gradient feature from season to season. The phytoplankton biomass remains low, but both the biomass and community structure can change a lot in response to the water column structure (e.g., occasional eddies) [24
Within the framework of a national key research program, this study was part of multidisciplinary investigations that included comprehensive physical, chemical, and biological studies undertaken in the study area. With a clear background in the northern SCS, the questions in this work are: (1) what are the short-term changes in AA-based molecular indicators in the northern SCS within 1 day, and based on the achieved short-term changes? (2) what is the DOM composition difference in surface waters (<200 m) between June and October and its implications? To answer the questions, time-series observations and section observations for the upper 200 m were conducted in the northern slope area of the SCS in October 2014 and June 2015, respectively. In this work, we first quantify the propagated errors in our laboratory measurements and evaluate the degrees of variability of the AA-based molecular indicators over 24 h. Then, the seasonal DOM composition difference in the upper 200 m along the section is characterized. At last, the potential implications, including the DOM sources and degradation status, are addressed based on the AA-based molecular indicators.
2. Materials and Methods
The benefits of multidisciplinary investigations and the physical, chemical, and biological background of this work are basically revealed [25
]. In June 2015, an eddy-pair was found during our diel observations, whereas in October 2014, no mesoscale process was detected (Figure 1
]. The observation sites were impacted by weak upwelling (Figure 1
]. Consequently, the surface phytoplankton biomass (0.12–0.48 μg L−1
) was higher in June 2015 than in October 2014, when the biomass was as low as 0.02 μg L−1
]. In June 2015, the phytoplankton community was dominated by Chaetoceros
, both of which are diatoms [27
]. Although their abundance remained unclear in previous work, when the bacterial diversity was studied in October 2014 the community was dominated by heterotrophs, and the abundance of autotrophic bacteria was rather low [28
2.2. Field Sampling
Two cruises were undertaken in R/V Nan Feng
on the northern slope region of the SCS during 10–30 October 2014 and 10–30 June 2015. The temperature (T), salinity (S), fluorescence, and photosynthetically active radiation (PAR; in June 2015 only) were measured with a CTD (Seabird, USA). Discrete water samples were collected in depth profiles (top 200 m) along a section that covers the SCS northern slope area (Figure 1
). In addition, time-series stations (TS2014 in 2014 and TS2015 in 2015) were also observed. For the time-series stations, sampling began at 06:00 and lasted for 1 day at TS2014. At TS2015, sampling began at 12:35 and also lasted for 1 day. In both cruises, the time-series observations were carried out 1 week later when compared to the corresponding section observations. Up to five profile samplings were conducted during the time-series stations in each of the two seasons. The sampling interval was approximately 6 h.
Immediately after collection, the seawater samples for the analysis of AAs and DOC were filtered with a clean nylon membrane (pore size 0.45 μm) and a syringe, and frozen (−20 °C) before analysis in the laboratory. Water samples for the analysis of dissolved nutrients were first filtered through acid-cleaned acetate cellulose filters (pore size 0.45 μM), treated with HgCl2, and stored at 4 °C in the dark before analysis. Nutrient samples were not collected during the last two samplings at TS2014.
At TS2014, water samples were further collected for the detection of bacterial abundance at 15:00 and 21:00 on 26 October (i.e., 3 h after the molecular indicators were sampled), whereas in late spring (June) at TS2015, the water samples for bacterial abundance were collected at the same time as the samples used to analyze the molecular indicators (one sampling point was missed: the 2nd sampling point). A volume of 47.5 mL of seawater was collected and fixed with 2.5 mL of formaldehyde for 15 min. The samples were stored at 2–8 °C and analyzed in the laboratory.
2.3. Laboratory Measurements
All the measured AAs (including both l
- and d
-enantiomers) and their abbreviations are shown in Table 1
. The THDAAs were analyzed with a previously described method [29
]. Briefly, the water samples were first hydrolyzed in sealed ampoules with 6 M HCl at 110 °C under N2
. For samples with elevated nitrate levels, ascorbic acid was also added [30
]. After hydrolysis, the samples were neutralized and the pH adjusted to 8.5. After precolumn derivatization with o-phthaldiadehyde (OPA) and N
cysteine, the AA enantiomers were separated and measured in the hydrolysates with high-performance liquid chromatography (HPLC; 1200 series Agilent, USA), with a fluorescence detector. During hydrolysis, asparagine (Asn) and glutamine (Gln) were deaminated to aspartic acid and glutamic acid, respectively. Therefore, “Asx” represents Asn plus aspartic acid (Asp), and “Glx” represents glutamine plus glutamic acid (Glu) (Table 1
). The DOC concentrations were measured with a total organic carbon (TOC) analyzer (TOC-LCPH
, Shimazu, Japan). Ammonium was measured manually by the sodium hypobromite oxidation method, with an analytical precision of 0.04 μM. Concentrations of the other four nutrients were determined with an auto-analyzer (AA3: SEAL Analytical, Mequon, WI, USA), with precisions for nitrate, nitrite, dissolved inorganic phosphorus (DIP, PO43−
), and silicate (SiO32−
) of 0.01, 0.003, 0.005, and 0.02 μM, respectively. Milli-Q water was used in all laboratory processing, and the chemicals were purchased from Sigma-Aldrich (Missouri, USA) or Merck (USA) and were of HPLC grade or above.
To determine bacterial abundance, 1 mL 4′,6-diamidino-2-phenylindole solution (DAPI; 10 μg mL−1
) was added to 30 mL of each sample, and the mixtures were incubated in the dark for 5–10 min. Each sample was then filtered through 25 mm, and 0.2 μm black Whatman®
Nuclepore™ Polycarbonate Filters were mounted on pre-wetted 0.8 Nuclepore Membra-fil®
cellulosic filters [31
]. The filters were then set onto glass slides with a drop of Leica type A immersion oil. The bacterial abundances were determined with an epifluorescence microscope (Leica, DM5000B).
2.4. The AA-Based Molecular Indicators and Propagated Error Quantification
Among the AA-based molecular indicators, AA carbon yield is the AA carbon (in nM) divided by bulk DOC (in nM). AA% is the mol percentage of a given AA in THDAA. The degradation index (DI) is a measure that tells the overall degradation status of organic matter. The DI, and its calculation equation, was first proposed by Dauwe and Middelburg (1998) [32
] for use in particulate and sediment studies. Later, it was applied to DOM with a revised constant derived from updated DOM samples [21
]. In this work we calculated DI following the method of Dauwe and Middelburg (1998) [32
], using the revised coefficients derived from DOM sample pools [21
To check the errors in the measurements of the AA enantiomers, a water sample (10L) obtained at a depth of 5 m near TS2014 was sealed and stored for 3 years at room temperature in the dark to ensure it reached a stable and homogeneous state. Three subsamples were then taken from the 10 L sample. The concentrations of the THDAAs in the subsamples were determined, using the sampling and laboratory procedures described above. The standard deviation of the three subsamples was regarded as the error of the primary measurement of DCAAs.
To quantify the errors in measuring DOC, we used the deep-sea water standard (CRM-batch 9) from the Consensus Reference Materials project led by Prof. Dennis Hansell, at the University of Miami (FL, USA). Up to eight samples were analyzed. Similarly, the standard deviation of the eight DOC values was regarded as the error of the primary measurement of DOC.
When deriving the AA-based molecular indicators (d/l
ratio, Gly%, percentage γ-aminobutyric acid (GABA%), AA carbon yield, and DI) from the primary measurements (i.e., individual AA concentrations and/or DOC concentration), the errors within the primary AA and/or DOC measurements were propagated to these molecular indicators. To obtain the equations for the propagation of uncertainties (as standard deviations), a statistical inference was conducted for each molecular indicator. The equations used to calculate the propagated errors are provided in the Supplementary Materials
5. Conclusion and Perspective
The THDAAs and their related molecular indicators reveal a bulk DOM diagenesis status that can correspond to thousands of years, e.g., [21
]. Against this background, short-term changes in THDAAs and AA-based molecular indicators, if they exist, would interfere with our understanding of long-term carbon cycling. Hence, the study of THDAAs and their molecular indicators’ short-term changes has a potential significance on long-time scale materials cycling.
Overall, our work indicates that care should be taken when using THDAAs and their related molecular indicators to determine the DOM’s degradation status. This is because some THDAA molecular indicators showed solid short-term changes within 1 day. The scale of variability within 1 day can be viewed as a resolution of the corresponding molecular indicators when elucidating the DOM composition changes over longer time scales. Such molecular indicators include the AA carbon yield and d/l Ala, which show 0.3–0.4% and 0.02–0.13 short-term changes, respectively.
Given these revealed short-term changes, we applied our findings to DOM seasonal composition comparison. We conclude that DOM in June 2015 is more fresh when compared to that in October 2014. This is quantified by the higher Ser% and lower GABA% and D/L Ala in DOM’s composition in June relative to October. The seasonal differences of these molecular indicators are larger than their corresponding short-term changes. The reason for such DOM composition seasonality is the fact that there was an eddy-pair event going on during our observation in June 2015. Accordingly, more upwelled deeper waters, better nutrient concentration, and higher phytoplankton standing stock are reported in June relative to October. Our finding of such seasonal DOM composition differences, confirms the seasonal DOM’s lability difference, which sheds light on the microbiol food loop and marine carbon cycling studies in the SCS.